Chapter 6
Hereditary Performing Communities in North India

One of the richest places to continue the search for a caste or occupational group called Kathak is the nineteenth-century British censuses and ethnographies of India, although the availability of this information is compromised somewhat by the context of the studies themselves. The anthropometric data (skull measurements, nasal indices and so on) that accompany some of the ethnographies and the consequent labelling of racial types are quite appalling to most present-day readers (see for example Risley 1891 and Crooke 1896). The methodologies for data collection were also problematic. Various provinces in India conducted censuses at different times and with different emphases, and the interests and priorities of individual officials shaped both collection and presentation of the material. Finally, not only did the British rely on Indian informants and employees for data collection, Indians on the whole became increasingly aware of the power wielded through statistics and eventually demanded input into the process (Barrier 1981). The censuses and resultant ‘Tribes and Castes’ volumes must therefore, like the travelogues and images in Chapter 5, be examined as products of the people who produced them. This is not to say that the information was necessarily falsified, but it needs to be realized that as well as documenting various population statistics, the categories and collection methods of the censuses and the subsequent presentation and interpretation of the data reflected the concerns, priorities and attitudes of both enumerators and informants. Yet, the censuses contain a type of information that is not necessarily available elsewhere. Although the earliest colonial source using the word ‘Kathak’ is a travelogue rather than a census, almost all the European documentation of Kathaks lies in these reports. Furthermore, as a number of the histories of kathak refer to figures in censuses as proof of the Kathaks’ status and occupation in the nineteenth century (see Hein 1972, Kothari 1989, Narayan 1998) the information contained in these sources merits thorough investigation.

Kathaks

Documentation before the first official British Imperial reports is scattered and somewhat unsystematic. The earliest is the aforementioned travelogue, and because the description invites comparison with later colonialist government publications I will begin with it. In The Costume, Character, Manners, Domestic Habits and Religious Ceremonies of the Mahrattas, Thomas Duer Broughton briefly described the fondness of Hindus for ‘exhibitions of dancing boys’ during the spring festival of Holi. The boys, he explained: ‘are called Kuthiks; and are, as well as their attendant musicians, always Brahmans. Their dress is nearly the same as that of the Nach girls; but their dancing and singing is generally much better’ (Broughton 1813: 94). The earliest censuses contain less detail and slightly contradictory information. James Princep’s article entitled ‘Census of Population of the City of Benares’, published in 1832, includes 118 Kathaks in a list of Shudras (the lowest vara or caste category) and describes their occupation as ‘Music and Dancing Masters’ (Princep 1832: 495). In Montgomery Martin’s 1838 compilation of Francis Buchanan’s earlier surveys of Eastern India, a table in the Appendix entitled ‘List of Artists in the City of Patna and Bihar’ contains ‘Katthaks’ listed between ‘Kalawangt and Dhari’ and ‘Yajak’. The record shows a total of 58 ‘Katthak’ houses. There is no accompanying description of any of the artists’ activities nor any identification of status, but the list also includes washermen, soap-makers and tailors (Martin 1976 [1838]: Appendix 35). John Beames revision of Henry Elliot’s study on the ‘Races of the North-Western Provinces and Oudh’ (Beames 1978 [1869]) and Kuar Lachman Singh’s statistical information on Bulandshahar, a district southeast of Delhi (Singh 1874) are similar. Elliot’s original glossary of castes included performing groups such as Bhand, Bhat and Bahrupiya (Beames 1978: 17–18), but not Kathaks. Following Elliot’s entry of the divisions of the Brahmanical order, Beames, in his 1869 revision, inserted a list summarizing the enumeration and classification of Brahmans in the 1865 census of the North-West Provinces; listed between Gandharb and Bhat, there are 78 ‘Kataks’ in Benares. No information regarding the occupation of ‘Kataks’ is provided (Beames 1978: 150–53). Singh did not provide a description of Kathaks either, but included them in the ‘Serving Castes’ along with ‘Dhati’, Dom Mirasi and Bhat (1874: 132).

After the uprising or ‘Sepoy Mutiny’ of 1857, an Act of British Parliament transferred responsibility for the governing of India from the British East India Company to the British Government. The first official census of India attempted by the British took place around 1872, but was unsuccessful and incomplete. The year 1881 then saw the first complete Imperial census and the beginning of regular decennial censuses that continue today (Martin 1981: 61–2). The official reports on each census are available, although wading through the charts and numbers takes some skill and patience. More interesting and accessible are the publications, usually written by the census commissioners, that interpret the figures and provide detailed descriptions of the people enumerated. It is to these ‘Tribes and Castes’ books that music and dance scholars refer, rather than the government reports.

