[“Caste in Indian History” is a brilliant essay on the evolution of caste in India, written by renowned Marxist historian Prof. Irfan Habib. It is the text of the Inaugural D.D. Kosambi Memorial Lecture delivered in Bombay in March 1985, and was published in Prof. Habib’s book ‘Caste and Money in Indian History’ (1987). It was reproduced in the collection of Prof. Habib’s essays titled ‘Essays in Indian History: Towards a Marxist Perception’ (1995), published by Tulika Books, New Delhi. Anticaste.in republishes this classic essay with the author’s permission, for the benefit of a larger readership. The text is from the seventh reprint (2007) of ‘Essays in Indian History’.]
Caste is the most characteristic — and many would say, unique — social institution of India. No interpretation of our history and culture can demand a hearing unless it encompasses the caste system. One of the abiding achievements of D.D. Kosambi’s scholarship was his ability to unite a lively spirit of anthropological investigation with a critical analysis of historical evolution. I hope that by choosing the role of caste in Indian history as the theme of this lecture, I may be able to touch on some of Kosambi’s most valuable insights.
Any such endeavour must, first, come to grips with the problem of definition. It is not surprising that this should be difficult, but perhaps a working definition could still be attempted to serve us as a point of departure. Caste, we may say, is a fairly well-marked, separate community, whose individual members are bound to each other through endogamy (and hypergamy), and very often also by a common hereditary profession or duty, actual or supposed. Many sociologists, however, appear to regard this definition as quite insufficient. They would add that we must also stipulate the existence of a perception of the rank of one caste in relation to other castes, a ranking which finds expression in the degree of ‘purity’ and ‘impurity’ of the other castes in relation to one’s own, and in specific rites and practices followed by, or assigned to, each caste. Louis Dumont, in his Homo Hierarchicus  considers, as the very title of the work shows, the hierarchical principle to be the very core and heart of the caste system; without it, there would be no caste.
Whether we should follow the very simple definition we have suggested, or the kind of definition that Dumont would approve of is not a mere matter of semantics: it is of crucial importance for understanding the history of our civilisation. I, therefore crave the reader’s indulgence for examining Dumont’s views in some detail.
To Dumont the caste system must be understood in terms of its essential religious (‘Hindu’) ideology, which pervades all the immense variety that it displays. It is reflected in the endless, complex, even conflicting, arrangements of ranks, the highest belonging always to the Brahmans, who are the ‘purest’ and command much of its ritual. The ranking does not originate in, or correspond to, the actual distribution of power or wealth, but arises, so to speak, out of the basic elaboration of the basic principle of purity or pollution. Thus neither are castes ‘an extreme’ form of classes, nor is the caste system a system of social stratification: it need not, and does not, correspond to the distribution of wealth or power. Dumont insists that caste must be understood as ‘part of the whole’ (a favourite phrase of his), which means that the entire society must be divided up among castes, and there must be no significant residue. Thus, in effect, caste must exist as the sole or dominant form of social organisation, or not exist at all.
If all this is to be accepted, if, that is, caste arose out of an ideology of ‘purity’ unfolding as an elaboration of hierarchy on the basis of relative ‘purity’, without any reference to economic phenomenon, then the economic impulse within Indian society must surely have been very weak. Further, if the caste system has given India an unchanging hierarchy, India can have had no history that one may recognise as such. Both these positions Dumont readily espouses.
‘I would like to raise, he says, ‘the very question of the applicability to traditional India of the very category of economics.’ He points, in justification, to the ‘elementary’ fact that ‘even in our own [western] society it was only at the end of eighteenth century that economics appeared as a distinct category, independent of politics’. The argument is so illogical that one hesitates over whether one has understood Dumont aright. The fact that there was at one time no science of sociology does not mean that there have been no societies before the arrival of that science; similarly, because economics did not exist as a science before the eighteenth century, one is not excluded from speaking of the economic factors behind the English Civil War, or any other earlier historical process or event.
So too is India to be deprived of history:
The indifference to time, to happening, to history, in Indian literature and civilisation in general, makes the historian’s task very hard- But under these conditions, is there a history of India in a sense comparable to that in which there is a history of Christian civilisation or even [!] China? 
In other words, shall we say, no biography can be written of anyone who has not authored an autobiography!
