Unlike the caste system specific to Hinduism, which has been the subject of many anthropological debates, “caste” in South Asian Islam has not received as much attention from specialists of the Indian sub-continent. And yet, in this region of the world, which currently has the highest concentration of Muslims (around 480 million), India stands out from those of its closest neighbours with a Muslim minority on account of its sheer demographic weight. Its Muslims number almost 140 million, in other words 13.4% of the population according to the 2001 census, making India the third Muslim country after Indonesia and Pakistan, where Islam is by far the dominant religion. 
The social organization of Indian Muslims, while sharing many general characteristics with Arab and Middle Eastern societies (patrilineality, lineage, marriage), can nevertheless be distinguished from them by the vernacular caste categories it reflects. However, it can only be partly likened to the Hindu caste system, which is primarily based on criteria of ritual purity and hierarchization particular to Hinduism. For a time, the opposition between egalitarian Islam and hierarchical Hinduism was the cornerstone of the debate between supporters and opponents of Dumontian theory. Many studies have shown that Muslim society was also strongly hierarchized and divided into social groups of varying status.
This article will review the apparent reappropriation of a specifically Hindu institution by the Muslims of South Asia. We shall first examine how this particular system of stratification by status works, and how South Asian Islam has built up a body of sources that legitimize the Muslim social order, before looking at the way in which it has given rise to a number of struggles over classification. Finally, we shall complete this introduction to Muslim castes by presenting the debates that centre on the future inclusion of Muslims in the reservation system (quotas) used by the Indian government.
Sociology of the caste and South Asian Islam
According to Louis Dumont’s theory, the religious ideology of Islam directly established a social order structured by castes (jati), which are determined according to the criteria of endogamy, hereditary professional specialization and hierarchical relationships defined by status, which together form a system. The author believes that this social order is not only fully established within the Hindu environment; externally, in areas far outside Hinduism’s sphere of influence, castes can exist but are often weak or incomplete.  This theory has fed many debates on the origin of caste in South Asian Islam, which centre on supposed structural analogies between the Hindu caste system, Middle Eastern society and even Central Asian society. However, while acknowledging that these forms of social organization share common features, and that the sociological specificities of South Asian Islam are the result of contact between them, here we shall show the underlying tension within Muslim society between a clearly hierarchized social system and an egalitarian view of Islam.
While Louis Dumont put forward religious Hindu ideology – particularly the opposition between the pure and the impure – as the primary foundational element of the caste hierarchy, normative Islam, on the other hand, today claims to have an egalitarian social system. In the Muslim world in general, as in India and South Asia, that view is legitimized by a Koranic concept frequently used in the practice of Muslim marriage. It is based on the legal principle of equivalence, compatibility or equality of status (kafa’ah) between the partners in a union. Determined by birth group or occupation, it also constitutes one of the basic ways of controlling social relations. Since equality between Muslims applies only if at least the father and grandfather are Muslims, marriage therefore becomes impossible between a young Muslim whose family converted fewer than two generations ago and another Muslim.
However, this egalitarian view of Islam contrasts sharply with the functioning of Muslim society in the Indian sub-continent, whether medieval or modern, in which principles of hierarchization and deeply unequal relations govern. In his study on a Muslim minority in Nepal, Marc Gaborieau suggests doing away with our theoretical standpoint that puts forward ideology as the only founding authority of social order. He proposes that we should focus on observing the sphere of social relations and therefore the vernacular categories it reflects, while seeking similarities or differences with the Hindu caste system.  For Gaborieau, there is indeed a caste system and hierarchy among the Muslims he studied, whereas Louis Dumont preferred to describe their social organization in terms of “groups of graduated status” as a means of distinguishing it from the Hindu system.  From a sociological standpoint, the opposition between an egalitarian Islam and a hierarchical Hinduism does not help us to understand contemporary Muslim society, especially since this egalitarian view has only been put forward by some Muslim ideologists since the end of the 19th century, during the socioreligious reforms.
Despite those who attribute this principle of equal status (kafā’ah) between Muslims to the Koran and certain hadiths, and who hold it up as a principle that was integral to the original Islam, it should be remembered that the war of succession to the Prophet in the 7th century was based on family, tribal and therefore political rivalries. From then on, belonging to the Prophet’s close family (ahl al-bayt), clan or tribe became criteria for social differentiation within Arab society. The exporting of those criteria beyond the borders of the Arab peninsula, particularly into the Indian sub-continent from the 8th century, gave rise to another element of distinction, this time between Arabs and non-Arabs, and within this latter group between the previously converted (khadim-al islam), from the time of the very first waves of Islamization, and the newly converted (jadid-al islam). Still today, these criteria for distinction divide the Muslim social space in the Indian sub-continent into the higher castes (unch zat), of Arab origin, and the castes of lower status (nich zat), made up of descendents of those who converted to Islam.
