Religion-making in South Asia: An interstitial perspective

Abstract

The study of ‘Hinduism’ in contemporary academia has generated considerable controversy. Many scholars have argued that the idea of a single ancient religion is difficult to substantiate based on the historical record. A common alternative position is that Hinduism is a colonial construct, without well-defined historical antecedents. This paper contributes to a scholarly middle ground, which provides an empirically based yet still contingent analysis of the evolution of ‘Hinduism’, by drawing on evidence from the Sikh tradition. In doing so, it also draws on approaches which interpret Islam as a discursive tradition, subject to contestable representations, shaped by conditions of knowledge and power, as well as by collective aspirations. Sikh attempts at self-definition included distinguishing their tradition from the two larger, pre-existing traditions of Muslims and Hindus in an explicit and self-conscious manner. In doing so, Sikh leaders recognised ‘Hindu’ as a religious category to some degree, well before the colonial period.

Introduction

The study of ‘Hinduism’ in contemporary academia has generated considerable controversy. Many scholars have argued that the idea of a ‘single ancient religion’ is difficult to substantiate based on the historical record. ¹ A somewhat polar alternative position is that ‘Hinduism’ is a colonial construct, without well-defined historical antecedents. ² This paper contributes to a scholarly middle ground—typified by David Lorenzen—which provides an empirically based, yet still contingent, analysis of the evolution of ‘Hinduism’. ³ Also relevant for this analysis are discussions of Islam as a clear tradition, but one subject to contestable representations, shaped by conditions of knowledge and power, as well as by collective aspirations. ⁴ We will employ the term ‘religion’ more expansively than authors such as Asad and King, though consistent with others such as Lorenzen.

This paper draws on evidence from the formation and evolution of the Sikh tradition—the interstitial referent of the paper’s title—which has been relatively less used in these debates. As a smaller grouping that arose comparatively recently in a region of South Asia where two (or more) larger, older traditions had interacted for centuries, the evidence embedded in its history has been relatively neglected. Of course, the term ‘interstitial’ does not imply symmetry in the interfaces or overlaps of Sikh tradition with Muslim and Hindu tradition(s). Nor, in keeping with all scholarly treatments, does it impose any kind of monolithic uniformity or fixity. While the Sikh community has been treated in detail, ⁵ its history has unexplored potential to inform debates about the nature of religion in South Asia before the colonial period. ⁶

In this paper, we embed what we term a Sikh project of religion-making in a historical milieu which includes the direct interplay of Hindu and Islamic traditions in South Asia. An important part of this process is the bhakti movement within which the Sikh tradition is typically placed. We consider the contemporaneous Dadu Panth as another interstitial attempt at differentiation, and draw parallels and contrasts with the Sikh case. A key point in considering the Sikh case is that it involves articulating Hindus and Muslims as somewhat parallel religious ‘others’ in the process of creating its own identity. Detailed consideration of the Sikh case from this perspective provides additional support for a middle ground with respect to the emergence of ‘Hindu’ as a religious identity, as opposed to the more popular ‘constructionist’ argument that ties that development to the effects of colonialism. ⁷ Lorenzen’s response to the constructionist approach is based on what he calls ‘the standard model’: historical narratives that organise components of what might make up ‘Hinduism’ according to some combination of metaphysical, historical and classificatory principles, while avoiding any characterisation as simplistic as a ‘single ancient religion’. ⁸

This paper refers to the process of the evolution of these traditions in South Asia as ‘religion-making’. This is a deliberate choice, differing from Asad and others, but consistent with Lorenzen, who defends the construction against explicit critiques from different sides. ⁹ Critics of the use of the term ‘religion’ argue that its Christian or ‘Abrahamic’ connotations lead to conceptual limitations and even misleading implications for the phenomenon being analysed, although constructionists such as Thapar use the term freely. ¹⁰ The position adopted here, that the term can be fruitfully used in a broad sense, is common in social science-based discussions of cultural evolution. ¹¹

Specifically, we argue, based on evidence embedded in the evolution of a Sikh tradition in the sixteenth century, that an idea of a distinct ‘Hindu’ tradition, including aspects of what we can term ‘religious’ beliefs, as well as societal norms and codes of conduct, is present at the beginning of the seventeenth century in the region of Punjab. The founder and subsequent leaders of the mainstream Sikh tradition saw it as a distinct community, having well-defined doctrines with respect to the divine, ¹² along with guides to ethical or moral conduct, which, while not necessarily coincident with Christian conceptions of ‘religion’, would fit into most notions of what constitutes a religious tradition. Their attempt at this definition included distinguishing it from the traditions of Muslims and of Hindus in an explicit and self-conscious manner. In doing so, they recognised ‘Hindu’ as a religious category to some degree, well before the colonial period. The manner in which this was done, giving the term ‘Hindu’ a substantive as well as contrastive meaning, will be central to our discussion. In doing so, we will still acknowledge the continuing multiple meanings assigned to each term, especially ‘Hindu’. ¹³

The next section of this paper provides a brief overview of recent academic perspectives on the concept of ‘religion’, including claims that the term is unavoidably associated with Christianity or with post-Enlightenment Europe. This is a deep issue, and tackling it fully is well beyond the scope of this paper, but we suggest that, while recognising the linguistic and conceptual pitfalls with using the word ‘religion’, one also must avoid downgrading non-Christian approaches to a range of beliefs and behaviours that can be associated with the term. The categorisation of Islam is particularly relevant for this paper, because of its major presence in South Asia and its contemporaneity with Hindu tradition(s). However, while acknowledging the arguments for not applying the term ‘religion’ to Islam, we argue that its use is still tenable. The section also summarises the different positions with respect to the history and meaning of the category of ‘Hinduism’, including the presence of Islam in South Asia and the rise of the bhakti movement. Some of the difficulties and ambiguities with conceptualisations of ‘religion’ become clearer when illustrated by the case of Hinduism, as differing positions on the validity of that label are connected to judgements about beliefs, practices, institutions and their heterogeneity and stability. A range of scholarly positions can be, and are, taken, based on this complex body of evidence.

