Taiwan Under Japanese Colonial Rule, 1895

This paper examines how the colonial administration was shaped in the specific context of Taiwan under Japanese rule. From the beginning of Japanese rule in Taiwan, the colonial government mapped, reworked, and created a series of organizations based on natural villages, and actively sought to integrate these colonial spaces, themselves structured and overlapping, into the hierarchy of the colonial administrative mechanism. The Japanese colonial bureaucracy imposed a discipline of order on Taiwan, and by the 1930s wartime concerns reshaped this order, thus turning Taiwan into not only a disciplined but also a disciplinary society.

To decipher the operational network of the bureaucracy in the government general of Taiwan (Taiwan sōtokufu ), I rely heavily on legal sources. My major source materials include various compilations of laws, ordinances, and archives concerning Taiwan. I begin with the statement that the administration of Taiwan paralleled that of modern Japan—with major and minor revisions adapted to local conditions. The crux of my interpretation lies in the techniques of governance and the art of political and social “grafting,” a process which I tentatively term “colonial engineering.” I conclude that this governance technique worked thanks to the creation of a hierarchically orchestrated network based on bounded spatiality.

By 1902 a centralized police system had been put in place, the hokō system had been built into it, and state and society were beginning to meet at the county level where the police system interacted with the administrative hierarchy. By 1920, when the colonial administration of Taiwan delegated part of its power to local governments—following Tokyo’s “extension of policies in Japan proper to colonies” (naichi enchō )—this process of structural integration by and large had been completed. Structural integration, however, did not necessarily result in social integration. It took the two decades of the interwar period, according to my ongoing research, for the colonial administration to appropriate Taiwanese society for its use. In the 1930s the Moral Suasion movement, which came to be called “village revival” (minfū sakkō ) in Taiwan, was key to this process of social grafting.¹ Wartime concerns reshaped this order, turning Taiwan into a disciplinary society.

The examination of Japan’s colonial administration in Taiwan points to the need to understand the informal bureaucracy; here the basic point is that the informal bureaucratic setup contributed to the “creation” of the “local.” For analytical purposes, this inevitably raises the embedded issues of colonial governmentality² and, to a limited extent, also identity politics. Japan’s colonial administrative initiatives in Taiwan illustrate the growing importance of colonial governmentality.

THE COLONIAL BUREAUCRACY

The administration of colonial Taiwan paralleled that of modern Japan itself—with major and minor revisions adapted to local conditions from time to time. As Japan’s first colony, Taiwan was also a laboratory of Japanese empire-building, so to speak. It has become generally accepted that the bureaucrats who were recruited into Taiwan followed an established order of official ranking, which determined salary as well as the package of career benefits and outlook. Much of this experimentation was taken to Korea after 1910; after 1932, it found its way to Manchukuo, then drifted even further into the “Greater East Asia Co-Prosperity Sphere” (Dai-Tōa kyōeiken ) during the Pacific War (Yamamuro 1998). This was the same source for the bureaucracy that the Taiwan government-general tapped into to run the colonial administration for half a century—with remarkable stability and efficiency.

In colonial Taiwan, as in Japan’s other colonies, the bureaucratic system relied greatly on native leadership and local initiatives for control and mobilization. And yet the number of bureaucrats in Taiwan at any time during the fifty years of Japanese rule remained very small. The key to the mechanism of Japanese colonial administration in Taiwan, as I will argue, lay not in the formal structure but in the extra-bureaucratic setup. This argument is not new, but previous scholars have neglected the nature of this “extra-bureaucracy”³ in a total empire of which Japan was the dominating part.

The modern Japanese bureaucratic system, like all bureaucratic organizations, was rigid in structure, and the law-making process could be painfully slow. Thus the tendency was for the bureaucracy to encourage central intervention at times, while making little room for local improvisation (Rueschemeyer and Evans 1985: 50–52). In Japan, the years between 1885 and 1900 witnessed the implementation of a series of reforms that ushered in a new civil bureaucracy (Silberman 1966 and 1974; Hata 1983). But from the beginning (1896–1912) of the Meiji period, an informal bureaucratic system had been created to help recruit personnel outside the formal bureaucratic structure (Yamanaka 1990; Ishikawa 1993 [1987]). In colonial Taiwan, as in Japan proper, the extra-bureaucrats and functionaries thus recruited were responding to the practical calls of the early colonial administration, and naturally fell into three major categories: technical support, administrative assistance, and the police force. Significantly, extra-bureaucratic personnel (such as shokutaku , koin , and junsa ) were eligible to become officials after the turn of the twentieth century, thus entitling them to enter into the formal bureaucracy.

After 1920, the extra-bureaucratic system continued to grow in variety and to expand in scale, as witnessed in the institutionalization of the system of the “temporary staff” of the Taiwan government-general. Meanwhile, the extra-bureaucracy was extended to local administration, as seen in the installation of the “local official-treatment staff” (chihō taigū shokuin ), which became a permanent feature internalized into, and yet remaining outside of, the formal bureaucracy. Equally important, it was later maximized to convert to Japan’s war efforts with remarkable efficiency (Ts’ai 2001).

THE POLICE IN LOCAL ADMINISTRATION

Japanese rule in Taiwan was characterized by a large police role in the colonial administration. It has often been claimed that the Japanese succeeded in penetrating one step deeper into hamlets in colonial Taiwan than had Qing China, thanks to the Japanese employment of the police force. But how exactly was this institutionalization of the village administration accomplished? The process can be identified in three stages, as witnessed in the 1901, 1909, and 1920 local reforms. By 1920, the village administration had been fully integrated into the colonial administration, a bureaucratic structure with which we are familiar from oral sources. It was upon this structure that war mobilization was later built, and it was by and large this structure that the Nationalist government took over in 1945 (Ts’ai 2000).

On the whole, the police force, with the aid of the hokō , was key to local administration. It was the 1901 bureaucratic reform that institutionalized the role of the police in the administration of local affairs. The bureaucratic system in prewar Japan operated primarily via civilian officials in the order of command, thus involving the authorization of power. A “local” reform had to be initiated by, and meanwhile was accompanied by, a change in the bureaucratic system (“local” is here defined as “prefectures and below”) in the administrative hierarchy; the term “bureaucratic” here is sometimes interchangeable with “local” or “administrative,” depending on the context.

How could this happen in a colony that was closely administered by a group of bureaucrats who jealously guarded the system? The answer can be found in the 1901 revision of the “bureaucratic system of the government-general of Taiwan” (Taiwan sōtokufu kansei). According to article 22 of the revision, “the commander-in-chief of police inspectors (keishi sōchō ) was to be supervised by both the governor-general and the chief of civil administration (minsei chōkan ) and, in case of emergency, was to be empowered to direct prefectural heads within the jurisdiction of his authorization [i.e., the police administration].” ⁴ In this way, the police force was not only internalized into, but also placed above, the civil administration.

Also, the police force was institutionalized in local administration. Significantly, the general board of the police (keisatsu honsho ) was independent of the bureau of civil administration (minseibu ), and this was to be a key feature of the Japanese rule in Taiwan. According to Article 17 of the partial revision of the draft for the 1901 bureaucratic system, the general bureau of the police was established within the bureau of civil administration—along with five other boards.⁵ Suffice it to say that the board of interior affairs (naimukyoku ) once challenged the installation of commander-in-chief of police inspectors, but the latter had the right to command prefectural chiefs, thus acting independently from both governor-general and chief of civil administration. The objection, however, was overruled.⁶ Thus lay, despite later administrative reforms, the legal foundation for the police involvement in local administration. In 1910 an administrative law was enacted in Japan authorizing the police force to assist local administrators in cases of emergency. The 1920 administrative reform of Taiwan upheld this principle, and secured the role of the police in local affairs.⁷

