How should a Hindu king expand his power into foreign land and how should he incorporate the conquered people into his kingdom? This is discussed by the eighth-century commentator on Manu, Medhātithi, in his commentary on MDh 2.23. The verse occurs in the first part of the second chapter, which deals with the sources of the law (dharma ), of which one is “the conduct of good people” (sadācāra ). That conduct is further defined both geographically and demographically (MDh 2.17–2.24). It is the conduct that is handed down through generations among the people who live in the sacred land of the Āryas, extending from the Himalayas in the North to the Vindhya Mountains in the South, and who belong to the four social classes and the “intermediate classes” (MDh 2.18), consisting of castes that were regarded as having their origin in a mixture of these four.
Manu (2.23) defines the heartland of this sacred territory as “the natural range of the black buck” and categorizes it as “fit for sacrifice,” thereby explicitly associating the country with Hindu rule. Beyond that land lie the countries of the barbarian foreigners (mlecchas ). Medhātithi, however, adds a rider:
If a good king from the warrior caste and the like gained victory over those foreigners and settled people from the four classes there, relegating the local foreigners to the status of caṇḍālas, just as in the land of the Āryas, then that land also would be “fit for sacrifice.” This is because land is not defiled by itself, for it becomes defiled through contact, as when it is sullied by something impure. Therefore, even apart from the regions specified, all people from the three upper classes must perform sacrifices whenever sufficient means are available even outside the natural range of the black buck.
(Medh 2.23 translated from Dave 1972 –85, vol.1: 200)
In other words, when foreign land is purified by the presence of people from the four classes settled there by a pious king, it becomes pure and fit for sacrifices. From then on, however, the local people must be regarded as caṇḍālas , the lowest caste in the Hindu society, which was considered untouchable, even according to the earliest Dharmasūtras. Shortly before this quote, Medhātithi had made it clear that “foreigners are known as people beyond the castes comprising the four classes, not even allotted the status of mixed castes of the reverse order” (Medh 2.23 in Dave 1972–85, vol.1: 199). 1
This quote and its context in Manu’s text give us a lot of information about the ideas behind the Hindu social structure as it was understood in Dharmaśāstra in terms of “class” (varṇa ) and castes regarded as “intermediate” (antarāla ) or as a “mixture of classes” (varṇasaṃkara ). It indicates that these categories were conceived by the Brahmin authors of these texts as both a political and a social structure closely connected with the ideology and aspirations of this priestly class. More precisely, it was a system meant to guarantee a special alliance between the Brahmins and the warrior rulers, including the specific privileges that such an alliance would yield. Hindu rule is made conditional on Vedic sacrifices, which is the specific sphere of expertise of Brahmins. The varṇa system is the prerequisite for Hindu rule, and it is the foundation of dharma . The conduct of the people from the four classes is one of its main sources, and the king will have to settle the four varṇas in the new territory in order to make it fit for the rituals that are the foundations of dharma .
The promotion of this special alliance between Brahmins and rulers is made explicit even in the earliest texts. According to Gautama:
There are in the world two who uphold the proper way of life—the king and the Brahmin deeply learned in the Vedas. And on them depend the life of the fourfold human race and of internally conscious creatures that move about, fly, and crawl; as well as their increase, protection, non-intermixture, and adherence to the Law.
The four classes (“the fourfold human race”) are the priestly class, that is the Brahmins (Brāhmaṇa ); the warrior class (Kṣatriya ), from which rulers are recruited or into which they are enrolled; the agricultural and merchant class (Vaiśya ); and the service class (Śūdra ), understood mainly as people engaged in various crafts and the working class.
Note that while Medhātithi advised the king to settle all four varṇas in the new conquered territory, he only called on the upper three of them to perform sacrifices there. This is because the varṇa and caste systems are also a system of marginalization. It was only the three upper classes that were allowed to have Vedic sacrifices performed for them; the Śūdras were excluded from such rituals and altogether from studying and using Vedic texts. Accordingly, boys from the three upper classes passed through an initiation ceremony (upanayana ) that involved training in recitation of the Vedas. This initiation was regarded as a symbolic new birth, and, therefore, these three classes were labeled collectively as “twice-born” (dvija ), whereas the Śūdras was labeled as “once-born” (ekeja ). 3
A further marginalization is indicated by the category of the “intermediate” classes that were understood as castes having their origin in a mixture of the four classes. Although these castes were not classified as belonging exclusively to any one of the four ideal classes but rather as the results of improper sexual mixing among them, they were still part of the Hindu social structure, collectively forming its hierarchical lower limit. This is expressed in precise terms by the fourteenth century commentator Kullūka: The conduct that is a source of dharma is that “of the classes starting with the Brahmins and extending as far as the mixed castes.” 4 Within this group of mixed classes, yet another hierarchical distinction is made. One group of castes are regarded as formed through hypergamous relations between two classes, say between Brahmin men and Vaiśya women, while another group is understood as formed through hypogamous relations, for instance between Kṣatriya men and Brahmin women. The principle behind the former is that the “father” must belong to a higher class than the “mother,” and behind the latter that the “father” belongs to a lower class than that of the “mother” (the quotation marks here indicating that the parents must be understood more in terms of some hypothetical past origin of these caste formations than as actual contemporary parents, although there are overlaps between the two cases, to which I shall return later). Generally, hypergamous relationships are tolerated, while hypogamous relationships are condemned. This is indicated by the Sanskrit terminology according to which the former type of relationship is “natural,” literally “with the hairs” (anuloma ), and the latter is “unnatural” or literally “against the hairs” (pratiloma ). We see here a blend of two hierarchical systems. One is the patriarchal ordering of gender (men higher, women lower), and another is the Brahmanical ordering of class (Brahmin higher, the other three gradually lower).
