Sponsored By: Infinity Foundation

Caste in Medieval India

Caste in Medieval India: The Beginnings of a Reexamination1
By Dileep Karanth


“Who says India says caste, or so it seems.” So wrote J. C. Heesterman in his essay “Caste, Village and Indian Society”2, underlining the centrality of the problem of caste in India. Heesterman points out the word caste started out meaning something like “tribe” or “race”, but in the nineteenth century it came to mean something very specific, a specifically Indian phenomenon. “caste began to loom large, until it became in our century a shorthand expression for Indian society at large: Indian society is caste.”

The inequalities of the modern caste system and the fissures in Hindu society resulting from it are too well-known to need elaboration. The caste system is so pervasive that it has become a feature of life of all religious groups that live in India. At least, that was the case when first contact with Europeans took place. Thus it is not surprising that caste and Hinduism have often been equated. Sir Denzil Ibbetson wrote of the “popular and currently received theory of caste” (which he would go on to challenge) as consisting consist of three main articles:

  1. that caste is an institution of the Hindu religion, and wholly peculiar to that religion alone;
  2. that it consists primarily of a fourfold classification of people in general under the heads of Brahman, Kshatriya, Vaisya and Sudra.
  3. that caste is perpetual and immutable, and has been transmitted from generation to generation throughout the ages of Hindu history and myth without the possibility of change.3

Caste Among Indian Muslims

The caste system, however, exists even among Indian Muslims. But many a scholar traces it to Hindu influence:

The Muslim caste system is a result of Hindu influence; the Indian Muslims have acquired the system, …, from the Hindus through constant and continuous culture contact; the system of caste groupings itself resulted in the concept of social distance between the two communities, the Hindus and the Muslims.4

In this paper we will try to collect some background information which will hopefully help in sparking a debate regarding caste. That a debate is necessary is clear from a recent book by Marc Gaborieau, Ni Brahmanes Ni Ancêtres, in which the author presents his detailed findings, after several years of field work in Nepal, studying the Curaute, a caste of banglemakers. The book is sure to revolutionize our understanding of caste. We will present several quotes from Gaborieau’s classic in this article.

Writing about the dominant trend in British ethnography, Gaborieau claims that the British took a simplistic view of castes and presented Hinduism, taken as a whole, as inherently hierarchical in structure, as opposed to Islam, taken as a whole, taken to be inherently egalitarian.5 Any elements of hierarchy in Islamic society is taken to be a relic from Hinduism, just as Ansari has done in the quote above. This idea has become very strongly rooted in the literature on caste. This idea has been championed particularly by Muslim scholars in the 19th century, as they defended their faith against criticism by Western scholars. The basic thrust of these arguments was that

far from bringing about forcible conversions as accused of by the British, the Muslim conquerors carried out peaceful conversions, notably by means of the Sufis. The chief reason for its success would have been the particular attractiveness of Islam, as an egalitarian faith, for the lower castes especially the untouchables. The cities and qasba established by the new conquerors would have been spaces of liberty; they would have permitted the most disfavored people to rise in the social hierarchy, by opening new economic outlets.

Gaborieau points out that the best example of this theory is the book by Arnold, The Preaching of Islam, for which the author collaborated with two influential Indian Muslim thinkers, Sayyid Ahmad Khan (d. 1898) and Shibli Nu’mani (d. 1914). The book, which was written in Aligarh and was first published in 1896, says among other things: “A Hindu will naturally be attracted by a religion which receives everybody with discrimination” (Arnold, 1965, 291-291); and: “It is this absence of class prejudice which constitutes the real force of Islam in India and which allows it to win so many converts from Hinduism” (pp. 118-119).

In Arnold’s book we clearly see the formulations of a theme that had been or would be elaborated by many other scholars, such W.W. Hunter6 and James Rice. In the context of conversions to Islam in Bengal, Rice wrote that the Islamic armies “were welcomed by the out-cast Chandals and Kaibarrta.”7 In the face of numerous such claims, it could be expected that modern Muslim society in Bengal would present an egalitarian picture. However, it turns out that such is not at all the case. No less a man than Dr. B. R. Ambedkar took cognizance of the existence of castes even in Muslim Bengal. Quoting the Superintendent of the Census for 1901 for the province of Bengal, Ambedkar noted:

“The conventional division of the Mahomedans into four tribes – Sheikh, Saiad, Moghul and Pathan – has very little application to this Province (Bengal). The Mahomedans themselves recognize two main social divisions, (1) Ashraf or Sharaf and (2) Ajlaf. Ashraf means ‘noble’ and includes all undoubted descendants of foreigners and converts from high caste Hindus. All other Mahomedans including the occupational groups and all converts of lower ranks, are known by the contemptuous terms, ‘Ajlaf’, ‘wretches’ or ‘mean people’: they are also called Kamina or Itar, ‘base’ or Rasil, a corruption of Rizal, ‘worthless’. In some places a third class, called Arzal or ‘lowest of all,’ is added. With them no other Mahomedan would associate, and they are forbidden to enter the mosque or to use the public burial ground.

