THE JAPANESE COLONIAL STATE AND ITS FORM OF KNOWLEDGE IN TAIWAN
The fact that societies can become the object of scientific observation, that human behaviour became, from a certain point on, a problem to be analysed and resolved, all that is bound up, I believe, with mechanisms of power.
—Michel Foucault
As alien rulers who hardly knew anything about Taiwan before they landed on the island at the end of the nineteenth century, the members of the Japanese colonial government inevitably encountered two fundamental difficulties, which can be encapsulated in the universal questions posed by Bruno Latour in another context: “how to be familiar with things, people and events which are distant,” and, in turn, “how to act at a distance on unfamiliar events, places, and people” (Latour 1987:220, 223). Latour’s answer to these questions is, of course, already well known: by appealing to “some mobile, stable and combinable means to bring home these events, places and people” (Latour 1987:223). This essay is an attempt to investigate the specific knowledge made possible by those “mobile, stable and combinable means” which from the end of the nineteenth century began to “bring home” everything about the faraway colony to the Japanese rulers.
One of my concerns here is with the form of colonial knowledge in Taiwan. “Colonial knowledge” involves two types of question: first, “who represents whom,” or “who is represented at the expense of whom” (i.e., the question of imagination or representation); second, and important, questions regarding what tool is used to understand and describe the colonized, what exactly the technique of the representation is, and what the relationship, if any, is between the specific technique and the colonial administration.
Mainly because of the great success and influence of Edward Said’s Orientalism (1978), which is almost exclusively focused on the former type of question, there has been relatively little attempt among scholars to explore the latter aspect of colonial knowledge. That is to say, colonial knowledge is now conceived of by most postcolonial theorists purely as a matter of discourse and representation. Moreover, these theorists overwhelmingly prefer the question of “what has been mis-represented” to the question of “what has been represented.” This is a serious theoretical inadequacy and bias. I am not suggesting, of course, that the “discursive approach” of Said and his followers is totally wrong. What is being suggested here is rather that, as far as colonial knowledge is concerned, we need to begin to see things from a governmentality approach. To do so, we must pay attention less to the “effect” of (mis) representation and more to those “humble and mundane mechanisms which appear to make it possible to govern” (Miller and Rose 1990:8). Again, to borrow Latour’s wisdom, “what is called ‘knowledge’ cannot be defined without understanding what gaining knowledge means” (Latour 1987:220; italics in the original).
Two general types of source material will be discussed here: maps and numbers. Both are mobile, stable, and combinable means, as suggested by Latour. Of course, I am not the first person to draw attention to those “humble and mundane mechanisms” in the colonial contexts. Matthew Edney’s Mapping an Empire (1997), for example, offers an excellent explanation of the link between cartography and British colonialism in India. As far as numbers are concerned, Arjun Appadurai’s essay “Number in the Colonial Imagination” investigates the agrarian surveys and censuses in colonial India, and relates them to the formation of Indian nationalism and the discipline of the colonized. Appadurai’s discussion is particularly useful and interesting for the current purpose, for he points out that Said’s Orientalism “does not specify how exactly the orientalist knowledge project and the colonial project of domination and extraction were connected” (Appadurai 1997: 115). In fact, what is missing in Said’s Orientalism is exactly the driving force of this article. As in colonial India, the precise and distinctive links between enumeration and government in colonial Taiwan have not yet been specified. This is indeed a perilous oversight, because if we overlook those tools employed by the colonial state to know, understand, study, scrutinize, and investigate the subjected people, we must inevitably give an incorrect interpretation of the way in which the colonized are governed and controlled.
From this perspective, we are in a position to consider an important question posed by Appadurai: “Is there any special force to the systematic counting of bodies under colonial states in India, Africa, and Southeast Asia?” (Appadurai 1997:114). That is, do colonialism and colonization contain some exceptional elements of quantification that cannot be found in other forms of controlling or governing? In other words, is there any difference here between governing the citizen and governing the colonized? Appadurai answers these questions in the affirmative. “I believe that the British colonial state employed quantification in its rule of the Indian subcontinent in a way that was different from its domestic counterpart in the eighteenth century and from its predecessor states in India” (115). Appadurai himself fails to tell his readers exactly what this difference is; nevertheless, such difference can be identified in the study of colonial Taiwan.
Unlike Appadurai, I will link the censuses and agrarian surveys in colonial Taiwan not so much to the development of nationalism, but to political rationalities and governmental technologies. In what follows, I will first demonstrate the colonial government’s infatuation with numbers and statistics. Second, I will discuss Gotō Shimpei’s “biological politics.” In doing so, I will reveal the political rationality behind the government’s effort to know and to govern the colonized. Finally, I will discuss land surveys and censuses launched by the colonial government. They were the key means by which the alien government was able to “make visible domains of life that were once invisible” (Murdoch and Ward 1997:308).
COLONIAL STATE
Before we discuss the Japanese colonial state’s obsession with numbers, let me give a quick review of the nature of the colonial state itself. Anne Phillips calls the colonial state a “facsimile of a state” (1989:11). J. C. Heesterman, a Dutch colonial historian, argues that the public domain of the colonial state “had been taken out and set apart from society” (1987:54). Partha Chatterjee, in The Nation and Its Fragments, therefore asks an important question: “Does it serve any useful analytical purpose to make a distinction between the colonial state and the forms of the modern state?” (1993:14). His answer is affirmative, although he does not have a lot to say in his book about the difference between the colonial state and the modern state. Addressing the same question, Jurgen Osterhammel, a German colonial historian, argues that “the colonial state was no simple extension of the metropolitan political system to overseas possessions, but a political form in and of itself” (Osterhammel 1997:57). As to “why” and “how,” however, Osterhammel, like Chatterjee, has very little to say, except this: “The liberal achievements of Europe did not usually extend to the colonized” (57).
In the introduction to their Tensions of Empire: Colonial Cultures in a Bourgeois World, Ann Laura Stoler and Frederick Cooper (1997) provide a comprehensive review of existing (post) colonial studies. The piece covers more than fifty pages, but one thing is conspicuous by its absence: discussions of the difference between governing a metropole and governing a colony.1 As far as colonial states and political economy are concerned, they seem to call for a rethinking of Paul Rabinow’s statement (1986:259): “We need a more complex understanding of power in the colonies. The state is something we need to know a lot more about” (Stoler and Copper 1997:20). Ironically, then, given its title, “Between Metropole and Colony” also offers us no research agenda for finding the difference in governmentality between metropole and colony.
Lastly, John Comaroff, an American anthropologist, claims that the difference between metropolitan and colonial governance is that “one depended, for its existence, on the ideological work of manufacturing sameness, of engendering a horizontal sense of fraternity; the other … was concerned with the practical management of difference” (1997:16). This is true. But on closer inspection, his theorization of the colonial state, in fact, adds only more abstract elements to the conventional framework.
A paradox: everyone (whether Dutch, German, British, American) agrees that the colonial state, colonial government, or colonial regime is special, but few find it necessary to push their arguments further. Matthew G. Hannah (2000) is an exception, though. In his Governmentality and the Mastery of Territory in Nineteenth-Century America he gives useful clues to thinking about the particulars of colonial government in concrete terms. He concludes that colonial regimes differ from metropolitan governments in three ways: first, “imperial state officials have frequently assumed greater freedom to experiment abroad than at home”; second, “they have also benefited from the relative weakness (at a national scale) of a Western-style civil society among colonized peoples”; and third, “colonial states have often been more bureaucratised and less tolerant of popular resistance” (Hannah 2000:114). He summarizes: “In principle, the kind of governmental thinking characterised by a more careful and solicitous approach to social control has been less prominent in imperial regimes” (114). To caricature somewhat, colonial governments seem good at only two things: unchecked rule by violence and endless bureaucratization.
