Common section


The Poetry of the Mahabharata


Dites, qu’avez vous vu?

Tell us, what have you seen?


ROBERT LOWELL, Imitations (1961)1

ONE OF THE pleasures of Carole Satyamurti’s retelling of the Mahabharata is that it pursues a variety of goals and accomplishes them with seemingly effortless skill. It is a contemporary poem in English that seeks to stand aesthetically on its own, to be valued for its craft, thematic significance, and imaginative scope and depth. At the same time, however, it is overwhelmingly concerned with representing another poem as transparently as possible, even though the latter is remote in time and place as well as language and culture, and embodies a very different set of shaping principles. On a different plane, Satyamurti’s poem sifts through the numerous interwoven stories of the original in order to fashion a cogent storyline, and creates a narrative momentum that will hold our interest continuously. But it also pulls us in other directions, as it flexibly accommodates a mass of material from Sanskrit, and absorbs an abundance of unfamiliar terms, concepts, and qualities. Even as it maintains balance and restraint, the book takes some remarkable risks: adapting iambic pentameter and English blank verse to its practical tasks, it achieves a monumental size of almost 27,000 lines and 200,000 words. More than two and a half times the length of Milton’s Paradise Lost and over three times the length of Wordsworth’s 1850 Prelude, it emerges modestly as the longest successful experiment in English narrative poetry in modern times.

Astonishing as Satyamurti’s technical accomplishments are, however, it is her desire to re-narrate an ancient Indian poem that defines her primary purpose in these pages. But what kind of work is the Mahabharata itself, and what are its attributes that a modern version ought to represent? How does this English poem actually relate to its largely inaccessible Sanskrit source? And what sort of world do the original Mahabharata and this innovative retelling open up for us, as cosmopolitan readers here and now? Wendy Doniger’s Foreword and Satyamurti’s Preface offer two kinds of answer to these and related questions; in this Afterword, I would like to explore a third angle of vision that complements their perspectives.


The Mahabharata became a subject of international interest beyond the borders of Asia almost two hundred and fifty years ago, when the typographer and philologist Charles Wilkins, working in Calcutta under the patronage of Warren Hastings, then governor-general of the East India Company’s Indian territories, started to translate it from Sanskrit into English. Like most other translators who have followed him, Wilkins was unable to complete the project, but he did publish his rendering of one part it, the Bhagavad Gita, in 1785, which proved to be both popular and influential in Europe.3 Ever since then, scholars and commentators have been divided into two main camps about the form and classification of the Mahabharata, especially in literary terms: one camp essentially views it as a “library,” or a loose-leaf “encyclopedia” at best, whereas the other regards it, first and foremost, as a particular poem in Sanskrit, with a well-defined structure and definite aesthetic properties. It may be difficult to pinpoint the work’s authorship, or to fix its date, place, and process of composition the way we can for modern works, but uncertainties of this kind do not deprive it of specificity as a Sanskrit poem. The poem’s unifying principle does not lie in a “coherent point of view” or a fixed set of themes on the surface, but lies instead, as A. K. Ramanujan also argued in the 1980s, in a multilayered integration of shape and substance, sources and ends, at a deeper level of organization.4 The aesthetic and imaginative aspects of the Mahabharata are vital factors in its reception in world literature today.

Around the turn of the twentieth century, Indian and Euro-American scholars came to generally agree that, given the complexity and importance of the Mahabharata, it was essential to establish a definitive text in its original Sanskrit form. After some delay, a team of Indian Sanskritists, led mainly by V. S. Sukthankar, took up the task independently at the Bhandarkar Oriental Research Institute at Poona (now Pune). They collated and calibrated 1,259 surviving manuscripts, from different parts of the subcontinent, and rigorously evaluated every word and every line in more than 89,000 verses attributed to the poem, before publishing its critical edition in 21 volumes between 1919 and 1969.5 In the course of the past eight decades, the international community of Sanskrit scholars has arrived at a clear consensus that the Poona critical edition gives us the best version of the Mahabharata as a poem that possibly can be reconstructed in modern times.


