Recent struggles and debates around the rather tentative concept of multiculturalism in the Western democracies have often fueled discussions of minority histories. The question has arisen in all democracies of including the histories of previously excluded groups in the history of the nation. Thus, the expression "minority histories" has come to refer to all those pasts that democratically-minded historians fought to include in mainstream narratives of the nation. Official or officially blessed accounts of the nation's past have been challenged in many countries by the champions of minority histories. Postmodern critiques of "grand narratives" have been used as ammunition in the process to argue that the nation cannot have just one standardized narrative, that the nation is always a contingent result of many contesting narratives.

Conceived in this way, "minority histories" are oppositional chiefly in the early part of their careers, when they are excluded from mainstream historical narratives. As soon as they are "in," the oppositional stance appears to become redundant. Begun in an oppositional mode, "minority histories" can end up being additional instances of "good history." One can ask legitimate Foucauldian questions about who has the authority to define what "good" history is or what relationships between power and knowledge are invested in such definitions, but let us put them aside for the moment.

The transformation of once-oppositional, minority histories into "good" histories illustrates how the mechanism of incorporation works in the discipline of history. The process of canon-formation in history-curricula in Anglo-American universities is different from the corresponding process in Literature/English departments. Any account of the past can be absorbed into, and thus made to enrich, the mainstream of historical discourse if two questions can be answered in the affirmative. First, can the story be told (or, phrased more directly, can you write histories of suppressed groups)? Second, does it allow for a rationally defensible point of view or position from which to tell the story? The authorial position can be an ideology, a moral position, or a political philosophy, but the choices here are not unlimited. A mad man's narrative is not history. Neither is an arbitrary or personal preference. The investment in a certain kind of rationality and in a particular understanding of the "real" means that history's, the discipline's, exclusions are ultimately epistemological.

I give two instances to show that so long as these two questions can be answered in the positive, the discipline can incorporate marginal or minority positions into itself, or even make them central to itself.

My first case is that of British social-democratic history or so-called history from below. History has not been the same since Thompson and Hobsbawm made the working classes the major actors in society, and feminist historians made us realize the importance of gender relations and the contributions of women to critical social processes. So the question whether such incorporation changes the nature of historical discourse itself can be answered simply: "of course, it does." But the answer to the question, Did such incorporation call the discipline into any kind of crisis? would have to be, "No." To be able to tell the story of groups hitherto overlooked, to be able to master the problems of crafting such narratives is how the discipline of history renews and maintains itself.

Rationality in the Discourse of History

The point about historical narratives requiring a certain minimum investment in rationality has recently been made in the discussion of postmodernism in the book Telling the Truth About History.1 The relationship between minority histories and postwar democracies is at the heart of this book written jointly by three leading feminist historians of the United States. To the extent that the authors read postmodernism as allowing for multiple narratives, they welcome the influence of postmodernism and thus align themselves with the democratic cause of minority histories. However, the book registers a much stronger degree of discomfiture where it encounters arguments that use the idea of multiple narratives to question any idea of truth or facts. For here the idea of a rationally defensible position in public life from which to craft even a multivocal narrative, is brought into question. If "minority histories" go to the extent of questioning the very idea of fact or evidence, then, the authors ask, how would you find ways of adjudicating between competing claims in public life? Would not the absence of a certain minimum agreement about what constitutes fact and evidence seriously fragment the body politic in the United States of America, and would not that in turn impair the capacity of the nation to function as a whole? Hence the authors recommend a pragmatic idea of "workable truths," based on a shared, rational understanding of historical facts and evidence. For a nation to function effectively, while eschewing any claims to a superior, overarching grand narrative, these truths must be maintained for institutions and groups to adjudicate between conflicting stories/interpretations.2What Joyce Appleby and her coauthors see as postmodern resistance to the idea of facticity thus does not meet the rationality condition for incorporating other narratives into the discipline of history.

Nevertheless, Telling the Truth About History is important for demonstrating the continuing relevance of the two questions about crafting and their connections to public life. So long as the two conditions can be met, "minority histories" can change the discourse of the discipline without having to practice any principle of permanent revolution. Successfully incorporated "minority histories" are like yesterday's revolutionaries become today's gentlemen. Their success helps routinize innovation.

