Japanese Buddhist images have been central to the religion’s dissemination since its establishment on the archipelago from at least the sixth century. Buddhism’s visual and material culture had the power to structure belief and practice, to serve as a mediator between the physical and intangible world, and to operate as a powerful source of legitimization for the devout. Even though Buddhist objects conceptually crossed the boundaries between the physical and the transcendent, they were nevertheless made by human hands and subject to the vagaries of time. Just as the Buddha taught, nothing lasts forever. Could an icon still function in the way that it was intended if it was damaged and did not appear in its original form? Were there notions of image maintenance in premodern Japan, and if so, what did they consist of? Given the sheer number of Buddhist objects in existence at any given time, how did a patron even decide what to preserve in the first place?

This article interrogates the practice of conserving Buddhist painting in premodern Japan and proposes the existence of three predominant ways to preserve images: to copy, to repair, and to re-create. Copying as a form of maintenance implies reproducing a composition due to the deterioration or destruction of the original medium. This practice pointedly differs from other forms of copying prevalent throughout Japanese artistic production that were carried out for different reasons.¹ As well, repairing an image in the context of premodern conservation refers to physically modifying the original. Re-creation combines these two modalities and entails copying an image while treating the newly made one as if it were the original and embodied the same historical, social, and religious associations.

Very few textual records exist that describe the exact treatments that premodern conservators executed in routine object care. Instead, inscriptions were customarily written on the back of a painting, its roller bars, or its storage box to commemorate the completion of a project and generally contained the names of the sponsors or a dedicatee and the date.² An investigation into the processes of preservation thus requires a look at these inscriptions to understand who was involved, as well as beyond them to other textual records and related extant objects in order to glean how an object was conserved to ensure its survival through generations.

I have derived the forms of image preservation presented above from three historical case studies that form the core of this article. The first investigates the preservation histories of the Mandalas of the Two Worlds (ryōkai mandara 両界曼荼羅; see fig. 1 for a seventeenth-century copy of these images; the originals are no longer extant) and the portraits of Shingon patriarchs that the eminent monk Kūkai 空海 (774–835) brought back from China to establish and propagate esoteric teachings in Japan (figs. 2–6). These sets of paintings were employed frequently together, and both met with significant damage around the same time in history. Yet, while the Mandalas of the Two Worlds were copied as new paintings, the original portraits of Shingon patriarchs were repaired. Analyzing the motivations behind the goals for these projects sheds light on how physically modifying paintings in premodern Japan for the sake of their preservation had the capacity to advance religious and institutional agendas in ways not possible through copying alone. The second case study examines the preservation history of the Takao mandara 高雄曼荼羅, a ninth-century version of the Mandalas of the Two Worlds executed in gold and silver ink on a dark background of purple-red twill weave (fig. 7). This object was copied anew and its original was also repaired, and provides additional insight into how its material qualities informed the divine presence contained within. The final object under consideration is the Taima mandara 当麻曼荼羅, an eighth-century textile depicting Amida’s Pure Land (fig. 8). This object was also poorly cared for in its early history, which consequently led to significant damage. Unlike the above objects, however, repairing the Taima mandara proved impossible due to limitations of technology, so the patrons decided to re-create it. This section of the article uncovers the nuances of re-creation as preservation and demonstrates how a painted version of the Taima mandara came to be understood as if it were one and the same as the eighth-century woven tapestry.

Before moving forward, it may be instructive to contextualize these analyses within two larger discursive fields: Buddhist materiality and studies of premodern Japanese Buddhist painting conservation. The former is an area that seeks to understand what the religion’s visual and material culture reveals about those who used and interacted with it, rather than how it reflects doctrinal content. For instance, Halle O’Neal has employed the framework of reuse, recycle, and repurpose to document transformative moments in an object’s biography. Studying afterlives in this way, she argues, provides a lens to understand how and why people over time inscribed new meanings and significance onto Buddhist artifacts.³ Other work in this sphere includes research by Caroline Hirasawa and Benedetta Lomi, who understand the study of materiality as an excavation of a network of relationships that place the object as the central node linking people, both past and present, and environment.⁴ Their work complements other divergent approaches including that of Fabio Rambelli, who exposes the porous boundaries between deity, living practitioner, and inanimate object, especially within esoteric Buddhist thought, to demonstrate the various ways in which all ultimately partake in the same universal principle.⁵ Further inquiries by O’Neal as well as Karen Gerhart seek to show how interaction with tangible objects enabled people to grieve the death of a loved one or gain control over the death process.⁶ All of these studies follow the pioneering work of Robert H. Sharf and Elizabeth Horton Sharf, who paved the way to understanding how objects functioned within belief and practice apart from their iconography or artistic style.⁷ In drawing on these studies, this article offers additional possibilities that consider how the physical condition of images allowed patrons to participate in Buddhist practice in highly personalized and personally meaningful ways not possible through ritual practice alone.

Studies in the conservation of premodern Japanese Buddhist paintings grew out of the custom of documenting processes for the preservation of a particular image. A recent report commemorating the restoration of the twelfth-century painting of the bodhisattva Fugen, now designated a National Treasure, brings together the expertise of scholars, curators, and conservators. After detailed analyses of the religious, historical, and cultural framework of the image and the painting techniques that artists used to communicate these ideas, the study delves into the choices modern-day conservators made for their treatments. Crucially, it emphasizes that they did not add anything extraneous to the visible surface and that they even went so far as to remove noticeable traces of earlier restoration projects. The result is an image that does not look brand new, but one that is instead free from debris and creases to allow, to the extent possible, the original artists’ hands to shine through.⁸ While premodern painting conservators undoubtedly shared the same concern to manage damage, they generally chose from a greater range of treatment options, as proposed above.