It follows logically from this system of class mixtures that the lowest group must be that resulting from the hypogamous union between Śūdra men and Brahmin women. This is the Caṇḍāla caste mentioned by Medhātithi in his comments on Manu 2.23 at the start of this chapter, and their lowness is emphasized by the idea that people from this caste are untouchable—members of higher castes have to take a bath in the event they come into physical contact with them. But notice that, according to Mehdātithi, the local people in the land that is conquered by a righteous Hindu king must be relegated into the status of the Caṇḍāla caste in order to make the conquered land “fit for sacrifice.” This means that, in spite of the low status of this caste, it is still an integrated part of the Hindu society that makes a land fit for the sacrifices performed by the three higher classes. In this respect the Caṇḍāla marks the final marginalization of this whole demographic ideology, that between the people belonging to such a Hindu social structure—including the Caṇḍālas at the bottom of it—and the “foreigners” (mleccchas ) outside it. To repeat, Medhātithi maintained that “foreigners are known as people beyond the castes comprising the four classes, not even allotted the status of mixed castes the reverse order” (Medh 2.23), that is, not even allotted a status as that of the hypogamous pratiloma castes of which the Caṇḍāla is the lowest.
In what follows, more will be said about each of these social categories, 5 but let me sum up the system for the sake of clarity:
The varṇa system consists of
(A) the three “twice-born” (dvija ) classes
Brāhmaṇa
Kṣatriya
Vaiśya
(B) beyond these three, the “once–born” (ekaja ) class
Śūdra
Each of these four varṇas contains a number of castes (jāti ), which also can be ordered hierarchically. Thus, there are high, learned Brahmins (bhrūṇas ) and lower Brahmins officiating in temples (devalakas ); there are “good Sūdras” (sacchūdras ) and “bad Śūdras” (asacchūdras ).
Beyond the four varṇas , there are the mixed classes (varṇasaṃkaras ): (A) hypergamous or anuloma castes, which are more or less tolerated; and (B) hypogamous or pratiloma castes, which are more or less condemned and looked down upon, of which the Caṇḍāla caste is regarded as the lowest and as untouchable. 6
This constitutes the Hindu society, and beyond that are the foreigners (mlecchas ). An obvious question that needs to be addressed is how far this ideologically motivated social classification existed as an empirical reality in ancient and medieval India. For sure, this question cannot be answered on the basis of Dharmaśāstra texts alone, but must involve a study of evidence outside this tradition. Patrick Olivelle has drawn the attention to the Aśokan inscriptions in which the word “varṇa ” does not occur at all; even the names of the varṇas are absent except for “Brāhmaṇa ” which however is mentioned as a religious community rather than as a class in a hierarchical social system (see Chapter 1 in the present volume). The grammatical literature is another source of evidence. Olivelle notices that the term “dvija ” (twice-born), which is a prominent classificatory concept in three of the four Dharmasūtras (from early second to first century bce ) is missing from Patañjali’s Mahābhāṣya , dated mid-second century bce , and from the literature before that period (Olivelle 2012a : 118–19). These findings indicate that these concepts (varṇa and dvija ) were promoted specifically by Dharmaśāstra authors.