Specifically, a mechanism had to be constructed so that the expansion of senior police officers was comparable to that of senior local officials. The 1901 local reforms witnessed the increase in the number of prefectures from seven (including four sub-prefectures) to twenty (chō , or prefectures). In this way the door for prefectural chiefs was opened for both incumbent senior civilian officials and junior officials who had served in their positions since the beginning of Japanese rule. This was achieved by means of a “special appointment order of local staff of the government-general of Taiwan ,” thus making way for qualified senior officials and for junior officials with a minimum of five years’ service to be nominated as prefectural chiefs (chōchō ). A similar device was applicable via the “special appointment order of police inspectors (keishi )” to qualified senior police officers and junior officers with at least five years’ service to become police inspectors; the number of police inspectors was then increased from one to three.⁸ Both junior police officers and junior local officials had to be at the rank of level two (and up) of hanninkan who also met the qualification of having served for five years or more. Hanninkan were the lowest-ranking officials, the other three being—in descending order—shinnin , chokunin , and sōnin . A police inspector had the rank of sōninkan , which was categorized as a higher-ranking official (kōdōkan ). Also ranked as sōninkan were keibuchō (heads of keibu), whereas senior police officers (keibu ) were hanninkan, lower-ranking officials. The commander-in-chief of police inspectors was ranked either as chokunin or sōnin.⁹

So far it seems that police officers followed different recruitment and promotion paths from administrative officials. The level of observation here, however, is not the government-general of Taiwan or prefectures, but sub-prefectures. In accordance with Article 17 of the finalized version of the “1901 Bureaucratic System,” a sub-prefecture chief was to be appointed from three sources, including senior police officers (keibu); the other two sources were zoku and gite ; zoku were junior civilian officials in general administration and gite, assistant engineers.¹⁰ The “special appointment order of local staff” mentioned above was also applicable to hanninkan incumbents (except for gite and tsūyaku , that is, interpreters) to be promoted to the rank of zoku.¹¹ The appointment of police officers to head sub-prefectures was an institutional invention carried on from the earlier period of bemmusho (1897–1907, a predecessor of the county system). The 1898 administrative reform, for example, also made officials of general affairs (shoki ) and of assistant engineers (gite) candidates for offices of sub-bemmusho ; the third source, of course, was again senior police officers (keibu).¹² The 1901 bureaucratic reform, however, institutionalized the practice, thus making it difficult in later years to separate the police administration from the general administration.

There were twenty prefectures (chō or ting) between 1901 and 1909, but the number was reduced to twelve during the period 1909–1920. Special channels for promotion for senior administrative officials and police officers remained opened in the 1909 bureaucratic reform.¹³ However, although the 1901 reform authorized the commander-in-chief of police inspectors to intervene in general administration only in emergencies, in the 1909 reform it was legalized as a general rule. According to the revised “1909 bureaucratic system,” the commander-in-chief of police inspectors was to “direct and supervise” prefectural chiefs and police officers in matters related to police affairs. This development to unify the police administration with the general administration was further enforced by the fact that the commander-in-chief of police inspectors was also to serve as the chief of civil administration. This was the case even when the general board of the police was abolished during the two years from 1909 to 1911.¹⁴

In the light of the dominant role played by the police force in local administration, it is understandable why the 1920 administrative reform did not go so far as to challenge the established practice of so-called “unification of the police administration with the general administration” (keisei gōichi ). Rather, an alternative was offered so that qualified police officers could be appointed as hannin officials,¹⁵ thereby completing the integration of the police system into the bureaucracy.

The county system that was formally institutionalized in 1920 was, in geospatial and administrative terms, based on the sub-prefectural (bunchō; or Ch. fenting ) system established in 1901. The sub-prefectural system had made it possible for police officers to be nominated to the position of sub-prefectural chiefs, hence imposing the police system upon the administrative system by legitimizing police involvement in local administration. Not surprisingly, after 1920 county seats housed the local police headquarters, supervising dispatched offices (or police boxes) within the jurisdiction of the county. As such, instead of being abolished in 1920, they were further institutionalized into an intermediary level of administrative hierarchy for coordination, although by legal definition it was only coordinative in nature. In this way, the county system helped shape local society in an unexpected way.

The police system thus constructed was a centralized one, independent of (and also superior to, when necessary) local administration. This was to characterize Japanese colonial rule first in Taiwan and later in many parts of the Japanese empire, Korea in particular. Shih Tien-fu , a geo-historian, recently proposed a three-division geo-spatial (Ch. kongjian ) framework in an attempt to reinterpret Japanese rule in Taiwan. In Shih’s model, local society in colonial Taiwan can be conceived of as being composed of three levels of communities. By and large following the local administrative hierarchy, the model refers to: (1) police officers (Ch. jingchaguan ), corresponding to the county (gun) level; people of towns and villages (gaishōmin ), equivalent to the gaishō level; and “hamlet people” (burakumin ), at the hamlet or natural-village level.¹⁶ Shih’s reconstruction of this local society in effect reconfirms the importance of the police in county administration.

Even granted that Shih’s geo-spatial framework is valid when referring to the early part of Japanese rule, however, it is also important to point out that from the perspective of the bureaucratic system, these three spaces were developed in three distinctive stages, shaped in 1901, 1909, and 1920. The 1901 local reforms set the framework for the police involvement in county administration. However, it was not until 1909 that the hokō (built upon natural villages) system was made to assist in the “sub-county” (ku ) administration, making the hokō an integral part of formal administration.

The 1920 local reforms completed the process of the institutionalization by converting big-wards into administrative villages (gai or shō ), making gaishō the lowest administrative hierarchy of the colonial bureaucracy. The 1935 local reforms did not change this setup, except for some minor alterations of legal definitions; for example, gaishō for the first time was recognized as a legal body. Essentially, the 1920 administrative structure lasted until the end of Japanese rule.

TABLE 5.1 Ratios of County Policemen to Prefecture-Level Policemen (1942)

Sources:”Shūchō ni okeru junsa no teiin” (April 1942; Kunrei no. 43) and “Taiwan Sōtokufu chihōkan kansei dai sanjūsan jō dai roku kō no kitei ni yori gun ni haichi subeki junsa no teiin” (June 1942; Kunrei no. 74), Taiwan Sōtokufu oyobi shoshoku kansho shokuinroku (1942), 56, 58.

The county (gun did not in itself function as a full-fledged local system, as it had no authority over local finance. The county government as created in the 1920 local reforms was composed of only two divisions: general affairs (shomuka ) and police (keisatsuka ). It was by nature coordinative, thus emerging as a lower center of administration for local coordination. In Taiwan, as in Korea, the county system was kept throughout the colonial period. By contrast, Japan itself abolished the county system in 1921, and in 1926 county heads also came to an end. Thereafter, the term “county” existed only as a geo-spatial denomination in Japan.

In Taiwan after 1920, the county was once again employed as the operational boundary for the local police force, very much in line with the tradition of subprefectures in which the police played a dominant role in local administration. According to Article 33 of the “local bureaucratic system” of 1920, “the governor-general of Taiwan was able to distribute to the county police inspectors (keishi), senior police officers (keibu), and assistant senior police officers (keibuho ), as well as policemen (junsa).”¹⁷ In 1942, for example, in accordance with the above-mentioned Article 33, roughly three-fourths to four-fifths of policemen in each prefecture or sub-prefecture were dispatched to the county level, except for Taipei prefecture, where a greater police force was concentrated at the prefectural level. The ratios of county policemen to prefectural (or sub-prefectural) policemen can be seen in table 5.1.

Not surprisingly, the county formed the highest level in the local command of police power. Police bureaus (in cities) and boxes (in rural areas) were only dispatched units within the jurisdiction of the county. The model of Shih’s three-division framework serves to highlight the local command of the police order. The burakumin simply refers to “hokōmin” , which literally means the “supporting force to the police system.” And the gaishōmin is a referential term, normally referring to a group of people who came to share an enlarged identity within the border of an administrative village. As such, Shih argues that the county embodied the integration of local administration with the police system.

By upholding the county as the pinnacle of his three-graded “imaginary communities,” however, Shih in essence suggests that the county exerted a visible imprint on local society. This argument needs qualification, I contend, especially for the period before the war. There were certainly signs of growth and extension in terms of the expansion of social boundaries to the county level, but the centrality of social life remained focused on towns and villages.

CREATING THE “LOCAL”

I will now turn to the examination of the way the “local” was created. By making the point that the “local” was “created” in colonial Taiwan, I aim to test the idea of the “local” as an invented identity. I argue that the Japanese attempted to build a modern system in the colony out of a traditional one by “collapsing the temporal with the spatial.” The geo-administrative system for administrative purposes was an institutional innovation developed over time in Japan’s colonial rule. The “ward” (ku) system is a good example of this process.

