Kanazawa Castle: The Work of Rebuilding

After fire destroyed Kanazawa Castle's Ninomaru Palace in 1808, one of the officials appointed to its reconstruction kept a detailed record of the work. Takabatake Atsusada's account runs to fifteen manuscript volumes. The castle research institute's introduction describes officials arranging construction, specialist trades being mobilized, and materials gathered from far beyond the city.^[1]

The record offers a different beginning from the usual portrait of a founding lord. Kanazawa was already an old seat of government. Rebuilding it meant making an established institution habitable and operational again, while coordinating projects that varied greatly in scale. A palace, a gate, a storehouse, and a retaining wall did not pose the same problem.

The buildings surviving or reconstructed in the park belong to different answers to that problem. Some replaced losses promptly; others followed after long intervals. Their differences are not incidental blemishes in a once-perfect design. They reveal how a castle was sustained through separate commissions, specialized knowledge, and repeated decisions about what should be built next.

Maeda Toshiie entered Kanazawa Castle in 1583. His family subsequently ruled the extensive domain commonly called Kaga, on the Sea of Japan side of central Japan. The castle associated with that long rule was repeatedly altered. Its early tower burned in 1602; a fire in 1631 destroyed the principal palace, after which the Ninomaru enclosure became the center of lordly residence and government.^[2]

The fire of 1759 removed much of the complex again. Restoring a seat of government did not require every lost roof to return on the same date. The lord occupied the rebuilt Ninomaru Palace in 1763. The surviving Ishikawa Gate dates to 1788. The Sanjikken Nagaya, an elongated building combining storage with a defensive boundary, was rebuilt only in the middle of the following century.^[3]

That last delay is especially revealing. The castle's current account dates the replacement longhouse to 1858 and describes a prolonged interval during which its stone base remained without the former building. A national heritage catalog instead gives 1860. Either dating places its return about a century after the 1759 destruction, far later than the renewed occupation of the palace.^[4]

The comparison does not, by itself, tell us which official rejected which expenditure. It establishes something more basic: an operating castle and a fully restored architectural ensemble were different conditions. The domain could resume using its principal residence while some structures remained absent. “Rebuilding Kanazawa Castle” was a succession of projects, not one contract with a single completion ceremony.

The Ninomaru's next reconstruction followed the same distinction at a smaller scale. After the 1808 fire, the lord entered the replacement palace in 1809, while the official chronology places completion of its construction in 1810.^[3] Occupation could precede the end of the work. The date of a return to residence therefore answers a different question from the date when the builders finished.

This gives the construction records their historical force. They concern the interval in which a damaged seat was becoming usable again, not merely the retrospective naming of a finished monument. Decisions about teams, materials, and individual buildings belonged to the activity of governing from the castle.

One small comparison makes that organization unusually clear. Architectural historian Tanaka Tokue studied copies of the ridge inscriptions recorded during the 1809 works. Such inscriptions identified the people associated with a building project. For the Ninomaru Palace and the Hashizume Gate, the administrative names and numbers matched. The technical contingent did not: eight domain master carpenters were named for the palace, three for the gate.^[5]

These were not totals for everyone carrying timber or working on a roof. They were named positions within a construction organization. Tanaka relates the difference to the projects' unequal scale: a palace demanded a larger technical team than a gate, while both could fall under the same supervising officials.

The combination is more informative than either list alone. Separate buildings shared an administrative structure, but their craft leadership was adjusted to the work. A government did not have to invent an entirely new authority for each roof it commissioned. Nor did a standard chain of office require an identical team at every site.

Takabatake's record, as described by its modern editors, also follows the movement of carpenters, plasterers, roofers, and painters. Timber, copper, and gold leaf came from a wider supply network, while stone was sought in Kaga and neighboring Echizen.^[1] A finished room could bring together materials whose procurement belonged to quite different trades and journeys.

That is the practical content hidden inside the phrase “the domain rebuilt its palace.” Appointing a commissioner did not make him a painter, and gathering timber did not produce a decorated interior. The project required decisions made in one part of the organization to become usable instructions and supplies in another. The common administrative names joined the commissions; the differing craft teams show where specialization entered them.

