A Genealogy of the Crow Letters of Kumano

The Crow Letters of Kumano

Talisman, Oath, and Story: Six Hundred and Fifty Years

熊野の烏文字 — 護符・誓約・物語の六五〇年

The three grand shrines of Kumano issue a talisman known as the Goōhōin — literally, the “Ox-King Treasure Seal.” At Hongū Taisha, eighty-eight crows interlock to form its design; at Hayatama Taisha, forty-eight; at Nachi Taisha, seventy-two. The birds assemble themselves into five characters, read as Kumano-san hōin (“Treasure Seal of Mount Kumano”) at Hongū and Shingū, and as Nachi-taki hōin (“Treasure Seal of the Nachi Falls”) at Nachi.¹ From the late Kamakura period onward, the reverse side of these sheets came to serve as the paper for kishōmon — written oaths — and by the Edo period the practice had grown into a pillar of warrior-class documentary culture, extending to the accession pledges sworn to each new Tokugawa shōgun and the oaths of office submitted within the great daimyō houses.²

Break an oath sworn to the Kumano Gongen, it was believed, and a crow — divine messenger of the god — would die: one crow, or by another account three. The oath-breaker, too, would vomit blood, die, and fall into hell.

But since when has this legend existed? When was the design in which crows assemble themselves into writing established? Is it an unbroken tradition reaching back to the Kamakura period, or the product of a particular moment?

This essay traces, against the surviving primary sources, how the crow letters of the Kumano Goōhōin — and the folk belief that “the crows die” — came into being. The conclusion may be stated at the outset. The crow-letter design was established at the end of the medieval period, between the mid-fifteenth and early sixteenth centuries. The specific folk belief that “the crows die,” by contrast, is first made explicit in the written record in the mid-Edo period, between the end of the Genroku era and the Kyōhō era (1705–1720). Between the two lies an interval of roughly two hundred years.

According to Aida Nirō’s comprehensive survey of the shinmon — the invocatory clauses of medieval oaths, which spell out the punishments awaiting the perjurer — no formula of dying crows can be confirmed in any oath from the medieval period through the early Edo years.³ Two centuries separate the birth of the design from its telling as a story. This essay follows the history that fills that gap, distinguishing throughout between what the sources document, what can be strongly inferred, and what remains unresolved.

I. 1266 — Before the Crows

The oldest surviving Japanese oath written on a goōhōin dates to 1266 (Bun’ei 3). The paper in that instance was a goō of the Nigatsudō hall at Tōdaiji; oaths written on Kumano goōhōin are thought to have begun appearing at roughly the same time.⁴

The Kumano goōhōin of this era was not the “crow letters” we see today. It followed what is called the jiyō, or “character-form,” style: the five characters simply written out in a relatively plain hand. The crows are not yet there.

Consider the typical shinmon of a medieval oath. In the oldest surviving example, from 1266, it unfolds in outline as follows:⁵

“Beginning with the Great Buddha and the Four Deva Kings, the Three Hachiman, Kasuga Gongen, protector of the Buddhist Law, the Eight Great Deities invoked at this temple, the great deity Amaterasu 天照, Kamo, Kitano, the Kumano Gongen, the Kongō Dōji of the ten directions … may [the undersigned] incur their divine punishment.”

Here the Kumano Gongen appears as one deity among the many invoked. Punishment is expressed collectively, as “the punishment of the assembled gods and buddhas” — and nowhere is there any specific mention of crows dying.

In his study “On the Goōhōin as Paper for Written Oaths,” first published in 1943, Aida Nirō examined more than one hundred oaths spanning roughly 350 years, from 1266 (Bun’ei 3) to the Genna era (1615–1623). Within that entire corpus, not a single shinmon contains the statement that one crow — or three — will die.⁶ The later, well-known legend that “the crows of Kumano die” simply did not exist within medieval oath culture. The crows were not yet present — not in the design, and not in the story of punishment.