Although Manu incorporated much material from the Arthaśāstra (Kauṭilya’s treatise on government), this text differs from Manu’s text in its view both on Brahmans and Śūdras. It does not eulogize the Brahmans as Manu and other Dharmaśāstra texts do, and it acknowledges the Śūdras as artisans (AŚ 1.3.8) and not merely as servants of the three upper classes like Manu. Kangle remarks, “This appears to be more in consonance with the actual state of things than the views of Smṛti writer like Manu” (1965 : 143). Finally, evidence from precolonial endowment records also give another picture than the one we meet in the Dharmaśāstra sources. In these records we get an impression of the identities that people ascribed to themselves. Cynthia Talbot concludes from material from Andhra Pradesh that “few of the donors of the endowments recorded in these documents choose to describe themselves in these terms” [varṇa and jāti ]. Instead, “the classical varṇa scheme was meaningful primarily to those who considered themselves brahmans.” Also the word jāti is “rarely found in the thirteenth-century inscriptions from Andhra, but there are also no references to specific subcastes by names.” (Talbot 2001: 50–2). All together, these different types of evidence confirm the hypothesis that the articulation and promotion of the varṇa system were primarily part of a priestly ideological strategy motivated by the specific interests of this class.
According to Manu, the four classes emerged together with the creation of the world. Like other species of living beings, they are each born with the inborn propensities (svabhāva ) that were placed in them by the Self-existent One at the time of creation. These specific propensities determine the activities (karmāṇi ) of each species, both animals and men. For men, however, “activities” in this context denotes first of all the specific occupation (svakarman ) and duty (svadharma ) prescribed for each class:
As they are brought forth again and again, each creature follows on its own the very activity assigned to it in the beginning by the Lord. Violence or non–violence, gentleness or cruelty, righteousness (dharma ) or unrighteousness (adharma ), truthfulness or untruthfulness—whichever he assigned to each at the time of creation, it stuck automatically to that creature. As at the change of seasons each season automatically adopts its own distinctive marks, so do embodied beings adopt their own distinctive acts. For the growth of these worlds, moreover, he produced from his mouth, arms, thighs, and feet, the Brahmin, the Kṣatriya, the Vaiśya, and the Śūdra.
(MDh 1.28–1.31 in Olivelle 2005a : 88)
The body parts mentioned here are obviously a reference to the twelfth stanza of the Puruṣasūkta (ṚV 10.90.12). The four varṇas , according to Manu, are not the product of any social negotiation or political reasoning but emerged out of the body of Cosmic Man and are, as such, direct manifestations of creation, just as the differentiations between horses, cows, and sheep (ṚV 10.90.10).
Kullūka spells out the significance of these verses of Manu just quoted. The violent acts of a lion are the manifestations of the violent inborn propensity of this animal, just as the gentle acts of priests and the violent acts of warriors are the direct results of the inborn nature of people born into these two classes (on MDh 1.29).
There can be no doubt that, according to these verses and according to the view of Dharmaśāstra in general, varṇa (and, accordingly, caste) is determined by birth; it is not the case that varṇa and caste can be decided solely on the basis of the character and skills of each individual (Smith 1994 : 28). This is because, ideally, birth (jāti ), inborn nature (svabhāva ), and work (svakarman ) will be harmoniously correlated as they were from creation. That said, however, other genres of literature, especially the Sanskrit epics, are full of examples of persons whose activities and wishes are in conflict with the norms of the class or caste they are born into. But eventually, dharma , that is the ideal order of birth, duty, work, and personal character, will prevail.
One episode from the Mahābhārata is especially telling. During the long instruction on dharma that king Yudhiṣṭhira receives after the war from the dying Bhīṣma, Yudhiṣṭhira asks the controversial question about how persons from the three lower classes can attain the same status as Brahmins. The answer he receives is that this status is unobtainable for these classes because it is the highest stage of the whole of creation. Only through innumerable births may people from the lower classes hope to be born one day in the Brahmin class (MBh 13.28.4–13.28.5).
The implicit premise is that birth determines class status. Consequently, differences and unequal possibilities in life are inherited through the generations, just as species transmit specific predispositions for wings or forelegs. To cross classes is like crossing species, a breach of the natural order established at creation. But even though crossing of classes (varṇasaṃkara ), unlike crossing of species, cannot always be known from their physical appearance, their identity will inevitably be revealed through the pattern of behavior that originates from their inborn propensities. 7 Bhīṣma illustrates this point by telling Yudhiṣṭhira the sad story of Mataṅga.
Mataṅga grew up as the son of a pious Brahmin. He was a good boy keen on living up to the expectations of his father. One day his father sent him to the town to get materials for a sacrifice. Mataṅga went on a cart pulled by a young donkey, which, however, did not want to leave its mother. Eager to fulfill his father’s wish, Mataṅga hit the donkey with the whip. The mother donkey comforted her crying child, telling him that such behavior is only to be expected from a Caṇḍāla like Mataṅga. Shocked by these words, Mataṅga asked the mother donkey how she knows about his identity. She answers that a true Brahmin embodies the quality of nonviolence (ahiṃsā ). 8 Therefore, this single act of violence inevitably reveals Mataṅga to be the bastard son of a Śūdra (in this case, a barber) with his Brahmin mother, in other words, an untouchable Caṇḍāla.
