Ramáyana Book I: Boyhood: Ancilliaries Introduction (page 3)

The Rāmāyaṇa is, at least in part, an exciting, moving, and beautiful poem. It is also the saga of the incarnate divinity’s battle against the forces of evil. Yet Indian literature is filled with poems and sagas such as this, none of which have achieved the prestige, diffusion, and influence of Vālmīki’s poem. The success of the work, like the prestige and even divinization of its hero, derives from its success in striking at the heart of one of the critical cultural problems of traditional India. Reading the Rāmāyaṇa; hearing it chanted, discussed, expounded, and analyzed; seeing it represented in plays, dances, paintings, sculptures, and films has enabled its audience, through their identification with Rāma and Sītā, to cope in their own lives with the problems that the epic poets have addressed. Like any piece of fantasy, the Rāmāyaṇa permits the reader partially to externalize and more completely master his most urgent anxieties and inner conflicts. Hundreds of generations of Indian children, urged to emulate Rāma and Sītā by elders eager to stifle rebelliousness and self-assertiveness, have submitted themselves to the sway of its powerful fantasies. In doing so they have made an adjustment that is central to the formation of the Indian personality, family, and society. As both the traditional literature and modern field observation show, many Indians, both men and women, act out in their own lives the central plot of the Rāmāyaṇa with all its negative entailments in the areas of sexuality, relation to authority figures, and emotional life in general. For them, and almost necessarily for their children, a fascination for the figures of Rāma and Sītā and their story is unavoidable. It is this, more than anything else, that is responsible for the extraordinary destiny of the Rāmāyaṇa.

3. Introduction to the Bālakāṇḍa

We may now turn our attention from the general problems of history and interpretation to a more detailed consideration of the form and contents of the Bālakāṇḍa and of their implications for the study of the Rāmāyaṇa as a whole.

Since the appearance of Jacobi’s seminal study, it has been generally accepted, at least among western scholars, that most if not all of the Bālakāṇḍa is a later addition to the central core of the poem. There is considerable evidence in support of this position. As early as 1841 Adolph Holtzmann pointed out a number of apparent contradictions in the early sections of the Bālakāṇḍa, remarking forcefully on the book’s stylistic inferiority to what he regarded as the genuine portions of the Ayodhyākāṇḍa. He also provided a cogent explanation for the inclusion in the Bālakāṇḍa of the miraculous origin of the hero and his early feats of prowess. [ Note 134 ] Holtzmann noted the evident confusion in the Bālakāṇḍa account of Daśaratha’s sacrifice to produce a son. There are at least two distinct rites involved, a putrakāmeṣṭi and an aśvamedha, the purposes of which appear to be identical. Moreover, as he suggests, there is some confusion and duplication in the roles of Vasiṣṭha, the official purohita of the king, and of Ṛśyaśṛṅga in the performance of the rites. Holtzmann was also the first to report that the contents of the first book, with its pastiche of myths, legends, genealogies, and other digressions, is in sharp contrast with the more coherent narrative of the middle books. Since Holtzmann’s time, this ‘purāṇic’ quality of the Bālakāṇḍa, as contrasted with the more ‘epic’ quality of Books Two through Six, has been noted by most writers on the Rāmāyaṇa. [ Note 135 ] Several of these authors have also remarked that the first sarga of the Bālakāṇḍa, as it now stands, summarizes the main events of the poem but makes no reference to the events of the book itself, a deficiency corrected in sarga 3.

These points are good ones, and it is impossible to avoid the conclusion that at least some significant portion of the Bālakāṇḍa is later, perhaps considerably later, than the bulk of Books Two through Six. The question of just how much later has been treated somewhat casually by many scholars. Most are content to remark that the book is later, and some, like Lesny, have stated simply that ‘ein grösserer Zeitraum,’ a considerable period of time, must lie between it and the five following books. [ Note 136 ] R. G. Tiwari argued, on the basis of a comparison of some of the mythological material from the Bālakāṇḍa with known data from political and religious history, that the book was composed sometime in the period from the latter half of the second century to the first half of the first century b.c. [ Note 13 7] His arguments are not wholly convincing, however, and since he offers no new evidence that bears on the date of the ‘genuine books,’ he brings us no closer to an informed estimation of the period that separates the Bālakāṇḍa from the older material.

The question of the date of the Bālakāṇḍa, like that of the epic itself, is greatly complicated by the fact that the text developed gradually, perhaps over a period of several centuries. As Jacobi convincingly argued, portions of the Bālakāṇḍa appear to be quite old and, undoubtedly, belong to the earliest strata of the text. [ Note 138 ] In fact, Jacobi went so far as to propose a reconstruction of the beginning of the Bālakāṇḍa from which he excised all but sixteen verses that set the scene in the prosperous Kosalan capital of Ayodhyā and introduced the principal characters. [ Note 139 ] Since he regarded these few verses as the original preface to what is now the Ayodhyākāṇḍa, Jacobi implicitly rejected most of the first book as part of the original work. In this, as we shall see, he was probably overly cautious.

In a brief article written in 1953 entitled ‘The Genesis of the Balakanda,’ Bulcke provides us with a conspectus of the book’s subject matter. Without subjecting the text to the minute verse-by-verse analysis that we find in Jacobi, he divides the book into five sections as follows:

  1. Introduction. Sargas 1-4.
  2. Daśaratha’s sacrifices. Sargas 5-17. A description of Ayodhyā; the horse sacrifice; *the Putreṣṭiyajña.
  3. Rāma’s birth and youthful exploits. Sargas 18-31. Birth of the four brothers; Viśvāmitra’s arrival; encounter with Tāṭakā, Mārīca, and Subāhu.
  4. *Pauranic stories. Sargas 32-65.
  5. Viśvāmitra’s family; sons of Sagara; churning of the sea; Ahalyā’s deliverance; the long story of Viśvāmitra becoming a Brāhmaṇa.
  6. Rāma’s marriage. Sargas 66-77.

The breaking of the bow; the marriage; *the encounter with Paraśurāma; return to Ayodhyā.

Bulcke’s asterisks indicate passages that he regards as later interpolations, so he appeared willing to allow that a substantial portion of the book, amounting to well over half the sargas of the constituted text, belongs to its oldest stratum. According to Bulcke, the early sargas are those dealing directly with the antecedents and early career of Rāma. Bulcke refers to Jacobi’s reconstruction and presumably accepts its validity, although he offers no explicit judgment of it. He appears to regard the first four sargas, [ Note 140 ] the upodghāta, as belonging to the ‘original Bālakāṇḍa.’ According to him, the oldest part of Book One is the introduction, the first sarga. This was followed — in response to questions asked by the poem’s original audience — by a description of Vālmīki and how he came to write his poem, and only then by the material about Rāma, his youth and marriage. After this material had been composed, continues Bulcke, bards added the puranic material that makes up the bulk of the central portion of the book.

There is nothing inherently unreasonable about Bulcke’s reconstruction of the history of the Bālakāṇḍa. But like Jacobi’s reconstruction of the kernel of the first book, it is based on a number of judgments concerning what the ancient bards and their audiences would or should have done. As such, it lacks any probative force. On the other hand, his arguments do force us to think of the first book analytically, rather than as an indivisible whole.

It is the failure to regard the book as having its own textual history that has led scholars such as Tiwari astray. [ Note 141] The geographical references in the portion of the Bālakāṇḍa concerned with Rāma’s journey appear to date from a time prior to the rise of Magadhan hegemony. If we take the earliest reasonable period for puranic reference to Śakas and Yavanas to be the late first century a.d., then it is evident that at least four hundred years must separate the earliest from the latest portions of the Bālakāṇḍa. These later portions are, as we shall see, heavily influenced by the Mahābhārata and older purāṇas.

As to the relative dates of the second and fourth sargas, the portion of the text dealing with the composition and early performance of the poem, it is difficult to say anything definite. The highly developed knowledge of traditional poetics and the sophisticated notion of art as a sublimation of emotion set forth in these sections suggests that their author or authors were at least familiar with the Nāṭyaśāstra; and although this text presents its own complicated problems of chronology, it can hardly be much older than the second century b.c. [ Note 142]

In discussing the nature and contents of the Bālakāṇḍa and its relation to the epic as a whole, earlier scholars have been, like Holtzmann, negative in their appraisal of the book. In their zeal to demonstrate the relative lateness of the book and its lack of an organic relation to the ‘genuine’ portions of the poem, a number of writers have tended to overstate their case. For example, as evidence for the tenuousness of the link between the first book and the core of the epic, Holtzmann claimed that in the second book, to which he ascribes real poetic merit, there are no references to the events described in the first. [ Note 143] Jacobi, Winternitz, and Bulcke all declared that this observation could be extended to include all the genuine books. [ Note 144] In conjunction with this point, these three scholars all point to what they regard as a clear contradiction between the Bālakāṇḍa’s description of Lakṣmaṇa’s marriage to Ūrmilā at 72.18 and Rāma’s famous assertion at 3.17.3 that his younger brother is unmarried.

The first and more important of these two points is simply not true. The ‘genuine’ books contain at least two references to the events of Book One. At 2.110.26-52, Sītā, responding to the questions of Anasūyā, gives a fairly detailed synopsis of the events surrounding her birth, svayaṃvara, and marriage. This synopsis follows closely the account of the concluding chapters of the Bālakāṇḍa, as we now have it, and mentions specifically that Lakṣmaṇa married Ūrmilā (verse 51). This passage, known to almost all of the subrecensions of the text, has nevertheless been overlooked or ignored by these scholars. [ Note 145]

Similarly, in verses 1 through 18 of the thirty-sixth sarga of the Araṇyakāṇḍa, the rākṣasa Mārīca, seeking to dissuade Rāvaṇa from his plan to abduct Sītā, tells him in some detail the story of his earlier encounter with Rāma that culminates at Bālakāṇḍa 29.7-21. In his preamble to the story of his being knocked unconscious by the force of Rāma’s weapons, the demon gives a concise account of Viśvāmitra’s persecution at his own hands and of the sage’s visit to the court of King Daśaratha to procure the services of the young prince, all matters treated in Book One. [ Note 146] It is of paramount importance to note that the passage not only summarizes the narrative of Book One with only minor changes, but it betrays, in at least one verse, a close textual affiliation with the Bālakāṇḍa passage, for lines 1.19.2ab and 3.36.6ab are almost identical. Clearly one of the passages is derived from the other.

These two passages, drawn from two of the central books of the epic, summarize the material of most of the Bālakāṇḍa passages dealing with the early career of Rāma. [ Note 147] It is thus apparent that the claim that the central core of the epic knows nothing of the events of the Bālakāṇḍa is without foundation. This fact does not, by itself, disprove the theory of the lateness of the Bālakāṇḍa. On the other hand, the parallels in text and content between the Bālakāṇḍa accounts of the youthful exploits of Rāma and their epitomes in the second and third books, coupled with the historical-geographical data offered by this same early stratum of Book One, make it clear that there could not have been, as has so often been argued, a very great lapse of time between the composition of Books Two through Six and the oldest portions of the Bālakāṇḍa, the portions roughly represented by sargas 5-8, 17-30, and 65-76. [ Note 148]

Another of the major charges in support of the claim for the book’s lateness and inferiority is the inconsistency involving the princess Ūrmilā who is said to marry Lakṣmaṇa in the Bālakāṇḍa. Not only do the ‘genuine’ books know nothing of Ūrmilā, so the argument goes, they contain a passage in which Lakṣmaṇa is explicitly said to be unmarried.

Once again Jacobi seems to have been the first to make this observation. He remarks that it is especially striking that Book Two contains no mention of Sītā’s sister, because one would have expected the poet, had he known of her, to have used her to intensify the touching farewells of the exiled heroes. In a footnote to this statement, Jacobi refers to the Ayodhyākāṇḍa passage mentioned above, but dismisses the passage as an interpolation on the grounds that the first book does not constitute an integral part of the Rāmāyaṇa. [Note 149] Moreover, Jacobi finds the verse in which Ūrmilā is mentioned to be spurious even by the standards of its own ungenuine context. He bases this argument on the grounds that the verse (2.110.51, Jacobi’s [Bombay edition of the Rāmāyaṇa] 2.118.53) falls between two verses that belong together. These arguments are invalidated by the fact that this passage, including the verse mentioning Ūrmilā, is found (with some variation in a few cases) in virtually every known subrecension of the text.

Nonetheless, Jacobi’s observation about the absence of Lakṣmaṇa’s wife from the elaborate scenes of leave-taking and departure that form so vital a part of the second book is interesting. It can be explained in the light of the Rāmāyaṇa’s consistent portrayal of Lakṣmaṇa as a man who suppresses his own emotional life in deference to that of his older brother. [ Note 150] The poet is at pains to develop Lakṣmaṇa as the archetype of the de-erotised and totally subservient younger brother that has become normative in traditional Indian culture. [ Note 151 ] The inclusion of any emotional leave-taking on the part of Lakṣmaṇa and Ūrmilā would therefore be wholly out of keeping with the carefully elaborated characterization of Rāma’s companion and alter ego. That Lakṣmaṇa as a Hindu prince should be married is to be expected, but omission of references to Lakṣmaṇa’s wife from the departure scenes of Book Two is appropriate to the poet’s design, and can tell us nothing about the date or genuineness of the Bālakāṇḍa.

Rāma’s well-known description of Lakṣmaṇa as akṛtadāra, or unmarried, at 3.17.3 has also been subject to misinterpretation on the part of many scholars. Jacobi remarked that Vālmīki does not make Rāma a liar by having him so describe Lakṣmaṇa to the infatuated rākṣasa woman Śūrpaṇakhā because, in the older books, his younger brother indeed has no wife. Winternitz and Bulcke concur. [ Note 152 ]

Here again, as in the whole question of Lakṣmaṇa’s supposed bachelorhood, these scholars have overlooked the cultural and narrative contexts of the passage. It is true that Rāma tells the lust-maddened demoness that she should choose Lakṣmaṇa, for he is unmarried and is seeking a wife (3.18.3-4). But how seriously are we to take his remarks here? Are we to believe, for example, that he really means his brother to marry this monster? Certainly not. Rāma is having a little joke, albeit a cruel one, with the foolish creature. When the jest goes too far, ending in an attack on Sītā and the savage mutilation of the rākṣasa woman, Rāma remarks casually to his brother, krūrair anāryaiḥ saumitre parihāso na kāryaḥ, ‘Saumitri, one really shouldn’t joke with these savage non-Aryans’ (3.18.19). Rāma is only teasing the wretched creature. The strict code of truthfulness of the Aryan warrior class, a virtue for which Rāma is famous, simply does not apply to dealings with such barbarians, any more than the code of kshatriya chivalry will apply, in the following book, to Rāma’s killing of the monkey Vālin from ambush. Certainly we can draw no serious inference about the relation of the older portions of the Bālakāṇḍa to the rest of the epic from this foolish and ultimately tragic joke. [ Note 153 ]

In conclusion, then, it would appear that previous scholarship has been overly harsh in its judgment of the genuineness of much of the Bālakāṇḍa. Examination of the first book shows that the fairly substantial portions of the text that deal directly with the early exploits of the epic hero are indeed known to the central books and are, on geographical and historical grounds, almost certainly products of the pre-Magadhan era. As such, it seems most probable that these portions of the Bālakāṇḍa formed part of the original stratum of the epic.

The remaining portions of the first book, comprising the prefatory material of sargas 1-4, the legend of the sage Ṛśyaśṛṅga and the account of his participation in Daśaratha’s sacrifices, Viśvāmitra’s retellings of sectarian and genealogical legends from the purāṇas, the lengthy and interesting account of Viśvāmitra’s struggle to achieve the rank of brahmarṣi, and the tale of the encounter of the two Rāmas, belong to later strata of the text. Let us now turn briefly to each of these sections in an effort to set them in their proper relation to the oldest material in the Bālakāṇḍa and, therefore, to the oldest core of the epic itself.

The opening sarga of the Bālakāṇḍa, as it now stands, presents us with a number of interesting puzzles. It consists for the most part of a detailed, if selective, synopsis of the epic story (verses 18-70) put in the mouth of the divine seer Nārada, who recites it in response to the questions of the seer Vālmīki concerning the existence of an ideal man (verses 1-5). Nārada begins his account with an elaborate catalog of Rāma’s physical, mental, and moral perfections (verses 6-18) and closes it with a brief description of the hero’s ideal reign (verses 71-74) and the various political, spiritual, and economic benefits that accrue to people who read the poem.

As has been remarked at various times, the first sarga does not appear to know Rāma as an avatāra of Viṣṇu. [Note 154 ] This inference is drawn on the basis of two references in the text. First, Viṣṇu is mentioned simply as one of those with whom the hero is compared (verse 17). Also at the end of the chapter, when Nārada is forecasting the future of Rāma’s rule, he states that, after having reigned for eleven thousand years, Rāma will ascend to the world of Brahmā (verse 76). Neither of these references seems consistent with the position that Rāma is regarded as an incarnation of Viṣṇu, and so it would appear that the section is older than the Uttarakāṇḍa and Bālakāṇḍa passages that posit this identification. In addition, the synopsis is strikingly free from allusions to the events of the Bālakāṇḍa proper, a fact that appears to support the theory that the substance of the Bālakāṇḍa is late. [ Note 155 ] In brief, then, it would appear safe to set up a chronological sequence according to which sarga 1, the so-called Saṃkṣipta Rāmāyaṇa, predates even the oldest stratum of the Bālakāṇḍa, which, in turn, would have to postdate the central body of the narrative that is summarized in this section. But although this conclusion appears to be unexceptionable, it may be open to serious question.

For one thing, the framing portions of the chapter, the first seven and the closing three verses, seem not to belong to the oldest strata of the text. The sketchy introduction of the sage-poet Vālmīki seems to link the chapter closely with the remainder of the upodghāta, a section that must be a subsequent addition to the core of the Bālakāṇḍa. The figure of Vālmīki is, as has often been noted, known only to the upodghāta and the Uttarakāṇḍa. He is unknown to the so-called ‘genuine’ books, and there is no allusion to him in the other parts of the Bālakāṇḍa. Moreover, the whole tone, style, and content of the opening sarga support the notion that it is an integral part of the preamble. The three closing verses of the chapter, the phalaśruti, make no explicitly Vaishnava reference, but do presuppose a powerful association with divinity on the part of the epic hero: the ancient audience of a pre-Vaishnava heroic ballad could hardly be expected to believe that hearing the piece would free them of all sin and conduct them with their descendants and retainers to heaven. [Note 156 ] The presence of a phalaśruti has been regarded as evidence of the late and interpolated character of the Bālakāṇḍa’s section telling of the descent of the Ganges, and there is no reason why this argument should not apply equally to this other set piece, the Saṃkṣipta Rāmāyaṇa.

Further evidence of Rāma’s identification with Viṣṇu in the first chapter is found in the last verse of the Bālakāṇḍa. Virtually every version of the text tells us that the newly married Rāma in the company of his wife was as radiant as the lord of the gods, Viṣṇu, in the company of Śrī. [ Note 157 ] Most scholars argue that the entire book is late, and particularly this verse with its longer meter and Vaishnava allusions. And yet, 1.76.18, like 1.1.17, merely compares Rāma with Viṣṇu; it does not identify the two. If, as Bulcke argues, we must take the verse from 1.1 as evidence of the first’s priority to the Vaishnava element in the Rāmāyaṇa, then we must make the same judgment with regard to the last verse, if not the last sarga of the text. In fact, it would appear that these allusions to Viṣṇu presuppose the identification of the god and the hero and are intended to suggest it in the context of the audience’s knowledge.

The failure of the first’s conspectus of the epic plot to include reference to the Bālakāṇḍa is evidence more suggestive than conclusive of the lateness of the book. The conclusion that the compiler of the Saṃkṣipta Rāmāyaṇa knew a Rāmāyaṇa that, as yet, lacked a Bālakāṇḍa is simple and appealing. It is not, however, necessarily correct. Bulcke argued that the first’s omission of allusions to the early career of Rāma is an ‘anomaly’ that has been ‘corrected by the insertion of a second conspectus (cf. sarga three) in which the subject-matter of the Bālakāṇḍa is incorporated.’[Note 158]

The ‘correction’ Bulcke refers to is a sarga of nearly sixty lines of which only three (1.3.4-5b) refer to the events of the Bālakāṇḍa. [ Note 159 ] Rather than compose an entire second table of contents, it would surely have been much simpler, had the redactors intended correction, for them to have inserted the necessary verse or two directly into the first chapter. The fact that such an insertion does not occur in any known manuscript of the text suggests that the first sarga is a fairly late set piece appended, not so much to the epic or even to the Bālakāṇḍa, as to the late upodghāta itself. Its function is not to provide the background to the story of Rāma, but to provide a narrative context and background for the story of the creation and dissemination of the poem. It does this by representing the legendary sage Vālmīki as the audience of Nārada’s terse puranic account of the virtues and career of Rāma who is represented as the contemporary ruler of Kosala. In order to illustrate the qualities about which the sage has inquired and of which, Nārada tells him, the Ikṣvāku prince is the unique repository, the divine seer summarizes the central, and original, portion of the epic in which they are most clearly made manifest.