II. 1394 to 1492 — Ordeals of Swallowing the Goō

Throughout this period, the Kumano goōhōin itself was gradually adopted as the paper of choice for written oaths. According to the table of medieval ordeals compiled by Ōkōchi Chie in 2021, multiple cases from the late thirteenth through the sixteenth century document a procedure in which the goō bearing an oath was burned to ash, mixed into sacred water, and swallowed.⁷

The earliest secure example is the seclusion oath of Kuun 救運, a monk of Tōji, in 1394 (Ōei 1). When rumors of “impurity” arose around him, an ordeal to determine the truth was held at his own request. Following the council’s decision, Kuun wrote two oaths on goō paper: one was sealed away in the Fudōdō hall; the other he burned and swallowed on the second day of his seclusion.⁸

Such goō-swallowing ordeals spread through the later medieval period and into the early Edo years. In 1492 (Entoku 4), at the Outer Shrine of Ise, an ordeal was conducted in which shrine attendants were made to swallow Kumano goō in order to identify a thief who had entered the offering hall.⁹

What deserves notice is that even at this stage, “the crows die” appears nowhere in the content of the shinmon. Procedures relying on the talismanic power of the goō were taking hold — but more than three centuries would still have to pass before that power found narrative expression in the death of crows.

III. 1505 — The Crow Letters of Kanshinji

The oldest surviving example of the crow letters is the joint oath of the scholar-monks of Kanshinji, dated the fourteenth day of the third month of 1505 (Eishō 2).¹⁰

The document was drawn up to resolve an internal dispute at Kanshinji, a mountain temple in Kawachi province (present-day Kawachinagano, Osaka). When a suit concerning permission for lower-ranking monks to undertake the nyūbu — the Shugendō mountain pilgrimage — was brought to the temple by the deputy governor, Yusa Jirōzaemon-no-jō, the scholar-monks rejected it as contrary to “the seals of successive provincial administrations” and “the immemorial regulations of the temple house.” Thirty-one monks set their signatures jointly to the oath, pledging solidarity.

Its physical construction is as follows: the oath was transcribed in reverse on the back of a Kumano goōhōin, which was then affixed at the place of the joint signatures. The shinmon at the document’s close unfolds thus:¹¹

“Beginning with Bonten and Taishaku, the Four Deva Kings, and all deities great and small throughout the three realms — and in particular the Kannon Kariteimo of this temple 觀音賀利帝 and the tutelary deity of Nishinomiya 西宮地主権現, together with the invoked Five Great Deities — may [the undersigned] incur their divine punishment; thus is the oath, as stated.”

Bonten, Taishaku, the Four Deva Kings, the deities of the three realms great and small; above all the Kannon Kariteimo enshrined at Kanshinji (Kariteimo — Hārītī, the child-guardian goddess), the tutelary deity of the West Shrine (Nishinomiya jishu-gongen), and the invoked Five Great Deities — theirs is the punishment the oath calls down. Here too, there is no phrase about crows dying.

What makes this 1505 document decisive, however, is that it physically preserves the oldest extant crow-letter Kumano goōhōin. By the end of the medieval period, the Kumano goō had already reached a major temple in Kawachi — evidence of nationwide circulation through the networks of the Kumano bikuni nuns and the oshi pilgrimage agents.

In terms of design, the studies of Shimazu Norifumi and Aida Nirō allow the Kanshinji document to be placed within the transition from the mid-fifteenth-century character-form style (last confirmed in 1445, Bun’an 2) to the “crow-dot jewel” style of the early sixteenth century.¹² The upper bound for the emergence of the crow letters is thereby fixed, for now, at 1505.

IV. From the Shinmon to the Goō — The Logic of the Turn of the Early Modern Era

Among the important arguments advanced in Ōkōchi Chie’s 2021 study is a shift in the agent of punishment.

In the medieval understanding, punishment was delivered by the gods and buddhas enumerated in the shinmon. The goō was the paper on which those deities were made present — a device for adding weight to the vow.¹³ With the coming of the early modern period, however, the goō itself came to be understood as the agent that punishes: a movement from “the punishment of the shinmon” to “the punishment of the goō.”