Now Mataṅga embarks on a severe regimen of extreme asceticism in order to change his destiny. Hard asceticism has the power to force the gods to appear and fulfill the wishes of the ascetic. Indra appears, promising to fulfill any wish that Mataṅga may have. However, confronted with Mataṅga’s only wish—the attainment of Brahmin status like that of his dear foster father—Indra has to refuse. That is beyond the possibilities of any god. When Mataṅga intensifies his ascetic exercises to the extent of almost dying, Indra reappears only to grant him the boon of being worshipped by women after his death, and eventually Mataṅga dies (MBh 13.28.7–13.30.16).
The most famous example, however, is that of Arjuna, caught in between his obligation with respect to his family and friends and his duties as a warrior. But the divine authority makes no exceptions:
Better to do one’s own duty (svadharma ) imperfectly than to do another man’s well; doing action intrinsic to his being (svabhāva ), a man avoids guilt. Arjuna, a man should not relinquish action he is born to (sahajaṃ karma ), even if it is flawed; all undertakings are marred by a flaw, as fire is obscured by smoke.
(BhG 18.47–18.48 in Miller 1986 : 149–50)
These two verses echo the view of Manu:
If a man of inferior birth out of greed lives by activities specific to his superiors, the king shall confiscate all his property and promptly send him into exile. Far better to carry out one’s own Law imperfectly than that of someone else’s perfectly; for a man who lives according to someone else’s Law fall immediately from his caste. (MDh 10.96–10.97 in Olivelle 2005a : 213)
Modern social reformers have seen this caste ideology as a serious hindrance to social mobility. My hypothesis is that this, precisely, has always been the intention of the Dharmaśāstra authors: to control upward social mobility.
Manu proposes a connection between the progression of the world ages (yugas ) and the gradual weakening of dharma , the ideal norm on which society rests:
In the Kṛta Age, the Law is whole, possessing all four feet; and so is truth. People never acquire any property through unlawful means. By acquiring such property, however, the Law is stripped of one foot in each of the subsequent Ages; through theft, falsehood, and fraud, the Law disappears a foot at a time.
(MDh 1.81–1.82 in Olivelle 2005a : 91).
What might these unlawful means possibly be by which people acquire property and thereby ruin dharma , and who are those “people”? We get a direct answer, I think, in a section of the text that sets the norms for the livelihood of Śūdras: “Even a capable Śūdra must not accumulate wealth; for when a Śūdra becomes wealthy, he harasses Brahmins” (MDh 10.129 in Olivelle 2005a : 214).
The religious idea behind this statement lies in the role of Brahmins as officiants of sacrifices. It is sacrifice that procures rain at the right time, producing rich crops and healthy cattle. As such, sacrifice is the foundation of all kinds of material wealth. The Brahmins are indeed self-assured about the significance of their role: “This whole world—whatever there is on earth—is the property of the Brahmin.…The Brahmin eats only what belongs to him, wears what belongs to him, and gives what belongs to him; it is by the kindness of the Brahmin that other people eat” (MDh 1.100–1.101 in Olivelle 2005a : 91–2, 242).
The economic reality of this idea is that sacrifice is also a source of income for Brahmins as receivers of donations. Since everything belongs to the Brahmin performers of sacrifice, what they receive in donations (dakṣiṇā ) at the end of the sacrifice has always been their property. While it is the prescribed “activity” (karman ) of all three twice-born classes to have sacrifices performed, to recite the Veda, and to offer donations in connections with these sacrifices and the other services from the Brahmins, it is the exclusive right and livelihood of the Brahmins to officiate at the sacrifices, to teach the Veda, and receive the donations from the members of all three upper classes. In addition, Kṣatriyas must protect the subjects by the use of weapons and be self-restrained, and Vaiśyas must look after cattle and live by agriculture, trade, and moneylending (MDh 1.88–1.90, 10.74–10.80). “A single activity did the Lord allot to the Śūdra, however: the ungrudging service of those very social classes” (1.91 in Olivelle 2005a : 91). The Śūdras, it seems, cannot be a direct source of income for Brahmins even though they might have accumulated wealth. They cannot study the Vedas, and they cannot have sacrifices performed; so they do not have any opportunity to offer donations. Whatever wealth they may possess therefore “harasses the Brahmins,” to whom it truly belongs. Besides, as servants they should not be in a position to accumulate wealth at all. Or, so it seems. 9
Reading behind the text and its strategies, however, we get another picture. One of these strategies is the notion of “times of adversity” (āpad ). Patrick Olivelle suggests in the first chapter of the present volume that Manu’s Dharmaśāstra was composed during a period with repeated invasions from the northwest that “created a cultural shock” among the Brahmin elite. These invasions caused the Brahmin authors to be explicit about the expectations of a true Hindu king, especially in regard to his relationships with the Brahmins. But even apart from these strained political conditions, and especially in the light of the nostalgic and utopian ideals of the perfect Hindu society envisaged in the texts, times are always adverse. We live in the degenerate Kali age, and āpad is more or less the new normal.