“Within these groups there are castes with social precedence of exactly the same nature as one finds among the Hindus.8

Bengal would become the seat of intense political activism and lobbying in the years after Ambedkar wrote these words. Caste would become a much talked-of political commodity, politicians would campaign for the loyalties of the masses, the province would go on to be partitioned, and yet even as late as 1973, caste would be an abiding feature of Bengali Muslim life. M.K.A. Siddiqui, who contributed an essay9 to an important book on the caste phenomenon among Indian Muslims,10 points out that there are several caste groups among the Muslims in Calcutta. Siddiqui discusses several different ways in which inequality manifests itself – restrictions on commensality, hypergamy, pollution by contact, etc. He divides the castes into three categories. The castes in any one category can accept food from the others in the category, but not from castes in lower categories.

The Dafalis who work as priests for the Lal Begis, or the Qalanders who sometimes live in their neighbourhood, refuse to accept food or water from Lal Begis.11

The groups are descent groups, “with or without occupational specialization”. For example, the Lal Begis (who roughly correspond to the Bhangi caste in Hindu society) are generally regarded as unclean on account of their humble occupation – “they often experience difficulty in getting their dead buried in the common Muslim burial ground.”

Hypergamy is widely practiced in the highest category, meaning that women from lower castes can be married into the higher castes (Sayyad and Sheikh), but not vice versa. The children of these mixed marriages are called “Sayyadzada” and “Sheikhzada” respectively. They do not attain the full status of their fathers, and are expected to make alliances of with people of their status.

Siddiqui regards the emphasis on birth as not being sanctioned by scripture, which he says, wiped out distinctions of colour and race. However, as Islam spread to distant lands, social stratification resulted as a result of historical developments and adjustments made to local traditions. Kinship with the Prophet became a new criterion of nobility. Siddiqui discusses a few other aspects of the caste structure, and, significantly notes that “the founders of the [Sufi] mystic orders belong exclusively to categories that claim foreign origin. Most of them are Sayyads.”

The situation in Bengal was similar to that in central regions of India, as shown by studies by Zarina Bhatty12 and by Imtiaz Ahmad. Bhatty studied the case of a village Kasauli in the state of Uttar Pradesh, and found the village society to be deeply caste-riven. At the top of the hierarchy was a lineage of Sayyads, and a subcaste of the Sheikhs, namely the Kidwais. These were the only Ashraf castes in the village. Elsewhere in India, the Ashraf castes include Sayyds, Sheikhs, Mughals and Pathans. These are communities claiming descent from population groups hailing from outside India. Bhatty points out that all four noble castes permit interdining, but commensality with the lower castes, consisting of groups descended from Indian converts, is not allowed. Also Sayyads and Sheikhs intermarry, but marriages between Sayyads and Sheikhs on the one hand, and Mughals and Pathans on the other, are not socially acceptable. In the village of Kasauli, there are eighteen other castes, consisting of groups defined by occupation. Closely linked to occupation is a notion of pollution, depending on the materials handled by persons following the occupation. A kind of hierarchy is defined, with castes who come into proximity with the Ashraf regarded as higher.

Nats, who skin dead animals and make drums, find a place close to the bottom of the scale while Julahas and Darzis are at the top end. Dhobis, who must wash soiled clothes, are closer to the Nats than to the Julahas.13

Bhatty discusses the interesting case of a divide in the musician community. The Mirasis who perform for the higher Ashraf castes, are regarded as superior to the Nats, who perform the same social functions, but for the public at large. The Mirasis have adopted the dress of the Ashraf, and have learned to speak Urdu, while the Nats converse in the local dialect. Thus the Mirasis have improved their social standing by the imitation of and association with the upper castes, who set the norms for the whole society.

In his article on the Siddique Sheikhs of Uttar Pradesh, Imtiaz Ahmed informs us of the various considerations taken into account when determining hierarchy within the status group called the Sheikhs. There are at least four of them:

  1. affiliation with an Arab tribe.
  2. descent from a person of Arab origin who is known to have close ties the Prophet.
  3. relationship to a place in Arabia or Persia.
  4. descent from someone who is said to have entered India along with the early Muslim armies.

According to Ahmed, the Sheikh subgroups emphasize their foreign origin and links to some Islamic personage of repute. The groups who claim to be descended from the Prophet’s own tribe, Quraish, are regarded as the highest. Then follow the descendants of first Caliph, Abu Bakr Siddique. Next in rank are those who count the next two Caliphs, Usman and Umar among their ancestors. They are followed by descendants of the close friends and associates of the Prophet. Descendants of other Persians or Arabs who may have come with the Muslim armies are ranked last.

As for the Siddique Sheikhs studied closely by Imtiaz Ahmed, they have only recently been recognized as descendants of Abu Bakr. Their Kayastha Hindu antecedents are quite well established, and their striving for recognition as Ashraf is a phenomenon quite well known all over India. Ahmed points out that the circumstances of the Siddique Sheikhs’ conversion is not known, but after conversion to Islam, they were allowed to retain their traditional occupation as land recordkeepers, a fact which is also attested to by the fact that the members of the caste also served as patwaris well after the annexation of the area by the British. (emphasis added)

Ahmed makes some other very interesting observations about the Siddique Sheikhs:

Convert groups to Islam are generally characterized as New Muslims and they are looked down upon by the social groups which are known to be descendants of foreign sources or who have succeeded in eliminating the stigma of recent conversion. This gave rise to certain differentiations in the adjustment of the Sheikh Siddiques after their conversion to Islam in the different villages. In villages that were largely or predominantly Hindu, the Sheikh Siddiques were excluded from the framework of interaction with the Hindu castes but they continued to enjoy a somewhat superior status as a Muslim group. But in villages where there were numerous other Muslim groups of superior status, the Sheikh Siddiques were not merely excluded from the social hierarchy of Hindu castes, but were also relegated to a somewhat lower position even within the hierarchy of Muslim castes.

