Beyond Turk and Hindu Copyright 2000 by David Gilmartin and Bruce Lawrence. This work is licensed under a modified Creative Commons Attribution-Non- commercial-No Derivative Works 3.0 Unported License. To view a copy of this license, visit http://creativecommons.org/licenses/by-nc-nd/ 3.0/. You are free to electronically copy, distribute, and transmit this work if you attribute authorship. However, all printing rights are re- served by the University Press of Florida (http://www.upf.com). Please con- tact UPF for information about how to obtain copies of the work for print distribution. You must attribute the work in the manner specified by the author or licensor (but not in any way that suggests that they en- dorse you or your use of the work). For any reuse or distribution, you must make clear to others the license terms of this work. Any of the above conditions can be waived if you get permission from the Uni- versity Press of Florida. Nothing in this license impairs or restricts the author's moral rights. Florida A&M University, Tallahassee Florida Atlantic University, Boca Raton Florida Gulf Coast University, Ft. Myers Florida International University, Miami Florida State University, Tallahassee University of Central Florida, Orlando University of Florida, Gainesville University of North Florida, Jacksonville University of South Florida, Tampa University of West Florida, Pensacola   Beyond Turk and Hindu Rethinking Religious Identities in Islamicate South Asia Edited by David Gilmartin and Bruce B. Lawrence University Press of Florida Gainesville . Tallahassee . Tampa - Boca Raton Pensacola - Orlando . Miami . Jacksonville . Ft. Myers  Copyright 2000 by David Gilmartin and Bruce Lawrence Printed in the United States of America on acid-free paper All rights reserved 05 04 03 02 01 00 6 5 4 3 2 1 Library of Congress Cataloging-in-Publication Data Beyond Turk and Hindu: rethinking religious identities in Islamicate South India / edited by David Gilmartin and Bruce Lawrence. p. cm. Includes bibliographical references and index. ISBN 0-8130-1781-5 (cloth: alk. paper) 1. Islam-Relations-Hinduism. 2. Hinduism-Relations-Islam. 3. India-Ethnic relations. 4. Ethnicity-India. 5. India-Religion. I. Gilmartin, David. II. Lawrence, Bruce. BP173.H5 B47 2000 297.2'845'0954-dc21 00-055977 The University Press of Florida is the scholarly publishing agency for the State University System of Florida, comprising Florida A&M University, Florida Atlantic University, Florida Gulf Coast University, Florida International University, Florida State University, University of Central Florida, University of Florida, University of North Florida, University of South Florida, and University of West Florida. University Press of Florida 15 Northwest 15th Street Gainesville, FL 32611-2079 http://www.upf.com  Contents List of Figures, Maps, and Tables vii Acknowledgments ix Introduction 1 David Gilmartin and Bruce B. Lawrence Section 1: Literary Genres, Architectural Forms, and Identities 1. Alternate Structures of Authority: Satya Pir on the Frontiers of Bengal 21 Tony K. Stewart 2. Beyond Turk and Hindu: Crossing the Boundaries in Indo-Muslim Romance 55 Christopher Shackle 3. Religious Vocabulary and Regional Identity: A Study of the Tamil Cirappuranam 74 Vasudha Narayanan 4. Admiring the Works of the Ancients: The Ellora Temples as Viewed by Indo-Muslim Authors 98 Carl W. Ernst 5. Mapping Hindu-Muslim Identities through the Architecture of Shahjahanabad and Jaipur 121 Catherine B. Asher Section 2: Sufism, Biographies, and Religious Dissent 6. Indo-Persian Tazkiras as Memorative Communications 149 Marcia K. Hermansen and Bruce B. Lawrence 7. The "Naqshbandi Reaction" Reconsidered 176 David W. Damrel 8. Real Men and False Men at the Court of Akbar: The Majalis of Shaykh Mustafa Gujarati 199 Derryl N. MacLean  Section 3: The State, Patronage, and Political Order 9. Shari'a and Governance in the Indo-Islamic Context 216 Muzaffar Alam 10. Temple Desecration and Indo-Muslim States 246 Richard M. Eaton 11. The Story of Prataparudra: Hindu Historiography on the Deccan Frontier 282 Cynthia Talbot 12. Harihara, Bukka, and the Sultan: The Delhi Sultanate in the Political Imagination of Vijayanagara 300 Phillip B. Wagoner 13. Maratha Patronage of Muslim Institutions in Burhanpur and Khandesh 327 Stewart Gordon Glossary 339 List of Contributors 345 Index 347  Figures, Maps, and Tables Figures 5.1 Jami Mosque, Amber 124 5.2 Fakhr al-Masajid from street, Shahjahanabad, Delhi 125 5.3 Shivalaya of Dhumi Mal Khanna, Katra Nil, Shahjahanabad, Delhi 128 5.4 Chandnee Chauk, Delhi 130 5.5 Entrance to the Sri Brijnandji Temple, Jaipur 131 5.6 Mosque of Maulana Zia al-Din Sahib, Jaipur 133 5.7 Interior, mosque at Dargah Zia al-Din Sahib, Jaipur 134 5.8 Interior courtyard of the haveli-style Ladliji temple, Katra Nil, Shahjahanabad, Delhi 137 Maps 10.1 Temple desecrations, 1192-1394 272 10.2 Temple desecrations, 1394-1600 273 10.3 Temple desecrations, 1600-1760 274 Tables 10.1 Instances of temple desecration, 1192-1760 274a 12.1 Narrative structure of Vijayanagara's founding 310a   Acknowledgments This volume originated with a conference held under the auspices of the Rockefeller Foundation. Without the encouragement and support of Lynn Szwaja of the Arts and Humanities Division, we would not have pursued our application to the foundation. It was a grant to the Triangle South Asia Consortium that made possible a three-year Rockefeller Residency Institute on South Asian Islam and the Greater Muslim World. The conference was held at Duke University in April 1995, and most of the papers included in this volume were originally presented on that occasion, although some have been significantly revised to suit the scope and theme of the current volume. Other scholars presented papers or made comments at the conference that could not be included in this volume. We would like to give special thanks to the following confer- ence participants: Simon Digby, Eleanor Zelliott, James Laine, Sanjay Subrahmanyam, Anisuzzaman, Philip Lutgendorf, Daniel Ehnbom, Carol Salomon, Gregory Kozlowski, and Tazim Kassam. John Richards of Duke University played a central role in conceptual- izing and organizing the 1995 conference. From 1994 to 1998, he served with Bruce Lawrence as codirector of the Rockefeller Residency Insti- tute, and he was instrumental in the early stages of this volume's prepa- ration. We would also like to thank both Katherine Ewing and Munis Faruqui, who not only participated in the conference but also helped to ensure its success. Among those who gave time and energy to planning the conference, Constance Blackmore of the Comparative Area Studies Program at Duke University deserves special mention. In preparing these papers for publication, we are indebted to the out- side reviewers from University Press of Florida. They read and com- mented on the initial draft in painstaking detail. It was due to their in- sight that we reorganized the essays and also reduced their number. Nor could we have produced a readable volume without the skill of John Caldwell in the initial scanning and editing of these essays and the labor of Rob Rozehnal in formatting the final version of essays to con- form to University Press of Florida guidelines. Patient in all aspects of  x I Acknowledgments drafting, copying, and preparing the manuscript for publication was Lillian Spiller, administrative assistant to the Department of Religion at Duke University. We acknowledge their collective assistance, even while absolving them of responsibility for the content of this manu- script. The editors are the sole custodians of blame, while all the above are to be thanked for making possible a volume that is at once unique in content and challenging to received wisdom about South Asia. Note on Transliteration A volume including papers drawn from such a wide variety of linguistic sources as this one presents unusual problems of transliteration. Since many words of Islamicate origin appear in variant forms in Indic lan- guages, we decided that it would violate the spirit of the volume to at- tempt to impose any standardized system of transliteration. We have thus left it to individual authors to use whatever system suited them. Some have employed a full range of diacritics while others have not. We have attempted to note in the glossary some of the more common vari- ant spellings that appear in the volume.  Introduction David Gilmartin and Bruce B. Lawrence Muslims have been an integral part of South Asia for over a thousand years. Why then is it so hard to define them as "indigenous"? Why are they not as seamlessly Indian as Sikhs, who have been there for less time, or as Jains, who have been there longer but in fewer numbers? Part of the difficulty lies with the stress on Islam and Hinduism as religious worldviews. Not only are Islam and Hinduism seen as alterna- tive belief systems but they are deemed competitive and irreconcilable in their differences. Here is how a prominent Egyptian jurist, Judge Muhammad al-Ashmawy, whose major life's work has been both to chart Islamic humanism and to demonstrate how compatible Islam is with Judaism and Christianity, describes Muslim-Hindu interaction: "On the Indian subcontinent the relationship between the Hindu (the majority population) and the Muslim (the minority population) forms a dark and disturbed chapter in the history of interreligious relationships. This is partially due to the great differences between Islam and Hindu- ism." He then goes on to enumerate in terms of belief, whether it is belief in a deity (Muslims do, Hindus do not) or belief in the next life (Muslims think it is judgment at the end of each life, while Hindus opt for reincar- nation).1 Judge Muhammad al-Ashmawy is not wrong to cite differences be- tween Islam and Hinduism, whether as worldviews or as belief systems. The differences he cites do exist, and they are not unimportant, espe- cially if one thinks of religion as, above all, belief or ritual. Yet religion also includes everyday life and social exchange; it elides with what is sometimes called "culture," and from the viewpoint of religion as cul- ture Judge al-Ashmawy has overweighted differences in belief as deter- minative of all other patterns of exchange between Muslims and Hin- dus. Still more serious is his presumption that Muslims and Hindus have always and everywhere been fixed as oppositional groups, each pitted  2 | Introduction irreconcilably against the other. The actual history of religious exchange suggests that there have never been clearly fixed groups, one labeled "Hindu," the other-both its opposite and its rival-labeled "Muslim." To open up the space between reductive religious orientations and mobile collective identities, one needs a new vocabulary that is not re- stricted to modern connotations of words such as Muslim and Hindu. It was to remedy the inadequacy of English popular usage that historian Marshall G. S. Hodgson coined the term Islamicate. For Hodgson, the neologism Islamicate allowed students of civilizational change to refer to the broad expanse of Africa and Asia that was influenced by Muslim rulers but not restricted to the practice of Islam as a religion.2 It is for the same reason, to suggest the breadth of premodern South Asian norms beyond Hindu doctrine or practice, that we employ the term Indic in the essays that follow. Both Islamicate and Indic suggest a repertoire of lan- guage and behavior, knowledge and power, that define broad cosmolo- gies of human existence. Neither denotes simply bounded groups self- defined as Muslim or Hindu. The goal of the contributors to this volume is thus contrarian: they do not accept popular notions, even those espoused by major and influen- tial world figures, that invoke identity as set, unchanging, and ex- clusive. Instead, the contributors have tried to understand within the frames of Indic and Islamicate norms those discrete processes of identity formation that shaped religious identities in precolonial South Asia. The aim is to move beyond a fixation with bounded categories, whether re- ligious or ethnic, Hindu or Turk, in order to pluralize the ways that these categories operated in varying historical contexts. While our goal is contrarian, we do not ignore common sense: both editors and authors recognize the pervasive importance attached to re- ligious systems that can be defined, pursued, and separated as Islamic and Hindu. Yet we vigorously contend that there is a larger point to make, namely, that the constant interplay and overlap between Islami- cate and Indic worldviews may be at least as pervasive as the Muslim- Hindu conflicts that Judge al-Ashmawy and others take to be symptom- atic of all life in the subcontinent. It is because the distinction between Islamicate and Muslim, Indic and Hindu, has been repeatedly obscured that not only South Asians but also scholarship on South Asians have been mired in controversy. If all history is present-minded, as Croce long ago asserted, then the histories of India and Pakistan are excessively so. South Asians of varying politi- cal persuasions have long searched for genealogies of modern identity  David Gilmartin and Bruce Lawrence | 3 that could be authenticated by being extended back to the precolonial era. The founder of Pakistan, Muhammad 'Ali Jinnah, thought that there had always been two nations in the subcontinent. He saw them sepa- rated primarily by cultural practices, which he took to be fixed markers between two distinct groups. In Jinnah's view, these markers justified the creation of a bounded nation: Pakistan. More recently, visionaries of India as a Hindu state have followed a line of reasoning similar to Jinnah's. They have seen every instance of Muslim political rule or mili- tary victory or architectural creation as evidence of a long struggle be- tween fixed Muslim and Hindu groups. Not local Indian rulers but Hindu norms were defeated in the period from the Ghaznavids to the Mughals, and they were defeated not by certain Muslim rulers but by Islam itself. Some historians of religion have noted the degree to which practi- tioners of their own craft have been implicated in popular quests for genealogies of religious identity. Some have suggested that nineteenth- century British categorization of major religions as global or world reli- gions played a formative role in facilitating the imagination in South Asia of fixed religious communities extending backward in time, each universal in scope and exclusive in loyalty. Tony Stewart, for example, describes in this volume how even many fair-minded historians have often read fixed religious categories back into history laden with mod- ern valences. Even when the categories palpably do not fit the evidence, scholars are often reluctant to jettison them, opting instead to suggest the existence of hybrid or syncretic forms, defined by the mixing of "irreconcilable" religions, or by the lack of those attributes that are thought to be essential to a given world religion. This is not to suggest, of course, that there is any unbridgeable divide between the operation of religious identities in modern and precolonial settings. While terms such as religion and nation are, in their common contemporary usage, laden with modern assumptions, the processes of identity formation described in this volume are ones that could be found in all historical periods. Terms of identity are inevitably shaped by the larger frames of knowledge in which they are embedded. While these frames of meaning differed in precolonial South Asia from those pro- vided by modern science, capitalism, and colonialism, the processes by which identities were forged were nevertheless strikingly similar. It is thus in the interaction between the particular and the general that we must embed the analysis of identity. While most contributors simply presume that identities are constructed in particular historical circum-  4 | Introduction stances, it is, above all, the particular meanings attached to the catego- ries Hindu and Muslim that must be understood in relation to the his- torical circumstances in which they existed. Such circumstances are both local and universal. They include the full range of other context-specific interests with which particular identities interacted as well as the larger Islamicate and Indic contexts that framed all categories of identity. Section 1: Literary Genres, Architectural Forms, and Identities In the first section of this book, literary genres and architectural forms are the topic of investigation. Contributors examine the texts and con- structed remains that scholars have used to determine the nature and scope of religious identities. We are confronted not only with texts but also with relationships between texts, and the recurrent desideratum is to understand the link behind a specific text and its multiple contexts. The initial three essays focus on the rhetorical strategies by which identities were created. The analysis of texts exposes processes of iden- tity construction that were at once complex and nuanced, for texts were not simply windows on identities but keys to the process by which iden- tities were generated. Most critically, an analysis of rhetoric suggests the critical interplay of difference and sameness in the construction of all identities. The construction of difference inevitably involved the simul- taneous construction of sameness, for difference could only be asserted in opposition to an "other" of like category. One thus finds categories of comparison closely allied to categories of opposition, with both being shaped by the forms and structures of textual presentation. As these essays show, the manipulation of genre was often as critical to processes of identity formation as were the precise labels of identity that authors used. Indeed, the interplay of genre and language is critical to the analysis of identity as it emerges in these essays. Those scholars who have stud- ied the vocabulary of religious identity in Indian texts have often found complex invocations of oppositional categories and meanings. As Carl Ernst has shown elsewhere,3 Arabic and Persian use of the term Hindu had a range of meanings that changed over time, sometimes denoting an ethnic or geographic referent without religious content. By the same to- ken, Indic texts referring to the invaders from the northwest used a va- riety of terms in different contexts, including yavanas, mleccchas, farangis, musalmans, and Turks. Such terms sometimes carried a strong negative  David Gilmartin and Bruce Lawrence | 5 connotation, but they rarely denoted a distinct religious community conceived in opposition to Hindus. In and of themselves, however, such terms tell us little. To understand the usage of these terms, one must move beyond the terminology it- self-beyond Turk and Hindu-to analyze the framing categories and generic contexts within which these terms are used. While genre itself is often elusive, it is this which links processes of identity formation in their local contexts to the wider worlds of knowledge that framed them and gave them meaning. Genre expectations are embedded in univer- salist frames of knowledge, yet they also provide a vehicle for express- ing the most particularistic identities. Uncovering the ambivalent, bidirectional force of genre in shaping identities is, in fact, one of the most significant contributions of this col- lection. The importance of linking the global to the local, and the local to the global, has been widely observed in the modern world but less de- veloped in the analysis of identities in precolonial South Asia. Consider the case of modern national identities. As many scholars have observed, national identity rests at once on an assertion of the irreducible differ- ences among nations and also on a recognition of the common participa- tion of all nations in an international system of order and knowledge. Nationalism could not function without a simultaneous expression of sameness and difference, at once powerful and ineffable. Even the most chauvinistic expressions of national distinctiveness imply-indeed, rely on-the commonality inherent in the generic category "nation." While this global framework of internationalism did not operate in the same way in precolonial South Asia, it was nevertheless the case that identities in Islamicate India were also constructed in relation both to particularistic categories and to larger framing systems of knowledge and order. Frames were at once Indic and Islamicate, defining the pa- rameters of the world of genres in which identities emerged. Essays in this section focus on elements of both these frames and on the interplay between genre and meaning, in both texts and architecture. The power of genre in exploring, and exploding, identity is evident in Tony Stewart's inquiry into Bengali literature in chapter 1. Stewart fo- cuses on the stories of Satya Pir, a mythical holy man of Bengal with both Hindu and Muslim identity markers. Some scholars have tried to make of Satya Pir a syncretic figure, since he was appealed to by Hindus and Muslims alike. But Stewart rejects the explanatory value of syncretism, since it reads modern definitions of religious identity into the past. In-  6 | Introduction stead, Stewart moves beyond the terminology of identity in order to call attention to ways that the narratives of Satya Pir themselves revealed distinctive identities. Identities, in Stewart's view, arise here not from distinct traits associated with differing groups but rather from the differ- ent "orientations to power" that mark the narrative structures of differ- ent types of stories. Stewart analyzes two distinct groups of stories. Each group illustrates a distinctive orientation to power, embodied in a distinctive "narrative code." Indeed, Stewart's approach suggests how identity is rooted in the simultaneous play of commonality and difference linking these groups of stories. The stories are alike in that they evoke Satya Pir as a figure of power capable of assisting the search for wealth and prosperity. All the stories are a product of the social and political pressures that redefined Bengal as a frontier society in the centuries following Mughal expansion eastward. And yet within this common frame, the two groups of stories are structurally quite distinct. One group drew on Vaisnava terminology and appropriated Satya Pir as an avatar of Visnu. The other group drew on Sufi terminology and offered Satya Pir as a figure demanding recog- nition from the common people. On one level, these structural differ- ences can be read as defining contrasting orientations toward power, orientations that can perhaps be linked loosely to terms such as Hindu and Muslim or, more narrowly, Vaisnava and Sufi. But such identities were inescapably framed by commonalities as well as differences. Both were products of a frontier society. As Stewart demonstrates, it is the interplay between narrative structure, on the one hand, invoking differ- ent notions about the operation of divine power, and common coping with the everyday world of Mughal Bengal, on the other, that empow- ered the Satya Pir narratives. The importance of genre as a framing context for a nuanced under- standing of terms and their relationship to identities is also highlighted by Christopher Shackle in chapter 2. He explores the qissa, or romance, among the most popular genres in the Punjab. Punjabi qissas drew heavily on both Islamicate and Indic images. Yet the major significance of this genre derived from its ability to tap into the tension between localized boundary markers and civilizational frames. Indeed, Shackle's analysis suggests the ways that the terminology of local social bound- aries-although critical to the qissa-was given a distinctive meaning through the framing structure of the qissa as a literary genre. Shackle quotes a verse from the Sufi poet, Bullhe Shah (1680-1758):  David Gilmartin and Bruce Lawrence | 7 Neither Arab am I nor man of Lahore Nor Indian from the town of Nagaur Neither Hindu am I nor Turk of Peshawar. The language here depicts a world where fixed identities mattered. Yet, as Shackle suggests, the very structure of the qissa, as Bullhe Shah's lines indicate, involved transcending such divisions. The best-known Per- sianate romances, in fact, commended the transgression of "almost ev- ery conceivable kind of social as well as psychic boundary," including those of tribe and status, "social and religious, sexual and spiritual standing." External boundary markers, such as Turk and Hindu, are not only invoked but are also necessary to the structure of these texts-but only to suggest that the real business of identity involves moving be- yond them. The main vehicle for boundary transgression in the qissa, of course, was love, which drew its power-and its symbolic meaning-from its links to the universal values of larger civilizational traditions. This was indicated by the lovers' inventories, which often opened such romances, inventories linking each romance to the larger Islamicate civilizational canon. But each qissa's appeal rested also on the localization of its set- ting, its creation of an imminent world replete with local actors and con- straints, hopes and fears, dalliance and delight. What we find here is the palpable juxtaposition of an idealized world of love, which knows no bounds, and the inescapable, multiply bounded everyday world of Punjabi society. Without the tension between these two poles, the qissa loses its bivalent appeal, but through the constant reiteration of this ten- sion it suggests the limited scope of either a purely universal or a nar- rowly local identity. To be real, to experience love, an individual's iden- tity must always be open to transgression. Viewed in its generic context, the language of identity thus takes on meanings in the Punjabi romance that are far more fluid than those implied by the fixed language of group definition, although this itself was critical to the genre. Similar processes are at work in the text described by Vasudha Narayanan in chapter 3. Her regional focus is Tamil Nadu, in South In- dia. Her text, the Cirappuranam, is a seventeenth- century poem in praise of Muhammad, composed by a Tamil Muslim. Since the Prophet is its subject, it belongs to an explicitly Muslim devotional genre, the sira, or life of the Prophet. But it is also firmly embedded in an Indic devotional genre, the purana. In the Cirappuranam, we thus confront a poem that  8 | Introduction manipulates genre to position the Prophet simultaneously within two worlds. Far more directly than in the texts analyzed by Stewart and Shackle, the Cirappuranam uses genre to undercut any notion of clear Hindu and Muslim boundaries. As in Stewart's analysis, it is not syncretism at work here but rather a bivalent process. The generic, elite literary conventions analyzed by Narayanan are at once resilient and adaptable; although Indic in origin, they translate the sira, a life of Muhammad, into a famil- iar Tamil cultural world. Here Tamil literary convention transcends any attempt to define clearly bounded Hindu and Muslim identities. Processes of identity formation can be traced, of course, not only in literary texts but also in material production. Indeed, few identity mark- ers have maintained a stronger hold on the imaginations of historians than the religious buildings-mosques, temples, shrines-that dot India's landscape. It was the destruction of the Baburi Masjid in 1992 that perversely revealed the multiple symbolic meanings attached to religious buildings in contemporary India. Yet the multivalent meaning of such structures must also be historically traced, through accounts of their construction and destruction, if we are to move beyond Turk and Hindu in looking at processes of identity formation in premodern South Asia. Carl Ernst reminds us at the outset of chapter 4 that such buildings are defined not just by their ritual use or iconic content but also by their historical location and political deployment. Ernst's essay focuses on Rafi ad-din Shirazi, an official of the Bijapur Sultanate. A prolific author, Shirazi included in his 1612 Persian history of Bijapur a description of the temples at Ellora. As Ernst notes, modern analysts of Ellora have often used the frame of religious identities to define Ellora's distinctive Hindu, Buddhist, and Jain features, seeing them as constitutive of the monument's significance. Yet Shirazi argues that the proper cultural frame for understanding structures like those at Ellora was not the frame of distinctive religious art forms, much less competing truth claims at the core of juxtaposed religions. Rather, in his view, the proper frame for understanding the Ellora complex was the competition for glory between kings, whether Persian or Indian. Ernst's essay shows how a Muslim official could at- tribute to Ellora a cultural meaning shaped by his view of dynastic ri- valry. Difference here was not primarily defined by opposing religions but by opposing polities, in this case the polities of royal monarchs com- peting for a greater historical legacy. As Ernst notes, Shirazi saw Indian  David Gilmartin and Bruce Lawrence | 9 and Persian monarchs as kindred spirits in their common quest for "rec- ognition through monumental art." Shirazi's interpretation thus under- scores both the role of the larger Islamicate framework in shaping gen- eric understandings of identity and also the contextual demands placed on Shirazi by his own position as a Muslim official in a state whose major edifice is an Indic temple complex. It suggests that in matters of material culture, as in the interpretation of literary texts, the determination of cultural identity is a fluid process, depending on both point of view and generic context. To understand the ground level, operative relationship between reli- gious structures and group identity in premodern North India requires, of course, a different sort of approach to material evidence. Catherine Asher provides just such an approach in her analysis of the role of reli- gious buildings within the context of urban space in North Indian cities. Asher argues that in North Indian cities, whether built under Muslim patronage (Shahjahanabad) or Hindu patronage (Jaipur), mosques and temples tended to occupy very different types of spaces. Mosques were visible public structures, while temples tended to be deliberately hidden and obscured. As Asher speculates, placement within urban space may well have defined the distinctive meanings associated with mosques and temples as foci for religious identity. Mosques and temples, although sharing a generic commonality as religious buildings, reflected in their placement and structure differing "orientations to power," just as did the contrast- ing narratives of Satya Pir analyzed by Tony Stewart. Mosques pro- claimed identity through their public presence in India's urban spaces, while temples offered a place to nurture the gods-and also maintain "one's identity among one's own community members"-in a space set off from the public realm. They thus occupied distinct, though not nec- essarily commensurate, places within a common urban framework. As Asher notes, historians need to do far more research into the Islamicate urban context, especially into the nature of urban patronage, if they are to understand how to interpret such religious structures as definers of collective identities. Section 2: Sufism, Biographies, and Religious Dissent The essays in section 2 move beyond texts and contexts to explore in more detail the operation of the Islamicate frame of social and intellec- tual ordering in India. However critical specific texts and contexts are to  10 | Introduction the understanding of the production of identities, it is critical also to understand the ways that broad civilizational traditions operated in In- dia, not as the foundation for generic identities but rather as frames shaping the articulation of, and the meanings attached to, more particu- lar identities. It has been one of the outcomes of modern structures of knowledge that civilizations have often been seen as fixed sources of bounded identity and culture. Yet the dynamic of identity construction that emerges from our analysis is one in which civilizations should prob- ably be seen more as frames shaping the language and meanings within which more particularized identities operate while allowing enormous flexibility to local actors, conditions, and contexts. As we noted earlier, it was Hodgson who first coined the term Islam- icate. He coined it over thirty years ago in order to suggest a structure or frame of moral reference that characterized the span of the Afro-Eur- asian world in which Muslims were major agents of exchange and con- trol. Islamicate denoted the moral values and cultural forms that spread through the world system of Muslim trade and power in the centuries following the rise of Islamic polities. Hodgson distinguished Islamicate from that which was strictly Islamic or Muslim, relating to the practice of Islam as a religion, whether through creedal, ritual, or juridical loyalty. Although Muslims did not make this distinction-they had no need to-the distinction between Islamicate and Islamic/Muslim is ex- tremely useful for us-moderns, or perhaps postmoderns, that we be. It is the term Islamicate which captures the civilizational dynamic for the framing of religious identities in India, including those of Muslims, that the authors in these sections attempt to capture. Particularly powerful in Islamicate India were the normative con- cepts of authority drawn from the larger Perso-Islamic tradition, analy- sis of which is at the heart of the essays in this section. But the Perso- Islamic tradition did not operate independently of more local and particular forms of identity. For most people, neither particularistic identities nor civilizational ones could be fully conceptualized without the other. As Christopher Shackle's essay suggested, the tension be- tween civilizational ideals and the realities of everyday divisions was often encapsulated in the structuring of regional Islamicate literary genres, such as the Punjabi qissa. But here the tension between these becomes the focus for a larger discussion of the rhetoric of religious identity. As the essays in this section suggest, it was the interplay be- tween the universal and the everyday-and the tensions it generated-  David Gilmartin and Bruce Lawrence | 11 that produced in Islamicate India the image both of civilizational identi- ties and of a world in which competitive particularities were central. Perhaps the most telling model for understanding this embrace of tensions within the Islamicate framework emerges from institutional Sufism. In chapter 6, Marcia Hermansen and Bruce Lawrence deal with the particular role of tazkiras, biographical literature on Sufi shaikhs, written by Indo-Persian elites. They project a model that helps us to understand the larger world in which Islamicate identities operated. Central to Indo-Persian culture, Sufism embraced many of the tensions that helped to define the structuring of Indo-Muslim identities more broadly. For Hermansen and Lawrence, two elements are especially determi- native: the relationship between personal authority and place in the structuring of identities, and the relationship between the distinctly In- dian and transregional Islamic frameworks. As Hermansen and Law- rence show, whatever its metaphysical foundations, Mughal Sufism was rooted in devotion to specific persons and tied to concrete places. Tazkiras were written to glorify and to legitimize distinctive spiritual genealogies. The ties of Sufi authors to particular places and regions informed the way in which they invoked larger frameworks of legiti- mation. The tazkira, as an Islamicate genre, laid claim to Muslim space in South Asia by inscribing on the subcontinent new spiritual and intel- lectual centers, largely through memorializing Sufis. Sufism, in this sense, provided a model for larger processes of Islam- icate ordering and identity formation. Sufism defined a language of identity and authority linked to hierarchy, charismatic genealogy, and the distribution of baraka (blessing), a language that, by extension, served, as Richard Eaton argues in chapter 10, to justify political author- ity as well. Yet there was no consensus on the precise contours of a global Sufi model of authority. Tazkira writers appealed to Sufism as an overarching Islamicate model for the operation of charisma and the le- gitimation of authority, yet their aim in doing so was to establish com- petitively the particular precedence and distinctiveness of their own orders, genealogies, and places. Such was the aim of dynastic political leaders as well. What Hermansen and Lawrence thus suggest is a critical but seldom noted paradox intrinsic to the Islamicate context itself. Par- ticipation in local status competition, entailing the assertion of narrow, parochial identities, was inextricably intertwined with participation in the larger structures of Islamicate ordering.  12 | Introduction This paradox can also give us insight into the possible meanings of the category "Indian" within the framework of Islamicate organization. As David Damrel indicates in chapter 7, scholars have often made sense of the reformist rhetoric of Indo-Persian elites, such as the Sufi mystic Shaykh Ahmad Sirhindi, by positing a fundamental opposition between Islamic and Indian as civilizational categories. The reformism of Sir- hindi's so-called Naqshbandi reaction is thus seen as a product of the unsustainable cultural tensions created by the attempts of rulers such as Akbar, or by Chishti Sufi masters, to "reconcile" Islam to the Indian en- vironment. But as Damrel argues, such an approach to Sirhindi's "re- formism" hardly fits with the evidence. The attempts by previous schol- ars to create an opposition between "syncretic" Sufi orders (such as the Chishti) and "purifying" Sufi orders (such as Sirhindi's suborder of the Naqshbandiya) were predicated on the problematic notion that Hindu- ism and Islam actually existed as pure civilizational essences. Tony Stewart challenges that assumption, and Damrel shows that Sirhindi's project was far from essentialist. The master bridge builder of a new suborder, Sirhindi was influenced by both Chishtis and Naqshbandis. However much he may have made use of "purifying" reformist rheto- ric, he produced in the end a distinctively "Indian" Sufi position. Sirhindi's concerns-about Sufi dhikr, prophetology, government em- ployment of non-Muslims, and many other issues-all were framed by major normative debates within the larger Islamicate world. And yet Sirhindi's particular position was also informed by a set of affiliations- at once distinctive and oppositional-that shaped competition internal to Indian Sufism. His efforts thus dramatized the extent to which par- ticularizing and universalizing identities fed off the other, creating ten- sion and debate, to be sure, but also ensuring an expanded form of Is- lamicate-Indic identity that would have been unimaginable without such competition. The crucial role of the rhetoric of corruption, purity, and reform in shaping identities is further explored, in a non-Sufi context, by Derryl MacLean. In chapter 8, MacLean analyzes the discourse of the debates at the Mughal court involving a prominent member of the millenarian Mahdavi movement in the late sixteenth century. Shaykh Mustafa Gujarati accused the Mughal ulama of moral impotence: they had failed to be "real men" by their pursuit of worldly advantage. Here, as in many of the later debates pursued by Sirhindi, competition for legitimacy was framed by an appeal to larger civilizational ideals, which were rhetori-  David Gilmartin and Bruce Lawrence | 13 cally configured in stark opposition to the corrupting influence of local and worldly interests. Yet the Mahdavi shaykh avoided making this appeal in terms of op- positions between Islam and Hinduism or even between Arab and In- dian. He did not want to compete with the ulama on their own turf. Instead, he deployed a rhetorical strategy at once bold and unexpected in its universalistic reach: He invoked gender norms. Accusing the ulama of lacking a "masculine" commitment to truth (evident in their slavish adherence not only to the political influence of the Mughal em- peror but also to the opinions of the ulama of the Hijaz), Shaykh Mustafa claimed a universalist cachet for his own Mahdavi identity, which was uncontaminated by such worldly connections. Only he and his follow- ers, by inference, could be considered "true men." In using such a rhe- torical ploy, he was able to claim a pure identity, closely aligned with an Islamicate language, without denying the distinctively "Indian" roots of his identity. Once again, what is critical to Shaykh Mustafa, as to Sirhindi in a later generation, is the dynamic between the universal and the par- ticular: Muslim protagonists asserted their own competitive identities and sought to undermine the identities of others, but always within the framing assumptions of Islamicate discourse. In sum, the chapters of this section indicate ways that conflicts for local status and influence often generated powerful images of civili- zational identity. This is not to suggest, of course, that the meanings attached to the universal were stable in Islamicate India. Islamicate rhetorics of identity were multiple. Nor did such frames operate inde- pendent of framing Indic idioms of identity. But at the heart of the analy- sis here is the way that the rhetoric of purity and reform, and of local competition for status, shaped an image of overarching civilizational identity. This was an image whose meaning and immediacy were in- separable from the reality of local particularized conflict and competi- tion. Section 3: The State, Patronage, and Political Order A central arena for this conflict and competition was provided by the institutions of the Islamicate state. One reason for this, of course, is that states were vital sources of the cultural patronage so critical to the gen- eration of identities. Equally important, however, the state stood at the nexus between the universal and the particular, between the legitimiz-  14 | Introduction ing language of civilizational allegiance and local structures of power and social ordering. The critical tensions shaping the operation of Muslim-ruled states in the Indian Islamicate context are suggested in essays by Muzaffar Alam and Richard Eaton. Alam focuses on the role of shari'a, or Islamic law, and its theoretical role in Islamicate political order in India. Symbolic deference to shari'a was among the most powerful markers of a Muslim ruler's claim to standing within the larger Islamic world. But, as Alam argues, in pragmatic terms, the concept of shari'a took on varying politi- cal meanings in India. Using evidence from Ziya ud-din Barani in the thirteenth century, he shows how, for some Muslim political theorists, shari'a took on a narrow juristic meaning, rooting the stability of Mus- lim community in the narrow adherence to Islamic law as interpreted by the ulama. But he contrasts this with another interpretation of the shari'a, strongly influenced by Persian akhlaq literature (and through it by Greco-Hellenic ideas and, perhaps, by Mongol political practice), which saw as the aim of shari'a the maintenance of proper order in the community at large by balancing the interests of differing groups and communities, including religious communities (and allowing, in this context, for considerable freedom of worship). This second vision ex- erted a powerful influence on the structure of the Mughal empire, he argues, shaping shari'a as a symbol linking the state to an international Islamic order and yet defining a structure of rule built on a recognition of the complex structures of division and difference that ordered Indian society. These visions are evident also in Eaton's analysis of the role of Mus- lim states in that centerpiece of modern identity polemics: the Muslim destruction of Hindu temples. For modern polemicists, past destruction of religious structures has provided grist for invoking abstracted reli- gious identities in order to wage modern-day warfare against "infidel" others. Yet as Eaton shows in chapter 10, Muslim temple destruction reveals patterns that are defiantly complex. His analysis operates on two levels. First, he draws the critical distinction between the rhetoric of temple destruction and its actual practice. For some Muslim sultans (as for some modern-day Hindu nationalists), an image of Muslim rulers as iconoclasts served legitimizing purposes, even when pragmatism dic- tated quite different policies. Stories of temple destruction, whether in Hindu sources or Muslim, thus require careful questioning. Second, and more fundamental to the argument, Eaton insists that even documented acts of temple desecration must be firmly grounded not in a narrative of  David Gilmartin and Bruce Lawrence | 15 religious competition but in the political history of Muslim state build- ing. Attacks on religious structures, he argues, gained significance in a world where the construction and maintenance of religious edifices were central to political legitimacy, for Muslim and Hindu rulers alike. Those Hindu temples that were attacked were thus attacked almost in- variably not as generic cultic sites but as symbols of Hindu royal author- ity. Desecration thus represented a critical symbolic act in delegitima- ting a rival sovereign in order to incorporate his territory into one's own realm. Eaton's essay underscores, once again, the pragmatic nature of Islam- icate rule in India. Both Hindu and Muslim places of worship shared a common value and vulnerability as focal points of spiritual devotion and political rule. Eaton's interpretation harks back to Ernst's essay on Shirazi's reading of the Ellora temple complex as above all a tribute to kingly ambition. Given this framework, however, Eaton must also ex- plain why Muslim rulers did not destroy in the same way the mosques and shrines patronized by their defeated Muslim rivals. Here he turns to a more essentialist argument, namely, that royal Hindu temples and Muslim mosques and shrines did embody in their different forms dis- tinct visions of the relationship between the divine and the political. Eaton's contention is thus not that differences between Hindu and Mus- lim are meaningless, even in matters of political legitimation. The prominence of symbolic religious acts in the rhetoric of dynastic legiti- mation makes this clear. Rather, his argument is that an examination of the norms and practices of Islamicate kingship provides a critical frame- work for the historical interpretation of these acts. The importance of this framework is evident also in the analysis of the final three chapters, which examine practices relating to Indic kingship in the era following the spread of Muslim power in India. If temple de- struction was, in certain circumstances, an important act of kingly legiti- mation for Hindu and Muslim kings alike, so was the patronage of his- torical chronicles legitimating dynastic authority as Talbot and Wagoner show. In chapter 11, Cynthia Talbot focuses on a sixteenth-century South Indian chronicle, the Prataparudra Caritramu, which deals with the last of the Kakatiya kings of Warangal. Written long after the end of the Kakatiya dynasty, the text is one intended, as Talbot sees it, to legitimate Telugu warrior influence within the framework of the new political or- der of the Vijayanagara kingdom in the early sixteenth century. Strikingly, the symbolic evocation of opposition between Hindu and Muslim plays little role in this text. As in Eaton's argument, this is not  16 | Introduction because the mobilization of this dichotomy was of no potential legiti- mizing importance but because, in the particular historical circum- stances in which the Prataparudra Caritramu was produced, the mobili- zation of such rhetoric served little purpose. Since the period was one in which alliances among Hindu and Muslim states in South India were as important as their oppositions, such religious oppositions held little appeal in legitimizing authority. This was hardly the case in all warrior chronicles of the period, as Talbot clearly notes. In some contexts, the language of resistance to Muslim "demons" could serve powerful legiti- mizing purposes, as it did, for example, among the warriors of Rajasthan whose epics of resistance to Muslim domination were ana- lyzed decades ago by Aziz Ahmad as part of a cycle of challenge and response, "epic and counter-epic." By comparing this text with others in its genre, Talbot underscores the critical role of context in framing lan- guage of identity-creation. As in Eaton's argument, the pragmatic needs of power defined the way that identity was constructed. The Prata- parudra Caritramu used a past Telugu (and Hindu) "golden age" to try to create a usable Telugu warrior identity within the increasingly Islami- cate framework of the Vijayanagara state. The meaning of the broader Islamicate context for the Vijayanagara state is explored more fully by Phillip Wagoner in chapter 12. Like the Maratha state (which is the subject of the final chapter, by Stewart Gor- don), the Vijayanagara state has sometimes been presented in the litera- ture as a "champion" of Hindu identity-the reviver and protector of "Hindu kingship" in the face of Muslim domination. However, as Wag- oner shows, Vijayanagara narratives of legitimation cannot be under- stood without also understanding the state's Islamicate context. Wag- oner examines sixteenth- and seventeenth-century narratives of the heroic fourteenth-century founders of the kingdom, Harihara and Bukka, and their relationship to the Delhi Sultanate. While standard modern narratives have seen Harihara and Bukka as apostates, convert- ing to Islam to serve the Delhi Sultanate before apostatizing to create a Hindu kingdom, Wagoner discerns in these Sanskrit narratives of Vijay- anagara's founding a very different story. Certainly, competition with Muslim states plays a significant role in the story of the founding of Vijayanagara. But critical to the story of Harihara and Bukka is their portrayal also as successors to the authority of the Delhi Sultanate, thus defining the Sultanate itself as one of the foundational, legitimizing sources of Vijayanagara's dynastic power. This was no accident, Wag- oner argues, since by the time these narratives were written, Vijayana-  David Gilmartin and Bruce Lawrence | 17 gara's competitive claims to regional power were rooted in its participa- tion in an Islamicate state system, which was reflected in myriad ways in the structure of the state itself. Its claims to power were thus critically framed by the need of the Vijayanagara kings to legitimate themselves both within an Islamicate order of states and with reference to an Indic system of legitimation. Nowhere was this captured more dramatically than in the appellation sometimes applied to the Vijayanagara king him- self, the "sultan among Hindu kings." Stewart Gordon's essay provides further analysis of such simulta- neous frames of legitimation through an examination of the Maratha state that arose during the declining years of Mughal power. Gordon, however, provides a different perspective on this phenomenon-a per- spective drawn not primarily from chronicles but from local eighteenth- century Maratha revenue documents that provide a window on the local flow of patronage within the Maratha system. While Maratha chronicles sometimes portrayed Shivaji as an inveterate foe of the Mus- lims and a champion of Hindu dharma, Gordon finds powerful Islami- cate structures of governance embedded in the most local operations of the Maratha state. Like the Mughals and the Deccan Sultanates before them, the Marathas' power depended heavily on their relations with local warrior elites, and like the Mughals and the Deccan sultans, they structured these relations around the reciprocal exchange of local, bu- reaucratically recorded revenue rights in return for state service. This structure of exchange provided the framework within which Maratha warrior families had long competed for honor under earlier Islamicate states, and it had now come to provide a critical frame for the validation of their local power, competition, and identity under the Marathas as well. It also provided the frame within which the Marathas, like Islam- icate states before them, patronized both Hindu and Muslim holy men, embedding into local society the institutions that supported their au- thority and power. Here, then, we have evidence of the power of an Islamicate framework both in shaping state legitimacy and in framing a more particularized structure of local identities in an Indic state. Most important in linking Indic and Islamicate framework was an ideology of "universal kingship," which stressed the importance of the maintenance of order and prosperity both as a dharmic duty and as the central legitimating function of kingship. This was a framework rooted in the pragmatic ability to maintain order and prosperity by balancing the interests of all groups, whatever their particular identities. Such a vision of kingship could, of course, be couched in either Indic or Islam-  18 | Introduction icate terms-and Gordon suggests how the competing frames of legiti- mation for the state sometimes framed the struggle for influence and power among competing elites deploying different sets of legitimizing terminology. But it nevertheless provided an integrating vision of state legitimacy that focused authority on the person of the ruler. Indeed, as all the chapters in this final section suggest, a focus on the integrating authority of the ruler is important for understanding the structuring of identities more generally. Individual religious differences between Muslims and Hindus (as between other generic religious cat- egories, like Saiva and Vaisnava, Sunni and Shi'a) were framed by their operation within a pervasive structure of personalized religious author- ity-a structure that, along with its bureaucratic technologies, defined the Islamicate state itself. This is not to say that marks of generic Hindu or Muslim identity were insignificant. But since religious virtue and spiritual power were embodied preeminently in holy individuals, reli- gious identity was defined primarily in relation to individual teachers, masters, or Sufi exemplars. The structure of Sufism (and of Hindu reli- gious and devotional lineages) represented, in some respects, an inte- grative cultural reflection of the assumptions about power ordering the larger Islamicate system. As markers of group identity or allegiance, the categories Hindu and Muslim were thus largely subsumed in more par- ticularistic structures of devotion. And yet networks of individual loy- alty and devotion were rarely constituted without at least some refer- ence to the legitimizing language of authority provided by these larger framing categories. Conclusion Two large conclusions thus emerge from these essays-conclusions which take us back to the problems in the academic study of religious identities in South Asia with which we began. First, focusing on the contexts that produced articulations of identity is critical to historicizing the vocabulary of religious identity and under- standing how it may have changed over time. As the essays in the last section have suggested, the meanings of identities can best be under- stood in relation to the operation of power. Seeing how identities relate to the structure of the state and to its networks of patronage is critical to understanding how identities gained meaning. But as several of the es- says have suggested, the authority of Islamicate states was itself embed-  David Gilmartin and Bruce Lawrence | 19 ded in larger frameworks of knowledge, and these frameworks defined the nature of power and its relationship to divine authority. In short, we need to historicize Islamicate identities and to see how they changed in response to India's changing position in larger cultural and economic worlds. We can then see these changes in relation to the changes that marked the succeeding colonial period as well. As many scholars have argued, British colonial rule was itself linked to a distinc- tive structure of knowledge, tied to science and capitalism. This framed identities in new ways, introducing into South Asian parlance, for ex- ample, the language of enumeration, of ethnic groups as territorially mapped entities, and of religions as fixed communities, susceptible to counting under the census. This in turn was linked to a new language of "majorities" and "minorities," which were then presumed to be the con- stitutive units of South Asian politics. All of this, of course, helped to define and to legitimize-even as it was in turn shaped by-the political practice of the colonial state. But if much of this was new, it nevertheless grew out of a structural relationship between state political practice, structures of knowledge, and the framing of identities that had marked the operation of identities in premodern Islamicate South Asia. Focusing on structural frameworks for identity thus helps us to escape a dichoto- mous view of the "modern" and the "premodern," and instead to see how structures of identity had long shifted in response to the shifting place of South Asia-and of forms of state authority-within a larger world. Second, and perhaps even more important, is the need to retain a vigilant eye on process, above all, the process of identity formation, for it is only through this process that we can see how identities were con- structed not simply through the opposition and juxtaposition of fixed categories but also through the tensions generated by the simultaneous deployment of framing categories of commonality and the assertion of particularities of difference. Many of the essays in the volume have highlighted this process of identity formation, both through an analysis of political structures and through the analysis of texts. The structuring tensions of many texts-tensions between generic identity and particu- larity, between commonality and difference-mirrored in critical ways the tensions shaping the larger political order. Critical to the structuring of the Islamicate world, then, has been at- tention to religion, but religion with movable parts and multiple forms. Hindus and Muslims alike experienced tension between universal ide-  20 I Introduction als and the modeling of the way the world actually worked. More often than not, they perceived such tension as a welcome venue for address- ing more complex ways of being both Indian and transregional in out- look. To grasp this process is to move beyond fixed identities such as "Turk" and "Hindu" in looking at premodern societies, and also their successors, in the subcontinent. Notes 1. Muhammad al-Ashmawy, Against Islamic Extremism (Gainesville: Univer- sity Press of Florida, 1998), 82. 