Hannah’s summary is only a different—more polite and less explicit—way of expressing Abdullahi Ali Ibrahim’s point: “the colonial state … retreated into a governmental ‘infantile disorder’ in which the art of government had still to be invented” (1997:14). For Ibrahim, “a colonial administration could never have been other than a pre-government institution qualified to run a piece of land not a population” (28). His judgment is wholly inapplicable to Taiwan, however. It is clear that we are not in the habit of seeing the colonial state as a “bio-political” government that takes charge of colonized lives.
What is, then, the form of knowledge accompanying the colonial state that “retreated into a governmental infantile disorder”? A passage by Albert Memmi will help clarify. He says, in his The Colonizer and the Colonized, that “the point is that the colonized means little to the colonizer. Far from wanting to understand him as he really is, the colonizer is preoccupied with making him undergo this urgent change” (1965:149). Similarly, D. K. Fieldhouse, one of the most famous colonial scholars in the English-speaking world, observes that as long as colonies provided advantages (strategic bases, raw materials, perhaps merely the satisfaction of imperial ownership), “few cared anything about what happened to them or their inhabitants” (Fieldhouse 1981:44). What emerges from Memmi and Field-house’s descriptions is a deep-seated indifference in the colonizer’s mind, or, to be more precise, inside European colonial governmentality: the colonizer is not interested in knowing and understanding the colonized in any great detail or with any great accuracy.
Yet the Japanese in Taiwan were in fact different. In his Colonial Development and Population in Taiwan, George W. Barclay observes of Japanese’s peculiar way of governing the colonized:
While under Japanese rule Taiwan probably had the distinction of being the most thoroughly inventoried colonial area in the world. Huge compilations of statistics and numerous special surveys were made from year to year. The economy, the terrain, the aboriginal tribes, the mineral wealth, the agricultural output, the industrial production and the foreign trade have all been studied and restudied until there is little to be added to this knowledge unless new evidence is uncovered that is not now available. (Barclay 1954:x)
Of course, having been “the most thoroughly inventoried colonial area in the world” was hardly a matter of pride for the Taiwanese. Why did the Japanese colonizer need such a wealth of knowledge when a more typical attitude was a certain degree of indifference to what was really happening in the colony? This is not the place to judge the arguments of Memmi and Fieldhouse, but to note that, whether or not their arguments hold in the cases of European colonization, they are hardly applicable to the Japanese colonization of Taiwan. Not only did the Japanese show great curiosity and enthusiasm about their new colony, but throughout the colonial period they were determined to know everything about the colonized as scientifically as possible. In short, the Japanese did not suffer from what might be called “the starvation of knowledge” of European colonialists. In fact, no other colonial power in the world invested more energy in knowing the colonized and, in turn, on contemplating and producing an “ideal” colonized—obedient, loyal, productive, diligent, healthy, and useful—than the Japanese. The Japanese colonial empire was indeed an “anomaly,” not only because of its lack of expansive capital, as Jon Halliday suggested (1975), but also because of its peculiar ways of governing the colonized.
COLONIAL NUMBERS
Among all the efforts made by the colonial government to understand its newly gained territory and people, the compilation of statistics was the most important. It must be noted that, according to Stuart Woolf, “the umbilical cord between statistics and modern state in the West was historically not always the case.” For instance, “the earliest mathematically sophisticated utilisation of statistics … was developed in England as a direct by-product of the insurance trade” (1989:588). Mary Poovey, in her A History of the Modern Fact (1998), claimed that the supremacy of statistics was a historically contingent phenomenon.
The situation in colonial Taiwan, however, was different because statistics were at the heart of colonial statecraft from the very beginning. This was mainly due to Japan’s “advantage” of being the “latecomer” among colonial powers. By the time the Japanese occupied the island, “the link between number, fact and good government” (Rose 1999:218) had already become commonplace, and it partly explained why there was hardly any debate in Japan about applying statistics to the governing of Taiwan. According to Ian Hacking, Western society between 1820 and 1840 witnessed a period of what he calls “the avalanche of printed numbers” (Hacking 1982:281). In Taiwan the period of “the avalanche of printed numbers” came much later than in the West. It was the Japanese colonial government that raised the curtain on “the era of enthusiasm for statistical data-collection” in Taiwan (Hacking 1982:281); or, to borrow a phrase from Mitchell Dean (1995), statistics were the episteme of colonial government.
This contention is best verified by The Statistical Summaries of Taiwan Province in the Past Fifty-One Years. This huge book, published in December 1946 by the Chinese Nationalist government after they took over the island at the end of the Second World War, contains more than 1,400 pages and 540 chronological statistical tables. For example, we can find in its table 245 the total number of cows in Taiwan’s farms from 1900 to 1945, while in its table 480 we can find the total number of books in Taiwan’s libraries from 1901 to 1942. The time spans covered by these tables are different, depending upon topic, purpose, availability, and government preference. For instance, numbers on colonial population were not available until after the first census was taken in 1905. The earliest statistics collected, accumulated, and systematically analyzed by the colonial government date back to 1897, two years after the Japanese landed on the island. These were statistics of public health, including the numbers of doctors, hospitals, and epidemic mortality. In this respect Taiwan resembles the European cases, especially that of France, where, according to Theodore Porter, the statistical initiative was taken chiefly by advocates of public health (Porter 1986:28). Yet in Taiwan these statistics were, in fact, an early sign of a special colonial governmentality that from the beginning of the twentieth century began to take the health of the colonized as one of its most important political agendas.
It is crucial to realize that all the numerical data in Statistical Summaries of Taiwan Province were collected by the Japanese colonial government. What the Chinese officials did was simply to compile, translate, and publish them in Chinese. The reaction of the Chinese Nationalist government is worth mentioning. They were stunned by the massive ocean of numbers accumulated by their Japanese predecessors. In the preface to the book there is an interesting statement by the Chinese official responsible for the compilation and translation.
We all know that the Japanese in Taiwan, in order to strengthen their colonial administration, maintained statistics every year; their accomplishment in this respect is even better than in Japan itself. When we came here to take over, we saw that they had a very well organized statistical department, a great number of books and documents, a fine legal record, an immense scope of scientific enterprise, and a proliferation of publications. All of the above can be taken as an indication that the Japanese had great respect for the statistics business. By maintaining rigid and correct statistical data, the Japanese then fulfilled their oppression plot. (iii; emphasis added)
Two major themes dominate the above remarks: (colonial) knowledge and (colonial) power. As the official said, numbers and statistics were nothing other than a colonial weapon. Unlike the other rulers prior to them, the Japanese governmentality was a number-, knowledge-, and truth-producing machine. Seen from this perspective, the historical moment of colonial occupation was thus not only a political rupture, but an epistemological rupture. What was involved in colonialism in Taiwan was, in other words, not only a change of rulers but a change of episteme. The numbers, facts, and knowledge all together generated a wealth of administrative records, which in turn created a sense of certainty and control in the colonial office. It is here that we can understand most clearly the argument of Jonathan Murdoch and Neil Ward that “an archetypal form of governmentality is statistics” (1997:313). Not only are statistics an archetypal form of governmentality, but they are also an archetypal form of colonial governmentality.