Although the 1872 census was never completed, enough information was gathered to allow publication of ‘Tribes and Castes’ volumes for some provinces including Benares, the North-West Provinces and Oudh, and Rajasthan. These books take the form of large glossaries: each volume is divided into chapters covering broad social categories within which groups are described in varying amounts of detail. In the volume on Benares one finds the first substantial account of the Kathaks, which is worth quoting at length:

The Kathaks are professional musicians. They are ‘to the manner born’, and form a distinct tribe or caste. The gift or inspiration of music is hereditary in this tribe, … The Kathaks, however, are only one of several tribes of Hindus devoted to music, dancing and singing; … They affect to be of high caste, equal in fact to the Rajpoots, and nearly equal to the Brahmans; and wear the janeo or sacred cord, which none but men of good caste are allowed to wear. Further west, the Kathaks do not make a salam in saluting any one, … but give their ashirbad, or blessing, like the Brahmans. Their women are not usually seen in public, but live in the retirement of the zanana, an additional testimony to the respectability of the tribe … The [male Kathaks] play on various instruments, and also sing and dance. They do not suffer their wives to appear on any other occasions [than performing at marriages]; yet women commonly accompany them to all musical festivals. Such women, who belong to many castes, come to the Kathaks’ houses for instruction in the art of singing and dancing. They are always and everywhere women of loose character. In India all professional singing and dancing, when performed by women, with very few exceptions, is performed by prostitutes … They are frequently hired together, the Kathaks to play on instruments, the women to dance and sing (Sherring 1974a: 273–4).

There are many interesting points in this paragraph to which I will return. Noteworthy for the moment is the timing of this description. Its publication in 1872 places it in close proximity to the Urdu sources discussed in Chapter 5, in which some of the earliest and most detailed Indian references to Kathaks are found. It is also interesting that this lengthy description of Kathaks occurs only in the Benares volume – there is no mention of Kathaks in Rajasthan (Sherring 1987) or Oudh (Sherring 1974b).

Next in this chronology should probably come H.A. Rose’s glossary on Panjabi castes and tribes (Rose 1970 [1911–1919]). Although published in the early twentieth century, it is a compilation of information of the censuses of the Panjab from 1881 by Denzil Ibbetson and 1882 by E. MacLagan. Here, Kathaks are included under the larger rubric of Mirasi (see below) and described in one sentence: ‘Hindus who teach singing and dancing to prostitutes’ (Rose 1970: III, 112). In 1885, however, another detailed description was provided by John Nesfield in his interpretation of the 1881 census figures for the North-West Provinces and Oudh. His words echo some of the information from Sherring, but his emphasis on status is worth noting:

The Hindu caste of musician called Kathak is entirely distinct, both in origin and character, from the preceding Muhammadan ones. While the latter have ascended from the Dom [low caste performers] and become Muhammadan, the Kathak has descended from the Brahman and remained a staunch Hindu … The tradition of their order has been preserved in various ways. They still wear the janeo or sacred thread; and in saluting any one they do not make a bow, as the other castes do, but pronounce the Asirbad or blessing like Brahmans. Though they have ceased to chaunt the Sama Veda in the presence of bloodstained altars, as their ancestors did …, they are still employed to chaunt sacred melodies before idols or other symbols in the temples. The modulations of the voice are now, as then, accompanied by certain gesticulations of the arm and movements of the body which are tantamount to dancing … But in these degenerate days … [Kathaks] are hired out to play and dance and sing at marriage festivals before large and mixed audiences; and their own wives occasionally, but rarely, sing in public. The men are generally accompanied on such occasions by women of loose reputation … They frequently take lessons in dancing and singing at the Kathak’s house and even from the Kathak’s own wife. This is a remarkable instance of extremes meeting (Nesfield 1885: 44–5).

Almost a decade later in The Tribes and Castes of Bengal, Risley also identified Kathaks as a sub-caste of Brahmans, but said that they ranked ‘very low’. He explained the term Kathak as ‘properly denoting a reciter of the Hindu books, [but now] also applied to musicians of any creed or caste who play the violin [sāragī]’ (Risley 1981 [1891]: 433). The last extensive description comes from Crooke who defined the Kathaks as ‘a caste of story-tellers, singers and musicians’ in The Tribes and Castes of the North-western Provinces and Oudh. He also described them as the teachers and accompanists of dancing girls and as ‘popularly regarded as Brahmins’ (Crooke 1896: III, 172–6). It is Crooke who is most frequently cited as an authority on North India castes, even in the late-twentieth-century Anthropological Survey of India (see Singh 1998). Through the censuses of 1911, 1921 and 1931, questions about caste and ethnography were gradually replaced by queries regarding economy and industry, and the census of 1931 was the last to report extensively on caste (Conlon 1981: 111). The last official reference to Kathaks is from the superintendent of the 1911 census: ‘These religious troubadours carefully preserve their ancient ballads, and allow nobody to tamper with them’ (Blunt 1931: 244).