The Semantics of Subordination

The debate about minority histories, however, allows for another understanding of the expression "minority." Minority and majority are, after all, no natural entities; they are constructions. The popular meaning of the words "majority" and "minority" are statistical. But the semantic fields of the words contain another idea: of being a "minor" or a "major" figure in a given context. For example, the Europeans, numerically speaking, are and have been a minority in the total pool of humanity today, yet their histories asserted norms that every other human society should aspire to, and compared Europeans to others who were still "minors" for whom they, the "adults" of the world, had to take charge. So numerical advantage by itself is no guarantor of a major/majority status. The problem of "minority histories" thus leads us to the question of which constructions and experiences of the past stay "minor" in the sense that their very incorporation into historical narratives converts them into pasts "of lesser importance" vis-à-vis dominant understandings of what constitutes fact and evidence (and hence vis-à-vis the underlying principle of rationality itself) in the practices of professional history. Such "minor" pasts are those experiences of the past which have to be always inferiorized as they are translated back into the historian's language, that is to say, as they are translated back into the phenomenal world the historian—as a historian, that is, in his or her professional capacity—inhabits. These are pasts which are treated, to use Kant's expression from his essay "What Is Enlightenment?", as instances of "immaturity" on the part of the historical agent, pasts which do not prepare us for either democracy or citizenly practices because they are not based on the deployment of reason in public life.3

Let me call these histories subordinated or "subaltern" pasts. They are not marginalized because anyone consciously intends to marginalize them but because they represent moments or points at which the very archive that the historian of a (marginalized) group mines in order to bring the history of that group into a relationship with a larger narrative (of class or of the nation, for instance), develops a degree of intractability with respect to the very aims of professional history. In other words, these are pasts that resist historicization just as there may be moments in ethnographic research that resist the doing of ethnography.

"Subaltern pasts," in my sense of the term, do not belong exclusively to socially subordinate or subaltern groups, nor to "minority" identities alone. Elite and dominant groups can also have subaltern pasts to the extent that they participate in subordinated life-worlds. Being a historian, however, I argue from a particular instance of it. My example comes from Subaltern Studies, the group with which I am associated, and from an essay by the founder of the group, Ranajit Guha.

Subaltern Studies is a series of publications in Indian history that was begun under the general editorship of Ranajit Guha in the early 1980s. Its explicit aim was to write the subaltern classes into the history of nationalism and the nation and to combat all elitist biases in the writing of history. These original intellectual ambitions were political; but they did not necessarily come from the lives of the subaltern classes themselves. Looking back, however, I see the problem of "subaltern pasts" dogging the enterprise of Subaltern Studies from the very outset: indeed it is arguable that what differentiates the Subaltern Studies project from the older tradition of "history from below" is the self-critical awareness of this problem in the writings of the historians associated with this group.

What Is (Historical) Truth?

Let me explain this with the help of Ranajit Guha's brilliant essay, "The Prose of Counter-Insurgency," published in an early volume of Subaltern Studies. A principal aim of Guha's essay is to use the Santal rebellion of 1855 in order to make the insurgent peasant's consciousness the mainstay of a narrative about rebellion. In pursuing the history of the Santal rebellion of 1855—the Santals are a "tribal" group inhabiting large areas of what is today Bengal and Bihar—Guha comes across statements by peasant leaders which explain the rebellion in "supernatural" terms, as an act carried out at the behest of the Santal god "Thakur." Guha himself draws our attention to the evidence and underscores how important this understanding was to the rebels themselves. Quoting statements made by the leaders of the rebellion, Sidhu and Kanu, to military interrogators wherein they explained their own actions as flowing from instructions they had received from their god (Thakur) who had also assured them that British bullets would not harm the devotee-rebels, Guha takes care to avoid any instrumental or elitist reading of these statements. He writes:

These were not public pronouncements meant to impress their followers . . . these were words of captives facing execution. Addressed to hostile interrogators in military encampments they could have little use as propaganda. Uttered by men of a tribe which, according to all accounts had not yet learnt to lie, these represented the truth and nothing but the truth for their speakers.4

A tension inherent in the project of Subaltern Studies becomes palpable here. Guha's idea of a truth which is only "truth for their speakers," is an act of taking critical distance from that which he is trying to understand. Taken literally, the rebel peasant's statement shows the subaltern himself disclaiming agency or subjecthood in action. "I rebelled," he says, "because Thakur made an appearance and told me to rebel." The subaltern is thus not necessarily the subject of his or her history but in the history of Subaltern Studies or in any democratically-minded history, she or he is. What does it then mean when we both take the subaltern's views seriously—the subaltern ascribe the agency for their rebellion to some god—and want to confer on the subaltern agency or subjecthood in their own history, a status the subaltern's statement denies?

