Kyoto’s Gion Festival in Late Classical and Medieval Times: Actors, Legends, and Meanings

Abstract

Kyoto’s Gion festival has arguably the best-documented history of all festivals (sairei) in Japan, and studies of its development have heavily influenced our understanding of festivals in general. Yet we must expect that our knowledge of this history is partial at most. Extant archives on its late classical and medieval history derive from a narrow group of festival actors, and are therefore intrinsically biased. This article looks at current reconstructions of the festival’s origin and development, addressing primarily the following questions: Which groups of actors are the historical record hiding from us? Is there a world of ritual action, beliefs, and concerns that we are missing entirely? Origin legends have been used throughout history to attribute meaning to the festival procedures. Today as in the past, these legends are always accompanied by narratives of continuity: at its core, it is implied, the festival remains unchanged. Such legends reflect the interests of actors and patrons of different ages, and changes in the festival’s context have required origin tales to be updated or even replaced. What do such narrative innovations reveal about the festival’s changing place in society at different historical junctures? Do such legends contain traces of the activities of actors who have since disappeared, taking their archives with them?

1. Introduction

In 2019, Kyoto celebrated the 1150th anniversary of the Gion festival with great pomp. The main commemorative event took place in the grounds of Nijōjō 二条城 castle, close to the ancient site of the Shinsen’en 神泉苑 palace garden. According to the Kyoto City Official Website, this garden was the place where the festival’s history had begun.1 The website cites what today is the official line, repeated in pamphlets, guidebooks distributed by Yasaka Shrine, the Floats Association (Yamahoko rengōkai 山鉾連合会), and Kyoto City—though not in publications by the Agency for Cultural Affairs (Bunkachō 文化庁), which is more bound to historical correctness. In 869, the story goes, a goryōe 御霊会 or “ritual [for the appeasement of] wrathful spirits” took place in the Shinsen’en garden. Sixty-six halberds (hoko 鉾) were set up to represent the spirits of Japan’s provinces, and people gathered in a grand ceremony to dispel the epidemics that were ravaging the country. This was the origin of the Gion festival on which the 2019 commemoration was based.

On 8 June 2019, a month before the main events of the Gion festival, eleven groups from different corners of Kyoto Prefecture carried their halberds to the Nijōjō. The halberd-bearers manipulated long poles topped with shiny blades and bells, called kenboko 剣鉾 or “blade halberds”. Skilfully balancing these poles as they walked, the men caused the blades to sway and the bells to chime. Kenboko halberds have no role in the Gion festival,2 but feature in a wide range of festivals in the Kyoto area. There, they typically lead the mikoshi processions that carry the gods to and from the otabisho 御旅所 (the sites of temporary enshrinement for the duration of the festival). The 2019 celebration reflected and reinforced common understandings of the origins and meanings of the Gion festival. In particular, it connected the floats parade with the sixty-six halberds that, allegedly, were central to the 869 goryōe. The largest Gion floats, after all, are also known as hoko (“halberds”), although the only feature that they have in common with the kenboko is their soaring poles. The hoko floats of the Gion parade are large carts designed around such poles (shingi 真木); but none of these poles are topped by a halberd blade.3

I would be the last to insist on historical accuracy as a prerequisite for festival innovation. The implication of the Nijōjō event—that the kenboko halberds are the “missing link” tying the 869 goryōe to the Gion festival of 2019—was certainly creative.4 From the perspective of an historian, such claims of longue durée continuity are both disconcerting and acutely suggestive. Disconcerting, because the assumption that the festival has maintained a continuous core of stable practice and meaning fails to do justice to its rich and complicated history; suggestive, because an event such as this allows us to witness a public and influential instance of imaginative resignification. Such moments of resignification have occurred many times in the festival’s history, and many of the most suggestive sources about the festival that have survived to this day were compiled or created in similar contexts. Such sources use (and, in many cases, rig or even forge) surviving scraps of evidence in a highly selective manner, with the express aim of constructing narratives of continuity and deriving meaning from origin stories—including the 869 legend.

In spite of the fact that Kyoto’s Gion festival is arguably Japan’s best-documented festival (sairei 祭礼), a hard look at extant sources reveals so many “known unknowns” that one cannot but wonder what crucial “unknown unknowns” we are missing. This is doubly problematic because few festivals have been studied in as much detail as the Gion festival. As a result, other festivals tend to be viewed through a Gion lens, as variants of or contrasts to this best-known specimen of the genre. If this lens is warped, our understanding of all festivals will be skewed. What is erased from the history of this particular festival tends to disappear from view also in studies of other festivals with even more fragmentary historical records.

Writing about the Ise Shrines, Fabio Rambelli notes that in mainstream accounts of the history of this shrine complex, “emphasis on tradition clearly belies multiple and conflicting palimpsests of re-invention of tradition” (Rambelli 2014, p. 222). He points out the tendency to claim a continuity of “signified (meanings)” on the basis of a perceived continuity of “signifiers (perceptual forms and actions)”. Rambelli concludes that Ise has throughout the ages functioned as a “floating signifier” that has been reconfigured whenever the cultural and discursive circumstances required resignification. Emphasis on continuity is equally striking in the literature on the Gion festival. Festivals are more fluid than shrines, making it even more difficult to maintain such a narrative of continuity. Not only ascribed meanings but also the “signifiers” themselves tend to shift more frequently and radically than the codified “forms and actions” of a site such as Ise—although Ise too has been subject to profound changes over time.