This century of documentation provides numerous points of comparison and contrast. The importance of Barrier’s caveat to assess all the data from as many angles as possible (Barrier 1981: vii; also see Buckley 1998) is clear – by choosing a small part of any one of these citations, one could ‘prove’ just about anything one pleased about Kathaks. There are, however, consistent elements. During the nineteenth century, there were people identified as Kathaks who were clearly occupied in the performing arts, specifically music and dance. They were associated with dancing girls and acknowledged as their teachers, but not related to them. Female relatives of the Kathaks did not perform, except perhaps at weddings. As a distinct community, Kathaks were most visible in and around Benares, although later in the century they appeared in the censuses of Oudh, the North-West Provinces and the Panjab. They do not seem to appear in the concurrent censuses of what is now Rajasthan. There is no clear connection of Kathaks to court performance. In only one document was there any connection of Kathaks to temple performance. Nowhere is there any reference to a dance called kathak.

One inconsistency visible in these documents is the Kathaks’ status, or vara. Whereas there are many thousands of individual jātis or birth groups variously identified as castes, tribes and hereditary occupations (the divisions between which are not always clear), they are typically organized into four varas. Highest are the Brahmans, the educated and priestly caste. Next are the Ksatriyas, the warrior caste which includes rulers and leaders like the Rajputs. Third are the Vaisyas, who are the artisans and agriculturalists, and last are the Shudras, the servant caste. Outside this organization are the so-called ‘Tribals’ or Adivasis and ‘untouchables’ or Dalits (see Bayly 1999 for more information and further discussion). In the censuses, the Kathaks’ vara ranges from Brahman in Broughton and Nesfield, to Shudra in Princep and Singh. Intriguing are the accounts like Sherring and Crooke, in which the Kathaks ‘affect’ to be or are ‘regarded’ as high caste, but are not clearly identified as such.

The Brahman question is one of the most controversial parts of any attempt to uncover the history of Kathak identity. Today, the Lucknow Kathaks identify themselves as Brahmans and much of the printed literature supports this. As early as 1914, in his article for Mask magazine entitled ‘Notes on Indian Dramatic Technique’, Ananda Coomaraswamy described Bindadin, the great-uncle of Birju Maharaj, not only as ‘a poet and dancer and teacher of many, many dancing girls’ but also as ‘a devout Brahman’ (Coomaraswamy 1914: 124). Articles from the 1930s identify kathak dance as ‘in the hands of the caste of Brahmans who are called Kathaks’ (Zutshi 1937) or ‘used by the sect of Brahmans who are called Kathaks’ (La Meri 1939: 18). Later research in Benares (Varanasi) by Frances Shepherd involved ‘musicians and dancers who are Brahmans and call themselves Kathak and claim to have descended from the Kathaka [the ancient expounders of the Purāas]. They seem to have no documented history but have kept their identity throughout the years by observing the Brahman kinship laws’ (Shepherd 1976: 13). Yet, in Marg magazine’s 1959 seminal issue on kathak, Nirmala Joshi wrote in her short introduction to the genealogical tables: ‘It is equally interesting to note that the families of the “Kathak gharana” claim to belong to a single clan, viz. Kathak, a sub-caste among Brahmans. In the opinion of many, this claim is open to question and further verification’ (Joshi 1963: 16).

Other Performing Groups

One of the reasons for examining the inconsistency of the Kathaks’ social standing is the general rarity of high status among hereditary performing groups in India and many other parts of the world. High status is usually only attained through association with individuals or activities that themselves have some greater value or power in society. As possible members of the priestly caste, Kathaks stand apart from almost all other hereditary performers except perhaps Ras Dharis (who are the musical directors of the devotional theatre called Rās Līlā) and Bhats (who are or were bards and genealogists for powerful rulers). Yet in the censuses the Kathaks are described over and over not as temple or court musicians, but as professionals who teach and accompany dancing girls. Adding to the confusion about both Kathaks as a group and their status are the sotto voce suggestions by Lucknow Kathaks that, while they are Brahmans, the Jaipur Kathaks are not only not Brahman but not even Kathaks and belong to another performing group, the Dholis. This is, on one hand, supported by one of Natavar’s Rajasthani informants who admitted to her ‘We people, we are Dholis’ (Natavar 1997, 151), but, on the other hand, refuted by Jaipur Kathak Rajendra Gangani who patiently explained to me in 2002 that Dholis only play the dhol (a barrel-shaped drum) and are not Kathaks. The puzzle is further compounded by various allusions connecting Kathaks with yet other groups: the Lucknow Kathaks were descended from Ras Dharis (Banerji 1982: 63); the Jaipur Kathaks used to be Bhands (Erdman 1985: 79); Kathaks are a community of Nats (Puru Dadhich, cited in Swarnamanjri 2002: 81). In addition, the curious absence of Kathaks in Wajid ‘Ali Shah’s book Banī – a book which includes performers called Bhands, Bhagatiya and Naqqals – demands an investigation and comparison of Kathaks with other hereditary groups of performers.

