Stereotype, n., fig. (Greek στερεό-ς solid + τύπος type)
a. Something continued or constantly repeated without change….
———Oxford English Dictionary, 2d ed. (1989).In the debate on “colonial knowledge” the argument for extreme rupture should have already run its course. If two decades back historians could confidently assert that what colonial rulers thought they knew about their subjects was mere imperialist myth, historically valuable only in reflecting colonial attitudes and power relations (e.g., Dirks Reference Dirks1987; Inden Reference Inden1990; Greenblatt Reference Greenblatt1991), few nowadays dare to venture this claim unqualified. Back then, when Edward Saïd's portentous Orientalism (Reference Saïd1978) still offered fresh inspiration, historians set out to expose the archives of European colonial powers for what Saïd said they really were: convenient fictions about the colonized Other. India's historians had at their disposal the world's richest colonial archive, full of startling claims about local society, and they became the noisiest advocates of this view.Footnote 1 From the mid-1980s, one social category and institution after another—caste, tribe, Hindu, Muslim, martial race, human sacrifice, the immolation of widows, and so on—was “deconstructed” until all colonial sociology appeared but an Orientalist ruse.Footnote 2 These “myths” were not idle fancies or misapprehensions, critics argued, but served the political and fiscal exigencies of the empire. As the movement's frontrunner, Nicholas Dirks, put it, “Colonial knowledge both enabled conquest and was produced by it; in certain important ways, knowledge was what colonialism was all about” (1996: ix). Colonial knowledge was somehow exceptionally political, more so than any other knowledge. In fact, it was a mere symptom of the will to rule. Colonial laws, ethnographic compendia, census reports, and even the study of Indian languages were instruments of “epistemic violence,” which glossed over and grossly disfigured the complex realities of local social life, reducing native identities to timeless essences like “caste” and “tribe.” Ronald Inden wrote that “Orientalists” imagined “an India kept eternally ancient by various Essences,” which were put to “invidious uses … to constitute the European world's Others” (1990: 1–2). For another critic, British officialdom “reduce[d] the natives to their racial essences to suit the exigencies of colonialism. Inherent in the production of this knowledge was the notion of the essential type, an object without history” (Nigam Reference Nigam1990a: 133).
Since then, this picture has grown more nuanced. Postcolonialism has been exhaustively criticized for ignoring the roles of indigenous actors, homogenizing “Orientalism,” and short-circuiting the relation of knowledge and power under colonial rule.Footnote 3 Historians showed that the aims and experiences of “Orientalists” often varied and sometimes conflicted sharply (e.g., Parry Reference Parry2004), that colonialists often drew extensively on indigenous sources, and that the production of “colonial knowledge” often involved a variety of Indian actors.Footnote 4 They showed that colonialists absorbed many indigenous concepts, practices, and institutions,Footnote 5 and that some ideas even flowed the other way, from colonies to European metropoles.Footnote 6 Many empirical infelicities also came to light. The references in Dirks' influential The Hollow Crown (1987) conflicted with cited sources, claims were unsupported by references, and few precolonial and non-European materials were consulted at all (see Guha Reference Guha2003).Footnote 7 It turned out that the caste-wise enumeration of population (the census), the empire's allegedly chief “technology of power” (Cohn Reference Cohn and Cohn1987; Appadurai Reference Appadurai, Breckenridge and van der Veer1993), was already used by Indian rulers in the late seventeenth century, over two hundred years before the first British census report (Peabody Reference Peabody2001). Communal conflict between Hindus and Muslims, staked as iconic artifact of colonial rule, was also there well before the British arrived (Bayly Reference Bayly1985; Talbot Reference Talbot1995).
Yet despite many devastating critiques, the force field of this “invention tradition” (Desai Reference Desai1993) has proven impossible to escape. By now, the claims of postcolonialist history seem so obvious, so deeply rooted in the historians' “unconscious” (Lazarus Reference Lazarus2011), that even critics do not question them. We take it for granted that studying colonial history requires us to focus on how its various aspects were “constructed,” “imagined,” or “re-imagined.”Footnote 8 And one still often comes across journal article titles like “Census in Colonial India and the Birth of Caste” (Samarendra Reference Samarendra2011). We assume that colonial sources and personages must be approached with special suspicion, as perpetrators of textual construction, epistemic violation, and misrepresentation of indigenous precursors, which were presumably less artful, fantastical, or politically fraught. More than three decades on, the specter of Orientalism is still with us. Only three years ago two historians tellingly felt the need to argue that it is possible to take colonial sources seriously, as more than insights into imperial psyche (Roque and Wagner Reference Roque and Wagner2012).Footnote 9 That Vivek Chibber's (Reference Chibber2013) recent criticism of postcolonial theory should have provoked so much bile and frenzy revealed once again that the movement's polemical value is yet to be outlived.Footnote 10
My aim here is not to examine the broader politics and dynamics that sustain this momentum. It is much more modest. I turn to the history of one presumptive colonial stereotype—the “criminal tribe”—which I show was a label of much older vintage on the Subcontinent. I show that while colonial uses of the stereotype add up to a lurid history of violence against people branded as congenital criminals in colonial law, the stereotype itself has a history stretching back far beyond British colonialism. My historical sketch points to two major errors in the historians' standard approach, one empirical and one analytic. The large and readily accessible body of precolonial references to robber castes, which I have space and competence enough to outline only roughly here, reveals a large blind spot in postcolonial historiography, an oversight that results from confusions about the nature of stereotypes and the relation of power to knowledge more broadly, problems to which I will return.
THE CRIMINAL TRIBE IN BRITISH INDIA
The criminal tribesman cuts an exotic figure with pride of place in the pageant of “Orientalist stereotypes.”Footnote 11 Colonial sources described criminal tribes as guilds of felons committed by upbringing and blood to the robber's trade; they had their own morality, modus operandi, and divine sanction, differing from Barbers or Brahmans only insofar as their trade required them to break the law. In the early 1870s, legal member of the Viceroy's Council, J. F. Stephen, described Criminal Tribes like this:
The caste system is India's distinguishing trait. By virtue of this system, merchants are constituted in a caste, a family of carpenters will remain a family of carpenters for a whole century from now, or five centuries from now, if it survives that long. Let us bear that in mind and grasp quickly what we mean here by professional criminals. We are dealing here with a tribe whose ancestors have been criminals since the very dawn of time, whose members are sworn by the laws of their caste to commit crime … for it is his vocation, his caste, I would go to the extent of saying his faith, to commit crimes (from Fourcade Reference Fourcade, Vidal, Tarabout and Meyer2003: 146).
