Toyok (Tuyugou) Cave 20: A Pure Land Cave Temple in the Desert with the Earliest Illustrations of the Visualization Sūtra

Abstract

In this paper, I examine the iconography, religious function and text–image relationship of one of the earliest illustrations of the apocryphal Visualization Sūtra, namely, the mural painting on the left (south) wall of Cave 20 at the Toyok Grottoes in Turfan created in the late sixth century. I study this mural with a structuralist approach to situate it within the overall pictorial program of the cave temple. I argue that this wall painting was designed as a visual and verbal cue to assist the meditating monks in situ to separately visualize the different individual visions constituting the “Thirteen Visualizations” taught in the sutra, excluding the five visions directly related to the Amitāyus Triad, namely, the eighth, ninth, tenth, eleventh and thirteenth visualizations. These five visualizations were likely facilitated by icons of the Amitāyus Triad placed in the center of the cave. The meditators started from the contemplations of the impure represented on the right (north) wall as preparational practices to eliminate their sins, improve their karma and thereby enhance their spiritual purity to a level appropriate for performing the Pure Land visualizations. They then turned to the left wall to perform the Pure Land visualizations represented there that end with the twelfth visualization: imagining oneself being reborn in Sukhāvatī. Continuing from the twelfth visualization, the meditators, in a state of meditational concentration, turned their faces to the rear wall to enter the Western Pure Land via the various lotus ponds represented on the rear wall.

This paper addresses the pictorial program of Cave 20 at the Toyok Grottoes, with an emphasis on the mural paintings on its left wall (Figure 1), which are among the earliest illustrations of Amitābha’s Western Pure Land (Sukhāvāti) and are believed to have been created based on the apocryphal Sutra on the Visualization of Amitāyus Buddha (hereafter referred to as the Visualization Sūtra).

By treating the pictorial program of Cave 20 as a whole, I further analyze the religious function of the Visualization Sūtra illustrations by situating them within the holistic architectural and pictorial program of the cave and thereby shed light on the religious activities performed in situ. By examining the iconography of each square-shaped unit of the painting on the left wall of the cave (Figure 2),1 I argue that this painting features the first seven visualizations, as well as the twelfth visualization, taught in the Visualization Sūtra. Moreover, the rear (west) wall demarcating the front and rear chambers of the cave (Figure 3), painted with various lotus ponds and rebirth figures, embodies the entrance to the Western Pure Land of Sukhāvāti. I argue that the five visualizations associated with the Amitābha/Amitāyus Triad, namely, the eighth, ninth, tenth, eleventh and thirteenth visualizations, were probably aided by portable icons of Amitābha, Avalokiteśvara and Mahāsthāmprāpta, which might have once been set up in the center of the front chamber. The contemplation of the impure, assisted by the painting on the right wall (Figure 4), was designed as a preliminary practice within the holistic Pure Land-entry meditational program performed in this cave. The contemplations of the impure transformed the bodies and souls of the meditators from impurity to purity, which enabled them to have proper visions, as instructed in the Pure Land visualizations, by improving their karma and moral purity. In the final step, the meditators confronted the main (west) wall to enter the Western Pure Land via the various lotus ponds depicted on it.

1. The Visualization Sūtra as Iconographic Source

The Visualization Sūtra is arguably the most influential text for believers of Amitābha’s Western Pure Land, as it is the only scripture that focuses on a concrete practice that leads to rebirth in Sukhāvāti within the three Pure Land sutras (Fujita 1990, p. 149).

The Visualization Sūtra comprises three components. The first part tells the narrative story of Queen Vaidehī, to whom the Buddha preached this sutra. In the second part, the Buddha instructs the queen on the practice of visualizing the thirteen different visions of Sukhāvāti to attain rebirth in this blissful realm. This part is commonly known as the “Thirteen Visualizations”. The objects of the “Thirteen Visualizations” are the Sun (1), Water (2), the Ground (3), Jeweled Trees (4), Jeweled Ponds (5), Jeweled Towers (6), the Lotus Seat (7), statues of the Amitāyus Triad (8), Amitāyus (9), Avalokiteśvara (10) and Mahāsthāmaprāpta (11), One’s Own Rebirth in Sukhāvāti (12) and the Amitāyus Triad in Sukhāvāti (13). The third part introduces nine different circumstances of being reborn in Sukhāvāti, arranged in a hierarchical way based on the varying karmas and religious endeavors of the reborn. This section is referred to as the “Nine Grades of Rebirth”. From the Tang dynasty (618–907 CE), this part was often combined with the second section, and together they are known as the “Sixteen Visualizations”, though the former is not intended to be objects of visualization.

2. Toyok Cave 20

The Toyok Caves are located in Shanshan County in the Turfan area, in the southeastern Xinjiang region, six miles to the east of the ancient capital of the Qocho Kingdom (mid-fifth–mid-seventh centuries). The site is divided into two precincts on the east and west cliffs of the Toyok Gorge. The west precinct comprises twenty-five rock-cut or brick-built caves, while the east precinct consists of twenty. These caves are arranged in four or five levels, many of which were designed as chambers for meditation.2 Cave 20 is located on the first level of the west group.3 It consists of a rectangular front chamber and a smaller back chamber. The back chamber was once numbered Cave 19, as the exact relationship between the two chambers could not be confirmed until recent excavations in the early 2010s.4 Cave 20 features what are probably the most significant wall paintings at the site of Toyok. Nowadays, the mural paintings there are in severe condition. Some of the paintings that will be discussed in this paper no longer exist in situ, both due to natural erosion and looting activities in the 19th and 20th centuries. At the beginning of the 20th century, German scholars Albert Grünwedel (1856–1935) and Albert von Le Coq (1860–1930) traveled to Turfan and surveyed a few cave sites there, including the Toyok Caves. Their investigation reports reveal significant information about the wall paintings in this cave more than a century ago (see Grünwedel 1912 and Von Le Coq 1913). Based on the results of carbon-14 dating, the paintings in Cave 20 are usually considered to be products of the sixth or early seventh century (Tulufanxue Yanjiuyuan and Tulufan Bowuguan 2017, p. 94) By comparing the stylistic features of the costume of the donor figure on the bottom of the left wall with corresponding images in the Mogao Grottoes at Dunhuang, archaeologist Jia Yingyi proposes the mid-to-late sixth century as a more probable date (see Jia 1995, p. 148).

2.1. The Mural Paintings in Cave 20

Cave 20 has a rectangular floor plan with a vaulted ceiling. The space in the cave resembles the shape of a half cylinder horizontally stacked atop a cuboid. The entrance generally orients to the east. This cave features the best-preserved mural paintings at the site of Toyok, even though they are still in quite poor condition. All murals in Cave 20 were actually painted on high places of the cave, at least 1.8 m from the ground level. Most of the murals on the rear wall are no longer there. Only a few square units, defined by green frames, each with a tree-like motif flanked by two small lotus flowers, remain on the southern half (left side) of the wall (Figure 3). On the right half of the rear wall there is a huge tree, the crown of which is filled with flowers or fruits, amid which a serpent-shaped water column goes up and down. There is an apsara flying downward to the tree, with his right hand holding a tray. Under the tree, a monk-like figure is painted in the meditation posture, with flames and water emanating from his shoulders and legs.

The mural painting on the left wall comprises four registers. The top register features three huge apsaras surrounded by suspended small lotuses (Figure 1). Close to the entrance side, there is a large moon motif encircled by a star atlas. Part of the second register has been purposefully peeled off; the remaining images show that this row consists of various meditating monks atop different miraculous mounts. These images of meditating monks riding miraculous mounts also exist on the right wall, seemingly to form an entirety with those on the left wall. The two lower registers depict various square units, within each of which there is a haloed monk who appears to be meditating on the image to his left-hand side. Based on currently available information, there are nine such units that remain generally legible: four on the third register and four on the bottom (Figure 2). The mural paintings on the right wall also feature four registers, suggesting a corresponding relation with those on the left wall (Figure 4). The top register also depicts apsaras. In contrast to the left wall, the two middle registers depict meditating monks who are mounted on miraculous vehicles. The bottom register is badly eroded, with only the two square units close to the side of the rear wall remaining legible. The monks in the two units are meditating on a bluish corpse lying on his belly with his unbound hair covering the floor and a female corpse with bundled hair whose arm is being pecked by a crow.