Why the events of the central story of the Bālakāṇḍa are omitted is not wholly clear. It may be, as has been claimed, that the author of the Saṃkṣipta Rāmāyaṇa did not know them. On the other hand, it may well be that, as in the case of the other famous ancient condensation of the Rāmāyaṇa, the Rāmopākhyāna, the author or authors were aware of the events of the Bālakāṇḍa and even the Uttarakāṇḍa but, in keeping with the special purposes of their texts, do not mention them. [ Note 160 ]

Sarga 3 was certainly not added as an afterthought or correction to sarga 1. On the contrary, it is an integral part of the upodghāta and, as such, is probably the older of the two condensations of the story and serves a different purpose. Sarga 3 is a conspectus of the contents of Vālmīki’s poem as known to the author of the upodghāta; sarga 1 is supposed to be the concise rendition of the Rāma legend upon which the poet based his work. Both accounts are doubtless drawn from the poem, but their authors regarded them as different sorts of condensation, and it is this that explains the epic’s need for the two passages and the differences between them.

Let us now turn our attention to the remainder of the upodghāta. The first four sargas of the Bālakāṇḍa, as it now stands, give us insight into the ancient Indian view of the poem, its origin, nature, and destiny. In its treatment of the epic as the world’s first piece of true poetry, and its author therefore as the ādikavi, or first poet, the preamble affords us our earliest glimpse of the tradition’s attitudes toward art, emotion, and aesthetics.

Sargas 2-4 of the Bālakāṇḍa constitute the heart of the preamble, recounting how the sage came to compose the poem and teach it to his disciples. In addition, this section provides some description of the poem in the technical terminology of poetics and music and concludes with a brief account of a command performance of the poem before Rāma himself by his twin sons Kuśa and Lava, disciples of Vālmīki. [ Note 161 ]

Sarga 2 is one of the most interesting, widely known, and frequently mentioned passages in all the epic literature. It is in this passage that we hear of Vālmīki’s discovery of the art of poetry and of how he comes to compose the Rāmāyaṇa. Immediately after hearing Nārada’s compressed and somewhat elliptical account of Rāma’s career, the sage, wandering through a charming forest glade with a disciple, is lost in rapture while contemplating a pair of mating birds. Suddenly a tribal hunter emerges from the cover of the trees and kills the cock with an arrow. The sage, seeing the bird in its death throes and hearing the piteous cries of its mate, is filled with compassion and spontaneously curses the hunter. [ Note 162 ] To the amazement of the sage, his curse bursts forth in metrical form. Upon reflection he realizes that his grief (śoka) over the suffering of the birds has somehow been transmuted into an aesthetic, rather than a purely emotional, experience and has expressed itself as poetry (śloka) (1.2.15-17). The verse itself (1.2.14), supposed to be the very first example of the poet’s genius, is rather a disappointment, for it appears to be almost entirely lacking in the qualities of sound, sense, and suggestion that form the basis for the major traditional schools’ critical assessment of poetry. Nonetheless, because of the place assigned the verse at the very wellsprings of poetic inspiration, it has, along with the entire episode, been accorded great significance by both literary critics and sectarian commentators. [ Note 163 ]

Vālmīki returns to his ashram, where he is visited by the great god Brahmā. The god tells the sage that it is he who has inspired him with poetic genius and commissions him to compose a major poetic account of the life of Rāma based upon Nārada’s tale, the gaps in which are to be made good through a special gift of insight (1.2.21-34). It is here that we find Brahmā’s famous prophecy concerning the longevity of the poem (1.2.35).

The sarga is interesting in many respects, but chiefly because of the light it sheds upon the close connection between emotional and aesthetic experience in early Indian thought. It is interesting also to note that in this legend of the creation of the first poetry, the underlying emotional states that give rise to the aesthetic experience are grief and pity. If we are to view this as reflecting a theoretical position — and this is by no means certain — then it is a provocative one that is not generally reflected in the massive technical writings of the followers of the ‘rasa school’ of Indian poetics. [ Note 164 ]

Nonetheless, the upodghāta’s evident familiarity with the theory and traditional enumeration of the rasas, and the implicit interpretation of the entire poem as having a single central organizing mood, strongly suggest that the preamble is later than the oldest portions of the epic, including the central portions of the Bālakāṇḍa. [ Note 165 ] Finally, after the conspectus of the poem’s contents that makes up the third sarga, the preamble addresses the issue of the bardic tradition of Rāmāyaṇa transmission. The poet is represented as teaching the poem and its mode of recitation to Kuśa and Lava, who are said to be twin sons of Rāma. It seems clear that these figures are a late invention designed to personalize the anonymous tradition of the bards, the kuśīlavas, of ancient India. [ Note 166 ] Through the creation of these two imaginary figures and their identification as both disciples of the legendary poet and sons of the epic hero, the authors have linked the origin of the poem with the career of its central character. In having the twins recite the epic before their father in Ayodhyā, they have set the stage for the beginning of the epic proper.

The saga of the Rāmāyaṇa begins, as Jacobi claimed, with the fifth sarga of the Bālakāṇḍa in which the audience is introduced to the royal house of Ikṣvāku, its scion King Daśaratha, and its hereditary seat Ayodhyā, capital of the fair realm of Kosala. Sargas 5-7 describe in some detail the glories of the city and the surpassing virtues of its inhabitants. The descriptions are elaborate and extremely hyperbolic, but this is in keeping with the style of the Sanskrit epics and lends weight to the Rāmāyaṇas claim of universal sovereignty for the princes of the Rāghava line.

Daśaratha is a fortunate, prosperous, powerful, and happy man. But as is so often the case with the great monarchs of Indian legend, his happiness is flawed by his lack of a son and heir. In order to rectify this deficiency, the king resolves to perform the great Horse Sacrifice (1.8.1-2). The staging of the elaborate Horse Sacrifice, normally employed in the epics to sanctify a king’s acquisition of sovereignty over his neighbors’ territories, is unusual for the purpose of procuring a son. Bhatt, in his notes to the critical edition, attempts to gloss over this peculiarity by referring to the vedic tradition, according to which ‘the performance of the Aśvamedha sacrifice secures everything for the performer. It is, therefore, performed even for a particular purpose (e.g. for getting a son).’ [ Note 167 ] This explanation is far from satisfactory. For one thing, there does not appear to be any other example in the extensive list of Aśvamedhas performed in the two Sanskrit epics of a king’s making such use of the rite. Moreover, as has been noted by other scholars, [ Note 168 ] the king performs at least one additional rite for the acquisition of a son, the Putrīyā Iṣṭi initiated at sarga 14. One of these rites would appear to be redundant, and in the light of the seeming inappropriateness of the Horse Sacrifice in this context and of the peculiar recruitment of the sage Ṛśyaśṛṅga as its chief officiant, I would be inclined to disagree with Bulcke’s suggestion that the superfluous rite is the one specifically designed to produce a son. [ Note 169 ] It would, on the whole, appear more probable that Daśaratha’s great Horse Sacrifice, which is described in far greater detail than any other ritual performance in the Rāmāyaṇa, is a later addition introduced with the purpose of firmly establishing in the mind of the audience the splendor and might of the Kosalan monarchy. By the period of the final shaping of the Sanskrit epics, the Aśvamedha had evidently become the great symbol and demonstration of Hindu hegemony. Although the Bālakāṇḍa and Uttarakāṇḍa together attribute the performance of the rite to no fewer than five Ikṣvāku kings and refer to at least eight performances of the ceremony in all, the central five books rarely, if ever, mention it.[Note 170 ] From this discrepancy, it would appear that the portions of Books One and Seven that mention this ritual, and especially the extremely detailed and elaborate account of Daśaratha’s somewhat otiose performance, are late additions to the text, introduced under the influence of the Mahābhārata/purāṇa tradition that sets such great store by the Aśvamedha.

The derivative and interpolated nature of the account of Daśaratha’s two sacrifices is further demonstrated by its use of an important figure drawn from a wholly unrelated mythic context — the sage Ṛśyaśṛṅga. This interesting figure, the story of whose seduction is one of the most widely distributed of Indian legends, had no original connection whatever with the tale of Rāma or the House of Ikṣvāku. [ Note 171 ] He is brought into it abruptly and in the manner of puranic narratives. Hearing that the king wishes to sacrifice in order to procure a son, his old retainer and counsellor Sumantra is reminded of an ancient story that he has heard from some sacrificing priests who tell him that the sage Sanatkumāra related it long ago (1.8.5-6). He then recounts to the king an abbreviated and relatively colorless version of the tale of the seduction of the innocent ascetic boy and his marriage to Śāntā, the daughter of King Romapāda of Aṅga. [ Note 172]

This perfunctory account of the unicorn-sage is introduced because Sanatkumāra had concluded it with the prophecy that Ṛśyaśṛṅga would somehow produce sons for Daśaratha. [ Note 173 ] Having heard this, Daśaratha asks Sumantra for a more detailed account of the career of Ṛśyaśṛṅga. Sumantra’s response is a slightly more elaborate but still highly compressed and pedestrian rendering of the charming story. [ Note 174 ] Sumantra, concluding the prophecy of Sanatkumāra, now reveals that the king will bring Ṛśyaśṛṅga to officiate in his sacrifice and that the result of this will be the birth of four mighty and renowned sons (1.10.1-11). Daśaratha journeys to the kingdom of Aṅga, brings back the sage and his wife Śāntā, and installs them in his womens’ apartments (1.10.13-29). This sets the stage for the performance of the rites of the Aśvamedha and the Putrīyā Iṣṭi. The description of the preparations for the execution of the first of these rituals occupies virtually the whole of sargas 11-13. The latter is undertaken at the beginning of sarga 14, but is interrupted immediately by the story of the petition, on the part of the gods, to Brahmā and then to Viṣṇu in their desire to find relief from the oppression of Rāvaṇa (1.14.4-15.6). Viṣṇu agrees to take birth as the sons of Daśaratha, and this resolution takes effect through the appearance of a celestial being who arises from the king’s sacrificial fire, bearing a vessel filled with a pudding infused with the god’s essence (1.15.8-22).

This dramatic event marks the culmination of the sacrificial interlude of the Bālakāṇḍa. After the king’s distribution of the divine food to his three principal wives, the Vaishnava interlude continues with a chapter in which the gods generate sons in the form of various kinds of monkeys in order to assist Viṣṇu in his mission (1.16). After this, the text returns at last to the tale of the birth of Rāma and his brothers, which occurs in due course after the completion of the king’s Horse Sacrifice (1.17.1ff.).

Clearly — as has been observed many times — the nine sargas from 8 through 16 are a somewhat diverse collection of materials added after the completion of the original story of Daśaratha and the birth of his sons. These materials may be subdivided into two originally unrelated elements, the sacrificial and the Vaishnava. The latter of these has long been known for a late addition to the original saga of Rāma, which appears to know nothing of its hero’s divinity. The former is probably older, but is itself made up of a number of heterogeneous elements. There appears to be some unresolved confusion and a certain degree of overdetermination in the relationship between the Horse Sacrifice and the Putreṣṭi. Moreover, the abrupt and clumsy introduction of the legend and the person of the sage Ṛśyaśṛṅga into the sacrificial material is evidently the result of a secondary manipulation of the text, which appears to have served two purposes. The legend of the seduction of the boy-sage, like the other puranic stories of the Bālakāṇḍa and Uttarakāṇḍa, lends to the Rāmāyaṇa something of the encyclopedic quality and prestige of the Mahābhārata. In this way the poem and its hero are glorified, as is the learning of the bards. More specifically, however, the Ṛśyaśṛṅga legend was, no doubt, thought apposite here because of the powerful association of the sage with sexuality and especially with great fecundity. The sage is perhaps viewed as serving the purpose of Vyāsa or the other Mahābhārata practitioners of the ancient custom of niyojana, or levirate, only through an act of sacrifice in place of direct sexual liaison with the king’s wives. [ Note 175]

The remaining mythological episodes of the Bālakāṇḍa may be generally classified according to two principal purposes that they serve. The first is to invest with a mythic and sacred significance the landscape through which Viśvāmitra and the two Ikṣvāku princes pass on their journey from Ayodhyā, via the ashram of the Perfected Being, to Mithilā. In this way the legend of Rāma is associated with and ultimately equated with the great myths of the brahmanic-puranic tradition. The second major type of legend has a similar purpose. It is to glorify the principal figures of the first book, Rāma and Viśvāmitra, through accounts of the greatness of their early deeds and those of their ancestors. Some important legends or sets of legends incorporate both elements.

Examples of the first type are Viśvāmitra’s version of the myth of the destruction of Kāma (1.22.10-14), his tale of the origin of the Ganges and birth of Skanda (1.34-36), and his narration, in connection with the cities of Viśālā and Mithilā, of the stories of the churning of the ocean, the origin of the Maruts, and the curse of Indra and Ahalyā (1.44-48).

Episodes of the second type, perhaps because the authors and audiences of the text were more interested in the heroic antecedents of their protagonists, tend to be more elaborate and, in fact, make up the most significant addenda. Two of these episodes combine genealogical with geographical interest. In response to Rāma’s questions concerning the region of Magadha through which they are passing, Viśvāmitra relates the genealogy of the legendary king, Kuśa. He tells of the cities founded by the king’s four sons and of the strange fate of the hundred daughters of his son, King Kuśanābha (1.31-32). Viśvāmitra concludes his tale with an account of his own descent from the line of King Kuśa and the extraordinary statement, in which the geographical and the genealogical elements are fully fused, that his sister Satyavatī has become the river Kauśikī (1.33.1-13).

As a sequel to his description of the Ganges and the birth of Skanda, the Kauśika sage tells Rāma and Lakṣmaṇa an elaborate and detailed saga of their own Ikṣvāku forbears. This tale, the story of the sons of Sagara, the digging of the ocean, and the descent of the Ganges, picks up from the geographically motivated tale that precedes it and presents, in seven chapters (1.37-43), one of the major legendary interpolations of the first book. The episode deals with the superhuman deeds of Rāma’s ancestors and with their close associations with divinity. It is just the sort of puranic set piece that one would expect in a work of this sort. [ Note 176]

The last significant legendary interpolation of this type in the Bālakāṇḍa is also the longest. This is the well-known saga, really an epic within an epic, of the career of Viśvāmitra and his struggle to achieve the status of a brahman seer within a single lifetime. This saga, which contains at least three separate legendary components organized around the common theme of the sage’s quest, occupies almost all of sargas 50-64.

While it is clear that this lengthy episode with its complicated sub-episodes must be a later interpolation in the Rāma story, it is equally clear that it is important in the minds of the authors of the first book. Viśvāmitra is one of the central characters of the book, and it was evidently felt that his importance demanded a major rehearsal of his legendary feats of asceticism and supernatural power. The significance of this exaltation of the sage and the consequent elaboration of the Mahābhārata tale of his conflict with Vasiṣṭha, here interspersed with the legends of Triśaṅku and Śunaḥśepa, is not wholly clear. Perhaps the Bālakāṇḍa reached its present form at the hands of bards or redactors associated with the Kauśika clan, who exerted a formative influence something like that exercised by the Bhārgava redactors of the Mahābhārata. [ Note 177]

The remainder of the Bālakāṇḍa consists of material directly connected with the early adventures of the epic’s hero and heroine. In sarga 65 we are given the tale of the origin of Sītā and one version of the history of the great bow that Rāma is to break. This is a necessary background to the central scenes of the testing of the bow and the hero’s marriage. If this material was added to the epic after the completion of the central tale, it could not have been much later, for, like the central portions of the Bālakāṇḍa, it fills lacunae in the audience’s grasp of the poem’s principal characters. Later on, the same motivation that led to the inclusion of the legend of Sītā will impel the authors of the Uttarakāṇḍa to provide their audience with the legends of the origins and early careers of such figures as Rāvaṇa and Hanumān.

Finally, a certain amount of originally unrelated and almost certainly later material has been juxtaposed with the central part of the Bālakāṇḍa. The most striking and important examples are the two dramatic encounters that frame Rāma’s adventures in the Bālakāṇḍa — the encounters with Tāṭakā and Rāma Jāmadagnya. The two episodes, whose psychological focus is the breaking of Śiva’s great bow, should be considered together; their placement at the very beginning and very end of the hero’s boyhood odyssey and their complementary emotive thrust whereby he overcomes first a terrifying mother figure and then the menacing aspect of the father, provide Rāma with the closest thing to a Bildungsroman the ancient Indian literature can show.

The literary or legendary sources of the Tāṭakā episode remain obscure. [ Note 178] Perhaps it is an adaptation of some older legend of a sphinxlike yakṣī who blocks a road at some dreadful enchanted forest and devours those who seek to pass. In any case, the destruction of this personification of the sexually charged devouring mother and the compensatory boon of the huge array of magical weapons offered by the surrogate father Viśvāmitra constitutes the first great rite of passage for Rāma from a pampered boy-prince to an Indian warrior hero. The lateness of the passage with respect to the central books of the epic is further suggested by the fact that in Books Two through Six the hero does not appear to use the weapons that Viśvāmitra so liberally bestows upon him at 1.26-27. [ Note 179 ]

The second of these episodes, the encounter with the dreadful brahman-warrior Rāma, is interesting not only from psychological and literary standpoints, but also with regard to the textual pre-history of the epic. Like the episode of Tāṭakā, it is introduced directly into the epic action and, indeed, forms the culminating event of Book One. Nonetheless, it is clearly a later interpolation, for the figure of Rāma Jāmadagnya is proper to the Mahābhārata in its expanded form and was a product of the Bhārgava redactors of that work. [ Note 180] Since the older portions of the Rāmāyaṇa are older than the Mahābhārata and the development of the figure of Rāma Jāmadagnya belongs to a relatively late stratum of the Bhārata corpus, it would follow that the episode of the encounter of the two Rāmas must be a late development in the Bālakāṇḍa. [ Note 181] Nonetheless, the episode has been used with great psychological and literary skill by the authors of the Bālakāṇḍa to complete their exposition of the character of Rāma as an irresistible warrior-hero who will earn his reputation not only for his skill at arms, but for his aptitude for renunciation and deference to authority.

The Bālakāṇḍa, like the character of Rāma himself, is, as we find upon close examination, considerably more complex than most scholars have hitherto led us to believe.

4. The Rāmāyaṇa Text and the Critical Edition

Sheldon I. Pollock

Despite its great antiquity, we probably know as much about the origin and development of Vālmīki’s epic as of any other ancient or early medieval work of Sanskrit literature. A substantial body of testimony and numerous parallel versions in addition to the long and self-conscious Rāmāyaṇa tradition aid us considerably in our effort to reconstruct its past. The publication at the Oriental Institute, Baroda, between 1960 and 1975 of the first critical edition of the poem — the basis of our new translation — has given us ready access to all of the manuscript evidence for the work that we are ever likely to have and enables us to draw some new conclusions about the nature of its transmission. It will be necessary to consider at some length the character of this edition, its rationale, value, and limitations. But before we do this, let us recall briefly what we know about the history of the poem beyond its strictly textual tradition.

In the late upodghāta, Vālmīki is represented as having created his masterpiece out of the terse narrative provided to him by the sage Nārada. He recasts this in a new metrical form and inspired by the god Brahmā, expands the story. ‘The whole Rāmāyaṇa poem’ is taught by Vālmīki to two disciples chosen because they are ‘retentive and thoroughly versed in the veda.’ They learn the poem by heart and perform it in public, singing it back ‘just as they were taught’ to the accompaniment of the vīṇā, or Indian lute. [ Note 182]

The tradition thus represents Vālmīki’s Rāmāyaṇa as an individual artistic elaboration of a pre-existing narrative, composed and transmitted orally in a more or less memorized form. There is little in this account that is not in keeping with the unitary character of most of the poem and with what we can infer about its sources. That the Rāmāyaṇa is an oral composition has now been statistically demonstrated, and indeed, as we shall see, our manuscript evidence implies a long antecedent period of oral transmission. [ Note 183]

The history of the Rāmāyaṇa in its written form effectively commences in the eleventh century. The probable date of our earliest exemplar, a palm-leaf manuscript from Nepal representing the northwest tradition, is a.d. 1020. No earlier manuscript fragments have been discovered. Ancient epigraphical documentation is wholly lacking except for the commemoration, in a Sanskrit temple inscription from Cambodia dating about a.d. 600, of the presentation of a Rāmāyaṇa codex. [ Note 184]

Between 1020 and the introduction of printing in India in the early nineteenth century, the Rāmāyaṇa was copied by hand repeatedly in all parts of the country, and at present more than two thousand manuscripts of the poem, in whole or in part, are known to exist. The sheer size of the text, the enormous number of manuscripts, and their often discrepant testimony, make for a text-historical problem equalled in complexity, perhaps, only by that of the New Testament.