2. Marshall G. S. Hodgson, The Venture of Islam: Conscience and History in a World Civilization (Chicago: University of Chicago Press, 1974), 57-60. 3. Carl Ernst, Eternal Garden: Mysticism, History, and Sufism in a South Asian Center (Albany: State University of New York Press, 1992), 22-37.  1 Alternate Structures of Authority Satya Pir on the Frontiers of Bengal Tony K. Stewart Interpreting Satya Pir Problems of Categorizing Inhabitants of the northeastern section of the subcontinent (the territo- ries of contemporary West Bengal, Orissa, and Bangladesh) have turned to the religious figure of Satya Pir to protect against the vagaries of ex- tended travel, to ensure the general weal of one's family, and perhaps most frequently, simply to get rich. The prominence of this latter feature led one public performer in the 1920s to satirize Satya Pir as "Lord of the Bazaar," the purveyor of dime-store religion.1 Yet Satya Pir, whose ru- bric embraces all forms of the somewhat older figure of Satya Narayana, numbers among his followers the populations of nearly all of the ethnic and religious communities of the region. Today he is most prominent among the middle and lower classes of both Hindus and Muslims. He has more Bengali texts dedicated to the telling of his stories than any other premodern mythic or historical figure, save the Vaisnava leader Krna Caitanya (1486-1533). Even though he has been the subject of cycles of popularity similar to other religious figures and to gods and goddesses in Bengal, his overall popularity seems to have grown steadily, perhaps peaking in the mid-nineteenth century and early twentieth century, but still finding new adherents today. Yet Satya Pir has attracted little scholarly attention.2 Part of the reason for this lies precisely in the fact that he is a figure who blurs the lines  22 | Satya Pir on the Frontiers of Bengal between Hindu and Muslim as religious categories. His historical ap- peal to both religious traditions is embodied in his name: satya is the Sanskrit and Bengali word for the "true" or "truth," while pir derives from the Persian, designating the Muslim spiritual guide who is re- nowned for wisdom and the ability to translate spiritual achievement into a practical power to aid supplicants. When scholars do address Satya Pir, he and the worship he engenders are generally damned to bear the label of "syncretism."3 Syncretism seldom deals directly with its object. It relies on metaphor to make its point, comparing the thing in question to some other entity that is impermanent, the most popular metaphors being organic (such as a hybrid or half-breed), alchemical (such as mixture or solution), or construction (bricoleur). The meta- phoric structure of that concept inevitably implies that no syncretistic entity is viable in its own right, for it combines elements that retain their identifier as discrete and mutually exclusive-in this case the categories of Hindu and Muslim-and because of that unholy alliance, it is artifi- cially created and destined not to endure. Yet for about five centuries endure is precisely what Satya Pir has done, even though he has cer- tainly been subject to cycles of popularity similar to other religious figures and to gods and goddesses in Bengal. If anything, his overall popularity seems to have grown steadily, perhaps peaking in the mid- nineteenth century and early twentieth century, but still finding new adherents today. The proliferation of manuscripts and printed books, the widespread familiarity with his image and tales, and the develop- ment recently and for the first time of permanent places of commemora- tion, including the establishment of his dargah-all the empirical evi- dence of his success-point to a scholarly failure to find an adequate explanation for him and what he means to his followers. It has become a truism of our contemporary scholarship to recognize that what we consider important is shaped, if not driven, by complex ideological concerns. The constructions of South Asian religious history and literature of the last century or so have frequently sought to read modern religious identities back into the histories of the subcontinent to generate seamless ideal histories.4 Ironically, in these many and complex and often subtly nuanced constructions, Hindu and Muslim all too fre- quently become monolithic in conception, and in this imposed unifor- mity they are assumed to be antithetical by nature. Satya Pir, however, violates that purity of conception and thus falls outside the structure of this idealized religious history. Further, with his appeal concentrated largely in the lower social strata, and with his activities frequently fo-  Tony K. Stewart | 23 cused on the pragmatic generation of wealth (especially as a prerequi- site for morality), his study has been viewed by many as having little connection to the "higher aims" of the religious "Great Traditions" in Bengal. This has further marginalized the study of Satya Pir, consigning him to the label "folk deity" and thus not legitimately part of the "proper" or high traditions. His study, however, may help us to gain new perspectives on the problems we face in constructing our religious categories. Rather than assume that Satya Pir represents a composite and therefore "unnatural" entity made from bits and pieces of two separate traditions-the syncre- tistic approach-perhaps we might more fruitfully ask how it is that people, whom we as scholars routinely designate by the terms Hindu and Muslim, can claim this religious figure without the overt conflict that one might predict based on contemporary political rhetoric. How is it possible that these individuals, no matter their label, can perform Satya Pir's vows and take refuge in his protection without running afoul of the theological positions and ritual injunctions internal to each of the Great Traditions designated by those very labels of Muslim and Hindu? To start, I would like to propose that the purveyors and consumers of the Satya Pir literature are not initially acting as members of either group. The common concerns framing the invocation of Satya Pir were not those defined by membership in a religious group of any sort but rather those defined by the context of life in early modern Bengal. If we are to understand the stories of Satya Pir, we must begin not with timeless religious categories but with context. What is important about Satya Pir, religiously and culturally (those obviously are not exclusive either), is that he deals with pragmatic concerns of survival-not overt ideology, theology, or ritual; people accept that he wields a power to make their lives better and that is good no matter how it is labeled. Put another way, the questions of pragmatic power cross whatever imaginary divide we construct between Hindu and Muslim. To enjoy the benefits of this gen- eral weal does not require group participation to be valid; most of Satya Pir's worship is, in fact, individual and ad hoc. To turn to Satya Pir is a matter of opportunity and convenience, not one that requires constant reminders of commitment (so even being a member of an imagined group of Satya Pir devotees seems to be limited to the time of actual invocation). If we can resist comparing the action described in the nar- ratives of Satya Pir-and the accompanying ritual, which is generally simple and unmediated-to the ideal standards of an Islam or Hindu- ism imagined in their pristine monologic purity and ideal praxis, we  24 | Satya Pir on the Frontiers of Bengal have then an opportunity to circumvent the tyranny of group inclusion as the dominant organizer of experience and the primary marker of identity. Yet, at the same time, the stories of Satya Pir do not suggest the com- plete irrelevancy of received religious categories. Rather, interpreting the stories requires that we carefully rethink their usage. Despite certain similarities, stories of Satya Pir fall into clearly differentiated groups, reflecting differing vocabularies, narrative styles, and orientations to- ward divine and worldly power. As we shall see, some stories (which we might loosely label as "Hindu") see Satya Pir as yet another incarnation of Visnu, especially suited to the disintegrating times of the Kali Age wherein dharma is at extreme risk. Another group (which we might loosely label "Muslim") portray him as but a pir, albeit a special one who resides in an ethereal Mecca and who can be conjured with a heartfelt call of his name; sometimes he is vaguely associated with the historical pir Husayn ibn Mansur al-Hallaj (d. 922) as the "True" Pir,5 but more often simply as a figure of local power, one pir among a group, including Gaji Pir, Manik Pir, Pir Badar, and Khizr Pir. Different genres of stories thus define in their narrative styles different orientations to power- orientations to power with links to the vocabularies of broader religious traditions. Yet the figure of Satya Pir himself provides the common focus that frames these differing orientations, and the different genres of sto- ries suggest not so much identification with different groups as they suggest differing visions of hierarchies of power within a common world. The stories illustrate how the same pragmatic (worldly) power can be mediated through different hierarchies of authority, and the pre- cise role of Satya Pir will vary depending on how the structure of author- ity is conceptualized and presented. Far from suggesting the existence of clearly distinct Hindu and Muslim groups, whose elements are com- bined in a "syncretic" cult, an analysis of the stories of Satya Pir thus suggests how the interplay of vocabulary and genre define the common concerns linking all the followers of Satya Pir, as a single figure, in the struggle to deal with worldly power, even as such an analysis shows, simultaneously, how narrative structure reveals the different orienta- tions to power characterizing the stories' diverse audiences. Frontier Narratives: Religious and Literary Typologies Satya Pir cannot be fixed historically in any temporal or geographic lo- cale-and in this sense he is mythic, so his "history" is not so much his but that of his followers' acceptance of him. For nearly five centuries,  Tony K. Stewart | 25 perhaps longer, he has rated as one of the most popular pirs of Bengal, and his legacy is captured in a corpus that is geographically dispersed throughout the region. His literature begins to emerge with regularity in the late sixteenth century, following the first known works by Phakir Rama, Ghanarama Cakravarti, Ramesvara, and Ayodhyarama Kavi.6 A number of seventeenth- and eighteenth-century texts relate new ex- ploits, but it is in the period of easy access to inexpensive printing and the concomitant creation of great entrepreneurial fortunes, the mid- nineteenth to early twentieth centuries, that his literature burgeons into one of the most prolific in Bengali. It is then that he moves into the rap- idly expanding metropolitan areas of Calcutta and Dhaka. With the po- litical realignments of the twentieth century, his popularity has shifted and today clings to the metropolitan fringe, while resurging in more rural areas that one might characterize as the new frontier of develop- ment, especially where population densities are still very much con- strained by geophysical barriers, such as the mangrove swamps of the Sunderband and the mountain fringes ringing Bangladesh from Chit- tagong in the southeast to Rangpur in the north. In these shifts, Satya Pir has retained on the surface his apparent dual character, which is reflected in his physical appearance. Krnahari Dasa describes it: He wears the dress of a fakir, the hair on his head the color of mud, the Prophet's patched scarf cinched at his neck. His lotus body shimmers brilliantly, Four times greater than a full moon perched above clouds thick with rain. The sacred thread drapes his shoulder, a chain belt hangs at his waist, in his hands tremble one's aspirations. A short string of anklets jingle in time with his dagger's clink to each clopping step of his wooden sandals.7 The appearance of this pir becomes an explicit visual metaphor in the way he combines key marks of a public Muslim and Hindu allegiance. It is not unusual for Satya Pir to approach significant religious figures in either community while quoting from the Qur'an and Bhagavata Purana. Deliberately conflating signs that would ordinarily be disjunc- tive endlessly amuses or annoys characters in the narratives-a clear  26 | Satya Pir on the Frontiers of Bengal indication that the authors deliberately count on the effect and play on these symbolic currencies. This play has a very serious side, for the nar- rative strategy of conflation serves to create momentary confusions among the characters that predictably elicit spontaneous, unreflective responses of ridicule and invective. These outbreaks create an opening for Satya Pir to instruct the naive in a way that is all the more compel- ling by virtue of the extreme situation he manipulates by playing on their prejudices, hubris, and ignorance to demonstrate their inappropriateness to the business of living. And to the delight of the listener or reader, he is not above resorting to more brutal magical persuasions to make his point. The content of these biting homilies varies dramatically, depending on the author's proclivity, for the narratives are anything but uniform in this regard. But apart from these occasional and short opportunities to lecture or preach, most of Satya Pir's message emerges through the reso- lution of predictable dramatic situations, and these stories account for nearly all of the written material that exists; there is no formal theology. The textual materials for glorifying Satya Pir are, then, almost exclu- sively literary narratives, ranging from the sophisticated poetic produc- tions of the royal courts of the eighteenth century to more rustic oral performances designed to be improvised and delivered by itinerant bards. While the authors limit overt and explicit theological speculation to occasional summary statements inserted extradiegetically into the narrative frame, they limit the inclusion of ritual materials even more. What ritual instructions we do have emerge most extensively during the frenzy of printing in the late nineteenth and early twentieth centuries, usually appended to the narrative as a result of deliberate attempts by certain individuals to Sanskritize the tradition by turning the simple offering of sirni (or sinni) into a more elaborate paja.8 In spite of these efforts to incorporate the tradition into the mainstream of the daily pajas, the ritual literature still accounts for less than 1 percent of all com- positions. The handwritten punthis or manuscripts are nearly devoid of such ritual instructions, with most taking the form of the orally deliv- ered p-ld- gana, puncali, and the increasingly popular vrata katha.9 While the vrata katha has become ossified in its thematic structure, the other two forms yield an impressive diversity, and when all types are enumer- ated, they are as prolific as they are diverse. In the repositories of West Bengal and Bangladesh, I have located approximately 750 manuscripts composed by more than 150 authors stretching over the last five centu- ries,10 some of which disappeared in the aftermath of the unfortunate  Tony K. Stewart | 27 confrontation at the Babri Masjid in December 1992.11 Printed texts are numerous, with the collections and markets in London, Calcutta, and Dhaka yielding more than 160 titles by more than 100 authors. But even though authorship is diverse, the narratives do show a strong thematic unity and, perhaps more significantly, a predictable set of narrative plots that yield equally predictable results. The situations described in the literature of Satya Pir constitute a fairly limited narrative domain, using small numbers of fixed character types, in a limited set of possible fictional predicaments, whose primary complications are generally permutations of a much smaller set of un- derlying or controlling themes, e.g., worship Satya Pir to get rich or to be rescued from trouble. These underlying themes, however, are not al- ways approached the same way, so to describe the strategies for negoti- ating these situations we will borrow the narratological term narrative codes.12 But in the case of Satya Pir, and in much of the popular religious literature of South Asia, the narrative codes are not simply shaping liter- ary fictions. They have a much more immediate connection to issues of everyday life, that is, they have a relevance to the way people live and come to understand how they should conduct themselves, how they might survive, in a world that does not always cooperate. In this, the narrative codes are different from their purely literary counterparts (if we can be allowed such a potentially artificial distinction for a "pure" vs. "practical" literature), and they reflect in every case the way actors mar- shal competing structures of authority to modulate the power of sur- vival represented by the protagonist, Satya Pir. This is a vital function because classification of these narrative codes will reveal something of the logic by which different people can and do think differently about the same contingent existence, interacting with the same figures in the same settings. Classification of these narrative codes, as indexes to the actors' "orientations toward authority," allows us to recognize other systems of signification, most obviously through intertextual references, both overt and implied (e.g., Skanda Purdhnza), or to other cultural insti- tutions (e.g., dargah), that are used to reinforce the orientation. Finally, because these individual items or subsets of alternate signification often stand in metonymic relation to the basic narrative code in the context of the narrative itself-they are often freely mixed and matched as ele- ments in the story-their differences will ultimately reveal that the structures of authority are considerably more complex and subtly nu- anced than the basic categories of Hindu and Muslim could ever recog-  28 | Satya Pir on the Frontiers of Bengal nize, and they often actively imbricate what are traditionally thought of as either exclusively Hindu or Muslim attitudes and acts. In short, the narrative structures adopted by individuals or groups serve as a basis for categorizing action and orientation-different ways of thinking about and negotiating the way power is wielded in this world-far more dynamically than the assignment of monolithic labels of static theological and ritual ideals; they reveal part of the decision- making process, the factors that are weighed in determining proper courses of action, and how individuals with different backgrounds can use the same situation in interaction with the same figures for different ends or can arrive at the same ends by different means. This is not to say that individuals who preserve and propagate these tales of Satya Pir would not recognize the content of the categories Hindu and Muslim, but those categories operate on a different level of experience most often associated with the symbolic posturing appropriate to the larger public sphere, and in that sphere they maintain a kind of consistency of image that everyone recognizes (e.g., the rules of public propriety, severely delimited ritual and symbolic action and dress, etc.). But in the narra- tives of Satya Pir, those kinds of distinctions do not play in the negotia- tion of the private vicissitudes of daily experience on the frontiers of Bengal, and they can be easily ignored or, as will become apparent, when invoked-as in the description of Satya Pir's garb noted above- they can be used as a foil to expose the ignorance of their improper ap- plication. In order to produce a workable sample, more than a third of the manuscripts and nearly all of the printed literature available have been analyzed.13 The narratives can be organized into three general types, sets that are determined by combination of the manifest identity of Satya Pir and his direct role (or absence) in the plot; the social standing and voca- tion of the protagonists other than Satya Pir; the nature and direction of instruction and subsequent conversion, if any; the occasional overt reli- gious point or more general moral of the story; and the audience for which the stories were apparently intended and which can be deter- mined only partially. Each set of characteristics contributes to the strat- egy that is adopted by the narrative, its narrative code, which ultimately defines its type. In characterizing these strategies, however, we will re- vert momentarily to the use of the general adjectives of Hindu and Mus- lim, but with the proviso that those be read as orientations (that are co- herently conceived, but not at all consistent) as the result of individuals  Tony K. Stewart | 29 responding to the pragmatic results of the orientation, rather than choosing to be included in a group that goes by that name. - Satya Pir as "Hindu" Vaisnava God: those tales that emphasize the Vaisnava identity of Satya Pir as the incarnation (avatara) of Visnu or Narayana, who has descended to right the dharma for the Kali Age. The narrative code operates according to strategies of "domestication" and "appropriation." - Satya Pir as a "Muslim" Moral Exemplar: those tales that feature Satya Pir as a pir or fakir, who challenges the hubris and exclusivity of a conservative brahmanical authority and the conniving ways of dishon- est and irresponsible individuals, regardless of religious persuasion. These tales promote an Islamic perspective on ritual, theology (when noted), and conversion, even though Satya Pir's persona often seems to invoke more the features of a Hindu deity. The narrative code dictates strategies of "recognition" and "accommodation." - Satya Pir as Personal Spiritual Guide: those tales that, at least on the surface, seem not to address any obvious religious issue, but focus in- stead on fundamental moral quandaries that ultimately lead to prag- matic resolutions of everyday problems, often through personal tests and unexpected alliances. Among these stories is a large subset of tales that focus on the acts of women who must survive compromising situa- tions where there are no clear guidelines for propriety. The narrative strategy is for "moral improvisation" and "alliance." What binds together all three of these narrative types is the common improvisation necessary to negotiate an often hostile or compromising environment using locally available sources of power, most notably the pir, but also committed or converted kings, and especially their entre- preneurial merchants. The environment of their setting is always some kind of frontier, so these are generally read as narratives of survival, and as Richard Eaton has clearly shown, the land of Bengal where these sto- ries proliferate has for centuries been conceived in just such terms.14 The frontier, however, is plural and shifting, for it is geographic, political, economic, and religious-and the stories of Satya Pir address them all. In these narratives, the frontier is an arena of human action that lies beyond the circumscribed limits of what is familiar and what constitutes the predictably settled world of "tradition." Therein lies much of the stories' interest and mystery, if not reason sufficient in itself to question the use of the larger categories of Hindu and Muslim which so often blur in these socially ill defined areas. These tales are a journey into the un-  30 | Satya Pir on the Frontiers of Bengal known, where dangers are manifold, not so much because they are in- herently threatening, although the tales are littered with episodes of real danger to the protagonists, but often simply because the modes of action that are considered normal do not always hold true in a land that is unfamiliar. Yet, for many of the people who listen to the tales of Satya Pir, that shifting frame of reference describes their Bengal precisely; it is a land of constantly renegotiated values, of improvisation, of attempts to impose stability. As a frontier it is a place where the social, political, and economic stakes are often high, with commensurate rewards for success or failure. In this formulation we discover part of the secret of Satya Pir's social mobility and appeal. Meeting the needs of the frontier has al- lowed Satya Pir to endure, for his pragmatic approach to the problems of the world is one that favors innovation and compromise in the pursuit of basic human needs, especially the elimination of penury and the quest for social dignity. His are the tales of survival in a world that does not always cooperate, and for many in Bengal, that is the commonplace of experience. Satya Pir as God: Vaisnava Tales Strategies of Domestication and Appropriation Those tales that we are inclined to read as unmistakably "Hindu" are better described as exclusively Vaisnava; and they place Satya Pir into the framework of a generic puranic avatara theory, part of its supple- mental signification. Although the nature of that avatara will vary, he is generally accorded the status of the yugavatara, Narayana's incar- national descent for the Kali Age. The logic of this characterization is quite understandable, for as is so well known and frequently cited, Narayana promises to descend whenever the dharma has languished (Bhagavad Gita 4.7-8) and in whatever form meets the needs of the people of that age (4.11). Sizable numbers of these texts frame the de- scent in classical terms by opening the narrative in the heavens where Narayana sleeps. Narada-that celestial gadfly who is as responsible for stirring up problems as for coming to everyone's aid-journeys to Narayana's court to alert him to the malaise that threatens to engulf civilization.15 After an exchange of traditional greetings, Narada invites Narayana to survey the situation and determine an appropriate re- sponse. As Narayana wakes up to the full extent of dharma's demise (those texts that begin in medias res generally begin here), Narada prods him to descend in a form people will understand, and because foreign-  Tony K. Stewart 1 31 ers alien to the traditional brahmanical homeland (madhyadesa) are ev- erywhere in power, the form of this particular descent, he reasons, should play on that familiarity. The prologue closes when Narayana takes the advice to heart and descends in the form of Satya Pir, overtly a pir, but in reality none other than the celestial Visnu. Even for those narratives that do not explicitly provide this narrative frame to justify the descent, some form of it is implied, for it everywhere replicates the puranic premise of the avatara. Where the justification does frame the story, it makes explicit several key features of the narratives that lend Satya Pir his broad appeal and certainly contribute to his endurance and adaptability. The texts unam- biguously identify the controllers of the land as "foreigners" (yavana); no other term is used until very late, well into the colonial age.16 While yavana is often translated as "Muslim," its derivation is a word indicat- ing "Ionian" or "Greek," with the implication that a yavana is someone who comes overland from the west (about the only direction from which new peoples entered Bengal in numbers until the colonial period). That they were almost always Muslim in this premodern age is only second- arily remarked, for usually the designations were more ethnically spe- cific (e.g., Pathan or Turk). The implication of the nonspecific term yavana operates on the controlling premise that someone whose ways are not of the traditional "Hindu" (the term is occasionally used adjecti- vally, but never nominally) has taken control of the countryside, and that in itself poses a threat to the stability of a common brahmanical culture, especially in the unsettled reaches of Bengal. It is easy to see how the yavana category as a generic "other" becomes associated with its alternative "phirirgi," applied specifically to the Portuguese (and French), but coming to designate all Europeans, many of whom arrived by ship through the Bay of Bengal. The land is controlled by yavanas (foreigners who look like us, but act differently) or by phiringis (foreign- ers who do not look like us and are even stranger and more unpredict- able and less trustworthy in their actions).17 Because of their generic nature and lack of historical specificity, Satya Pir's narratives easily function with both connotations. When the texts do adopt these desig- nations, they follow a sequence that begins with the earliest stories rec- ognizing an initial opposition that establishes a brahmanical, specifi- cally Vaisnava, cultural norm (the term Hindu is never used here) against yavana (the term Muslim or any equivalent is likewise never used). By the end of the Vaisnava cycle of narratives-and often other tales as well-the stories articulate a pragmatic alliance of Vaisnava and yavana  32 | Satya Pir on the Frontiers of Bengal (both of whom are now considered indigenous) vying for power with an implied phiringi or other oppositional category, such as Saiva or Sakta. This somewhat unexpected alliance of Vaisnava and yavana is possible because the Vaisnavas domesticate the yavana through their theological constructs, their religious institutions, and their developing rituals. Domestication is precisely Narayana's strategy in these narratives, but to domesticate, one must first appropriate. With the decision to use the familiar form of the yavana in order to reveal a new dharma that will unite the yavana with the Vaisnava, Narayana follows Narada's prompting by choosing to appropriate the form of the wandering men- dicant. The samnyasi, of course, was a familiar figure to the countryside, and there were a number of them who routinely found their way through Bengal during the period. They were homeless, and they had aban- doned personal possessions in pursuit of a religious ideal, which was often attained through yoga. These samnyasis were figures of consider- able power, much of it magical, thanks largely to those arcane yogic practices. (In fact, the yogi, who is strongly associated with the Nathas, is perhaps a more ubiquitous and ambivalent figure during this period.) The samnyasi (or yogi) is the first obvious analog to the image adopted by Satya Pir; this analog is all the more compelling when he is desig- nated a fakir, for the fakir is a homeless itinerant who has taken a vow of poverty for religious ends. Here the samnyasi and fakir serve as general institutional equivalents, utilizing in the narrative the expectations in- voked by their different externally grounded signification systems. But Satya Pir is actually often understood to be permanently attached to a place-one of his most popular homes is Mecca, although other candi- dates in and around Bengal and Orissa qualify-and he is called upon to guide his followers in a stable environment over time, a function much more consistent with established sheikhs or pirs who tend not to be wandering mendicants. In this persona, his image bears all the marks of the god he ostensibly incarnates (now drawing on a conflation of signi- fying systems), with his personal abode in Mecca serving as an ersatz heaven in these Vaisnava versions. But his historical equivalent may be more aptly found in the vairagi, who is the Vaisnava alternative to the samnyasi, frequently married, and only infrequently itinerant if at all, but who in his rustic form often develops the kinds of powers associated with the yogi. This Vaisnava image of piety provides a much closer ana- log to the pir, although the associations are as much implied as explicit. Yet either association of samnyasi, yogi, or vairagi makes the external form of the pir already familiar, and when taken together they make it  Tony K. Stewart 1 33 intimate. The familiarity is made comfortable because both such sets of individuals were reliable institutions of local power in their ability to advise and guide, to help their followers negotiate the trials and tribula- tions of this world, and when truly necessary, to use their considerable extraordinary powers for mundane as well as spiritual ends. This seems to be key, for Satya Pir exhibits traits common to all of these figures by his constant concern to meet the immediate needs of his constituency. The religious practices he proposes and the demands he makes are very much of this world. They do not promise futures in heaven, union with or annihilation in God, or escape from the cycle of life. They only prom- ise basic prosperity, safety, and weal in this uneasy land. The authors of these narratives oriented toward a Vaisnava sensibil- ity feel compelled to justify the decision made by Narayana to appropri- ate the image of the pir to Vaisnava ends, for it is clear that the pir's form represents something other than what is traditionally acceptable to brahmanical culture. Obviously, in public image any pir is Muslim and not Vaisnava. It is not enough that the puranic frame of the tale explains Narayana's motivation. More is needed to convince the Vaisnava audi- ence, so these narratives almost always appear in a trilogy designed to persuade the audience in a step-by-step fashion of the necessity and efficacy of the act. Before total domestication is possible, the form of this pir must not only be recognized as comfortably familiar but must also be made legitimate. Legitimation is the linchpin to the process of appro- priation, for if the new form is to endure as a viable and appealing future alternative, it must be grounded in an unassailable logic of possibility; that is, it must be made to conform to expectations in a way that is unde- niably appropriate to the Vaisnava conception of, or at least orientation to, the world-and that is precisely where the narratives begin. The pro- cess of legitimation starts by having an experienced brahmana-the representative of traditional society, but a society that has failed to sup- port him-recognize the form of Satya Pir by affirming his "true" iden- tity as Narayana. From this simple beginning the pir's form is gradually valorized throughout the whole of brahmanical society, which is "docu- mented" in the set of three stories-and that set is the overwhelming favorite form for practicing Vaisnavas. They tell of the conversion of (a) the old brahmana and his wife, (b) the local woodcutters, and (c) the merchant and his family, directly or indirectly ending with the local king himself.18 While the final tale is occasionally the subject of an entire work,19 nearly three-quarters of all manuscripts and printed texts are devoted to this complete three-part strategy precisely because its effec-  34 | Satya Pir on the Frontiers of Bengal tiveness lies in its progression; the two versions among them attributed to the Bengali poets Sankaracarya and Ramesvara prove most popular.20 Not surprisingly, it is this trilogy which is read directly into the puranic material in the Skanda Purana (5.233-36) and Bhavisya Purana (3.2.24-29), and which forms the basis for incorporation into the monthly vrata cycle of the wider Hindu households of Bengal,21 another result of the Sanskritizing effort.22 The tales can be summarized as fol- lows. The brdhmana's tale The tales begin with the saga of the old brahmana who is reduced to utter penury. He resides in Varanasi, that center of traditional piety, but cannot even beg a day's worth of alms to feed his wife and himself. He is distraught over his prospects because the downward spiral conspires to keep him from being productive as a priest, for the poorer he be- comes, the less likely his employment. When his prospects dim to the point where he can no longer offer a viable service to the competitive world of that metropolis, he finds himself in the unthinkable horror of being pushed to the very edges of civilization, east into the wilds of Bengal.23 In this pitiful state, he is approached by Satya Pir, who holds out one last alternative. "Offer sirni to me," he commands, "and your wishes will be fulfilled." Ever polite and sorely tempted, the brahmana resists the cry of his stomach and refuses to jettison the last remnants of his dignity as a brahmana, demurring on the grounds that Satya Pir is yavana and such worship would be improper. Satya Pir acknowledges the brahmana's piety and instructs him to pay close attention. He gently suggests to that good but poor brahmana that he must never be fooled by outward appearance, for Satya Pir is really none other than Na- rayana-incarnate. The brahmana is skeptical and asks for proof, which Satya Pir provides by displaying his six-armed form as Visnu, the Satya Narayana. Satya Pir, he explains, is but an avatara. Having witnessed with his own eyes, the brahmana happily acknowledges the revelation, proffers the sirni precisely as instructed, and in an instant grows wealthy, all to the extreme pleasure and benefit of him, his wife, and others around him. In every version of the story he does, in fact, live quite happily ever after. The woodcutters' tale Numerous woodcutters reside in the same area as the brahmana, and it falls to them to clear land for cultivation and provide wood for fuel in  Tony K. Stewart 1 35 this expanding economy.24 They have grown accustomed to passing the old brahmana beside the road as they make their daily trips deep into the forests. When the brahmana's fortunes abruptly change, they are astounded, for the transformation is both miraculous and nearly instan- taneous; overnight he becomes successful and highly esteemed. Natu- rally, they want to know the source of his good fortune, and when they inquire, the brahmana proves himself worthy of Satya Pir's trust. Being ever grateful to that mysterious pir who has so dramatically secured his future, he does just as he has been instructed and shares the secret. He is blunt: "Sincerely worship Satya Pir with sirni, and you too will become rich." Not slow to recognize the opportunity, the woodcutters follow the injunction and within a very short time they become controllers of fabu- lous wealth. So successful are they that they can build large fortresses on the tracts of land they clear, their estates expanding rapidly, while the frontier they are taming extends further eastward. Inevitably, their suc- cess brings more land under cultivation and makes it fit for habitation by traditional brahmanical society, for not only is it cleared but it is filled with moral people, including law-abiding kings to rule and brahmanas, like the one who shared his secret, to ensure propriety.25 The merchant's tale As the settlements develop, local rulers require certain royal items, both luxury and symbolic, to assert their status and claim to power, that is, simply to be kings of these new lands. To bring the requisite and rare goods to court, each king finds himself in need of reliable merchants, who, if they are successful, become fabulously wealthy and powerful in the process. Procuring these unusual items, however, entails great risks, for their source invariably lies beyond the seas, and any venture onto the ocean is risky. Through their own devices or with the financial backing of the king, the merchants set off to adventures only imagined by ordi- nary people. Their ships glide effortlessly through the familiar waters of Bengal, out into the Bay of Bengal and the Indian Ocean. When they dare to venture away from land, they cannot but encounter threats found only deep at sea, for instance, the report of Dayala, who records "a tomb of marble floating on the sea with girls dancing around it to the musical accompaniment of celestial kimnaras, and deerskins spread like carpets on the surface of the waters, with four fakirs saying their namaz facing West."26 Because of such reports and with a practical estimate of their own limitations, they more often prefer to hug the coast as they work their way south. They stop periodically at cities and lands of decreasing  36 | Satya Pir on the Frontiers of Bengal familiarity until they reach the furthest outposts of civilization, Kalinga, then the Dravida region, and even the isle of Sri Lanka, which is always populated by demons and monsters, who just as predictably protect great wealth. To offset the dangers, the merchants turn to Satya Pir, for the creator of instant wealth can likewise be counted on to watch over its acquisition. Thus Satya Pir comes to be the protector of merchants and travelers in general. To ensure this success, the merchants promise to worship Satya Pir to a degree commensurate with their acquired wealth. But if wealth and good fortune can be created at a stroke, so too can it be removed and destroyed; failure to maintain that promise to worship Satya Pir will only result in disaster. Sometimes it is the merchant or his accompanying sons whose greed causes one of them to withhold the worship, which in turn precipitates the ship's foundering or which lands one of them in jail. In those vile dungeons they may languish for years with no hope of escape until they belatedly remember Satya Pir. Equally disastrous is the negligent action of the merchant's wife who has remained at home, or more frequently it is the action of the selfish daughter-in-law, who offends Satya Pir so that success is denied even as the ships sail back into view after years abroad, sinking in the estuary as they come to dock. The variations are many, but the theme is monoto- nous: if you fail to make good on your contractual promise to worship Satya Pir in exchange for his protection, you are doomed. But here, when the worship is properly discharged, or the mistakes are acknowledged and corrected with appropriate humility, the merchant is successful: the king receives those goods he requires to maintain his status as rightful and just ruler of the land, the merchant accrues wealth and status for his reliable delivery, his wife and daughters-in-law receive appropriate pro- tection of their fidelity in the merchant's absence, and the society as a whole confirms the validity of its attempt to maintain stability and or- der-all because Satya Pir is widely worshiped. In short, dharma pre- vails, everyone prospers, and, say the stories, if you have paid attention, you can prosper, too.27 Seeking Equivalence: Pragmatic Implications of Vaisnava and Sufi Theology It is no surprise that of all the Hindu communities enjoying a substantial following in Bengal during the last several centuries-many different forms of Saiva, Sakta, Natha, Sahajiya, and Kart-a Bhaja, to mention only primary groups-it is the Vaisnavas (and later Bauls) who attempt to appropriate a figure who is clearly Muslim, for they alone can easily  Tony K. Stewart 1 37 justify the action through their ever-expanding avatara theory, which claims virtually any popular figure as its own. As becomes apparent through the other narrative types, the Vaisnava model of God's descent, the avatara, and the Islamic institution of the pir, can be allied not only because the respective images of the holy man-pir (and fakir) and vairagi (and samnyasi)28-coincide so conveniently as metaphors of the embodiment of power, but because there is a basic theological compat- ibility that undergirds both conceptions of divinity to which they refer, and this consonance will generate apposite orientations toward author- ity that will prove their coherence in the narratives of Satya Pir. Like the vairagi, the pir does not prescribe the esoteric practices re- served for adepts like himself, but simpler and more popular forms of piety appropriate to the average follower; much of his guidance falls into the adjudication of everyday problems, marital issues, arbitration of disputes, and so forth. The image of divinity associated with these simpler prescriptive rituals and instructions will run the full gamut of experiences, just as they do in the Vaisnava order. Not only are the insti- tutional structures of the pir and vairagi, then, analogous in a general way, but their operational and theological underpinnings are closely equivalent, and this is borne out in comparisons of both general and historically specific dimensions of theology, such as the nature of the godhead and the injunctions to ritual practices. While it is easy to specu- late in purely intellectual or theological terms why these two traditions may be inclined to find a mutual alliance, it is their operational dimen- sion that bears out the practicality of it-and that allows the Vaisnavas to appropriate the image of Satya Pir with virtual impunity-in fact, one might even argue, with a very unsurprising anticipation if not expecta- tion of its inevitability. Given the similarity of the functions of the Vaisnava and Sufi spiritual guides and the theological parallel, it is ultimately the fact that Satya Pir is a mythic figure that effectively eliminates any possible challenge to the narratives' veracity, for no historical documentation of the pir's life and teachings aligns him with any particular sectarian group.29 This in- dependence of the narrative from historical verification dramatically aids the process of appropriation by enabling the Vaisnava to sanitize it. In this, Satya Pir's image is plastic and malleable in the manner of a puranic figure and, indeed, he quietly slips into the purarhas as just another form of Narayana. This same kind of plasticity likewise extends to the use of the narratives, for it enables them to be applied to a wide range of generic situations, again quite apart from any explicit historical  38 | Satya Pir on the Frontiers of Bengal event. Each of the Vaisnava episodes deals tacitly, if not explicitly, with generalized processes of reclamation-geographical and cultural- making habitable a land that had been off-limits to brahmanas and therefore problematic for establishing a proper brahmanical society.30 Because of its lack of specificity, the nature of that rehabilitation can be adjusted to the user's immediate circumstance. The progress docu- mented in the trilogy of Vaisnava tales parallels the historical events of the settling of Bengal. As the Ganga shifted steadily to the east, the lim- its of what defined the traditional heartland or madhyadesa of brahmanical culture could be extended, but only if brought under proper control. Making good use of the available powers, one agent of that Hindu domestication became the pir, for the pir could actually do what brahmanas themselves could not: inhabit a wild land to tame it. Ironically, the pir is often the very same agent for the analogous pro- cesses of Islamization, for the Sufi guide as pir or fakir is often the first into new countries and the first to convert the local population so that land may be brought into the line of traditional Islamic culture-and here Bengal was no exception. The same figure of the pir serves two religious orientations in nearly exactly the same capacity. Herein may lie the most important reason for Vaisnavas to appropri- ate the pir's image, for by doing so they not only unquestionably ac- knowledge the presence of Islam as a legitimate social organization and religious option in the region, but they also acknowledge that the pir works as an effective source of local power. It is an act of a pragmatic "Realpolitik" in that the Vaisnavas adopt a stance toward their rulers' culture and religion that does not try to wish away the reality of that rule but attempts to adapt to its presence and co-opt its power by appropri- ating it: they take one of the most effective tools of conversion and re- valorize its image to their own ends. It comes as no surprise, however, that even though Satya Pir is embraced, the embrace is not unmitigated or unconditional, because the Vaisnavas do not elevate him to the level of their adored Krma, but ultimately absorb him into the lower strata of the brahmanical hierarchy, placing him squarely in the women's ritual cycle of the vrata, which is dominated nearly exclusively by lesser images of divinity, especially the benign household goddesses, such as Sasthi, Laksmi, et al., who are petitioned to make life easier and more fruitful. Satya Pir proves his worth by doing much of the "dirty work" of making the land habitable and ensuring the wealth and weal of the family-the mundane role of lesser celestials-and in that proves his expediency. But in spite of the "official" recognition, he must remain  Tony K. Stewart 1 39 a marginal figure at the lower end of the Vaisnava and brahmanical world.31 Satya Pir as Islamic Exemplar: Seeking Accommodation and Demanding a Place The tales of Satya Pir that are Islamic in their provenance and orientation take a decidedly different tack to the power of the pir and the dynamics of interacting with the local populace. For obvious reasons, no time need be spent justifying Satya Pir's existence, as was necessary for the Vais- navas, for pirs are part of the everyday world. Nor is there any attempt to equate Satya Pir with a samnyasi or vairagi or yogi, even though the authors routinely refer to these figures in ways commensurate with the analogs of fakir and pir, and in so doing draw upon the associations of their underlying signification systems. Because the form of the pir func- tions in Bengal's culture as a source of local power and moral fortitude, any pir would be an obvious choice for literary interest. But document- ing Satya Pir's triumphs as a way of celebrating his superiority is clearly subsumed to the larger interest of proving or confirming that he is wor- thy of a following in the first place. These triumphs are not always nar- ratively sequenced as they are in the Vaisnava trilogy, nor are they or- dered for consumption in any way similar to the incorporation of his tales into the vrata cycle. Most are independent or only loosely related to others, but the liveliest coordinated group can be found in one ex- pansive collection, Batla satya pira o sandhydvati kauiydra punthi of Krsna- hari Dasa,32 which is structured in the form of an anecdotal hagiography of the hero. This lengthy book opens by invoking the glory of Allah and the Prophet and describing the wonders of Behest (paradise). Because a cer- tain Hindu (but not Vaisnava) king named Maidanava had been perse- cuting pirs and fakirs indiscriminately, God ordered the goddess Cand- bibi to descend to earth to initiate a plan that would reestablish a just society. When Candbibi accepted that order, she began to fulfill a long- held prophecy that Satya Pir would be born in the Kali Age to save hu- manity (note the Hindu cosmology), and so she was born to Priyavati, the wife of the evil king, as his daughter Sandhyavati. When she was bathing in the river one day, she picked up and smelled a flower petal that had been floating downstream, and with that inhalation the just prepubescent girl immaculately conceived Satya Pir. Her mother was distraught and tried to force an abortion, but to no avail. When she was  40 | Satya Pir on the Frontiers of Bengal about to give birth, her mother banished her to the forest, where she was taken and left to die. Her cries alerted Allah, who sent an angel to protect her, while from the womb, Satya Pir called on Lokamana Hakim33 to build her a suitable abode. When she gave birth, there was no child but a clot of blood, which she sadly consigned to the river. But a sin-filled tortoise swallowed that blood and was instantly transformed. The tor- toise gave birth to Satya Pir and retired directly to Paradise. Satya Pir returned to his mother after five years of extended study with various famous murshids. She was still alone in her palace in the woods, so he used his persuasive powers to relocate an entire population into the Jharikhanda Forest,34 clearing massive areas of land and establishing a community. Here he came into conflict with local kings, whose inhabit- ants he stole, beginning the long saga of righting the wrongs that had been perpetrated against good and pious pirs and fakirs. The tale then traces Satya Pir's exploits through his youth and adult life, each tale adding to the strength and depth of his miraculous powers and his ever-expanding circle of influence. The book is of special interest because it attempts to create a "life" for Satya Pir on the order of the hagiographies devoted to historical figures of the premodern period; in fact, this would appear to be the only hagiography devoted to a "mythic" pir among the dozen or so who are popular throughout Ben- gal. But being anecdotal, it is only loosely organized with no ending and is, therefore, infinitely expandable. The range of exploits is considerably greater than its more tightly controlled and limited Vaisnava counter narratives, whose protagonists are other than the pir, for in those tales, as we saw above, the pir primarily serves as a catalyst for action and the object of worship, but he is never the direct protagonist of the story. The function of Satya Pir in the Vaisnava tales is similar to his function in the third type of apparently nonsectarian tale where his role is to initiate action, complicate the plot, or provide the raison d'etre for the protago- nists' adventures. In contrast, the Muslim-oriented tales focus on Satya Pir as the hero, but it is perhaps significant that the opening gambit is precisely the same impulse that operates in the Vaisnava stories, for Satya Pir begins the process of reclaiming the forests of Bengal while he is still gestating in the womb. Furthering the contrast with the timeless- ness of the puranic-style Vaisnava trilogy (previously analyzed) or with the nonsectarian type (not analyzed here),35 the Muslim tales appear to be somewhat more historically fixed in the immediately precolonial and early colonial world. There are, for instance, references to historical fig- ures, such as Satya Pir's encounter with Man Singh late in the opening  Tony K. Stewart 1 41 book, Malanca pala.36 Satya Pir encounters individuals who wield Euro- pean rifles and cannons. It is reported that he is strapped to a cannon and blown to bits, only to miraculously rematerialize before the eyes of the miscreants, who then receive their much deserved punishment. Other tales bear witness to a phiririgi presence, whereas the Vaisnava tales never finger the phiringi directly, but leave it to be supplied by the audi- tor as appropriate to an immediate crisis. Some of the more moralizing tales, however, do exhibit some ambiguity of historical location, and that allows their messages to be transferred and adapted more easily to immediate or generalized exigencies regardless of time or place. A good example is Satya Pir's instruction to the greedy and selfish Dhananjaya, a prosperous milkman,37 summarized as follows. Dhananjaya's tale As Satya Pir wandered through the delta, he approached the expansive home of Dhananjaya the milkman. He sat down and recited the names of God and the Prophet, then called out for food, but Dhananjaya, being a mean and selfish man, was ungracious and ordered him to wait for some leftovers. Satya Pir was annoyed at the affront and quickly cursed him: "You are doomed, for you have no faith in fakirs or sanraysis. From this day your house will be abandoned by Laksmi, the goddess of wealth and prosperity." Dhananjaya scoffed at his anger and called him a raving (paigala) fakir, which he attributed to starvation from fasting. He announced defiantly that no matter how much Satya Pir tried to make him suffer, he would never beg. "That wealth was given by God, and none other can take it away. If God protects me, who can hurt me?" Satya Pir restrained himself long enough to offer Dhananjaya the opportunity to repent by lecturing him on the sins of hubris and greed and on the sin of serving polluted garbage, especially to good Musal- mans: "Anyone who offers contaminated garbage to any living being plunges into the bottomless pit of hell! You, like a haughty brahmana, greet the Musalman's salam with the raised hand of false sincerity and then to that same Musalman you have dished out rancid, polluted gar- bage. He who gives the Musalman such garbage must, in the final ac- counting, stand before the Prophet. You have ground your dharma into dust. In your next life you will be born as a jackal or a dog and will eat the garbage you have distributed in this. Your sin can be expiated only after you yourself have consumed the leftover garbage from twelve different social groups (jati)." The milkman was indifferent and mocked him as a crazy fakir made  42 | Satya Pir on the Frontiers of Bengal daft from starvation. But his final mistake was to punctuate this con- tempt by turning his back on the pir as he was speaking. The pir grew furious. He commanded a high-flying kite to snatch the plate from Dhananfjaya's hand. Because the command of a pir must be obeyed, the kite instantly dove from the sky, plucked the plate in his beak, and soared back to the heavens whereupon he dropped the plate on the milkman. The plate hit its mark, smashing into his head like a bolt of lightning. The spurts of blood and cries of agony only intensified Satya Pir's anger. The moment the pir set foot in the milkman's home, Laksmi fled in fear. The milkman's pots of money were soon ground into shards, and his pots of milk were spilled onto the ground. His rice granary crackled into a fiery conflagration that resembled hell itself. The fructify- ing cows that packed his sheds were soon transmogrified into deer, who fled deep into the forest, terrified of the raging fires. As the milkman lay comatose and bleeding, Satya Pir took his leave, and soon bands of thieves looted whatever was left. Dhananjaya, along with his four sons, was reduced to utter penury, to the very begging he swore he would never do. It took nearly six months for Satya Pir's anger to subside, but being an ocean of mercy, he eventually felt a twinge of compassion for Dhananjaya. The fakir proceeded to Dhananfjaya's house, ready to for- give. He chanted the customary dhikr as he approached. The milkman was terrified, for it was precisely that sound which had preceded his downfall. Dhananjaya did not recognize Satya Pir, but was quick to ac- knowledge the power of all such pirs. He fell at the pir's feet, rubbing his face in the dust, and with all humility implored him to be merciful. He related the sad tale of his previous stupidity and ugly behavior toward some anonymous pir. He observed that even though he was now a worthless beggar, he had learned hard the lesson of his pride and prom- ised to make an offering a hundred times over anything the pir might be pleased to ask. Satya Pir questioned the integrity and sincerity of the milkman, for, as he noted, as far as he could see, God had already been very kind to him, gracing him with large herds of cattle. Dhananjaya was nonplussed, but sensing perhaps an opportunity to regain his wealth, he promised Satya Pir that he would offer sirni made from the milk of a hundred cows, should that former wealth be restored. Because it was clear that the milk- man meant what he said, Satya Pir