2.2. The Left Wall: Eight Visualizations

2.2.1. Controversies in Current Scholarship

Scholars generally agree that the painting on the left wall of Cave 20 is related to the Visualization Sūtra, although the exact meanings of the different units within it are still subject to debate. There are four inscriptions written in the tall, rectangular cartouches separating the different units of this wall painting (I1–I4).5 These inscriptions are highly similar to some sentences in the Visualization Sūtra but not identical. Therefore, Ning Qiang argues that there might have once been another version of the Visualization Sūtra circulating in the Turfan area, where the scripture is believed to have been compiled (Ning 2007, pp. 133–42). Miyaji Akira is one of the earliest scholars who systematically studied these wall paintings. He identifies six visualizations taught in the sutra from this wall painting, namely, the Ground (V3), Jeweled Tree (V4), Jeweled Pond (V5), Jeweled Tower (V6), Lotus Seat (V7) and One’s Own Rebirth (V12) visualizations, and he emphasizes it as a rare example of the early mode of Sukhāvāti visualization pictures in Central Asia, before the arrival of the pictorial tradition of the Tang dynasty (618–907 CE), which conquered the Qocho Kingdom in 640 CE.6 Nowadays, Chinese academia generally accepts this theory.7

In contrast, Yamabe Nobuyoshi challenges the role of the Visualization Sūtra in the composition of the murals on the left wall in Cave 20 (see Yamabe 1999, pp. 38–44; Yamabe 2002, pp. 123–52; Yamabe 2004, pp. 401–7; Yamabe 2014, pp. 217–62; Yamabe 1997, pp. 251–80). He adopts a different approach to interpreting the flame patterns that are frequently depicted in this painting by looking at a few other meditation scriptures that are also believed to have been compiled in the Turfan area, like the Guanfo Sanmei Haijing (Sutra on the Ocean-Like Samādhi of the Visualization of the Buddha, hereafter referred to as the Ocean Sutra), Chan miyao fa jing (A Manual of the Secret Essentials of Meditation, hereafter referred to as Secret Essentials), Wumen chanjing yaoyong fa (The Essence of the Meditation Manual Consisting of Five Gates, hereafter referred to as Five Gates) and Zhi chanbing miyao fa (The Secret Essential Methods to Cure the Diseases Caused by Meditation, hereafter referred to as Methods for Curing), to argue that the flame patterns frequently appearing in this wall painting was depicted based on this group of meditation sutras rather than merely the Visualization Sūtra. He also brings up a significant issue: the different visualizations in this wall painting were depicted in neither a complete nor an ordered manner; thus, he goes one step further to suggest that the Visualization Sūtra might not have been standardized into the current version at the time when the mural painting was produced. Various oral teachings might have been active in this region that mingled the teaching of the Visualization Sūtra with other meditation texts also born in the Turfan area. Yamabe’s provocative theories further complicate the understanding of this wall painting, as well as the history of the compilation and transmission of the Visualization Sūtra more broadly. More recently, Eric Greene proposed a new theory on the function and nature of these wall paintings in Cave 20. He considers them persuasive depictions of “confirmative visions” assuring certain meditational attainments, mainly based on the way in which the meditation monks are delineated: they are depicted as haloed monks emitting fire and water from their bodies.8

In this section, I will first examine the iconography of the lower two registers of the left wall mural and the inscriptions in them to argue that they were painstakingly designed as visual and verbal cues to aid the Pure Land visualization practices performed in this cave. The images and inscriptions were based on the Visualization Sūtra, yet they are not simply a pictorial transliteration of the sutra together with textual excerpts from it. Instead, they were carefully curated by the designers to cater to the practical needs of the different visualizations entailed in the varying, complex mental processes involved here. I will examine each square unit from the side of the rear wall (west) to the entrance side (east). The two square units on the western end of the left wall are the least controversial ones. I will use the numbering system given by Miyaji to refer to each square unit (After Miyaji 1995a, plate 3).

2.2.2. Jeweled Tower Visualization (V6)

On the western end, the seated monks in II. 1 and III. 1 are no longer completely legible; a crack on the wall surface runs downward through their bodies. However, the scenes in front of them are still legible. The scene in II. 1 (Figure 5) depicts a Chinese-style, three-story pavilion with an entrance on the bottom level, railings on the middle level and cross-shaped windows on the top level. There is also trapezoidal infrastructure painted in blue below the pavilion, seemingly alluding to the suspended quality of the Jeweled Towers in Sukhāvāti. On the left of the pavilion, a Chinese harp and a lute were painted next to the top and middle levels of the structure. This seems to be direct evidence pointing toward the sixth visualization taught in the Visualization Sūtra: the Jeweled Tower, as the scripture describes, “There are also musical instruments suspended in the sky, which, like those on the heavenly jeweled banners, spontaneously produce tones even without a player”(Taishō, v.18, No. 365, p. 345).9

2.2.3. Jeweled Pond Visualization (V5)

Below II. 1, III. 1 (Figure 6) features an image rarely seen in Buddhist art outside the Turfan area: four round, green buds arranged in two rows, each supported by a light-brown stem. There are two rectangular bars below each row of green buds, one painted in green and the other left in white but filled with whirly patterns resembling water waves. Small lotus buds and flowers were painted inside the two white bars. Within each row, a zigzagging white band also marked with wave patterns is depicted as “flowing up and down between the green buds”. Although there is no visual precedent for such a depiction, Miyaji suggests that it might be a representation of the fifth visualization: Jeweled Ponds (see Miyaji 1995a, pp. 29–30). He points to a statement in the sutra that reads, “In each stream there are sixty koṭis of lotus flowers of the seven kinds of jewels, which are round and symmetrical, measuring twelve yojanas in diameter. The water from the mani-gem flows among the flowers and meanders between the trees” (Taishō, v.18, No. 365, p. 345). I think Miyaji’s suggestion is insightful, as this sentence is the only content in the Visualization Sūtra that corresponds to the serpent-shaped water column depicted in the mural. Building on his identification, we could further interpret the green and white bars as the gold banks and the water bodies of these eight ponds in the Western Pure Land.

The trees were depicted quite differently from the other trees in the same wall painting. Those in II. 3, II. 5 and III. 3 all have pointed tops, with multicolored flowers on their crowns, in contrast to the monochromatic, roundish “crowns” in III. 1. According to the text quoted above, the trees are not supposed to be in the ponds. Instead, they are amid the eight ponds and connected with the latter by the serpent-shaped water columns springing from mani-gems in the ponds. Within the white bars below the two rows of trees, there are many small lotuses depicted, some with their buds open and some with them closed. Again, those white bars embody the zone of water ponds. In this case, the trees serve as a background for the water columns coming out of the ponds; therefore, the detailed features of the trees are omitted. The water columns seem to be the center of visualization. According to the sutra, as the columns ripple, they produce exquisite sounds, which proclaim the truths of suffering, emptiness, impermanence and no-self, and of the pāramitās as well as praise the physical characteristics and marks of the buddhas (Inagaki and Stewart 2003, p. 71). This simple motif seems to evoke quite rich visual and audio information for the meditators. The combination of trees, water columns and ponds also appeared on the mural of the rear wall, which will be analyzed later.