The third section of the paper examines the category ‘Hindu’ by approaching it from the interstitial perspective of the Sikh tradition. We provide a brief background of the tradition as a prelude to the main argument of the paper, which is to draw inferences about ‘Hindu’ as a category from the canonical writings of early Sikh leaders. We emphasise the Sikh goal of delineating a distinct tradition, which includes articulating contrasts with the existing, larger traditions of Muslims and Hindus, including sacred texts, rituals and social practices. This evidence provides support for the ‘intermediate’ position of scholars such as Lorenzen.

The penultimate section discusses the scope and limits of the evidence with respect to this intermediate position. We note some of the ambiguities of what constituted the Sikh categorisation of ‘Hindus’, and continued efforts by Sikh leaders to articulate a different identity, both through their writings and through concrete actions to sharpen the boundary of Sikh identity, especially towards the end of the seventeenth century. The delineation of Sikhs and Hindus remained subject to ambiguities and contestations, but some of them were not dissimilar to issues of heterogeneity in Islam. We also incorporate related evidence from two strands of the bhakti movement, the followers of Kabir and, especially, Dadu, both of whom articulated interstitial positions between the communities of Muslims and Hindus. The latter comparison is particularly revelatory about why the Sikh movement achieved a durable distinctiveness, whereas the Dadu Panth, despite having a Muslim founder, eventually became part of a larger Hindu grouping.

The paper concludes with a summary of our arguments, which are consistent with an understanding of ‘Hinduism’ as neither a ‘modern myth’ nor a ‘single ancient religion’. The interstitial position of the Sikh tradition, caught between two much larger, older traditions in a particular geographic space, leads to a quest for a distinct identity, that includes differentiating itself from those ‘others’. This process of ‘religion-making’ is not embedded in colonial-era concepts and institutions. While we follow other scholars in applying the term ‘religion’ broadly, the argument survives even if a different term is used; we are not making any case for fixity, uniformity or narrowness of societal scope.

Concepts of Religion and Histories of Hinduism

It may seem unsatisfactory to discuss religion-making in South Asia without a definitive statement of what ‘religion’ is, but given the lack of scholarly agreement, and the variety of human behaviours and beliefs that can be encompassed within the term, we avoid a fixed definition of the term. A generic definition as ‘a socio-cultural system of beliefs and practices that relates humanity to supernatural elements’ does not come to grips with scholarly claims that the word itself connotes a Western—or more precisely, Christian—perspective. ¹⁴ This set of arguments against applying the term ‘religion’ too broadly also associates its current scholarly application with an intellectual and societal project, rooted in Europe, of defining the religious in opposition to the secular. ¹⁵ In the South Asian context, it is claimed that the term is misleading, because it refers to something historically absent from that context. ¹⁶ This stance is also associated with what seems to have become a popular conceptual division between ‘Abrahamic’ and ‘Dharma’ traditions, with the former including Judaism, Christianity and Islam and the latter including Hinduism, Buddhism, Jainism and Sikhism. ¹⁷ The dichotomy of these groupings is associated not just with geography but also with differences and commonalities in prescribed behaviour (food taboos, ritual observances, marriage prescriptions and so on) and especially in metaphysical or supernatural concepts (the nature of the divine, of creation, of time and of the progression of human lives). Nevertheless, it is clearly recognised as an oversimplification in the case of the first three categories. For example, membership of the Jewish community has traditionally relied much more on ethnicity than the other two traditions in this grouping have. And the commonality of (Western) Christianity and Islam has been explicitly questioned. For example, Asad asserts:

[A]nthropologists who would study, say, Muslim beliefs and practices will need some understanding of how ‘religion’ has come to be formed as concept and practice in the modern West. For while religion is integral to modern Western history, there are dangers in employing such a normalizing concept when translating Islamic traditions. ¹⁸

This passage is quoted in King, who goes on to state somewhat sweepingly that Asad’s claim can be ‘applied to the study of Asian culture in general’, which illustrates the pitfalls of categorisations such as these. At the same time, he uses the category ‘Judeo-Christian’ which has its own problematics, beyond the obvious one of very different concepts of membership in the respective traditions. ¹⁹

Differences among the four categories in the second grouping, the so-called ‘Dharma’ traditions, are less well-acknowledged. Indeed, issues relating to this second classification are closely related to the arguments in this paper. In particular, proponents of Hinduism as a single ancient religion have been among the staunchest advocates for this grouping, and its opposition to the ‘Abrahamic’ group. King’s strictures, although they have their own problematics, will be relevant to the consideration of differing concepts of Hinduism.

Asad traces the evolution of religion as a concept in the context of Western Christianity. Key ideas of his analysis can be found in earlier writers, including the separation of religion from politics, law and science, but what is distinct about Asad’s treatment is his discussion of Islam. He asserts that ‘The attempt to understand Muslim traditions by insisting that in them religion and politics…are coupled must, in my view, lead to failure’, and the ‘separation of religion from power is a modern Western norm, the product of a unique post-Reformation history’. ²⁰ Asad directly takes on various problematic views of Islam, as an empty theoretical object, as an academic label for a heterogeneous set of items, or as ‘a distinctive historical totality which organises various aspects of social life’. ²¹ This discussion of the problems with academic treatments of Islam will be relevant for our subsequent argument and will be invoked at suitable points.

