Sakamoto Castle: Mitsuhide’s phantom fortress and the 2026 dig

For decades, beginners who typed "Sakamoto Castle" into a search bar got a frustrating answer: Akechi Mitsuhide built a famous lakeside fort on Lake Biwa, Jesuit observer Luís Fróis ranked it just behind Oda Nobunaga's Azuchi, and then the place more or less disappeared. A park, a statue, a few shoreline stones when the water drops. Locals and historians nicknamed it a phantom castle (maboroshi no shiro) because the fame outran the ruins.

That story is changing in pieces, not with a Hollywood rebuild. In late May 2026, Otsu City announced that a dig in Shimosakamoto 3-chome had exposed a short row of granite foundation stones—about three meters of wall base, roughly ten stones, with fist-sized backing gravel behind them. City staff and castle historian Nakai Hitoshi read the line as the southern edge of the honmaru, the main enclosure. Tile scraps matched patterns from older honmaru trenches. The trench was wet enough to need pumps. Sand piled around the stones suggested an old channel reused as a moat. Then, because the pit sat on a busy road edge, crews covered it again the same afternoon. Archaeology often ends that way: a day of clarity, a backhoe of dirt, a press release, and a better map in someone's notebook.

This article is for readers who want the beginner version of that map. How do you reconstruct a Sengoku castle when drawings are missing and later builders stole the good stones? What is a honmaru versus a ninomaru when you cannot walk a tourist keep like Himeji? And why did a water fort on Biwa matter so much to Mitsuhide and Nobunaga in the first place?

Why Mitsuhide built on Lake Biwa

In 1571, after Nobunaga's forces destroyed the Enryaku-ji complex on Mt. Hiei, Mitsuhide received the Sakamoto area as a strategic charge. The slope between mountain and lake controlled approaches to Kyoto and the western Biwa shore. Boats moved bulk goods faster than packhorses; a fort that sat in the water could watch docks, tax cargo, and threaten anyone who tried to slip past by lake. Sakamoto was not a mountain eagle's nest. It was a checkpoint with a wet moat that never needed rain to fill.

Fróis's praise—second only to Azuchi in magnificence or fame, depending on the translation you meet—tells you the castle looked finished and expensive to foreign eyes, not that we have his floor plan. War chronicles care about who held a place and when it burned. They rarely list bailey widths in meters. So the written record gives politics: Mitsuhide's base, the 1582 crisis after Honnō-ji, a fire associated with the Akechi collapse, rebuilding under Niwa Nagahide, then abandonment around 1586 when Toyotomi Hideyoshi pushed power toward a new Ōtsu Castle and reused materials. The dirt has to finish the architecture lesson.

If you are new to castle politics, pair this with castle governance. A Sengoku lord did not only need walls; he needed a place to store ledgers, host allies, stage messengers, and show rank. Sakamoto's short life (about fifteen years of active use) still packed those jobs into a shoreline footprint that later towns paved over.

What a lakeside castle is trying to do

Japanese castle talk sorts sites by terrain. A mizujiro (water castle) leans on rivers, lagoons, or lakes the way a mountain castle leans on cliffs. At Sakamoto, Lake Biwa was not scenery. It was the eastern ditch, a supply route, and a psychological barrier: attackers who hated boats had to hug the landward gates where Mitsuhide could stack more walls and retainers. That logic sits inside the wider toolkit of siege warfare. Defenders still feared fire, starvation, and betrayal. Water only changed which approaches hurt more.

Low stone lines that appear in the lake during droughts—the famous kochu ishigaki, or mid-lake stone walls—fed the phantom nickname. They prove someone armored the shoreline. They do not, by themselves, tell you where the kitchen stood or how wide the outer bailey ran inland. Beginners sometimes treat those stones like a complete blueprint. Think of them as one edge of a torn photograph.

• Wet defense: Moats tied to the lake stay full; attackers wade or boat under fire.
• Logistics: Grain, timber, and people move by water; the castle watches the pier as much as the road.
• Reuse risk: Shoreline stone is easy to barge away when a new fort needs fill—exactly the fate Sakamoto likely met when Ōtsu rose.

Castle zones and how sure we can be

Read the certainty column as a weather report, not a courtroom verdict. Otsu's cultural-properties staff and Nakai framed the 2026 stones as a strong estimate for the honmaru south line—aligned with older guesses, backed by tiles and guri-ishi—not as a laser survey of every corner. That honesty is the point. Phantom castles become real through stacked probabilities.

How archaeologists rebuild a plan when drawings are gone

Modern visitors expect blueprints. Sengoku builders left few. Even when a screen painting survives for another castle, artists flatten space for drama. At Sakamoto the method looks almost boring until you see how each clue locks the next.

1. Start with documents for people and dates. Who ordered the fort? Who burned it? Who dismantled it? Texts set the time window. They do not draw the wall line.
2. Find edges, not fantasies of keeps. A foundation row with a consistent face direction is worth more than a rumor of a seven-story tower. Height estimates for missing courses stay soft; the line on the ground is the hard fact.
3. Match materials across trenches. Same granite family as the sannomaru wall. Same tile stamp family as the 1979 honmaru dig. Shared kits imply shared campaigns.
4. Read absences as possible recycling.If only base stones remain, ask where the upper courses went. For Sakamoto, the working answer is Ōtsu Castle's construction appetite around 1586.
5. Protect and publish even when you cannot leave a hole open. Backfilling feels anticlimactic. It is how you keep foundation stones from being stolen, cracked by traffic vibration, or turned into a roadside hazard.