Furthermore, the criteria for dividing Muslim society in India were not incorporated into the corpus of vernacular Islamic sources until fairly late on. The distinction between the Ashraf nobility, in other words the social class of individuals belonging to a lineage dating back to the time of the Prophet, and the lowest stratum (Arzals) only appeared in a number of texts from the 13th century onwards, when a sustainable political power was established in the region through the Delhi Sultanate. Subsequently, during the prosperous time of the Mughal Empire, from the 16th century to the end of the 18th century, the sovereigns (Sultans) who succeeded one another, all from a high caste, adopted very different positions with regard to the social hierarchy and discrimination involved in the recruitment of administrative officers, as well as the subject of forced conversions to Islam. Throughout that period of Muslim presence, many Muslim chroniclers recounted episodes of social discrimination based on whether or not people belonged to the Prophet’s lineage. 
In modern-day Muslim society, the principle of kafa’ah has become a factor causing social division and hierarchization by status. Certain reformist schools of thought such as the Deobondi and Barelvi movements have often made use of this concept even today to legitimize the importance of caste in South Asian Islam. Given that kafa’ah is hereditary, it enables the superiority of Ashrafs in relation to other groups within the Muslim social space to continue, beyond its role in drawing up marriage contracts.
The multiplicity of categories for social identification
Far from being homogeneous, for over 100 years the largest religious minority in India has been structured by three main schools of thought represented by the Barelvis, Deobondis and Ahl al-Hadiths, as well as the Twelver and Ismaili Shiite groups.  In the Muslim social organization of South Asia, three main hierarchical divisions emerge (Ashraf, Ajlaf, Arzal), within which we find many social units that are interdependent, more or less endogamous, of varying size, similar to Hindu castes and sub-castes, and unevenly distributed across Indian territory.
At the top of the hierarchy are the Ashrafs (nobles), of Arab, Persian, Turkish or Afghan origin, who lay claim to a prestigious lineage that can sometimes be traced back to the Prophet (in the case of Sayyids) or his tribe (in the case of Qureshis) , and recognized as such by society. The Shaikhs (descendants of the Prophet’s companions), Pathans (descendants of migrants from Afghanistan), and even Mughals (originating in Central Asia and Iran) can also be included in this group. Many Ashrafs are either ulamas in the case of the Sayyids, or else landowners, merchants or business people. One’s birth group constitutes a major criterion for defining social status, and the distinction between Arabs and non-Arabs remains fundamental; scholars from the Hanafi school of jurisprudence followed by the Shafi’i school , approved this principle of differentiation between groups at the turn of the 20th century.
At the middle level, Ajlafs (low-born) represent the masses, whose status is defined by both their profession (pesha) – unlike the Ashrafs – and their identity as descendents of converts to Islam. Many castes of intermediate status fall into this category, such as farmers, traders and weavers (Ansari and Julaha). The latter category is often mentioned in recent textual sources legitimizing the Muslim social order, particularly those written by doctors of the law or theologians (ulemas) who usually belong to the Sayyid caste, or else it is used pejoratively in daily life to indicate the entire Ajlaf caste. Among the social elite, many Ashrafs – still today in rural areas – even believe that this category is not part of the Indian Muslim community (millat) and should remain outside of any emancipation process.
At the bottom of the social scale come the Arzals (vile, vulgar), in other words a group comprising non-Untouchables and converted “Untouchables” who, as in Hinduism, practise supposedly impure trades. This was the case of slaughterers, laundrymen (Dhobi), barbers (Nai, Hajjam), tanners (Chamar) and so on. It is important to highlight the fact that, like the Hindu caste society, relations between Muslim social groups are governed by a series of social taboos (sharing a table, marriage, sociability) and spatial restrictions (access to domestic areas and places of prayer, segregation in cemeteries and neighbourhoods). However, these taboos, which aim to distance the high castes from the low, are not based on the notion of ritual contamination in the strictest sense by which it is defined in Hinduism.  They are relaxed considerably in urban areas.