Consider an example from 1628 (Kan’ei 5). When the Shintō priest Shinryūin Bonshun had a sword stolen, he had the servant arrested as the culprit write an oath, made him swallow a burned Kumano goō, and then released him. To administer the oath and the goō after the theft had been admitted is a procedure aimed at preventing future breach. The underlying idea is that punishment resides in the goō itself as a charged object, and that taking it into the body secures the deterrent.¹⁴

In 1634 (Kan’ei 11), the characters Nihon daiichi — “First in Japan,” the same phrase we see today — appear on the Kumano goōhōin for the first time. The formularization of the title “First Great Site of Miraculous Power in Japan” is thought to reflect the Shugendō networks that had persisted since the Warring States period; by 1709 (Hōei 6), the present style is considered to have been essentially fixed.¹⁵

In short, the consolidation of the crow-letter design, the recognition of the goō as the agent of punishment, and the settling of the “First in Japan” formula that survives today all took place between roughly 1505 and 1709 — a span of some two hundred years.

V. At the Institutional Core of the Tokugawa State

What matters here is that the goō oath of the early Edo period has often been described in terms of “formalization” and “decline” — and that this description is contrary to fact. As Ōkōchi Chie demonstrated in her 2014 study, the Tokugawa shogunate, from the reigns of Ietsuna and Tsunayoshi onward, institutionalized the submission of oaths on Kumano goō at each shogunal succession and at the appointment of daimyō and shogunal retainers to office.¹⁶

Within the daimyō houses as well, retainers assuming office submitted oaths to their lords written on Kumano and other goō, as a matter of institutional practice. Assessing Ōkōchi’s finding, Hara Michio writes that “the basic governing apparatus of the Edo period — the Tokugawa shogunate and the daimyō houses, together with their bureaucracies — was, to put it a little grandly, supported in each of its principal details by oaths written on goōhōin.”¹⁷

The goō oath of the Edo period was no marginal religious custom of warrior society; it was a documentary practice constituting the core of the machinery of rule. The Kumano goōhōin stood at the intersection of ritual, politics, and law.

VI. The Sign at the Nuns’ House — Edo, 1714

The circulation of the goō was not confined to warrior institutions. Within the urban culture of Edo, the Kumano goō reached even the pleasure quarters and the theater.

An illustration in the ukiyo-zōshi novel Aikyō mukashi irogonomi, published in 1714 (Shōtoku 4), depicts a bikuni house in Izumi-chō, Edo. Posted on its lattice is a sign reading “Kumano goō available.” The picture shows a bikuni attended by two young novices, making her procession to an assignation house in the manner of a Yoshiwara courtesan.¹⁸

The Kumano bikuni combined three roles: religious specialists who performed pictorial recitations of the Kumano foundation legends; proprietors of houses of assignation engaged in prostitution; and vendors of the goōhōin. Working in concert with the networks of the Kumano oshi and the yamabushi, they sold goōhōin throughout the city of Edo.

This bikuni network would later prove significant. By 1714, the ground had already been prepared for the goōhōin to become the paper on which courtesans and their clients exchanged vows.

VII. The Crows Enter the Story — 1705 and 1720

Once the goō oath entered the pleasure quarters and the theater, a dramatic question surfaced: what happens when the vow is broken? It is here that the belief that “the crows die” appears in the written record in verifiable form.

The clearest primary source is Chikamatsu Monzaemon’s Shinjū ten no Amijima (The Love Suicides at Amijima), first performed at the Takemoto-za in Osaka in the twelfth month of 1720 (Kyōhō 5). In the final act, Jihei, resolved on the double suicide, speaks to Koharu:¹⁹

“Listen — do you hear them? The crows, come to guide the two of us to the land of the dead. They have said since long ago that each time a single sheet of goō paper is written, three crows of Kumano die on the sacred mountain. You and I, at the fresh dawn of the new year, wrote our first oath of the year…”

Here for the first time, the specific number and mechanism — “each time a single sheet of goō paper is written, three crows of Kumano die on the sacred mountain” — are made explicit in a written text. Note the phrase “they have said since long ago.” Chikamatsu has his character speak of this not as a novel invention, but as a tradition the audience already knows.