2.2.4. Lotus Seat Visualization (V7)

Depiction III. 2 (Figure 7) features three stems of lotuses framed by suspended ribbons forming an inverted, heart-shaped banner. A radiating jewel was painted blue and brown with wings-like flames slightly above the right side of the banner. An ink inscription was written on the right of III. 2. The inscription can be translated as “the practitioner visualizes a jeweled bannered supported by four columns on a dais, with a jeweled canopy atop the banners resembling the palace in the Yāma Heaven”.10 For this reason, III. 2 is usually identified as a representation of the Lotus Seat Visualization. The images in III. 2 further support this idea: the presence of a suspended canopy formed by banners with radiating jewels on it is described in the seventh visualization.11 A minor difference from what is described in the sutra is that the banner here is supported by three lotus stems rather than four columns on a dais. This seems to refer to an earlier step of this particular visualization describing the large lotus from which the dais emerges: “This lotus flower has eighty-four thousand petals. Between the petals, there are a hundred koṭis of king maṇi-gems as illuminating adornments”.12 As shown in III. 2, there is a round motif comprising a brown rim and an inner oval marked by a few black lines depicted inside the bud of each lotus. I suspect that the round motif should represent such maṇi-gems within the lotus petals, though they are depicted in a way different from the polygonal gems in II. 2, II. 4 and III. 3, as well as the round jewel with flames on both sides resembling candy in III. 2 and III. 4. According to the Visualization Sūtra, the jewels within the lotus petals, the gems forming the dais and those adorning the banners are all different.13 It seems that the artists were intentionally distinguishing these different jewels, though the polygonal gems were most frequently seen in Central Asian Buddhist art. In this case, the round motif in III. 2 should represent the maṇi-gems within lotus petals. The polygon gems in II. 2 represent the different types of jewels forming the dais atop the lotus: the polygon shape seems to be deliberately chosen to embody gems forming a rectangular ground, like the case in II. 4 (Jeweled Ground). The candy-like image on the top of III. 2 represents the excellent jewels adorning the four-columned banners.

The coexistence of lotus, dais and jeweled banners in the mural seems to be a juxtaposition of the visual hints of two consecutive steps within the Lotus Seat Visualization. Therefore, the painting seems to be a detailed step-by-step manual or guide for the visualization practices taught in the Visualization Sūtra. The key difference between the inscription mentioned above and the corresponding text in the sutra is the addition of the starting phrase that reads, “the practitioner visualizes (xingzhe guan)”, whose practical nature seems to support this assumption further.14

Above III. 2, the three stems of lotuses growing from a pond with wavy water and two waterfowls depicted in II. 2 (Figure 8) seem to be related to an earlier step of the same visualization (the Lotus Seat). This image is generally identified as a generic representation of “lotus flowers in a water pond”, not corresponding to any particular visualization among the thirteen. However, I think it is quite likely that II. 2 represents the inaugurating step of the same visualization aided by the image in III. 2: the Lotus Seat. The Lotus Seat Visualization opens with the following instruction: “Those who wish to see that Buddha should form an image of lotus flower on the seven-jeweled ground”.15 As previously analyzed, the three lotuses in II. 2 feature radiating polygon gems forming the dais that holds the banners.16 Both III. 2 and II. 2 feature three lotus stems arranged almost in the same way. It seems that they are ontologically the same. The number three might allude to the Amitāyus Triad, which will be visualized next. If this wall painting is a step-by-step visualization guide, then it is possible that II. 2 represents the first step of the Lotus Seat Visualization, whereas III. 2 represents the second and third steps. While looking at II. 2, the practitioners would likely have focused on the emergence of the lotus flowers and the radiating lights emitted by the maṇi-gems within their petals. After that, the practitioners switched to III. 2 to visualize the formation of the dais by the various polygon gems and the rising banners against the lotuses. Most scholars have a tendency to link each square unit in this wall painting with a particular visualization and thereby ignore the possibility that, in real practice, a more complicated visualization could entail the visual assistance of multiple images. As will be discussed in the next paragraph, the images in II. 3 (Figure 9), III. 3 (Figure 10) and II. 5 (Figure 11) are all related to a single visualization, the fourth: Jeweled Trees.

2.2.5. Jeweled Tree Visualization (V4)

By studying the images in III. 3, II. 5 and II. 3, which all feature the tree motif, I argue that they embody the three steps of the Jeweled Tree Visualization. I think the number of images assisting each visualization entails the varying complexity of the different visualizations. Moreover, the inscriptions are curated as verbal cues to aid visualizations rather than as direct copies of excerpts from the Visualization Sūtra.

III. 3 depicts a tree with three flowers on its crown, flanked by two lotuses with radiating jewels on the top. More importantly, there are two flying pennants extending from the crown, seemingly blown by the wind to both sides of the tree. An inscription accompanies the images in III. 3, which can be translated as “the practitioner visualizes the tree leaves: each tree leaf (should be) contemplated as a hundred types of jewels; on the two sides of each tree, there are two jeweled banners”.17 The inscription corresponds quite well with the images in III. 3, especially in terms of the two flying banners extending toward both sides of the tree crown. Scholars have been unable to find a particular sentence in the Visualization Sūtra that matches this inscription. As mentioned earlier, the inscription in III. 2 also slightly varies from a similar sentence in the same sutra. Therefore, Ning Qiang suggests that there might have been a different version of the Visualization Sūtra circulating in sixth-century Turfan (Ning 2007, p. 137).

However, a hidden presumption here is that the inscription composer intended to exactly copy certain excerpts of the scripture onto the wall painting. Such an assumption cannot be taken for granted, as these inscriptions are more likely to be fragmented, verbal cues assisting the visualization practices rather than products of a holistic sutra-engraving project intended to preserve and disseminate Buddhist teaching. Such projects were widely sponsored in northern China in the mid-to-late sixth century under the anxiety of the Buddhist eschatology “the Ending of the Dharma”, but they are rarely seen in the Turfan area.

I argue that the inscription composer had the active agency to curate the written inscriptions, tailoring them to the needs of the particular visualizations performed in the cave shrine rather than simply choosing complete sentences from the sutra text. A detailed analysis of these inscriptions clearly supports this idea. First, despite the fact that the sutra uses different wording to open the Buddha’s instruction on how to perform each distinct visualization, all four legible inscriptions on this wall painting start with the phrase “xingzhe guan”, which means “the practitioner visualizes …”. Although the term “xingzhe” appears a few times in the Visualization Sūtra, it mainly exists in the part of “Nine Grades of Rebirth” but never in the instructions on the Pure Land visualizations represented here. The deviation from the sutra text aiming at creating a standard subject for the inscriptions not only alludes to the inscriptions’ nature as pragmatic, verbal hints for visualization practices, but they also reveal the existence of the subjective discretion of the composer. These ink-written inscriptions are short, usually about twenty characters. For some relatively more complicated visualizations, copying any single sentence from the sutra would not be sufficiently helpful to aid the visualization practice. In other words, the composer needed to select a few different phrases dispersed in the passages on each visualization to form a new, twenty-word sentence in order to aid the meditator in the most effective way. For example, the inscription in III. 2 is a combination of two short sentences in the sutra on the “banners supported by four columns” and “jeweled canopy atop the banners”, but the sentence between them, which assimilates the various four-columned banners with the myriad Mount Sumeru, was intentionally omitted.18

In this case, it seems that the composer was well-versed in the sutra text but considered the middle sentence to be a repetitive description of the four-columned canopy and deliberately omitted it. Returning to the inscription in III. 3, it is often considered to be the most unusual one in this cave, as it seems to be different from any particular sentence in the Visualization Sūtra. The first part of the inscription, “the practitioner visualizes”, invites the practitioner to visualize the tree leaves, which is followed by a more detailed instruction that teaches the practitioner to contemplate these leaves as hundreds of types of jewels. This part seems to be rephrased from a sentence describing the leaves of the jeweled trees in the Visualization Sūtra: “each blossom and leaf has the colors of various jewels”.19 Notably, this inscription includes a second verb other than the standard “visualize (guan)” in the beginning, “contemplate (nian)”, which further displays its function as a verbal cue for meditation. The second half, “there are two jeweled banners on both sides of each tree”, embarks on a different object of visualization, the jeweled banners (baochuang). This part is likely to be related to a sentence at the end of the passage on the Jeweled Tree Visualization: [the fruits emitting great lights] transform themselves into banners and innumerable jeweled canopies(Taishō, v.18, No. 365, p. 345).20 Therefore, the inscription in III. 3 also seems to be a combination of two separate sentences in the Jeweled Tree Visualization, with some degree of rewording.