Masuki Ryūsuke has argued for two approaches to Buddhist painting conservation before the modern era. Patrons could discard an old painting and request the production of an entirely new copy, which Masuki notes was entirely common through at least the twelfth century, or they could repair the original.⁹ At first glance, his argument on copying seems to run counter to injunctions in sutra literature that equate the repairing of “reliquaries, images, and scriptures that are ramshackle or ruined” to an act of great merit.¹⁰ As Benedetta Lomi further elaborates, however, the choice of when and how to repair a damaged or destroyed object was not always straightforward in premodern Japan. Although she notes a great deal of concern for the upkeep of stupas and sculptures, Buddhist painting presented a completely new set of challenges due to the comparatively fragile nature of the objects.¹¹ If keeping an object looking new or presentable was crucial for religious practice, copying a painting as a form of maintenance as Masuki describes most likely occurred when it was deemed that no amount of repair or modification could render the image functional again.

The case studies explored below build on these fields and demonstrate that while all three proposed modalities of preservation served to keep the image in a good state for viewing, studying, or ritual use, repairing and re-creating an image could additionally serve as a means for the production and diffusion of Buddhist cultural knowledge—although in different ways. Finally, special consideration is given to how the materiality of the object—its medium, pigments, or other materials that it comprises, as well as any manifest traces of engagement, either by makers or later viewers—structured the understanding of its meaning and function.

An illustration from the Miraculous Legends of the Kasuga Shrine Deity (Kasuga Gongen genki-e 春日権現験記絵, fourteenth century; fig. 9), depicts a typical Japanese esoteric Buddhist worship space. A court official dressed in black bows before the Diamond World Mandala, placed just out of view, while an identical altar to his right contains an image of the Womb World Mandala, identifiable by the central eight-petaled court. Together, these paintings comprise the Mandalas of the Two Worlds and represent the totality of the esoteric Buddhist teachings, in which the Diamond World Mandala denotes the cosmic nature of the Buddha and the Womb World Mandala the manifestation of the Buddha in the natural world. In the right corner of the composition, a half-hidden folding screen reveals three of the twelve celestials (jūni ten 十二天) who guard the ritual space, and in the left corner are portraits of three Shingon patriarchs, with more located outside the picture plane.

The Mandalas of the Two Worlds and the five portraits of Indian and Chinese Shingon patriarchs were among the more than two hundred objects that Kūkai brought back to Japan from China to help establish and propagate the esoteric teachings.¹² Both sets of images were given to Kūkai by his teacher Huiguo 恵果 (746–806) and produced in the workshop of the noted Chinese Buddhist painter Lizhen 李真 (dates unknown).¹³ Even though Kūkai received both sets of images from Huiguo, their inclusion in one space as illustrated in the Miraculous Legends, for instance, was Kūkai’s own design.¹⁴ His understanding of the Shingon ritual arena thus consisted of two core elements: the totality of the esoteric Buddhist teachings represented by the mandalas and its authenticity as demonstrated by the lineage of patriarchs represented in the portraits.¹⁵ For Kūkai, both sets of images were of equal importance for the successful spread of esoteric practice.

Upon Kūkai’s initial return to Japan in 806, he presented the objects he brought back with him, including the two mandalas and the portraits, to the court of Emperor Heizei 平城天皇 (774–824). Three years later, Heizei abdicated the throne due to illness and returned Kūkai’s objects.¹⁶ The monk then presumably deposited them at Jingoji 神護寺, a temple atop Mount Takao 高尾山, located to the northwest of the Heian capital, where he took up residence that same year. Although at this early date Kūkai had yet to codify what became a typical esoteric ritual space as exemplified in the Miraculous Legends, he must have used the mandalas and the portraits together and with a similar degree of frequency because both sets of images started to deteriorate around the same time. In the case of the Mandalas of the Two Worlds, Kūkai notes in his “Vow of the Four Debts of Gratitude in the Making of the Mandalas of the Two Worlds” (Shion no ontame ni nibu no daimandara o tsukuru gammon 奉為四恩造二部大曼荼羅願文):

The silk is damaged, and the colors are flaking [so much] that the faces of the deities have started to change. Students will express lament viewing [the paintings in this way] and it is a pity that [the damaged mandalas] will bring misfortune to all living beings. . . . Reverently, beginning in Kōnin 12 (821), fourth month, third day and until the end of the eighth month, I had the Mandala of the Womb of Great Compassion made up of eight widths of silk, the Diamond World Mandala made up of nine widths of silk . . . and portraits of Nāgārjuna and Nāgabodhi all newly painted.¹⁷

Kūkai bemoans the fact that the mandalas’ deterioration had rendered the iconography illegible, which he fears would hinder proper teaching and even cause misfortune for others. Yet the monk does not stop at requesting copies of the paintings; he commissions entirely new objects, including two additional patriarchal portraits of Nāgārjuna and Nāgabodhi, the significance of which I will return to below.¹⁸ Although it is impossible to speculate what the actual condition of the objects might have been at this time, copies of Kūkai’s mandalas were produced throughout the premodern period, with the oldest extant example dating to 1191 (fig. 10).¹⁹ Recent restoration work has sought to stabilize the two paintings, which have been heavily damaged and fragmented, especially in the lower third of the image.²⁰ Even if the deterioration of Kūkai’s images at that time was only a fraction of that seen today on the twelfth-century version, it is not difficult to imagine that any distortion of the detailed iconography, however major or minor, could render the images ineffective for teaching or even ritual viewing in the eyes of a master attempting to establish veneration of the objects in Japan for the first time.²¹