All the information we can gather and all the inferences and assessments we can make indicate that the poem reconstructed in the critical edition was composed collectively in a preclassical variety of Sanskrit by successive generations of poets between about 400 BCE and 400 CE, on the Gangetic plains in north India, mostly under imperial regimes.6 However, in view of the astonishing connectedness, consistency, and cogency the poem achieves on such a temporal and textual scale, it is conceivable that, at the end of the compositional cycle, the text may well have been assembled, edited, and integrated by a single group of poets, possibly working under one master editor, on the eve of India’s classical age (which runs roughly from 400 to 1200 CE). Given our bias as modern readers—that a poem is “never finished, only abandoned”—it is plausible that the canonical Sanskrit form of the Mahabharata that we have today is the form in which that final editor or group of poets “abandoned” it to the future accidents of history some sixteen hundred years ago.

The text itself says that it is the work of a rishi, or visionary sage, of brahmin patrilineal descent named Krishna Dvaipayana, whose extraordinary life span stretches across several generations of principal characters in the poem, and who is an eyewitness to its events as well as a seminal participant in them. Krishna Dvaipayana—not to be confused with Lord Krishna, the very different divine-human character—is also addressed as “Vyasa” (literally, redactor or editor), and appears forty-four times in the poem’s action. Once he has completed his great poem, Vyasa teaches it to others, including his preeminent pupil, Vaishampayana, who becomes its principal transmitter. Vyasa’s text is broadcast for the first time in his presence and under his supervision, when he authorizes Vaishampayana to recite it in its entirety to King Janamejaya, at a “snake sacrifice,” a Vedic ritual of cosmomoral significance, which the latter sponsors in order to set the world in order early in his reign.

Among those attending this public event is a bard named Ugrashravas, whose father, Lomaharshana (literally, “the teller of hair-raising tales”), is the most famous bard of the times. Ugrashravas carries the vast narrative Vaishampayana delivers at King Janamejaya’s sacrifice to another notable event, a conference of all the hermits who live in Naimisha Forest—a region that, famously, is as much a spiritual and ecological retreat from worldly human society as a sequestered celestial zone of magic and fabulation. At the invitation of the hospitable hermits, who are eager to hear his “tales of wonder,” Ugrashravas recites the text he had heard from Vaishampayana; and, as the Mahabharata’s opening chapter informs us, Ugrashravas’s renarration in the enchanted forest—at two removes from Vyasa—is the version that reaches the rest of humanity as the canonical form of the poem. Thus, unlike the Homeric epics, which offer barely a glimpse of the supposedly blind poet of Greek antiquity, the Sanskrit poem provides us with a full meta-narrative about its origins and transmission.


What makes the Mahabharata a poem, and how is it put together? If we were to answer this question as fully as possible, we would arrive at an account of the Sanskrit work that would serve a purpose similar to that of Aristotle’s Poetics,7 which is concerned with poetry in the ancient Greek world, especially in the genre of drama and the subgenre of tragedy. In order to explain—theoretically and practically, descriptively and prescriptively—how tragic drama is composed and how it works in the theater, Aristotle breaks it down into six constituents and their mutual relations and functions in an artistic whole. These six components are muthos, plot, narrative; ethos, character and its environment; dianoia, thought, theme, meaning; lexis, language, diction; opsis, spectacle onstage; andmelopoeia, melodic composition, lyric set to musical accompaniment. For the Mahabharata, we have no such formula or ready-made analytical framework for a comprehensive explanation. However, keeping Aristotle’s paradigm in mind, along with a range of premodern Indian and modern Euro-American literary scholarship, we can formulate a basic poetics of the Sanskrit poem, by focusing on four of its aesthetic elements: its handling of characters and characterization; its method of emplotment, of plotting its multiple storylines; its treatment of narrators, narration, and narratives; and its conceptualization of its own form and genre.


The Mahabharata probably has more than five hundred human characters who appear as distinct, active figures in its narrative. The majority of them have names and play well-defined roles in the action, even if they appear only briefly; scores of them are one-dimensional and unnamed, and serve mainly as functional devices to get the story told. This rough estimate does not include the unspecified members of groups, assemblies, entourages, communities of slaves and servants, and gathered armies described or invoked in the text; nor does it include the numerous personages who are mentioned by characters and narrators in passing, but who do not participate in the action. The estimate also excludes all non-human characters or participants; if we count the many gods, celestial beings, mythological figures, anti-gods, demons, subhuman creatures, supernatural animals, and creatures belonging to other categories and to mixtures of categories, who have names and play specific roles in the narrative, then the final number would be much higher. As the foregoing catalog indicates, the Mahabharata is a cosmic poetic tale in which almost everything in the human and natural orders is interconnected (and also transmuted in various ways), and both are intimately related to the underworld as well as the world of the gods. The size and spread of this elaborate cast of characters have several consequences for the poem, for its architecture as well as its meanings.