Dumont offers here and there examples of how one can interpret India’s ‘non-history’. The most impressive example is his exposition of the rise and fall of Buddhism. The ideology of the caste system, he says, requires individuals’ renunciation of society. Some of ‘the renouncers’ begin competing with the Brahmanas. Out of such competition, the Buddhists and the Jainas expounded the doctrine of ahimsa and condemned animal slaughter and meat eating as polluting acts. This led the Brahmanas to give up animal sacrifice and stress vegetarianism to a greater or more systematic degree than even their challengers. The Buddhists were thus thwarted, vegetarianism became yet another symbol of purity, and the Brahmanas slept more easily until the next round of renouncers (e.g., the Lingaits) came round with some other eccentric competing propositions. Kosambi may point to the shift from pastoralism to agriculture, R.S. Sharma to the rise of towns and growth of commerce, in order to explain the success of early Buddhism; but their attempts are vain. The only factor behind it was a more successful appeal to the ‘idiom of purity’; and what made such an appeal at all possible was, again, the phenomenon of ‘renunciation’.
If such is to be history of India, to fit a contemporary western sociologist’s image of the caste system, is it not more likely that there is something wrong with this image rather than with Indian history? It may, in fact, well be that there is a good historical explanation for Dumont’s excessively narrow view of caste. During the last hundred years and more, the hereditary division of labour has been greatly shaken, if not shattered. As a result, this aspect has increasingly receded into the background within the surviving domain of caste. The purely religious and personal aspects have, however, been less affected. (One can see that this is by no means specific to India: religious ideology survives long after the society for which the particular religion had served as a rationalisation has disappeared.) It is obviously tempting to take the caste system’s surviving elements (mainly religious) as the sole or crucial elements, and the declining aspects (economic) as secondary and even superfluous. Dumont not only falls to the temptation, he builds a whole theoretical structure on a false premise to explain what India is. But then what he postulates about the hierarchical man in India is, perhaps, as difficult to accept as his other belief that western society today is ‘egalitarian’.
If then, Homo Hierarchichus fails to convert us, from where are we to begin? I think it is important to use the approach that Kosambi explicitly and consistently followed, the one that was introduced by Karl Marx. Caste should be viewed primarily in its role in different social formations that have arisen in a chain of sequence. A social formation, in so far as it is based on the form of the ‘labour process’, arises after the producers in society are able to provide a ‘surplus’. It is vain to expect a social institution like caste to exist before this stage has arrived. Indeed, Dumont himself recognises this, for he admits that the emergence of castes presupposes division of labour which cannot be found in primitive societies. The purusasukta in the Rigveda, the original statement for the four varnas, is more a description of social classes than of castes: the rajanyas, aristocracy, the brahmanas, priests, the vis, people at large (mainly peasants), and the sudras, springing from the dasyus, servile communities. There is no hint yet in Vedic times of either a hereditary division of labour or any form of endogamy. The varnas thus initially presaged very little of the caste system that was to grow later.
Kosambi, in An Introduction to the Study of Indian History, offers the view that castes did not arise out of any internal division of the varnas in the original Vedic society, but from an external process altogether: ‘The entire course of Indian history shows tribal elements being fused into a general society. This phenomenon . . . lies at the very foundation of the most striking Indian social feature, namely, caste.’ For this insight one can adduce confirmation from the use of the word jati. When the Buddha is spoken of as belonging to the Sakya jati the word obviously means a tribe. When, in the same literature, we also read of ‘excellent as well as low’ jatis, castes are clearly implied. Tribes are often rigorously endogamous: thus the Buddha’s story of the Sakya brothers who married their own sisters in order to avoid marrying outside the tribe. Can we suppose that as the tribes entered the ‘general society’, they carried their endogamous customs into that society? If the tribe was already an agricultural community, it would simply turn into the peasant caste of its territory.
However, the tribes ‘entering the general society’ would include a large number of primitive hunting or food-gathering tribes living in forests, who would be subjugated by the advancing peasant communities. This may be illustrated by the struggle between the Sakyas and Kolis. Kosambi has a long passage on the Nagas, the forest folk, who retreat before the Aryan advance, but leave their traces behind in brahmanical lore and later Vedic ritual. As the food-gatherers were subjugated they were reduced to the lowest jatis, so low as to be outside the four varnas altogether. The enumeration of the ‘mixed jatis’ in the Manusmriti shows a preponderance of such communities: the Sair Andhra ensnare animals, the Kawarta are boatmen; the Nisadas pursue fishing; the Medas, Andhras, Chunchus and Madgus live off the ‘slaughter of wild animals’; the Kshattris, Ugras and Pukkasas by ‘catching and killing [animals] living in holes’; the Karavara and Dhigvanas by working in leather; and the Pandusopaka by dealing in cane (Manu, X, 32, 34, 36-37, 48-49). The Chandalas and Nisadas both appear as hunters in Buddhist texts. These were the original ‘untouchable’ castes. Since they were excluded from taking to agriculture, and their own original or altered occupations were of minor or seasonal importance, they became a large reservoir of unfree, servile landless labour available for work at the lowest cost to the peasants as well as superior landholders. It is difficult to avoid the view that the bitter hostility which the rest of the population has displayed for these menial jatis had derived from this fundamental conflict of interest. Concepts of ‘purity’ and ‘pollution’ were a rationalisation of this basic economic fact.