Masuki Ryūsuke has argued that the practice of commissioning a copy of a painting instead of restoring the original was an altogether common practice before the thirteenth century.²² Nonetheless, the deterioration of the Mandalas of the Two Worlds did not simply spur Kūkai to request unblemished copies of the iconography; it also allowed him to refine and implement his vision for systematizing the totality and demonstrating the authenticity of the esoteric teachings, as indicated by his commission of the two new portrait paintings.²³ Abé Ryūichi has postulated that one of the reasons Kūkai was so vested in this project was that he was planning to implement a grand initiation rite in 822, the first of its kind in the archipelago, for Heizei and Prince Takaoka 高丘親王 (799–865) that was to be sponsored by the reigning emperor Saga 嵯峨天皇 (786–842).²⁴ A rite of the utmost importance required the most luxurious objects, and Kūkai must have been keen to present the mandalas, and by extension the Dharma, in spectacular fashion. In preparation for the very same event, however, Kūkai made a completely different decision when faced with the deterioration of the set of five portraits of Shingon patriarchs. He describes the damaged portraits in a letter addressed to the Minister of the Left Fujiwara no Fuyutsugu 藤原冬嗣 (775–826):

The fog atop the mountain [i.e. Mount Takao] is thick; the figures [of the patriarchs] are damaged and everything is covered with dust. I am afraid they are not adequate for imperial viewing. My request today is to ask for a specialist to immediately repair [the paintings]. I would be most grateful for this.²⁵

Kūkai requests that the damaged images be restored, not copied. Moreover, he took advantage of the occasion and added the two newly painted portraits of Nāgārjuna and Nāgabodhi (figs. 11, 12) to the original set of five. Why did Kūkai choose to copy one set of images while repairing the other, and what might the distinction have been between these two modes of preservation?

While Kūkai’s concern with the mandalas was the legibility of the represented deities, his interest in the portraits lay in the information they conveyed about the sitters rather than the images of the sitters themselves. Kūkai’s goal for restoring the five patriarchal portraits and adding two new ones was twofold: to prepare the paintings for an imperial viewership, as he states in his letter to Fuyutsugu, and also to involve Emperor Saga in the completion of the project. Each of the now-seven portraits is nearly identical in size and scope. That of Śubhakarasiṃha is a representative example (fig. 3). The image illustrates the patriarch in three-quarters view atop a dais. His name in both Sanskrit and Chinese characters appears to his right and left, respectively. Below, a brief biography summarizes the figure’s major accomplishments. The portraits of Nāgārjuna and Nāgabodhi that Kūkai commissioned to complete the series follow the same conventions. They depict the sitter facing left or right on a platform and contain their names in Sanskrit and Chinese characters as well as their biographies below. The five portraits that Kūkai brought back from China, however, initially lacked text. During their repair, craftsmen appended new strips of silk to give space for Kūkai to compose biographical entries for each sitter. The two new portraits were created with space to inscribe the relevant information so that all in the set would be identical. Kūkai also added the names of the sitters on the five original portraits. He then asked Emperor Saga to inscribe Nāgārjuna and Nāgabodhi’s names on the newly completed paintings.²⁶ The act of repairing these paintings effectively transformed them into images that additionally transmitted information about the patriarchs, including their accomplishments and their ethnicity.²⁷ Saga’s involvement in the completion of the set extended the line of transmission from the Shingon masters through Kūkai to the emperor himself as the head of the realm.

Saga’s act of inscribing the names of Nāgārjuna and Nāgabodhi takes on additional significance when considering the calligraphic style he employed for this task. For example, in the upper right corner of the portrait of Śubhakarasiṃha that Kūkai completed, the monk wrote “Indian name 梵號” in Chinese characters and below that the sitter’s name “Śubhakarasiṃha” in Siddhaṃ script. On the left, the word “Chinese name 漢名” appears at the top, and then the figure’s name “Shanwuwei” in Chinese characters below.²⁸ All of these characters are composed in flying white (hihaku 飛白), a technique that renders the vertical lines thick and horizontal lines thin, sometimes with ribbonlike strokes as most clearly visible in characters in the upper right and left corners. Emperor Saga followed Kūkai’s precedent and completed his paintings by writing characters in flying white as well.

Both Kūkai and Emperor Saga were noted calligraphers in their own right; Saga studied with Kūkai, who was widely versed in both writing and calligraphic theory. In particular, Kūkai understood that characters written in the style of mixed script (zattaisho 雑体書), which encompassed flying white, had the power to communicate ideals of the natural world, peaceful rulership, political power, and the protection of the state. Flying white in particular was also a technique favored by Chinese emperors to mark the auspicious. Kūkai had given Saga manuals of mixed script and flying white to study, and the emperor must have been well attuned to the potential of these techniques to project imperial authority.²⁹ Thus, when Saga completed the repair of the patriarchal paintings by inscribing the two masters’ names, he was simultaneously marking the auspicious nature of the initiation rite to be held for Heizei—when the renewed paintings would be displayed—while at the same time proclaiming a foundation for a harmonious and prosperous realm. Saga’s involvement also served Kūkai’s goals of establishing and disseminating Shingon Buddhism. His participation was concrete proof that he now belonged to the lineage of figures memorialized in the portraits and had inherited their teachings.

The choice to copy the Mandalas of the Two Worlds and repair the portraits of the patriarchs reveals distinctions about these two modes of preservation and the role of each set of images in the ritual arena, especially given the fact that both most likely started to deteriorate around the same time, as previously noted. Creating a clean copy of the mandalas ensured the images would be suitable for viewership both by the emperor as well as by students who studied the teachings, while preserving the portraits facilitated the creation and dissemination of religious and cultural knowledge, which here consisted of codifying a lineage of Shingon masters in Japan. Additionally, both modes of preservation were intertwined with the patron’s authority. Kūkai capitalized on the occasion of copying the mandalas by requesting wholly new patriarchal portraits to be produced, in order to more fully express his vision of the Shingon teachings and their diffusion from India, China, and Japan. Emperor Saga’s participation in repairing the older portraits legitimized Kūkai’s interpretation and expression of esoteric teachings and history.

The Takao mandara refers to a version of the Mandalas of the Two Worlds rendered on purplish-red twill with gold and silver paint (fig. 7). The paintings were created sometime between 829 and 833 under the patronage of Emperor Junna 淳和天皇 (786–840) and were installed at Jingoji, albeit after Kūkai had left the temple and taken up residence at Tōji 東寺 in the capital. Although Kūkai himself did not play a role in the production of the Takao mandara, the fact that the paintings were produced at Jingoji rendered them conceptually associated with the monk soon after their production.

