Caste system in India

This article is about socio-political stratification in Indian society. For socio-religious Hindu stratification, see Varna (Hinduism).

"Castes in India" redirects here. For the book, see Castes in India: Their Mechanism, Genesis and Development.

In India, social classification based on caste has its origin in ancient times. It was transformed by various ruling elites in medieval, early-modern, and modern India, especially in the aftermath of the collapse of the Mughal Empire and the establishment of the British Raj. Caste is traditionally associated with Hinduism, but is more pervasive in extent: an estimated 98% of contemporary Indians, including Hindus, Muslims, Christians, Sikhs, Buddhists, and Jains, identify with a caste.

Beginning in ancient India, the caste system was originally centered around varna, with Brahmins (priests) and, to a lesser extent, Kshatriyas (rulers and warriors) serving as the elite classes, followed by Vaishyas (traders and merchants) and finally Shudras (labourers). Outside of this system are the oppressed, marginalised, and persecuted Dalits (also known as "Untouchables") and Adivasis (tribals). Over time, the system became increasingly rigid, and the emergence of jati led to further entrenchment, introducing thousands of new castes and sub-castes. With the arrival of Islamic rule, caste-like distinctions were formulated in certain Muslim communities, primarily in North India. The British Raj furthered the system, through census classifications and preferential treatment to Christians and people belonging to certain castes. Social unrest during the 1920s led to a change in this policy towards affirmative action.

Caste-based differences have also been practised in other regions and religions in the Indian subcontinent, like Nepalese Buddhism, Christianity, Islam, Judaism and Sikhism. It has been challenged by many reformist Hindu movements, Buddhism, Sikhism, Christianity, and present-day Neo Buddhism. With Indian influences, the caste system is also practiced in Bali, Indonesia.

After achieving independence in 1947, India banned discrimination on the basis of caste and enacted many affirmative action policies for the upliftment of historically marginalised groups, as enforced through its constitution. However, the system continues to be practiced in India and caste-based discrimination, segregation, violence, and inequality persist.

Definitions and concepts

Varna, jāti, and caste

Varna

Main article: Varna (Hinduism)

Varna, meaning type, order, colour, or class  is a framework for grouping people into classes, first used in Vedic Indian society. It is referred to frequently in the ancient Indian texts. There are four classes: the Brahmins (priestly class), the Kshatriyas (rulers, administrators and warriors; also called Rajanyas), the Vaishyas (artisans, merchants, tradesmen and farmers), and the Shudras (labouring classes). The varna categorisation implicitly includes a fifth element, those deemed to be entirely outside its scope, such as tribal people and the Dalits (also known as "Untouchables").

Jati

Main article: Jāti

In ancient texts, Jati, meaning birth, is mentioned less often and clearly distinguished from varna. There are four varnas but thousands of jatis. The jatis are complex social groups that lack universally applicable definitions or characteristics and have been more flexible and diverse than was previously often assumed.

Certain scholars of caste have considered jati to have its basis in religion, assuming that the sacred elements of life in India envelop the secular aspects; for example, the anthropologist Louis Dumont described the ritual rankings that exist within the jati system as being based on the concepts of religious purity and pollution. This view has been disputed by other scholars who believe it to be a secular social phenomenon driven by the necessities of economics, politics, and at times geography. Jeaneane Fowler says that although some people consider jati to be occupational segregation, in reality the jati framework does not preclude or prevent a member of one caste from working in another occupation.

A feature of jatis has been endogamy, in Susan Bayly's words, that "both in the past and for many, though not all Indians in more modern times, those born into a given caste would normally expect to find marriage partner" within their jati.

A 2016 study based on the DNA analysis of unrelated Indians determined that endogamous jatis originated during the Gupta Empire. An earlier 2013 study from Tamil Nadu has suggested older (4–6 Kya) origins of endogamy based upon male lineages.

Jatis have existed in India among Hindus, Muslims, Christians, and tribal people, and there is no clear linear order among them.

Caste

Main article: Caste

The term caste is derived from the Portuguese and Spanish word casta, meaning "race, lineage, breed" and, originally, "'pure or unmixed (stock or breed)". Originally not an Indian word, it is now widely used in English and in Indian languages, closely translated to varna and jati.