The image in II. 5 is also associated with the Jeweled Tree Visualization, as it features an inscription reading, “each jeweled tree is veiled by seven layers of nets. Between each of the nets …”(Taishō, v.18, No. 365, p. 345)21. Except for the standard opening phrase, “the practitioner visualizes”, this inscription is almost identical to the corresponding sentence in the sutra text (Taishō, v.18, No. 365, p. 345).22 Therefore, the image in II. 5 is often identified as a representation of the Jeweled Tree Visualization (for an example, see Miyaji 1996, pp. 55–56). On the visual side, II. 5 depicts a large tree in a pond, with four flowers on its crown, segregated by two intersecting ribbons embedded with diamond-shaped jewels. The decision to depict four flowers on a tree instead of three, as seen in II. 3 and III. 3, might be related to a sentence directly following this inscription’s counterpart in the sutra that describes four kinds of colors of flowers on the Jeweled Trees.23

More interestingly, there are two flames burning from the tree. Based on these fires, Yamabe argues that this wall painting was not painted solely based on the Visualization Sūtra but was also informed by the teachings of several other meditation sutras popular in the Turfan area during the sixth century, as he considers it unreasonable for the Jeweled Trees in the Western Pure Land to be on fire (Yamabe 1999, pp. 38–44). However, scrutinizing the passage on the Jeweled Tree Visualization in the Visualization Sūtra, it seems that the intention to depict flames here was not to represent real fire. This passage includes a sentence that states, “These trees have marvelous blossoms which are in the color of gold from the Jambu River, resembling spinning fire wheels” (Taishō, v.18, No. 365, p. 345).24 The flames in II. 5 are depicted in a more realistic manner than other flames in the mural. It is likely that the flames on the tree here were formed by the analogy of the flaming wheel made in the sutra than that they represent real fire. There is also no change of condition due to the effect of burning represented on the tree.

In contrast to III. 3 and II. 5, II. 3 has no inscription written next to it, although it also features the tree motif. More specifically, it depicts two trees grown from a pond with two swimming waterfowls. For this reason, II. 3 is also identified as an aiding image for the Jeweled Tree Visualization (see Miyaji 1996, p. 54). However, based on this situation, Yamabe again questions the wall painting’s relevance to the Visualization Sūtra from a different perspective: the images seemingly associated with the “Thirteen Visualizations” taught in the sutra were depicted in an unbalanced and chaotic way (Yamabe 1999, p. 44). In other words, some visualizations are missing, and some are represented repetitively. In later sections, I will argue that images related to all “Thirteen Visualizations” might have all been embodied in this cave, although not completely on the left wall.

Here, I want to propose an alternate logic according to which the designer of the wall painting determined the arrangement of images related to different visualizations: the level of complexity and difficulty of each visualization. Simply put, from a practical perspective, the technical difficulty of constructing an eidetic image of the object of each visualization is not the same: some visualizations feature more objects to be visualized in a multistep process. The complicacy of different visualizations can be roughly evaluated based on the length of their respective passages of instruction in the sutra. Among the visualizations represented in this wall painting, the sutra text on the fourth visualization, Jeweled Trees, is the longest: 411 characters. The seventh visualization, the Lotus Seat, follows closely: its corresponding text is 335 characters in length. This might possibly explain why the Jeweled Tree Visualization has three related images (II. 3, III. 3, II. 5) with two inscriptions, and the Lotus Seat Visualization has two related images (II. 2, III. 2) with one inscription. It could also be the reason why inscriptions were not written next to every image, facilitating Pure Land visualizations on the mural. In the case of the Jeweled Tree Visualization, the three images address different objects of the visualization. Depiction II. 5 focuses on the flowers and the seven-layered nets housing trillions of palaces on the Jeweled Trees. Depiction III. 3 targets the leaves and fruits on the Jeweled Trees, and II. 3 provides a panoramic view of the various Jeweled Trees in Sukhāvāti. The images of the Lotus Seat Visualization reveal a two-step visualization process: depiction II. 2 aids with visualizing the myriad lotuses supporting a dais made up of various maṇi jewels on the flowers of these lotuses, and III. 3 assists with visualizing the four-columned banners on this dais. Admittedly, looking at the text length is a quick-and-dirty way to assess the complexity of each different visualization, yet it enlightens us on the perspective of the practical need of the meditating practice in the approach to the image arrangement on this wall painting.

2.2.6. Water Visualization (V2) and Ground Visualization (V3)

The images in II. 4 (Figure 12) are divided into two parts, separated by a horizontal, white plank. The upper section shows two white, hexagon-shaped gems that seem to be on fire and fuming. The lower section depicts a chessboard-like motif comprising various diamond-shaped units painted in red, blue and green. These diamond units are demarcated by alternating white lines. Within each unit, a diagram made of four small rectangles resembling the Chinese character “tian 田” is painted in ink. Depicting a chessboard-like motif was a common way to represent the beryl ground in Sukhāvāti in the Xinjiang area. Based on a similar wall painting collected by British archaeologist Aurel Stein (Figure 13) during his visit to the Toyok Caves in 1907, which features the inscription, “the meditation master visualizes the seven-jeweled ground”, Miyaji associates II. 4 with the third visualization, the Ground Visualization (Taishō, v.18, No. 365, p. 345).25 Of note, the Ground Visualization passage in the Visualization Sūtra is quite brief and void, stating that if you attain the state of samādhi (meditational concentration), you will see the ground so clearly and distinctly that it will be impossible to describe it in detail.26 A concrete description of the beryl ground is actually in the passage of the second visualization, the Water Visualization. It states, “On this beryl ground, golden paths intercross like a net of cords. The Land is divided into areas made of one or the other of the seven jewels, so the partitions are quite distinct”.27 The alternating white lines on the chessboard motif in II. 4 should represent those golden ropes on the beryl ground described above. The diamond units painted in different colors seem to refer to the seven jewels filling the beryl ground.

The flames on the gems representing the basic units of the beryl ground have caused many controversies. If the chessboard motif in the lower section represents the beryl ground of Sukhāvāti, then the two gems on the top seem to be a zoom-in vision of the various jewels making up the ground. However, Yamabe again points out the irrelevancy of the flames and smoke on the gems (Yamabe 1999, p. 40). I propose that these flames represent the radiating lights of the gems. A fragmented wall painting also collected from Toyok by Aurel Stein (Figure 14) features the exact same type of gem with flames (See Miyaji 1995a, p. 38). The inscription accompanying this painting is “the meditation master visualizes the light of the jewels”.28 In the mural paintings in other caves at the same site, we can also see many bodhisattvas wearing the same type of polygon gem on their crowns, emitting such flames.29 Using flame patterns to depict the dazzling, radiating light of a buddha or bodhisattva was a common practice in Central Asia as well as Medieval China. Here, the jewels are simply depicted as emitting lights.

