The Ōmiya Incident: Reading the Records of Ryōma’s Last Night

On the fifteenth day of the eleventh month of Keiō 3, in 1867, Sakamoto Ryōma and Nakaoka Shintarō were attacked at the Ōmiya in Kyoto. The National Diet Library’s biography of Ryōma supplies the corresponding date in the Gregorian calendar: December 10. The two ways of writing the date refer to the same night.

In accounts of Ryōma’s life, the incident forms the final scene. Yet a detailed reconstruction of that scene brings established points into contact with questions that must be considered through testimony. Who carried out the attack? What was said? The explanations published by historical museums leave some of these questions open.

Here the focus moves from Ryōma’s entire life to this particular incident. Establishing its location and date is the starting point. From there, we can follow how accounts of the attack reached later generations.

Acknowledging what remains unknown helps distinguish the event’s outline from explanations added afterward. It allows a history of the incident to retain both its established framework and the questions that the surviving accounts do not settle.

The two men were attacked in the same place, but what followed was not identical. The Sakamoto Ryoma Memorial Museum explains that Ryōma died almost immediately, while Nakaoka Shintarō lived two days longer.

The National Diet Library’s biographies record Ryōma’s death on the fifteenth day of the eleventh month and Nakaoka’s on the seventeenth. Their Gregorian equivalents are December 10 and December 12, 1867. Treating an attack on both men as a simultaneous death erases this distinction.

An account attributed to Nakaoka after the attack cannot be treated as the same kind of evidence as something written by Ryōma himself after the event. The circumstances under which an account could have been transmitted matter to its interpretation.

Even when describing the shared attack, the two men’s deaths should not be collapsed into a single moment. Separating their timelines helps explain the place occupied by testimony that was subsequently passed down.

The memorial museum’s FAQ places three works beside one another when discussing Ryōma’s final words: Tani Kanjō’s posthumous writings, published in Meiji 45; the autobiography of Baron Abo Kiyoyasu, published in Taishō 8; and the autobiography of Count Tanaka Mitsuaki, published in Taishō 15. The museum gives these accounts some credibility as reports transmitted by people who arrived soon after the attack.

Publication dates deserve separate attention. NDL Search records the lower volume of Tani’s writings as edited by Shimauchi Toshie and published by Seikensha in 1912. It records Abo Kiyoyasu as the author of his autobiography, published by Abo Kiyokazu in 1919. These are bibliographic facts about the books; publication is not necessarily the moment when a recollection was first heard.

The three passages introduced by the museum do not describe words and actions in identical terms. Rather than taking the most memorable sentence as a complete account, it is more informative to retain the identity of the recollection and the period of the publication that carries it.

This article does not reenact dialogue or present a new critical edition of the original recollections. It identifies the works and their publication periods and uses the museum’s explanation to show that the incident has been transmitted in more than one form.

A notice for a 2024 lecture at the Ryozen Museum of History describes the theory that the Kyoto Mimawarigumi carried out the assassination as a leading view. The Sakamoto Ryoma Memorial Museum’s FAQ likewise gives weight to the involvement of Imai Nobuo and others from the Mimawarigumi, while explaining the problems with testimony.

The similar direction taken by these two institutions helps readers understand how the incident is currently presented. It does not authorize a reconstruction that identifies, as settled fact, the person wielding a sword at each instant.

Calling an explanation a leading theory preserves both comparison with other possibilities and the limitations of the records. If an article communicates an institution’s assessment, it should retain the language expressing the strength of that assessment.

Thus the Mimawarigumi theory can be described as influential without turning it into a complete identification of every participant and action. How far the evidence supports that next step is a separate question.

A shop at night, a light, stairs and a sword: imagining the incident readily brings concrete objects into view. Their familiarity can make a scene seem complete before its historical basis has been examined.

Each element nevertheless raises a question about sources. Which account supports it? Is an action described in one recollection also presented in the same way in another?

Joining striking details from different testimonies without checking their positions and sequence can produce a scene that no single source describes. This is a problem for visual storytelling. Even an explanatory illustration should not become a supposedly established diagram of the number of attackers or their movements indoors.

Using the location and date as a framework, while placing differences in testimony alongside one another, keeps the evidentiary problem visible. A vivid picture should not conceal the uncertainty that the account is trying to explain.

The outline can now be separated from the unresolved details. The attack took place at the Ōmiya in Kyoto on the fifteenth day of the eleventh month of 1867 in the historical calendar.

Ryōma and Nakaoka were both attacked, but their survival after the incident differed. These points form the framework supplied by the institutional explanations examined for this article.

Accounts of final words and details of the perpetrators require comparison of testimony transmitted afterward. The assessment that the Mimawarigumi theory is a leading view also remains distinct from a conclusive identification of perpetrators or instigators. Areas of uncertainty need not make every part of the incident equally uncertain.

Established points provide a footing for considering what the testimony can tell us and where a conclusion remains unavailable. In this reading, the event retains a recognizable outline while continuing to raise specific questions.