The continued prestige of the Siddique Sheikhs in their native villages even after conversion can probably be explained by the fact that they were already a community which enjoyed prestige among the Hindus. After all, they were allowed to retain their prestigious occupation as land recordkeepers. But in Muslim dominated villages, the Siddique Sheikhs commanded little prestige among the Muslims, since they were not Ashraf. This is an example of conversion from Hinduism which has obviously not been motivated by a desire to escape the disabilities of the Hindu caste system.

Ahmed’s observations regarding the inferior status of “New Muslims” seems to be applicable widely in India. We find confirmation of this generalization is places as far removed from Uttar Pradesh as the Moplah-dominated regions of Kerala. The hierarchies in Moplah society have been studied by Victor D’Souza.14 He reports that there are five distinct sections among the Moplahs: Thanghals, Arabis, Malbaris, Pusalars and Ossans. The Thangals who are at the top of the pyramid, are a small group of people who trace their descent to the Prophet, through his daughter Fatima. The term Thangal is a respectful term of address, usually applied to Brahmins in Kerala. The Arabis are a group of people mostly concentrated in Quilandy (a town north of Calicut), who are descendants of Arab men and local women, but who have preserved the memory of their descent. The association of the Arabis with Arabia entitles them to a respect in Moplah society second only to that of the Thangals. The Malabaris also claim descent from Arabs, but they are those who followed a matriarchal system – the so-called “mother-right” culture. As for the Pusalars and the Ossans, D’Souza writes:

The so-called Pusalars are converts from among the Hindu fishermen, called Mukkuvans. Their conversion took place relatively late. Because of their latter conversion and their low occupation of fishing they are allotted a low status in the Moplah society. The Pusalars are spread all along the coastline of Kerala and they still continue their traditional occupation of fishing.

The Ossans are a group of barbers among the Moplahs and by virtue of their very low occupation they are ranked the lowest. Their womenfolk act as hired singers on social occasions like weddings.

The hierarchies in Moplah society also show a tendency to accord the highest place of honour to the Sayyads, and lowest place to new converts and despised groups, such as barbers. The motive for conversion could hardly have been the keen desire to escape the disabilities of the Hindu caste system. Why would converts voluntarily accept similar disabilities in the new society they were going to join? The realities of Indian Muslim society flatly contradict claims such as the one made by the renowned Islamic thinker, and rector of Deoband, Shibli Nu’mani:

If an Asiatic converts to Christianity, he does not obtain the rights (huquq) of the European community (qaum), even though he may have the same religion; even on the plane of religious rights, he cannot be equal (ham-sar) to the Europeans. In contrast, the “communitarian identity” (qaumiyyat) of the Muslims is not bound to the country, nor to the race (nasal), nor to lineage, nor to any other criterion – it is only bound to religion. Whether one is Persian (‘ajami), Indian (hindi), European or Asiatic, as soon as one enters the Muslim community, by the sole fact of conversion, one immediately (daf’atan) becomes the equal (barabar) of other Muslims sous le rapport de all rights: As soon as he has pronounced the profession of the faith, a tanner (Camar) can take a place in the first row in the mosque, thus putting himself on a rank equal to that of the [Ottoman] Sultan ‘Abd al-Hamid Khan; the Sultan cannot then claim to dislodge him from that place (cited in Shams-i-Tabriz Khan, 1983, 58-59).15

Gaborieau points out that this ideal picture of Muslim society fails to correspond with reality. Nu’mani aimed to fight Western scholars with their arguments. He wished to show that while conversion to Christianity did not guarantee the Asiatics racial equality with Europeans, such was not the case within Muslim society. Muslim society, in Nu’mani’s romantic view, was even ahead of the Christian West. However, Gaborieau points out that equality did not fall to the lot even of Nu’mani himself:

His immense learning was recognized the world over: but he lacked any mystic aura. Most importantly, he was descended from converts; and what is worse, his lineage lacked prestige locally. He was a pseudo-Rajput, in fact, from an obscure caste of peasants. He was perfectly conscious of this stigma. This is what led him, like Iqbal, to overemphasize learning and the role of Islam, and to be extremely determined to exact the mark of respect due to him in the capacity of scholar. He never succeeded in gaining admittance into the intellectual and religious establishment of the day. He was not able to establish himself in the two prestigious institutions where he taught, (Aligarh and the Nadwatu’l-‘ulama’ of Lucknow) which had been founded by genuine Sayyids and were populated with Ashraf; in both cases he had to resign (Metcalf, 1982a, 340-342).16

As we have seen before, scholars such as Ansari regard manifestations of societal inequality within Islamic society as an inheritance from Hinduism. In regions where the demographics have not shifted in favour of Islam, the influence from the ambient Hindu society on the Muslim minority is indeed strong. However, the practice of segregating the lower castes has continued in regions where Hindu political power, and possibly Hindu demographic preponderance, has long vanished. As discussed before, in the case of Bengal, castes like the Lal Begis have been discriminated against long after the majority of the population turned Muslim.