Yet why did numbers and numerical representation appeal to a colonial government in particular? Because, first of all, statistics are a mobile, stable, and combinable means of “government at a distance.” They are a means of representation by which Taiwan could be brought back to Japan with minimal dispute. As Poovey puts it, “Quantification was different from qualitative descriptions in being less subject to controversy or dispute” (Poovey 1998:122; italics in the original). Second, statistics are particularly suitable for those who must find a quick and convenient way of “studying mass phenomena profitably without first having to attain detailed knowledge of the constituent individuals” (Porter 1986:6). Confronted by a different race, culture, and language in the colony, the Japanese colonial government obviously faced this task. Third, statistics can ensure that the “confusion of politics could be replaced by an orderly reign of facts” (Porter 1986:27). The harsh colonial encounter in Taiwan created political confusion and tension that could be resolved only by certainty produced at a different level, namely, the statistical.
It takes only a glance at the categorization in the first few pages of Statistical Summaries of Taiwan Province to understand the increasingly “profitable dialectic of information and control” (Said 1978:36) in the colony. The 540 tables have been divided into 24 categories: calendar and weather, land, population, administration organization, justice, agriculture, forestry, marine products, animal husbandry, mineral resources, industry, labor, business, finance, monopoly, monetary, postal and cable service, rail, highway, shipping administration, education, hygiene, philanthropy and religion, and police. As a result, multifarious colonial phenomena ranging from the sunrise times to the total number of sheep; from the forest area to exact colonial population levels; from the total number of factory workers to the number of postal packages sent by the colonial people; from the total number of prostitutes examined by hospitals to the number of temples, monks, nuns, and religious followers—all are documented in the book. In short, every aspect of colonial society was counted. The Japanese showed remarkable and consistent energy, passion, and patience in their effort to keep all sorts of colonial phenomena under tabs. Those astonishingly all-embracing statistical tables represented the fact-hunting nature of the colonial state, and indicate that the colonized in Taiwan were observed, recorded, analyzed, and then governed in every way and on every side. “Nothing was left untouched by the statisticians” (Hacking 1982:280).
The purposes of the colonial state’s determination to numeralize everything in the colony were probably several: some had a utilitarian and pragmatic purpose, such as public health measures, crime prevention, food distribution, taxation, and natural disease prediction. In addition, the Japanese may simply have wished to show off to the West the results of their successful colonial administration. The important point here is that, whatever the purpose, all were tools of colonial government. If nothing escaped the eyes of Japanese colonial statistical analysis, then nothing would be able to escape from the colonial administration. Here I take a view opposite that of Appadurai. In the essay mentioned earlier Appadurai argues:
Although early colonial policies of quantification were utilitarian in design, I would suggest that numbers gradually became more importantly part of the illusion of bureaucratic control and a key to a colonial imaginary in which countable abstractions, of people and resources at every imaginable level and for every conceivable purpose, created the sense of a controllable indigenous reality. (Appadurai 1997:117; italics added)
I have a problem with Appadurai’s word “illusion.” Numbers are of greatest use to colonial social control in terms of their administration rather than their imagination. It is here that we see Said’s great influence on Appadurai (although he leaves a strong anti-Saidian impression). A technique as pragmatic as statistics is predominantly conceived of by Appadurai as part of the general framework of colonial fabrication, illusion, and imagination. That is, Appadurai, like many of Said’s followers, draws too heavily on a conception of number as a cultural product rather than a governmental technology. This is not the place to judge whether Appadurai’s argument on the “illusion of bureaucratic control” is sociologically or historically correct, but I want to point out that, judging from the scope and behavior of the colonial administration, what was created by number in colonial Taiwan was not an illusion but a sufficient authority; not a controllable indigenous reality but a controlled indigenous reality.
BIOLOGICAL POLITICS
What then was the politics, if any, accompanying the proliferation of scientific colonial knowledge? The answer: biological politics. For Gotō Shimpei 
the main advocate of the idea, “colonisation policy was biology” (Tsurumi 1967:107). And “government of a colony cannot go beyond biological laws” (Takekoshi 1907:97). But what are “biological laws”? Gotō gives a first definition in his famous Emergent Proposal for Governing Taiwan, which he offered to the authorities in Japan before he took his position in Taiwan’s government-general in 1898:
Any scheme of colonial administration, given the present advances in science, should be based on principles of Biology. What are these principles? They are to promote science, and to develop agriculture, industry, sanitation, education, communications and police force. If these are satisfactorily accomplished, we will be able to preserve in the struggle for survival and win the struggle for the “survival of the fittest.” Animals survive by overcoming heat and cold, and by enduring thirst and hunger. This is possible for them because they adapt to their environment. Thus depending upon time and space, we too should adopt suitable measures and try to overcome the various difficulties that confront us. (Gotō, cited in Chang and Myers 1963:438)
These principles could be seen as the epitome of the political rationality of colonial government. The entire colonial question for Gotō came down to this: in what way could the Japanese adapt to the harsh and hostile colonial environment? Gotō’s primary solution was to promote science, namely, knowledge. A “regime of truth” was quickly established in the colony, and its task was to forge a link that could not otherwise have been made: mutual enforcement between a disciplinary colonial power and scientific investigation. Successful adaptation to the harsh colonial environment “could be achieved only after careful study of the colony” (Chang and Myers 1963:439). “More than anything else, this meant obtaining sound and relevant information on which to base policy, information to be derived from careful research” (Peattie 1984:84). In short, biological politics for Gotō was a politics of knowledge, of know-how. “Each individual problem demanded a separate investigation” (Tsurumi 1967: 108). Gotō insisted that colonial policy had to be “based on thorough research,” so he “promoted extensive research in many fields” (108).
If taken at face value, there is hardly any similarity, save the etymological, between Gotō’s “biological politics” and Foucault’s idea of “bio-politics.” The former is largely based on social Darwinism while the latter designates a special historical period in which bio-power “brought life and its mechanisms into a realm of explicit calculations and made knowledge-power an agent of transformation of human life” (Foucault 1979b:143). However, on closer inspection, these two terms—invented by different people, in different parts of the world, in different periods of history, and for different reasons—are in fact highly congruent. The secret of this unlikely congruence lies in its relation of power and knowledge. Just as demography found its raison d’être in the era of bio-politics, as Foucault suggests, so too a variety of colonial knowledge came to be formed and institutionalized in the Taiwanese colonial era of “biological politics.” In both cases, we witness an explosion of knowledge and techniques, including reports about inhabitants and resources, and numbers and tables analyzing trends. Such knowledge-producing activities, such techniques of numbers, classifications, and measurement, were by no means outside the colonial power relation. They were rather at the heart of colonial governmentality.
The second meaning of Gotō’s “biological politics” concerns the moral justifications for particular ways of exercising racial politics in the colony, as evidenced in the following statement:
In governing Taiwan, first of all we must investigate scientifically the local customs and institutions, and not adopt any policy that provokes the locals. If we transplant what we have in Japan to Taiwan without any careful investigation beforehand, it is no different from the naive idea of transforming the eyes of flatfish to the eyes of scup. It would be a typical behavior only from someone who does not understand the true meaning of politics. (Gotō, cited in Chen Yen-hong 1986:48)
Like the previous quotation, this is an important statement of the political rationality of colonial government. Not only does Gotō explain to his countrymen how to govern Taiwan—investigate before action—but he also points out the limits of governing the colonized—the eyes of flatfish cannot be transformed to the scup’s. On Japan’s own peculiar “fishery” scale, the scup is far more expensive than the flatfish, so when Gotō said that it would be naive to transform one kind of fish to another, he was actually indicating an insurmountable racial hierarchy, that is, a gap between Japanese and Taiwanese, colonizer and colonized, ruler and ruled, civilized and uncivilized. In short, “biological politics” was a racial politics. It served simultaneously as a political rationality and as a moral discourse in order to make sense of and/or justify the colonizer’s particular ways of governing the colonized.
