How Did Japan Know a Document Was Genuine?

In a letter written in 1590, Date Masamune added an unusual reassurance. He had made a mistake in drawing his kaō, the stylized signature used to identify and authorize a document. Despite the faulty mark, he explained, the letter was indeed his own.^[1]

Masamune was a powerful northeastern warlord, writing home after meeting Toyotomi Hideyoshi during the campaign against Odawara. The surviving letter's precise addressee is unknown. Its postscript nevertheless preserves a small problem that its writer thought worth addressing: the mark intended to establish authorship had come out wrong.

The explanation was another statement inside the very letter it sought to authenticate. A recipient still needed reasons to accept the letter as a whole. The problem was larger than recognizing a shape.

Across Japanese history, documents carried instructions, permissions, agreements, and claims beyond the presence of the people who made them. Their credibility depended on several connections: between a person and a mark, between a decision and its authorized messenger, and between a surviving sheet and the wording originally approved. Seals and signatures helped establish those connections. They worked within arrangements that could be examined—and could fail.

A document from 1072 joins a temple's application to the response it received. Gufukuji, also known as Kawaradera, requested confirmation of tax-exempt treatment for its lands in Yamato Province. The application listed the holdings and carried the signatures of three temple officials. At its end, the provincial governor added a ruling recognizing the land as temple property.^[2]

The completed scroll is more than three meters long. The provincial seal appears twenty times, including on the ruling's text and at joins between sheets. The placement brought official endorsement into contact with the physical structure of the record. Someone handling the assembled paper could see more than a single mark detached from the passages it purported to authorize.

A seal across a join is particularly revealing. Its impression belongs to two adjoining surfaces. Separating them or substituting another sheet would disrupt that relationship unless the mark, too, were reproduced. This is a practical implication of the arrangement, not evidence that no one could defeat it.

The temple's signatures and the governor's seals also served different positions in the transaction. One party made an application; another supplied a decision. Preserving both together made it possible to relate the endorsement to the request that had prompted it. A seal's significance came partly from whose act it identified and where it stood in the document.

That distinction protects the historical reading as well. The record establishes an official recognition; it does not give us an independent survey of every field. Authenticity concerns whether the document is what it claims to be. The accuracy of information supplied within an authentic application remains a further question.

Already, then, there were several things to secure. The parties had to be identifiable, the ruling had to belong to the application, and the sheets had to remain connected. Calling all the marks proof of authenticity compresses these separate tasks into one word. Their particular positions explain what work they could actually do.

Under the fifteenth-century Muromachi shogunate, authorization and dispatch could leave separate records. Materials from 1430–1432 preserve stages of business conducted before the shogun Ashikaga Yoshinori.

The legal historian Nitta Ichirō explains the procedure: officials, usually working in pairs, submitted a proposed disposition to the shogun. After approval, they prepared a bugyōnin hōsho, a document conveying the decision to the relevant parties. One surviving record contains Yoshinori's own kaō authorizing decisions. A companion record of documents issued survives in a later medieval copy.^[3]

The arrangement gave handwriting a defined location in a longer process. The shogun could personally approve an entry without personally writing every document that carried its consequences outward. Officials did the work of turning the approval into an instruction for recipients.

An authorized document written by someone else was therefore not necessarily a substitute of inferior status. Its authority depended on the writer's role and on the decision being conveyed. Conversely, a ruler's genuine mark on an internal record did not mean that every required communication had already been issued.

This helps explain why the distinction between an autograph and a forgery is insufficient for understanding documentary government. There was also authorized production by intermediaries. To test a suspect document of that kind, knowledge of the ruler's handwriting alone would not answer the question. One would need to understand which officials could communicate a decision and what form their communication should take.

The two surviving records do not provide a complete, matching register of all business in those years. Their more specific contribution is to reveal an arrangement in which approval, drafting, dispatch, and retention could be separated. That made action at a distance possible while multiplying the places at which a mistake or an unauthorized intervention could enter.

The kaō gave a document a recognizable personal mark without requiring a plainly written name. Its execution could also signal the individual's participation in an act of communication.