restored everything as it had been with the simple wave of his left hand. With his sons in tow, Dhananjaya hustled to herd the cows, for they were in sore need of milking, as Satya  Tony K. Stewart 1 43 Pir smiled and slipped quietly away. The author concludes: "I have come to the end of this tale, meditating on Radhakanta the Tolerant. Muslims call him Allah, while Hindus call him Hari." The author's assertion that Allah and Krsna are but two names of the same God introduces a new level of ambiguity by failing to designate a clear sectarian orientation; the adjectives are likewise derived from both traditions-Radhakanta, "the beloved of Radha," and ksanta, "the toler- ant," a standard attribute of Allah. The frame created by the author's signature line (bhauita) at the close of the narrative-a common tech- nique in Bengali poetry from the period-creates the illusion that the two traditions are somehow not different. The effect, however, is to in- vite the reader to insert his or her own god as the one to whom the other god has been assimilated, so Vaisnavas will read it as a confirmation of the truth of the already established trilogy that makes Satya Pir an avatara of Narayana (even though Satya Pir is not mentioned by name in the bhanita). Conversely, Muslims can read the text to interpret Satya Pir as the pir he is, while acknowledging that Vaisnavas are sincerely reli- gious, even though they do not recognize the full truth of God. Given the ambiguity of this double reading, it is easy to see why this tradition is given the label of syncretic, but that is not at all what the author pro- poses. Consistently through the more than two hundred pages of this text describing scores of adventures, Satya Pir demonstrates an Islamic orientation toward divinity and worldly power, and just as the opening frame story suggests, he is intent on establishing that in the world. When he actually converts the wayward (as opposed to simply making them recognize his power and give respect to pirs), it is always a conversion to Islam, usually initiated by the recitation of the kalima. The author seems to be equally comfortable articulating a cosmology that encompasses Laksmi, the Hindu goddess of wealth, and the Prophet, as a celestial figure; so, too, an Islamic judgment day side-by-side with references to transmigration. The author makes no attempt to rectify these and other apparently contradictory references because they are not being used to construct a consistent syncretic cosmology. They function to demon- strate certain equivalences, acting as metaphoric alternatives or simply different ways that people have of describing the same reality, while still acknowledging differences among the religious perspectives, not trying to fuse them. But these, too, lend themselves to the same convenient double reading as the bhanita, which disguise the thrust of the narrative. Even the author's own name gives one pause, for Krsnahari Dasa is a Vaisnava epithet.38  44 | Satya Pir on the Frontiers of Bengal What exactly the author intended by this dual strategy of obfuscation and metaphoric equivalence we can only speculate, but the effect of this strategy we can certainly gauge. Disguising the Muslim orientation of the narrative by suggesting that the Muslim God and the Vaisnava God are not different, but simply approached differently, functions as a plea for recognition. The stories confirm it by having Satya Pir routinely hu- miliate his opponents into recognizing his right to be honored, and this humiliation is effected by an awesome display of magical prowess. Demonstrating unchecked power guarantees attention, and this is how Satya Pir ensures that he will be taken seriously. Satya Pir projects this power in two ways: to convince the skeptic and to assure the already convinced, who in turn mirror the two audiences. His kardmat or miracle working (which is often described as a kind of sakti) enables him to twist nature to his own ends in a way that is possible only by the most accomplished of spiritual adepts-and this power is generally marshaled for the sake of the noncommitted, the skeptic, or the utter kafir (infidel). The demonstrations, as Dhananjaya attests, can be ex- tremely violent, but may just as easily be used to counter someone else's ill will, as will be the case in the story of King Kasikanta below. When he has gained their undivided attention, he achieves the recognition of his power and the truth of the Islamic cosmology that makes it so, but that acknowledgment does not necessarily require conversion. Recognition is, however, a necessary prerequisite for an ultimate goal of accommo- dation, to find a place in the shared cosmology of Bengal. Satya Pir seeks to maintain a permanent position in this world by extending his power to protect those who recognize him. The texts will generally use the term baraka (benevolent blessing) to describe the power that blankets the right-minded and morally pure in a general weal that is measurable in terms of increased wealth or rule in a kingdom of peace. The stories abound with the results of this protection, but to be literarily effective, that is, to make the causal relationship of acknowledgment and benefit unequivocal, they must be more dramatic and entertaining. Like all such hagiographical tales designed to glorify the hero, the authors exaggerate the blindness and stupidity of the antagonists in the face of Satya Pir's obvious superiority. This narrative strategy provides the opportunity for Satya Pir to enact a decisively dramatic conversion of the antagonist to drive home the point. The summary of the conversion of King Kasikanta is illustrative.39  Tony K. Stewart 1 45 The Tale of King Kasikanta One day Satya Pir wandered into a brahmana village dressed half as a Vaisnava brahmana and half as a Muslim fakir, carrying a string of tulasi beads, a chain around his waist, ash smeared on his forehead, and san- dal paste on his feet. In this garb he headed for the Sanskrit school. Al- though students at the school taunted and insulted him, Satya Pir was undaunted as he asked for food, specifically requesting unboiled milk, banana, honey, and rice flour.40 He also requested a parasol, so that he could be seated with them, and a fresh sacred thread, all of which prompted a fusillade of imprecations. Enduring the invectives, Satya Pir eloquently rejoined, cursing to illiteracy for seven generations one par- ticularly dull and arrogant young brahmana, who had dared to insult him in pidgin-Sanskrit. Somewhat mollified at the spectacle, he then retired to meditate under a tree he conjured. As he sat deep in meditation, Satya Pir summoned the sacred threads of those arrogant brahmanas. One after the other, the threads snaked down the road to join their master.41 Trailing behind was an equally long line of dejected and obsequious brahmanas who by then had had their pride curbed and their curiosities piqued. They plaintively inquired just who he was and why he tormented them so. Satya Pir responded, "You may be brahmanas but you are no different from the rest, for the serpent of Time and Death bites equally. Be respectful of all samnyasis and fakirs, treat them with kindness, lest they show you to be nothing but students of Sayatan." He then revealed his true identity as a pir favored by God, and the brahmanas not only submitted to his authority but made restitution by offering the sirni they had previously denied him. When their king, Kasikanta, heard of this strange behavior, he raised the cry of blasphemy and summoned the brahmanas to account for themselves by bringing Satya Pir to demonstrate his power. Because he had no faith, he stupidly challenged the pir to do something extraordi- nary, something which could demonstrate that he was more powerful than the king himself. Satya Pir quietly replied that that should be easy, for a king who could not even control his own wives could not wield too much might. The king's ire was sparked, and he demanded an instant apology, but Satya Pir demurred, preferring to offer proof of his conten- tion. He transformed himself into a white fly to wing his way unmo- lested into the queens' quarters. There he began to incite them, gently at  46 | Satya Pir on the Frontiers of Bengal first, but with increasing pressure, to dance. The queens in turn were struck suddenly with mysterious and outrageous sexual urges, causing them to writhe in uncontrollable lust. Like a contagion, the undulations swept through the zenana, and these queens began to dance, giving vent to their true natures, base and lascivious. In a frenzy, the women broke out of their quarters and violated the public hall, where they enraptured the audience but enraged and humiliated the king with their salacious advance. Unable to stop her, King Kasikanta looked on in absolute hor- ror as his beloved chief queen disgraced herself by performing a strip- tease in front of the throne. Finally Satya Pir sent them scurrying to re- cover their modesty. The completeness of the king's humiliation made him all the more obdurate, and he refused to capitulate. Instead of acknowledging Satya Pir's power, he ordered him hurled into the deepest well in the palace. Satya Pir pulled the king down with him, his sacred thread having snaked around the king's neck. Try as he might, the king could not break the thread to stop his descent. Finally, recognizing his defeat, he allowed Satya Pir to climb the thread and drag them both to safety.42 In spite of the outcome, the king's submission was initially grudging. Satya Pir accepted it nonetheless, lecturing his now captive audience on the na- ture of royal propriety and the modes of dharma and proper action. He initiated the king into the recitation of the kalima, transforming Kasi- kanta into a God-fearing, law-upholding Muslim king. As the benefits became clear, the king enthusiastically honored Satya Pir, who turned away and headed home. In the narratives, Satya Pir demands that people publicly acknowl- edge his legitimacy as an effective source of moral power, and from that the legitimacy of his worship. For those who do so, numerous benefits accrue. His relentless insistence and the concomitantly harsh forms of persuasion suggest that this acceptance was hard won, that coercion was in fact on occasion necessary. These struggles and their inevitable confrontations complicate the plots, and their consistently antibrah- manical tenor is often undisguised. But in spite of that, the condemna- tion is selectively, not universally, applied. It is not that brahmanas are inherently bad but that brahmanas are too frequently blinded by their own hubris, which results from an overvaluation of their social stand- ing; that is, they confuse the privileges of their rank with an inalienable birthright, rather than seeing it as a fragile commodity that must be maintained through virtuous conduct. Significantly, the brahmanas in Kasikanta's tale are not forced or even asked to convert. They are asked  Tony K. Stewart 1 47 simply to honor and respect Satya Pir, for which they will be restored to their position of respectability. Brahmanas are just another social group in a Muslim cosmology, but as the top of a Hindu hierarchy they metonymically represent the whole of society, just as we saw in the opening episode of the Hindu trilogy. Kings who support such brah- manas are a more serious target because they perpetuate this arrogance and misuse of status, ensuring it as the norm for society. The need to convert King Kasikanta in no way challenges his right to rule by replac- ing him with another individual chosen by Satya Pir, who in this and other narratives could easily do so. The conflict is adjudicated on a moral battleground, so Satya Pir proves that the king has but a frail hold on dharma by effortlessly undermining the moral integrity of his own palace and family. Since, in the traditional constructions of dharma, pro- priety traditionally flows from the king into his realm, the corruption of his personal life will inevitably be manifest in the society at large. Be- cause the Hindu model has been shown deficient here, its dharma un- stable and easily subverted, Kasikanta must be converted to a just and moral order of kingship in an Islamic mode. Then goodness and mercy will undoubtedly reign-and pirs and fakirs can practice their Kraft un- molested. There is no way to determine just how closely any of these tales may approximate historical circumstances, for fictions at best can only al- lude, but we can see in them the imagination that presents an idealized perspective geared to a pragmatic survival. In these stories, Hindus of any type who acknowledge Satya Pir can and do retain their Hindu sta- tus and prosper. The stories teach that everyone must recognize and demonstrate a sincere respect for Satya Pir, if not all pirs, and that defer- ence will invariably result in worldly success. But its moral is also clear, for status and wealth once gained carry special responsibilities and must redound to the greater good of society. The proper conduct common to both the Vaisnava and Islamic God, as articulated by the Muslim por- trayal of Satya Pir, hinges on humility and benevolence, not exclusion, persecution, or greed.43 It is no coincidence, then, that the Vaisnava tales of Satya Pir assume a correlative position to the Muslim tales by arguing that a condition of utter penury obviates any chance for an individual to act in a morally responsible way, hence Satya Pir's concern to provide wealth. And here we have a strong indication that even though the Vaisnava and Islamic narratives target different audiences, they manage somehow to articulate a very closely related set of religious, or perhaps more basic existential, concerns.  48 | Satya Pir on the Frontiers of Bengal The world of Satya Pir's narratives is one where pragmatism takes precedence over idealism in matters of social organization and religion. This literature makes clear that in the minds of its authors, the encounter of people we might tend automatically to label as Hindus and Muslims in premodern Bengal does not automatically produce conflict, nor does it require differences among people to be effaced. In this world, the "other" is clearly recognized and often easily tolerated, if not embraced. The narratives typical to our sample depict similar dramatic acts viewed from religiously opposing positions in a way that acknowledges the le- gitimacy of both religious traditions without threatening either, which suggests that it is not the Great Traditions that are at stake here but lower-order symbolic concerns. The reason is simple: everyone needs a place for a figure who, in the all-important task of generating wealth for survival, comes to the aid of everyone regardless of social status and religious orientation. But this place is very pointedly low in the cosmo- logical hierarchies of both communities. For Hindus, he fits into the women's world of household concerns and is petitioned for what he can give for success in this life, not distantly future gains. For Muslims, he ironically idealizes the acquisition of wealth, the very thing that is often seen to hinder spiritual development among Sufis, but epitomizes the survival techniques valued by traders and rulers. It is perhaps here that the narratives have given us a clue regarding his insatiable desire for sirni, for even though Sufis have historically vacillated between fasting and satiety, poets have for centuries referred in derogatory terms to the halvah-sucking mendicant, "the Sufi with milk-white hair who has made the recollection (dhikr) of sugar, rice, and milk his special litany"44-the very ingredients of sirni. In these stories, filling this pir's stomach equates indirectly with being moral, for those are the two things he re- wards with wealth. He instructs in morality, but does not give explicit religious instruction to a following of spiritual adepts, disciples (murid or sisya), as we might expect. In this he is unlike his historical counter- parts. His vanity and thirst for recognition prompt him to keep his magi- cal powers on prominent display, and these are directed toward issues of pride and place, consistent with the desires of his lay following for mi- raculous intervention in life's demands. His niche has become secure at the bottom of both religious hierarchies as a metaphor for getting by in a tough world-and in this, while valuing Satya Pir's acts differently, these two orientations to authority appear so closely related as to be but complementary dimensions of one Bengali world.  Tony K. Stewart 1 49 Notes I would like to thank Richard J. Cohen, Margaret Mills, Robin C. Rinehart, Carl W. Ernst, and Edward C. Dimock Jr. for their close reading and suggestions. But a special thanks goes to David Gilmartin, with whom many of the ideas con- tained herein were fleshed out over months of sustained discussion. 1. Sanjivakumara Bagachi, Satyandrdyana: ndtya kdvya (Dinajapura: Kalipada Bagachi and Ranajit Kumara Bagachi, 1334 B.s. [1926-27]), preface. 2. Compared with other figures in premodern and contemporary Bengal, the amount of scholarship is grossly disproportionate to the manuscript and printed material devoted to him. When I first began this research in the mid-1980s, fewer than two hundred pages in all languages of academic writing had been focused on him, and nearly all of that was a simple recounting of his tales. Several new works have started to remedy the situation, however. The most complete ac- counting is in Girindranatha Dasa, Bangld pira sdhityera kathd (Kajipacja, Bara- sata, 24 Paraganas: Sehid Laibreri, 1383 B.S. [1975-76 C.E.]). The first extended set of translations has been published, all attributed to the Oriya poet Kavikarna; see Bishnupada Panda, ed. and trans., Pdlds of Sri Kavi Karna, Kalamalasastra Series, vols. 4-7 (New Delhi: Indira Gandhi National Centre for the Arts and Motilal Banarsidass, 1991). Some of the tales are in Bengali or in a mixed Oriya and Bengali idiom in manuscripts that are written in Bengali script, while many of the Oriya tales overlap with Bengali counterparts in plot and theme. Cashin has also included a translation of Vallabha's Satyandrdyanera punthi in his chap- ter on "The Cult of the Pir," in David Cashin, The Ocean of Love: Middle Bengali Sufi Literature and the Fakirs of Bengal, Skrifter utgivna av Freningen for Orien- taliska Studier no. 27 (Stockholm: Association of Oriental Studies, Stockholm University, 1995), 251-82. 3. See, e.g., Asim Roy, The Islamic Syncretistic Traditions of Bengal (Princeton: Princeton University Press, 1983), esp. 214-18, and his earlier article, "The Pir- Tradition: A Case Study in Islamic Syncretism in Traditional Bengal," in Images of Man: Religion and Historical Process in South Asia, ed. Fred W. Clothey (Madras: New Era, 1982), 112-34, esp. 129-32. See also the recent article by Kanai Lala Raya, "Satyapir," Barld EkdlemiPatrika 36, no. 2 (Sravana-Asvina 1399 B.S. [1991- 92 C.E.]): 71-82. For a critique of the concept of syncretism, see Tony K. Stewart and Carl W. Ernst, "Syncretism," in South Asian Folklore: An Encyclopaedia, ed. Margaret A. Mills and Peter J. Claus (New York: Garland, forthcoming). 4. For the litany of the scholarly constructions of India, see Ronald B. Inden, Imagining India (Oxford: Basil Blackwell, 1990). 5. This popular story is asserted by the editor in Kavivallabha, Satyandrdyana punthi, ed. Munsi Abdul Karim, Sahitya Parisad Granthavali no. 49 (Calcutta: Ban giya Sahitya Parirat by Ramakamala Simha, 1322 B.S. [1914-15 C.E.]), 7, and then repeated frequently in the secondary literature as "hearsay." The most ex- plicit connection is proposed by Louis Massignon, La Passion de Husayn Ibn  50 | Satya Pir on the Frontiers of Bengal Mansar Halldj, new ed., 2 vols. (Paris: Gallimard, 1975), 2:299-302. I have not seen any primary documentation of this association. The same goes for Satya Pir's identity as the son of the daughter of the famous ruler of Bengal, Husain Shah (r. 1493-1519), which is frequently repeated; for the earliest citation, see Dineshcandra Sen, The Folk Literature of Bengal (Calcutta: Calcutta University Press, 1920), 100, who credits manuscripts of Kavi Aripha and Sankaracarya. I have been unable to confirm the passage in any version of either text. 6. Sukumara Sena, Bandhlar sdhityera itihasa, 4 vols. in 6 pts. (Calcutta: Eastern, 1963), vol. 1, pt. 2: 471. 7. Krsnahari Dasa, Bacla satya pira o sandhydvati kanydra punthi (Calcutta: Nuraddin Ahmad at Gaosiya Laibreri, n.d.), 214. 8. Worship is an aniconic form that involves the heartfelt offering of sirni in a simple mixture of rice (or rice flour), sugar, milk, banana, and spices. During the first decade of the twentieth century, the Satyandrdyana o subhacanira kathd, 2d ed., ed. Syamacarana Kaviratna (Calcutta: by the editor through Gurudasa Cattopadhyaya at Bengal Medical Library, 1315 B.S. [1907-8 C.E.]), gave seven detailed pages of instruction just for making the offering of sirni, which now includes twenty-eight ingredients. In the same year, the Satyandrdyana vrata- kathd, edited with Bengali translation by Rasaviharisanikhyatirtha (Murshida- bad: Ramadeva Misra for Haribhaktipradyinisabha of Baharamapura at Rad- haramana Press, 1315 B.S. [1907-8 C.E.]) gives twelve pages of the same. A decade later, Ramagopala Raya's version, Satyamangala bd satyandrdyana devera vrataka- thd o pujdpaddhati (Calcutta: Jayakrsna Caudhuri, 1835 Saka), contains twenty- two detailed pages for performing the offering. In a book that was probably published during the 1970s, fifteen pages are devoted to the offering of the paja, including illustrations of thirteen hand madras; see Ratnesvara Tantrajyotisasas- tri, ed., $risrisatyandrdyana o subhacuni pajdpaddhati (Calcutta: Puspa, n.d.). 9. The pancali and pala gana are set to music for public performance, while the vrata katha is a story told at the time of women's household ritual vows, into which cycle Satya Pir has been incorporated. 10. For the most complete listing of these manuscripts, see Jatindra Mohan Bhattacharjee, Catalogus Catalogorum of Bengali Manuscripts, pt. 1 (Calcutta: Asi- atic Society, 1978). There are many other manuscripts in private hands and in collections whose catalogs had not been compiled when Jatindra Mohan com- piled his monumental catalog. 11. The Srihatta Sahitya Parirat in Bangladesh was razed in retaliation for the toppling of the Babri Masjid; there have been other unconfirmed reports of manuscripts being destroyed. 12. Gerard Genette, Narrative Discourse: An Essay in Method, trans. Jane Lewin (Ithaca: Cornell University Press, 1980). 13. In order to maximize the use of manuscripts, I generally read only com- plete versions of texts and no more than three versions by any one author (there were only minimal variations), and I surveyed as many authors as possible, starting with the oldest texts available, but I tried to maintain a balance of names  Tony K. Stewart 1 51 that appeared to represent the general distribution of "Hindu" and "Muslim" names. The latter proved to be misleading, for the names do not necessarily reflect the author's religious preference, confirming secondarily the inappropri- ate assumption of "naming" and "belonging" noted earlier. 14. Richard Eaton, The Rise of Islam and the Bengal Frontier, 1204-1760 (Berke- ley: University of California Press, 1993). 15. Occasionally the interlocutor will be Yudhisthira or some other celestial figure, e.g., the prominently titled work by Dvarakanatha Pala, Satyndrdyanera paculi: krsna yudhisthirera sambuda o kaldvatira updkhydna (Dhaka: Lachamana Basaka at Dhaka Bangla Press, n.d. [1285 B.S., or 1877-78 C.E.]). 16. As far as my survey can determine, the texts do not use the Bengali term Musalman until very late in the nineteenth century. And that term is itself not entirely unambiguous, but seems generally to refer to Muslims who are not of Arabian origin but who follow a culture rooted in Perso-Arabic ideals. A "muslim" is one who "submits" (aslama) to the will of God (Allah), and in the Indian context historically it refers to those who can trace their direct ancestry to the tribes of Arabia at the time of Muhammad in the seventh century. The word Musalman, while often confused with Muslim, more connotatively refers to those whose lineages originate in South Asia or outside of Arabia but who converted to Islam. For more on the term, see the authoritative Persian dictionary, Laghut nama by 'Ali Akbar Dikhuda, 15 vols. (Tehran: Mu'assasah-'i Intisharat va shap- e danishgah-i, 1993-95), s.v. "musalman" (fasc. 211, pp. 428-29, microfiche 113:1). I am indebted to Carl W. Ernst for first pointing out the potentially pejorative reading of Musalman and for the reference to Dikhuda. 17. For a very useful and pointed analysis of the nature of such comparisons, especially as they apply to religious situations, see Jonathan Z. Smith, "Adde Parvum Parvo Magnus Acervus Erit," History of Religions 11, no. 1 (August 1971): 67-90. Starting with the basic division of "we" vs. "they," he extends the formu- lation by (apparently) running it through the transformations dictated by the semiotic square. 18. Occasionally a fourth tale will be appended making the connection with the king explicit; it is the story of a king who loses his sons after failing to join the worship of Satya Pir by a group of cowherds he meets in the wilderness. See the previously noted work edited by Rasaviharisankhyatirtha (supra n. 8), who refers to it as the Tungadhvaja gopa samvdda; and Satyandrdyana vratakathd, com- piled by Meghanatha Bhattacarya (Calcutta: Saniskrta Press Depository, 1306 B.S. [1898-99 C.E.]), who calls it the Vansadhvaja gopa samvada. 19. See, e.g., the elegant tale of Vikrampura poet Lala Jayakrsna Sena, Harilild, ed. Dinesacandra Sena and Basantaranjana Raya (Calcutta: Calcutta University, 1928), who finished the text in 1772 (p. 7), and the powerful narrative of Ka- vivallabha in his aforementioned Satyandrdyana punthi (supra n. 5), which was composed earlier in the eighteenth century (p. 15). Both of these texts are sub- stantially larger than the standard Hindu trilogy taken as a whole. 20. These two texts are available in multiple Battala editions and have been  52 | Satya Pir on the Frontiers of Bengal printed together as many times as they have been issued separately. I have per- sonally examined more than fifty such publications. Typical among them are $arkaracarya and Ramesvara, $risrisatyandrdyanera pacali: lildvati kaldvati daridra brdhmanera updkhydna (Calcutta: Taracanda Dasa and Sons, n.d.); Sarikaracarya and Ramesvara, $risrisatyandrdyanera pacali: lildvati kaldvati daridra brdhmanera updkhydna (pajddravya pajdvidhi, dhydna o prandma sambalita), 3d ed., comp. and ed. Avinasacandra Mukhopadhyaya, rev. Surendranatha Bhattacarya (Calcutta: Calcutta Town Library by Karttika Candra Dhara, 1360 B.S. [1952-53 C.E.]); and $arkaracarya and Ramesvara, $risrisatyandrdyanera pacali: lildvati kaldvati daridra brdhmanera kdhini (pajddravddi o pajdvidhi sambalita), ed. Gaurangasundara Bhattacarya (Calcutta: Rajendra Library, n.d.). 21. For translations of different versions of these three tales from the Vaisnava vrata kathus, see Tony K. Stewart, "Satya Pir: Muslim Holy Man and Hindu God," in Religions of India in Practice, ed. Donald S. Lopez Jr. (Princeton: Princeton Uni- versity Press, 1995), 578-97; selections come from Sarikaracarya, Dvija Rama- bhadra, Bharatacandra Raya, and Ayodhyarama Kavicandra Raya. For the San- skrit versions and an analysis of their paja, see Gudrun Bnhnemann, "Examples of Occasional Pajds: Satyandrdyanavrata," in Pajd: A Study in Smarta Ritual, De Nobili Research Library Publications, vol. 15 (Vienna: Institute for Indology, University of Vienna, 1988), 200-213. For a contemporary version of the story and an account of the paja, see Anoop Chandola, The Way to True Worship: A Popular Story of Hinduism (Lanham, Md.: University Press of America, 1991). 22. The prefaces and introductions to a number of editions at the turn of the century document the widespread desire to Sanskritize the tradition. For ex- ample, in 1904, Priyanatha Ghosala complained bitterly about the sorry state of the textual materials and paja instruction. He attempted to remedy the situation by producing a clear and properly edited Sanskrit text and paddhati after consult- ing numerous printed texts and unpublished manuscripts; Priyanatha Ghorala, Satyandrdyana vratavyavasthd, pajdpaddhati o pancavidha mdhdtmyakathd (Calcutta: Patrick Press, 1310 B.S. [1902-3 C.E.]). The previously noted edition by Syama- carana Kaviratna (supra n. 8), with its twenty-eight sirni ingredients, was writ- ten with the express intention of eliminating the use of popular and misleading pdicali texts in a very self-conscious effort to clean up (Sanskritize) the tradition. In a different type of foliation, one author notes the injunction in the Skanda Purana to make music and dance (nrtyagitadikancaret) while offering the paja has prompted him to adapt the offering of sirni to a musical mode, including exten- sive musical notation, in the text. By his own admission, he also takes the oppor- tunity to "correct" common mistakes in theology in an effort to universalize the message; see Suranatha Bhattacarya, $risrisatyandrdyana vratakathd, with the ba- sic text, paja instruction, Bengali verse translation, and songs for accompani- ment (Calcutta: B.P.M.'s Press, n.d. [132? B.S.]). 23. It is interesting that the eastern reaches of the delta region have always provided last-ditch money-making opportunities for poor brahmanas, for the  Tony K. Stewart 1 53 dearth of brahmanas in the region puts their services at a premium; even Krsna Caitanya made the journey when his family was in financial straits. Being mo- mentarily itinerant in the region does not seem to overly affect the status of the brahmana, but residence in the region during this period does seem to compro- mise status, for most of Bengal sits outside the boundaries of madhyadesa, the traditional brahmanical homeland, and therefore lies beyond the reaches of civi- lization, a barbaric frontier; it is, then, the ideal place for a pir to exercise his power. 24. It should be noted that the woodcutters' tale is always the shortest of the set, often reduced to a few lines, yet never eliminated completely, apparently because it is necessary to complete the progressive appropriation of Satya Pir following the logic of the need to domesticate the land in stages. The extension of habitable land, then, includes social, economic, agricultural, and, with the concern for kings and righteous rule demonstrated in the final tale and other non-Vaisnava versions, political dimensions. 25. This eastward push parallels the eastward shift of the Ganga River. But it contrasts with the next category of "Muslim" tales, which articulate a different frontier, and one where they are already present when the Vairnavas arrive, which affects their narrative strategy or code. 26. Sukumara Sena, Bd'hldr sdhityera itihdsa, vol. 1, pt. 2: 474-75. 27. Sukumara Sena completely ignores the woodcutters' tale, while declaring the merchant's tale to be an unimaginative recapitulation of the Dhanapati khullana in the Candi mangala; ibid., 471. The merchant's tale is indeed suffi- ciently close to be called a variant, but the question of historical priority-that is, whether Satya Pir's story or Cancji's story is earliest-is never considered. 28. And we can easily add the Natha yogi and popular (but in Bengal, not necessarily Sufi) dervish to this set. 29. I have been unable to locate a single instance of a historical pir or other Muslim figure being appropriated by the Bengali Hindu traditions; all other adoptions have been mythic or legendary figures. For a summary of some of these important figures, see Girindranatha Dasa, Bungld pira sdhityera kathd, which gives stories of thirty-three historical pirs and nine legendary pirs in Ben- gal during the nineteenth and early twentieth centuries. 30. Ronald B. Inden has argued that in previous centuries the genealogical histories include several mythic episodes for the royal importation of brah- manas with a proper Vedic knowledge to people the land and make it properly habitable; the last of these kings fades into the historical figure of Vallala Sena. See Inden, Marriage and Rank in Bengali Culture (Berkeley: University of Califor- nia Press, 1976), 49-82. It should be noted that Hindu Bengal has been, including in the myths, a two-varia society, composed of brahmanas and sadras. 31. Ironically, but not at all surprisingly, the scholarship from Hindu nation- alists treats Satya Pir essentially the same way, perhaps taking its cue from the obvious use to which he is put in the society.  54 | Satya Pir on the Frontiers of Bengal 32. Krsnahari Dasa, Bada satya pira o sandhydvati kanydra punthi, 214. 33. Lokamana Hakim can be identified as Lugman Hakim, the legendary Arab sage who is mentioned in the Qur'an (31:11-19). 34. The Jharikhancja Forest traditionally extends through the wild regions of southwest Bengal, south of Visnupura, and runs into the Midnapur districts and the northern reaches of Orissa and southeastern Bihar on part of the Chotan- agpur plateau. Much of it remains a frontier today. 35. See Tony K. Stewart, "Surprising Bedfellows: Vaisnava and Shi'a Alliance in Kavi Ariph's 'Tale of Lalmon,"' International Journal of Hindu Studies (forth- coming). 36. Dinesh Candra Sen sees this as corroborative evidence to the report of Satya Pir being the son of the daughter of Husain Shah; see his Folk Literature of Bengal, 102-3. 37. Krsnahari Dasa, Bada satya pir, 214-16. 38. Dinesh Candra Sen pronounces Krsnahari Dasa unequivocally to be a Muslim, in spite of his name (Folk Literature of Bengal, 101 ff.); based on the con- tent of the narrative and the direction of the action, I am inclined to agree with the general orientation (with the proviso previously stated regarding inclusion in a group). Girindranatha Dasa refers to him as a Baul-Daravesa [=Dervish] (Bdngld pira sdhityera kathd, 470), which is, of course, simply another form of syn- cretism. It is perhaps significant that the woodcuts in the printed version show the image of Satya Pir very much in the mode of a Baul, but given the mixing of sectarian emblems of hair, clothing, bag, and so forth, it might well be that Satya Pir proves to be more the prototype for the Baul than the other way around, i.e., he is a generic "holy man" with associations of samnyasi, vairagi, dervish, pir, fakir, etc. 39. Krsnahari Dasa, Bacla satya pir, 206-14. 40. This is sirni, although he does not call it that in this passage, for then it might well be interpreted by the antagonists as an offering, and that would be premature for its place in the plot. But the refusal to give him such food is a direct refusal to offer worship to Satya Pir, and that is an offense that requires punish- ment. 41. This is a variant of the old rope trick. 42. This is the rope trick. 43. It is perhaps ironic that it is the exclusive intolerance of the brahmana that is brought into question, rather than that of the Muslim, who is all too frequently characterized in scholarship and the contemporary press by that charge. 44. Schimmel, Mystical Dimensions of Islam, 117, quoting 'Abal-Majd Majdad Sana'i's Hadiqat al-haqiqat wa shari't at-tariqat.  2 Beyond Turk and Hindu Crossing the Boundaries in Indo-Muslim Romance Christopher Shackle If there is a discrete South Asian context shaped by geography as well as by history, then the Panjab resembles other parts of the subcontinent while also projecting the cultural influence of its own topographical fea- tures. Above all, we are concerned with cultural geography, and it is our argument that the topographical features of the Panjab provide a back- drop that fosters a strategic tension, otherwise seen as a fluidity of meta- phor, that characterizes the literature of this region at the same time that it influences Indo-Muslim identity. Dichotomies and Unions What are those features of Indo-Muslim identity within the premodern world of Islam that have been shaped by cultural geography? At one level, they embrace physical setting and material culture, yet they also include that great range of nonmaterial phenomena which, whether as customs and attitudes or as languages and legends, form everyday cul- ture, and it is the interplay of physical setting with its connotative, ev- eryday expression that helps make Indo-Muslim identity at once differ- ent from, but allied with, its counterpart: Indic identity.1 Above all, identity is shaped by what Tony Stewart has called "prag- matic concerns of survival," and while these concerns, in his apt words, "cross whatever imaginary divide we construct between Hindu and Muslim," they, in turn, are shaped by class markers and expectations. In India, and especially in the Panjab, the cultural disjunctions between  56 | Crossing the Boundaries in Indo-Muslim Romance elite and native, shari f and desi, are too marked to be ignored. Nowhere are they more apparent than in the area of language, with its exception- ally marked cultural diglossia between Persian as a widely used elite standard language, both imported and pan-Indian, and localized Indo- Aryan or Dravidian languages. Immediately associated with this contrast is the presence of two types of creative literature. The cultivation of Persian poetry was always a central marker of the cultural identity of the elite,2 whose efforts to dis- tance themselves from Indic cultural associations led to the formation of an elaborately self-contained symbolic system underlying the inter- linked genres of gasida, ghazal, and masnavi. Such purism was not lim- ited to Muslim elites; it also extended to the practice of Hindu poets from the Persianizing classes.3 Increasingly from the late premodern period, we find more popular types of lyric and narrative poetry being culti- vated by Muslim as well as non-Muslim authors in indigenous lan- guages. Such verse proliferates in styles sometimes entirely indigenous and at other times a blend of these with elements from the Persian tradi- tion. In other words, the wide-ranging and profoundly differentiated views of the premodern period are not associated with the Hindu-Mus- lim divide itself; they are marked more by class than by creedal separa- tions.4 And among the genres of poetry that attempt to cross class boundaries we also find crucial examples of the interplay between Mus- lim and Hindu sensibilities. None is more intensive than Panjabi love lyrics. It is in these lyrics that we find the theme of love between Muslim and Hindu (or between Turk and Hindu). Such love is at once transgres- sive and assimilative, for at the same point that Panjabi poets highlight it as illicit love, they also undermine the very categories Muslim and Hindu as oppositional or incommensurate. The Transcendence of the Lyrical Setting Annemarie Schimmel has explored the opposition of Turk and Hindu in classical Persian poetry,5 but it takes a different turn in South Asia pre- cisely because of the cross-class and cross-creedal intensity of flowering regional verse, such as the Panjabi love lyrics that are our principal sub- ject. Consider the lyrics of the Qadiri poet Bullhe Shah (1680-1758). In the famous kdfi whose refrain asks, "Bullha kih jdndn main kaun?" (Bullha, how should I know who I am?) the poet answers his own ques-  Christopher Shackle | 57 tion in one verse. Its rhymes tumble, in English as also in the original Panjabi: Neither Arab am I nor man of Lahore Nor Indian from the town of Nagaur Neither Hindu am I nor Turk of Peshawar.6 With a complementary appropriateness, the same simple idea recurs in Bullhe Shah's poems composed in a "Hindi" style reflecting Hindu reli- gious vocabulary,7 as in the kafi whose refrain opens, "Hindu, no! nor Musalman," with the final verse: Bullha, once God filled my thoughts, Hindus, Turks, I quit both sorts. Through their continually vivid repetition of such fundamentals,8 Bullhe Shah and the other great local Sufi lyricists of the later Mughal period, like their nirgun bhakti and Sikh predecessors and contemporar- ies from the other side of the Muslim-Hindu divide, have continued as mother tongue literary classics, but they have also molded in the Panjab a diffuse conception of South Asian religious identity that is as im- mensely influential as it is dimly understood. A major difficulty is translation, not just from Panjabi to English but from the idiom of the aorist tense, which denotes past continuous action, without limits in either time or space, to the unmarked past tense in English. In the original Panjabi, both the above quotations from Bullhe Shah confirm that the simplest of truths are aorist in their expression. Once the inner meaninglessness of outward religious and social distinc- tions is grasped, it can be grasped forever, and the preferred practice is to internalize such insights repetitively and permanently through song.9 The message and its form reinforce each other: Just as the lyric naturally elides with the aorist tense, so the primary status of the lyric in the pro- vincial as well as the Persianate literatures of South Asia derives from this intrinsic timelessness of the genre.10 It is particularly suitable for singing, which makes its rhyming verses even easier to memorize. Ironically, it is just this universalizing quality of timelessness which makes the Sufi lyric an ideal justification for academic generalizations about Islam, generalizations that dwell on the eternal constancy of its irenic strands.11 What is important to note, however, is its originary im- pulse: not that it is eternal but that it evokes the eternal or the timeless by its ability to move between different registers of verse and connotation.  58 | Crossing the Boundaries in Indo-Muslim Romance Without actually doing so, it seems to cross all dividing lines-time and place, creed and class. It is this boundary crossing that gives Sufi poetry its power in shaping a distinctive language of identity. The existence of worldly identities, rooted in the realities of everyday life, is essential for its structure and message. These identities-whether of religion, ethnicity, occupation, or class-provide the critical backdrop that creates the literary form. The evocation of these identities is in many respects central to the evocation of place, to the distinctive cultural geography that shapes an image of the Panjab. And yet it is central to the genre-and to the qissa, the Panjabi love lyrics of which we will have more to say-that such identities are repeatedly transgressed by love. Indeed, it is the constant tension be- tween these that defines a framework for understanding identities, a framework whose contours can only be understood with the generic contours of this literary production. The interplay of these elements in defining a distinctive regional ethos was suggested in Persian verse by the Nairang-e Ishq (1683) by Ghanimat of Kunjah near Gujrat, which became the most popular mas- navi to have been produced by a provincial Persian poet from the Panjab.12 Why?' Aziz (Noble), son of the local governor, falls in love with Shahid (Beauty), the orphaned child of poor parents who has been trained by a troupe of traveling entertainers (tava'if). 'Aziz confounds all expectations of class and custom by installing Shahid as his live-in partner. Shahid is even described in such tantalizingly gender-neutral fashion that her feminine identity only becomes apparent to the reader after she has finally left 'Aziz. Their separation occurs on a hunting trip, where Shahid meets and falls instantly in love with the handsome coun- try lad Vafa (Fidelity). After a succession of events perhaps too swift not to have been drawn from life, Beauty marries Fidelity. Deftly the poet downplays this crossing of the urban-rural divide, even though it would have been evident to all listeners. Instead, he focuses on the affair be- tween the noble born (sharifzada) and the dancing girl (ta'ifa). Yet the verse itself makes clear the extent to which Ghanimat's world becomes very much one of town rather than country, nearer to the Hira Mandi than to Hir, in the grace of its opening invocation of the Panjab as "land of love": No land so irresistible I've seen None matches fair Panjab's delightful scene ...  Christopher Shackle | 59 To glimpse Panjab its only aim- Kashmir at heart is turned to shame ... In all its cities beauties throng the mart In eagerness to buy a lover's heart.13 This prologue undoubtedly had a special resonance for the poem's original local audience,14 but by the mid-nineteenth century, when Nairang-e Ishq, along with so much Persian poetry, was translated into Urdu, it added no more than an exotic touch to the opening pages of the version, less graceful both in its meter and in its enforced preservation of the masculine gender for Beauty, which was produced far to the east in Avadh by Bhagvant Ra'e "Rahat" of Kakori as the Nigdristan-e Ulfat (1852): The land of Panjab is so fair- Canals with water flow there. A country so cool and so clear Whose breezes chill even Kashmir. Its beauties wait ready to take The goods of both Brahmin and Shaikh.15 There is a somewhat pallid exoticism here, perhaps to be expected in a fairly faithful adaptation. But in its exhibition of this quality, Rahat's "Gallery of Intimacy" typifies the story of the Urdu masnavi in northern India. Apart from copies and translation, it is dominated by the reli- giously neutral fairy-tale mode of its only acknowledged masterpieces, the Sihr ul Baydn of Mir Hasan (d. 1786) and the Gulzar-e Nasim (1833) of Pandit Daya Shankar Nasim (d. 1843).16 The Lovers' Inventory I draw attention to the fate of one lyric in later Urdu rendition, in con- trast to its original Persian/Panjabi composition, in order to demon- strate how place, though no more than a seeming backdrop, is nonethe- less crucial to the everyday expectations of poet and listeners alike. Few genres show a more powerful attachment to the specificities of place than the Panjabi verse romance called the qissa (plural qisse),17 which was largely the creation of Panjabi Muslim poets.18 Here the attachment to place supersedes not only class and creed but also gender markings. Sometimes the Panjabi kafis revolve around key aspects of individual  60 | Crossing the Boundaries in Indo-Muslim Romance stories, giving ample scope for the poet to make affecting use of the well- known Indic preference for the feminine persona as the voice of the lyric. Elsewhere, use is made of the favorite topos of the "lovers' inventory," in which whole lists of lovers are lined up as collective testimony to the power of love. Like most such topoi, this has its classic Persian exem- plars, but it brings to the fore one obvious criterion for distinguishing Indo-Muslim literatures from those of other parts of the Islamic world- their use of local stories. Whereas the Urdu ghazal relies for its characters on the old stories that were given their definitive narrative shape in the Persian masnavis of Nizami (d. 1199), Khusrau (d. 1325), and Jami (d. 1494), the Panjabi kafi is characterized by an abundance of references to local romantic legends. The lovers' inventory, while grounding Panjabi stories in a literary genealogy suggesting the power of love as a defining civilizational frame, underscored as well each story's specific place and context. As always, Bullhe Shah provides a particularly long and fine ex- ample, embracing both Islamic and Indic mythology, as well as some of the most famous local lovers, in Kafi no. 65. First love the mighty came on Hir And then her Ranjha pierced his ear To wed his Sahiban so dear Was Mirza sacrificed. Losing Sassi in the desert hot Drowning Sohni on her unbaked pot Love for Rod -a did destruction plot- It had him chopped and sliced. Such lovers' inventories, moved from the international sets found in the Persian masnavi to include local tales, are found in the Panjabi qissa; as I have suggested elsewhere, these came by the later Mughal period to be a more or less consciously articulated part of a region's "vernacular literature capable of re-articulating the Muslim identity of its inhabit- ants in local terms by drawing on the deepest range of their cultural roots."19 Although they could become a rather mechanical device in the qisse of inferior later poets, they continued to be used to powerful effect in the works of the best poets, even into the nineteenth century. A good example is provided by that last master of the qissa genre, the great Panjabi poet and Qadiri saint Miyan Muhammad Bakhsh (1830-1907) of Khari in the Mirpur district of Jammu, now in Azad Kashmir.20 With a fine symbolic appropriateness, the completion of his first mas-  Christopher Shackle | 61 terpiece almost exactly coincided with the British defeat of the rebels outside Delhi at Badli ki Sarai on 8 June 1857, which won them the Ridge whence their catastrophic assault on the old capital of the Indo-Muslim world was soon to be launched. This was Miyan Muhammad's qissa on the story of Sohni Mahinval, finished on the afternoon of Wednes- day, 12 Shavval A.H. 1273. The section of its prologue devoted to descrip- tion of the havoc-wreaking workings of love contains the following finely planned inventory: Love caused the sight of Laila's face to be the cause of Majnnn's pain It used Shirin's sweet lips to steal Parvez's heart and kill Farhad From Ynsuf's coat too it displayed a dream to snatch Zulaikha's heart Jalali's fire made Rocla burn, love dazzled Raijha with Hir's flash Beneath bright Chandarbadan's sun, Ma'yar's green garden turned to ash By Sohni's spark from this same fire was Mahinval reduced to ash Encamped within illusion dwells the essence of the Absolute.21 Here, as often in such inventories, the primary reference is to the uni- versal romances of the Persianate world, whose three most famous rep- resentatives are continually exploited core elements of the poetic lan- guage of all genres. Collectively they embrace the three symbolic worlds of Arabia with the Laila-Majnun story, of Iran with Shirin's rival lovers Parvez and Farhad, and of the Qur'an with the tale of Yusuf and Zu- laikha, as mediated through Jami's version.22 Yet these three great ro- mances-even disregarding all their subsidiary characters and sub- plots-embrace almost every conceivable kind of social as well as psychic boundary crossing, and it is this suspension of the everyday which evokes their unusual force. It also mirrors the madness of love, which may be exacerbated by tribal rivalry, as in the case of Laila and Majnun, or abetted by the disparities of status between technician and prince, as happened for Farhad in his rivalry over Shirin with Khusrau Parvez, or challenged by social and religious, sexual and spiritual hier- archies, which must have separated a passionate pagan lady and chastely enslaved prophet in the tale of Zulaikha and Yusuf. If it is this supreme transregional trinity which provides the arche- typal frame for the modeling of the local romances, the latter-in all but  62 | Crossing the Boundaries in Indo-Muslim Romance the feeblest adaptations-also preserve an independent vitality by vir- tue of their separate local origin, and while one must stress both continu- ity and change, it is important to note how pragmatic concerns for ev- eryday vitality constitute the heart of the literary "matter of Panjab." Like most things Indo-Islamic, these local romances participate pro- foundly in both the Islamic and the Indic worlds but also emerge as entities sui generis independent of either precisely because the everyday world is not lost but heightened in their nimble narratives. At the same time, of course, the lovers' inventory also sets up a ten- sion between the universal claims of the great romances of the Islamicate world and the more localized significance of the settings tied to the Panjab. Many poets of the first order of creativity ignored local stories in their lists in favor of the international romances of the Persianate world. The very embeddedness of Panjabi stories in the local and everyday life made them not only less desirable for some elites but also less accessible to neat categorization in such lovers' inventories. That these lists were far from being fully standardized is suggested by the differences be- tween Bullhe Shah's list and Miyan Muhammad's. But it is significant that both these inventories did include several stories with Panjabi set- tings along with those of more universal Islamicate provenance, includ- ing the story of Sohni and Mahinval; the tale of Mirza and Saliban, typo- logically important for its bridging of the romantic and the heroic;23 and the story of Sassi and Punnnn , the only one of these four with a non- Panjabi setting.24 Perhaps most prominent generally in such inventories was the story of Hir and Ra jha. Lying at the very heart of the local romantic canon, this story furnishes much of Bullhe Shah's symbolism. Crossing the Boundaries First, let us consider the hero of these romances. All have a Muslim youth as their hero, the Ego of their psychic universe. The hero is defined through his adventures, inspired by a love that must involve suffering even if it does not end in tragedy. Both the object of the hero's love and the secondary characters with whom his adventures bring him into con- tact collectively include many varieties of Other, different from the hero in sex, status, age, origin, or belief (or in various combinations of these). The hero's quest for his beloved and his encounters with secondary characters accordingly involves him in the crossing of boundaries, lead- ing to at least partial loss of his initial identity-typically through his  Christopher Shackle | 63 becoming a faqir or yogi-and his partial assimilation of a different iden- tity. This new identity at least prepares him for his destined union with the Other, which is as likely as not to be finally achieved only after the further loss of identity consequent upon physical death of both Ego and Other. This shift in register to underscore the value of fana, or annihila- tion of Ego, is especially crucial for explicitly Sufi poets like Miyan Mu- hammad, but Sufi categories also appear in other qisse, even those by non-Muslim poets. What are the typical implications of these changes for the hero's iden- tity as a Muslim? While an additional poignancy may be conveyed to a romance by a religious incompatibility between the lovers, Muslim- Hindu relationships are no more central to most of the Panjabi romances than is the contrast between Turk and Hindu to most Indo-Persian po- etry, or indeed than romantic connections with Muslims would appear to be in the largely Hindu-inspired narrative poetry of classical Hindi literature.25 Differences of status or origin within Indo-Muslim society are of greater concern, and the recurring dynamism of the poetry is its ability to reflect the tension within sharp class divisions without violat- ing their actual boundaries. Fixed identities are underscored even as they are transcended by the structure of the genre. Consider the most popular story of the premodern Panjabi Muslim romance, Hir and Rarijha, which reaches its climax in the great qissa by Varis Shah.26 When the halfway point of the story is reached with Hir's marriage to Saida the Khera, Rarjha has to undergo the usual hero's transformation into an ascetic before he will be able to win her back. A Muslim pir may have played this role, but the Muslims of premodern India had an even more striking model of asceticism across the religious frontier in the yogis of Hinduism,27 and it is into a yogi that Rarijha is regularly described as having been transformed. Already in the earliest extant Panjabi version, supposedly composed before 1650 by Damodar Gulati, Rarjha follows Hir's written instructions and seeks initiation from the spiritual chief of the great center of the Gorakhnathi Kanphat ("split-eared") yogis on the lofty summit of the Tilla Jogian in District Jhelum.28 In the description of their straightforward encounter,29 ample employment is made of that useful stylistic device always so readily available to writers in chronically multilingual India. The Hindified ex- pressions put into Ran jha's mouth demonstrate his readiness for yogic initiation: Then Ran jha made entreaties with humility:  64 | Crossing the Boundaries in Indo-Muslim Romance From you, whose favor is salvation's guarantee, For yoga I ask like Gopichand and Bharthari.30 In Varis Shah, of course, the implications of all this are magnificently teased out,31 and along with a dalliance across religious boundaries he introduces a sexual undercurrent to this supposedly ascetic episode. It is given expression by Balnath's jealous disciples: shocked to see the favor- able impression Ran jha has made, they exclaim: "A taste for boys affects those yogis, whose wits God has confounded!" Whatever else its multiple implications, Ran jha's yogic transforma- tion is not about actual conversion, any more than the pretended hostili- ties of his subsequent courtship of Hir are literally about interfaith rela- tions, although they may be taken as such metaphorically. At least for the period from 1650 to 1850, when the Panjabi qissa tradition was at its most creative, the literary evidence would not suggest that relations of this kind were of such profoundly overriding concern to Muslims of the Panjab that they needed to be explored through creative writing. There is only one well-known Panjabi qissa from the period that does explore this theme explicitly, the exception that perhaps proves the rule. This is the tale of Chandarbadan and Ma'yar, frequently cited in the inventories of other qisse. It is exceptional not only in its theme but also, tellingly, in that its origin is Indian but not Panjabi; its provenance is the Deccan. The story was first treated as the report of an incident that actu- ally occurred in the time of Ibrahim 'Adilshah (1580-1626) in the artless Dakani Urdu masnavi by Muqimi (d. c. 1665), where the lover is called Mahyar.32 Its contents may be conveniently recalled in an adaptation of Schimmel's deft summary: A Muslim merchant, Mahyar, fell in love with Chandarbadan, the daughter of a Hindu raja. When she undertook a pilgrimage to Kadrikot he confessed his love to her. Being rejected, he spent a whole year as a hermit in the jungle. The following year, when Chandarbadan visited the temple again, he cast himself at her feet; but she turned away, amazed that he was still alive. Dismayed, he commit- ted suicide. Ibrahim ordered that Mahyar should be given an honorable burial. The funeral procession with the bier stopped at the princess's mansion and could not move farther. Deeply moved by Mahyar's love which lasted even beyond the grave, Chandarbadan embraced Islam, clad herself in pure white garments, and placed herself beside him on the bier. The two lovers were buried together.33 "By no means an out- standing work of art," as Schimmel charitably remarks, Muqimi's poem is chiefly remarkable for its naivete. Only occasionally does this perhaps  Christopher Shackle | 65 bestow a certain pathos, as on Chandarbadan's tentative first steps to Islam: She said: "How should I purify myself? I know not how to purify myself."34 Transmitted to the northwest by one of those routes which always look as if they should be far easier to document than in fact turns out to be the case,35 the Chandarbadan story resurfaced in Panjabi, perhaps receiving its first full-length treatment in the qissa by the prolific Al.mad Yar (d. 1848).36 The most popular published version, however, was that by Imam Bakhsh (d. 1863).37 Chandarbadan is firmly located in Hindu India as the daughter of Rangapati, raja of Patna! Ma'yar, still a wealthy merchant's son in spite of the slight change to his name, first sees her in a picture painted by an artist friend who glimpsed her on a visit to a temple. It takes him a year to come before her. By now a faqir who rejects her offers of money, he entreats her as a humble Farhad to her royal Shirin. She, however, will have none of him on religious grounds: "I am a Hindu, you a Muslim-we have no ties of faith." To which he retorts: "Love cares not for attributes, nor lovers for creeds and faith, 0 queen!"38 There are many more exchanges of this type, whose frequency in most types of premodern Indian literature is actually far less than the predilections of many modern historians sometimes lead them to sug- gest. The story follows a somewhat tangled course, partly because of well-meaning interventions by the Muslim king who adopts Ma'yar as his pet madman; it involves various yearlong journeys from one distant city to another. Finally, on the insistent prompting of Ma'yar's stationary bier, Chandarbadan is granted the doubly joyous fate prescribed for Hindu lovers in the Muslim romance: conversion and burial with her beloved. As the poet observes: The whole world knows that it will never do, 0 king, to couple Muslim with Hindu.39 From Mughal to Mahinval The Chandarbadan story expands the spectrum of Panjabi romantic verse, but in the premodern period its subject matter was nowhere near as compelling as that of the principal romances of the Panjab. Why? Because the geographical universe of these core romances is symboli-  66 | Crossing the Boundaries in Indo-Muslim Romance cally established by the river glades where Hir first comes to Ranjha and the Tilla where he becomes a yogi, and then almost equally important, by the desert sands in which Sassi dies in search of Punnnn , and by the rivers that give the Panjab its name. In this context, the greatest of these is the Chenab, the Panjab's great "river of love," and the love story to which it is most central, the legend of Sohni and Mahinval. Although this romance furnishes significant symbolical material to earlier lyric poetry, notably to Bullhe Shah,40 full-length versions do not seem to predate 1800.41 Of the pre-1850 versions, Ahmad Yar is said to have composed a qissa on this theme, as did Hashim Shah and Qadir Yar. All these, however, were completely overtaken in popularity by the inflated version produced in 1849 by Fa2al Shah (1828-90).42 Besides be- ing a gifted wordsmith adept at the flashy verbal effects so favored by Panjabi audiences, Fa2al Shah was fortunate in his close association with Lahore at a period when the rise of the publishing industry in that city was to reduce most other centers of Panjabi literary creation to relative insignificance. His version of Sohni Mahinval thus came to enjoy an immense reputation in the late nineteenth century. The classic story line common to all these versions tells how 'Izzat Beg (Sir Noble), the son of a wealthy merchant in Bukhara, comes to Delhi with a caravan of goods to trade. On its return trip, the caravan halts at Gujrat by the Chenab. There 'Izzat Beg falls in love with Sohni (Beautiful), the lovely daughter of Tulla the potter. Bidding his compan- ions farewell, he spends all his money buying pots so as to have a pretext to keep on seeing her, until he is forced by poverty into looking after Tulla's buffaloes as Mahirval (Herdsman). He becomes a faqir, with a cell out in the wilderness. Sohni slips out to see him there, using a pot as a float to get her across the river. Mahirval feeds her fish kebabs, for which once-when bad weather makes fishing impossible-he offers a piece of his own thigh. Eventually Sohni's sister-in-law discovers her secret and substitutes an unbaked pot. Unwittingly using this to cross on the following wild and stormy winter's night, Sohni is drowned. As soon as he hears this dreadful news, Mahinval plunges in to join her in death. Such climactic episodes act as the focus of attention for most of the qissa writers, along with the lyricists and the artists who produced vi- sual representations of the story. In that qissa of 1857 by Miyan Mu- hammad, however, with which all this exploration of an inventory began, other emphases are also notably at work. Like all Miyani Muham-  Christopher Shackle | 67 mad's work, his Sohni Mahirval has a meditative quality entirely in keeping with its profoundly Sufi focus. Along with this overtly mystical dimension, there also emerges one that appears to address one of the essential issues of Indo-Muslim identity: the relation between the maj- esty of an authority that has come from abroad and the realities of indig- enous existence. This relation between foreigner and native is typically visualized as one between male and female, which may of course have its own overtones for modern Indian critics,43 but equally crucial is the way in which the theme can be updated and criss-crossed from one his- torical context to another. Although this theme was certainly available to earlier Indo-Muslim romance, the Chandarbadan story chose instead to become stuck with Ma'yar's bier. Only when Indo-Muslim political authority was defini- tively lost, as it had been in the Panjab and Kashmir, first to the Sikhs and Dogras, then to the British, did it perhaps become possible to begin un- dertaking a fuller imaginative exploration of how the actual history of Muslims in the Panjab reflected, or deflected, their present and future identity. It may be therefore plausibly argued that the Sohni Mahirval story emerged into full popularity at just that time in the mid-nineteenth century because its shift in register suited the crisis of declining Mughal elites. Although Miyani Muhammad's version is too rich a poem to be prop- erly analyzed here, some impression of its scope may be conveyed through sketching the hero's progress from Mirza 'Izzat Beg the wealthy Mughal to the humble Mahiival. Along with frequent prophetic glimpses of what is to come and retrospective glances at what has occurred ear- lier in the story, besides a continual swirl of lovers' inventories around its every point, one of the features that helps to make Miyari Muham- mad's poem a far more substantial creation than might be indicated by its relatively brief physical length is the solidity of its geographical set- ting. After 'Izzat Beg has left Delhi, city of royal palaces and site of the holy tomb of Khwaja Nizam ud Din (the Chishti pir and paragon of Indo-Muslim piety), he comes to Lahore, where he visits the tomb of Jahangir and enjoys its splendid gardens. He then approaches the Che- nab (24-25): The Mirza passed the Beas and Ravi, brought to the Chenab by fate Whose magic waters steal all sense, and love's delirium instate.  