THE LAND SURVEY
Driven by the principle of “biological politics,” in September 1898 the colonial government established the “Temporary Land Survey Bureau” in order to carry out a thorough land survey of colonial Taiwan. There had been some cadastral surveys conducted by the Qing government; according to John Shepherd, the first could be dated back to 1728 (1993:138). But the colonial government’s land survey was special in that it was the first undertaken by any ruler on the island seeking to understand Taiwan and its inhabitants scientifically, and can therefore be seen as the beginning of the singular colonial power/knowledge. The duties of the land survey bureau were “to survey the whole island, to estimate the produce, to make maps, ledgers, and land registers” (Takekoshi 1907:126). It was written in the “Cadastral Regulations in Taiwan,” enacted in 1901, that “apart from rail and sewage, each and every piece of land must be given a number, and thus be registered” (Chen Chao-ming 2000:47).
This massive survey was divided into three parts: land registration, trigonometrical surveys, and topographical surveys. It was the aim of the colonial government that each and every inch of land in the colony be investigated and thus regisered.2 It took the Japanese seven years in total (1898–1905) to complete this survey, during which more than 5,225,000 yen were spent and more than 1,670,000 personnel were mobilized and involved (Tu 1981:37). The colonial government investigated a total of 777,850 chia of land, produced 37,869 maps of villages and cities, compiled up to 9,610 cadasters, and accumulated 5,624 volumes of land rent registration (Chen Yen-hong 1986:73). The massive expenditure and mobilization were deemed by the colonial government to be justified by the increased understanding of, and the correspondingly increased control over, the land and its inhabitants.
Much has been written about the contribution of the land surveys to Japanese colonial rule in Taiwan. Takekoshi, for instance, observed: “The whole area of Formosa has now been accurately measured, its hills and valleys carefully surveyed, and its productive capacity ascertained. This is of untold value to the military and civil administration” (Takekoshi 1907:129). And: “The land survey business had played a decisive role in helping the government-general in Taiwan to control the Han people as well as consolidate its foundation for governing” (Fujii 1997:140).
Yanaihara Tadao 
’s understanding of land survey is the most comprehensive. The land survey in Taiwan served at least three major purposes, he said: “to promote security and peace by clarifying the topographical situation, to increase the land-tax revenues by finding unregistered or ‘hidden’ lands and simplifying land-ownership, and to make transactions of land safe and secure by ascertaining and simplifying land-ownership” (Yanaihara 1985: 17; see also Tsurumi 1967:112). In short, the land survey served a threefold purpose: first, to ascertain and understand the topography of the colony; second, to clarify land ownership, and thus to build up a clear and official relation between the land and its inhabitants, as well as increase land revenues from taxation; third, to pave the way for the later expropriation in which the state’s capital investment (mainly in sugar factories) could claim monopoly status over the land of the colonized.
Yanaihara’s points are worth dwelling on for a moment. According to him, the first purpose of the land survey was to “promote security and peace.” This had been achieved simply by “clarifying the topographical situation.” During their first few years of colonization, the Japanese had been deeply frustrated by rebellious Taiwanese bandits and guerrillas who had been forced to take refuge in rural villages, where the Japanese had real difficulties in finding them, let alone in enforcing the peace. But after the survey, the guerrillas could run but they couldn’t hide. The colonial government, in the process of investigation, had drawn numerous maps on which the tribes, villages, rivers, and mountains were all clearly depicted. This was exactly the moment, to put it in Latour’s phrase, when “the implicit geography of the native is made explicit by geographers; … the fuzzy, approximate and ungrounded beliefs of the locals are turned into a precise, certain and justified knowledge” (Latour 1987:216; italics in the original). His implication is clear enough: if everything was shown on maps, then nothing was escaping the colonial government. Of course, the pacification of the Taiwanese guerrilla depended largely on the superior military force of the Japanese colonizer, but the land survey performed a role that guns and tanks alone could not.
So, beyond Said, what is at issue here is by no means a question of “imaginative geography and its representation” (Said 1978:49). It is instead a practical question of obtaining a “real” geography. The land survey gives us a clear example of the coincidence and mutually enforcing relationship among geography, mapping, knowledge, and government. On this point, just consider Takekoshi’s personal experience during his visit to Taiwan. Due to its importance, I quote his description at some length below:
When I inspected the different maps and ledgers in the Bureau of Surveys, and had the pleasure of seeing the officials at work, I could not but admire the vastness of the undertaking, and the scientific way in which it was being carried through. There is no town or village on the island the exact position of which has not been determined, no field or plantation, however small, which will not be found upon one or other of the prepared maps; in short, precise information is at once obtainable as the size, etc., of any piece of ground down to the smallest rice field. When we arrived at the godown reserved for ledgers, Baron Gotō asked one of the officials to show me the map of a certain village. The officials referred to the index, and speedily spread a map out before us. … On the village map we saw rice fields, tea plantations, brooks, hills, and woods, all drawn with a precision trigonometrical survey alone can secure. … With one of these maps and a pencil it was easy at once to ascertain the size of each village. In that godown it would be almost as easy to examine the physical contour and geographical features of Formosa as it is to study the palm of one’s own hand. (Takekoshi 1907:128–129)
Notwithstanding Takekoshi’s exaggeration and overconfidence about Japanese colonial rule, the accomplishment of the land survey offered the alien government the opportunity to know the territory and its productivity without overlooking any detail; and, more important, the opportunity to exercise its governing power at minimal cost: a map and a pencil. Takekoshi’s confidence perhaps sounds bizarre, but it is telling. It underlines the fundamental determination of a colonial government obsessed with complicated enumeration and detailed accuracy in measurement. The result: the colonizers were able to examine the colony as easily “as it is to study the palm of one’s own hand.” This is what Shih Tien-Fu 
has described as “govern the land by maps, and people by land” (Shih 2000:13; 2001:4).
Timothy Mitchell comments in Colonising Egypt that “colonial power required the country to become readable, like a book” (1988:33). The Japanese set for themselves a “higher” standard; in a similar context, Edney’s argument in Mapping an Empire is intriguing. In defining the scope and nature of his study of the surveys and maps in British India, Edney remarks:
It is a study of how the British represented their India. I say “their India” because they did not map the “real” India. They mapped the India that they perceived and that they governed. … India could never be entirely and perfectly known. The British deluded themselves that their science enabled them to know the “real” India. But what they did map, what they did create, was a British India. (Edney 1997:14–15; italics in the original)
Inevitably, an important question has to be raised: what then is the “real” India? Edney does not offer us the answer. My aim here is, of course, not to judge whether Edney’s interpretation of British colonialism is correct. Nor is it my aim to squabble with the postcolonial theorists over the meaning of the “real.” My concern is with whether maps can achieve their governmental ends. Takekoshi and perhaps most of the administrators in Taiwan would have been sure that they could. When compared with Edney’s phrase “India could never be entirely and perfectly known,” Takekoshi’s metaphor seems highly significant. In the eyes of the colonial government Taiwan was by no means an imaginary objectification. In fact, throughout the whole colonial period the colonial government hardly invented or created an imaginary Taiwan in order to “override” the real Taiwan. Quite the contrary: the whole purpose of this colonial “cadastral politics” was to replace the imaginary colonial contours with the real things, to replace inaccurate numbers with the correct ones.