Ghurye's 1932 description

The sociologist G. S. Ghurye wrote in 1932 that, despite much study by many people,

we do not possess a real general definition of caste. It appears to me that any attempt at definition is bound to fail because of the complexity of the phenomenon. On the other hand, much literature on the subject is marred by lack of precision about the use of the term.

Ghurye offered what he thought was a definition that could be applied across India, although he acknowledged that there were regional variations on the general theme. His model definition for caste included the following six characteristics:

• Segmentation of society into groups whose membership was determined by birth.
• A hierarchical system wherein generally the Brahmins were at the head of the hierarchy, but this hierarchy was disputed in some cases. In various linguistic areas, hundreds of castes had a gradation generally acknowledged by everyone.
• Restrictions on feeding and social intercourse, with minute rules on the kind of food and drink that upper castes could accept from lower castes. There was a great diversity in these rules, and lower castes generally accepted food from upper castes.
• Segregation, where individual castes lived together, the dominant caste living in the center and other castes living on the periphery. There were restrictions on the use of water wells or streets by one caste on another: an upper-caste Brahmin might not be permitted to use the street of a lower-caste group, while a caste considered impure might not be permitted to draw water from a well used by members of other castes.
• Occupation, generally inherited. Lack of unrestricted choice of profession, caste members restricted their own members from taking up certain professions they considered degrading. This characteristic of caste was missing from large parts of India, stated Ghurye, and in these regions all four castes (Brahmins, Kshatriyas, Vaishyas and Shudras) did agriculture labour or became warriors in large numbers.
• Endogamy, restrictions on marrying a person outside caste, but in some situations hypergamy allowed. Far less rigidity on inter-marriage between different sub-castes than between members of different castes in some regions, while in some, endogamy within a sub-caste was the principal feature of caste-society.

The above Ghurye's model of caste thereafter attracted scholarly criticism for relying on the census reports produced by the colonial government, the "superior, inferior" racist theories of H. H. Risley, and for fitting his definition to then prevalent orientalist perspectives on caste.

Ghurye added, in 1932, that the colonial construction of caste led to the livening up, divisions and lobbying to the British officials for favourable caste classification in India for economic opportunities, and this had added new complexities to the concept of caste. Graham Chapman and others have reiterated the complexity and they note that there are differences between theoretical constructs and the practical reality.

Modern perspective on definition

Ronald Inden, the Indologist, agrees that there has been no universally accepted definition of "caste". For example, for some early European documenters it was thought to correspond with the endogamous varnas referred to in ancient Indian scripts, and its meaning corresponds in the sense of estates. To later Europeans of the Raj era it was endogamous jatis, rather than varnas, that represented caste, such as the 2,378 jatis that colonial administrators classified by occupation in the early 20th century.

Arvind Sharma, a professor of comparative religion, notes that caste has been used synonymously to refer to both varna and jati but that "serious Indologists now observe considerable caution in this respect" because, while related, the concepts are considered to be distinct. In this he agrees with the Indologist Arthur Basham, who noted that the Portuguese colonists of India used casta to describe

... tribes, clans or families. The name stuck and became the usual word for the Hindu social group. In attempting to account for the remarkable proliferation of castes in 18th- and 19th-century India, authorities credulously accepted the traditional view that by a process of intermarriage and subdivision the 3,000 or more castes of modern India had evolved from the four primitive classes, and the term 'caste' was applied indiscriminately to both varna or class, and jati or caste proper. This is a false terminology; castes rise and fall in the social scale, and old castes die out and new ones are formed, but the four great classes are stable. There are never more or less than four and for over 2,000 years their order of precedence has not altered."

The sociologist André Beteille notes that, while varna mainly played the role of caste in classical Hindu literature, it is jati that plays that role in present times. Varna represents a closed collection of social orders whereas jati is entirely open-ended, thought of as a "natural kind whose members share a common substance". Any number of new jatis can be added depending on need, such as tribes, sects, denominations, religious or linguistic minorities and nationalities. Thus, "Caste" is not an accurate representation of jati in English. Better terms would be ethnicity, ethnic identity and ethnic group.