The badly eroded image in II. 6 seems to be related to the Ground Visualization. Although the images in II. 6 are barely legible now, fortunately, using an image-synthesizing technique that relies on digital means to integrate the different layers of images generated by visible lights, infrared light, ultraviolet-induced luminescence and visible-light-induced luminescence, Yamabe in collaboration with Academia Turfanica, enhanced the visibility of the devastated images in II. 6 (Taishō, v.18, No. 365, p. 345).30 Yamabe partially identified the inscription in II. 6 as “xingzhe [][][] chuang qing di”.31 The Water Visualization passage contains a related sentence that means “supporting the beryl ground from below are golden banners made of diamond and the seven kinds of jewels”.32 Based on the stereotype of the inscriptions observed in the previous part and the fragmented traces left on the wall, I suspect that this inscription might be “xingzhe guan qibaochuang qing di”, meaning “the practitioner visualizes the banners made of seven kinds of jewels supporting the (beryl) ground”.33

The Water Visualization and Ground Visualization are mingled together in the sutra text: the appearance of the beryl ground is mainly introduced in the passage on the Water Visualization as a background of the omnipresent water in Sukhāvāti. Considering that the text description of the “seven-jeweled banners” mentioned in the II. 6 inscription is prior to the text describing the “gold-rope demarcated ground” represented in II. 4, within the Water Visualization passage of the Visualization Sūtra, it seems more likely that the image in II. 6 aids the Water Visualization and that of II. 4 assists the Ground Visualization.

2.2.7. The Worshipper’s Own Rebirth in the Pure Land Visualization (V12)

The image in III. 4 (Figure 14) is the most enigmatic one among all those painted in this cave, especially in terms of the relationship between its lower and upper parts. The lower part appears to be a small house, with a disproportionately huge gate painted in brown whose two doors are opening inward. The roof of the house overlaps with the green oval with a brown rim in the upper part. There are nets hanging down from the green oval. Within the oblate, green oval, there is a small figure seated in a semi-open lotus flower, which appears to be growing from the roof of the house. Three gems with flaming lights resembling the shape of a wrapped candy, two inside the green circle and one outside, are painted in the upper part.

Pioneered by Grünwedel, who described this image as “a tree blocked by a small house” in his 1902 travel report, many scholars hold such an identification (Grünwedel 2007, p. 590). Rather than viewing the brown structure on the bottom as a house, it might be more accurate to call it a gate pavilion: the only emphasis of the brown structure appears to be the opening doors. Noteworthily, the nets hanging down from the green oval are partially blocked by the brown house, especially when they overlap with the upper beam of the gate. This seems to be a visual device to allude that the green oval might be behind the gate. The roof of the gate pavilion also partially blocks the bottom edge of the green oval, suggesting the same spatial logic. Admittedly, this spatial relationship is suggested by minor visual details. However, considering that these murals are meditation-facilitating images painted in high places (about two meters from the floor), it seems impractical to place part of the image entirely behind the other, which jeopardizes the legibility and clarity of both. Based on previous analysis of the image assisting the Water and Ground Visualizations (II. 4), it seems that when a square unit comprises two vertically divided parts, the upper part might be a zoom-in vision of the key element in the world embodied by the lower part.

I argue that the huge gate in the lower section might have been designed to aid the mental process of imagining oneself being reborn in the Western Pure Land. The inscription in III. 4 means “the practitioner should give rise to his own mind and imagine that (he) is reborn in a lotus in the Land of Bliss to the West”.34 This highly resembles a sentence in the twelfth visualization: One’s Own Rebirth in the Pure Land visualizations.35 Therefore, it seems easy to identify the upper section as a “rebirth figure seated in a closed lotus bud”, but the huge gate below becomes an enigma, as the Visualization Sūtra does not describe anything similar to such a gate. Considering that these images were painted as visual aids for this particular visualization, this gate should represent either a part of the content of the twelfth visualization or one of the few steps of this visualization. Because no relevant texts on this gate can be found in the Visualization Sūtra, it seems more likely that it represents and facilitates a vision that functions as a transitional experience during the visualization process. Fourth-century Chinese monks often compared the experience of rebirth in Sukhāvāti to that of “pushing in the heavenly gate”.36

Though there is no textual evidence stating that this idea also existed in sixth-century Turfan, the image of the heavenly gate had been incorporated into contemporaneous Buddhist art at Dunhuang. The image of the heavenly gate with the iconography of paired gateposts (known as que) widely exists in the funerary art of the Han dynasty (202 BCE–220 CE), embodying the entrance to the Chinese Heaven above the axis mundi known as Mount Kunlun (Zhao 2023a, pp. 85–94). In fifth-century Dunhuang, this iconography was adopted to represent the entrance to [the palace in] Tuṣita Heaven dwelled by Maitreya Bodhisattva (see Figure 15).37 In a mid-sixth century cave at Dunhuang, a gate pavilion with both doors opening inward, very similar to the Toyok piece, was depicted as the entrance to [the palace] in Trāyastriṃśa Heaven, above Mount Sumeru, the Buddhist axis mundi (Figure 16) (see Shen 2023, pp. 4–13). Zhao Yi argues that early clerics in China identified Mount Sumeru with Mount Kunlun, Trāyastriṃśa Heaven with the Chinese Heaven (Zhao 2023a, pp. 99–103). It seems both the idea and motifs of the heavenly gate had already penetrated the Buddhist iconography.

There is psychological consistency between the action of entering a gate and rebirth into a new world. Considering that the lotus bud containing the reborn figure seems to be intentionally painted “in a space behind the huge gate”, it is likely that the visualizer needs to imagine himself pushing the doors and entering the gate to arrive at the lotus bud and continue the following visions. If this is the case, the gate facilitates the initial step of visualizing oneself being reborn in Sukhāvāti. The green oval seems to represent the second step, in which the reborn is sealed in the closed lotus bud. Moreover, the flaming jewels depicted inside and outside the lotus bud seem to mark a particular moment when the lotus is about to blossom, as the sutra states, “as the lotus flower opens, five hundred rays of colored light illuminate your body” (Taishō, v.18, No. 365, p. 345). In addition, this representation of a small lotus flower in a larger lotus bud seems to again embody the idea of the existence of different kinds of lotuses in the jeweled ponds of Sukhāvāti.

The final step of the twelfth visualization is represented in III. 5. The images in III. 5 have been terribly eroded and are not legible at all. Again, thanks to the image-synthesizing technique mentioned earlier, Yamabe offers us a multi-spectrum fusion picture, showing the general silhouette of the III. 5 images (Figure 17). This picture clearly indicates that a haloed figure with radiating lights on both sides is depicted seated on a T-shaped platform emerging from a fully blossomed lotus flower. This image seems to eloquently represent the third and final step of the twelfth visualization: imagining that, after the blossoming of the lotus flower, the newborn opens his eyes to see paradise without obstruction. The fragmented inscription in III. 5 further echoes this identification: the still-legible parts read as “xingzhe diguan zi [][] lian”, which is undoubtedly talking about someone seeing himself in the lotus.38

So far, I have identified images related to seven visualizations among the thirteen taught in the Visualization Sūtra, namely, Water (V2), Ground (V3), Jeweled Trees (V4), Jeweled Ponds (V5), Jeweled Towers (V6), the Lotus Seat (V7) and One’s Own Rebirth in Sukhāvāti (V12). Among the twelve images on the middle and lower registers of the left wall, eleven images have been studied, with III. 6 the only exception, which seems to have been completely devastated even by the time of Gründewel (Grünwedel 2007, p. 588). Local archaeologist Jia Yingyi, based on his firsthand observation, proposes that the remaining trace on the wall appears to resemble a round object, possibly representing the sun (see Jia 1985, pp. 281–83). The first visualization taught in the Visualization Sūtra is the Sun Visualization, which instructs the visualizer to imagine a red sun about to fall, like a suspended drum (Taishō, v.18, No. 365, p. 345). If III. 6 represents the Sun Visualization, then the painting on the left wall should feature all the initial seven visualizations and the twelfth visualization. Noteworthily, the remaining five visualizations share a key commonality: they are all related to the images of the Amitāyus Triad. As will be discussed in the following section, I argue that they were facilitated by icons of the Amitāyus Triad, possibly placed at the center of the cave [front chamber].