Ideally, Śūdras do not count among the paradigmatic givers, but during āpad Brahmins may receive gifts from anybody (MDh 10.102). 10 However, while “officiating at sacrifices and teaching always pertain to those who have undergone consecratory rites [whose donations therefore are pure],…accepting pertains even to a lowest–born Śūdra” (MDh 10.110 in Olivelle 2005a : 213), meaning that, although gifts from Śūdras who maintain their lives by other means than by serving Brahmins are not desirable, they are nevertheless received. Manu is at pains to ensure that this should not tip the balance of dependency: While “a man who knows the Law should never beg money from a Śūdra” (MDh 11.24 in Olivelle 2005a : 216), a Brahmin whose sacrifice is interrupted for want of an item “may freely take two or three items from the house of a Śūdra; for a Śūdra has nothing to do with sacrifices” (MDh 11.11 in Olivelle 2005a : 215). Begging is one thing; taking from a Śūdra’s property what truly belongs to the Brahmins is quite another.
What were the services that Śūdras had Brahmins performing for them, and what made the Śūdras wealthy enough to pay for these services in the form of donations? The historical development of the economic situation of the Śūdra class and the religious possibilities connected to it cannot be deduced from the Dharmaśāstra texts alone, but requires a study that includes many other types of sources, literary as well as epigraphic and archaeological. R. S. Sharma’s Śūdras in Ancient India (1990) is an attempt in that direction. Sharma describes a gradual improvement of the economic situation of Śūdras through the Gupta dynasty and early medieval period including the times when the late Dharmaśāstras of Yājñavalkya and others were composed. He concludes:
Perhaps the reason for the broadening of the religious rights of the śūdras lay in the improvement of their material conditions, which enabled them to perform sacraments and sacrifices by paying for priests. For the ability to sacrifice was rightly believed to be intimately connected with the ability to pay.
(R. S. Sharma 1990 : 312–13)
The “sacraments and sacrifices” that gradually became part of the religious rights of Śūdras were those such as prāyaścitta (penance), śrāddha (ancestral offerings), mahāyajñas (Great Sacrifices), certain saṃskāras (life cycle rites), pūjā (worship), vratas (votive observances), and dāna (donations). 11 Commenting on the statement that the wealth of Śūdras harasses the Brahmins (MDh 10.129), the eighth-century commentator Bhāruci rejects the idea that receiving the gifts of Śūdras should amount to receiving gifts from unworthy givers (asatpratigraha ). For, if it would be wrong for Brahmins to accept the donations from them, Śūdras would not be able to perform śrāddhas , which involve presenting the priests with a donation, obviously an indication that Brahmins performed śrāddhas for Śūdras. Penance is perhaps the most significant case, for here the inequality between Brahmins and Śūdras in terms of religious merit and material wealth is spelled out. The lower the varṇa of a penitent, the lesser the amount of the observance (typically fasting) but the more the value of the donations—often in the form of cattle—that must be given to the Brahmins who advised the penance. Thus, Brahmins observe harder penance but give a small donation, while the opposite holds for Śūdras. 12
Another of Manu’s āpad rules, which probably prescribes what already had become the normal state of affairs, allows Śūdras who are “unable” (or perhaps unwilling) to provide for their families by serving twice-born people to earn their living as craftsmen, provided they work in a craft that “best serves the twice-born” (MDh 10.99–10.100 in Olivelle 2005a : 213). And according to Viṣṇu, who composed his Dharmaśāstra five hundred years or so later than Manu (i.e., seventh-century ce ), these crafts are not restricted to times of adversity but do constitute the livelihood by which Śūdras fulfill their duty to serve the twice-born (ViDh 2.8 and 14). The śūdravarṇa comprised many crafts that were necessary for big building enterprises like royal temple building projects, crafts such as those of blacksmiths, carpenters, bricklayers, and artists (Sharma 1990 : 262–3). With increased trade, furthermore, the demand for goods produced by Śūdra artisans may have grown.