Like the polychrome Two Worlds Mandalas, the Takao mandara was copied due to damage. This act allowed the patron to display the images in a superior condition as part of his plan to revive Jingoji and elevate the temple. Unlike the polychrome versions, however, the gold and purple examples were also repaired on a separate occasion, not for religious use but in an attempt to preserve their perceived link to Kūkai. In this case, the Takao mandara’s gold and silver ink on a dark background stand out as a key element that structured the understanding of the images’ cultural significance, solidifying their inextricable relationship to the esoteric monk. A comparison between the driving factors behind the decisions to copy the Takao mandara versus to repair it reveals the role these special pigments played in informing the historical legacy the patron wished to retain through its restoration.

The story of these mandalas’ preservation begins roughly three hundred years after their production, when the Jingoji monk Mongaku 文覚 (1139–1203) sought to revitalize the temple after it fell into a period of neglect.³⁰ Part of Mongaku’s push to revive Jingoji included reinstalling the Takao mandara in the Initiation Hall (built 823–35), as they had been evacuated from the temple for safekeeping sometime prior.³¹ Mongaku petitioned Retired Emperor Go-Shirakawa 後白河法皇 (1127–1192) to bring the objects back, arguing not only that they properly belonged at their place of origin but also that it was particularly appropriate given that he believed they were painted by Kūkai himself. Mongaku writes:

The Mandalas of the Two Worlds in gold paint were painted by [Kōbō] Daishi himself . . . long ago when Daishi was still alive, these objects were the original mandalas enshrined at this temple. . . . [I], Mongaku, have revived this temple. On the day when the rituals of old are carried out again, the mandalas should be installed at this temple as they were. I urgently make this appeal [to bring the paintings back].³²

Mongaku’s assessment that the Takao mandara was painted by Kūkai is significant, for the same designation was not extended to the polychrome mandalas, which have the same composition and were also inextricably connected to the monk. For instance, the Illustrated Activities of Kōbō Daishi (Kōbō daishi gyōjō ekotoba 弘法大師行状絵詞, fourteenth century) depicts Kūkai, identifiable by a three-pronged vajra on the table in front of him, overseeing Lizhen and his workshop painting the Mandalas of the Two Worlds for the monk to bring back to Japan (fig. 13). While the image depicts Kūkai’s involvement in the production of these paintings, he is not the figure who is making them. Mongaku thus distinguished the Takao mandara from other versions by positing them as the manifest traces of Kūkai himself.

The petition to return the gold-inked paintings to Jingoji was successful, yet in lieu of displaying them in their original home, Mongaku opted to place them in a box and shut them away in the storehouse. He then proceeded to commission a copy of the paintings instead.³³ Mongaku’s decision to copy the Takao mandara while placing the originals out of view may seem like a striking decision considering his aspirations to return Jingoji and all its constituent elements, images included, to their ninth-century state. Mongaku’s treatment of the objects, however, is consistent with attitudes toward damaged iconography raised in the previous discussion, namely, that instead of repairing an icon for posterity, it was simply rendered afresh. Doing so reflected the patron and his agenda in superb light and demonstrated Mongaku’s success at returning Jingoji to its perceived original state.

Over one hundred years later, in 1308, Retired Emperor Go-Uda 後宇多法皇 (1267–1324) visited Jingoji and viewed the original ninth-century mandalas. He was intensely aware of their reputed connection with Kūkai, but whereas Mongaku chose to copy the paintings and take the old ones out of circulation, Go-Uda commissioned their repair. He documented the process, along with his sentiments, in a record titled Account of Repairing the Takao Mandara (Takao mandara go shūfukuki 高雄曼陀羅御修復記, 1309), where he first reiterates his understanding of the paintings’ importance:

The Mandalas of the Two Worlds housed in the Initiation Hall at Jingoji consist of the true brushstrokes of the founder [Kōbō] Daishi. It is the chief object of worship for all of [Shingon Buddhism].³⁴

It was also perhaps because of the mandalas’ perceived provenance that Go-Uda expressed dismay to see them in a damaged state. He goes on to write:

When the venerable monk Mongaku revived the temple long ago, a new copy of the Mandalas of the Two Worlds [was made] and venerated. The originals were put in storage and have not yet been repaired. Moreover, when I visited this temple on the twenty-fourth day of the third month in the first year of Engyō [1308], I viewed these sacred objects. There were many visible signs of damage, and fragments [of paint] had fallen off and are gathered at the bottom of the storage box. The exalted deities [appear] lacking and incomplete, and it is difficult to identify the names of the assemblies. The icon has been destroyed and is imperfect, and the true forms of the deities are deficient. With all due reverence, even though these are the brush traces of the esteemed [Kōbō Daishi], the object has been hidden away and one can only feel pity about its damage.³⁵

Emperor Go-Uda, like many sovereigns before him, recognized the unique power of Shingon Buddhist ritual to protect and fortify both imperial rulership and the realm. Shingon Buddhist practice had splintered in Japan after Kūkai’s lifetime, and Go-Uda differed from his predecessors in that he sought initiation in multiple lineages, believing ritual mastery could be used as tool for political legitimization.³⁶ Although each system of practice had its own distinct protocols, all were united through Kūkai and the Mandalas of the Two Worlds. When Go-Uda viewed the Takao mandara and lamented its physical state, his concern was not simply a pair of paintings that had become damaged over time; they were the images at the very center of Kūkai’s teachings, and thus the object from which the emperor drew his authority to rule. Furthermore, Go-Uda’s description of the inability to recognize the iconography due to deterioration echoes the way Kūkai recounts the condition of the mandalas that he requested to be copied. While Kūkai understood the afflicted state of the mandalas to hinder ritual practice and be an obstacle in the propagation of the teachings, Go-Uda felt the loss of pigmentation to be regrettable because of their inability to facilitate a direct connection to Kūkai by proxy through his supposed brushstrokes.

