Skip to content

Reflections on Looking Again at the Aryan Question

Romila Thapar

A Harappan excavation site, from The Telegraph Archives

I would like to begin by conveying birthday greetings to Professor Irfan Habib. All good wishes Irfan. You have in your wisdom provided us with excellent historical vision and insight, and you have received the extensive admiration and affection of colleagues and friends. I wish you yet more of this in the years to come.

There is in all of us an anxiety about the times in which we live. Your choice of such an appropriate subject for this evening – the defence of history — is an indication of your concern, as it is of many of us. A spark of hope however lies in the conviction that no matter how fierce the attempt to annul the history we study and write, ultimately the explanation of the past will have to come from rational and logical argument, drawing from reliable sources of information. But the ultimate may be distant.

So I thought it might be worth exploring briefly one theme of current interest to illustrate what I mean. Answers to what happened in the past are never single and uniform. Their very complexity and diversity demands further exploration. One has to recognise how even one theme has undergone changing explanations.

I would like to speak about what used to be called the Aryan question – a subject that has come alive in contemporary politics and has become highly controversial. Today it seems to go in two different directions. Popular opinion is obsessed with the origins of the Aryans and with pushing their date back to earlier and earlier times. Professional historians are focusing on the relatively new sources and the fresh dimensions that these sources suggest, in order to work out a more precise history and chronology of this period. 

The history began to be written in the nineteenth century. When the Vedic texts came to be studied by European Sanskritists, it was discovered that Vedic Sanskrit had close affinities with other related languages, hence the label Indo-European languages. The ancestral language is referred to as Proto-Indo-European, the knowledge of which is based on a linguistic reconstruction. These languages had spread from a location in the proximity of Central Asia and northern Iran going westwards to eastern Europe. The speakers of the earliest form of Indo-Aryan were thought to be genetically related to the speakers of early Indo-European and Vedic Sanskrit evolved from this. 

What was called ‘race-science’ emerged in the late nineteenth century. Language and race were equated and there were theories about the superior Aryan race. It was often forgotten, as it still is, that the term Aryan identifies language and not race. Any number of racially diverse cultures can pick up the same language in a given historical situation. (An illustration of this could be the contemporary example of the use of American English in the USA).

Historians held that in the early second millennium BC, groups of agro-pastoralists from Central Asia migrated, probably in search of pastures. A few went to north-eastern Iran and a few to nearby north-western India and the Panjab. A sequence of migrations is listed in the Iranian Avesta. Could it be the same ? The Iranian-Aryans had close language affinities with the Indo-Aryan speakers. The texts of the two groups have some striking similarities and some curious inversions. 

Colonial scholarship attributed the origins of Hindu culture and religion to this earliest stratum of what was described as Vedic Aryanism. Max Mueller, among others, argued to this effect in the nineteenth century, projecting the idea of a superior people. The term arya was more often used as a qualifier with reference to persons, actions, thoughts, in ancient Iranian and Indian texts, to mean that which is respected. Thus there are references to the arya varna in the Vedas as the people who have status and the same qualifier is applied in Buddhism to the four noble truths, referring to the much respected teachings of the Buddha. 

These ideas were also propounded by the influential Theosophists, such as Col. Olcott, who were based in India in the late nineteenth century. They agreed with the usage of arya as people but disagreed with the theory of origins and added the argument that the Aryans originated in India. Aryan culture was centre-stage in this theory as it was in the Arya Samaj, which was initially and for a short while, closely associated with the Theosophists. 

This simple and comfortable reconstruction of the past was disrupted in the 1920s with the discovery of the Indus civilisation, dating to the Pre-Vedic third millennium BC. This was primarily a sophisticated urban culture, familiar with literacy. Their language has still to be deciphered from the pictograms they used. Both these features – cities and literacy – are characteristic of an urban civilisation, and are absent in the agro-pastoral Vedic culture. 

The Aryan presence was a subsequent and dissimilar culture from the Harappan. However, some attempts are made these days to continue to project an Aryan foundation by insisting that the Harappans were also Aryan, although there is no evidence for saying so. Labelling everyone as Aryan is an easy way of avoiding the investigation of a complex period of history, requiring an understanding of a range of divergent sources. It also annuls the richness that comes from the interface of multiple cultures and the complexities involved in comprehending their origin and growth. 

