In contemporary Malaysia, the Malay identity is a socially constructed identity that lumps together a number of people of a wide range of regional and national descents. Even legally, the definition of Malayness refers to practice, language, and self-identity. The percentage of the population of peninsular Malaysia that identifies as indigenous but not Malay is vanishingly small; and among that population, the vast majority identify as orang asli, Thai, or as members of one of the bumiputera ethnic groups from Sabah and Sarawak.

I have long wondered if this was always the case. One place to look is censuses, which Charles Hirschmann argued three decades ago is a particularly productive way to observe how political actors surveyed the landscape of people in front of them. The feature of censuses in Malaysia, both under British rule and after British rule, is the urge to classify people into types. Uniquely, British colonial census officials often reflected upon the nature of that task. Those reflections contain within them evidence of what they believed to be the structure of the society in front of them.

And contrary to the widely held assumption among many scholars of colonialism, and certainly among most scholars of Malaysian history, the British census reflects both thoughtfulness and uncertainty about how to classify the people of British Malaya.

We can actually read from as early as 1921 about the challenges of defining what a Malay is. From the Census of British Malaya, 1921, page 22.

Under the heading “Malays” are grouped all the native peoples of the Malayan Archipelago and 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.” The difficulty does not arise in respect of the Malayan peoples from the islands in the South and East of the Archipelago. The Japanese [sic], the Banjarese and Dyaks [sic] from Borneo, the Boyanese from Bawean and the Bugis from Celebes are distinct races with separate languages and customs and, after emigrating to this country, they preserve their distinctive features for generations and are not merged in the native Malay population. The child born in British Malaya of Javanese and Boyanese parents would invariably be described as Javanese or Boyanese on the Census schedules. It is in dealing with the Malays of Sumatra that a decision is difficult.

If we can take the census author’s words as reflecting something of the social reality of the time, we see that as late as 1920 “Malay” as an ethnic identity did not encompass all indigenous peoples of the region. Malay could be used as a short-hand for a perceived racial or descent category, but this would be different than an ethnic identity. It was therefore possible to speak of industrious “Javanese Malay” in the late 19th century (for examples, see The Myth of the Lazy Native, p. 74).

Now let us look to a more recent census, from Malaya: A Report on the 1947 Census of Population, p. 72.

The distinction between the Malays and the Other Malaysians is not very great and is, indeed, ignored in the compilation of Malayan vital statistics; for the Malays themselves are, to a large extent, descended from the Malays of the east Coast of Sumatra from whom they…are ethnographically indistinguishable. For that matter, on the occasion of the 1921 census when the classification employed was based on ethnographical rather than social consideration, all Sumatran Malays exception the Achinese, Korinchi, and Mendeling were actually included with the Malays proper in the specific race tables; while through the remainder of the tabulation all Malaysians were lumped together under the head “Malays.”

As for the Other Malaysians, all but a handful of them are of the Muslim faith and speak the Malay language and are, although to varying degrees, readily assimilable with the Malays of the peninsula with whom, in fact, they tend to form a single community.

Just twenty-six years later the sharp distinctions between Javanese, Boyanese, Bugis, etc. and Malay are beginning to fade. They are still tabulated separately—the State of Selangor in 1947 was inhabited by 53,859 Javanese versus 103,456 Malays, plus another 14,322 Banjarese—but the task of distinguishing them is recognized to be difficult and imperfect precisely because of changing local understandings of identity.

This 1947 census is particularly interesting as it is completely and utterly constructivist with respect to questions of “race” or ethnicity. Take a look at this, from p. 71, my emphasis added:

Enumerators were instructed that, unless it was manifestly absurd or impossible, the statement made by the person questioned should be accepted….Very few Indonesian immigrants, for example, would claim to be Malays unless they were accepted as such by the village community. So, too, the girl of Chinese blood adopted in infancy by a Tamil would be entered as a Tamil unless she insisted upon being described as a Chinese. These were merely the natural consequences of insistence upon community; for the Chinese girl in the illustration would almost certainly speak Tamil and no 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 in which it is used in this context.

By the 1970 and 1980 censuses, carried out by the independent Malaysian government, there was indeed no separate category at all for Javanese, Boyanese, or any of the other non-Malay groups meticulously recorded by the British. There is only a subcategory within the general Malay category of “Indonesian,” numbering 228,126 in Peninsular Malaysia in 1980, and it is not clear who would have so identified.

By 1991 things have changed yet again. Now there is a category termed “other Bumiputera,” and the category “Indonesian” has moved under the general category “Other,” alongside Europeans and Vietnamese. “Other Bumiputera” in peninsular Malaysia number 120,685 in 1991, far less than 228,126 “Indonesians” of 1980, whereas “Other” number 397,394, far higher than the total of 68,910 of 1980. This is might be evidence that those groups identifying as Indonesian in 1980 and counted under the category of “Malay” have preserved that under the category of “Other” were it not for the fact of migration from Indonesia. To be clear, the figures I am citing are only among Malaysian citizens, but it is difficult to know how many of these citizens are recent migrants who have taken citizenship versus descendants of migrants from generations ago.

I would note as well that neither the 1980 nor the 1991 census contains any of the reflexive discussion about defining identity that the two censuses of British Malaya did.

The message in all of this is necessarily incomplete, because we are looking only at figures and how a couple of colonial administrators described their tasks. Yet with those caveats, we can begin to see evidence that the porousness of what has become the modern Malay identity is a relatively recent phenomenon. We also see evidence of constructivists thinking on ethnicity and identity by colonial bureaucrats, long before constructivist approaches to identity gained ascendance in academia.