In our history of the Ise Shrines, John Breen and I seek to pinpoint and analyse eight historical moments of radical resignification (Teeuwen and Breen 2017). We point out that resignifications typically occurred when central agents were marginalized or even banned from the site for economic or political reasons. We propose that eight times in Ise’s history, the arrival of new actors, introducing new models of patronage and addressing new audiences, occasioned major changes in the shrines’ signifiers and meanings. Such historical watersheds have often become invisible by the disappearance, or active erasure, of the practices of actors who are no longer present. Also, these moments of change usually led to the creation of new origin stories about the shrines, borrowing from older narratives where possible but shifting the emphasis in a manner that suited new groups of actors and patrons and fitted new socio-economic circumstances. Naturally, archives tend to perish when the practices that they document and the actors who attached value to those practices disappear from the scene. As with Ise, the Gion festival was prestigious enough to survive numerous crises. This resulted in a historical record that is full of both holes and fanciful apocrypha, some of which continue to be cited to this day as though they were reliable primary sources. The 869 legend is a telling example.

In this article, I explore this and similar origination legends in order to excavate the actors and agendas that gave rise to them. In the process, I hope to peel away the mainstream continuity discourse and gain at least a fleeting impression of the roles of once central groups of actors that have since become defunct. I will start by providing a brief summary of the roles of different actors in the Gion festival prior to its discontinuation in 1467. Then, I take stock of the main archives on which our understanding of the festival’s early history is based. To shed light on the processes of selection and manipulation that have shaped the body of sources on which we must depend, I will return to the narrative of the 2019 commemoration. The 869 legend that now has become established as the festival’s official origination narrative came about as a result of a sequence of disparate agendas. The multi-stage process that led to the creation and selection of this legend helps us understand how such agendas lifted certain historical sources to prominence, while silencing others.

Finally, this discussion will lead me to the question what medieval constituents of the festival have been obscured as once prominent groups of actors were superseded and deposed by others. In particular, I highlight the otabisho as formative sites of the classical and medieval Gion festival. The otabisho, their priests, and their networks served as the interface between the shrine and the city. Yet we know very little about their role in different phases of the festival’s development, since the otabisho lost their autonomy in the fifteenth and sixteenth centuries. My hypothesis is that we would have had a much better understanding of the festival if we had known more of what went on at the otabisho before this loss. The otabisho, one might argue, are an obvious “known unknown” that is all but absent not only from extant archives of primary sources, but also from the now dominant 869 origination legend.

2. The Pre-Modern Gion Festival and Its Actors

Who were the main actors of the Gion festival, and what sources did they leave us? In a rough outline, we can divide the pre-modern development of the festival into three phases.5

In its first phase, the festival developed from interim performances to an annual event, structured around three dates. This transition likely happened in the 970s. On the seventh day of the sixth month, three mikoshi palanquins carried the gods of the Gionsha (the pre-Meiji name of Yasaka Shrine) to the Ōmandokoro 大政所 otabisho, which took up a city block along Takatsuji and Higashi-no-Tōin streets.6 This event was called mikoshi mukae 神輿迎, “welcoming the mikoshi”. The Gionsha gods had many identities, but were most commonly conceived of as the male Tenjin 天神or (later) Gozu Tennō 牛頭天王; his wife Harisainyo 頗梨采女; and their sons, the Hachiōji 八王子 or “Eight Princes”. In the 1130s at the latest, a second otabisho was built further north, on the intersection between Reizei and Higashi-no-Tōin streets. The mikoshi of Harisainyo was rerouted to this site, called the Shōshōi 少将井 otabisho.

A week later, on the fourteenth day of the sixth month, the mikoshi returned to the Gionsha. This was referred to as the goryōe proper. This second mikoshi procession grew into an elaborate parade thanks to ever more lavish court sponsorship. Nobles were ordered to contribute extravagantly decked out horse riders (uma no osa 馬長 or mechō no warawa 馬長童), each accompanied by a group of retainers. Goups of dengaku 田楽 dancers, either formed by nobles themselves or sponsored by them, became a prominent part of the procession in the eleventh century. On the fifteenth day, finally, court offerings were presented at the Gionsha by an imperial envoy. On this day court nobles sponsored Gagaku dances (azuma asobi 東遊) and horse races (tōtsura 十列) at the shrine. The procedures on the fifteenth were referred to as rinjisai 臨時祭 or “interim rites”; they retained this name also after they had become an annual event.

The second phase began in the Insei period, when the court became caught up in internal strife. The festival had experienced its first heyday under the patronage of Retired Emperor Shirakawa around 1100, but as tensions within the court escalated, nobles’ interest waned. In 1157 Retired Emperor Go-Shirakawa sanctioned an already existing system of raising funds for the festival by granting three “sickle halberds” (kamahoko 鎌鉾, presumably one for each of the three mikoshi) to the Gionsha’s “shrine lineages” (shake 社家)—a term we will return to below. Every year, these halberds were assigned to “rich houses in the city” who were then obliged to pay a charge called “horse chief duty” (bajōyaku 馬上役).7 Similar sanction had already been given to Enryakuji on Mt Hiei some decades earlier. There, the bajōyaku was linked to a festival of the Hie Shrines, kosatsukie 小五月会, which was most renowned for its horse races (Shimosaka 2001, p. 261, note 1).

In this same period, Enryakuji developed a new method to defend its interests by carrying the mikoshi of Gion, Kitano, and Hie into the city. Such “deity attacks” often ended in pitched battles (Adolphson 2000, chp. 6). The Gionsha, especially, served as a vanguard base of Enryakuji in such conflicts. The twelfth century also saw the emergence of merchant guilds under the authority of temple overlords (Gay 2001). Enryakuji was the largest actor in this field, and membership of its network was rapidly becoming a prerequisite for any aspiring business. In his 1157 order, Go-Shirakawa presented the bajōyaku system as a means “to adorn the goryōe and enhance Tenjin’s powers”. In effect, the expansion of this Hie system to the Gion festival marked an exchange of actors, with Enryakuji and merchants connected to Enryakuji through guild structures entering the stage. These new actors put their hold over the festival to good use. In 1300, for example, the Gion festival was postponed not because of an incidence of impurity (as often happened with all festivals), but due to an unresolved lawsuit filed by Enryakuji, quite unrelated to the festival itself.8 Right up to Oda Nobunaga’s burning of Mt Hiei in 1571, the Gion festival was held hostage by Enryakuji in this manner.