In the 1930s, when the Hindutva version of Indian history was constructed it was based on two major colonial theories : one was that of the Aryan origins of Hindu culture and of the Hindus, and the other was the two-nation theory of antagonistic Hindu and Muslim nations, as propounded by James Mill and other colonial historians. In the last half-century however, professional historians have shown the fallacy of these two theories and much else in colonial history. Consequently there is today a strong disagreement between the Hindutva view based on colonial theories and that of professionally trained historians. The latter have raised a different set of historically challenging questions, the answers to which provide more viable explanations of what might have occurred at that time.

My intention is to mention some of the inquiries that professional historians are making, in trying to reconstruct the events of the second millennium BC And explain why they differ from the Hindutva view. This requires a reference to the important aspects raised by these historically more relevant questions. 

Let me begin by saying that the Vedic texts remain central to this study. But there are now new ways of obtaining fresh, important information from these texts. New disciplines, sources, methods, have surfaced that have changed the format of investigations. I shall only mention those that are at the moment, under active discussion : geography, archaeology, linguistics and archaeogenetics — genetics linked to archaeology.


The Harappa culture was geographically far more extensive than the initial Vedic culture. Harappan settlements are found from the Pamir mountains in the north, coming south along the wider Indus plain and across the Persian Gulf to Oman in the Arabian peninsula. Towards the western end they went roughly from Baluchistan eastwards towards the upper Doab. Contacts were nurtured with the Gulf and Mesopotamia, with the Harappans supplying them, it is thought with lapis lazuli and copper and possibly ivory. Thus the geographical orientation of the Harappa culture tended to be westwards and differs from that of the Vedic. 

The Rigveda knows the upper north-west coming south-eastwards to the Doab. The migration of Aryan speakers within the sub-continent was eastwards from the Doab into the Ganges plain. The Shatapatha Brahmana narrates how Videgha Mathava travelled east with his people, and carried Agni in his mouth – a beautifully symbolic thought. The land in the middle Ganges plain was cleared of marshes and settled. 

As for links with areas outside the subcontinent, the only substantial one was with the adjoining area of North-eastern Iran where the Iranian Aryans settled. The use of the term arya to identify language and culture was used in Iran into historical times under the Achaemenid dynasty. But there is little reflection of a large-scale migration westwards in Vedic sources. Mention of moving west from the Indus stops with the Indian borderlands. This had suggested early on that the Iranian and Indian aryas were two initially linked but gradually differentiated cultures. Had they migrated from India to Iran there would have been mention of this is some detail in Vedic sources especially as the Iranian Aryans had differences. The Vedic texts mention the eastern movements of the aryas more fully. Nor would there have been an inversion of concepts and meanings of terms on such a substantial scale as there is in the Avesta and the Rigveda for instance, had the Iranian Aryans been migrants from India. The languages too would have been closer in form. It would seem that the two went their separate ways fairly early on. 

A far more fleeting and very brief presence of Aryan usage was apparent in west Asia in the mid-second millennium BC but disappeared fairly soon. It is thought to have been a visitation of Aryan speakers from the north of Iran following the southern Caspian route.


In the nineteenth century Vedic textual studies focused on the language of the texts and the rituals they described. The structure of grammar was studied as well as philology and phonetics. Vedic Sanskrit, because it was linked to the sacred, was assumed to have been isolated from other languages but this assumption was soon questioned.

In the twentieth century linguistic methods began to be applied to Vedic Sanskrit. In studying the evolution of the language, comparable forms from other languages used in the vicinity began to be recognised. Speakers of other languages can be inferred even from an analytical investigation of the Vedic texts. Any language in the proximity of another language cannot remain unaffected by the other. Contemporary non-Aryan languages in North India came to be known and dated. Vedic Sanskrit began to be seen as having evolved in the midst of other languages. Some imprint from these would be evident. What were these languages? The two likely ones were Dravidian and Munda. The Dravidian presence has drawn more attention among scholars. 