Like the Mahābhārata, the second great epic of ancient India, the Rāmāyaṇa has been handed down in two principal recensions, one from northern and one from southern India. [ Note 185] These recensions consist of often heterogeneous versions written in the various regional scripts. Manuscripts of the northern recension come from Gujarat, Rajasthan, Kashmir, Nepal, Bihar, and Bengal; those of the southern recension from Kerala, Andhra Pradesh, and Tamil Nadu, with Devanāgarī manuscripts variously affiliated to the northern and/or southern tradition. Unlike the Mahābhārata (and this is of primary significance for the text criticism of our poem), the recensions of the Rāmāyaṇa display disagreements of a sort that cannot be accounted for by the inevitable accidents of written transmission. [ Note 186]

Although the phenomenon of recensional divergence has long engaged the attention of scholars, adequate scrutiny has become possible only with the appearance of the critical edition. Our understanding of the complicated character of the variations is still imperfect. One explanation that has come to have wide scholarly currency since Jacobi first offered it in 1893 is that the northern recension represents a purification, a polishing of an archaic southern recension. [ Note 187] According to this theory, the northern schoolmasters or learned reciters were the custodians of a pristine Sanskrit tradition. They held the Rāmāyaṇa to be not so much a sacred document as the archetypal poem and expected it to observe all the canons of linguistic and rhetorical usage that had come to be regarded as standard in post-epic times. When the Rāmāyaṇa departed from these norms, the northerners were prepared to alter it. [ Note 188]

The basic suggestion — that the northern recension presents some sort of revision — we feel to be correct, but not necessarily for the reasons usually given. For the argument supporting that theory is based on the preservation in the southern recension of grammatical irregularities and no longer seems tenable. [ Note 189]

If we closely examine the northern recension, we observe two phenomena that are far more common than any attempt to bring the poem into conformity with the rules of classical grammar or rhetoric and tell us a great deal about the history of the poem’s transmission and the value of the northern recension in the reconstruction of the original. First, the wording of the northern recension frequently differs from that of the southern without appreciably altering the text’s grammatical regularity or poetic acceptability. The northern recension, moreover, often tends toward a popularization or glossing of the southern text.

The critical apparatus on virtually every page of the Rāmāyaṇa indicates how the northern recension rephrases the southern recension almost gratuitously, without eliminating solecisms or enriching the poetic quality of the text. The density of this divergence is highly variable, anywhere from 0 to 66 percent for different sections of the poem. It seems that the only way we can account for these variations is to posit a long period of oral transmission after, as well as before, the split in the tradition had occurred. [ Note 190 ]

Although this first feature has been appreciated to some extent by other scholars, the apparent tendency of the northern recension to gloss southern recension readings appears to have gone unnoticed. The text came to be viewed as obscure in places, as the learned medieval commentators amply attest. The northern singers seem to have been particularly sensitive to this, and in the course of centuries, they evolved a somewhat simpler idiom, vulgarizing Vālmīki’s poem for the sake of their audiences. [ Note 191 ] Instances of this are very common. [ Note 192 ] This tendency does not, however, generally involve an effort to regularize grammatical usage. It is rather a simplification of the text, a transposition into a more popular idiom, a close paraphrase of passages that, although grammatically correct, are nonetheless difficult or obscure for lexical, syntactical, or other reasons. In fact, in many ways, the northern recension acts as our oldest commentary on the Rāmāyaṇa. [ Note 193 ]

This tendency of the northern recension to modernize and gloss a text perceived as archaic offers decisive support to the position adopted by the editors of the critical edition that the southern recension preserves an older state of the text, and consequently must serve as the basis for any reconstruction. We have been able to find no passages that would indicate such a tendency on the part of the southern recension. Indeed, we would appear to have in the type of variation found in the northern recension the first sign of the popularizing impulse that leads ultimately to the great vernacular translations and adaptations of medieval times.

How these recensions are related to one another, or, indeed, whether they are related at all, forms the central problem of Rāmāyaṇa textual criticism. With the publication of the northeastern and southern recensions in the mid-nineteenth century and the north-western version in the early twentieth, a fairly complete picture of the text’s history began to emerge and with it a certain pessimism about the possibility of recovering the original poem. Thus Hopkins argued that ‘all our classical notions of a fixed original from which manuscripts vary by the slightest alteration vanish into thin air before such freedom of transmission as instanced here. … The hope of getting at any ādi-[original] Rāmāyaṇa by working back from the textual variations handed down in the several recensions is quite vain. There can be no plausible original reconstructed.’[Note 194]

Other scholars, although they acknowledge textual fluidity, have argued in just the opposite way on the grounds of the remarkable congruence that often does appear between recensions. Jacobi, for instance, maintained that the various local versions must ‘all have descended from an old recension, and one can adduce no reason why this Ur-recension should not have been one that was set down in writing.’ [ Note 195 ]

On the one hand, then, we have the denial that the Rāmāyaṇa ever existed in any stable form, and on the other, the assertion that not only was its form stabilized at an early date but it was fixed in a written archetype. Each position has some truth in it, but obviously both cannot be wholly correct.

Disagreement among the recensions, as we have noted, is sometimes stark — in fact, irreducible. Nonetheless, the different versions of the Rāmāyaṇa are unquestionably versions of the same poem. This is the basic postulate that underlies the critical edition. [ Note 196 ] Although substitutions do occur, and although their density sometimes reaches two lines in three, it frequently drops to as low as 2 or 3 percent, or disappears altogether. In some places we find dozens of consecutive verses or even whole chapters for which there are no significant parallel passages. Thus, though variable to a degree, agreements between the recensions in wording, sequence of verses, chapters, and incidents are often remarkably close, and the only way to account for this continuous concord is to posit a common descent. This in turn implies that the source must be to some extent recoverable.

But if convergence is too marked to deny a genetic relationship between the recensions and thus the possibility of reconstruction, divergence is likewise too pronounced to allow the assumption of a written archetype. [ Note 197 ] Moreover, were such an archetype admitted, we should expect the original to be potentially always recoverable, which is patently not the case. [ Note 198 ] We must, therefore, postulate a mode of transmission that can account for both features of the Rāmāyaṇa textual tradition. The recensions must have been handed down through oral transmission — perhaps influenced in a distinctive way by the vedic mnemonic tradition — from the oral composition attributed to Vālmīki, that is, the monumental poem that was a remaniement of an ancient Rāma story. The resulting versions were then independently fixed in writing at different times and places. [ Note 199 ] This hypothesis alone would allow for both the divergences and agreements, and although it is not consistently upheld by the editors of the critical edition, [ Note 200 ] it is what study of the critical apparatus clearly and emphatically confirms.

Under these special conditions of textual transmission, stemmatic analysis is clearly inappropriate. For the many verses in irreducible disagreement of a neutral sort (that is, in the absence of linguistic, stylistic, contextual, or historical features that would allow discrimination), an a priori choice on the basis of the generally best version is not only admissible but necessary. [ Note 201 ] But the absence of stemmatic compulsion also requires that where the choice between versions is not neutral, we must review the recensions with care: for if they all ultimately derive more or less independently from the same oral source, then the correct reading in any given case may be preserved by any one of them.

In countless instances it appears that the ordering of the verses and the readings of the southern recension are far more intelligible and authentic than those of the northern recension, while its transmission, in general, seems considerably more uniform. [ Note 202 ] And thus, despite some literary and historical arguments that have been made to the contrary, it recommends itself as the basis of a critical edition.[Note 203] But the southern recension, too, is marred by corruptions, false emendations, accretions, and the like, and does not invariably give us the right text. The northern recension can help correct it and thereby reveal the oral original.[ Note 204 ]

We can show the truth of this at every level of the text in the case of individual words and phrases as well as large interpolations. One small paradigmatic example may serve as demonstration. In 2.63.4 we read in the vulgate:

vādayanti tathā śāntiṃ (lāsayanty api cāpare

nāṭakāny apare prāhur …)

The reading śāntiṃ is that of the entire southern recension. The commentators try desperately to explain its sense: ‘Some caused śānti, peace, to sound (others danced or staged dramatic pieces …)’ but obviously, without success: for here the word has no sense. It is a stop-gap emendation, an early one, faithfully reproduced throughout the whole southern tradition. Northern manuscripts, for the most part, offer:

(avādayañ) jaguś cānye …

replacing the meaningless śāntiṃ with ‘(some made music) and sang.’ The northern recension is, in fact, glossing an obsolescent verbal form preserved for us in three other northern manuscripts, one from the northeast, one from the northwest, one from the west:

(vādayanti tathā) gānti …

This form (classical gāyanti), as we now know, was current in the epic dialect, for the critical edition of the Mahābhārata repeatedly attests to it. [ Note 205 ]

Even such a minor example should suffice to answer Hopkins’ complaint that ‘no comparison of the varied readings of the two versions will enable one to discover the ādi-form.’ If we multiply this type of evidence many times over, in the case of word, verse, or chapter, we can get some sense of the text-critical value of the northern recension and the reality of the critical edition’s reconstruction.

Perhaps the most dramatic results of the critical edition can be seen in the treatment of interpolated passages. We must bear in mind that committing the versions to writing in no way arrested their growth. New material of a mythological, sectarian, or simply expansive nature continued to be added nearly equally in the different recensions and versions throughout the period of written transmission, just as we suppose happened in the period of purely oral transmission. The principle developed to deal with these interpolations is similar to the one used for the critical edition of the Mahābhārata: A passage missing in any of the recensions or versions as a whole, or in uncontaminated manuscripts of these (in a descending order of probability, with due attention paid to contextual requirements), is suspect and eliminated from the critical text. In practice as well as in theory the principle has proved to be sound.

At first glance, this may seem like an artificial formula that might have disastrous consequences in application. It is, of course, a natural corollary of the hypothesis of common origin, which is probable on other grounds. But one might expect it to be too crude to deal with, for example, the tendentiousness and wilfulness of scribes so often demonstrated in the western literary tradition. We do well here to recall the remarkable, perhaps unparalleled, fidelity of the Indian copyist to his exemplar. As Edgerton describes it, ‘it appears that no scribe, no redactor, ever knowingly sacrificed a single line which he found in his original … there is certainly not a shred of evidence for a single deliberate omission, and I do not believe it ever took place.’ [ Note 206 ] In fact, when the interpolations of the Rāmāyaṇa are excised, a perfectly smooth text usually does result. The editors may sometimes have erred either way in their application of it, but the principle itself repeatedly demonstrates its validity. And the result is remarkable: a full 25 percent of the vulgate (the southern recension) has been eliminated as not deriving from Vālmīki’s monumental composition.

The critical edition, then, we believe, puts us in possession of the most uniform, intelligible, and archaic recension of the Vālmīki Rāmāyaṇa, corrected and purified on the basis of the other recensions and versions that are descended from the common oral original. Although the reader of this translation thus has access to a more authentic text than has hitherto been available in translation, we are aware of the fact that those familiar with the Rāmāyaṇa may miss favorite or well-known passages that have become established in later tradition, particularly those that belong properly to the vulgate. For this reason, we have translated or summarized in the notes such passages as we thought significant. On the other hand, where the text-critical principle has been applied with less consistency and rigor (as in the Bālakāṇḍa), such material as ought properly to have been excised has nonetheless been included in the translation, since our primary purpose was translating, not editing. An examination of the notes to the translation will show which passages we regard as possible interpolations and why.

The critical edition has in general followed the methods and fulfilled quite admirably the expectations that Johnston wrote of fifty years ago:

The proper procedure would be to collect and collate the oldest and most representative MSS. from the various parts of India and Nepal and prepare from them a composite text. After excising obvious interpolations, there would remain a number of passages in substantial agreement and probably original in the main, and secondly, many much expanded passages in which the MSS. would differ greatly and which would require skilled handling. According to all appearance we have lost little of Vālmīki’s work, and it is a question in the main of determining which passages or verses are original. In the end it should be possible to obtain a coherent text which, though constructed by subjective methods, would not differ so very much from the poem as it left Vālmīki’s hands; and such a version would have the supreme advantage that, stripped of most of the accretions of later times, it would reveal to us in precise detail the genius of the greatest figure in Sanskrit literature. [ Note 207 ]

But, of course, although a certain degree of scientific precision can be attained in application of the critical method, manuscript testimony can be inconclusive, and subjective decision is sometimes the only recourse available to the editors. But editors, as one textual critic puts it, ‘are not always people who can be trusted, and critical apparatuses are provided so that readers are not dependent on them.’ [ Note 208 ] For these reasons we have carefully scrutinized the sources of the constituted text and have never followed it where we felt it was in error. When textual emendation was unavoidable we have emended. But again, given our main task, this has been kept to a minimum, and for the most part, we rest content with registering and explaining our disagreements in the notes.

5. Translating the Rāmāyaṇa


The aim of every serious translator of poetry is to stay faithful to the original and yet create something like its quality in the receiving language. This ambition is always tempered by a disheartening awareness of the difficulties that inhere in the task. Difficulties exist even when the transaction is between languages that have a good deal in common; but they increase as the linguistic and cultural distance that separates the original from the target language grows greater. It should come as no surprise, then, that the task of rendering an ancient Sanskrit epic into contemporary English is more than usually daunting; for the distance here is virtually antipodal.

Consider what we now have come to regard as authentic poetry: personal utterance in the poet’s own voice, dense with the inner reality of private experience, and largely free from the formal patterning of calculated art. What, then, is a contemporary audience to make of the Rāmāyaṇa, a poem for public recitation, emphatically impersonal in tone, and whose author was committed to his culture’s unwavering faith in an immutable social, moral, political, and aesthetic order?

We are, for the most part, silent and solitary readers, appreciating poetry in the quiet of our homes and only insofar as it resonates through our own inner lives. The Rāmāyaṇa was meant to be heard at gatherings, to be chanted like liturgy — a poem that, early in its history, promises its audience not only aesthetic rapture but salvation. And though we call this poem an epic, there is little, as we shall see, in a reading of the “Iliad” or “Paradise Lost” that can adequately prepare the western reader for the movement, tone, and style of Vālmīki’s masterpiece. There is little, for example, in the most widely read western narrative poetry to compare to the Rāmāyaṇa’s long and repetitive passages that seem to lack all poetic or dramatic function. That such passages do have some function is attested by their preservation, but the translator must despair of making them palatable to his audience.

It is bad enough to have to deal with difficulties that arise out of extreme cultural disparities. But such difficulties are not the worst kind. There are other, less tractable problems, problems of style and of feeling, of trying somehow to capture and transmit the flavor of the stately, scarcely changing, and sonorous drone of the Indian oral poet. Many of our solutions to these problems were ad hoc — what seemed to answer most effectively to a particular need. But certain general principles were fundamental to the endeavor. Above all, we decided that it was essential to follow, as closely as possible, the relative simplicity of Vālmīki’s syntax and diction, and to adhere, as strictly as we could, to the poem’s verse-by-verse narrative movement.

The notion of attempting a verse translation was early abandoned because there is no equivalent of chanted prosody in modern English poetry, and anything less would have yielded the sort of doggerel employed by some earlier translators. An alternative could have been a more elaborate meter and syntax, but this would have been a fatal distortion of the uncomplicated surface of the original. Where the Rāmāyaṇa does rise to something that the modern reader might recognize as poetry, we have tried to follow with an enriched prose. But for the most part we have aimed at a diction that would unobtrusively carry the narrative without jarring either with archaism or colloquialism.

If the music has been lost and with it something of the magic the poem held for its original audience and their heirs, we think we have kept the grand outline of the conception and the large verbal gestures that create a sense of the idealized world of the poet and his characters. To this degree we have kept faith with the original. It must remain for our readers to decide whether or not we have given them a living Rāmāyaṇa in their own tongue.

6. The Translation and Annotation

On the basis of the foregoing discussions of the antiquity of Rāmāyaṇa and of its unrivaled popularity and influence in the traditional culture of India and much of contiguous Asia, one need hardly offer an explanation for undertaking to produce a new translation of this important work. Translations into several European languages of the entire poem exist, including at least three into English. A few of these, like Gorresio’s Italian rendering of the Bengal Recension and Roussel’s French translation, are elegant and readable, while others, such as Shastri’s English rendering and the much older translation of Dutt, are helpful, perhaps, to the beginning Sanskrit student, but are stylistically unpalatable to the English reader. Whatever their literary merit, and this varies radically, all of the translations so far made have been based upon a single version or recension of the text, either the vulgate or the Bengal version, and their authors have not been in a position to judge accurately the vital text-critical problems that the epic presents. None of them, moreover, has attempted to put before its audience the results of a close and critical reading of the extensive and important commentarial literature that has grown up around the poem. As a result, readers of previous translations have had to accept without criticism the translators’ judgments on often very difficult questions of interpretation without even knowing where the problems are to be found, much less the issues raised by attempts to solve them.

It is principally these two shortcomings, which diminish the value of all previous translations of the Vālmīki Rāmāyaṇa, that we seek to remedy in the present work. We wish to set before the general, as well as the scholarly, reading public an accurate and readable translation of the epic that is as faithful as possible to the style, tone, and general feeling of the original. As an aid to the understanding of the poem and our translation, we have provided a fairly dense annotation in an effort to lay before the reader the issues involved in each one of the numerous textual, interpretational, and stylistic problems that the work presents. In this way, we hope to enable our readers to judge for themselves the success of our efforts at the solution of these problems.

A point that must inevitably strike the reader upon first examination of this translation is that it is in prose. Surely, it will be argued, one cannot hope to capture the feeling and flavor of the original if the very mode of composition is abandoned. The Rāmāyaṇa is unquestionably regarded by the majority of exponents of the Indian tradition of literary scholarship as kāvya, or poetry. [ Note 209 ] Nonetheless, for this tradition the quality of poetry in a literary work does not inhere in its metrical form. Prose can be judged as kāvya, whereas much metrical composition, such as the great mass of śāstras, kośas, purāṇas, and so on, represents, in the main, the antithesis of poetry.

In any case, despite its traditional reputation as the ādikāvya, and with the exception of a number of well-known and undeniably beautiful passages, the Rāmāyaṇa is often far from poetic. The major thrust of the text is narrative. The bulk of the poem, moreover, consists of short syntactic units, generally coincident with the basic metrical unit, the śloka, the effect of which is most accurately represented to the modern ear by simple, rather paratactic, English prose. [ Note 210 ]

Moreover, it was ultimately agreed by all the members of the translation consortium that, as this was to be the first translation of the critical edition of the Rāmāyaṇa, its principal concerns should be accuracy and readability. It was agreed that these two goals would be best served by a prose diction as uncomplicated and straightforward as that of the original. Moreover, in view of the decision that a small group of Sanskritists would each prepare the translation of a different book of the poem, it was soon realized that stylistic consistency from kāṇḍa to kāṇḍa could be maintained most easily in the medium of prose.

The most general and by far the most important issue confronting someone who contemplates an English translation of the Vālmīki Rāmāyaṇa is that of the style, tone, and underlying ethos of the poem. [ Note 211 ] The educated western reader’s conception and expectation of what an epic poem is and should be are largely derived from Homer. In his still-illuminating essay, “On Translating Homer”, Matthew Arnold called attention to what he saw as the four principal stylistic virtues of the Homeric epics. Homer, he remarks, is rapid, plain, and direct in both his language and thought, and noble. [ Note 212 ] It is precisely these qualities of the Greek poet that we have come to regard as the standard for all epic poetry. But although many passages in the Rāmāyaṇa can be said to share something of Homer’s rapidity, directness, and elevation of style, the feeling of the whole poem is quite different from that of the “Iliad”. True, both poets rely heavily on the use of elaborate and highly formalized speeches. Both make heavy use of epithets, the choice of which is largely governed by meter. Both poets delight in the liberal use of descriptive adjectives and poetic figures, particularly the simile. Both authors are oral poets employing densely formulaic metrical discourse to sing of the quests and battles of ancient warrior-heroes. And yet, the feeling one gets from reading the two poets is very different.