By 1871 this stereotype was formalized in the Criminal Tribes Act (Regulation XXVII of 1871), aimed to “control and reclaim” communities “addicted to the systematic commission of non-bailable offences.” The idea of a congenital criminal had mileage. In the nineteenth and twentieth centuries descriptions of the criminals' secret argots, omens, customs, ordeals, and bloodthirsty sacrifices filled novels, police reports, and voluminous caste compendia, reaching us in Stephen Spielberg's 1984 film Indiana Jones and the Temple of Doom.Footnote 12
This imagery strikes lay Westerners as a product of an overheated imagination. But so do many other aspects of Indian life. The same shock and horror prevails among today's historians, who portray the criminal tribe as an altogether “preposterous notion.”Footnote 13 As one critic put it, the robber guilds of colonial description were scarcely “more real than the orientalist fantasies or prosecutorial zeal of their hunters” (Schwartz Reference Schwartz2010: 14). “The double qualification ‘criminal tribe,’” writes another, “has no equivalent in any Indian language and was in fact imported into India by British administrators and jurists in the nineteenth century, who applied it to those whose traditional occupations they viewed as ‘predatory’ or ‘delinquent’” (Fourcade Reference Fourcade, Vidal, Tarabout and Meyer2003: 143). In the aftermath of the 1857 Rebellion, it became a “convenient expedient to label bothersome or unproductive lifestyles, like those of small itinerant peddlers” (ibid.: 167).
True, the label had many usages in late colonialism, especially following 1857: from advancing individual officers' careers to controlling redundant or inconvenient groups; combating insurgency or rural banditry; and more broadly expanding imperial legal, policing and penal institutions.Footnote 14 Indeed, in the late nineteenth and early twentieth centuries itinerant craftsmen, traders and entertainers, displaced peasants, and the indigenous “robber police” were increasingly labeled as congenital criminals.Footnote 15 Between 1871 and 1949, an ever-growing number of communities were systematically registered as “criminal tribes,” and settled in labor colonies where they were subjected to special surveillance and penal measures that included roll call, raiding, absentee passes, warrantless arrests, and thrashings. In 1947 India was home to some 128 communities and 3.5 million people classed as innately criminal by law (Major Reference Major1999). In 1949, the Criminal Tribes Act was replaced with the Habitual Offenders Act, which closely mimicked its predecessor: more than half a century after the label of criminality was formally removed and the colonies disbanded, the lists of “habitual offenders” in police stations remain full of “denotified” or “ex-criminal tribesmen” subject to the old set of special policing and penal measures (Piliavsky Reference Piliavsky2011a; Reference Piliavsky2011b; Reference Piliavsky and Gellner2013a). The story of lives crushed by the runaway juggernaut of criminal tribe legislation must not be forgotten, and several historians have already commemorated it in detail.Footnote 16
What remains abstruse is the deeper history of the idea of criminal tribe, which, as some historians have already suggested, stretches far beyond the colonial archive (Gordon Reference Gordon and Yang1985; Dundas Reference Dundas1995; Wagner Reference Wagner2007). But first, a few words of background.
A “CRIMINAL TRIBE” TODAY
I am a social anthropologist and my interest in criminal tribes first arose in the field. Since 2005, I have been conducting research in the North Indian state of Rajasthan among a people called Kanjars, known locally as a “caste of thieves” (choroñ ki jāt). Robbery, mainly housebreaking and cattle rustling, is not just their regular business. It has totemic significance, like pottery for potter castes or drumming for drummers, who identify as such even if they have never touched a pot or a drum in their lives. Before I ever met a Kanjar, they were described to me as a “criminal caste” (aparādhi jāti or juraim pesā qaum), a concept in wide current circulation on the Subcontinent (Dumont Reference Dumont1957; Shulman Reference Shulman1980; Viramma and Racine Reference Viramma and Racine1997; Pandian Reference Pandian2009). Such assertions often trailed off into tales of Kanjars as a fierce and dangerous people who live in the jungly wastelands, rear lizards for wall-scaling burglary, know secret tongues, and possess hidden treasures and magical healing techniques. People told me that Kanjars defecated ceremonially on the site of their burglary, sacrificed children to bloodthirsty goddesses, outran police jeeps, and could even vanish magically on the spot. They were a closed, secret society, which no ordinary human could hope to penetrate (Piliavsky Reference Piliavsky2011a).
When I first arrived among Kanjars, I expected them to deny all this. Instead, they confirmed and elaborated on these wild fantasies. Their tales were also full of wall-climbing lizards, bloodthirsty goddesses, secret tongues, hidden treasures, and magical powers they claimed to possess. Kanjars maintained that they hailed from a long larcenous lineage and that they inherited the extraordinary skills and cryptic knowledge that makes them the most highly skilled thieves. One conversation with my Kanjar host, whom I will call Gopal, went like this:
AP: Is there anyone in this village who can tell me about Kanjar history?
Gopal: Yes, I can tell you all about it myself! We, Kanjars, are a very old caste.
AP: What kind of work did your people do in the past?
Gopal: Thieving (gaimi) is our old vocation, [emphatically] we are old-time thieves. I am a thief and my father and my grandfather's grandfather were all thieves.
AP: But don't people of other castes steal as well? Say, if I went now and stole some sugar from your wife's shop, wouldn't I also be a thief?
Gopal: Naturally, people of all castes may steal, but they are all new players who know nothing about the thieving business. They steal in the daytime and they get caught. They are never good thieves because theft is not their khandān (caste business/heredity), it is not in their blood. The youngest of our boys are better thieves than these jokers.
AP: How is it that Kanjars are so much better at stealing than others?
Gopal: [laughing] Don't you understand? How can I explain this to you? Look, you know the old cobbler who sits in the bazaar? He is an old man. He is blind and deaf and you know that when we go to him, I have to shout into his ear so he can hear me. But when he makes shoes, they shine. You and I could not make shoes like that. Shoemaking is in the old man's blood. The cobblers have their own knowledge. This is why they make excellent shoes. It is this way with us, Kanjars. Everyone knows that we are a caste of thieves and we have our Mother's [Goddess'] special blessing.
If their neighbors dwelt on how low, lewd, and dangerous Kanjars were, Kanjars themselves stressed their strength, valor, and wit. Yet both agreed that theft was the Kanjars' age-old vocation, which defined their caste and cast them outside the ordinary social, moral, and legal order. This fact could not have been more starkly displayed. Kanjars almost never live in proper, multi-caste “villages” (gāoṅs), but rather in single-caste “camps” (deṛās) or “settlements” (bastīs), which are only rarely connected by paved roads and electrical wiring to the wider world. By repute and often in practice, Kanjar ghettos are the dens of Brahminical vice: meat eating, drinking, and thievery. Respectable people avoid Kanjar settlements or visit them surreptitiously to get some freshly brewed liquor (dāṛū) or meat. Local Kanjars practice theft as a regular, expert trade. In the sixteen months I spent in their encampment, few nights passed without at least one gang venturing on a thieving trip. Every night, stolen goats and rams were brought to the camp and, because I lived with a gang leader, every night I ate meat.