Complexity and flexibility

Research on caste systems across the Indian subcontinent during the latter 1900s revealed that caste was far more complex and dynamic than previously thought. While British colonial authorities had portrayed it as a uniform, rigid system fundamental to Indian society, studies showed that caste's significance and structure varied considerably between regions. Rather than being a fixed hierarchy, caste functioned as one of several possible forms of social organization and identity. People could maintain multiple community affiliations, with caste sometimes taking precedence and other times being secondary to different social bonds. This flexibility allowed caste to serve as one way of creating social cohesion while leaving room for other types of community ties to flourish.

Sociologist Anne Waldrop observes that while outsiders view the term caste as a static phenomenon of stereotypical tradition-bound India, empirical facts suggest caste has been a radically changing feature. The term means different things to different Indians. In the context of politically active modern India, where job and school quotas are reserved for affirmative action based on castes, the term has become a sensitive and controversial subject.

Sociologists such as M. N. Srinivas and Damle have debated the question of rigidity in caste and believe that there is considerable flexibility and mobility in the caste hierarchies.

Contemporary extent and enumeration

Between 2019 and 2020, Pew Research interviewed 29,999 Indian adults, finding that 98% of Indians identified as members of a caste when asked.

The question of if and how to enumerate caste is politically fraught in its own right. No census under the Republic of India has enumerated caste, on the premise that doing so would be socially divisive; the last Indian census to do so was in 1931, under British rule. The 15th census in 2011 did not count caste, but an associated Socio Economic and Caste Census was conducted outside the scope of the 1948 Census of India Act. The 16th census, which as of 2026 is scheduled to enumerate population data in 2027, is anticipated to ask each individual to declare a jati.

Origins

Caste system in 19th century India

Hindu musician Muslim merchant Sikh chief Arab soldier Pages from Seventy-two Specimens of Castes in India according to Christian Missionaries in February 1837. They include Hindu, Muslim, Sikh and Arabs as castes of India.

Perspectives

There are at least two perspectives for the origins of the caste system in ancient and medieval India, which focus on either ideological factors or on socio-economic factors.

• The first school of thought focuses on ideological factors that are claimed to drive the caste system and hold that caste is rooted in the four varnas. This perspective was particularly common among scholars during the British colonial era and was articulated by Dumont, who concluded that the system was ideologically perfected several thousand years ago and has remained the primary social reality ever since. This school primarily justifies its theory by citing the ancient law book Manusmriti and has been criticized for disregarding economic, political and historical evidence.
• The second school of thought focuses on socioeconomic factors, claiming that those factors drive the caste system. It believes caste to be rooted in the economic, political and material history of India. This school, which is common among scholars of the post-colonial era such as Berreman, Marriott, and Dirks, describes the caste system as an ever-evolving social reality that can only be properly understood by the study of historical evidence of actual practice and the examination of verifiable circumstances in the economic, political and material history of India. This school has focused on the historical evidence from ancient and medieval society in India, during the Muslim rule between the 12th and 18th centuries, and the policies of the British colonial government from the 18th century to the mid-20th century.

The first school has focused on religious anthropology and disregarded other historical evidence as secondary or derivative of this tradition. The second school has focused on sociological evidence and sought to understand the historical circumstances. The latter has criticised the former for its caste origin theory, claiming that it has dehistoricized and decontextualised Indian society.

Ritual kingship model

According to Samuel, referencing George L. Hart, central aspects of the later Indian caste system may originate from the ritual kingship system prior to the arrival of Brahmanism, Buddhism and Jainism in India. The system is seen in the South Indian Tamil literature from the Sangam period (3rd BCE-3rd c.CE). This theory discards the Indo-Aryan varna model as the basis of caste, and is centred on the ritual power of the king, who was "supported by a group of ritual and magical specialists of low social status", with their ritual occupations being considered 'polluted'. According to Hart, it may be this model that provided the concerns with "pollution" of the members of low status groups. The Hart model for caste origin, writes Samuel, envisions "the ancient Indian society consisting of a majority without internal caste divisions and a minority consisting of a number of small occupationally polluted groups".