After the emperor articulated his sorrow at having seen the paintings in a less-than-spectacular state, he then decided to sponsor their repair. In his Account, he writes that he consulted with the monks of the temple to find a priest from another institution who excelled in painting restoration. The repairs were then completed after a period of five and a half months. Upon the completion of the project, Go-Uda returned to Jingoji to view the paintings for a second time. The emperor concludes the Account overjoyed by asserting:

The forms of the exalted deities look [now] as they did in the past and they glitter again as they did in the old days. [It was as though] Kūkai has reappeared in our time through his brushstrokes; it bestows his blessing and the inconceivable protection of [the deity] Hachiman! . . . What is even more joyful is that this is the object that allows karmic connections to form to the teachings [i.e., it allows the teachings to prosper].³⁷

In the very same year that Emperor Go-Uda requested the repair of the Takao mandara, he also patronized the revival of Tōji, the temple where the copies of Kūkai’s polychrome mandalas were held and where Kūkai spent the bulk of his career after he left Jingoji. The emperor’s main goal for the restoration of Tōji, however, was to enhance the education of monks by building residences and expanding facilities for lectures and other religious events.³⁸ While the restoration of both the Takao mandara and Tōji’s temple compound was meant to facilitate a greater access to the esoteric teachings, Go-Uda understood that only the paintings, after being properly repaired, were able to channel Kūkai’s numinous powers.

The motivations behind commissioning the repair of the Takao mandara mirror those for the restoration of the Shingon patriarchal portraits. The patrons of both projects aimed to foster connections to the past for self-legitimization, and to gain cultural authority as well as religious benefits. Whereas Kūkai enabled Emperor Saga to become a participant in the lineage of Dharma transmission by inscribing the sitters’ names on two of the paintings, Emperor Go-Uda brought Kūkai back to life through his purported brushstrokes, simultaneously enabling anyone who viewed the paintings to form a connection to the monk and to the teachings. Doing so also served to elevate his own rulership, as he placed himself as close to the apex of Shingon Buddhism as possible. The question remains: why did the Takao mandara capture Mongaku’s and Go-Uda’s imaginations in a way that the more colorful mandalas did not? Why did they believe the former to contain Kūkai’s true brushstrokes without applying the same designation to the latter, even though it had the same iconography and was also intimately associated with Kūkai and his teachings?

The Takao mandara’s medium of gold and silver paint on a dark ground played a role in structuring the perception of the paintings’ historical and cultural significance and allowed the emperor to feel especially close to Kūkai. This occurred in two ways. First, the color scheme of gold, silver, and purple or indigo at once signified the realm of the ethereal, approximated the lands of the deities, and, more broadly, suggested a world wholly removed from the present. These associations easily aligned with the fact that Kūkai had long been dead and existed only as an intangible religious and cultural force. Second, the gold and dark blue colors were closely related to the practices of calligraphy and sutra copying. The Takao mandara is clearly not an example of writing, but both Mongaku and Go-Uda describe them as consisting of Kūkai’s brush traces as if they were. Kūkai’s talents as a calligraphic master were widely known while he was alive. By the fourteenth century, legendary stories recast Kūkai’s calligraphy as possessing wondrous qualities in much the same way as a Buddha image, and reimagined Kūkai as effecting miracles through the power of his calligraphy alone. The combination of a set of paintings produced at Jingoji during Kūkai’s lifetime (even though it was after his residence there) with gold and silver pigments refigured them into objects that were particularly suited to transmit Kūkai’s aura of authority.

Images in gold and silver paint on a dark background were known in Japan as early as the eighth century. The earliest record of a painting with these characteristics dates to 754 and describes “a purple hanging scroll of Amida’s Pure Land with gold and black paint brought over by the great Tang monk [Jianzhen 鑑真, 688–763].”³⁹ Paintings employing this combination of pigments, however, seem to have been rare, and none are extant. On the other hand, sutra transcriptions completed in gold and silver on indigo-dyed paper were far more common. An account describing a copy of the Lotus Sutra produced for Emperor Kanmu 桓武天皇 (735–806) in 826 captures how the medium of gold ink on an indigo background elevated the object into the realm of the extraordinary:

It was gold letters on indigo paper and the rollers were made of jewels with an embroidered case. Everything down to the smallest detail [was exquisite, and the characters had] body and power. [They looked like] a string of jewels and a row of stars! It glittered and filled the eyes to the brim [with beauty]. A viewer exclaimed, “What a precious object! The calligraphy is nearly equal to [the great Chinese calligraphy masters] Zhong Yao (151–230) and Wang Xizhi (303–361).”⁴⁰

The night sky is a powerful analogy suggesting that the golden characters glittering on a dark background had their own light-emitting properties similar to a celestial body. This visual property served to suffuse the entire object with an otherworldly character that was made even more palpable by the jeweled rollers and embroidered case.⁴¹ Furthermore, the commentary that the calligraphy resembled that of the lauded masters from the distant past not only rooted the object in the cultural authority of China, it also provided a way for the author to remark on how the object appeared as if it could exist in a world that was superior to the present.