When I arrived in the field, I was certain that tales of Kanjar criminality were a postcolonial rumor, echoed by colonial subjects long after their masters had gone. Yet my informants were adamant that what they described was an ancient tradition stretching back as far as memory reached. My suggestions that British authorities created “castes of thieves” met with a mixture of amusement and affront. An elderly woman explained: “Child, the English made thief colonies where in the time of kings the Rajputs [ruling class] settled thieves. We have been thieves for a long time, child. Our fathers and grandfathers made us into thieves; it was not the work of the English.”
On my return to Oxford, where I was completing my doctorate degree, I set out for the Indian Institute Library to see if I could find references to castes of thieves in precolonial sources. Expecting to find little, I quickly came upon a vast array of writings, from ancient legal treatises and medieval epics to early modern travel accounts, full of references to robber tribes described in very familiar terms. And it became apparent that even preliminary research would consume an entire summer. It took far longer and, but for lack of time, could have gone on for months. The sketch I offer here is by no means a consecutive or a comprehensive history of the idea, but a selection of references from sources readily available in the English language. I chose my sources to range widely across time, space, genre, and purpose so as to give readers a sense of the idea's reach on the subcontinent.
BRAHMANS' BRIGANDS
South Asian literati have long thought that thieving was a communal trade.Footnote 17 Brahmanic narratives have almost never portrayed robbers as solitary operators, but invariably as “banded, cartelized and organized groups that live together” (Bloomfield Reference Bloomfield1926: 205–6). Most Brahmanic sociology of thieving is contained in the narrative literature: the Vedic “wilderness books” (āraṇyakas), the great epics Rāmāyaṇa and Mahābhārata, the vast corpus of ancient and medieval stories (kathās), the dramas, and Jātaka parables of the Buddha's lives. In these texts, the roadsides, mountain passes, and forest tracks teem with robber bands that prey on merchants and travelers. The robbers, whose vague outlines usually appear in the narrative middle distance, almost always come out of the “jungle,” a place beyond the ordinary social, political, and moral pale.Footnote 18 In Brahmanic cosmology the jungle is the periphery of civilization, the cosmic fringe, the wasteland, a socially negative space home to various outsiders to ordinary moral, ritual, legal, and social life.Footnote 19 The jungle is a place of danger, desire and magic, austerity and intrigue, extreme purity and extreme pollution, intense virtue, and equally intense vice, the home of gods, vagrants, exiles, and holy men, of fugitives and renouncers, saintly wives, half-breeds, and deposed kings.Footnote 20
The narratives make it clear that these robbers are beyond the fold of acceptable social and political life. They are pariahs. Manu, India's famous ancient lawgiver, wrote that robber tribes lived “outside the village,” on “burial grounds, on mountains and in groves,” and that they wore “the garments of the dead,” ate “their food from broken dishes,” and “always wander[ed] from place to place” (Bühler Reference Bühler1886: 50–52). The kathā (story) literature equally locates robbers outside: in secret lairs, underground dens, and on hillsides. Vedic accounts, like the kathās, repeatedly describe robber tribes as “riteless, void of sense, inhuman,” “frightful and terrible,” “flesh-eating,” “wine-drinking,” “bird-hunting,” “parrot-roasting,” and “human-sacrificing” predatory brutes.Footnote 21 In the epic of Mahābhārata, when the sage Gautama begets “godless progeny” with a woman from a robber tribe, he goes straight to hell (12.173). When in one kathā story robbers are asked whether virtue or vice ensures success, they choose, quite expectedly, vice (Tawney and Leumann Reference Tawney and Leumann1895: 163).
The names of robber tribes that appear in the texts (including Dāsyu, Mleccha, Drāviḍa, Pulinda, Śabara, Bhilla, and Barbara) are epithets of exclusion, which connote lawlessness, depravity, and chaos at large (e.g., Thapar Reference Thapar1971). In Vedic and later writings, “Dāsyu” (or Dāsa) referred to brutes, barbarians, demons, outcastes, robbers, and beasts. Like its Greek cognate, barbaros, “Barbara” referred to “babblers”: foreigners, hillmen, and fools who spoke in strange tongues (Monier-Williams Reference Monier-Williams1876: 674; Thapar Reference Thapar1971). In Brahmanic cosmology to “babble” was not to be merely unintelligible, but unable to recite Sanskrit formulae and thus excluded from sacrificial practice—the hearth of Brahmanic civilization. The babbling “Barbaras” were therefore “sinful, low and barbarous” (Thapar Reference Thapar1971: 410–11). The name “Pulinda” (also Pulindī), used for jungly robbers throughout the medieval Kathākośa collection of tales, meant simply a “ruffian” or a “boor” (Bloomfield Reference Bloomfield1926: 212). Perhaps most plainly, “Bhilla,” a name still used by some tribal groups, in Sanskrit meant simply “separate” or “outcaste” (Bloomfield Reference Bloomfield1899: 81; Doniger and Smith Reference Doniger and Smith1991: 113). Like late Victorians, the authors of ancient texts often described this “low-born” (Sanskrit [Skt.] apasada) folk as “delinquent-born” (Skt. apadhvaṃsaja) (Daniélou Reference Daniélou1991: 141). Manu wrote that forest tribes (dāsyus) and the brutal, heretical, and low-living servants (śudras) were naturally drawn to “forbidden occupations,” like theft (Bühler Reference Bühler1886: 4.61, 5.131, 9.225, 8.66). Furthermore, thieves were seen as external to mainstream morals and way of life, the “common dharmic order” of sādhāraṇa. As outsiders, thieves were neither subject to law nor protected by it. Thus, Brahmin lawmakers stipulated that thieves could neither bear witness nor take part in sacrifice (Ganguli Reference Ganguli1883–1896: 8.66–67, 3.150). They also encouraged their princely precepts to take freely whatever belonged to the jungly thieves (Olivelle Reference Olivelle2004: 191).