Vedic varnas

The varnas originated in late Vedic society (c. 1000–500 BCE). The first three groups, Brahmins, Kshatriyas and Vaishya, have parallels with other Indo-European societies, while the addition of the Shudras is probably a Brahmanical invention from northern India.

The varna system is propounded in revered Hindu religious texts, and understood as idealised human callings. The Purusha Sukta of the Rigveda and Manusmriti's comment on it, being the oft-cited texts. Counter to these textual classifications, many revered Hindu texts and doctrines question and disagree with this system of social classification.

Scholars have questioned the varna verse in the Rigveda, noting that the varna therein is mentioned only once. The Purusha Sukta verse is now generally considered to have been inserted at a later date into the Rigveda, probably as a charter myth. Stephanie Jamison and Joel Brereton, professors of Sanskrit and Religious studies, state, "there is no evidence in the Rigveda for an elaborate, much-subdivided and overarching caste system", and "the varna system seems to be embryonic in the Rigveda and, both then and later, a social ideal rather than a social reality". In contrast to the lack of details about varna system in the Rigveda, the Manusmriti includes an extensive and highly schematic commentary on the varna system, but it too provides "models rather than descriptions". Susan Bayly summarises that Manusmriti and other scriptures helped elevate Brahmins in the social hierarchy and these were a factor in the making of the varna system, but the ancient texts did not in some way "create the phenomenon of caste" in India.

Jatis

Jeaneane Fowler, a professor of philosophy and religious studies, states that it is impossible to determine how and why the jatis came into existence. Susan Bayly, on the other hand, suggests that the jati system emerged because it offered a source of advantage in an era of pre-Independence poverty, lack of institutional human rights, volatile political environment, and economic insecurity.

According to social anthropologist Dipankar Gupta, guilds developed during the Mauryan period and crystallised into jatis in post-Mauryan times with the emergence of feudalism in India, which finally crystallised during the 7th–12th centuries. However, other scholars dispute when and how jatis developed in Indian history. Barbara Metcalf and Thomas Metcalf, both professors of History, write, "One of the surprising arguments of fresh scholarship, based on inscriptional and other contemporaneous evidence, is that until relatively recent centuries, social organisation in much of the subcontinent was little touched by the four varnas. Nor were jati the building blocks of society."

According to Basham, ancient Indian literature refers often to varnas, but hardly if ever to jatis as a system of groups within the varnas. He concludes that "If caste is defined as a system of group within the class, which are normally endogamous, commensal and craft-exclusive, we have no real evidence of its existence until comparatively late times."

Untouchable outcastes and the varna system

The Vedic texts neither mention the concept of untouchable people nor any practice of untouchability. The rituals in the Vedas ask the noble or king to eat with the commoner from the same vessel. Later Vedic texts ridicule some professions, but the concept of untouchability is not found in them.

The post-Vedic texts, particularly Manusmriti mentions outcastes and suggests that they be ostracised. Recent scholarship states that the discussion of outcastes in post-Vedic texts is different from the system widely discussed in colonial era Indian literature, and in Dumont's structural theory on caste system in India. Patrick Olivelle, a professor of Sanskrit and Indian Religions and credited with modern translations of Vedic literature, Dharma-sutras and Dharma-sastras, states that ancient and medieval Indian texts do not support the ritual pollution, purity-impurity premise implicit in the Dumont theory. According to Olivelle, purity-impurity is discussed in the Dharma-sastra texts, but only in the context of the individual's moral, ritual and biological pollution (eating certain kinds of food such as meat, going to bathroom). Olivelle writes in his review of post-Vedic Sutra and Shastra texts, "we see no instance when a term of pure/impure is used with reference to a group of individuals or a varna or caste". The only mention of impurity in the Shastra texts from the 1st millennium is about people who commit grievous sins and thereby fall out of their varna. These, writes Olivelle, are called "fallen people" and considered impure in the medieval Indian texts. The texts declare that these sinful, fallen people be ostracised. Olivelle adds that the overwhelming focus in matters relating to purity/impurity in the Dharma-sastra texts concerns "individuals irrespective of their varna affiliation" and all four varnas could attain purity or impurity by the content of their character, ethical intent, actions, innocence or ignorance (acts by children), stipulations, and ritualistic behaviours.