Apart from its contribution to ensuring peace, the land survey also fulfilled a financial function in the colony. It was an effort to ascertain the relation between the inhabitants and their land: Who owns it? Who is responsible for paying taxes on it? A passage from George Kerr in his Formosa: Licensed Revolution and the Home Rule Movement, 1895–1945 gives us a clear description of this function:
The inaccuracies of nineteenth-century Chinese official records were now revealed in their true dimensions. According to the Chinese land registers surrendered to the Japanese, a total of 867,000 acres were yielding revenue, whereas Gotō’s men discovered that the figure should have been 1,866,000 acres. This means that when the new records were complete, tens of thousands of Formosan landholders were compelled to pay taxes on land never before taxed. (Kerr 1974:86)
Attention should first be paid to Kerr’s statement that the numbers, records, and registration of land under the Qing government were inaccurate, and that these inaccuracies were not modified and corrected until the Japanese colonial government launched its massive land survey. The Qing government in Taiwan, in other words, failed to understand what is now seen as one of the most taken-for-granted motifs of political rationality: to govern territories, one must know them. Table 2.1 gives a rough idea of the improved cadastral enumeration and classification system introduced in Taiwan by the Japanese colonial government.
It is here that we can see most clearly the operation of “governing by numbers” (Rose 1991). What changed, of course, was not the actual area of the land itself, but the techniques of measurement and, more important and fundamental, the governmentality of the ruler. Or, in the words of Bernard Cohn, “they [the colonizer] … invaded and conquered not only a territory but an epistemological space as well” (Cohn 1996:4). In fact, as the Taiwanese case demonstrates, the colonial territory actually became “bigger” as a result of the change in the epistemological space.
The huge changes illustrated in the above table were a consequence of the general project of colonial enumeration and classification. The Japanese “discovered” that the revenue-yielding area was much larger than had been thought before, and after the survey many more Taiwanese people were forced to pay taxes on their land. Revenue greatly increased, which helped ease the financial difficulties facing the colonial government. According to Tu, total land revenue in 1903 was about 920,000 yen, but in 1905 the figure surged to 2,980,000 yen (Tu 1991:39)· However, to suggest that the sole purpose of levying more land taxes was to increase the colonial state’s revenue is to misunderstand the essence of the land survey and, more generally, the whole project of colonial enumeration and classification. The land survey project was not so much a method of finding new sources of land revenue as it was a method of overcoming and controlling potential threats to the colonial government: lack of knowledge about the colony (its size, shape, fertility, productivity), uncertainty over the land’s relation to its inhabitants, and a resulting interstice between the colonial power and its people. In short, what was created by the land survey was not an illusion (as suggested by Appadurai concerning the Indian colonial situation) but a certainty, a controlled cadastral reality.
TABLE 2.1 Results of the Japanese Land Survey of Taiwan
Source: adapted from Chen Yen-hong 1986:72.
Yanaihara’s third purpose of the land survey—simplifying land ownership—is perhaps the most complex. There is no room to discuss this issue in great detail here. The landholding system in Taiwan before Japanese colonization was indeed very complicated.3 This was mainly the result of the intriguing triangular relationship among plains aborigines, Chinese settlers, and the state under the Qing’s frontier policy (see Shepherd 1993). The multitiered land right in Taiwan can be described in the following way: “The tenant worked the land and paid an annual fixed rent in grain called shosō 
(small rent) to a holder who possessed the right to the shosō. The holder in turn paid an annual grain rent to the holder of taisō (great rent), who had originally received the land in trust from the government, an aborigine, or an aborigine tribe. The taisō-holder was the acknowledged ‘owner,’ insofar as ownership could be said to have existed” (Tsurumi 1967:113). In short, the majority of taxes in the agricultural sector were not paid directly from actual landowners to the colonial government. As far as the colonial administration was concerned, this was something that had to be changed.
According to Tu’s research on land ownership in Taiwan, the class structure of the agricultural population in 1905 was as follows: 38,000 taiso-holders, 300,000 shoso-holders, and 750,000 tenants who actually worked the land. In 1904 the colonial government abolished the right of the taiso-holder, and gave to the shoso-holder the formal and legal right to the land. The taiso-holders were compensated with government loan bonds. Thereafter, the new landholding system in Taiwan was established: the shoso-holders paid their taxes directly to the government-general without the intermediation of taiso-holders. Once the question of ownership was legally clarified, the purchase of land was legally protected and the way was paved for the massive investment of Japanese capital and the establishment of its monopoly. The Japanese capitalists began to realize that all of the colony’s land was now at their disposal. This is why Yanaihara claims that “the land survey became the necessary precondition and basic foundation for Taiwan’s capitalization” (Yanaihara 1985:18).
In terms of landholding relations, then, the old order had been destroyed in Taiwan. (The right of the taiso-holders had been annulled, and in place of them now were just the shoso-holders, more numerous but each with much less land.) With no real delay the old order had been replaced by the new. This new order was established via a regime of cartography and statistics. But, whatever the form of ownership, for the vast majority of farmers in Taiwan throughout the colonial period the general conditions remained the same. The only difference was that they now found themselves governed by a far more scientific, efficient, and complicated system of domination since their land had been completely and inexorably scrutinized by the foreign ruler.
THE CENSUS
The colonial government in Taiwan spared no expense or energy in finding out everything it possibly could about the colonial territory. It also spared no expense or energy in finding out everything about its subjects—namely, the population of the colony. In contrast to the Qing government, which governed Taiwan for more than two and a half centuries without taking a census and therefore without knowing exactly how many people lived under the its regime, the Japanese colonial government was very quick to understand a “common sense” dictum of twentieth-century political rationality: “the people should be counted” (Hacking 1982:290).
Like the land survey, the census represented another aspect of the general project of colonial enumeration and classification. It was therefore a genuinely political calculation and political measure. According to Barclay, among all the statistics collected by the colonial government, “the numerical data relating to the population are the best of all, since their collection was most directly linked to the very effective system of control imposed by the Japanese” (Barclay 1954:x). Thanks to the census, for the first time in Taiwanese history the inhabitants of the island had been transformed into a governable population. Afterward, the natives became identifiable, classifiable, and describable by government. Cohn has noted that “the taking of the census was based on administrative necessity” (Cohn 1987:242). And he is absolutely correct in saying that “the history of the Indian census must be seen in the total context of the efforts of the British colonial government to collect systematic information about many aspects of Indian society and economy” (231). Quite clearly, although Taiwan was colonized by a different colonizer, the reason for the census-taking in the colony was no different from that in India.
