Macro Power 2: Dogo^[*] Power and the Dominant Class

The mid seventeenth century was not the first time dogo^[*] struggled against extramural interference. Two generations earlier, around the turn of the century, they had even organized armed resistance against land surveys in a number of places. In Higo province (Kumamoto) they mobilized twenty thousand peasants in opposition to Sassa Narimasa, who was in charge of the land survey. Similar rebellions took place in the north, where dogo^[*] power was strongest. As Mizubayashi points out (and as I shall discuss shortly), this contrasts with the economically more developed Kinai region, where since the time of Oda Nobunaga (1534-82) this dogo^[*] class had spontaneously begun to split into those who sought fortune primarily in military campaigns and those who sought it on the land. In the "backward" areas, however, the rural magnates clearly felt threatened by the warriors and their land surveys.^[61]

This struggle they lost. The rural magnates could stop neither the warriors nor the surveys, but they survived quite well the restructuring of political power that was taking place over their heads. It is important to clarify the changing structural position of this dogo^[*] class in order to understand how the dogo^[*] made the transition into the Tokugawa period, when they functioned as the overlords' local agents for over a century, until the bakufu dissolved their power in the first decades of the eighteenth century in part as a response to abuses.

Under warlords of the sixteenth century like the Rokkaku, the traditional control the dogo^[*] , as often semimilitary rural magnates and

[61] Mizubayashi, Hokensei^[*] , 145. On dogo^[*] uprisings, see Donald Burton, "Peasant Struggle in Japan, 1590-1760," Journal of Peasant History 5, no. 2 (1978): 144-47.

fief holders, had over their retainer/servants (hikan ) and dependents remained firm. When they and their followers were called upon for military or other tasks by these warlords, they were recruited in toto as small coherent units. Oda Nobunaga, however, started to sever the dogo's^[*] ties to their subordinates when he used his formidable authority to bypass these local bosses by recruiting some of their dependents directly. Nobunaga presented projects such as the building of Azuchi castle in 1576, employing this labor force as kuniyaku, or "national service," and he indemnified the masters of these recruits with some tax exemption.

A second way in which the ties between dogo^[*] and their followers were dissolved was through transfers of warlords. When Nobunaga assigned a new territory to a daimyo, he ordered him to take his dogo^[*] along, as he did when he sent the Maeda to Noto.^[62] These local magnates thus lost their power base of land and followers/servants and, as full-time retainers, became completely dependent upon their lords.

Many dogo^[*] were jizamurai, or landed men of arms; in the Kinai region, however, many of them had considerably reduced the portion of their holdings that they worked themselves. Thus loosening their ties to the land on their own initiative, many had joined Nobunaga's armies without giving up their landed property, expecting greater gains from military campaigns, in the form of increased holdings, than from working or managing the land they owned. Hideyoshi, as is well known, institutionalized this initial voluntary and incomplete removal of armed men from the land through his Sword Hunt Edict (1588) and his "Status Regulations" (1591).^[63] Henceforward armed men would be clearly distinct from peasants. This had a profound effect on the social position of these landed samurai or armed rural magnates, splitting this class into two since it forced them to choose either to take the warrior road or to remain in their mini-dominions working the land. Under Nobunaga such a choice had been neither total nor final; under Hideyoshi it was both.

It is important to stress that either way, as new, full-fledged retainers

[62] Mizubayashi, Hokensei^[*] , 109. For a detailed discussion of Maeda's case, see Brown, Central Authority , chap. 5.

[63] For the text of these decrees, see Tsunoda Ryusaku, William Theodore de Bary, and Donald Keene, comps., Sources of Japanese Tradition , 2 vols. (New York: Columbia University Press, 1964), 1: 319-22. The reason for quotation marks around Status Regulations will be explained later.

without land in the castle towns or as rural magnates without arms in the country, they did not become men without power in the new order. This class was reconstituted at the lower echelons of the ruling hierarchy, where it performed a number of essential functions with or without arms. The fact that some were now legally samurai and others were not should not obfuscate the fact that the new rulers encompassed nonsamurai and the ruled included peasants with a samurai background. Without a significant component of nonwarriors in the ruling cadres the system would have been unable to function, since almost all warriors were now separated from their subjects. Peasants were thus not only tribute producers, separated socially and geographically from the samurai; they were also essential to the operation of the system, in times of war, as we have seen, and in times of peace.