Guha's strategy for negotiating this dilemma unfolds in the following manner. His first move, against liberal or standard Marxist historiography, is to resist analyses that see religion simply as the nonrational expression of a secular-rational nonreligious entity, relationship (such as class, power, or economy), or consciousness. But in spite of his desire to listen to the rebel voice seriously, Guha cannot take it seriously enough, for there is no principle in an "event" involving the divine or the supernatural that can give us a narrative strategy that is rationally defensible in the modern understanding of what constitutes public life. The Santal's own understanding does not directly serve the cause of democracy or citizenship or socialism. It needs to be reinterpreted. Clearly, in the narrative of the rebels, the event (the rebellion) was not secular; in our language, it included the supernatural. The supernatural was part of what constituted public life for the nonmodern Santals of the 19th century. This, however, simply cannot be the past in the language of professional history in which the idea of historical evidence cannot admit of the supernatural except as part of the nonrational (that is, somebody's belief system). Fundamentally, the Santal's statement that God was the main instigator of the rebellion has to be anthropologized (that is, converted into somebody's belief or practice) before it finds a place in the historian's narrative. Guha's position with respect to the Santal's own understanding of the event becomes a combination of the anthropologist's politeness—"I respect your beliefs but they are not mine"—and a Marxist (or modern) sense of frustration with the intrusion of the supernatural into public life. "[I]n sum," Guha writes, "it is not possible to speak of insurgency in this case except as a religious consciousness," and yet hastens to add, "except that is, as a massive demonstration of self-estrangement (to borrow Marx's term for the very essence of religiosity) which made the rebel look upon their project as predicated on a will other than their own."5

Here is a case of what I have called "subaltern pasts," pasts that cannot ever enter history as belonging to the historian's own position. These days one can devise strategies of multivocal histories in which we hear subaltern voices more clearly than we or Guha did in the early phase of Subaltern Studies. One may even refrain from assimilating these different voices to any one voice and deliberately leave loose ends in one's narrative (as does Shahid Amin in his Event, Metaphor, Memory).6 But the historian as historian, unlike the Santal, cannot invoke the supernatural in explaining/describing an event.

The Past Has Many Voices

In other words, the act of championing "minority histories" has resulted in discoveries of subaltern pasts, constructions of historicity that help us see the limits to the mode of viewing embodied in the practices of the discipline of history. Why? Because the discipline of history—as has been argued by many (from Greg Dening to David Cohen in recent times)—is only one particular way of remembering the past.7 The resistance that the "historical evidence" offers in Guha's essay to the historian's reading of the past—a Santal god, Thakur, stands between the democratic-Marxist historian and the Santals in the matter of deciding who is the subject of history—is what produces "minor" or "subaltern" pasts in the very process of the weaving of modern historical narratives. Subaltern pasts are like stubborn knots that break up the otherwise evenly woven surface of the fabric. Between the insistence of the Subaltern Studies historian that the Santals are the agents of their own action and the Santals' insistence that it was to their god Thakur that such sovereignty belonged, remains a hiatus separating two radically different experiences of historicity, a hiatus that cannot be bridged by an exercise that converts, however understandably from the point of view of the historian, the Santals' statements as evidence for anthropology. When we do "minority histories" within the democratic project of including all groups and peoples within mainstream history, we both hear and then anthropologize the subaltern. We treat their beliefs as just that, "their beliefs." We cannot write history from within those beliefs. We thus produce "good," not subversive, histories. However, historians of indigenous peoples throughout the world have reminded us that the so-called societies without histories—the object of contempt for European philosophers of history in the 19th century—cannot be thought of as societies without memories. They remember their pasts differently. Why must one privilege the ways in which the discipline of history authorizes its knowledge?8