The term “ward” was a sub-village unit used for administrative purposes. Much of the administrative system in colonial Taiwan was patterned after that of Japan proper, and the ward system was no exception—with historical variations and local adaptations. The ward system in Japan evolved from the ōaza , or mura after the turn of the century, and its counterpart in Taiwan was largely built upon the hokō (or hamlets) after 1920. Three major revisions took place over the course of the institutionalization of the ward system. Just as the implementation of the ōaza in Japan was antedated by a “big ward” system prior to the 1889 local administrative reforms, the ward system in Taiwan too was proceeded by a “big ward” system effective from 1909 to 1920. The 1920 local administrative reforms witnessed the creation of a “small ward” system, with ward representatives (kusōdai ) in 1935 renamed ward committee members (ku-iin ). Moreover, the history of the ward reflects both the evolution of Taiwan’s local administration in general and changes in local settlements in particular (Ts’ai 2001).

More importantly, while the creation of the ward system in modern Japan was administrative and oriented toward state-building, it also provided links of social instrumentality for regional identity and local reformation built along geo-spatial boundaries. The local administrative system, such as the ward, worked not only because it was a fully integrated state mechanism, but also because the mechanism tapped into local sources (historical roots, social needs, and romanticization of “natural community”), which offered a vital means for organization. In this way “local” identities were bridged, although it remains to be seen to what extent such an enlarged regional identity was shaped over time. Moreover, the ward system was less normative in Taiwan than in Japan proper, and the hokō system in Taiwan worked both to regulate and to complement the ward system in a way unique in the Japanese empire.

In Taiwan, a “big ward” (ku) system was carried out between 1909 and 1920, as opposed to a “small ward” (also named ku) system for the second half of Japanese rule. A “big ward” of pre-1920 consisted of several hamlets and was equivalent to an administrative village of post-1920. A post-1920 “small-ward” was subdivided from an administrative village, but the system should not be confused with an earlier small-ward system implemented in the early years of Japanese rule. For a period from 1897 to 1905, small-wards (hamlets, adopting same name of ku) under the “big-district” (gai-shō-sha ; that is, sub-county) system were enumerated. The year 1897 witnessed the implementation of the “three-division” (plains, mountains, and in-between areas) administrative suppression strategies initiated by Governor-General Nogi Maresuke (1849–1912), and 1905 was the year when the island-wide land survey was completed except for mountainous areas. The early small-ward system was carried out at a time when anti-Japanese movements threatened the colonial administration and when land and population surveys were still in flux. Accordingly, the enumerated small-ward system preceded a well-surveyed land and population as well as a stable social and political order.

The first period of the small-ward system survived the “three-division” system to the end of the Kodama (Gentarō )-Gotō (Shimpei ) administration (1898–1906). By 1915, these small-wards had come to be named in principle after one of the leading settlements, and the system paralleled much of the big-ward system. By 1920, the local system had been by and large institutionalized, and for the next quarter century until 1945 the “wards” were recognized as sub-village (towns included) administrative units under the gaishō (towns and villages), coexisting with the hokō, which was revived by the Kodama-Gotō administration in 1898. Both post-1920 small-wards and the hokō shared same bounded units of natural settlements. In this way, the “local” was created and the sub-village networks were both bound and bounded. As Mary Douglas notes from the perspective of institution-making,

To recognize a class of things is to polarize and to exclude. It involves drawing boundaries, a very different activity from grading. To move from recognizing degrees of difference to creating a similarity class is a big jump. The one activity can never of itself lead toward the other, any more than institutions can evolve toward a complete organizing of information by beginning from spontaneous self-policing conventions. (Douglas 1986:60)

The creation of the “neighborhood” was another example. The prevailing view of neighborhood organization stresses the role of the state, in terms of control and mobilization, thus placing “neighborhood” in the service of Japan’s “mobilization machine.” An alternative view, adopted originally by a few Japanese anthropologists and geo-historians, approaches neighborhood organization from a geo-spatial perspective and zeroes in on the study of “spheres of worship.” A revisionist argument, as I contend, is that “neighborhood” is a mediating paradigm that accommodates the two perspectives and provides a comparative framework of analysis.

The crux of my operational analysis of neighborhood lies in the “nexus” where the state met society. The “nexus-of-operation” approach highlights structural similarities, and it also underscores functional variations. In Japan, the development from buraku (or hamlets) to burakukai (, wartime sub-village units organized along the line of “buraku”),¹⁸ for example, exemplified constant tensions built up within the structure of Japanese-style bureaucratic rule with overlapping, but often also conflicting, norms and values. As such, the concept of buraku (and for that matter, also kyōdōtai ) needs to be contextualized in its historicity.

In Shōwa Japan, buraku was generally employed as an equivalent of ōaza (son ) or ku (sub-village divisions),¹⁹ and the watershed occurred in 1889 when Meiji law provided for the system of towns and villages. As a discourse, postwar Japanese scholarship on hamlets has been impressively rich and well articulated (for example, Ōishi and Nishida 1994 [1991]; Ōshima 1980 [1977]). It will suffice to borrow the following generalization from John Embree’s work:

Before the Meiji Restoration there were about seventy thousand mura in Japan, whereas today there are less than ten thousand. For administrative purposes many groups of two or three mura were consolidated, and, as the mura now became in part self-governing and self-supporting, it was often necessary to have a larger geographical unit with greater economic resources. When this occurred, the old mura, as included in the new larger political units, were called ōaza. Naturally such a mura covers a comparatively large area. … Each ōaza continued to function as a mura, much as it did before the consolidation. (Embree 1964 [1939]:22–23)

Ku and buraku, referring to settlements of people within a village, are not strictly geographical terms. In Suye village, where John Embree conducted fieldwork in Japan in the 1930s, there were seventeen buraku, “natural communities of about twenty households each,” and many kumi , “groups of three to five houses.”

Historically the social and economic unit is this buraku. It has its own head (nushidōri) and takes care of its own affairs, such as funerals, festivals, roads, and bridges, on a co-operative basis. … While for official elections involving governmental matters all men over twenty-five vote, in all local affairs each house has but one vote, a household being the political unit in the buraku. Life in the buraku is notable for its lack of bosses, and the nushidōri is not so much a chief as a caretaker of buraku affairs. Formerly nushidōri also supervised agricultural matters. During the last ten years, with the formation of the Agricultural Co-operative Association (sangyō kumiai []) in each buraku, these functions of the nushidōri have been greatly reduced. Rivalry exists between the buraku. … Formerly buraku rivalry was more manifest than it is today. Buraku lines are to a certain extent giving way to social class lines. (Embree 1964 [1939]:26–29)²⁰

A geographical sub-buraku unit for administrative purposes was aza , a basic division in geo-administration and household administration. Some buraku consisted of but one aza, but most buraku included several aza. Many aza were forest or paddy fields with no houses, thus often uninhabited. Like buraku, each aza had its own name. Another geographical division came to be called shikona (an equivalent of what we call tumi , that is, local names, in Chinese), which was not recorded on any official maps or geo-administrative registers. As small areas, shikona usually included only very few houses, although many of them had popular names. The words mura, aza, and buraku were often used synonymously, but mura was the most commonly used term (Embree, 1964 [1939]:24–26).

Briefly, mura, ku, buraku, and kumi were social and political divisions; ōaza, aza, and shikona were geographical terms. Both mura and son share the same Chinese character, which literally means “village”; hence a “hamlet” (mura) refers to a “natural village,” while a “village” (son) points to an “administrative village.” Thus, in modern Japan buraku in geo-administrative structure equals ōaza, which, in turn, also means mura, that is, “natural villages” (Torigoe 1985:74–75).

Tonarigumi (neighborhood organization ) was an economic, as well as a social, system dating back to the days of the Tokugawa period. Once Japan’s “old standby,” the tonarigumi was revived to cope with wartime mobilization at home, and the employment of the tonarigumi enabled Japan to maintain an effective national front. In wartime Japan, tonarigumi functioned not only to maintain neighborly harmony and daily needs, but also performed wartime duties such as neighborhood air defense, fire-fighting, and first-aid training. The basic problems that confronted a nation at war—food, clothing, fuel, shelter, and its bureaucratic continuity—had a vital bearing on the maintenance of daily life in Japan proper. In operation, the neighborhood association was really a simple affair. Circulating bulletins (kairanban ) passed around from one household to the next served as tonarigumi’s voice and its only medium of control. In addition, monthly meetings of tonarigumi provided member households a place to air complaints over privations. Significantly, tonarigumi were mostly run by women, as men were engaged in other war-defense related work (Satō 1944:779–787).