One can distinguish three ways in which the dogo^[*] were incorporated into the new power structure. Some became district intendants (daikan ) or their assistants (tedai ), others became village group headmen (ojoya^[*] ) in charge of a number of villages, and still others turned into a kind of contractor for "public works," especially land reclamation projects. We shall first take a look at the third group. Our example is from Kita-Saku district, a few kilometers east of Makibuse (see map 2).

As is well known, the seventeenth century witnessed the greatest boom in land reclamation in premodern Japan. Between 930 and 1450 Japan's arable increased by 10 percent. Another 70 percent was added by 1600. Equating the arable of 1450 with the value 100, it was 90 in 930, 100 in 1450, and 173 in 1600; then it almost doubled, to 314, by 1720 (and it was only 322 in 1874), as table 10 shows. Unlike the land reclamation of the eighteenth century, most of which was financed by merchant capital, the much larger projects of the seventeenth century were initiated and funded by the daimyo and the shogun. Often dogo^[*] and their bands of dependents were employed in these projects. This was the case with four such shinden (new paddies) projects in Kita-Saku.^[64]

The most famous of the four shinden is Gorobe-shinden, the others being Mikage-, Shiosawa-, and Yaehara-shinden. They were developed between 1630 and 1662. The irrigation networks for the four projects, totaling some 175 km, took twenty years to build. The ducts that brought

[64] Oishi^[*] , Kinsei sonraku no kozo^[*] , chap. 3.

the water to Gorobe-shinden alone were about 20 km long (including some 2 km hewn in rock, nine tunnels totaling 1 km, and a duct crossing a river) and passed through nine villages.^[65] All four projects were undertaken by descendants of local vassal and subvassal houses of Takeda Shingen (1521-72). The warrior ancestors of these developers had continually sided with losing parties in the wars of the late sixteenth century. For instance, the Ichikawa house, to which Gorobe belonged, was first a subvassal of the Uesugi, then of Takeda Shingen, and subsequently of Oda Nobunaga and the Hojo^[*] when its holdings were greatly reduced. The remaining Ichikawa property, mountain forest and some undeveloped grassland in the Kita-Saku plain, was finally divided among followers of the house (between fifty and one hundred). Gorobe, reluctant to leave the area, refused twice to enter Tokugawa Ieyasu's service; instead, he took the opportunity given by Ieyasu's decision to develop new arable. In return, Ieyasu acknowledged Gorobe's jurisdiction over his numerous hereditary vassal-servants (fudai and genin). In 1642, Gorobe received from the lord of Komoro domain 150 koku of the new fields as his "fief." Then the fifty or so fudai literally came out of the woodwork, writing to Gorobe, "We are sons of fudai

[65] The calculations are based on data given by Saito^[*] Yoichi^[*] in his Gorobe-shinden to hisabetsu buraku (San'ichi shobo^[*] , 1987), 1, 2, 56-57. The most detailed history of a shinden mura in English is the work of an anthropologist, Jennifer Robertson's Native and Newcomer: Making and Remaking a Japanese City (Berkeley: University of California Press, 1991), 76-103, 136-143. The remaining information on Gorobe-shinden and the developers of the other shin-den that follows is again taken from Oishi^[*] , Kinsei sonraku no kozo^[*] , chap. 3.

who for generations have been in your family and fought together until your father's generation, but then we moved to the mountains not so long ago because there was no land left in the family. Now, however, that you have received 150 koku, we shall remain loyal in any battle that may lay ahead." They worked for Gorobe again, first as vassals, then for wages, and they were finally set free in 1713. Meanwhile, the Ichikawa had left the area in 1670 and become absentee landlords.