Perhaps the most striking consequence of such a large, variegated roster of characters is that, unlike other epics from the ancient world, the Mahabharata is not centered on a singular hero. It is fundamentally a narrative poem with a disorientingly large group of protagonists and antagonists—twenty-four to be exact—none of whom can be omitted without altering the tale significantly.


The reason the Mahabharata has two dozen leading characters at its core lies in the magnitude of its action, and in the fact that its action is manifold. From the verbal surface as well as the organization of the Sanskrit poem, it is evident that it seeks to develop a comprehensive account of all the events in its narrative, making them plausible and explicable at one level and emotionally and imaginatively resonant at another. The events it represents span the space, time, and causal structure of the cosmos. It uses this enormous scaffolding to narrate the history of the race of its heroes (the Bharatas), together with the history of their land (Bharata-varsha). For its poetic potential and explanatory power, the war of succession to the throne of Hastinapura is the defining moment in both these histories, and it also becomes the master-node upon which the diverse quests and destinies of all the protagonists and antagonists converge.

In the case of an Athenian tragedy, Aristotle’s Poetics conceives of muthos or plot as a single, continuous, and plausible order of events, with secondary episodes integrated with the main developments, and the whole adjusted in scale for representation on the stage. In this conception, the plot is a concrete sequence of events as a spectator perceives it in the course of a performance, from a vantage point “outside” the depicted characters. Simply because of its scope, the action of the Mahabharata cannot be treated on a par with action represented in an amphitheater: if chanted by a relay of reciters at a steady rate of two verses per minute for about eight hours a day, the text of the critical edition would take nearly seventy-five days to deliver in full. But the Indian poem also does not fit into Aristotle’s prescription for the epic—which is supposed to offset the magnitude of its action (“unlimited in time span”) with an “unchanging meter”—because its Sanskrit text does not use prosodic uniformity to contain thematic diversity. To explain how the Mahabharata actually manages its narratives, it therefore may be more productive to turn to the Natyashastra, the earliest work on poetics in the Sanskrit tradition, which began to circulate in India around 300 CE and provides an alternative account of plot.8 In one of the Natyashastra’s perspectives, every character in a narrative is motivated by a desire to attain a specific goal or set of goals; and the character’s actions in pursuit of those goals, at particular times and in particular places and situations, constitute his or her distinctive plotline, which has the shape of a quest. In a work’s overall narrative, each significant character has a specific plotline with a beginning, a middle, and an end determined by the origin, evolution, and conclusion of a quest, and especially by his or her point of view on the quest’s stages and progress. The reader or spectator views this quest empathetically from the character’s point of view, and not from some disembodied or impersonal vantage point “outside” the action. The aggregate plot of a poem, then, is the intertexture of the individual plotlines that unfold in the fictional lives of its principal characters—however many they may be—with each evolving plotline viewed from the “inside,” as it were. The Natyashastra complicates this picture by arguing that a character’s quest and motivation emerge from an initial situation, a bija (seed or germ) that serves as an origin, and whose offshoots comprise the primary and secondary strands in a plotline; and that the offshoots spreading out from the germ are brought into line by a bindu, a “centering principle” that ensures structural as well as discursive cohesion. Moreover, a narrative germ produces a plotline in which a thesis gives way to an anti-thesis, and both yield to a synthesis; the synthesis then bundles the various offshoots of the initial situation into a decisive moment of crisis and a subsequent resolution.