Historian Kubota Masaki distinguishes that participation from the use of a seal that could be applied by an authorized subordinate. In his account of medieval documents, the choice between handwriting and a seal can help reveal the relationship between sender and recipient and the handling of the message.^[4] The distinction concerns how authority was expressed; it should not become a rule that every handwritten document was sincere or every sealed document impersonal.

Masamune's postscript belongs in this setting. A signature mattered enough for an irregular version to need an explanation. Yet his explanation also asked its recipient to look beyond the irregular mark. Recognition could involve the rest of the writing and the circumstances of communication, rather than a mechanical requirement that each stroke duplicate an earlier example.

By the late Tokugawa period, another practice deliberately combined stamping and writing. The National Archives describes kagoji signatures on lords' letters to the shogunate: the outline of a kaō was stamped, then filled with ink. Examples survive among records transmitted through the shogunate's Tamon Yagura archive.^[5]

The result complicates a modern attempt to sort every mark into either a handwritten signature or a seal impression. A stamped framework could be part of making a formal signature. That method tells us how the mark was produced; by itself, it does not identify every person who handled the document or establish whether the communication was authorized.

The recurring requirement was consequently familiarity with accepted practice. A perfectly unfamiliar method might arouse doubt; an accepted method could include assistance or mechanical reproduction. The physical trace had to be interpreted in the documentary world to which it belonged.

There was no single, timeless Japanese preference for one technology of trust. The provincial impressions on the Gufukuji record, the shogun's internal approval, Masamune's faulty kaō, and a later outline stamp established different relationships between a person, an office, and the page. Their meanings depended on those relationships, not simply on the presence of ink.

A seal could only be recognized if someone knew what to recognize. A small document preserved in Nagano shows an institution supplying that knowledge.

It records the names and seal impressions of Hayashi Jūzaemon and Kishi Chūbei, officials concerned with a lord's journeys. The accompanying communication was sent to three local contractors involved in providing transport services. The library catalogue tentatively dates it to the Kansei–Kyōwa period, 1789–1804.^[6]

The sheet's business was the specimen itself. Rather than waiting for recipients to meet an unfamiliar impression on an operational document, the administration sent examples connected to named officers. Those examples provided a reference against which later instructions or receipts could be compared.

This was a different solution from expecting everyone to recognize a powerful person's handwriting. Knowledge could be supplied in advance to particular people who needed it. A contractor did not have to acquire a general education in every official's seals; the relevant names and impressions could be transmitted through the existing relationship.

The reference also had its own vulnerability. A recipient had to trust the source of the specimen. Matching a later mark to an unauthorized example would reproduce the original error rather than expose it. The logic of authentication therefore extended backward into the delivery and custody of the reference sheet.

The surviving object does not include a sequence of contractor decisions showing how often a comparison was made. What it preserves is a means of making comparison possible. It connects a mark to a named officer and places that information in the hands of someone expected to handle documents from the administration.

The case identifies an important development in scale without requiring a national system of identical registers. Personal recognition could be supplemented by a maintained reference. Authority traveled more dependably when the receiver had something besides the arriving document against which to judge it.

Meiji notarial rules addressed the problem through a designated public office. The regulations issued in 1886 prescribed how a notary should establish who had requested an instrument, what the instrument should contain, and how it should be retained. The procedure began before a signature reached the page.^[7]

Article 28 required the notary to know the requesting person's name and face. If that personal acquaintance was absent, the rules supplied alternatives: certification from specified local officials, or identification by at least two adults whom the notary knew. The rules also required an adult witness for the making of the instrument. The following article excluded relatives and employees of the notary or requesting person, among others, from serving as that witness.

These requirements did not simply replace personal knowledge with bureaucracy. They organized ways of extending personal recognition. Someone unknown at the notary's desk could be identified through people or an official certificate that the procedure accepted. The question was who would stand behind the identification.

The document then had to record more than its substantive bargain or declaration. Article 30 required identifying details for the participants, the place and date of production, and relevant information about representation. A person acting for someone else needed a basis for that authority; identifying the agent did not establish the principal's instruction by itself.^[8]

The rules also treated the wording as something that could be corrupted after an apparently proper beginning. Blank spaces within connected writing were to be filled with ink lines. Additions and corrections had to be located and recorded. Deleted characters were to remain readable, with the deletion documented and sealed. The prescribed procedure preserved a visible history of amendment instead of allowing the altered page to masquerade as untouched.