The founding families of the other new villages have similar histories. Kashiwagi Koemon, founder of Mikage-shinden, having also sided with losing lords, went to Suruga for a while to take an office under Tokugawa Ieyasu, but soon he was back in Kita-Saku in charge of bringing new land under cultivation. He eventually received eighty-three koku in tax-exempt new fields, enjoyed labor services from the new settlers, and wielded considerable power in the new village. He passed on to his descendants the right (exercised as late as 1819!) to veto any election of village officials by the titled peasants and the right to keep all the records, which he lent to the officials at his own discretion. For the first fifteen years there was not even a village headman; when one was finally appointed, in 1665, he came from a branch house, and the office remained in that family for half a century.

A second road taken by dogo^[*] in the seventeenth century, more frequently by dogo^[*] remaining on the land, led to the new rural office of ojoya^[*] (village group headman). They thus became charged mainly with collecting tribute from a number of villages, their area of jurisdiction often coinciding with the pre-Tokugawa districts (go^[*] ).^[66] In this way, they supplemented their economically based and informal patriarchal power as local magnates with a new official, administrative dimension. Just as the sixteenth-century warlords used "the state" (which they built) to continue furthering their interests, the landed dogo^[*] added the same surplus political value to their traditional authority.

It should be remembered that the purpose of Hideyoshi's land surveys, conducted and registered at the village level, was not tribute exaction but the mobilization of a national army. For taxation purposes, Hideyoshi relied on the old local administrative units (often the go^[*] , encompassing a number of villages), which often became the juris-

[66] For this paragraph and the following one, see Mizubayashi, Hokensei^[*] , 136-37. Brown also stresses the indispensability of nonsamurai at the district and intendant level in Kaga domain (Central Authority , 114-41, 193, 205).

diction of a village group headman. Thus, one of the two dogo^[*] from Toyota village, mentioned earlier, filled the post of village group headman for thirteen villages.^[67]

The village group headmen, as stipended agents for the city-bound warriors, were an integral part of the dominant class in not only structural but also economic terms, for in both status and privilege they were considerably removed from the ruled in the villages. Located administratively beyond the village and performing their corvée or service to the overlords as officeholders, in their village of residence they were exempt from the obligations shouldered by regular villagers.

As developers of new paddies, the dogo^[*] expanded the overlords' tribute base, and as village group headmen they channeled this tribute upward. In addition, they functioned at the next higher administrative level also, as district intendants or, more often, as their representatives or assistants. One usually thinks of intendants as samurai, but especially in the early decades of the Tokugawa period a good number of them were dogo^[*] . Why?

The indispensability of the disarmed fraction of the dogo^[*] class to the new military rulers was not only a consequence of the concentration of the samurai in urban centers. It was also the result, in some important cases, of massive daimyo transfers to new territories. After the Hojo's^[*] defeat in 1590, Tokugawa Ieyasu was ordered by Hideyoshi to leave the five provinces he controlled and take charge of the eight provinces in the Kanto^[*] plain, most of which constituted the former Hojo^[*] domain. This transfer affected the dogo^[*] of Ieyasu's old and new domains in different ways.

In his old domain, the transfer put into practice, in one stroke, the separation of warriors from peasants, splitting the dogo^[*] class into those who took their arms and followed Ieyasu to the Kanto^[*] and those who stayed behind. It was a move that also strengthened the integration of his vassal band, which, cut off from its economic base, now became totally dependent on its lord. In his new domain, however, the local dogo^[*] class became indispensable for ruling the countryside, since Ieyasu concentrated most of his army in and around Edo, granting many (but not all) rural administrative posts—hatamoto, or bannerman, fiefs in the "Tokugawa houseland" and daikan or gundai intendant offices for

[67] For the system of village group headmen in the Ueda domain, see Bix, Peasant Protest , 62-64.

the remaining portion of the Tokugawa domain—to only a small fraction of the thirty-five thousand warriors under his direct command.^[68]