Erich Auerbach, in his learned and provocative study of realism in western literature, argues that Homer’s well-known predilection for narrative digressions and epithets is a result of the poet’s ‘need for an externalization of phenomena in terms perceptible to the senses.’ [ Note 213 ] Auerbach argues persuasively that the epic style, as exemplified by Homer, knows no depths of narrative perspective but rather strives to present a world in which all events, thoughts, and actions are uniformly foregrounded in a uniformly illuminated present. Upon first consideration one might be tempted to say that this quality is shared equally by the poetry of Vālmīki and his followers. For Vālmīki, like Homer, is at pains to hold back nothing from his audience. Every thought, speech, and action of the epic characters is carefully set in the context of its antecedents and its consequences. Like Homer, the poets of the Rāmāyaṇa are deeply concerned with the visual description of their characters, their actions, and the places in which these actions take place. Yet a sense of stylistic parallel is diminished by a careful reading of the Rāmāyaṇa and the Greek epics.

Vālmīki is not so digressive as the poets of the Mahābhārata, and yet the pace of the poem is often slower than that of the greater epic and of the bulk of Homer. As with any work of originally oral poetry, the flow of the Rāmāyaṇa’s narrative is often impeded by formulaic phrases conveying only the information that one character has heard the speech of another and will now reply; but the occurrences of such phenomena seem especially numerous in the Rāmāyaṇa. Moreover, the Indian poet’s diction is not infrequently repetitive, relying on participles and gerunds that echo a previous finite verb to lend an often dreary continuity to the tale. Features such as this, which originally served both to provide time to the oral poet to compose successive verses and to provide continuity for a shifting audience, tend to become tedious to a literate audience of a printed version of the poem. [Note 214 ] Again, whereas Homer, in the manner of oral bardic poets, employs many metrically conditioned epithets, Vālmīki seems to take an unusual delight in their proliferation. Thus, a character in the Rāmāyaṇa may be burdened with three or more epithets or patronymics in a single verse. In many cases, a single such term, such as kausalyānandavardhana, or a term paired with a proper noun, such as lakṣmaṇo lakṣmīvardhana or rāvaṇo rākṣasādhipa, will occupy a full quarter of a verse. This sort of thing taken cumulatively slows the progress of the poem enormously and although it enhanced the appreciation of the original audience, often stands in the way of a contemporary reader’s enjoyment.

Aside from these tendencies, both the style and the content of the Rāmāyaṇa are, in the main, quite simple. Partly because of the absence of ambivalence in the characterizations of the epic’s central figures, they almost invariably think, speak, and act simply and directly. This simplicity is matched by the poet’s diction. The style tends to be paratactic and the periods are usually short. Only occasionally does a syntactic unit extend over one and a half or, at most, two of the thirty-two-syllable śloka verses that constitute the bulk of the text. The major exception to the rule of simple diction is the verses composed in meters longer than the śloka. Aside from the frequent usage of the forty-four-syllable upajāti meter, which varies the text somewhat but often modifies the style only insofar as it permits the packing of still more adjectives and epithets into a verse, these longer lines occur in two main contexts. Such verses occur sporadically in groups or singly at the end of sargas, where they slow the progress of the narrative and mark closure. Also, in the middle six books of the poem, longer verses appear in clusters, sometimes constituting an entire sarga in which the poet gives rein to his powers in creating highly wrought descriptions rich in rhetorical figuration. These form many of the poetic set pieces of the poem and lend it much of its charm. In both cases, the longer lines tend to slow the already stately progress of the story and interrupt somewhat the monotony of the various ślokas.

In our translation, we have striven to reproduce the paratactic style of the original on a verse-by-verse basis, while remaining within the limits of what we judged to be contemporary readability. Thus, we have rendered the basic śloka passages into straightforward English prose, avoiding, wherever possible, long periods and intricate constructions. When, as in the case of sentences that run over verse boundaries, we felt that the style of the original warranted it, we have used a somewhat more complex English syntax.

Another aspect of the Rāmāyaṇa that distinguishes it from the western epic is the way in which the poet acts as a mediator between the objects and events he describes and his audience. Homer’s descriptions, even though he makes frequent use of figures, are characterized by the accumulation of minute, objective, and almost clinically observed detail. As Auerbach notes, Homer is never far from the world of sense perception. As a result, Homeric description serves to focus the attention of the audience upon the perceptible nature of the person, thing, or act described. By way of contrast, Vālmīki’s descriptions, partly because of their unremitting accumulation of similes and partly because of the poet’s general preference of moral to physical attributes, have the effect of interposing the poet between the event and the audience. On the one hand, the imagination of the reader is constantly diverted from the objects in the story to the things with which they are compared; on the other hand, the text’s ethical and moral descriptions tend to filter the audience’s perception of the epic events through the implicit moral judgment of its author.

In order to illustrate this difference, it will be helpful to quote briefly two passages that describe the ends of the duels between Diomedes and Pandaros [ Note 215 ] and between Rāma and Kumbhakarṇa. [ Note 216 ] In the Greek poem, Diomedes, wounded by Pandaros’ arrow, parries a spear throw and hurls his own spear:

At this he made his cast,
his weapon being guided by Athena
to cleave Pandaros’ nose beside the eye
and shatter his white teeth: his tongue
the brazen spearhead severed, tip from root,
then plowing on came out beneath his chin.
He toppled from the car, and all his armor
clanged on him, shimmering. The horses
quivered and shied away; but life and spirit
ebbed from the broken man, he lay still.Vālmīki renders Rāma’s killing of Kumbhakarṇa in this way:

Then Rāma took up the arrow of Indra. Sharp, well-feathered, and perfect, it shone like the rays of the sun, had the speed of the wind, and seemed like the staff of Brahmā at the time of the world’s end. Its feathers were made beautiful with gold and diamonds, and it shone like the glare of the blazing sun. Its striking power was like that of great Indra’s mace or a lightning bolt. Rāma shot it at that roamer of the night, and once set in motion by Rāma’s arm, it looked like the flame of a smokeless fire as it sped on its course, like the thunderbolt of Śakra, lighting up the ten directions with its inherent luster. Just as Indra, smasher of citadels, once long ago severed the head of Vṛtra, the arrow cut off the head of the rākṣasa lord. Its teeth bared and its gorgeous earrings wildly swinging, the head looked like the peak of some huge mountain. Struck off by Rāma’s arrow, the rākṣasa’s head, which looked like a mountain, fell smashing roads, buildings, and gateways and knocking down the lofty ramparts of the city. But the rākṣasa, with his huge body that looked like the Himalayas, fell into the sea where, crushing sea monsters, shoals of great fish, and serpents, he at last entered the earth.

The passages deal with similar events but, obviously, quite differently. Where Homer is spare, Vālmīki is diffuse and repetitive, relying on the massive accumulation of adjectives and similes for his effect. Homer, who is in general by no means averse to the use of simile, in this stark passage is at pains to render the nature of Pandaros’ injury in unemotive, objective, and even clinical terms. The effect is powerful, even overwhelming, for the reader is forced to imagine, almost to experience, the agonizing passage of the spearhead through the stricken man’s skull. Homer achieves his effect using only two simple adjectives, ‘brazen’ and ‘white,’ both of which make direct appeal to our senses and with no similes. Elsewhere in his poem he delights in the extended simile and in the piling up of masses of descriptive adjectives, but, nonetheless, the effect is always to bring the audience close to an objective and sensorially grounded perception of objects, actions, and characters.

How different is the effect of the Rāmāyaṇa passage. Here the objects that principally absorb the attention of the poet, Rāma’s arrow and Kumbhakarṇa’s head, are all but obscured by a cloud of qualifiers and rhetorical figures. The first three verses of the passage quoted convey only three pieces of narrative information: Rāma selects an arrow and shoots it at his opponent, and the arrow flies on its way. This is communicated in some six or seven short words; the great bulk of the twelve lines of poetry consists of thirteen adjectives, seven of which function within similes. In the two following verses Rāma’s target, the demon’s head, is severed and falls. The head is qualified by five adjectives of which two express similes, while the act of decapitation itself is made the subject of a non-adjectival simile based on a conventional mythic reference. Finally, the fall of the now separated head and body (modified by an adjectival simile) is rendered in a straightforward but elaborate and extremely hyperbolic fashion. Only three or four of all the adjectives in the passage are simple words. The great majority of them are nominal compounds, many of four or more members. In one case, the adjective sabrahmadaṇḍāntakakālakalpam, ‘seeming like the staff of Brahmā at the end of the world,’ constitutes in itself a full eleven-syllable pāda or quarter verse.

The mere density of the modifiers and the richness of the figuration are not all that distinguish the passage from Homer’s. There is a fundamental difference in the kinds of adjectives and similes that the two poets have chosen and in their purposes in choosing them. Where the Greek poet strives to speak to the sensory experience of his audience, the Sanskrit bard aims to stir the emotions of awe and amazement with a dense texture of descriptions and comparisons that move us away from the world of mundane experience to that of the supernatural. Vālmīki tells us very little about the physical nature of Rāma’s arrow and what he does tell us is of small help to those who would visualize the weapon. We are told only that it is sharp and flawless [ Note 217 ] and that it is well-feathered. Later it is said that the feathered portion of the arrow is adorned with gold and gems. But all arrows are sharp, and even the second reference to the feathers leaves much to the imagination. What is more, it tends to remove us at once from the world of ordinary experience. For who, in real battle, uses arrows wrought with gold and diamonds? To know that the arrow was associated with the god Indra and that it is set in motion by Rāma’s arm does not enhance our ability to picture it. For the rest, we find that two qualities of the arrow engage the interest of the poet. He is concerned with the speed of the arrow, for it is twice said to have the swiftness of Indra’s lightning bolt and once that of the wind. He is also eager to represent the dazzling brilliance of the arrow; for he compares it first to the blazing sun, and, later, to a smokeless fire, mentioning further that its luster illuminates the ten directions. Once more, the emphasis is on exaltation and exaggeration, with the repeated mention of the vedic gods and the celestial powers serving to divert the mind of the audience from a mundane weapon to more lofty realms.

Related to the poet’s use of simile to draw the hearer from the world of ordinary experience is his pervasive use of hyperbole. Not only are Kumbhakarṇa’s head and body likened to mountains, but their fall is disastrous. The monster’s severed head destroys roads, buildings, and palisades, while his falling body wreaks havoc among the creatures of the sea. Yet there is nothing of Rabelais here. It does not appear that Vālmīki is aiming to be either comic or grotesque. The real function of passages such as this — and they constitute a considerable portion of the poem — is to create and sustain for the audience a sense of events utterly removed from everyday experience.

Homer is not averse to hyperbole, and he uses it frequently to enhance the grandeur of his subject. Thus, he refers often to the prodigies of strength of his heroes whom he represents as men of an earlier and more heroic age. Yet, here again, there is a marked difference between the two poets. When Homer has his heroes heft and throw with ease a boulder such as ‘no two men now living could lift,’ he not only refreshes for the audience its sense that these are not ordinary men, but through the very restraint of his hyperbole, he gives us a boulder that we can easily visualize and whose weight we can intuitively judge. How different is the feeling we get from Vālmīki’s statement that Rāma, while still a boy, lifts, wields, and snaps a divine bow that five thousand powerful men are barely able to drag on a wheeled cart. The very exuberance of the exaggeration serves to remove both the object and the event from the realm of our sense experience. The unrestrained and overwhelming nature of Vālmīki’s use of hyperbole, the creation of a world in which tens of millions of monkeys aid a prince in battle, in which a monkey can leap across the ocean holding a mountain peak, and in which the sixty thousand sons of a king can dig the ocean may — at least for a modern audience — lead to a deadening rather than a heightening of effect.

On the basis of Vālmīki’s fabulous treatment of the battle scenes, as compared to their relatively realistic representation in the Mahābhārata, Macdonell argued that the poet had never witnessed a real battle. [ Note 218 ] This conclusion is not, in fact, justified by the evidence. We may be quite certain that the epic bards had ample opportunity to observe real monkeys, yet their own needs and those of their audience led them to the creation and elaboration of the fantasy figure of Hanumān. [ Note 219 ] We cannot then assume, with Macdonell, that the poet’s recourse to exaggeration and fantasy indicates any lack of opportunity to observe real phenomena. On the contrary, we should seek to understand the motives that lead them consistently to idealize and exaggerate the phenomena they did observe, distancing them as much as possible from the realm of sense perception and reality. [ Note 220 ]

Even the profusion of descriptive epithets that cluster around the characters of the Rāmāyaṇa fail, in the main, to appeal to the perception of the senses. Homer’s epithets, as often as not, allude to a man’s particular skill, the loudness of his war cry, the swiftness of his feet, or some other physical attribute, so that they are, as Auerbach has remarked, ‘in the final analysis … traceable to the same need for an externalization of phenomena in terms perceptible to the senses.’[Note 221] But although a number of Vālmīki’s epithets, such as ‘great-armed,’ are similar to those of Homer, they are vastly outnumbered by epithets that allude to moral, ethical, intellectual, and emotional rather than physical qualities. Principal, secondary, and peripheral figures alike are endlessly described as righteous, great, mighty, truthful, noble, and the like. These qualities are applied almost indiscriminately and are generally vague, often obscure, and always sensorially opaque. Such physical descriptions of characters and places as the text does contain tend to be conventionalized, as in the description of Ayodhyā, the forest of Tāṭakā, and Rāma; fantasied, as in the description of the ten-headed Rāvaṇa or his colossal brother, Kumbhakarṇa; or difficult for the modern audience, at least, to visualize, as where Rāma is said to be ‘shell-necked,’ ‘great jawed,’ or to have ‘a hidden collarbone.’ [ Note 222 ] A fair number of epithets are employed simply because they echo the name of the character to whom they are applied. Thus, Lakṣmaṇa may be called lakṣaṇasampanna, ‘endowed with (auspicious) signs,’ or Rāma described as lokābhirāma, ‘delight of the world.’ Such terms merely heighten the modern reader’s sense that the poet is verbose. [ Note 223 ]

The point of this somewhat protracted comparison between Vālmīki and Homer is not to denigrate the literary achievement of the former. For, as Ingalls has so passionately argued, it would be unfair to expect the literature of one tradition to conform to the aesthetic norms of another.[Note 224] If Homer’s intended audience might have found Vālmīki verbose, sentimental, nebulous, and boring, so might an ancient Indian have found the Greek poet terse, crude, ignoble, and lacking in religious and moral sensibility. The heirs of Homer’s literary sensibility must be aware of these differences if the poem is to be read on its own terms without an effort to adapt it to western tastes.

We have therefore consistently avoided the temptation to improve upon Vālmīki in an effort to make the poem more attractive to a western audience. Like Griffith, we have attempted to ‘give the poet as he is,’ rather than ‘to represent him as European taste might prefer him to be.’ [ Note 225 ] Where the poem is repetitive, our translation will be repetitive, and where it is tedious, our translation will, I fear, be equally dull. Yet many of the virtues for which the poem is prized in India should survive both translation and shift in the intended audience. The poem is remarkable for its even tone and diction, the nobility of its characterizations, the beauty of many of its conventionalized descriptions, and above all, its deeply felt rendering of the emotions of grief, sorrow, and a sense of loss. But our readers should not expect in this translation a work that will strike them at once with its universal appeal or uncomplicated charm. The purpose of our efforts was not to render an amusing story into easily digestible form. That has been attempted more than once in the case of the Rāmāyaṇa with indifferent results. Rather, we have sought to put before a scholarly and a general audience as accurate and as readable a rendition of this great epic as was possible, retaining as much as an alien language and idiom would permit of the measured pace, lofty tone, and gravity of our original. In our efforts to accomplish this goal, we have had to come to terms with a great number of features of the Sanskrit text that, although they present no apparent difficulties to Indians who are able to read the poem for pleasure and/or religious purposes, are often extremely problematic when the work is subjected to the close reading necessary for a scholarly translation. [ Note 226 ]

The special difficulties that the Rāmāyaṇa presents to a western translator are of two kinds. The first are the problems that result from its being a product of a time and a culture far removed from ours. Some of these problems involve simple questions of realia — the weaponry, architecture, costume, flora, and fauna of ancient India. In some cases the language of the poet is so imprecise that neither we nor the traditional commentators can be absolutely certain that we fully understand what is happening in the story. [ Note 227 ]

The second, and to my mind the most pervasive, significant, and intractable set of problems derives from the poet’s general disinclination to address the perceptions of the senses. The innumerable epithets and descriptive adjectives based on moral rather than physical attributes of the characters in the epic drama are difficult to conceptualize accurately and therefore hard to translate with any conviction that one has found an appropriate English equivalent. This problem is aggravated by the fact that our contemporary literary idiom has nothing like the vast Sanskrit lexicon of words suggestive of virtue and vice. Thus, for example, although the epic bards were quite at home with derogatory and pejorative nouns such as durātman, ‘wicked man,’ most of the simple nouns with this connotation in English are either archaic or colloquial. Neither of these modes of diction, we feel, adequately represents the tone of the original. The common adjectives and epithets based on the word tejas, ‘brilliance,’ are another good example — terms such as tejasvin and mahātejas. The word tejas may mean in the epic ‘brilliance,’ ‘moral, spiritual, or physical power,’ or ‘semen.’ We have decided that it is best in most cases to keep the ambiguity of the term and have generally rendered the epithets as ‘powerful.’

Another problem of this type is presented by the epic poets’ seemingly limitless fondness for compound epithets ending in ātman, ‘soul, self, mind, or body.’ It has become a common practice in translating Sanskrit to render this term as ‘Soul.’ Thus, the common epithet mahātman is often rendered as ‘Great-Souled one.’ We think this is a mistake. Not only is it vapid and inelegant English, but the translation is based on a technical application of a general term made in the upaniṣads and texts on vedānta, in a way for the most part alien to the intentions of the bardic narrators of the epics. If the philosophical term for ‘Soul’ were really intended here, one would hardly expect it to be modified, as it is invariably in these compounds, by adjectives. In what sense can the Soul of the vedāntins be great, accomplished, unskilled, etc.? In general we regard this term, when it appears at the end of compounds, as a reflexive marker such as is widely used in Sanskrit to strengthen simple adjectives by turning them, through the formation of bahuvrīhi compounds, into true epithets. When used in this way, the term is essentially untranslatable, and, in fact, unnecessary to translate. Thus, we translate mahātman as ‘great,’ dharmātman as ‘righteous,’ kṛtātman as ‘accomplished,’ and so on.

One characteristic use of epithets in Sanskrit texts is the substitution of an attributive term — patronymic, descriptive, or referent to some legendary feat — for a proper name. Whereas Homer never employs a bare epithet, Vālmīki often uses one or more in lieu of a name; the effect is that of a kenning. A major mythological figure such as the god Indra may be referred to by one or more of dozens of possible epithets, the choice being largely governed by metrical, alliterative, or other formal considerations. Although all of these epithets are either known or transparent to an audience steeped in Hindu mythology, they can only frustrate and baffle the average western reader. But to substitute more familiar, or at any rate more pronounceable names, like Rāma and Indra, would be not only to misrepresent the original but also to deprive the reader of the variety and richness of epithets that lend color to the text. On the other hand, the practice adopted in other translations of Indian epic texts of providing footnotes with unfamiliar epithets would only serve to clutter the page and reduce rather than enhance readability. Our solution to the problem of independent epithets is to provide the name in the running text of the translation together with the translated epithet. Thus, for example, the epithets sahasrākṣa and purandara have been translated as ‘thousand-eyed Indra’ and ‘Indra, smasher of citadels,’ respectively. Only in the case of recurrent compound epithets used to designate the central characters in the epic story was it thought unnecessary to provide the more familiar proper nouns. In the case of janakātmajā, for example, it was felt that ‘Janaka’s daughter’ was sufficient and did not require the name Sītā to complement it.

In the case of shorter, uncompounded epithets, particularly patronymics and those that indicate a country or city of origin, it was decided to leave the original terms untranslated on the grounds that, on the one hand, collapsing the names back to familiar ones would only diminish the color and variety of the text, whereas translation would only yield the equally unfamiliar name of an ancestor. Thus, although the compound epithets of Rāma, daśarathātmaja and raghunandana, are rendered as ‘the son of Daśaratha’ and ‘delight of the Raghus,’ respectively, uncompounded names such as Jānakī, Saumitri, Vaidehī, Dāśarathi, Vāsava, and the ubiquitous Rāghava have been left untranslated. By this set of modest compromises, it is hoped that something of the flavor and the variety of the original has been preserved without compromising the reader’s ability to follow the story.

Those who are able to follow the Sanskrit of the critical edition will find that we have taken a few additional liberties with the text for the sake of intelligibility. It was frequently felt advisable to break up long syntactic units into shorter English sentences. In so doing it was sometimes necessary to shift or repeat, with or without changes, a single verb that in the Sanskrit is made to govern an extremely long clause or series of clauses. An example of this may be seen at 1.5.1-3. Here, as elsewhere, we have departed from our normal practice of translating on a verse-by-verse basis and numbering each of the verses to correspond to the constituted text of the critical edition. In these cases we provide only the inclusive numbers of the verses in question.