68 | Crossing the Boundaries in Indo-Muslim Romance This is where his fate is to be decided (25): The waters of the Chenab-crossing made the noble Mughal ponder. For it is here that Sohni's chastity has been preserved for him (27): Who'd touch the Mughal's trust? What God had kept fault-free and pure? When he meets Sohni in Gujrat, he abandons his home outside India forever (34): The Mughal, slain by love, cast off his Balkh-Bukhara for Gujrat. The cost of loving her is the sacrifice of all his former status and wealth (36): The Mughal now was Mahinval, once rich, impoverished by love. Even his name and title are lost in this fated demotion from the ranks of the foreign-born ashraf (38): Once Mirza 'Izzat Beg, now victim of her eyes, her Mahinval. There is, of course, much more to this complex narrative, which abounds with subtleties of detail.44 Overall, however, it might be said that Miyan Muhammad's achievement in his richly reflective picture of the changes undergone by Mahinval is to have shown how Grandeur is fated to be lost through the pull of the Other, leading to Disempower- ment, even to Annihilation (fana). In its own way, this is a spectacular attempt to wrestle with the ambiguities and perplexities of premodern Indo-Muslim identity, and it is in the multifaceted appeal of this as a story both local and universal that the Panjabi romantic verse can be said to lie enduringly at the heart of the "matter of Panjab." Notes 1. The use of everyday here is suggestive of an approach counter to Freudian and Marxist analyses alike. It elides more closely with the reflections of Bakhtin, or at least the early Bakhtin, who argued that time and the world become histori- cal not just in the festival but in the tension between the anticipation of the fes- tival and the commonplace expectations of the everyday; in premodern India, as in premodern Europe, poetry bridges the register of the carnival and the ha-  Christopher Shackle | 69 bitual. For a general discussion of some of these issues, but from an entirely Eurocentric viewpoint, see Peter Osborne, The Politics of Time: Modernity and Avant-Garde (London: Verso, 1995), esp. chap. 5. 2. See Francis Robinson, "Perso-Islamic Culture in India from the Seven- teenth to the Early Twentieth Century," in Robert L. Canfield, ed., Turko-Persia in Historical Perspective (Cambridge: Cambridge University Press), esp. 106-7. 3. Most recently surveyed in Sa'id 'Abdullah, Adabiydt-e farsi dar miydn-e hinduvan, trans. from Urdu into Persian by Muhammad Aslam Khan (Tehran, 1992). 4. See Carl W. Ernst, Eternal Garden: Mysticism, History, and Politics at a South Asian Sufi Center (Albany: State University of New York Press, 1992), 22-37. 5. Annemarie Schimmel, "Turk and Hindu: A Poetical Image and Its Applica- tion to Historical Fact," in Speros Vryonis, ed., Islam and Cultural Change in the Middle Ages (Wiesbaden: Otto Harrassowitz, 1975), 107-26. 6. Kulliyydt-e Bullhe Shdh, ed. Faqir Muhammad Faqir (Lahore: Panjabi Adabi Academy, 1960), kafi no. 27. Due to the limits of space, and in order to project the broader points of this theme, I have not included the original Panjabi verse. For those who seek the text, either consult the just cited edition or contact me at the School of Oriental and African Studies. 7. For further examples, see Denis Matringe, "Krsmaite and Nath Elements in the Poetry of the Eighteenth-Century Panjabi Safi Bullhe Sah," in R. S. McGregor, ed., Devotional Literature in South Asia: Current Research, 1985-1988 (Cambridge: Cambridge University Press, 1992), 190-206. 8. It may, however, be observed that secondary sources reflecting twentieth- century enthusiasms for earlier statements of communal harmony (see note 21 below) make it easy to overestimate the actual frequency of such repetitions. In Bullhe Shah, for instance, other notable instances of "neither Turk nor Hindu" are substantially confined to isolated verses in kafi nos. 21, 22, 48, 90, 95,118, and 120. 9. Here, as Annemarie Schimmel's many wide-ranging surveys have so well demonstrated, it must be remembered that negations of formal religious differ- ences have always been characteristic of Sufi poetry throughout the Islamic world, so that the negation of the Turk-Hindu distinction is no more than an expansion-albeit one with a particular local relevance-of such already well established denials of difference as the equations between Ka'ba and idol- temple. See the Urdu Sufi verses cited in Christopher Shackle, "Urdu as a Side- line: The Poetry of Khwaja Ghulam Farid," in Shackle, ed., Urdu and Muslim South Asia: Studies in Honour of Ralph Russell (London: School of Oriental and African Studies, 1989), 82-83. 10. For the general dominance of the lyric in Asian poetic systems, see Earl Miner, Comparative Poetics: An Intercultural Essay on Theories of Literature (Prince- ton: Princeton University Press, 1990), esp. 9, 82. 11. This does little justice to the Sufi lyricists' emphasis, in full conformity  70 | Crossing the Boundaries in Indo-Muslim Romance with universal human experience, that the transcendence of otherness is inher- ently fraught with pain. A particular role here is played by the interpretations of Hindu writers, especially by Lajwanti Rama Krishna, Panjdbi Safi Poets, A.D. 1460-1900 (Calcutta: Oxford University Press, 1938), a doctoral thesis long over- due for replacement as a standard work of reference in English, although not by S. R. Sharda, Sufi Thought: Its Development in Panjab and Its Impact on Panjabi Literature (New Delhi: Munshiram Manoharlal, 1972). The necessary corrective is provided by the best Pakistani critics, but unfortunately it is not available in English; see, e.g.,' All cAbbas Jalalpuri, Vahdat ul Vujad te Panjdbi Shd'iri (Lahore: Pakistan Panjabi Adabi Board, 1977). 12. See Christopher Shackle, "Persian Poetry and Qadiri Sufism in Late Mughal India: Ghanimat Kunjahi and His Mathnawi, Nayrang-i Ishq," in L. Lewisohn and D. Morgan, eds., Late Classical Persianate Sufism (Oxford: One- world, 1999). 13. Nairang-e 7Ishq (Lahore: Panjabi Adabi Academy, 1962), 8-9. 14. Yet it does not convey the nationalist sentiments doubtless intended to be evoked by the later inclusion of the passage in a Pakistani anthology, Muham- mad Ikram, ed., Armaghan-e Pak, 2d ed. (Karachi: Idara-e Matba'at-e Pakistan, 1953), 249-50. 15. Nigdristan-e Ulfat (Lucknow: Gulzar-e Avadh Press, 1899), 5-6. 16. The traditional comparison between the two is presented in supercilious summary by Muhammad Sadiq, A History of Urdu Literature (London: Oxford University Press, 1964), 138-42. Besides the Urdu critical studies cited in the preceding note, see D. J. Matthews, C. Shackle, and Shahrukh Husain, Urdu Literature (London: Urdu Markaz, 1985), esp. 28 ff., 66-69, for a brief account of the Urdu masnavi in English. 17. Described with many further references to the secondary bibliography in Christopher Shackle, "Transition and Transformation in Varis Shah's Hir," in C. Shackle and R. Snell, eds., The Indian Narrative: Perspective and Patterns (Wies- baden: Otto Harrassowitz, 1992), 241-63, esp. 243-48, 263. 18. For reasons yet to be investigated in that comparative framework of Indo- Muslim literary studies which has still to be properly established, the romance also appears to be a genre of relatively much greater importance in Panjabi than in such typologically similar Indo-Muslim literatures as those produced in Sindhi or Kashmiri, as described, for example, in Annemarie Schimmel, Sindhi Literature (Wiesbaden: Otto Harrassowitz, 1974), and Braj B. Kachru, Kashmiri Literature (Wiesbaden: Otto Harrassowitz, 1981). 19. "Early Vernacular Poetry," 288. 20. I am engaged in a fuller study of Miyai Muhammad, whose pivotal position for generic studies of masnavi and qissa may be indicated here only in passing by a bare mention of his unique Panjabi version of the Nairang-e Ishq. Lack of space, unfortunately, prevents a fuller discussion of this most sugges- tive poem.  Christopher Shackle I 71 21. Muhammad Bakhsh, Sohni Mahinval (Jhelum: Malik G hulam Nur, 1964), 11-12. 22. The poem is known to have been a particular favorite of Miyan Muhammad's. See the abridged English prose translation by David Pendlebury (London: Octagon Press, 1980). For Nizami's versions of the other two ro- mances, which were probably a good deal less familiar to most qissa poets, it would be sufficient to cite here the English summaries in Peter J. Chelkowski, Mirror of the Invisible World: Tales from the Khamseh of Nizami (New York: Metro- politan Museum of Art, 1975), 21-67. 23. The most famous version is attributed to Pila (c. 1600) and is recorded in R. C. Temple, The Legends of the Panjab, 3 vols. (Bombay: Education Society's Press, 1884-1900), 3:1-23. 24. The classic version of this story, set in the river and deserts of Sind, is the short qissa (c. 1800) by Hashim Shah, translated by Christopher Shackle (Lahore: Vanguard Books, 1985), and given in an incomplete bardic version in Temple, Legends, 3:24-37. The differences in the structuring of Panjabi and Sindhi iden- tity might be partly mapped in terms of the inventory of such narratives and the frames of their transmission. Even a superficial comparison will show quite marked differences in Sindhi, for example, in the stories drawn on for the ency- clopedic lyrical treatment (which is quite without real parallel in Panjabi) of the classic Risalo of Shah Abd ul Latif (d. 1753), most recently treated in English in Durreshahwar Sayed, The Poetry of Shdh Abd al-Latif (Jamshoro/Hyderabad: Sindhi Adabi Board, 1988). 25. Little evidence to the contrary, at least, is forthcoming from R. S. McGregor, Hindi Literature from Its Beginnings to the Nineteenth Century (Wies- baden: Otto Harrassowitz, 1984). The Sufi romance in Avadhi is, of course, an important case apart but one which cannot be properly addressed here. 26. See further Shackle, "Transition and Transformation." 27. This is shown by Simon Digby in a series of chronologically and geo- graphically wide-ranging studies. 28. The qissa of Paran Bhagat is more closely concerned with yogis than the Hir story and perhaps for this very reason is popular across the communal boundary in the Panjab, even though by far the best-known version is by the Muslim Qadir Yar (1802-91). See especially M. Athar Tahir, Qddir Ydr: A Critical Introduction (Lahore: Pakistan Panjabi Adabi Board, 1988), 97-99, plates IX-XIII. 29. Gurdit Singh Premi, ed., Damodar Rachnavali (Patiala: Bhasha Vibhag, 1974), 192-96, whence the following quotation from pauri 673. 30. See Temple's introductory note to the "Legend of Raja Gopal Chand" (in the dramatic format of a Hariyanvi svang) in Legends, 2:1-77. The Gopichand story is an "anti-romance" quite as important for the understanding of premodern Hindu cultural identity of the Panjab as the romances considered in this essay are for its Muslim counterpart. 31. See my "Transition and Transformation," 255-59.  72 | Crossing the Boundaries in Indo-Muslim Romance 32. Muhammad Akbar ud Din Siddiqi, ed., Masnavi Chandarbadan-o Ma'ydr (Hyderabad: Majlis-e Isha'at-e Dakani Makhtutat and Dakhini Sahitya Prakashan Samiti, 1956). 33. Annemarie Schimmel, Classical Urdu Literature from the Beginning to Iqbal (Wiesbaden: Otto Harrassowitz, 1975), 138. 34. Muqimi, 116. 35. Neither the literary histories nor the British Library catalogs appear to provide evidence of North Indian Urdu versions of the Chandarbadan story. Gyan Chand Jain, Urdu Masnavi Shimali Hind men (Aligarh: Anjuman-e Taraqqi- e Urdu [Hind], 1969), 128, is much taken with its Liebestod motif of "union in death," repeated as the conspicuously similar climax of Mir Taqi Mir's well- known Darya-e cIshq, which is summarized in Khurshidul Islam and Ralph Russell, Three Mughal Poets (1969; Delhi: Oxford University Press, 1991), 101-3, but his discussion has nothing to say about the difference of the lovers' religions. 36. The earlier literary histories (see, e.g., Kushta, Tazkira, 144-49) provide intriguing notices of Ahmad Yar, clearly a major literary figure of his time and one who was associated with the court of Ranjit Singh. Earlier accounts have been superseded by the extended doctoral study of the poet published as Shahbaz Malik, Maulavi Ahmad Ydr: fikar te fan (Lahore: Meri La'ibreri, 1984), which suggests (89 ff.) that Al.mad Yar's Chandarbadan may be no longer extant, like his Sohni Mahirval and quite a number of other qisse. 37. See Kushta, Ta2kira, 150-51. His poem is available in a number of good Gurmukhi editions, including Divan Singh and Roshan Lal Xhja, eds., Sohni Mahirval Fatal Shdh, rev. ed. (Jalandhar: New Book Company, 1976). 38. Imam Bakhsh, Chandarbadan ba-zabdn-e Panjdbi (Lahore: Matba'-e Sultani, 1876), 8. 39. Ibid., 16. 40. Especially kafi no. 42, with the refrain Tangh mahi di jaliydu (I burn with love, my dear herdsman). 41. See the detailed account of the many available Panjabi versions in M. S. Amrit, Panjdbi Sohni-kuvi du Alochndtmak Adhiain (Amritsar: Guru Nanak Dev University Press, 1989). Amrit's introduction disposes of the connection sup- posed to exist between this story and the legend of Chandarbadan and Ma'yar, on the confusing basis of the similarity between the latter's name and that of Mahinval's Sindhi counterpart, Mehar, on which see Nabi Bakhsh Baloch, ed., Mashhar Sindhi Qissa: Ishqiya Dustun 2, Suhni-Mehar ain Nari Jdm Tamdchi (Hyderabad: Sindhi Adabi Board, 1972). Neither of these works explains the puzzling discrepancy between an apparent lack of Urdu or Persian treatments of the romance and its great popularity in the Avadh area from the eighteenth cen- tury as a subject for paintings, as described in Stephen Markel, "Drowning in Love's Passion: Illustration of the Romance of Sohn! and Mahinwal," in P. Pal, ed., A Pot-Pourri of Indian Art (Bombay: 1988), 99-114. The earliest Persian masnavi mentioned in Baqir, Panjdbi Qisse, 193 ff., is that by Salih, which dates only from 1841.  Christopher Shackle I 73 42. On whom see Kushta, Tazkira, 191-96. 43. Particularly notable in this regard is Sudhir Kakar and John Munder Ross, Tales of Love, Sex, and Danger (London: Unwin Paperbacks, 1987), whose discus- sion of the HIr and Sohni stories is much colored by its vision of Islamic patriarchalism and its neurotic need to control female sexuality (see esp. 60, 67). More straightforward modern English retellings of the romances show how much is lost by just sticking to the story line, whether this is done by Pakistani authors, as in Zainab Ghulam Abbas, Folk Tales of Pakistan (Karachi: Pakistan Publications, 1957), or Masud-ul-Hasan, Famous Folk Tales of Pakistan (Karachi: Ferozsons, n.d.), or by Indians, as in Laxman Komal, Folk Tales of Pakistan (New Delhi: Sterling, 1976), which is Sindhi-based, given the companion volume by Mulk Raj Anand, Folk Tales of Panjab (New Delhi: Sterling, 1974), although the latter tellingly omits the romances in favor of more socially meaningful animal fables. Similar considerations apply to modern retellings in Panjabi prose, e.g., Haribhajan Singh, ed., Kissa Panjdb (New Delhi: National Book Trust, 1972). 44. Like Tulla's use of the money cIzzat Beg paid for his pots to provide a dowry for Sohni to be married to another (45).  3 Religious Vocabulary and Regional Identity A Study of the Tamil Cirappuranam Vasudha Narayanan India is our motherland. Islam is our way of life. Only Tamil is our language. Song from a Tamil cassette, "Makka nagar Manapi" We emphatically say that we who live in the south are the oldest Muslims in India. We take pride in that. K. P. S. Hamid, 1973 Muslims from South India pride themselves on being descendants of people who converted to Islam while the Prophet was alive and in thus being the oldest among the Muslim communities of India. Of the 2.5 million Muslims in Tamilnadu, about 1.7 million are said to speak Tamil.1 Their spoken and written Tamil contains many Arabic and Per- sian loan words, yet it is closely aligned with Standard Tamil and bor- rows from Sanskrit as well. Most Tamil Muslims see themselves as participating in Tamil literary history. There is a long tradition of Muslim scholarship on both secular and sacred forms of Tamil literature, Islamic and non-Islamic. Muslim men and women have been among the most eminent scholars, for ex- ample, in interpreting the ninth-century Tamil Ramayana composed by Kampan (known as Iramavataram or the Kampa Ramayanam). M. M. Ismail, the former chief justice of the Madras High Court and noted scholar of the Tamil Ramayana, who has written almost forty books on the subject, remarked with justifiable pride that in every generation  Vasudha Narayanan | 75 there is at least one Muslim who is an authority on the Tamil Ramayana.2 Muslims also participate in the Festival of Kampan (Kampan Vila), an annual celebration devoted to the scholarship on this poet.3 Generally speaking, Tamil Hindus have not paid the same scholarly attention to Islamic literature in Tamil. Instead, their encounter with the Islamic tradition has been more on the level of myth and ritual. The Hindus of this region incorporated some Muslim saints and teachers into their pantheon, made pilgrimages to their tombs, and wove stories of Muslim devotees into the legends of Hindu gods. For instance, Lord Ranganatha, the manifestation of Vishnu in Srirangam, has a Muslim consort, and there is a special shrine for her in the temple complex. This pattern is repeated in several other Vaishnava shrines. Performers of classical South Indian "Carnatic" music also incorporated what were perceived to be Muslim melodies into the traditional raga structure of classical South Indian music, which itself shows significant Persian in- fluence.4 But the interactions that have shaped the distinctive character of Tamil Muslim identities have only just begun to be studied.5 Muslim authors have expressed their understanding of Islam through a variety of literary genres that have defined their Islamic identity and their "Tamilness" as well. This essay will begin to address this complex pro- cess of identity construction among Tamil Muslims by examining one important seventeenth-century text, the Cirappuranam (Life of the Prophet). It will highlight the importance of literary vocabulary, literary images, and literary conventions in shaping cultural values and expec- tations shared by Muslims and Hindus, even in a work whose purpose was to underscore the distinctive claims of Muslims to participate in a religion of foreign origins. The Origins of Muslims in Tamilnadu Many Tamil Muslims understand that they are descendants of seafarers who encountered Islam and converted to the new religion.6 Muslims in some areas of Tamilnadu have the last name Marakkayar. This name derives from the Tamil word marakkalam (ship) and means "shipmen." Although the Tamil lexicon states that the origin of the word may possi- bly go back to the Arabic markab, the name Marakkayar (shipmen) at- tests to the Muslims' belief that their ancestors were sailors. There is also a legend of a king in the present state of Kerala who witnessed a miracle and who, after learning that the prophet Muhammad was responsible  76 | A Study of the Tamil Cirappuranam for it, converted to Islam.7 It is important to note that the Marakkayars believe that their ancestors either came directly from Arabia or were Tamil natives who accepted Islam after direct contact with Arab traders within a few years of the Prophet's death (and by some accounts during his lifetime), not after the conquest by Muslims from northern India. In other words, they believe that they are descendants of early Tamil con- verts or of Arab traders who settled down in the Tamil-speaking areas in the seventh and eighth centuries C.E. Command of Tamil literature and language are thus marks of their claims to an "early" origin that brings them close to the time of the Prophet. K. P. S. Hamid argues that the Tamil Muslims are the oldest Muslims in India. He relates an incident at the Second World Tamil Conference on Tamil in 1968. His friend, Dr. K. K. Pillai, professor of history, apparently said in a panel called "Milestones in South India" that while some Mus- lims in North Arcot, Thanjavur, Tirunelveli, and Kanyakumari Districts were of Arab descent, many of them were descendants of those who converted to Islam after the time of Tippu Sultan (c. 1749-99). "We live with such ignorance of our ancient history," Hamid laments, "that we think that Muslims came to South India just a century ago, just after the time of Tippu Sultan. A society which forgets the pride of its ancient history is a society that has forgotten itself."8 Hamid goes on to quote Colonel Wilks's History of Mysore to say that Islam came to the southern tip of India during the reign of one Hajjaj ibn Yusuf of the Hashim clan (kulam). Muslims escaped his persecution and settled in Kanyakumari district in the seventh century C.E. These earliest Muslims called themselves Lappai, Marakkayar, Malumikal, and Nay- inar. For Hamid, the arrival of Islam in India was thus only one part in the larger narrative of the merchants who, "with the companionship of the south-east winds and the north-west winds," took the religion to Lanka, Malaysia, Indonesia, and China. As proof, he refers to a small mosque in Tirucchirapalli. This city (then called Uraiyur) was the capital of the ancient Chola empire. Hamid says that the small mosque is simi- lar to Jain and Buddhist places of worship and has an Arabic inscription dating it to 738 C.E. (Hijri 116). Hamid reminds us of the Muslims' pride in their ancestry and the antiquity of their residence in Tamilnadu.9 Tamil Works on Islam and Tamil Works by Muslims Muslims in the state of Tamilnadu have composed hundreds of works in the last thousand years, and participation in the larger world of Tamil  Vasudha Narayanan | 77 literature has represented an important marker of their identity. Not all these works deal with Islam. The earliest work is a partial preservation of a poem written between the twelfth and fourteenth centuries. Eight verses from the canta Palcantamalai (Garland of many metric verses) are preserved in a longer commentary.10 This work seems to focus on inner love (akam) in the style of the classical Tamil genre of akam poetry. The Palcantamalai is only the first poem in a long line of works on Islam in Tamil. Over the centuries, Muslims in Tamilnadu have studied both secular works and Hindu religious poetry and utilized many of the traditional Tamil literary conventions with great skill in their religious writings. Religious works in Tamil written by Muslims include the fol- lowing: (a) Several kappiyam (Sanskrit: kavya, epic poems) of which the Cirappuranam, a seventeenth-century biography of the prophet Muhammad, is the best known. (b) Hundreds of devotional Tamil poems about many Muslim leaders, including the Prophet, the early caliphs, the prophets, the Prophet's grandsons, and many Muslim walis or saints. Some are addressed to a goddess and written in a mystical Sufi genre. Some of the Tamil literary genres adopted by Muslims in writing about Islam include kirtana and sintu (different kinds of songs in South India), kummi (folksong for a dance by girls, sung to the clapping of hands), ammanai (sung by girls while playing certain games, throwing stones or balls in the air and then catching them), ecal (songs that insult another person), temmanku (rural songs), and ti- ruppukal (sacred praise). (c) Miscellaneous works, including folklore that portrays a shared world of metaphors between Tamil literature and Arab sto- ries.11 (d) Descriptions of holy places like Nagore in South India. (e) Arabic genres adapted to the Tamil language. These include kissa (the most famous works being one on Joseph and one on Ali and Zaytun), pataippor (leading an army into battle), nama, and masala. Umaru Pulavar The most famous work on Islam in Tamil is the Cirappuranam of Umaru Pulavar (Omar the Poet), who lived in Kilkarai, the site of many recent Tamil Islamic conferences. Umaru's date of birth, 1665, as with other  78 | A Study of the Tamil Cirappuranam dates in premodern and early modern India, is disputed, and some scholars say he was born in Hijri 1052, on the ninth day of the waning moon in the month of Shabban, and they calculate this to be November 2, 1642. There is no disagreement about his death, which is said to have occurred on July 28, 1703.13 His composition made its debut around that time but was first printed only in 1842 by Ceykku Aptul Katir Neynar (Sheikh Abdul Khader Nayanar). There is no written evidence of Umaru Pulavar's life, and almost all that we know about him is from oral tradition. His name is not men- tioned in his epic work. He was apparently born in Ettayapuram.14 His father, Ceku (Sheikh) Mutali, was a dealer in spices and perfumes. His ancestors were apparently Arab merchants who settled down in Tamil- nadu.15 Umaru is thus said to be of the Conakar (foreign, especially Greek or Arab) community. He married and lived in Kilkarai; in honor of his residence there, whenever a Muslim marriage takes place there, the families donate money to charity known as "the poet's share." Umaru's brilliance impressed Citakkati (1650-1715), a Muslim phi- lanthropist who was a patron of Hindu and Muslim scholars. Citakkati's real name was Sheikh Abdul Qadir, and he was the financial adviser of Vijaya Raghunatha Cetupati, the ruler of Ramnad.16 Citakkati appar- ently asked Umaru to compose a work on the life of the Prophet. Umaru then went to Lappai Ali Hajjiyar to learn about the Prophet's life from Arabic and Persian sources. The teacher did not accept Umaru initially because Umaru was dressed like a Hindu.17 The Prophet then appeared in the dreams of both Umaru and Lappai, and Umaru was directed to Lappai's brother in Parankipettai (meaning "town of foreigners").18 The first public reading of the Cirappuranam, according to some ver- sions, took place under the patronage of Apul Kacim Marakkayar (Abdul Kasim Marakkayar), after the death of Citakkati (although the traditional dates ascribed to Umaru and Citakkati cast doubt on this story). Umaru mentions Apul Kacim Marakkayar as his patron in the Cirappuranam and praises him twenty-two times. It is said that Umaru lived in Apul Kacim's house while he composed the Cirappuranam. Since the patron's name is not mentioned in the last third of the poem, some scholars speculate that either the patron withdrew his support (rather unlikely) or Umaru did not believe in exaggerated praise of the patron. It is also possible that his patron died.19 There are other stories that are part of oral tradition, which scholars agree need extensive research if their accuracy is to be validated. For example, one story is that the patron's wife was so entranced by the  Vasudha Narayanan | 79 poem when she heard the section on the birth (avatara) of the Prophet that she did not direct the milk properly to her nursing child, and the child died (possibly of choking). Not wanting to interrupt the flow of the recitation, the mother kept quiet, holding a dead child in her arms until the poem was recited in full.20 Umaru also composed two small works. One was a kovai,$ a poem dealing with various aspects of love in honor of his earlier patron, Citakkati, and the other was Mutumolimalai (Gar- land in the old/mature language), eighty-eight stanzas modeled on ear- lier bhakti poems on the prophet Muhammad. The Cirappuranam The Tamil title of Umaru Pulavar's Cirappuranam is indicative of the blending of genres in its text. Cira is the Tamil form of the Arabic sirah, a word used for hagiography, specifically the biography of the Prophet. Purana (Tamil Puranam) is a genre in Hindu literature, a Sanskrit term that occurs in Tamil and other Indian vernaculars. Puranas include pious accounts of the salvific deeds of a divine being, sometimes seen as an incarnation of the supreme deity, and they contain long poetic accounts of this person's wondrous qualities. Thus, there are puranas addressed to the Hindu gods Vishnu and Shiva, to the goddess Durga, and so on; the purana addressed to Vishnu speaks of his various incarnations to save human beings. Tamil puranams generally deal with deities, saints, or the sanctity of a sacred place. Unlike the Tamil epics like Civika Cin- tamani, the Ramayanam, or the Cilappatikkaram, Tamil puranams usually do not focus on one character but are narrative accounts of gods or saints. Sanskrit puranas deal with the creation of the world, evolution of this creation, genealogy of the gods or divine beings, world "history," and history of royal families. Famous biographical puranas in Tamil fo- cus on Skanda (Murukan) as well as on Saiva saints. Calling the life of the prophet Muhammad a purana, therefore, predisposes one to have certain expectations of the central figure in the text. The combination of a foreign (here Arabic) word with a Sanskrit one in the title gives us a hint of what is to follow: the presentation of a "foreign" religion in a genre predominantly used by Hindus-a genre shaping a vocabulary of praise and devotion shared with Muslims. The Cirappuranam thus incor- porates Tamil literary conventions and customs and the Tamil landscape into the description of the lives of the Prophet and members of his fam- ily. The author shows exquisite knowledge of earlier Hindu devotional literature in Tamil and seems to be acquainted with the ninth-century  80 | A Study of the Tamil Cirappuranam Tamil version of the Hindu epic Ramayana (Story of Rama) composed by Kampan as well as the tenth-century Civika Cintamani. The Civika Cin- tamani (Jivaka, the wish-fulfilling gem) was composed by Tirutakka- tevar (Sri Daksa Deva), a Jain monk, and deals with the love life and conquests of Jivakan. The 5,028 verses of the Cirappuranam are presented as three cantos: "Vilatattuk Kantam" (24 chapters; 1,240 verses) "Nupuvat Kantam" (21 chapters; 1,105 verses) "Kicurattu Kantam" (92 chapters, 2,683 verses) The names of the cantos derive from Arabic words. Viladattu means birth, nupuvat is from nubuvat (prophethood), and kicurattu is the Tamil form of hijrat. Cirappuranam's Opening Chapter: "Praising the Lord" The Cirappuranam begins with salutations to God and the prophet Mu- hammad. The first verse describes God as tiruvinnun tiruvai (being the tiru of Tiru). The Tamil word Tiru is the equivalent of Sri. One way in which the word Sri is used is as a name for the goddess Lakshmi. In more general terms it means "auspicious," "fortune," "wealth," or "sacred." Thus, the names of many of the Tamil sacred compositions begin with the word Sri or tiru (wealth). In Vaishnava literature, both in Sanskrit and Tamil, one finds words strikingly similar to Umaru Pulavar's begin- ning words, tiruvinin tiruvai. In the ninth century, Tirumankai Alvar re- ferred to Vishnu in the town of Terazhundur22 as "becoming the tiru of even Tiru" (tiruvukkum tiruvakiya). Vishnu was called the wealth of the goddess who personifies all fortune. The Cirappuranam thus begins with auspicious words in the Tamil language. Umaru then venerates Muhammad: He, the Handsome One, appeared as the light of the four Vedas which showed the path in the world. Those who keep the words of this leader ever in the center of their mouths will be celebrated by poets and praised by all. They will know the Truth so doubts are slashed and their ears are appeased.  Vasudha Narayanan | 81 Thoughts that give rise to evil deeds will go away.2 The verse itself is important in many ways. It speaks of the four Vedas and the words of the Prophet. The commentator expounds on the four Vedas: the Taurat (Torah) given to Musa (Moses), the Capur (Zabur) given to Tavoot (David), the Injil (Gospel) given to Isa (Jesus), and the Purukan (Furqan or the Qur'an) given to Muhammad.24 The words that one is supposed to keep constantly on one's tongue are called the mula mantra (primary mantra). This mula mantra, he says, is the shahada, or "La illah illallah.. ." In these and other verses, the framing vocabulary is shaped by Hindu tradition-witness the commentator's explanations of the four Vedas and the mula mantra-but the exegesis is clearly Islamic in character. In the opening chapter of the Cirappuranam, after eight verses prais- ing the Lord and Muhammad, the poet pays his respects to the first four Caliphs, the wali Mukiteen (Muhiyudin Abd al- Qadir al-Jilani, 1078- 1166) of Baghdad, and then his teacher Capatullah Appa. He reveren- tially places their feet on himself. The verse addressed to Uthman is a typical example: Uthman decreed that one form of the Sacred (tiru) Veda which came from the tongue of the Prophet whose effulgent body makes the moon cringe sweep through this world. Uthman holds as his life those who know the four great Vedas the elders, and the young ones. Not ever forgetting him let us place his twin feet firmly within us.25 The chapter concludes with a sense of humility, and the poet ex- presses his unworthiness to compose this work. Beginning a work with praise of gurus and a confession of unworthiness is typical of the stotra (panegyric) genre in South Indian Vaishnava works in Sanskrit; the works of Yamuna and Kurattalvan in the eleventh century and of Vedanta Desika in the thirteenth century bear testimony to this style. Umaru writes:  82 | A Study of the Tamil Cirappuranam Like a little ant, grown weak from hunger exhaling its breath In front of the squalls and gales that churn the seven seas and storm the mountains as though their very nature were to change I compose my poem in front of the exalted Tamil poets. Line by line, I see nothing but fault in all that I compose. Step by step, the exalted poets of yore have obtained knowledge. To compose in front of them is to measure the noise that comes when I snap my fingers with the sound of rolling thunder.26 In earlier stotra literature in Sanskrit, we see verses like the following: Though I know of my ignorance, I am shameless enough to wish to string together these words of love for the feet of our Lord; for even when the river Ganga which is naturally pure is licked by a dog, it is still known as "holy water."27 Although the sentiments are similar, we see how Umaru gets our atten- tion with simple and unpretentious similes. In other verses of the first chapter, Umaru Pulavar pays reverence to or refers to the twin feet of the exalted teachers, and one can easily recognize that these references are typical of Hindu devotional literature where a poet reveres the sacred feet of the teacher or the deity. The Tamil Landscape in Arabia: The Use of Literary Conventions from Cankam Poetry As did the Tamil epic poets Kampan and Tirutakkatevar, Umaru Pula- var gives extensive descriptions of the country and the city where the Prophet is to "descend" (avatara). This is followed by a list of the ances- tors of Muhammad. Umaru apparently never traveled to Arabia, and his description of the country is a description of Tamilnadu. In the utiliza- tion of this method, too, he has a predecessor. Kampan, the author of the  Vasudha Narayanan | 83 Tamil Ramayana, transposes the Tamil landscape to Ayodhya in northern India. Descriptions of the river Kaveri are transferred to the river Sarayu. Umaru Pulavar also transfers the Tamil landscape to Arabia. A typical feature of Tamil classical poetry, especially the puram verses (typically dealing with chivalry, kings, and war), was description of the wonders of a king's land. In the Tamil verses of classical (Cankam) po- etry, we find roaring cascades (the presence of water indicated prosper- ity in South India where drought was all too common), fertile fields that are well irrigated, lush fields of paddy and cane sugar, blossoming lo- tuses, and bees sucking nectar from flowers redolent with honey. The waterfalls and rivers carry precious gems fallen from the jewelry worn by people who bathe in them-obviously indicating that the king's land is filled with rich people. This is also seen in a description of Lord Mu- rukan's domain in the fifth-century (?) poem Tirumurukarruppatai: The cataracts of the mountains look like varied, waving flags of glory ... they spill ... sweet-smelling, huge honeycombs built upon lofty hills that kiss the sky.... The cascades gush along ... the falling waters bear in their bosom the pearl-bearing white tusks of huge elephants; the torrents leap along with fine gold and gems shining on their surface, washing aside glittering dusts of gold ... the hills abound in groves with ripening fruits. God Muruga is lord of such hills.28 The cities are also described in considerable detail in Cankam literature: prosperous seaports, terraces looming like mountains, palaces stretch- ing to the sky. They are centers of culture where bards and courtesans flourish. The wealth of a nation rests on its ability to produce food, and this depends on rainfall. Poets describe the prosperity of a land by the abun- dant rain it receives. Umaru Pulavar talks of the white clouds drinking up the seawater, becoming dark and heading for land (Arabia). The clouds cover the mountains, and storms rage. The storms abate, but the heavy rains continue, flooding the place, and it becomes chilly. El- ephants, lions, and other animals feel the cold and, forgetting the enmity between them, go to one place. Elephants, deer, squirrels, tigers, bisons, giant lizards, monkeys, lions, spotted deer, anteaters, lemurs, bears, wild dogs, buffalos, porcupines, humped bulls, and other animals huddle in the cold, shivering. Because of the high winds, trees on the mountains fall. Flocks of birds fly, frightened, and floods of water come down the emerald slopes of the mountains onto the plains. The flooding  84 | A Study of the Tamil Cirappuranam streams approach the houses of the gypsy women (with wide eyes and red lips) who live on the foothills. The streams drop as water falls over the emerald mountains knocking down the banana trees and wood- apple trees. The floods sweep away gems from the mountains just as a courtesan embraces a king and sweeps away his gold, gems, and price- less pearls.29 Kampan, the author of the Ramayana, also uses this analogy: Like a courtesan embracing her lover his head, his body, his feet, as if in desire all for a minute [and fleeing with his ornaments] the floods embrace the peak, the slopes, the foothills and sweep away everything.30 Compare this with the following verses from the Cirappuranam: Like a courtesan embracing the mountain-like king giving him pleasure, and sweeping away gold which gives us prosperity, precious gems, pearls, and all splendid things, and flees the frontiers, the floods flow carrying with them all riches.31 A waterfall carrying gems is also a traditional image in Tamil litera- ture. Nammalvar, a ninth-century poet, speaks of the waterfalls of Tir- upati hills: Lord of Venkata hill where clear waterfalls crash spilling gems, gold, and pearls.32 This is, of course, part of the wealthy, fortune-filled land that is being described. The waters rush like an elephant. The streams, when they slow down, look like lovely girls. Their white froth looks like white gar- ments worn by maidens; the dark silt resembles dark hair; the fish look like the eyes of a girl; the water bubbles seem to be like breasts, and the whirlpools circle like the navel. These descriptions and analogies are generally necessary in Tamil poetry to prove one's mettle as a poet. Like a Vaisya merchant, the streams carry sandalwood, ivory, pearls, and gems.33 Here, too, Umaru Pulavar follows Kampan: Carrying the pearls, gold, peacock feathers, beautiful white ivory from an elephant, aromatic akil wood,  Vasudha Narayanan | 85 sandalwood, matchless in fragrance, the floods looked like the vaniya merchants.34 In the Cirappuranam we find the following: Carrying the fallen sandalwood, branches from the dark akil tree, pearls from the broken elephant's horn, white ivory, more precious than these, red rubies, radiant in three ways, carrying these all towards the sea, the stream laden rich bamboo, looked like a vaniya merchant.35 The river flows through the Kurinci (mountainous) land, presumably of Arabia, through the desert (a recognized category in the landscapes of Tamil poems), and into the forests. Reaching marutam (cultivated land), it fills the lakes, ponds, and tanks. The streams break through the lakes and approach farmlands. They sweep through the sugarcane planta- tions and slush up the ponds where the fragrant lotus flowers bloom. The water is then contained and used for irrigation. The single body of water held in many tanks, ponds, lakes, and areas where lotus bloom is compared by Umaru to life (Tamil: uyir), which appears in hundreds of millions (Tamil: cata koti from Sanskrit Satha koti) of beings. This idea reminds one of the Advaitin notion that a single soul (atman) appears in many forms and bodies and seems to be many. While Umaru does not elaborate on his analogy, it is striking that he seems at home with these Vedantic ideas.36 Where did all these descriptions come from? The earliest Tamil litera- ture composed in the earliest centuries of the Common Era recognizes five landscapes. The Cankam poems (also known as the poems of the classical age or the "bardic corpus"), dealing with romantic or heroic themes, refer to five basic situations. These situations correspond in poetry to five landscape settings (tinai), birds, flowers, times, gods, etc. The five basic psychological situations for akam poems are lovemaking, waiting anxiously for a beloved, separation, patient waiting of a wife, and anger at a lover's real or imagined infidelity. These correspond to the mountainous (kurinci), seaside (neytal), arid (palai), pastoral (mullai), and agricultural (marutam) landscapes.37 More specifically, Umaru's de- scriptions closely resemble the descriptions of Kampan in the first two chapters of his version of the Ramayanam. However, even though the details are exquisitely similar in spirit and in concept, each poet has his own inimitable style. Reading both descriptions is similar to listening to the same raga played by two maestros.  86 | A Study of the Tamil Cirappuranam After describing the mountainous regions, Umaru Pulavar speaks of the wealth of the cultivated land and the beauty of its women. He projects the practices of the (presumably Hindu) farmers from Tamil- nadu onto the fertile rice fields of Arabia. In a striking verse, he talks about the farmers' worship of the sun and the earth before they sow the seeds in the field, and then he waxes eloquent about the beauty of the women: Wearing jewels, quaffing toddy, Worshiping the Sun with their hands, and then Worshiping the god of their clan, Milling in crowds, those who labor in the fields gather and praise Earth. Their right hands shake the sprouting seeds and scatter them thick on the ground. They fall like golden rain on earth.38 Modern Muslim commentators interpret this verse in two ways. Jus- tice M. M. Ismail merely says that the poet is conversant with the farm- ing practices of South India and then projects them on to Arabia. How- ever, in his detailed exegesis of the Cirappuranam, Kavi (poet) Sherip (Sherif) says that these practices may have existed in Arabia before the time of the Prophet.39 Sherip also takes a more literal view of Umaru Pulavar's description of the mountains and streams of Arabia and says that this kind of landscape can be found in Yemen. Descriptions of women are found soon after these verses; in Tamil poetry, a flourishing land is filled with voluptuous women wearing daz- zling jewelry. These women have participated in the sowing of the seeds and are walking through the well-tilled, fertile fields, which are prosper- ously slushy with life-giving waters: With twin eyes made red by drinking palm-toddy, with slender bamboo-like shoulders heaving, the women walk with drunken steps, swaying softly like a swan. Their feet tread the well-tilled land, Mud splashes on their breasts that soar like the tusks of a lusty elephant. Their breasts speckled with slush  Vasudha Narayanan | 87 look like the tender buds of lotus flowers swarming with tiny bees.40 There are half a dozen verses like this describing their teeth, coral lips, victorious demeanor, etc. While these descriptions may seem somewhat startling in a work that purports to be the life of the Prophet, the phrases and general tone of the verses are almost standard fare in any self- respecting Tamil poem and not at all unusual in Tamil literature. Umaru Pulavar also follows the conventions of Tamil poetry in de- scribing the city of Mecca (Tamil: Makka). Takkatevar, the author of the Jain epic poem Civika Cintamani, and Kampan described the towns and urban culture in considerable detail. The city of Mithila, hometown of Sita, is described with great poetic skill by Kampan. Following the de- scription of the countryside, Umaru Pulavar describes Mecca, begin- ning with a number of Tamil/Hindu cosmological details. The verses are strikingly "Hindu" in origin but again a part of the common lore to which Umaru Pulavar had access: I shall now expound in brief on the great city of Mecca. The expansive lakes filled with radiant conches brimming with pearls seem like a moat with waves. Many kinds of lotus blossoms filled with lustrous gems ring the town. The prosperous fort appears like a lotus flower with golden petals. Wealth and luxurious fortune ever increase in this prosperous town. The seven isles invite the northern mountains to surround them on one side and install them there. A tall mountain they establish as the crest of the crown and surround it on all four sides with fortresses.  88 | A Study of the Tamil Cirappuranam The sea surrounds the land on three other sides like a moat. The city of Mecca appears like the gem on the sacred head of the King of Serpents [Adi Sesa].41 This city resounds with the noises of busy life, horses running swiftly, chariots, elephants trumpeting like thunder, such that even the sea is afraid of making noise. This is the standard city of Tamil poetry: for- tresses, lakes, lotuses, horses, chariots, elephants, and so on. Almost ev- ery literary convention in the description of the prosperous city is in- cluded in Umaru Pulavar's portrait of Mecca. Umaru also refers to Hindu deities occasionally. Abu Talib, he says, is brave, knowledgeable about the arts, and ever triumphant. Lakshmi, the goddess of good for- tune, reigns victorious at the portals of his house.42 But apart from a few allusions like this (such as references to the path of Manu, the generosity of Surabhi, the wish-fulfilling cow, and so on), there is no attempt made to incorporate Hindu deities within the worldview of Islam in either a positive or a negative fashion. The appropriation of the Prophet into a Tamil world shared by Muslims and Hindus alike is accomplished through the generic use of convention and language.43 The Chapter on Fatima's Wedding Other Tamil literary conventions are used with equal skill by Umaru Pular. The long chapter on Fatima's wedding (Pattima Tirumanappat- alam) contains a beautiful description of Ali's procession through the city of Medina that parallels Rama's procession through Mithila in Kampan's Ramayanam. Let us consider the swarming of women in Mi- thila rushing to get a glimpse of Rama: Like a herd of deer closing in Like a flock of peacocks wandering Like a shower of brilliant meteors [or "a galaxy of splendrous stars"] Like flashes of lightning coming close With bees that flutter around their garlands humming in tune With bands of anklets and rings tinkling, Women, their hair adorned with flowers soaked [with honey] swiftly thronged around.  