The records also prevent every visible building from becoming a personal achievement of the lord whose reign supplied its date. His government commissioned the work, but the means of realizing it included offices and technical appointments whose responsibilities could be compared across projects. Kanazawa's repeated rebuilding depended on that capacity to organize particular tasks within a continuing institution.

The organizational chart is only the first view. Researcher Shōken Hiroshi compared entries concerning two domain carpenters, Ōnishi and Inoue, during the 1809–1810 reconstruction period. In the passages he classified, fifteen entries concerning Ōnishi dealt directly with construction instructions or reports; sixty-nine concerning Inoue did so. Yet Ōnishi reached the senior office of head of the employed carpenters before Inoue.^[6]

These figures count a historian's categories of diary entries, not hours worked or a complete balance sheet of two careers. Their value is that they separate formal position from the activities recorded for a particular commission. Starting with the highest title and attributing the project's daily direction to its holder would flatten a more distributed arrangement.

A ridge inscription and a work diary therefore supply different views of the same construction. The inscription records named participation in an ordered group. The diary can show where particular instructions, submissions, and reports appeared during the process. One should not be asked to perform the other's task.

The technical staff also produced records that extended beyond individual buildings. The castle research institute's investigation of an 1850 revised castle plan identified domain carpenters among the people responsible for its preparation.^[7] Their work included representing the larger site on paper as well as constructing parts of it in timber.

This enlarges what “craft” means at Kanazawa. Technical knowledge could involve a joint or roof, the supervision of a commission, or the measured representation of the grounds. Those activities required different scales of attention. The buildings belonged together because the institution could connect those scales, not because every component had been conceived in one moment by one designer.

The surviving warehouses are particularly useful here. Unlike the vanished palace, they allow some of the technical choices recorded in descriptions and investigations to be attached to an actual standing structure. The next question is what those choices were designed to do.

Kanazawa's pale roofs can make its buildings look more uniform than their construction warrants. The characteristic lead tile is a composite: a wooden form covered with sheet lead, rather than a solid piece of cast metal. The modern reconstruction program's technical account explains how the shaped core and metal covering work together to produce the roof's profile and surface.^[8]

The principle joins two kinds of fabrication. Timber provides the form; metal supplies the outer skin. Seeing only a silvery roof suppresses that relationship. It also conceals the need to coordinate the roof's covering with the carpentry beneath it. The reconstructed buildings make this combination visible, but their published workshop procedures are modern records of traditional construction, not eyewitness accounts of the 1809 site.

Below the roofs, white plaster and tiled lower walls answer another set of requirements. The castle's technical explanation describes earthen layers applied over a lattice and finished with plaster. On the lower wall, tiles and thick, raised plaster joints form the pattern called namako, after its resemblance to sea cucumbers. The assembly protects vulnerable surfaces against rain and snow while giving the building its distinctive appearance.^[9]

The Tsurumaru Storehouse uses a different combination. Built in 1848 as an arms store, it was the work of the domain carpenter Yamamoto Katsusaemon. Its national cultural-property description emphasizes the large interior and construction suited to storage, including stone facing along the lower exterior.^[10]

Its roof is covered with ceramic tiles, not lead. An inner earth-coated enclosure lies beneath a separate outer roof, a construction described as a saya-yane, or covering roof.^[11] The outer shelter and the enclosed storage space are related but distinct parts of the structure. A warehouse did not have to reproduce the roof assembly of a gate to belong to the same castle.

The distinction brings the institutional story back to materials. Organizing carpenters, plasterers, and roofers was not a matter of summoning interchangeable labor. Their work had to meet at particular junctions: wall and roof, timber and metal, an enclosed store and its outer shelter. The building's use helped determine which combination was appropriate.

The stonework offers a similar lesson about choice. Much of Kanazawa Castle's stone came from the Tomuro area southeast of the city. Its reddish and blue-gray varieties appear in walls ranging from relatively rough assemblies to closely fitted masonry. The site's stonework account distinguishes those treatments rather than describing the whole castle as one method of construction.^[12]

The Ishikawa Gate is a particularly compact example. Its entrances, walls, and turrets enclose a boxlike space, the masugata, through which access is controlled. The standing gate buildings date to 1788, but the stonework within the enclosure was substantially refaced in 1765. Rougher worked stone on one side and finely cut masonry on others coexist within that arrangement.^[13]

It would be misleading to turn the contrast automatically into an evolutionary diagram, with a primitive wall replaced everywhere by an improved one. At this gate, different finishes participated in the same setting. Cutting stone to close joints and choosing the treatment of its exposed face were decisions about the finished surface as well as the wall's mass.