Dumont, in his later publications, acknowledged that ancient varna hierarchy was not based on purity-impurity ranking principle, and that the Vedic literature is devoid of the untouchability concept.

History

Early history

The existence of jati and the precursor of caste has been found in the Indus Valley Civilisation (3300–1700 BCE). Sociologist S C. Malik writes Indus Valley Civilization saw "perpetuation of caste status by birth" and "caste-class patterns" were found in Indian society since this period. Charles Maisels finds caste stratification to have arisen from occupational groups upon the devolution from urban Indus Valley society. Romila Thapar finds possibility of a caste as pre-Vedic element, and notes that Jati pre-dated Vedic varna. Thapar further notes that Jatis were derived from clans of Indus Valley Civilisation which saw emergence of different occupations that were inherited and became hierarchically organised with unequal access to resources with stringent marriage regulations and rituals becoming rigid system over a period of time. Archaeologist M.K. Dhavalikar has also supported existence of caste system in Indus Valley Civilisation.

Early Vedic period (1500–1000 BCE)

During the early Vedic period in northern India, when the Rigveda was composed (1500–1200 BC), there were only two varnas in the Vedic society: arya varna and dasa varna. The distinction originally arose from tribal divisions. The Vedic people were Indo-European-speaking tribes who migrated over a period of several centuries into northern South Asia from the Bactria-Margiana, and mixed with the "indigenous Dravidic-speaking populations", but regarded themselves as superior. The Vedic tribes regarded themselves as arya (the noble ones) and the rival tribes were called dasa, dasyu and pani. The dasas were frequent allies of the Aryan tribes, and they were probably assimilated into the Aryan society, giving rise to a class distinction. Many dasas were, however, in a servile position, giving rise to the eventual meaning of dasa as servant or slave.

The Rigvedic society was not distinguished by occupations. Many husbandmen and artisans practised a number of crafts. The chariot-maker (rathakara) and metal worker (karmara) enjoyed positions of importance and no stigma was attached to them. Similar observations hold for carpenters, tanners, weavers and others.

Towards the end of the Atharvaveda period, new class distinctions emerged. The erstwhile dasas are renamed Shudras, probably to distinguish them from the new meaning of dasa as slave. The aryas are renamed vis or Vaishya (meaning the members of the tribe) and the new elite classes of Brahmins (priests) and Kshatriyas (warriors) are designated as new varnas. The Shudras were not only the erstwhile dasas but also included the aboriginal tribes that were assimilated into the Aryan society as it expanded into Gangetic settlements. This class-distinction is still reflected in the fact that the upper castes have a relatively higher steppe ancestry than the lower castes.

There is no evidence of restrictions regarding food and marriage during the Vedic period. According to Moorjani et al. (2013), co-authored by Reich, extensive admixture took place between 2200 BCE and 100 CE (4200 to 1900 before present), whereafter India shifted to "a region in which mixture was rare". In southern India, endogamy may have set in 1000 years earlier.

Later Vedic period (1000–600 BC)

In an early Upanishad, Shudra is referred to as Pūşan or nourisher, suggesting that Shudras were the tillers of the soil. But soon afterwards, Shudras are not counted among the tax-payers and they are said to be given away along with the land when it is gifted. The majority of the artisans were also reduced to the position of Shudras, but there is no contempt indicated for their work. The Brahmins and the Kshatriyas are given a special position in the rituals, distinguishing them from both the Vaishyas and the Shudras. The Vaishya is said to be "oppressed at will" and the Shudra "beaten at will".