For now, using the term ‘religion’ in a broad sense, we turn to the case of Hinduism and outline contrasting perspectives as to its meaning and its history. ²² Interestingly, King, having quoted Asad and while highlighting the role played by indigenous elites in constructing the modern myth, almost completely ignores Islam in South Asia, focusing instead on Western writers from the nineteenth century to the present. ²³ Romila Thapar frames the construction of Hinduism in terms of the political implications and uses of religious identity, and summarises as follows:

The modern description of Hinduism has been largely that of a brāhman.a-dominated religion which gathered to itself in a somewhat paternalistic pattern a variety of sects drawing on a range of Buddhists, Jainas, Vaiṣṇavas, Śaivas and Śāktas. The texts and the tradition were viewed as inspirational, initially orally preserved, with multiple manifestations of deities, priests but no church, a plurality of doctrines with a seeming absence of controversies and all this somehow integrated into a single religious fabric. Differences with the Semitic religions were recognised and were seen as the absence of a prophet, of a revealed book regarded as sacred, of a monotheistic God, of ecclesiastical organization, of theological debates on orthodoxy and heresy and, even more important, the absence of conversion. ²⁴

Importantly, sources of the early period, in which Hinduism is supposed to have emerged, instead refer to two prevalent religious groups, brahmanas and śramanas. In these sources, the two groups are sometimes jointly considered as relative elites, distinct from the common people, but not only is their difference recognised but also their mutual hostility. Early Brahmanism separated the ‘twice-born’ upper castes from the rest of society and gave prominence to ‘śruti – the Vedas and…smr.ti – the auxiliary texts to the Vedas and particularly the Dharmaśāstras’. In consequence, ‘Dharma lay in conforming to the separate social observances and ritual functions of each caste. The actual nature of belief in deity was left ambiguous and theism was not a requirement.’ ²⁵

Śramanism, a less common term than Brahmanism, includes Buddhism, Jainism and other traditions: Social hierarchies were accepted, but separation in social observances was not accepted, and this group of traditions sometimes attracted lower castes. Buddhism and Jainism had founders, but those founders were not considered incarnations of a deity. They also had their own ecclesiastical organisations. Most importantly, śramanas denied the foundational importance of the śruti and smr.ti accorded to those texts by the brahmanas. The dichotomy between Brahmanism and Śramanism is therefore an important corrective to claims of a singular Hindu tradition, in which Buddhism and Jainism were mere offshoots. ²⁶

The high economic and social status of the brahmanas may have encouraged ‘the modern idea that Brahmanism and Hinduism were synonymous’. ²⁷ At the same time, Brahmanism engaged in a dialectic with groups outside its core, alternately absorbing and seeking purity by exclusion. By the end of the first millennium ce, Brahmanism had largely displaced Śramanism in South Asia, and further, had melded with the increasingly popular puranic traditions of Śaivism and Vaiṣṇavism. Subsequent devotional movements, grouped under the generic term ‘bhakti’, drew on Śaivism and Vaiṣṇavism, but were also ‘to varying degrees the inheritors’ of Śramanism, by virtue of their opposition to Brahmanism. This opposition included attempts to ‘transcend caste and differentiated social observances’. ²⁸

Bhakti attitudes to caste can be connected to the presence of Islam in South Asia. ²⁹ Muslims of Afghan or Turkic origin ruled significant portions of South Asia from the beginning of the thirteenth century ce. ³⁰ It was these foreign arrivals who applied the term ‘Hindu’ to the non-Muslim inhabitants of the conquered lands, essentially as a geographic designation. The court chronicles of the Muslim rulers ‘spoke of Hindus sometimes in the sense of the indigenous population, sometimes as a geographical entity and sometimes as followers of a non-Islamic religion…. Possibly the germ of the idea of a Hindu community begins when people start referring to themselves as Hindus, perhaps initially as a concession to being regarded as “the other”’. This usage in non-Islamic sources dates back to the fifteenth century and includes Kabir, an important figure in the bhakti grouping (and likely of Muslim heritage), who ‘counterposes Hindus and Turuṣkas in a religious sense’. ³¹

The precise nature of Islamic influence on indigenous traditions, including bhakti, has been analysed and debated. The spread of Islam in South Asia was mediated by state power, religious scholars (‘ulama) and sufis, who sought to operate simultaneously at a more metaphysical and experiential level. ³² These were not separate factors or even represented by distinct individuals, but part of a process of negotiation of societal influence and governance. Furthermore, the spread of Islam in India was not a process of formal conversion, but one of gradual adoption of beliefs, customs and practices, sometimes without completely abandoning earlier ones. ³³ This process was strongest in regions such as Punjab and Bengal, on the periphery of the geographic core of Brahmanism. There was a process of interaction, exchange of ideas, and even influence between Muslims and groups with indigenous beliefs and practices—sufis and yogis, for example, as considered by Ernst—but also a recognition and defence of distinct identities, despite internal heterogeneities. ³⁴

The main constructionist argument, however, is that the heterogeneity of religious traditions that existed in South Asia, even during the period of Islamic rule and the spread of Islam, does not allow one to apply with confidence the category ‘Hinduism’ to any subset of those traditions. Instead, there was a ‘multiplicity of cults and sects’ as well as of beliefs. ³⁵ These smaller groupings were the vehicles of social and political life, while an overarching organisational or conceptual system was notably absent. In particular, caste continued to play an important role in defining the relevant groupings for religious life. There were significant differences across groupings in the doctrines of karma and rebirth, which have been adduced as key elements of Hinduism, with conduct (including rituals) not being determinative of religious identity. While acknowledging structural similarities in social conduct, Thapar concludes, ‘But nevertheless it is different from a shared creed, catechism, theology and ecclesiastical organization’, and:

Identities were, in contrast to the modern nation state, segmented identities. The notion of community was not absent but there were multiple communities identified by locality, language, caste, occupation and sect. What appears to have been absent was the notion of a uniform, religious community readily identified as Hindu. ³⁶

In this situation, the construction of Hinduism was an outcome of the colonial encounter, including nationalist responses: colonialism had significant and lasting impacts. However, the logic of this conclusion has been challenged, on the basis that the concept of ‘religion’ that is used is itself anachronistically narrow. ³⁷ Early European descriptions of South Asia used ‘religion’ broadly, along with terms such as ‘sect’ and ‘nation’, to describe the plurality of religious affiliations in the region, and these accounts do not impose a more modern, reified concept of ‘religion’ on the region. As Sweetman puts it, scholars such as King ‘deny that Hinduism is a religion in terms of the same normative paradigm that they detect in the writers they criticise’. ³⁸ Keeping this critique in mind, we turn to what might be considered a middle ground between characterisations of ancient origins and modern construction.