Matters were hardly different at the other end of the Indian subcontinent, in Baluchistan. The Census Superintendent of Baluchistan wrote in 1931 that members of the Chuhra caste or tribe, who identified themselves variously as Hindu Balmiki, Hindu Lal Begi, Musalman Lal Begi, Musalman Balashai, Sikh Mazhabi or simply as ‘Chuhra’ were ‘without exception … not allowed to drink from wells belonging to real Hindus, Muslims or Sikhs’ and were ‘not permitted to enter their places of worship’.17

The disabilities inflicted on the lower castes in Kerala’s Muslim (and of course Hindu) society also continued in the Laccadive Islands, after the islands’ link with Hindu society had been severed. The caste hierarchies prevailing in the islands have been studied in some detail by Leela Dube, one more contributor to the book edited by Imtiaz Ahmed. The aristocracy, called Karnavars are descendants of Nambudiris and Nayars. They are also referred to by the respectful appellation, Koya, which means a religious dignitary. It was this class that monopolized land- and boat-owning. The Malumis or Urukkars formed the sailor-caste, and the Melacheris (literally, tree-climbers) formed a class of serfs, who earned their livelihood by “plucking coconuts, tilling their lords’ lands, rowing their boats.” The number of castes or classes varied from island to island, some places having four instead of three. One island, Agatti, was regarded as a Melacheri island.

In view of evidence such as has been adduced above, Gaborieau argues that a hierarchical ordering is quite characteristic of Islamic society. He points out that Hindu society is not unique in holding some occupations in ritual opprobrium:

The influence of the Indian context is often invoked to explain the opprobrium suffered by the barber, the weaver, the butcher, the tanner, the sweeper,18 etc. These are professions stigmatized as being low and even impure. The last two are even regarded as untouchable. But this interpretation is short-sighted: the Hindu ideology only confirms prejudices formerly more widespread spread throughout the world. The tanner who handles animal carcasses is universally detested; the Hajjam who is the scarifier (and, among the Muslims, the circumcizer), as well as the barber are looked down upon in the Talmud as well as under the Sassanians, among Hindus and among Muslims. The contempt towards weavers also dates to remote times. Here we are dealing with traditions which date back to prehistoric times and can be found very different civilizations (for all these questions, see the important documentation analyzed by Brunschvig, 1962, 46-57.) This fact demolishes the contrary argument according to which Islam in India is supposed to have rehabilitated the depressed castes – the professions which are cited as examples, the barbers, weavers, and tanners, are precisely those whose inferior status is most explicitly affirmed in the Islamic tradition. The stereotypes attached to them in India are widely attested in Islam and in the most ancient traditions: the barber is greedy and arrogant; the weaver – wretched, stupid and treacherous. The conversion to Islam of those who practiced these professions unquestionably opened for them new economic outlets – and this is precisely the reason for their conversion; their conversion in no way improved their status.19

In addition to the material already available in the literature, Gaborieau brings to the discussion the results of his own long years of field work in Nepal. He has seen Muslim Curautes redoing their ritual ablutions if they happened to touch a [Muslim] untouchable by mistake. He has also studied the phenomenon of ritual uncleanness associated with some professions, and social hierarchy based on profession, at work in Muslim society. One of his examples concerns the Kashmiri Muslims in Nepal who pass for Ashraf. Periodically, these high born Muslims send for a barber from India, at great expense. However, the barber becomes wealthy, and turns his back on the profession in favour of something more respectable. He refuses to perform circumcisions, and the need for another barber is acutely felt. A new barber is sent for, and he despised, he faces the same stereotypes, and the cycle is repeated. The stereotypes, applied to barbers and weavers, are an old Islamic tradition.

While individual social mobility is attested, collective mobility is virtually impossible, because there is a kind of “barrier” separating the Ashraf castes from the artisan castes:

Nowhere have I sensed this barrier as strongly as in my field work in Nepal. The oldest Ashraf of Kathmandu, the Kashmiri, traditionalists and devoted to the cult of the saints, totally refuse all socially significant transactions with other Muslims of low status from the valley, who are collectively called Hindustani and who are recruited from various artisan castes. At the most they sometimes accept, under the rubric of hypergamy, some of their daughters as secondary wives who are never any more than concubines. A primary marriage would be unthinkable. The Kashmiris have always been opposed, even in multiple proceedings in front of Nepalese tribunals composed of Hindus, to having common mosques and even a common cemetery with the Hindustani . This is a clinching argument when we remember that the total number of Muslims in the valley does not ever exceed twelve hundred persons. The Hindustani may well be reformed, instructed in religion and devout, they can never cross the barrier (Gaborieau, 1977a, 52-54).20

Gaborieau’s studies of conversions into the Muslim Curaute caste contradict the theory that conversion has taken place in India selectively from the lowest orders. The Curaute admit conversions into their caste from Hindus of higher castes such as Chettri and Gurung. But they do not accept untouchables into their caste.