The third phase was marked by changes in both the format of the festival and its relation to power. In the early fourteenth century, the uma no osa and dengaku groups gradually disappeared. They were initially replaced by “halberd groups” (hokoshū 鉾衆), drummers and dancers converging around a halberd that may or may not have resembled the kenboko halberds of later times. These groups not only accompanied the mikoshi but also made their own way through the city, at times getting caught up in fights. Halberd groups soon became part of most festivals around the city. It is hard to tell whether they originated at Gion, although it is there that they get their first mention (in 1321).9

It is also unclear how these groups relate to the pre-existing halberds of the bajōyaku and to the later hoko floats. In 1345, a diarist notes that due to rain on the seventh, “hill floats and other contraptions” (yama ika no tsukurimono 山以下之作物) were paraded through the streets on the eighth day of the sixth month, the day after the mikoshi mukae procession.10 The yama floats depicted theatrical scenes, artfully arranged on platforms carried (or sometimes wheeled) by a group of men. One imagines that the “contraptions” included other types of decorated floats and parasols (kasa 笠・傘), accompanied by costumed dancers and musicians. Such floats were now paraded through the streets before the mikoshi processions on the both the seventh and the fourteenth—though not when it rained.

The development of a separate parade of hoko halberds, yama floats, and more accelerated in the 1370s. In 1369, the mikoshi of both Hie and Gion were lost in a famous conflict between Enryakuji and the new Zen temple Nanzenji, which was closely associated with the Muromachi shogun (Adolphson 2000, pp. 307–15). It took decades before the mikoshi were replaced because Enryakuji demanded that the shogunate would foot the bill. Meanwhile, the festival continued without mikoshi processions. Shogun Ashikaga Yoshimitsu 足利義満 (1358–1408) made it an almost annual routine to put himself on display in a pavilion along the route of the Gion parade for formal “viewings” (kenbutsu 見物). This inspired the merchant sponsors of the parade to outdo each other with extravagant float designs. When the mikoshi were finally revived, the parades remained separate from the now less extravagant mikoshi processions.

Meanwhile, the shogunate sought to wrest control over Kyoto’s merchants from Enryakuji. A series of laws issued by Yoshimitsu in 1393 ended the power of temple overlords to grant their protégés exemption from taxation by others (Gay 2001, p. 81). Yoshimitsu also succeeded in limiting Enryakuji’s control over the Gionsha. In 1385, he installed his personal “prayer master” (oshi 御師), a Gionsha priest called Kenshin 顕深, as Gionsha’s manager (shigyō 執行), decreeing that this office was henceforth to remain in Kenshin’s lineage. Kenshin’s cloister at the Gionsha, the Hōjuin 宝寿院, would go on to dominate the Gionsha as the seat of the shigyō managers until Meiji.

In spite of these successes, however, Enryakuji maintained a solid grip on the Gionsha’s mikoshi. Its central weapon was the bajōyaku. This levy was revived in a new format in the 1380s, also under Yoshimitsu. The bajōyaku now became the permanent responsibility of a group of leading merchant houses known as the “bajō association” (bajō isshū 馬上一衆). This association consisted of the largest moneylenders and financiers of the city. These members collected contributions from hundreds of lesser businesses, ranging from sake brewers and miso manufacturers to bathhouse keepers. Under this new system there was only one bajōyaku that was shared between the Hie Shrines and the Gionsha, with Hie consistently receiving a much larger share. Whenever the bajōyaku was not paid in full, Enryakuji prevented the Gion mikoshi from leaving the shrine gates with military might. The same merchants who held the bajōyaku halberds as members of the bajō association were also responsible for collecting shogunal taxes, and the shogunal Board of Retainers (samuraidokoro 侍所) made sure that both taxes—the bajōyaku and the shogunal tax—were duly paid. In this scenario, the merchants of Kyoto, who both funded and provided manpower for the festival, found themselves navigating between their old overlord on Mt Hiei and the newly assertive Muromachi shoguns.

In the course of the Muromachi period, shogunal viewings grew into a ceremonial demonstration of warrior control over the city (Futaki 1985; Kawauchi 2012, chp. 1). The parades now outshone the mikoshi processions, while the classical rinjisai died a quiet death. Many of the floats that made up the parades were designed with the shogunal viewings in mind. Most displayed scenes from warrior lore, celebrating heroic deeds of the Minamoto and their allies, or from Noh, an art form that was favoured by the shogunal court. Others drew on episodes from Chinese classics, perhaps with a nod to the rise of shogunal Zen temples. Merchants participated in the festival for many reasons beyond faith in its deities. It was their time-honoured duty as jinin 神人, “service people” of the Gionsha, and it was expected of them as members of the powerful Enryakuji network of trade and finance. Also, the parades offered a unique opportunity to display their capital (economic, social, and cultural) in front of the shoguns and other leading warlords—and, of course, the crowds of fellow commoners that lined the streets.

The main actors that stand out in the festival’s pre-Ōnin history are the imperial court and the retired emperors, the Gionsha, Enryakuji, the shoguns, and, not least, the merchants who were assigned bajōyaku halberds in the Kamakura period and paraded their floats in the Muromachi period. Our reconstruction, however, is merely a reflection of available sources—available, thanks to the fact that they have survived in the archives of elite groups that lived on into modernity. As a first step towards conjecturing what we are missing, we must gain an understanding of the biases and blind spots of those archives.

3. Archives

Primary sources related to the Gion festival are relatively rich from the sixteenth century onwards. Particularly striking is the stunning record of high-quality images of the Gion festival on gilded byōbu screens, starting in the mid-1500s. In the Edo period there were illustrated guidebooks for visitors, detailed procedures and account books kept by the townsman officials of float streets, records of conflicts adjudicated by Kyoto’s city magistrates, and much more. Earlier periods, however, are less easily recoverable. There are occasional mentions in court histories and works of literature, and even an Edo-period copy of a screen depicting some of Kyoto’s festivals, including Gion (Tsukinami saireizu byōbu 月次祭礼図屏風). Aside from such isolated references and images, there are two main bodies of primary sources, each with its own merits and biases.