Śūdras were also allowed to trade according to Yājñavalkya (YDh 1.120) and other late Dharmaśāstras (Sharma 1990 : 267), and, apart from the Śūdras’ exclusion from Vedic recitation and rituals, the borderline between Vaiśyas and Śūdras became gradually more porous. R. S. Sharma (1990 : 322) speaks of medieval India as a “vaiśya-śūdra society”: “Since the social fabric of ancient India was based on the vaiśya tax and the śūdra labour, it may be called a vaiśya-śūdra society, but from the ideological and ritualistic point of view it may be called a brāhmaṇical society.” This is an indication of the pyramid structure of the Indian social hierarchy even from early times: the number of producing, trading, and laboring castes at the relative bottom of the Brahmanical hierarchical model outnumbered by far the warrior and priestly castes at the top.
With the growing economy and improved religious rights of the Śūdras, starting during the Gupta period and continuing into medieval India, their upward economic mobility could not be prevented. This was noticed by the Brahmin authors of Dharmaśāstra works, who reluctantly allowed Brahmins to receive as donations a part of the accumulated Śūdra wealth.
The Dharmaśāstra doctrine asserting that the vast number of low castes below the four classes have originated from a mixing of those classes (varṇasaṃkara ) has puzzled scholars including me. 13 In line with the preceding part of this chapter, however, I think it is reasonable to regard this doctrine as basically another attempt by the Brahmin authors to control upward social and economic mobility, in this case, not the mobility that took place through work and trade but the more tangible one that occurred though sexual relations. Whereas economic mobility created better possibilities and perhaps increased influence, it did not, from the point of view of Dharmaśāstra, change class and caste identities. 14 But that was possible through intermarriage. Intermarriage could have been initiated by the dissatisfaction of artisans and workers who protested against the occupational obligations they were assigned according to the Dharmaśāstra rules, and it could have been initiated by peasants who protested against the taxes that were demanded from them by the king (Sharma 1990 : 261).
Manu’s terminology on mixed classes is not entirely clear, but theoretically, it should be possible for descendants of mixed relationships to raise their status to that of the highest of his/her parents within a certain number of generations. That is, if the child of a mixed relationship marries upward, and the same happens in the next generations, at a certain time, the progeny will have the caste or class of the highest of the parents (MDh 10.64–10.65). 15 The practical realities behind this rule can only be guessed, but it seems to indicate that, after a certain amount of time, descendants from mixed relations would be accepted (or their descent would be forgotten) and assimilated into the superior class, as it actually happens when distinct ethnic bodily features gradually vanish through generations of mixed marriages.
The comparison with ethnic features is not far from the sense of the texts. We saw that varṇas were correlated with zoological species when they emerged from the body of Puruṣa. The idea is the same with mixed varṇas . This is made very clear by Kullūka in his comment to Manu’s overview of the varṇas , which states: “Three classes—Brahmin, Kṣatriya, and Vaiśya—are twice-born; the fourth, Śūdra, has a single birth. There is no fifth” (MDh 10.4). Kullūka explains:
Moreover, there is no fifth. As with a mule, the mixed castes do not belong to any class because their caste is different from both of the different castes of the parents. And this clarification in the text about these other castes is for the sake of the regulation of mutual interactions. (on MDh 10.4) 16
Caste definitions were meant as a way of regulating the occupations of individual castes and their mutual interactions, although occupation and caste definition are mutually dependent: “From the function pursued, the caste as laid down by the śāstra can be inferred. And by indicating the caste they can be enjoined to perform their functions” (Bhāruci on MDh 10.40 in Derrett 1975 , vol. 2: 310).
But to define castes “genealogically” as descendants from former mixed-class relations seems artificial. When did these relations take place? Should they merely be regarded as speculative myths of origin? The question gets complicated because the notion of varṇasaṃkara is used in the texts in two different ways. One is as caste definitions, that is, in terms of some hypothetical past origin of the caste. Another is with regard to contemporary mixed marriages or sexual relations as when specific punishments are laid down for men and women who have such illegitimate relations, 17 or when the texts lay down exact rules about inheritance for descendants of mixed marriages. 18 In these latter cases, Dharmaśāstra texts do not label these mixed relations by the caste names known from the caste definitions. Probably the children of such relations were not regarded as actually belonging to the castes that the relations between their unequal parents would suggest according to the genealogical-mythical caste definitions. As we saw with the Mataṅga story, it is different with the epics, whose object is not to make prescriptions for actual real-life situations but more to exemplify moral dilemmas in literary form.
Connecting certain tribes and castes, whose life styles and livelihood appeared appalling to Brahmin sentiments, with illegitimate sexual relations between persons from the four varṇas accomplished a double aim. It marginalized these castes and tribes thereby expressing the exclusivity of the original classes from the mouth, arms, thighs, and feet of Puruṣa, at the same time, connecting them to these very classes and thereby including them in the emerging larger Hindu social order.
