Intersectionality and belonging: Muslims in the census of British Asia

Abstract

The British territories of greater Southeast Asia were administratively connected to London and Calcutta, and while local censuses show that these centres could exert some influence at the furthest peripheries of the Empire, a close analysis of the ways in which race and religion were approached in the classification of colonial subjects in Southeast Asia shows peculiarities specific to the region.

In this article I argue that the demographic and socio-political contexts of British Burma and Malaya (with references to Hong Kong) led to a framing of ‘race’ that challenged European ‘scientific’ definitions and embraced instead the interweaving of multiple aspects of an individual's identity, most prominently religion. This shift, potentially empowering as reflective of local understandings of belonging, and an improvement from the period's anthropometric framework, was to backfire, however. With the emergence of nationalism, majoritarian identities came to be homogenised in these ethno-religious intersectional communities, marginalising and excluding those who did not fit.

[W]hen a census officer comes around and asks the Burma Muslim his name, if he says Abdulla he will be classed under the category of Indians, but if he gives his name as Maung Gye he is put under Zerbadies [meaning a Muslim born in Burma of Burmese parents].¹

A Chinese convert [in Malaya] invariably assumes a Malay name and may easily be mistaken for a Malay, even by officers of long experience. Religion and race are to a Malay overlapping terms, and a ‘Maalap’ [a Chinese convert to Islam] would be apt to describe himself in the schedule as ‘Malay’ by race.²

[T]he girl of Chinese blood adopted in infancy by a Tamil [residing in Malaya] would be entered as a Tamil unless she insisted upon being described as Chinese. These were merely the natural consequences of the insistence upon community; for the Chinese girl in the illustration would almost certainly speak Tamil and no[t] Chinese, would live on a typically South Indian diet and be bound by Tamil custom and would, in due course, marry a Tamil according to the Hindu rites. To argue that she was a Chinese would be wholly to misconceive the meaning of ‘race’ in the special sense [of community] in which it is used in this context.³

British Burma and Malaya,⁴ like Hong Kong, emerged as thriving colonial possessions because of their strategic locations on trade routes; more fundamentally, it was thanks to individuals’ mobility as sailors and traders, military personnel and bureaucrats, miners and planters who moved across Asia. Some moved voluntarily, others were encouraged, and more were forced to. The result was an extremely diverse population, that at times continued to move, and at times settled in their new homes.

The enumeration of people was as important to the British Empire as the cataloguing of land and natural resources. For decades, colonial administrators collected information on their subjects, asking for details about their origins, religious beliefs, and ‘nationality’, the latter defined as a compound of political affiliation to a nation-state, cultural affinity, or descent. Often their primary concern was to understand fertility rates and population flows, thus leading to the categorisation of individuals (and entire groups) as either natives/indigenous or immigrant/foreign.

Echoing James Scott, in the censuses of British Asia “the state's shorthand formulas through which tax officials must apprehend reality are not mere tools of observation […] they frequently have the power to transform the facts they take note of” with the purpose of making “society legible” and the state governable.⁵ As concluded by A. J. Christopher, the Imperial census “represented an attempt by the Colonial Office to obtain a view of the Empire as a whole, as an aid to its efficient administration, although the precise use of the census was never explicitly stated”.⁶ Whatever the original intentions, it is clear that census practices—and attempts at broad-enough categories to make the data comparable—had long-lasting effects across imperial subjects. Bernard Cohn has argued that the census was key to “making objective to the Indians themselves their culture and society”.⁷ Arjun Appadurai has pointed to the importance of colonial enumeration exercises to “creat[ing] the sense of a controllable indigenous reality”.⁸ And Benedict Anderson has suggested that census classifications contributed to the shaping of ‘imagined communities’ in the late- and post-colonial processes of nation-building.⁹

There were Imperial priorities and policies, but practices and implementations often varied. While London demanded standardisation, Calcutta continued to show much diversity in data collection, and when the General Register Office focused on race, the Colonial office in Calcutta asserted caste as the primary lens of tabulation. British possessions in Southeast Asia were connected to both centres. But their census reports show yet a third variation in approach, as the interweaving of colonial practices and local enumerators’ interpretations led to the shaping of primarily ethno-religious intersectional identities.

This article explores three interconnected lines of inquiry: first, the intellectual and practical trajectories of the European category of race as it pertained to census enumeration and tabulation in the colonies, with a focus on London's General Register Office, British India, Burma, and Malaya.¹⁰ Second, an exploration of the categories of indigeneity and foreignness. These categories, inherited from earlier concerns over population growth, became politically significant in the twentieth century as colonised subjects began to define their proto-national identities along lines of belonging. Third, I turn to the issue of intersectionality: whereas in Euro-America race was being increasingly defined according to physiognomic and anthropometric paradigms, in 1910s-1930s Southeast Asia, racial census groups were constructed through a combination of criteria, amongst which religion and ethnicity (i.e. geographical origins, language, and customs) were key.

This interweaving of multiple aspects of an individual's identity contributed to a more complex configuration of ‘race’, but it also came to impose a constructed homogeneity over an originally diverse reality. This latter development led, on the one hand, to the disappearance of the ‘statistically insignificant’, and on the other, to the hardening of boundaries between indigenous and foreign groups.