The 1979 excavations already found burned soil, foundation stones for buildings, a well, drains, and wall footings in the presumed honmaru. Otsu's city pamphlet still cites that work. Newer science on tile color complicates the old "everything red means the 1582 fire" story: some reddish tiles may simply be how they left the kiln. That is a useful beginner lesson. Dirt remembers chemistry as well as drama.

What changed in May 2026

According to Sankei, Kyodo, Mainichi, and Otsu's own public write-up, the new trench produced a south-facing stone row about three meters long. Stones ran roughly forty centimeters across. Backing gravel sat behind them—the textbook signature of a stacked wall base, not a decorative rock garden. City archaeologist Okada Yuya noted how wet the cut stayed and how much sand hugged the line; he floated the idea that an older stream had been drafted into moat duty, with several courses of wall once rising above the surviving base. Kyodo relayed a city estimate on the order of 1.2 meters for the original wall height if missing upper stones are counted—treat that as a reconstruction guess, not a measured standing face.

Nakai's comment matters for beginners learning how experts weigh certainty. He argued that tiles found with the stones make a Sakamoto identification even firmer than the lake-bed walls alone, because tile kits tie the trench to known castle roofing. He also linked the lonely base course to the reuse story: if upper stones left for Ōtsu Castle, you would expect exactly this kind of truncated wall. In other words, absence can be evidence of recycling, not evidence that the wall never existed.

Context around the dig helps too. The early-2020s sannomaru discovery drew thousands to a public explanation day and pushed preservation talks that helped land a national historic-site designation in September 2025 for part of the estimated castle area. The 2026 honmaru-edge find sits next to that protected core on private land tied to ongoing development archaeology. Japan's castle revival is often this unglamorous: a housing plan triggers a legal dig, a wall appears, a memorandum saves one stretch, a later trench clarifies another edge. Phantom castles fade when construction law meets careful trowel work.

Stone reuse: why castles eat each other

Quarrying, shaping, and hauling granite costs labor. When Hideyoshi's network favored a new lakeside strongpoint at Ōtsu, Sakamoto's cut stone was inventory. Timber, tiles, and fittings moved the same way. This is normal Sengoku and early-modern practice, not a special insult aimed at Mitsuhide's memory. It does mean tourist expectations fail. People want a ruined keep photogenic against the lake. They get foundation ghosts and a statue in a park.

Compare that mentally with Himeji, where Tokugawa-era investment and later conservation kept a vertical silhouette. Sakamoto's lesson is the opposite half of castle history: forts were tools. When the tool's job moved down the shore, the hardware followed. Archaeology then becomes forensic accounting—tracing which stones stayed as trash in the dirt and which became someone else's curtain wall.

• What stays: Awkward base stones sunk in wet ground, broken tiles, postholes, burned lenses.
• What leaves: Clean ashlar, reusable timber, metal fittings, anything easy to barge.
• What confuses tourists: Park monuments that mark estimated baileys without showing the dig trench that justified the label.

Beginner mistakes this site wants you to skip

First mistake: treating Fróis as a surveyor. He tells you the place impressed him. He does not hand you bailey lengths. Second: assuming every Japanese castle looks like the Himeji poster. Many important sites are parks, foundations, or lake scars. Third: reading "phantom" as "legendary fake." Sakamoto was real enough to burn, rebuild, strip, and designate as a national historic site. The phantom label is about visibility, not existence.

Fourth mistake: ignoring the town. A castle without a working settlement is a lonely barracks. Sakamoto sat on a boat hub; merchants and temples kept living after the keep vanished. That is why castle-townarticles pair with castle articles. Digging only for the lord's yard misses the economic reason the yard existed.

Fifth: expecting every 2026 headline to open a tourist pit. Otsu canceled on-site explanation meetings for the May trench because of road safety. Public history sometimes happens in a symposium hall—Otsu promoted talks with historians such as Owada Tetsuo and Nakai—while the stones sleep under fill. That is not a cover-up. It is how you keep a foundation from becoming a pothole.

Mitsuhide, memory, and an empty shoreline

Popular culture loves Mitsuhide as traitor or tragic intellect, depending on the drama. Sakamoto complicates the cartoon. Before Honnō-ji he was a trusted builder of Nobunaga's western approach, parked on a lake that fed the capital region. After Yamazaki, tradition says he never made it home; the castle's Akechi chapter ends in fire and suicide stories around adoptee Hidemitsu. Later captains reused the shell until Hideyoshi's map no longer needed it. The shoreline park statue is memory theater. The 2026 stones are closer to Mitsuhide the engineer—someone who spent granite and tile on a wet perimeter, not only ink on a coup.

If you visit Otsu, go for the lake view and the historic-site panels, not for a climbable keep. Check whether any temporary exhibit or symposium is running. Walk nearby temple areas with Akechi graves and attributed relocated gates as tradition, not as courtroom exhibits. Then reread the siege and governance pieces on this site with a wet ditch in mind: defense is logistics plus geometry, and geometry sometimes only survives as a row of base stones beside a prefectural road.

FAQs