Research into the activities, history and status of North India’s hereditary musical communities has been conducted by a number of scholars including Neuman (1990), Schreffler (2002), Brown (2003) and Lybarger (2003). An authoritative taxonomy still eludes all scholarship and one is often left with a picture of a rather fluid social layer wherein groups shift, migrate and adopt new names and identities as circumstances dictate. This fluidity is ongoing and current, and any attempt at a definitive glossary of performing groups would be far, far beyond the scope of this study and my – and perhaps anyone’s – knowledge and fieldwork. Placing the Kathaks in context of other castes and groups of hereditary performers is necessary however, and not the least because it is often insisted that they are somehow different from the others (as in Nesfield above).

A convenient starting point for the contextualization of Kathaks is The Life of Music in North India, Daniel Neuman’s study of musicians in Delhi in the 1970s. In it he reported on the social divisions he observed between hereditary communities of soloists, whom he identified as Kalawants, and accompanists, called Mirasis. According to Neuman, Kalawants are vocalists who trace their descent from court singers and are accorded high status among musicians. Mirasis, on the other hand, comprise families of accompanists who traditionally play tablā and sāragī, although some members of Mirasi families do perform as vocal or instrumental soloists. Both Kalawants and Mirasis are Muslim and neither community identifies itself using these names. The title Kalawant, which translates simply as ‘artist’, can be found not only in the sixteenth-century Ā’īn-i Akbarī, but also in mid-nineteenth-century documents like Maūdan al-Mūsīqī. Mirasis, on the other hand, are not among the artists listed in Ā’īn-i Akbarī or Maūdan al-Mūsīqī, and the name itself does not refer specifically to any musical occupation. Neuman theorized that in the climate of rampant social change in the nineteenth century, the term Dhari (referring to a low-caste performing group) became ambiguous: he observed that just as Mirasis appeared in the censuses of the second half of the nineteenth century (as in Sherring 1974a [1872]: 275), the Dharis (or Dhadhis, who can be found in Ā’īn-i Akbarī and Maūdan al-Mūsīqī), began simultaneously to be enumerated as a sub-caste of Mirasi. Some Dharis moved up the social ladder by becoming Kalawants, others were absorbed into the new social category of Mirasi (for more information and clarification see Neuman 1990: 85–144).

The multiple descriptions and sub-groups of the Mirasis have been explored by Lowell Lybarger in his ethnography on Pakistani tablā players (Lybarger 2003). Yet, the polysemy that Lybarger observed in the uses of the term Mirasi is not unique, but applicable to most of the performing groups. Although Mirasis are generally accepted to be Muslims, Rose included Kathaks under the Mirasi rubric in his Glossary of Panjabi tribes and castes (Rose 1970: 112). In Maūdan al-Mūsīqī, Neuman found evidence of Kalawants and Dharis intermarrying, and a Kalawant (Himmat Khan) who played the sāragī, an accompanying instrument (Neuman 1990: 132, also see Vidyarthi 1959a: 23–4). There are also four vocalists with the surname Dhari listed under Kalawant (Vidyarthi 1959a: 16). Yet, the name Mirasi occasionally appears as Dom Mirasi, which connects Mirasis to another low-caste performing group, the Doms (see for example Singh 1874: 132, Nesfield 1885: 6, and Rose 1970: 106). The Anthropological Survey of India offers no further insight, finding Mirasi in Uttar Pradesh known also as Dhadi, Pakwaji, Kalawant and Ranwal and in Delhi known as Charan, Charan Bhat, Tasbirdar and Merasi (Singh 1998: 2302–3).

Other than in Rose’s Glossary, there is only a footnote in Natavar commenting: ‘sometimes Kathaks are called Mirasis’ (1997: 148fn). Neuman, however, makes a clear point of separating the Muslim Mirasis from ‘the only other major communities of accompanists’ who are Hindu dancers, tablā and sāragī players from the ‘Katthak’ caste (Neuman 1990: 129). This is interesting because it throws a different spin on the social division of families of hereditary accompanists and soloists: if Kathaks as dancers are soloists, but as tablā or sāragī players are accompanists, where do they fit in this scheme? Or could one claim that the social organization of Hindu and Muslim hereditary musicians is significantly different? What is most probable is that the divisions between accompanists and soloists within families and communities are much murkier than it would seem from the neat division between Kalawant and Mirasi in Neuman. Yet more intriguing is the observation that both Kathaks and Mirasis seem to ‘appear’ in the written record – Kathaks at the beginning of the nineteenth century and Mirasis around 1870, while many other performing groups are documented back into the seventeenth century.