Buraku, like such terms as hokō of colonial Taiwan, served as basic units for administration in imperial Japan. Until the 1920s, buraku was a derogatory term, an abbreviation for special buraku, meaning people who were by profession involved in socially contemptible careers (such as leather-making). In the 1930s, however, along with the unfolding of the rural revival movement, the term buraku was increasingly employed to refer to natural settlements, invoking the image of hamlets. It was based on this conception of buraku as natural settlements that hamlet associations (burakukai) were later to be created in the 1940 “new polity” (shintaisei ) movement in Japan. Specifically, it was upon this structural basis of administrative unification that the movement of Imperial Subjects for Patriotic Services (kōmin hōkō , hereafter ISPS), an umbrella organization for wartime mobilization, was launched in 1941 in Taiwan (Nakai 1996 [1945]).

A further step down, and at the bottom of the administrative hierarchy, was the neighborhood organization. Broadly defined, the term “neighborhood,” in accordance with the change of time and place, has been conceived as a settlement (shūraku), a natural village (sonraku ), a buraku, a section of an administrative village (ōaza), a tithing structure of households (hokō), or a neighborhood unit of ten to twenty households (tonarigumi); in popular terminology both ōaza and buraku were often called mura (Embree 1964 [1939]:23). Narrowly conceived, it refers to the Japanese term tonarigumi, a tithing organization of ten to twenty households in pre-1945 Japan.²¹ As a tithing structure, tonarigumi had a well-defined geographical boundary.

In Taiwan, the tonarigumi was structurally an equivalent of the kō of the hokō. Thus, the evolving transformation of the functional hokō made possible an alternative conceptualization of “neighborhood.” For colonial Taiwan, however, the term buraku is not so clearly identified. As a term, buraku did not begin to sink into popular consciousness until the early 1920s, as social reform movements began. This does not mean that buraku was not introduced to the island earlier, but it does mean that it was a borrowed term, imported directly from Taishō Japan and later tailored to the tastes, problems, and therefore commands of Shōwa Japan.

In practice, due to the implementation of the hokō, the Japanese version of the age-old Chinese system for social control, the geo-administrative structure of buraku or ōaza was much more complicated than in Japan proper. The hokō had evolved on the geographical boundaries of “natural villages” that had existed long before the Japanese takeover of 1895, despite the fact that many “natural villages” were merged or redefined over time. It is perhaps not an exaggeration to claim that in colonial Taiwan the hokō mediated between state and society by constantly modifying the boundaries of natural settlements.²² More often than not, a hokō merged with other hokō units to form a buraku, although in remote areas it also existed as a single-hokō buraku. In Taiwan, many buraku formed in this way were re-merged into the boundary of ōaza after 1920, when the ōaza began to take on a definitive shape as a synonym for the divisions of “administrative villages.”²³ Thus, a hokō was almost always smaller in size than an ōaza, and a buraku in colonial Taiwan could be a hokō or an ōaza, but more likely it was something in between (Ts’ai 1998:82).

FORGING AN “ADMINISTRATIVE FRONT”

This picture began to change quickly after the war began.²⁴ A closer look at Taiwan and Korea reveals that mobilization in the form of social reform began earlier and more intensively in the colonies than in Japan proper, suggesting that the centralized police system in the colonies served as a backup force. The 1920s in both colonies witnessed campaigns for acculturation and the promotion of the Japanese language in the name of social reform. By the mid-1930s campaigns for social reform had been institutionalized, and the campaigns were translated into movements.

The key to the transformation in imperial Japan has to be found both in the Great Depression and in the escalation of war with China. The Great Depression of 1929 led in 1932 to an empire-wide movement of economic regeneration centered on Japan. It was against the background of this “movement of rural revival” (nōsan gyōson keizai kōsei undō ) that the “campaign for spiritual mobilization” (kokumin seishin sōdōin ) was launched in 1937. Meanwhile, the escalation of war with China in 1931 was followed by the creation of the “Nation of Manchuria” in 1932, and eventually total war with China in 1937.

In Japan proper, in view of the outbreak of the Sino-Japanese War, the local bureau (chihōkyoku ) of the Ministry of Interior Affairs (naimushō ) set up in August 1937 a “survey committee of local system” in an attempt to reform the local system. Following the initiative of the local bureau, the Konoe (Fumimaro , 1891–1945) cabinet enacted a “revised outline of village self-rule system” on June 30,1938. The revised outline made villages and towns (chōson ) all-inclusive in terms of local mediation, and—among other things—turned buraku into cell units in the ward administration. The Ministry of Interior Affairs managed to limit the functions of chōson councils and strengthen central control over them, in order to imitate the movement of rural revival of 1932 and the “purification movement of elections” of 1935. It was hoped that by doing this, the ministry would be able, first, to integrate the administration of towns and villages, making it possible to control local economy; second, to expand the reach of town and village administration, with buraku as bases; and third, to facilitate local administration by strengthening the power of town and village heads. The purpose of all these attempts was to enhance local efficiency and finance.

The range of reform proposed in the outline was so wide that it in effect suggested the overhaul of Japan’s local system. Not surprisingly, this plan provoked strong reactions, in particular protests from chōson councils as well as agricultural and industrial cooperatives. Meanwhile, the integration of local agricultural groups into the governments of towns and villages also involved the restructuring of agricultural administration, which remained under the jurisdiction of the Ministry of Agriculture and Forest, which in turn opposed the plan. A compromise was reached so that the local reform was moderated, and the emphasis was now on coordination rather than integration. For the first time, buraku were to be institutionalized into local administration. On the one hand, the legal acknowledgment of buraku as sub-village units necessitated a redefinition of buraku previously taken as natural settlements, thus involving the mapping of geo-spatial boundaries among buraku, administrative wards, ōaza, natural settlements, and agricultural cooperative groups. On the other hand, it facilitated the efficiency of the integration, after the merging of towns and villages. However, opposition continued to grow, causing the entire plan finally to be abandoned.²⁵

In 1940, the Imperial Rule Assistance Association (taisei yokusankai , hereafter IRAA) was created in Japan, aimed at direct participation of all imperial subjects for national mobilization. After the war with China erupted in 1937, Japan began to claim that it was creating a “New Order in East Asia (Dōa shin-chikujō ).” This slogan was later modified into a “New Order of Greater East Asia .” On August 1,1940, the term “Greater East Asia Co-Prosperity Sphere” was officially used at a press conference for the first time. Meanwhile, all political parties were dissolved. The IRAA was an umbrella organization built along the all-embracing networks of the administrative system. At the substructure of the IRAA were auxiliary bodies of gender-specific or age-specific organizations such as women’s associations (fujinkai ), youth corps (seinendan ), and vigilance corps (keibōdan ), and the like). In this way, the entire population was regimented into bounded networks and merged into the administrative hierarchy. Neighborhood associations turned out to be at the smallest and lowest, all-inclusive, multifunctional level of the Japanese administrative machinery. Apart from being a social organization, they played a crucial role in the controlled economy of wartime Japan, channeling the rationing of daily necessaries.²⁶

More importantly, the ōaza (or buraku) began to serve as the connecting nexus of power in the hierarchical transmission of wartime administration. A directive issued on September 11,1940, by the Ministry of Interior Affairs in response to the IRAA movement for the first time acknowledged hamlets (chōkai and burakukai) as auxiliary administrative units; meanwhile, it further subdivided chōkai and burakukai into tonarigumi and rinpo (neighborhood associations).²⁷ From a geo-administrative unit of pre-1940, the ōaza was turned after 1940 into a crucial part of Japan’s wartime mobilization machine.²⁸ As Gregory J. Kasza maintains (1988:281), the “New Order” movement of 1940–1941 brought about a “structural revolution of administration” comparable to the Meiji Restoration and the postwar Allied Occupation of Japan.

