Public intellectual, journalist and parliamentarian Takekoshi Yosaburō visited Taiwan in 1904 to write a book celebrating Japan’s colonial achievements. A committed Whig historian, Takekoshi was the embodiment of progressive thought in Meiji Japan.^[47] In his largely hagiographic portrait of Gotō’s reforming administration, Takekoshi cited Inō Kanori as his authority on conditions in Aboriginal country, while praising Gotō Shinpei’s broader vision for developing the island. Takekoshi’s ambivalent report on indigenous administration, I believe, is evidence that Inō was still articulating a vindicationist rhetoric in his face-to-face dealings with other Japanese, even if his ethnological and historical writing under the auspices of the Government-General deployed the “trope of the savage headhunter” as an explanatory device. Takekoshi’s view of the situation as it existed around 1904 is more candid and less reticent than Inō’s 1905 digest as to why the “northern tribes” had been singled out for the policy of “intimidate and coerce” after 1903:

In Takekoshi’s analysis, “the policy of moderation and goodwill” is none other than the Bukonsho ideal of kyōka, or “moral suasion”. The Bukonsho’s charter stated that officers should learn Aboriginal languages, study their customs and enact the policy of kyōka in accordance with this hard-earned knowledge. In fact, the wording of the Bukonsho’s charter (March 1896) resembles Gotō’s 1901 rationale for a survey of customary law: good policy is based on accurate knowledge of local conditions. But as the Bukonsho project was launched, the enormity of the task as outlined became apparent. The variety and difficulty of local languages, the complexity of the “late imperial frontier economy”^[49] in the borderlands, and the resistance of armed Taiwanese all revealed that the enactment of “moral suasion” would be a long-term approach. Japanese officials, with few exceptions, did not understand Aboriginal languages, and even lacked accurate information about the location of many northern villages. To learn these languages, and map this terrain, local alliances would be required, and these were always slow in the forging.^[50]

For officials in Taipei, however, the rhythms of compound interest on public bonds, Japanese election cycles and the challenges of international diplomacy set the timetable for action. Despite protestations that severity was required as a response to “headhunting”, it is clear that the “metropolitan clock” indeed ticked loudly in Gotō’s ears as the Japanese public grew weary of colonial debts to the mother country and stories of rampant corruption in the management of the island. To solve the fiscal problems of empire, Gotō instituted a camphor monopoly in 1899, and it began to pay quite handsomely by 1901. As the Government-General became addicted to this new income stream, policy indeed tipped away from the kyōka faction and towards the chōbatsu faction. As Antonio C. Tavares has demonstrated, harvesting and processing of camphor under the old system of traditional fees to indigenous strongmen was too slow and complicated to accommodate the high-velocity commodity flows that were now required to balance the colonial books. Thus, the Government-General began to support, rather than restrain, Japanese camphor companies who flouted local conventions and thereby exacerbated frontier skirmishing over access to resources.^[51]

Takekoshi’s blunt statement of an “us or them” mentality indicates a way out of a key dilemma posed by Japanese survey anthropology under the Gotō regime. That is, it clarifies how Gotō and his chosen coterie could at once be the very emblems of Foucauldian “governmentality” in their zeal for colonial research, infrastructure programmes and public health policy while, at the same time, remaining ignorant and even hostile to survey anthropology, even if done in the name of the state. Yao Jen-to argues that post-colonial critics are off the track when they study literature and fiction to understand colonialist discourse as a series of misreadings, errors and silences. Instead, Yao insists, the statistical-bureaucratic-legal machine that was the Taiwan Government-General is better understood in terms of what it did know about the population it constituted through its statistical compendia, surveys and development projects. In this analysis, the Taiwan Government-General was the quintessential biopolitical regime, because it “forced the Taiwanese to become healthy”; not for humanitarian reasons, but to grow a large labour force of “docile bodies” as the engine of a colonial economy.^[52]

Interestingly, Yao considers the 1905 census of Taiwan to have been a precociously detailed and fine-grained example of surveillance, an example of how far the Japanese state reached into the lives of the populace. However, Yao fails to mention that 60% of Taiwan’s territory, the abode of the Aborigines, was excluded from the census because of its sparse population, poor infrastructure and unsettled political conditions. How can we consider a regime as the essence of “biopower” and “governmentality” when it avoids surveying 60% of its territory? The answer lies in Takekoshi Yosaburō’s summary remarks on indigenous administration around 1904. If the “biopolitical” body, the population whose increase and health the government seeks to further, is construed as the whole empire, the home islands of Japan (naichi) and all of Taiwan, then the government’s willingness to confine, embargo and slaughter Atayal villagers is logical within a governmental framework. That is to say, from Takekoshi’s perspective, the northern tribes were jeopardizing the increase, wealth and survival of Japan’s “45 millions”.

From Inō’s, Torii’s and Mori’s centre of calculation, Aboriginal territory loomed large; it was a treasure-house of ethnic abundance and a forbidding terrain that might require decades to survey properly. For the survey anthropologist, space was everything. For Gotō and Takekoshi, however, population was more important than space; from their centre of calculation in Tokyo, biopolitical considerations doomed the ethnologist’s vindicationist rhetoric to the dustbin of history. And yet, as Kobayashi Gakuji has perceptively argued, Inō’s taxonomic work was not a completely innocent exercise. In Kobayashi’s analysis, it was Inō and the survey anthropologists, with their modern theories of race and ethnicity, who drew the sharp conceptual line between “Chinese” and “Aborigine” in Taiwan.^[53] This epistemic operation, to use Nicholas Thomas’s language, indeed “partitioned the human species” in such a way as to enable the Japanese officials analysed above to treat camphor-related violence as a distinct category of trouble; namely, the “Aboriginal Problem”.