Some narratives give further substance to the stereotype of born felons, most of them describing banditry as a family business. Several canonical Jain texts and Buddhist Jātaka Tales refer to “settlements of thieves” (caurapallī) where robber families live (Das Reference Das1977: 19; Chalmers and Cowell Reference Chalmers and Cowell1895, vol. 4: 268).Footnote 22 In one eleventh-century folktale, a merchant abducted by a band of Śavara-robbers finds himself in their hidden village where he meets their kinfolk (Tawney Reference Tawney1923: 141–42). A fifteenth-century Jain narrative, The Adventures of Rāuhineya, gives a fuller description of a “thief settlement”:
In this country of Magadha, on the banks of the Ganges, there was situated a beautiful town, named Rājagṛha, adorned with wealthy inhabitants. Nearby was the mountain of Vāibhāra, delightful with its plateaux, which was ever a place of repose for both thieves and ascetics. The mountain—where thousands of lions and tigers roared by day, while (by night) it was terrifying with the howls of jackals and the hooting of owls—was resplendent with vanaspatī [trees or tree-ornaments] measured by eighteen bhāras (a large weight; or, a load), and with cascades like marvelous ropes of pearls. By virtue of magic charms, amulets, and simples the young of the thieves habitually played there with the young of the lions. Many ascetics, who lived on bulbs, roots, and fruit, dwelt in the woods around the mountain and performed manifold penance; and hundreds of families of thieves dwelt in the caves, which, shut in by bamboo network, were in the recesses of the mountain (Johnson Reference Johnson1920: 165–66).
The protagonist, Rāuhiṇeya, describes himself clearly as a “caste thief,” “the scion of a distinguished thief-family, proud of its reputation and position among fellow thieves … a thief, sprung from a thief-family, of pure thief-lineage on both my father's and my mother's side, uncontrollable even by the gods” (ibid.: 160, 189). In another eighteenth-century adventure tale the hero-thief is heir to his father's craft (Passi Reference Passi2001; Reference Passi2005: 514); and in several other folktales thieves work alongside their fathers, mothers, sisters, and mothers' brothers.Footnote 23 The idea of a professional robber guild appears also in the ancient (forth century BCE–forth century CE) Treatise on Politics (Arthaśāstra), whose author urges the king to employ “veteran thieves” to destabilize forest tribes (Olivelle Reference Olivelle2013: 234).
Many ancient texts also describe robbery as a specialist trade with professional knowledge and know-how. A twelfth-century Jain text, Upadeśamālā (Garland of instruction) describes “a horde of thieves (dhāṭī) intensely well-practiced in Ṭhagavidyā,” the tricksters' wisdom and magic spells used to break locks on houses (in Dundas Reference Dundas1995: 282). This specialist thieves' knowledge exercised the imagination of ancient India's scholars, who generated a substantial body of “thievery manuals” or texts of “larceny lore” (steya-śāstras, cora-śāstras or steya-sūtras).Footnote 24 Three of these texts are available in English translation. The Skandayāga or the Dhūrtakalpa (Skanda-sacrifice or Rogue ordinance) is a Vedic manual of sacrifice for Skanda, “the god of cunning and roguery” and the patron of thieves (Goodwin Reference Goodwin1893: vi). The Ṣaṇmukhakalpa (Dieter Reference Dieter1991) is a manual, written sometime in the seventh–ninth centuries, of thieves' potions, poisons, incantations, invisibility ointments, sleeping draughts, and the like. And Dharmacauryarasāyana (Elixir of dharmic thievery) is an eighteenth-century satire, adventure tale, and practical manual of the robber's craft (Passi Reference Passi2001). These and other texts describe elaborately formalized sets of specialist knowledge and skills, including cryptic omens, magic spells, thieves' signals and argot, and special dress codes.Footnote 25 Several texts include pedantic prescriptions of tunnels and breaches, the keystone of a thief's craft;Footnote 26 a second-century BCE Sanskrit drama even provides a list of seven canonical breach shapes: “blown like a lotus,” “sun,” “crescent moon,” “cistern,” “extended,” “cruciform,” and “full pot”' (Ryder and Lanman Reference Ryder and Lanman1905: 47–48).Footnote 27
These descriptions are astonishingly like the Kanjars' proud tales of making wall breaches. Gopal, my Kanjar host, once gave me a tour of breaches freshly made by “his boys” (Image 1). We can only marvel at the links between ancient authors' and modern thieves' keenness on breaches. What is important is the Sanskrit authors' and the Kanjars' shared insistence on the professional standard of thieves' craft, practiced canonically just like weaving, pottery, or Sanskrit chanting. Moreover, like any other caste profession in South Asia, robbery requires a patron-deity's sanction. As Wilhelm Halbfass and Paul Dundas have pointed out, the idea of divinely sanctioned robbery dates back to at least the seventh century (Halbfass Reference Halbfass1983: 10–15) and was certainly in circulation in learned discourse almost a century and a half before the earliest British writings on communal criminality (Dundas Reference Dundas1995: 284). In ancient India the robbers' divine patrons included the Vedic thunder-bearing god Rudra, the “Lord of Thieves” (stenānām pati) (Vājasaneyi-saṃhitā 7.20ff, in Falk Reference Falk1986: 60–65), Rudra's son Skanda, the “father” and patron of thieves,Footnote 28 and most importantly, the many forms of the Goddess (śakti).Footnote 29
Like other professional guilds, robbers were also said to possess their own dharma, “the dharma of thieves” (Skt. cauryadharma), the “identity” or “moral code” that determines the thieves' physical, mental, and moral qualities, their modus vivendi and place in broader society, as well as their ethical code (Doniger Reference Doniger1976: 94–95). A twelfth-century Jain text lists the robbers' moral qualities: “pride, egoism, greed, craftiness and so on, ability to adopt various outward shapes, habitual stealing of vast quantities of wealth, lack of concern for military strength used against them, contempt for divine and temporal power, and skill in burglary” (in Dundas Reference Dundas1995: 282). The Arthaśāstra instructs the king to gather intelligence through an army of spies “appearing as ascetics, doctors, astrologers, thieves, and the like” (Olivelle Reference Olivelle2013: 43), implying that thieves can be identified physically, presumably through bodily signs or clothing.
The author of the earlier-mentioned Dharmacauryarasāyana expounds on the ethics of thievery, whose technical mastery is its virtue: a good thief can skillfully assess, navigate, and outwit his circumstances because stealing is only wrong “when one gets apprehended or killed” (Passi Reference Passi2001: 2.38–39, 1.42–47; 2005: 519). There is also a strict code of ethics. The same author tells us that virtuous theft is pure of greed and socially correct: the righteous thief steals from the rich and gives to the gods and the poor; “the ultimate end of dharmic stealing is to make off with a part of incalculable wealth” obtained from “men of incalculable wealth” (Passi Reference Passi2001: 2.29), but not from “those who have acquired wealth dishonestly, or possess limited … wealth”; stealing from “polluted” persons like “women, merchants, śūdras [servants], pimps, debtors, paupers, and whores” also threatens the integrity of the trade (ibid.: 2.35, 2.61, 2.64).
The short set of references I have provided reveals an old and pervasive stereotype of a “robber caste” which comprises an association that (1) is socially peripheral and (2) professional, (3) with its own dharma (intrinsic nature, modus vivendi, and ethical code) and (4) a set of specialized knowledge and skills, which its members (5) transmit and practice by heredity within closed family circles (6) under the tutelage of patron-gods.