Indeed even conversion effected by the Sufis does not seem to wipe out the stigma of untouchability. Gaborieau points out that

the Ashraf monopolize the so-called orthodox Sufi brotherhoods (ba-shar’), as opposed to the “heterodox” brotherhoods (be-shar’) who are relegated to the level of the lowest castes (Gaborieau, 1986c)21

and also that

the heterodox brotherhoods are lower than and subservient to the former, so much so that the musicians (qawwal), who pronounce the mystic chants (qawwali) and play the drum, are in fact untouchable musicians (as in the Hindu temples) even though they claim affiliation with a Sufi order. The heterodox Sufis and these musicians are relegated to the far corners of the shrines dedicated to Muslim saints.22

A caste-like phenomenon exists in the Punjab with the Chishtis forming a hereditary clan, controlling not only the tomb of their ancestor Baba Farid-ud-din (d. 1265) at Pak-pattan, but also the lands and the cultivators surrounding it.23 Clan members had justified their hypergamous codes of marriages even in front of British courts. The noted historian Richard Eaton rejected these claims as being contrary to Islamic tradition, and as reflecting the influence of the ambient Hindu culture. Gaborieau disagrees with his famous colleague, pointing out that such hypergamous traditions are completely in consonance with worldwide Islamic practice. According to Gaborieau, Robert Brunschvig (1962, 55) had long ago compared the laws of the Manu Smriti to Islamic tradition:

If a young girl likes a man of a class higher than her own, the king should not make her pay the slightest fine; but if she unites herself with a man of inferior birth, she should be imprisoned in her house, and paced under guard. A man of low origin who courts a maiden of high birth deserves a capital punishment. (Laws of Manu (VIII, 365-366))

Following Brunschvig, Gaborieau claims that this law is exactly that which the dominant Hanafite tradition of jurisprudence would require,24 as has been spelled out on in the famous compendium of the Moghul period titled Fatawa-i-‘alamgiriyya. The idea that a woman can only marry a man of equal or higher status has been upheld as late as the twentieth century by scholars even of the eminence of Ashraf ‘Ali Thanawi (d. 1943). The marriage of a woman to a man of lower status can be annulled by the family’s request, or by a Qazi’s judgment.

Gaborieau calls for a frankness in studying the phenomenon of caste in Indian Muslim society. The Muslims who entered did not seem to be shocked by the institution of caste, and if they were not shocked by it, it must be that they were not unfamiliar with such arrangements themselves. Even writers such as Ansari (whom we have quoted on the first page), who trace the caste inequalities in Indian Muslim society to Hindu influence, admit however that Islam was not egalitarian when it entered India.

The ideal of equality among Muslims was practicable only in the then prevailing conditions of Arabia. In the course of the expansion of Islam and its contact with other complex cultures the democratic forms of political organization and social equality within the community gradually disappeared.25

He then traces the origin of caste to the “Indo-Iranian community”. Ansari declares that though “Islam proclaimed the message of equality and universal brotherhood”, “the established and deep rooted institution of social segregation in Persia” eventually won out.

Even the reputed Muslim scholars of Persia, like Nasir-ud-Din at-Tusi preached the division of society; his classification of society remained the same as it was during the Sasanian period. In his book, Akhlaq-i-Nasiri(which was finished shortly before the fall of the Caliphate), at-Tusi considers that each of the social classes should be kept in its proper place. A seventeenth-century work, Jami-i-Mufidi, again retains the same four-fold division of society, but it puts forward a slight change in giving precedence to warriors at the top and reducing the relative rank of priests to that of second in the hierarchy. In addition to these philosophers, the noted statesman of Persia, Nizam-ul-Mulk, in his Siyasat Nama, instructs his subordinates to maintain the people in their proper ranks.26

The idea of social hierarchy, Ansari says, had already become part of Islamic society by the time it entered India in the twelfth century. Over the centuries attitudes only hardened, until at last even untouchability entered Islamic society. The plight of Muslim untouchables is described by Ansari in moving detail:

A Bhangi, either Muslim or non-Muslim, is not permitted to enter a mosque no matter how clean he may be at the time. Although in theory a Muslim Bhangi or Chamar is allowed to offer his prayer[s] in a mosque, but in usual practice their entrance into such pious places as mosques and shrines of Muslim saints is socially disapproved and thus it is resisted. Even if they could get into a mosque or shrine, provided they have had a bath and are dressed in clean clothes, they do not usually proceed beyond the entrance steps. In contrast to the Hindu caste system, Muslim Bhangis are allowed to learn the Quran, but they are not expected to teach it.

It is a common practice observed in almost all the households of Ashraf, Muslim Rajputs, and the clean occupational castes, that Bhangis, either Muslim or non-Muslim, are generally served food in their own containers. If they do not have their own bowls they are served in clay pots which are not again used to serve clean caste members. Bhangis are given water to drink in such a way that the jar does not touch even their lips.27

However Ansari never explains how caste structures in India can be attributed to Hindu influence alone, if Muslim society had also stratified into hierarchies, before Islam’s advent in India. Several such problems in the literature need explanation. Gaborieau again offers perspective:

… While we have good contemporary studies of Hindu untouchables, no work was done on Muslim untouchables during the colonial period. The absence of work on this key point deserves reflection. This refusal to consider the reality is understandable on the part of Muslim scholars; the problem of untouchability clashes against their ideological convictions on the ecumenical character of Islam. And what is more, any conversion even of untouchables involves burning political complications. On the part of western researchers, this omission is less excusable: I regard it as a manifestation of the prejudice according to which Muslim social order must necessarily obey a different logic than the Hindu social order, and also by the illusion of believing that the enumeration of castes is done from top down, whereas in reality it happens from the bottom up, starting from the untouchables.28

Gaborieau also explains that advent of Islam did not spell the demise of hierarchical structures in Indian society. The Muslims allowed the hierarchical structures to remain, and were not even shocked by it; not only that, they occupied the apex of the pyramid (in the form of the Ashraf castes), without otherwise undermining it.29

Caste in the Islamic World

We have already seen from the examples of the Ashraf‘s practices regarding marriage, or admittance to mystic brotherhoods, etc., the Ashraf also retained their own stereotypes and prejudices which cannot be traced solely to Hindu influence.