Vasudha Narayanan | 89 Not attending to their hair that loosened and cascaded down Not heeding their waist belts that broke loose Not pulling up the flower-soft clothes that slip away Not pausing to rest their tired waists they closed in [on Rama]. Coming close they cried, "Make way, make way!" Women who lend splendor to the city swarmed around him like bees tasting honey.44 Women in Medina overflow from the balconies trying to catch a glimpse of this handsome bridegroom, and when they see him, they are filled with longing and wonder: Is all this charm and beauty to be mo- nopolized by just one woman? Describing the women of Medina, the poet adopts a strategy common to Sanskrit and Tamil literature known either as padadi kesa varna (description from the feet to the hair) or kesadi pada varna (description from the hair to the feet). Umaru thus describes the women of Medina: The anklets swirled on the feet, the golden belts around their hips tinkled. The ornaments on their radiant breasts- breasts as sharp as the tusks of a lusty elephant- flashed. Their hair, adorned by flowers fragrant and fresh, dripping honey spilled out from their constraints. Like many moons flowering on the ocean, young maidens, thronged around.45 The women in Medina, watching Ali's procession, are seen to be wear- ing jewels (anklets, mekalai, or waist belt) like Tamil women, and tradi- tional descriptions of their breasts and hair follow. Their breasts are pointed and sharp, their hair is long and collected together in swirls of fragrant flowers filled with honey. Their hair, which is to be demurely gathered together, loosens in their mood of abandon. The words suggest an erotic mood, appropriate for the wedding. They are filled with long- ing at the sight of Ali. According to Tamil literary conventions, a woman's body becomes pale when she is parted from her lover or hus- band. This special lovesickness signified by a pallor is seen in the young women of Medina. They swarm around like an ocean, says Umaru, and then he uses metaphors connected with the sea to describe them:  90 | A Study of the Tamil Cirappuranam Eyes like darting fish, necks, exulting like conches; teeth like white pearls, smile through parted lips, which flash like corals. With golden skins growing pale with longing flowers loosened like shining foam, A sea of women, swarm thick without any gap between them.46 The poet uses the sea as the primary metaphor and says the women splash forth like the ocean. The poet comes from and lives on the sea- shore and is lavish in his use of metaphors from the sea. Fish, conches, corals, pearls, and sea foam are all used as elements of comparison in describing these women. A poet's skill was frequently seen in the use of metaphor and similes, and descriptions of the human body-for both men and women-were particularly relished by the audience. Umaru certainly excels according to all Tamil standards in these areas. Apart from literary convention, the poem utilizes words extensively found in other forms of religious literature in Tamil. Many are Sanskrit loan words. We have already noted the use of the word Veda. The Qur'an is called Veda and frequently spoken of as the marai ("mystery," Tamil synonym for Veda).47 The word marai means "hidden, that which is a secret, a mystery." The Vedas and the Upanishads were called marai in the Tamil texts written in the first five centuries of the Common Era. We find this word in the Tolkappiyam, the Paripatal, and other Cankam works. The word marai functioned as a literal translation of Vedas; thus the village of Vedaranyam (Forest of Vedas) in Tanjore District was called Maraikkatu in the Tevaram. The Upanishads were called marai cirai (the head of the Vedas). The Lord who reveals the Qur'an to Muhammad is called "Srutiyon." Sruti is Sanskrit for "that which is heard" and another synonym for the Veda. We can certainly see how apt the use of the word Sruti is to the Qur'an, which was heard by Mu- hammad. The deity who reveals this is called srutiyon (he of the Sruti). What did these words mean for the Hindus writing in Tamil? Several philosophers, especially those from the Nyaya (Logic) and Mimamsaka schools, had already discussed the term Veda in some detail, and the Vedanta teachers including Ramanuja, the most important Srivaisnava teacher, had clearly discussed the trans-human nature of the Sanskrit Vedas in commentaries on the Vedanta Sutras. All schools of thought agreed on the transcendental aspect of the Vedas and their authoritative  Vasudha Narayanan | 91 nature, but differed on what was meant by the trans-human (apauruseya) nature of their composition. The followers of the Nyaya school believed that God was their author, and since God was perfect, the Vedas were infallible. The Mimamsakas, on the other hand, starting from at least the second century B.C.E., said that the Vedas were eternal and authorless.48 The Vedic seers (rsi) saw the mantras and transmitted them; they did not compose or author them. Calling the Qur'an a Veda, therefore, includes at least some of these meanings as understood in Hindu writings. Other theologically loaded words are used. For instance, the birth of the Prophet is referred to as avatara ("Vilattattu Kantam, Napi Avatara Patalam"). Avatara (descent) is used in Hindu theology to refer to the incarnations of Vishnu, who descended into this world "to save the good and destroy evil" (Bhagavad Gita 4:9). The use of this term for Muhammad, therefore, exalts him as more than human. This word is generally not used for the birth of human beings in Tamil literature. He is also called the tiru tutar (sacred messenger). Tutar is the Tamil form of the Sanskrit dhuta, which means "messenger." While the word is used in secular and sacred language, in Hindu religious literature it is Krishna who is associated with the word dhuta and is frequently called "Pandava dhuta" or "the messenger of the Pandavas" during the Mahabharata war. Muhammad's mother is called the abode of dharma, and this word is used frequently in the biography. Abucakal is said to follow "the path of Manu" (manu neri) ("Nupuvattuk Kantam, Matiyai Alaitta Patalam," v. 152). The use of this religious vocabulary thus helped to shape a dis- tinctive Tamil appreciation of Islamic theology. While not implying any sort of self-conscious accommodation to Hindu ideas, this vocabulary provided the conceptual framework in which much of the Prophet's life was explained and understood. Modern Implications: The Testimony of Tamil Songs Written by Muslims None of this implied, of course, that such writings provided a charter for Muslim and Hindu participation in common ritual practices. To the con- trary, at least on the level of popular practice, it was more common for Hindus, in spite of their general lack of interest in Islamic literature, to incorporate Muslim holy men into their devotional exercises than the reverse. As Susan Bayly has demonstrated, Muslim saints and their shrines came to be important sites in the Tamil sacred landscape, fre- quented by Hindus and Muslims alike.49  92 | A Study of the Tamil Cirappuranam The significance of a work such as the Cirappuranam, however, lies in a different direction; it illustrates how the generic conventions of Tamil literary production have defined a framework for Muslim participation in the Tamil religious world. This was the case even though the focus of devotion was a figure who lived in a foreign land, the prophet Muham- mad. On the one hand, as a sirah, or life of the Prophet, the Cirappur- anam linked Tamil devotees of the Prophet generically to a wider Is- lamic world; the text defined clearly the connections of Tamil Muslims to a world of devotion to the Prophet whose boundaries were far wider than either Tamil vocabulary or Tamilnadu. On the other hand, the poet Umaru's claims to recognition depended on his skill in manipulating a Tamil devotional idiom defined by the text's generic claim to be a puranam. The conventions and vocabulary of the text thus rooted devo- tion to the Prophet in a Tamil conceptual world-a world shared by both Hindus and Muslims. It was generic conventions that helped to con- struct a framework for identity that was simultaneously Muslim and Tamil. The tensions in the process by which such identities are formed also have their modern sequel in Tamilnadu. Analysis of the Cirappuranam provides historical perspective on the more recent claims by many Tamil Muslims to superior status as Muslims because of the antiquity of their connections to Tamilnadu. Familiarity with Tamil literary conventions, like early links to Arabia, does provide evidence of their comparative antiquity as an Indian Muslim community. But how did Tamil Muslims relate to other South Asian Muslims who were not as ancient as them- selves? Tamil Muslims had, of course, long adapted Arabic and Persian forms of literature to Tamil genres, and they showed reverence to Mus- lim saints from other parts of India and the Middle East. At the same time, they did not identify strongly with Muslims in the rest of India, whether in Kerala, Hyderabad, or the north, nor did the trauma of par- tition seem to affect the deep south. Yet the logic of remaining connected with, but aloof from, other Mus- lims of the subcontinent has recently been challenged with the rise of Hindu nationalism. The belligerent stance of some Hindu nationalists in the 1980s and 1990s has prompted a growing sense of insecurity among Tamil Muslims. Their anger, sorrow, and bewilderment have now sur- faced in public through new Tamil Muslim songs sold in prerecorded audiocassettes. Packaged along with standard songs on the glory of Muhammad and Mecca are songs that are very patriotic, some filled with distress, others filled with rage. These cassettes go under titles such  Vasudha Narayanan | 93 as "Makkanakar Manapi" (Great Prophet of Mecca), "Makkavai Nokki" (Looking toward Mecca), or "Pallivacalil kutuvom" (Let us gather at the gates of the mosque). The lyrics of two songs from "Makkanakar Manapi" suggest new frameworks for identity: Song 1 India is our motherland Islam is our way of life Tamil is our language ... Who is it who said we are enemies? Our forefathers worked and fought for freedom Muslims fought to get rid of the nation's sorrow. They have forgotten gratitude. Is this betrayal? Are these the sins we have done? The blood spilled by the Mappillas in Kerala is not yet dry The wounds we got in Bengal are not yet healed Mysore's lion Tipu's bravery will not change Your heart cannot bear the grief of Bahadur Shah Do you know we have ruled India for eight hundred years? The Taj Mahal and the Kutb Minar are witnesses, have you for- gotten? You cannot deny it, You cannot conceal truth. Song 2 I swear on the earth I swear on the heavens I swear on the mother who bore me I swear on God who created [all]. We will not lose our faith as long as we are alive We will not weaken in resolve as long as we are alive Leaving the Land of India, forgetting its glory We will not flee in fear We will not flee in fear of anyone. This is the land where Muslim Kings ruled for eight hundred years. Say, does anyone have this pride [of rulership] other than us? We have never betrayed this country. Muslims have served this land without end.  94 | A Study of the Tamil Cirappuranam Our crowd rose first to seek freedom. That is why blood began to flow in Kerala. Will my heart forget the sacrifice of the Mappilas? We will do countless sacrifice again, for the country. Think of the sacrifice of Bahadur Shah who ruled Delhi when enemies gave him his son's head on a platter.... There is no one equal to us in devotion (bhakti) to the country. He was born as the brave son of Haidar, Mysore's king.... Tipu gave his life in war. He tried hard to free his mother-country. He bore endless grief in the British prisons. Maulana Muhammad Ali Saukat died pining for freedom. Two strategies are immediately visible from these verses. First, the composers are now aligning themselves with Muslims from all over In- dia and not just the south (although a southern emphasis is evident in the references to the Mappilas and Tipu). Second, the verses remind the listeners of the sacrifices made by Muslims all over India during the independence movement. Since the simplistic war cry of the aggressive Hindu is to tell the Muslim to go to Pakistan, the Tamil Muslim songs emphasize that India is their home. This patriotism is woven with songs on the Islamic faith. The songs suggest a new aggressiveness and hardening of religious boundaries, as Hindus and Muslims compete for the right to speak for the nation. Yet they also suggest the ways in which popular literary or musical genres continue to define commonalities and a common identi- fication with Tamil language, even as they shape new conceptions of difference. Although far removed from the shared generic literary con- ventions that defined the Cirappuranam, the language of these songs projects another type of shared generic convention-a devotionalism, now linked to country, that is rooted in a long generic heritage of devo- tional poetry shared by both Hindus and Muslims. Indeed, many of the audiocassettes on which these songs appear continue to be simply cat- egorized by the stores that sell them, along with songs in praise of the Prophet, as "Muslim Devotional (bhakti) songs."  Vasudha Narayanan | 95 Notes 1. Statistical Abstract: India 1990 (Central Statistical Organisation, Ministry of Planning, Government of India), 33; see also Muslim India (January 1984), 18, quoted by Syed Shahabuddin and Theodore P. Wright Jr., "India: Muslim Minor- ity Politics and Society," in John L. Esposito, ed., Islam in Asia (New York: Oxford University Press, 1987), 167. According to Shahabuddin and Wright, the total number of Muslims in Tamilnadu in 1981 was 2.5 million. 2. Interview with Justice M. M. Ismail, July 1993. 3. Looking through the programs on the debates and discussions in the Fes- tival of Kampan held between 1991 and 1993, one finds names like Parveen Sultana, Abdul Khader, and Abdul Karim. 4. In music, the categories of Muslim and Hindu ragas are misnomers; we may more accurately speak of Indian and Persian forms of music. Because geo- graphic origin was often associated with religious affiliation (the word for Mus- lim in Tamil is tulukka from turka or Turkish), music from Persia was character- ized as Islamic. It has also been claimed that Sufi writings in Tamil affected Tamil Siddha poetry. This is probably true, but it has not yet been demonstrated. 5. Paula Richman discusses possible reasons for the neglect of Islamic Tamil literature in the appendix of "Veneration of the Prophet Muhammad in an Is- lamic Pillaitamil," Journal of the American Oriental Society 113.1 (1993): 57-74. The encounter between Hindus, Christians, and Muslims in Tamilnadu and Kerala has been discussed in Susan Bayly, Saints, Goddesses, and Kings (Cambridge: Cambridge University Press, 1989). See also David Shulman, "Muslim Popular Literature in Tamil: The Tamimancari Malai," in Yohanan Friedmann, ed., Islam in Asia, vol. 1, South Asia (Boulder, Colo.: Westview Press, 1980). 6. Interview with Justice M. M. Ismail, July 1993. See also Mattison Mines, The Warrior Merchants: Textiles, Trade, and Territory in South India (Cambridge: Cam- bridge University Press, 1984), and Susan Bayly, Saints, Goddesses, and Kings. 7. Annemarie Schimmel, And Muhammad Is His Prophet (Chapel Hill: Univer- sity of North Carolina Press, 1985), 70. 8. K. P. S. Hamitu (Hamid), "Tennakattil Islattin tonmai," Islamiyat Tamil Araycci Manatu-1 (Research Conference on Islamic Tamil) (Tiruccirapalli: Islami- yat tamil ilakkiyat kalakam, 1973), 51. 9. Ibid., 52-53. On the existence of this small mosque, see Bayly, Saints, God- desses, and Kings, 87. 10. Ma. Mu. Uvais (Uwise), Islam Valartta Tamil, 16-17, quoting Vaiyapuri Pillai, Tamil Ilakkiya Varalaru-14an nurrantu (Cennai [Madras]: Ulakat Tamil- araycci niruvanam, 1984), 335. 11. See Shulman, "Muslim Popular Literature in Tamil." 12. On the maritime importance of Kilkarai and the fact that it is home to many prominent Muslims, see Susan Bayly, Saints, Goddesses, and Kings, 81-85. 13. Uvais, "Cirapuranamum Umaruppulavarum," in Cirappuranam, ed. M. Ceyyitu Muhammatu "Hasan" (Cennai [Madras]: Maraikkayar Patipakkam, 1987).  96 | A Study of the Tamil Cirappuranam 14. K. Zvelebil, Lexicon of Tamil Literature (Leiden: E. J. Brill, 1995), 721. 15. Uvais, "Cirapuranamum Umaruppulavarum," v. His father's name was probably Sheikh Muhammad Ali, which was changed to Ceku (Sheikh) Mutali. Mutali is a respectful suffix added to the names of Hindu devotees, and it is noteworthy that Umaru's father's name was transformed in this way. Others say his name was Mappilai Mukammmatu Nayinar. 16. For a discussion of 'Abd al-Qadir as Citakkati, see V. Narayana Rao, David Shulman, and Sanjay Subrahmanyam, "On the Periphery: State Formation and Deformation," in Symbols of Substance: Court and State in Nayaka Period Tamilnadu (New Delhi: Oxford University Press, 1992), 292-304. 17. K. Zvelebil, Lexicon of Tamil Literature, 721. Uvais, "Cirapuranamum Uma- ruppulavarum," viii, gives the teacher's name as Catakkatulla Appa, as does Umaru himself. 18. Uvais, "Cirapuranamum Umaruppulavarum," viii. Either Lappai Ali Hajjiyar or his brother was the teacher, according to this account. Yet another version, again received by oral tradition, says that Umaru, through the grace of Nakur Cakul Amitu Antavar (Shahul Hamid, a saint buried in Nagore), received the ability to sing. 19. Ibid., xii. 20. K. Zvelebil, Lexicon of Tamil Literature, 722. 21. A kovai poem may be composed on God, a chieftain, or a king. The word literally means "string" or "arrangement." A kovai poem frequently has four hundred verses, each dealing with an aspect of love. The most famous is the Tirukkovaiyar of Manikkavacakar. 22. This town was the birthplace of Kampan, who wrote the Tamil Ramayana. Umaru Pulavar, of course, is very knowledgeable in the Tamil Ramayana, and it is a striking coincidence that he begins his poem with the very words used to describe the deity of the place where Kampan was born. 23. Cirappuranam, "Vilatattuk Kantam, Katavul Valttu Patalam" (The chapter on praising the Lord), 6. 24. Cirappuranam, "Vilatattuk Kantam," pt. 1, ed. with commentary by Kal- aimamani Kavi Ka. Mu. Sherip (Sharif), commentary on v. 6, p. 17. 25. Cirappuranam, "Vilatattuk Kantam, Katavul Valttu Patalam," 11. 26. Ibid., 19. 27. Kurattalvan (twelfth century), Vaikuntha Stava, 8. 28. Translated by R. Balakrishna Mudaliyar, The Golden Anthology of Ancient Tamil Literature, 3 vols. (Tirunelveli and Madras: South Indian Saiva Siddhanta Works Publishing Society, Tinnevelly, 1959-60), 2:20. 29. Cirappuranam, "Vilattatuk Kantam, Nattu Patalam" (The chapter on the countryside), 2-9. 30. Kampa Ramayanam, "Bala Kantam, Arruppatalam," 6. 31. Cirappuranam, "Vilatattuk Kantam, Nattu Patalam," 9. 32. Nammalvar, Tiruvaymoli, 6.10.3. 33. Cirappuranam, "Vilatattuk Kantam, Nattu Patalam," 10-11.  Vasudha Narayanan | 97 34. Kampa Ramayanam, "Bala Kantam, Arruppatalam," 7. 35. Cirappuranam, "Vilatattuk Kantam, Nattu Patalam," 12. 36. Ibid., 12-17. 37. To these five situations of love two more are added: peruntinai and kaikkilai, which have no corresponding landscape. Peruntinai indicates mis- matched love, and kaikkilai unrequited love. For discussions of the landscapes, see A. K. Ramanujan, trans., The Interior Landscape: Love Poems from a Classical Tamil Anthology (Bloomington: Indiana University Press, 1967), 104-12; K. Zvelebil, Tamil Literature, 98-99; K. Zvelebil, Smile of Murugan (Leiden: Brill, 1973), 85-110. 38. Cirappuranam, "Vilatattuk Kantam, Nattu Patalam," 26. 39. Cirappuranam, "Vilatattuk Kantam," pt. 1, p. 80. 40. Literally, "covered with six-legged beetles." Cirappuranam, "Vilatattuk Kantam, Nattu Patalam," 30. 41. Cirappuranam, "Vilatattuk Kantam, Nakara Patalam" (The chapter on the city), 1-3. 42. Cirappuranam, "Vilattattu Kantam, Pukaira Kanta Patalam," 2. 43. The pattern is quite different here from the incorporation of Krishna in Bengali Muslim literature described by Asim Roy, The Islamic Syncretistic Tradi- tion in Bengal (Princeton: Princeton University Press, 1983). 44. Kampan's Ramayanam, "Bala Kantam, Ulaviyar Patalam," 1-2. 45. Cirappuranam, "Kicurattu Kantam, Pattima Tirumanappatalam" (The chapter on Fatima's wedding), 132. 46. Ibid., 134. 47. Marai is frequently used by the alvars and in the classical Cankam poetry to refer to the Vedas. It literally means "mystery." 48. Jaimini composed the sutras in the second century B.C.E. and Sabara com- mented upon them around the second century C.E. On the views of Jaimini, see Francis X. Clooney, Retrieving the Purva Mimamsa of Jaimini. 49. For a discussion of dargas in Tamilnadu, see Bayly, Saints, Goddesses, and Kings, chaps. 3 and 4. Among the most important dargas was that at Nagore, which I have in "The Zam Zam in Nagore, Tamilnadu: Shahul Hamid in the Tamil Landscape," paper presented at the national meeting of the American Academy of Religion, New Orleans, November 1996. All translations in this paper are mine, except where other sources are cited. Tamil words are transliterated according to the style accepted by the Tamil lexi- con. This is not very helpful in pronouncing the words. Tamil has one letter to denote ka, kha, ga and gha, one for ca, cha, ja, and jha, one for ta, tha, da, and dha, one for ta, tha, da, and dha, one for pa, pha, ba, and bha. A native speaker would know how to pronounce the words, which may also vary by community and by region. Centamil or pure Tamil also lacks the letters sa, sha, and ha; these letters are borrowed from Sanskrit. Thus Sirah is written as Cira but pronounced as Sira. Tamil does not have an equivalent for the English f sound and substitutes the letter p. Thus, Fatima is written as Patima in Tamil.  4 Admiring the Works of the Ancients The Ellora Temples as Viewed by Indo-Muslim Authors Carl W. Ernst One of the recurrent problems in the interpretation of Indo-Muslim identity is the attempt to ascribe a consistently Muslim attitude toward Hindu temples. This problem arises initially with the incorporation of building materials from Hindu temples in the construction of mosques or other buildings commissioned by Muslim patrons. Although the evi- dence for the significance of this kind of recycling is sometimes later and retrospective, it is hard to avoid the conclusion that this phenomenon involves the triumphal political use of trophies. Perhaps the most no- table example is the Quwwat al-Islam (or Qubbat al-Islam) mosque near the Qutb Minar in Delhi, which contains numerous columns with par- tially effaced Hindu caryatids and Jain (or Buddhist?) figures, as well as the famous Iron Pillar.1 This kind of triumphal reuse of temple materials and ancient royal monuments has been seen since British times as evi- dence of the insatiable propensity of Muslims to destroy idols at every opportunity. Today it affords ammunition to the Hindu extremists who led the attack on the "Baburi" mosque at Ayodhya; the supposition is that the mosque not only rests on the site of the birthplace of Rama but also took the place of a preexisting temple. There are, of course, competing theories of the exact relationship be- tween the Ayodhya mosque and any preceding temple. Some believe that the mosque was built of the remains of the temple and that the construction of a mosque thus required the demolition of a temple; the reverse of this zero-sum game is that the erection of a temple on that spot would require the destruction of the mosque, as indeed took place in  Carl W. Ernst | 99 December 1992. Others like P. N. Oak assume that Muslim buildings are only partially defaced Hindu structures, so that in theory only a slight amount of restoration would presumably be required to return them to their original functions, rather than full-scale destruction and recon- struction; this has the appearance at least of a less costly program. The problem arises, however, when these modern interpretations of Muslim iconoclasm deduce Muslim attitudes from an essential definition of Is- lam rather than from historical documentation of the significance that particular Muslims attached to Hindu temples. Attempts to describe Muslims as essentially prone to idol-smashing are confounded by the historical record, which indicates that Muslims who wrote about "idol temples" had complex reactions based as much on aesthetic and politi- cal considerations as on religion. The concept of unchanging and mono- lithic Muslim identity accordingly needs to undergo serious revision. This article is an attempt to fill out the historical dossier, by present- ing a translation and analysis of a brief text in which a Muslim author, Rafi' al-Din Shirazi, has set forth a striking interpretation of one of the jewels of Indian architecture, the Ellora cave temples. Shirazi viewed Ellora not as religious architecture but as a primarily political monu- ment, which fit best into the category of the wonders of the world. When Shirazi's reaction to Ellora is compared with other accounts of it by Muslim authors, with Muslim accounts of other "pagan" monuments in Egypt, and with descriptions of Ellora by early European travelers, his aesthetic and political reaction does not seem very unusual. This ac- count is another reminder that, for premodern Muslims, the monolithic Islam defined by twentieth-century discourse was far from being the only or even the primary category of judgment. The text in question is Tadhkirat al-mulak (Memorial of kings), a Per- sian history of Bijapur written by Rafi' al-Din Shirazi in 1612.2 The au- thor (born in Shiraz in 1540) had a long career in Bijapur government service, from the age of thirty serving Sultan 'Ali' Adil Shah as a steward and scribe. In 1596, Sultan Ibrahim 'Adil Shah appointed him ambassa- dor to Ahmadnagar, and he also held posts as governor of the Bijapur fort and treasurer. Shirazi witnessed many important events over more than half a century in the Deccan, and he was also steeped in the tradi- tion of Persian historical writing, having written abridgements of stan- dard court chronicles such as Mir Khwand's Rawdat al-safd' and Khwand Amir's Habib al-siyar. His history is an important independent historical source comparable to the chronicle of Firishta.  100 | Ellora Temples as Viewed by Indo-Muslim Authors In the handwritten edition of Khalidi, the outline of the text is as fol- lows, divided into an introduction, ten parts, and an appendix: Introduction (1-15) I. The Bahmani dynasty (15-35) II-V. The 'Adil Shahi dynasty (36-83) VI. Dynasties of Gujarat, Ahmadnagar, and Golconda (84-156) VII. Various events in the Deccan (157-96) VIII. Ibrahim 'Adil Shah, the author's patron (197-269) IX. The Mughals (270-93) X. The Mughals and Safavids (294-496; in some MSS this lengthy section is divided in three parts to make twelve parts in all) Appendix. On Wonders and Rarities (497-566) The section under discussion occurs toward the end (476-83) of the tenth part, and although its title includes the phrase "wonders and rari- ties," it does not fall into the appendix proper; instead, it is sandwiched between accounts of military campaigns of the Safavids and the Mughals. The appendix consists of a series of accounts of mirabilia of the 'aj 'ib genre of wonders long established in Arabic and Persian litera- ture.3 Some of these wonders are related by others, although a few were seen by the author himself. These include narratives based on the Per- sian Book of Kings by Firdawsi (497-517), travelers' tales of strange is- lands (517-32), and accounts of the rivers and geography of India (532- 43), followed by brief reports of natural wonders (544-66). Shirazi's location of his account of Ellora in the dynastic history proper, and not in the appendix on wonders, suggests that he wished to treat it as a serious political concern, framed around a legendary Indian monarch named Parchand Rao. It thus remains separate from the super- ficially similar stories about fabulous islands and idol temples that occur in the appendix. Those remain comfortably in the realm of two-headed calves and other marvels, but the serious point that Shirazi wanted to make about art and royal monuments required that he situate the story of Ellora amidst similar political and military narratives. In this kind of arrangement Shirazi resembles the Egyptian chronicler of the pyramids, al-Idrisi, who kept his meticulous measurements and historical ac- counts of the pyramids in one chapter and saved the bizarre and the miraculous for the last chapter of his book.4 Shirazi's chapter has, how- ever, been circulated separately as a "Treatise on Wonders and Rarities," and in this form it would not have taken on the political coloring af-  Carl W. Ernst | 101 forded by its contextual position in the larger history.5 Here follows a translation of the extract: Description of the Wonders and Rarities of the Building of Ellora in Daulatabad, Which Parchand Ra6, the Emperor of India, Built Nearly 4,000 Years Ago 1. Parchand Rao was an emperor. With great majesty, he had brought under his control all the land from the border of Sind, Gujarat, the Deccan, and Telingana to the limit of Malabar, and most of the neighboring kings were his subjects. He was noble, just, and upright, and he lived in harmony with the people. The peasant and the soldier in the days of his reign were in all ways happy and free from worry. They passed all their lives in happi- ness, joy, contentment, and pleasure. 2. In the springtime, when the climate was perfectly mild, Parchand Rao would go on a tour of the kingdom, and he let the people partake of his magnificence. He made every effort to bring about justice and fairness. In every place that he saw abundant water, greenery, and good climate, he laid foundations for build- ings, and he supplied the officials of the kingdom every resource for completing them. In this way, having traveled through the en- tire kingdom three times, he constructed and brought to comple- tion lofty idol-houses (but-khana) outside the buildings just men- tioned throughout most of his kingdom. 3. Now as for the famous Daulatabad-fine and elegant fabrics were available there, and in the neighborhoods and environs mer- chants brought them as gifts and donations, and they still are ac- tive and do so; wealthy merchants full of tranquillity are always dwelling in that city, both Muslim and Hindu. Every year nearly a thousand ass-loads of different kinds of silken and gold-woven fabrics are brought to its neighborhoods and environs, and general welfare prevails. The same Parchand Rao made Daulatabad his capital, and people from the four corners of the world headed in the direction of Daulatabad. Most of this multitude came to a place that was nearly five or six farsakhs away, and having built houses and gardens, they settled there; tall houses were set up with some difficulty. 4. One day in the assembly of Parchand Rao there was a discus- sion of the construction of buildings and abodes, and the king said,  102 | Ellora Temples as Viewed by Indo-Muslim Authors "During my reign, I have built and finished many buildings in my dominion, but these ordinary buildings do not have much perma- nency. I want a building that will be truly permanent, so that it will be spoken of for years afterward, and there should be wonders and rarities in it so that it will endure and remain lasting for long years and uncounted centuries, and its construction will be famed and well known throughout the world." 5. Some of the architects, engineers, and stoneworkers were dedicated to the emperor and spoke his language, because of the many buildings that they had made. They said, "In the region of this very city there is a mountain that is unlike any of the moun- tains of the world. This is because the mountains that we have seen and see today are mostly of this kind: part is bedrock, and part is soft and has cracks and fissures. In this city there is a great and lofty mountain that has absolutely no cracks, joints, fissures, or rubble. In this way, one can make a great and lofty house, which every great king can do, for lofty buildings have been repeatedly built. If one brings together all the eighteen workshops of the realm, which are famous and well known, so that the supervisor does not need to have any other building built, and he has the capacity and basis associated with that workshop, and the quantity of men and ani- mals necessary for those workshops, then they will prepare every- thing from stone: the assembly of the king's realm, the private pal- aces, the soldiers in attendance on the king-all will be carved in stone, so that each will be established in the proper place. Until the dawn of the resurrection, that court, those workshops, and those people will all be preserved, each in the proper place. Such a court as this, this foundation, and this army will all be in five or six sec- tions of stone, with the human and the other animals of proper proportions in the same form and size in which they were created, neither larger nor smaller." 6. The emperor said, "This account that you have given, if it is possible and can indeed take form, is a wonder. By all means, let them make a model from wax or chalk so that I can have a look." When the artisans, engineers, and stoneworkers heard that the emperor asked for a model, they had to come to agreement and make a completed model such as the emperor had asked for from brick, clay, and chalk. When they invited the emperor to their pre- mises [to see the model], he became very happy, and he consented with delight.  Carl W. Ernst 1 103 7. Beginning from the middle of the mountain, they made a great open space in the palace, which they call the retreat (khilwat- khana). On all four sides of the open space, they cut open spaces (sar-saya, lit. "shades") in the stone, perfect in height, width, and length, with a polished and proportioned foundation. In most places these are carved in the fashion of great arches (taq) needing no pillars. The carving is extremely even and polished, or rather, is even given a luster. In some of these open spaces there are alcoves (bahl, usually bahla, lit. "purse") with caves. Their ability reaches such subtlety that if the master artist wished to paint one with a brush made from a single hair, nowhere would it be easy for him [to match their skill]. In some of the arches there is a string of cam- els, and in some a stable of horses. Some are with saddle, and some with colored blankets. There is no need to mention the extraordi- nary workmanship and subtlety again. One should compare the alcove with the palace; in each one of these palaces there are some human forms in the attitude of servants, which are necessary in those palaces. One would say that all are standing ready to serve, while some appear in such a way that one would say they are in the act of being rejected. The remaining animals, wild beasts and birds ... are everywhere in the manner of delivering an obligatory reply to a question. The forms of armed and equipped soldiers, to the number of one or two hundred, are as if ready for service, each one established in his own place. On the courtyard in front of the pal- ace gate, here and there several large and small elephants are standing in order. Around each elephant a few attendants stand in their regalia. Description of the Foundation of the Palace Fort and Its Capacity 8. Four arches (taq) cut from stone are on one side of the court- yard, and within, two shorter ones are in the place of the gate. Symmetrical in height, breadth, and length, these four are linked by a single roof. Two great benches (suffa) are built into the great arches, as a seat for servants, for the servants of the fort and the courtyard are within. Nearly five or six hundred people are sitting in their places, some standing fully armed. Outside of that, many weapons are carved in various places, such as swords, daggers, dirks (Hindi kat-ara), spears, bows, quivers, and arrows. One re- mains in astonishment at the subtle and painstaking work. In the  104 | Ellora Temples as Viewed by Indo-Muslim Authors courtyard, inside the four arches, are benches, porticoes (ayvan), and rooms carved and hollowed out in the same style. On one side are the imperial workshops, such as the armory, stable, water- works, kitchen, storehouse, and wine cellar. In every one of these palaces there are at least fifty or sixty human forms, each one of which appears to be in the act of performing something. The skill of each workshop is cut into rock to such a degree that the human mind cannot imagine it. Everyone who goes there says that the people [in the stone reliefs] are having a party. One should spend several days at the palace if one wishes to see them all, and to understand them fully a long lifetime would be needed. Many wild beasts and birds have also been added to these festivals to adorn the palace. 9. Proceeding behind this palace, there is a fort and some other palaces pertaining to the previously mentioned palace. Here too a multitude of figures is made in the form of servants, done with great workmanship, in a more prominent position, and the court- yard of this is greater than that in the previous palace. Some work- shops are set up in this palace, and benches, arches, and porticoes have been raised up to heaven. By way of workshops, things such as the bachelor quarters (dar al-'azab), goldsmith shop, fountain shop, wardrobe, treasury, and the like [have been made] with such subtlety and workmanship that a hair of a single brush could not have rendered it. The attendants of the workshops, their trade, tools, and basis of each workshop have been made to the necessary extent, each one being made in the performance of [the appropri- ate] action, and each servant of these palaces has been made firm in the proper position. A Hint of Conditions of the Court and the Arrangement of the Place of the Workmen and Attendants 10. Having made another palace with the arch and portico in perfect proportion, and having placed some smaller palaces to the sides with workmanship and beauty, and the imperial throne at the front of the portico, they fixed the portrait of the emperor upon it, depicting that amount of ornament on the limbs of the emperor that is customary among the people of India, some sculpted and some in relief. Its painstaking subtlety is beyond description. To the left and right of that throne, half-thrones have been prepared with solid foundations, and on each of these they have sat princes  Carl W. Ernst 1 105 and nobles of the realm. Behind the head and shoulders of the emperor are servants, friends, and relatives, each in the proper place. There are some watchmen holding swords with handker- chiefs in their hands, in the Deccan fashion. Waterbearers in their own manner and order hold vessels of water in their hands, and waiters (shira-chi) hold a few flagons with cups in their hands. Winebearers, by which I mean betel-leaf servers, hold trays of betel leaf in their hands with suitable accompaniments, some trays hav- ing sweet-scented things, for in each tray are cups of musk, saffron, and other items. The saucers in those trays are made in the fashion of cups, with pounded ambergris, sandalwood, and aloes, and aro- matic compounds are set forth, and trays full of roses. This portico, which is subtler in arrangement than a rose, is such that the de- scription, beauty, workmanship, and subtlety of workmanship of that assembly do not fit into the vessel of explanation. 11. In front of that portico of the court, the chief musician (sar-i nawbatan) and the court prefect (shihna-i divan) stand in the proper arrangement and position in their places. On both sides, nearly 2,000 horsemen, extremely well executed, are in attendance in the proper fashion. In the courtyard of the court and the portico, across from the emperor, there are several groups of musicians, each standing with his own drum and lute; one would say that they are dancing. In the same courtyard, tumblers, jesters, wres- tlers, athletes, and swordsmen exhibit their skill. One would say that each group in its particular area and assembly is right in the middle of its activity. Several famous and large elephants, which were always the apple of the emperor's alchemical eye, are in his presence, and several head of elite imperial horses, which were always present with the court drum, are present in the customary fashion. 12. So many beautiful and well-wrought things are in those buildings and courtyards that, if one wished to explain them all, he would fail to reach the goal. The listener should prepare for fatigue of the brain! 13. Outside this assembly, several other small banquet assem- blies have been made and constructed, which tongue and pen are unable to explain. Three or four private palaces have been built, and in each palace are the private inhabitants, who are the women and eunuchs-more than one or two hundred. Each one is in a distinct style and position, and a detailed account of the motions  106 | Ellora Temples as Viewed by Indo-Muslim Authors and postures of those palaces would not be inappropriate; it can be generally summarized in a few words. In each of these palaces some obscene activities-none repeated-are taking place. 14. In general, of that which is actually in existence at Ellora, not one part in a thousand has been mentioned. Few people have reached the limit of its buildings, and those who have [come] sim- ply take in the generality of it with a glance. What is presently observable and displayed takes up nearly two farsakhs. Even fur- ther, there are places with buildings and hunting lodges, but a wall of chalk and stone has been firmly set up, so no one goes past that place. It is famous. 15. There is a smaller building like this in a village at least fifty farsakhs from Ellora. It is said that in every place palaces, build- ings, and hunting lodges have been built in the same fashion, and it is still in existence. But God knows best as to the realities of the situation. Description of Various Matters on the Same Subject 16. There are several constructions of similar form in the neigh- borhood of Shiraz, and that region is called Naqsh-i Rustam and The Forty Towers (Chihil Sutun, i.e., Persepolis). In The History of Persia it is well known that there were four such towers that Jamshid had made, and on top of all the towers he had made a single tall building, so that these towers were pillars for that build- ing. He spent most of his time in that building sitting on the seat of lordship and holding public audience. The people from below bowed to him and worshiped him. In that building of Ellora, most places are roofed and dark. Some places are made with illumina- tion from windows, and most rooms have no roof and are perfectly illuminated. Since this was three or four thousand years ago, and in that time lifetimes were long, and humans were mighty of frame and full of power and strength, such places as have been written of [above], which they made-if anyone of this age wished to make them, and had a thousand people and a period of a thousand years, it is not known whether it could be carried out to completion. In fact, the intellect is astonished at that construction. 17. There was always a joke about that building which was shared between the former Burhan Nizam Shah and Shah Tahir. The Nizam Shah used to say that sodomy was brought to the Dec- can during the present time by foreigners [i.e., Persians]. Shah  Carl W. Ernst | 107 Tahir objected that this practice is immemorial in this kingdom. Once when they went to visit the buildings of Ellora, Shah Tahir saw a depiction of two men embracing each other. He took the hand of the Nizam Shah and brought him near that depiction, say- ing, "Have foreigners brought this also?" By this example, he re- moved the Nizam Shah's displeasure with foreigners. Description of the Idol Temple of the Town of Lakmir 18. In the neighborhood of Bankapur is a town called Lakmir. In ancient times, it was the capital of one of the great emperors of unbelief. With the greatest architectural skill, the emperors, princes, and pillars of the realm built many idol temples in imita- tion of one another, extremely large and well built. Years passed, and most of the buildings fell into ruin, and only a few were still inhabited. But four hundred idol temples remained perfectly sound, having been constructed with the utmost of painstaking and elegant workmanship. At the time when we saw it, we saw many wonders and rarities, and astonishment increased upon as- tonishment. Out of all those, we saw one idol temple with dimen- sions of seventy cubits by fifty cubits. Both inside and outside of it a trough (taghari) had been cut in relief. Its subtlety was to the degree that in the space of a hand, in natural proportions, the forms of ten men had been made, along with the forms of ten or fifteen animals, both beasts and birds, in such a way that the eyelashes and fingernails were visible. On the border were roses, tulips, and trees of the locality, about the size of one hand. This degree of art- istry has been forgotten. 19. Imagine how much work has been done on the inside and outside of all the idol temples, and how many days and how much time it took to complete them. May God the exalted and transcen- dent forgive the World-Protector [i.e., 'Ali' Adil Shah, d. 988/1580] with the light of his compassion, for after the conquest of Vijay- anagar, he with his own blessed hand destroyed five or six thou- sand adored idols of unbelief, and ruined most of the idol temples [at the battle of Talikota or Bannihatti, January 1565]. But the lim- ited number [of buildings] on which the welfare of the time and the kingdom depended, which we know as the art of Ellora in Dau- latabad, this kind of idol temple and art we have forgotten. There are several striking aspects to this text. First of all, Shirazi makes hardly any reference to Indian religions in his description of  108 | Ellora Temples as Viewed by Indo-Muslim Authors Ellora. Second, he appreciates the monument on an aesthetic level, and he explains its origin in political terms. For him, Ellora is a royal monu- ment that depicts the court life of an ancient king of India, making it comparable to pre-Islamic Persian monuments such as Persepolis.6 The statue of Shiva in the Kailas temple is explained as a royal portrait. Third, and most unexpectedly, he only makes a strong bow to religion when he calls upon God to forgive his former patron, Sultan 'Ali 'Adil Shah, for destroying the temples of Vijayanagar. This last gesture turns the stereotype of Muslim iconoclasm on its head. Shiraz- acknowledges that temple destruction has taken place in military and political contexts of conquest, but he deplores it as a violation of beauty and, ultimately, as an offense against God. Although he does not mention it, the temple at Bankapur, which he also admires, was evidently the "superb temple" that 'Ali' Adil Shah destroyed and replaced with a mosque when he took the city in 1575.7 Shirazi's strong emotional and religious reaction against the destruction of temples is all the more noteworthy in view of his basically conservative Muslim attitude; his account of the religious innovations of the Mughal emperor Akbar is highly critical, closely re- sembling Bada'nni's negative view of Akbar rather than the universalist perspective of Abu al-Facjl.8 Shiraz- was not the first Muslim to appreciate the importance of Ellora. The Arab scholar Mas'ndi (d. 956) spent several years as ambas- sador to the powerful Rashtrakuta empire, under whose auspices some of the temples of Ellora were constructed; the Rashtrakutas had friendly relations with the Arabs, whom they viewed as allies against the Gur- jaras of northern India.9 In his Meadows of Gold, in the context of a lengthy disquisition on temples of the ancient world, Mas'ndi briefly describes the temple of Ellora in the following passage, noting that in another place (unfortunately, a lost work) he has more fully discussed "the temples (hayakil) in India dedicated to idols (asnam) in the form of Bud- dhas (bidada), which have appeared since ancient times in the land of India, and information about the great temple which is in India, known as Ellora; this is an object of pilgrimage (yuqsadu) from far distances in India. It has a land endowment, and around it are a thousand cells, where monks supervise the worship (ta :im) of this idol in India."10 It should be noted that in this account, Mas'ndi does not distinguish between Hindu, Buddhist, or Jain temples and images; the words for "idol" in Arabic (bidada) and Persian (but) were in fact derived from Buddha (he immediately follows this reference with a vague note about  Carl W. Ernst 1 109 the temple to the sun in Multan).11 Indian temples are viewed here in a continuum with Roman, Egyptian, and Sabian temples, a point to which we shall return. Later references to Ellora by Muslim authors belong to the period after the Turkish conquest of the Deccan, when the temples of Ellora had ceased to function as an active religious center. According to Firishta, it was during some unofficial sightseeing at Ellora in 1307 that some Turkish soldiers stumbled across the Hindu princess Dewal Rani, whom they captured and brought to Delhi as a bride for Khidr Khan.12 In 1318, Sultan Qutb al-Din Mubarak Shah Khalji spent a month at Ellora awaiting the return of his general, Khusraw Khan, from campaigns in Warangal.13 A tradition related in a current gazetteer maintains that 'Ali' al-Din Hasan, founder of the Bahmani dynasty of the Deccan, visited Ellora in 1352, "taking with him those who could read the inscriptions and understand the significance of the frescoes and statuary on the walls."14 We have seen above how the ruler of Ahmadnagar, Burhan Nizam Shah, and his Persian minister, Shah Tahir, used to visit Ellora for pleasure. The most surprising of all the admirers of Ellora is none other than the Mughal emperor Awrangzib, who spent years in the Deccan, first as governor under Shah Jahan, and later as emperor reducing the Deccan sultanates and quashing Maratha rebels. He was buried in 1707 in the Chishti shrine complex at Khuldabad, just a few miles down the road from Ellora. In a letter, Awrangzib recorded a visit to Khuldabad, Dau- latabad, and Ellora, describing the latter as "one of the wonders of the work of the true transcendent Artisan (az 'ajd'ibat-i sun '-i sani'-i haqiqi subhanahu)," in other words, a creation of God.15 The tourist visiting Ellora today is inevitably informed that half-ruined elephants, etc., are due to Awrangzib's fanatical destruction of idols, but there is no histori- cal evidence to indicate that the emperor engaged in any destruction there, or why he would have stopped with so much left undone. J. B. Seely, a British soldier who spent several weeks on furlough at Ellora in 1810, recorded many reports from local informants on idol smashing and cow slaughter by Awrangzib at Ellora, but he viewed them with the same skepticism that he reserved for tales of Portuguese doing the same.16 Catherine Asher has pointed out that the reports of Awrangzib's iconoclasm in the Deccan are typically from late sources that may reflect nothing more than legends that were hung on Awrangzib; his docu- mented acts of temple destruction were almost all associated with put- ting down political rebellions.17 Ironically, some of the examples of  110 | Ellora Temples as Viewed by Indo-Muslim Authors Awrangzib's temple destruction given by these late sources are failed attempts, frustrated by snakes, scorpions, or a deity. It seems that temple destruction is viewed as an essential characteristic of Awrangzib, re- gardless of whether he succeeded in actually carrying it out.18 The reaction of Shirazi to the destruction of Vijayanagar's temples can be compared to that of certain Muslim writers in Egypt in the thir- teenth century, who were enthusiastic admirers of the great pyramids at Giza. As Ulrich Haarmann put it, they were "deeply disturbed by the brutal demolition of intact pharaonic remains and the mutilation of pa- gan pictorial representations in the name of Islam, yet in reality all too often out of a very mundane greed for cheap and at the same time high- quality building materials."19 Similarly one may quote the physician 'Abd al-Latif, who in 1207 made the following remarks about Egyptian temples: "It is useless to halt to describe their greatness, the excellence of their construction and the just proportion of their forms, this innumer- able multitude of figures, of sculptures both recessed and in relief, and of inscriptions that they offer to the admiration of spectators, all joined to the solidity of their construction and the enormous size of the stones and materials in use."20 The literature of Muslim travelers in fact contains much of this kind of admiration for ancient "pagan" monuments. The non-Islamic origin of these temples does not seem to have been a particularly big stumbling block to Muslim tourists. Some, like Shirazi, simply found religion irrelevant to their appreciation. Others were able to assimilate the non-Islamic religious traditions to acceptable catego- ries. A number of Muslim authors interpreted the religion of the ancient Egyptians as forming part of the Sabian religion, an obscure Qur'anic term which permitted groups such as the Hellenistic pagans of Harran to function as "people of the book" for centuries.21 Popular Coptic my- thology combined with Hermetic lore permitted Muslims to identify the great pyramids as the tombs of Agathodaimon (Seth), Hermes (Idris), and Sab, founder of the Sabeans, or else as the constructions of the Arab ancestor Shaddad ibn 'Ad.22 Further examples can be added to the dossier of Muslim tourists who wrote appreciatively of Indian temples. The Timurid ambassador 'Abd al-Razzaq Samarqandi, who visited Vijayanagar at the order of Shah Rukh in 1442, reported with delight on the functioning temples he vis- ited en route near Mangalore and Belur. He compared these temples to the paradisal garden of Iram mentioned in the Qur'an, and remarked that they were covered from top to bottom "with paintings, after the manner of the Franks and the people of Khata [Cathay]."23 Another in-  Carl W. Ernst | 111 stance is the Afghan traveler Mahmad ibn Amir Wall Balkhi, who wrote a Persian narrative of a journey from Balkh to India and Ceylon and back, completed after seven years' travel in 1631. He traveled for plea- sure only, and on his return to Balkh he was appointed to a librarian's position. He has described at length, though with some disparagement, the rituals performed at the Krishna temple constructed by Raja Man Singh near Mathura. More entertainingly, he has related his own partici- pation in the festival at the Jagannath temple in the city of Puri, where by his own admission he doffed his clothes and joined the throng of pil- grims, thus participating in the dramatic rituals firsthand.24 There are undoubtedly other similar accounts. Shirazi's aesthetic delight in Ellora places his reaction in a category separate from the moralizing reactions to vanished earthly glory, the theme of ubi sunt qui ante nos in mundo fuere. Shirazi would have been familiar with the great Persian poem of Khaqani (d. 1199) on the ruins of the ancient Persian palace at Ctesiphon, the famous Tuhlfat al-'iraqayn. Unlike Khaqani and the Egyptian al-Idrisi, Shirazi does not draw an admonition ('ibrat) from the fall of kingly power.25 In his view, the de- struction of the temples of Vijayanagar is a cause for meditation not on the vanity of human wishes but rather on the tragedy of the loss of beauty. Shirazi's perspective contrasts with that of figures such as the Naqshbandi Sufi leader Almad Sirhindi, whose anti-Indian attitude led him to regard the ruins scattered over India as evidence of divine pun- ishment for failure to pay heed to divinely inspired messengers.26 Later Muslim tourists at Ellora would combine moralizing reflection on the decline of ancient pagans with enjoyment of the beautiful natural and artistic setting. Here is how this kind of reflection is presented in the Ma athir-i 'Alamgir, a history of Awrangzib's reign, completed in 1711: A short distance from here [i.e., Khuldabad] is a place named Ellora where in ages long past, sappers possessed of magical skill excavated in the defiles of the mountain spacious houses for a length of one kos. On all their ceilings and walls many kinds of images with lifelike forms have been carved. The top of the hill looks level, so much so that no sign of the buildings within it is apparent [from outside]. In ancient times when the sinful infidels had dominion over this country, certainly they and not demons (jinn) were the builders of these caves, although tradition differs on the point; it was a place of worship of the tribe of false believers. At present it is a desolation in spite of its strong foundations; it  112 | Ellora Temples as Viewed by Indo-Muslim Authors rouses the sense of warning [of doom] to those who contemplate the future [end of things]. In all seasons, and particularly in the monsoons, when this hill and the plain below resemble a garden in the luxuriance of its vegetation and the abundance of its water, people come to see the place. A waterfall a hundred yards in width tumbles down from the hill. It is a marvelous place for strolling, charming to the eye. Unless one sees it, no written description can correctly picture it. How then can my pen adorn the page of my narrative?27 In this passage the moralizing tone is almost a perfunctory note, inserted in what is for the most part an enthusiastic report. To modern Muslim scholars, Ellora provides a very different sort of lesson. Now equipped with the religious analysis that separates Hindu- ism, Jainism, and Buddhism, the contemporary Iranian Indologist Jalali Na'ini cites Ellora as one of a series of Indian monuments that form an outstanding ancient example of that modern religious virtue, religious tolerance. "Apparently, prior to the edict [of Ashoka] in the Indian sub- continent, as early as the Vedic age, there was a kind of tolerance and patience between followers of various religions in terms of differing beliefs. Support for this assumption includes the hymns of the Veda and the caves of Ajanta and Ellora. In these caves the temples of three reli- gions-Hindu, Jain, and Buddhist-are located in the bosom, the very heart of the mountains and hills of the Vindhya mountain range, about 60 miles from Aurangabad, and they can be taken as a clear sign of reli- gious freedom and the search for peace and tranquillity among the fol- lowers of the three indigenous religions of India."28 The vocabulary and conceptual apparatus of this remark derive from the European enlightenment rather than from medieval Islamicate cul- ture. Nonetheless, one might characterize it as yet another Muslim reac- tion to Ellora, which puts the cave temples into a historical sequence constructed in terms of the relations between religions. It is also interest- ing to consider the estimate of Ellora by the former head of the archeo- logical service of Hyderabad state, the well-known Muslim scholar Ghulam Yazdani: "At Ellora the religious fervor of the followers of the Brahmanical faith has carved out in the living rock temples which might well have been considered to be the work of gods not only by the vota- ries of that religion but also by the most discerning critic of the period, because they are unique specimens of this kind of architecture in the world."29 The British, in contrast, tended to be reassured by looking at  Carl W. Ernst | 113 these monuments, since they saw no one in India capable of building such grandeur who thus might prove an obstacle to their plans.30 As Seely put it, "Surely these wonderful workmen must have been of a different race to the present degenerate Hindoos, or the country and government must have been widely different from what it is at the present day."31 We would doubtless ascribe this reaction to the colonial mentality rather than to any internal imperative derived from Chris- tianity. Today every Indian schoolchild is taught the names of the ancient and medieval kings of India. Harsha and Candragupta Maurya are at least as well known as Alexander and Caesar are to western history texts. It is often forgotten that before the nineteenth century, and the prodigious antiquarian efforts of early orientalists and the Archeological Survey, these names had vanished from living memory. The rise and fall of multiple dynasties had erased the meaning of many monuments that dot the Indian landscape. Oral narratives were bound to replace lost traditions with plausible tales about the mighty men of old capable of building such wonders. We do not know what stories were told to Bija- pur officials by local dwellers in the vicinity of Daulatabad about the impressive temples of Ellora, but they may well have been connected to images of the Daulatabad fort, which has notable stylistic similarities with the construction of Ellora.32 Shirazi's political interpretation of the monument does not seem strange when compared with the explana- tions that were offered to Seely by his guides in 1810. Large guardian figures were still being identified with Persian terms from Indo-Muslim court life, such as chabdar (mace-bearer) and pahlavan (wrestler).33 It is hard to recall that, before the age of modern tourism, travelers were not likely to see evidence of what we would call a foreign culture. The first European explorers of Asia and the New World went equipped with fantasies like The Travels of Sir John Mandeville, and they saw the cannibals, Amazons, and giants that they were prepared to see. Early European engravings of Indian idols have more than a passing resem- blance to Roman deities. When the Portuguese explorer Vasco da Gama and his crew arrived in India in 1498, so great was their relief in seeing buildings that were evidently not "Moorish" mosques that they ac- cepted the Hindu temples of Calicut as Christian churches, kneeling in prayer before goddesses that they described as images of the Virgin Mary and the saints (they were evidently unconcerned by the unusually large teeth and extra arms of these images).34 Seely notes that the first Indian soldiers sent to Egypt, in British military expeditions to combat  114 | Ellora Temples as Viewed by Indo-Muslim Authors Napoleon, announced in amazement that the ancient Egyptians clearly worshiped Hindu gods in their temples; this was probably the first In- dian hermeneutic of pharaonic antiquities.35 In a sense the response of the sepoys was a repetition of the reactions of early visitors from Hero- dotus onwards, who described the gods of Egypt in terms of their own theologies. When Shirazi saw Ellora as analogous to Persepolis, he was only making a natural comparison from his own experience of ancient monuments. Seely did much the same when he described what he saw as Sphinxes at Ellora.36 Muslims were not the only ones to reinvent Ellora's significance along new lines. When the Rashtrakutas conquered the Chalukyas and took over power in the Deccan in the seventh century, in addition to adding new Hindu monuments such as the Kailas temple, they con- verted Buddhist viharas into Hindu temples, chiseling out many Bud- dha images at Ellora and covering or replacing some with images of Vishnu.37 Architectural guidebooks unfortunately do not indicate what essential characteristic of Hinduism caused this extreme form of renova- tion. The Yadavas of Deogir were not a direct extension of the Rash- trakutas, and they must have formed their own interpretations of the meaning of Ellora, a monument near the center of their empire. While we can only speculate about the way the Yadavas positioned themselves in relation to Ellora, their interpretation must have reflected their own self-interpretation as a successor-state to the Rashtrakutas. The founder of the Mahanabhuva sect, Cakradara, is said to have briefly established a new form of worship in Ellora that was completely unrelated to the Shaiva, Buddhist, and Jain traditions of earlier eras.38 Ellora evidently took on a new significance among the elites of the Marathas, starting from the sixteenth century. As James Laine points out, Maloji, grand- father of the Maratha warrior Shivaji, is buried in an Islamicate tomb in the village of Ellora.39 In the eighteenth century, Ellora evidently received further patronage from the ruling Maratha family of the Holkars, who must have interpreted the monuments in terms of their own political and religious position.40 European travelers such as Anquetil du Perron in 1760 and Seely in 1810 were informed by local brahmins that Buddha images in some of the caves actually represented Vishvakarma (a form of Vishnu), and Seely was given conflicting opinions about the meaning of Jain figures in a cave that the guides regarded as dedicated to Jagannath (another form of Vishnu).41 These Hindu names for Buddhist and Jain temples are  Carl W. Ernst | 115 still used in current guidebooks. Anquetil was also told that a number of Ellora temples were the various tombs of Vishnu; his brahmin infor- mants said that other cave temples near Bombay had been built by Alexander.42 Goddess figures at Ellora were always identified for Seely as Bhavani, following her ascendancy in modern Maratha culture. Seely occasionally caught his guides changing their identifications of images, but this he attributed to the confusion inherent in Hindu mythology rather than to any other cause.43 Col. Meadows Taylor, author of Confes- sions of a Thug, claimed that a Thug told him that the Ellora caves con- tained depictions of all the methods of murder employed by the Thugs.44 All this goes to say that Ellora, like any ancient monument, has not had a single fixed meaning over time. The precincts were constructed over centuries, with multiple religious patterns that we today distinguish by the categories of Hindu, Buddhist, and Jain. Different generations of patrons contributed their own interpretations with their commissions and constructions. Just as the monuments themselves are subject to physical modification by later visitors and patrons, so their meaning has been adjusted to the symbolic parameters of new civilizational orders. As far as the question of Muslim iconoclasm is concerned, the evi- dence of Muslim travelers who visited Hindu temples does not provide justification for assuming that idol-smashing activity is easily detect- able, much less the visceral instinct that it is often assumed to be. The examples cited above are not random or selective, but constitute the results of a fairly extensive search for textual reactions by Muslims to Hindu temples. Why should we assume that Muslims are by nature and training iconoclastic, and when they do violence to idols or temples, why do we assume that this behavior is rooted in Islamic faith? Take the example of Babur, in an incident that took place near Gwalior in 1528. On that occasion, he recorded a bout of severe opium sickness with much vomiting. The next day, he saw some Jain statues, which he de- scribed as follows: "On the southern side is a large idol, approximately 20 yards tall. They are shown stark naked with all their private parts exposed. Around the two large reservoirs inside Urwahi have been dug twenty to twenty-five wells, from which water is drawn to irrigate the vegetation, flowers, and trees planted there. Urwahi is not a bad place. In fact, it is rather nice. Its one drawback was the idols, so I ordered them destroyed." The following day, he visited Gwalior fort. "Riding out from this garden we made a tour of Gwalior's temples, some of which are two and three stories but are squat and in the ancient style with dadoes en-  116 | Ellora Temples as Viewed by Indo-Muslim Authors tirely of figures sculpted in stone. Other temples are like madrasas, with porches and large, tall domes and chambers like those of a madrasa. Atop the lower chambers are stone-carved idols. Having examined the edifices, we went out."45 At that point he enjoyed an outdoor feast. What part of Babur's behavior during these three days was Islamic? On day one, he was hung over from drug intoxication, on day two, he destroyed two naked Jain idols, and on day three he enjoyed a pleasant excursion to Hindu temples with the governor of Gwalior fort and left the idols there intact. Why did he destroy idols on one day and enjoy them the next? His good mood on the third day may have had some- thing to do with either his recovery from hangover or the embassy of submission he received that morning from a major Rajput ruler. Alterna- tively, he may have considered it ill-mannered to destroy part of a monument he was being shown, in a fort that one of his subordinates was in charge of. In any case it is clear that it is highly problematic to predict political behavior (such as destruction of temples) from the nominal religious identity that may be ascribed to an individual or group, without reference to personal, political, and historical factors. Above all, it is noteworthy that the occasions when Muslim writers have invoked God and religion in relation to Hindu monuments have been when they have been awed by the creation of beauty. While Rafi' al- Din Shirazi in a sense reduced the significance of Ellora to the familiar terms of imperial monuments, he was also stirred to protest on religious grounds against the iconoclasm of his imperial patron. It does not seem accidental that at the moment of praising the extraordinary, even in what seems the stereotyped convention of the wonders of the world, the emotion of reverence should take control. It would be a shame if contem- porary ideological conflicts blinded us to the perception of the profound admiration that Indian monuments like Ellora have evoked in Muslim visitors. More to the point, accounts like Shirazi's indicate that Muslims had complex reactions to non-Muslim religious sites. Their responses could be dictated by a variety of factors, including their education and temperament, the political situation, and whether the building fell into the category of ancient wonder or living temple (Muslims seem to have enjoyed both). The popular one-dimensional portrait of Muslim icono- clasm survives as a durable stereotype because it does not acknowledge its subjects as actors in historical contexts. The iconoclasm stereotype derives not from the actual attitudes of Muslims toward temples, but from a predetermined normative definition of Islam. The reasons for the  Carl W. Ernst | 117 appeal of such religious stereotypes, ironically, will need to be sought elsewhere. Notes 1. To this category of the trophy belongs the transport of Ashokan columns and other ancient pillars, of which the Iron Pillar of Delhi is but one example. These trophies may be found in royal mosques of the Sultanate period at Hisar and Jaunpur as well as the Quwwat al-Islam mosque of Delhi, and possibly at Tughlugabad as well. Cf. Mehrdad Shokoohy and Natalie H. Shokoohy, "Tugh- luqubad, the Earliest Surviving Town of the Delhi Sultanate," Bulletin of the School of Oriental and African Studies 57 (1994): 548. 