3. The Right (North) Wall: Impure Visualizations

Similar to the left wall, the right wall of the front chamber also comprises three horizontal registers. The two upper registers depict various haloed monks seated in meditating positions, emitting fire and water from their knees and shoulders. Different real and miraculous animals are painted slightly below the mediating monks, suggesting that they serve as the mounts for the meditators. There are four groups of animals that remain legible on the top register of the right wall, namely, ducks, vultures, gold-wing rocs and winged horses (from rear wall side to front wall). Differently, the haloed monks in the middle register feature no mounted creatures but a flying lotus pedestal instead. As previously mentioned, similar meditating figures were also painted on the top of the left wall, which are hardly legible now. Based on the report of Gründewel, these monks were depicted riding white oxen, flowing clouds, trees, flying lotus flowers and suspended thrones decorated with mountains, serpents and lions (Grünwedel 2007, p. 589).

3.1. The Story of Sumāgadhā

In this section, I argue that these mounted, meditating monks are a decontextualized representation of the narrative story of Sumāgadhā, created to accentuate the magic power of spanning distance and traversing realms that are achieved in meditation. These types of mounted, meditating monks are usually identified as the various disciples manifesting their miraculous power to fly to the city of Pundravardhana, as described in The Story of Sumāgadhā.39 This story was incorporated into many different sutras, with minor variations in the plots. It stands alone as a scripture known as Xumotinü jing (Sumāgadhāvadāna), first translated by Zhi Qian in the Wu Kingdom period (Taishō, v.2, No. 128, p. 835). The story has been briefly summarized by Yamabe:

“Sumāgadhā was the beautiful and devout daughter of Anathapindika, a famous patron of Buddhism, who married a son of a rich merchant in the distant city Pundravardhana. Her husband’s family were followers of Jainism, and in order to convince them of the Buddha’s greatness, Sumāgadhā offered incense to invite the Buddha and his disciples to a meal the following day. The smoke of the incense miraculously travelled to Śravasti (the capital of Kosala kingdom), where it appeared in front of the Buddha, who immediately understood the message, and instructed those of his disciples who had acquired miraculous powers to fly to Pundravardhana. The next morning, these great disciples, and finally the Buddha himself, flew on various animals and other vehicles to Pundravardhana, where gave teachings and converted the family” (Yamabe 1999, p. 41).

Though scholars generally agree that these mounted, meditating monks were created based on The Story of Sumāgadhā, the reason for painting this narrative of converting non-Buddhists to Buddhism in this Pure Land-dominated cave chamber remains controversial. As part of his larger argument to connect various motifs in this cave with the Ocean Sutra, Yamabe considers that the creation of these mounted monks was informed by the motif of flying monks mentioned twice in the sutra (ibid., pp. 42–43). Miyaji simply considers it as a motif related to the general practice of meditation (Miyaji 1995b, p. 28). Building on this, Eric Greene views them as symbols of the miraculous power attained through achieving meditational concentration (samādhi) (Greene 2021, p. 104). The flame and water emitted from these meditating monks seem to support this idea, as depicting the emissions of flame and water from legs and shoulders was a common way to embody the miraculous power of a certain religious master, first seen in Gandhāran art.40 Not only those traveling disciples but also all the monk-like meditating figures painted in Cave 20 are depicted spraying fire and water, as well as with halos, which seem to emphasize the efficacy of meditation practices in the pursuit of gaining magical power and spiritual advancement. The Buddha also instructed on the magical power of controlling fire and water gained from the meditations of fire-king samādhi and water-king samādhi, respectively, in The Story of Sumāgadhā(Taishō, v.2, No. 128, p. 835). Ren Pingshan further argues that displaying the miraculous power acquired by performing Buddhist practices played an indispensable role in attracting followers of Buddhism (Ren 2015, p. 232).

The subject of The Story of Sumāgadhā also appears in the contemporaneous cave shrines in Kizil and Dunhuang, usually painted across the central ridge of the cave (Ning and Fang 2021, p. 5). At Toyok Cave 20, these meditating monks are also depicted on the upper section of the mural that is close to the ceiling. The lofty location and the scroll-like, rectangular composition of The Story of Sumāgadhā illustrations seem to allude to the “flying and traveling nature” of this story: the disciples of the Buddha managed to traverse incredibly long distances within short periods of time by manifesting the miraculous power gained from meditation. Distinctly, at Toyok Cave 20, the illustration of The Story of Sumāgadhā only depicts these disciples in the moment of meditating and traveling and completely omits other figures, even the protagonists in the story, including the Buddha, Sumāgadhā and his father-in-law. Moreover, the Toyok paintings seem to portray the disciples in a more generic way, without emphasizing the particular attributes of each disciple, as described in The Story of Sumāgadhā. Only the disciples of the lion (Cūḍapanthaka), gold roc (Rāhula), mountain (Maudgalyāyana), elephant (Subhūti) and serpent (Mahākāśyapa) are identifiable, and the remaining can just be described as “flying monks in meditative concentration”.41 In general, the effort to marginalize the narrative side of the story seems to be clear: the key intention is to emphasize the disciples’ state of meditation and the magical traveling prowess achieved through it.

Finally, I argue that this illustration was painted in Cave 20 to emphasize the efficacy of meditation in terms of actualizing cross-distance and trans-realm teleportation, an issue essential for meditators visualizing the Land of Bliss to the far west. The depiction of these monks displays a combination of static and in-motion statuses, as embodied by the disjuncture between the meditation postures of the monks and the moving actions of the mounts. This fusion is realized in quite a stiff and inharmonic way: the monks were not really depicted as riding the mounts but were partially overlapped by them while retaining the meditation postures. The intention to represent the concurrency of meditation and travel seems to be clear. The mind of the meditator can travel a great distance while in meditational concentration. For example, the mind of the meditators on Maitreya could rise up, passing by Mount Sumeru and the first three heavens in our realm to meet the Maitreya Bodhisattva in Tuṣita Heaven (Sponberg 1988, pp. 94–109). Based on existing materials, it seems all contemporaneous caves featuring Sumāgadhāvadāna images have Maitreya and Tuṣita Heaven as their central themes, which might not be a coincidence.42 Sukhāvāti and Tuṣita Heaven were comparable paradises competing to be the most desirable place for rebirth in the sixth century.43 The practice of the “Thirteen Visualizations” is, in essence, a visualization of the key components of the Western Pure Land. As mentioned in the earlier discussion on the “huge gate in III. 4”, the visualizer needs to be there in the Western Pure Land in order to see it. Meditational teleportation seems to be a hidden premise for the visualization practices on Sukhāvāti.

3.2. Contemplations of the Impure

The bottom register is in very poor condition now: only the square units close to the western end are still legible (Figure 18). The part representing the upper torso of the meditating monk on the very left has fallen off the wall, yet it can still be observed that he is meditating on a naked corpse painted in light blue. The hair of the corpse tilts upward to the left in an unnatural way, seemingly suggesting that this corpse is lying on its front on the ground but is rotated vertically to fit in the square unit. In the other legible unit on the right, the meditating monk appears to be contemplating a naked female corpse painted in light brown, raising her right arm to touch her hair, which is partially blocked by her own back and shoulder. It seems that this corpse is lying on the ground, with her breasts facing upward. A small crow is pecking her flesh near her right elbow.

In this section, I argue that the contemplations of the impure assisted by the images on the right wall were performed as preliminary practices for the Pure Land visualizations in the cave, in terms of improving the karmas of the meditators as well as elevating the purity of their spiritual quality to the level of a newborn in Sukhāvāti.