The separation of the peasant and menial jatis represents a division of labour in a very generalised form. But R.S. Sharma has called attention to a second urban revolution (the first being represented by the Harappan culture) which took place on the eve of the rise of Buddhism. This implies that a multiplicity of productive skills must have developed. Gordon Childe has stressed the importance of ‘new tools and labour-saving devices, such as hinged tongs, shears, scythes, rotary querns’ for the emergence of ‘a number of new full-time specialists’. Some of these tools (shears, rotary querns) appear by the first century AD at Taxila. The Jatakas introduce us to the ‘manufacturers’ villages’, peopled by smiths and carpenters. It is possible that tribes brought wholesale into the general society began to throw off splinters under the pressure of the emerging division of labour. Different craftsmen isolated from the original tribe were formed into specific jatis. Thus Manu (X, 47-48) includes among the ‘mixed jatis’ those of carpenters, charioteers and physicians. A similar process of differentiation, based on the growth of commerce, led to the mercantile castes, which are quite prominent in the Jatakas. In time, they would make the vaisya varna exclusively their own.
The position of the brahmanas in the caste framework derived naturally enough from their priestly functions, and their guardianship of the dharma protecting the caste system. Kosambi suggests that part of their position also derived from their grasp of the calendar, which was so essential for regulating agricultural operations.
The one segment of caste structures most vulnerable to change was that of the ruling and warrior class, the kshatriyas (rajanyas). Invasions and rebellions made a hereditary monopoly of armed power extremely difficult, as the Puranas amply bear witness to. Thus where, logically, the caste system should have been strongest, in actual terms, it was the weakest- namely, in the stability of the ruling community. The entire caste structure has thus supposed a system of exploitation whose major beneficiaries, by its own terms, have so often been usurpers or outsiders.
Almost everyone seems agreed that in universalising the caste system within India, Brahmanas have played a key role, and that by integrating the caste doctrine into the dharma, brahmanas made the caste system and brahmanism inseparable. One result of these assumptions has been that the role of Buddhism in the process of caste formation has often escaped notice.
To anyone who reads Kautilya’s Arthasastra with its heavy stress on the varna system, and then turn to Asoka’s edicts, the contrast is a striking one. The word varna (or jati) never appears in Asoka’s texts; obedience to the varna rules does not form even implicitly a part of the dharma that Asoka propagated and whose principles he inscribed on rock and pillars. In so far as Buddhism rejected the religious supremacy of the brahmanas, it necessarily questioned the legitimacy of the varna division inherited from the Vedas.
And yet it may be asked whether Buddhism did not have its own contribution to make to the development of the caste system. The karma doctrine, or the belief in the transmigration of souls which formed the bedrock of Buddhist philosophy, was an ideal rationalisation of the caste system, creating a belief in its equity even among those who were its greatest victims. In the Manusmriti (XI, 24-26) it already appears as a firm part of the caste doctrine.
Second, there was the stress on ahimsa. Kosambi attributed the stress on avoidance of animal killing in Buddhism to the irrationality of large scale slaughter of livestock for sacrifice by Brahmanas, once settled agriculture had replaced pastoralism. Kosambi did not of course, intend to disparage the sincerity of the Buddha’s disapproval of violence or cruelty (and, after all, Asoka condemned the massacres by his army in Kalinga). What he implied was that any criticism of the large scale animal sacrifices would be popular among the ‘cattle raising vaisya’. But I would like respectfully to suggest what seems to me to be a more plausible reason why ahimsa should have become a popular doctrine. It provided reason for the subjugation and humiliation of the food-gathering communities. The Asokan edicts contain injunctions against hunting and fishing, and the Buddhist texts look down on ‘animal-killing jatis’ as much as the brahmanical texts do.
Indeed, here Buddhism also contributed to the ultimate denigration of the peasantry in the varna structure. R.S. Sharma’s exposition of how the sudra and not the vaisya varna came to be regarded as the category to which peasants must belong is practically definitive. In this denigration the ahimsa doctrine too was made to play was part. Manu (X, 84) condemns the use of plough for the injury that its iron point causes to living creatures. This is echoed in later Buddhism; I-tsing says that the Buddha forbade monks from engaging in cultivation because this involved ‘destroying lives by ploughing and watering field’.