A middle ground accepts heterogeneity and change in the constitution of a ‘religion’, while not abandoning the term. Much of the South Asian evidence concerning differentiating group labels can be examined to reach a different conclusion than the constructionists. For example, Lorenzen asserts that

passages…of this type generally emphasise the differences in religious customs and rituals rather than in philosophy or theology, between Hindus and Muslims, although differences in religious texts and ideas are also noted. Customs and rituals are, after all, the primary practical means that ordinary people use to establish and mark off separate religious identities. ³⁹

The passages referred to make the case for ‘Hindu’ as an identifier of a religious category (in the broad interpretation of that term), despite its decentralised and heterogenous character, and include a wide range of seventeenth- and eighteenth-century European sources. Pre-1800 ‘vernacular Hindu literature’ also has illustrations of a religious self-identity associated with the term ‘Hindu’. Lorenzen gives three examples: Ekanath, a sixteenth-century Maharashtrian brahmana; Anantadas, a late sixteenth-century bhakti follower of Ramanand; and Kabir, who came a little earlier. These examples mostly pair Hindu and Turk, but Lorenzen argues that ‘Turk’ in these contexts clearly represents Muslims. ⁴⁰ He also gives examples from the thirteenth through fifteenth centuries, predating Nanak and (sometimes) Kabir. For example, verses of Vidyapati from the early fifteenth century pair the terms Hindu and Turk, using the term ‘dharma’ in a sense close to ‘religion’ or at least a set of religious customs. ⁴¹

Rather than documenting how the term ‘Hindu’ was used in the medieval period, a different approach to investigating its historical content examines attempts at a more unified set of religious doctrines. In particular, Nicholson argues that self-consciousness of a ‘Hindu’ religious identity must have been spurred by the long and pervasive presence of Islam in South Asia, and traces that impact through the work of Vijñānabhiksu, a philosopher from Bihar of the fifteenth or sixteenth century, as well as other philosophers of this period, going back to the twelfth century. In this analysis, a widespread doxographic effort precedes and supports an intellectual effort to unify Hinduism, by connecting some schools of thought as orthodox, while excluding others (Buddhist, Jain and Islamic) as heterodox, through this classificatory enterprise. ⁴² The basis of establishing this connection is treating the diverse teachings of the Upaniṣads, Purāṇas, epics and other texts as a single whole. This exercise, therefore, is quite different from Lorenzen’s, focusing on philosophy rather than customs and rituals, though orthopraxy is sometimes also involved in this alternative approach.

Despite these multiple approaches to articulating a case for a middle ground, scholarly opinion is not unanimous in accepting ‘Hindu’ as a coherent religious category until the colonial period. What the case of the Sikhs can tell us about this conclusion is investigated in the next two sections.

Sikh Tradition(s): Origin and Implications

The argument we develop is based on selections from the primary Sikh sacred text, the Guru Granth Sahib (GGS), but some brief background is helpful. Sikh tradition begins with Nanak (1469–1539), who was born in Punjab in what is now Pakistan. His family was relatively well-off and literate. His father worked for a local Muslim government official. Nanak married, secured a job and had two sons. At the age of about 30, he had a mystical experience, which he described as a divine revelation, and thereafter travelled extensively for many years, spreading his spiritual message, composing hymns and interacting with representatives of different spiritual traditions. He left traces of his travels and his teachings in many parts of South Asia. For the last two decades of his life, Nanak settled in Kartarpur, on the banks of the river Ravi, where he founded a new community of disciples.

Nanak refers to his disciples as ‘sikhs’, which means students or disciples, and they considered him their ‘guru’, or spiritual preceptor. Such communities were not uncommon in South Asia at the time, and Nanak is often viewed as a part of the bhakti movement. ⁴³ Since he was preceded by several other significant figures in this category, he is often considered to be a successor to some of them (Kabir in particular) in terms of thought, though not in any sense of formal lineage. Some Western scholars have placed Nanak in a so-called ‘Sant’ tradition, which refers to a strand of nirguna bhakti, distinguished from saguna traditions (such as Vaisnavism) which involved worship of divine incarnations. But there is no evidence that Nanak saw himself that way. ⁴⁴

Several factors distinguish Nanak and the community he founded from similar phenomena of the period. He collected his verses in written form, in a bound volume and passed this on to an appointed successor, in a formal ceremony. ⁴⁵ His successor, Angad, also wrote verses, signing them as Nanak, with authorship differentiated by the designation ‘mahalla dooja’ (palace the second) with the implication that the guiding spirit of the successor guru was still that of Nanak, though its bodily residence had changed. ⁴⁶ From slightly later accounts (GGS verses), it is known that Nanak’s first successor ran a langar (community kitchen), with sharing of consecrated food, but it is likely that this institution began in Nanak’s Kartarpur community.

The next successor to the position of Sikh guru, Amar Das, considerably expanded the corpus of verses meant to provide devotional material for the community, including verses of various bhaktas (or ‘bhagats’ in the regional rendering), as well as a number of his own compositions. ⁴⁷ He created a formal hierarchical organisational structure, which provided local spiritual guides for geographically dispersed Sikh congregations. This institutional step in particular distinguishes the Sikh community from many others of the time. Therefore, by the 1570s, the community already had a central text, a recognised founder and line of succession, and an organisational hierarchy, thus meeting many of the criteria for a single religious community, as described in the previous section. ⁴⁸

Nanak’s next two successors, Ram Das and Arjan, further consolidated these structures. Ram Das developed a new town—Ramdaspur, now Amritsar—which became a sacred centre for the community. Akbar, then the Mughal emperor, had a policy of tolerance and engagement with non-Muslims, which benefited the growing Sikh community. ⁴⁹ Ramdaspur was governed by the Sikh gurus relatively autonomously of the empire’s local representatives nearby Lahore. ⁵⁰ The community became more prosperous and more diverse in its social composition. Arjan also constructed a central place of worship (Harmandir Sahib) in Ramdaspur.