In the case of the Mahabharata, the large roster of protagonists and antagonists means that the poem’s overall emplotment has some two dozen distinct plotlines, each centered on a particular character pursuing a distinctive set of goals in interaction with other characters. Readers and listeners who grow up with the Mahabharata in their inherited cultural environment learn, from others around them, to imaginatively traverse the poem’s aggregate plot many times over, and to grasp the totality of the represented action from the different points of view of different principal characters. Thus, on one occasion, one may follow Kunti’s plotline from beginning to end, and grasp the epic as a whole from her point of view as Pandu’s wife and widow, as the biological mother of Yudhishthira, Bhima, and Arjuna and of Karna (each fathered on her by a different god), as the foster mother of Nakula and Sahadeva, as Draupadi’s mother-in-law, and as Duryodhana’s, as well as Lord Krishna’s, aunt; and on another occasion, one may view everything afresh through the eyes of, say, Karna or Duryodhana or Arjuna, and discover how different that character, as well as the whole epic, looks. Many Indians, perhaps the majority of those familiar with the Mahabharata, internalize this (non-Aristotelian) process of shifting interpretation over a lifetime, and hence entertain multiple perspectives on the characters, events, and meanings of the epic. Given the range of characters and the number of protagonists and antagonists, as well as the scope of the represented action, no one character’s quest, point of view, or plotline defines a true or complete interpretation. That is, the absence of an omniscient narrator and of an absolute frame of reference in the Mahabharata implies that the coexistence of multiple perspectives in a multiplex plot is fundamental to its poetics. At the same time, the protagonists’ and antagonists’ divergent quests and points of view are held together by the bija, or germ, from which all of them spring (who should inherit the kingdom, under what conditions and why), by the progression from thesis to anti-thesis to synthesis in this central emplotment (from dishonorable disinheritance to failed diplomacy to inescapable war), and by the establishment of a just society and kingdom in the war’s aftermath.


If the action represented in the Mahabharata is inherently manifold, and its emplotment in the poem is necessarily multiplex, then the technical challenge facing the poets who finalized the Sanskrit text had to be the same that still faces every storyteller in modern times: since no story is ever one story only, and the telling of one story always spills over into the telling of other stories, what is the most efficient and effective way to tell multiple stories together? Historically speaking, even though the Ramayana had experimented earlier with an answer on a moderate scale, and the Buddhist Jataka moral tales and the Hindu Panchatantra animal fables had developed basic strategies for short narratives, the Mahabharata was the first work in world literature to confront this technical problem head-on, and to invent a solution that remains the most comprehensive one to this day.

The Mahabharata’s strategy for multiple narratives and narrations is to maximize the possibilities opened up by the device of narrative framing. At its simplest, a narrative frame is defined by a narrator and his or her narration and narrative; whenever a narrator is explicitly identified in a poem, and relates a particular emplotted sequence of events, the resulting narrative defines the extent of his or her narrative frame. If a narrator stops narrating events, and a different narrator takes up the narrative (or starts a different one), then the poem changes its narrative frame, and we, the audience, adjust our expectations and interpretive procedures accordingly. A poem with multiple framed narratives has to employ devices that inform its audience about who the narrator is at any given moment, what his or her ethos or character and setting are, and what we may expect from him or her; about the junctures at which a particular narrator’s narrative begins and ends, or various narrative frames open and close; and exactly what the structural relationships are between one frame and another, and among all the frames and narrators that may be housed inside an aggregate plot.

The Mahabharata exploits the potential of the last of these devices fully, by employing all the vital relationships among narrative frames, and all their variations and combinations. One closed frame may be fully contained inside a larger one; one frame may be contiguous with another, so that the latter opens as soon as the former closes; one larger frame may contain two or more closed frames that are not contiguous with each other; and in the most complex interrelation, two frames may significantly overlap, so that a new frame opens well before—and continues well after—the previous one closes. Moreover, the epic does not hesitate to superpose these structures of containment, contiguity, discrete serialization, and intersection upon each other, or to use them in tandem. The variety of frames and framings produces the thick interweave that constitutes the actual text of the Sanskrit poem.