In his Colonial Development and Population in Taiwan, George Barclay points out that:
Taiwan is the only place where it is possible to study the processes of change of a Chinese population over a substantial period of time on the basis of excellent data. … Under Japanese rule an unusually accurate statistical record was maintained in Taiwan from 1905 to 1943, providing an unbroken series of data that is longer, and in some respects richer, than that for Japan proper. (Barclay 1954:v; emphasis added)
In using the words “unusually accurate,” he appears to imply that there is no reason for any given colonizer to extract such a great deal of accurate information from its colonized. Hidden in Barclay’s argument is a long-term stereotype about colonization and colonialism—that there is no need to know the colonized people—but this need not concern us here. Here I focus on why the Japanese colonizers, against the norm, were so keen on knowing the colonial population. This is partly explained by Barclay’s subsequent sentence, suggesting that the data on the colony is “in some respects richer than that for Japan proper.” Such an assertion may seem surprising at first, but if we consider that the earliest census was actually taken in the colonial context (Hacking 1990:17), then it is more plausible. As Rose points out, “by and large the first to be enumerated were the most dominated—the inhabitants of the colonies” (Rose 1999:215). It would be wrong to infer that the same infatuation with censuses in Japan did not exist. Because the colonized are more difficult to tame and to govern, richer data and greater knowledge are inevitably required by the colonizer.
After three years of preparation the first census in Taiwan was taken in October 1905 (ten years earlier than the first census in Japan itself), involving 842 supervisory staff, 1,339 assistants, and 5,224 census-takers (or “enumerators”) (Tsurumi 1967:113). These numbers have little meaning, though, until we consider that the population to enumerator ratio in this census was as low as 582 (3,039,751÷5,224). Each enumerator in Taiwan was responsible for counting 582 people on average. Compare this to the situation in nineteenth-century America in which, according to Hannah, “enumeration districts could engross as many as 20,000 people” (Hannah 2000:118). Clearly, the ratio in colonial Taiwan was very low, and this is highly significant. While the land survey was a starting point of the special colonial power/knowledge, the first census was no doubt the starting point in Taiwan of what has been called by Foucault “the governmentalisation of state” (Foucault 1979a:20) in Taiwan. It was, in other words, a historical moment, when the government’s knowledge of the population was transformed from estimation and imagination to calculation and classification.
In 1915 the colonial government organized the second census in Taiwan, and one took place every five years thereafter until 1940. During the fifty years of colonization, people in Taiwan were counted seven times by the alien government! The census, which in turn led to a new kind of calculating rationality in colonial Taiwan, made it possible to know the number, identity, sex, occupation, age, marital status, language, birthplace, education, wealth, and production of the colonized. This series of censuses, described by Barclay as a “complete monument to their [i.e., the colonizers’] enumerative capacities,” comprises “the only trustworthy information that we have about the numbers and kinds of people in Taiwan at any time” (Barclay 1954:10). Again, as he remarks, the Taiwanese censuses
all maintained uniformly high standards in the completeness of reporting, in the consistency of defining ethnic classes, and in the accuracy of information such as age reporting. In these respects they often ranked in quality well above the more advanced censuses of the Western world. The first two censuses, in 1905 and 1915, were carried out before a genuine enumerative census was ever conducted in Japan Proper. (Barclay 1954:10)
In order to highlight the extraordinary effort of the colonial government and the satisfactory results of the census, it might be helpful here to take the colonial census in India as a comparison. Cohn notices that the Indian censuses of 1871–1872 were not satisfactory and that “such imperfections, both in administration and in conception, developed that not much reliance was put in the census at the time” (Cohn 1987:238). In contrast to the Indian census organized by the British, the Taiwanese ones—right from the first—were highly accurate. According to Cohn, one of the main reasons for the inaccuracy of the Indian census was that the British colonial state had no correct knowledge and information about the colonial territory. Cohn says, “for twenty years starting with the census of 1871–2, there were difficulties in Bengal in getting an accurate list of villages or, for that matter, even defining what a village was” (239).
Yet thanks to the land survey completed in 1905 (not coincidentally, the year of the first census-taking), the Japanese colonial state in Taiwan was free of the uncertainty that afflicted the British census-taker in India. Perhaps nothing could be more remote from the ethos of the Japanese colonial governmentality than to imagine that a colonial ruler, after governing the colony for more than a century, still lacked knowledge about the subject territory. Not only did they know what a village was, but in most cases there was a map at their disposal. According to Hannah, the land survey was a prerequisite for census-taking: “It is impossible to undertake an accurate census unless there is some geographical framework on which to define and precisely locate enumeration districts which exhaust the territory without any overlap” (Hannah 2000:118). Perhaps it is now impossible to identify the exact person in the colonial government responsible for making the decision to organize the land survey prior to the census, but what is clear is that, at the turn of the century, the colonial government began, perhaps only gradually, to build up a power/knowledge network in which the scientific knowledge obtained by different institutions for different purposes was nevertheless highly complementary.
Of course, such high-quality data collection in the census was not a result simply of the curiosity of the colonizer; it stemmed rather from the need for registration, administration, and an effective system of control. One of the most important purposes of these censuses was to make a register of the population under the colonizer’s control. And “this was undertaken for purposes of regulation, and not to satisfy disinterested curiosity about the inhabitants” (Barclay 1954:10). Everyone in the colony had to be clearly identified, by name, age, sex, family, marriage, address, ethnicity, occupation, and so on. This was another act of the colonizer aimed at eliminating uncertainty in the colonial situation. All of the colonized could fit into one of the categories (male/female; Taiwanese/Chinese, etc.) on the census form. On this point, Benedict Anderson notes something important for us. In his second edition of Imagined Communities he writes:
One notices … the census-makers’ passion for completeness and unambiguity. Hence their intolerance of multiple, politically “transvestite,” blurred, or changing identifications. Hence the weird subcategory, under each racial group, of “Other”—who, nonetheless, are absolutely not to be confused with other “Others.” The fiction of the census is that everyone is in it, and that everyone has one—and only one—extremely clear place. No fractions. (Anderson 1991:166–167)
What is described here by Anderson is an important aspect of the colonial subject, an aspect that disappointingly is more often than not neglected by contemporary (post) colonial theorists. When we are considering the keywords used by Anderson—”completeness,” “unambiguity,” “extremely clear place,” and “no fractions”—we see how different they are from (post) colonial jargon such as “mimicry,” “ambivalence,” or “hybridity.” The image of the colonized in the census is indeed in conflict with that drawn from other colonial discourses such as novels. The census represents a colonial governmentality, whereas other discourses perhaps represent colonial hybridity. But what is important is that, confronted by such a power-knowledge combination as the census, nothing in the colony was ambiguous, and nobody in the colony had the privilege of remaining anonymous. Given this, one would wonder where the idea of “ambivalence” or “the mark of the plural” (Memmi 1990:151) comes from, or how it is that “mimicry emerges as one of the most elusive and effective strategies of colonial power and knowledge” (Bhabha 1994: 85). To what extent can it be true that “both coloniser and colonised are in a process of miscognition” (Bhabha 1994:97)? In what sense is it that “the colonial authority is best understood and resisted through its ambivalence” (110)? And how can the colonized be “a chaotic, disorganized, and anonymous collectivity” (Hartsock 1990:160)?
One minor episode from 1915 can best describe the impact of the census-taking and thereby the whole power/knowledge complex in the colony. On October 1, 1915, when the second census was due to be taken, all trains before midnight were jammed with panicking Taiwanese passengers: they desperately wanted to rush back to the places where they had originally registered before the census started. At that time, there was a rumour spreading widely among the natives: if they failed to show up at home to answer the census-taker’s questions, they would be deemed and classified as anticolonial “bandits,” with consequences that are not difficult to imagine. So people tried everything that they could to rush home. Those jammed trains were a sign of a certain form of governmentality. And those panicking nocturnal travelers were nothing but objects of knowledge who were determined to meet this singular “regime of truth” punctually, and thus subjected themselves to the successful administration of colonial power/knowledge.