To show the resilience of a stereotype is not to claim the actual existence of groups that might have matched this description. For all we know, the texts in question are no more than records of Brahmins' imaginative form. This imaginative form, the idea of hereditary, collective, kin-based brigandage, is exactly what I try to show here. We will never know what popular stereotypes of robbers were abroad through millennia on the subcontinent, but the texts offer as good an insight as we can hope to get into what they might have been.
CASTES OF THIEVES IN EARLY MODERN INDIA
In this section I draw on writings of authors ranging from Mughal emperors to Venetian travelers and Jesuit priests, all of whose accounts precede British mention of congenital criminals by at least a century. The idea of a robber tribe thrilled early modern chroniclers no less than ancient Indian authors. In the early 1700s Father Bouchet, a Jesuit missionary stationed in Madurai, wrote in a letter: “I met with a Custom, which surprised me, in one of the Indian Castes, the Caste of Thieves … the Indians of this Caste, really rob in a very licentious manner.… Though I have found but one Writer more who speaks of this Caste of Thieves. I yet don't doubt the Existence of such a Caste: a Sett of Wretches, who may be tolerated in some Measure, and yet heartily despised by their Countrymen, as the Gypsies, &c. among us” (Lockman Reference Lockman1762, vol. 2: 261).
Father Peter Martin, another Jesuit writing from Madras in 1700, described his excursion “into the Country of the Caste of Thieves,” noting, “Tho' most of the People in question are turned Christians, and detest every Thing that has the least Tendency to Theft, they yet retain their former Appellation; and Travellers are afraid of passing through their Forests” (ibid., vol. 1: 452). A few years later Venetian adventurer and writer Nicolò Manucci noted the “customs of the caste of Thieves” he encountered in northern India (Manucci and Irvine Reference Manucci, Irvine and Irvine1907 [1708], vol. 3: 69), writing: “The ground-work of their manners corresponds very well with their caste name; for living almost entirely in the open country and paying no heed to magistrates, they come forth nightly and rob at will, even in the towns and at the courts of kings and princes. As for the helpless traveller, he is not only robbed but murdered, or else he receives so many blows on his legs that he is almost always left for dead” (ibid.).Footnote 30 Manucci went on to describe the caste's bizarre customs, including its strange wedding rituals and the “barbarous practice” of divorce which could be initiated, to his astonishment, by both women and men (ibid.: 69–70).
Most Early Modern accounts by European and Indian chroniclers are much vaguer than this. As in the ancient tales, in later narratives the tribes of robbers all resided in the jungles and hills, from which they descended occasionally to raid the Maratha, Mughal, and Rajput polities on the plains (e.g., Zaidi Reference Zaidi1989).Footnote 31 By the early sixteenth century, several such groups (including Gujars, Jats, Kolis, Minas, Bhattis, Mewatis, and Bhils) had acquired the repute of robber castes. The first Mughal Emperor Babur (r. 1526–1530) lamented in his journal, “If one go into Hindūstān the Jats and Gujūrs always pour down in countless hordes from hill and plain for loot in bullock and buffalo. These ill-omened people are just senseless oppressors!” (Beveridge Reference Beveridge1921, vol. 2: 454). By the mid-seventeen century this reputation was formalized. In his 1672 decree (farman) on the treatment of criminals Mughal emperor Aurangzeb described all Grasia tribes as “habitual robbers” and “wicked men” (Sarkar Reference Sarkar1935: 85–86).
In the 1670s French traveler Jean Chardin mentioned a certain people called Kolis, whom he described as “a race of robbers” who were “all arrant rogues and thieves” (1811, vol. 7: 479). Around the same time (1656–1686), François Bernier, a French physician in the Mughal Court, described “Messieurs the Koullys, or robbers” as “the greatest robbers and altogether the most unprincipled people in the Indies” (1891: 88–89, 91). And in 1708 Manucci wrote that Mewatis and Bhattis were the “great thieves and plunderers of the roads and villages” (Manucci and Irvine Reference Manucci, Irvine and Irvine1907, vol. 2: 457–58). A century later, in 1809, William Broughton, a British mercenary in the Maratha army, described a tribe of robbers who claimed to have “sprung from one family, the founders of their tribe.” He wrote, “They profess to be thieves and robbers and think it no more harm to maintain themselves by plunder than by entering into any military service” (1892: 105).
The idea of a robber tribe also informed judicial and penal practice in precolonial polities, where robber castes were often subjected to more severe punishment than “ordinary offenders.” As early as 1290, Jalal uddin Firz Khilji, the Sultan of Delhi, deported a thousand men from a “fraternity of thags” (biography of Jalal uddin Firz Khilji, cited in Elliott Reference Elliott1867–1877, vol. 4: 54).Footnote 32 We do not know whether these were the same “thugs” we find later in colonial sources, but the Sultan certainly treated them as a criminal fraternity, shipping them off en masse “to be conveyed into the lower country to the neighbourhood of Lakhnauti” (ibid.). Mughal rulers likewise distinguished ordinary thieves from professional robbers. Whilst the first were often simply reprimanded and released for stealing a length of cloth, “professional thieves” were often banished or had their hands cut off for the same crime. An ordinary person accused of robbery by strangulation was often simply chastised and imprisoned till repentance, but if they were “known professionals,” they were liable to be executed, and were often sentenced to death (Sangar Reference Sangar1967: 78). In 1672, Aurangzeb's decree on procedures to follow in criminal cases specified that while an ordinary “strangler” (phansigar) was to be “reprimanded” and “confined till he repents,” a “habitual strangler” should be put to death (Sarkar Reference Sarkar1935: 85–86).
The same principles operated in western Indian Rajput and Maratha polities. A seventeenth-century text from a court in South Rajasthan describes local Bhil and Mina hillsmen as inherently violent and deserving of especially severe punishment (Kothari Reference Kothari1985: 142). In a memoir written in 1817–1821, the Scottish soldier and historian John Malcolm reported that landowners in central India had “the power of putting to death the offender, if he is of a tribe of noted and avowed thieves (of which there are many in Central India); but not if he belongs to an industrious class” (1832: 563). While historically the business of robber castes was probably a much less stable trade than the narratives claim, the stereotype of the robber caste remained in steady use. As in British accounts, in early modern India the stereotype appears to have had many uses, ranging from a colorful flourish in a travel tale to condemnation of political enemies or inconvenient groups.