It would, therefore, be wrong to suppose that the caste ideology has been exclusively brahmanical in its development.
The period from the rise of Buddhism (c. 500 BC) to the Gupta age (fourth and fifth centuries AD) may, then, be supposed to be the period of the formation of the caste system and its supporting ‘ideology’. It is significant that outsiders were struck not by the ‘hierarchy’ of the system, but by its hereditary occupations. Megasthenes (c. 300 BC), with his listing of the seven castes, and Yuan Chwang, both make unqualified statements in this respect, as do later foreign observers like Babur and Bernier.
Being a relatively rigid form of division of labour, the caste system formed part of the relations of production. But the caste system operated in two different worlds of labour, and these two must be distinguished in order to better understand both the caste system and the social formation of which it was a part. Marx derived a very important insight from Richard Jones, when he distinguished the artisan maintained by the village and the artisan of the town, wholly dependent on the vagaries of the market. In one case the caste labour belonged to a natural economy, in the other to a commodity or monetised sector.
Those who are familiar with Marx’s writing on the Indian village community may remember that he locates the base of its economy on two opposite elements existing side by side: ‘the domestic union of agricultural and manufacturing pursuits’, limiting thereby the domain of exchange within the village, and ‘an unalterable division of labour’, with the artisans and menials belonging to particular castes as servants of the village as a whole, maintained through customary payments in kind or land allotments, dispensing, again, with commodity exchange. Max Weber gave to this kind of caste-determined labour the name of demiurgical labour.
Modern sociologists since W.H. Wiser have been sprinkling cold water on this view. To them, the actual system was not of demiurgical labour, but the jajmani system, that is, a system of the artisans serving only particular families. For Louis Dumont, as one would expect, it becomes immediately a matter of a ritualistic relationship between certain upper caste families and the ‘purity’ specialists, viz. the brahmana and the barber, which was then extended to relationships with other artisans and labouring castes. So, we are told: ‘In the last analysis, the division of labour shows not a more or less gratuitous juxtaposition of religious and non-religious or “economic” tasks, but both the religious basis and the religious expression of interdependence. Further it deduces interdependence from religion.’ Dumont has apparently let his edition run on without paying much attention to current historical work. In 1972 Hiroshi Fukazawa published the results of his investigations in the eighteenth century records, in which Maharashtra is so rich. His definite conclusion, closely based on documentary evidence, was that the jajmani theory was applicable only to the family priesthood; the traditional twelve balutas (carpenter, smith, potter, leather-worker, barber, etc.) were basically village servants, paid through land allotments (watan) and out of peasant crops. So strong is Fukazawa’s evidence that he is led to the perceptive comment that if modern village artisans appear to be servants of certain families alone, this is due to the decay of the old system under modern conditions.
Historical evidence for the village servants in fact goes back to a fairly early period. Kosambi cites epigraphic evidence attesting to carpenters’ plots in north Indian villages going back to the fifth century. Similarly, B.N.S. Yadava draws attention to the Lekhapaddati documents (Gujarat, c. 1000), which speak of the five village artisans (pancha karuka), viz. the carpenter, ironsmith, potter, barber and washerman, entitled to receive handfuls of grains from the peasant. The balahar, or the village menial, appears as the lowliest landholder in Barani’s account of Alauddin Khalji’s taxation measures (early fourteenth century).
The hereditary artisan and servant, thus, was of crucial importance in sustaining the self-sufficiency as well as the internal natural economy of the village. Such self-sufficiency not only isolated the village, but enlarged its capacity to deliver a larger part of the surplus to the ruling class, since it did not need much extra produce to exchange for its own imports.
As the surplus was taken out of the village, it entered the realm of commodity exchange, as Marx particularly noted in his classic passage (already cited) on the Indian village community in Capital, Volume 1. Outside the village the artisan appears as an individual selling his wares on the market. The hereditary occupation by caste was necessary to enable ‘special skill’ to be ‘accumulated from generation to generation’. The hereditary transmission sustained skill while excluding even horizontal mobility. In addition, the caste system possibly created another element of advantage for the ruling class, by giving a lowly status to many artisan castes. Artisan castes already appear among the mixed jatis in Manu; and in the eleventh century Alberuni classes eight professions, including those of weavers and shoemakers, among the outcaste antyajas. Their depressed status and lack of mobility must surely have helped to curtail the powers of resistance of the artisans and so to keep wage costs low.
The caste system, in its classic form, could therefore function with as much ease in a natural economy as in a market-oriented one. In either case it helped essentially to maintain not a fabric of imagined purity (if it did, this was incidental), but a system of class exploitation as rigorous as any other.