As nationalism emerged, these newly created intersectional identities were eventually deployed to define who the native majority was, from Malaya's Muslim Malays to Burma's Buddhist Burmans. It is because of the double edge of ethno-religious identities as both potentially empowering (as a push-back against ‘scientific’ racism) and a source of exclusion (in the process of national-identity formation), that I find the concept of intersectionality coined by Kimberlé Crenshaw extremely apt.¹¹

‘Race’ from London to Asia

As post-Enlightenment governments had already done across Europe¹² in the nineteenth century, the British Empire embarked on the census classification project as a necessary component of colonial rule. Statistics became key to maximise the Empire's possessions’ productivity; and data about people—from land-ownership and occupation to racial and religious affiliations—were needed to govern colonial subjects.¹³ Even though the British claimed that their classification efforts were a descriptive exercise, through the decades the range of categories used shrank, forcing reality within new boundaries.

For the analysis of the colonial census in Southeast Asia, it is worth not only focusing on London, but also taking the conversation to Calcutta, as British possessions in Southeast Asia were often seen—and treated—as appendages of India. After all, Burma was a province of British India until 1937, and the Straits Settlements (comprising Singapore, Malacca, and Penang) had been administered through Calcutta until 1867, when they became a Crown Colony directly managed by the Colonial Office.¹⁴ Ties were also reinforced by colonial labour policies, which encouraged (sometimes forced) migration across the Subcontinent, China, and Southeast Asia. Population data from Burma was included in a dedicated volume of the census of British India, following its structure and emphasis on caste. At the same time, the censuses of the Straits Settlements, the Malay states, and Hong Kong regularly separated the tables for their Indian population, influencing—although rarely mandating—their respective general approach. Calcutta stood out as an alternate point of reference for Southeast Asia, and one that was not necessarily fully modelled along London's directives. As pointed out by Arjun Appadurai, London was struggling to impose a degree of “standardization against the on-the-ground variation” in India between the 1840s and the 1870s.¹⁵

The British territories of Southeast Asia come to stand as yet a third model. The census models of the territories connected to both centres, and were further modified through local adaptations and interpretations. In the very early enumerations of the Straits Settlements the main concern of census-takers was geographical origin and racial affiliation of the territory's residents. Racial affiliation was sometimes defined as “nationality” and religion occasionally emerged as a characterisation of race too. Such examples were the “Native Christians”, Parsees, and Jews of Penang (1822),¹⁶ Singapore (1827, 1829, 1830),¹⁷ and Malacca (1833).¹⁸ These enumerations show no attempt at homogenisation, as the hand-written lists include a vast, and often changing, set of “peoples”. Notably this approach was to disappear in the second half of the nineteenth century.

In India as across Southeast Asia, local superintendents and enumerators had a fair degree of agency in the compilation of the census, but it is evident that London's efforts to compile a comparable Imperial census had a strong impact on the colonies. The desire for comparability had been expressed to the colonies as early as the 1840s, when recommendations were issued to the colonies in preparation for the 1851 census, stating that “in completing the questionnaire, enumeration by age and race was necessary”.¹⁹ Requests for standardisation become stronger on the eve of the 1871 enumerations, meeting with resistance as close to the metropole as Ireland: “nothing could be less worthy of statistical science than condescension to the pedantries of a forced uniformity”.²⁰ With the first synchronous census of the British Empire, undertaken in 1891, colonies across Asia begun to integrate the terminology of race.

The 1857 Census of Penang presented its population as belonging to the “Europeans and descendants”, “East Indians”, “Malays”, “Chinese”, and “India Proper” groups, but no descriptions or definitions were provided. On the occasion of the 1871 census, as it was reported in the Census of England and Wales, the population was divided between “Europeans and Americans”, “Eurasians”, and “Native Population &c.”, which included Malays, Chinese, Indian military, Klings (“or immigrants from Southern India”), local prisoners, and prisoners from British India.²¹ In Burma, officers pushed back against the India practice of collecting data for caste (the primary site of India's census tabulation until 1931²²), as it was “determined to be useless” in the province.²³ In 1872, at the first all-India census, the represented ‘races’ of Burma were listed to include “Europeans and Americans” as well as Chinese and Afghans, Hindus and Muslims, Burmese, Arakanese, Shans, and Hill Tribes.²⁴ Ethnicity, then, was mixed with nationality and religion. In 1881 “the mother-tongues and birth-places of the people were taken as the best tests of nationality or race”.²⁵

By 1891 it was agreed that although “caste were not practicable”, “race at least should be returned” as neither birth-place nor religion could be taken as proxy indicators of race. Race and nationality were thus introduced “to clear away the ambiguity caused by an immigrant population and consequent confusion of tongues”.²⁶ But all sorts of labels were taken to identify racial groups, from Karen to Chinese and Hindu; by the turn of the twentieth century, Burma's “peoples” were grouped under labels that referenced multiple facets of an individual identity: mother-tongue, place of birth, religion, and more.