Most importantly, Arjan compiled and personally approved a definitive text, mainly of the verses of the Sikh gurus, but also including some bhagats such as Kabir, bards who were part of the gurus’ court and one sufi Muslim (Shaikh Farid). ⁵¹ The volume begins with a liturgical section, containing verses that are specifically meant to be recited at different times of the day. ⁵² This volume subsequently became (with a few additions in the late seventeenth century) the Guru Granth Sahib, the continuing spiritual guide of the Sikh community. That formal transition occurred in the early eighteenth century, but the compilation in 1604 was immediately followed by the ceremonial installation of the volume in the Harmandir Sahib (deepening the sacred meaning and authority of that site) and the extensive scribing of authoritative copies of the Granth for use by dispersed Sikh congregations. These developments solidified the distinctiveness of the community and set it apart from other religious groupings of the time. ⁵³

While the community at the beginning of the seventeenth century may not have satisfied all the requirements of a modern ‘religion’, arguably, neither did any other tradition in this period. If this is accepted as a pre-colonial case of religion-making, then we have an expanded and more nuanced understanding of the process of religion-making in South Asia. To complete the argument, we need to consider the relationship of this new formation to Islam, as well as to the non-Muslim groups that may or may not have constituted ‘Hinduism’.

We focus in this section on the latter, extending some of the evidence considered by constructionists and those occupying a middle ground, as outlined in the previous section. We present statements relating to religious formations that were contemporary with, or temporally preceded, the 1604 volume that became the GGS. ⁵⁴ These are the views of Nanak and his successors, as well as those of the bhagats and bards that were included in the volume, with respect to the religious milieu in which they lived. They pertain to caste, ritual functionaries and sacred texts. We discuss the role of the bhagats in the Sikh corpus and the use of the term ‘Hindu’. Importantly, we make the case that the term ‘Hindu’ was not just a social/kinship designator that named in/out groups but represented both shared beliefs and identifiably ‘religious’ practices. Therefore, this discussion buttresses the evidence and conclusions of Lorenzen, from a relatively less-studied source.

Caste

The varna system is clearly recognised by the Sikh gurus, which suggests that it was an active model of social organisation, but one that the gurus deemed irrelevant in the context of connection with the divine. For example, Nanak, after saying that the gurmukh (one facing the Guru) can hold the divine light within, adds that ‘kshatriyas, Brahmanabrahmanas, sudras, and vaishyas cannot find its value with myriad calculations’. ⁵⁵ Ram Das and Arjan echo this view, in verses that say all four of the varnas can seek and achieve salvation if they follow the path of the Guru. ⁵⁶ Going beyond this, each of them extends this possibility to untouchables, who are outside the varna system, and remained outside most conceptions of who was a Hindu until the twentieth century. For example, Ram Das writes, ‘those who seek the sanctuary of the holy are saved, kshatriya, brahmana, sudra, vaishya, or the most untouchable untouchables’. ⁵⁷

The varna system was not a theoretical construct, but a reality of the society that the Sikh gurus were engaging with, but they rejected varna as a basis for spiritual uplift. Kshatriyas and brahmanas—particularly the latter—were criticised for their conduct, and for falling short of the qualities that supposedly accompanied their status. Nanak was particularly trenchant in condemning the hypocrisy of brahmanas: ‘the brahmana takes life, then has cleansing baths’. ⁵⁸ He commented sardonically on the way that brahmanas took advantage of people’s faith: ‘Rice balls are offered to gods and dead ancestors, but it is brahmanas who eat them.’ ⁵⁹ The Sikh gurus therefore recognised caste as a societal phenomenon in the sphere of religion, broadly conceived.

Ritual Functionaries

It is reasonable to relate this religion to the Brahmanism which Thapar views as a major and persistent phenomenon in South Asian history. This system has ritual roles, and the Sikh gurus also questioned the value and sincerity of the key figure of the pandit as someone who possessed the learning to perform rituals and ceremonies. The GGS refers to rituals performed by brahmanas, which the pandits invariably were. Direct references to pandits are more numerous. Many of these references criticise the pandits’ empty scholarship, and Amar Das, in particular, made numerous statements along these lines: ‘Outwardly, they call themselves pandits, but their minds are foolish and ignorant.’ ⁶⁰ In many cases, the learning of pandits is explicitly mentioned as rooted in the Vedas and Puranas: ‘Pandit, your filth shall not be lifted, even by reading the Vedas for four ages’ ⁶¹ ; as well as ‘The pandits, teachers and astrologers endlessly read the Puranas, but do not know the divine is hidden deep within themselves.’ ⁶²

Sacred Texts

The Vedas, Puranas and other sacred texts are referenced even when not tied to pandits or brahmanas. Nanak refers to the Vedas, Puranas, Shastras and Smritis in his composition Jap [ji], which begins the GGS, but later on, is clear on their limits, ‘The Vedas speak and expound on the divine, but do not know its expanse.’ ⁶³ These texts are subordinate to the Naam, the manifestation of the divine. ⁶⁴ Thus, ‘The Naam is the support of the Smritis, Vedas and Puranas.’ ⁶⁵ Other references are more favourable towards these texts, but still make clear that they are not a path to salvation, nor are brahmanas and pandits as intermediaries and interpreters likely to be helpful in that direction. The path being offered by the Sikh gurus is distinct from the alternative, not specifically named in these verses, but plausibly labelled ‘Brahmanism’. Yet, when this alternative path is named in other verses, its followers are referred to as ‘Hindu’.