The Takao mandara was not the only version of Two Worlds iconography executed in silver and gold. At least three other versions are extant, including a version owned by the Smithsonian’s National Museum of Asian Art (fig. 14), and evidence for several more, now-lost copies are found in the historical record.⁴² Most significantly, a copy of the Mandalas of the Two Worlds in gold and silver ink on a purple woven background with mother-of-pearl inlayed rollers was produced in 1152 as part of the celebrations for Emperor Toba’s fiftieth birthday. For the occasion, the paintings accompanied a number of other objects such as Buddhist statues, bronze ornaments in the shape of tortoises and phoenixes, and embroidered textiles with Chinese twill.⁴³ Of note is the fact that the mandalas were hung and venerated in a space completely different from a typical esoteric ritual altar, a space that, as discussed above, included the Shingon patriarchal portraits. The mandalas here were not employed for rites of initiation into a Shingon lineage, but instead took on a celebratory character that required the use of precious materials to match the event of an emperor’s fiftieth birthday, a milestone not experienced in every generation. The medium of gold and silver ink on a purple textile raised the object up to a rarefied world filled with auspicious forces.⁴⁴

The Takao mandara’s specific color scheme and its close associations to sutra transcriptions take on greater significance in light of Kūkai’s fame as a calligrapher. One of the best-known examples of calligraphy with gold and silver ink on indigo-dyed paper associated with Kūkai is a copy of the Heart Sutra penned by Emperor Saga. Photographic reproductions of this particular object are prohibited, but later emperors followed in Saga’s footsteps and produced their own copies of the Heart Sutra in exactly the same format; an image of the work by Emperor Go-Hanazono 後花園天皇 (1418–1471) is shown here for expediency (fig. 15).⁴⁵ Saga completed his copy of this short text, which was accompanied by a frontispiece depicting the Buddha’s assembly, in an attempt to slow the spread of an epidemic that caused widespread suffering. Accounts describe that Kūkai then sought to amplify Saga’s act by reciting a commentary on the sutra, which later became codified in a new text titled The Secret Key to the Heart Sutra (Hannya shinkyō hiken 般若心経秘鍵, early ninth century).⁴⁶

Later legendary tales systematize this relationship into a wondrous exchange, clarifying that while Saga’s copying of the sutra was meritorious in and of itself, it was Kūkai’s exegetical performance of it that caused miracles to occur. For instance, the Illustrated Acts of the Great Master of Mount Koya (Kōya Daishi gyōjō zuga 高野大師行状図画, fourteenth century) depicts two scenes related to this event. In the first, Emperor Saga sits behind a desk with his newly completed, indigo-dyed scroll, which rests between a brush and an inkstone (fig. 16). Kūkai sits across from the emperor, and ritual implements placed in front of him indicate that he is in the process, or has just completed, his lecture on the Heart Sutra. The next scene focuses on a purported miracle that Kūkai raised the dead as a result of this lecture. The scene depicts the monk in a cemetery; figures in the background who were previously deceased begin to stir while Kūkai bestows perfumed water to alleviate the sufferings of others (fig. 17).⁴⁷

Other narratives describe Kūkai’s writing itself as having the ability to spontaneously effect miracles, much in the same way as Buddhist icons. One such story in the Illustrated Acts, for instance, tells of how Kūkai offered his calligraphy as compensation for a ferry ride. Long after he disembarked, other passengers physically ate fragments of the character he gifted to the captain as it became known to cure illness.⁴⁸ Stories such as this did not simply exist in the realm of fiction but were indicative of broader beliefs. Allegedly, Emperor Go-Uda also kept a fragment of Kūkai’s calligraphy in his personal medicine cabinet.⁴⁹

The above evidence demonstrates that Go-Uda’s conflation of the Takao mandara with Kūkai’s brushstrokes was rooted in the paintings’ place of origin at Jingoji as well as its medium of gold and silver ink on a dark background. The combination of the optical effect generated by the paintings’ colors, which imparted a sublime quality, and the elevation of writing associated with Kūkai to an otherworldly realm must have served as a powerful motivation for the emperor to request the repair of this specific object instead of, for instance, the polychrome mandalas. Of note are Go-Uda’s goals for this project. As an emperor who legitimized his rule by receiving ordination rites in competing lineages of Shingon Buddhist practice, he had also become an expert at invoking Kūkai to promote himself as the unifier of these traditions to solidify his authority. Go-Uda was involved in numerous projects to achieve these aims. He wrote a laudatory text on a portrait of Kūkai, for instance, that posits Tōji as the center of all esoteric ritual (fig. 18).⁵⁰ His sponsored restoration of Tōji, introduced above, also placed renewed focus on it as a place for the performance of rites dedicated to the protection of the state. The repair of the Takao mandara occurred as part of Go-Uda’s concerted effort to consolidate his reign around Kūkai and esoteric practice. Rather than copying the object, as Mongaku did to demonstrate his success in reviving Jingoji, Go-Uda requested its repair to further strengthen his privileged relationship to the ninth-century monk, and to emphasize his benevolence by preserving Kūkai’s perceived miraculous brush traces for the sake of the realm.

The final example of the preservation of a Buddhist image pivots from Kūkai and the Mandalas of the Two Worlds to examine a very different but equally important iconography in the premodern Japanese Buddhist landscape: the Taima mandara, an eighth-century tapestry consisting of a representation of Amida’s Pure Land (fig. 8). Little is known about the object’s early ritual use. By at least the thirteenth century, however, veneration of the tapestry, either by listening to pictorial recitations explicating its iconography (etoki 絵解き) or by contemplating and directing prayers to it, came to be widely understood as particularly conducive to aiding in the achievement of a Pure Land rebirth. Furthermore, the object’s medium—woven tapestry—rendered it efficacious for women. The Taima mandara’s status as an object tailored to women’s devotional practice is captured in the Miraculous Origins of the Taima mandara (Taima mandara engi emaki 当麻曼荼羅縁起絵巻, second half of the thirteenth century). The tale describes how the bodhisattva Kannon descended to earth disguised as a woman of the court to weave it at the request of a princess-turned-nun named Chūjōhime 中将姫. The miraculous narrative concludes when Chūjōhime, after a lifetime of learning about the tapestry through pictorial recitation and veneration, dies and achieves rebirth in the Amida’s Land of Bliss.⁵¹