A similar rearrangement of the elements of the original has been made in cases where a single Sanskrit sentence contains more epithets, adjectives, or vocatives than its English counterpart can comfortably hold. In such instances our concern for maintaining a readable English style has led us to separate elements that were originally juxtaposed, turn vocatives into phrases, and generally distribute qualifiers in a way that seemed to us more intelligible to the English reader. In passages where the original uses a pronoun without a clear antecedent, and yet there is no serious doubt as to which character is intended, we have inserted the name, without indicating the insertion either by brackets or notes. In those few cases in which an antecedent is really uncertain, the question is discussed in the notes.

Another problem the translator of Sanskrit epic poetry has to confront is that of the inevitable loss of the kinesthetic effect this poetry had upon its original audience. The Rāmāyaṇa, as the second and fourth chapters of the Bālakāṇḍa tell us, was intended to be sung to the accompaniment of musical instruments, and the pleasure in hearing this song-poem derived as much from the sonorous chant as from its edifying story. The work is, after all, the prototype of what the Indian tradition calls śravyakāvya, ‘poetry that is intended to be heard.’ Even today, the poem, like its vernacular derivatives, is normally read aloud to a variety of gānarītis, or traditional musical chants, and is not infrequently actually sung to classical rāgas. Hearing the poem chanted, or chanting it oneself, produces an extremely pleasant effect, impossible to duplicate through silent reading. [ Note 228 ]

This characteristic droning effect is enhanced by a certain formulaic repetitiveness inherent in the bardic style of the work. In the poem’s extensive passages of straightforward narrative, for example, verse after verse begins with the adverbial form tatas, ‘then.’ Moreover, continuity between verses is often sustained by beginning a verse with a participial form of a verb expressed or implied at the end of the preceding verse. If the translation were to attempt to render these repeated expressions whose function can only be appreciated in the context of oral composition and recitation, it would result in an effect far from pleasurable.

Consistent translation of specific Sanskrit terms posed another problem. In general the substitution of a given English word for each occurrence of the corresponding Sanskrit term was one of the goals of this, as of any serious translation. Yet the nature of the Sanskrit lexicon, the epic no less than the language of the classical poets, is such as to doom any attempt at mechanical correspondences to failure. On the one hand, Sanskrit shows an extraordinary — perhaps unequalled — degree of true synonymity. In many cases there can be dozens of words that signify the same thing with no perceptible nuance of meaning. [ Note 229 ] It is often impossible to find a discrete and acceptable English term for each of the many synonyms of such words as ‘son,’ ‘king,’ and the like. In some cases, where particular synonyms lend themselves to analytic or etymological representation in English, this has been done in order to retain some sense of the variety of the original. Thus the term nandana at the end of compounds, where it idiomatically signifies ‘descendant’ or ‘offspring,’ has been given its etymological value and translated as ‘delight of.’ Similarly, a standard synonym for ‘king,’ the compound pṛthivīpāla, has been rendered literally as ‘protector of the earth.’

Synonymity is not, however, the only problem to confront the translator. Just as many different words may have the same meaning in Sanskrit, so may a single word have a bewildering variety of meanings. Some of traditional India’s most highly charged terms, words such as dharma, karma, tapas, and so on, have wide ranges of meaning, and it is, as the pandits say, possible to determine the exact meaning only from the specific context. In many cases, where terms do not exhibit this polysemic character, a one-to-one translation is both possible and desirable. Thus, we have rendered every occurrence of the term ṛṣi with the English word ‘seer’ and each appearance of the partially overlapping term muni as ‘sage,’ in order to reflect the distribution of these two common terms. The important term dharma, however, has a number of related but quite different meanings. When it is used generally, as in compounds and derivatives like dharmajña and dhārmika, ‘knowing dharma’ and ‘dharmic,’ we have translated it, depending on the English context, as ‘righteousness,’ ‘what is right,’ and so on. On the other hand, the term has various other meanings. At 1.24.16 it refers to duty, at 2.7.19 it means nothing more than ‘custom,’ and at 2.23.4 it signifies ‘traditional insignia.’ In different contexts the term tapas may refer to the practice of bodily austerities or the supernatural powers that are thought to derive from them. But at 2.23.3 the term tapasvinī, ‘woman possessed of tapas,’ clearly has the idiomatic sense of ‘poor thing.’

One could multiply examples indefinitely, but the point should by now be clear. We have striven, through our close reading of the text and our study of the commentaries and relevant testimonia, to represent accurately the meanings of these multivalent terms, not by mechanically applying a set of unvarying lexical approximations, but by giving careful consideration to their various and often subtly differing contexts.

One of our fundamental decisions concerning this translation of the Rāmāyaṇa was that it should contain a minimum of untranslated words. To avoid providing an appropriate English equivalent for terms such as dharma is to abdicate the translator’s responsibility. Accordingly it has been our consistent practice in this translation to render all Sanskrit words and terms that have appropriate English equivalents into English.

Nonetheless, it is neither necessary nor desirable in the translation of a document of a culture so far removed from ours in space and time wholly to eliminate words in the original language. For one thing, such words remind the reader that this is a poem of an alien culture rooted in an unfamiliar landscape. Then too, certain terms are either awkward or unwieldy in translation, whereas others — if only a few — do not require it. The most numerous category of words untranslated is that of proper nouns. We have said that epithets in the form of descriptive adjectives are regularly translated even when they come to be treated, as, for example, daśagrīva, ‘ten-headed Rāvaṇa,’ as alternate proper nouns. On the other hand, principal names for characters are never translated, even when, as in the case of Daśaratha (‘having ten chariots’), they are of the same form as the translated epithets. This holds true also for epithets derived from the names of characters and countries. Thus, for example, the text’s numerous patronymics have been left in Sanskrit. Also untranslated are the numerous terms for fauna and especially flora for which no readily recognizable English name is available. This practice has also generally been followed in the case of technical terms, such as, for example, the mārga mode of ancient Indian music and, more generally, in the case of culture-bound terms for important classes of supernatural beings when possible English terms seemed overly restrictive or misleading. Thus the terms rākṣasa, yakṣa, gandharva, apsaras, and so on, have been left untranslated, whereas others, such as deva, nāga, cāraṇa, and siddha, have been rendered by ‘god,’ ‘serpent,’ ‘celestial bard,’ and ‘perfected being,’ respectively. Sanskrit words that have come to be generally recognized by English speakers are translated without diacritical marks, such as ‘guru,’ ‘ashram,’ ‘kshatriya,’ and ‘brahman.’ In a very few cases technical or semi-technical Sanskrit terms have been rendered by English equivalents chosen more for their transparency to the reader than because they represent precise equivalents. Thus, for example, the text’s two common terms for measures of distance, yojana and krośa, have been represented by the English words ‘league’ and ‘mile,’ respectively, for these latter words represent the relationship and something of the feeling of the originals, a feeling that would have been obscured or lost in an effort to calculate and represent the actual distances.

As Professor Nathan reminds us in his brief prolegomena to the translation given above, the Vālmīki Rāmāyaṇa is preeminently a work intended to be heard, and heard, as I have suggested, with a kind of relaxed absorption into its sonorous drone and its strange and yet compelling reality. It was our task to present this oral poem of ancient India to an audience of modern readers of English in the guise of a written prose-epic. In doing so we have constantly striven to create as easy a flow of contemporary diction as close adherence to the verse structure of the original would permit. For, we realized, only in freeing ourselves to a certain extent from the letter of the oral style would we retain any hope of transmitting some echo of its spirit to those who would try to grasp at first hand this poem that lies so near the heart of India.

One additional feature of the present work is its extensive annotation. In this the translation is virtually unique. Some previous editors and translators of the poem provided sporadically dense annotations of one or sometimes two kāṇḍas of the poem, but those who have translated the entire work have been generally satisfied with an extremely sparse set of notes most of which simply provide cultural data for the reader unfamiliar with traditional India. Gorresio, with his two hundred and thirty odd notes for the Bālakāṇḍa, and rather fewer for succeeding books, has as copious an annotation as any and yet his notes, like those of his colleagues, tend to concern themselves chiefly with cultural, mythological, and botanical information for the edification of the European audience. [ Note 230 ] Few scholars, aside from Peterson and a few other authors of school-texts of the Bālakāṇḍa, [ Note 23 1 ] have attempted to address systematically the numerous serious problems that the text of the Rāmāyaṇa presents to the scholar. Moreover, even such noble, if partial, efforts as these were seriously hampered by the absence of a critical edition of the text or a readily available means of making important text-critical judgments.

The issue of close textual analysis is extremely important for any attempt at translation of the epic. For although the style, diction, and lexicon of the poem are for the most part simple, the exact sense of a particular word, phrase, or verse is frequently quite difficult to ascertain. Paradoxically, or so it seems, we have found it more difficult to be sure of the meaning of a Rāmāyaṇa passage than of passages in far more elaborate and difficult genres of Sanskrit literature. Thus we agreed that the value of a translation of the critically edited text would be materially diminished unless it were accompanied by annotations that attempted to deal with every unclear and difficult passage in the poem. Wherever such a passage was encountered, the translation represents our final decision as to the most likely meaning. Those who are interested in the issues taken into consideration in arriving at these decisions, or in disputing them, will find in the notes the textual, recensional, commentarial, and other material that were judged relevant to the problem.

In general, the major thrust of the annotation is to aid in the reading and comprehension of the translation. General cultural, geographical, and botanical matters have been ignored except where a particular item bears on our understanding of the relevant passage or has some direct bearing on a textual or recensional question. Culture-specific phenomena whose significance is not readily apparent are generally glossed in the translation. Thus, for example, the common gesture of deference suggested by the term prāñjali is translated as ‘cupping the hands in reverence,’ thus obviating the need for a note. The names of mythological and legendary figures appearing in the text are identified briefly in the Glossary and therefore also do not require annotation.

We have occasionally departed from this practice of elucidating only the text. When the explanation of a passage, problematic or not, by one or more of the commentators is sufficiently interesting, we have included it in a note. This has been done for two reasons. First, the ten or so Sanskrit commentators whose works we have read in connection with this translation represent, collectively, the closest approximation available to us of the audience for which the poem was intended. The reaction of these commentators — even when they are, in our opinion, wrong in their interpretations — often sheds valuable light on the reception of the epic in traditional Hindu circles and constitutes a part of our understanding of the role and destiny of this crucial text. Second, although general cultural, religious, mythological, historical, and literary information about traditional India is readily available in various translations and scholarly studies, the learned and often highly illuminating contributions of the Rāmāyaṇa commentators are both unknown and largely inaccessible to all but a handful of Sanskrit scholars. None of their works has ever been translated nor, in all probability, will they ever be. Limitations of time, space, and readership naturally preclude our inclusion of much of the contents of these invaluable works. Nonetheless, we felt that we would be providing a real service to the general as well as the scholarly audience by providing a sample of the more interesting observations and arguments of the exponents of the rich tradition of the Rāmāyaṇa commentaries.

A detailed study of the works and interrelationships of the major commentators will have to await a future volume of Rāmāyaṇa studies planned as a companion volume to the translation. Nonetheless, a few words on this subject are in order here. In the preparation of the translation and annotation of the epic, ten Sanskrit commentaries have been examined. These are, in roughly chronological order: the Vivekatilaka of Varadarāja Uḍāli, the Rāmānujīya attributed to Rāmānuja, the Tattvadīpikā of Maheśvaratīrtha, the Amṛtakataka of Mādhavayogin, the Rāmāyaṇabhūṣaṇa of Govindarāja, the Rāmāyaṇatilaka of Nāgeśa Bhaṭṭa, the Rāmāyaṇaśiromaṇi of Vaṃśīdhara (Bansidhara) Śivasahāya, and the Dharmākūṭam of Tryambaka Makhin. [ Note 232 ] In addition, extracts from the commentaries of Satyatīrtha, [ Note 233 ] Munibhāva, and Sarvajñanārāyaṇa have been used.

These works range in date from around the middle of the thirteenth century a.d. (in the case of Varadarāja) to the eighteenth (in the case of Tryambaka), and vary from the sparse glosses of the former to the dense and often copious commentaries of Mādhavayogin, Govindarāja, Nāgeśa, and Śivasahāya, who have at least something to say on almost every verse in the poem. These works are immensely valuable and, indeed, it would be rash to attempt any serious reading of the poem without reading at least five of the more extensive commentaries. Nonetheless, one obvious problem with the surviving commentaries is the fact that the earliest of them dates from a time nearly two millennia removed from that of the probable composition of the core of their text. As a result they are often as puzzled as we by certain bits of material, and there are places where it seems clear that the original meaning of the text has been lost. In spite of this, the unusually conservative nature of traditional Indian culture and of the Sanskrit language in particular, coupled with the maintenance of a long and unbroken tradition of reading and discussing the poem, make the interval of two thousand years less significant than they would be in less stable cultural contexts.

One additional factor must, however, be taken into consideration. The authors of all the surviving commentaries on the Rāmāyaṇa were devout Vaishnavas for whom the primary significance of the poem was theological. For these men, most of whom were products of the passionate and compelling world of south Indian devotionalism, every action in the epic — every line — was fraught with clear or hidden theological significance. But, as has been argued above, it is clear that with a few exceptions, most of which are to be found in the later portions of the Bāla and Uttara Kāṇḍas, the text posits no special relationship between its hero and Viṣṇu, much less their identity. In keeping with their passionately held view of the nature and significance of the Rāmāyaṇa, the commentators very frequently see as possessing profound and often arcane theological meaning verses and passages that, we believe, had originally nothing whatever to do with theology. This tendency, we consequently hold, causes a consistent overinterpretation and even falsification of the text. [ Note 234 ]

If such theological interpretations and sectarian pleading were all there was to the Sanskrit commentators, one could safely ignore them, at least as serious aids to a better understanding of the Rāmāyaṇa. But like many of the best minds of traditional India, the commentators were more than special pleaders. They were scholars, connoisseurs of poetry, and grammarians, and a great deal of their exegesis shows little or no concern with theology. The commentaries frequently disagree among themselves, and often a single scholiast offers a series of alternative explanations of the same passage. In passage after passage the best of the commentators, particularly Govindarāja, Maheśvaratīrtha, and Nāgeśa, provide learned, judicious, and convincing interpretations of readings that without their help would have remained obscure to us. We gratefully acknowledge the enormous contribution to Rāmāyaṇa scholarship of the commentators, and as a perusal of our notes will show, have relied heavily, although never uncritically, upon their learning.