Bhagats

All the preceding quotes are from the Sikh gurus. The verses of the bhagats in the GGS reinforce the Sikh rejection of Brahmanism. ⁶⁶ Kabir is the most prominent of these bhagats; he had a significant following in his time, but did not appoint successors, or systematise any collection of his verses. The GGS collection of his verses is a unique part of his corpus and was likely chosen to support the message of the Sikh gurus. ⁶⁷ For example, he states, ‘I have searched all the Vedas, Puranas and Smritis, but none of these can save anyone’, ⁶⁸ and, ‘Listening to all the teachings of the Vedas, Puranas, and Smritis, I wanted to perform religious rituals. But seeing all the wise people caught by Death, I arose and left the pandits, and am free of that desire.’ ⁶⁹ Among other bhagats included in the GGS, Ravidas, from an untouchable background, expressed similar sentiments against status tied to birth: ‘I am a great poet of noble heritage, I am a pandit, I am a yogi and sanyaasi, I am a spiritual teacher, a warrior and a giver—such thinking never ends. Says Ravidas, no one understands; they all run around, deluded like madmen.’ ⁷⁰

Hindus

We can now return to the usage of the term ‘Hindu’ in the middle of the second millennium CE. As Thapar notes, Kabir counterposes Turks and Hindus, as in, ‘The Turk knows the Turkic way of life; the Hindu knows the Vedas and Puranas.’ ⁷¹ The usage here, while referring to one group (Turks) by geographic origins or ethnicity, implies some religious identity as well, since the ‘Hindu’ (being compared to the Turk) is connected to the sacred texts of Brahmanism. There are similar implications in the statement ‘Worshipping their idols, the Hindus die; the Turks die bowing their heads.’ ⁷² But Kabir also compares Hindus with Muslims rather than Turks: ‘The Hindu utters the name of Raam, the Muslim has one Khudaa.’ ⁷³ Namdev, another bhagat, makes a similar comparison: ‘The Hindu worships at the temple, the Muslim at the mosque.’ ⁷⁴

Nanak, too, made the same pairing, stating that ‘The Creator is everywhere. There are as many travellers as there are living beings. When one’s death warrant comes, there is no delay. One who knows the divine here, does so there as well. Others, Hindu or Muslim, are just babbling.’ ⁷⁵ Arjan used the Hindu–Turk pairing in the following lines, which are part of a composition that emphasises that only union with the divine transcends impermanence. ⁷⁶ ‘Castes, races, Turks and Hindus; beasts, birds and countless beings and creatures; the entire world and the visible universe—all forms of existence shall pass away.’ He used the same Hindu–Turk pairing in another composition, but later in that piece, stated: ‘I am not a Hindu nor a Muslim; to Allah/Raam my life-breath and body belong.’ ⁷⁷

To summarise, the Sikh sacred text represents a coherent collection of ideas and commentary with respect to the religious milieu of South Asia at the beginning of the seventeenth century. Centuries of Muslim presence in the region and the influence of the bhakti movement (itself influenced by the Islamic presence) are represented in this text. There is a clear message in the text that rejects features of social organisation and ‘religious’ rituals tied to Brahmanism, which was a long-present social phenomenon in South Asia. Furthermore, there are verses in the text supporting the view that by this period ‘Hindu’ was at least partially being used as a term commensurate with ‘Muslim’. This last point uses evidence from the GGS to extend typical considerations of the ‘Hindu–Turk’ pairing. These examples are therefore consistent with a pre-colonial concept of ‘Hindu’ as a religious category, contrasted with that of ‘Muslim’.

The distinctive feature of the evidence from Sikh writings is that this evidence does not come from outsiders, such as European observers, or from insiders seeking to create a coherent community of ‘Hindu’ belief and practice. Instead, the evidence comes from a project of self-differentiation, what we can think of as a praxis of ‘religion-making’ indigenous to South Asia. This project has many complexities and contestations, but that is arguably true of any religion—as long as one is willing to accept the broad use of that label. Given its interstitial position and temporal location, the definition of Sikh identity includes (but is not limited to) articulating its differences from ‘Muslim’ and ‘Hindu’. To further understand the scope and limits of this argument, in the next section, we consider the further evolution of the Sikh tradition and compare it with other ‘religion-making’ projects of the same period.

Scope and Limits

We have argued that the Sikh tradition followed conceptualisations that emerged from Islamic writers and from the bhagats, in beginning a transition in the use of the term ‘Hindu’ to mean a religious formation and not just a geographic designation. Moving away from the latter was necessitated by the fact that followers of Islam in South Asia were no longer completely embedded in ‘Hindu’ civilisation. The bhagats, somewhat sporadically, and the Sikhs, more decisively, moved in this direction of differentiation by beginning a shift in the application of the term ‘Hindu’.

One must acknowledge the strong overlap of ‘Hinduism’ in this sense with Brahmanism, as illustrated by another example from Sikh writings. In one verse, Nanak identifies a third grouping in addition to Hindus and Muslims—that of various ascetic sects, the yogis. ⁷⁸ While Van der Veer treats such orders as within the category ‘Hindu’, ⁷⁹ and in Thapar’s schema they would likely be closer to Brahmanism than Śramanism, here they are a third category, neither Hindu nor Muslim:

The Hindu comes to the house of a Hindu. He puts the sacred thread around his neck and reads the scriptures. He wears the thread, but does evil deeds. His bathing and cleansing will not be accepted. The Muslim glorifies his own faith. Without a guru or a spiritual guide, no one is accepted. They may be shown the way, but few go there. Without the right actions, heaven is not attained. One goes to the yogi’s abode to be told their skills. They wear ear-rings to show the way. ⁸⁰

The message of this verse is the importance of good actions rather than superficial claims to identity, but what is relevant here is that three religious formations are being named and compared. This indicates a somewhat more complicated picture than the dichotomous examples of the previous section but does not invalidate the general argument, insofar as ‘Hindu’ is used to describe a follower of an identifiable set of beliefs and practices.