By 1720, in other words, the legend was current enough among Osaka audiences to pass as something “already known.” Reckoning backward, the belief must have formed earlier.

How far back can that “earlier” be traced? In his 2015 study, Hara Michio takes up Chūjōhime mittsu no kuruma, staged at the Nakamura-za in Edo in 1705 (Hōei 2). The play’s script does not survive, but an illustration and commentary remain in the actor-critique booklet Yakusha sanzesō, published in the fourth month of that year. Deciphering the illustration, Hara reconstructs the staging as follows: in a scene where the samurai Kume Hachirō, taken captive, is forced to write a false confession as an oath upon a Kumano goō, his wife Shiratama bites through her finger to supply the blood seal in his place — whereupon Shiratama is tormented by three crows.²⁰

Hara himself frames this reconstruction cautiously — “it appears,” “it may reasonably be supposed.” With the script lost, it cannot be asserted that the words “three crows die” were spoken as dialogue. That a scene of three crows tormenting an oath-breaker was realized on stage, however, is strongly supported by the illustration.

The performance records of Chūjōhime plays compiled by Torii Fumiko in 1987 lend Hara’s reading further support. Following the revival of the Neri-kuyō procession ceremony at Taimadera in 1700, Chūjōhime plays were staged in great numbers in the puppet theater and kabuki — and 1705 falls at the height of that vogue.²¹

|—|—|—|

For at least two decades, audiences had lived with the Chūjōhime plays. It is natural to suppose that the motif — a false oath written on a goō, punished by crows — was presented within that context.

What must be stressed is this: even granting a staging in which three crows torment the perjurer, it was a representation formed within the space of urban theater. The goō oath as institutional document of warrior society — the shogunal succession pledges, the daimyō oaths of office — never incorporated the language of dying crows. The crow-death belief was the product of a particular cultural sphere, crystallizing where urban theater and the urban bikuni network intersected.

VIII. Texts of Authorization — 1732 to 1822

The three-crows motif made explicit in Chikamatsu’s text would be reproduced again and again through the eighteenth and nineteenth centuries.

In 1732 (Kyōhō 17), the Shugen koji binran, a compendium of Shugendō lore, saw its first printing.²² The kabuki play Koibikyaku Yamato ōrai, in its Kameya act, staged in 1795 (Kansei 7), contains the line:²³

“Drink the ashes of a burned Kumano goō upon a lie, and you will straightway vomit blood and die.”

Here it is not the crows that die, but the person, vomiting blood, who stands in the foreground. Within the same complex of belief, who dies — the crows or the oath-breaker — shifts with the context.

In 1822 (Bunsei 5), Fujiwara Michisuke of Sonshōin, a pilgrims’ lodge at Nachi, compiled an origin account entitled Kumano-san Nachi karasu goō kiryū narabini meigi (“The Establishment and Meaning of the Crow Goō of Nachi, Mount Kumano”).²⁴ It ties the design of the crow goō to the legend of Emperor Jinmu’s eastern campaign, setting it within the sacred lineage of the Yatagarasu, the great crow of myth. Viewed critically, however, this is retrospective meaning-making performed in 1822 — an after-the-fact interpretation that seeks to push the origin of the crow letters back into mythic time. The 1822 account cannot be used to claim an ancient origin for the design.

These texts progressively authorized the crow letters as “a sacred design of ancient origin.” From the standpoint of documentary history, however, the oldest verifiable example of the crow letters is the Kanshinji document of 1505, and the oldest explicit statement of the crow-death legend is Chikamatsu’s text of 1720. The 1822 engi postdates the design’s emergence by some three hundred years, and the legend’s first attestation by about one hundred.

IX. The Crows in Folk Memory — Minakata Kumagusu

By the modern era, the crow-death belief was no longer confined to certain plays and sermons; it had settled into the everyday memory of the people of Kishū — Wakayama.