Despite the object’s sacred iconography and perceived miraculous origins, records indicate that the Taima mandara was preserved poorly in its early history; from at least the twelfth century, the lower register had already become abraded to the point of being illegible.⁵² Unlike the versions of the Mandalas of the Two Worlds discussed above, the Taima mandara presented a unique challenge for those who wished to preserve it: sometime between the late twelfth and early thirteenth centuries, the monks at Taimadera 当麻寺, the temple where the mandara was kept, pasted the tapestry onto a wooden backing board in an attempt to prevent further deterioration.⁵³ Any restorations that were to take place would have to contend with the additional elements such as glue and wood that were introduced to the object’s storage environment. Despite these challenges, three major attempts to restore it took place in the premodern period, in 1217, 1505, and 1677. In each case, patrons opted to re-create the mandara to preserve its connection to Chūjōhime and its association with women, as they may have determined that repairing the original was too difficult. As stated above, re-creation as a third category of preservation combines the practice of copying an image afresh with the conceptual concerns of transmitting Buddhist cultural knowledge that could accompany repair projects. Within these three types of interventions, the efforts that took place in 1505 stand out as particularly illustrative of re-creation as a means of image maintenance and demonstrate how a newly made object could come to embody the same cultural connotations and numinous properties as the original miraculously created centuries before.⁵⁴

This repair project began with a nun named Kakuen 覚円 (dates unknown), who lived at Saihōniji 西方尼寺, a convent known for its dedication to Pure Land devotion and practice.⁵⁵ Saihōniji also housed a smaller version of the Taima mandara that is extant today (fig. 19). Despite Kakuen’s access to the iconography of the Pure Land, the nun expressed regret that the original tapestry was continuing to deteriorate, so she set off to raise funds to preserve it. Kakuen’s efforts are recorded in Getsuan’s Record of Soberness and Intoxication (Getsuan suiseiki 月庵酔醒記, sixteenth century), a literary work written by Isshiki Naotomo 一色直朝 (Getsuan, d. 1597?). Getsuan was a samurai who served in present-day Ibaraki Prefecture, but his vocation was acquiring knowledge of events in the capital.⁵⁶ Getsuan writes about Kakuen and notes her visit to Kyoto-area weavers to solicit threads as part of her fundraising efforts. According to Getsuan, Kakuen expresses sorrow regarding the fact that the mandala’s threads had begun to rot and decay. A weaver then responds to her sentiments: “Shouldn’t it be known that long ago, Chūjōhime, with the help of the Buddhas and the bodhisattvas of the three ages, set up a loom and wove [the Taima mandara] with hand-spun thread?” This directive prompts Kakuen to state her resolve: “No matter what, I will weave the object.”⁵⁷ The author then asserts that the nun set out to re-create the mandala using the same weaving techniques as employed for the original. While the resultant object is the same size as the eighth-century tapestry, it is executed entirely by a brush and in ink, color, and gold pigments (fig. 20).

The conflation of Kakuen with Chūjōhime, as described in the above account, is reflective of a broader effort to refigure the painted 1505 copy into an object that embodied the cultural authority of the original. The nobleman Sanjōnishi no Sanetaka 三条西実隆 (1455–1537) and the reigning emperor Go-Kashiwabara 後柏原天皇 (1464–1526) also engaged with the sixteenth-century version as if it were couched in the same religious framework and miraculous account that defined the tapestry. For instance, the original object used to contain an inscription at the center of the bottom register recounting the story of how Kannon came down to earth to weave it. As noted above, by Sanetaka’s time the Taima mandara’s entire lower register had become abraded and illegible; moreover, none of the extant copies made in the centuries beforehand bore a full inscription. Sanetaka remarks: “None of the [copies] of the mandara in wide circulation have a complete [inscription]; the characters appear only in places. This time, just as in the original [at the time of its creation], all will be recorded.”⁵⁸ As there was no way to verify the contents of the original inscription, Sanetaka inscribes a version of the story of Chūjōhime as it grew out of oral lectures and written commentaries beginning in the twelfth and thirteenth centuries.⁵⁹ The tale he recorded at the bottom of the painted copy could not have appeared on the original. It was through the process of defining the story and recording it on the sixteenth-century copy, however, that he solidified the painting’s connection to Chūjōhime and reinforced the work as created by divine intervention.

Inscriptions found on the roller bars of the 1505 version also offer a glimpse of the perceived efficacy of the re-creation itself. They note Kakuen as the primary patron and detail that, in addition to Sanetaka, the temple also sought Emperor Go-Kashiwabara’s hand to finish the calligraphic passages found in the cartouches around the edges. It further notes that the emperor intended for his involvement to serve as a memorial offering to his late mother, stating, “the new mandara’s inscriptions were intended for the salvation of [the emperor’s] late mother and were completed on the 20th day of the seventh month.”⁶⁰ Just as Chūjōhime achieved birth in the Pure Land for her devotion to the Taima mandara, so too did Go-Kashiwabara wish for his mother’s religious awakening as he demonstrated his devotion by contributing to the new version’s completion. The process of rendering the woven tapestry anew by re-creating its form in ink and color and appropriating its history provided the opportunity for the emperor to draw on the proven efficacy of the original to direct women’s prayers to his mother’s postmortem well-being.

While the Taima mandara’s restoration project of 1505 was closely linked to the desire to preserve the object’s association to the legendary Chūjōhime, the act of re-creating a sacred icon for reasons other than conservation had long been a common practice in Japanese Buddhist art production. The best-known example is the statue of Shakyamuni housed at Seiryōji 清凉寺. The tale of this statue begins with the Indian king Udayana, who sent his artisans up to the heavens to observe the historical Buddha preaching, whereupon they sculpted his true likeness out of fragrant sandalwood. Sometime later, the statue was evacuated from India to China due to political unrest, at which time a Japanese monk by the name of Chōnen 奝然 (938–1016) came across it. Wishing to possess an exact copy of the legendary artwork, Chōnen commissioned a second object. When the monk was making preparations to return to Japan, the two sculptures miraculously switched places and Chōnen ended up bringing the version supposedly carved at the behest King Udayana back with him to Japan.⁶¹ Scholars have determined that the extant “original” now housed at Seiryōji was most likely produced in the tenth century and sculpted out of cherrywood, not sandalwood as the story states. Nevertheless, documentation found deposited inside the sculpture refers to the image as a “sandalwood image.”⁶² In the same way as a premodern viewer might understand the sixteenth-century Taima mandara to be one and the same as the eighth-century original, the conflation of the sculpture at Seiryōji with the legendary Indian object occurred by imagining its materials to be those that originated within its perceived production context.