But that is not the whole story. Even if the caste structure was largely a relic from the pre-Islamic past, new castes also sometimes came into existence. The Maratha Bugtis in Balochistan are an interesting case of what may be a caste forming even under Islamic rule. Theirs is a clan claiming descent from Marathas captives of war brought back by members of the Bugti tribe, who served the armies of Ahmad Shah Durrani (Abdali) after the fateful battle of Panipat.

In time they underwent ‘Bugti-ization’and became Muslims. Although for all practical purposes they may now be considered Bugtis, and are even in the forefront in education and employment, they were once considered little better than bonded labour. They could not own or buy land. Up to two generations ago they could be ‘bought’ for twenty or thirty rupees. Their women were fair game for Bugtis.

The Maratha Bugtis took jobs as unskilled labourers, which their tribal overlords disdained. Over the years they have come to occupy higher positions, and their prosperity is resented by the Bugtis.30 It is interesting to note that this caste-like phenomenon has endured for more than two centuries, even in a region largely devoid of Hindus.

The Maratha Bugtis were not alone in their position as a group living in the Islamic world, with their inferior position determined by heredity. The Haratin31 or Harratin of southwestern Morocco and Mauritania are “a socially and ethnically distinct class of workers”. They are descended from slaves, but are now serfs, “without the privileges of freedom”. (One of the people who is trying to help them to become independent is Abdel Nasser Ould Yessa, whose life and work is discussed at the following web site:

The facile practice of regarding all hierarchies in the Islamic world as a substratum from pre-Islamic societies does not always work. Hierarchies (in other words, castes) exist even in places like Yemen and the rest of the Arabian peninsula.

As a perusal of the informative entry on “Bedouin” in the Encyclopaedia Brittanica32 reveals, Bedouin society in twentieth-century Arabia was also divided into various groups. While the nomads have been settled after the formation of the modern states, the societal hierarchical and patriarchal structure has been retained. The Bedouin tribes were classified on the basis of the species of animal on which they depended. Camel nomads were highest in prestige. They were spread on extensive territories in the Sahara, Syrian and Arabian deserts. Sheep- and goat-herding nomads, rank below, and live closer the cultivated zones in Jordan, Syria and Iraq. The noble tribes are proud of their ancestry, and are divided into “Qaysi” (northern Arabian) or “Yamani” groups. In addition to the noble elements, the Bedouin society also includes vassal tribes, which are “ancestorless” (i.e., tribes whose heredity is not prestigious). These groups are subservient to the noble tribes and include professional groups such as artisans, blacksmiths, entertainers, etc.

Caste-like phenomena are attested in other regions of the Arabian peninsula, even among the sedentary populations. Paul Dresch has studied the situation in Yemeni tribal society at the beginning of the twentieth century.33 He observes that two groups of people are widely regarded as not belonging to the tribe, but are still endowed with rights and obligations. The first of them is the Sayyids – a group claiming descent from the prophet, and the Qadis. (The Qadis are also a group defined by heredity. While elsewhere in the Islamic world the title Qadi refers to judges, in Yemen it only denotes a member of this class, whether judge or not. The Qadis or mashaykh are also said to be descended from the Prophet Hud. The mashaykh do not enjoy as much prestige as the Sayyids.34) Below the tribesmen rank the ‘weak’ people (dua’fa) (sing. da’if). Weak people have no prestige. They include people of various trades, some respectable and some not so respectable.

Artisans and merchants in the traditional towns tend to be highly organized into castelike guild groups that are ranked largely according to the nature of their craft. In many areas those who ply so-called respectable trades are sharply differentiated from the bani khoms, or sons of the five, practitioners of the five despised trades of barber, bloodletter, butcher, bath attendant, and tanner. In the Hadramaut artisans who handle clay, such as masons and potters, also fall into the despised group, as do sweepers, fishermen, and some others, depending on locality. Poor farm laborers also occupy a low status, but it is higher than those of the despised crafts.

The akhdam, in many areas the lowest group, are so isolated from society that they have been compared with the untouchables of India. Found especially along the Tihama coast and in southern Yemen (Sana) but also in the Hadramaut, they are often distinguished socially by their negroid appearance and often follow the despised trade of sweeper. The akhdam appear to be descendants of slaves, although not all former slaves occupy such degraded positions. Slavery existed in the territories of the Aden Protectorate until the 1930s and persisted in Yemen (Sana) until 1962.35

The Sayyids in Yemen did not allow intermarriage with other Yemeni castes. This superiority was challenged only by expatriates in Singapore in 1905, and again under the Irshadi movement in Java in 1915.36

It would thus seem that the practice of forming hierarchical structures is quite a widespread phenomenon. In fact, not only hierarchies, but the specific practice of untouchability is attested in Burma and Japan. The idea of pollution by contact is attested in Qajar Iran, to cite but a single example.37 An exhaustive comparative study of all the different phenomena that are usually subsumed under the notion of caste is yet to be undertaken. Home hierarchichus is not endemic only to India; Homo sapiens everywhere has mostly been Home hierarchicus.