In many ways the beginning of thirteenth century marks a ‘break’ in Indian history. This break arises not only from the intrusion of Islam: we begin to see a social formation which is at last close to Marx’s ‘Oriental despotism’ as against the preceding age of ‘Indian feudalism’ delineated by Kosambi and R.S. Sharma. We must, however, allow for a much larger extent of commodity production and urbanisation than Marx seems to have visualised for pre-colonial India.
The caste structure in both villages and towns continued essentially to be the same as in the earlier period. As will be seen from what we have said in the foregoing section, the evidence for hereditary caste labour in villages and towns is practically continuous from ancient India to the eighteenth century. It is true that Islam in its law recognises differences based only upon free man and slave (and man and woman); caste, therefore, is alien to its legal system. Nevertheless, the attitude of the Muslims towards the caste system was by no means one of disapprobation. When in 711-14, the Arabs conquered Sind, their commander Mohammad Ibn Qasim readily approved all the constraints placed upon the Jatts under the previous regime, very similar to those prescribed for the Chandalas by the Manusmriti. Muslim censures of Hinduism throughout the medieval period centre round its alleged polytheism and idol worship, and never touch the question of the inequity of caste. The only person who makes a mild criticism of it is the scientist (and not theologian) Alberuni (c. 1030) who said: ‘We Muslims, of course, stand entirely on the other side of the question, considering all men as equal, except in piety.’ But such an egalitarian statement is almost unique; the fourteenth century historian Barani in his Tarikh-i iruz-Shahi fervently craved for a hierarchical order based on birth, although he was thinking in terms of class, rather than of castes, and does not appeal to the Hindu system as a suitable example.
In so far as the caste system helped, as we have seen, to generate larger revenues from the village and lower the wage costs in the cities, the Indo-Muslim regimes had every reason to protect it, however indifferent, if not hostile, they might have been to Brahmanas as the chief idol-worshippers. (Does not this also mean that the supremacy of the Brahmanas was by no means essential for the continuance of the caste system?) Nevertheless, the caste system had to undergo certain adjustments and changes, which must be recognised as important, not as a result of the policy of the Sultans, but of the new circumstances.
In the first place, the new ruling classes and their dependents brought not only demand for new products and new kinds of services, from their central and west Asian backgrounds, but also a fairly wide range of new craft technology. Kosambi, with his usual perceptiveness, spoke of the ‘Islamic raiders — breaking hidebound custom in the adoption and transmission of new techniques’. Among the technological devices which came early (thirteenth and fourteenth centuries) were right-angled gearing (for the final form of the Persian wheel), the spinning wheel, paper manufacture, vault construction, use of bitumen and lime-cement, iron horse-shoe, and so on. These necessitated, in some cases, the creation of new professions (e.g., paper-makers, lime-mixers) and, in others, the learning of new devices; in general, one can postulate the need for a considerable expansion of the artisan population, to accompany what, in fact, may be designated the third ‘urban revolution’ in Indian history.
The pressure of the new circumstances led initially to large scale slave- trading and the emergence of slave labour as a significant component of urban labour during the thirteenth and fourteenth centuries. The numbers of slaves in the Sultan’s establishments were very high (50,000 under Alauddin Khalji, and 1,80,000 under Firuz Tughluq). Barani judges the level of prices by referring to slave prices, and the presence of slaves was almost all-pervasive. Slaves were, in effect, deprived of caste and, converted to Islam, could be put to almost any task or learn any trade. Manumitted in course of time, they probably created, along with artisan immigrants, the core of many Muslim artisan and labouring communities. There was also in time a conversion of free elements, possibly in many cases sections of castes splitting off from their parent bodies in search of higher status or willing to take to occupations of practices not permitted to them previously. There thus arose, in course of time, a substantial Muslim population.
Caste undoubtedly continued to exercise its influence on these communities. To judge from their practices as reported since the nineteenth century, weavers, butchers, barbers and others had strong tendencies to be endogamous. The menial castes duplicated themselves as kamin communities among Muslims, not untouchable but still kept separate and held in contempt. Nonetheless, it would still be correct to say that sectors of Muslim populations remained outside the caste framework even in its most rudimentary form; and in any case, the framework remained weak, since both shifts of occupation and deviations from endogamy could occur. In other words, there was always a much greater degree of mobility.