Furthermore, this tripartite distribution is not found in all regions of India. In Tamil Nadu , in southern India, several authors agree that there was no opposition between Ashrafs and Ajlafs. Mattison Mines, for example, believes that the hierarchy of status between the Muslim social groups of Tamil Nadu is not so much the result of belonging to a caste as the level of socioeconomic development of a group.  Other factors involved in hierarchization, apart from belonging to a birth group, can therefore come into play when defining the social status of individuals and groups, such as the economic situation of a household and level of education. Similarly, in many other regions such as Kashmir and Uttar Pradesh, the three aforementioned categories do not appear to form part of local sociological vocabulary; they therefore tell us very little about the functioning of Muslim society. 
Indeed, these three main hierarchized categories then subdivide into many groups and subgroups, which make up true categories for social identification at local level, and whose denomination and proportion also vary by region. These social groups, which can be referred to as castes given that they have a strong hierarchy among themselves according to geographical origin, birth group and professional occupation, are identified by a series of terms, often of Arab or Persian origin and usually interchangeable: jati (“hope”, birth group for Hindus), or zat (identical in Urdu), qaum (clan, community, lineage, tribe, nation) and jama’at (group, community, association). At local level, the terms khandan (lineage, family, dynasty) and nasab (lineage, line or any group founded on blood ties) certainly remain those most used by individuals to distinguish themselves in the social space, given that lineage constitutes the reference unit for choosing the partners in a union. Several lineages combined can form a biraderi in one or more localities, in other words a “marriage circle” within which marital relationships are formed.
The sociological content of these categories together has varied a great deal over the years, particularly since the Mughal period and British colonization. The social usage made of them today usually depends on the context in which they are expressed and on political, local or regional issues.
Strategies for symbolic social advancement
“The first year, I was a weaver (julaha),
The following year, I was a Shaikh,
This year, if prices rise, I will be a Sayyid” 
In this social order, which is strongly characterized by discrimination according to a person’s affiliation with their group of origin, the dichotomy between high and low castes polarizes vertical relations in the Muslim social space even today, particularly in rural areas. Many different strategies have been implemented at various times in India’s history in order to enable people to improve their status in the hierarchy and thus ensure that they can reposition themselves, usually in a related area, within the social space.
A kind of upward social mobility is nevertheless possible within Muslim society. By analogy with “sanscritization” as defined by the Indian sociologist Srinivas at the beginning of the 1970s and applied to Hindu society, the process of “Ashrafization” saw individuals or whole groups adopting new social and ritual practices, taking on names and titles from high castes, sometimes rewriting the group’s history and giving themselves a new community genealogy as a way of symbolically improving their social status, here defined by their affiliation with a caste. This was particularly the case with many Muhajirs, after the Muslims left India following the 1947 partition. When they arrived in West Pakistan, they hastened to change their surnames, taking on the names Sayyid or Qureshi, for example, in order to lay claim to a more prestigious lineage. Others played on the proximity of the names Ansars (descendents of Medina, ashrafs) and Ansaris (caste of weavers, ajlafs). The mass migrations such as those caused by the partition brought about quite exceptional circumstances that facilitated the implementation of these strategies.
In the Muslim world in general, as in Muslim India and South Asia, endogamous marriage is predominant, practised within one’s group of origin (zat or biraderi), between first or second parallel cousins. Endogamy also usually takes on a territorial dimension if the “marriage circle” is limited to one or several localities, which form a territory structured by networks of marriage alliances and over which the authority of a caste council (zat panchayat) can be exercised. In most cases, endogamy takes place within the limits of each of the three main abovementioned categories, especially among ashrafs, which does not exclude the possibility of intermarriages between members of different castes within those categories. This type of marriage was inevitable between, for example, Arab merchants and Indian women who spoke Tamil or indeed Urdu when those women came from the north of India, following the expansion of the Mughal Empire towards the south from the 17th century onwards. Intermarriages were also common between Pathans who had migrated from their territory of origin to northwest Pakistan, where territorial endogamy was strictly observed for women, and non-Pathans who had settled in the south of India. Moreover, hypergamous marriage for ashrafs, in other words the union of a woman with a man of higher status such as a Sayyid, was a factor that improved one’s position in the social hierarchy. J. B. P. Moore showed this in his study on the Sayyids of Madurai in Tamil Nadu. Marrying a Sayyid enabled ties with the Prophet’s bloodline to be re-established, but more than the purity of the bloodline it was hypergamous marriage and the male descendants that determined a Sayyid’s status.  Today there are not enough studies on the practices of Muslim marriage in the Indian subcontinent. However, some practices have already shown a certain level of permeability, even extension, as regards the boundaries of caste endogamy. This can be seen from the practice of exogamous marriage, which is not unusual among castes of low status in urban areas.