At times the stereotype could also be appropriated. The reputation of congenital thieves could be an important resource for tribal and itinerant groups, to whom it gave license to burgle, levy dues, and claim the patronage of local chiefs who employed them as go-betweens, watchmen, raiders, and escorts.Footnote 33 Since at least the ninth century, tribes of Meenas, Kolis, Gujars, and Bhils had the reputation of the fiercest robber groups who gained employment by raiding those whose patronage they sought. By the early nineteenth, many of them were patronized by village communities, landholders, heads of states, and increasingly by British officers, who employed them as watchmen, escorts, retainers, and guards. During the nineteenth century more and more of these hill tribes became gentrified, acquiring land grants and titles in a process that historians call “Rajputization” (Sinha Reference Sinha1962). A similar process has been described in South India, where over time thieves became watchmen, brigands became headmen, and bandits became kings, as in the famous case of the Tamil caste of Kallars (Dumont Reference Dumont1957; Dirks Reference Dirks1987; Pandian Reference Pandian2009).
In the north, robbers who captured steady patronage slowly joined the ranks of respectable, landed society. Many received titles and land grants, took up respectable trades, commissioned illustrious genealogies, built forts, and even married minor aristocrats, until some ultimately joined the Rajput elite, and later the British state (Broughton Reference Broughton1892: 85–86, 105, 233; Sinha Reference Sinha2000).Footnote 34 Some remained in the robber's business, but were set adrift to seek new employment as colonial policies disenfranchised their landholding masters.Footnote 35 Nevertheless, by the second decade of the nineteenth century it was noted that Bhils “have not yet abandoned their habits, but their robberies [were] upon a very limited scale to what they were a few years ago” (Malcolm Reference Malcolm1832, vol. 1: 525). In the course of the nineteenth century, as the hill tribes gentrified, new groups gradually took over the trade.Footnote 36
The new men were the various “vagrant” bhantu communities—itinerant craftsmen, peddlers, entertainers, and multi-professional castes like Sansis, Kalbeliyas, Berias, Moghias, and Kanjars.Footnote 37 To the newly rajputized tribesmen they became what the tribesmen themselves were once to the older landed elites: escorts, spies, hunters, retainers, watchmen, and genealogists. More and more bhantus started claiming the label of professional robbers and the rights to employment that it potentially entailed (Piliavsky Reference Piliavsky2011b: ch. 2).Footnote 38 In Rajasthan, Kanjars boast of long-standing, special ties to Bhils, Gujars, Jats, Meenas, and Kolis, the hill tribes that appear time and again in precolonial lists of “born robbers,” and whom Kanjars, too, name as the old, “true castes of thieves” (sacce corõ ke jāt).
WHOSE STEREOTYPE AND FOR WHAT PURPOSE?
Colonial practice vacillated between attempts to bring robber groups into the government's fold and altogether expunge them (Brown Reference Brown2004). By the turn of the twentieth century, British authorities came largely to favor the complete removal of robber castes from polite life. Millions were locked up in ghettoes where they were “reclaimed” or quarantined indefinitely, or deported to the Andaman Islands. Earlier, however, robber castes were integrated into colonial bureaucracies. In the early nineteenth century, the British employed hill tribes in western and central India as watchmen, soldiers, intelligence agents, and hunting assistants (e.g., Hunter Reference Hunter1843), and late in that century incorporated many into the newly formed colonial police (Arnold Reference Arnold1986; Piliavsky Reference Piliavsky2013b).Footnote 39 British officers often replicated indigenous styles of patronage, sealing their employees' loyalties with customary gifts of land, titles, and food (Russell and Hira Lal Reference Russell and Lal1916, vol. 2: 375). William H. Sleeman, the early nineteenth-century persecutor of “Thugs,” cast himself in the role of a Rajput pater familias (Freitag Reference Freitag1991: 236).
This very Sleeman is the key colonial source of the idea of criminal tribe, an idea dating back to the early nineteenth century and the persecution of Thugs, or robbers who purportedly strangled their victims to death on the roadside, and whom Sleeman claimed to “discover” and later to obliterate.Footnote 40 Sleeman described Thuggee as a “pan-Indian fraternity of felons” united by profession, heredity, and divine sanction in pursuit of their macabre trade. By the 1840s, the “cult of Thuggee” was pronounced dead, but as we have seen, the notion of congenital criminality lived on in later criminal tribe legislation whose advocates drew liberally on Sleeman's reports. The archives of his Thuggee and Dacoity Department were later repeatedly cited as prima facie evidence for the existence of criminal tribes (Piliavsky Reference Piliavsky2013b).
There is no way to tell how exactly Sleeman derived the concept of a “murderous fraternity” and later “criminal tribe.” Like all colonial officials, he claimed direct experience and native accounts as his source. But, as historians rightly note, his reports are subject to interested distortion since the power to criminalize groups en masse had many appeals. If nothing else, the dependence of Sleeman's career on criminal fraternities, in which he claimed exclusive expertise, could well have inclined him to creativity (e.g., Singha Reference Singha1993). Thus historians have dismissed his accounts and reports on the interrogation of Thugs (1836; 1839; 1849) as tales that tell us nothing about what the Thugs themselves may have thought and only what Sleeman himself wished his readers to believe (Chatterjee Reference Chatterjee1998: 125–44; Brown Reference Brown2002; Lloyd Reference Lloyd2008). They say Sleeman's reports are so hopelessly contaminated by imperial politics that it is impossible to tell fiction from fact (e.g., Freitag Reference Freitag, Anderson and Guha1998; Singha Reference Singha1998). They argue that even if Sleeman did not forge his evidence, his “approvers” (informers) were probably far too frightened to do any more than echo his words. Thus we can learn nothing of the Thugs' own conceptions, motivations, and rhetoric.
Here I will take a brief detour into Sleeman's archive to suggest that Thugs themselves had reasons to claim hereditary professionalism, informing Sleeman's accounts much more than we allow. A report on Thugs' interrogation in his Ramaseeana (1836) is representative and I will cite it here at some length:
Sahib [Muslim Thug].—We suppose that all Thugs originated by descent or initiation from the Delhi clans….
Q [Sleeman].—What do you think, Sahib Khan, am I right in thinking that we shall suppress Thuggee, or is Nasir right in thinking we shall not?
Sahib.—There have been several gurdies (inroads) upon Thuggee, but they have ended in nothing but the punishment of a few; and, as Nasir says, we have heard our fathers and sages predict these things as punishments for our transgression of prescribed rules….
Q.—Do you never feel any dread of punishment hereafter?
Sahib.—Never; we never murder unless the omens are favorable; and we consider favorable omens as the mandates of the deity.
Q.—What deity?
Sahib.—Bhowanee [Goddess]….
Q.—How can you murder old men and young children without some emotions of pity—calmly and deliberately as they sit with you and converse with you,—and tell you of their private affairs,—of their hopes and fears,—and of the wives and children, they are going to meet after years of absence, toil and suffering?
A.—From the time that the omens have been favorable, we consider them victims thrown into our hands by the deity to be killed; and that we are the mere instrument in her hands to destroy them….