It is plausible, as has been suggested by many scholars, that the notion of varṇasaṃkara arose in the process by which various groups of the indigenous population were gradually included in the networks of transactions with the āryas . 19 Some of the names of these groups are known from earlier texts, but it is the invention of the early Dharmaśāstra texts to stipulate a specific varṇasaṃkara genealogy to each of them. They do this in the context of marriage or they speak of these groups as “sons” from men and wives from different varṇas . 20 Perhaps this idiom was even used as a deliberate expression of inclusion.
It is in these early texts we get the distinction between relations in “the direct order” (anuloma ) and relations in the “reverse order” (pratiloma ). The latter is “outside the law” according to Gautama and cannot improve their status by intermarrying upward (GDh 4.25). An explicit distinction is also made by Yājñavalkya: “Those born in the reverse and those born in the direct order are known respectively as bad and good” (YDh 1.95c–d).
The notion of pratiloma as known from other textual genres is also used in contexts where a norm has been violated, where a situation is unnatural, or where a relationship has been reversed. When, according to the Bṛhadāraṇyaka Upaniṣad , Gārgya, the Brahmin, admitted his ignorance and asked to be instructed by King Ajātaśatru, the latter remarked, “Isn’t it a reversal of the norm [pratilomaṃ ] for a Brahmin to become the pupil of a Kṣatriya?” (BṛU 2.1.15 in Olivelle 1998 : 63). But eventually he taught Gārgya about the states of consciousness during dream and deep sleep. Purāṇic instructions for royal astrologers regarding favorable and unfavorable omens constitute another case. The former belong to the anuloma class of signs, whereas the latter belong to the pratiloma signs. These latter would typically be related to various classes of unnatural events (Inden 1985 : 32). The rhetoric about varṇasaṃkara , like the worst of these omens, is not without a touch of apocalyptic associations: “Wherever these delinquent-born individuals, who corrupt the social classes, are born, that realm quickly comes to ruin together with its inhabitants” (MDh 10.61).
In relation to the Dharmasūtras, Manu’s discussion of mixed classes was innovative in that he stipulated specific occupations for each of the varṇasaṃkara castes (see, e.g., MDh 10.46–10.49). I think this is an expression of an increased inclusion of these groups in concrete transactions and, consequently, increased social complexity. Whatever was the ethnic or other demographic identity of the groups mentioned in the Dharmaṣūtra lists of mixed classes, they are now regarded as occupational castes with which the people from the four varṇas make transactions. The number of castes has also increased considerably, and added to the earlier listed varṇasaṃkaras we get varṇasaṃkaras of varṇasaṃkaras , many of whom are associated with specific occupations (MDh 10.26–10.39). Clearly, the system tended to proliferate and multiply itself expanding the bottom of the pyramid.
One feature that clearly distinguished the bottom of the caste system from the layers above it was that of permanent untouchability. In the early texts, this feature did not yet have its own technical term but was articulated as the need for other people to have a bath in case they had been in physical contact with these persons. Untouchability, however, was not limited to persons from low castes but was a temporary condition attached to various persons in the domestic sphere, such as menstruating women and those in a period of impurity resulting from a birth or a death (see Chapter 17 on impurity). What made the untouchability attached to a caste special was that it was made permanent, inborn, and professional in the sense that it involved certain scorned occupations. 21
I have already mentioned the Caṇḍāla caste as untouchable, but the name was generic, and, like the term Śūdra, it could apply to various despised groups at the lowest rungs of society. Alternative names are used, such as Śvapaca or Śvapāka, Paulkasa or Pulkasa, and Divākīrti. The explicit rule that prescribes a bath after having been in physical touch with a Caṇḍāla is recorded in all four Dharmasūtras (ĀpDh 2.2.8; GDh 14.30; BDh 1.9.5; VaDh 23.33). It is repeated by Manu (5.85) and taken for granted in all medieval Dharmaśāstric texts along with the varṇasaṃkara definition of the Caṇḍāla as the pratiloma progeny of Śūdra men with Brahmin women. 22 Indeed, he is “the worst of all men” (MDh 10.12). As such, he is also classified as one of seven antyāvasāyins together with other pratiloma castes; often the term antyāvasāyin is just another generic term for untouchable castes.
As the term (antyāvasāyin ) indicates, Caṇḍālas were associated with remoteness in space or sequence (antya meaning remotest, last, or lowest). This reflects the fact that Caṇḍālas were geographically segregated and forced to live in areas outside or at the edge of villages and cities (MDh 10.51). 23 But antya applies to an end in more than one sense. Already in the Chāndogya Upaniṣad , we learn that persons of “foul behavior can expect to enter a foul womb” after they have died—“like that of a dog, a pig, or an outcaste [Caṇḍāla ] woman” (ChU 5.10.7 in Olivelle 1998 : 237). In a structurally similar manner, the twice-born family man is instructed to throw some food on the ground to Caṇḍālas, dogs, and birds outside his house at the end of the domestic vaiśvadeva offerings (ŚāṅkhGṛ 2.14). 24 Caṇḍālas are at the end, together with village dogs, pigs, and birds, not only in the spatial sense of the home and the village areas. As the lowest of men, they also mark the boundary in saṃsāra between human and animal existence.