Page 1 | Page 2 | Page 3


1  Macdonell 1919, p. 574.
2  Schlegel 1829. Schlegel’s work was not the first piece of western Rāmāyaṇa scholarship. Carey and Marshman had published a rather confused edition of the text as early as 1808, and Frederich Schlegel had offered some tentative translations from Vālmīki in the closing decade of the eighteenth century. European orientalists and travelers must have known of the poem even earlier.
3  One could compile a quite extensive negative bibliography of Rāmāyaṇa studies. A serious analysis of these often amusing works would constitute an important contribution to our understanding of the peculiar effect the poem has had on its audiences. Some striking examples of such works are: Mehta 1941, Buck 1976, and Iyer 1941.
4  For a discussion of the manuscripts of the Rāmāyaṇa, see Bhatt 1960, pp. xiii-xxix.
5  Ibid., p. xix.
6  On the description, genesis, and interrelation of the recensions, see Bhatt 1960; Ruben 1936; Jacobi 1893, pp. 1-23; Bulcke 1949 and 1951; and van Daalen 1980, pp. 13-14.
7  There appears to be in addition a western subrecension (W), heavily contaminated by both S and NW and represented in the Bālakāṇḍa by four Devanāgarī manuscripts (Bhatt 1960, pp. xiii-xiv; xxxii). Wirtz’s dissertation on the ‘Westliche Rezension des Rāmāyaṇa’ (Wirtz 1894) is, as Bhatt has noted (Bhatt 1960, p. xxi), concerned with what is, in fact, northwestern. Although Bhatt, in his introduction to the Bālakāṇḍa, originally accepted a separate western subrecension, he later, in his introduction to the Araṇyakāṇḍa (Bhatt 1963, pp. xxiii-xxv), felt that there was not enough manuscript evidence to support an independent western version. Other editors of the critical edition, however, have disagreed with him. Vaidya (1971, p. xv) apparently accepts a separate western recension, despite Bhatt’s arguments. Mankad (1965, pp. xxvi-xxviii) felt that evidence tended to support a separate recension for the Kiṣkindhākāṇḍa. Jhala (1966, pp. xxvii-xxx), like Mankad, felt that an independent western version ‘would be justified.’ Shah (1975, p. 23) briefly summarizes the different positions of these editors and feels that various manuscript evidence of the Uttarakāṇḍa supports the existence of an independent western recension.
8  Macdonell 1900, pp. 304-305; Jacobi (1893, p. 4) calculates on the basis of a thirty-sarga sample of the fourth book that his versions C and B (our S and the Bengal version of NE) have, respectively, only 57 percent and 66 percent of their text in common.
9  Cf. Bhatt 1960, p. xx and van Daalen 1980, pp. 5-8. Hopkins (1926, p. 19) is overly pessimistic in his feeling that no textual reconstruction of the text is possible. For further and more detailed discussion of the textual issue, see the section below on text history and the critical edition.
10  In fact, many of the subsequent versions of the Rāma story — Hindu, Buddhist, and Jain — condense the narrative considerably. In connection with the development of the Rāma cult, there developed a tradition of extreme compression of what is considered the essence of the tale into the space of a hundred verses, ten verses, one verse, and finally, into the recitation of the single saving name, Rāma. For a further discussion of this phenomenon, see our treatment of the first sarga of the epic, the so-called Saṃkṣipta, or abridged Rāmāyaṇa, below; and Masson 1980, p. 100 note 10.
11  Cf. Macdonell 1900, pp. 312-13.
12  It is interesting that, although Rāma provides a casuistic and, finally, unconvincing series of justifications for this seemingly cowardly act (4.18.18-39), the tradition has never been wholly comfortable with what it continues to regard as a stain on the hero’s character. See Masson 1976 and 1980, pp. 95-966.
13  The Sanskrit word sundara means ‘beautiful.’ The title was therefore taken by Jacobi (1893, p. 124) and, after him, by Winternitz (1927, vol. 1, p. 490) to mean, ‘The Beautiful Book.’ This, however, makes little sense, for it does not appear that Sundara is so different in style, tone, or content from the two preceding books to justify its being singled out for its beauty by any criterion. Moreover, the titles of the other books refer to either subject matter or location, and it seems unlikely that the type of kāṇḍa name would change for this one book. Some writers have argued that the word Sundara is a place name, a reference to the Sunda islands or straits of Southeast Asia (see, for example, Mehta 1941, pp. 187-89). But this explanation, too, lacks any strong evidence to support it. I am inclined to agree with Jhala 1966, p. xxii that so far the question lacks a solution.
14  This passage, like several of its type in Sundara, appears to have been interpolated at some point in the textual prehistory of the epic, for the incidents described in it seem to be unknown to the characters in subsequent passages, a fact that has been noted by several scholars. See Jacobi 1893, pp. 33-34.
15  The two issues are to some extent connected in that both the poem (in the form in which we have it) and the unanimous tradition of India represent the sage Vālmīki, the legendary author of the work, as a contemporary of its hero and, indeed, a participant in the epic events. One of the first public recitations of the poem is said to have been given by Vālmīki’s disciples, Rāma’s sons Lava and Kuśa, in Rāma’s presence. Cf. 1.4 and 7.85.
16  See Ramaswami Sastri 1944, pp. 23-24.
17  See Weber 1870, pp. 44-63, and Ramaswami Sastri 1944, p. 39. Gorresio (1843, p. xcix) cites the opinions of Jones, Bendy, and Tod, who place Rāma in 2029, 950, and 1100 b.c., respectively. Gorresio, himself a keen student of the epic, dated the poem to the thirteenth century b.c. (1843, p. c). These early opinions are based on often faulty premises. Nonetheless, as we shall see, some of the positions taken with regard to the absolute and relative chronology of the Rāmāyaṇa are still defended today.
18  Jacobi 1893, pp. 50-59. Jacobi developed a suggestion first made by Holtzmann (1841, pp. 36ff.), whom he quotes. van Daalen (1980, pp. 1-2, 223) sustains Jacobi’s views on the Bāla, that is, that some portions of it belong to the oldest stratum of the text. On the basis of his own text-critical studies, he posits a single poet for the older portions of Bāla. There are a number of reasons for regarding Uttara as a later work even than Bāla (see Guruge 1960, p. 32). A number of scholars, particularly in India, have, however, seen some portions of Uttara as belonging to an early stratum of the text on literary critical grounds. See Kibe 1947, pp. 321ff. and Guruge 1960, p. 32.
19  Some of the more useful studies in this area are: Gorresio 1843; Michelson 1904; Roussel 1910; Brockington 1969b and 1970; Vrat 1964; the numerous valuable articles of N. M. Sen (see bibliography); and van Daalen 1980.
20  Most recently van Daalen (1980) has argued that the ‘irregularities’ are characteristic of late portions of the text and were largely alien to the original poet.
21  van Daalen’s disclaimer, ‘“Irregularities” have not been defined in this study in terms of forms contravening particular rules of Pāṇini’ (van Daalen 1980, p. 37) is confusing, especially in the light of his statement on the same page, that ‘the above does not mean that Pāṇini is not the frame of reference with previous collectors of irregularities and, consequently, implicitly in this study in many cases, since the collection of §4 is the summation of the previous collections as far as the items chosen for study are concerned.’ In any case, the analysis of ‘irregularities’ is problematic without some standard for what is regular. See Goldman 1982.
22  See Roussel 1910, p. 6, and Macdonell 1900, pp. 311-12.
23  For discussions of the common features, interrelationships, mutual influence, and metrical and common stylistic features of the two epics, see Hopkins 1901, Kane 1966, and Sukthankar 1941.
24  The tradition is not unanimous in making this distinction. The MBh, for example, refers to itself as kāvya. See, for example, the famous passage at the beginning of the Ādiparvan where, in an encounter parallel to that of Vālmīki and Brahmā at Rām 1.2. Vyāsa and the god discuss the creation of the former’s poem (MBh, crit. ed., Ādiparvan App. 1). Some writers on alaṃkāraśāstra regard Rām as an example of itihāsa and group it with the MBh for purposes of criticism. For a discussion of Abhinavagupta’s handling of the epics, see Masson 1969, pp. 78-84 and 103-12.
25  See Macdonell 1919; Winternitz 1927, vol. 1, pp. 324, 506.
26  In response to this objection, Winternitz was driven to the somewhat circular argument that ‘The Purāṇas have always retained these prose formulas in order to preserve the appearance of antiquity’ (1927, vol. 1, p. 506 note 1).
27  In any case, as we shall see, this ‘advance’ on the part of the Rāmāyaṇa poets often has the effect of slowing the epic text by dogging it with tedious formulae that merely mark the end of one speech and the beginning of another, whereas the ‘archaism’ of the Mahābhārata permits the poets to free themselves and their text of what is often a stylistic disaster. It seems to its an equally tenable argument that the poets of the MBh have made a stylistic advance over the Rām.
28  See, for example, the elaborate and useful treatment of Guruge 1960, and the works of Khan 1965, Vyas 1967, and Sharma 1971, which cull the text for information on the subjects of realia, social, economic, and political life, the arts, religion, and so on.
29  In fact, the practice is mentioned only once in the Rām, in a passage in the late Uttarakāṇḍa (7.17.13), where Vedavatī, who is represented as a prior incarnation of Sītā, tells Rāvaṇa that her mother had burned herself on her husband’s funeral pyre. On this point see Lassen 1858, p. 592; Hopkins 1889, pp. 116-17, 314-15; 1901, p. 149, Guruge 1960, pp. 203-204; Sharma 1971, pp. 95-98; and Meyer 1952, pp. 412-14.
30  See Guruge 1960, p. 201; Sharma 1971, p. 437; and Meyer 1952, pp. 165-73. At one point the MBh ascribes the practice of niyojana to the Ikṣvāku dynasty: at MBh 1.168.11-23, the sage Vasiṣṭha, the hereditary purohita, or family priest, of the Ikṣvākus, is said to have fathered an heir for King Kalmāṣapāda upon his queen at the king’s own request (see Goldman 1978, pp. 356-57). Although this practice is apparently eschewed in the royal family of the Rām, there is some evidence that it may lie at the bottom of the Bālakāṇḍa’s rather awkward introduction of the sage Ṛśyaśṛṅga into the account of Daśaratha’s efforts to procure an heir. If this is so, then the lack of instances of the levirate in the Rām cannot be used as an argument of the epic’s priority to the MBh.
31  See, for example, Guruge 1960, pp. 197-99.
32  Jacobi 1893, pp. 100-107. Jacobi’s arguments put to rest the issue of a post-Buddhistic date for the Rām raised by scholars such as Wheeler (1867-81, vol. 2, p. lxxiv and Weber (1870, pp. 1-13). For this reason, we shall not discuss the theses that they put forward. Moreover, the critical edition has shown the epic’s one explicit reference to Buddhism to be an interpolation. See Ayodhyākāṇḍa, crit. app., *2241, 14-15.
33  See Guruge 1960, pp. 51-80. It is interesting in this connection that, although the poem shows, in a textually well-supported passage from what is generally regarded as its oldest stratum (2.62.9-15), at least a general familiarity with north central and northwestern India, including sites central to the story of the MBh such as Hastinapura, Pañcāladeśa, and Kurujāṅgala, its authors nowhere show any familiarity with the characters or events of the longer epic. The absence of references to the MBh in the Rām has been noted many times before and has been urged as evidence of the priority of the latter (Jacobi 1893, p. 70). On the other hand, the significance of the Rām’s knowledge of Kurukṣetra has been perhaps insufficiently appreciated. It would appear to lend additional support to the supposition that the Rām was substantially completed at a time before the epic of the Kauravas had gained very wide currency in north India, for, had that not been the case, we would expect at least some passing reference to the events of the Bhārata war that took place in the regions named.
34  Jacobi 1893, pp. 101-102.
35  Cf. the treatments of the foundations of Kauṣāmbī, Mahodaya, Dharmāraṇya, Girivraja, and Vasumatī at 1.31.3-8, where the region of the latter is called Māgadhī.
36  Cf. Law 1951, pp. 23-27. One must treat such an argument from ignorance with great care. Nonetheless, the fact that the Bālakāṇḍa poet is at such pains to provide the legendary and historical traditions connected with the sites encountered during Rāma’s journey from Ayodhyā to Mithilā lends credence to the proposition that, if an important urban site at Pāṭaliputra had been known, it would have been mentioned.
37  7.98.5. Here Śrāvastī (crit. ed., Śrāvatī) is established as the capital of North Kosala, to be ruled by Rāma’s son Lava. South Kosala is to be governed from Kuśāvatī by Rāma’s other son, Kuśa. This division of the ancient janapada of Kosala into two parts, a northern and a southern, is well attested in inscriptional and literary sources, as shown by Sarma (1927). It is, however, a division that is only to be found in quite late material. Sarma’s contention that this distinction is known to the Bālakāṇḍa (1927, p. 70) is based on the vulgate passage at 1.13.26, according to which Daśaratha includes Bhānumant, the king of Kosala, among those he invites to his horse sacrifice. The passage is, however, a spurious one and known only to a very few Devanāgarī manuscripts. See crit. ed., 1.*373.
38  Jacobi 1893, pp. 104-105.
39  See Law 1951, pp. 4-6; Mookerji 1951, pp. 22-31, 36-38.
40  Cf. Pargiter 1922, pp. 90-95, who argues that the lists are often questionable and that the Rām version of the genealogy of the solar kings is much less reliable than the versions of the MBh and the purāṇas.
41  Cf. Johnston 1936, pp. xlvii-xliix, and Thomas 1927, pp. 5-15.
42  Cf. Gorresio 1843, pp. cx-cxi, basing his judgment on the traditional Tretā Yuga date and the Rājataraṅgiṇī legend that King Dāmodara, traditionally dated to remote antiquity, heard the Rām recited, puts the composition of the poem in the thirteenth or fourteenth century b.c., a date that he himself regards as conjectural. Weber (1870, pp. 61-62) has shown the unlikelihood of Gorresio’s argument, although, as noted above, his own views of the date of the core of the epic are equally unconvincing.
43  These dates, like so many in pre-Alexandrian Indian history, are ultimately based on the generally accepted date of the death of the Buddha in 486 b.c. If one accepts the Sinhalese tradition of the death of the Buddha in 544 b.c., then all the dates of the late Kosalan and early Magadhan period, and with them the latest date for the composition of the Rām, must be shifted back by some sixty years. For a discussion of the problem of the date of the Buddha’s death, see Raychaudhuri 1923, p. 184.
44  It has been generally accepted that tales of Rāma and the ancient royal house of Ayodhyā must have been current for some time before the composition of the epic itself. This is more than probable. But in the absence of any corroborative evidence, our conception of these tales must be purely speculative.
45  See Guruge 1960, pp. 7-9, and Jacobi 1893, pp. 127-32. Of these perhaps the most elaborately described in the vedic literature is the philosopher-king Janaka of Videha who plays an interesting and important role in the brāhmaṇas and upanishads.
46  Thus Jacobi’s arguments about the mythological connection between the epic heroine Sītā and the like-named vedic goddess, personification of the plowed furrow, are highly questionable. The former is doubtless named and perhaps inspired by the latter, but the superimposition of the epic story on the vedic myth of Indra and the rains is unwarranted by the evidence.
47  For discussions of these lists and the light that they are thought to shed on the Rām, see Lesny 1913; Pargiter 1910, 1913; and Smith 1973.
48  As noted above, Pargiter found the Rām lists to be inferior to the one generally common to the purāṇas. lf the Rām lists are not — as may well be the case — themselves late and abridged versions of a received genealogy, then the purāṇa list may simply be inflated with additional names as a result both of the conflation of the dynastic lists of various families that derived from or claimed to derive from the ancient solar race and of an effort to dignify the family further by extending its lineage. I am inclined toward this last position and am skeptical of Pargiter’s faith in the vaṃśa lists and his cavalier dismissal of the Rām genealogy as the product of brahmans who ‘notoriously lacked the historical sense’; see Pargiter 1922, pp. 93, 119-25.
49  See, however, Guruge 1960, p. 35 for reference to a tradition of Rāma’s having lived in the Dvāpara Yuga.
50  Abrupt if one disregards the Bāla. The subject matter of the first book with its demons and demonesses, its supernatural weapons, superhuman feats, and marvelous legends, is in many ways more like that of Books Three through Six than that of the more sedate and realistic Ayodhyā.
51  Macdonell 1900, pp. 312-13.
52  Gorresio 1843; Lassen 1867, p. 535; Weber 1870, pp. 8-9; Wheeler 1867-81 (vol. 2); Mehta 1941.
53  Gorresio 1847, p. iv. Citing Gioberti, he notes that an epic is nothing other than a system for representing poetically what philosophical systems express theoretically (p. v). The suggestion here is that the epic represents not the external historical data of an area, but rather the collective inner ‘history’ of a culture. A somewhat similar notion is expressed by B. J. Chatterjee 1956, p. 117, who quotes R. C. Dutt as follows: ‘To trace the influence of the Indian epics on the life and civilization of the nation, and on the development of the modern languages, literatures, and religious reforms, is to comprehend the real history of the people during the three thousand years.’ The point is well taken, although here, once again, the use of the word ‘history’ is questionable.
54  See Jacobi 1893, pp. 89-90; Macdonell 1900, p. 313.
55  It has been argued, for example (Bhatt 1960, pp. 446-47), that Indian tribal people may have had monkeys’ faces and tails, and it has even been proposed that the lord of the rākṣasas, the ten-headed Rāvaṇa, suffered from a birth defect that, incidentally, accounted for his hostility (Iyer 1941, pp. 55-56). Even so scientifically oriented a scholar as Mankad, critical editor of the Kiṣkindhākāṇḍa, argues that Rāma’s legendary feat of hurling the carcass of an enormous theriomorphic demon (a man, according to Mankad) a distance of considerably more than a mile with a single blow of his foot is well within the realm of possibility (Mankad 1965, p. 457).
56  See, for example, Pargiter 1894; Guruge 1960, pp. 51-80: Kibe 1914, 1928, 1941a, and 1947; Shah 1976; Sankalia 1973; Ramdas 1925 a and b. 1928, and 1930.
57  Jacobi 1893, p. 103. Cf. Macdonell 1900, p. 313, who remarks, ‘The poet knows nothing about the Deccan beyond the fact that Brāhman hermitages are to be found there. Otherwise it is a region haunted by the monsters and fabulous beings with which an Indian imagination would people an unknown land.’
58  Macdonell 1919, p. 574.
59  For more detailed discussion of the Ceylon issue, see Jacobi 1893, pp. 89-93; Keith 1915, p. 324; and Guruge 1960, pp. 67-69.
60  For a brief survey of these texts see Bulcke 1958, 1959.
61  Cf. Jacobi 1893, pp. 66-67; Weber 1870, p. 210; Winternitz 1927, vol. 1, p. 475; and Bulcke 1958, p. 121. Also cf. von Schroeder 1887, pp. 454ff., and Glasenapp 1929, p. 89, both cited in Bloch 1964, pp. 81-82 note 1. But see also Renou and Filliozat 1947, vol. 1, p. 404, who, like the above authors, see no reason to doubt the attribution of the poem to a single author named Vālmīki, but are not convinced of the poet’s contemporaneity with the protagonist of the epic story. For a lengthy discussion of this question, see Bloch 1964.
62  Even the stories of the two sages’ creation of their respective masterpieces through the inspiration of the god Brahmā, as given at MBh Ādiparvan, App. 1 and Rām 1.2, are closely parallel, and it is clear that one of them has been heavily influenced by, if not actually derived from, the other. Most likely the MBh episode, rejected as an interpolation by Sukthankar, was modeled on the famous story of Vālmīki and the origin of poetry.
63  The famous reference to Vālmīki’s ashram in Book Two has been shown to be an interpolation by the critical edition. For a discussion of the passage, see Bulcke 1958 and 1959.
64  See van Daalen 1980; Bulcke 1958 and 1959; and Bhatt 1960, p. 425.
65  Weber 1870, pp. 22-36; see, for refutation. Jacobi 1893, pp. 94-99, and Macdonell 1900, pp. 309-10.
66  Weber 1870, pp. 1-32.
67  See Lassen 1874, pp. 102-103, and Jacobi 1893, pp. 84-89. Cf. also Winternitz 1927, vol. 1, p. 510, and Macdonell 1900, pp. 308-309. For a comprehensive survey of the question of the Dasaratha Jātaka and the Rām, including summaries of the arguments of the principal writers on the problem, see Bulcke 1950, pp. 84-105.
68  On the date of the Rām, see Keith 1915, pp. 318-28. Keith, incidentally, does not accept the theory that the Rām is post-Buddhist. On the other hand, he denies the validity of Jacobi’s historical argument and does not accept a date for the text earlier than the third century b.c. On the date of the jātakas, see Rhys Davids 1903, p. 103, and Sen 1920, pp. 4-9. The theory was very strongly argued, for example, by Sen in his collection of lectures on the Bengali Rāmāyaṇas (Sen 1920, pp. 4-23). Here he adduces evidence from other jātakas that contain characters or themes in common with the Rāmāyaṇa in support of this hypothesis.
69  See, for example, the short but cogent remarks of the great epigraphist Sircar (1976, pp. 50-53). See also Kane 1966, pp. 45-46. There seems to be little likelihood of, and less evidence for, Keith’s suggestion (1915, p. 323) that the Dasaratha Jātaka and the epic are independently derived from a common source. Bulcke (1952, pp. 102-103) concludes his discussion of the jātaka problem by saying that although a narrative poem on the subject of the Rāma story was current at the time of the compilation of the Pali Tipiṭaka, the composition of the Rāmāyaṇa was not yet complete. This amounts to the same argument as Keith’s. The argument that the Jātaka is older than the epic because it would surely have included the story of the abduction of Sītā had its author known of it, still finds adherents among learned authorities on ancient Indian cultural history. The difficulty is that it is a purely negative argument that presumes a knowledge that we do not have of the motives and methods of the authors and compilers of the jātaka collection. We cannot say with any certainty that the author of the Dasaratha Jātaka would have included any portion of the Rāmāyaṇa that did not serve his needs, even if he knew it intimately. The jātakas are short stories, each of which is designed to illustrate some specific Buddhist virtue through an episode drawn from the career of some creature — animal, human, or supernatural. In the Dasaratha Jātaka, the virtue is steadfastness in the face of emotional trauma, and it is on the basis of his reputation as a paragon of steadfastness that Rāma is chosen to be its exemplar. The legend of the abduction of Sītā is irrelevant to this purpose. If anything, Rāma’s reaction to the loss of his wife is the very antithesis of that expected of the steadfast and self-controlled hero. The authors of the jātakas chose various legends from the Rāmāyaṇa and other stories as it suited their purposes. The fact that the author of the Dasaratha Jātaka restricted himself to one highly modified episode from the Rāma legend is to my mind no proof that he did not know the tale in its entirety.