The figures are quite striking. Forty percent of the Tokugawa domain was entrusted to some twenty-two hundred fief-holding bannermen, each with his own retainers, but the overwhelming majority of the latter were stationed in Edo, not in the fiefs themselves. For instance, the Sengoku bannerman, who in the 1790s held a fief of two thousand koku (in eight villages) in Shinano and an additional seven hundred koku in two other provinces, had thirty-two retainers in Edo but only four in Shinano. His Shinano office was headed by someone with a dogo^[*] background who also relied on locals: two wariban (the equivalent of village group headman) and the eight village headmen.^[69] The remaining 60 percent of the Tokugawa domain was ruled by a mere forty intendants. They were assisted by helpers, whose total number in 1839 was less than a thousand, slightly over half of them based in Edo. Under Hideyoshi the intendants had been recruited from the ranks of vassals and dogo^[*] . In the early Tokugawa period they were former retainers from defeated daimyo (the Imagawa, the Takeda, the Hojo^[*] ) or local dogo^[*] who had served under them. This local entrenchment, perpetuated through heredity, led to numerous abuses that were further facilitated by the way these intendants were paid, namely, by allowances they themselves took from the tribute they collected. In the 1680s the majority of the intendants (thirty-five) were purged, and after 1725 intendants had to forward the total amount of tribute to the center, which substituted fixed stipends for their allowances. In addition, their posts were now increasingly being filled by staffers from

[68] Mizubayashi, Hokensei^[*] , 143. The figure 35,000 comes from Conrad Totman, Politics in the Tokugawa Bakufu , 1600-1843 (Cambridge: Harvard University Press, 1967), 135. By the end of the eighteenth century only 44 percent, or 2,264, of the bannermen held fiefs, with a total kokudaka of 2.6 million koku (see KDJ 9:381, s.v. "chigyoseido^[*] "; for similar figures, see also Totman, Politics , 135), spread over forty provinces, and governed from 3,677 local offices because many fiefs consisted of separate territories, 43 percent of these fiefs having parcels in from two to six different provinces. The portion of the Tokugawa domain that was not parceled out in fiefs (each less than 10,000 koku) to bannermen (3.2 million to 4.2 million koku—the amount fluctuates over time) was Tokugawa "houseland" (tenryo^[*] ), administered by intendants, called daikan (50,000 to 100,000 koku) or gundai (over 100,000 koku), whose number stabilized at around forty (KDJ 4:1040-44; see also Totman, Politics , 66-85).

[69] Mizubayashi, Hokensei^[*] , 268-69.

the bakufu's central Finance Commission. These purges and reforms signaled the end of those dogo^[*] who had survived as intendants.

Many of the intendants who were not dogo^[*] but samurai took up residence in Edo and, like the bannermen, maintained only skeletal staffs in their country and Edo offices. The two intendant offices in Shinano—Nakanojo^[*] (69,000 koku) and Nakano (54,000 koku)—had staffs of only three and four men, respectively, while their Edo offices had nine and eight. (Shinano also counted two branch offices, one in Mikage with a staff of three and one in the highway station of Oiwake with only one post.) Their small staffs consisted of helpers called tedai and tetsuki . The latter, shogunal retainers, were introduced only in the 1790s; the former were local peasants.^[70] The situation was slightly different in the domains, for there the offices were often overstaffed. The small domain of Tanoguchi (12,000 koku) in Saku district, for instance, counted only twenty-five villages, yet there were three intendants.^[71]

Given the sparse use of samurai for rural administrative purposes, these intendants and fief-holding bannermen had to be assisted by locals both in their small territorial offices and in the villages, where they were assisted by village group headmen and village headmen. Moreover, although one might expect responsibility for the expenses, if not for the personnel then at least for the upkeep of these offices, to have been assumed by the mighty bakufu, that was not the case. Routine maintenance expenses were shouldered by villagers, who were also charged with the corvée duties of cleaning, changing the paper of sliding doors, operating the hot bath, and so on. They also had to provide messengers to communicate between the rural and Edo offices of these administrators.^[72]

[70] This information on the daikan and gundai was taken from KDJ 4:1040-44; see also Mizubayashi, Hokensei^[*] , 270-71, and Totman, Politics , 66-85. In 1853 the bakufu employed 640 tedai (291 of these in Edo) and only 225 tetsuki (146 in Edo). Thus, when in the 1790s the bakufu created the new post of tetsuki, to be filled by retainers, next to the identical post of tedai, filled by commoners, not only were its numbers significantly lower than that of the tedai but fewer than 100 retainers were deployed in the countryside (KDJ 9:885, s.v. "tedai").