Untouchable castes were typically assigned jobs connected with pollution in the form of death (cremation workers, executioners), products from dead animals (leather workers, drummers), or trash (cleaning the streets of villages). They came to form a large and indispensable unskilled labor force, not only in such traditionally assigned jobs but also as day laborers in the fields and on building sites. Unlike the Śūdras, they could be prevented from upward mobility by a systematic series of discriminating practices, all prescribed in Dharmaśāstra texts. 25 These included isolation in terms of sexual relationships, sharing food and food vessels, 26 communication, habitation, use of wells, participation in religious practices including temple worship, and, significantly, giving gifts (gift exchange remained crucial for economic transactions and mobility). In addition, they were forced to make themselves both visible and audible at markets and in the streets. It was not until the period of social reform movements during the nineteenth and twentieth centuries that some of these discriminations were slowly but effectively lifted.
In fact, the attempt to restrict upward social and economic mobility, which I think runs all through the varṇa and caste rules in the Dharmaśāstra texts, did not go unchallenged, even during the history of that literature. Buddhist and other early renouncer movements, medieval ascetic and tantric groups, bhakti sant poets of both nirguṇa and saguṇa orientation, and Sikhs who included their poetry in the Ādi Granth, all challenged the Brahmin social order. The marginalization of Brahmins in the South with the rise of Maharashtrian and Dravidian self-assertiveness lead by Jyotirao Phule and E. V. Ramasamy, respectively, marked a definite turn of authority, but perhaps a certain Brahmin frailty already lay behind the opulent self-promotion that we read in large parts of the Dharmaśāstra texts. After all, Brahmins have always been a service class dependent on kings and local landowners.
This, I think, comes through in some of the recent debates on caste theory. These debates took their departure in a critique of Louis Dumont’s hierarchical model. According to this model, the social system is oriented top-down, from the Brahmin to the Untouchable. Status, defined in relation to an ideology, is hierarchically superior to power. Therefore, Brahmanical values of personal and inherent ritual purity are the parameters in relation to which all else is defined (Dumont 1970: 36–42).
Critiques of this Brahmanical-inspired hierarchical model, such as those of Gloria Goodwin Raheja (1988 ) and Declan Quigley (1993), were inspired by the alternative model of the British ethnographer A. M. Hocart. Unlike Dumont, who saw society as structured from a hierarchical top, Hocart saw it as organized around centers of power. The basic idea in Hocart’s theory is that the state is a ritual organization that includes the services of the varṇas and castes. The occupations assigned to these groups are primarily ritual services. Drummers may make their primary living as day-workers rather than by playing drums, but they are known as drummers because this is their ritual service at funerals and weddings. The king, as the paradigmatic sacrificer (yajamāna ), is the pivotal character of the whole organization. However, the system is multicentric, reaching down through the society, as well as up through the skies:
The King’s state is reproduced in miniature by his vassals: a farmer has his court, consisting of the personages most essential to the ritual and so present even in the smallest community, the barber, the washerman, the drummer and so forth.… The temple and the palace are indistinguishable, for the king represents the gods.…The god in his temple has his court like the king in his palace: smiths, carpenters, potters, all work for him. (Hocart 1950: 68)
The centrality model highlights the true relationship between Brahmins and kings. Brahmins may be sacrificers themselves, but ultimately, they depend on other sacrificers for their living. The praise of the righteous Hindu king who expands his kingdom and makes the conquered land “fit for sacrifice” by settling the four varṇas on the new land is also an appeal for support and moral-social leadership, which is a basic motivation behind the Dharmaśāstra texts.
An unknown man without the proper complexion [varṇa ], born from a squalid womb, a non–Ārya with some measure of Ārya features—one should detect such a man by his activities. Un–Ārya conduct, harshness, cruelty, and the neglect of rites reveal in this world a man who is born from a squalid womb. He will possess the character of either his father of his mother, or of both; a man born from an evil womb is never able to conceal his nature. (MDh 10.57–10.59 in Olivelle 2005a : 211)
See also the note to 10.57 at page 337 about the translation of varṇa as “complexion.” Outward appearance and color are among the meanings of the word varṇa , and the connection between complexion and class is made explicit by the grammarian Patañjali on Pānini 2.2.6.