Among the more striking elements of Dravidian in the earliest Indo-Aryan seems to have been phonetic. Characteristic of the Dravidian language are the retroflex sounds, not found in any other Indo-European language, not even Iranian Aryan, but present in Indo-Aryan and located in the alphabet. This has suggested the interface between the two languages – Indo-Aryan and Dravidian – and therefore between the communities that spoke the languages. There is also some vocabulary in Vedic texts and other linguistic elements that point to a possible proximity of Dravidian-speakers. Given this the question to be asked is whether there was the presence of other different cultures alongside the aryas? References in Vedic texts to peoples and cultures that are said to be alien to the arya, need to be examined as differentiated cultures, but nevertheless proximate.

There are many references to non-aryas. The term asuras for instance, has been translated as demons, but could in some contexts refer to hostile others of a different culture. Among their alien cultural habits are that they bury their dead in graves. This was a differentiating comment from those who more frequently cremate their dead. In other contexts the asuras are said to be good spiritual beings, as for example among the Iranian Aryans where they are referred to as ahuras, the ‘s’ of Indo-Aryan becoming the ‘h’ of the Iranian-Aryan.

But the most oft-repeated reference to the culturally “Other” is of course the word dasa. The arya varna is differentiated from the dasa varna. The dual division means that the dasa was not marginal. The dasas are adeva, do not worship the same gods so are said to be without gods, they are avrata, because their rituals differ so they are said to have none, and they are mridhravach since they either do not speak Indo-Aryan or do so incorrectly. Some dasas are quite rich in cattle wealth and are therefore raided by the aryas

There is even an occasional incorporation of a dasa into arya society when a situation requires it, as for example, when something is needed from the dasa. This is illustrated by an interesting story in the Aitareya Brahmana. It refers to a curious social category, that of the dasyah-putrah brahmana or dasi-putra brahmana, literally, the brahmana who is the son of a dasi

On one occasion at the end of a ritual, the sage Kavasha, was dismissed by the brahmanas. They refused to let him eat with them, he being the son of a dasi. As he was going away he prayed to the river Sarasvati and the river started following him. So the brahmanas realised that he was someone special and quickly invited him back, not only to join them, but also conceded his superiority. Two questions become obvious. Was there some aspect of knowledge that the brahmanas did not have but were willing to be taught by those who had it, even if such people were initially despised? Was the category of brahmana a more open category than is stated in some of the later Vedic texts? 


There has not been any decisive evidence from archaeology that can be identified as Aryan, despite many excavations strewn across south Punjab and Haryana and the upper Doab, suggesting provocative questions. 

Various post-Harappan sites have been excavated, revealing the presence and interface of a number of cultures rather than a single culture conforming in its entirety to Vedic society. The origins of these cultures appear to be diverse, nevertheless some suggest a possible coming together. This is often so when cultures emerge from the demise of a possibly centralised system and struggle to create their own identities. Clearly Post-Harappan cultures have to be examined more fully, both for continuities from the Harappan, and as mixed cultures that incorporate elements from the variant societies of the region, and even beyond. 

The huge cemetery and elaborate burials from Sinauli in western UP for instance, are posing some tantalising questions. Among the more obvious is that the centrality of the ritual of burials and the richness of the grave furniture differs from the comparatively simple Harappan burials. Neither do Vedic texts refer to graves as the norm, and more so not graves in which chariots and coffins heavily decorated with copper are also buried, not to mention other objects including a range of pottery. The juxtaposition evokes elements of Harappan together with non-Harappan, reflecting what may be local traits and some perhaps from more distant cultures. Identifying the strands and placing them in context, may be complicated by the mixture of finds, and may require some new ways of contextualising the data. 

Archaeological reconstructions of the northern sub-continent give little attention to the distinctively different archaeological cultures of megaliths that dominate the peninsula from the late second-early first millennia BC. Occasional megalithic sites surface north of the Vindhyan borders, but should there be evidence of contact south of the border, the cultural communication might hold much potential.


The most recent source likely to provide clues of a different kind is the evidence from archaeogenetics, using some data from archaeology. Its information comes from DNA generally taken from skeletal material. DNA studies are of recent vintage. Samples are difficult to come by as many get contaminated, having been buried for many centuries. 

The genetic evidence has suggested that the people of the Harappa Culture may have originated from hunter-gatherers and farmers from Iran. Sometime just after about 4000 years before the present, that is after 2000 BC, new population strains entered north India. Some links with Central Asia are present. This has been read by some as evidence of migration from Central Asia into north India. Attempts have been made to suggest that these were the Aryan-speakers, but this can only be tentative for the moment. 