The rise of Joseph Chamberlain as Secretary of State for the Colonies in 1895 meant a further reinforcement of this approach Empire-wide. The General Register Office thus undertook the task of compiling a Report on the Census of the British Empire for 1901, which included comparative tables and commentaries. This had been a post-facto attempt at comparison, which inevitably encountered major difficulties. It was likely to have been the outcome of Chamberlain's doing that pushed for a more detailed classification of the residents in the Straits Settlements. Although called “nationalities” in 1891, the six main groups—“Europeans and Americans”, “Eurasians”, “Chinese”, “Malays and other Natives of the Archipelago”, “Tamils and other Natives of India”, and “Other”²⁷ —were, in the 1901 census, adapted to London's terminology. This latter census deployed the term “race” interchangeably with “nationality”.²⁸ Moreover, race had become the primary classificatory category in the census of the Federated Malay States that year, and was retained in the following enumerations.²⁹

In preparation for the census of 1911, colonial governors were asked for “as much uniformity as possible”.³⁰ And even if by the time the report was published (after World War I authorities in London questioned whether the exercise “serve[d] any very useful purpose”),³¹ London renewed its request for a return for race, offering specific grouping suggestions in the 1920s. This was also the high point of the race discourse in Europe. By now, the all-India census had achieved some degree of homogeneity, even though it remained characterised by internal contradictions between an “urge to specificity and to generalizability”.³²

As ‘scientific’ definitions of race had taken hold of Europe and the United States, the intellectual framing of the census of British India (if not the actual enumeration of its subjects) was in-step with anthropometric understandings of race and caste. Hence, in 1931, the census of India published B. S. Guha's essay “Racial Affinities of the Peoples of India”, written under the mentorship of Earnest Hooton, Guha's PhD adviser at Harvard University and a prominent representative of ‘scientific’ racism.³³

Southeast Asia then stands clearly as an outsider. Instead of focusing on ‘scientific’ theories of race, racial groups—and thus claims to indigeneity—begun to emerge as defined by, and sometimes defining of, religious identities. Beginning with the 1921 census of Malaya, individuals who performed their identity as Malays were classified as Malays—including individuals of Chinese descent (the Maalap)—and were tabulated as Muslims by default; similarly, all Europeans and Eurasians were deemed Christians.³⁴ Racial affiliation carried an embedded religious identity, and vice versa. The report for 1921 Burma stated that “Persons who are partly of Hindu and partly of Burmese descent” or of mixed Burman-Chinese ancestry (where Burman usually meant Buddhist, and at times Animist), could be entered according to racial self-identification: “if they are in doubt record the race of the father or of the mother according to the customs in which the person is brought up” [emphasis mine]. But for Muslims this was not an option: those whose “father is a Mahomedan of any race and the mother Burmese” had to be classified as Zerbadis, a foreign Indo-Burman race.³⁵ In the next section I explore the implications of this shift for the construction of national identities and belonging, as in late- and post-colonial Burma, Muslims of mixed heritage were deprived of the opportunity to self-determine or assimilate, ultimately being marginalised and excluded.

Early census reports in the nineteenth century reflected on the nuances and complexities of identities in Asia's diverse British colonies as geographical origin, rather than physiognomics or anthropometrics, was a crucial identifier of colonial subjects’ race. But London's requests for standardised and comparable classifications dictated change. Notably, though, in Southeast Asia the social, political, and economic contexts stirred the conversation on subjects’ classification in other directions. Census superintendents were first challenged by, and then became receptive to, local understandings of race as a cultural phenomenon (later labelling it as “community”). While B. S. Guha published his essay grounded in scientific racism, C. A. Vlieland's report on the census of British Malaya suggested that race should be considered as a “judicious blend” of geographic origin, ethnic descent, nationality, and customs.³⁶

Indigeneity and foreignness

Entangled with the emergence of the terminology of race, which was further politicised in early twentieth-century Europe, was the question of who was ‘native’ to a given territory and who had been a migrant or foreign settler. Whereas this question was not applied to white settlers, it assumed increasing political relevance as the Great Depression opened the conversation on decolonisation across Asia. If in a first instance the classification of the Empire's population in racial groups was aimed at distinguishing the white population from the “coloured races”, it later became a tool to differentiate between native and foreign non-whites.³⁷

The census of Penang had already deployed the term “natives” in 1857, and this was not unique. In 1881, the Chinese population of the Straits Settlements was enumerated separately from those born in the Straits. In 1901, the report of the Federated Malay states differentiated between “Malays” and the “Natives of the Archipelago”. In Burma, various Muslim sub-groups—depending on place of enumeration, place of birth, or even language—were classified as either indigenous or aliens. In Hong Kong, almost the entire population was considered ‘foreign’ by the British, who propagated the myth of pre-colonial Hong Kong as a barren rock.³⁸ But through the decades, while certain Chinese groups were seen as settled, Euro-Americans and Indian peoples retained their foreign status as members of the civil or military administration. Being classified as ‘native’ or ‘foreign’ had little bearing on individuals during the decades of colonial rule, but on the eve of colonial independence it came to mean inclusion or exclusion from the new nation-states.

The Muslims of Arakan

Upon their take-over of Arakan in the 1820s, the British found the region promising, with natural resources, but scanty in potential labourers. Building on their experience in Singapore, the British devised plans for diverting Chinese and Indian migrants away from the Straits Settlements and towards Lower Burma instead. By 1881, the population of the province had more than tripled, but not without hiccups. In 1835, Chittagonian coolies showed that they were not so keen on relocating to the jungles of Arakan, and many preferred being imprisoned for breach of contract to life in that insalubrious region.³⁹

