114 SOCIAL SCIENTIST "literatures" may not be coeval with their languages. This becomes more comprehensible if "beginning" and "literary" are clarified. I am not unaware of the enigmas that beset the very idea of "beginnings," and the complicated issues in historiography, and indeed, in epistemology, ontology, and ideology, that beginnings raise. A beginning is always provisional until something earlier can be found, and therefore it always bears the trace of an absence. There are discursive and categorical complications to beginnings: what in fact is "Old Gujarati" or "Maithili," and how do we decide the former is not Apabhramsa, and the latter not Bengali or Hindi? What is "the novel"? There are ideational complications to beginnings: agents may think that they are making a new beginning when (to other eyes) they are not; or they may imagine that they are simply reproducing the old when (to other eyes) they are making the new. Moreover, beginnings are often only what traditions choose to make into beginnings; they may remember or erase one beginning in favor of another; indeed, it is through the very denial of the possibility of beginning—the delusions of autochthony and primordiality—that traditions like nations often constitute themselves. In regard to South Asia historiography moreparticulariy, lhavenoted a stubborn if usually tacit insistence that real new beginnings are somehow conceptually permissible only with cegard to colonialism. Colonial critique derives its power largely from the assumption of the sharp discontinuity and new beginning (in consciousness, power, culture) that colonialism, it is thought, uniquely wrought. Similar beginnings seem ex hypothesi excluded for precoloniality, which therefore is left without a history.3 None of these enigmas, however, is fatal to the historiography of the "literary" in South Asia. They are, to repeat, the very problems we need to subject to historical analysis. This historicization gains special purchase from taking seriously the representations of beginnings within literary traditions themselves. How people have thought, that they have made history with literature—emic literary histories— is as important a component of the history of literary cultures as any positivist facts we pride ourselves on recovering. And from this perspective what agents thought counted as "Kannada" or "the novel" is itself a historical truth, which linguists or literary historians may be right to challenge but cannot ignore. One way to think about beginnings in the history of literary cultures is to investigate the role of writing. For in certain crucial ways the category of the "literary" itself is linked, conceptually and historically, emically and etically— both outside and inside the indigenous systems of the literary—with the category of writing. Historically the most important way both of making history with "literature" and of making "literature" as distinct from something else, has been to write it down. It is "literizadon" as such, and not so much its different modalities (manuscript, print), that seems to mark the great discontinuum in culture, history, and above all, power. And it is power we shall fail to grasp if the role of writing in making the history of "literature" is overlooked. Literatures commence, as I want to understand those terms, is ways that seem to conform to a strong cultural tendency. This forms the third topic I want to address, the role of superposed cultural formations: Literatures develop in response to other,