The genetic composition of the North Indian Ancestry is not a single pure genetic identity but a mix of genetic strains. More detailed evidence and definitive testing will clarify these hypotheses. At the same time those who insist on the Aryans being indigenous to India and refute any migrations into India, describe the genetic evidence as unreliable. Equally vehemently the geneticists maintain that it is reliable. 

In the study of the spread of the Proto-Indo-European language from near central Asia westwards to Europe, there is an increasing interest in the data from archaeogenetics and the investigations that these suggest. Migrations, whether within a defined area or beyond it, lead to questions related to demographic history. These concern the numbers that migrated as also the social categories. A frequently asked question is whether a migration was male-based. This is important to know because if it was, then it would suggest that the migrating men would tend to marry the women of the area where they settled. This in turn could point to marriages between people of different cultures and origins.

Would this for instance throw light on the puzzling category of men referred to in the Vedic texts as the sons of dasis. In the Brihaddevata, the Rigvedic seer Kakshivant is said to be the son of the sage Dirghatamas by his dasi wife Ushij and this also explains why Kakshivant is also called Aushija after his mother. This is not the only reference to such ancestry.

Other questions concern technology. Does a migration bring new technologies that might necessitate discontinuing some of the existing ones and adopting the new. The Indo-European migration to Europe is thought to have introduced Europe to horse-riding, the Indo-European language, and elaborate burial practices. Could the change from bronze technology to the selective use of iron in some areas have been the result of migrants bringing in a new technology ? Archaeogenetics is a new discipline and its study so far has been suggestive of providing some useful data.

Some Thoughts

The evidence from new categories of sources requires us to ask new questions about this period of history. We don’t have to discard the old questions but we have to consider the additional questions that have now become pertinent and relevant and the old may well require modification. Reconfigurations and discussion are called for. Can the old reading be reconstructed in the light of the new ? This applies as much to popular takes on history as it does to professional history. But of course since professional history has kept abreast of the new sources it will require less reconfiguration. 

To take an example from the constantly propagated popular history : V.D. Savarkar, when establishing the concept of Hindutva, defines the Hindu as the one who can claim India as the land of his ancestors – pitribhumi – and as also the land where his religion originated – punyabhumi. Since the Aryans were claimed as the ancestors of the Hindus, they had to be indigenous to India and their religion had also to have originated in India. Therefore, Hindutva cannot accept the fact that the Aryans originated anywhere else and migrated into India. The Hindutva definition of the Hindu is of course an invention of modern times and is not found in ancient texts. The Hindutva version of ancient history drawing on colonial theories of the nineteenth century is not acceptable to contemporary historians.

The point that I am trying to make is that we have now entered an era where a person does not become a historian by reading six books on history. To be a historian two requirements are vital. 

A historian has to be properly trained in the methods of researching history and understanding the social sciences. Each piece of evidence whether a text or an object, has to be analysed, and its reliability established. Many more questions are being asked these days given the range of new sources and methods of analyses. One has to be trained to test the reliability of a source and this training differs according to the source and requires expertise. Oral history is analysed differently from literary sources. A numismatist asks different questions of his data from an epigrapher. The logic and reasoning of the explanation is clarified. Only then are those explanations selected that seem most convincing. More than one explanation is debated in understanding historical situations.

Historical explanations therefore, tend not to be single answers based on a single source. Cultures and civilisations as patterns of living and thinking, evolve from the interaction of ideas and activities of consent and of dissent, manifested in a variety of facets pertaining to the participating societies. Civilisations are no longer seen as unique, self-contained and limited to an elite. Nor are they labelled by religion or dynasty or language since a civilisation can be confined to one qualifier. More often than not it incorporates more than one religion, dynasty or language in a crucial role. Civilisations are socially and economically porous, incorporating the migration of peoples and ideas as well as the interface of cultures. They implicitly reflect the participation and contribution of all levels of society in varied ways. The description of a civilisation is no longer confined to the activities of the elite. It has to include other sections of society and their patterns of living. If we today are truly civilised then we should acknowledge this and search for our multiple origins. 