This suggests that the kind of disciplinary consensus around the historian's methods that was once represented by "theory" or "methods" courses which routinely dished out Collingwood or Carr or Bloch as staple for historians has now broken down. This does not necessarily mean methodological anarchy or that Collingwood et al., have become irrelevant; but it does mean that E. H. Carr's question "What Is History?" needs to be asked again for our own times. The pressure of pluralism inherent in the languages and moves of minority histories has resulted in methodological and epistemological questioning of what the very business of writing history is all about. And the question of including "minorities" in the history of the nation has turned out to be a much more complex problem than a simple operation of applying some already-settled methods to a new set of archives and adding the results to the existing collective wisdom of historiography. This additive approach to knowledge has broken down. What has become an open question is: Are there experiences of the past that cannot be captured by the methods of the discipline or which at least show the limits of the discipline? Fears that such questioning will spread like a dark death-inducing disease through Historyland, seem extreme, for the discipline is still securely tied to the positivist impulses of modern bureaucracies, judiciaries, and to other instruments of governmentality. Minority histories, if they are going to be about inserting hitherto neglected identities into the game of social justice, must also be "good," and not subversive, histories, for history here speaks to forms of representative democracy and social justice that liberalism or Marxism have already made familiar.

But minority histories can do more than that. The task of producing "minority" histories has, under the pressure precisely of a deepening demand for democracy, become a double task. "Good" minority history is about expanding the scope of social justice and representative democracy, but the talk about the "limits of history," on the other hand, is about struggling for non-statist forms of democracy that we cannot yet either completely understand or envisage. This is so because in the mode of being attentive to the "minority" of subaltern pasts, we stay with heterogeneities without seeking to reduce them to any overarching principle that speaks for an already-given whole. There is no third voice which can assimilate into itself the two different voices of Guha and the Santal leader, we have to stay with both, and with the gap between them that signals an irreducible plurality in our own experiences of historicity.

We can—and we do usually in writing history—treat the Santal of the 19th century to doses of historicism and anthropology. We can, in other words, treat the Santals as signifiers of other times and societies. This gesture maintains a subject-object relationship between the historian and the evidence. In this gesture, the past remains genuinely dead; the historian brings it "alive" by the telling of the story. But the Santal statement "I did as my god told me to do" also faces us as a way of being in this world, and we could ask ourselves: Is that way of being a possibility for our own lives and for what we define as our present? Does the Santal help us to understand a principle by which we also live in certain instances? This question does not historicize or anthropologize the Santal, for the illustrative power of the Santal as an example of a present possibility does not depend on the particular period or society from which the illustration is drawn. In this mode of understanding the Santal stands as our contemporary and the subject-object relationship that normally defines the historian's relationship to the archives is dissolved in this gesture. To stay with the heterogeneity of the moment when Guha the historian meets with the Santal the peasant, is then to stay with the difference between these two gestures: one, that of historicizing the Santal in the interest of a history of social justice and democracy, and the other, that of refusing to historicize and of seeing the Santal instead as a figure throwing light on a possibility for the present. When seen as the latter, the Santal puts us in touch with the heterogeneities, the plural ways of being, that make up our own present. The archives thus help bring to view the disjointed nature of our own times. That is the function of subaltern pasts.

Attending to this heterogeneity could take many different forms. Some scholars now perform the limits of history by fictionalizing the past, by experimenting to see how films and history might intersect in the new discipline of cultural studies, by studying memory rather than just history, by playing around with forms of writing, and by similar other means. Such experiments are welcome; but the fact that there are subaltern pasts, unassimilable to the secular narratives of the historian, allows us to see the complex understanding of time—treated as invisible in most historian's writing—that must underlie and indeed make possible the secular chronology of historical narratives without which there cannot be any historical explanation.

