The following commentary comes from Chapter 2 ("Three Hundred Ramayanas: Five Examples and Three Thoughts on Translation") of Many Ramayanas: The Diversity of a Narrative Tradition in South Asia, ed. by Paula Richman. The section immediately preceding the text excerpted below ends as follows: "Some shadow of a relational structure claims the name of Ramayana for all these tellings, but on closer look one is not necessarily all that like another. Like a collection of people with the same proper name, they make a class in name alone." The texts Ramanujan refers to as "tellings" are a set of alternate versions of segments of the Ramayana.
That may be too extreme a way of putting it. Let me back up and say it differently, in a way that covers more adequately the differences between the texts and their relations to each other, for they are related. One might think of them as a series of translations clustering around one or another in a family of texts: a number of them cluster around Valmiki, another set around the Jaina Vimalasuri, and so on.
Or these translation-relations between texts could be thought of in Peircean terms, at least in three ways.
Where Text I and Text 2 have a geometrical resemblance to each other, as one triangle to another (whatever the angles, sizes, or colors of the lines), we call such a relation iconic . In the West, we generally expect translations to be "faithful," i.e. iconic. Thus, when Chapman translates Homer, he not only preserves basic textual features such as characters, imagery, and order of incidents , but tries to reproduce a hexameter and retain the same number of lines as in the original Greek—only the language is English and the idiom Elizabethan. When Kampan retells Valmiki's Ramayana in Tamil, he is largely faithful in keeping to the order and sequence of episodes, the structural relations between the characters of father, son, brothers, wives, friends, and enemies. But the iconicity is limited to such structural relations. His work is much longer than Valmiki's, for example, and it is composed in more than twenty different kinds of Tamil meters, while Valmiki's is mostly in the sloka meter.
Very often, although Text 2 stands in an iconic relationship to Text ! in terms of basic elements such as plot, it is filled with local detail, folklore, poetic traditions, imagery, and so forth—as in Kampan's telling or that of the Bengali Krttivasa. In the Bengali Ramayana , Rama's wedding is very much a Bengali wedding, with Bengali customs and Bengali cuisine. We may call such a text indexical : the text is embedded in a locale, a context, refers to it, even signifies it, and would not make much sense without it. Here, one may say, the Ramayana is not merely a set of individual texts, but a genre with a variety of instances.
Now and then, as we have seen, Text 2 uses the plot and characters and names of Text 1 minimally and uses them to say entirely new things, often in an effort to subvert the predecessor by producing a countertext. We may call such a translation symbolic . The word translation itself here acquires a somewhat mathematical sense, of mapping a structure of relations onto another plane or another symbolic system. When this happens, the Rama story has become almost a second language of the whole culture area, a shared core of names, characters, incidents, and motifs, with a narrative language in which Text 1 can say one thing and Text 2 something else, even the exact opposite. Valmiki's Hindu and Vimalasuri's Jaina texts in India—or the Thai Ramakirti in Southeast Asia—are such symbolic translations of each other.
One must not forget that to some extent all translations, even the so-called faithful iconic ones, inevitably have all three kinds of elements. When Goldman and his group of scholars produce a modern translation of Valmiki's Ramayana , they are iconic in the transliteration of Sanskrit names, the number and sequence of verses, the order of the episodes, and so forth. But they are also indexical, in that the translation is in English idiom and comes equipped with introductions and explanatory footnotes, which inevitably contain twentieth-century attitudes and misprisions; and symbolic, in that they cannot avoid conveying through this translation modern understandings proper to their reading of the text. But the proportions between the three kinds of relations differ vastly between Kampan and Goldman. And we accordingly read them for different reasons and with different aesthetic expectations. We read the scholarly modern English translation largely to gain a sense of the original Valmiki, and we consider it successful to the extent that it resembles the original. We read Kampan to read Kampan, and we judge him on his own terms—not by his resemblance to Valmiki but, if anything, by the extent that he differs from Valmiki. In the one, we rejoice in the similarity; in the other, we cherish and savor the differences.
One may go further and say that the cultural area in which Ramayanas are endemic has a pool of signifiers (like a gene pool), signifiers that include plots, characters, names, geography, incidents, and relationships. Oral, written, and performance traditions, phrases, proverbs, and even sneers carry allusions to the Rama story. When someone is carrying on, you say, "What's this Ramayana now? Enough." In Tamil, a narrow room is called a kiskindha ; a proverb about a dim-witted person says, "After hearing theRamayana all night, he asks how Rama is related to Sita"; in a Bengali arithmetic textbook, children are asked to figure the dimensions of what is left of a wall that Hanuman built, after he has broken down part of it in mischief. And to these must be added marriage songs, narrative poems, place legends, temple myths, paintings, sculpture, and the many performing arts.
These various texts not only relate to prior texts directly, to borrow or refute, but they relate to each other through this common code or common pool. Every author, if one may hazard a metaphor, dips into it and brings out a unique crystallization, a new text with a unique texture and a fresh context. The great texts rework the small ones, for "lions are made of sheep," as Valery said. And sheep are made of lions, too: a folk legend says that Hanuman wrote the original Ramayana on a mountaintop, after the great war, and scattered the manuscript; it was many times larger than what we have now. Valmiki is said to have captured only a fragment of it. In this sense, no text is original, yet no telling is a mere retelling—and the story has no closure, although it may be enclosed in a text. In India and in Southeast Asia, no one ever reads the Ramayana or the Mahabharata for the first time. The stories are there, "always already."