Even the institution of caste may have served its purpose in India. A.L. Basham regards the institution of caste as having been directly responsible for the survival of skills, handed down from generation to generation in India, whose counterparts were lost in other countries.38 Yet over the years it has become an oppressive institution, and a great obstacle to human freedom and national integration.

While studying caste it is certainly important not to lose one’s perspective in the zeal to undermine the caste system. It is counter-productive to demonize one religious community as inherently caste-ridden and inegalitarian, and absolve other hierarchically structured religious groups of their responsibility. Indeed, the exclusive identification of caste with Hinduism has caused the situation in Pakistan and in Indian Muslim society to be largely neglected. Several important facts have gone virtually unnoticed.

Gaborieau points out that Syed Ahmad Khan, spearhead of Muslim thought in the Indian subcontinent in the last century, was an Ashraf working for the welfare of the Ashraf. He used to say that his Aligarh college was not for weavers. The Muslim League’s social program was copied off the Congress’ program, and made no radical improvements. The Congress however, has been labeled a baniya party, and that is how it has been portrayed to generations of Pakistani students. Land reforms in Pakistani Punjab (1959, 1972) have not been as successful as in Indian Punjab, and large landholders still have a disproportionate share of the land.39 A human rights commentator points out that Bhutto’s land reforms were ‘cosmetic’, because landowners had been previously warded to transfer their lands to their family members.40 This has not provoked the public outrage that it should have. The same commentator also points out that the ulama have not campaigned for the eradication of feudalism. Thus, even in the year 2002 the situation seems to be no different that which obtained in the early forties, when the peasants of Punjab and especially of Sindh were

“under the spell of the pirs,” and “had imbibed the doctrine of taqdir (fate) from the constant preachings of the pirs”, … whose message was “He is low forever because God has made him so.””41

The advent of Islam has not automatically ensured equality. Indeed, the example of former Prime Minister Bhutto shows that inequality continued to be rampant. Bhutto’s family owned, according to his own admission, hundreds of thousands of acres of land in Sindh and for generations. His ancestor Sheto had obtained tax exemptions and other benefits, including the title of khan, from Aurangzeb.42 Populists such as Bhutto have been able to get away with the rhetoric of musawat (equality), which he held out in his electoral promises, while doing nothing to promote social reform.

In a future article we will look at the issue of the development of vernacular languages on the subcontinent. The idea that Hindu society is inherently more hierarchical has led to the blanket assertion that Sanskrit stifled the growth of regional languages, which needed the stimulus of Islamic society before they could come into their own. This claim is also quite untenable. But here we content ourselves with the remark that in the subcontinent it was again India that took the lead in promoting regional languages. The regional languages in Pakistan have not been able to acquire parity of status with the Ashraf-imposed Urdu.43

It was left to the outspoken Dr. Ambedkar to point out that responsibility of fighting the iniquities of the caste system on the subcontinent devolve equally on Muslim and Hindu:

The existence of these evils among the Muslims is distressing enough. But far more distressing is the fact that there is no organized movement of social reform among the Musalmans of India on a scale sufficient to bring about their eradication. The Hindus have their social evils. But there is this relieving feature about them – namely, that some of them are conscious of their existence and a few of them are actively agitating for their removal. The Muslims, on the other hand, do not realize that they are evils and consequently do not agitate for their removal.44

The situation has changed much since the days when Dr. Ambedkar wrote these words. The battle against inequality on the subcontinent, however, is far from won. There is every indication that the battle will yet prove to be long and costly. As a very first step, it is hoped that intellectuals will rise up to examine the institution of caste with an unbiased mind, and rid us of all illusions in this matter.


1. This study is the work of secondary scholarship. The author claims no originality, as the reader can see from the many quotes. A reexamination of the problems of the caste system has long been underway. It is the author’s hope that this reexamination will now be debated vigorously on the internet.

The casual reader no less than the specialist will notice that the word caste is used throughout in a loose sense, meaning a population group that is defined by heredity. We have not examined very minutely whether the principle governing caste relations are based, for instance, on notions of purity/pollution, or occupation. Problems such as a precise definition of ‘caste’, the question whether the Ashraf are really a caste, or a comparison between notions of purity and pollution among Hindus and among other religious groups, will be discussed in future articles.

2. J.C. Heesterman, The Inner Conflict of Tradition, The University of Chicago Press, 1985.

3. Caste in the Punjab, From the Census Report of the Punjab, 1881, by Sir Denzil Ibbetson, K.C.S.I.

4. Muslim Caste in Uttar Pradesh (A Study of Culture Contact), Ghaus Ansari, Lucknow, 1960, Page 66.

5. Ni brahmanes ni ancêtres: Colporteurs musulmans du Nepal by Marc Gaborieau, Nanterre, Société d’ethnologie (It must be pointed out that there have been several British scholars who did not take this view. Sir Denzil Ibbetson is one. Gaborieau’s point is valid broadly speaking.)

6. See, for instance, Sir W.W. Hunter, “The Religions of India”, The Times, (London), February 25, 1888.

7. Dr. James Wise, “The Muhammadans of Eastern Bengal”, J.R.A.S.B, (Calcutta: Asiatic Society of Bengal, 1894), No. 1, p. 32.