Q.—You think that a Kuboola or tyro [amateur] could not any where form a gang of Thugs of himself?
Sahib and Nasir.—Never; he could know nothing of our rules of augury, or proceedings; and how could he possibly succeed? Does not all our success depend upon knowing and observing omens and rules?
Q.—It would therefore never be very dangerous to release such a man as a Kuboola?
Sahib and Nasir.—Never; unless he could join men better instructed than himself. Everyone must be convinced that it is by knowing and attending to omens and rules that Thuggee has thrived….
Q.—Have you any … instances [of Goddess's protection]?
Inent.—Hundreds! When Madhajee Seindheea [a Maratha ruler] caused seventy Thugs to be executed at Mathura, was he not warned in a dream by Davey [Goddess] that he should release them? And did he not the very day after their execution begin to spit blood? And did he not die within three months? …
Sahib Khan.—In the Duckun [Deccan] they [Thugs] are almost all composed entirely of Burkas—men well born, staunch and able; above all the men of Arcot.
Feringeea.—And the Hindoo Thugs of Talghat upon the Krishna river?
Sahib Khan.—Yes; they are extraordinary men.
Feringeea.—They have three painted lines on their foreheads extending up from a central point at the nose. I served with them once for two months.
Sahib Khan.—Yes; they have those lines.
Q.—But do not all Hindoos in that quarter wear the same marks?
Sahib Khan.—All Hindoos put them on occasionally, but they [the Hindu Thugs of Talghat] always wear them….
Feringeea.—You may hear and say what you please, but your funeral and marriage ceremonies indicate that your ancestors were nothing more than Khunjurs and vagrants about the great city?
Inaent.—It is impossible to say whether they were really what is described in these ceremonies, or pretended to be so; that they performed these offices for a time is unquestionable, but I think they must have been assumed as disguises.
Feringeea.—But those who emigrated direct from Delhi into remote parts of India, and did not rest at Agra, retain those professions up to the present day; as the Moltanies?
Sahib Khan.—True; but it is still as disguises to conceal their real profession of Thuggee.
Feringeea.—True, and under the same guise they practised their trade of Thuggee round Delhi before the captivity, and could never have had any other (Sleeman Reference Sleeman1836: 144–62).
When I first read this account, I was struck by parallels between Sleeman's reported dialogues and the many discussions I had with my Kanjar hosts. Like my informants, Sleeman's approvers claimed a pure and ancient pedigree, divine patronage, special skills, magical powers, and their own ethical code.Footnote 41 Parallels between my conversations and his interrogation ran further to Sleeman's reported reluctance to believe his approvers and my own initial refusal to take what the Kanjars told me seriously.Footnote 42 The parallels with Sleeman, the archetypal colonial brute who spun hideous yarns about natives to advance his career and the empire he served, were disconcerting to say the least. Was I too a closet imperialist, complicit with the colonial project? Did I force my own fantasies onto my informants? And, most unnervingly, what was the worth of what I thought I had learned in the field, if Kanjars in recounting their own history were mere ventriloquists of the Raj? But before the train of anxious inference runs away, let us return to the historical context in which Sleeman and the Thugs found themselves in 1836.
In 1809, when the twenty-one-year-old Sleeman left his native Cornwall for India, where he remained for most of his life, Lombroso's L'uomo delinquente (Reference Lombroso1876) was not yet published,Footnote 43 and late Victorian notions of criminal classes, often cited as key sources for the idea of criminal tribe, were nowhere in sight (e.g., Verma Reference Verma and Knafla2002: 124; Anderson Reference Anderson2004: 181ff). This was still the case when Sleeman died in 1856. Even in the late nineteenth century, when the notion of criminal classes was in circulation in London, British officials who were not well familiar with India often demurred at the notion of a “criminal tribe.” Even they found this extent of social determination implausible. The archive of criminal tribe administration reveals that most initiatives to classify “criminal tribes” came not from above, as has been argued, but from field-level officers and their native assistants.Footnote 44 Resistance to the idea among authorities in Calcutta, Delhi, and London meant that early attempts to police “criminal tribes” took a long time to get off the ground (Brown Reference Brown2004; Piliavsky Reference Piliavsky2013b).Footnote 45 When they were finally approved, the criminal tribe initiatives (like the Thuggee campaign before them) were perennially under-funded. Drastically understaffed reformatory colonies often dispersed within weeks of being formed, and the whole venture was perennially vulnerable to fiscal and administrative collapse (Singha Reference Singha1993; Brown Reference Brown2002: 84; Piliavsky Reference Piliavsky2013b). By the 1920s, the Government of India substantially washed its hands of the dubious enterprise, subcontracting most of it to the Salvation Army (Tolen Reference Tolen1991; Radhakrishna Reference Radhakrishna2000).Footnote 46 Sleeman certainly tried to persuade audiences of his cause. But why should we assume that he was less taken aback by the claim that robbery is a caste trade than were British officials in the late nineteenth century, Jesuits and European travelers before him, and I in 2008?
Why should we reject the possibility that Thugs too had reasons to convince? After all, as I noted earlier, the early nineteenth century was a time of great political changes when many robber groups lost patrons and were set adrift. British administration held out a promise of employment to the many robber groups that had lost indigenous patrons. These included both hill tribes and the bhantu vagrants who pepper the lists of Sleeman's Thugs. As prospective clients, the robber bands would have been keen to present themselves to potential patrons in the best possible light: as “proper,” hereditary, caste professionals. This would have been especially so with Sleeman. His patronage not only saved them and their associates from the scaffold, where many masterless Thugs perished, but also offered employment—and respectable standing—in the bureaucracy of the Raj. Sleeman himself acted like a Rajput patron: he provided for his Thug informers an income and a home. As Sandria Freitag points out, “The similarity between the spatial and psychological configurations of his compound and those of thag-landlord relations in a village is not coincidental” (Reference Freitag1991: 236).
I myself experienced something like this: a pursuit of my patronage among Kanjars who were keen to impress on me the idea of a thief caste. Kanjar hosts were for a long time unconvinced that “research” was what a young “English girl” was pursuing when she lived in their ghetto, away from her family, for months on end. They suspected I was a government agent and hoped I would hire them as informers and bodyguards, just as local landlords, policemen, and petty politicians do (Piliavsky Reference Piliavsky and Gellner2013a). Even after my hosts conceded the possibility that I might indeed be studying their “history and culture,” I retained the alarming nickname “Sarkar” (boss, government, state). Gopal, who took me into his home, introduced himself as a pedigreed thief because he hoped to secure patronage (protection and income) from me. He later told me so himself. The hope for my patronage never left the village and several of my Kanjar hosts still boast of being my “bodyguards.” Kanjars also advertised their pedigree whenever we visited their current or prospective patrons in nearby villages or the police.