Reflecting on Lower Burma's population trends in the 1883 Gazetteer, Major Douglas MacNeill pointed at the heavy flow of coolie migrants from Upper Burma, Madras, and Chittagong into Arakan, Pegu, and Tenasserim.⁴⁰ As “immigrants from India came in the first instance without their wives”,⁴¹ the gender balance of Muslims and Hindus in Lower Burma was heavily skewed; and yet Arakan was singled out as an exception. Already in 1872 Muslims were considered indigenous to the region. Alongside the many migrants that continued to arrive into Arakan from Cox's Bazaar every year, the census also reported that “the Mussulman population of Akyab, however, is not, as elsewhere in the province, alien, as they have for the most part been settled in the province for many generations”.⁴² The Arakan Muslim community “has been so long in the country that it may be called indigenous”.⁴³

A decade later, the Superintendent stated: “There are few or no indigenous Hindus, though there is a considerable indigenous Mahomedan population in Arakan.”⁴⁴ In 1891, renewed waves of immigration from across the Bay of Bengal were defining the character of Islam as Indian and foreign, to the extent that census Superintendent Eales asserted that Islam was “of comparatively recent introduction”. But indigenous Muslims retained their turf in other areas as “Burman-Muslims”. Villages in Ramree Township (Kyaukpyu District) had a significant portion of their Muslim population—sometimes its totality—indicating Arakanese as their parent tongue.⁴⁵ In Sandoway District, Muslims in several villages in the Thade and Zadibyin circles (Central Township) spoke Burmese.⁴⁶ In Hanthawaddy District, at least two villages had their Muslim population identifying Burmese as their parent tongue.⁴⁷

The structure of the following census, in 1901, affirmed the primacy of religion, and clearly separated indigenous and foreign groups. Whereas “Buddhist and Animist” could be of either kind, Superintendent C. C. Lowis made it clear that for him Hinduism, Islam, and Christianity were ‘foreign’ religions, and its affiliates were to be considered outsiders to Burma's social fabric. Recognising that “the Muslim strain is at times of greater antiquity” descending from Muslim immigrants from northern India at the time of King Alaungpra (or Alaungpaya) in the mid-eighteenth century, and from prisoners caught by King Mindon a century later,⁴⁸ Lowis concluded that Islam was the “principal” among the “non-indigenous religions of the province”.⁴⁹ The more census superintendents drafted ‘big picture’ tables, the more this extraction of general principles went at the expense of minority groups.

Morgan Webb, census superintendent in 1911, offered yet another window on the specificities of certain sub-districts. In his report, Webb declared: “The Hindu and Mohammedan religions are professed principally by recent immigrants from India and their descendants, though there exist several communities of both religions who have been established in the province for many generations”.⁵⁰ Akyab and Mergui divisions, for example, hosted a “large indigenous Mahomedan population”.⁵¹ Akyab held the largest proportion of Muslims across Burma at 33.66 per cent of the population,⁵² and 15.34 per cent of Burma's total Muslims (second only to the City of Rangoon, with 26.86 per cent). Indeed, “in Akyab [the Muslims] are indigenous”;⁵³ they are “scarcely differentiated from the neighbouring Arakanese or Burmese in dress and speech and customs”.⁵⁴

In 1921, Muslim Burmans were still present in districts all across the province,⁵⁵ and at least two ‘indigenous groups’ listed Muslim members.⁵⁶ The census, then, had not made a univocal determination as per whether there could be ‘indigenous’ Muslims, but the primary lens was no longer place of origin or birth, rather religious affiliation. The parameter to adjudicate indigeneity to Burma, then, was religious affiliation, as Hindu and Muslim ‘immigrants’ from across the Bay of Bengal were differentiated from the ‘native’ Buddhists and Animists. Through the decades, the focus shifted towards “race”. The classification of “mixed Indo-Burman races” soon came to be unharnessed from genealogical descent, especially when Islam was part of the picture.

In the 1920s-1940s, Islam was increasingly seen as not belonging to Burma's cultural landscape, and as relating almost exclusively to “Indians”. Zerbadis and Arakanese Muslims both came to be considered groups of the “Indo-Burman Races”. I pursue this thread of compounded racial and religious identities in the next section of this article, but what needs to be addressed here is that the census enumerations and reports were not operating in a political vacuum. In his 1921 report, Grantham explained that his choice of setting the Muslims apart was inspired by his perception that the Zerbadis “regard[ed] themselves as a distinct race”,⁵⁷ leading him to assume that self-identification as a distinct race meant setting themselves aside from society as foreigners. In fact, the very foundation of Zerbadi interest groups was to affirm their “Burmese-ness” and advance their rights as Burmese nationals, not as Indians or mixed subjects.

In a memorandum sent to the Public Service Commission in 1913, the Rangoon Zerbadi Community drew a firm line between themselves and “the Indians who come to this country for a short time”. This claim was made on the grounds that the Zerbadis represented the descendants of the “many Indians, both Hindus and Mahomedans [who] emigrated to Burma” before its annexation by the British, and pointed at their assimilation through language, marriage, and customs, and the fact that they knew no other home but Burma. Ultimately, “in everything except religion Zerbadies are to all intents and purpose Burmans”.⁵⁸

In 1931 the census recorded Muslim subjects in several more race-groups than 1921,⁵⁹ and both the Zerbadis⁶⁰ and the “Indo-Burman Races”⁶¹ were included as “Indigenous Races”. Maybe this was J. J. Bennison's top-down attempt at integrating the Muslim population that for so long had inhabited Burma, as the province was about to become autonomous from British India. Concerns over the rights of the Indian minorities became stronger in the late 1920s and 1940s.⁶² The voice of Burma's Muslims became louder,⁶³ protesting against Burma's separation from British India, requesting a separate Muslim electorate and the protection of Urdu language education to safeguard religious traditions.⁶⁴ Nevertheless, these demands were never separate from the community's self-proclamation as an “indigenous minority”, as stated in the early 1910s⁶⁵ and reaffirmed in 1935.⁶⁶