One group that has been consistently documented since John Fryer mentioned it in his seventeenth-century travelogue (Fryer 1967 [1912]: 38) is the Bhat. Bhats are described as bards and genealogists. They are not called Kathaks nor connected with them, but they are also given Brahman or pseudo-Brahman status in some of the censuses (Nesfield 1885: 45). There are a number of sub-sections of Bhats recorded in both Uttar Pradesh and Rajasthan and their varying status seems to be, logically enough, connected to the status of their patron group. Bhats are connected to another bardic group primarily centred in Rajasthan, the Charan (see Sherring 1974a: 271–2, Sherring 1987: 53–4, Beames 1978: 17–19, Singh 1998: 118–121). Bhats seem at first of great interest since, like Kathaks, they are Hindu, they are the performing group most visible in what is now Rajasthan, and they are connected to the Rajput courts where the Jaipur gharānā of kathak dance is said to have developed. The one extant early reference to Kathaks in Rajasthan, however, equates them not with Bhats but with yet another performing group, the Bhand (Erdman 1985: 79).

Bhands are one of the groups included in Ā’īn-i Akbarī, where they are described as mimics who sing, play dhol and tāla, and perform various circus-like feats. Like Bhats, they are in the British censuses, but, unlike Bhats, they are in Muraqqa‘-yi Dehlī and Banī. They had low status, Mughal noblemen in the seventeenth century were warned to avoid their performances (Brown 2003: 138), and by the censuses of the mid-1800s were described as ‘mimics, buffoons and jesters’ who were predominantly Muslim (Sherring 1974a: 276). In Sarmāya-yi Iśrat, however, one dance gat, the ‘male’ gat, is to be danced by both Kathaks and Bhands (Sarmāya-yi Iśrat: 173), and some Company-style paintings show Bhands dancing in kathak-like postures (see Miner 1993: Fig. 16). It is intriguing, therefore, to find the names Kathak and Bhand linked by a hyphen in tables showing budgets for the Jaipur court musical establishments (Guījankhānā) between 1883 and 1933 (Erdman 1985: 81–2). Erdman pointed out that the number of Kathaks in the employment of the court increased during these decades and suggested that ‘Kathaks as individuals became separated from their Bhand groups during this time, indicating a separation of the specialized dancers of Kathak from the general play of the Bhand performers’ (1985: 79).

The term Bhand, however, may not refer to a specific caste or community, but instead be an occupational category, simply indicating people who performed in plays. This is certainly true of the terms nā (actor) and naqqāl (mimic), although one sometimes finds these groups referred to as if they were castes. Thus, Dadhich’s statement that ‘Kathak denotes a caste (a social class or community) of Nats’ (cited in Swarnamanjri 2002: 81) and Erdman’s table linking Kathaks and Bhands may simply indicate that Kathaks are or were actors. Perhaps, therefore, the absence of Kathaks in Wajid ‘Ali Shah’s book, Banī, can be explained by assuming that they were included either under the rubric of Naqqal or Bhand. Yet, the gat dance items in Banī that are clearly part of today’s stage kathak are not connected to the activities described in the chapter on naqqāl (Shah 1987: 115–72), but rather to the Rahas performances presented by the women of the court. This brings us back to the familiar association of Kathaks with dancing girls, but no closer to understanding their place among other performing groups.

In her fieldwork in Rajasthan during the 1990s, Mekhala Natavar found an elderly Kathak who described his community as Dholi, although this confession made his relatives very uncomfortable (Natavar 1997: 151). In the widest and most general of definitions, Dholis ‘are a community of musicians and drummers … named after their traditional musical instrument called the dhol. … They trace their origin from the Rajputs’ (Chakraborty 1981: 343). Natavar wrote that Dholis are ‘mostly Hindu’ and ‘prefer to be called Kathaks, as they are the hereditary community responsible for the development of the North Indian dance style called Kathak’ (Natavar 1997: 147). Both these sources firmly place the Dholis in Rajasthan, but the MA thesis of Gibb Schreffler on Panjabi Bhangra contains descriptive information about Dholis in both the Indian and Pakistani Panjabs that includes Muslim ustāds and identifies families containing both Hindu and Sikh names (Schreffler 2002: 82–9). Once again one finds a supposed caste designation that, rather than identifying a discrete group, points to a social stratum that can only be defined as a caste or community within a given geographical area and perhaps a specific time period.

Another group of performers with whom Kathaks are associated is the Ras Dhari, as there are claims that the Lucknow Kathaks ‘belonged to the Rasdhari tradition’ (Banerji 1982: 63, see also Natavar 1997: 147fn and 154). Ras Dharis are the musical leaders and directors of the devotional folk theatre depicting the life of Krishna called Rās Līlā. They and the boys who perform Rās Līlā are said to be Brahman, but whether they belong to a discrete hereditary performing group is unclear and to my knowledge the Ras Dharis themselves claim no consanguine or other affiliation with the Kathaks. There are a number of studies on Rās Līlā (see among others Hein 1972, Thielmann 1998 and Mason 2002) but none explore the ethnography of the participants. Although there are any number of choreographic similarities between kathak and Rās Līlā dance that make this possible ethnographic connection all the more intriguing, opinion is divided as to which one influenced the other and most scholars admit it is probably impossible ever to know (Awasthi 1963, Natavar 1997: 59–60 and Mason 2002: 9).