The point is especially helpful beside the roofs and plasterwork. A more elaborate finish in one place does not prove that the builders lacked another technique elsewhere. The castle could contain several treatments because their positions and tasks differed. A practical interpretation begins by asking where a wall stood and how it related to the building above it.

The Ishikawa Gate also preserves a gap between two rebuilding campaigns: the renewed stone enclosure preceded the surviving wooden superstructure. Like the Sanjikken's long interval without its upper building, that sequence makes the castle's assembly legible. Foundations, walls, roofs, and rooms could have different construction histories while functioning together.

Those histories complicate any effort to assign the visible castle a single date. They also explain why rebuilding needed records. Repairing an inherited structure required knowledge of what was already there, not simply a new commission to reproduce a familiar outline.

Even the Sanjikken Nagaya's lead covering belongs to a later stage than its nineteenth-century rebuilding. The castle's account says the replacement initially had a boarded roof, subsequently received tiles, and acquired its present lead covering during conservation in 1969.^[14]

The standing longhouse is genuine late-Edo fabric, but that description does not make every part of its present appearance late Edo. Its roof history is an especially useful reminder because the pale covering now helps make it look consistent with other buildings in the park. The visual unity is stronger than the chronological unity.

Other familiar buildings are recent reconstructions throughout their timber structure. The Hishi Turret, Gojikken Nagaya, and adjoining Hashizume turret complex were rebuilt in 2001. Their predecessors, reconstructed after the 1808 fire, burned in 1881 when the site was in military use. Old photographs, drawings, and archaeology informed the modern project; traditional timber construction was combined with contemporary seismic provisions, fire-suppression equipment, and accessibility measures.^[15]

Those additions answer a new brief. An Edo-period storehouse or defensive turret and a building admitting modern members of the public do not carry identical obligations. Reproducing a historical form does not remove the responsibility to make its new use workable.

The modern restoration policy deliberately favors a late-Edo appearance compatible with the surviving buildings.^[16] It does not attempt to place Toshiie's sixteenth-century castle, the palace after every fire, and the military installation on the same footprint at once. Choosing a period is necessary to construct a coherent ensemble; distinguishing its new work from its survivors remains necessary to explain that ensemble honestly.

Kanazawa's present unity is therefore a managed historical presentation. It brings selected buildings back into relation while leaving their separate histories available for examination. The old construction records help recover those relations, but they do not turn new timber into an unbroken survival.

The Ninomaru Palace introduces a further limit. Surviving plans do not all depict the same arrangement. The prefecture's restoration committee found that the later public and ceremonial portion changed relatively little, while the lord's residential rooms and the inner household quarters were altered more substantially with their occupants. A newly identified plan also supplied information about interior finishes and dimensions.^[17]

This uneven evidence affected what could responsibly be reconstructed. In its 2020 findings, the committee considered work on the omotemuki—the public-facing ceremonial and administrative portion—feasible, while the residential and inner quarters required further investigation. The current project concentrates on principal parts of that public section. Its first phase centers on the entrance area.^[18]

As of 16 September 2026, this is construction in progress, not a completed palace. The project's update of 3 September describes reinforcement work for modern concrete foundations beneath the future building. That present-day engineering belongs to the reconstruction's own material history.^[18]

The choice of the public front is important without implying that the less recoverable household rooms were less consequential. A surviving record can make one part of a lost institution more buildable than another. What returns first may reflect the evidence available as well as the proposed use of the new building.

Takabatake's commission and the present project consequently share a site but not a purpose. The earlier work restored accommodation and facilities for an operating domain government. The present work gives a historical institution a form that can be studied and experienced. Its success as history depends on keeping that difference visible.

The deeper continuity lies in the need to coordinate separate kinds of knowledge. Kanazawa's castle was maintained through officials who could commission work, specialists who could give it form, and records that connected a project with what preceded it. Its long history was not the preservation of one complete design. It was the repeated work of deciding how an inherited seat could be used again.