By the beginning of the seventeenth century, the Sikh community—self-identified as distinct, with a founder, a line of successors, a sacred text and a sacred centre—had followers from diverse social backgrounds and dispersed locations. This situation was destabilised by the events that ended the life of Guru Arjan. The growth of the community sparked a response from Jahangir, who had succeeded Akbar as emperor. His memoirs refer to Arjan by name and find exception to the fact that he had been attracting Muslim followers. ⁸¹ The emperor noted that the community had developed over several generations but was propagating a false message, and ended by describing the intention to confiscate Arjan’s property and possessions and have him executed. ⁸² After Guru Arjan’s execution, the physical circumstances of the Sikh community changed dramatically. They were economically more straitened, lacked a secure and permanent central site, rival claims to the Guruship sharpened, ⁸³ and conflicts with the imperial authorities continued throughout the seventeenth century.

In these circumstances, Gurdas Bhalla, a relative of Guru Arjan and the scribe of the base manuscript for the GGS, wrote a large, separate collection of verses, emphasising the community’s distinctiveness from those who follow the Vedas or Quran. ⁸⁴ Gurdas explicitly compared Hindus and Muslims as religious communities by employing parallel characteristics of pilgrimage (Varanasi versus Mecca), bodily markers (forehead marks versus circumcision) and sacred books (the Vedas versus the Quran and Judaic sacred texts, which were collectively called kateb). ⁸⁵ He then proceeded to use the descriptions of these two larger ‘others’ to contrast with his own community, that of ‘Gurmukhs’ or ‘Gursikhs’. The Gurmukhs are socially diverse but equal within the congregation, they serve the community, recite, sing kirtan, meditate and reject other affiliations. This last characteristic, even if normative rather than observed in practice, is particularly strong evidence against blanket arguments that there was no ‘Indic’ concept of exclusivity. Indeed, Gurdas mentions Jews and Christians, along with more detailed commentary on Hindus and Muslims, suggests that the way of the Gurmukhs is preferable, and invites these others to join the Gurmukhs. ⁸⁶

The seventeenth-century struggles culminated in a sharpening of the boundary of the Sikh community at the century’s end, driven by heightened conflict with the Mughal empire, and the public execution of Nanak’s eighth successor, Guru Tegh Bahadur, in Delhi, the imperial capital. ⁸⁷ The executed guru’s son, Guru Gobind, introduced a new initiation ceremony, with new rules of conduct for the initiated, called the Khalsa. The likely goals included ending the intermediation of local representatives of the Guru’s authority for dispersed Sikh congregations, reaffirming the close bond between the Guru and his disciples, emphasising the egalitarian aspects of that bond and creating a physical identity that also emphasised equality and self-affirmation of adherence to this spiritual community. At this time, he also enjoined his followers not to associate with followers of rival claimants to the succession to Nanak. ⁸⁸

Arguably, the explicit attempts to sharpen a distinct Sikh identity, including Gurdas’s writings, and Guru Gobind Singh’s creation of the Khalsa—both of which emphasised a degree of exclusivity—were crucial to the success of the religion-making project that we have articulated. The insistence on a definitive or canonical sacred text was also crucial to this definition, although the text could be, and was, drawn on by multiple groups who did not follow the line of gurus through to Guru Gobind. In the court of Guru Gobind, his followers were instructed to distinguish their conduct from that of Muslims and Hindus. ⁸⁹

By contrast, there was no central Kabir text that would correspond to the Sikh GGS, and ‘There is no evidence to suggest that Kabir himself founded any institution.’ ⁹⁰ By the late sixteenth century, Nabhadas, a follower of the Vaishnav bhakta Ramanand, was already claiming that Kabir was also a follower of Ramanand. ⁹¹ Lorenzen captures this trajectory of the followers of Kabir in his title ‘The Kabir Panth: Heretics to Hindus’.

The case of the Dadu Panth is a more informative comparison. Dadu Dayal was an almost exact contemporary of Guru Arjan, but, like Kabir, a Muslim by birth, though he was arguably more familiar with Islamic practices. ⁹² Many of his teachings were similar to those of Kabir, and he, too, stated that ‘No Hindu am I, nor yet a Musalman’ ⁹³ and ‘The ignorant Hindu and Muslim are both in error.’ ⁹⁴ As in the case of Kabir and of the Sikh gurus, Dadu made parallel critiques of practices that are identifiable as those of Muslims and Hindus. ⁹⁵ While the followers of Dadu compiled a sacred text and a central location was established, the solidity of the organisation was perhaps not comparable to that inherited and further developed by Guru Arjan.

However, the Dadu Panth was far from inconsequential as an institutional presence, in a region neighbouring where the Sikh community arose. They possessed a militarised wing as well, attracted Muslim followers, and benefited from the patronage of local rulers. ⁹⁶ Arguably, this patronage, along with the Vedantic turn chosen by a disciple, Sundardas, and interactions with the Vaiṣṇava Niranjani Sampradaya, all helped to push the Dadu Panth in a direction that was similar to that of the Kabir Panth, and very different from Dadu’s Muslim heritage. ⁹⁷ These subsequent developments may also have been encouraged by earlier differences in the Dadu Panth’s self-definition. In particular, Dadu very much saw himself as following Kabir’s legacy, whereas the structure of the GGS and specific verses within it imply the primacy of Nanak and his successors over the Bhagats. ⁹⁸

An additional external factor that may have shaped the differences in the trajectory of the Sikhs and the Dadu Panth could have been the different social composition of the areas in which they arose. Eaton has emphasised that agriculturist Jats in western Punjab slowly adopted Islam from the thirteenth century onwards, partly because they were peripheral to the Hindu social system that was most firmly established in the Gangetic plain. ⁹⁹ The Sikh movement, coming slightly later, but operating in a similar social setting, would have had similar attractions to local Jat clans, whereas the Dadu Panth in Rajasthan did not have as advantageous a geographic position. This bottom-up driver of growth in the Sikh community would be complementary to the top-down project of religion-making, which was more robust, with a stronger institutional identity, than that of Dadu and his followers. As a result, the execution of Guru Arjan and the continual contests for primary authority in the seventeenth century could be overcome, though not without somewhat radical change (the institution of the Khalsa) at the end of that period.