2. Iam basing this analysis on the critical edition of the text established by the late Aba Narr Khalidi, which has been entrusted to me by his son, Omar Khalidi, to see through the press; it is to be published by the Islamic Research Foundation of the Asitan-i Quds-i Rizawi in Mashhad, Iran. For further information on this author, see my articles "Shirazi, Rafic al-Din," in The Encyclopaedia of Islam, ed. H. A. R. Gibb et al., 2nd ed. (Leiden: E. J. Brill, 1960-), IX, 483 (cited henceforth as E2), and "Ebrahim Shirazi," in Encyclopedia Iranica (Costa Mesa, Calif.: Mazda Publishers, 1986-), VIII, 76. 3. C. E. Dubler, "'Adja)ib," E2, I, 203-4. 4. Ulrich Haarmann, "In Quest of the Spectacular: Noble and Learned Visi- tors to the Pyramids around 1200 A.D.," in Islamic Studies Presented to Charles J. Adams, ed. Wael B. Hallaq and Donald P. Little (Leiden: E. J. Brill, 1991), 65-66. 5. Haji Muhammad Ashraf, Catalogue of Persian Manuscripts in the Salar Jung Museum and Library, vol. 2, Biographies, Geography, and Travels (Hyderabad: Salar Jung Museum and Library, 1966), 277, no. 643. 6. On Persepolis, see M. Streck [G. C. Miles], "Istakhr," E2, IV, pp. 219-22. It is worth noting that the author of this article attributes the defacement of human figures at Persepolis to "Muslim fanaticism," something that calls for further analysis. 7. Mahomed Kasim Ferishta, History of the Rise of the Mahomedan Power in India, Till the Year A.D. 1612, trans. J. Briggs, 4 vols. (London, 1829; reprint, Lahore: Sang-e Meel, 1977), 3:84, dates this to 1573, but epigraphic evidence places this conquest in December 1575; see H. K. Sherwani and P. M. Joshi, eds., History of Medieval Deccan, 1295-1724, 2 vols. (Hyderabad: Government of Andhra Pradesh, 1973-74), 1:335. 8. Iqtidar Alam Khan, "The Tazkirat ul-Muluk by Rafi'uddin Ibrahim Shirazi: As a Source on the History of Akbar's Reign," Studies in History 2 (1980): 41-55. 9. Andre Wink, Al-Hind: The Making of the Indo-Islamic World, vol. 1, Early Medieval India and the Expansion of Islam, 7th-11th Centuries (Delhi: Oxford Uni- versity Press, 1990), 303-9, esp. 305. 10. Abu al-IHasan 'Al! ibn al-Husayn ibn 'Al! al-Mas'adi, Muraj al-dhahab wa  118 | Ellora Temples as Viewed by Indo-Muslim Authors ma adin al-jawdhir, 4th ed., ed. Muhammad Mul.yi al-Din Abd al-Hamid, 4 vols. (Egypt: al-Maktaba al-Tajariyya, 1384/1964), 2:262; cf. Mas'adi, Les Prairies d'or, trans. Barbier de Maynard and Pavet de Courteille, ed. Charles Pellat, Collection d'Ouvrages Orientaux (Paris: Societe Asiatique, 1965), 2:547, §1424, correspond- ing to 4:95-96 in the nineteenth-century edition of the Arabic text by Barbier de Maynard. There are problems in the Arabic text published in Egypt; I have fol- lowed the French translators in reading bidada rather than badra (which would result in "the form of the moon" rather than "in the form of Buddhas"), and Ellora (AlUrd) rather than the anomalous MS readings al-adri and bildd al-ray. Both Arabic editions are in error, however, in reading jawarin ("female slaves," pl. of jdriya) in place of jiwarun ("resident pilgrims," pl. of jar, probably in this case meaning Jain monks); this led the French translators to render the last phrase as "jeunes esclaves destin es aux pelerins qui viennent de toute l'Inde pour adorer cette idole." From what we know of Ellora under the Rashtrakutas, it would have functioned as a monastery rather than as a massive devaddsi center. 11. The British traveler Seely, too, was fairly vague about the relations be- tween Hinduism and Buddhism; see J. B. Seely, The Wonders of Ellora or the Nar- rative of a Journey to the Temples or Dwellings Excavated out of a Mountain of Granite at Ellora in the East Indies (London, 1824), 197-98. 12. Aba al-Qasim Firishta, Gulshan-i Ibrahimi (Lucknow: Nawal Kishor, 1281 / 1864-65), 1:117; Mahomed Kasim Ferishta, History of the Rise of the Mahomedan Power in India, 1:210. 13. Banarsi Prasad Saksena, "Qutbuddin Mubarak Khalji," in Mohammad Habib and Khaliq Ahmad Nizami, ed., A Comprehensive History of India, vol. 5, The Delhi Sultanate, (A.D. 1206-1526) (New Delhi: People's Publishing House, 1970; reprint, 1982), 436. 14. Aurangabad District, Maharashtra State Gazetteers, 2d ed. (Bombay: Gazet- teers Department, Government of Maharashtra, 1977), 88. This information is apparently drawn from an important modern Urdu history of the Deccan, Muhammad Abd al-Jabbar Mulkapuri, Mahbab al-watan, tazkira-i saldtin-i Dakan, vol. 1, Dar baydn-i salatin-i Bahmaniyya (Hyderabad: Matba'-i Fakhr-i Ni- zami, n.d.), 147-50, which is followed by a lengthy and enthusiastic appreciation of the Ellora caves. 15. Inayatullah Khan Kashmiri, Kalimat-i-Taiyibat (Collection of Aurangzeb's Orders), ed. and trans. S. M. Azizuddin Husain (Delhi: Idarah-i Adabiyat-i Delli, 1982), 27 (English), 13 (Persian). 16. Seely, The Wonders of Ellora, 150, 165, 202, 245, 345. 17. Catherine B. Asher, The New Cambridge History of India, 1:4, Architecture of Mughal India (Cambridge: Cambridge University Press, 1992), 254. As an ex- ample of later sources on Awrangzib's temple destruction, she notes Jadunath Sarkar, History of Aurangzib (Bombay: Orient Longman, 1972), 3:185 (not 285), who cites a Marathi source dated Saka 1838 (1916 C.E.). 18. Seely's brahmin informants told him "that if Aurungzebe actually did not commit the atrocious act himself, he allowed his court" (241).  Carl W. Ernst | 119 19. Haarmann, "Quest," 65. See also Haarmann, ed., Das Pyramidenbuch des Abu Oafar al-Idrisi (st. 649/1251), Beiruter Texte und Studien, 38 (Stuttgart: Franz Steiner, 1991). 20.' Abd al-Latif, Relation de l'Egypte par Abd-allatiph, trans. Silvestre de Sacy (Paris, 1810), 182, quoted in Gaston Wiet, L'Egypte de Murtadi fils de Gaphiphe (Paris: Librairie Orientaliste Paul Geuthner, 1953), introduction, 98. 21. Wiet, L'Egypte de Murtadi fils de Gaphiphe, 60. 22. Ibid., 2, 87-88. Further on the Arabic Hermetic histories of pre-Islamic Egypt, see Michael Cook, "Pharaonic History in Medieval Egypt," Studia Islamica 57 (1983): 67-104. 23. R. H. Major, ed., India in the Fifteenth Century, Being a Collection of Narratives of Voyages to India . .., Works Issued by the Hakluyt Society, 22 (London: Hakluyt Society, 1857), 20-21; cf. C. A. Storey, Persian Literature: A Bio-bibliographical Sur- vey, 2 vols. (London: Luzac, 1927-71), 1:293-98. 24. Malhmad ibn Amir Wall Balkhi, Bahr al-asrdr fi mandqib al-akhydr, ed. Riazul Islam (Karachi: Institute of Central and West Asian Studies, 1980), 13-16, 32-38, of the Persian text. See Iqbal Husain, "Hindu Shrines and Practices as Described by a Central Asian Traveller in the First Half of the Seventeenth Cen- tury," in Medieval India I: Researches in the History of India, 1200-1750, ed. Irfan Habib (Delhi: Oxford University Press, 1992). 25. Haarmann, "Quest," 58. 26. Yohanan Friedmann, Shaykh Ahmad Sirhindi: An Outline of His Thought and a Study of His Image in the Eyes of Posterity, McGill Islamic Studies, 2 (Montreal: McGill-Queen's University Press, 1971), 71. 27. Saqi Mustcad [sic] Khan, Ma'dsir-i-' Alamgiri: A History of the Emperor Aurangzib-' Alamgir (reign 1 658-1707 A.D., trans. Jadu-nath Sarkar, Bibliotheca In- dica, no. 269 (Calcutta: Royal Asiatic Society of Bengal, 1947), 145 (passage dated 1094/1683); this translation is superior to that in H. M. Elliot, The History of India as Told by Its Own Historians, ed. John Dowson, 8 vols. (London, 1867-77; reprint ed., Allahabad: Kitab Mahal, n.d.), 7:189-90. 28. Muhammad Dara Shikah, Majma' al-bahrayn, ed. Mul.ammad Ricja JalalI Na'ini (Tehran: Nashr-i Nuqra, 1366/1987-88), introduction, v-vi. 29. Ghulam Yazdani, "Fine Arts: Architecture," in Yazdani, ed., The Early His- tory of the Deccan, 2 vols. (London: Oxford University Press, 1960), 2:731. 30. Seely, The Wonders of Ellora, 230, quoting Lieutenant Colonel Fitzclar- ence. 31. Ibid., 258. 32. Seely (145-47) was informed that the Ellora caves were excavated by the Pandavas prior to the main action of the Mahabharata. 33. Ibid., 139, 299. Seely also records that "two colossal figures resting on large maces" were called dewriesdars (172), apparently from the Hindi term deorhi (door) plus the Persian suffix -ddr (holder); cf. Sarkar in Madsir-i- 'Alamgiri, 325. Modern scholars unselfconsciously go back to the classical Sanskrit term dwarapala to describe the massive doorkeepers at Ellora (Surendranath Sen, ed.,  120 | Ellora Temples as Viewed by Indo-Muslim Authors Indian Travels of Thevenot and Careri, Indian Records Series [New Delhi: National Archives of India, 1949], 320 n. 6). 34. K. G. Jayne, Vasco da Gama and His Successors, 1460-1580 (London: Meth- uen, 1910), 55. 35. Seely, The Wonders of Ellora, 156-67. It was particularly representations of the bull (i.e., Nandi) and of serpents that aroused recognition among the "Bom- bay Siphauees." 36. Ibid., 156-58. 37. Yazdani, The Early History of the Deccan, 2:731. 38. T. V. Pathy, Elura: Art and Culture (New Delhi: Humanities Press, 1980), 4. 39. James Laine, "The Construction of Hindu and Muslim Identities in Ma- harashtra, 1600-1810," paper presented at conference on "Indo-Muslim Identity in South Asia," Duke University, May 1995. 40. Seely, The Wonders of Ellora, 152. 41. Anquetil du Perron, Le Zendavesta, 3 vols. (Paris, 1771), 1:ccxxxiii, cited in Jean-Luc Kieffer, Anquetil-Duperron: L'Inde en France au XVIIIe si cle (Paris: Societe d'Edition "Les Belles Lettres," 1983), 347-63 (Duperron's map of the caves, with identifications proposed by his informants, is in Bibliotheque Nationale, Nouvelles acquisitions frangaises, Fonds Anquetil-Duperron, 8878); Seely, The Wonders of Ellora, 205 ff., 238-39. 42. Partha Mitter, Much Maligned Monsters: History of European Reactions to Indian Art (Oxford: Clarendon Press, 1977), 107-8. 43. Seely, The Wonders of Ellora, 286. 44. L. F. Rushbrook Williams, A Handbook for Travellers in India, Pakistan, Bangla- desh, and Sri Lanka (Ceylon), 22d ed. (New York: Facts on File, 1975), 149 n. 1. 45. The Baburnama: Memoirs of Babur, Prince and Emperor, trans. Wheeler M. Thackston (Washington, D.C.: Freer Gallery of Art/Arthur M. Sackler Gallery, 1996), 406-7; cf. Zahiru'd-Din Muhammad Babur Padshah Ghazi, Babur-ndma (Memoirs of Babur), trans. Annette Susannah Beveridge (1922; New Delhi: Ori- ental Books Reprint Corporation, 1979), 608-13. Beveridge notes that Babur's destruction amounted to cutting off the heads of the idols, which were restored with plaster by Jains in the locality.  5 Mapping Hindu-Muslim Identities through the Architecture of Shahjahanabad and Jaipur Catherine B. Asher As C. J. Fuller has rightly observed, scholars interested in temples and temple ritual tend to focus on South India, where presumably there is relatively little impact of Muslim conquest and rule. As a result, he notes, we know very little about temples in North India, even those at the most important pilgrimage sites on the Gangetic Plain.1 Slowly we are moving away from this direction; for example, the final volume of the Encyclopaedia of Indian Temple Architecture will examine North Indian temples constructed into the fourteenth century.2 Some scholars have looked at post-Muslim conquest temples in Bengal, and some are now interested in temples built during the Mughal period.3 With few excep- tions, these studies consider temples as individual works of art, never as a focus for ritual activity in a larger social context. The situation in some ways is not all that different for North Indian Islamic architecture. True, we know a good deal about individual monu- ments, but to date there has been little attempt to examine how they fit into the larger urban fabric or how they coexisted with contemporary Hindu or other non-Muslim sacred structures-temples, shrines, dhar- mashalas, or schools.4 In many ways this is understandable, since rela- tively little was known even about the basic buildings. Events at Ayod- hya culminating in the 1992 destruction of the so-called Babri masjid stimulated me, at least, to think increasingly about how Muslims in what were predominately Hindu cities or cities constructed by Hindu rajas, such as Jaipur, understood their built environment and how Hin- dus in seemingly Muslim cities, such as Shahjahanabad (Delhi), ex- pressed their own religious identity through structures. Much of what I will present here is the result of very recent work thatI have just begun.5  122 | Mapping Hindu-Muslim Identities through Architecture Even if I am not always yet ready to draw conclusions from the observa- tions I present here, I believe that probing the visible landscape can help us understand the complex issue of Hindu-Muslim identities in pre- modern and even early modern India. Little is known about temples in North India built after the Muslim conquest of Delhi; attention tends to be drawn to Islamic monuments, often on those parts constructed from Hindu temples, thus skewing our perceptions of Muslim relations with India's majority population. A rare statement of political victory-for example, the use of temple pillars in Aibek's Jami mosque in Delhi-is seen as a universal Muslim mode of building in India. Archaeological reports repeatedly present mosques that are claimed to consist of elements from destroyed (read wantonly) Hindu temples.6 Yet examining these monuments at the site-for ex- ample, the Jami mosque at Kannauj widely believed to be constructed from reused material-could go a long way in dispelling that view.7 In fact, it is generally assumed that no temples were constructed in Mus- lim-ruled domains, whereas Islamic structures were built and survive in great numbers. Ironically, to cite one example, we have about a hundred inscriptions telling about once existing pre-Mughal Islamic monuments in Bihar, but only one survives.8 By contrast we have about six extant fourteenth-century temples.9 Such tabulations may help us understand how a Muslim elite or religious devotees perceived themselves, but it does little to further our understanding of identity perception across or among communities. More useful is to look at structures in a larger con- text. For this purpose I will focus on cities whose original setting is more or less intact: Shahjahanabad, Jaipur, and Amber, among others. It might be useful to see first what can be gleaned about Hindu-Mus- lim identity in the built environment before the construction of planned cities such as Shahjahanabad and Jaipur. Since relatively few towns or extensive sites built before the seventeenth century survive intact, an examination of paintings provides insight. The distant backgrounds of some Mughal miniatures give us a sense of a shared Hindu-Muslim landscape. Domes and shikharas, possibly mosques and temples, are depicted, although in a rather fuzzy manner.10 But such depictions hardly can be taken to mean that these are religious structures or that they literally sat side by side, for the Mughal artist painted the known world, not the literally observed world. We may conclude, though, that both Hindu and Muslim monuments were part of the larger landscape. All the same, we have virtually no surviving temples at Mughal pal- aces or palace towns, even those as well preserved as Fatehpur Sikri, to  Catherine B. Asher I 123 suggest what might have been the situation. Rohtas fort, the seat of Raja Man Singh when governor of Bihar, is an exception. There temples and mosques are grouped in separate areas of the hill fort, although this may be as much a result of chronology as it is of community.11 I have argued elsewhere for a strong sense of Rajput identity in the Mughal governor's palace adjacent to the mosques and tombs there, but the palace has no sectarian overtones-either Hindu or Muslim.12 We know that under Akbar the construction of monumental Hindu temples proliferated. A case in point is Vrindaban, where Akbar himself used imperial funds to support the construction and maintenance of Raja Man Singh's Govinda Deva temple, among others.13 But the site is essentially a temple enclave and a pilgrimage center (tirtha); there are no Muslim monuments in town.14 Thus two other contemporary towns, Amber, Raja Man Singh's watan jagir in Rajasthan, and Rajmahal, his capital as Akbar's governor of Bengal,15 each intended for Kachhwaha residence and administration, might be more useful for understanding issues of Hindu-Muslim identity in the sixteenth and seventeenth centu- ries. Amber is famed for its palace built by this raja and further expanded in the seventeenth century by Mirza Jai Singh (1622-1667), but many other contemporary structures grace the site that largely go unnoticed by today's visitors. For example, on the main Delhi road lies a mosque originally constructed in accordance with Akbar's order in 1569-70. It is now rebuilt, but the locale and space occupied remain constant, as is shown on an early eighteenth-century map of Amber.16 In Rajmahal the Jami mosque, built in this case by Raja Man Singh, is in a similar location, that is, on the main road.17 The small temple also reputedly provided by Raja Man Singh is behind the mosque, not visible from any distance. How are we to read this placement? Does this mean that Islamic identity always sublimates a Hindu one even in cities built by a Hindu prince? Or does it mean that location of temples to Hindus in premodern North India had a very different meaning than it might today? Some answers may be gleaned by further examining Amber, the Kachhwaha seat of authority until the construction of Jaipur in 1737. Upon entering Amber's Delhi gate, one encounters a kos minar, an official Mughal distance marker, and the very prominent Jami mosque on the main road (fig. 5.1).18 Yet the town also has temples dedicated to a variety of deities including Jinas, the goddess, Shiva, and Vishnu. Some of these temples predate the Mughal period. They are all small structures (for example, the one dedicated to Ambikeshvara, the town's  124 1 Mapping Hindu-Muslim Identities through Architecture tutelary deity). The most important of these temples are located close to a large tank, or kund, and a small stream. They sit at the edge of a hill- the traditional seat of Rajput retreat and safety-not on the plain of the town. Others spill into the town but are away from the main Delhi road, where the Jami mosque is located. About 1600 Raja Man Singh added to Amber his magnificent Jagat Siromani temple, but like most other temples it is located in the town's interior close to the palace and an even older Kachhwaha haveli.19 It is not visible to the casual passerby or to those adhering to the main road only. In some ways this arrangement recalls those Jain temples at Abu, Ranak- pur, and even in nearby Sanganer designed to resemble mansions; from the outside, they do not look like temples.20 There is another model for the post-twelfth-century North Indian temple, one that probably seems more familiar, namely, temples that are prominently and centrally located in an urban setting.21 This type is not limited to the Rajasthani examples but is found across much of the Doab in the late eighteenth and nineteenth centuries. Pilgrimage maps of premodern India confirm that this type existed widely.22 If there is a significant distinction between these two basic temple types-one whose presence is overtly manifest and the other more obscured-it re- mains unclear. 5.1. Jami mosque, Amber.  Catherine B. Asher 1 125 5.2. Fakhr al-Masajid from street, Shahjahanabad, Delhi. In some cities, Varanasi or Lucknow, for example, mosques dominate the landscape.23 Even in Lucknow today, temples are not particularly visible, although many do exist.24 In Varanasi, of course, deemed by many the Hindu city par excellence, small temples literally dot the ghats and city, although most of them date no earlier than the late eighteenth century.25 It is particularly interesting that Rani Ahilya Bai Holkar's newly constructed Vishvanath temple, the focal tirtha in all Varanasi, is notably smaller than the adjacent mosque constructed during Aurang- zeb's reign from the spoils of an earlier Vishvanath temple.26 Yet the Rani was a woman of considerable resources, and the temple was built in 1777 when Hindu political power dominated in Varanasi.27 Had she wished to build a larger temple, rather than one almost lost in the inte- rior gullies of Dasashvamedh Ghat, she could have done so. To understand this pattern of dominant mosques and small temples, we may turn to Shahjahanabad, where nuance between Hindu and Muslim society is still evident today. Shajahanabad is considered by most scholars an Islamic city; in fact, by many it is believed to have once been the subcontinent's leading Islamic city.28 One highly visible feature is its mosques commencing with the enormous Jami Masjid, followed only by somewhat smaller ones such as the Fatehpuri mosque, the mosque of Zinat al-Nisa, or the no longer extant Akbarabadi masjid.29  126 | Mapping Hindu-Muslim Identities through Architecture There are then those that dot the main roads of every mahalla, for ex- ample, the Fakhr al-Masajid (fig. 5.2) or even the more humble Muh- tasib's mosques.30 Still today in the overcrowded and overbuilt walled city, these mosques are evident. For example, if one proceeds from Hauz Qazi to Khari Baoli, a distance of less than a mile, on the west one would see the following in this order: the mosque of Mubarak Begum, Sirki Walan's mosque, Sabz Mosque, the mosque of Tahawwar Khan, and a few other small ones.31 Then veering a little to the right, one of the city's major landmarks, Masjid Fatehpuri, comes into view. One also sees the entrances to havelis-not so much the havelis themselves-but it is the mosques that draw attention.32 This tour includes structures built from the seventeenth through nineteenth centuries provided by begums, queens, nawabs, and landholders; at any one historical moment the pic- ture would be a little different, but not much. Any major thoroughfare of the city would reveal similar structures. True, before the city was rather radically changed by the British after 1858, there was more space for fewer people and more gardens, and Chandni Chowk from Masjid Fatehpuri to the Red fort was a tree-lined street through whose center ran a canal.33 The best comparison would probably be contemporary Isfahan. Before the nineteenth century we have no figures for the breakdown of Delhi's population in terms of Muslims and non-Muslims.34 In 1845, however, it was about equal,35 suggesting that the Hindu population since the inception of Shahjahanabad in 1639 was always sizable. Even in Shah Jahan's time, highly desirable plots in the Chandni Chowk vicin- ity had been allocated to Hindu and Jain bankers and merchants.36 Wealthy Khattri Hindu merchants and Jains, including one branch of the Jagat Seth family,37 played a role in the city's economic well-being. So it is not surprising that between 1639 and 1850 Hindus and Jains built over a hundred temples that still survive; others must have been destroyed, for example, in the massive rebuilding of Faiz Bazaar.38 Yet for the most part, these temples are almost invisible to the casual visitor. The ques- tion is why? These temples, unlike Delhi's mosques and tombs, rarely bear dated inscriptions.39 Basing my fieldwork on the 1916 List of Muhammadan and Hindu Monuments: Shahjahanabad, I have started to study these temples, although a good deal more work still needs to be done.40 Even though I have yet to establish a chronology based on style, it is apparent to me that a substantial number of these temples were founded in the seven- teenth and eighteenth centuries and then rebuilt. I do not think the  Catherine B. Asher I 127 exact date of each construction or reconstruction is critical to my obser- vations. Rather, location and scale-factors that have remained con- stant-are important. It might be useful to consider these temples in light of their better- known contemporary Islamic counterparts, although doing so should not suggest that there were distinctive religious styles or that one was derivative of the other. The eighteenth- and nineteenth-century mosques built throughout the walled city are easily visible from a dis- tance. The largest take up considerable space, while the smaller ones are located on the second story above a main street intended for both ve- hicular and pedestrian traffic. The very smallest are in gullies and kunchas (a linear street similar to a gully). By contrast, as at Amber, temples in Shahjahanabad are not as immediately visible. They are al- most never located on a main road. Rather, they are found in the city's interior lanes. Often they are essentially openings in a wall, appearing little different from a shop. Some are so small that one never enters; others are entered and vary in size from a small room to perhaps the size of two rooms. In some mahallas, these temples are inside small private courtyards just off the narrow pedestrian lanes of the city (fig. 5.3). In these courtyard types, the temple area is public, but the dwellings within which they are located are private.42 None of these temples is surmounted by a high shikhara (superstruc- ture), which we so often associate with traditional temple construction.43 Many lack any superstructure, especially those similar to recessed shops. Others with domes are low and small. In fact, most of these temples are only apparent during the timings for darshan, that is, when the doors to the courtyards are open, a situation today that probably reflects original use. I found, for example, that as an outsider it was vir- tually impossible to locate these temples between noon, when darshan ceases, and 5 P.M., when it recommences. Rarely was there an indication that behind a pulled shutter or a locked arched door was a temple, not a house or shop. Moreover, these temples are never very large. Even in areas such as Katra Nil, which always housed a predominant Hindu community, as many as six or seven small obscure temples might be found on a single short lane.44 There are areas, according to the List, where mosques stand, but there are virtually no premodern temples, for example, in Daryaganj, Bazaar Chitliqabr, and Bazaar Churiwalan. However, rebuilding after 1857 may have distorted the accuracy of this picture, since, as we know, there were havelis of Hindus in some of these areas.45 All the same, the List indicates  128 1 Mapping Hindu-Muslim Identities through Architecture 5.3. Shivalaya of Dhumi Mal Khanna (List 1:359), Katra Nil, Shahjahanabad (Delhi). that there is no mahalla in which there are temples yet no mosques. An excellent case in point is at the intersection of Khari Baoli and Lal Kuan. A cluster of temples is there,46 and just to the south is the mosque of Tahawwar Khan, while to the east is the Fatehpuri mosque. The mosques are highly visible; the temples are apparent only if one enters the mahalla's interior. Even in Katra Nil, considered the richest quarter of Delhi in the early nineteenth century,47 there is a huge concentration of temples, but there is also a small mosque.48  Catherine B. Asher I 129 The only temples located prominently on a main street are the Digam- bara Jain temple, called today the Lal Mandir but known until recently as the Urdu or Camp temple, and Gauri Shankar temple next to it. The Gauri Shankar temple was built in 1761 during a period of Maratha su- premacy by Appa Gangadhara, a Maratha Brahman in the service of the Scindia family.49 The temple, like most of them in Delhi, has been considerably enhanced throughout the twentieth century. Its modern shikhara does not reflect its original appearance. The bulk of the Jain Lal Mandir was built between 1835 and the 1870s;50 portions are more re- cent, but reputedly its position on this site dates to Shah Jahan's time.51 Inside the original portion of this restored temple are three white marble Jina images, each dated the equivalent of 1492 C.E.,52 suggesting veracity to the claim that the site is of considerable antiquity. While much can be said for this prime location, the relevant question for our purposes is: How visible were these temples and others before the late nineteenth century?53 Were they the exceptions to the rule in Delhi, or were they as virtually invisible as Delhi's other Hindu and Jain religious edifices? Maps dating between 1751 and 1850 suggest that these temples, no matter their initial construction date, were not part of the visible landscape. The area north and east of the Jami mosque was then residential but leveled after 1857. A map datable to about 1850 indi- cates a dharmasala where the Jain and Gauri Shankar temples are lo- cated, apparently subsuming the two temples under the common term dharmasala.54 They were not sufficiently significant to be separately iden- tified. Illustrations by two British artists, each done in the early nine- teenth century, indicate no visible temple on the skyline, but the Fateh- puri mosque is present in each, as is evident in the one reproduced here (fig. 5.4).55 Even an illustration of Shahjahanabad made by a Kotah artist about 1842-43 does not show these temples, indicating that the domi- nant appearance they have today is not an original feature.56 How can we understand the prominence in Delhi's landscape of con- struction associated with Islam and the almost invisible presence of structures associated with non-Islamic traditions? Blake, for example, equates the lack of monumental temple building in early eighteenth- century Delhi with the "economic and political impotence of [Delhi's Jain and] Khattri merchants."57 But if this is the sole reason, then why do we not see much temple building on a main street until about 1900 when we know that they had become powerful in the late eighteenth century and remained so into the early twentieth century? Even around 1900, temples constructed along the British-created Esplanade were small and  130 1 Mapping Hindu-Muslim Identities through Architecture 5.4. Chandnee Chauk, Delhi. From John Luard, Views in India, St. Helena and Car Nicobar: Drawn from Nature and on Stone (London: J. Graf, 1838). Courtesy of the Ames Library, University of Minnesota. lacked the traditional signifiers indicating temple presence.58 For ex- ample, little differentiates them from any shop facade on the street; there are no shikharas or domes or even large-scale images of deities to draw attention to their sacred status. To understand the dominance of mosques and the surprisingly low visibility of Shahjahanabad's temples, it is instructive to look at temple and mosque construction in a city planned and ruled by Hindu mon- archs, that is, Jaipur in the eighteenth and nineteenth centuries. Com- menced by Sawai Jai Singh in 1727, the city today boasts more temples than any other but Varanasi.59 Most of these date to the eighteenth and nineteenth centuries, that is, exactly contemporary with temples in Delhi's walled city. Let me interject my own story here. I have visited Jaipur numerous times, but never noticed many temples in the city be- sides the eighteenth-century Govinda Deva temple, actually part of Sawai Jai Singh's City Palace complex, and the modern Birla one. So what, you might say, but I am an art historian, and it is second nature for me to notice monuments. Why did I not see them until I consciously sought them out armed with lists of temples in Jaipur's various bazaars? Even though Jaipur was built by a Hindu ruler and, many have ar-  Catherine B. Asher 1 131 gued, as a Hindu city,60 temples are no more visible within the confines of Jaipur's walls than they are in eighteenth- and nineteenth-century Shahjahanabad. Only one inside Jaipur's walls, the Kalkiji temple, built by Sawai Jai Singh in 1740, bears a shikhara.61 While the entire temple is easily visible from its platform on the second story above shops, it is not readily visible from the busy main street in the major Sireh Deori Bazaar. The other two temples with dominant shikharas visible from a distance probably were not built until the late nineteenth century.62 The rest are, like those in Delhi, within courtyards. Examples include the Ramachan- draji temple (1854) located above shops in Sireh Deorhi Bazaar and Shri Brijraj Behariji's temple (1813) in Tripolia Bazaar. Today signs in Hindi indicate the presence of some; others have none. For example, there is no indication that an almost unnoticeable entrance off the main Jalab Chowk leads to Shri Brijnandji's temple provided by Maharaja Sawai Pratap Singh in 1792 (fig. 5.5).63 Of course, we do not know how they were marked at the time of construction, but chances are there were few written indications to tell of the temples' existence.64 While the temples are larger than those in Shahjahanabad, they are no more visible from the outside. We can therefore conclude after examin- ing temples in Jaipur that the Delhi ones are obscure to those who do not know their location not because their builders and patrons sought to 5.5. Entrance to the Sri Brijnandji Temple, Jaipur.  132 | Mapping Hindu-Muslim Identities through Architecture hide them from Muslim rulers and Muslim neighbors but for some other reason. What about the situation of mosques in Jaipur?65 This is a little more difficult to analyze, since almost every mosque has been refurbished recently with a new facade or is currently undergoing remodeling. But if we focus on location and general patterns of appearance, we can at- tempt to understand mosque construction here. First, though, we need to recognize that Jaipur has a sizable Muslim population. Just as we have no statistics for Delhi's Muslim-Hindu breakdown before the nineteenth century, so the same holds true for Jaipur. Today the Muslim population of Jaipur is about a fifth of the total; this is about the same as in Lucknow, which is so often imagined to be a Muslim city.66 Muslim military personnel and craftsmen were employed by Sawai Jai Singh and his successors. For example, in 1788 Tirandaz, a predomi- nately Muslim mahalla of the city, was established within Ramganj Ba- zaar to house Jaipur's Muslim archers originating from Lahore.67 Today Muslims live mainly in the areas between Suraj Pol and Johari Bazaar, apparently reflecting patterns since at least the late eighteenth and nine- teenth centuries.68 In this area are a large number of mosques. But Mus- lims also live around Chand Pol gate. In fact, their businesses and mosques are found throughout much of the city. For example, there is a small mosque adjacent to Ajmeri gate, and the city's Jami mosque occu- pies prime land where the Bari Chaupar intersects with Johari Bazaar- just a few hundred meters away from the Kachhwaha palace, the seat of the Jaipur Maharaja.69 Other major mosques line the main street that intersects the city on an east-west axis with most of them nearer the Suraj Pol end of town. Smaller mosques are located on the main streets within the mahallas themselves. In short, just as mosques tend to be located on the main streets of Delhi, some close to the ruler's palace, so the situation seems parallel here, even though Jaipur is the seat of a Hindu prince. If the location of these mosques follows a pattern similar to the one in Delhi, what about their appearance? In Delhi, we argued, the prominent facades of mosques on the main street made them a highly visible fea- ture of the landscape. It is more difficult to discuss the appearance of mosques in Jaipur, especially those on the main roads, since they have all undergone remodeling. We need to examine those inside lanes and even in dargahs, where less damage to the original appearance has been done. For example, although Masjid Maulana Zia al-Din Sahib in Mahalla Hadipura (a subdivision of Ramganj) was being refurbished during my visit in February 1995, it had enough of the original construc-  Catherine B. Asher 1 133 - 5.6. Mosque of Maulana Zia al-Din Sahib, Mahalla Hadipura, Jaipur. tion visible to understand it and other contemporary mosques in the city (fig. 5.6). Similar to contemporary Mughal mosques in Delhi, it is sur- mounted by three domes. Each dome marks an interior bay. Arches are defined by faceted stucco work; exquisite arabesques are painted on fine chunam. I found similar painting elsewhere in Jaipur, for example, on the mosque of Dargah Zia al-Din Sahib (fig. 5.7), which the local inhabitants assured me is one of the oldest mosques in the city.70 All the mosques on Jaipur's main thoroughfares have new high fa- cades. Only at the mosque of Bilor Khan on Ghat Darwaza Rasta does  134 1 Mapping Hindu-Muslim Identities through Architecture 5.7. Interior, mosque at Dargah Zia al-Din Sahib, Jaipur. any semblance of the original appearance remain. The mosque was originally a three-domed structure, not dissimilar to those on the back streets. The point here is that not only in location but also in overall appearance do the mosques of Jaipur built in the eighteenth and nine- teenth centuries relate well to those of Delhi. What pertains elsewhere in North India? Are dominant mosques and almost invisible temples a consistent pattern? Although my research to date is preliminary, it indicates that the situation in Jaipur and Shahja- hanabad is not universal. Certainly in Bengal, where the most innova- tive and widespread tradition of temple building developed after the sixteenth century, temples continue to be a highly visible part of the land- scape vying in competition for visibility with contemporary mosques.71 The Bengal temples, however, generally are not part of the urban fabric, more often in the domain of a zamindar than in Bengal's premier cities. So too in tirtha sites such as Ayodhya or Varanasi, temples are readily seen as are a number of mosques.72 One issue here, though, is that with some exceptions, for example, Chait Singh's temple in Ramnagar or the famous Vishvanath temple in Varanasi, we do not know enough about these structures to differentiate between those of the eighteenth and nineteenth centuries.  Catherine B. Asher I 135 We know that some temples with shikharas were built in the late eighteenth century, for example, the one at Dev ki Nandan ki Haveli in Varanasi.73 Since the temple is enclosed within high walls, only its shik- hara is visible. Others are more elusive: for example, the Jagannath temple reputedly provided by Nawab Asaf al-Daula in a village known as Serai Shekh, about twenty kilometers from Lucknow.74 The temple is enclosed in a high walled courtyard. A small dome, virtually indistin- guishable from those used on mosques and tombs, marks its presence. In short, this eighteenth-century structure is clearly a religious building, but nothing from a distance further defines its sectarian affiliation. So too the domed temple at Chakiya, about forty-five kilometers from Var- anasi, resembles a Muslim tomb.75 Similarly, the famous Kalkaji temple in south Delhi originally had no features that would indicate its reli- gious affiliation, even though it was always a prominent structure.76 When the temple was first constructed in 1764, it was a flat-roofed twelve-sided structure; the shikhara, now prominent, was not added until 1816, when Mirza Raja Kedar Nath, peshkhar of Akbar II, provided it.77 Shahjahanabad's neighborhood mosques, regardless of mahalla, al- most invariably follow a single pattern. That is, they are single-aisled multi-bayed structures surmounted with domes. Usually they are above shops on a main street, thus enhancing their visibility (fig. 5.2).78 The temples show a greater variety of types. For example, in Katra Nil most of the temples are small domed Shivalayas situated within an open courtyard (fig. 5.3), while in nearby Balli Maran, Hauz Qazi, and Sita Ram Bazaar, Shivalayas are incorporated into walls almost as if they were shops. I am not sure what conclusions I can draw from this, but the variety of temple types seen from mahalla to mahalla has nothing to do with sect. In the eighteenth and nineteenth centuries, most temples in Delhi were dedicated to Shiva.79 Many of these are domed. The few dedi- cated to Vishnu, more specifically to Radha and Krishna, are what I call haveli types.80 That is, they are flat-roofed temples that are located within high enclosure walls and have an open courtyard in their center recalling a traditional house. One creative example, the Temple of Charan Das, founded by an eighteenth-century Delhi reformer, is essen- tially a haveli type, but in the open courtyard a domed structure that resembles a typical Shivalaya of the Katra Nil variety serves the samadhi (memorial) of Charan Das and enshrines impressions of his footprints, as if they were Vishnu pada.81 In contrast to the variety of temple types, the high degree of unifor-  136 | Mapping Hindu-Muslim Identities through Architecture mity found in mosques-whether they are in Shahjahanabad, Jaipur, or elsewhere-may reflect a conservative adherence to a pattern of impe- rial patronage established earlier in the Mughal period. For Hindu temples in Delhi, at least, there had been no recent imperial patronage and so no uniform models to emulate. Thus buildings constructed by prominent individuals in a single mahalla may have served as models for subsequent work, resulting in narrowly localized styles. The re- served appearance and diminished scale of Old Delhi's temples may reflect patterns of use that in turn reflect the sociological and economic makeup of the mahalla. For example, Katra Nil, at least by the early nineteenth century the wealthiest area of the city,82 had the largest num- ber of temples as well as those that occupy the greatest amount of physi- cal space. Nevertheless, even these temples remain essentially hidden. Although a number of temple types in Delhi and elsewhere were built throughout the eighteenth and nineteenth centuries, the two most common types in the walled city are the domed circular or polygonal ones (fig. 5.3) and the haveli type (fig. 5.8). How the North Indian temple was transformed from large structures with porches and shikharas that we might perceive as textbook examples to the essentially interiorized haveli or domed Shivalaya type is not clear. The small domed Shivalaya seems to have had a long albeit obscure history. For example, a small domed temple enshrining a lustrated Shiva linga is depicted in the 1591 Chunar Ragamala, a manuscript commissioned by the Bundi raja.83 Since our knowledge of seventeenth-century temples is restricted to a few large-scale ones, the development of the domed Shivalaya is un- clear.84 However, this domed Shivalaya type relates closely to two sorts of memorial structures: the Muslim tomb and the Hindu memorial chattri. I am not trying to suggest that Shivalayas are transformed me- morials, Hindu or Muslim, but rather that the domed chattri-like struc- ture simply was associated with the visual vocabulary of religion in gen- eral. The haveli temple style in Delhi is particularly interesting for two reasons. One is the origins of its appearance, and the other is its sectarian affiliations. The haveli type probably derives from an imperial audience hall inside which the ruler sits in his throne known as a jharoka-i darshan. In a temple, the deity is similarly situated for darshan.85 A good example of such a figure is the Govinda Deva image installed in the Jaipur temple in 1734; it is probably the first temple of the haveli sort.86 In Shahja- hanabad a haveli-style temple is the one of Charan Das in Hauz Qazi, originally founded in the eighteenth century.87 A good deal of work  Catherine B. Asher 1 137 h .. 5.8. Interior courtyard of the haveli-style Ladliji temple in Katra Nil, Shahja- hanabad, Delhi. needs to be done on the development of the haveli temple type, but tentatively I am inclined to think that it is closely associated with Jai Singh's concept of regnal authority as validated by the divine.88 To ob- tain darshan, be it divine or imperial, it is necessary to approach a structure's threshold.89 In the case of a ruler, Hindu or Muslim, the jharoka must be approached for darshan; in the case of the divine, the garbhagriha (inner sanctum) must be approached. The haveli temple, at least in Delhi and Jaipur, is associated with Vaishnavism, the sect favored by wealthy Hindu bankers and mer- chants. What is puzzling is why we have so few Vaishnava temples in the walled city. It is true that they are considerably larger than most Shivalayas; it is also true that the two main ones-the Ladliji temple in Katra Nil (fig. 5.8), claimed to have existed since the seventeenth century,90 and the eighteenth-century temple of Charan Das in Balli- maran91-are in areas long associated with concentrated merchant and banker wealth. Nearby are six Jain temples, each claiming some antiq- uity.92 Since many Shivalayas in the city also feature images of Vishnu, his consort, Hanuman, and other deities and since each home would have an interior shrine, very likely dedicated to Krishna, perhaps the seemingly small number is deceiving to modern eyes.93  138 | Mapping Hindu-Muslim Identities through Architecture Structures originally associated with Mughal authority (the Diwan-i Aam cum jharoka) and religious commemoration (the domed chattri) become appropriate forms for temples. The surface decoration of temples, too, is similar to contemporary Muslim architecture: for ex- ample, there is no difference between the arches and baluster columns on the Charan Das temple and those on the mosque of Tahawwar Khan.94 They belong to a common visual vocabulary; for example, these architectural features also resemble the ones from an eighteenth-century raja's palace in Dig (Rajasthan).95 It is clear that identity is not sought through individual architectural members or stylistic components, but rather in the manner that they are combined and, perhaps more signifi- cantly, displayed to the faithful.96 We now may return to why the temples of Shahjahanabad and Jaipur are essentially invisible to the uninformed. Blake's suggestion that the lack of visible temple construction reflects the economic and political impotence of the merchant class might be true for Delhi in the early eighteenth century, but it does not explain why temples remain obscure in the late nineteenth century, when Delhi's Jain and Khattri bankers are powerful.97 How, in turn, do we explain the situation in Jaipur, where many of the temples are actually the products of imperial patronage? How do we explain in Varanasi the fact that the city's major temple is small and tucked inside a gully? Instead of a single answer, I would like to suggest that several forces are at play here. We might recall that the dominant sixteenth-century mosque at Am- ber was on the main road, while the near-contemporary temples were in the interior. This is similar to the situation both in Jaipur and Delhi later in the eighteenth and nineteenth centuries. Thus I would suggest that the arrangement, location, and prominence of religious structures are factors of community identity-though not communal identity. In fact, if we think back to North Indian temple architecture before the Muslim conquest of Delhi, we must realize that large-scale temples provided imperially such as those at Khajuraho are the exception, not the rule. Temples in North India tend, in fact, to be small and often in areas that today seem distant from well-established centers of population. They, in fact, are often in areas that are intentionally remote, that is, the tirtha sites.98 Their patronage is unknown; they do not adhere to the more com- mon South Indian patterns of enormous structures sponsored by the king. Today's large prominent temples in North India-for example, the recently constructed Birla one in Jaipur-may reflect modern concepts of Hindu identity, not premodern ones.  Catherine B. Asher I 139 Delhi's obscured temples may have yet an additional explanation. Bayly points out the conflict many merchants, especially Khattris and Jains, experienced. Hindu social code demanded a frugal lifestyle, yet a business code with Islamic rulers demanded a display of opulence and wealth.99 Dual lifestyles developed. For example, the exterior of a haveli might be splendid, but the interior rooms were austere. Jewelry suitable for Mughal court attendance was worn in public, while at home more traditional jewelry and garb were donned. Often money was spent on religious structures, for example, the Jain Digambara Lal Mandir and others within the walled city in Delhi. Their carved and inlaid marble surfaces are covered lavishly with gold gilt.100 Concurrently, reform movements instigated by men such as Charan Das urged a return to a simpler and more austere form of religion.101 The tendency to look in- ward, whether for spiritual purposes or to maintain one's identity in the community, is clear. So I think we can argue that the very form, appear- ance, and location of Delhi's religious Hindu architecture reflects these same values. Visible mosques and less visible temples, Hindu or Jain, need not be seen as a matter of Islam subsuming Hindu or other non-Muslim iden- tities. Rather, these patterns of mosque and temple construction in Delhi, Jaipur, and other places in North India follow long-standing prac- tice. Since the construction of its earliest extant monument, the Dome of the Rock, Islam has manifested itself boldly in urban settings. This is not true for North Indian temples, which tend to be situated in those settings deemed ideal by Sanskrit texts-among trees, hills, and bodies of wa- ter.102 In the urban context this would be translated as in a courtyard garden (figs. 5.3 and 5.8). The surrounding walls of the haveli temple or the Shivalaya are the hills, and the potted flowers are the greenery that can only thrive on water. This is the setting, so the Brhatsamhita tells, where the gods dwell and frolic. Notes 1. C. J. Fuller, "The Hindu Temple and Indian Society," in Michael Fox, ed., Temple in Society (Winona Lake: Eisenbrauns, 1988), 58. 2. Volumes of the Encyclopaedia of Indian Temple Architecture have been pro- duced since 1986 by the American Institute for Indian Studies' Center for Art and Archaeology under the editorship of M. A. Dhaky and Michael W. Meister; later volumes are being produced by M. A. Dhaky. 3. For examples, see George Michell, ed., The Brick Temples of Bengal (Princeton: Princeton University Press, 1983); Catherine B. Asher, "The Architec-  140 | Mapping Hindu-Muslim Identities through Architecture ture of Raja Man Singh: A Study of Sub-Imperial Patronage," in Barbara Stoler Miller, ed., The Powers of Art: Patronage in Indian Culture (New Delhi: Oxford University Press, 1992), 183-201; Asher, "Authority, Victory, and Commemora- tion: The Temples of Raja Man Singh," Journal of Vaisnava Studies 3 (summer 1995): 25-36; and Asher, "Kachhvaha Pride and Prestige: The Temple Patronage of Raja Man Singh," in Margaret Case, ed., Govindadeva: A Dialogue in Stone (New Delhi: Indira Gandhi National Centre for the Arts, 1996), 215-38. 