[71] NAK-KS2 (1): 486. One of these intendants, appointed in 1851, kept a detailed list of all incoming and outgoing presents for his first year in office: he himself gave gifts on 188 occasions, 4 times to headmen from five villages, and received gifts on 596 occasions! (Ichikawa Yuichiro^[*] , Saku chiho^[*] , 220-47).

[72] KDJ 8:711, s.v. "daikansho"; 7:940, s.v. "jinya."

The system was extremely cost-effective and user-friendly for the samurai because the peasants, both elites and others, not only paid for it but ran it as well. The analogy of firing squad victims who first have to pay for the bullets and dig their own grave comes to mind. This analogy might seem far-fetched, bur in Tokugawa times villages that for one reason or another had a member in prison had to pay the "boarding" costs.^[73] Notwithstanding Confucian rhetoric about peasants' being "the basis of the country," the samurai were unwilling to dirty their hands through contact with peasants, not even for the purpose of levying tribute.

Samurai rural administrators, a rather rare breed in the first place, ranked low in the prestige and stipend hierarchy. Ogyu^[*] Sorai (1666-1728) wrote that "the intendants' sole preoccupation being to levy tribute, they appointed people just because they could write and count and they themselves were of low status ... which was the reason why no samurai with a pedigree became intendants, and this in turn was why no samurai, including the upper ranks of counselors, had any real sense of the people."^[74] The fundamental division between manual and mental labor stressed by Karl Marx and Max Weber^[75] is basic to the separation between warriors and peasants, although in Tokugawa Japan, as Ando^[*] Shoeki^[*] noted in the eighteenth century, it was more a question of manual versus no labor.^[76]

The nonsamuraized fraction of the pre-Tokugawa dogo^[*] class was thus reconstituted as an essential lower fraction of the dominant class and maintained local power well into the eighteenth century. Its demise, however, is not to be attributed only to bakufu initiative against abuses that diminished the flow of tribute to the upper fractions of the domi-

[73] Those confined to an intendant's jail whose provenance was known (i.e., those who were not vagrants, mushuku ) had to arrange for their own food, make payments for their expenses to the prison guard, and provide even their own lamp oil. The situation was slightly different in Edo (see Hiramatsu, Kinsei keiji soshoho^[*] , 936 ).

[74] Ogyu^[*] Sorai, Sorai sensei tomonsho^[*] (1727), in Ogyu^[*]Sorai , ed. Bito^[*] Masahide, Nihon no meicho, 16 (Chuokoronsha^[*] , 1983), 319, 320-21.

[75] Karl Marx and Frederick Engels, The German Ideology (New York: International Publishers, 1970), 51; Max Weber, Economy and Society , ed. Guenter Roth and Claus Wittich, 2 vols. (Berkeley: University of California Press, 1978), 2:936.

[76] Maruyama Masao, Studies in the Intellectual History of Tokugawa Japan (Princeton: Princeton University Press, 1974), 252-53.

nant class, a story of rationalization and bureaucratization; there was also a political side to this story.

Protests and petitions had led to limitations on headman authority in the first decades of the Tokugawa period. Then these dogo-headmen^[*] were forced to share power with a class of new landholders, often former dependents. Moreover, since the late seventeenth century their authority had been further checked by a representative from the other peasants, a hyakushodai^[*] (a new village office to be discussed below). The position of village group headman was abolished in the bakufu domain in 1713 as a response to a violent peasant rebellion in the Murakami fief in Echigo. Similar rebellions against abuses by village group headmen followed in daimyo domains (Kanazawa, 1712; Hiroshima, 1718; Aizu, 1720; Kurume, 1754; Fukuyama, 1789), leading there also to the abolishment of the office. The bakufu and domains, however, could not long do without some conduit between their regional offices and the villages, hence the on-again, off-again status of the office of the village group headman (on again in the bakufu in 1734, on again in Fukuyama in 1791 and off again in 1823). In the bakufu the office was now staffed on a rotation basis by headmen from the villages under its jurisdiction.^[77]