This is not to assign any greater independent reality to either Sleeman's or the Thugs' purported claims, but to recognize a more complex choreography of motivation reflected in his reports. Nor is it to elide the shifts introduced by colonialism, any more than it is to say that history brings no change. Before the British arrived in India there were no reformatory colonies or attempts to “reclaim” robber guilds. The distinctive idea of a “criminal” who is not only outside the ordinary order, but also intrinsically in conflict with it, certainly shaped the way robber castes were treated during the Raj. As “criminals,” robber castes could be either exterminated or converted. In South Asia today, as in precolonial India, people acknowledge that robber castes are a reality they must reckon with. This does not mean that the robber castes ever were a plainly “legitimate” part of the social mainstream. In Brahmanic cosmology, while robber castes had a place in the range of dharmic possibility, they nevertheless remained outside the social, moral, and legal mainstream—sādhāraṇa or the “common dharmic order.” In every source they appear as delinquent and dangerous, if excitingly strange or magical—what one could call exo-castes. Many robber groups were employed as retainers and watchmen, but sometimes they were also banished or their members subjected to brutal punishment. What colonialism did was not to introduce the idea of a robber guild or an outlaw tribe. Rather, it brought an evangelical attitude that still prevails in NGO circles and among upper-caste activists who wish to “uplift” the likes of Kanjars.
CONCLUSION
The stereotype of “criminal tribes” has a deep history on the subcontinent. I have offered neither a consecutive account nor even its small segment, but a comment on the strange blindness among historians to this history. Scholars of Sanskrit and historians of ancient and medieval India know that Indian literature is peppered with descriptions of outlaw tribes.Footnote 47 Nor have historians of late precolonial India, including postcolonial critics, been unaware that certain communities have been thought of as, and styled themselves, robber castes (e.g., Dirks Reference Dirks1987; Pandian Reference Pandian2009). And yet, whenever academics now mention “criminal tribes,” they almost always reiterate the story of colonial fiction. Last year I received two articles on “criminal tribes” to review for esteemed academic journals, and each repeated that it was a colonial stereotype. The hypnotism of this bias is extraordinary and leaves scores of learned and talented historians tone-deaf to voices that were there before, during, and after the Europeans reached, and quit, the subcontinent.
To transmit these voices is no mean task, empirically every bit as much as analytically. First, empirically: studies of European colonialism dominate the historiography of India and of other former European colonies. Historians of colonialism, or postcolonial historians, do not only drastically outnumber medieval and ancient historians; postcolonial historians are usually the ones who make grand statements about the arc of history, however poorly their own research equips them for the task (often confined to colonial archives and sources in European languages). They set the tone for debates in which all historians must engage (if their work is to get noticed) and generate metanarratives that circulate far beyond their region and discipline. The core metanarrative—that European colonialism was an epistemic watershed—is an obstinate stereotype, repeated by students and young historians as if by rote. The result is methodological solipsism: history becomes confined to the European archive, which offers its circular confirmation, an effect reinforced by the dense population of historians of the colonial era and the sheer volume of their output. As Peabody remarked, the irony of this approach is that by presenting colonialism as the source of most (if not all) substantive ideas, postcolonial critics “may come closer to success where colonialism itself tried and failed; that is, in suppressing native agency” (2003: 3).
Concerning analysis, the trouble runs cognitively deeper than Eurocentric methods. It stems from a presumption that knowledge is necessarily determined by power. Historians who tilt the causative weight in the knowledge/power relation onto the side of politics, trace their intellectual genealogy generally to Foucault's work. Foucault wrote a great deal about the relation between knowledge and power, but he considered both sides of the relation (see for instance Foucault Reference Foucault and Gordon1980). Foucault read Francis Bacon's claim that “knowledge is power” as meaning both that power shaped knowledge and that power required real knowledge, not mere prejudice or invention. Saïd and his followers have run the equation in one direction: the effect of power on knowledge, attenuating knowledge to the status of a rhetorical residue of political process. From this angle, ideology, or indeed “culture,” is no more than a reflection of politics. Rulers' ideas reign and the ideas of subjects are subjugated to their masters' political schemes. Thus Saïd refused to examine the truth-value or indeed any content of Orientalist knowledge, polluted as it was by the colonial project. This set an analytical pattern which prevailed, making it nearly impossible to establish continuities of knowledge or indeed say anything of substance about anything other than the colonialists' conceptions and political schemes. It was important to draw attention to the politics of knowledge formation under the Raj, as Bernard Cohn did long before Saïd. But after Orientalism, when “power” was adopted as a kind of natural force in analyses (Sahlins Reference Sahlins2004: esp. 145–47), our capacity to apprehend the subtleties of what people think they know, how they come to know it, and why their knowledge takes one or another form has been severely undermined.
But let me return to the problem of stereotypes. By showing that the stereotype of criminal tribe stretches into the past far beyond European colonialism in South Asia, I hope to have more than redressed one error in historiography, but to have indicated a greater degree of communication, and perhaps even comprehension, between colonial masters and subjects. Colonial officers certainly put the idea of criminal tribe to their own different uses, but the stereotype itself was neither new nor were they the only ones using it. Authors of ancient treatises, Mughal rulers, European travelers in precolonial India, and itinerant groups (today and in the past) all called on the idea of hereditary robber tribes to pursue a wide range of distinctive purposes.
The rhetorical force of this and other stereotypes derives from their being just that—stable ideas. While setting in motion the mobile machinery of political life, stereotypes themselves remain static ideas that imbue momentary strategic decisions with the moral authority of fixed, eternal, or commonly recognized “truths” (see Herzfeld Reference Herzfeld1990). Stereotypes are what Thomas Trautmann called “locational technologies,” or cognitive frameworks that remain exceptionally stable and which often come from the deep past (Reference Trautmann2006: Introduction). To be effective, stereotypes must have cultural purchase: they must be widely and readily recognized. It was because British rulers and their Indian subjects occupied some shared cultural spaces that indigenous stereotypes were more efficacious for classifying groups, justifying legal action, or indeed communicating in any way at all, than any foreign fancy would or could have been. Relatively simple ideas, which can absorb shifts of emphasis and different shades of meaning, while still retaining a familiar shape, are the best—the most resilient—stereotypes. The idea of a criminal caste is like that. One can inflect it to put greater stress on heredity or criminality, or underscore its religious aspects (as did Sleeman), but the basic, recognizable outline—a community of hereditary, professional robbers with special skills, power and a code of ethics, operating under divine mandate—retains its shape. Those who see such ideas as mere traces of nefarious imperial bewitchment endorse their historiography's stereotype. And whenever they do, they can only do so at the cost of comprehension.