70  MBh 3.257-76.
71  See Weber 1870, pp. 64-75; Hopkins 1901, pp. 58-84; Kane 1966, pp. 11-58; Winternitz 1927, vol. 1, pp. 500-507; Jacobi 1893, pp. 69-84; Sukthankar 1911, pp. 472-87, and 1939, pp. 406-15; Holtzmann 1846; Holtzmann 1892; Vaidya 1971, pp. xxxi-xxxvi: van Buitenen 1975, pp. 207-14; Shah 1975, pp. 29-53; and Raghavan 1973, pp. 2-31.
72  Weber 1870, pp. 68-71.
73  Sukthankar 1941, pp. 472-87 and 1939, pp. 406-15. Raghavan 1973, pp. 11-25. Sukthankar’s work has been updated with reference to the critical edition of the Rām by Jhala 1968, pp. 295-98. Jhala makes some interesting observations on the question of the recension of Rām known to the authors of the Nalopākhyāna. For a critique of this view and a discussion of the parallel passages of the two epics from the point of view of their grammatical irregularities, see van Daalen 1980, pp. 42-56. Also see Shah 1975, pp. 29-53.
74  For some dissenting views, see Hopkins 1901, pp. 62-64. Lüdwig 1894, pp. 30ff., and, more recently, Vaidya 1971, pp. xxxi-xxxvi, and van Buitenen 1975, pp. 207-14.
75  For a brief summary of some of the less recent opposing views and the rationales behind them, see Bulcke 1950, pp. 53-54. Bulcke himself is of the same opinion as Jacobi and Sukthankar.
76  Vaidya 1971, p. xxxii. xxxvi.
77  van Buitenen 1975, p. 214.
78  Ibid., p. 213.
79  Vaidya 1971, p. xxxii.
80  Ibid., pp. xxxiv-xxxv.
81  Raghavan 1973, pp. 12-13 note 3.
82  Vaidya 1971, p. xxxii.
83  van Buitenen 1975, p. 212. It is odd that in attempting to strengthen his sense of the contrast between Avindhya and Trijaṭā he should describe the latter as ‘young.’ Virtually the only thing we are told about the demoness in the Rāmāyaṇa is that she is vṛddhā, ‘old’ (5.25.4).
84  5.35.12-13 and 6.25.20 (v.l. Aviddha). Vaidya’s error was first pointed out by Raghavan (1973, p. 18). It is particularly unfortunate for van Buitenen’s argument that it is so heavily dependent on this error of Vaidya.
85  See his point 6 about the magic water (p. xxxii) and Raghavan’s discussion (Raghavan 1973, p. 22).
86  van Buitenen 1975, p. 209.
87  MBh 3.266.67, erroneously cited in van Buitenen (1975, p. 210) as 3.266.15.
88  van Buitenen 1975, p. 210.
89  Ibid.
90  For an extensive and comprehensive survey of virtually all of the major literary versions of the Rāma story in both Indian and non-Indian languages, see Bulcke 1950, pp. 1-285. Bulcke (pp. 281-285) also gives a short survey of early western writers on the Rāmāyaṇa. An early survey of the Indian Rāma literature is to be found in Baumgartner 1894, pp. 235-330. See also N. M. Sen 1956b, pp. 95-100. An interesting discussion of epic and puranic versions and variants is provided by Raghavan 1973. For a treatment of some of the less well-known and fragmentary Rāma plays, see Raghavan 1961. On Jain Rāmāyaṇas, see Narasimhachar 1939, pp. 575-94; Kapadia 1952, pp. 115-18; and Kulkarni 1959, pp. 189-204, 284-304. Sen 1920 has an interesting account of the nature and extent of the Bengali Rāmāyaṇas. For the expansion and development of the Rāma story beyond India, see the important and scholarly work of Sweeney on Malaysia (1972), as well as Hooykaas on Java (1958) and Sahai on Laos (1976). Raghu Vira and Yamamoto 1938 have presented the text and translations of two Rāmāyaṇa-derived jātakas found in China. For a general survey and discussion of the non-Indian versions of the Rāma story, see Raghavan 1975.
91  For treatments of important examples of artistic representations of the Rāma story see the select bibliography in Mittal 1969, p. 67. A sizable filmography of Indian cinematic representations of the Rām could be compiled and would be both interesting and useful.
92  Cf. Thompson 1957, vol. 5, p. 300, for references to folktales involving a wicked stepmother. If we can see in the unflattering portrait of Kaikeyī the wicked stepmother of the fairy tale, we may perhaps also perceive in Mantharā, her deformed and malicious maidservant, a variant of the fairy-tale figure of the wicked witch or evil fairy.
93  3.11.29-34. It is not unlikely that this encounter with the sage Agastya is the source for the much more elaborate episode of Viśvāmitra’s gift of divine weapons to Rāma at 1.26-27.
94  See, for example, Thompson 1955-1958, vol. 1, p. 450 for references to the motif of an animal helping in the quest for a vanished wife. For a thoroughgoing analysis of the poem as a ‘Märchen-epic,’ see Gehrts 1977.
95  1.1.1-18. This catalog of Rāma’s virtues is probably not the oldest of its type in the epic. At the beginning of the second book (2.1.15-27) there is a similar list that, if our understanding of the history of the text is correct, is very likely the source of the Bāla passage.
96  The Yuddhakāṇḍa chapter (6.105) in which Brahmā reveals to Rāma his divine nature and praises him as the supreme lord Nārāyaṇa has good manuscript support. It is, however, virtually the only unambiguously Vaiṣṇava passage in Books Two through Six, and like the devotional portions of Books One and Seven, it is almost certainly a relatively late addition. Several authors have remarked that the closing verses of the late first chapter in all but a few Devanāgarī manuscripts have Nārada tell Vālmīki that, at the end of Rāma’s eleven-thousand-year reign, he will go to Brahmā’s world and not Viṣṇu’s (1.1.76), whereas earlier in the same passage the hero is compared to Viṣṇu for his valor (1.1.17). These references would seem to support the theory that the identification of Rāma with Viṣṇu is not known to some strata even of the Bālakāṇḍa. Ruben 1936, p. 63. argued that the issue of the hero’s identification with Viṣṇu has little bearing on the question of the relative age of the Bālakāṇḍa, since we cannot explain the fact that Books Two through Six, which have numerous late interpolations, have still no Vaiṣṇava passages. Therefore, Ruben argues, the central books and the bulk of the Bālakāṇḍa may deliberately withhold allusions to Rāma’s divinity. I find Ruben’s argument to be implausible, not least because it would seem to imply some unspoken but uniformly observed agreement among generations of Rāmāyaṇa scribes and reciters. As to why so few of the later interpolations in Books Two through Six are of a Vaiṣṇava cast, there is little that can be said with any certainty; but there is little evidence to oppose the theory that an explicit Vaiṣṇava reference is a sign of a relatively late stratum of text formation in the Rāmāyaṇa. The question of the textual history of the Bālakāṇḍa will be taken up in greater detail below. For a provocative and dissenting discussion of the divinity of Rāma in Vālmīki, see Pollock’s forthcoming Introduction to Volume 3 of this translation and forthcoming article in JAOS.
97  1.1.77-79.
98  GPP 6.128.105-122, crit. app. *3703. Similar passages at the end of Uttara are, likewise, relegated to the apparatus.
99  6.105. This is the only passage accepted by the critical edition that resembles a classic Vaiṣṇava devotional hymn in both form and content. The hero, hearing this hymn of praise from the mouths of the gods, is puzzled and remarks, ‘As far as I know I am a man, Daśaratha’s son.’ Brahmā then intervenes, explaining to Rāma that he is, in fact, Nārāyaṇa. He identifies Rāma with various of the avatāras of Viṣṇu, including Kṛṣṇa, and this fact alone suggests the late date of the passage. The testimony for this passage is unusually uniform, and the variants are remarkably few and insignificant. In this case, as in many others in the epics, textual homogeneity indicates not antiquity but a late and sectarian passage accepted with little change by all scribes. One is reminded of the situation of the Bhagavad Gītā in the recensions of the Mahābhārata. Passages such as this must make us very wary of claims, such as those put forward by van Daalen 1980, pp. 5-13, that universal testimony is a sign that passages belong to the archetype or even to the corpus of the original poet. See Goldman 1982.
100  At 1.75.17-19. This episode, with its preamble (1.74.14-20), is one of the few truly sectarian passages in the epic. The Bhārgava Rāma, who in the purāṇas and very late strata of the MBh is elevated to the status of an avatāra, shows no evidence of divinity in this passage, even though the episode is a relatively late addition to the Bālakāṇḍa and presupposes knowledge of the MBh tradition of the warrior-sage. The Bāla episode (1.48.12-21) in which Rāma releases Ahalyā from her long curse — an episode that in the hands of Tulsi Das (Rāmcaritmānas 1.210-211) and other poets of the bhakti movement becomes the archetypal demonstration of the lord’s saving grace — is in Vālmīki handled with no reference to the divinity of the hero.
101  There can be little doubt that the HariVaṃ is later than the bulk of the Rām. For a discussion of this point, see Ingalls 1967, pp. 393-94. If, as seems likely, the HariVaṃ was the first great Sanskrit epic poem to concern itself centrally with the life of a divine incarnation, it is at least plausible to see the Vaiṣṇava element in the Rām as derived from the poetic life of Kṛṣṇa. The date of the HariVaṃ is uncertain, but it would appear unlikely that, in the absence of any new evidence, the core of the work, the Viṣṇuparvan, much predates the beginning of the Christian era. If all of these suppositions are correct, we can take the first century a.d. as the rough date of the introduction of the Vaiṣṇava portions of the Bāla and Uttara Kāṇḍas.
102  The HariVaṃ knows Rāma as an incarnation of Viṣṇu (HariVaṃ 65.42-43). The significance of this, however, is unclear. The HariVaṃ reference to Vālmīki as a poet of Vyāsa’s eminence (see Ingalls 1967, p. 393) has been relegated to an appendix by the compilers of the critical edition. See HariVaṃ App. No. 8.30.
103  See Ingalls 1967, for a discussion of the poetry of the HariVaṃ. There is even some evidence that specific incidents in the HariVaṃ account of Kṛṣṇa’s career may have been drawn from the Rāmāyaṇa. Thus the story of how Kṛṣṇa and Balarāma come to Mathurā to see the great bow of Kaṃsa, which Kṛṣṇa then breaks (HariVaṃ 71.27-46), must certainly be an adaptation of the story of Rāma’s breaking of Janaka’s bow that forms the climax of the Bālakāṇḍa.
104  See Bhandarkar 1913, pp. 1-4, 30-48. On the basis of literary, inscriptional, and other evidence it is clear that the cult of Vāsudeva must date from at least the third century b.c., and that the identification of this figure with the cowherd-god Kṛṣṇa cannot be much later than the beginning of the Christian era. On the other hand, despite the evidence of the Bāla, Uttara, and other stray references in the Rām and the inclusion of Rāma Dāśarathi in epic and puranic lists of avatāras, an organized cult of Rāma does not appear to have existed much before the eleventh century a.d.
105  Additional evidence of this may be found in the Uttarakāṇḍa’s knowledge of Mathurā as an important site, and its appropriation on behalf of Rāma’s brother, Śatrughna, of the glory of having founded that important city.
106  See 1.2.32-34 where, through the grace of Brahmā, Vālmīki is granted knowledge of all the actions and thoughts of the epic characters — the public and private, the known and the unknown.
107  Thus, for example, Wheeler felt he could discern four distinct historical stages of religion and civilization in the epic, whereas Lassen and Weber saw the poem as a sort of chronicle of Aryan expansion into peninsular India. These interpretations and others offered by nineteenth-century scholarship are no longer generally accepted, but the idea that the Rām developed around a historical core is still current.
108  Cf. Jacobi’s amusing deflation of Wheeler’s interpretation of the epic as a historical allegory: ‘Imagine! Vālmīki, the greatest poet of the pre-classical age composed an allegory that no one understood until a European nineteen centuries later came upon the obscure secret!’ (1893, p. 90). The point is a good one and ought to be brought to bear on all interpretations that seek to ‘discover’ a concealed meaning in the epic. It is, ironically, equally damaging to Jacobi’s own cherished interpretation.
109  Jacobi 1893, pp. 126-39. It is probable that the figure of Sītā is somehow derived from the vedic divinity of the plowed field, but we are in agreement with Winternitz 1927, pp. 515-16 that Weber 1891, p. 818 was right in his feeling that a wide gulf separates the vedic from the epic legends. Moreover, there is evidence to indicate that the Rāma story does not proceed originally from the vedic legends of the asuradeva conflict. See Goldman and Masson 1969, pp. 95-100.
110  Narain 1957, pp. 111-15. Despite Narain’s statement, Rāma is subject in certain circumstances to almost total loss of self-control, a loss accompanied by unrestrained anger and sorrow; see Goldman 1984 pp. 166-67. This will be discussed further below.
111  This difference has long been a subject of discussion among scholars. Gorresio summed it up in a sort of aphorism: ‘accordo dell’ uomo col creato in Grecia, lotta dell’ uomo col creato nell’ India’ [Harmony of man with nature in Greece, struggle of man with nature in India] (Gorresio 1847, p. vii).
112  Bhīṣma, like Rāma, is deprived of his rightful succession to the throne through his father’s infatuation for a beautiful and ambitious woman. His sacrifice is, however, in many ways more complete than that of Rāma: he relinquishes his claim to the throne voluntarily. His father does not have to ask or order him to do so. Moreover, his loss of royal power is permanent, unlike Rāma’s, which has a fixed term. Finally, Bhīṣma also renounces both the pleasures of sex and the hope of progeny in an act of self-denial so dreadful that the gods, seers, and heavenly nymphs cry out ‘bhīṣmo ’yam,’ ‘He’s awesome!’ thus giving the hero his best known epithet; see MBh 1.94.86-90. For a discussion of some of the psychological implications of Bhīṣma’s filial devotion and that of similar MBh figures such as Pūru and Rāma Jāmadagnya, see Goldman 1978, pp. 338-47.
113  BhagGī 1.25-47, 2.4-8.
114  2.91.6-8. In this sarga, many southern manuscripts and the vulgate have a passage of some twenty lines in which Rāma, in a virtual paraphrase of the comparable Gītā passage, expresses his horror at the thought of fighting his kinsmen (Ayodhyākāṇḍa, crit. app. 2112*.1-12; GPP 2.97.2-8). Like Arjuna, Rāma argues that there is no point in killing his kinsmen for the kingdom, since it is only for their sake that he desires wealth, power, and pleasure. Like Arjuna, he too claims that, although he could easily gain the earth, he would not wish even the lordship of the gods (śakratvam), if it should come to him through foul means. He too likens things acquired through the killing of his friends and kinsmen to the consumption of tainted food (bhakṣyān viṣakṛtān iva). Although several factors make it difficult to determine the relative priority of the two passages, it seems probable that one is a paraphrase of the other. It is especially interesting that the tradition sees the two confrontations as similar, even though their resolutions are radically different.
115  In the Rām, Bharata, although he is Rāma’s junior and must therefore defer to his elder brother, is viewed by the latter as the representative of Daśaratha, since he is now the appointed heir.
116  2.16.27-61.
117  2.98-104.
118  6.113.12-16. Rāma’s unwavering devaluation of the throne relative to what he perceives as his duty to his father and such abstractions of paternal authority as dharma and truth is closely paralleled by his recurrent undervaluation and repeated rejection of Sītā in favor of his male kinsmen and such abstractions as duty and reputation. For a detailed discussion of this see Goldman 1980.
119  For a detailed study of this phenomenon in the Rām with special reference to the epic’s composite hero, see Goldman 1980.
120  See Goldman 1978; Devereux 1951; and Carstairs 1961, pp. 159-60.
121  MBh 1.94.86-88. It is interesting to note that in a Jain account of this episode, found in the Pāṇḍava Purāṇa of Vāḍicandrasūri (1.105-106), Bhīṣma is said to have lent substance to his famous vow by actually castrating himself. I am grateful to Professor P. S. Jaini for pointing out this reference to me.
122  MBh 1.79.27-29.
123  MBh 3.116.13-14. See Goldman 1978, pp. 342-44.
124  See 1.76 where the blossoming of this love is charmingly described; and 3.59-61, 4.1. etc. for his grief.
125  For a detailed discussion of this, see Goldman 1980, pp. 160-63.
126  For a dramatic illustration of this attitude in Indian society, see Gandhi 1960, pp. 43-46.
127  1.25.2-4.
128  1.25.11-12. In the vulgate and several southern manuscripts, Rāma’s reluctance is so great that the sage has to order him once more, rather sharply, to get on with it. See App. I, No. 5.15-20.
129  MBh 3.116.18. This is one of the several ways in which the MBh accounts for Rāma Jāmadagnya’s reputation for being a great master of the science of arms. For a discussion of this and other legends bearing upon the Bhārgava Rāma’s acquisition of his military skills, see Goldman 1977, pp. 99-112 and 1982a.
130  1.26-27. Bhīṣma and Pūru, the other major epic exemplars of filial subservience, are also given boons by their fathers in compensation for their renunciation of sexuality. Śantanu grants Bhīṣma the power to choose his time of death (MBh 1.94.94), whereas Yayāti makes Pūru, his youngest son, his heir and successor (MBh 1.79.30).
131  1.66.15-23.
132  1.73.
133  A detailed psychological analysis of the Rāmāyaṇa, although desirable, is beyond the scope of this introduction. For psychological discussions of particular characters and episodes in the epic, see Goldman 1978 and 1980; and Masson 1975 and 1980, pp. 80-109.
134  Holtzmann 1841, pp. 36-38.
135  See Lesny 1913, p. 497; Macdonell 1900, pp. 304-307; Winternitz 1927, vol. 1, pp. 495-96; Bulcke 1952/1953, pp. 327-31; and Jacobi 1893, pp. 50ff.
136  Lesny 1913, p. 497.
137  Tiwari 1952-53, pp. 9-17.
138  Jacobi 1893, pp. 55-59.
139  Jacobi 1893, p. 59.
140  Actually sargas 1, 2, and 4: see Bulcke 1952/1953, p. 329. The exclusion of sarga 3, the second conspectus of the epic poet, is evidently based upon the fact that it knows the events of Bāla: the journey with Viśvāmitra, the marriage, the breaking of the bow, and the confrontation with Bhārgava Rāma (1.3.4-5). This exclusion is problematic, however, in the light of Bulcke’s acceptance of these events as belonging to the oldest portion of Bāla.
141  Tiwari’s arguments as to the late date of the Bālakāṇḍa (Tiwari 1952/1953, p. 15) are, in fact, relevant only to the later portions of the text. His remarks about the Śakas, Yavanas, and other barbarians produced by Viśvāmitra’s cow — remarks similar to those of earlier writers on the date of the Rām — apply only to the passage in which the reference occurs, a fact first pointed out by Jacobi (Jacobi 1893, pp. 94-95).
142  Ghosh 1967, pp. xlix-lii.
143  Holtzmann 1841, p. 39.
144  Jacobi 1893, p. 52 note 1; Winternitz 1927, vol. 1, p. 496; Bulcke 1952/1953, p. 328.
145  See crit. app. for 2.110. The passage appears in the vulgate and is accepted by the critical edition. It is, in fact, very widely distributed and is omitted only in a few B and D manuscripts. Jacobi was aware of the passage and mentioned it specifically in a note (Jacobi 1893, p. 53 note). Yet so firm was he in his belief in the lateness of the first book that he regards the episode as interpolated.
146  3.36.3-9.
147  This is precisely the material that Bulcke regards as constituting the oldest stratum of the Bālakāṇḍa. Regardless of the direction of the borrowing, it is clear that all of this material is early, and it is probable that these portions of the Bālakāṇḍa are contemporaneous with or not very much later than the ‘genuine’ books. The detailed and highly colored account of the confrontation between the sage and the king over the issue of sending Rāma to fight Mārīca and Subāhu is found at 1.17.22-1.21.6. Although Mārīca’s account in Araṇya is concise and omits many details, it covers the whole period described at Bāla 17-29. Similarly, the tale Sītā tells Anasūyā at the end of Ayodhyā summarizes the events detailed in Bāla 65-76.
148  Some of the passages and episodes within these sargas have been introduced at later periods. The most significant of these, such as the story of the encounter of the two Rāmas, will be discussed individually below.
149  ‘Diese Episode scheint eingeschoben zu sein, alt das erste Buch noch keinen festen Bestandteil der Rāmāyaṇa bildete.’ Jacobi 1893, p. 53 note 1.
150  See Goldman 1980, pp. 167-70.
151  Cf. Carstairs 1961, pp. 69-70. The culturally enjoined suppression of a younger brother’s reference to his sexual life, even mention of his wife’s name, is amusingly illustrated in the context of the Rāma story by the great Sanskrit playwright Bhavabhūti in his famous Uttararāmacarita. After verse 18 in act 1, Sītā who, with Rāma, is being shown a group wedding portrait of the brothers and their brides asks Lakṣmaṇa to identify Ūrmilā. This embarrasses the prince so much that he quickly skips to another picture to cover his confusion and avoid having to mention his wife’s name or openly acknowledge her existence in the presence of his elder brother and sister-in-law. See Belvalkar 1915, p. 20. For a discussion of this incident and this phenomenon in the traditional literature, see Goldman 1978, pp. 327-29 and note 21.
152  Jacobi 1893, p. 53; Winternitz 1927, vol. 1, p. 487 note 2; Bulcke 1952/1953, p. 328.
153  For a detailed discussion of this incident and its implications both for the development of the epic story and for our understanding of the relationship between Books Two through Six and the Uttarakāṇḍa, see Goldman and Masson 1969, pp. 95-100.
154  See Bulcke 1952/1953, p. 330.
155  Ibid., p. 328.
156  It is interesting that the Bālakāṇḍa, although it has often been regarded as a late addition to the poem, does not, according to the best reconstruction of the manuscript evidence, end with a phalaśruti.
157  1.76.18; the reference to Śrī and the identification of Viṣṇu, rather than Indra, as the ruler of the gods is a sure sign of sectarian influence.
158  Bulcke 1952/1953, p. 328.
159  These events — the journey with Viśvāmitra, the breaking of the bow, the marriage, and the encounter with the other Rāma — are mostly those that constitute the older strata of the first book. In addition, the chapter contains a couple of lines mentioning some of the events of the Uttarakāṇḍa (1.3.28).
160  As discussed above, it is clear that the Rāmopākhyāna is later than and derived from the Vālmīki Rāmāyaṇa. That its authors knew the Uttarakāṇḍa is demonstrated by their familiarity with the legend and antecedents of Rāvaṇa and by the fact that they quote the latter book. See Jacobi 1893, pp. 74-75, and Raghavan 1973, pp. 11-12.
161  This episode, in which Rāma and his sons first meet and the king is overwhelmed by emotion at hearing the moving ballad of his tragic career, has fascinated the transmitters of the Rāmāyaṇa. They return to it at the end of the Uttarakāṇḍa where an elaborate and highly dramatic version culminating in Rāma’s recognition of his sons and the return and final disappearance of Sītā is made to occupy five sargas (7.84-89). Most of the dramatic events in this version, including the encounter of Rāma and Vālmīki and the return of Sītā, are unknown to the Bālakāṇḍa passage.
162  1.2.8-13. For a brief discussion of the underlying psychological significance of this episode and its relation to similar legends in the Sanskrit epics, see Goldman 1978, p. 392.