The Mahabharata as a whole is solidly encased in three outer narrative frames, nested successively inside one another, that are never dislodged over its entire length. For us, as an audience, the outermost frame always belongs to the “voice” that tells us the story, in a recitation or on the page, here and now; inside this anonymous frame is Ugrashravas’s frame, retelling the epic to the sages gathered in the Naimisha Forest; and inside that frame is Vaishampayana’s frame, in which the latter recites the entire Mahabharataat Janamejaya’s snake sacrifice. All the events we read about or hear being narrated in the epic are always inside Vaishampyana’s frame—except for the meta-events that reach us from Ugrashravas’s outer frame. My rough estimate is that, inside the three outermost frames, we encounter some four hundred distinct narrative frames that fall along a spectrum, one end of which is marked by narrators who appear many times, and the other end by narrators who appear only once each, to convey a single fable, explanation, vignette, or thumbnail sketch.

This vast structure of emboxed narrative frames—conceptually resembling a set of Russian dolls or Chinese boxes, fitting snugly one inside another—allows the Mahabharata to create the most compact configuration for the integrated narration of multiple stories, in which both the causal interrelations among events and the mimetic force of their representation in the poem are articulated clearly. The tightness of this structure in each recension, as in the critical edition of the poem, cannot be overemphasized. In the various Sanskrit recensions we have inherited, each narrative frame is duly opened and closed, and the poem stands consistently as a well-crafted whole, its great diversity of themes held together by the underlying grid of narrative frames.


The Mahabharata derives one kind of structural integrity from its scaffolding of framed narratives, but its poetic qualities depend as much on what it communicates as on its mode or manner of communication. What, then, is the text’s shaping principle with respect to its raw material, and what kind of poem does it become in the process of representing its content? In the past two centuries, Euro-American scholars have often found it difficult to respond to the Sanskrit work as an aesthetic object; as a consequence, they generally have ignored its structural integrity, and have focused instead on its uses as a religious text and its usefulness as a cultural document for outsiders, especially as a source of information about ancient Indian society, religion, politics, law, and morality. For a much longer time, Indian audiences have also turned to the Mahabharata primarily for its content, and perhaps secondarily for its aesthetic qualities—but this is so because, in the practice of Hinduism, the Sanskrit text has the status of scripture or quasi-scripture, next in spiritual authority only to the four Vedas. As a “fifth Veda” or body of authoritative knowledge and discourse, it is valued for its guidance on duty and ethics, its teachings about right and wrong, its arguments about the human self and the ends of life, its analysis of war and justice, and its vision of a good ruler and a good society. At the same time, however, the subcontinent’s artists have long affirmed the epic’s power as a work of beauty and imagination—not apart from, but in addition to, its power as a text linked to scripture revealed in “the language of the gods.”9

Nevertheless, just as the Mahabharata provides a picture of its author, its authorship, and its transmission, it also offers commentary on its form and genre. One of its classifications of itself is as an akhyana, which identifies its mode of presentation and its poetic function. This label means that the poem sees itself as a telling, a narration, or an informative communication; that what it relates is specifically an old tale or a legend; and that its preferred mode of delivery is speech or oral performance rather than writing. The Sanskrit text also describes itself as an upakhyana which, as a nuance of akhyana, means that it is a retelling, rather than an original narration, of a tale heard earlier from others.

When the Mahabharata uses a different self-descriptive label, samvada, it characterizes its own form as dialogue. This is a precise specification because, in recitation and on the printed page, the Sanskrit text presents itself, from beginning to end, as a vast dialogue involving primary and secondary (and sometimes also tertiary) characters, nested inside and outside its numerous narrative frames. The poem, in fact, is fundamentally dialogic, not only in the sense of being cast in the form of a verbal exchange, but in the more robust sense of belonging to “the dialogic mode.”10 In Bakhtin’s theory, this is the mode in which individuals exist in human society in constant interaction with others, so that any one person’s thought and speech on any given occasion are always already “in dialogue with” some other—or someone else’s—prior thought and speech. The poets of the Mahabharata explicitly understand and acknowledge dialogism as the shaping principle of their poetry, and they systematically flesh out the entire poem as a vast, multivoiced, ongoing samvada about everything on earth and in the universe, inside as well as outside the borders of human experience, so that, as they claim, “Whatever is found here may be found somewhere else, but what is not found here is found nowhere else.” As a dialogic poem that goes well beyond the surface form of verbal exchange, the Mahabharata is an active response to the states of affairs it depicts, and its narrative therefore is always a multifarious modification of plain mimetic representation.