The 1943 local reforms legally defined chōnaikai and burakukai as cell organizations for lower-level local administration (namely, cities, towns, and villages). In the meantime, the authorization of power relegated from the central government to higher-level local administration (that is, prefectures of fu and ken ) was greatly expanded. The new local system was aimed at simplifying the administrative process in the face of total war. The idea was to streamline and centralize the bureaucratic system in an attempt to establish an effective administrative hierarchy. Compared with the revised outline of 1938, the 1943 system all the more favored the bureaucratic rule. In this way, the local system was restructured in wartime in a way Fujita Takao aptly terms “a process of bureaucratic rationalization” (1944b and 1944c).

In Taiwan, by comparison, the hokō constituted a viable bounded network. It was fundamental to the Japanese control and consolidation of power over rural areas. Conventional wisdom has it that it worked because of the threat of authority and sometimes also violence. From a control system in the employ of the police, the hokō evolved during peacetime into the basic infrastructure of colonial local administration and after 1932, but particularly after 1937, was quickly transformed into a vehicle for Japan’s wartime mobilization (Ts’ai 1990). This picture began to change quickly after the war set in.²⁹

So far as Taiwan was concerned, it was the 1936 movement to “remake local customs” that laid the foundation for the movement for general mobilization in 1938. The 1936 customs-reform movement in Taiwan began in 1933 with initiatives taken by prefectural governments for economic regeneration (the five-year economic plan of industry). The implementation of the plan in 1936 culminated in a social reform program. Thus, a key to the 1936 movement was the examination of hamlet-level agricultural units (nōgyō jikkō kumiai , or agricultural implementation associations), which involved both corporate bodies and individuals.

At the hamlet level in Taiwan in the mid-1930s, the movement for social reform took the form of buraku shinkōkai (, sub-village revival associations). It is commonly assumed that the sub-village revival associations were organized along the lines of the hokō. While this observation remains largely valid, a closer look reveals that the picture was somewhat more complex. Briefly, ōaza, along with hokō, are two key concepts for conceptualizing the spatial structure of the sub-village level administration in rural Taiwan during the latter part (1920–1945) of Japanese colonial rule—depending on regional variations. In principle, the 1936 movement was organized—as were many other movements within the Japanese empire—along the administrative hierarchy of prefecture (shū ), county (gun), and towns or villages (gai or shō); in addition, buraku shinkōkai were set up at the ōaza or hokō level. In reality, the implementation varied in accordance with locality.

In this way, various agricultural groups were brought under the unified supervision of buraku shinkōkai, most of which were newly created. However, where the majority of agricultural implementation associations were created before the buraku shinkōkai, such as in Kaohsiung prefecture, agricultural implementation associations rather than the buraku shinkōkai played the key role in the movement. The case of Taichung prefecture, by and large as a result of compromise, suggests another form the movement took. And implementation in eastern Taiwan was only loosely observed. Significantly, the buraku shinkōkai was created partially to reinforce—not replace—the coordination of existing local groups; thus transportation, social order, and sanitation, traditionally under the police administration, remained the responsibility of the hokō. As such, the 1936 movement was designed mainly to improve local enlightenment,” despite its intention of being all-inclusive (Ts’ai 1998).

NANSHIN, TOTAL WAR, AND ICHIGENKA

In the wake of Pearl Harbor, the notion of abolishing Japan’s overseas territories” was first seriously voiced in terms of the incorporation of Japan’s direct colonies, mainly Taiwan, Korea, and Karafuto. Ichigenka as engineered by the government-general in Taiwan, for example, was a movement that had a clearly identifiable pattern of development, each stage corresponding to Japan’s overall military condition in the war. The first stage, “administrative incorporation” (nai-gaichi gyōsei ichigenka ), in 1942, aimed at the administrative integration between gaichi and naichi. The second stage, “administrative speedup” (fusei minsokuka), in 1943, concerned the question of bureaucratic modernity geared toward “decisive battle.” And the third stage, “improved treatment” (shogū kaizen ) of the colonized, in 1944, rested on the assumption of a binary framework with “cultural integration” posted against national integration,” as Komagome Takeshi (1996) has argued. On March 26, 1945, use of the term “overseas territory” (gaichi ) finally ended.

Events in Taiwan cannot be understood without reference to events in Japan proper and other colonies and territories. The outbreak of full-scale war in 1937 changed the situation and brought Japanese plans for total war into reality. The “law of air raids” came into force as early as November 1937, and soon Taiwan became a “stepping-stone” for Japan’s “Southward Advance” (nanshin ).³⁰ The government-general of Taiwan took upon itself the mission of “becoming the south-bound advance base of the empire,” providing technology, talent, and materiel for administration, industry, finance, and propaganda. Its “cooperation” region covered Indo-China, Thailand, the Malay regions, the East Indonesian archipelago, and Burma; that is, what was identified by the Japanese as “the South of the Greater East Asia Co-prosperity Sphere” (Nanpō kyōeiken ).

In June 1941, Germany entered into war with Russia. Later in the same month, the Japanese cabinet passed a resolution to enhance the role of Taiwan in terms of policy implementation in Southeast Asia. Taiwan was given the mission of providing military supplies by stepping up military-related industrialization, chemicals in particular, and by quickly expanding or constructing air bases. From a “stepping-stone,” Taiwan was now poised to become an “unsinkable aircraft carrier.” The Pearl Harbor incident (December 8,1941, Pacific time) led to war between the United States and Japan. Total war diminished useful manpower and precious resources, and the crisis Japan faced demanded an overhaul of an imperial government that valued efficiency by means of centralized control. It was against this background that a major administrative reform began in 1942. Given the fact that this reform was empire-wide, it inevitably posed the question of how the colonies were bureaucratically linked to Tokyo. It also illuminates the dilemma Japan faced in redefining where its colonies should be located in the wartime empire.

Taiwan, for example, became a focal point of competition between the colonial government and the central government of Japan. To embrace trade with “the South,” the government-general had to restructure Taiwan. On the one hand, to pursue the goal of total-war mobilization, the unification of administration had to be maximized, or rationalized, thus making Taiwan a potential part of Japan proper. This was clear from the decision to shift the supervision of the government-general of Taiwan from the minister of the colonial affairs to the minister of interior affairs in 1942. On the other hand, Taiwan had traditionally been an “exterior” (gaichi), outside of Japan proper, thus practices, practical considerations, and ethnic prejudices all worked to dictate that Taiwan remain an “exterior” of Japan. It was against this background that the issue of the abolition of the hokō system, among other things, reemerged in the final years of Japanese rule in Taiwan.³¹ Meanwhile, the hokō remained under the jurisdiction of the police system, thus staying outside the realm of the 1942 and later administrative reforms.³²

There is also the issue of colonial engineering in terms of wartime mobilization. The county system began to take on an increasingly important role during wartime, especially after the ISPS movement was launched. This development was both an embodiment and an endorsement of a contemporary political slogan, “general mobilization within the county” (gunka sōdōin ).³³ Furthermore, the enhanced role of the county government imposed a permanent imprint on the structural formation of postwar Taiwan, whereby the prefecture was abolished in 1945 and the county has survived in a modified form.

Once Japan turned down the road toward the Pacific War in 1941, the scale of mobilization was quickly enlarged. The “army special volunteer system of Taiwan” was instituted in April 1942, and the “navy special volunteer system of Taiwan” in August 1943. As Japan’s losses mounted, the Koiso [Kuniaki, 1880–1950] cabinet began taking improved treatment” of the colonized seriously in 1944. Indeed, the fundamental point is that, so far as modes of colonial power were concerned, military laborers were chiefly mobilized without any great changes to existing colonial political and social structures. In Taiwan, this was made possible partly by indoctrination: mass media were key in shaping an image of loyal subjects in such a way that being a military laborer was praiseworthy. Wartimelabor mobilization was also enhanced by a shift from requisition or recruitment to both “volunteerism” and organized local groups. Volunteers came to be called “glorious military laborers” (homere no gunpu ). With the populations of Korea and Taiwan together constituting one-fourth of the empire, Tokyo was well aware that positive inducements were necessary.