Second urbanisation (500–200 BC)

Knowledge of this period is supplemented by Pali Buddhist texts. Whereas the Brahmanical texts speak of the four-fold varna system, the Buddhist texts present an alternative picture of the society, stratified along the lines of jati, kula and occupation. It is likely that the varna system, while being a part of the Brahmanical ideology, was not practically operative in the society. In the Buddhist texts, Brahmin and Kshatriya are described as jatis rather than varnas. They were in fact the jatis of high rank. The jatis of low rank were mentioned as chandala and occupational classes like bamboo weavers, hunters, chariot-makers and sweepers. The concept of kulas was broadly similar. Along with Brahmins and Kshatriyas, a class called gahapatis (literally householders, but effectively propertied classes) was also included among high kulas. The people of high kulas were engaged in occupations of high rank, viz., agriculture, trade, cattle-keeping, computing, accounting and writing, and those of low kulas were engaged in low-ranked occupations such as basket-weaving and sweeping. The gahapatis were an economic class of land-holding agriculturists, who employed dasa-kammakaras (slaves and hired labourers) to work on the land. The gahapatis were the primary taxpayers of the state. This class was apparently not defined by birth, but by individual economic growth.

While there was an alignment between kulas and occupations at least at the high and low ends, there was no strict linkage between class/caste and occupation, especially among those in the middle range. Many occupations listed such as accounting and writing were not linked to jatis. Peter Masefield, in his review of caste in India, states that anyone could in principle perform any profession. The texts state that the Brahmin took food from anyone, suggesting that strictures of commensality were as yet unknown. The Nikaya texts also imply that endogamy was not mandated.

The contestations of the period are also evident from the texts describing dialogues of Buddha with the Brahmins. The Brahmins maintain their divinely ordained superiority and assert their right to draw service from the lower orders. Buddha responds by pointing out the basic facts of biological birth common to all men and asserts that the ability to draw service is obtained economically, not by divine right. Using the example of the northwest of the subcontinent, Buddha points out that aryas could become dasas and vice versa. This form of social mobility was endorsed by Buddha.

Early Hinduism (200 BC–320 AD) and Classical period (320–650 AD)

See also: Dharmaśāstra and Hindu synthesis

According to Moorjani et al. (2013), endogamy set in after 100 CE. According to Basu et al. (2016), admixture between populations was "rapidly replaced by endogamy [...] among upper castes and Indo-European speakers predominantly[...] almost simultaneously, possibly by decree of the rulers, in upper-caste populations of all geographical regions, about 70 generations before present, probably during the reign (319–550 CE) of the ardent Hindu Gupta rulers". Johannes Bronkhorst, referring to Basu et al. (2016) and Moorjani et al. (2013) states that "it seems safe to conclude that a shift to endogamy took place during the first half of the first millennium CE, at least in northern India", due to the growing influence of Brahmanism. This shift is attested in the Manusmriti (1st to 3rd century CE), which "explicitly forbade intermarriage across castes".

The Mahabharata, estimated to have been completed by the end of the fourth century CE, discusses the varna system in section 12.181, presenting two models. The first model describes varna as a colour-based system, through a character named Bhrigu, "Brahmins varna was white, Kshatriyas was red, Vaishyas was yellow, and the Shudras' black". This description is questioned by Bharadvaja who says that colors are seen among all the varnas, that desire, anger, fear, greed, grief, anxiety, hunger and toil prevails over all human beings, that bile and blood flow from all human bodies, so what distinguishes the varnas, he asks. The Mahabharata then declares, "There is no distinction of varnas. This whole universe is Brahman. It was created formerly by Brahma, came to be classified by acts." The epic then recites a behavioural model for varna, that those who were inclined to anger, pleasures and boldness attained the Kshatriya varna; those who were inclined to cattle rearing and living off the plough attained the Vaishya varna; those who were fond of violence, covetousness and impurity attained the Shudra varna. The Brahmin class is modeled in the epic as the archetype default state of man dedicated to truth, austerity and pure conduct. In the Mahabharata and pre-medieval era Hindu texts, according to Hiltebeitel, "it is important to recognise, in theory, varna is nongenealogical. The four varnas are not lineages, but categories".

Late classical and early medieval period (650 to 1400)

Scholars have tried to locate historical evidence for the existence and nature of varna and jati in documents and inscriptions of medieval India. Supporting evidence has been elusive and contradictory evidence has emerged.