Another form of the “Ashrafization” process takes place through the establishment of caste associations that aim to defend the community’s interests, as many Hindu castes of low status did in the early part of the 20th century, following the first censuses recorded at the end of the 19th century. However, Muslim social groups did not use this method of repositioning themselves in the social and economic space until much later, and in fewer numbers. The case of the Khojas, a merchant caste of Ismaili Shiites well established in the Delta area of Sindh and in Karachi, the economic capital of Pakistan, illustrates this phenomenon particularly well. The historian Michel Boivin has clearly shown how this caste was forced to reorganize itself over the course of the 20th century in order to maintain its social autonomy when faced with the religious authority of the Ismaili Imam, the Aga Khan, while pledging their religious allegiance to him. One of the strategies used was to create an association providing community support, the jama’at, a term widely used today by members of the community instead of the word zat. These associations, established mainly by groups like the Khojas but also Memons and Bohras in Sindh (Pakistan) and Gujarat (India), helped to renew traditional networks of social solidarity while integrating lower castes, and to strengthen their position in the social space thanks to their newfound economic influence. 
With the exception of this last example, and other isolated cases of social success, the attempts at upward social mobility described above have not necessarily resulted in a higher standard of living or social recognition at local level. Whether descendants of converts who have added the title of Shaikh to their name, or Muslim outcasts or “Untouchables” practising so-called impure trades, it still seems difficult to be free of one’s socioprofessional identity linked to one’s group of origin, particularly in rural areas. For example, many Shaikh communities have seen their status relegated to the rank of service castes when they had previously enjoyed a far higher status among ashrafs. Since many Hindus who converted to Islam took on the name Shaikh when they were required to register with colonial census officials, the entire community was dragged further down the social ladder, which shows the disjunction that sometimes exists between a group’s theoretical level in the caste hierarchy and its social status.
Finally, most castes belonging to the Muslim “Untouchables” category (arzals) have remained on the fringes of the social advancement dynamic – albeit relative –during the post-colonial era. These different elements partly explain the forcefulness of current debates on the integration, whether full or partial, of low-level Muslim castes into categories of the population that benefit from certain social advantages.
Religious minorities, caste and reservations policy
These debates focus on the goal of justice and equality between groups and communities, both social and religious, that could be reached by the affirmative action policy implemented at the end of the colonial period; this was a set of measures giving the most impoverished a number of social benefits (rationing cards, jobs in public services and educational institutions, etc.). India was the first country to enshrine a policy of preferential advantages in its Constitution (1950). However, a number of problems still persist when identifying groups that may be eligible to benefit from quotas and reservations, and therefore from the social tools and categories to be used in order to determine the best scope of action for that policy.
The colonial practice of census-taking established at the end of the 19th century, whose main goal was to be more familiar with Indian society (caste, religion, level of education, etc.) in order to control its dynamics, first of all created a discrepancy between the structures of the social, local or regional space and the simplified, rigid and abstract categories of the population that were created by the census. Following the first recording campaigns, many Muslim social groups of lower status asked to be recorded as Ashrafs, a category that was given emphasis by the census officials, subsequently causing a major distortion in the histograms of the Muslim population in some regions, as was the case in Bengal. The enactment of Morley Minto’s reforms (1909), which saw the creation of separate electorates for Hindus and Muslims, only compounded the process of division in the social sphere and, at the same time, intensified concurrent nationalistic ideologies which ended in the creation of two separate States, India and Pakistan. In this context of increasingly polarized identities, castes began to form associations in order to claim the right to access social and political resources within the political sphere that were hitherto inaccessible. A policy of quotas was then implemented in the colonial period as a way of enabling the “Untouchables” to benefit from certain advantages. The “Scheduled Castes” category (SCs) was thus created in 1936 for Hindu “Untouchables”, excluding de facto the outcasts of other religions.