After June 1942 Japan’s war strategy began to turn from offense to defense, and by the fall of 1943 Tokyo turned to the policy of the “sphere of absolute national defense” to defend the main islands from Allied attacks. In Taiwan, the colonial government instituted a policy for strengthening the decisive battle of Taiwan” on October 19,1943. Its goal was to rapidly expand materiel and food for military use, as well as providing supply. In March 1944, expecting that the Allies would land on Taiwan, the government-general promulgated the outline for implementing extraordinary measures in decisive war.” At this point Taiwan entered the final defense stage: in January 1945, to expand the “volunteer” system as Japan was clearly losing the war, conscription was carried out in Taiwan. By 1945, therefore, Taiwan was not only a disciplined society—it was a disciplinary one (Ts’ai 2005).

THE EXTRA-BUREAUCRACY

As of September 1,1945, in the wake of Japan’s defeat, the size of the staff serving in the government-general of Taiwan was 117,231. Of them, 110 (0.1%) were chokuninkan; 2,070, soninkan (1.8%, or 1.9% if including the equivalents); 20,909, hanninkan (17.8%, or 27.1% if including the equivalents). The rest were support staff, totaling 83,100, or 70.9%. Race was no doubt a significant factor in bureaucratic employment. Broken down by race, then only 1 Taiwanese (Tu Ts’ungming ) served as chokuninkan (or 0.9% of the rank); 27 Taiwanese as sōninkan (1.3% of the rank), or 51 (15.4%) if including the equivalents. Even at the lowest official rank where the overwhelming majority of Taiwanese held their official posts, only 3,673 Taiwanese served as hanninkan, or 17.6% of the rank; if another 5,177 Taiwanese taigūsha who served at the equivalent level are added, then the ratio rises to 47.6%. Simply put, the majority of the Taiwanese worked as the supporting clerks.³⁴

The Japanese “created” the “local” as a colonial space and, after 1920, further institutionalized it, thus integrating the local into the bureaucratic structure. This bureaucratic structure was deeply rooted in the bounded and overlapping networks of natural villages, which were built into semi-official local organizations. From this aspect, the formation of a dual bureaucratic structure in colonial Taiwan was not much different from its Japanese counterpart. What distinguished the colonial model from the mother country was mainly that in Taiwan, the county—rather than ōaza (sub-village administrative units), towns, or villages—was created as a colonial space parallel to the police system, thus forming a solid boundary of sociopolitical space for local Taiwanese.

The development of this colonial dual structure also enabled Japan to tap into Taiwan’s manpower and natural sources outside of its formal bureaucratic structure, set up initially to serve as an intervention mechanism for the colonial government. The integration of the local into the colonial administrative structure after 1920 along the colonial space, for example, significantly improved the coordination between the colonizer and the colonized, but the continued existence of such a dual-structured colonial space in local administration could also complicate coordination problems in unexpected ways, especially during wartime. The integration problem of the hokō into the umbrella organization of the ISPS provides only one such example during the war years.

For all its harshness and racial discrimination, the hokō now symbolizes in an idealized measure the “good old days” of order, discipline, and community identity. This reconstruction of a Japanese rule of order and discipline was shaped over the course of fifty years of Japanese rule. Moreover, it was reinforced by the civil war-cum-cold war political structure in East Asia in which Taiwan has been struggling to search for a self-identity. Thanks to the hokō’s tie to the police system, the gaishō heads did not become the targets of resentment, either before or after the war. Rather, resentment was directed at the police system and, to some extent, the hokō system.³⁵ The hokō was not improvised by the Japanese as a makeshift wartime measure—it was rooted in traditional Han Chinese society and evolved over four decades, from a policing system for social control to an all-embracing cell organization for local administration and wartime mobilization (Ts’ai 1990).

The extra-bureaucracy of the Japanese colonial administration allowed the government-general of Taiwan to respond promptly to Japan’s wartime demands with minimal bureaucratic red tape. This may help explain, if only partly, why wartime mobilization was carried out earlier and more effectively in Japan’s colonies than in Japan proper. The Japanese made their administration of Taiwan work, in essence, by turning the civil bureaucracy into a disciplinary institution for effective administration. Colonial governance also worked to shape the colonial administration into a disciplined tool for social control.

NOTES

*I would like to thank Dr. Peter Zarrow of the Institute of Modern History, Academia Sinica (Taiwan) and Edmund Fung of the University of West Sydney (Australia), for their critical remarks and editorial help in this and earlier versions of this paper.

1. By social “grafting,” I mean a certain mixing of Taiwanese and Japanese, such as the revived hokō (baojia , a tithing organization of households for social control) system in colonial Taiwan.

2. “Governmentality,” referring to “techniques of government,” is a neologism created by Michel Foucault; see his article on “governmentality” (1991). The term is employed in this essay to “pose the question of the epistemological and technical conditions of existence of the political, to analyse the historical a priori by which we construct politics as a domain of thought and action, and to analyse the instrumentation, vocabulary and forms of reason by which this is done,” see Dean 1999:47.

3. The term “extra-bureaucracy,” rather than “informal bureaucracy,” is employed here with a broader definition, which not only embraces the informal bureaucracy but also extends to a gray area, of which semibureaucratic organizations such as the Youth Corps and the hokō were important components. It is important to note that while in the study of modern Japan the term “extra-bureaucracy” often applies to political parties, there was no such mechanism in colonial Taiwan.

4. “Taiwan sōtokufu kansei chū kaisei” (November 9,1901), Chokurei no. 201, Kōbun ruishū [Hōseikyoku, ed.] file no. 25 (1901), vol. 4.

5. The five boards are, respectively, board of general affairs (sōmukyoku), board of finance (zaimukyoku), board of communication (tsūshinkyoku), board of industry (shokusankyoku), and board of public works (dobokukyoku).

6. “Naimu daishin seigi Taiwan sōtokufu kansei chū kaisei” (October 1, 1901; Naikō no. 204), and Kōbun ruishū file no. 25 (1901), vol. 4.

7. “Keisatsu kanri shokumu ōen ni kansuru ken” (November 9, 1910; Chokurei no. 427 of Japan) and same law (November 26, 1920; Chokurei no. 553 of Taiwan), Kōbun ruishū file no. 44 (1920), vol. 10.

8. “Qualified senior officials” here refers to incumbent high-ranking administrative officials (kōdō gyōseikan ) at the time of the 1901 reform, and “qualified senior police officers,” to keibuchō (), chief keibu of a prefecture or a sub-prefecture (benmusho , a dispatched unit of a prefecture). Article one of “Taiwan sōtokufu chihō shokuin tokubetsu ninyōrei” (November 1901; Chokurei no. 214), and “Taiwan sōtokufu keishi tokubetsu ninyōrei” (November 1901; Chokurei no. 212), see Kōbun ruishū file no. 25 (1901), vol. 4.

9. Articles 7 and 8 of “Taiwan sōtokufu chihōkan kansei” (March 1896; Chokurei no. 91), Kōbun ruishū file no. 21 (1897), vol. 11; article 19 of “Taiwan sōtokufu kansei chū kaisei” (November 9,1901; Chokurei no. 201), Kanpō no. 5508 (November 11, 1901), 209.

10. “Taiwan sōtokufu chihōkan kansei” (November 9, 1901; Chokurei no. 202), Kanpō no. 5508 (November 11, 1901), 210.

11. Article two of “Taiwan sōtokufu chihō shokuin tokubetsu ninyōrei” (November 1901; Chokurei no. 214), Kōbun ruishū file no. 25 (1901), vol. 4.

12. Article 38 of “Taiwan sōtokufu chihōkan kansei” (June 1898; Chokurei no. 108), Kōbun ruishū file no. 25 (1901), vol. 4.

13. “Taiwan sōtokufu chō jimukan oyobi chō keishi tokubetsu ninyōrei” (October 25, 1909; Chokurei no. 287), Taiwan hōrei shūran (1918), part 1,129.

14. “Taiwan sōtokufu kansei,” Taiwan sōtokufu shokuinroku (1910), 2. Incidentally, the chief of civil administration was to head the bureau of civil administration from 1901 through 1919. And, from 1920 till the end of Japanese rule, it was renamed sōmu chōkan (chief of the bureau of general affairs), while the board of general affairs (sōmukyoku) came to be abolished. See the revised bureaucratic system of 1919, Taiwan sōtokufu shokuinroku (1920), 23.