In his 1916 (Taishō 5) essay “The Name of the Goō and the Folk Beliefs of the Crow,” Minakata Kumagusu records his own childhood memories:²⁵

“I saw the Kumano goō often as a child, and whenever something went missing at primary school we were threatened with being made to swallow it…” “To detect a thief, I heard, one burns it to ash and drinks it in water; then the crows that dwell at the Kumano shrine die in the same number as the sheets burned, and the punishment strikes home in the guilty one, who at once vomits blood.”

Minakata was born in 1867 (Keiō 3); his childhood falls around 1875. In the primary schools of the Wakayama castle town, children were threatened with “being made to drink the goō” whenever an item went missing — the swallowing of the goō invoked as a technique for smoking out a thief. The formula “as many crows die as sheets are burned” is simpler than Chikamatsu’s “three die for each sheet written,” and posits a more direct correspondence.

Minakata’s account shows the belief written into Chikamatsu’s play in 1720 fixed, some 150 years later, as oral tradition in the Kishū region. A theatrical motif inherited as local folk memory — and at the end of that path stands Minakata’s childhood.

X. One Crow, or Three?

Finally, let us return to the question with which this essay began. When an oath is broken, does one crow die — or three?

The current official explanation of Kumano Hongū Taisha reads: “Break a vow sworn to the Kumano Gongen, and one crow — messenger of the great deity of Kumano — dies; the oath-breaker too vomits blood and falls into hell.”²⁶ The official accounts of Hayatama Taisha and Nachi Taisha broadly agree.

Yet Chikamatsu’s 1720 text states explicitly: “three die each time.” And the 1705 Chūjōhime mittsu no kuruma, if Hara’s reading is followed, brought three crows onto the stage. Plainly, the early modern urban-theatrical lineage presupposed three.

Hara notes in his 2015 study, parenthetically, that “there appear to be regions where the count is ‘one’ rather than ‘three.’”²⁷ This is a critically important observation. The one-crow type and the three-crow type coexisted, varying by region and context.

One possible configuration is this: in its initial form, the belief circulated as local oral tradition in both one-crow and three-crow variants; the three-crow variant was dramatically amplified in eighteenth-century urban theater; and when the shrines later codified their official explanations, the account converged on the one-crow type. Hongū Taisha’s modern explanation of “one crow” may thus continue the lineage of regional tradition, formed along a path independent of the Chikamatsu-kabuki line.

Another possibility is that Chikamatsu’s “three” was devised as a symbolic number corresponding to the three shrines — Hongū, Hayatama, Nachi — and took root in the world of the theater. This remains conjecture.

To the question “one crow or three?” there is at present no settled answer. What this shows is that the traditions surrounding the crow letters are not a homogeneous system with a single origin, but a cultural phenomenon built of accumulated layers, in which multiple lineages run in parallel across region, class, and medium.

Conclusion — What the Crow Letters Ask

Tracing six hundred and fifty years in the making of the crow letters of the Kumano goōhōin, several historical facts come into relief.

First: the crow-letter design does not descend from antiquity. It is a comparatively recent creation of the late medieval period, its oldest surviving example dated 1505.

Second: the specific belief that “the crows die” appears nowhere in the shinmon of medieval oaths. Its oldest explicit attestation in writing is Chikamatsu’s Shinjū ten no Amijima of 1720.

Third: roughly two hundred years separate the design’s emergence from the belief’s attestation. That gap is filled by several movements at once — the Tokugawa institutionalization of the goō oath, the expansion of the urban bikuni networks, and the revival of the Taimadera Neri-kuyō with the theatrical flowering of the Chūjōhime plays that followed.

Fourth: the difference between “one crow” and “three” reflects the coexistence of two lineages — regional tradition and urban theater. It should be understood not through the search for a single origin, but as a process in which multiple layers accumulated into the present.

The Kumano goōhōin is not a fixed relic of the past. From the medieval period into modernity it has gone on transforming — its meaning renewed again and again — as talisman, as oath paper, as narrative motif, as folk memory. A single sheet of paper with so many-layered a cultural career is rare anywhere in the world.

To look at the crow letters is to look at those layers, one resting upon another.

Notes