Once prepared, the instrument was to be read aloud to the parties, and that act recorded. Signatures and seals followed. Joins were also to be sealed. Each operation addressed a different risk: a mistaken identity, an unauthorized agent, an unnoticed change of wording, or the substitution of a sheet.^[8]

These were legal requirements, not evidence that every notary always complied. Their value for the historical comparison lies in the way they distributed the work. The maker of the instrument had duties toward the identification of people, the recording of consent, and the physical integrity of the result.

A notary's seal consequently meant more than the recognition of a particular design. It was supposed to indicate that specified work had been done before the finished document left the procedure. The reader of the paper could not personally repeat every encounter. The office and its obligations gave those earlier acts a continuing point of reference.

The Notary Act of 1908 expressly included authentication of private documents among notaries' functions. It distinguished that task from drawing up a public notarial instrument. A privately prepared paper could acquire a certificate concerning its execution without being rewritten as the notary's own instrument.^[9]

Article 58 specified the occasions for authentication. The person could sign or apply a seal before the notary, or acknowledge an existing signature or seal. The notary recorded which act had occurred. For a copy, the task was comparison with the document and a statement that the two corresponded. Visible alterations, damage, or conspicuous grounds for suspicion had to be noted in the authentication.^[10]

That distinction made the certificate's object important. Witnessing someone acknowledge a signature establishes a relationship between that person and the document. It does not turn every assertion in the text into an independently investigated fact. Comparing a copy establishes correspondence with another document, rather than resolving every question about the underlying transaction.

The act also recorded the authentication outside the authenticated sheet. An entry in a register identified the document, relevant participants, method, and date. The document and register were to be linked by a seal impression. A certificate was therefore part of a recorded event, not merely an attractive mark added to someone else's paper.^[10]

This offers a useful comparison with the earlier specimen sheet. There, the receiver obtained something against which to recognize an official impression. Here, the notary's register preserved information about a particular act of authentication. The former established an expected mark; the latter supplied a record of what had been done with a particular document.

Both arrangements depended on information beyond the page in immediate circulation. They differed in the information retained and in who was responsible for it. The Meiji statute made those responsibilities part of an office whose procedures could themselves be examined.

Preservation introduced a problem that neither an accurate signature nor a properly performed ceremony could eliminate. Paper could disappear. An authentic instrument might be needed after its maker died, the office changed hands, or its original was destroyed.

The 1908 act required original instruments, attachments, and designated records to remain in official custody, subject to stated exceptions. It also prescribed controlled registers. Before use, a register's number of leaves was to be recorded by the district court's chief, and its joins officially sealed. Entries then followed the sequence in which instruments were made.^[11]

The register was itself being protected against alteration. A list of authentic acts would be less useful if pages could be removed or inserted without disturbing anything. Sealing the record's structure extended the same practical concern visible in the eleventh-century document into another institutional setting. Similar treatment of paper did not mean the two legal systems were identical.

Most revealingly, Article 42 provided for the loss of an original. The notary was to obtain an already issued authorized copy and, with approval from the chief of the district court, preserve it in place of the lost instrument. The replacement had to record the approval and its date, with the notary's signature and seal.^[12]

The law did not pretend that this made the first sheet reappear. It established a controlled substitution. The copy's earlier authorized production, its identification, and the new approval allowed a surviving document to assume a role previously held by another.

Continuity also required someone to take over the files. Further provisions regulated sealing and transferring records when a notary died or left office. Successors had to identify their capacity when acting upon documents made by predecessors.^[12] The memory of an act was being entrusted to an office and a sequence of custodians, rather than to the lifespan of its original witness.

Japan's documentary practices therefore supplied no single test that made fraud impossible. They assembled checks around different vulnerabilities: identifying an author, recognizing an official act, keeping approved wording together, and maintaining a record beyond the immediate exchange. Which check mattered depended on what the document was supposed to accomplish.

Masamune's awkward signature remains a useful warning against expecting the mark to do all that work. A genuine communication could contain an imperfect sign. A perfectly reproduced sign could still be used without authority. Dependability came from the connections that made a document answerable to people, procedures, and other records—not from the shape of the ink alone.