15. “Hannin bunkan tokubetsu ninyōrei” (August 20,1920; Chokurei no. 357), Kōbun ruishū file no. 44 (1920), vol. 10.

16. Shih 2001. This case study, however, refers only to the last two decades of Japanese rule.

17. “Taiwan sōtokufu chihōkan kansei” (July 27, 1920; Chokurei no. 218), Taiwan hōrei shūran (1918), part I, 47.

18. Buraku as a term in Japanese history means both natural settlements (shūraku) and special buraku (tokubetsu buraku). Buraku (a hamlet or natural village) was a term popularly used in both prewar and postwar Taiwan. While it remains to be examined, this term seems to have first been used by the Japanese, consciously or unconsciously, to identify natural settlements in “undeveloped” regions, in contrast with regions well “developed” in terms of modernity. It was reified by anthropologists working in Taiwan in the early twentieth century, and since then has been taken for granted by the academy until today.

19. Torigoe 1985:72–74. Buraku also refers to the meaning of “outcastes” or “the lowly” in modern Japan, but it is the geo-historical, and not the social, implication that concerns us here.

20. As of 1936, there were 9,724 mura, 1,693 towns (machi), and 129 cities (shi) in Japan (Embree 1964 [1939]:35).

21. In practice, moreover, it referred to neighborhood associations in cities and urban areas only, as neighborhood associations were called rinpo when applied to rural Japan. Nevertheless, “neighborhood associations” in English generally refer to both tonarigumi and rinpo, as well as similar organizations such as hokō.

22. For excellent research on the theoretical formulation of traditional Taiwanese settlements, see Shih 1996.

23. Again, this does not mean that prior to 1920 the concept of aza and ōaza was not introduced to Taiwan—it was. However, the pair of vocabulary items existed only in a geographical sense whereby a variety of colonial administrative purposes (such as investigations of land and population, and thus household registrations and taxation systems, as well as social controls) were made possible.

24. In this paper, the “war era” refers to the period from 1931 through 1945; that is, the “Fifteen Years’ War.”

25. Fujita 1944a. Fujita Takeo headed the survey section of the “survey committee of metropolitan Tōkyō administration” at the time of the writing.

26. Ralph Braibanti, however, pointed out the fact that the distribution of rationed commodities in Japan did not actually begin until October 29, 1942, some two years after the implementation of the tonarigumi system, and he thus cautioned against remarks made in some recent studies that “the tonarigumi was reorganized for the purpose of distributing rations” (1948:150).

27. For chōnaikai, see also Taisei yokusankai chōsakai daikyū iinkai 1943.

28. Nakatsuka 1978 and 1983. Nakatsuka Akira is one of very few Japanese scholars today who treat the ōaza seriously in terms of its role in Japan’s war mobilization.

29. “Southward Advance,” as opposed to “Northward Advance” (hokushin ), was not formally endorsed as Japan’s national policy until 1936. In Taiwan, this policy was further elaborated, three years later, and put into a slogan: “Japanization, industrialization, and Southward Advance” (kōminka , kōgyōka , and nanshinka ).

30. The abolition issue of the hokō was once raised in the self-rule movement of the 1920s. For details, see Ts’ai 1995.

31. Since the hokō system was racially directed against the Han Taiwanese, one wonders if the hokō agents (hosei or baozheng in particular) ever became a source of resentment after the colonial rule came to an end. This, however, does not seem to have been the case in the immediate period of postwar Taiwan. The surveys I conducted as a form of oral history in the early 1990s revealed that the elders who had lived through the two regimes indeed embraced mixed feelings toward the hokō system (Ts’ai 1994).

32. It is intriguing to note that the county (gun) system in Japan was abolished in 1921, followed by the abolition of county chiefs in 1926. In this way, gun was abolished as a political division, thus placing mura directly under the prefectural government. The county remained, however, a geographical and social unit, as people continued to be conscious of their identity with a particular county, and many agricultural and business organizations were built upon a county-wide basis (Embree, 1964 [1939]:22). Debates over whether the county system should be abolished included attacks on its ineffectiveness and financial insufficiency, as well as on county heads who had supported anti-establishment movements, and so forth. Yet such debates may help explain why at roughly the same time the county system was reenforced in Taiwan.

33. Taiwan sōtokufu 1945:8; Kondō 1996:432. The term “employees for both government offices and public organizations” (kankōga shokuin) refers not only to civilian officials of four formal ranks of statute (shinninkan, chokuninkan, sōninkan, and hanninkan) and their equivalents, taigūsha, but also all support staff who worked as clerks (riin), commissioned (part-time) employees (shokutaku), daily-waged employees (yatoi) and office runners (jimu yatoi), and the like.

34. In the comparison between Taiwan and Indonesia, the former under Japanese colonial rule and the latter under Japanese wartime occupation, historical contingency matters. In Indonesia under Japanese rule, for instance, the peasants’ hatred was directed immediately at the village and hamlet heads and other lower-level administrators, mainly because of the latter’s identification with the strict conscription of labor and food (rice in particular) (Kurasawa 1981).

REFERENCES

Braibanti, Ralph J. D. 1948. “Neighborhood Associations in Japan and Their Democratic Potentialities.” Far Eastern Quarterly 7.2:136–164.

Burchell, Graham, Colin Gordon, and Peter Miller, eds. 1991. The Foucault Effect: Studies in Governmentality. Chicago: University of Chicago Press.

Dean, Mitchell. 1999. Governmentality: Power and Rule in Modern Society. London: Sage Publications.

Douglas, Mary. 1986. How Institutions Think. Syracuse: Syracuse University Press.

Duus, Peter, and Hideo Kobayashi, eds. 1998. Teikoku to iu gensō: “Dai-Tōa Kyōeiken” no shisō to genjitsu [The illusion of empire: Ideology and practice in Greater East Asia]. Tokyo: Aoki Shoten.

Embree, John F. 1964 [1939]. Suye Mura: A Japanese Village. Chicago: University of Chicago Press.

Evans, Peter B., Dietrich Rueschemeyer, and Theda Skocpol, eds. 1985. Bringing the State Back In. Cambridge: Cambridge University Press.

Foucault, Michel. 1991. “Governmentality.” In The Foucault Effect, ed. Burchell, Gordon, and Miller, pp. 87–104.

Fujita, Takeo. 1944a. “Chihō seido no kakkiteki tenkai (1): chōsonsei kaisei mondai no shinten [The epoch-making development of the local system (1)].” Toshi mondai [Urban problems] 38.5:25–44.

———. 1944b. “Chihō seido no kakkiteki tenkai (4): Chihō seido no kakkiteki kaikaku no jitsugen (2) [The epoch-making development of the local system (4)].” Toshi mondai [Urban problems] 39.2:29–39.

———. 1944c. “Chihō seido no kakkiteki tenkai (5): Chihō seido no kakkiteki kaikaku no jitsugen (3) [The epoch-making development of the local system (5)].” Toshi mondai [Urban problems] 39.3:32–54.

[Hōseikyoku]. 1882–1945. Kōbun ruishū [Categorized archives of government documents]. Tokyo: Kokuritsu Kobunshokan.

Hata, Ikuhiko, ed. 1983. Kanryō no kenkyū: fumetsu no pa-wa (1868–1983) [A study of bureaucrats: The immortality of power]. Tokyo: Kōdansha.

Ishikawa, Hisao. 1993 [1987]. Kindai Nihon no meibōka to jichi [Men of high repute and self-rule in modern Japan]. Tokyo: Bokutakusha.

Kasza, Gregory J. 1988. The State and the Mass Media in Japan, 1918–1945. Berkeley: University of California Press.

Komagome, Takeshi, 1996. Shokuminchi teikoku Nihon no bunka tōgō [The cultural integration of Japan’s colonial empire]. Tokyo: Iwanami Bookstore.

Kondō, Masami. 1996. Sōryokusen to Taiwan: Nihon shokuminchi hōkai no kenkyū [Total war and Taiwan: A study of the collapse of Japan’s colonies]. Tokyo: Tōshui shoten.

Kratoska, H. Paul, ed. 2005. Asian Labor in the Wartime Japanese Empire: Unknown Histories. New York: M. E. Sharpe.