The first consists of brief notes about festival performances in diaries of court nobles. These diaries, known collectively as kokiroku 古記録 (“old records”), are often very terse, revealing little more than the mere fact that the festival took place. Sometimes, however, they convey impressions of the mikoshi processions, the rinjisai horse races, or the float parades. Scattered entries report incidents or discuss duties that the author must perform in connection with the festival. Diarists note that they attended the rinjisai (a), served as imperial envoys there (b), provided horses and riders for the races (c), or were prevented from doing so due to some form of pollution (d). We learn that they received orders to provide uma no osa for the goryōe on the fourteenth day (e), and at times were greatly inconvenienced by this (f). Some diarists went to great lengths to avoid crossing paths with the mikoshi in fear of impurity (g). Others note that they customarily ate fish after the mikoshi had passed (h). A few entries record private visits to the two otabisho, sometimes bringing offerings (i). Some mention that they viewed the parade on the fourteenth from a roadside pavilion (j).11

The main body of kokiroku is accessible through a searchable database, hosted by the Historiographical Institute of the University of Tokyo.12 This database, which is far from complete, contains well over two hundred entries related to the Gion festival between 982 and 1467. Coverage is concentrated in the years 1091–1117, 1197–1257, 1283–1307, 1363–78, and 1416–60; some of the gaps can be filled in with the help of diaries that are not included in the database. A fuller overview for the years 1321–1602 has been compiled by Kawauchi (2012, pp. 33–54).

The second rich archive consists of diaries, documents, and compilations kept at the Gionsha. Yasaka Shrine (the new name given to the Gionsha in 1868) has been particularly active in publishing extensive selections from its records. Yasaka jinja kiroku was first published in 1923, followed by the even larger Yasaka jinja monjo in 1939–40 (Yasaka Jinja Monjo Hensan Iinkai 2002). More materials from the Yasaka archive were added in 2014 and 2016 (Yasaka Jinja Monjo Hensan Iinkai 2014, 2016). Together, these collections offer sources that shed light on a stunning array of topics related to the Gionsha complex13.

Yasaka jinja kiroku sets out with Gion shigyō nikki 祇園執行日記, the diary of the Gionsha “shrine managers” (shigyō 執行) Kensen 顕詮 and Seiken 晴顕. It covers the period from 1343 to 1372 in great detail and also includes a selection of older documents kept at the Gionsha. This is followed by Gionsha ki 祇園社記, a compendium compiled by the shigyō Gyōkai 行快 (dates unknown) in the Hōreki period (1750–65). This work, which is only partly preserved, contains a fifteenth-century engi (origination myth) about the Gionsha gods; various collections of letters and documents (mostly from the Muromachi and early Edo periods) on appointments, landholdings, payments, incidents and conflicts; Edo-period maps of the Gionsha and its cloisters; and more. Yasaka jinja monjo contains 2301 documents from the archives of the Hōjuin cloister, where, as noted, the Gionsha managers (shigyō) were based from the 1380s onwards. These documents are organized by topic in nine sections (each further subdivided), and thoroughly indexed. Section 3.15 contains 93 documents related to the Gion festival, dated between 1409 and 1754; only eight, however, are pre-Ōnin. Section 3.16, titled “bajōyaku expenditure” (bajō ryōsoku 馬上料足), consists of 368 documents that record payments to various actors from the bajōyaku levy, almost entirely in the form of receipts. Of these, 352 are from the years 1397–1466; the final 16 refer to 1502. Other sections are organized by groups of actors. These include various categories of shrine monks (shasō 社僧, Sections 8.1–20) and lay “shrine people” (shajin 社人, Sections 9.1–28), the latter including miko 神子, otabisho priests (kannushi 神主), mikoshi bearers (kayochō 駕輿丁), and others.

These sources offer a mosaic of scraps of information about the people, resources, patronage networks, and conflict lines in and around the Gionsha. It is worth remembering, however, that in all of their diversity, these collections of documents all stem from the archives of the Hōjuin cloister. They have been preserved because the Gionsha managers based there regarded them as potentially useful. These documents were kept first and foremost as future evidence of land and income rights, privileges granted by power holders, customary practices, and settlements between the many groups of actors at the Gionsha, for use in lawsuits or as normative precedents for various eventualities. There are no ritual protocols here, and much less explorations of theological, ideological, or cosmological “meanings”. There are no esoteric readings or initiatory revelations about hidden truths, or even explanations of how any of the procedures of the Gion festival, or other Gionsha rituals, were thought to work as means to achieve particular ends. If we, for example, wonder what the halberds meant or how their ritual functions may have been understood, we search in vain for even the most indirect hints.

Of course, this does not mean that pre-modern actors had no ideas about the meanings or functionalities of the festival and its constituting parts. It is just that if they did, their thoughts are not expressed in extant sources. Or perhaps, in one sense they are. The Gionsha archive in itself serves as a genealogy of the shrine and its rituals, including the Gion festival. By copying and systematizing selected documents, the archivists collected tools for the production of meaning.

4. The 869 Goryōe as an Origination Legend

The 2019 commemoration attributed meaning by celebrating a defining moment of origin, construed with the help of these archival tools. The legend of the 869 goryōe, after all, is not just an origination legend but also a narrative that gives the festival meaning. On closer inspection, the creation of this legend and its selection as the festival’s “official” narrative was the result of many acts of signification, by a series of actors at different historical junctures, leading up to its post-war canonisation. The process that produced this legend bears closer consideration because it reveals much about the process of creating and highlighting documentary evidence as raw material for new narratives.