The Presence of the Past

The broad statement that the Santal had a past in which events could belong to the order of the supernatural does not appear as something completely beyond our own experience. Why? Because the principle is not completely strange to us. We have a pretheoretical, everyday understanding of it precisely because the supernatural or the divine, as principles, have not disappeared from the practices of the modern. We are not the same as the 19th-century Santal. One could even easily assume that the Santal today would be very different from what they were in the 19th century, that they would inhabit a very different set of social circumstances. The modern Santal would have the benefit of secular education and may even produce their own professional historians. No one would deny these historical changes. Nor does one have to reduce the 19th-century Santal to the one statement quoted here. But astrological columns in the newspapers, the practices of "superstition" that surround the lives and activities of sports fans, for example, and all the deliberately "cultic" expressions of religiosity that have never gone away, go to show that we are all, in principle, capable of participating in supernatural events and the sense of the past they help create. The 19th-century Santal—and indeed, humans from any other period and regions—are thus in a peculiar way our contemporaries; that would have to be the condition under which they become intelligible to us. Thus the writing of history must implicitly assume a plurality of times existing together, a discontinuity of the present with itself. Making visible this discontinuity is what "subaltern pasts" allow us to do.

An argument such as this is actually at the heart of modern historiography itself. One could argue, for instance, that the writing of "medieval history" for Europe depends on this assumed contemporaneity of the medieval, or what is the same thing, the noncontemporaneity of the present with itself. The medieval in Europe is often strongly associated with the supernatural and the magical. But what makes the historicizing of it at all possible is the fact that its basic characteristics are not completely foreign to us as moderns. Historians of medieval Europe do not always consciously or explicitly make this point but it is not difficult to see this operating as an assumption in their method.

The double move of historicizing the medieval and of seeing it at the same time as contemporary with the present can be seen in the work of Jacques Le Goff. Seeking to explain an aspect of the European-medieval, Le Goff writes:

People today, even those who consult seers and fortune-tellers, call spirits to floating tables, or participate in black masses, recognize a frontier between the visible and the invisible, the natural and the supernatural. This was not true of medieval man. Not only was the visible for him merely the trace of the invisible; the supernatural overflowed into daily life at every turn.9

On the surface the passage is about what separates the medieval from the modern. Yet this difference is what makes the medieval an ever present possibility that haunts the practices of the modern. The people who consult seers today are modern in spite of themselves, for they engage in "medieval" practices but are not able to overcome the habits of the modern. Le Goff rescues the present by saying that even in the practice of these people, something irreducibly modern lingers—their distinction between the visible and the invisible. But it lingers only as a border.

Subaltern Pasts and Disciplinary Limits

Subaltern pasts are signposts of this border. With them we reach the limits of the discourse of history. The reason for this is that subaltern pasts do not give the historian any principle of narration that can be rationally defended in modern public life. Going a step further, one can see that this requirement for a rational principle, in turn, marks the deep connections that exist between modern constructions of public life and projects of social justice. That is why a Marxist scholar like Fredric Jameson begins his book The Political Unconscious with the injunction: "Always historicize!" "This slogan," writes Jameson, "the one absolute and and we may even say "transhistorical" imperative of all dialectical thought—will unsurprisingly turn out to be the moral of The Political Unconscious as well."10 Historicizing is not then the problematic part of the injunction; the troubling term is "always." For the assumption of a continuous, homogeneous, infinitely stretched-out time which makes possible the imagination of an "always," is put to question by subaltern pasts that makes the present, as Derrida says, "out of joint," noncontinuous with itself.11

One historicizes only insofar as one belongs to a mode of being in the world that is aligned with the principle of "disenchantment of the universe" which underlies knowledge in the social sciences (and I distinguish knowledge from practices). It is not accidental that a Marxist would exhort us to "always historicize," for historicizing is tied to the search for justice in public life. This is why one welcomes "minority histories," be they of ethnic groups, gay-rights activists, or of subaltern social classes. Here the historical discipline enriches itself by incorporating these histories but its very methodological dominations create what I have called subaltern pasts.

For the "disenchantment of the world" is not the only principle by which we world the earth. There are other modes of being in the world—and they are not necessarily private, the superstitious acts of sports fans, for example, being often public. The supernatural can inhabit the world in these other modes and not always as a problem or result of conscious belief or ideas; the supernatural or the divine can be brought into presence by our practices. These modes are not without questions of power or justice but these questions are raised—to the extent modern public institutions allow them—on terms other than those of the political-modern.