These are the only two images preserved in the bottom register of the right wall painting. I consider them evidence to argue that the bottom register depicts images facilitating the contemplation of the impure: a meditative practice centered on corpses and skeletons.44 The contemplation of the impure is an ancient type of meditation practice that originated in Theravāda Buddhism.45 Its key goal is to trigger the awareness that the attachment to oneself has to be forsaken. It reaches this goal by letting the meditators see the process of corpses decaying into bones, which thereby allows them to realize the impurity and impermanence of their own bodies. Furthermore, it teaches the practitioners to mentally peel off their skin, flesh, bones and organs, like peeling an onion, and thereby awaken to the truth that the ego, or self, does not exist at all, as it cannot be equated to any parts of the body being stripped. Later, it was incorporated into various meditation scriptures, including many Mahāyāna sutras. Scholars tend to associate the contemplation-of-the-impure images depicted in Toyok with three meditation sutras that are believed to have been compiled in the Turfan area: Secret Essentials, The Abridged Essence (Siwei lueyao fa) and the Ocean Sutra (see Yamabe 2002, pp. 123–43; also see Miyaji 2000, p. 281, and Greene 2021, pp. 75–94), which all feature instructions on visualizing corrupting bodies. To give an example, the Ocean Sutra provides instruction on how to meditate on the “Nine Forms of Corpse”, namely, fresh corpses, corpses with blue bruises, corpses with purulent blood, corpses with flowing crimson liquid, corpses being eaten by animals, corpses with only tendons connected to the bone, corpses with only detached bones, empyreumatic corpses and eroded bones turning into soil.46 The corpse painted on the western end could possibly be identified as a “corpse with blue bruise”, as it was intentionally painted in light blue. The other corpse on its right seems to represent a “corpse eaten by animals”, as we see a crow pecking the flesh from its right arm. It is hard to know whether the right wall once featured all nine forms of corpses taught in the Ocean Sutra, as most of the images on the bottom register no longer exist.

I would like to argue that there should be at least one image depicting a pure skeleton, once painted on the right wall, as this image plays a particular transition role between the contemplation of the impure and the Pure Land visualizations. Secret Essentials first instructs on ten approaches to visualizing corrupted corpses as part of the contemplation of the impure. The white bone is a symbol of purity within the contemplation of the impure. In the first section, the sutra instructs the practitioner to contemplate their organs and then imagine vomiting these organs out of their throat onto the ground. As a result, the practitioner will see feces, urine and bodily worms crawling all over each other on the ground.47 This seems to be a mental process of “purification”: after peeling away the organs and related filths, the practitioner will see their own pure skeleton. Based on the karmic merit of the practitioner, he or she will see a white skeleton, a yellow skeleton or even a black one. People who can see their own skeletons in the color of white can progress to the next step of contemplation, whereas those who see a yellow or black one need to improve their karma by performing additional repentance. This process elevates the level of purity of the meditators, which enables them to be transformed into beings reborn in the Western Pure Land, actualized in the final visualization of the Pure Land represented on the left wall: One’s Own Rebirth into Sukhāvāti. In the tenth visualization in Secret Essentials, the meditator is taught to imagine many skeletons being disassembled in front of him (Taishō, v. 15, No. 613, p. 247). Then, in the subsequent eleventh visualization, the sutra teaches the meditator to see white lights radiating from the joints between the bones. The scripture states, “When he attains this contemplation, he will suddenly see a sixteen-foot buddha within this sun light. Its halo measures eight feet horizontally and eight feet vertically. Its body is golden, radiating white light all over, and emitting heat [approaching the meditator]”.48

I think that the Buddha in the sunlight, though not specified in the sutra, could be Amitābha, the Buddha of Enormous Light, due to its close connection with the sun and light. As discussed earlier, the initiating visualization taught in the Visualization Sūtra is to imagine oneself seeing the falling sun in the west. Also, there is a group of scholars who believe that the Amitābha Buddha was modeled after the Celestial Mithra, the Iranian god of the rising sun (Fujiwara 1974, pp. 1–9). Stronger evidence comes from the method to meditate on Amitāyus Buddha taught in the Abridged Essence, which instructs the practitioners with dull faculties and inadequate karma to first meditate on the red bone of him or herself, then mentally turn the red bone into white bone emitting white lights, to eventually see Amitāyus Buddha within the lights (Taishō, v. 15, No. 617, p. 299). It seems the process is not too different from the white bone visualization in Secret Essentials, but the name of Amitāyus is clearly mentioned. Cave 42 at Toyok, opened at about the same time as Cave 20, is normally considered a sister cave of the latter, as it features a highly similar pictorial program. In Cave 42, there is a wall painting featuring a seated monk contemplating a half-skeleton, half-fleshed corpse (Figure 19), among other images of corpses and skeletons, which is usually considered as a representation of the “white-bone visualization”, the transitional meditation between the contemplation of the impure and the Pure Land visualizations (see Lai 2002, p. 125).

Michel Mohr summarizes the contemplation of the impure as a three-stage process: seeing the impurity, the morphing of the impurity into purity and, eventually, the manifestation of purity as a Pure Land. More specifically, he states that the meditation begins by visualizing impure or unpleasant scenes, including one’s own body and decaying corpses, and ends with a perception of light and purity. The final phase coincides with a transformation from the perception of light and purity to visions evoking a Pure Land. He also argues that the Pure Land evoked here is likely to be Amitābha’s Western Pure Land, as the earliest renowned follower of Sukhāvāti, Lushan Huiyuan, also practiced the contemplation of the impure and even wrote the preface for scripture on it: Meditation Sutra of Dharmatrāta (Damoduoluo chanjing) (Mohr 2020, pp. 17–19).

Lai Wen-ying notes that the appearance of images representing both the contemplation of the impure and the Pure Land visualizations within a single cave was a unique invention at Toyok, which embodied the local trend to connect Theravāda meditation practices with Mahāyāna ones via the bridge of four immeasurable states of mind (si wuliang xin) (Lai 2002, pp. 114–35). In addition, the scriptural foundations for these two types of meditation practices have a very intimate, kindred relationship. Stylistic and linguistic evidence suggests a clear link between Secret Essentials, Methods for Curing and the body of apocryphal fifth-century Chinese Buddhist texts known to modern scholars as the Contemplation Scriptures, the most prominent member of which is the Visualization Sūtra (Greene 2021, p. 75). Furthermore, many scholars believe that these scriptures were compiled in the Turfan area on the Silk Routes, a place where many Indian and Central Asian Buddhist teachers associated with these texts are said to have dwelled for a time (see Greene 2021, pp. 100–2 and Fujita 1990, pp. 156–59).

4. The Rear (West) Wall: A Threshold

In this section, I argue that the rear wall featuring representations of various water ponds with rebirth figures functions as an entryway to the Western Pure Land. As discussed earlier, the last visualization performed by meditators while facing the left wall is the twelfth visualization, during which the meditator imagines himself reborn in Sukhāvāti. The paintings in III. 4 and III. 5 jointly depict the three steps of this visualization: pushing the doors to enter Sukhāvāti, seeing oneself seated in the enclosed lotus bud and watching the lotus blossoming. I think the painting on the rear wall (west wall) of Cave 20 is the final episode of this process.

Most of the paintings on the rear wall have been peeled off. On the bottom right corner of the rear wall mural, a haloed monk was depicted seated in meditation, with flames and water around him, indicating a state of meditational concentration. Above the monk, there is a big tree with a serpent-shaped water column flowing among the flowers on it, a reminder of the image painted in III. 1 on the left wall, representing the eight Jeweled Ponds in Sukhāvāti. Further above, there is an apsara flying downward with a tray in his left hand and a stick in his right hand. The only images remaining on the left part of the rear wall are a few square units with a round, tree-like motif in the center, resembling the images in III. 1.