Finally, the Sanskrit poem also describes its genre in terms of its theme or content. Almost two millennia ago, the poets who finalized its text cast it as an epic, but they did not have a label for their enterprise. Itihasa was the everyday word they adopted to name the genre with which they could represent the absolute past of the people and the land of which they were the imaginative inheritors. At least two thousand years old, the word itihasa literally means “thus it was” or “so it happened,” and is the exact Sanskrit precursor of the German phrase wie es eigentlich gewesn ist, “as it has been actually.” By claiming to be an itihasa, the Mahabharata seems to assert its function as a mimetic representation of events past. Following this implication, Indian as well as Euro-American scholars since the nineteenth century have tried to interpret the Mahabharata as history—often too literally, and usually with absurd outcomes. Any encounter with the Mahabharata’s narrative, however, indicates immediately that the poem does not define itself as a history in the modern sense. Its poets are under no illusion that they somehow are composing a factual, empirically verifiable, or documentary account of the past, and they do not wish to impose any such illusion on their audience. Instead, they seem to focus on events that are long over, even for them, to which they themselves no longer have any real access (through direct experience or personal memory or reliable eyewitness reports), and which they know they can memorialize only poetically—events that can be “recovered” solely by imagining and reimagining, narrating and renarrating what Goethe, with a fellow poet’s acuteness, called “the absolute past.”

Bakhtin, again, explains what this means in a way that fits the Mahabharata with surprising precision. The genre that is designed to represent the absolute past is the genre that we now designate as the epic which, for the Russian theorist, has “three constitutive features”: it seeks to represent the “epic past” of a nation; it draws on “national tradition” for its narrative, and not on its author’s “personal experience and the free thought that grows out of it”; and, in “Goethe’s and Schiller’s terminology,” it establishes and works across “an absolute epic distance” between “the time in which the singer . . . lives”—or the worlds in which the author and his audience exist—and “the epic world” that it depicts. These three features, Bakhtin argues in his emphatic style, are interrelated:

The world of the epic is the national heroic past: it is a world of “beginnings” and “peak times” in the national history, a world of fathers and founders of families, a world of “firsts” and “bests” . . . The epic was never a poem about the present, about its own time . . . [It] has been from the beginning a poem about the past, and the authorial position immanent in . . . and constitutive for it . . . is [that] . . . of a man speaking about a past that is to him inaccessible, the reverent point of view of a descendant . . . Both the singer and the listener, immanent in the epic as a genre, are located in the same time and on the same evaluative (hierarchical) plane, but the represented world of the heroes stands on an utterly different and inaccessible time-and-value plane, separated by epic distance. The space between them is filled with national tradition. To portray an event on the same time-and-value plane as oneself and one’s contemporaries . . . is . . . to step out of the world of epic into the world of the novel.11

The Mahabharata displays all the characteristics that Bakhtin attributes to an epic. It combines the epic mode with myth and romance, allegory and high mimeticism; and it perfects the method of narrative framing, using it to contain multiple plotlines and divergent points of view, as each of its main characters pursues his or her own quest. It is at once a telling and a retelling, a dialogue and a history—but it is especially a poem on a grand scale about a clan divided by hatred, a queen molested and avenged, a just war against an unjust disinheritance, a victory that is indistinguishable from defeat, and the death of an old order and the birth of a new one.


Satyamurti, an established British poet with a connection to India, approaches the Mahabharata as a modern poet responding poetically to an ancient poem. In its most direct form, her objective is to understand the Sanskrit work, to capture her understanding of it in a poem of her own, and to place her text before contemporary readers in English—for whom it then becomes a means to experience the poetry of the Mahabharata, though necessarily at a remove. Her concern is with poetry at every stage in an idealized circuit: as an object of understanding, as a mode of understanding, as a medium of expression, as a vehicle of communication, and as an outcome of the process as a whole. Poetry thus becomes a seamless continuum enveloping her enterprise, an analogue of the Irish poet W. B. Yeats’s “hermetic egg,” which needs nothing outside itself in order to give birth to new life.