The first census of ‘separated Burma’ was held in 1941, and due to the war, its report is a bare-bone affair. However, instructions sent out in early 1940 were clear in seeking “information regarding permanent and temporary Indian, Chinese and Gurkha immigrants” while fertility rates were to be “confined to indigenous races”.⁶⁷ The distinction between indigenous and foreign races was further reinforced in the post-War era. The 1953 census report explained how several of the “minor groups” were included in larger “basic” groups of indigenous races: “These main groups speak different languages. Minor groups within a main group may have different dialects but this is more emphasised in remote hill districts. Though habit and custom of the people in the main groups are different from one another, there is a tendency for these differences to disappear”. Indians, Pakistanis, Chinese, Europeans, and Americans were instead all labelled as “foreign races”.⁶⁸

What had started as a historical colonial conversation in the 1800s over whether Islam and Muslims should be considered ‘indigenous’ or not to Burma, had evolved in a deeply politicised direction in the 1920s-1930s. Proving indigeneity at the time of the British arrival became necessary to advance claims for citizenship, when the nationalist discourse was taking shape around the idea of Burma as a Buddhist nation. Within this vision, being Burmese (Bar-Ma) meant being Buddhist, and being Muslim meant being foreign.

Maritime Southeast Asia

The Indo-Malay Archipelago had been rich in maritime connections, with ‘local’ and ‘foreign’ sailors, traders, and pilgrims connecting Java, Sumatra, and the Malay Peninsula to South Asia and beyond, since well before the European arrival. These categories remained fairly fluid and ultimately irrelevant through the centuries, until the British establishment of the Straits Settlements in 1826, and the beginning of enumeration and ‘classification’. These processes became all the more important as the British contributed to the diversification of their territories through migration policies that specifically brought in Chinese and Indian subjects to work in Southeast Asia's plantations and mines.

Initial distinctions were made based on geography (“East” versus “Proper” Indians; Malays versus Buginese, etc.) or between civilians, military men, and prisoners.⁶⁹ But Lionel Mabbot Woodward, Deputy Superintendent of the Census of Penang in 1891, soon expressed his dissatisfaction with the methodology applied to the counting of the Indian population in the Settlements, and his thoughts were revived by J. R. Innes, Straits Settlements Superintendent in 1901. Similarly, Superintendent Hare of the Federated Malay States for the census of 1901 made explicit the distinction between ‘native’ and ‘foreign’ Malays.⁷⁰ Starting in 1911, in both the Straits Settlements and the Federation, the census reports began to reflect on who belonged to each sub-category.

Straits-born Indians were separated from India-born Indians in 1911.⁷¹ Such accommodation was only temporary, though, as the 1921 census of British Malaya—the first report combining the Malay States and the Straits Settlements—did away with the ‘Straits-born’ category, for Indians and Chinese alike. On this occasion, “careful scrutiny of the information supplied in the birthplace, language and religion columns of the schedules made it possible, in most cases, to classify Indians correctly according to race”.⁷² On the one hand, all Indians (and Chinese) were marked as foreigners; on the other, race was presented as a compounded category, inclusive also of religion.

The year 1911 was a turning point also for the way that Malays were enumerated in the peninsula, after the 1901 report for the Federated Malay States had advanced a suggestion for refining racial groups.⁷³ Hence, as A. M. Pountney complained that “many foreign-born Malays were included in the returns of Malays published in the 1901 census report”,⁷⁴ the census of 1911 applied subcategories identifying place of birth or origin for the ‘foreign’ Malays (e.g. Javanese, Buginese, Sumatrans, etc.), geography and language for the Indians (overwhelmingly Tamils from Madras, but also Telugus, Punjabis, Bengalis, Malayalis, Hindustanis, Afghans, Gujaratis, Maharattas, etc.), and tribe or language for the Chinese (Hakka, Hokkien, Cantonese, Hailam, Tie Chiu, etc.).⁷⁵

Through the decades, superintendents regularly voiced the difficulty of enumerating Malays correctly and attempted different strategies to differentiate between native and foreign Malays. Echoing A. M. Pountney, J. E. Nathan explained how “considerable difficulty was experienced in coming to a decision as to which of these races should be tabulated separately, and which amalgamated under the heading ‘Malay’”. Whereas “the Japanese [sic. Javanese?], the Banjarese and D[a]yaks from Borneo, the Boyanese from Bawean and the Bugis from the Celebes are distinct races with separate languages and customs” who preserve them after migrating, “it is in dealing with the Malays of Sumatra that a decision is difficult.”