Kurasawa, Aiko. 1981. “Ja-wa no sonraku ni okeru shakai henyō no ichi kōsatsu: Nihon gunseika no momi kyōshutsu seido to sono eikyō [A survey of the social changes in the villages of Java: The supply system of unhulled rice and its influence under Japan’s military administration].” Tōnan A-ji-a kenkyū [Journal of Southeast Asian studies] 19.1:77–105.

Nakai, Atsushi. 1996 [1945]. “Hōmin hōkō undō no shinden [The development of the movement of imperial subjects for public services]. In Taiwan keizai nenpō, 1944 [The yearly news on the Taiwan economy, 1944], vol. 4, pp. 251–310. Taipei: Nan-t’ien, 1996; originally published in 1945 by Taiwan shuppan bunka kabushiki kaisha, Taihoku, 4 vols (1941–1945).

Nakatsuka, Akira. 1978. “Aru ‘mura’ no Taiheiyō sensō [The Pacific War as seen from a “natural village”].” Kenkyū nenpō [Annual report of studies in humanities and social sciences, Nara Women’s University] 22:69–86.

———. 1983. “Aru ōaza ni miru tennōsei shihai no tenkai [The development of the domination of the emperor system as seen from an ōaza].” In Kindai Nihon kokka no hō kōzō [The legal construction of the state of modern Japan], ed. Yamanaka, pp. 11–36.

Ōishi Kaichirō and Nishida Yoshiaki, eds. 1994 [1991]. Kindai Nihon no gyōseison: Nagao-ken Hanishina-gun Goka-son no kenkyū [Administrative villages in modern Japan: A case study of Goka village, Hanishina county, Nagano prefecture]. Tokyo: Nihon Keizai Hyōronsha.

[Ōkurashō kanpōka, ed.] 1885–1945. Kanpō [Official gazette]. [Tokyo: Ōkurashō Kanpōka].

Ōshima, Mitsuko. 1980 [1977]. Meiji no Mura [Natural villages in the Meiji period]. Tokyo: Kabushiki Kaisha Kyōikusha.

Rueschemeyer, Dietrich, and Peter B. Evans. 1985. “The State and Economic Transformation: Toward an Analysis of the Conditions Underlying Effective Intervention.” In Bringing the State Back In, ed. Peter B. Evans, Dietrich Rueschemeyer, and Theda Skocpol, 44–77 (Cambridge: Cambridge University Press).

Sato [Satō] Kennosuke. 1944. “How the Tonari-Gumi Operates.” Contemporary Japan: A Review of East Asiatic Affairs 13.7–9:779–787.

Shih, T’ien-fu. 1996. Lanyang pingyuan de chuantong juluo: lilun jiagouyü jiben ziliao [The traditional settlements of the Lanyang plain: Theoretical constructs and the basic archival material], vol. 1. Yilan: Yilan Xianli Wenhua Zhongxin.

———. 2001. “Rizhi shidai Taiwan diyu shehui de kongjian jiegou jiqi fazhan jizhi—yi Minxiong difang weili [Spatial structure and the developmental mechanism of Taiwan’s territorial society during the Japanese era: The case of Ming-hsiung].” Taiwanshi yanjiu [Taiwan historical research] 8.1:1–39.

Taisei yokusankai chōsakai daikyū iinkai, ed. 1943. Chōnaikai, tonarigumi no seibi kyōka ni kansuru chōsa [A survey of the enforced preparation (for war) of urban and rural neighorhood organizations]. [Tōkyō]: Taisei Yokusankai Chōsakai.

Taiwan shifan daxue, lishixue xisuo. 1995. Jiawu zhanzheng yibai zhounian jinian xueshu yantaohui lunwenji [Conference volume in memory of the centenary of the Sino-Japanese War, 1894–1895]. Taipei: Department of History, National Taiwan Normal University.

Taiwan sōtokufu, comp. 1910. Taiwan Sōtokufu shokuinroku (1910) [A directory of government officials and the staff of the government-general of Taiwan, 1910]. Taihoku: Kabushiki kaisha Taiwan nichi nichi shinpōsha.

———. 1918. Taiwan hōrei shūran [1918] [A directory of government officials and the staff of the government-general of Taiwan, 1918]. Tōkyō: Teikoku chihō gyōsei gakkai.

———. 1920. Taiwan sōtokufu shokuinroku (1920) [A directory of government officials and the staff of the government-general of Taiwan, 1920]. Taihoku: Kabushiki kaisha Taiwan nichi nichi shinpōsha.

———. 1945. Taiwan tōchi gaiyō [An outline of (Japan’s) rule in Taiwan]. Taihoku: Taiwan sōtokufu.

Taiwan sōtokufu nai Taiwan jihō hakkōjo, comp. 1942. Taiwan sōtokufu oyobi shoshoku kansho shokuinroku (1942) [A directory of government officials and the staff of the government general of Taiwan and its attached offices and burueas, 1942]. Taihoku: Taiwan nichinichi shinpōsha.

Torigoe, Hiroyuki. 1985. Ie to mura no shakaigaku [The sociology of the family system and natural villages]. Tokyo: Sekai shisōsha.

Ts’ai Hui-yu Caroline. 1990. “One Kind of Control: The Hokō System in Taiwan Under Japanese Rule, 1895–1945” (Ph.D. diss., Columbia University, New York City).

———. 1994. “Japanese Rule in Taiwan as Oral History: Findings of Hokō Questionnaire Investigations [sic], 1992–1993.” Xingda Lishi Xuebao [Chung-hsing journal of history] 4:121–144.

———. 1995. “Baojia zhidu chefei lun: Cong ‘yu zizhi yu chefei’ dao ‘yu chefei yu zizhi.’ [On the abolition of the hokō system: From “promoting self-rule by means of advancing the abolition of the hokō system” to “advancing the abolition of the hokō system by means of promoting self-rule]” In the conference volume of Jiawu zhanzheng yibai zhounian jinian xueshu yantaohui lunwenji, ed. Taiwan Shifan Daxue, 617–639.

———. 1998. “1930 niandai Taiwan jiceng xingzheng de kongjian jiegou fenxi: yi ‘nongshi shixing zuhe’ weili [A spatial analysis of the administrative foundation in Taiwan in the 1930s: A case study of “agricultural implementation associations”]” Taiwanshi yanjiu 5.2:55–100.

———. 2000. “Rizhi Taiwan jiezhuang xingzheng (1920–1945) de bianzhi yu yunzuo: jiezhuang xingzheng xiangguan mingci zhi tantao [Township administration in Taiwan under Japanese rule (1920–1945): A study of related terms],” in Taiwan falüshi yanjiu de fangfa, ed. Wang, 95–159; originally published in Taiwanshi yanjiu 3.2 (1996 [1998]): 93–141.

———. 2001. “Rizhi Taiwan zhiguan bianzhi de faling jichu: jianlun lianghua de mixi [The legal construction of the bureaucratic system in Taiwan under Japanese rule: Notes on quantitative methodology],” Research Seminar of Thematic Projects Sponsored by the National Science Council, Nankang, Taipei, Academia Sinica, June 28.

———. 2004. “Total War, Labor Drafts, and Colonial Administration: Wartime Mobilization in Taiwan (1936–1945).” In Asian Labor in the Wartime Japanese Empire, ed.Paul H. Kratoska, 106–126, 365–372 (New York: M.E. Sharpe).

Wang T’ai-sheng, ed. 2000. Taiwan falüshi yanjiu de fangfa [Approaches to the study of the legal history of Taiwan]. Taipei: Xuelin Wenhua, 2000.

Yamamuro Shin’ichi. 1998. “Shokumin teikoku, Nihon no kōsei to Manshukoku: Tōchi yōshiki no sen’i to tōchi jinsai no shūryū [The colonial empire, the construction of Japan, and the nation of Manchuria: Shifting modes of rule and the circulation of people with special talent or skills].” In Teikoku to iu gensō, ed. Duus and Kobayashi, 155–202.

Yamanaka Einosuke, ed. 1983. Kindai Nihon kokka no hō kōzō [The legal construction of the state of modern Japan]. Tokyo: Bokutakusha.

———. 1990. Kindai Nihon no chihō seido to meibōka [The local system and men of high repute in modern Japan]. Tokyo: Kōbundō.