The first account of a goryōe occurs in Nihon sandai jitsuroku 日本三代実録, the official “Record of the Three Reigns [of the Emperors Seiwa, Yōzei, and Kōkō]”. This goryōe, held in the Shinsen’en garden on the twentieth day of the fifth month of 863, involved offerings to six wrathful spirits, sutra recitations, lectures, and dances.14 It featured no mikoshi and no halberds, and of course bore no relation to the Gionsha, which did not yet exist. Yet there is one detail that connects this event to the Gion festival. In 865, the court issued a ban on “private” goryōe, horse races, and horseback archery contests.15 Such spontaneous goryōe were protests as much as they were rituals. Involving armed displays of prowess, they were an obvious cause of court concern. The 863 goryōe may well have been designed to offer an alternative to such private events, which could easily trigger riots. Strikingly, the 865 ban was issued on the fourteenth day of the sixth month—the day on which the Gion goryōe would be held a century later. In the middle of the sixth lunar month, the monsoon ended and the heat of summer turned the soaked city into a hunting ground for pestilence deities. Perhaps 14.6 was a significant date in this connection long before the Gion goryōe originated.

In contrast to the event in 863, there is no mention of a goryōe in 869 in any contemporary or even medieval source. It appeared for the first time in a documentary history of Gion titled Gion hon’en zatsu jikki 祇園社本縁雑実記 (Miscellaneous true records of Gion’s origin), dated to the 1670s:

In Jōgan 11 (869), there was a great epidemic in the realm. To enhance the glory of the Throne, bring peace to the people, dispel the sickness and appease [the spirits], [Emperor Seiwa] gave orders by edict to Urabe Hiramaro 卜部日良麻呂. On the seventh day of the sixth month, [Hiramaro] set up sixty-six halberds, each two jō (c. six metres) high. On the fourteenth, he led men from the city and farmers from beyond the city boundaries to the Shinsen’en garden, carrying mikoshi to worship [the halberds] there. This was called the Gion goryōe. Ever since, this has been a customary practice performed on the seventh and fourteenth days of the sixth month of each year.16

The same tale is referred to in Gionsha hon’en zatsuroku 祇園社本縁雑録 (Miscellaneous record of the Gionsha’s origin), a similar work that is also dated to the 1670s. Here, the emphasis is on the origin of the shrine rather than the festival:

In Jōgan 11 (869) there was a great epidemic in the realm. To enhance the glory of the Throne, bring peace to the people, dispel the sickness and appease [the spirits], [Emperor Seiwa] gave orders by edict to worship the pestilence spirits in Yasaka Village at a site called the Goōji-sha 護[牛]王地社, the Ox King [or King-Protecting] Shrine. (It is also said that Urabe no Hiramaro performed these rites on imperial orders.)17

The point of this version is that the 869 rituals proved so effective that a palace building was moved to this site in 876, marking the founding of the Gionsha. The author of Gionsha hon’en zatsuroku (likely a Hōjuin shrine monk) notes with some scepticism that “this theory is a tale (kōjitsu 口実) that has been used from ancient times until today by the lay shrine lineages (shake 社家)”. His doubt is understandable. The name Goōji refers to Goō amulets (goō hōin 牛王宝印), most commonly used to add potency to oaths and pledges. Although Goō amulets usually derived from Kumano, association with the deity Gozu Tennō at the Gionsha allowed for the amulet business to thrive here as well. In the Edo period, the sale of Goō amulets was an important source of income for the Gionsha’s shake.

It would appear, then, that the 869 legend was popular in shake circles, rather than among the shrine monks who topped the Gionsha hierarchy. In contrast to shrine-temple complexes such as, for example, Enryakuji-Hie and Kōfukuji-Kasuga, the Gionsha had no lineage of shrine priests. It did, however, accommodate a broad array of lay retainers, ritualists, performers, and guards, who were collectively referred to as shajin 社人, “shrine people”. As noted, Yasaka jinja monjo (vol. 1) has its own section on various categories of shajin. Among them are the priests of the Ōmandokoro and Shōshōi otabisho (called the “left” sahō 左方and “right” uhō 右方 kannushi, from the viewpoint of the Gionsha); guild-like groups of female and male miko dancers and oracle-givers, some of whom were the wives and daughters of shrine monks; lesser miko who kept the grounds clean and performed kagura dances and other services for common believers (called katahaya-shū 片羽屋衆 or “the people who live in the ‘half-winged’ shacks”);18 different groups of mikoshi bearers; and outcasts called inu 犬 (“dog”) jinin, who mainly served as guards.

The katahaya-shū are of particular interest here. In medieval times, these miko were perhaps comparable to the miyako 宮籠 of the Hie Shrines: beggar-like figures who settled in the spaces below the shrine floors. There, they earned a living by addressing the needs of pilgrims while making themselves useful to the shrine priests by performing menial tasks and passing on a cut of their earnings (Kuroda 1999, p. 181). By the late sixteenth century, however, the katahaya-shū had greatly improved their lot. Now known as “kagura performers” (kagura-yaku 神楽役), they had expanded their range of activities considerably and at times refused to set aside time for the menial chores of their forebears.19 Among the services they offered were naming new-born babies, running lotteries (tomi 富), and even making rounds of patron believers in the provinces. Selling Goō amulets was also part of their repertoire.20 It is likely, therefore, that this group of kagura performers were the “shrine people” to whom Gionsha hon’en zatsuroku referred as the main propagators of the 869 legend.