One final group, the Bhagat, deserves mention, not because any literature clearly associates Kathaks with it, but because Wajid ‘Ali Shah included Bhagats in his book along with Naqqals and Bhands. In Ā’īn-i Akbarī, the Bhagatiya are similar to two other performing groups, the Kurtaniya and Bhavaya. The Kurtaniya ‘dress up smooth-faced boys as women and [make] them perform, singing the praises of Krishna and reciting his acts’ (Ā’īn-i Akbarī, vol. III ‘On the Classes of Singers’). The Bhagatiya perform similar material at night and the Bhavaya perform during the day. A later Persian manuscript, the eighteenth-century Bar al-Ma‘ānī, describes ‘Bhagat-Bazan’ as Hindu performers specializing in scenes from the life of Krishna. The Bhagats are also present in Muraqqa‘-yi Dehlī, although Katherine Schofield argued that its author Dargah Quli Khan mixed them up with Bhands, a rather interesting mix-up considering this investigation (Brown 2003: 138fn). In Rose, the ‘Bhagatia’ are ‘musicians who accompany dancing boys’ (1970: II, 83) but in Sherring they are ‘a caste … of loose people who pass their time in buffoonery, singing and dancing’ (1974a: 276). Yet by the 1881 census, they seem to have disappeared: Nesfield (1885) made no mention of them and in Crooke the name is applied primarily to a religious sect, with a single sentence advising that ‘the name is also applied to a class of dancing girls’ (Crooke 1896: I, 252–3). They appear in the Marwari census of 1891, as the male family members of the Bhagtan, a class of unmarried dancing girls (Singh 1998: 140). Risley’s Glossary does not contain any separate entry for Bhagat, but under the entry for Kathak mentions a Kathak who ‘went about … with a troop of Mahomedan dancing boys (Bhagtiya)’ (Risley 1981: I, 433).

Kathaks Reconsidered

If Dharis became Mirasis who are sometimes called Doms, but other Dharis remained in Panjab; if Lucknow Kathaks are descended from Ras Dharis but Ras Dharis are not Dharis and also do not consider themselves Kathaks; if Jaipur Kathaks sometimes identify themselves as Dholis but are occasionally called Mirasis and perhaps used to be Bhands, one is left either with a complete muddle or else with a picture of a very fluid social layer of musicians and dancers who have continually migrated, adapted, shifted, adopted identities and had yet other identities thrust upon them. Furthermore, although one can certainly argue that the census officers made errors and very likely misunderstood the finer points of the social organizations they were documenting, one needs also to take into account that the groups being enumerated were aware of the importance of names and increasingly aware of the power of the census to validate identities.

Caste in India is not, and probably never has been, the immutable system of social barriers it often seems, and it is undeniable that the British efforts to list and categorize social groups contributed to a perception of solidity that probably never was accurate. Scholars re-examining the censuses have found numerous examples of flexibility in caste designation and self-identification. Schwartzenberg identified what he terms ‘category climbing’ among agricultural people (1981: 46) and Conlon wrote of the ‘contentious representations’ of caste categories created by ‘upwardly mobile castes seeking to raise their rank’ (Conlon 1981: 109). Jacob Pandian, in his study of the raised status of the Nadar caste in Tamil Nadu observed that ‘it is not uncommon for divisions in caste groups to occur, with successful members of the caste group forming a new caste group’ complete with a new name and higher social standing. Such caste shifts often occur in relation to increased wealth or economic influence and involve the caste group not only ‘discarding its former lifestyle and adopting a lifestyle that is associated with ritually high-caste groups’ but also revising its history to ‘exclude any reference to [its] former inferior status’ (Pandian 1995: 71–2).

Considering the almost ubiquitous low status of hereditary performing groups, it would seem illogical to insist that attempts to gain higher status were never made by any of the above-mentioned groups. In one of the first written descriptions of Mirasis, Sherring wrote: ‘The men are said to be highly respected by Mahomedans, though for what especial reason I have been unable to learn’ (Sherring 1974a [1872]: 275). It would seem that this esteem did not last, but if, as Neuman asserts, Mirasis did arise from Dharis around this time, perhaps the ‘new’ caste initially attempted to present itself as respectable. Perhaps because the name shift was not accompanied by economic gain, or because the economic gain that precipitated the name change was fleeting, the group slid back into lower status. Subsequent reports make no reference to respect for Mirasis. The question, of course, is not what may or may not have happened to Mirasis, but whether this hypothesis of a caste shift can be applied to Kathaks and what evidence exists to support it.