One commentator in Taiwan observes, “When these people were rushing on their way home in the night, they had unwittingly entered a brand new world of time” (Lu 1998:56). Lu is right to point out the temporal dimension of colonial discipline, but it makes more sense to say that the people rushing home had unwittingly entered a new world of colonial governmentality in which knowledge had become an effective technique of power, and power became increasingly reliant on accurate systems of knowledge. As Rose and Miller point out, “making people write things down and count them—register births, report incomes, fill in censuses—is itself a kind of government of them, an incitement to individuals to construe their lives according to such norms” (Rose and Miller 1992:187).
In 1901 the Japanese colonial government organized the “Temporary Bureau for Investigation of Traditional Customs,” which was arguably “the first investigation institution of human sciences in Taiwan” (Tu 1991:34). The results of its investigations, according to some commentators, were the first organized books about the tradition and customs of the Han people (Chen Chi-liu 1980:7). In 1905 (when the land survey was completed), ten years after the Japanese colonizer occupied Taiwan, the first accurate, complete, and detailed map of Taiwan was drawn up by the colonial government. Also in 1905, the Japanese colonial government held the first census in Taiwan and estimated the total population in Taiwan to be 3,039,751 (Taiwanese: 2,973,280; others: 66,471). There were on average 6.2 people per household; the ratio of men to women was 112.7:100; and there were 85 people per square kilometer (Chen Yen-hong 1986:75). For the first time, the island’s rulers knew the exact numbers of its population. “The Manchu [Qing] government did not even know the number of its subjects, far less their behaviour with respect to births and deaths. The colonial regime founded by the Japanese started a thorough system of vital statistics only after a decade of control” (Barclay 1954:145). In 1908 the Japanese colonial government began its study of the aborigine people in Taiwan. It found nine main groups: Taiyal, Saisett, Ami, Bunun, Tsuou, Piyuma, Tsarisen, Paiwan, and Yami. For the first time in Taiwan’s history the aborigine people on the island had all been put under scientific study and clearly identified. And, in 1911, some Japanese botanists published a book entitled The Dictionary of Vegetation in Taiwan, in which for the first time in Taiwanese history the plants of the island—which have no direct governmental meaning—were clearly identified, described, and categorized in print. More examples could be given.
So what kind of conclusion can be drawn from these “firsts”? Seen individually, these events are no more than five minor episodes drawn from a long colonial history, but if seen from a panoramic point of view, these separate events have a previously unseen feature. They all demonstrate the close relation between power and knowledge, between political rationality and governmental technologies; more precisely, they are all products of a certain form of colonial “governmentalization of the state.”
Yet the peculiarity of the Japanese form of colonial knowledge should not be exaggerated. The Japanese were by no means the only colonizer that launched massive colonial surveys, nor had the Japanese “invented” the link between science and colonial administration. “Knowing the colonized” is also an essential element of other colonial governments. Zaheer Baber, for instance, claims that “British India proved to be a good testing ground for a number of experiments in the application of science and technology by the colonial state. What followed in nineteenth-century India was one of the largest state-sponsored scientific research and development activities undertaken in modern time” (Baber 1996:185–186). I do not claim that the Japanese had the most “accurate” colonial knowledge in the world. Nor do I deny entirely the “scientific” knowledge of the European colonial powers. What is proposed here is a rather simple theoretical point: the form of colonial knowledge—whether knowledge is imaginary, and whether colonial knowledge entails an imaginary governance—is not an epistemological question; it is above all a governmental question. In fact, it is those concrete governmental technologies—police and hokō system, for example—that determine the scope, nature, and “usefulness” of colonial knowledge.4
In 1978 Edward Said asked a famous question in his Orientalism: can there be a true representation of anything? Said answered in the negative, at least as far as Islam was concerned. He asserted that “a representation is eo ipso implicated, intertwined, embedded, interwoven with a great many other things besides the ‘truth,’ which is itself a representation” (1978:22). I agree. Thanks to Said, we now understand that every form of colonial representation is bound up with the colonial power itself. However, Said’s disillusionment has ironically created a new illusion: after Orientalism, colonial knowledge and colonial discourse in general became asymmetrically conceived of as a matter of imagination and mis-representation, as if the whole scope and purpose of colonial knowledge and colonial discourse were simply to distort the colonized. In the words of David Ludden, “the particulars that connect histories of imperialism and knowledge are missing” (Ludden 1993:250). This is an inadequate interpretation of colonial knowledge. I am not denying, of course, Said’s contribution, nor am I suggesting that colonial discourse contains no element of mis-representation and distortion. What I am suggesting, rather, is that we should stop naively equating a colonial government’s “regime of truth” with an enterprise of distortion. To say that colonial power is always based on imagination and illusions, and always attempts to distort and mis-represent the colonized, is to conceal the most subtle technologies and techniques of colonial government.
The land survey and census discussed in this article both had their importance for the colonial governmentality. In the land survey, the colonizer showed its thirst for knowledge of its newly gained territory. Maps and numbers were produced to represent what was to be governed. Similarly, in the census we witness one of the most accomplished feats of data-collection from any population, at any given time, anywhere in the world. After being carefully counted, not only were the natives construed to be a governable entity but they began to enter a sphere of governmental calculation.
So, can there ever be a true representation of anything? As far as governmentality is concerned, Said obviously asked the wrong question and then searched for his answer in the wrong places: novels, operas, and so on. To understand the way in which the colonized were actually governed, we have to answer the question posed by Dean: what forms of thought, knowledge, expertise, strategies, means of calculation, or rationality are employed in practices of governing? (Dean 1999:31). We should talk less about how representations of colonial subjects have been distorted, and more about how the alien government came to be able to know the colonized and their colony as well as they knew the palms of their own hands.
NOTES
1. Another significant omission is Japanese colonialism. Japan as a colonial power is not mentioned in this introduction.
2. The land survey focused only on the west side of Taiwan, and left the east coast and the mountain area untouched. It was not until the forest survey started in 1910 that the entire island was thoroughly investigated.
3. For a discussion of this point, see Ka 1995.
4. For a detailed discussion on the relation between knowledge and governmental technologies in colonial Taiwan, see Yao 2002.
REFERENCES
Anderson, Benedict. 1991. Imagined Communities: Reflections on the Origin and Spread of Nationalism. London: Verso.
Appadurai, Arjun. 1997. “Number in the Colonial Imagination.” In Modernity at Large: Cultural Dimensions of Globalization. Minneapolis: University of Minnesota Press.
Baber, Zaheer, 1996. The Science of Empire: Scientific Knowledge, Civilization, and Colonial Rule in India. New York: State University of New York Press.
Barclay, George. 1954. Colonial Development and Population in Taiwan. Port Washington: Kennikat Press.
Bhabha, Homi. 1994. The Location of Culture. London: Routledge.
Chang Han-yu and Ramon Myers. 1963. “Japanese Colonial Development Policy in Taiwan, 1895–1906: A Case of Bureaucratic Entrepreneurship.” Journal of Asian Studies 22.2.
Chatterjee, Partha. 1993. The Nation and Its Fragments: Colonial and Postcolonial Histories. Princeton: Princeton University Press.
Chen Chao-ming 
. 2000. “
. [Some preliminary accounts of Taiwan’s land survey under Japanese rule].” 
[Modern land administration] 228.
Chen Chi-liu 
. 1980. “
[Anthropological studies in Taiwan].” Chinese Cultural Renaissance Monthly 11.6.