While the reference to Goō amulets betrays the perspective of this group of actors, the legend also contains elements that point in different directions. One is the mention of Urabe no Hiramaro. Hiramaro (807–881), an expert on tortoise shell divination from Izu, was the legendary ancestor of the Yoshida 吉田 lineage of court priests. First mention of this figure in connection with the Gion festival can be traced to Nakatomi harae ge 中臣祓解 (“Exegesis of the Nakatomi Purification Formula,” 1523), a work by Kiyohara Nobukata 清原信賢 (1475–1550).21 In this text, Nobukata records teachings that he received from his father, Yoshida Kanetomo 吉田兼倶 (1435–1511). Kanetomo experienced both the discontinuation of the Gion festival in 1467 and its revival in 1500—a feat that was made possible by decisive action on the part of Hosokawa Masamoto 細川政元 (1466–1507), the shogunal deputy (kanrei 管領) and military strongman of the day. It would appear that in this context, Kanetomo stressed the role of his ancestor Hiramaro in this festival, which clearly was so important to the new hegemon. Nakatomi harae ge, however, dates Hiramaro’s goryōe to 876, the year of the founding of the Gionsha, and makes no mention of the sixty-six halberds. Those halberds entered the story in yet another context. They appear for the first time in a compilation of poetry titled Chōmei shiki monogatari 長明四季物語 (Chōmei’s tale of the four seasons), which is attributed to Kamo no Chōmei but likely dates from the late fifteenth century.22

Honda (2014, p. 25) has traced how explanations of the festival’s origin evolved in Edo-period guidebooks and gazetteers. Early guidebooks cited 970 or 974 as the festival’s year of origin, based on sources that dated the founding of the Ōmandokoro otabisho to that year (e.g., Fusō keikashi 扶桑京華志, 1665). References to 876, under the influence of Nakatomi harae ge, first appeared in the early 1700s (e.g., Sanshū meisekishi 山州名跡志, 1711) and became standard in the mid-1800s. The sixty-six halberds made it into such guidebooks only in the eighteenth century, notably in the particularly popular Gion’e saiki 祇園会細記 (1757).23 Honda fails to find any mention of 869 before the Meiji period (Keika yōshi 京華要誌, 1895). In general, he notes that the trend was towards an ever earlier dating of the festival’s beginning, with ancient dates first being mentioned as exciting though not entirely convincing possibilities, to be promoted to the main theory in later works. What strikes me as particularly significant, however, is the fact that the dating was changed from the date of the origin of the otabisho to that of the goryōe and the Gionsha.

The 869 theory gained its current status as the festival’s semi-official legend only in the 1950s. This was the result of yet another particular set of circumstances. Between 1947 and 1952, the Gion festival was gradually revived with vigorous backing of the Kyoto city authorities. During the war the festival stood in the token of the war effort: in 1942, large banners hung from the hoko floats and thousands of inscribed paper lanterns proclaimed that it was dedicated to “eternal victory for the Imperial Army” (Teeuwen 2020b, p. 227). A new narrative was needed if the festival was to find a natural place in the new context of post-war democracy. It was in this setting that Hayashiya Tatsusaburō 林屋辰三郎, a Marxist historian based at Ritsumeikan University, revalidated the Gion festival as an expression of the peaceful culture of Kyoto’s commoners (Hayashiya 1953). Hayashiya laid out his innovative reading of the festival in Gion matsuri, a publication of the Association of Scientists for Democracy, describing it as a monument to the emancipation of the masses from warrior oppression. The association of the festival with alleged traditions of Kyoto pacifism and inclusive equality was most timely, and it would evolve into the dominant narrative about the festival in the 1960s and beyond (Teeuwen 2020a).

Looking back on the festival’s long history, Hayashiya referred to a text titled Gionsha hon’en roku 祇園社本縁録 (Record of the Gionsha’s origin), which quotes the passage in Gion hon’en zatsu jikki translated above.24 In Hayashiya’s paraphrase, an epidemic struck the realm in 869, and sixty-six halberds were set up at the Shinsen’en garden on the seventh day of the sixth month. Convinced that this calamity was caused by the anger of Gozu Tennō, men from the entire city then carried mikoshi to this garden and staged a goryōe there (Hayashiya 1953, p. 64). It was not the early date that was important to Hayashiya, nor was he interested in the role of the imperial court or in Hiramaro’s alleged leadership. Instead, he stressed the agency of “men from the entire city”. In Hayashiya’s reading, Gionsha hon’en roku shows that the festival was initiated by the commoners of Kyoto. Their effort then allowed the festival to develop into a vivid expression of the autonomy and creativity of what Hayashiya called the “people of the streets” (machishū 町衆). His notion of the Gion festival as an expression of the “folk culture” of Kyoto’s commoners has since gained official status in the discourse of heritage preservation—e.g., in the description of the festival’s float parades as UNESCO Intangible Cultural Heritage (in 2009).

What, in the end, was the 2019 commemoration celebrating? The 869 narrative is a work of fiction, written by many hands. It was inspired by the Nihon sandai jitsuroku account of the 863 goryōe. The Shinsen’en, with its imperial cachet, was kept, but different narrators added their own ingredients. Yoshida Kanetomo inserted his ancestor Hiramaro, the miko at the Gionsha appended their Goō amulets, and the guidebook authors of Kyoto’s printing houses were attracted both to the court connection and to the lure of ancient origins. The sixty-six halberds stemmed from the imagination of a poet. Hayashiya added a new twist with his emphasis on the agency of the city “masses,” inspiring a new storyline that stages the festival as an expression of machishū autonomy and creativity. The whole story is an example of how sources create history, rather than the reverse.

5. Another Set of Actors, Another Legend: The Vanished Otabisho Priests

The 869 legend has a rival that is prominently displayed as part of the festival itself. Since the Meiji period, the mikoshi processions are headed by a group of Yasaka parishioners (the Miyamoto-gumi 宮本組) who carry “divine treasures”. Foremost among those treasures is a six-foot long lacquered board inscribed with the imperial decree that, allegedly, marks the beginning of the festival. In its earliest recorded version, the text of this “decree board” (chokuban 勅板) was as follows:

Office of the Kanjin’in 感神院政所

In the late fifth month of 974 (Ten’en 2), during the reign of Retired Emperor En’yū, an oracle proclaimed that there was to be a divine progress to the residence of our ancestor Sukemasa 助正, on the intersection of Takatsuji and Higashi-no-Tōin streets, which is to serve henceforth as the otabisho. It was found that a spider’s thread stretched from a mound in the back garden of that residence to the [Gionsha] shrine hall. The shrine officials, who found the thread strange, followed it to Sukemasa’s residence. [The Gionsha gods] appeared to Sukemasa in a dream, announcing that they wished to stay [in his residence] for seven days. The matter was brought before the Throne, and it was decided that henceforth Sukemasa’s residence was to serve as an otabisho, with Sukemasa himself as its priest. Imperial decree (senji 宣旨), 7.6.974.25