Nathan eventually took an ‘inclusivist’ approach, concluding in 1921 that the term “Malay” should apply to all Peninsular Malays and all Sumatran Malays except “Achinese, Korinchi and Mendeling” because “Sumatra was originally the home of the peninsular Malay, and linguistically, ethnically, and ethnologically the Malays of British Malaya and the Malays of Jambi, Kampar, Siak, Menangkabau and the other districts of Sumatra are one race. No fundamental error is involved in their tabulation under one head.”⁷⁶ But his successor, C. A. Vlieland, saw things differently as he intended to reflect the persistence of “race consciousness, customs and language” as well as political status, in the criteria for race. Hence, in 1931:

The term ‘Malaysian’ is used to include all indigenous peoples of the Malay peninsula and archipelago, and the term ‘Malay’ to include only those ‘Malaysians’ (excluding aboriginals) who belong to British Malaya. No immigrants born in Java, Sumatra or other parts of the Malay Archipelago, are numbered amongst ‘Malays’, and the children and later descendants of the original immigrant are only treated as ‘Malays’ when definitely so returned … the children born in Malaya of Sumatran parents are normally returned as ‘Malay’ and so classified.⁷⁷

In 1947, Del Tufo introduced a seventh “main race”, namely the “Other Malaysians”. Even though he was not too convinced by the effectiveness of this distinction himself—“the term other Malaysian … might well be dropped in favour either of ‘immigrant Malaysian’ or of ‘Indonesian,’ the latter, preferably, since it has the merit of meaning something”—Del Tufo appeared torn between the ability of “Other Malaysians” to assimilate by virtue of their religious and linguistic affinity to the “Malay proper”, and the fact that “differences do exist”.⁷⁸

Intersectionality: the ethno-religious identity compound

In Burma, views over whether Muslims could be indigenous, or if they were all ‘foreigners’ from India, had alternated. Similarly, census reports and tables from Malaya and the Straits Settlements show that superintendents and enumerators had attempted to capture as many combinations of ethnic and religious identities as possible. But as local concerns over who was native and who was foreign met with imperial trends towards standardisation and the creation of comparable identities, the final outcome of these competing and interweaving priorities was the emergence of combined ethno-religious intersectional identities in the late colonial period. These ‘pre-packaged’ identities had been constructed by government officials who had come to understand race as a broader concept than physiognomic genetic features, but who also thought of religion as an all-encompassing affiliation.

Hong Kong is another important piece of this puzzle, for at least two reasons: first, it is rarely included in historical analyses of Southeast Asia; secondly because, as rarely, is ‘religion’ considered a relevant category of identity for analysis. As I have argued elsewhere, this illusion is created by the very fact that colonial policies gave shape to intersectional ethno-religious identities.⁷⁹ As hinted at by Hong Kong New Territory's district officer S. B. C. Ross in 1911, the British government saw conversion and distance from traditional beliefs as a possibly “harmful” change, as it would have upset the status quo.⁸⁰ Within this perspective, the Chinese population, and especially those who lived in the rural areas, was expected to retain its allegiance to Animism or Confucianism, with some exposure to Buddhism. Christianity was to remain the purview of the English and some of the Chinese elite, while Islam and Hinduism characterised the (transient) Indian population. In this three-partite division of society, ethno-religious affiliation reflected different degrees of indigeneity and foreignness, and implied social status.

The British retained a strongly stratified society, in which racial belonging determined one's place in the settlement, as a living individual and a dead body. Whereas racially determined urban planning was a feature of British rule across the colonies,⁸¹ in Hong Kong such strictures applied in even greater detail to cemeteries. These had been set since before the earliest census, and still remain in operation today. Since the mid-nineteenth century, Hong Kong's burial grounds were set up to host specific ethno-religious communities. The Colonial Cemetery—also known as the Protestant or Anglican Cemetery—was opened in Happy Valley in 1845, and the Catholic Cemetery was opened, right next to it, in 1848. The Zoroastrian, Jewish, Muslim, Hindu, and Sikh burial grounds followed suit in the 1850s-1880s to accommodate wealthy merchants (as South Asian troops started to arrive in Hong Kong only later on).⁸² Chinese immigrants were buried along the island's hillsides until 1871, when the first Chinese cemetery was opened in Kowloon; and although in the early days a few Christian Chinese were buried in the Colonial Cemetery—as conversion to Christianity was the one avenue towards assimilation and privilege⁸³ —two Chinese Christian cemeteries were authorised in 1882.⁸⁴

Burmese are Buddhists

In Burma, Eales had stated that religion could not be a proxy indicator for race, but the introduction of race as the primary classificatory principle for the 1891 census set up a structure for the narrowing of options in terms of religious affiliation. At this point, most likely to conform to London's request, the notion of race had been introduced in the summative tables. Whereas Eales had stated that this was meant for clarity,⁸⁵ the new tables for “Race, Tribe and Caste” only created more confusion. Burma's subjects were classified as Hindu, Muslims, Burmese, Chinese, Chin, Karen, Miscellaneous, and European Races. Within this scheme, each category was then sub-divided in specific groups, such as individual Hindu castes, Muslim tribes, Burmese racial groups, European nationalities, etc. What can be surmised from the various tables and the final “Index of Castes, Tribes, and Races”, is that the term “Hindu Castes” and “Musalman Tribes” had relatively little to do with the religious affiliation of the individuals enumerated, but were rather treated as racial classifications. The “Hindu Castes” Dhobis, Fakirs, Gujaratti and Gurkha all included Hindus as well as Muslims; the Jats and Lohars listed Hindus and Sikhs. The “Musalman Tribe” of the Pathans included Muslims and Christians. And many of the Burmese and cognate races—Burmese, Arakanese, Karens, Kathe, Malays, and Shans—had Muslims among them.⁸⁶

The following census, in 1901, took a reversed approach, affirming the primacy of religion. In this framework, religious belonging was taken as an indicator of either indigeneity or foreignness. Moreover, as race and religion started to interweave as classificatory headings, the population of Burma was gradually channelled in prescriptive sub-groupings. Superintendent C. C. Lowis concluded that “Buddhism is the sole religion of the provincials [and] none but the indigenous profess it.” But he also candidly admitted that “Neither assumption is, strictly speaking, correct but, for the purpose of comparison such as it is here desired to give, the non-Buddhistic people of Burma may be treated as a negligible quantity.”⁸⁷It is thus that the statistically insignificant Burman-Muslims were relegated to the margins of the British census.