It is questionable whether the presence of such relatively caste- free populations at all undermined the caste system. Such populations might indeed have reinforced it by providing reserve labour for new professions or occupations without causing any disruptions to the structure of existing castes. But it is also doubtful whether the caste system was so completely devoid of capacity for mobility as has been assumed by Max Weber. Morris D. Morris in a notable paper argues that in actual practice the caste system has been vastly different from how one thinks it should have operated on the basis of the law book — or what Dumont calls its ‘ideology’. Castes divided to enable one section to take to new professions: Fukuzawa draws attention to a well-documented caste from eighteenth century Maharashtra, where a section of tailors took to dyeing and yet another to indigo-dyeing and set up as endogamous sub-castes. A historically singular case is that of the Jatts, a pastoral Chandala-like tribe in eighth century Sind, who attained Sudra status by the eleventh century (Alberuni), and had become peasants par excellence (of vaisya status) by the seventeenth century (Dabistani-i Mazahib). The shift to peasant agriculture was probably accompanied by a process of ‘sanskritisation’, a process which continued, when, with the Jat rebellion of the seventeenth century, as section of the Jats began to aspire to the position of zamindars and the status of Rajputs.
Moreover, where sanskritisation failed or was too slow a process, hearing began to be given to monotheistic movements, which condemned the ‘ideology’ of the caste system. It may be that the monotheistic belief of Islam and the legal equality of the Muslim community exercised a certain influence on these movements. But their stress on equality and condemnation of caste and ritual observance was certainly much greater than is to be found in any contemporary Islamic preaching. Most of its great teachers belonged to the low jatis: Namdev, a calico printer; Kabir, a weaver; Raidas, a scavenger; Sain, a barber; Dadu, a cotton-carder; Dhanna, a Jat peasant. In beautiful verse, composed in the name of Dhanna Jat, the fifth Guru of the Sikhs (Arjun) insists on God’s special grace for such lowly worshippers.
In these communities (panths) the doors were open to people of all castes. The Satnami sect (which arose in the seventeenth century, with some allegiance to Kabir), contained goldsmiths, carpenters, sweepers and tanners, according to one account, and peasants and traders of small capital, according to another. The Sikh community in the seventeenth century consisted in bulk of Jat peasantry; early in the next century, the complaint was being made that authority could be given among them to ‘the lowliest sweeper and tanner, filthier than whom there is no race in Hindustan’. The practices of these panths forbade caste distinctions within the community, and, there was a tendency in the communities, as with the Satnamis, amongst whom this was prescribed by scripture, to become endogamous. The net result was the creation of religious communities which drew their following from the caste framework but which ultimately returned to that framework though usually at a higher ‘rank’ than at the time of their departure from it. This had happened before, as in the case of the Lingayats in Karnataka; and these movements of the sixteenth and seventeenth centuries similarly made necessary adjustments in the caste system, without however subverting it.
The caste system, therefore, remained an important pillar of the system of class exploitation in medieval India. As we have said before, its chief beneficiaries could only be the ruling classes; in the medieval case, these were, first, the nobility and, second, the rural superior class, the zamindars. To the extent that the political structure was sustained by the zamindars, caste again was important, since the zamindars, by and large, belonged to the ‘dominant castes’ which maintained their position by force. It is worth remembering that when Abu-l Fazl in his detailed statistical tables of the Mughal Empire in the A’in-i Akbari (1595-96) gives the caste of zamindars of each locality, this information is followed not by the area of land they held but by the numbers of their retainers, horse and foot. There was, therefore, an undoubted connection between caste dominance and military power.
Barrington Moore Jr expresses some surprise that in his detailed descriptions of the Mughal Indian economy, W.H. Moreland should have had so little to say on caste. This applies to some of my own work as well, the reason, perhaps, is that when one looks at the specific relations, such as those of the peasant and the tax appropriator, or the petty producer and the merchant, caste is not immediately visible. What it did mainly was to provide a large part of the setting for these relations. It divided the agrarian classes into two antagonistic camps, the caste peasants and the menial labourers; and it stabilised the division of labour in petty production. But it is questionable if these functions were crucial enough for us to propound that caste defined the form of the labour process in medieval India (c. 1200-1750). The factors of mobility and competition were present to a certain degree, as we have seen. Iran was very similar to India in its economic and political organisation in medieval times, but without the benefit of the caste system. Can we nevertheless say that Mughal India and Safavid Iran belonged to two separate social formations, just because one had caste and the other lacked it? Any comment on the matter can at present be only tentative, and one may look forward hopefully to more discussion on the subject.