It was this tale that inspired the 974 dating of the festival’s origin in guidebooks such as the 1665 Fusō keikashi, mentioned above. This dating is more in tune with extant primary sources: the court record Nihongi ryaku 日本紀略 first mentions rinjisai offerings to the Gionsha on 15.6.975, while in court diaries, similar entries begin appearing from 982 onwards.26

In the Edo period, this chokuban board was kept at the Ōmandokoro otabisho, where it was known simply as “the tablet” (ofuda 御札). According to a 1617 document, carrying this tablet at the front of the mikoshi processions was among the tasks of shrine servants (miyaji 宮仕) based at the Ōmandokoro.27 The tablet text connects the origin of the festival not to an imperial goryōe or to the Gionsha, but to the founding of the otabisho. This is stated explicitly in an older version of this same myth, included in Gionsha ki with the dating 27.11.1385 (Shitoku 2). Here, the text continues: “This is the origin of the festival. Since [974], there have been thirteen generation of [Ōmandokoro] priests; the position has never passed to another lineage [than Sukemasa’s]”. After this, the names of the thirteen incumbents up to that time are listed.28

This legend ended up in Gionsha ki in the context of a conflict over the control of the Ōmandokoro otabisho. As we have already seen, Ashikaga Yoshimitsu had installed his oshi, Kenshin, as Gionsha manager in the same year 1385, as part of a larger strategy to undermine Enryakuji’s power in the city, including its control over the Gionsha. After all, the Gionsha served as an important base for Enryakuji activities in the city, from mikoshi attacks to guild-based moneylending and trade. In subsequent years Yoshimitsu strengthened Kenshin’s hand by granting various possessions and privileges at the Gionsha to Kenshin’s cloister, the Hōjuin. As part of this scheme, in 1397 Ashikaga Yoshimitsu transferred the priestship of the Ōmandokoro otabisho from Sukemasa’s descendent Sukesada 助貞 to the Hōjuin. Sukesada appealed to the Tendai abbot, who in 1411 granted the otabisho to Sukesada’s successor Kōyashamaru 幸夜叉丸; but in 1412, when a new abbot took over, control over the otabisho was returned to the Hōjuin. In 1431, another heir of Sukesada called Shōjumaru 松寿丸 appealed to shogun and won the otabisho back—only to lose it again in 1443 (Shimosaka 2016a, pp. 848–49). This proved to be the end of Sukemasa’s lineage of Ōmandokoro priests. The juridical details in these conflicts involved both claims of ancient precedent and accusations of murder, but the outcome was likely based on the ever-shifting balance of power between (factions at) Enryakuji and the Muromachi shoguns. The final winner was the Hōjuin, thanks to its close links with the shogunate as the seat of the shogunal oshi.

In 1591, Toyotomi Hideyoshi moved the Ōmandokoro to its present location on Shijō street. At the time, this place was a cul-de-sac, an anonymous bamboo grove on the banks of Hideyoshi’s famous odoi お土居 embankment (Tsuchimoto 1994). The second otabisho, Shōshōi, was simply confiscated and never revived. In the Edo period, the Shijō otabisho was no longer referred to as a separate shrine, and its caretakers were not priests but mere “keepers” (miyamori 宮守 or tanamori 棚守), appointed by the Gionsha’s shigyō manager. If there was an Ōmandokoro archive in the fifteenth century, it did not survive the demise of Sukemasa’s priestly lineage and the 1591 relocation, and all that is left are a few scattered references in Hōjuin documents.

The protracted battles over the otabisho that took place between 1397 and 1443 call our attention to the functions of these city shrines in the festival. The Ōmandokoro, in particular, was located in the heart of the Lower City, where it occupied an entire block. It would appear likely that the otabisho priests played a much larger role in the festival proceedings than extant sources reveal. Court diaries and Gionsha documents—our two main archives—focus on the return of the mikoshi to the Gionsha on the fourteenth day of the sixth month and the court offerings and horseraces there on the fifteenth. They contain no information whatsoever about what went on at the otabisho during the week that the gods resided in the city. If bringing the gods into the city streets was the point of the enterprise, we must expect that the procession on the seventh day and the week spent at the Ōmandokoro and Shōshōi shrines constituted a ritual world of their own. However, the triple whammy of the demise of the otabisho priests, the discontinuation of the festival due to the Ōnin war, and the 1591 demise of both otabisho as independent shrines erased most traces of that world. What happened in the medieval Lower City between the seventh and the fourteenth days of the sixth month is simply unknown, to the extent that even the question has hardly been asked.

One scholar who has drawn attention to the role of the otabisho is Seta Katsuya (Seta 2009). Seta invites us to consider what the festival may have meant to the commoners who invited the Gionsha deities into their community and invested both effort and economic resources in the festival in the shadow of the court and the shrine. The otabisho, he proposes, may have served as sites where this part of the festival was coordinated. Seta discovers some hints to that effect in the documents pertaining to the conflicts of 1397–1443. Particularly telling is a 1431 list of the entitlements that accrued to the otabisho priests—the concrete stakes of the conflict. These included rent from the inhabitants of houses built on otabisho land, levies on three guilds engaged in various branches of the textile business, another levy on sellers of second-hand goods, and 150 kanmon out of the bajōyaku levy.29 This document implies that before 1397, these entitlements had been part and parcel of the otabisho priestship—and now, thirty-four years later, they were (briefly) restored.30 150 kanmon was half of the portion of the bajōyaku that was allocated to the Gion festival; another 150 kanmon went to the Gionsha. Perhaps this suggests that the otabisho priests, or the community of guild merchants around the otabisho, once had a central role in allocating the bajōyaku as well? At the very least, the fact that the Ōmandokoro served as the overlord and protector of three guilds and another more informal group of merchants serves as an indication of its position in Lower Kyoto’s commoner society.