Counting the number of civilisations across the world is now something that few historians would endorse. Similarly we prefer not to speak of Hindu, Muslim, Confucian and Christian civilisations, since the concept of civilisation has become much broader than these single entity descriptions, that are in any case inadequate in describing historical change. Even the geographical or locational terminology that was preferred until recently is now being questioned given the fact that what we regard as a civilisation evolved out of many migrations and much inter-communication across many frontiers. 

Since the concepts used by historians are being redefined by historians in order to make them more precise and to draw in a larger body of sources and consequent explanations, then it follows that some of the generalisations that earlier were regarded as fundamental would also have to be reconsidered. Let me take an example linked to what I have been speaking about which is now looked at somewhat differently.

A century ago the study of religion referred mainly to theology and rituals. Today religion is viewed as part of a larger culture and assessing significance implies considering its relations to society. Are they marginal or do they encompass many activities? How does a religion acquire legitimacy among those that become its followers ? Does it have to do not only with what it teaches but also with who its patrons are and their status? This can be illustrated by observing which deity is worshipped among those of the varna society with a caste status, and which among those of the avarna society with no caste status. It can also note which castes are allowed to enter the place of worship and which are kept segregated. Does anyone annul these practices of worship and if so who does? When a religion becomes wealthy and established, how does it maintain a hold on society? This can be done through rituals, or associating superstitions with rituals and with promises of betterment both in this life and in the vision of the next life. More concretely it is done through setting up centres for education and through congregational worship. A community may crystallise around a sect.

These are among the questions that historians ask today not because they are hostile to religion but because it is now recognised that all religions in all societies also play a social role, and historians have to understand and analyse this role. This also enables the historian to comprehend why particular social groups prefer particular kinds of religious articulation, and why religions differ. It also enables a historian to figure out the process by which a religion comes into existence and has a history of its own.

History when viewed from the Hindutva perspective presents serious objections and these keep growing with further research. Since Hindutva history is rooted in colonial interpretations of the Indian past, it has now to contend with the fact that the colonial interpretation of Indian culture and religion is currently being seen as largely erroneous and therefore gets discarded. Some of the basic tenets of the colonial history of India in which Hindutva history is rooted, are untenable. The Aryan-speaking people were neither indigenous nor of a single ancestry. Their origins lie in the interface of peoples and cultures. From many points of view this makes their history culturally far richer than the projection of a single, narrowly confined origin. 

This brings me back to the question that historians working on the early religions of South Asia are now asking, namely, to what extent was Vedic Brahmanism the sole foundation of Hinduism ? Vedic hymns and rituals are essential to Vedic Brahmanism but the evolving of the Hindu religion required far more than this. It has been said quite justifiably by a few, that the foundations lay less in the Vedic hymns and more in the dialogues encapsulated in the Upanishads and similar texts, some of which seem to have been reflections on the teaching of the Shramanas, as for instance those of the Buddha as set forth in texts such as the Dhammapada

The philosophical dialogues are what we most admire in early Hindu religious teaching and these emerged from a range of sects that cut across ideas and explanations, sects such as those of the Buddhists, Jainas, Ajivikas and even the Charvaka. These in themselves were not foundational to the Hindu religion but the debates that they nurtured had a role in giving form to the religions that emerged. The source therefore was a mélange of potentially rich ideas rather than a single foundation. The historical perspective drawing in the variance of views is inevitably different from the study of a religion in isolation. 

One of the striking features of the period from the Mauryas to the Guptas is that it witnessed many debates and dialogues about religion. These were mainly between Brahmanism and the Shramana religions such as Buddhism and Jainism. The Buddha for instance did not accept the idea of deity, nor the sacrificial ritual as the link to deity, nor therefore the notion of revealed texts. The emphasis was on social ethics and human behaviour. The differences between the two were easily recognised. As Patanjali the grammar puts it, the relations between the two are comparable to those between the snake and the mongoose. What emerged from this activity was not the resurgence of Vedic Brahmanism but rather a different formulation of the Hindu religion, which some have called Puranic Hinduism. 