But one final word, before we leave the question of caste in medieval India. Any class formation like medieval Indian society was bound to generate internal tensions, finding expression notably in the struggles of the oppressed. India has a history of peasant uprisings, going back to the revolt of the buffalo-riding Kaivartas of Bengal in eleventh century. But it is from the seventeenth century that we get perhaps the richest evidence of peasant uprisings. One great weakness of these uprisings, when compared with those of Europe or china, is the rebels’ extremely backward class consciousness. Peasant rebels appear as zamindars’ followers (Marathas), or members of religious communities (Sikhs, Satnamis) or of castes or tribes (Jats, Afghans); they fail to see themselves as peasants or to raise economic or social demands for any section of the peasants. It seems to me that caste provides part of the reason for this failure. It prevented peasants of one caste from finding common ground with those of another, and so all the time undermined the growth of self-awareness of the peasantry as a class.
The history of caste in the foregoing pages has been brought close enough to modern times. I do not intend to pursue it further, but a few concluding remarks may still be offered.
In 1853, discussing the results of British rule, Marx predicted that ‘modern industry, resulting from the railway system will dissolve the hereditary divisions of labour upon which rest the Indian castes, those decisive impediments to Indian progress and Indian power.’ It has become customary to deride this statement (vide Louis Dumont) as having been too optimistic. But it will be futile to deny that modern conditions have gravely shaken the economic basis of the caste system. This is not only because workers of several castes come together on the factory bench; and once this happens the traditional division of labour begins to collapse. Even more important has been the fact that industrial production has destroyed the crafts of a whole series of professional or artisanal castes. This process began with the import of Lancashire cloth, even before Marx was writing. The basis of the caste division of labour has weakened due to one further factor. As Surendra J. Patel has pointed out, commercialisation of agriculture has converted large numbers of peasants into landless labourers, so that no longer does landlessness remain a monopoly of the ‘menial’ castes, though they still form in most areas its largest contingent.
But if the economic base of the caste system has been shaken, can the same be said of its ideology? Endogamy continues to reinforce caste; and there has been a process of territorial enlargement of castes through mutual identification and absorption. ‘Sanskritisation’, which modernisation at one level strongly fosters, converts the erstwhile victims of the system into its votaries. So long as the conflict of interest between landless labour and landholding classes remains, there is an incentive for all castes to combine against the untouchables, whom we euphemistically call the scheduled castes. Caste still remains perhaps the single most important divisive factor in our country.
For all those who wish to see the Indian people united in a struggle for their material and spiritual liberalisation, it is of utmost importance that there be a renewed effort to eradicate the sway that caste continues to hold over the minds of our people. What Marx called the decisive impediment to Indian progress could only then be removed, and caste at last relegated to history, to which it properly belongs.
 See J.H. Hutton, Caste in India, fourth edition, Bombay, 1969 (reprint), p. 71 and passim.
 All my references are to the Paladin edition, London, 1972.
 Homo Hierarchicus, p. 300. (‘Hierarchy culminates in the Brahman.’)
 Ibid., pp. 288ff.
 Ibid., p. 300.
 Ibid., pp. 262, 274. Accordingly, to Dumont, ‘Hindus and Muslims form two distinct societies’ (p. 257).
 Ibid., p. 209.
 Ibid., p. 242
 Ibid., pp. 230-231.
 Ibid., pp. 192-95.
 Dumont shows an almost total lack of awareness of this development in his remarks on the unchanging caste framework. Ibid., pp. 265-66.
 So Dumont can now say: ‘In the caste system the politico-economic aspects are relatively secondary and isolated’. Ibid., p. 283.
 Ibid., pp. 260, 331-32. Compare Kosambi’s dictum that ‘caste is class on a primitive level of production’. Culture and Civilization of Ancient India in Historical Outline, London, 1965, p. 50.
 D.D. Kosambi in An Introduction to the Study of Indian History, Bombay, 1956, p. 25.
 Narendra Wagle, Society of the Time of the Buddha, Bombay, 1966, pp. 122-23.
 Ibid., pp. 103-4.
 Kosambi, Introduction, p. 122.
 Ibid., pp. 121-23.
 See Kosambi, Culture and Civilization of Ancient India, p. 15.
 See Vivekanand Jha in Indian Historical Review (IHR), II(1), pp. 22-23.
 R.S. Sharma’s review of A. Ghosh, The City in Early Historical India (Simla, 1973), in IHR, I (1), pp. 98-103.
 V. Gordon Childe, Social Evolution, edited by Sir Mortimer Wheeler, 1963, p. 110.
 See Sir John Marshall, Taxila, II, Cambridge, 1952, pp. 486 (rotary querns) and 555 (scissors).
 R. Fick, The Social Organization of North East India in Buddha’s Time, English translation, Calcutta, 1920, pp. 280-85.
 Kosambi, Introduction, pp. 236-37.
 Ibid., pp. 158-59.
 Recorded on Rock Edict XIII.
