Let me make it clear that the raging Medusa of cultural relativism is not rearing her ugly head in my discussion at this point. To allow for plurality, signified by the plurality of gods, is to think in terms of singularities. To think in terms of singularities, however—and this I must make clear since so many scholars these days are prone to see parochialism, essentialism, or cultural relativism in every claim of non-Western difference—is not to make a claim against the demonstrable and documentable permeability of cultures and languages. It is, in fact, to appeal to models of cross-cultural and cross-categorical translations that do not take a universal middle term for granted. The Hindi pani may be translated into the English “water” without having to go through the superior positivity of H2O. In this, at least in India but perhaps elsewhere as well, we have something to learn from nonmodern instances of cross-categorial translation.
I give an example here of the translation of Hindu gods into expressions of Islamic divinity that was performed in an eighteenth-century Bengali religious text called Shunya-puran. (The evidence belongs to the “history of conversion” to Islam in Bengal.) This text has a description, well known to students of Bengali literature, of Islamic wrath falling upon a group of oppressive Brahmins. As part of this description, it gives the following account of an exchange of identities between individual Hindu deities and their Islamic counterparts. What is of interest here is the way this translation of divinities works :
Dharma who resided in Baikuntha was grieved to see all this [Brahminic misconduct]. He came to the world as a Muhammadan … [and] was called Khoda… . Brahma incarnated himself as Muhammad, Visnu as Paigambar and Civa became Adamfa (Adam). Ganesa came as a Gazi, Kartika as a Kazi, Narada became a Sekha and Indra a Moulana. The Risis of heaven became Fakirs… . The goddess Chandi incarnated herself as Haya Bibi [the wife of the original man] and Padmavati became Bibi Nur [Nur = light].
Eaton’s recent study of Islam in Bengal gives many more such instances of translation of gods. Consider the case of an Arabic-Sankrit bilingual inscription from a thirteenth-century mosque in coastal Gujarat that Eaton cites in his discussion. The Arabic part of this inscription, dated 1264, “refers to the deity worshiped in the mosque as Allah” while, as Eaton puts it, “the Sanskrit text of the same inscription addresses the supreme god by the names Visvanatha (‘lord of the universe’), Sunyarupa (‘one whose form is of the void’), and Visvarupa (‘having various forms’).” Further on, Eaton gives another example : “The sixteenth-century poet Haji Muhammad identified the Arabic Allah with Gosai (Skt. ‘Master’), Saiyid Murtaza identified the Prophet’s daughter Fatima with Jagat-janani (Skt ‘Mother of the World’), and Saiyid Sultan identified the God of Adam, Abraham, and Moses with Prabhu (Skt. ‘Lord’).”
In a similar vein, Carl W. Ernst’s study of South Asian Sufism mentions a coin issued by Sultan Mahmud of Ghazna (c. 1018 C.E.) that contained “a Sanskrit translation of the Islamic profession of faith.” One side of the coin had an Arabic inscription whereas the other side said, in Sanskrit : avyaktam ekam muhamadah avatarah nrpati mahamuda (which Ernst translates as, “There is One unlimited [unmanifest?], Muhammad is the avatar, the king is Mahmud”). Ernst comments, expressing a sensibility that is no doubt modern : “The selection of the term avatar to translate the Arabic rasul, ‘messenger,’ is striking, since avatar is a term reserved in Indian thought for the descent of the god Vishnu into earthly form… . It is hard to do more than wonder at the theological originality of equating the Prophet with the avatar of Vishnu.”
The interesting point, for our purpose and in our language, is how the translations in these passages take for their model of exchange barter rather than the generalized exchange of commodities, which always needs the mediation of a universal, homogenizing middle term (such as, in Marxism, abstract labor). The translations here are based on very local, particular, one-for-one exchanges, guided in part, no doubt—at least in the case of Shunya-puran—by the poetic requirements of alliterations, meter, rhetorical conventions, and so on. There are surely rules in these exchanges, but the point is that even if I cannot decipher them all—and even if they are not all decipherable, that is to say, even if the processes of translation contain a degree of opacity—it can be safely asserted that these rules cannot and would not claim to have the “universal” character of the rules that sustain conversations between social scientists working on disparate sites of the world. As Gautam Bhadra has written : “One of the major features of these types of cultural interaction [between Hindus and Muslims] is to be seen at the linguistic level. Here, recourse is often had to the consonance of sounds or images to transform one god into another, a procedure that appeals more … to popular responses to alliteration, rhyming and other rhetorical devices—rather than to any elaborate structure of reason and argument.”
One critical aspect of this mode of translation is that it makes no appeal to any of the implicit universals that inhere in the sociological imagination. When it is claimed, for instance, by persons belonging to devotional traditions (bhakti) that “the Hindu’s Ram is the same as the Muslim’s Rahim,” the contention is not that some third category expresses the attributes of Ram or Rahim better than either of these two terms and thus mediates in the relationship between the two. Yet such claim is precisely what would mark an act of translation modeled on Newtonian science. The claim there would be that not only do H2O, water, and pani refer to the same entity or substance but that H2O best expresses or captures the attributes, the constitutional properties, of this substance. “God” became such an item of universal equivalence in the nineteenth century, but this is not characteristic of the kind of cross-categorial translations we are dealing with here.
Consider the additional example Ernst provides of such nonmodern translation of gods. He mentions “a fifteenth century Sanskrit text written in Gujarati for guidance of Indian architects employed to build mosques. In it, the god Visvakarma says of the mosque, ‘There is no image and there they worship, through dhyana, … the formless, attributeless, all-pervading Supreme God whom they call Rahamana.’” The expression “supreme God” does not function in the manner of a scientific third term, for it has no higher claims of descriptive ability, it does not stand for a truer reality. For, after all, if the supreme One was without attributes, how could one human language claim to have captured the attributes of this divinity better than a word in another language that is also human ? These instances of translation do not necessarily suggest peace and harmony between Hindus and Muslims, but they are translations in which codes are switched locally, without going through a universal set of rules. There are no overarching censoring/limiting/defining systems of thought that neutralize and relegate differences to the margins, nothing like an overarching category of “religion” that is supposed to remain unaffected by differences between the entities it seeks to name and thereby contain. The very obscurity of the translation process allows the incorporation of that which remains untranslatable.