163  The episode has been regarded by many literary critics as foreshadowing the tragic events of the central epic narrative, while the Vaishnava scholiasts have interpreted the verse, ingeniously if not wholly convincingly, as encoding allusions to many of the episodes in the epic and their underlying theological significance. For a discussion of the interpretations of the episode at the hands of Indian writers on poetics, see Masson 1969. For some examples of the treatment of the verse on the part of the Rāmāyaṇa commentators, see the notes to 1.2.14 below.
164  The position that the upodghāta appears to hold regarding the priority of the karuṇarasa is strikingly articulated (in a non-technical context) by the great dramatist Bhavabhūti who in the Uttararāmacarita (3.48) has one of his characters remark à propos the pathos of the Rāma story, eko rasaḥ karuṇa eva nimittabhedād, bhinnaḥ pṛthak pṛthag iva āśrayate vivartān, āvartabudbudataraṅgamayān vikārān, ambho yathā salilam eva hi tat samastam. ‘There is really only one aesthetic mood, that of pity. It is only through its different modifications and manifestations that it appears to have different forms. It’s just like water that, although it may take on the forms of whirlpools, bubbles or waves, is still, in the end, just water.’
165  See 1.4.8 where, in a reading regarded as uncertain by the editor, the poem is said to possess the traditional rasas, of which seven are named in the S manuscripts and nine in the N manuscripts. For a discussion of the commentators’ views on the Rāmāyaṇa’s rasas and the text-historical implications of the difference between the N and S recensions, see Bālakāṇḍa, critical notes, at 1.4.8.
166  It is evident that the names of Rāma’s sons are derived from the term for bard, not vice versa, as tradition holds. See Bālakāṇḍa, critical notes, at 1.4.3, for further details.
167  Bhatt 1960, p. 437.
168  See Holtzmann 1841, pp. 36-37, and Bulcke 1952/1953, pp. 330-31.
169  Bulcke 1952/1953, p. 331. Bhatt, who has argued for the propriety of performing the Aśvamedha as a device for producing an heir, appears willing to accept the traditional view that the purpose of the former sacrifice was merely to remove the sins that were obstructing the birth of a royal heir (Bhatt 1960, p. 446).
170  Aside from the detailed description of Daśaratha’s sacrifice, the poem mentions those of Sagara (1.38-40), Ambarīṣa (1.60), Saudāsa (7.57.18), and, of course, Rāma (7.82-83). By way of preamble to the performance of Rāma’s sacrifice, the Uttarakāṇḍa briefly mentions two legendary performances of the rite for the purpose of freeing someone from an impure undesired state. At 7.77.8-10 the gods perform the rite on behalf of Indra in order to free him from the taint of brahmahatyā. At 7.81.12-20 the sage Marutta and a company of brahmans perform the sacrifice on behalf of King Ila who has been transformed into a woman by Śiva. The rite does not operate directly, as in the case of Indra, but rather serves to ingratiate the officiants with Śiva who, as a favor to them, agrees to restore the king’s masculinity. At 7.23.7-8 the Aśvamedha is mentioned as one of the seven sacrifices completed by Rāvaṇa’s son, Meghanāda.
171  For a discussion of this fascinating character, the probable prototype of the European unicorn, whose legend has permeated the literatures of Hinduism, Buddhism, and of areas far beyond the borders of India, see Lüders 1897 and 1901, and the forthcoming study of the legend by Masson and Goldman.
172  The northern recensions of the epic, and much of later Indian literature, regard Śāntā as actually a daughter of Daśaratha given in adoption to his friend and ally Lomapāda (Romapāda). On the basis of careful textual analysis, Asoke Chatterjee has shown this tradition to be a later invention of the northern redactors, owing its existence to their confusion of the Aṅga monarch, Daśaratha or Lomapāda, mentioned in several puranic genealogies, with the Kosalan Daśaratha. See A. Chatterjee 1954 and 1957.
173  1.8.22. In a passage substituted in a number of manuscripts of various recensions it is specified that the sage will accomplish this after haying offered oblation in the sacrificial fires. See crit. app. *298.
174  1.9. The story, like all but a few of the puranic narratives inserted into the Bālakāṇḍa, is brief to the point of opacity, and largely lacks the charm and power that characterize many other versions. This is especially noticeable in the Bālakāṇḍa versions of the stories of the churning of the ocean (1.44) and the birth of Kumāra (1.36), which are terse, perfunctory, and obscure, comparing unfavorably with the Mahābhārata versions.
175  Although I have as yet come across no assertion on the part of any scholar that the sacrificial role of Ṛśyaśṛṅga is, in fact, a disguised form of niyojana, I think that the evidence of the text leads us to serious consideration of such an underlying element. Other scholars may have advanced this thesis, for we find that one V. Panoly takes issue with them in the strongest and most colorful language. See Panoly 1961, pp. 17-19.
176  The story is popular in the epic and puranic literature, and Sagara is one of the cherished dynasts of the Rāghava House. Its somewhat separate origin is suggested by its ending in a widely distributed phalaśruti (1.43.20).
177  Cf. Sukthankar 1937, pp. 1-76, and Goldman 1977.
178  It is probably related to other epic accounts of the confrontations between heroes and rākṣasa women such as those between Bhīma and Hiḍimbā in the Mahābhārata (1.139-144) and between Rāma and Śūrpaṇakhā in our text (3.16-17). Those episodes are, however, typologically different from the slaying of Tāṭakā.
179  In fact, the conferral of the magic weapons upon Rāma by the Kauśika sage, although it is a version of a motif common to the epics, is probably a transposition and expansion of a similar conferral of divine weapons upon the prince by the seer Agāstya at 3.11.29-34. It is these latter weapons, most particularly the great Vaishnava bow, that Rāma appears to use in his battles with the rākṣasas in much of the remainder of the poem.
180  See Goldman 1977.
181  It should be noted here that if we do not presuppose the Mahābhārata or a still later puranic account of Jāmadagnya and his extermination of the kshatriyas, the figure introduced in the Bālakāṇḍa would be utterly mysterious, for the Rāmāyaṇa nowhere recounts his proper legend, trusting to its audience’s knowledge of the story from the greater epic and puranic tradition. The encounter of the two Rāmas, which appears to be a Rāmāyaṇa application of a figure drawn from the Mahābhārata, in turn appears in a late stratum of the latter text where, in a Vaishnava passage known to the vulgate and a few northern manuscripts, a version of it appears. See MBh 3, crit. app. I, No. 14.20-84 (Citrashala ed. 3.99.40-71). Sukthankar 1937, pp. 20-21, has clearly shown the Mahābhārata passage to be the work of a late and ‘ignorant’ northern interpolator. In this way the Mahābhārata has borrowed a Rāmāyaṇa episode in which one of its own heroes is humiliated by the hero of the older epic.
182  See 1.1ff. on Nārada; 12.40 on the new metrical form; 1.3 on Brahma; 1.2.40, 7.84.5 on the whole Rāmāyaṇa poem. In the northern variant of 1.3.1 Vālmīki is said to discover the other events in Rāma’s life — to supplement the account of Nārada — ‘from the world,’ though we need not, with Agrawala, view the poet as an early folklorist ‘who collected the several versions of the legend from what was current as folklore’ (Agrawala 1962, p. 578). See 1.4.5 on the performers and 1.4.12, 7.84.9,16 on memorization. The singers are to recite twenty chapters per day (7.84.9) or about some 1,200 sixteen-syllable lines; the performance is said to take many days. (Excluding Book Seven, there are 500 chapters, according to 7.85.20, and thus, the performance would extend over approximately a month.) For the various types of duo oral recitation, cf. Chadwick 1932, p. 574; Lord 1960, p. 125 and note. It is not clear from the Rāmāyaṇa itself how we are to picture the recitation.
183  See P. A. Grintser 1974 (English summary, pp. 416ff.) on the genesis of the oral poem. That the transmission of the Rāmāyaṇa cannot be reconciled with the image of a wholly memorized original is not a serious contradiction. Although exact reproduction is an ideal that performers of oral poetry envision, in reality a certain amount of personal modification occurs in any given performance.
184  See Shah 1975, pp. 50-51 and references. The Buddhist poet Aśvaghoṣa (fl. a.d. 50) might have known a written Rāmāyaṇa, for the close agreement in verbal and narrative detail between his Buddhacarita and the Rāmāyaṇa argues for the kind of ‘consultability’ that only a written text allows. Cf. also Gawrónski 1914-1915, pp. 280-81.
185  This was already apparent to the editors of the incomplete editio princeps, Carey and Marshman, in 1806, see Gorresio 1843, p. xx. Sometimes it appears as if we must speak rather of three recensions, distinguishing a NW (Kashmir and west) from a NE (Nepal and east), cf. especially Shastri 1940, pp. 58 and 75. But there is so much contamination among N manuscripts that it is difficult to decide for certain. Ruben, additionally, wished to divide the southern recension into two, one represented by the commentators Vaṃśīdhara Śivasahāya (Rāmāyaṇaśiromaṇi), Maheśvaratīrtha (Tattvadīpikā), and Govindarāja (Bhūṣaṇa), and the other by Kataka Mādhava Yogīndra (Kataka) and Nageśa Bhaṭṭa (Tilaka). The editors of the critical edition are not unanimous in their understanding of these problems, and, in fact, the whole notion of ‘recensions’ with regard to the Indian epics is somewhat indeterminate (see Johnston 1933, pp. 182-83).
186  Ruben 1936, pp. ix, xi, and Bhatt 1960, p. xxxiv, do not adequately appreciate this signal difference. Bhatt’s editorial practice, in fact, contradicts his theoretical statements; contrast his remarks in volume 1 with Vaidya’s statement (Vaidya 1971, p. xxx).
187  The most recent major work on the subject, van Daalen 1980, takes issue with the theory but not, in our opinion, in an adequate fashion. For a detailed discussion, see Goldman 1982.
188  See Jacobi 1893, p. 9; repeated with approval by Bhatt 1960, p. xv, Burrow 1959, p. 78, Renou 1963, p. 283. and, most importantly, Bulcke 1955, p. 92, and 1960, p. 38. The editors of the critical edition, when they do not simply parrot this theory (as Mankad 1965, p. xxiv; Jhala 1966, p. xxiii; Vaidya 1971, p. xxx), have only trivial examples to offer in support of it (Bhatt 1960, p. xiii, Suppl. Intro.).
189  It is not clear how much reliance is to be placed on the so-called linguistic archaisms preserved in the southern recensions. as indicating an earlier date. A very large percentage of the archaisms that have been examined (Böhtlingk 1887, 1889; Michelson 1904; Roussel 1910; Satya Vrat 1964, pp. 173ff.; N. M. Sen, all items in bibliography; van Daalen 1980, especially pp. 72-117) are contained only in the first or seventh book, of which substantial portions are later additions, and, more remarkably, in passages that the critical edition excludes from the constituted text as more recent interpolations. The northern recension, moreover, frequently preserves archaisms that appear in the southern recension, and quite often ‘archaizes’ where the southern recension does not. See also van Daalen 1980, pp. 27-32.
190  In essence, the critical edition of the Rāmāyaṇa has collected the fullest record anywhere of the stages of growth and development of a great oral epic tradition.
191  In the south the religious significance imputed to the text lent it an almost scriptural status, insulating it to a greater extent from alteration. The commentators, attracted to the text for this same reason, would have been particularly instrumental in preserving the poem in its archaic state. Just the opposite is true of the Mahābhārata. There the southern recension revises rather freely, whereas the northern recension (the NW version in particular) preserves more authentically the tradition of the archetype.
192  See Pollock 1981. For examples, see the notes on 1.8.9, 2.17.7, 21.11, 47.26, 94.49 (lexical glosses); 2.24.7, 51.42 (syntactical glosses).
193  Like any commentary, the northern recension must be used with discretion as a gloss. The glossers were not invariably right, though as participants in a continuous and ancient tradition of recitation they can claim weightier authority than our medieval commentators. The general editor of the critical edition seems to have been aware of this feature, but the examples he provides are trivial (e.g. dhanuḥ replacing śarāsanam, Bhatt 1960, p. xxxiii).
194  Hopkins 1926, pp. 206, 219. As early as 1870, Weber argued that ‘there are as many Rāmāyaṇas as there are manuscripts’ (p. 21; cited in Ruben 1936, p. x; cf. Burrow 1959, p. 78). Recent work on oral poetry might appear to support Hopkins’ impression. For example, Lord concludes that ‘it is impossible to retrace the work of generations of singers to that moment where some singer first sang a particular song. … There was an original, of course, but we must be content with the works that we have and not endeavor to “correct” or “perfect” them in accordance with a purely arbitrary guess at what the original might have been’ (Lord 1960, p. 100). It seems to us, however, that the type and quality of manuscript congruence in important sections of the Rāmāyaṇa suggest that the sort of transmission here may be of a different order from what we see, for example, in Slavic or French literary history. Very possibly the mnemonic tradition of vedic transmission exerted some influence upon the performers of secular heroic poetry. The text may preserve an historical reminiscence when it states that the first performers of the Rāmāyaṇa were deeply grounded in the vedas (1.4.5). In any case, broad arguments from the nature of oral poetry in general should not be applied uncritically to the Indian evidence, where a reconstruction may not be ‘purely arbitrary.’
195  Jacobi 1893, p. 11 (this position is somewhat contradicted by what we find on pp. x, 5). Lévi also speaks of a written archetype. ‘Our Rāmāyaṇa, composed at a still undetermined period, derives in its multiple recensions from an edition published sometime around [the commencement of] the Christian era’ (Lévi 1918, p. 150). Ruben’s Studien are predicated on the existence of an archetype: Agrawal, too, assumes one, without explanation (1963, p. 577).
196  See Bhatt 1960, p. xxx.
197  The agreements among the recensions in the Sundara passage noticed by Jacobi (1893, pp. 17ff.), for example, can be as conveniently explained by postulating an oral transmission, which saves us from the real contradictions involved in the archetype theory. We may then interpret the data in Jacobi’s passage by the special dynamic of an oral tradition, which in one place gives rein to variation, in another inhibits it, which permits deviation in wording to some extent but demands conservation of the significant structures of significant passages.
198  One need only glance at Ruben’s Textproben to confirm this (1936, pp. 84-222).
199  Such is also more or less the opinion of Bulcke 1955, p. 66. and 1960, pp. 37-38. The versions continued to grow, perhaps orally, and to interact throughout the period of written transmission, both within and, to a lesser extent, across recensional boundaries. A number of passages that on the grounds of higher criticism must be considered quite late additions to the text are sometimes, especially when they have a powerful sectarian thrust, unusually well represented in all the recensions, with a minimum of variation. A good example is Brahmā’s hymn in praise of Rāma as Viṣṇu at 6.105.
200  Bhatt 1960, p. xiv and particularly Vaidya 1971, p. xxx understood this. Contrast however Mankad (1965, p. xxiv) and Jhala (1966, p. xxvi).
201  This principle was clearly enunciated by Sukthankar with respect to the Mahābhārata. ‘The peculiar conditions of the transmission of the epic force upon us an eclectic but cautious utilization of all manuscript classes. … Each variant has to be judged on its own merits.’ But where the tradition is irreducibly divided, a choice on the basis of otherwise generally best versions must be followed (Sukthankar 1944, pp. 243, 249). It is even more compelling in the case of the Rāmāyaṇa than the Mahābhārata, for which a written archetype must have existed.
202  Granted the circularity involved in applying standards of authenticity to correct a text from which those sane standards are derived, nevertheless, as Kenney puts it, ‘critical argument is by its very nature circular,’ and it is not ‘necessarily vicious, providing, as Lachmann said, that the circle is trodden with care and discretion’ (Kenney 1974, pp. 126, 135). Ruben adduces other grounds for the relative antiquity and sincerity of the southern recension, such as the agreement in parallel passages of the Mahābhārata with the southern recension against the northern recension (Ruben 1936, pp. 47, 54. etc.; but n.b. his caution, p. 55).
203  Sylvain Lévi, in a fascinating article on the geographical data of the Rāmāyaṇa, determined that a text of the Kiṣkindhākāṇḍa (39-43) was used by a Buddhist work, the Saddharmasmṛtyupasthānasūtra, which was translated into Chinese in a.d. 593; and since, he says, the translator only worked with materials of great authority, the Sūtra must be far older than that (Lévi 1918, p. 15; Lin Li-Kouang, however, has shown that the Sūtra is a composite work and that chapter VII, the one in question, is the latest, see 1949, pp. 111-12). He concluded that, although the southern recension alone does preserve some readings and details that are in harmony with the Sūtra, the northwestern recension is in fact closest to it (p. 135), and he considers this fact ‘the most ancient datum with regard to the recensions and a datum decisively in favor of the western recension’ (p. 14). It is only reasonable, however, that the Sūtra should employ the version current in the area in which it itself was composed. This would apply also to the arguments adduced in favor of the northwestern recension by Słuszkiewicz 1938, pp. 266-73. Furthermore, the evidence can only serve to confirm the fact that the split in the transmission of the Rāmāyaṇa occurred at a relatively early date; it cannot prove which branch of the tradition was more conservative.
204  Because the northern recension transposed or vulgarized in one place it does not mean that it did so in another, nor is the southern recension’s conservation is absolute. Furthermore, the problems inherent in transcribing an oral poem would affect the southern recension no less than the northern.
205  See the note on 2.63.4; unfortunately the editor of the Ayodhyākāṇḍa mistook this variant for a corruption. One serious error of the critical edition is its failure fully to exploit the northern recension and to realize that a reading that is not utterly impossible (Bhatt 1960, p. xxxiv) does not, therefore, become probable.
206  Edgerton 1944, p. xxxiv. Rāmāyaṇa commentators continue to transmit passages even when they themselves consider them interpolated.
207  Johnston 1933, p. 183.
208  West 1973, p. 9.
209  Some influential authors, however, such as Abhinavagupta, refer to the poem as itihāsa, or traditional history, whereas others, such as Madhusūdana Sarasvatī, regard it as śāstra.
210  The few actual attempts to capture the flavor of the epic in English verse are particularly unpalatable to the modern reader. Thus, Griffith’s noble effort (Griffith 1870-1874) in rendering the entire poem in rhymed octosyllabic couplets does not, as he himself warned, ‘bear reading through’ (p. viii).
211  I am referring here to attempts at actual translation, but similar problems have beset the authors of the numerous retellings and ‘transcreations’ (to use P. Lal’s infelicitous term) of the poem and its derivative works that have appeared in recent years. These authors, such as Buck (1976), Menen (1954), and Narayan (1972) have had the advantage of being able to select and modify the episodes that appeal to them and to render them in their own words. Their efforts have, therefore, produced works which, if occasionally entertaining, bear no resemblance whatever to the style or the feeling of their originals. We believe, however, that the task of representing Vālmīki in a readable guise to the English reader is not an impossible one and have striven to avoid what Griffith feared, that his version might actually enhance the tediousness of some of the original. (See Griffith 1870-1874, p. vii.)
212  Arnold 1905, pp. 41ff.
213  Auerbach 1953, p. 6.
214  Cf. Griffith 1870-1874, p. viii. In parts of the Bālakāṇḍa, for example, nearly one-quarter of the verses contain formulaic elements. For an analysis of the first book from the standpoint of the study of oral poetry and formulaic analysis, see Nabaneeta Sen 1966, pp. 397ff.
215  “Iliad” 5.259-268; Fitzgerald 1974, pp. 118-19.
216  6.55.120-125.
217  Or, according to Cg, inimical (to foes); neither rendering of ariṣṭa helps us form a visual impression of the arrow.
218  Macdonell 1919, p. 574.
219  For a discussion of Hanumān as an imaginary companion, see Masson 1981.
220  For a discussion of some of the uses of fantasy and hyperbole in the Rāmāyaṇa, see Masson 1980, pp. 80-109.
221  Auerbach 1953, p. 6.
222  1.1.9-10.
223  For a sympathetic and learned discussion of the Sanskrit poets’ preference in general for idealization and impersonality as opposed to realism and kinesthesia see Ingalls 1965, pp. 1-29.
224  Ibid., pp. 49-53.
225  Griffith 1870-1874, p. viii.
226  In 1974-1975 I had the privilege and pleasure of reading the Bālakāṇḍa with Pandit Śrīnivāsa Śāstri of the Deccan College in Poona. He had read the entire epic through several times and knew it intimately. Day after day he would clearly and brilliantly elucidate for me words, phrases, and passages that had seemed utterly opaque. Yet, not infrequently, even he would find a passage lucid at first glance, remarking without hesitation, ahaṃ vacmi, ‘I’ll explain it,’ only to stare at it, examine four or five Sanskrit commentaries, and conclude by saying, īśvaro veda, ‘God knows!’
227  Thus, for all the importance that the Vaishnava poet attaches to the exact apportionment of the divine pāyasa, or porridge, among the three wives of Daśaratha, it is all but impossible to make out exactly what fraction of the stuff goes to each woman. This imprecision, in turn, leads to a traditional debate of some theological significance. For a discussion of this point, see the notes to the translation below at 1.17.
228  The effect is somewhat monotonous, especially since the reciter rarely varies the cadence or intonation except where the meter changes. The Rāmāyaṇa of Vālmīki is recited today mostly for religious purposes, its place in popular entertainment having been taken by its vernacular derivatives.
229  Through the use of compounds it is possible, as Ingalls has noted, to generate hundreds of synonyms for some words. See Ingalls 1965, pp. 6-8.
230  Cf. Gorresio 1847 and Griffith 1870-1874, pp. 527-565.
231  See Peterson 1879 and Bhandare 1920.
232  See Raghavan 1950 for the date of Uḍāli. See also Bhatt 1964. Bhatt’s Appendix III in Shah 1975 at pp. 655-64 provides a list of commentaries, including Aufrecht’s list (Catalogus Catalogorum, pp. 523-24) as well as references to and extracts from a number of unpublished commentaries. The question of the chronology of the commentators and of the ways in which they were influenced by their predecessors is a complicated and interesting one, with important implications for our understanding of the history of Vaishnavism, and it will be taken up in the forthcoming volume of Rāmāyaṇa studies.
233  Unavailable for the Bālakāṇḍa.
234  See, for example, the notes to 1.2.14 and 1.42.17.