4. One notable exception is the work of Atillio Petruccioli, "Geometry of Power: The City's Planning," in Michael Brand and Glenn D. Lowry, eds., Fatehpur-Sikri (Bombay: Marg, 1987), 49-64. Also see his edited issue of Environ- mental Design on "Mughal Architecture: Pomp and Ceremonies," 1-2 (1991). 5. For an even more exploratory article on this work, see my forthcoming article, "Piety, Religion, and the Old Social Order in the Architecture of the Later Mughals and Their Contemporaries," in Richard B. Barnett, ed., New Perspectives on Early Modern India (Columbia, Mo.: South Asia Books, forthcoming). 6. A good example is the Arhai Din ka Jhompra Masjid in Ajmer, which is commonly believed to be composed of spolia; however, Michael Meister, "The 'Two and a Half Day' Mosque," Oriental Art 18, no. 1 (1972): 57-63, has argued convincingly that the bulk of its columns and ceilings are newly carved. Many other monuments need to be reevaluated in light of Meister's argument. 7. This view was published by 1891 by A. Fuhrer, The Monumental Antiquities and Inscriptions in the Northwestern Provinces and Oudh, vol. 12a of the New Impe- rial Series, Archaeological Survey of India (reprint ed., Varanasi: Indological Book House, 1969), 21, and repeated by James Fergusson, History of Indian and Eastern Architecture, vol. 2 (London: John Murray, 1910), 68. These views may seem antiquated, but nothing has been published to reverse them. A case in point is that two leading scholars who finished a draft of an otherwise superb book on Islamic art and architecture had intended to include this mosque and point of view in their survey; luckily, they showed me the draft, and the mosque under discussion was dropped from the text. 8. For the inscriptions, see Qeyamuddin Ahmad, Corpus of Arabic and Persian Inscriptions of Bihar (Patna: K. P. Jayaswal Research Institute, 1983), 1-118. Catherine B. Asher, Islamic Monuments of Eastern India and Bangladesh (Leiden: Inter Documentation Co. on behalf of the American Committee for South Asian Art, 1991), provides illustrations. 9. See D. R. Patil, Antiquarian Remains in Bihar (Patna: K. P. Jayaswal Research Institute, 1963), 134, 139, 141, 213-15, 580-81; and Frederick M. Asher, "Gaya: Monuments of the Pilgrimage Town," in Janice Leoshko, ed., Bodhgaya: The Site of Enlightenment (Bombay: Marg, 1988), 77. 10. An example is the background of a depiction of Saint Jerome in Amina Okada and Francis Richard et al., A La Cour du Grand Moghol (Paris: Bibliotheque Nationale, 1986), 118. 11. For example, the Rohtas temple is the earliest one and situated so it can be  Catherine B. Asher I 141 seen from below. Man Singh built the Harischandra temple next to it for political reasons; on this part of the hill there was inadequate room for a palace. The mosques and tombs are situated near the palace, and even though they predate the palace, this same area likely always was used for habitation. For basic cover- age of Rohtas fort, see Muhammad Hamid Kuraishi, List of Ancient Monuments Protected under Act VII of 1904 in the Province of Bihar and Orissa, New Imperial Series, vol. 51, Archaeological Survey of India (Calcutta: Government of India, Central Publication Branch, 1931), 146-83. Also see Catherine B. Asher, Architec- ture of Mughal India (Cambridge: Cambridge University Press, 1992), 67-74, 90- 92, and "The Architecture of Raja Man Singh," 187-91. 12. Catherine B. Asher, "Sub-Imperial Palaces: Power and Authority in Mughal India," Ars Orientalis 23 (1993): 287. 13. Tarapada Mukherjee and Irfan Habib, "Land Rights in the Reign of Akbar: The Evidence of Sale-Deeds of Vrindaban and Aritha," Proceedings of the Indian History Congress, 50th Session (Delhi: Indian History Congress, 1989-90), 236-55, and the same authors, "Akbar and the Temples of Mathura and Its Environs," Proceedings of the Indian History Congress, 48th Session (Panajim: Indian History Congress, 1988), 234-50. 14. Mons. Victor Jacquemont's comment, "I could not discover in [Vrinda- ban] a single mosque," in F. S. Growse, Mathura: A District Memoir (reprint, Ahmedabad: New Order, 1978), 188. This is an interesting contrast to another tirtha, Pushkar, where there is a mosque near the water's edge. 15. While today Rajmahal is in Bihar, in the Mughal period it was in suba Bengal. 16. For the mosque's foundation inscription, see S. A. Rahim, "Nine Inscrip- tions of Akbar from Rajasthan," Epigraphia Indica, Arabic and Persian Supplement, 1969,55-56. See Susan Gole, Indian Maps and Plans (New Delhi: Manohar, 1989), 170, for a portion of a map of Amber in the National Museum, New Delhi [56.92.4] showing the mosque. 17. Kuraishi, List of Ancient Monuments Protected under Act VII of 1904, 217-19, for the mosque; the temple is in Asher, "The Architecture of Raja Man Singh," 192. 18. There is little scholarly discussion of Amber. B. L. Dhama, A Guide to Jaipur and Amber (Bombay, 1948), 68-82, is a good source for the town. The best map is provided by George Michell, Buddhist, Hindu, Jain, vol. 1 of The Monuments of India (New York: Viking Press, 1989), 290. 19. B. L. Dhama, A Memoir on the Temple of the Jagatshiromani at Amber (Jaipur: Chiranji Lal Sharma, 1977), provides a good description supported by drawings and photographs as well as a map indicating its location. For analysis of the temple's meaning, see Asher, "Kachhvaha Pride and Prestige," 221-26 and plates 10.3-10.24. 20. For example, see Susan L. Huntington, The Art of Ancient India (New York: Weatherhill, 1985), 490.  142 | Mapping Hindu-Muslim Identities through Architecture 21. See the Jagdish temple in Udaipur in H. B. Pal, The Temples of Rajasthan (Alwar and Jaipur: Prakash, 1969), plate 22. 22. Gole, Indian Maps and Plans, 172-73, shows pilgrimage towns with such temples depicted on the walls of the Bhojan Shala of Amber palace. 23. For the mosques of Lucknow, see B. N. Tandan, "The Architecture of the Nawabs of Avadh, 1722-1856," in Robert Skelton et al., eds., Facets of Indian Art (London: Victoria and Albert Museum, 1986), 66-75. For those in Varanasi, see Pierre-Daniel Coute and Jean-Michael Leger, Benares (Paris: Editions Creaphis, 1989), 71-81. Perhaps the most famous drawing of the so-called Alamgir mosque dominating the riverscape is in James Prinsep, Benares Illustrated (Calcutta: At the Baptist Mission Press, 1833). I still need to examine the prominence of mosques in cities such as Udaipur or Kotah. 24. Some temples are cited in Yogesh Praveen, Lucknow Monuments (Luck- now: Pnar, 1989), 1-23. 25. Diana Eck, Banaras, City of Light (Princeton: Princeton University Press, 1982), 90. 26. For a view showing this juxtaposition, see Rajesh Bedi and John Keay, Banaras, City of Shiva (New Delhi: Brijbasi Printers, 1987), 37. 27. Coute and Leger, Benares, 54; Eck, Banaras, City of Light, 120, 248, where the queen is credited with building a temple at the Manikarnika Cremation ghat in Benares. F. M. Asher, "Gaya," 74, indicates she also built the Vishnupad temple in Gaya. 28. For example, see Eckart Ehlers and Thomas Krafft, "Islamic Cities in In- dia? Theoretical Concepts and the Case of Shahjahanabad /Old Delhi," in Ehlers and Krafft, eds., Shahjahanabad/Old Delhi: Tradition and Colonial Change (Stuttgart: Franz Steiner, 1993), 9-26, and Jamal Malik, "Islamic Institutions and Infrastruc- ture in Shahjahanabad," 43-64 of the same volume. Also see Stephen P. Blake, Shahjahanabad: The Sovereign City in Mughal India, 1639-1739 (Cambridge: Cam- bridge University Press, 1991), for the city as essentially an Islamic one. 29. Blake, Shahjahanabad, 51-55, and Asher, Architecture of Mughal India, chaps. 5-7 for plates. These mosques were built by queens or high-ranking women. 30. Asher, Architecture of Mughal India, 298-99, for plates and text on these two mosques. 31. See List of Muhammadan and Hindu Monuments: Delhi Province, 4 vols. (Calcutta: Superintendent, Government Printing, 1916-22), 1:168-225, for these and other structures in the vicinity. Hereafter this work will be known as List; the numbers correspond to the catalog number, not the page. See Asher, Architecture of Mughal India, 198, 256, for plates of the mosque of Tahawwar Khan and the mosque of Mubarak Begum (Lal Kunwar). 32. Illustrations to the Zeenat Mahal haveli are in Pavan K. Varma and Sondeep Shankar, Mansions at Dusk: The Havelis of Old Delhi (New Delhi: Spantech, 1992), 78-80. For a good illustration showing the Fatehpuri mosque in its original state, see Emily Bayley and Thomas Metcalfe, The Golden Calm: An  Catherine B. Asher I 143 English Lady's Life in Moghul Delhi, ed. by M. M. Kaye (New York: Viking Press, 1980), plate facing p. 168. 33. Blake, Shahjahanabad, 55-56, portrays Shahjahanabad through the early eighteenth century. Narayani Gupta, "Military Security and Urban Develop- ment: A Case Study of Delhi 1857-1912," Modern Asian Studies 5, no. 1(1971): 61- 77, is a good source for changes made after the Uprising of 1857. 34. Blake, Shahjahanabad, 174. 35. Naryani Gupta, Delhi between Two Empires, 1803-1931 (Delhi: Oxford Uni- versity Press, 1991), 46. 36. Ibid., 49; Hamida Khatoon Naqvi, "Shahjahanabad: The Mughal Delhi, 1638-1803: An Introduction," in Delhi through the Ages, ed. R. E. Frykenberg (Delhi: Oxford University Press, 1986), 144. Also see Susan Gole, "Three Maps of Shahjahanabad," South Asian Studies 4 (1988): 14-17, where she discusses a map in the IOL (AL 1762) of Chandni Chowk which gives the names of many Hindus' havelis and quarters. 37. Naqvi, "Shahjahanabad," 144. 38. List, 1:33-39; these monuments are in the area formerly called Faiz Bazaar. Gole, "Three Maps," 18-23, cites Hindus' havelis on the IOL map of Faiz Bazaar (AL 1763). Since we know, for example, that the massive Akbarabadi mosque was destroyed after 1857, small temples doubtless were as well. 39. This is only true for premodern additions. Now Hindi inscriptions abound on very recent restorations or constructions of temples in Shahja- hanabad. 40. This work was commenced in the summer of 1994 and has been ongoing over subsequent summers. 41. Volume 1 of the List cites a number of temples believed to date to the seventeenth or eighteenth centuries. Given the location and antiquity of these structures, I see no reason to doubt this. Blake, Shahjahanabad, 111, assumes there are only two temples built before 1707, since only these bear dated inscriptions; however, in general very few temples in North India bear dates; thus this argu- ment lacks substance. 42. Conversation with Sri Dalip Sharma, the son of the pujari of a Shivalaya in Katra Nil in February 1995. My treatment upon visiting all these temples would also confirm this-I was always most welcome. 43. In the entire walled city, I saw only one temple with a shikhara that ap- pears to date before 1900; it was in the area of Sitaram Bazaar. Today it is en- closed by walls and appears not to function as a temple. It cannot be examined. 44. Thomas Krafft, "Contemporary Old Delhi: Transformation of an Histori- cal Place," in Eckart Ehlers and Thomas Krafft, eds., Shahjahanabad/Old Delhi: Tradition and Colonial Change (Stuttgart: Franz Steiner, 1993), 81; List, 1:351-61. 45. Gole, "Three Maps," 18-23. 46. This is Mahalla Nayabans. See List, 1:219-21. Temple no. 220 reputedly dates to Shah Jahan's time, although it has been rebuilt in the late nineteenth or  144 | Mapping Hindu-Muslim Identities through Architecture early twentieth century. Consult the map provided at the volume's back for the location of these temples and the following mosques. 47. Gupta, Delhi between Two Empires, 23. 48. List, 1:355. 49. Ibid., 1:334. 50. Gupta, Delhi between Two Empires, 50, and Chakrash Kumar Jain Bijli Wale, "Sri Digambara Jain Lal Mandir, Delhi," in Grismahkalin Naitik Siksan Sibir (New Delhi: Digambara Jain Naitik Siksa Samiti, 1995). This three-page contribution to this summer school's souvenir program sums up the oral tradition I had been told in August 1995 by the temple's authorities. 51. List, 1:333, and Marie-Claude Mahais, Dilivrance et Convivialit: Le System Culinaire des Jaina (Paris: Editions de la Maison des Sciences de l'Homme, 1985), 20-22. Mahais, who worked in the temple for some time, is not aware that it was rebuilt in the nineteenth century. This is understandable, since her study has nothing to do with architecture. This information is verified in Balabhadra Jaina, Bharata ke Digambara Jaina Thirta, 5 vols. (Bombay: Bharatavarshiya Digambara Jaina Tirthakshetra, 1974-88), 1:288, and by the temple authorities. For a brief description of the temple's history preserved by its authorities, see Bijli Wale. 52. Both the List, 1:333 and Bijli Wale state these images are dated samvat 1548. The images are kept in a locked glass case with numerous smaller metal images on tiered steps before them; unfortunately, these metal images obscure portions of these donative inscriptions, but I could read enough of the numerical section of date on two of the images to assume that 1548 is correct. 53. This prime location reflects a relationship between the Jain community and the Mughal court, for in this same location lived the jewelers many of whom were Jain. While John F. Richards, "Mughal State Finance and the Premodern World Economy," Comparative Studies in Society and History 23 (1981): 285-301, argues that Khattri and Jain merchants were not vital to the Mughal court before the eighteenth century, the location of land allotted to them suggests they played a role of some substance in the maintenance of Mughal culture. I would like to thank John Cort for discussing the role of the Jain community in premodern North India with me. 54. See Gole, "Three Maps," 13-27, and Eckart Ehlers and Thomas Krafft, Shahjahanabad/ Old Delhi, facsimile map of Shahjahanabad (IOL Records X, 1659). 55. H. H. Wilson, The Oriental Portfolio: Picturesque Illustrations of the Scenery and Architecture of India (London: Smith, Elder, 1841), plate: The Chouk, Delhi; John Luard, Views in India, Saint Helena and Car Nicobar: Drawn from Nature and on Stone (London: J. Graf, 1838), plate: Chandnee Chauk, Delhi. Luard's illustration also includes the Sunhari mosque. 56. Stuart Cary Welch, India: Art and Culture, 1300-1900 (New York: Metro- politan Museum of Art, 1985), 429-33. 57. Blake, Shahjahanabad, 111. The words in brackets are my additions, for Blake ignores the religious activity of the Jain community, also merchants who lived around Chandni Chowk.  Catherine B. Asher I 145 58. There are no published references to or illustrations of these temples, in- cluding the Jagannath temple, dated 1864. 59. Ashim Kumar Roy, The History of the Jaipur City (New Delhi: Manohar, 1978), 29. 60. Joan L. Erdman, Patrons and Performers in Rajasthan (Delhi: Chanakya, 1985), 28, is the only scholar to give a balanced reading indicating Jaipur's debt to Mughal tradition. 61. While this appears true for the temples in the city itself, those on the surrounding hills do bear shikharas including the Ganeshgarh temple with which the Govinda Deva temple and City Palace are aligned. The Surya temple, provided by Jai Singh's minister in 1734, also bears a shikhara. The significance of this is not clear, although it may be that these are not Vishnu/Krishna temples; in Jaipur, Krishna temples are inevitably constructed as haveli type temples. For distant views of the Ganeshgarh temple and others, see Aman Nath, Jaipur: The Last Destination (Bombay: India House Pvt. Ltd., 1993), 4-5, 72, 73, 147, 198. See Sten Nilsson, "Jaipur, in the Sign of Leo," no. 1 of Magasin Tessin (Lund: Wallin and Dalholm, 1987), plate 32 for a plate of the Surya temple. There is no pub- lished photograph of the Kalkiji temple proper, but Nath, Jaipur: The Last Desti- nation, 162, illustrates its small fore shrine, and G. H. R. Tillotson, The Rajput Palaces: The Development of an Architectural Style, 1450-1750 (New Haven: Yale University Press, 1987), 168, shows a view of the bazaar taken from the Hawa Mahal, that is, a structure used by the imperial women; the shikhara visible is that of the Kalkiji temple, but access to this view would have been highly lim- ited. I spent considerable time checking its visibility from the street; its shikhara and only this portion of the temple is visible when standing directly in front of it; it is simply invisible from any other location on the street. 62. These include Ramchandraji's temple in Chandrapole Bazaar, which was built in 1894, and the Lakshmi Narayana temple in Bari Chaupar. This temple may possibly be earlier; since it bears no date, it needs to be analyzed in terms of its style. However, so far there is no systematic understanding of the eighteenth- century North Indian temple and its stylistic development. 63. These temples are listed by Roy, The History of the Jaipur City, 229-31, but I know of no published illustrations of them. The exception is a detail of the 1854 Ramchandraji temple's exterior in Nath, Jaipur: The Last Destination, 68. 64. Roy, The History of the Jaipur City, 125, indicates the general population was illiterate or barely literate until the first half of the twentieth century, suggesting written signs would have little value. 65. This portion of the paper is based solely on fieldwork I did in 1995 and 1996. To date, I have found virtually no secondary information on any of Jaipur's mosques. 66. Savritri Gupta, Jaipur, vol. 26 of Rajasthan District Gazetteers (Jaipur: Direc- torate, District Gazetteers, Government of Rajasthan, 1987), 64-65, although Roy, The History of the Jaipur City, 104, indicates that in 1901 a quarter of Jaipur's population was Muslim. For the Muslim population of Lucknow, see V. C.  146 | Mapping Hindu-Muslim Identities through Architecture Sharma, Lucknow, vol. 37 of Uttar Pradesh District Gazetteers (Lucknow: Govern- ment of Uttar Pradesh, 1959), 73. 67. Nath, Jaipur: The Last Destination, 163. 68. Information given to me by local informants. Also see Roy, The History of the Jaipur City, 48. 69. Since I have found no written material on the mosque, I asked every Muslim I met in Jaipur which mosques were the oldest. The unanimous re- sponse was the mosque at Amber and then the Jami mosque in Jaipur. It has been completely rebuilt, although its location remains unchanged. Gopal Narayan Bahura and Chandramani Singh, Catalogue of Historical Documents in Kapad- Dwara Jaipur, vol. 2: Maps and Plans (Jaipur: Maharaja of Jaipur, 1990) published a number of maps of Jaipur, but no mosque or dargah was included on them. However, this does not mean that there was no mosque in this position. There is no surviving map of Johari Bazaar, where the Jami mosque is located. Moreover, few Hindu temples are included on any of these Jaipur maps (see nos. 205 and 208 for rare exceptions). No Jain temple, for example is cited on the extant pub- lished maps, yet K. C. Kasliwal, Jaina Grantha Bhandars in Rajasthan (Jaipur: Shri Digamber Jain Atishaya Kshetra Shri Mahavirji, 1967), 43-59, indicates a num- ber of Jain temples were established in the eighteenth century within the city's walls; many of them are still extant in Johari Bazaar. 70. Annual Report for Indian Epigraphy, 1975-76, D. 155 gives the date on en- trance to dargah as A.H. 1213/1798-99 C.E. This probably is the same date as the mosque; it is not the oldest dated structure, although it may be the oldest intact one, for D. 154 of the same volume indicates an inscription dated A.H. 1190/1776 C.E. on Jaipur's Rahamiyya mosque. This mosque, situated not far from the city's Jami mosque, has recently been completely remodeled. 71. For examples, see Michell, The Brick Temples of Bengal, plates 165, 169, 185, 605-6. 72. See Coute and Leger, Benares, 51, 78, for examples. 73. Ibid., 52. 74. Praveen, Lucknow Monuments, 17-19, provides a description but no illus- trations. 75. Coute and Leger, Benares, 55. 76. List, 4:13; Sayyid Ahmad Khan, Asar al-Sanadid (reprint ed.; Delhi: Central Book Depot, 1965), 337-40; Carr Stephen, The Archaeology and Monumental Re- mains of Delhi (Ludhiana: Mission Press, 1876), 28-29. 77. Stephen, Archaeology and Monumental Remains, 29, indicates that he pro- vided the shikhara without much enthusiasm. 78. In addition to examples provided in note 31, see Asher, Architecture of Mughal India, 275, 297, 302-5. Imperially sponsored mosques in Delhi are larger and necessarily adhere to a slightly different imperial plan. See Ebba Koch, Mughal Architecture: An Outline of Its History and Development, 1526-1858 (Munich: Prestel, 1991), 119, for an example.  Catherine B. Asher I 147 79. See List, vol. 1, for the temples that are almost all primarily dedicated to Shiva. See note 93 of this paper for reference to other deities enshrined with these Shiva lingas. 80. The two major ones are the temple of Charan Das (List, 1:243) in Hauz Qazi (although the List includes it under Balli Maran) and the Ladliji temple in Katra Nil (List, 1:357). 81. See Christopher A. Bayly, Rulers, Townsmen, and Bazaars, rev. ed. (Delhi: Oxford University Press, 1992), 388; H. H. Wilson, Hindu Religions (reprint, Calcutta: Society for the Resuscitation of Indian Literature, 1899), 117-19; A. W. Entwistle, Braj, Centre of Krishna Pilgrimage (Groningen: Egbert Forsten, 1987), 213. 82. Gupta, Delhi between Two Empires, 23. 83. The manuscript is discussed in Milo Cleveland Beach, Mughal and Rajput Painting (Cambridge: Cambridge University Press, 1992), 41, 44-47. However, this particular page is illustrated in Douglas Barrett and Basil Gray, Painting of India (Lausanne: Skira, 1963), 143, although the date is incorrectly given. A simi- lar and probably earlier Shivalaya is depicted in an undated illustration of Bhairava Raga in Beach, 47. 84. In Vrindaban are several large temples reputedly erected during the sev- enteenth century, for example, the Jugal Kishore in Asher, Architecture of Mughal India, 164, 166. Pal, Temples of Rajasthan, plate 22, provides a view of the mid- seventeenth-century Jagdish Temple in Udaipur. 85. Asher, Architecture of Mughal India, 194-96. 86. For a distant view of the temple and its interior image, see Nath, Jaipur: The Last Destination, 146-47. Erdman, Patrons and Performers in Rajasthan, 37, re- iterates a story that the temple was first constructed as Jai Singh's sleeping quar- ters, but after divine revelation in a dream the king gave the structure over to Krishna as Govinda Deva. It is not possible to trace the antiquity of this belief. 87. List, 1:243. 88. See Erdman, Patrons and Performers in Rajasthan, 27-44, for the interaction of religion and authority in the city's construction. 89. See Diana L. Eck, Darshan: Seeing the Divine in India, 2d ed. (Chambers- burg, Penn.: Amina Books, 1985), 3-4. 90. List, 1:357. This was confirmed by the mahant in August 1994 and again in March 1996. This temple, which belongs to the Lalita Sampradaya, is discussed by Babulala Gosvami, Lalita Sampradaya: Siddhanta aura Sahitya (Pilani: Kamala Prakasana, 1991), and Entwistle, Braj, Centre of Krishna Pilgrimage, 213. 91. List, 1:243; Wilson, Hindu Religions, 119. Charan Das was born in 1703, and he died in 1783. 92. See List, 1:285, 287, 296, 322, 323, 333. 93. A tabulation of the secondary images included in Shahjahanabad's Shivalayas based on material provided in the List indicates that about half of them have Vaishnavite images. John Cort has suggested to me that the Shivalaya  148 | Mapping Hindu-Muslim Identities through Architecture is appropriate for a public shrine but that Krishna worship is more suitable for the private domain, and thus he feels that the home shrine probably was dedi- cated to Krishna. 94. See Asher, Architecture of Mughal India, 299-300, for Tahawwar Khan's mosque. 95. Tillotson, The Rajput Palaces, 190. 96. See Asher, "Piety, Religion, and the Old Social Order," forthcoming. 97. Blake, Shahjahanabad, 111; Bayly, Rulers, Townsmen, and Bazaars, 371-72. 98. Examples would include those Gupta period temples at Maria, Bhumera, and Natchna. Others might include those at Menal, Rajasthan, which is clearly a tirtha site. 99. See Bayly, Rulers, Townsmen, and Bazaars, 384-90. 100. Examples include the Naya Mandir in Mahalla Dharmapura (List, 1:310) and Jain temple in Kuncha-i Seth (List, 1:322). Even today continual restoration of the gold is undertaken in these temples; the authorities tell me with great pride that it is real gold. 101. Bayly, Rulers, Townsmen, and Bazaars, 388. 102. Varahamihira, Brhat Samhita, 2 vols., trans. M. Ramakrishna Bhat (Delhi and Varanasi: Motilal Banarsidass, 1981), 1:499, 516-38; Varahamihira, India as Seen in the Brhatsamhita of Varhamihira, trans. A. M. Shastri (Delhi and Varanasi: Motilal Banarsidass, [1969]), 268-69, 273, 395.  6 Indo-Persian Tazkiras as Memorative Communications Marcia K. Hermansen and Bruce B. Lawrence Naqsh faryadi hai kis ki shukhi-yi tahrir ka kaghazi hai pairhan, har paikar-i tasvir ka [Of whose careless recording does the inscription complain, For every representation wears a paper shirt.] Ghalib In a volume that focuses on Muslim identity in general and Indo-Mus- lim identity in particular, we have a treasure trove to help us understand how South Asian Muslims identified themselves. It is the tazkira, or bio- graphical compendium. The tazkira is a genre of literature produced by elites for other elites. Its primary linguistic expression in the subconti- nent is Persian, at least during the Mughal or Indo-Tirmuri period, which we emphasize here. Only in the late Mughal and colonial periods of Indo-Muslim history do we find both Urdu translations of Indo-Per- sian tazkiras and Urdu sequels, with companion compositions in Pun- jabi, Sindhi, and Gujarati. The tazkira is a staple of Indo-Muslim cultural production, and as such it demands closer analysis than it has so far received if we are to grasp the relationship between personal authority and place in structuring Indo-Muslim identity. Or, rather, Indo-Muslim identities, since the range of elite literary re- flection is very broad, and Indo-Persian tazkiras encompass a wide va- riety of contexts. Above all, the tazkira traces memory through the lives of heroes, both lyrical and spiritual, and in doing so, it raises a number of interesting theoretical questions. How are its heroes selected, and who gets to tell their story for which audiences? While the heroes may display courage, evoke hope, and elicit loyalty, how do their lives relate to the unheroic, to the everyday, to the lives without traces?  150 | Indo-Persian Tazkiras as Memorative Communications A common complaint by critics has been the lack of any consistent principles of selectivity, critical evaluation of facts, or analytical frame- work in the tazkira tradition.1 An anthologist has noted "their seemingly irrelevant style of diction."2 Some scholars of contemporary third world literature have even suggested that we cannot go beyond describing these texts. They are resistant to critical theory because they are drenched in the minutiae of local detail. Yet we propose a method of reading tazkiras that acknowledges the ambiguity at their core. As the poet Ghalib implies in the epigraph for this essay,3 there is no "careless recording." Each representation that wears a paper shirt is intended to communicate to others the quality, the cultural residue that commends its content to would-be readers. Tazkiras are not mere mnemonic repetitions. They are conscious remem- brances, and therefore they are both cultural artifacts and cultural recon- structions. In Walter Benjamin's language, "the 'after' is precisely not the 'again."' The "after" requires "the destruction of the illusion of its continuity with the past," for "only thus can the past be 'put to work' in the present as remembrance."4 If it is possible for tradition to appear in the guise of cultural history, and if that history is marked above all by narrative forms, then the Indo- Persian tazkiras of South Asia exemplify what could be labeled, in a paraphrastic gloss from Benjamin, "memorative communication."5 At the vertical level, these memorative communications reflect the divine favor conferred on worthy Indo-Muslim emissaries-in this case, saints and poets-just as these same emissaries reflect the divine impress on their own creatively courageous lives. At the horizontal level, Indo-Per- sian tazkiras project a collective testimony for others who also locate themselves in the same subcommunity of South Asian Muslims. These are literary works that both remember and communicate. They concen- trate the readers' focus on their heroic subjects at the same time that they disperse, or redeploy, that focus to present-day concerns and contingen- cies. Although they draw from the past, they are not commemorative; they do not recall the past for its own sake or for the sake of the heroes whom they exalt. They are memorative, relying on memory and remem- brance to communicate with the living the legacy of prior Indo-Muslim exemplars. Indo-Muslim exemplars are Muslim as well as Indian, and so we find that certain general cultural themes typical in Islamic civilization persist in the tazkira genre in Muslim South Asia. For example, the moods of nostalgia or boasting can be traced back even to the Jahiliyya poetry of  Marcia K. Hermansen and Bruce B. Lawrence | 151 pre-Islamic Arabs.6 The genealogical preoccupation of the Arabs merged with the formulation of a sacred history embodied by the early Muslim community in early biographical compendia. Works written in Arabic, such as the Tabaqat of Ibn Sa'd (ca. 784), feature the practice of exten- sively listing very ordinary participants in the Muslim community, as if to somehow represent its existence and significance by remembering even the names of who had been present, in other words, to provide a trace for those who otherwise would remain traceless, unacknowl- edged, forgotten.7 But much more important in terms of thematic characteristics is re- gional and urban location. The nature of Islamic civilization, at least from the perspective of the celebrators of its intellectual vibrancy, had an overwhelmingly urban focus, and we will demonstrate from several cases how the memorialization of cities came to characterize this genre as it expanded into the space of Muslim South Asia.8 The urban notables who abound in the pages of Indo-Persian tazkiras are rarely rulers, sometimes religious scholars, but more often urban intellectuals. It is, above all, poets and saints who become the principal subjects memorial- ized in Islamic biographical literature generally but even more fre- quently in the Indo-Persian tazkiras of South Asia. Common to each entry is attention to the rank, affiliation, profession, and year or century of death of the person remembered. But no less important is the locality of the individual's primary activities, with the double message that the urban setting enhanced the spectrum of possible activity for the de- ceased at the same time that the achievements of this notable brought to that city a fame or spiritual bounty it did not, or had not, enjoyed before his lifetime. Nowhere is the accent on place and the reciprocal importance of he- roes and homes more pivotal than in a distinctly biographical genre developed in India due to the influence of institutional Sufism. That genre is the malfuzat, or recorded conversations, of spiritual masters. It is developed by the most notable Indo-Muslim brotherhood of the premodern period, the Chishtiya, and from its earliest appearance in North India to its later proliferation throughout Hindustan, it is identi- fied with specific saints whose places shape their audience and whose responses, or question-and-answer sessions with close disciples, en- hance the places where they settle and teach, advise and warn, fast and pray, and meditate and eventually (usually after long lives) die.9 From the major Chishti exemplar of his day, Shaykh Nizam ad-din Awliya (d. 1325), in Delhi to his disciple, Shaykh Burhan ad-din Gharib, in the  152 | Indo-Persian Tazkiras as Memorative Communications Deccan and especially in Burhanpur (named after him), we find the genre of malfuzat fostering an intensely localized memory, even when it is a "false" memory. "False" memory? Yes, for as Carl Ernst has noted, we should not ap- proach the malfuzat only as historical purists. Although many later malfuzat either were spurious or projected major distortions of their core contents, they still reinforced both the authoritativeness of their saintly subjects and the place where they presided, then expired. What interests us is that both kinds of malfuzat became included in the canon of South Asian Sufi memory.10 The lesson from malfuzat is extended in the genre of tazkira. It is attention to place, or to relocation of place, from a real or imagined Cen- tral Asian or Arabian homeland to a new South Asian primary home, that governs the principle of biographical collection and the inscription of an altered sense of identity. Even the focus on individual heroes is subordinated to a still larger purpose: the collective display of groups of individuals. It is as groups of heroes, sharing a common identity and a convergent legacy, that saints and poets reinforce the value of a Muslim presence in South Asia-not just anywhere in South Asia, but in South Asian urban settings, in premodern cities. It is bards and Sufis together who authorize the cultural symbolism of South Asia as an urban, and also an urbane, Muslim realm. Especially during one period, the Mughal or Indo-Tirmuri, from the early sixteenth century to the mid-nineteenth century, we find this pro- nounced tendency to compile tazkiras both to honor heroes and to au- thenticate Muslim urban spaces. For that reason, while our essay will look at multiple points of the millennial long history of South Asian Islam, we will give particular stress to those poetic and saintly tazkiras that emerged from Mughal India.11 Tazkiras, as their name suggests, both memorialize individuals and communicate their legacy to a new generation. They are, to repeat our title, "memorative communications," projecting the worth of individual heroes as icons of urban Indo-Muslim collective identity. If there is a key word for understanding poetic and Sufi tazkiras, it is memory. Memory has a long etiology in western thought, which is bound up with its invo- cation in Islamic contexts. Both need to be considered before applying our analysis to Indo-Persian figures from South Asia.  Marcia K. Hermansen and Bruce B. Lawrence | 153 The Concept of Memory in Muslim South Asia A study of memory in medieval European culture, The Book of Memory by Mary Carruthers,12 suggests some interesting approaches to looking at memorialization in premodern Muslim South Asia. For example, con- sider the civilizational context of memory. In developing an Islamic con- cept of memory, a root image is established by the Qur'anic metaphor of the source of primordial reality in a "preserved tablet."13 The "pen" that inscribes is also mentioned in the Qur'an (68:1), and this imagery has been drawn upon throughout the interpretive tradition of Muslim Neo- platonic philosophy and in the philosophical and poetic tradition fol- lowing Ibn Arabi's Sufi metaphysical system.14 The world itself and hu- man, religious, and political destinies are conceived as a sequence of articulations of what has been written on this primordial preserved tab- let (al-lauh al-mahfuz). Embodied in mahfuz is a double meaning: What has been "preserved" has also been "memorized." Annemarie Schimmel, discussing "the pen" in a section on "letter symbolism in Sufi literature," notes: The mystics have dwelt on another aspect of pen symbolism as well. There is a famous hadith: "The heart of the faithful is between the two fingers of the All-compassionate, and he turns it wherever He wants." This hadith suggests the activity of the writer with his reed pen, who produces intelligible or confused lines; the pen has no will of its own, but goes wherever the writer turns it. . . . The hadith of the pen has inspired the poets of Iran and other coun- tries-they saw man as a pen that the master calligrapher uses to bring forth pictures and letters according to his design, which the pen cannot comprehend. Mirza Ghalib, the great poet of Muslim India (d. 1869), opened his Urdu Diwan with a line that expresses the complaint of the letters against their inventor, for "every letter has a paper shirt."15 We should not be surprised that the same Aristotelian categories of form and matter which structured medieval European thinking about the process of creation, whether divine or human, were also inherited by the Islamic tradition. The imagery of wax which takes on the shape of the mold or the signet ring was a way of describing the creative process as it was channeled through retrieved memory. Tracing the develop- ment of this model in Socrates and Plato, Carruthers observes, "In fact, Socrates is at some pains to say that his way of describing the memory as  154 | Indo-Persian Tazkiras as Memorative Communications being like seals made by a signet ring is not new, but really is very old. This is important because it is a model based on how the eye sees in reading, not how the ear hears. In recollection, one looks at the contents of memory, rather than hearing or speaking them; the mediator is vi- sual."16 Coexisting with and sustaining this centrality of memory in clas- sical Islamic civilization were the now almost forgotten technologies of memorizing and then retrieving memorized information in speech and writing.17 In the experience of madrasa study, for example, the visual arrange- ment of text and commentary on the written or printed page was piv- otal, as was the tradition of transforming essential texts or principles into rhyme. Al-Suyuti's Alfiyya,18 for instance, was composed in order to make the rules of hadith criticism accessible for memorization, and Shah Wali Allah of Delhi undertook a rhyming translation into Persian of the standard manual of rhetoric, the Sarf-i Mir.19 Memory was closely linked to remembrance for South Asian Muslim elites. It included various acts or practices such as dhikr (ritualized reci- tation of pious phrases that are sanctioned by Qur'anic injunctions to remember God in all situations),20 yad dasht,21 tasawwur,22 and hiif:. In Sufi practice, moreover, memory was ritualized in the recitation of shajaras or khatms.23 On the occasion of celebrations of a saint's marriage to God ('urs), or at other notable moments, the lineage of a Sufi order would be ritually recited, with the belief that the spirits of the departed saints would present themselves and bestow blessings on those assembled. The Sufi ritual of dhikr is a more specialized form of a basic Islamic practice of the remembrance of God through recitation and repetition; its opposite, ghaflat, that is, "forgetting" or "negligence," is both a moral shortcoming in terms of religious piety and a personal affront to the beloved in the tradition of poetic love. This premodern sense of memory in Islamic civilization included an appreciation both of memory as "recollection," which constituted a powerful tool for self-awareness and creativity, and memory as connec- tion, which incorporated the emotional dimension of the act of memor- ialization. Thus the inscription of memory as a cultural activity involved both an appropriation of power over a space and the creation of an emo- tional investment in it. It was, in every sense, a memorative communica- tion, not merely a commemoration of bygone saints/poets or nostalgia for their heroism, whether courageous spirituality or lyrical fervor. Ham parvarish-i lauh-o-qalam karte rahenge joo dil peh guzarti hai raqam karte rahenge  Marcia K. Hermansen and Bruce B. Lawrence | 155 We will safeguard the tablet and the pen, We will record whatever the heart experiences. Faiz Ahmad Faiz24 Language and writing then inscribe what is essential; they also project values that must be reappropriated in each successive genera- tion of readers, almost invariably urban elites who descended from other urban elites. "Literature was thought to contribute to the ethical life of the individual and the public memory of society," writes Carruthers in reference to medieval Europe. In the case of South Asian Muslim tazkiras, their writing suggests the further intention and effect of "making Muslim space." From multiple regions of Hindustan, their authors appropriate urban places and authorize them as the sites of Indo-Muslim cultural memory. In studying the Chishti shrine complex at Khuldabad, for instance, Carl Ernst recounts two stories connected with memory that link the classical Islamic tradition with South Asian tazkira compositions. The Chishti saint Nizam al-Din Auliya' (d. 1325) in commending the efforts of the poetic compiler of his malfuzat, Hasan Sijzi, suggests a parallel to Abu Hurayra (ca. 678), the most prolific transmitter of Prophetic hadith according to the Sunni tradition. "Nizam al-Din said that the Prophet told Abu Hurayra to extend the skirt of his garment whenever the Prophet spoke, then slowly gather in the garment when the words were finished and place his hand upon his breast; this routine would enable him to memorize Muhammad's words."25 This same motif of extending the skirt of the garment to collect words of wisdom and guidance is later echoed in another malfuzat, Khayr al-Majalis, where Hamid Qalandar applies this method in recording the sessions of Nizam al-Din's succes- sor, another Delhi Chishti master, Nasir al-Din Chiragh-i-Dihli.26 The symbolism of extending and pulling in the skirt of a garment when applied to a literate tradition evokes the process of interpretation through reading inward from the commentaries on the margins of texts to the texts themselves in order to gain access to the core meanings. This also sets up a resonance with the emotional quality of memorizing and preserving the words of an individual. It underscores the link between memory and personal devotion. Grasping the daman or skirt of a gar- ment is, in fact, the gesture of the petitioner or supplicant, resonant with the paper shirt worn by the plaintiff of Ghalib's couplet. It also under- scores the need of each person in the present to account for the represen- tation offered, that is, to be an active agent in the process of commemo- rative communication.  156 | Indo-Persian Tazkiras as Memorative Communications Patterns in Poetic and Sufi Tazkiras The tazkiras of both poets and Sufis evidence the changing shape of Muslim identity in South Asia through their principles of ordering as well as their thematic concerns. We have stressed their commonality, but one might also note their difference. While the role of saints is to sanctify the new soil of Hindustan and, above all, its cities, the function of the poets and poetic tazkiras lies elsewhere. In the case of the poetic tazkiras, the language and imagery of a city's poets inscribes another sort of privileged space; most often it sets the scene for a particular "state of mind" associated with that place,27 and to understand that link we will examine the poetic tazkiras before turning to their saintly counter- parts. Poetic Tazkiras So extensive has been the scholarship on tazkira as a genre applied to the lives of Indo-Muslim poets that it merits brief review. Some of the basic work in the literary history of this genre was carried out by the Pakistani scholar Farman Fatehpuri in a special 1964 issue of the journal Nigar and then in a 1972 monograph entitled Urdu Shuara kB TazkirB aur Tazkira Nigari.28 Fatehpuri suggests that the model for subsequent poetic tazkiras was the early Persian biographical compendium Lubab al-Albab, composed in A.H. 618/1282-83 C.E. by Muhammad Awfi.29 A glance through Awfi's chapter headings indicates that the principle of organization of this tazkira was primarily chronological rather than spatial. E. G. Browne, who edited this work, evaluates it as primarily an anthology and disappointing in biographical particulars.30 The early tazkiras in South Asia were written in Persian, even those of Urdu poets. A glance through Fatelpuri's catalog of tazkiras indicates that the three earliest tazkiras of Urdu poets were written in Persian in the same year, 1165/1752.31 Perhaps a conclusion might be drawn re- garding the connection of tazkira preparation with patronage networks. The Mughal empire at this point was on the verge of takeover by Euro- peans; it was also experiencing a major shift in its revenue system. One might speculate that part of the motivation for compiling such compen- dia was to draw the attention of potential patrons and gain reward for the needy litterateur. The prospect of European rule could have also produced another motivation: faced with the loss of social as well as  Marcia K. Hermansen and Bruce B. Lawrence | 157 political and economic power, Indo-Persian elites may have intensified their memorative communications, with poets high on the list of those deemed to be endangered species! According to Fatehpuri, until the Ab-i Hayat (Water of life) of Mu- hammad Husain Xzad" in 1880, most of the tazkiras were written in Persian following the old formula.33 Azad's tazkira has been celebrated as a breakthrough or watershed in the genre that, according to Ralph Russell, a noted scholar of Urdu, helped "lay the foundation of modern literary criticism in Urdu." Russell cites Azad's readiness to learn from the British methods, that is, in employing historical critical standards. He explains how Azad advanced on the traditional tazkira form, which had been essentially that of a biographical dictionary, providing "the poet's name, his takhallus (pen-name), the city of his birth, his patrons, the date of his death, a description of the quality of his poetry, couched in rather conventional terms, and one or two specimen couplets from his ghazals." Azad's principal contribution was to introduce a periodi- zation of Urdu poetry into five periods based on chronology and the use of language by the poets.34 Two scholars of Persian and Urdu have attempted to sort traditional tazkiras according to two rather unrevealing types, the general (in which the time frame is not limited) and the particular (centered on a particular period).35 Fatehpuri raises a more interesting tension regard- ing this tradition by inquiring whether these biographical works were primarily composed as a showcase for the poetry, or whether, in fact, the biographical component was the primary motive for composition.36 The French scholar Garcin de Tassy took an interest in the tazkira form, compiling an extensive list of tazkiras available to him in the mid- nineteenth century and making synopses of their notices on Hindu and Muslim poets of Hindustani, together with representative translations of their work in three volumes. He speaks of the poetic notices falling into three basic categories, based primarily on the quality and extent of a poet's production, so that some figures merit only a brief notice, whereas poets at a middle rank who have produced longer collections known as divans or kulliyat receive an "honorable mention." The highest category, in his estimation, are those notices of poets or authors whose works have been given specific titles.37 In exploring the motivations for the composition of tazkiras, Fateh- puri makes the following proposals: 1. The memorialization of the compiler or others might have moti- vated tazkira writing.  158 | Indo-Persian Tazkiras as Memorative Communications 2. Biaz nigari, or the citation of favorite or thematically coherent poetic couplets, might have been the true motivation behind bio- graphical compendia.38 3. Another motivation could be the urge to discuss the personal traits and rivalries of poets. 4. There was also appeal of such works during that period of arti- ficiality in speech and imagination; they could appeal as a way of denying political and material decline by retreating to an interior world. 5. One also had to consider the increasing popularity of poetic gatherings (musha'iras) and the publication of more and more po- etic anthologies based on the couplets recited at a particular one.39 6. Finally, they might assist the movement to establish Urdu over against Persian.4* By contrast, the contemporary critic Muhammad Sadiq evaluates tazkiras solely by the standards of historical accuracy. The history of the tazkiras reveals a more serious approach and a greater desire for authenticity and fair play as time passed. The earlier tazkiras, for the most part, confined themselves to notes on the poets and drew heavily on their predecessors. Subsequent writers enlarged the sphere of their research by including discus- sions on prosody, diction, and the history of the Urdu language, some of them discarding the alphabetical order in favor of the chronological. They also tried to establish contact with their con- temporary poets, and obtained first-hand information about them from their friends and relatives. We may say, therefore, that the history of tazkiras shows a steady advance in research; and what was once a pastime, a desire for personal recognition, or a means of expressing one's approval or disapproval, became a really respon- sible undertaking.41 Not till very recently, though, have we seen the student of poetic tazkiras shift to the interest of our essay: the use of heroes and homes in a reciprocal form that enhanced the benefit of each for Indo-Muslim ur- ban identity. It is the American linguist Carla Petievich who has traced the organizing principle of locating the Urdu poetic tradition. It has shifted from the space of the markaz or city-based circle, centered around an ustadh (or master), to the region, which became defined by the scope of princely patronage, before it finally became linked to the school,  Marcia K. Hermansen and Bruce B. Lawrence | 159 based on canons of European literature as taught in the new universities modeled on the British system. Ironically, although this latter, accompa- nied by the inculcation of Victorian literary and moral standards, was an artificial fit with the reality of the South Asian Urdu poetic tradition and its lines of influence, it did become the generally accepted way of distin- guishing the "Dihlavi" from the "Lakhnavi."42 In other words, the very accent on competing traditions legitimated these major North Indian cities as spaces of urbane Indo-Muslim cultural expression.43 What resulted was more than an innocent competition of mutually reinforcing minority identities. The intense urban/regional focus of later poetic tazkiras became a crucial marker of regional identity. Tazkiras highlighted linguistic variants, rivalries over poetic eloquence, or even the correctness of local expressions. They thus served to define a certain space in terms of a "state of mind," with Lucknow, for example, seeking the highest ground as the epitome of refinement, "nazakat." Carla Petievich has criticized this "two school" construction of Urdu poetry in the late eighteenth and nineteenth centuries.44 Lucknow po- etry, she observes, is too often represented as decadent: "It is acclaimed as the quintessential symbol of what Muslim culture in India achieved, yet it is simultaneously denounced today for the societal immorality, waste and decadence of its past." She goes on to lament that "in the case of Lakhnavi poetry, most critics have described the society of Lucknow as leisured, pleasure-loving and courtesan pursuing."45 Other character- istics of Lakhnavi style cited in the critical literature include effeminacy, sensuality, frivolousness, a certain vulgarity, amorous repartee, and a lack of the traditional ambiguity regarding the beloved's gender and whether a divine or human beloved is really the object of the poet's address.46 Petievich, however, tries to rescue Lucknow from the slanders of its detractors: she analyzes samples of poetry from various cities in order to disprove the applicability of these stereotypes solely to Lakh- navi poetry. Nor is the shift in cultural memory as represented in the tazkira tradi- tion limited to Lucknow. Another city-based tazkira from the twentieth century is the Kamilan-i Rampur. Written by Hafiz Ali Khan in 1929, it was reissued by the Rampur library in 198647 with the addition of a postnationalist preface to the original composition. In this the writer of the new preface, Abid Rida Bedar, develops the concept of Rampuri Urdu poetry as a "third school" aside from Delhi and Lucknow.48 He evokes symbols of Rampuri identity such as a Rampuri cap (topi), a knife, and a particular style of knife fighting. And he also argues for the inclu-  160 | Indo-Persian Tazkiras as Memorative Communications sion of the Ali brothers, activists in the independence movement, who had connections to Rampur but were omitted from the original tazkira. In other words, the tazkira takes on still another urban role: to demon- strate the emergence of a proto-nationalist space within the Muslim memory of particular cities, in this case, Rampur. The principle for inclu- sion is widened beyond saints and poets to include other notables, espe- cially those identified with the emergence of independent India. Sufi Tazkiras Sufi tazkiras provide a genre parallelism to the poetic tazkiras but with some distinct differences. In the case of the Sufi tazkira tradition as in the poetic one, inspiration was drawn from Persian models, in this instance, Attar's famous Tazkirat al-Auliya as well as from the Naqshbandi- inspired models, Abd al-Rahman Jami's Nafahat al-uns49 and Kashifi's Rashahat.50 Also, Sufi tazkiras, like the poetic tazkiras just examined, are closely linked both to the institutional formation and the collective memory of Indo-Muslim elites. A kind of partial tazkira is suggested by a section of the earlier Kashf al-Mahjub of Hujwiri (d. 1074) that establishes the idea of tariqa or Sufi order. Subsequently, the schema of fourteen Sufi lineages or families became the organizing feature for most later tazkiras, despite its often poor fit, which generated many anomalous categories. One possibility is to contrast the tariqa-based tazkiras with those that attempt to catalog all orders, since multiple affiliation among Sufis had become more com- mon by the sixteenth century.51 But even more important is to connect the writing of these tazkiras with the places and the contexts, either re- gional or urban but most often both, that generated the need for memorative communication at the heart of tazkira writing. It is for this reason that we draw special attention to Indo-Persian production from the high Mughal or Indo-Timuri period. It epitomizes the elite literary activity that we are analyzing throughout this essay, and at the same time it gives an extended case illustration of one of the most powerful processes of cultural production at any point in the his- tory of Islamicate South Asia. Each Indo-Persian tazkira from premodern South Asia illustrates concern for one's own silsilah as the declared motivation for authorship. There is one exception: Abd al-Qadir al-Badauni's Muntakhab at-taw- arikh, the memorative communication of a pious but independent-  Marcia K. Hermansen and Bruce B. Lawrence | 161 minded courtier that throws light on the patronage that shapes the pro- duction of other Mughal period Sufi tazkiras.52 But to understand the rareness of Badauni's project, one must first note other tazkiras that con- firm the pattern of privileging one's own order in telling, or retelling, the entire drama of Muslim saintly labor. Let us consider the ill-fated older son of Shah Jahan, Dara Shikoh (d. 1659). Before he was executed by his younger brother, Aurangzeb, Dara was both a Sufi adept and a Sufi tazkira author. Dara composed not one but two Sufi tazkiras, fulsome dictionaries of antecedent Muslim spiri- tual heroes. Dara seems to have been motivated by a concern for getting the record straight, but it is a surface concern. Dara's apparent concern masks his overriding goal: not only to affirm Abd al-Qadir Jilani as the foremost Sufi exemplar and the Qadiriya as the paramount Sufi brother- hood but to undergird his own authority vis-a-vis rival claims to Qadiri spirituality. His was not the first Indo-Persian biographical dictionary written by a Qadiri. He was preceded by the formidable scholar of hadith, himself a Qadiri adept, Shaykh Abd al-Haqq Muhaddith Dihlawi (d. 1642). Abd al-Haqq's Akhbar al-akhyar, completed in 1618, had al- ready gained considerable fame by 1640, and Dara models many of his own entries on Indian saints after the longer, fuller entries of Akhbar al- akhyar. Yet in presenting the Qadiriya, he bypasses the lineage traced by Abd al-Haqq, acknowledging only that line of Qadiri affiliation trace- able through Abd al-Qadir ath-thani to Abdallah Bhiti to Miyan Mir (d. 1635) and then to his own preceptor, Mulla Shah (d. 1660). The significance of the Islamic past for Dara Shikoh is functional: its retelling helps to affirm his own status as a Qadiri adept. The tazkira, in his imaginative plane, becomes the ideal tool of memorative communi- cation. Giants of Persian Sufism like Ala al-Dawla Simnani and Jalal al- Din Rumi, when mentioned, are accorded half a page, consisting mostly of a cursory recap of standard biographical, travel, and literary data. While their inclusion affirms Dara Shikoh's awareness of the long tradi- tion in which he stands, their sole purpose is to provide a backdrop for the stage onto which he parades as central exhibit the Qadiriya, espe- cially his own immediate spiritual mentors, and their location in the region of Lahore. Is it mere coincidence that his disagreement with the Delhi author of Sufi tazkiras, Shaykh Abd al-Haqq, has to do with the region of their respective Qadiri affiliations? It would seem that even within the domain of elite Indo-Persian cultural production the impor- tance of space, specifically urban sacred space, was determinative, even if unstated.  