However, the relation between what I have called "subaltern pasts" and the practice of historicizing is not one of mutual exclusion. It is because we always already have experience of that which makes the present noncontemporaneous with itself that we can actually historicize. Thus what allows medievalist historians to historicize the medieval or the ancient is precisely the fact that these worlds are never completely lost. It is because we live in "time-knots" that we can undertake the exercise of straightening out some part of the knot (which is what a chronology is).12 Subaltern pasts—aspects of these time-knots—thus act as a supplement to the historian's pasts and in fact aid our capacity to historicize. They are supplementary in a Derridean sense—they enable history, the discipline, to be what it is and yet at the same time help to show forth what its limits are. But in calling attention to the limits of historicizing, they help us distance ourselves from the imperious instincts of the discipline—the idea (of Haldane's or Jameson's for example) that everything can be historicized or that one should always historicize—and returns us to a sense of the limited good that modern historical consciousness is. Gadamar once put the point well in the course of discussing Heidegger's philosophy. He said: "The experience of history, which we ourselves have, is . . . covered only to a small degree by that which we would name historical consciousness."13 Subaltern pasts persistently remind us of the truth of this statement.

An earlier version of this essay was presented at the 1997 annual meeting of the American Historical Association. The author acknowledges with gratitude the criticisms and comments received there. A subsequent presentation at the University of Colorado at Boulder at the invitation of copanelist Patricia Limerick also benefited from the criticisms of those present. He thanks Sandria Freitag for her original invitation prodding him to think through this topic, and Anne Hardgrove, Uday Mehta, Faisal Devji, Homi Bhabha and David Lloyd for more recent conversations on the subject.

Notes

1. Joyce Appleby, Lynn Hunt, and Margaret Jacob, Telling the Truth About History (New York: Norton, 1994).

2. Cf. Georg G. Iggers, Historiography in the Twentieth Century: From Scientific Objectivity to the Postmodern Challenge (Hanover and London: Wesleyan Univ. Press, 1997), 145: "Peter Novick has in my opinion rightly maintained that objectivity is unattainable in history; the historian can hope for nothing more than plausibility. But plausibility obviously rests not on the arbitrary invention of an historical account but involves rational strategies of determining what in fact is plausible." Emphasis added.

3. Immanuel Kant, "An Answer to the Question: What is Enlightenment?" in Immanuel Kant, Perpetual Peace and Other Essays, trans. Ted Humphrey (Indianapolis: Hackett, 1983), 41–8.

4. Kant, 80.

5. Kant, 78.

6. Shahid Amin's book, Event, Metaphor, Memory: Chauri Chaura, 1922–1992 (Berkeley: Univ. of California Press, 1995), is an excellent illustration of the points being made in this essay and shows a self-conscious appreciation of the problems of translating the subaltern's voice into a modern political project (without giving up the socially necessary attempt to translate).

7. See Greg Dening's essay, "The Poetics of History" in his Performances (Chicago: Univ. of Chicago Press, 1995) and David Cohen's The Combing of History (Chicago: Univ. of Chicago Press, 1994). Ashis Nandy's essay, "Themes of State, History, and Exile in South Asian Politics," in Emergences, 7/8 (1995–96), 104–125, makes comparable points.

8. A sensitivity to the question of alternative pasts is increasingly becoming visible in the work of many historians: my own very personal and random list of such scholars would include Saurabh Dube, Carolyn Hamilton, Christopher Healy, Patricia Limerick, Iain McCalman, Stephen Muecke, Klaus Neuman, Sumathi Ramaswamy, and Ajay Skaria.

9. Jacques Le Goff ed., The Medieval World, trans. Lydia G. Cochrane (London: Collins and Brown,1990), 28–9.

10. Preface to his The Political Unconscious: Narrative as a Socially Symbolic Act (Ithaca: Cornell Univ. Press), 9.

11. See Jacques Derrida, Specters of Marx: The State of the Debt, the Work of Mourning, and the New International, trans. Peggy Kamuf (New York and London: Routledge, 1994).

12. I owe the conception of time-knots to Ranajit Guha.

13. Hans-Georg Gadamar, "Kant and the Hermeneutical Turn" in his Heidegger's Ways, trans. John W. Stanley (Albany: SUNY Press, 1994), 58. Emphasis in the original.