This radical change was contested, especially by higher-caste Sikhs. The Khalsa rejection of caste boundaries and proscriptions, including eating restrictions, was closer to Islamic practices such as those of the various sufi orders of the time, though the ideology of the Khalsa also employed puranic mythology. ¹⁰⁰ Some Sikhs were reluctant to embrace the new requirements, even to the point of hostility, because the new practices sought to minimise or even erase the caste identities that still dominated marriage and other social ties. ¹⁰¹ While proponents of the Khalsa order in the late eighteenth and early nineteenth centuries emphasised its distinction, and even its supremacy over Hindus and Muslims as antagonists, this was not a unanimous viewpoint, and compromises were continually made. ¹⁰² These complications are among many that exist for any simple linear narrative of the evolution of the Sikh community, but are consistent with contingent framings of Islam and Hinduism, as evinced in the scholarship of Talal Asad and David Lorenzen, respectively.

The Khalsa gained political ascendancy in the eighteenth century but made compromises to achieve and maintain this position. The practices of Maharaja Ranjit Singh, who nominally ruled Punjab in the name of the Khalsa, often involved monarchical conventions that were at odds with the teachings in the GGS. ¹⁰³ One response was that some Sikhs split off in new sects at the beginning of the nineteenth century, claiming to reconnect with those teachings. ¹⁰⁴ At this time, the British colonists controlled much of South Asia, but not Punjab, and Sikh ‘reform’ movements or schisms (depending on one’s point of view) were driven by the community’s own history and circumstances, rather than by colonial influence. Sikh responses to colonial and Christian influence only began in the last quarter of that century, at about the same time as another ‘Hindu’ reform movement, the Arya Samaj, emerged. ¹⁰⁵ These responses were significantly shaped by the actions of the Arya Samaj but were grounded in the Sikh history of religion-making.

The colonial period modified the idea of ‘religion’ in South Asia, especially because of its implication in colonial political and legal structures. ¹⁰⁶ But this paper suggests that political and legal structures that incorporated religious identity were present in pre-colonial India, even if not formalised in the same way. ¹⁰⁷ In the colonial context, both Sikh and Hindu identities and their boundaries were vigorously contested, but neither was wholly new. The Sikh process of identity formation began with its founder, Nanak, and proceeded with a project of institutionalisation and growth, up to the point of conflict with a dominant imperial authority. The initial process of self-definition and differentiation included articulating the identity of Hindus as a religious community, distinct from Sikhs in matters such as the authority of the Vedas and of brahmanas, and in some customs and rituals. The imperial threat was met with further attempts at self-definition—the Khalsa—and the collapse of the Mughal empire created space for this new Sikh identity.

This does not imply that all three traditions—Islamic, Sikh and Hindu—were on the same footing in the sixteenth and seventeenth centuries. Islam had existed for almost a millennium before Guru Nanak’s birth, and there were markers of belief and practice that set it apart relatively clearly, although disputes about heterodoxy and boundaries were present in South Asian Islam at that time. ¹⁰⁸ The Sikh tradition had some of these features but permitted a fuzzier boundary until the colonial period. ¹⁰⁹ And what is now termed Hinduism was arguably the least well defined of these three religious groupings, and its modern meaning is more attributable to a response to colonialism, as compared to the Islamic and Sikh traditions.

Conclusion

The origins and meaning of the categories ‘Hindu’ and ‘Hinduism’ continue to be subjects of scholarly debates. Two polar positions are that as ‘religious’ categories, they are a colonial-era construct, and on the other hand, that they can be applied to an ancient and coherent system of beliefs and practices and its followers. A range of intermediate characterisations are also possible, which incorporate the complexities, ambiguities and temporal variation of the categories, while simultaneously making the case that they are appropriate labels for societal formations that can be termed ‘religious’ well before the colonial era in South Asia. In both the constructionist and intermediate positions, the term ‘religion’ does not have to be restricted to being applicable only to the colonial era and its aftermath. ¹¹⁰

In this paper, we have provided additional evidence for the intermediate position by systematically analysing how the category ‘Hindu’ was used in canonical Sikh writings of the sixteenth and seventeenth centuries. Our analysis builds on previous work which adduces the long presence of Islam in the region, the nature of reactions to that presence in shaping a self-consciousness of a distinctive identity, and the pre-colonial observations of European and Muslim outsiders, as well as of representatives of indigenous traditions. The Sikh case is similar in its characterisation of ‘Hindus’ as a religious other but is based on intentional institution-building and attempts to achieve coherence and visible boundary markers. The process is imperfect but recognisable, and it succeeds where other attempts at creating a third path in distinction to Muslims and Hindus—particularly the followers of Dadu—have not.

The issue of the evolution of what we call Hinduism (and, similarly, of Sikhism) is entangled with the meaning and applicability of the term ‘religion’. We have used the term ‘religion’ in line with Lorenzen and others such as Thapar and Murphy, but the main conclusions of the analysis would apply even if a different term, such as ‘way of life’, is used. There is a set of social practices that are broadly identifiable as distinct from other aspects of culture or ‘civilisation’, even if they are heterogeneous over time and across societies. ¹¹¹ Many of the features of Lorenzen’s analysis, as well as of this paper, are consistent with Asad’s characterisation of Islam, even though he prefers not to call it a religion. The case of the Sikh tradition, because of the geographic and historical circumstances—what we have termed an interstitial position—can offer some insights into the broader debates, including commonly used binaries such as West–East and Abrahamic–Dharmic. These issues are beyond the scope of the current discussion, except to note that both binaries are subject to problems, highlighted by the history of pre-colonial South Asia, but the features and histories of other traditions as well. ¹¹² All of these issues of nomenclature and categorisation deserve a more detailed comparative analysis.