There is, however, one practical rupture in this continuum. Whenever a modern work sets out to represent an older work, it can only do so in a genre that shades off into other genres performing similar functions but in other ways. A text that represents another text may be a translation (“faithful”), an adaptation (somewhat “loose”), a retelling (relatively “free”), or even a reworking (“creative”). In the past two decades, literary theorists have argued persuasively that all these categories can be placed on a single conceptual gradient called “translation” in the broadest sense, which moves from the most literal rendering of a text at one end to the most approximate at the other. In the late seventeenth century, John Dryden suggested brilliantly that the three defining positions on this spectrum be labeled “metaphrase,” which is a word-for-word or interlinear version; “paraphrase,” which deviates from the letter of a text, but not its spirit; and “imitation,” which ignores the letter, and also merely “strives after” its spirit. In general, translators may pursue any of these shades of rendering legitimately, provided they identify the genre of their output without ambiguity—that is, as a translation, a paraphrase, an adaptation, an imitation, and so on. For whichever genre they practice, translators can choose one of two overall methods: the direct method, where the translator adequately knows and uses both the languages involved, the one from and the one into which she is translating; and the indirect method, where she knows only the language into which she is translating, and has access to the original text solely through preexisting intermediary renderings and other resources.

In Satyamurti’s project, the break appears at the link of genre of representation and choice of method. On the general spectrum of translation, her retelling can be seen as a paraphrase. Her method, however, is the indirect method, because she is not a scholar of the Sanskrit language, and has access to the original text of the Mahabharata only through intermediary English translations and commentaries. Since her poetics of retelling rests on the assumption that her English poem is a reliable representation of the Sanskrit poem, she has to ensure that her intermediary resources are as trustworthy as possible. Her version of the epic therefore is based on the most literal and scholarly renderings available in English, together with the most dependable and informative commentaries, selections, and condensations.

In its practical application, Satyamurti’s indirect method is as elaborate and systematic as it is fine-tuned. Her retelling is not a mere versification of an existing prose translation or someone else’s gloss or crib. Depending on the material and her selection, she chooses to narrate some parts in detail, some in a condensed form, and some only in brisk summary. The Sanskrit original and its prose metaphrases in English may handle a given episode in uniform depth, but Satyamurti omits some parts altogether and treats other parts differentially, guided by her sense of their poetic value and meaning, their significance in the larger narrative, and their potential imaginative impact on her readers. Even as she goes by the emotional flow of sounds, rhythms, images, characters, and events in her own verse, however, she works strictly within the limits of her intermediary resources, which determine the reliability of her rendering. Given the meticulousness of her craft and her strong sense of balance and proportion, restraint and understatement, her retelling emerges as a neatly scaled “miniature representation” of a gargantuan whole.

The proportionality that Satyamurti is able to maintain follows largely from her strategies and decisions at the level of craft. For one, she pays close attention to the poetic organization of the Mahabharata. She keeps all 18 major books in her version, but she treats them astutely as “accordion structures” that can be expanded and compressed to fit narrative exigencies; she represents the minor books as “chapters,” and reduces their number to 60. Each of her chapters corresponds to one or more minor books and, within its framework, she carefully selects the characters, episodes, stories, and themes that will advance the narrative effectively. For whatever she chooses to highlight in a chapter, she usually combines detailed narration with condensation and précis to keep up the proportionality between retelling and original. Her poem is only one-twelfth the size of the Mahabharata, but it still gives us a detailed and balanced picture of the whole.

For another input at the level of craft, Satyamurti brings extraordinary discipline and inventiveness to her versification. Unlike the Iliad and the Odyssey, each of which is composed uniformly in a single meter, the Mahabharata is a composite poem in Sanskrit. In its critical edition, the text contains 73,821 numbered units; almost 99.5 percent of these are in verse, while 385 of them are prose passages, the latter being distributed over 12 “cantos” in 6 minor books. Of the 73,436 units in verse, 4,426 are composed in meters belonging to the trishtubh class, whereas 68,858 verses are in meters of the anushtubh class. The common form of this last category is the shloka, which consists of 32 syllables arranged in two equal “lines,” each divided at its midpoint by a caesura; structurally, we can view the shloka as either a couplet (with 16 syllables per line) or a quatrain (with 8 syllables per line). In Sanskrit, a “verse” is syntactically closed, so it ends in a period; the end of a “line” coincides with a syntactic break, and hence coincides with the end either of a clause or of a sentence; and a caesura within a line, whether in a symmetrical or an asymmetrical position, coincides with the end either of a phrase or of a clause but not of a sentence. The short and long pauses around the middle and at the end of a line, and the long pause at the end of a verse, reinforce the rhythm arising from the metrical pattern, which gives the verse an incantatory or lyrical quality. One consequence is that, whenever a Sanskrit text composed in verse is delivered orally, it is not merely recited or read aloud, it is either chanted or sung. The conceptions involved here, especially of lyricism, differ from those in a language such as English, because Sanskrit verse forms do not require end rhyme or internal rhyme. The convention of chanting and singing applies as much to a short poem, a ritual mantra, and a philosophical treatise, as to an epic. Given the general prosodic features of Sanskrit, and given that the bulk of its text (about 75 percent) is composed in the shlokaverse form, the Mahabharata appears to graft a “lyrical” texture onto a metrically variable narrative of massive size.