While the Indian Constitution gave the Scheduled Castes category official recognition and simultaneously created that of “Scheduled Tribes” (STs), the commission chaired by Kaka Kalelkar was responsible for evaluating and identifying other “backward” groups. Although the commission’s recommendations regarding the updating and centralization of the SC and ST lists were applied in 1956, those concerning the creation of a new category, that of “Other Backward Classes” (OBCs), were rejected. The Constitution nevertheless specified that other “Backward Classes” (BCs), including a number of groups and castes that sometimes belonged to religious minorities, could benefit from preferential treatment if they were considered to be socially and educationally backward.  However, the recommendations made by the Mandal commission (1979) regarding the expansion of reservations in public services (State institutions) to include the OBC category were not implemented by the Singh government until 1990.
Since 2006, when the central government in New Delhi decided to grant 27% extra places to OBCs in higher education institutions, a confrontational power game was established, highlighting the social and political dynamics of the reservations system between the central government in New Delhi and the States of the Union. Up to that point, the SC and ST lists, which had remained almost unchanged since Independence, were centralized in New Delhi and applicable to the entire national territory. Conversely, the list of groups eligible to join the OBC category was established directly at regional level, and the number of benefits was proportional to the size of the groups in question, and therefore liable to be reassessed at the end of each electoral period. The OBCs had not been clearly defined socially or numerically  and were therefore the focus of social struggles and constant confrontations in the political sphere.  With the new law on the extension of reservations to OBCs, the power games played out between the government and the States are evolving. The new law applies to central educational institutions, in other words universities, schools of engineering (IITs), management (IIMs) and medicine (AIIMS) run directly by the central government. These establishments are located both in Delhi and other cities around the country. The 27% of extra OBC places are in addition to SC and ST quotas, reaching a total that flirts with the legal limit of 50% of reservations. There are now lists of OBCs prepared at State level and lists of OBCs fixed by the National Commission for Backward Classes in Delhi.
The decision to extend reservations to OBCs in central higher educational establishments – which was part of a series of debates and controversies  concerning the possible inclusion of Muslim castes in the quotas policy – is relevant to religious minorities on several different levels. First of all, the Constitution specifies that the State cannot discriminate in the area of employment and recruitment on the grounds of social or religious criteria (articles 14 and 15); it focuses mainly on the most disadvantaged groups (article 16). So far, only Hindus – as well as some groups of Sikhs (1956) and Buddhists (1990) – were legally included in the SC category. However, despite limitations set by the Constitution and by Supreme Court orders not to exceed the legal limit, the total reservations in Tamil Nadu  today stands at 69%. Almost all of the Muslim castes and converted Christians in the region form part of that total, although they do not benefit from reserved positions in public administration. This apparent contradiction is due to the fact that the Hindus who have converted to a minority religion can only take advantage of reserved posts or places under specific legislation on minorities, and not legislation related to the OBCs.
The Indian Muslim community remains deeply divided over the issue of reservations. On the one hand are those who believe that the Muslim community as a whole should be included in the OBC category, since Islam does not recognize caste as a factor in social discrimination. Their opponents, on the other hand, believe that not all Muslim groups should necessarily benefit from preferential treatment. They believe that the position allowing the community as a whole to be eligible to benefit from the reservations system, mainly defended by high castes, would result in the appropriation of posts reserved by Ashrafs, a sector of the population that is largely in the minority and considered economically as a “creamy layer”.  The defenders of this position include the Muslim Dalits  who claim to be more in favour of including Arzal castes either in the OBC or SC quotas. Although the Supreme Court decided to exclude the creamy layer from the OBC quotas in 2008, the question of whether or not religious minorities can benefit from compensatory measures is currently being studied by the Indian parliament, taking inspiration from the system of quotas established in the four States of the South of India.  This would require amending the Constitution.
The debate linking caste, religious minority and reservations policy has persisted for at least three decades. Election campaigns and internal discussions on minority groups held in committees and institutions producing statistics are occasions to revive the debate on the inclusion and exclusion of Muslim castes in the categories that are granted the right to preferential treatment, whatever they may be. The publication of the 2011 census will provide an excellent sample of points of view and arguments because, for the first time since 1931, it will include a list of castes. Does this recording process, different from the traditional gathering of information, show innovation or a kind of regression? How will the castes be registered, counted and categorized? What status will be given to castes from religious minorities such as Christians or Muslims? While some fear that the publication of the results will lead to new waves of social and intercommunity tension, as was the case during the colonial period, representatives of the low-caste parties, on the other hand, see it as an opportunity to better count the groups that are eligible to benefit from compensatory measures.