The census of 1921 described the two main groups of Burma Muslims—the Zerbadis and the Arakanese Muslims—as “mixed Indo-Burman races which in some ways seem to attach themselves to the Burmese but in other are rather Indian”.⁸⁸ In the classification of “Race-Groups by Religion and District”, which followed the linguistic classification, Zerbadis and Arakanese Muslims were listed as “Indo-Burman Races”,⁸⁹ one of “five artificial groups” compounding the “foreign population” of Burma.⁹⁰ This approach was reconfirmed when “Arakanese Muslims” were listed as a racial group separate from the Arakanese Buddhists, as well as from the Chittagonian and Bengali Muslims.⁹¹ Grantham saw Zerbadis’ associationism, shown in the formation of a Burma Moslem Society as a sign of rising “racial consciousness”. He further reflected on the position of Arakanese Muslims, observing that “racially … [they] did not associate with [the Arakanese] at all … marry almost solely among themselves and have become recognised locally as a distinct race”, thus justifying the fact that “The Arakanese Buddhists in Akyab asked the Deputy Commissioner there not to let the Arakan-Mahomedans to be included under Arakanese in the census”.⁹² Religious affiliation had become an indication of racial identity.

A newly introduced table, distributing the population “by people and religion”, made Christianity, Islam, Hinduism, Buddhism, and Animism into mutually exclusive religious affiliations for any given “people”.⁹³ Under this heading, Karens belonged to different groupings depending on whether they were Christians or not; Burman Muslims were separated from “Other Muslims”, but also from “Burmans” who had returned their religion as Hindu, Buddhists, or Animists. The goal of this table was to specifically rebalance the relationship between race and religion. Grantham was convinced that in Burma race was more suitable a basis of classification than religion (as it was used in India),⁹⁴ but in fact his tabulations only led to religion becoming enshrined into definitions of race. Its political implications are still relevant today.

… and Malays are Muslims

At the beginning of the twentieth century, categorisation by race and religion was taken as a useful metric beyond its own immediate scope, reinforcing what the colonials thought they knew already about their subject population. They had concluded that Confucianism could be considered the “national religion of China”, “if that can be strictly styled a religion”,⁹⁵ because most local Chinese returned themselves as Confucian; similarly, Superintendent A. M. Pountney concluded that “Mohammedanism is the national religion of the Malays”.⁹⁶ Even though some Malays had indicated that they practised Christianity, Buddhism, “other religions”, or no religion at all, Islam had claimed the lion's share of their returns.⁹⁷

A mere decade later, the Advisory Committee “considered that several of these tables [dealing with religion and published in 1911] were of such little value as not to warrant the labour and expense involved in their compilation”. As “the number of non-Christians among the Europeans and Eurasians and the number of non-Muhammadans among the Malays are so small”, only the tables dealing with the Chinese and Indians should be published in the report. What is more, within this already limited scope, Nathan quoted Pountney in determining that as “it is difficult to say what is the religion of those Chinese who are neither Christian nor Muhammadan … no attempt has been made in this report to differentiate between the religions of those Chinese”.⁹⁸

Despite these simplifications, difficulties were still emerging. Nathan's main disconcertment in 1921 emerged upon reflecting on the small number of Chinese who returned themselves as Muslims “open[ing] to doubt whether the number is not in reality considerably higher”.

A Chinese shopkeeper will settle and open a small shop in a remote Malay village, where perhaps he is the only one of his race. He learns the language and customs of the people of his village, grows friendly with them and in time wishes to take a wife from among them and make his permanent home there. The difference of religion is, however, an insuperable bar: a Malay woman cannot marry any but a Muhammadan, and so in some cases the Chinese will embrace Islam and became what is known in Malay as a ‘Maalap’.⁹⁹

Maalap was not an official census category, and the Chinese man described above would have self-identified as Malay. The Jawi-Pekan, as a Muslim Indian married into a Malay family, would have done the same.¹⁰⁰ Becoming Muslim was cognate with becoming Malay.

In the subsequent census, in 1931, Vlieland observed that “The difficulty of achieving anything like a scientific or logically consistent classification is enhanced by the fact that most Oriental peoples have themselves no clear conception of race, and commonly regard religion as the most important, if not the determinant, element.” Thus, he had suggested that ‘race’ should have been deployed in the Malay census as a “judicious blend, for practical ends, of the ideas of geographic and ethnographic origin, political allegiance, and racial and social affinities and sympathies”.¹⁰¹

In 1947, Superintendent Del Tufo suggested that instead of race the census should “gradually” replace this term with “‘community’ (although even that term is not free from objection) to connote groups whose members are bound together by a community of interests, that is to say by common ties of language, religion, custom or allegiance”.¹⁰²

By the mid-twentieth century the census had stopped enquiring about religion, and ‘race’ had been changed to ‘community’. In Del Tufo's words:

Past experiences having shown it to be of little value in Malaya where the entire Malay population is Muhammadan, practically every European and Eurasian is a Christian and the great majority of Chinese hold to the national religion of China which some describe as Confucian and others prefer to regard as ancestor-worship. Significant results would have been obtained only in the case of the Indians … but as their number was expected not to exceed one-tenth of the total population the inclusion of this enquiry … was not considered to be justified.