Satyamurti makes no attempt to reproduce or even approximate this system of versification, which is impossible in any case because of fundamental differences in phonology, syllabification, and prosody. Instead, she chooses boldly to invigorate English blank verse, developing a flexible line of nine to eleven syllables, with an average of five stresses. She brings it forward into the twenty-first century, not only by reducing its mechanical quality, but specifically by blending its metrical pattern with the syntax and vocabulary of contemporary “middle diction.” Satyamurti’s blank verse avoids poeticisms as well as archaisms, syntactic inversions as well as semantic simplifications, without becoming prosy or prosaic. Her sentences flow smoothly through enjambments and shifting caesuras, providing a medium for the narrative that is both transparent and dynamic—as amenable to the conversation of the gods and the fury of battle, as to philosophical disquisition, theological debate, deathbed speech, domestic vignette, emotional outburst, and evocation of landscape. A light and elastic surface of this kind stands in contrast to the Mahabharata’s verbal texture, but it is the ideal vehicle in English for a multiplex narrative that stretches out to 27,000 lines. It transports us back to a past that is cut off from us, as in an epic, but it also brings that past alive, here and now, as though it were a novel in finely crafted verse.


1. Robert Lowell, Imitations (1961; New York: Farrar, Straus and Giroux, 1990), p. 68

2. Detailed information on the epic’s textual history, print publication, and critical edition appears in the general introduction to The Mahabharata, vol. 1: The Book of the Beginning, translated and edited by J. A. B. van Buitenen (Chicago: University of Chicago Press, 1973), pp. xiii–xliv; see especially pp. xxiii–xxxix.

3. J. J. Clarke, Oriental Enlightenment: An Encounter between Asian and Western Thought (New York: Routledge, 1997), pp. 58–59 and 85.

4. On classical arguments and his own position, see Ramanujan’s “Repetition in the Mahabharata,” in The Collected Essays of A. K. Ramanujan, edited by Vinay Dharwadker (New Delhi: Oxford University Press, 1999), pp. 161–83, especially p. 163.

5. V. S. Sukthankar et al., Critical Edition of the Mahabharata, 21 vols. (Poona, Maharashtra, India: Bhandarkar Oriental Research Institute, 1919–69). The edition includes 18 volumes for the 18 parvans, 2 volumes for the Harivamsha, and 1 volume for the critical apparatus. The editors consulted 1,259 manuscripts, and processed over 89,000 verses attributed to the epic.

6. The case for multiple authors of the Mahabharata is made in van Buitenen, op. cit., “Introduction.”

7. Aristotle, Poetics, edited and translated by Stephen Halliwell, Loeb Classical Library 199 (Cambridge, Mass.: Harvard University Press, 1995), pp. 47–55.

8. Manomohan Ghosh, trans., The Natyashastra, vol. 1: chaps. 1–27 (Calcutta: Royal Asiatic Society of Bengal, 1950). Plot and emplotment are covered in Chapter 21, pp. 380–400. Ghosh’s translation is unfortunately opaque on this topic; my overview uses the Sanskrit text.

9. “The language of the gods” translates an ancient epithet for Sanskrit as a sacred medium.

10. See M. M. Bakhtin, Dialogic Imagination: Four Essays, edited by Michael Holquist (Austin: University of Texas Press, 1981), pp. 426–27.

11. Bakhtin, ibid., pp. 13–14.

If you find an error please notify us in the comments. Thank you!