The Takeda Sites: What Conquest Left in Use

When Takeda Katsuyori left Shinpu Castle in 1582, he was leaving a second seat behind him. He had already moved out of his family's established residence at Tsutsujigasaki to occupy the newly built castle. Shinpu burned before his departure. After his death, both places would be used again: the new castle as a military position, the older residence as a continuing base of government.^[1]

The two departures make a compact account of dynastic collapse. The subsequent occupations open a different question. What could later users retain from places whose original ruling household was gone, and what did they need to change?

The answer survives in alterations to entrances, boundaries, and ground surfaces. It also reaches beyond the residences. At Tano, where Katsuyori died, a memorial temple gave the defeated family a continuing place of observance. The former governing seat, the burned castle, and the place of death acquired different futures because they offered different things to the people who came afterward.

In 1519, Takeda Nobutora established his residence at Tsutsujigasaki, in the northern part of today's Kōfu. This became the seat of three generations: Nobutora, his son Shingen, and Shingen's son Katsuyori. The house governed Kai, the province occupying much of modern Yamanashi. Its residence joined the family's accommodation to the business of territorial rule.^[2]

The principal enclosure was broadly rectangular, bounded by a moat and substantial earth banks. Additional compounds accumulated beside it. One expansion has a particularly clear family setting. Kōfu's conservation plan cites a chronicle connecting the western compound with the 1551 marriage of Shingen's heir, Yoshinobu, to a daughter of Imagawa Yoshimoto. A new household could be accommodated alongside the established core rather than by replacing it.^[3]

The arrangement made expansion visible in the ground plan. The western compound occupied a separate, adjoining space, itself divided into terraces. An enlarged residence could retain internal distinctions among its parts. Its development was not simply a matter of spreading one building across more land.

Security also extended beyond the lower enclosure. In 1520, Nobutora began a refuge castle on Yōgaizan, the mountain to the north. The mountain site used a sequence of enclosed terraces and defended entrances along an ascent that repeatedly changed direction. Near the summit, a cut across the ridge left a narrow earthen connection; other cuts ran down the slopes.^[4]

These works addressed different approaches. A cut across a ridge interrupted movement along its top. A cut descending the slope impeded movement across that slope. The surviving combination made access a succession of problems rather than one climb ending at a single wall. Height supplied an advantage that the shaped ground developed further.

Tsutsujigasaki and Yōgaizan therefore offered different kinds of occupation. The lower residence accommodated an expanding ruling household and its continuing business. The mountain provided a more restricted defensive refuge. Neither had to perform every function of the other.

This distributed arrangement is important to the later history. The Takeda had already invested in several kinds of usable ground before Katsuyori commissioned a new seat. A change at the head of the house could leave earlier buildings, boundaries, and approaches available for another assignment.

Katsuyori began Shinpu Castle in 1581. Its name referred to a new provincial seat, distinguishing the project from the family's established center at Kōfu. He subsequently moved his residence there. This was an attempted relocation of the ruling establishment, not simply the addition of another distant lookout.^[5]

Shinpu occupied the western edge of the Shichiriiwa plateau near Nirasaki, northwest of Kōfu. The area's routes connected toward Suruga, Suwa, and Saku. The builders took advantage of the plateau's steep edge while shaping the more approachable ground into enclosures, banks, and ditches.^[6]

The plan joined substantial occupied areas to defended entrances. South of the principal enclosure lay further compounds and an approach protected by an outwork and a crescent-shaped ditch. Projecting works on the northern side addressed another part of the perimeter. These features belonged to the control of movement around the site, not just to the protection of one central building.^[7]

Shinpu's short life has encouraged the description “unfinished.” The excavators offer a more discriminating account: the earthwork layout had been substantially carried out, even though the whole building program need not have reached completion. A castle could contain considerable finished preparation without being a mature seat of government.