8. Pakistan or The Partition of India, B. R. Ambedkar, Thacker & Co., Ltd., Bombay, pp. 218-219

9. Caste among the Muslims of Calcutta, M.K.A. Siddiqui, in Caste and Social Stratification among the Muslims, Imtiaz Ahmad (ed.) (see next footnote).

10. Caste and Social Stratification among the Muslims, Imtiaz Ahmad (ed.), Manohar, 1973

11. M.K.A. Siddiqui, in Caste and Social Stratification among the Muslims … op.cit., pp. 149-150.

12. Status and Power in a Muslim Dominated Village of Uttar Pradesh, Zarina Bhatty, in Caste and Social Stratification among the Muslims … op.cit.

13. ibid., p. 95.

14. Status Groups among the Moplahs on the South-west Coast of India, Victor S. D’Souza, in Caste… op.cit., pp. 45-60.

15. Gaborieau, op.cit., p. 266.

16. Gaborieau, op.cit., p. 378.

17. Quoted in Caste in India: Its Nature, Function, and Origins by J.H. Hutton, Oxford University press, 1963, p. 219.

18. In Pakistan many Christians from farming communities became landless after independence, and had to become sweepers by profession. This has caused them to be further stigmatized, indicating that dignity of labor is not yet widely upheld, even though the region is nearly devoid of Hindus. (Religious Minorities in Pakistan, Dr. Iftikhar H. Malik, 2002, Minority Rights Group International, p. 12)

19. Gaborieau, op.cit., p. 370.

20. Gaborieau, op.cit., p. 383.

21. Gaborieau, op.cit., p. 378.

22. Gaborieau, op.cit., p. 354.

23. Gaborieau, op.cit., p. 291, 356.

Gaborieau also gives us the example of the Nizami clan, which manages the tomb of Nizamu’d-Din Auliya’. This clan claims descent from and intermarries with high-born Sayyid. The clan has inherited the mystique and sanctity attached to the founder of the hospice. (Gaborieau, op.cit., p. 375).

It appears that the phenomenon of Pirs and Pirzadas is caste-like, in that sanctity and prestige are inherited by birth. In Caste in India, J.H. Hutton gives the example of Pir Pagaro in Sindh, who is a “hereditary religious leader descended from a family which entered Sind with the Arabs in AD 711.”

24. Gaborieau, op.cit., p. 356.

25. Ansari, p. 28.

26. Ansari, p. 30.

27. Muslim Caste in Uttar Pradesh (A Study of Culture Contact), Ghaus Ansari, Lucknow, 1960, Page 50.

28. Gaborieau, op. cit., p. 393.

29. Gaborieau, op. cit., p. 387, p. 415.

30. Marginality and Modernity: Ethnicity and Change in Post-Colonial Balochistan, Paul Titus, (ed) , Oxford University Press, Karachi, 1996, pp. 54-55.

31. “Haratin” http://www.britannica.com/eb/article?eu=40019

32. <http://www.britannica.com/eb/article?eu=14268>

33. Tribes, Government, and History in Yemen, Paul Dresch, Clarendon Press, Oxford, 1989, p. 117.

34. Area Handbook For The Yemens, Richard F. Nyrop, et al., 1977, p. 74.

35. Area Handbook For The Yemens, Richard F. Nyrop, et al., 1977, p. 77.

36. Tribes, Government, and History in Yemen, Paul Dresch, Clarendon Press, Oxford, 1989, p. 27.

37. Iran and the Muslim World: Resistance and Revolution, Nikki R. Keddie, Macmillan Press Ltd, 1995, p. 147.

38. Certainly this is true of music in the lands west of India. Until very recently recording music has not been possible. The musicians in Baluchistan, Iran and Afghanistan happen to be hereditary caste-musicians (Doms, Jugis, Loris) all with Indian caste counterparts. These musician castes will be studied in a future article. It will be seen that they are in all probability descendants of Indians brought as slaves or serfs. These castes completely dominate the folk music scene in Baluchistan and Afghanistan.

39. Pakistan: Nationalism without a Nation?, Christophe Jaffrelot (ed.), Manohar, New Delhi, 2002, pp. 59-60.

40. Religious Minorities in Pakistan, Dr. Iftikhar H. Malik, 2002, Minority Rights Group International, p. 7.

41. Politics in Pakistan: The Nature and Direction of Change, Khalid B. Sayeed, Praeger Special Studies, 1980, p. 7

42. Zulfi Bhutto of Pakistan: His Life and Times, Stanley Wolpert, Oxford University Press, 1993, pp. 3-4.

43. As late as the 1901 census, it was only the Ashraf who claimed Urdu as their mother-tongue, even in the Gangetic belt. The others spoke one of many regional languages which have all been declared ‘dialects’ of Hindi/Urdu. (In fact, an authority on the Muslim League estimates the population of Ashraf to be about 7 percent of the total Muslim population, based on the number of people claiming Urdu as their language. See The All India Muslim League, Muhammad Saleem Ahmad, 1988, Ilham Publishers, Bahawalpur.)

44. Pakistan or The Partition of India, B. R. Ambedkar, Thacker & Co., Ltd., Bombay, p. 223.

Dileep Karanth dskaranth@yahoo.com