While the 1397–1443 conflicts have left us with at least some hints about the Ōmandokoro, we know even less about the Shōshōi otabisho. A shrine hall was built here in 1234; according to the court chronicle Hyakurenshō 百錬抄, “locals” (zaichinin 在地人) were the driving force behind its construction and embellishment.31 This site also had its own kannushi, who likewise ran into trouble in the 1440s. In 1441, the position of Shōshōi priest was bought by the large and well-known financier Zenjūbō 禅住坊, a merchant house with close links to Enryakuji.32 What the tasks of these kannushi consisted of, however, is unknown.

The Shōshōi also had a guild of miko, whose leader was officially appointed (or recognised) by the Gionsha since 1335 at the latest. At some point prior to 1500, this position had been acquired by another group of miko, based at the Goryōsha 御霊社 (today’s Kami Goryō Jinja) some 2.5 Km further north. This shows that the business of the miko at this site was part of a city-wide market, as was that of the Shōshōi priest. Rights to siphon off some of the earnings of these miko were bought and sold between competing investors, until the festival lost its foothold in the Upper City with the 1591 demise of the Shōshōi.

The Goryō women had also acquired the so-called komagashira 駒頭, a sculpted horse-head worn on the chest of a youngster. The horse-head youngster (komagata chigo 駒形稚児) joined the mikoshi processions on horseback. The seller of this horse-head was a “komagata guild,” which likely derived some form of trading privileges from its participation in the procession before financial difficulties forced it to sell the horse-head to the Goryō miko. In 1500, when the Gion festival was revived, these miko drove a hard bargain, making good on their earlier investment in this crucial object of ritual power.33 Negotiations with the Goryō miko to have the horsehead returned needed to be brought to a conclusion before the mikoshi processions could be restored.

Taken together, these tantalizing traces of otabisho activity suggest that they were important sites of festival activity. Clearly, the priests and miko of the otabisho shrines were central actors, especially from the perspective of city commoners who left us no records. Perhaps the otabisho were once communal centres where groups of priests and miko orchestrated a range of Gion practices that are no longer part of today’s festival. The 974 origination legend, displayed prominently on the “decree board” that is carried in the mikoshi processions to this day, is a relic from a different age. Its meaning is lost, since the actors who devised it have disappeared. This throws up the question what else might be lost to us, without even leaving a vestige to make us wonder.

6. Conclusions: Between Actors, Origination Legends, and Meanings

Rambelli raised the question of the relation between signifier and signified, forms and meanings. He argues that the Ise Shrines are only one example of a site where the correct execution of ritual forms has taken precedence over the formulation of clearly defined meanings. This leads him to wonder whether ritual practice without explicit exegesis was part of a “culture” that emphasizes rules and precedents over content and meaning. He also points out that apparently empty signifiers are uniquely versatile: if there is no predefined content, this allows for a proliferation of meaning that can enhance sacredness (Rambelli 2014, pp. 236–37).

The Gion festival is perhaps even more devoid of “content” than the rites of Ise; at least, there is a complete absence of pre-modern exegesis. The closest one comes intimations of the festival’s meaning are the origination legends that have served as precedents. Those legends, strikingly, never refer to the myths of the Gionsha deities. The famous tale of Gozu Tennō, his punishment of rich but stingy Kotan Shōrai, and his protection of poor but generous Somin Shōrai appears to be only tangentially relevant to the festival.34 Rather, the origination legends of Kyoto’s Gion festival refer to ancient decrees and oracles related to the beginnings of the goryōe.

These legends deal not with the meanings of festival procedures, but rather with particular actors who received sacred privileges in a distant past. They lead us away from questions of “why” and “how” to “who” and “where”. They tell us about the fortunes of those whose archives have survived, and also about those who mined these archives for their own purposes. The Hōjuin monks emerge as the main gatekeepers, while we are allowed only glimpses of other groups—the lay “shrine people” (shajin) at the Gionsha, the priests and miko at the two otabisho, and, further afield, the merchant guilds who were arguably the most active among the festival’s actors. We know about some of the forms that imperial and shogunal patronage took, but there is not a single source that expands on the meanings that any of these actors ascribed to the festival as a whole, or to its parts.

Rituals are of course always polysemic, carrying many meanings at the same time. Festivals, as public events with many actors, participants, contributors, and observers, are never simple systems of symbols with set meanings. They serve as stages for social action where selected actors perform prescribed acts that mean different things to different people, but are sensed by all to carry a particular importance. Michaels (2006, p. 261) argues that “the significance of rituals lies in the fact that they often create an auratic sphere or arena of timelessness and immortality—at least in religious or semi-religious contexts”; once that sphere is in place, “rituals can indeed do without any specific meaning”. Origination legends, claiming continuity since ancient times, likewise convey a sense of timelessness without imposing a concrete symbolic or functional meaning on the festival proceedings. The act of creating origination narratives, or of pretending that they are true, has the effect of enhancing the “auratic sphere” that gives the festival its value.

In the post-war period, ethnology (minzokugaku) theories have democratized the notion of precedent by adducing timeless “folk culture” as the festival’s source of meaning, rather than classical imperial decrees. While such theories configure ancientness differently, they perform much the same function as the origination legends discussed here. Yet there is no evidence in any source that, for example, the floats were at any time understood as yorishiro 依代—material mediums—designed to attract and remove pestilence spirits from the streets of the city. In fact, that term did not exist until Orikuchi Shinobu 折口信夫 created it in a 1915 essay.35 The 2019 celebration of the hoko as emblems of the festival’s original meaning elaborated on the modern “folk yorishiro” legend by pretending that the floats of today are the descendants of the spirit-catching halberds of the 869 goryōe. Whatever historians may say, media reports suggest that this event succeeded in adding colour to the festival’s present-day aura.