The north entrance to the principal enclosure gives that preparation a small, precise focus. Investigation recovered stones that had supported the gate, with stonework where it met the neighboring bank. An earthwork castle still needed particular materials at places where timber structures met the ground. Calling Shinpu a castle of earth should not make those stone foundations disappear.^[8]

In 1582, an invasion by Oda and Tokugawa forces overtook the new establishment. Katsuyori set fire to Shinpu and left. At Tano, farther east near Tenmokuzan, he, his wife from the Hōjō family, and their son Nobukatsu died. The Kai Takeda ceased to rule as an independent territorial house.^[9]

The movement from a new residence to Tano gives the sites a compressed dynastic sequence. Yet it does not exhaust their material history. The western plateau still existed; the banks still divided its surface; the older residence still occupied its ground at Kōfu. Fire and defeat had changed what could be used, and by whom, rather than removing every previous investment at once.

The excavation at Shinpu's northern entrance, carried out between 2019 and 2021, exposed that difference in successive surfaces. Investigators first encountered the hard, compacted ground and then the earlier gate foundations beneath it. Their published drawings separate the gate-bearing phase from the later leveling.^[10]

A new surface over the foundations changes the meaning of the stones. During the first phase, they supported an upright structure. In the later arrangement, they were covered by ground made suitable for use. The same position could remain within an occupied enclosure while ceasing to work as the same built entrance.

The excavators found little intervening accumulation and concluded that the clearance followed relatively soon after abandonment. They associate it with Tokugawa use during the conflict with the Hōjō over the former Takeda territories later in 1582. Their account also cites the diary of Matsudaira Ietada, which records a Tokugawa encampment at Shinpu. The identification of the leveling as Tokugawa work is the archaeological interpretation connecting that record with the sequence below ground.^[11]

This was a different use from Katsuyori's proposed provincial seat. An army needed ground on which it could establish itself. It did not necessarily need to recreate the residential arrangements that had burned. Clearing and leveling could bring part of the enclosure back into use without rebuilding every lost component in its earlier form.

The distinction makes Shinpu's brief Takeda occupation more intelligible. A hurried departure did not prove that all the labor spent on its banks and terraces had become useless. Some of that work could serve a force entering the same district for a different conflict. The site retained capabilities beyond those required by its original occupant.

It also explains why the objects found near the gate do not all belong to one frozen catastrophe. Nails could derive from its destroyed timberwork; ceramics or wall material could arrive in earth moved during clearance; other objects could belong to later activity. The investigators explicitly retain those different possibilities while studying the finds.^[12]

The key result is the sequence itself. Destruction was followed by work. Someone selected what to clear, covered what no longer needed to stand exposed, and made another usable surface. At Shinpu, reuse appears as an alteration to the ground rather than as a replacement castle tower. Its modest physical trace records a substantial change of assignment.

At Tsutsujigasaki, reuse operated on a larger scale. Katsuyori's departure for Shinpu had involved partial dismantling of the old residence. After the Takeda collapse, men serving successive Oda, Tokugawa, and Toyotomi authorities used it as a base for governing Kai. The site did not pass directly from Takeda residence to abandoned ruin.^[13]

Kōfu's conservation plan describes the method of rebuilding: retain the earlier residence as a core, add compounds around it, and alter entrances and internal divisions, including through stone construction. New authorities could change the site's organization while continuing to use its established center.^[14]

The eastern side provides a clear example. An outer enclosure developed in front of the old principal approach, joining defenses around the east and north. This placed another defended space outside an existing entrance. The earlier boundary remained part of the complex, but its relationship to the outside had changed.

The Baiō compound, south of the western enclosure, was another addition assigned by historical mapping to the period after the Takeda. Its position enlarged the arrangement around the older core rather than substituting a new residence somewhere unrelated to it. These changes addressed the edges and connections of an inherited establishment.^[14]

Inside the main enclosure, the evidence is differently dated. The conservation plan identifies records assigning a dividing stone wall to 1591–1592, when Katō Mitsuyasu held the site. At the northwest corner, high masonry associated with a tower platform is also post-Takeda, but its exact construction date remains unresolved.^[15]

The two features must therefore keep separate dates. Their shared use of stone does not place them in one campaign, and their presence within the Takeda residence does not make them Shingen's work. The dividing wall is especially revealing because it altered the inside of an enclosure whose overall position endured. A new occupant could reorganize access and separation without redrawing the entire outer rectangle.