Del Tufo thus offered estimates, aggregating the perceived religious affiliation of various racial groups: the “Malays” counted as Muslims; the “other Malaysians” as mostly Muslims, with the noted exceptions of “the handful of Balinese [who] are probably Hindus, and there may also be a few Christians”; the “Aboriginal Tribes” as pagan, considering that those who had converted to Islam would return themselves as Malays; “Chinese” and “Indians” were returned according to the averages from 1921 and 1931; and all “Europeans” and “Eurasians” were assumed to be Christians.¹⁰³

The suggestion to use ‘communities’ was embraced by the last colonial census, held in July 1957. Almost quoting verbatim from Vieland's 1931 Report, H. Fell sentenced: “Race is used in the sense in which is understood by the man in the street … a term which blends, for practical purposes, the ideas of geographic and ethnographic origin, political allegiance, and social affinities and sympathies”.¹⁰⁴

This shift from a colonial physiognomic construct to a locally understood group of belonging, could have been an empowering transformation of the term race. However, in later years this reformulation came to impose a constructed homogeneity over a diverse reality (as not all Malays were Muslim), and cemented religion as an interlocked aspect of language and customs.

Conclusions

Burma's nationalist debate, still strong in the 2010s, took shape in the 1920s as “Burma for the Burmese Buddhists”,¹⁰⁵ and Buddhist lay persons and monks were actively opposing the possibility for Muslims and Hindus—or, as they called them, “Moulvi and Yogi Weda”—to feel at home in the new Burma, fomenting dissent (as at the Shwe Dagon Pagoda in July 1938) and causing clashes (both in Rangoon and Mandalay in late 1938).¹⁰⁶ In the first post-colonial census of Malaysia (in 1970), the report explained that the only non-Muslim Malays were from among the Orang Asli (“Aborigines” as categorised by the British) and foreign Indonesians;¹⁰⁷ beginning in 1991 the option of being both “Malay” and non-Muslim had disappeared altogether.¹⁰⁸ In 2013, Hong Kong's Chinese Church Alliance could still refuse the request for burial of a permanent resident of Indian ethnicity and Christian religion, because “by law the cemeteries under the organisation only take ethnic Chinese applicants”.¹⁰⁹ The Alliance controls all of the city's active Christian cemeteries, which were established in the nineteenth century.

In his work on ethnic conflict in the Balkans, Brubaker advanced the suggestion that there “ethnopolitical entrepreneurs” contributed to ethno-religious tensions and conflict by making abstract categories real:

By invoking groups, they [ethnopolitical entrepreneurs] seek to evoke them, summon them, call them into being. Their categories are for doing—designated to stir, summon, justify, mobilise, kindle and energise. By reifying groups, by treating them as substantial things-in-the-world, ethnopolitical entrepreneurs can, as Bourdieu notes, ‘contribute to producing what they apparently describe or designate’.¹¹⁰

British Asia is a good comparative ground for Brubaker's analysis. First, the British contributed to the region's diversification through their migration and labour policies. And next, census superintendents and enumerators produced new groups, qualitatively significantly different from what they were initially only supposed to describe, preparing the grounds for post-colonial societies to emerge as structured along ethno-religious identities.

Census practices relating to the classification of colonial subjects’ religious and racial affiliation varied across territories, yet in each of the cases studied here I have shown the sustained creation, through census practices, of new intersectional ethno-religious identities. These practices encouraged the narrowing of recognised groups, and empowered superintendents and other officers (both colonialists and local)¹¹¹ to draw simplistic conclusions on the demographic and sociological outlook of each territory. Such reliance on quantitative data, combined with prioritizing colonial mapping and boundary-drawing, favoured majority groups, while leading to the creation of bold demarcation lines where more porous borders had previously existed, and eventually to the elision of marginal minorities. By the turn of the twentieth century, Buddhism was seen as the sole religion of Burma's indigenous inhabitants; Islam was identified as the core characteristic of the inhabitants of the Peninsula, i.e. the Malays. Alongside a fixation with numerical majorities, post-colonial identities were so fixed.

This article has offered a historical key to better understand processes of identity formation in late- and post-colonial Southeast Asia, where exclusionary identity politics based on ethno-religious identities has continued to escalate for decades. The close reading of census reports and data from British Burma and Malaya—two countries (Burma/Myanmar and Malaysia) where contemporary politics remains today deeply entrenched in ethno-religious identities—was supported by data from the Straits Settlements and Hong Kong to diffuse potential arguments for the exceptionality of Burma/Myanmar and Malay(si)a. This analysis highlighted changing framings and census tabulations of ‘race’, and reflected on the interweaving of London's efforts to standardise data with ongoing conversations in the colonies.

The importance of honing in on intersectionality, rather than isolating religious or ethnic identities became evident as, on the one hand, individuals were stripped of the opportunity to self-determine their religious affiliation, and, on the other hand, states drew the boundaries of indigeneity according to such intersectionalities, and based on their imagination of a given territory's ‘authentic identity’.