Setting Broughton’s travelogue aside for a moment, the first census table enumerates Kathaks as Shudras (Princep 1832). In the next three documents (Martin 1976 [1838], Beames 1978 [1869], and Singh 1874), Beames includes Kathaks in a table of Brahmans, but Martin places them on a list that includes washermen and Singh on a list of ‘Serving Castes’ (which is the definition of Shudras). In Sherring’s descriptive paragraphs from 1872, Kathaks ‘affect’ to be Brahmans and act like them, wearing the sacred cord and offering blessings. In Nesfield’s 1885 account, almost the same language is used to describe the sacred cord and blessing behaviour, but the Kathaks are now clearly identified as Brahmans and the description includes a lengthy explanation of how they descended from a now extinct order. According to the ethnographies from the last decade of the nineteenth century, Kathaks were a low-ranking sub-caste of Brahmans in Bihar (Risley 1981 [1891]: I, 433) and ‘popularly regarded’ as Brahmans by the lower castes in the North West (Crooke 1896: III, 174). At the beginning of the twentieth century, Ananda Coomaraswamy made a point of identifying Bindadin as a ‘devout’ Brahman (Coomaraswamy 1914) and Sharar, although he never specifically wrote that Kathaks were Brahmans, associated the two by writing that ‘dancing in India was always connected with divine worship …. The people who practised it were Brahmans and their centres were Ajodhya and Benares where the kathak dancers lived’ (Sharar 1975: 141). By the 1930s the dance is plainly associated with a group of Brahmans calling themselves Kathaks (Zutshi 1937, La Meri 1939) and the Kathaks themselves are described as ‘religious troubadours’ in the census report (Blunt 1931: 244).

It is logical, in the face of this evidence, to suggest that the Kathaks of North India raised their status by adopting Brahmanical behaviour, attempting to dissociate themselves from the dancing girls by emphasizing their devotional repertoire, and eventually identifying themselves as Brahmans – particularly, it would seem, to census officers and dance researchers. It is possible that the early census officials simply got it wrong and, misunderstanding the culture, mistakenly enumerated Kathaks in the wrong categories by associating them with the other low-status performing groups. Yet, the evidence of their activities teaching and accompanying dancing girls belies this. A great deal of ink has been spilled, from Nesfield to Natavar, explaining how the priestly Kathaks lost their status and became associated with base activities. But what if one suggested that they may instead have gained status? This proposed status shift also appears in the census reports just after the Kathaks’ reported successes in the court of Lucknow, thus meeting Pandian’s criterion that caste mobility occurs in relation to increased politico-economic status (Pandian 1995: 72). Furthermore, the current tension between the supposedly Brahman Lucknow Kathaks and the perhaps Dholi Jaipur Kathaks combined with Natavar’s observations of the Rajasthani Kathaks’ concern about the ambiguity of their caste points to a similar shift in process in Rajasthan (Natavar 1997: 151–2). Although, to my knowledge, the Jaipur Kathaks themselves do not say they are Brahmans, Reginald Massey has already asserted strongly in print that ‘the Kathak gurus of North India (whether of the Jaipur, Lucknow or any other gharana) are all Brahmins’ (Massey 1999: 115).

One of Pandian’s other indicators of a caste shift is a name change. Although perhaps not quite as visible as the group’s rising fortunes, the hypothesis that Kathaks adopted the name at some time is not difficult to support. Kathaks are not among the performing groups in Ā’īn-i Akbarī. This could be because they were originally centred around Benares and had no contact with Akbar’s court. They are also not in Banī, yet there is no doubt that they were active in Lucknow. The possibility that the predecessors of the Kathaks were indeed present in both contexts but went by another name or were part of another group seems very likely. The theatrical element in stage kathak, forming such an important part of the activities of the Uttar Pradesh Kathavacaks, and suggested by Dadhich’s connection of Kathaks and Nats, leads one to consider Bhands and Bhagats as possible contenders. Of the two, the Bhagats are the stronger, not only because they are Hindu (except in Risley 1981 [1891]), but also because they seem to lose both presence and status in the censuses as the Kathaks gain prominence through the nineteenth century. There is a common connection to dancing boys. There is also the Krishna element, which was the main feature of the Bhagats’ repertoire in Mughal times and, in spite of the connection to dancing girls, seems to have been an important part of the Kathak’s repertoire at least since the time of Bindadin (Coomaraswamy 1914). Certainly, it is made much of now. The connection in Ā’īn-i Akbarī between Bhagats and Kurtaniyas who direct boys dressed as women in plays about Krishna also seems to reinforce the link between kathak and Rās Līlā, Kathaks and Ras Dharis, and Kathaks and dancing boys. Moreover, Kurtaniya can be translated simply as one who sings a kurtana. The kurtanas, devotional songs about Krishna, contain dance bols (Vyas 1963: 6–7). One cannot say with any certainty that Kathaks were indeed Bhagats. There are enough connections to make the suggestion, although this still does not account for their role as the accompanists of dancing girls, a role filled by Mirasis in other contexts. Nevertheless, one must entertain the idea that the reason for their absence in the documentation before 1800 is that they arose from some other performing group.