More recently, Samuel Morse has demonstrated similar logics at play in the re-creation of the Eleven-Headed Kannon at Hasedera 長谷寺, first carved in 727.⁶³ While the original statue was thought to have special miracle-granting powers, it was destroyed by fire; copies were created including one example by the renowned sculptor Kaikei 快慶 (d. before 1227) and another sculptor in his artistic lineage named Chōkai 長快 (act. mid-thirteenth century), in an attempt to reproduce the statue’s unique properties. Kaikei’s sculpture burned in 1238, but Chōkai’s is extant. Morse details how Chōkai sought to endow the new statue with the same numinous presence as the original Eleven-Headed Kannon. The sculptor did so, not by creating an exact copy of the appearance of the original, which had long disappeared, or of Kaikei’s version of it, but rather by incorporating fragments of wood and iconographic features associated with the original’s legendary history. In particular, Morse describes how the Eleven-Headed Kannon came to be associated with wood that came from trees struck by lightning. Subsequent sculptors sought out this kind of wood when fashioning their copies to demonstrate that their object comprised the same materials as the original.

Like the statue of Shakyamuni housed at Seiryōji and the Eleven-Headed Kannon at Hasedera, the patrons of the Taima mandara repair project in 1505 sought to transmit the object’s history and perceived efficacy for women by engaging the painted image as if it were woven tapestry thread. To be sure, any consecrated Buddhist object could serve as the seat of the deity in ritual practice, no matter its material. The above examples suggest, however, that Buddhist objects were not simply vectors for the deity; they also had the potential to transmit information about their own cultural and religious histories, institutions of which they were a part, and historical persons with whom they were associated. For instance, while multiple versions of the Mandalas of the Two Worlds executed in blue and gold were produced, it was only the Takao mandara created in the ninth century that had the power to transmit Kūkai’s powers by means of his purported brushstrokes. Likewise, although the nun Kakuen had access to a version of the Taima mandara at Saihoniji, where she practiced, only the eighth-century original was able to fully express Chūjōhime’s divine vision.

It is clear from Kakuen’s encounter with Kyoto-area weavers that she was highly cognizant of the legendary story of Chūjōhime and the miraculous conditions under which the original tapestry was understood to be made. Since repairing the eighth-century tapestry was not possible due to its unique storage conditions, Kakuen made use of a third mode of preservation, already in practice in premodern Japan, and re-created it by copying the image but treating it as if it were the restored original. Through the process of its re-creation, Kakuen and others involved in the project referenced both the narrative framework and the materials of the original: the nun commissioned a specially woven piece of silk, the same size as the tapestry, on which to paint the iconography; she imagined the painted surface to be woven; and Sanetaka inscribed an expanded version of the Chūjōhime story to emphasize the legendary princess’s role in the creation of the mandara. Blurring the boundaries between copy and repair, re-creation ensured the continued existence of Buddhist objects and their associated miracles and divine qualities.

Today, the conservation of Buddhist painting has largely become divorced from premodern modes of thought and instead centers on tasks such as removing creases or debris and stabilizing the object by remounting it to prevent further destruction. In the examples examined above, premodern patrons of Buddhist art were also concerned with the condition of objects, but unlike today, the preservation techniques that they drew upon also provide a valuable lens through which to examine the range of attitudes that existed toward these religious objects, especially for the elite. Patrons who sponsored the copying of an image, for instance, sought pristine iconographies to keep them in a superior state for viewing, studying, or ritual use, which also positively reflected larger agendas; at the same time, the preservation of an existing image also entailed the curation and perpetuation of cultural knowledge, such as information about Dharma transmission in the case of the Shingon patriarchal portraits and Kūkai’s role in legitimizing political authority as demonstrated by the Takao mandara. Re-creating Buddhist images combined the practices of copy and repair, and also transmitted information about production contexts, such as the efforts of the devout Chūjōhime, whose actions served as a paragon for women’s engagement in Pure Land practice as evidenced by the Taima mandara.

Crucially, the choice to copy, repair, or re-create an object necessarily entailed an engagement with its material properties, which played a significant role in structuring understandings of its meaning and function. The Takao mandara’s colors, for instance, enveloped the object in the glow of the divine. Its close associations to the practice of sutra copying, along with both Kūkai’s reputation as a famed calligrapher and the understanding that his calligraphy effected miracles, positioned the mandara as uniquely suitable for conveying Kūkai’s authority. The eighth-century Taima mandara could not be repaired, so patrons produced a painted re-creation of it in 1505 while imagining it to be, and engaging with it as if it were, produced from the same materials and circumstances as the original. Emperor Go-Kashiwabara’s prayer for the salvation of his mother also drew upon Chūjōhime’s precedent of venerating the tapestry, as the textile medium was particularly suited for women’s religious practice.

It is notable that the patrons of all three case studies examined in this article acknowledge the practice of calligraphy as central to a project’s completion. Emperors Saga and Go-Kashiwabara partially inscribed the patriarchal portraits and the 1505 version of the Taima mandara, respectively, and the Takao mandara was lauded for representing Kūkai’s brushstrokes. Writing allowed for privileged interaction with these images that was not necessarily possible through ritual practice. While these three examples are not representative of every instance of Buddhist painting repair in the premodern period, they are indicative of the ways in which the act of preserving an object could lead to highly personal forms of engagement with one’s set of beliefs. Premodern Buddhist painting repair went far beyond fixing damage. It provided opportunities to engage with and reaffirm worldviews and practices.