What Is the Bansenshūkai? Its Twenty-Two-Scroll Structure

The Bansenshūkai was completed in 1676, in the fourth year of Enpō. It is conventionally attributed to Fujibayashi Yasutake and associated with Iga. Those facts locate a compilation made decades after the great civil wars, not a field notebook written during one famous infiltration.

That distinction is the best place to begin. The work gathers teachings, classifies them, gives them an order, and presents them as knowledge worthy of controlled transmission. A prescription may preserve older practice, but its presence in a 1676 manuscript does not by itself date that practice or prove how frequently anyone performed it. The book is evidence first of what its compiler wanted to record and organize.

A title in which streams enter the sea

The four characters of the title are commonly read Bansenshūkai. They evoke many streams gathering into the sea. That image suits the stated scale of the project: teachings attributed to different lines are assembled into one large body. It should not be mistaken for proof that every contributor, school, or technique can now be reconstructed as an independent historical unit.

Compilation changes its materials. A compiler selects, copies, abbreviates, groups, and labels. The arrangement therefore deserves as much attention as the spectacular items inside it. When a method appears under overt infiltration, covert infiltration, weather, or equipment, that address tells us how the compiler understood its function.

Why “twenty-two scrolls” needs a footnote

The National Archives of Japan catalogues its witness as twenty-two scrolls with one additional volume of military secrets, Gun’yō Hiki. The same catalogue describes the physical item as eleven manuscript books. Those numbers count different things. A scroll is a textual division; a bound book is a material container. Two scrolls may occupy one physical volume.

Iga City’s digital museum describes another witness as twenty-six books. It also lists surviving units separately, including Seishin I, Yōnin upper and middle, Innin III, and several equipment scrolls. The difference between eleven and twenty-six books is not evidence that one catalogue has miscounted the work. It warns us that copies can bind, divide, and preserve the textual plan differently.

For a reader, the safest wording is therefore precise: the work is organized as twenty-two scrolls, while surviving manuscript sets may have a different number of physical books and may include an additional associated text. “Twenty-two volumes” is convenient English, but it can blur that material distinction.

The sequence begins with rule, not equipment

The opening division, Seishin, addresses the ordering of the heart or mind. It is followed by Shōchi, knowledge expected of a commander. Only after those foundations does the work proceed through operational classifications. This order prevents us from treating the collection as an illustrated hardware catalogue.

Yōnin concerns methods in which the practitioner appears or works in a visible social role. Innin treats covert entry and activity. Other divisions address timing and weather, terrain and direction, and the construction or use of equipment. The equipment material itself is organized by purpose, including climbing, water, opening or breaching, and fire. Classification is one of the work’s central acts.

The categories also correct a common modern assumption. Concealment need not mean physical invisibility. A person may conceal intention while appearing as an ordinary traveler, worker, religious figure, or guest. Conversely, a hidden entry depends on preparation, information, and an exit plan rather than on darkness alone.

Fujibayashi Yasutake: attribution and caution

Catalogues at the National Archives and Iga City attribute the work to Fujibayashi Yasutake. The Iga catalogue, however, notes that its particular witness gives only a recluse named Fujibayashi in the relevant statement. The expanded personal name belongs to the received attribution and should be reported as such, not imagined as a modern signature page common to every copy.

That caution does not erase authorship. It describes the level at which it is known. Early-modern works often survive through copies whose titles, colophons, annotations, and bindings differ. A responsible account distinguishes what a given manuscript says from what later catalogues and scholarship identify across the transmission.

The same discipline applies to Iga and Kōka. The compilation is often presented as bringing together traditions connected with both regions. That is useful orientation. It does not license assigning every passage to a named historical practitioner or turning textual labels into a complete genealogy.

Geographical labels also belong to the transmission history. Iga and Kōka were not timeless organizations with fixed membership lists. They were places with villages, households, local alliances, and changing relations to wider political authorities. When the compilation claims or implies a regional lineage, the statement must be compared with documents produced outside the manual before it becomes a history of actual service.

Catalogues help by keeping material facts visible. A shelf mark, number of books, script description, and institutional provenance may seem less exciting than a fire device. They answer essential questions: which object was examined, where it is preserved, how it is divided, and whether another reader can inspect it. A modern transcription can make the wording searchable, while manuscript images disclose diagrams, marginal notes, damage, and the placement of headings.

Neither form is sufficient for every question. Searchable text may silently normalize variant characters. Images may be difficult to decipher or omit a binding context from the screen. Strong claims should therefore record the witness and the editorial layer used. That habit is especially important in a work whose popular reputation encourages fragments to circulate without an address.

How to use a technical manual as historical evidence

A manual is prescriptive: it says what should be known, made, carried, or done. It is not automatically a report that a method succeeded on a particular date. Tool diagrams establish that an object was represented and explained. They do not establish production numbers, routine use, or archaeological survival.

Normative passages require the same care. The work distinguishes authorized service from action driven by private desire. That distinction tells us how the compiler justified secret methods. It cannot prove that every practitioner obeyed the rule. The gap between a rule and conduct is a historical question, not a defect to be filled with legend.

Close reading therefore moves in stages. Identify the manuscript or edition. Restore the scroll and section. Read the nearby classification. Separate wording from inference. Compare another witness when a disputed character matters. Then bring in letters, domain records, material objects, or other manuals if the question extends beyond the book.

Readings in this series

Sources and scope

This guide uses the National Archives of Japan catalogue record for the date, attributed author, twenty-two-scroll plan, associated Gun’yō Hiki, and eleven-book manuscript description. It also uses Iga City Digital Museum’s catalogue and its content guide for the twenty-six-book witness and its surviving divisions. This page is a guide to the source as a whole. It quotes no selected passage and adds nothing to the excerpt ledger.