 Especially P.E.V. and the bilingual Qandahar inscription.
 R.S. Sharma, Sudras in Ancient India, Delhi, 1958, especially pp. 232-34.
 A Record of the Buddhist Religion as Practised in India and the Malay Archipelago, translated by J. Takakusu, Oxford, 1986, p. 62.
 For Megasthenes see R.C. Majumdar, The Classical Accounts of India, Calcutta, 1960, pp. 224-26, 263-68; Yuan Chwang, Buddhist Records of the Western World, I, translated by S. Beal, London, 1884, p. 82; Babur, Baburnama, translated by A.S. Beveridge, London, 1921, p. 520; Francois Bernier, Travels in the Mogul Empire, 1655-68, translated by A. Constable, edited by V.A. Smith, Oxford, 1916, p. 259.
 Karl Marx, Theories of Surplus Value, Vol. III, English translation, Moscow, 1971, p. 435.
 Capital, I, edited by Dona Torr, translated by Moore and Aveling, London, 1938, p. 351.
 Dumont, Homo Hierarchicus, p. 150.
 H. Fukazawa, ‘Rural Servants in the Eighteenth Century Maharashtrian Village – Demiurgic or Jajmani System’, Hitotsubashi Journal of Economics (HJE), XII(2), 1972, pp. 14-40.
 Kosambi, Introduction, p. 312.
 B.N.S. Yadava, Society and Culture in Northern India in the Twelfth Century, Allahabad, 1973, p. 267.
 Ziauddin Barani, Tarikh-I Firuz-Shahi, edited by S.A. Khan, W.N. Lees and Kabiruddin, Bib. Ind., Calcutta, 1860-62, p. 287. On balahar, see H.M. Elliot, Memoirs of the History, Folklore and Distribution of Races in the North-Western Provinces, II, edited by John Beames, London, 1869, p. 249; and Irfan Habib, Agrarian System of Mughal India, pp. 120-21.
 Marx, Capital, I, pp. 331-32.
 Edward C. Sachau, Alberuni’s India, I, London, 1910, p. 101.
 Anonymous, Chachanama, Persian version of the thirteenth century, edited by U-Daudpota, Hyderabad-Dn., 1939, pp. 214-16 (also see pp. 47-48). A later Arab Governor insisted that the Jatts should, as mark of identification, be always accompanied by dogs. See Elliot and Dowson, History of India as told by its own Historians, I, London, 1867, p. 129.
 Alberuni’s India, I, p. 100.
 Kosambi, Introduction, p. 370.
 See Irfan Habib, ‘Changes in Technology in Medieval India’, Studies in History, II, NO. I (1980), pp. 15-39.
 This paragraph is based on evidence already presented by me in Cambridge Economic History of India, Vol. I, Cambridge, 1982, pp. 89-93, where detailed references to sources will be found.
 See D. Ibberson’s remarks on conditions in western Punjab ‘where Islam has largely superseded Brahminism’, Punjab Castes, Lahore, 1916, pp. 10-11.
 Morris D. Morris, ‘Values as an Obstacle to Economic Growth in South Asia,’ Journal of Economic History (JEH), XXVII, pp. 588-607.
 HJE, IX(1), (1968), pp. 39ff.
 This evidence is examined in my paper ‘Jatts of Punjab and Sind’, Punjab Past and Present: Essays in Honour of Dr Ganda Singh, edited by Harbans Singh and N.G. Barrier, Patiala, 1976, pp. 92-103.
 Saqi Musta’idd Khan, Ma’asir-I Alamgiri, edited by Agha Ahmad Ali, Calcutta, 1870-73, pp. 14-15.
 Khafi Khan, Muntakhabu’l Lubab, edited by Kabiruddin Ahmad, Calcutta, 1860-70, Vol. II, p. 252.
 Muhammad Shafi Warid, Miratu’l Waridat, British Library (London), MS., Add. 6579, f. 117b.
 The Satnami scripture, containing these injunctions, is titled Pothi Gyan Bani Sadh Satnami, and is preserved in Royal Asiatic Library, London, MS. Hind. No copy is known.
 This point is lost in Jarrett’s translation of the A’in-i Akbari, since the translator has altered the arrangement of the columns. See my Agrarian System of Mughal India, Bombay, 1963, p. 139.
 Social Origins of Dictatorship and Democracy, 1977, pp. 317-18 n.
 R.S. Sharma, Indian Feudalism, second edition, Delhi, 1980, p. 220.
 Homo Hierarchicus, p. 265.
 Surendra J. Patel, Agricultural Labourers in India and Pakistan, Bombay, 1952, especially pp. 9-20, 63-65.