The other evidence that points to a name change is the name itself. Even today, there is some contention as to whether it is simply a descriptive term meaning ‘storyteller’ or belongs to a specific caste group. In addition to the ongoing questions regarding the relationship of Kathaks from Rajasthan and Uttar Pradesh, there are scattered references to people called or calling themselves Kathaks in other parts of Northern India. In her study, Dance Legacy of Patliputra, Shovana Narayan documented people she identified as Kathaks throughout Bihar, but although she linked them to the Lucknow family through discipleship, she did not clarify whether they are in fact part of the same birādarī (Narayan 1999). There is some evidence that the name could be adopted and that at least up to the 1950s male dancers trained in kathak could call themselves Kathaks even if they were not from Kathak families (Hein 1972: 33fn). More recently still, when I spoke to Jaipur Kathak Rajendra Gangani in 2002 about Kathaks in other parts of India, he told me that performers in Panjab put ‘K’ before their names to indicate ‘Kathak’, although it would seem highly unlikely that they would be related to the clan in Uttar Pradesh. The fact also that the rural members of the Kathak birādarī identify themselves as Kathavacaks rather than Kathaks, although the two names are arguably synonymous, indicates that there is to this day considerable flexibility regarding ‘Kathak’ identity.

A final aspect in this examination of the name Kathak, so crucial to the identity of contemporary dancers, arises from the dictionary definitions discussed in Chapter 4. There is nothing in the nineteenth-century dictionaries that connects the word kathak to a community; a Kathak at that time was a reciter, narrator, actor or bard. The final definition in Platts also defines a Kathak as a dancing boy. The connection between Kathaks and dancing boys has been as consistent as the connection with dancing girls (who, it must be remembered, were actually young women). Broughton wrote in 1813 about dancing boys called ‘Kuthiks’ he observed in what is now the Central Indian state of Maharashtra. He also mentioned their accompanists, who were undoubtedly their male relatives, but whom he did not call ‘Kuthiks’. It seems most logical in this case to suggest that the boys, who were the performers, were called Kathaks because they were telling stories through their dance, whether or not they were somehow related to the community documented in the Benares census of 1832.

The historical link of Kathaks to the area around Benares allows one further conjecture. Although the location of the initial caste shift cannot be confirmed, the documented concentration of Kathaks in this area combined with the city’s political and economic climate in the second half of the eighteenth century makes it a logical suggestion. In 1740, Chait Singh, the son of a local landowner, made himself Raja of Benares. Although tributary at first to the Nawab of Awadh and subsequently to the British East India Company and Raj, the area became an important commercial capital with a unique culture created by the upstart Raja and a successful merchant class of bankers and soldier-traders. Patronage of the arts was an important part of a legitimization process for the Raja and the new elite. In particular, the Raja became an important patron of the Hindu theatre tradition Rām Līlā, ‘[reflecting] among other things [the] need to cultivate an explicitly Hindu symbol of royal legitimacy, and thus achieve ideological as well as political independence from the Nawabs’ (Lutgendorf 1989: 41; also see Freitag 1989 for further information). The ongoing connection of the Kathak community to acting as well as music and dance suggests that they may well have been centrally involved in these productions and gained both wealth and status through their performances. Furthermore, the context of a purposeful Hindu renaissance in the midst of Nawabi rule in addition to the concurrent development of devotional storytelling traditions like mānas-kathā in the same area (see Lutgendorf 1991) makes the adoption of the arguably Sanskritized name of ‘Kathak’ in late eighteenth-century Benares very plausible.

It is very probable, therefore, that the caste now called Kathak arose through a process of ‘category climbing’ that began when a group of performers in what is now eastern Uttar Pradesh began to call themselves ‘Kathaks’. Sometime in the late eighteenth or early nineteenth centuries, they gradually claimed ownership of a term that had been broad-ranging, applicable to members of any number of performing groups. The location of this shift was perhaps the regional Hindu court of Benares. Whether the Kathaks initiated their status shift by riding this early wave of political change and religious revival is difficult to prove, but surely possible. The promise of further patronage combined with the possibility of wealth and status offered by the larger Nawabi court at Lucknow drew members of the birādarī there. When the Kathaks appear in the census reports and the mid-nineteenth-century Urdu treatises, they are playing tablā and sāragī, singing, dancing and accompanying courtesans, but are still associated with acting in the form of bhāv. In a process that would still seem ongoing, certain families within the birādarī specialized in response to various opportunities for patronage and advancement, moving away from vernacular theatre and focusing on the more specialized genres of tablā, sāragī and vocal music, and the more refined gestures of court dance. Most probably they initially learned from the Muslim court musicians and dancers, and subsequently found a place accompanying and eventually teaching female hereditary performers in artistic centres like Varanasi and Lucknow. Throughout the nineteenth century, armed with a new name and eventually the support of further success in Lucknow, they began to adopt the ritual behaviours associated with Brahmanhood. They seem to have removed their women from public performance at least partly in order to further differentiate their community from the female hereditary performers they accompanied. The nineteenth-century census officials offered a validation in print of this new identity, and the dance historians of the next century wrote its supporting history.