This pattern differs from the clearance at Shinpu. There, the investigated trace concerns a return to military occupation after fire. Here, several administrations continued to operate from a former governing seat and enlarged it. Both processes made use of existing ground, but they required different amounts and kinds of alteration.

Eventually the regional headquarters shifted to Kōfu Castle. Work on that separate castle began after 1590, and the city's plan places its substantial completion around 1600.^[16] Tsutsujigasaki's period under successor authorities was thus part of a longer transition in the location of government. It was not merely a momentary occupation on the way through a defeated province.

Tano acquired a different future. It had not been the established residence that a successor government could take over. Its association was with the deaths of Katsuyori and his remaining party. The institution subsequently founded there gave that association a sustained religious purpose.

Yamanashi's account relates that Ieyasu directed former Takeda retainers to establish a temple for Katsuyori's commemoration and granted land at Tano to it. The temple came to be called Keitoku-in, taking its name from Katsuyori's Buddhist posthumous name. Its initial complex of buildings was completed in 1588.^[17]

The combination of an instruction, people to carry it out, and a land endowment distinguishes a continuing institution from a mark placed at a death site. The endowment gave the temple a material basis; buildings provided spaces in which its work could continue. Ieyasu's sponsorship did not restore the Takeda's territorial government. It established provision for the defeated within the order that followed them.

That relationship need not be explained through a guess about the patron's private feelings. The documented institutional account already identifies a concrete outcome: former retainers were charged with commemoration, and a temple was established at the place associated with their lord's end. Government and memorial care could address different consequences of the same conquest.

Keitoku-in also gave Katsuyori a different kind of name within the landscape. The residence and new castle had referred to places from which he governed. The temple's name referred to his religious commemoration after death. Its work did not require a return to the political circumstances in which he had lived.

The institution's buildings suffered later fires and rebuilding. The present temple should consequently not be read as a complete surviving installation from Ieyasu's order. Its continuity had to be carried through further acts of construction and observance. Evidence beneath its memorials makes that later work unusually tangible.

During conservation in December 2006, stones bearing writing were found inside the bases of the grave monuments at Keitoku-in. Kōshū's education authorities investigated the monuments and their deposits over the following years. Among the finds was a stone bearing the Buddhist posthumous names of Katsuyori, his wife, and Nobukatsu.^[18]

The published investigation account records 5,275 inscribed stones. Many carry multiple characters on several surfaces, including passages from the Lotus Sutra and Prajñā scriptures. Dated examples belong to the eighteenth century. The objects record a substantial undertaking in writing and deposition long after the family's deaths.^[19]

Their placement is important. The words were put within the memorial bases, not arranged solely as text for passersby to read. They belonged to an act of religious dedication. A stone bearing the family's posthumous names joined that dedication to these particular dead, while copied scriptures gave the work a form extending beyond a biographical label.

This evidence changes the scale at which continuity can be seen. A land grant and temple foundation established an institution. Later people gathered stones, wrote on them, and deposited them in renewed memorial settings. The later undertaking was not a reconstruction of Katsuyori's last movements. It was work performed for the dead by people whose own historical circumstances were different.

At Tsutsujigasaki, another religious institution eventually occupied the old governing center. Takeda Shrine was established in 1919 and enshrines Shingen. Its buildings stand within the principal enclosure, surrounded by earthworks and other remains with much earlier histories.^[20] The shrine makes the Takeda association conspicuous, while the archaeology preserves uses that extended beyond Takeda rule.

The connected places therefore retain several different kinds of continuation. Yōgaizan explains the refuge belonging to the original residential system. Shinpu preserves a new capital's prepared ground and evidence of reuse after its burning. Tsutsujigasaki records continued government and alteration under successor authorities. Tano preserves the establishment and renewal of memorial care.

They do not form one unchanged Takeda landscape. Nor do they become meaningless once the dynasty ends. Each subsequent use required something different from the place: defensible ground, an established governing enclosure, or a location at which the dead could continue to receive observance. Conquest broke the connection between the Takeda family and territorial rule. What remained was not simply a set of empty sites, but ground, structures, and associations that other people could put to work.