162 | Indo-Persian Tazkiras as Memorative Communications Dara Shikoh's Safinat al-auliya also contrasts with another biographi- cal dictionary from Mughal India. While much has been written about Safinat al-auliya, mention is seldom made of the Chishti master, Shaykh Abd ar-Rahman (d. 1094/1683), or his Sufi tazkira, Mir'at al-asrar. Al- though it appears in several published catalogs, Mir'at al-asrar has never generated a fraction of the interest directed to Safinat al-auliya, yet the two works merit comparison, if only because their authors were near contemporaries, they employed the same inclusive method of tazkira writing, and, above all, they were both preoccupied with the relation- ship of personal authority to place. In Mir'at al-asrar, after noting the twelve family clusters into which Sufi brotherhoods may be parceled, Shaykh Abd ar-Rahman reviews no less than twenty-three generations of spiritual exemplars. He brackets the prophet Muhammad and his three immediate successors as the first generation, followed by 'Ali and the other eleven Imams in the second generation. He continues in this manner until he reaches the tenth gen- eration in which the first Chishti master is said to have lived and died in Syria (ca. 328/940). Appearing in the same generation with him were his contemporaries Shibli and Hallaj. By the time of the fourteenth genera- tion when Qutb al-Din Mawdud (d. 537/1132) became the successor at Chisht, he counted among his contemporaries Muhammad and Ahmad Ghazzali as well as Ayn al-Quzzat Hamadani. Successive generations boasted still more illustrious names, so that by the sixteenth generation when Uthman Harvani (d. 607/1210) became the Chishti standard- bearer, he welcomed as fellow mashaikh both Abd al-Qadir Jilani and Abu Madyan Maghribi. Abd ar-Rahman's primary purpose is to retell the saga of Persian/ Indo-Persian Sufism as a single dramatic endeavor shaped by the Un- seen for the benefit of humankind. Yet from the nineteenth generation on, that is, from the time of Shaykh Farid al-Din Ganj-i Shakar (d. 664/ 1265), a major Chishti saint in the Sultanate period, to the end of Mir'at al-asrar, the Indo-Persian actors begin to overshadow their Persian pre- decessors. After the eighteenth generation, few if any non-Indian saints are even mentioned, and the reason is directly connected to place and its importance for structuring collective identity. Shaykh Abd ar-Rahman is not only a Chishti master; he is also the incumbent of a shrine in Avadh, well to the east of Delhi in modern-day Uttar Pradesh. Authenticating Avadh as an urban Muslim realm is as delicate as it is crucial for Shaykh Abd ar-Rahman. He traces his own spiritual lineage  Marcia K. Hermansen and Bruce B. Lawrence | 163 back through the Sabiri rather than the Nizami branch of the Chishtiya. That lineage is beset with chronological difficulties that cloud its initial years. Its eponymous founder was one Shaykh Ala al-Din Ali ibn Ahmad Sabir, who died in 691/1291 in Kalyar, a town in northern Uttar Pradesh. He is said to have been identical with the Shaykh Ali Sabir, who is briefly mentioned in Siyar al-auliya as a disciple of Shaykh Farid al-Din Ganj-i Shakar. No less an authority than Shaykh Abd al-Haqq, however, questions the conflation of the two names and persons. Even if it were to be accepted, there seems to be more than a generation between Ali Sabir's successor, Shams al-Din Turk Panipati (d. 718/1318), and his successor, Jalal al-Din Panipati (d. 765/1364). Further comprising the historical markings of the lineage is the fact that Ahmad Abd al-Haqq (d. 837/1434), who succeeds Jalal al-Din and is the biological as well as spiritual ancestor of Abd ar-Rahman, was not born until ca. 751/1350.13 Yet our concerns with chronological plausibility and historical accu- racy were not Abd ar-Rahman's. Instead of lingering on these hiatuses and discrepancies, he paints a colorful canvas of spirituality that in- cludes all the major figures of the Nizami branch of the Chishtiya as part of his own mystical legacy. Unlike Dara Shikoh's brief reminders, these are full, vivid accounts of both Persian and non-Persian saints of earlier eras. The organization by successive tabaqat or generations, despite the chronological discrepancies, draws attention to the preeminent Sufi au- thority (the "axis" or qutb) of each age. From the perspective of Abd ar- Rahman's lineage, the qutb of each age, since the appearance of Shaykh Ali Sabir, had to be, and has been, a Sabiri Chishti master. Yet his is not a partisan view arguing for Sabiris over Nizamis, Chishtis over other Sufis, Sufis over other Muslims, or Muslims over Hindus. Instead, he shows a wide acquaintance with classical Persian Sufism and an appre- ciation for the luster that its exemplars bring to his own generation and to his own place. While each generation is marked by a qutb, he is situ- ated among, not apart from, other Sufi masters. Although he stands at their head, they add to his preeminence. By this ingenious artifice the author of Mir'at al-asrar accomplishes a double purpose: (1) he makes clear how vital was the connection to a Persian Sufi tradition for all Sabiri Chishtis while (2) conferring the highest spiritual rank on a hand- ful of obscure saints, most of whom lived and toiled and died in north- eastern India, specifically in the region of Avadh. What is evident in Abd ar-Rahman's wide-ranging account of his saintly forebears is also discernible in the very different project of  164 | Indo-Persian Tazkiras as Memorative Communications Badauni. Although he devotes but one section of his Muntakhab at- tawarikh to saintly biographies, it would be hard to overemphasize their significance for him. Not only do they exude a freshness lacking in the comparable section of Ain-i Akbari, itself part of the Akbarnama, the most famous commissioned history of Mughal India, but they also indicate the variety of spiritual endeavors that were taking place outside the royal court. Unlike the narratives of Shaykh Mustafa, the Mahdavi mas- ter whom Derryl MacLean analyzes in a later essay of this volume, none of the endeavors depicted by Badauni were in explicit competition with the imperial cult increasingly focused on Akbar after 1574, that is, for almost the entirety of his reign at Fatel.pur Sikri. Badauni was a maverick intellect. He had no illusions of obtaining a reward for his book. He did not write to please his powerful patron. At most, he may have entertained the hope of some historical redress. Above all, he wanted to acquit himself at the court of Divine Justice, as is clear from his final supplication: "[If it] please God this work will, for a while, be preserved from the treachery of lack of preservation, of faith- lessness, or of evil guardianship ... and being constantly hidden under the protection of God's guardianship, will receive the ornament of ac- ceptance."54 Yet, even if one discounts the author's special pleading for the authority of his own experience, the sum total of these individual accounts provides an independent profile of Indo-Muslim identity as shaped through institutional Sufism, and it confirms both the resilience of the orders in their regional manifestations and also the significance of local, often urban Sufi lodges. For Badauni, the strongest claimants to spiritual authenticity were those Shaykhs who combined a grounding in the traditional religious sciences of Sunni Islam with an attachment to mystical pursuits. Two exemplars from less well known urban sites are Shaykh Nizam al-Din Ambethi and Shaykh Daud of Chati. In both cases Badauni dwells on noble ancestry, pursuit of learning, and calm judgment under fire. The present-voice narrative infuses his account of these and other saints. Both Shaykhs come alive as holy men constantly being tested, whether by jealous notables, a distant sultan, or a persistent visitor. With Shaykh Nizam al-Din, it is Badauni who is the overzealous guest, making a ver- bal faux pas that seems to doom him never to obtain the saint's favor. But in the case of Shaykh Daud, it is the saint himself who is set up to be the victim of a court conspiracy against Sufi masters (perhaps because of his Mahdavi persuasions). His gracious manners and sound learning not only rescue him. They actually turn the tables on his would-be persecu-  Marcia K. Hermansen and Bruce B. Lawrence | 165 tors and secure fame both for him and for the minor urban location of Chati where he taught and prayed and was buried. The importance of place in Badauni's Sufi biographies becomes still clearer when his vignettes are intercalated with the acknowledged mas- ter of Mughal hagiography, Shaykh Abd al-Haqq Muhaddith Dihlawi, whom we noted above. Abd al-Haqq is himself the subject of one of Badauni's sketches, confirming that the production of saintly biogra- phies was a proven means of securing memorialization in one's own right! Like Badauni, Abd al-Haqq was among the Indo-Persian urban elite of the late sixteenth century: even though he survived well into the reign of Shah Jahan, his most famous tazkira, Akhbar al-akhyar, was writ- ten during the third phase of the Akbar period, ca. 999/1591. Also like Badauni, he was not beholden to the new imperial ideology constructed by Abul-Fazl and advocated by Akbar in the late 1570s, for although he studied at Fatelpur Sikri as a teenager, by age twenty-one (1572) he opted to return to Delhi, where he had been born and reared, where his parents still resided, and where he could teach in his father's madrasa. Unlike Badauni, however, Abd al-Haqq is clearly writing his work for public dissemination. In the light of Badauni's fears, his literary strategy has to be subtly shaped, at once revealing and concealing his true inten- tions. Unable to disagree with the emperor directly, he also cannot fol- low the not so subtle pattern of Badauni's clandestine work: to criticize those who were the confidants of the emperor, especially Faizi and Abul-Fazl. Instead, Abd al-Haqq constructs his work in such a way that it both supports Akbar's imperial agenda and offers an alternative set of spiritual authorities. He lauds the Chishti epigones of virtue but does not dwell on Shaykh Salim. Rather, he adopts a diachronic scheme that begins with the Chishtis and so with Muin al-Din and then progresses generationally through the Delhi Sultanate to the Akbari era. The saints who merit most extensive attention and whose biographies mirror Abd al-Haqq's disposition are the later Qadiris. They were the spiritual pre- cursors of his father, Sayf ad-din, and also his own mentor, Abd al- Wahhab. So generous does he appear to be toward all saints that a censor would have been hard-pressed to fault him on either his organizational strategy or his more than 250 individual entries. In short, Abd al-Haqq attempted to be more than a pawn in the grand design for expanding Mughal hegemony that Akbar, with assistance from his courtiers, di- rected. Yet the Delhi savant could not operate outside the constraints of a bureaucratic structure that dominated, even as it animated, all aspects of the emergent Indo-Persian culture complex, and he himself was  166 | Indo-Persian Tazkiras as Memorative Communications prone to privilege those saints whose labor confirmed the region of northern India where he himself lived and labored, prayed and fasted, and died.55 The very process of memory and recording had its constraints: Not all Indo-Persian memorative communications had the freshness of Badauni or the comprehensiveness of Abd al-Haqq. While the period of Mughal imperial expansion made possible the concept of a pan-Indian scope in historical writing, as we have seen above,56 most later tazkiras tended to have a more limited scope, because of the restricted audience for whom their authors were recording as they tried to memorialize the saints of earlier epochs. Rather than the analytical study of tazkiras, which might generate new categories, one too often finds a replication of the genre, still another tazkira of tazkiras rather than a creative or locally derived approach to memorialization. In the colonial period we find the routinization of tazkira writing taken to new depths of serial logic. The idea of the comprehensive or cataloging tazkira, one which listed Sufis of all orders in tabular form, became prevalent, in part due to the influence of maps and census tak- ing.57 Examples of such compendia abound, the most notable being the Masalik al-Salikin: Tazkira al-Wasilin of Mirza Muhammad Abd al-Sattar Baig (Agra: Matba' Faid, n.d.) (Urdu) and Hadiqat al-Asrar fi Akhbar al- Abrar of Imam Bakhsh (Lahore, 1364/1944). More interesting for our general thesis is the way that Indian cities become Muslim holy spaces for certain Sufi tazkira authors. From an early date Ajmer had been recast as Madina58 in the biography of Muin al-Din Chishti, but more extensive still was the new topography of holy cities charted in Kalimat al-Sadiqin of Muhammad Sadiq Dihlavi Kash- miri Hamadani.59 The work is a tazkira of the Sufis buried in Delhi up to the year 1023/1614. The author, a student of Baqibillah, the Delhi-based Naqshbandi Shaykh, whose prize disciple was Almad Sirhindi, dis- cussed in the next essay by David Damrel, claims to have modeled his work on the Rashahat of Kashifi. Consider how the author depicts Delhi as Muslim "sacred space." In his preface, he asks God to protect Delhi from calamities, and then continues: Know, may God support you with the light of gnosis, that Delhi is a very large and noble city and that certain of the saints of the nation (ummat) have said things about it like, "One in a thousand and very few out of the multitude recognize its greatness." Thus, whoever has the least understanding and the slightest knowledge  Marcia K. Hermansen and Bruce B. Lawrence | 167 will surely recognize that after the two holy cities (of Mecca and Madina), if there is any nobility to be found in a place or greatness in a land, it is in this noble land which is distinguished completely over the rest of cities and is exceptional. Therefore it is said by the common folk that Delhi is a little Mecca and even the elite have no doubt of its greatness. Everyone asserts its exaltedness, whether due to the fact that the great ones of the religion, the ulema among the people of certainty, the great shaikhs, the reputable wise men, the powerful rulers, and the exalted nobles have filled this city and have been buried here, or due to its fine buildings, delightful gar- dens and pleasant localities.... According to some esteemed personages, since one of the people of mystical intuition said in elaboration, "All of Delhi is declared to be a mosque," all of this city is distinguished from other places by its greatness and nobility. In summary, these verses of Khwaja Khusrau inform us of the greatness of this city and certain of its sites. Noble Delhi, shelter of religion and treasure, It is the Garden of Eden, may it last forever. A veritable earthy Paradise in all its qualities May Allah protect it from calamities. If it but heard the tale of this garden, Mecca would make the pilgrimage to Hindustan.60 Yet Delhi was not the sole claimant for divine favor as the urban Muslim capital of South Asia. "Even more than the Akhbar al-Akhyar, which is a Delhi-oriented work, Khazinat al-Asfiya is Lahore directed, including the entire region to the North and North west of Delhi."61 It was in the late 1800s that Mufti Ghulam Sarwar Lahori wrote his mas- sive and impressive Sufi tazkira, Khazinat al-Asfiya. It was a memorative communication that privileged Lahore over all other Indo-Muslim capi- tal cities, and in the postcolonial period Lahore acquired a renewed im- portance with the creation of Pakistani sacred space. Since Ajmer and, of course, Delhi remained within the Republic of India, the sacral role of Lahore became upgraded through its "patron," 'Ali Hujwiri, whose tomb, the Data Darbar, has been increasingly celebrated during the last half century. The Data Darbar underscores what has been hinted at but not devel- oped in the literary focus of this essay. Heightening the power of taz- kiras to both create and sustain Muslim cultural memory in urban South  168 | Indo-Persian Tazkiras as Memorative Communications Asia was the cemetery. All tazkiras took note not only of death dates but also of burial places, so that the symbolic resonance of the cemetery was crucial for Indo-Muslim urban identity. On the one hand, the tomb-cults were transitional spaces between the higher world and this one, but on the other hand, and with increasing emphasis, they were symbols of a distinctly Muslim identity in the Indian context (since Hindus cremate their dead). Graveyards as sites, then, are both a locus of inscription for local communal memory and the means of this inscription.62 Conclusions: Space and Identity Muslims over time imagined their space in South Asia differently as their sense of identity changed in the light of social and political devel- opment. This change may be traced in the organizing and structuring principles of the tazkira genre. The frame for this genre is memorialization, or better, memorative communication. One key element in this is inscription, which is done through the writing of memory on new spaces whose imagined shape is also subject to reconfiguration. Critical also in the South Asian tazkira tradition is the language of inscription, which serves to define a space even as it is the medium for writing it. In the course of this process, spaces have expanded from cities to regions to nations, while the principles of affiliation have loosened: no longer direct initiation, or even continuity in space and time, they have relied on a sense of "imagined community," as suggested by Benedict Anderson in his classic study of the construction of nationalist identi- ties.63 While the production of books generally encouraged a mnemonic reflex, it went well beyond memorializing dead heroes, whether poets or saints, in the Indo-Persian tazkira genre favored by South Asian ur- ban Muslim elites. Whatever their location or their authors' motivation, the premodern tazkiras laid a claim to Muslim space in South Asia. They did so by Islamicizing the soil, by creating a "new" home, by config- uring "new" spiritual and intellectual centers, and also by laying out "new" circuits of pilgrimage.64 The late Mughal and early modern period introduced a different tone in tazkira writing. While the urge to celebrate cities did not disappear, it appeared in a new guise. The desire to project memorative communica- tion that is felt in times of expansion takes a different turn in times of crisis or despair. Among Indo-Muslim versifiers it has been reflected in the laments over chaos in the poetic shahr ashob tradition.65  Marcia K. Hermansen and Bruce B. Lawrence | 169 Is ahd ko na janiyee agla sa ahd Mir voh daur ab nahin, voh zamin asman nahin These times are nothing compared to the old days, Mir; That age has passed, that heaven and earth are no longer. Mir Beyond Mir but through his agency, what we discover as the final accent on tazkiras is the possibility of mapping the altered sense and shape of urban Hindustan. Tazkira writers continue to project an inscribed space and identity in the colonial and post-independence periods of South Asian history, but the modern/postmodern reflex traces a more solemn sense of spatial orientation and organization of collective memory in contemporary tazkiras. The modern/postmodern space is one of aggressive retrieval of memory, for example, the proliferation of translations of old tazkiras from Persian into Urdu in Pakistan, as well as attempts to erase it. In the case of the poetic tazkiras, new canons of literary appreciation66 and even an altered mode of eloquent expression have rendered them obso- lete. In Ab-i Hayat, Azad mourns the fact that "the page of history would be turned-the old families destroyed, their offspring so ignorant that they would no longer know even their own family traditions." Pritchett observes, "The critical attitudes and vocabulary used by the tazkiras are all but unintelligible to most scholars-and in fact arouse considerable disdain."67 It is, above all, the threat of chaos68 that looms in the remarks of the late tazkira author Muhammad Din Kalim.69 In contrast to the hope of relocation, which marked many of the Indo-Persian tazkiras from Mug- hal India that we examined above, Kalim sees dislocation, even erasure, as the theme for his own memorative communication. Commenting on the contemporary situation in Lahore, he laments: Wherever you see an old grave, the keepers or greedy persons have spent quite a bit of money on fixing it up, popularizing it, and giving it some name which is unknown in the old sources so that they make it a means of earning money. [He then lists several such shrines saying, "God knows who is really buried there."] Nowadays the style of constructing new tombs has incorporated a lot of use of marble and other expensive stones and even the use of inlaid mirrors in some, so that you don't feel that you are in a graveyard but rather in a Shish Mahal. These tombs have prolifer- ated to the point that they are found in every lane, street, bazaar,  170 | Indo-Persian Tazkiras as Memorative Communications field, government park, and even in cinemas and government of- fices etc. even though there is no historical mention of them.... For some years I have been shocked by the lamentable situation that certain dissolute persons have pitched tents in the public grave- yards out of which they deal in drugs.70 In response, Kalim writes of the special features of his work. He per- sonally visited the shrines he writes about, he investigated the accurate names of the persons whom he mentions, and he reports the names of pirs falsely attributed to shrines when no such individuals were ever known to exist. Even while decrying "the lamentable situation" he con- fronts, he finds in the act of writing a recuperation of the past for the benefit of responsible mediation in the present; he remains a memor- ative communicator. And so there is a link between the oldest and the most recent phases of tazkira writing among South Asian urban Muslim elites. Kalim, as a contemporary tazkira writer, finds himself responding to an imminent threat of chaos, yet his remains a quest for the recovery of history, not a repetition of the past. Like the poet Ghalib, he struggles to understand how the act of erasure still retains its quality of a trace, a reminder, an emberlike hope. As Ghalib himself attests, with the fullness of his own sense of irony and place: ya rabb zamana mujh ko mitata hai kis li'e lauh-i jahan peh l.arf-i muqarrar nahin hun main O Lord, why is time erasing me? I am not a repeated letter on the tablet of this world. Whether one accents the trace that is never a repeated letter, or bemoans the self erased by time, one acknowledges in both cases the power of the Lord, the One who can both erase and re-create all that is. The poet, like the saint, locates his faith in language and in space. Ghalib implores the Omnipotent through Urdu (or Indo-Persian), and he implores Him from a familiar place, the still sacralized though much reduced Muslim space of Delhi. Memorative communication thus becomes more than a trace. It continues to embody hope; it projects the erased self in a reduced space as the servant before the Lord, the letter of Muslim identity etched not on the tablet of this world but on the Tablet that is both preserved and memorized, al-lauh al-mahfuz.  Marcia K. Hermansen and Bruce B. Lawrence | 171 Notes 1. M. Garcin de Tassy, Histoire de la litterature hindouie et hindoustanie (Paris: A. Labitte, 1870-71); Farman Fatel.puri, Nigar, 1964 Annual (Tazkirah Number). 2. George Morrison, Persian Literature from the Earliest Times to the Time of Jami (Leiden: E. J. Brill, 1980), 14. 3. Annemarie Schimmel comments on this and other couplets of Ghalib that are apposite to the present topic in her chapter "Poetry and Calligraphy," in A Dance of Sparks: Imagery of Fire in Ghalib's Poetry (Delhi: Vikas, 1979), 112-36. 4. Peter Osborne, The Politics of Time: Modernity and Avant-Garde (London: Verso, 1995), 179. There are other aspects of Benjamin's project, especially his dialectical sparring with Marx, Heidegger, and Freud, that lie beyond the scope of our article, but his accent on time as nonlinear, at once disruptive and recu- perative, is key to our own project. 5. The actual phrase "memorative communication" is coined by Peter Osborne in trying to distinguish the benefit of Benjamin's approach to tradition from the flawed approaches of Gadamer and Ricoeur. Insofar as "memorative communication" projects historical narrative in its living relationship to the present, the two categories elide, but memorative communication accents the role of the one who both remembers and uses memory to communicate; hence its special benefit for interpreting the catalytic role of the Indo-Persian tazkira. 6. Discussed, for example, in Jaroslav Stetkevych, The Zephyrs of Najd: The Poetics of Nostalgia in the Classical Arabian Nasib (Chicago: University of Chicago Press, 1993); and Michael Sells, Desert Tracings: Six Classical Arabian Odes (Middletown, Conn: Wesleyan University Press, 1989). 7. Noting here H. A. R. Gibb's statement that "the biographical dictionary is a wholly indigenous creation of the Islamic community": Gibb, "Islamic Bio- graphical Literature," in B. Lewis and P. M. Holt, Historians of the Middle East (Leiden: E. J. Brill, 1968), 54. Reflecting on the inclusion of very ordinary persons in the biographical dictionaries (tabaqat), Gibb further observes that the history of the Islamic community is essentially the contribution of individual men and women to the building up and transmission of its specific culture. That is, it is these persons (rather than the political governors) who represent or reflect the active force of Muslim society. 8. "Urdu literature, almost from its very beginnings, has been concerned with city life. The language has functioned for a long time as an urban-centered, but non-regional language": Leslie Fleming, "Two Pakistani Women Writers View the City: The Short Stories of Bano Qudsiyya and Farkhandah Lodhi," Journal of South Asian Literature 25, no. 1 (1990): 1, quoting A. K. Ramanujan, "Towards an Anthology of City Images," in Urban India: Society, Space, and Images, ed. Richard G. Fox (Durham, N.C.: Duke University Press, 1970), 224-44. 9. Bruce B. Lawrence, Notes from a Distant Flute: Sufi Literature in Pre-Mughal India (Tehran: Imperial Iranian Academy of Philosophy, 1978).  172 | Indo-Persian Tazkiras as Memorative Communications 10. Carl W. Ernst, Eternal Garden: Mysticism, History, and Politics at a South Asian Sufi Center (Albany: State University of New York Press, 1992), 62-84. 11. A discussion of the inscription of memory on nonwestern sites and the emotional and ephemeral qualities of memories inscribed there may be found in Susanne Kuchler, "Landscape as Memory: The Mapping of Process and Its Rep- resentation in a Melanesian Society," in Landscape Politics and Perspectives, ed. Barbara Bender (Oxford: Berg, 1996), 85-86. 12. Mary Carruthers, The Book of Memory: A Study of Memory in Medieval Cul- ture (Cambridge: Cambridge University Press, 1990). 13. Qur'an 85:22. Annemarie Schimmel's discussion of tablet and pen imag- ery in the Qur'an, hadith, and poetic tradition is quite helpful here: Mystical Dimensions of Islam (Chapel Hill: University of North Carolina Press, 1975), 414- 16. 14. For Ibn Arabi's influence in South Asia, see William Chittick, "Notes on Ibn al-Arabi's Influence in the Indian Sub-Continent," Muslim World 82 (July- October 1992): 218-41. 15. Schimmel, Mystical Dimensions of Islam, 414-15. This is the very couplet cited at the outset of the current chapter. 16. Carruthers, The Book of Memory, 21. 17. A work that stimulated scholarly awareness of technologies of memory was Francis Yates, The Art of Memory (London: Routledge and Kegan Paul, 1966). On the role of memory in a world of few books and the relationship of trained memory or "memoria" to literacy, see Carruthers, The Book of Memory, 7-15, passim. 18. Al-Suyuti, Alfiyya fi Ilm al-hadith (Cairo: al-Babi al-IHalabi, 1934). 19. Cited in J. M. S. Baljon, Religion and Thought of Shah Wali Allah of Delhi, 1703-1762 (Leiden: E. J. Brill, 1986), 12. 20. Qur'an 3:191. 21. Literally, "keeping in mind," or as one Sufi puts it, keeping the heart in the presence of God in all situations (yad means "memory"), this refers technically to one of the steps of practice in the Naqshbandiyya Sufi order that might be sum- marized as remaining aware at all times. See Fritz Meier, Zwei Abhandlungen aber die Naqshbandiyya (Istanbul: Franz Steiner, 1994), 44-46. 22. The term tasawwur evokes the element of visual memory. The Sufi practice of "tasawwur-i shaykh" is calling to mind the image of a person's spiritual mas- ter. It is also specifically related to the Naqshbandi practice known as "rabita," or developing a bond with the spiritual preceptor. Hamid Algar, "Devotional Prac- tices of the Khalidi Naqshbandis of Ottoman Turkey," in The Dervish Lodge: Ar- chitecture, Art, and Sufism in Ottoman Turkey, ed. Raymond Lifchez (Berkeley: University of California Press, 1992), 209-27. Tasawwur also means "representa- tion" in terms of the logic of formulating propositional statements, as opposed to a statement of "verification" (tardiq). See Roy Mottahedeh, The Mantle of the Prophet: Religion and Politics in Iran (New York: Simon and Schuster, 1985), 72-74, for a brief discussion of this system.  Marcia K. Hermansen and Bruce B. Lawrence | 173 23. These consist of ritualized recitations of rhymed spiritual genealogies of previous saints in a particular Sufi lineage. 24. Cited in V. G. Kiernan, Poems by Fai2 (London: George Allen and Unwin, 1971), 128-29. 25. Ernst, Eternal Garden, 67; Nizam ad-Din Awliya: Morals for the Heart, trans. Bruce B. Lawrence (New York: Paulist Press, 1992), 214. 26. Ernst, Eternal Garden, 69. 27. The phrase was used by Stetkevych in The Zephyrs of Najd, 121. 28. Farman Fatehpuri, Urdu Shuara k@ Tazkir aur Tazkira Nigari (Lahore: Majlis-i Taraqqi Urdu, 1972). An earlier work on this topic is Sayyid Muhammad Abdullah, Shuara-yi Urdu k@ Tazkire awr Tazkirah Nigarika Fann (Lahore: Maktaba Khiyabane-Adab, 1952). 29. Muhammad Awfi, Lubab al-Albab (Tehran: Kitabkhanah-yi Ibn Sina, 1957), partially translated in R. A. Nicholson, Studies in Islamic Poetry (Cambridge: Cambridge University Press, 1969). Fatehpuri, Nigar, 9-11. 30. Edward G. Browne, ed., Lubab al-Albab, vol. 2 (London: Luzac, 1906), 12. 31. Fatehpuri, Nigar, 14-15. They were Nakat ash-Shuara of Mir Taqi, Gulshan- i Goftar of IHamid Aurangabadi, and Tuhifat ash-Shi' r of Afcjal Beg Qaqshal. The latter two were produced in the Deccan, but they mentioned northern poets as well. 32. Muhammad Husain Azad, Ab-i Hayat (Lucknow: Uttar Pradesh Urdu Academy, 1982). Reprint of the 1907 Newal Kishore edition. Azad as a critic is discussed in Frances Pritchett, Nets of Awareness: Urdu Poetry and Its Critics (Ber- keley: University of California Press, 1994), 46-59. 33. Fatehpuri, Nigar, 7. A review of previous tazkiras of poets and their pos- sible influence on Azad is Aslam Farrukhi, Muhammad Husayn Azad: Hayat awr Tasanif (Karachi: Anjuman-i Taraqqi Urdu, 1965), 2:28-58. 34. Ralph Russell, The Pursuit of Urdu Literature (London: Zed Books, 1992), 121-22. 35. Abd al-Sattar Siddikqi, Oriental College Magazine, 1927, and Gulchin-i- Ma'ani, Ahmad, Tarikh-i-Tazkiraha-yi Farsi (Tehran, 1970). 36. Fatehpuri, Nigar, 19. 37. Garcin de Tassy, Histoire de la litterature hindouie et hindoustanie, 57. 38. Pritchett describes "bayaz" as "the ubiquitous little notebook that lovers of poetry carried around with them for recording verses that caught their fancy": Nets of Awareness, 66. 39. These collections are known as guldastas (bouquets). On the guldasta see Pritchett, Nets of Awareness, 74. On musha'iras see Pritchett or Carla Petievich, Assembly of Rivals: Delhi, Lucknow, and the Urdu Ghazal (Delhi: Manohar, 1992). 40. Fatehpuri, Nigar, 17-21. 41. Muhammad Sadiq, History of Urdu Literature, 2d ed. (Delhi: Oxford, 1984), 40. 42. Petievich, Assembly of Rivals, 204. 43. Pritchett also notes the role of tazkiras as being "the most important genre of literary record and commentary that existed in Urdu": Nets of Awareness, 74.  174 | Indo-Persian Tazkiras as Memorative Communications 44. Two works specifically organized according to this paradigm are Nur ul IHasan Hashmi, Dilli ka Dabistan-i Sha'iri (Delhi school of poetry) (Lucknow: Uttar Pradesh Urdu Academy, 1971) and A. L. Siddiqi, Lakhnau ka Dabistan-i Sha 'iri (Lucknow school of poetry) (Lahore: Urdu Markaz, 1955). The paradigm was criticized by Ali Javad Zaidi, Do Adabi Iskul (Two literary schools) (Luck- now: U. P. Urdu Academy, 1970), a work which is examined in Petievich, Assem- bly of Rivals, 89-99. 45. Petievich, Assembly of Rivals, ix, xiii. 46. Ibid., 13-15. 47. Hafiz Al.mad Ali Khan Shauq, Kamilan-i Rampur (Patna/Delhi: Khuda Bakhsh Oriental Public Library, 1986). 48. Tajammul Hussain Khan, Urdu Sha'iri ka tisra Skul (Karachi: Jauhar Acad- emy, 1976); Kalb-i Ali Khan Fa'iq, "Rampur ka Shaciri Skul," in Ma'arif, October 1955, 285-94. 49. Abd al-Ral.man Jami, Nafahat al-uns min ha2arat al-quds (Tehran: Intisharat- i ittilacat, 1994). 50. Jo-Ann Gross, "Khoja Ahrar: A Study of the Perceptions of Religious Power and Prestige in the Late Timurid Period," Ph.D. diss., New York Univer- sity, 1982, discusses the tazkira of Ali ibn al-IHusayn Kashifi: Rashahat (Lucknow: Newal Kishore, 1890). 51. Ernst, Eternal Garden, 89-90. 52. Abd al-Qadir al-Badauni's Muntakhab at-tawarikh (Summation of histo- ries) was translated into English at the turn of the century; see G. S. A. Ranking (vol. 1), W. H. Lowe (vol. 2), and Wolseley Haig (vol. 3) (Calcutta: Asiatic Society of Bengal, 1884-1925; reprint, Delhi: Idarat-Adabiyat-i Delli, 1973). Although volume 3, with which we are concerned here, abounds with infelicities, it is still remarkable to have an English version of this maverick biographical review of Sufi masters provided by Badauni. 53. The problems of this lineage have been traced with singular clarity and characteristic understatement by Simon Digby in "'Abd al-Quddus Gangohi (1456-1537 A.D.): The Personality and Attitudes of a Medieval Indian Sufi," Me- dieval India-A Miscellany 3 (Aligarh, 1975): 4-5. 54. See Badauni, Muntakhab at-tawarikh, 3:535-36. 55. The Shaykh's experience at the Mughal court was interrupted by a five- year absence (1587-92), most of it spent on a "pilgrimage of penance" to the Hijaz, where he furthered his own studies of hadith with the Indian alim in exile, Shaykh Abd al-Wahhab Muttaqi, who also receives prominent attention in Akhbar al-akhyar. See N. H. Zaidi, "Abd al-Haqq Mohaddeth Dehlavi," in Encyclopaedia Iranica (London: Routledge and Kegan Paul, 1982): 1/2: 113-14. 56. Ernst, Eternal Garden, 90. 57. For example, the extensive tradition of compiling gazetteers of regions under British control, studied by Robert C. Emmett, "The Gazetteers of India," M.A. thesis, University of Chicago, 1976; Henry Scholberg, The District Gazet- teers of British India (Zug: Inter Documentation, 1970). 58. Lawrence, Notes from a Distant Flute, 20.  Marcia K. Hermansen and Bruce B. Lawrence | 175 59. Muhammad Sadiq Dihlavi Kashmiri Hamadani, Kalimat al-Sadiqin, ed. Muhammad Saleem Akhtar (Islamabad: Iran Pakistan Research Center, 1988). 60. Ibid., 5. Amir Khusrau, Qiran al-Sa'dayn (Lucknow: Newal Kishore, 1875), 22-23, is the mathnavi from which these verses are taken. Other features of Delhi, such as the Friday mosque and minaret, are also praised in the same section. 61. Bruce Lawrence, "Biography and the Seventeenth-Century Qadiriyya of North India," in Islam and Indian Regions, ed. Anna Libera Dallapiccola and Stephanie Zengel-Ave Lallemant (Stuttgart: Steiner, 1993), 399-415, 402. The entire chapter has developed some of the points of this essay, but without an appreciation of the crucial role of place, an appreciation that only emerged later, due to the Rockefeller Residency Program on "South Asian Islam and the Greater Muslim World, 1993-1997," convened through the Triangle South Asian Colloquium. 62. A further dimension of the relationship between gardens, tombs, and Is- lamic cosmology has been explored by art historian Wayne Begley in "The Myth of the Taj Mahal and a New Theory of Its Symbolic Meaning," Art Bulletin, no. 61 (March 1979): 7-37. 63. Benedict Anderson, Imagined Communities: Reflections on the Origin and Spread of Nationalism (London: Verso, 1991). 64. The incorporation of such pilgrimage circuits into calendars of ritual ob- servances of saints' anniversaries is discussed in Carl W. Ernst, "An Indo-Per- sian Guide to Sufi Shrine Pilgrimage," in Manifestations of Sainthood in Islam, ed. Grace Martin Smith and Carl Ernst (Istanbul: Isis, 1993), 43-68. "Whenever one comes to a town, the first thing one has to accomplish is to kiss the feet of the saints who are full of life, and after that, the honor of pilgrimage to the tombs of saints found there. If one's master's tomb is in that city, one first carries out the pilgrimage to him; otherwise one visits the tomb of every saint shown him" (61). Quoted from Simnani, Laa'if al-Ashrafi. 65. Carla Petievich, "Poetry of the Declining Mughals: The Shahr Ashob," Journal of South Asian Literature 25, no. 1 (1990): 99-110. 66. Gauri Vishvanathan, Masks of Conquest: Literary Study and British Rule in India (New York: Columbia University Press, 1989); Sara Suleri, The Rhetoric of English India (Chicago: University of Chicago Press, 1992). 67. Pritchett, Nets of Awareness, 75. 68. In his study of "place" in religion, Jonathan Z. Smith writes concerning the Jewish and Christian understandings of sacred centers in Jerusalem, "For each there was a triumphant, ideological literature that perceived in their con- struction a cosmogonic act. For each, there was a literature of indigenous lamen- tation ... that found, in the destruction or loss of the sites, a plunge into chaos." Smith, To Take Place: Toward Theory in Ritual (Chicago: University of Chicago Press, 1987), 3. 69. For Kalim's biography see Narr Iqbal Qureishi, "Mu'arrikh-i Lahur Mian Muhammad Din Kalim Qadiri," "Arafat, Lahore (November/December 1989): 128-38. 70. Muhammad Din Kalim, Madinat al-Auliya (Lahore: Ma'arif, 1982), 78-79.  7 The "Naqshbandi Reaction" Reconsidered David W. Damrel The event of the "Naqshbandi reaction" has occupied a special place for a generation of historians of religion seeking an explanatory paradigm for the history of Islam in early seventeenth-century South Asia. There is no question that at the center of this "reaction" stands the prominent Naqshbandi scholar and sufi Shaykh Almad Sirhindi, the Mujaddid-i alf- i thani, "renewer of the second millennium" (d. 1624). But the nature of his movement, its long-term effects, and even the "crisis" that initially prompted the reaction remain subjects of intense controversy. Aziz Ahmad coined the phrase "Naqshbandi reaction" over thirty years ago in his well-known Studies in Islamic Culture in the Indian Envi- ronment. There he argued that the Mughal emperor Akbar's neglect of Islam and experiments with "imperial heresy" spurred Sirhindi and the Naqshbandis to answer with an orthodox reformation of Sufism de- signed to promote the general "rehabilitation of Islam in India."1 Fazlur Rahman surely had this in mind when he suggested that "the Sirhindi- led movement was a successful reaction" against Akbar's much- discussed Din-i Ildhi. It was also effective in "counteracting the antinom- ian tendencies" within Indian Sufism.2 Annemarie Schimmel agreed that the movement was the struggle of the Naqshbandi order against "Ak- bar's syncretism and against the representatives of emotional Sufism." Expanding on this theme, she asserted that the later Naqshbandiyya played "a remarkable role in all parts of Muslim India," especially as a "defense against syncretism."3 It is clear that this issue of syncretism explicitly and implicitly in- forms much of the writing about the Naqshbandi reaction. In seeking the spiritual and intellectual roots of the movement, many scholars in- terpret the dramatic rise of the Naqshbandi line at the start of the seven-  David W. Damrel | 177 teenth century as a critical moment in a long Indian encounter between two competing Islamic mystical traditions: wahdat al-wujad (unity of being) and wahdat al-shuhad (unity of appearance).4 When paired as op- posites, the former esoteric philosophy typically is seen as a door to religious syncretism, while the latter is seen as a defense against syn- cretic interaction. Fazlur Rahman describes the concept of wahdat al- wujud formulated by Ibn al-'Arabi (d. 1240) as "pantheistic mysticism" and argues that when these ideas were introduced into South Asia they readily found "a strong ally in the Vedantism of orthodox Hinduism."5 The perceived mutual affinities between these two doctrines-wahdat al-wujud and Vedantism-created an arena for syncretistic interaction between mystically minded Muslims and Hindus. Aziz Ahmad argued that Sirhindi feared the syncretism that pantheistic "heterodox Sufism" invited; these "syncretisms" actually threatened "the disintegration of Islam in India and its gradual absorption into Hinduism."6 In this view, Sirhindi's revitalized doctrine of wahdat al-shuhud presented a defen- sive mystical barrier, even a corrective, to such exchanges between Hindu and Muslim "spiritual athletes." Driven by such concerns, inter- pretation of the Naqshbandi reaction suddenly involves much more than the dispute between the Mughal Padishahs and a Sufi order over religious practice at court. It becomes a battle between syncretism and exclusivism, religious tolerance and intolerance, and, for some, nothing less than the defining moment in the course of Hindu-Muslim relations to this day.7 Our goal is not to revisit these controversies-which comprise for some the central elements of the Naqshbandi reaction-but rather to ask a question more closely related to the issue of Muslim identity in South Asia. Simply put, what are the discernible Naqshbandi elements in Sir- hindi's "Naqshbandi reaction"? Can the actions that he took and the attitudes that he held unquestioningly be ascribed to his Naqshbandi persona and not to other spiritual affiliations and religious influences on him? In particular, what effect, if any, did his Chishti affiliations play in the worldview that he developed in his writings and praxis? If, painted in such broad strokes, the Naqshbandi-Mujaddidi order that crystallized around Shaykh Sirhindi is seen as the great champion of wahdat al-shuhud, then the long-established Chishti order must be counted to represent the alternative view, that of wahdat al-wujad.8 Notable even in a long tradition of so-called wujadi, Chishti shaykhs, the famed Chishti-Sabiri Sufi Shaykh 'Abd al-Quddns Gangohi (d. 1537) in particular has long been recognized as "a vigorous advocate of the doc-  178 I The "Naqshbandi Reaction" Reconsidered trine of wahdat al-wujad."9 Further, Shaykh 'Abd al-Quddus is also identified as a leading Indian Muslim syncretist between Islamic mysti- cism and Nathapanthi Yogic traditions.10 At one level then, Shaykh 'Abd al-Quddus, the ecstatic wujadi Chishti syncretist, could easily be con- strued as the antithesis of Shaykh Ahmad Sirhindi, the sober shuhadi Naqshbandi exclusivist. But, as will be explored below, the lives of these two medieval North Indian Sufis contain similarities and convergences that challenge such simplistic constructions of identity and offer insight into the more complex configuration of Sufi identity and praxis in the first century of Mughal rule in India. This chapter explores certain elements in the identities of these two Sufi shaykhs and attempts to demonstrate a thread of continuity in their actions, attitudes, and behavior that undermines the exaggerated con- flicting identities often ascribed to them in later literature. In particular, their attitudes toward Hindus and Hindus in state service will be dis- cussed, as well as certain elements of each man's spiritual beliefs and practices. Shaykh Qutb al-'Alam' Abd al-Quddus b. Isma'il b. Safi al-din Hanafi Gangohi was born in 1456 in Rudauli, now a village in modern Uttar Pradesh. Born into a family of ulema, 'Abd al-Quddus abandoned an agenda of formal studies early on in favor of spiritual pursuits. When he was still quite young, he presented himself to the sajjada-nishin at the khanaqah of Shaykh Ahmad 'Abd al-Haqq Rudaulvi (d. 1434) for mysti- cal instruction. This sajjada-nishin was Shaykh Muhammad, son of Shaykh 'Xrif and grandson of Shaykh 'Abd al-Haqq. This Chishti family represented what became known as the Chishti-Sabiri line, so named because of their spiritual descent from Shaykh 'Ala' al-din 'Ali b. Ahmad Sabir (d. 1291 at Kalyar). Shaykh' Ala' al-din' Ali b. Ahmad Sabir was a khalifa of the prominent Chishti pir Shaykh Farid al-din Ganj-i Shakar (d. 1265 at Pakpattan).11 Shaykh 'Abd al-Quddus nominally accepted Shaykh Muhammad as his pir, but he apparently received much of his spiritual instruction from Shaykh Piyara, an elder trusted companion of Shaykh 'Xrif. Moreover, in an episode that hints of a strong Uwaysi influence, 'Abd al-Quddus also claimed to have received "spiritual grace" directly from the rah of the deceased Shaykh 'Abd al-Haqq.12 These spiritual ties were aug- mented with family links following his marriage to Shaykh Muham- mad's sister, the granddaughter of Shaykh 'Abd al-Haqq. In Rudauli, Shaykh 'Abd al-Quddus practiced the rigorous devotions for which he later became famous, including namaz-i ma 'Cs ("inverted  David W. Damrel | 179 prayer," that is, prayer performed while suspended upside-down) and gained a reputation for asceticism and mystical intoxication. He also developed a substantial following among the Afghan soldiers posted to the Awadh region with the Lodi armies. Lodi conflict with the Sharqi dynasty in Jaunpur through the last half of the fifteenth century caused periodic devastation in Rudauli and, in 1489, the town briefly fell into the hands of the Bachgoti Rajputs. Two years later 'Abd al-Quddus ac- cepted the invitation of one of his prominent Afghan disciples and left Rudauli entirely, moving himself and his family almost five hundred miles north and west to Shahabad, in eastern Punjab, near Gangoh. He remained in Shahabad for over thirty years, strengthening his ties with the Lodi Afghan nobility while writing and instructing his numerous disciples. 'Abd al-Quddns's intimate links with the Afghans served him poorly in the tense years surrounding Babur's initial Mughal incursions against Sultan Ibrahim Lodi. Internal dissent led some Lodi nobles in the Punjab to contact Babur in Kabul, and the Mughal conquest of northern India began in earnest late in 1525. The Mughal invasion must have been widely anticipated in the Punjab, for even before Babur's army actually advanced, 'Abd al-Quddus relocated to Gangoh, a village some forty miles from Shahabad and presumably safe from any army marching on Delhi. In 1526, as the Mughal and Lodi armies assembled for what would be the decisive battle at Panipat, 'Abd al-Quddus and his family attempted to flee. But at Kutana, near Panipat, Sultan Ibrahim Lodi pre- vailed on the elderly shaykh to remain with his army and provide spiri- tual support. 'Abd al-Quddus did so, but he also sent his family south. Following the Lodi disaster at Panipat-in which Sultan Ibrahim him- self was killed-Shaykh 'Abd al-Quddus was captured by the Mughals and marched forty miles to Delhi. The length of his captivity in Delhi is unspecified, although it was probably brief. Released in Delhi, 'Abd al-Quddus returned to Gangoh, where he remained until his death in 1537. During this period of early Mughal rule,' Abd al-Quddus retained contact with his defeated Afghan disciples and cultivated respectful if spare relations with Babur and then Humaynn. These relations will be discussed at length below. Shaykh 'Abd al-Quddns's writings are many and diverse. His most important works include Rushd-Nama (Murshid-Nama), a work in Per- sian for the spiritual preparation of his disciples that contains frequent and detailed allusions to Yogic practice; a no longer extant commentary on Ibn al-'Arabi's Fusas al-Hikam; and an Arabic commentary on Shihab  180 | The "Naqshbandi Reaction" Reconsidered al-din Suhrawardi's 'Awdrif al-Ma'drif. There are also two collections of correspondence, the minor Muntakhab-i Maktat-i Quddasi and the much more important Maktat-i Quddasi.s3 Recently, I. H. Siddiqui an- nounced the discovery of another work by 'Abd al-Quddus, entitled Sharh-i Risala-i Lama'at.14 This work is a commentary in Persian on the Lama'at of the widely traveled, ecstatic wujadi Suhrawardi mystic Shaykh Fakhr al-din 'Iraqi (d. 1289). 'Abd al-Quddus also produced a significant body of verse, including a partial translation into Persian of an Awadhi romance and, in Hindi, several verses that survive mainly as marginalia in the Rushd-Nama. In addition, his son and sajjada-nishin Shaykh Rukn al-din (d. 1576) left the Lata 'if-i Quddasi, a compilation of richly detailed biographical anecdotes about 'Abd al-Quddus that still serves as the best, if incomplete, introduction to the shaykh's life.15 Despite his distinctive and innovative spiritual practices, it is impor- tant to note that Shaykh 'Abd al-Quddus considered himself (and was considered by others) to be impeccable in his observance of the shari'a. As Shaykh Rukn al-din declares, perhaps defensively, "It is apparent that in conforming to the example of the Prophet and in the observance of the shar' of the Prophet he was so strict that he did not allow the most minute departure from it to be permissible in either exterior or interior matters-as regards himself or others. If he got to know of any departure from the shar' by anyone he showed dissatisfaction and wished to avoid his company." Shaykh Rukn al-din adds that Shaykh 'Abd al-Quddus, despite associating "with all kinds of men," was unaffected by "com- pany which was hostile to the faith" and in fact helped return his asso- ciates to "the narrow path" through his companionship.16 'Abd al-Quddns's important contact with Yogic thought and practice should also be noted. In his various techniques designed to produce mystical ecstasy, particularly involving breathing exercises, he drew upon Nathapanthi Yogic traditions; in explaining these practices he employed spiritual-physiological constructs that also reveal Yogic ori- gins. He had a lifelong enthusiasm for Hindi love poetry (which often produced ecstatic mystical states in him), and he produced some verse in Hindi under the Hindi pen name Alakhdas, "servant of the Invis- ible."17 These borrowings and inspirations from Yogic tradition have earned him a lasting name as a leading syncretist between Hindu and Muslim esotericism.18 In sum, Shaykh 'Abd al-Quddus was an influential North Indian Sufi who left a wide range of writings and numerous initiates. He instilled new vigor into what later became the powerful Sabiri line of the Chishti  David W. Damrel | 181 order, and his life spanned the political transition from the last of the Delhi Sultanates, the LOdis, through the early years of Mughal rule. A staunch supporter of wahdat al-wujud with an inclination for ecstasy, he was also well versed in the intimacies of Yogic practices. The striking later success of the Chishtiyya-Sabiri lineage after Shaykh 'Abd al- Quddns need not detain us here, except to note two of his key spiritual descendants. The first was his principal khalifa, the famous Shaykh Jalal al-din Thaneswari (d. 1582), who gained prominence in Akbar's reign. A chain of Chishti-Sabiri initiates from Jalal al-din onward has proved im- portant in South Asian Muslim circles ever since.19 The second impor- tant spiritual descendant was a young Sufi who first met 'Abd al- Quddas while the great pir resided at Gangoh. The novice learned dhikr from the shaykh but actually took bay at into the Chishti-Sabiri line from Shaykh Rukn al-din, 'Abd al-Quddns's son and sajjada-nishin. This ini- tiate, named Shaykh 'Abd al-Ahad (d. 1599), then settled in Sirhind to a life of scholarship and piety. Shaykh 'Abd al-Ahad is, of course, the father of Shaykh Almad Sirhindi. He is also, significantly, Sirhindi's pir in the Qadiri and the Chishti-Sabiri turuq. Sirhindi, born in 1564 in the same Punjabi village where his father had settled, is too well known to deserve more than passing treatment here. However, certain aspects of his life and career require more critical attention in light of his early and persistent connec- tions with the Chishti-Sabiri order. From his early religious and spiritual education in Sirhind, and through his further studies of hadith and fiqh in Sialkot, his brief and much debated tenure at Akbar's court in Agra, and his fateful trip to Delhi in 1599, Sirhindi followed the path of a Chishti Sufi. More than that, in his later years he portrayed himself in this period of his life as an ecstatic mystic who, under the influence of wahdat al-wujud, could not distinguish Islam from infidelity.20 And, although his career as a Chishti pir was brief, he did take over the instruction of several of his aging father's disciples, and he actually enrolled several Chishti disciples of his own. At one point he refers to specific practices he learned from Chishti mentors: "I also developed a taste for supererogatory works (nawifil), particularly nafl prayers, from my father, who got it from his teacher, a Chishti saint."21 The sources note that even while he was an active Chishti pir, Sirhindi chose not to participate in the sama'C assem- blies that are so closely identified with Chishti practice.22 Like many of the ulema in North India at the time, Sirhindi was initiated into several orders-the Chishti, Qadiri, and, by one account, the Suhrawardi-but  182 | The "Naqshbandi Reaction" Reconsidered it is clear that, by choice and family tradition, his Chishti affiliation was paramount. The change came in 1599, less than half a year after the death of his father. That year Sirhindi undertook the Hajj and visited Delhi en route. There he met the itinerant Naqshbandi Sufi Khwaja Baqi Billah Birangi (d. 1603). Following an intense three-month discipleship, he abandoned his pilgrimage and returned home to Sirhind as Baqi Billah's Naqshbandi khalifa. For the next two decades he established himself as a Naqshbandi shaykh, mostly through a substantial literary output that included several short treatises and his famous correspondence, more than five hundred letters collected in three volumes known as the Maktabat-i Imam Rabbani. In addition, anecdotal literature about Sirhindi appeared early and continued well after his death, devoting special at- tention to his miraculous works and political activities.23 Based in Sirhind, the Mujaddid-i alf-i thani oversaw a network of khulafa' that, according to the Mughal emperor Jahangir, was active in every town and city in the empire.24 The critical events of his public life-such as his precise role in the accession of Jahangir in 1605, his spiritual impact on the nobles at the Mughal court, his brief imprison- ment by Jahangir in 1619, and his subsequent influence on that em- peror-still spark controversy and debate. More important for our theme is the emergence of his religious persona and the development of his style of mystical practice. That Sirhindi asserted and maintained his Chishti ties even after he became a Naqshbandi is quite clear in his Maktabat, and from the follow- ing it is also clear that he ranked his Naqshbandi affiliation above his membership in both the Qadiri and the Chishti: "I am a disciple of Muhammad connected with him through many intermediaries: in the Naqshbandi order there are twenty-one intermediaries in between; in the Qadiri, twenty-five; and in the Chishti, twenty-seven; but my rela- tionship to God as a disciple is not subject to any mediation, as has al- ready been related."25 There are a number of important points on which Shaykh Almad Sirhindi and his Chishti-Sabiri predecessor Shaykh 'Abd al-Quddus Gangohi show remarkable similarity and continuity. The first issue that we will examine is the nature of the correspondence each mystic con- ducted with the ruling houses of his day. Shaykh 'Abd al-Quddus Gangohi wrote to various nobles, at first the Lodi and then the Mughal courts. He also addressed letters directly to  David W. Damrel | 183 Sikandar Lodi, Babur, and Humaynn during their respective reigns.26 The earliest of these letters was an appeal to Sikandar Lodi (reigned 1489-1517). Although he praised his Chishti-Sabiri predecessor Shaykh 'Abd al-Haqq Rudaulvi for a pointed indifference toward political af- fairs and patronage,' Abd al-Quddus himself is conspicuous for his will- ingness to petition the imperial court for financial relief. His letter to Sikandar Lodi warns of the calamities that will befall the state because of his cancellation of permanent stipends (waza'if) to the Muslim religious elite.27 This letter, stressing how generosity to the ulema and Sufis strengthens an empire, is a plea for the emperor to restore his monetary support for the Muslim religious classes; if they are neglected, 'Abd al- Quddus obliquely warns, there will be a cry for redress.28 The emperor's reaction to this letter is unrecorded, but 'Abd al-Quddns's interest in matters that are explicitly financial is notable. In a letter to Babur soon after his accession in 1526, 'Abd al-Quddus reiterates this emphasis on the imperial obligation to nurture the fuqard i, the ulema, and the masha'ikh. In this epistle his immediate aim is to end the imposition of 'ushr, a tithe that Babur had enacted on the revenue- producing lands that supported the various Muslim religious classes. And again 'Abd al-Quddus warns Babur of the danger of a "cry for redress" from the fugara' if their needs are ignored.29 This letter is interesting in many ways, for it marks the Chishti recog- nition of a new, non-Indian Muslim dynasty that would have had little special reason to honor the Chishti order. The Mughals, as K. A. Nizami has pointed out, from Timur downwards maintained "an unbroken tra- dition of respect, devotion and attachment to the Naqshbandi saints."30 Babur certainly demonstrates this attachment, and while the Baburnama is full of repeated and copious notice of the Naqshbandi order and Babur's own affiliations with Naqshbandi pirs, there is no mention in the work of the Chishti silsila at all.31 Indeed, Babur's sole knowledge of the Chishti silsila might well have been that members of the order-and onetime Mughal captive Shaykh 'Abd al-Quddus himself-had in fact supported the Lodis against him. But neither of these experienced politi- cal figures was willing to let the circumstances of conquest interfere with the expediencies of rule: 'Abd al-Quddus noted with approval Babur's early provisions to sponsor the religious classes, and Babur in turn was generous to the religious elite he inherited from Sultan Ibrahim LOdi.32 After noting his concerns about the burden of 'ushr on the religious classes (ta 'ifa-i 'ulama'), 'Abd al-Quddus then offers Babur advice on  184 I The "Naqshbandi Reaction" Reconsidered how he should rule in India. Much of his counsel, in the tradition of Islamic polity, is standard and seemingly formulaic: royal justice should prevail, the army should be firm in its attachment to the shari'a, muhtasibs should be appointed to inspect the bazaars and enforce order, congregational prayer should be enjoined, and the state should support the people of learning and faith. 'Abd al-Quddus also details who should serve in the Muslim govern- ment in India. Muslims "of pure and zealous faith" should occupy posts of authority in the countryside. No non-Muslim (he uses the term kafir) may serve in a Muslim administration or receive an assignment of rev- enue. Non-Muslims should not be employed in bureaucratic offices, and they should not be tax collectors or local commanders. They should be forced to pay jizya and should not be allowed to dress like Muslims or to practice their faith ostentatiously and publicly.33