Chen Yen-hong 
. 1986. “
[Gotō Shimpei’s colonial policies in Taiwan].” M.A. thesis, Japanese Research Centre, Dan-Chiang University.
Cohn, Bernard. 1987. “The Census, Social Structure, and Objectification in South Asia.” In An Anthropologist Among the Historians, and Other Essays. Delhi: Oxford University Press.
———. 1996. Colonialism and Its Forms of Knowledge: The British in India. Princeton: Princeton University Press.
Comaroff, John. 1997. “Reflections on the Colonial State, in South Africa and Elsewhere: Factions, Fragments, Facts, and Fictions.” Bulletin of the Institute of Ethnology, Academia Sinica 83:1–50.
Dean, Mitchell. 1995. “Governing the Unemployed Self in an Active Society.” Economy and Society 24.4:559–583.
———. 1999. Governmentality: Power and Rule in Modern Society. London: Sage.
Edney, Matthew. 1997. Mapping an Empire: The Geographical Construction of British India, 1765–1843. Chicago: University of Chicago Press.
Fieldhouse, D.K. 1981. Colonialism, 1870–1945: An Introduction. London: Macmillan.
Foucault, Michel. 1979a. “On Governmentality.” Ideology and Consciousness 6:5–22.
———. 1979b. The History of Sexuality: An Introduction. London: Allen Lane.
Fujii, Shizue 
. 1997. 
: 
(1895–1915) [The aborigine policy of the Taiwan government-general in the period of Japanese dominance, 1895–1915] Taipei: Wen-Yin-Tong.
Hacking, Ian. 1982. “Biopower and the Avalanche of Printed Numbers.” Humanities in Society 5:279–295.
———. 1990. The Taming of Chance. Cambridge: Cambridge University Press.
———. 1991. “How Should We Do the History of Statistics?” In Graham Burchell, Colin Gordon, and Peter Miller, eds., Foucault Effect: Studies in Governmentality. Chicago: University of Chicago Press.
Halliday, Jon. 1975. A Political History of Japanese Capitalism. New York: Pantheon.
Hannah, Matthew. 2000. Governmentality and the Mastery of Territory in Nineteenth-Century America. Cambridge: Cambridge University Press.
Hartsock, Nancy. 1990. “Foucault on Power: A Theory for Women?” In Linda Nicholson, ed., Feminism/Postmodernism. London: Routledge.
Heesterman, J.C. 1978. “Was There an Indian Reaction? Western Expansion in Indian Perspective.” In H.L. Wesseling, ed., Expansion and Reaction: Essays on European Expansion and Reaction in Asia and Africa. Leiden: Leiden University Press.
Ibrahim, Abdullahi Ali. 1997. “Tales of Two Sudanese Courts: Colonial Governmentality Revisited.” African Studies Review 40.1:13–33.
Ka Chih-ming. 1995. Japanese Colonialism in Taiwan: Land Tenure, Development, and Dependency, 1895–1945. Boulder, Co.: Westview Press.
Kerr, George. 1974. Formosa: Licensed Revolution and the Home Rule Movement, 1895–1945. Honolulu: University Press of Hawaii.
Latour, Bruno. 1987. Science in Action. Cambridge: Cambridge University Press.
Lu Shao-li 
. 1998. 
[Whistle from the sugarcane factory: The transition of time cognition and rhythms of social life in Taiwan under Japanese rule, 1895–1945]. Taipei: Yuong-Liu Publications.
Ludden, David. 1993. “Orientalist Empiricism: Transformations of Colonial Knowledge.” In C.A. Breckenridge and P. Van Der Veer, eds., Orientalism and the Postcolonial Predicament: Perspectives on South Asia. Philadelphia: University of Pennsylvania Press.
Memmi, Albert. 1965 [1990]. The Colonizer and the Colonized. London: Earthscan.
Miller, Peter, and Nikolas Rose. 1990. “Governing Economic Life.” Economy and Society 19.1:1–31.
Mitchell, Timothy. 1988. Colonising Egypt. Berkeley: University of California Press.
Murdoch, Jonathan, and Neil Ward. 1997. “Governmentality and Territoriality: The Statistical Manufacture of Britain’s ‘National Farm.’” Political Geography 16.4:307–324.
Osterhammel, Jurgen. 1997. Colonialism: A Theoretical Overview. Trans. Shelley L. Frisch. Princeton: Markus Wiener Publishers.
Peattie, Mark. 1984. “Japanese Attitude Towards Colonialism, 1895–1945.” In R. Myers and M. Peattie, eds., The Japanese Colonial Empire, 1985–1945. Princeton: Princeton University Press.
Poovey, Mary. 1998. A History of the Modern Fact: Problems of Knowledge in the Science of Wealth and Society. Chicago: University of Chicago Press.
Porter, Theodore. 1986. The Rise of Statistical Thinking, 1820–1900. Princeton: Princeton University Press.
Rabinow, Paul. “Representations Are Social Facts: Modernity and Post-Modernity in Anthropology.” In J. Clifford and G. Marcus, eds., Writing Culture: The Poetics and Politics of Ethnography. Berkeley: University of California Press.
Rose, Nikolas. 1991. “Governing by Numbers: Figuring Out Democracy.” Accounting Organizations and Society 16.7:673–692.
———. 1999. Powers of Freedom: Reframing Political Thought. Cambridge: Cambridge University Press.
Rose, Nikolas, and Peter Miller. 1992. “Political Power Beyond the State: Problematic of Government.” British Journal of Sociology 43:173–205.
Said, Edward. 1978 [1991]. Orientalism: Western Conceptions of the Orient. London: Penguin.
Shepherd, John Robert. 1993. Statecraft and Political Economy on the Taiwan Frontier, 1600–1800. Stanford: Stanford University Press.
Shih, Tien-fu 
2000. “
[The lineage constitution of traditional settlement in Taiwan: A research method].” 
[The Yilan journal of history] 47:3–47.
———. 2001. “
[Spatial structure and the development mechanism of Taiwan’s territorial society during the Japanese era: The case of Ming-hsiung].” 
[Taiwan historical research] 8.1:1–39.
Stoler, Ann Laura, and Frederick Cooper. 1997. “Between Metropole and Colony: Rethinking a Research Agenda.” In L. A. Stoler and F. Cooper, eds., Tensions of Empire: Colonial Cultures in a Bourgeois World. Berkeley: University of California Press.
Statistical Office of Taiwan’s Administrative Bureau 
. 1946. 
[Statistical summaries of Taiwan province in the past fifty-one years]. Taipei.
Takekoshi, Yosaburo. 1907. Japanese Rule in Formosa. Trans. George Braithwaite. London: Longmans, Green and Co.
Tsurumi, Patricia. 1967. “Taiwan Under Kodama Gentaro and Goto Shimpei.” In Albert Craig, ed., Papers on Japan 4. Harvard East Asian Research Center, Cambridge, Mass.
Tu Zhaoyan 
1991. 
[Taiwan under Japanese imperialism]. Taipei: Jen Chian Publications.
Woolf, Stuart. 1989. “Statistics and the Modern State.” Comparative Studies of Society and History 31.1:588–604.
Yanaihara Tadao 
. 1985. 
[Taiwan under imperialism]. Trans. S. W Cho. Taipei: Pamier Books.
Yao Jen-to. 2002. “Governing the Colonised: Governmentality in the Japanese Colonisation of Taiwan.” Ph.D. diss., University of Essex.