Vedic Brahmanism was based on an oral tradition for many centuries and large parts of the texts are concerned with the correct performance of ritual and its efficacy. The foci of worship were huge sacrificial performances – the yajnas – at which much wealth was expended. They required a large area of land where a huge altar was constructed at one end. The ritual could last from two weeks to two years and involved a minimum of seventeen categories of priests. The deities worshipped were Mitra and Varuna and later Indra, Agni and Soma. The oft-repeated message was reciprocation – dehi me dadhami te / (you) give me, I give you. Many were involved in providing the necessary wealth to ensure the efficacy of the ritual. 

The teaching of the Shramana religions had to be contested by the Brahmana religion and this resulted in the evolution of a rather different articulation of the Brahmanical religion in the first millennium AD. Prominence was given to the worship of a new set of deities and forms of worship – Shaiva, Vaishnava and Shakta. The permanently built temples became the new places of worship and replaced the temporary yajna / sacrificial sites. Temples housed the deities that now took a new form, that of icons. This recalls the earlier chaityas of the Buddhists and the worship of the Buddha image. One of the earliest temples was a small structure of the Gupta period built in the vicinity of a Buddhist stupa at Sanchi. The Vedas continued to be sacred texts studied largely by brahmanas, but the effective texts with all the required information on these new forms were the Puranas. With the hardening of social boundaries there was a demarcation between those that could enter the sanctuaries to worship and those that could not. The theory of karma became basic to social ethics. This inevitably justified caste distinctions, these latter becoming more central to religion from this time. What finally emerged and gave the contours to Hinduism as we refer to it in modern times was therefore something of a departure from the essentials of Vedic Brahmanism. 

What seems evident is that Hindu belief and worship reached its peak not in the Gupta period but the late first millennium and the early second AD. This was the time of the great mammoth temples such as those of the Cholas and others elsewhere, when institutions and functions linked to the temple approximated royal power. Donations were forthcoming from royalty and the wealthy. Temples became symbols of state power and an enhanced elite. This was also the period when there were intense philosophical debates, together with a maximum of theories in mathematics, astronomy and medicine as well as some of the most impressive departures in architecture, art and Sanskrit literature. It was the threshold of entering into the universality that was taking shape in so many parts of the world.

Then came the third phase, that of Bhakti teachings by various sants that went even further away from Vedic Brahmanism. Bhakti teachings, ranging from those who focused on a single deity of choice to be worshipped in any way one pleases, to those who sang of the centrality of a social ethic. There were new manifestations of existing Puranic deities and mutations of these. These and many other phases and forms were articulated through multiple sects where the new ones were juxtaposed to the existing ones and some of the latter were gradually marginalised. 

Some Bhakti sects encouraged the merging and mingling of belief and worship across the formal religions. Brahmana texts however excluded the Shramanas, the Charvakas and the Turushkas/Muslims as nastikas or non-believers. If the study of religion in India were to start with the study of the sects going back to earliest times, we would have a better historical understanding of religion in Indian society instead of looking at each religion as a single unitary monolith. At the level of popular religion the characteristic of Indian religions has been their ability to absorb and reformulate belief and worship of varying kinds over the centuries.

Attempts are being made to legitimise the currently popular distortions of the history that have official backing, in order to defend the current official political ideology. This fantasised history is being projected in multiple ways: through social media, TV channels and glossy magazines, all locations where none are bothered to separate fact from fake; but also through more systematic ways such as through education. The biggest fear is that the one freedom that education ensures, the freedom to think freely, will be disallowed. Indications of this are apparent. We have to insist that it is not authority that is at a premium but reliable evidence, and the reading of evidence. 

Considerable attention therefore has to be given to the curriculum and texts for schools. If we have to prevent history from being used as political propaganda then we have to insist on the right to critique textbooks and to have the freedom to present alternate explanations where these are required. This cannot be treated as an anti-national act but as an asset in the discussion on what is being taught. The defence of history is an imperative and immanent requirement, if we are to return to being what we once were in the early years of independence – a thoughtful, humane, secular, society. 

This is a slightly edited transcript of Romila Thapar’s speech at the symposium, ‘In Defence of History’, held on August 12, 2021, in honour of Irfan Habib on his 90th birthday.

Romila Thapar (born in 1931) is an Indian historian whose principal area of study is ancient India. She is the author of several books including the most recent, Voices of Dissent: An Essay (2021).