4.1. The Water Ponds as Entryway

In this part, I argue that the various water ponds depicted on the rear wall surrounding the entrance to the back chamber function as the entryway to Sukhāvāti for the meditators. Yamabe identified a fragment of the mural painting, now in the State Hermitage Museum in Saint Petersburg, Russia (Figure 20), as a piece peeled off from the left part of this wall, mainly based on the identical square units shared by both as well as relevant visual descriptions given by Grünwedel who visited this cave in 1902 (Yamabe 1999, p. 240). Based on this detailed image of the two square units (Figure 21), it is observable that each unit is defined by a square frame painted in green, surrounded by blue, swirled lines resembling water waves. Combined with the tree motif with serpent-like water on the right side of the rear wall, it seems clear that these square units represent the various jeweled ponds in Sukhāvāti.

Different from the case in III. 1, within each jeweled pond on the rear wall mural, there are two naked figures seated in the crown of the round tree. These can be interpreted as the many heavenly boys who reside in the miraculous palaces amid the nets on the Jeweled Trees, as described in the Visualization Sūtra (Taishō, v. 18, No. 365, p. 345). Noteworthily, the figures were depicted as joining their palms in the añjali mudra. The lower body of each figure was replaced by a small, open lotus flower. This was a typical way to represent the figures reborn in Sukhāvāti on lotuses in various sites along the Silk Routes during the fifth and sixth centuries. It seems that although the relevant description in the “Jeweled Trees” might have served as the scriptural ground for the imagery, the artist’s intention to evoke the theme of rebirth in Sukhāvāti is evident.

There are also rebirth figures in much smaller sizes, seated in semi-open lotuses growing from the watery space amid the square units. There is no direct representation of “Nine Grades of Rebirth” in any pre-Tang Buddhist art in Xinjiang. Fujita Kōtatsu found convincing evidence showing that there are strong Chinese elements in the terminologies, literary style and narration of the Visualization Sūtra, which supports the Chinese compilation theory, though he also acknowledged that the core of the sutra transmitted a form of meditation that was then practiced somewhere in Central Asia, likely in Turfan (Fujita 1990, pp. 149–73). Yamada Meiji also noted that in the “Nine Grades of Rebirth” section, the name used to refer to the main Buddha in Sukhāvāti is Amituo (Amitābha), different from Wuliangshou (Amitāyus) used in the “Thirteen Visualizations” part (Yamada 1984, p. xxi). Going one step further, Julian Pas considers the particular part of “Nine Grades of Rebirth” as a later Chinese interpolation into the Visualization Sutra (Pas 1995, pp. 15–20). On one side, I think the absence of images depicting Nine Grades of Rebirth in Cave 20 seems to support his argument. On the other side, the distinctions among the different circumstances of rebirth in Sukhāvāti might be already present in the Turfan area, as shown by the rebirth figures of varying sizes seated in the round trees and the small lotuses.

Because beings are reborn in Sukhāvāti on lotuses, the Jeweled Ponds where the myriad lotus flowers grow could be understood as gateways to the Western Pure Land. This idea is inspired by the research of Anne Feng, who convincingly argues that Cave 171 at the Dunhuang Grottoes was designed as a lotus pond, a site of transformation and an entry point to the Pure Land (Feng 2018, pp. 195–234). The idea of the “lotus pond as entrance to Sukhāvāti” is supported by the context of Cave 20. First, the rear wall is the west wall. While facing it, the meditator faces the direction of the Western Pure Land. This might not be a coincidence. A similar painting on the rear wall of Cave 42 (Figure 22) gives scholars some sense of the ways in which the various square units were arranged on the rear wall of the cave. The general look of each square unit can be better observed in a mural at Cave One (Figure 23). Such murals are often considered to represent the “Seven Rows of Tree (qichong hangshu)” described in the “Jeweled Trees” part, as each side of the mosaic square comprises seven units (for an example, see Miyaji 1995a, pp. 29–31). However, as previously analyzed, these trees with roundish crowns were created to serve as the background for the Jeweled Ponds. Based on similar murals, Yamabe digitally reconstructed the rear wall, restoring the mosaic-like squares. It is observable that the various jeweled ponds were painted on a wall featuring an entrance to the back chamber. At Toyok, such mosaic-like water pond images were always painted on the rear wall above the entrance to the rear chamber.49 The entrance on the wall surrounded by various water ponds seems to further echo the gateway nature of the latter. Few scholars challenge the connection between the water ponds imagery and Sukhāvāti, yet such pictorial mode of representing Sukhāvāti seems to be an invention at Toyok as it is seen nowhere else, like the case of serpentine-shaped water column in III. 1. Rather than representing the environment and landscape of Sukhāvāti, it seems the designers at Toyok focused on the action and result of entering the Western Pure Land, the final step of meditation performed in this cave.

4.2. The Rear Wall Meditating Figures

There are two huge, seated figures depicted on the two bottom corners of the rear wall mural (see Figure 3). In contrast to the meditating monks painted on the bottom right, a strange figure wearing some kind of hood, with exaggeratedly elongated earlobes, is depicted seated on a Central Asian-style chair in the bottom-left corner (see Figure 19). Grünwedel described the figure as “meditating on a chair” (Grünwedel 2007, p. 588). Among the existing images painted on the left wall, the meditator of the twelfth visualization (III. 4) is unique. Though not fully legible, it seems clear that the meditator depicted in III. 4 is not a monk. Some scholars consider it a representation of Queen Videhī, to whom the Buddha preached the Visualization Sūtra (See Miyaji 1996, pp. 66–67; also see Yamabe 1999, p. 40). This is a reasonable inference, but no representation of Queen Videhī has ever appeared in any Buddhist art pre-dating the Tang dynasty. I think the hood worn by the meditating figure in III. 4 is quite similar to the one on the rear wall (on the State Hermitage Museum fragment). Coincidentally, like the meditator in III. 4, the rear wall figure turns his head toward his left, facing a gateway on the rear wall, embodying the entrance to Sukhāvāti.50 A being reborn in Sukhāvāti will become a non-regressive bodhisattva. It could be possible that this image represents such a bodhisattva in Sukhāvāti. Moreover, slightly above the meditating figure, there are two flowers suspended in the air and a bird with a human head. The bird should be Kalaviṅka, the bird of splendid sound and music residing in Sukhāvāti, as described in the Shorter Sukhāvatīvyūha Sūtra (Taishō, v. 12, No. 366, p. 346). Further above the Kalaviṅka bird, there is a suspended handbell. Together, they represent the raining flowers and splendid music in Sukhāvāti, as described in all three Pure Land sutras. It seems clear that the atmosphere in Sukhāvāti was evoked around this meditating figure. The twelfth visualization is special in the sense that it asks the visualization performer to imagine himself turning into a figure reborn in Sukhāvāti. The change in the depiction of the meditator seems to mark the realization of the transcendence of life level from human to bodhisattva, whereas the meditation posture reminds the means to achieve such transcendence.

5. Cave 20 as a Meditation Space

5.1. The Order of Meditations

Based on the architectural structure and pictorial program of Caves 40–42, Xia Lidong reconstructs a four-step ritual and meditational practice performed in these caves comprising meditating on the auspicious symbols (laksana) of the Buddha, contemplating the impure, visualizing the Pure Land and receiving confirmative visions (Xia 2022, pp. 30–34)51. I think Cave 20 conveys a similar message. The layout and pictorial program of Cave 20 suggest a clock-wised, circular order of meditation performed in the cave, starting from the meditating monk on the bottom-right of the rear wall mural and culminating at the bodhisattva figure on the opposite side.

The pictorial program of Cave 20 displays a strong circular feature. The structure of this cave is not symmetrical at all. The entrance to the cave is located almost at the left end of the front wall (see Figure 24 and Figure 25). A clay, wind-blocking wall was built in the front of the entrance, whose width is about two-fifths of the whole chamber. Correspondingly, a huge window was opened on the right part of the front wall. As a result, after entering the cave, people would naturally make a right turn to avoid the barrier and gather at the right half of the chamber, which has a better illuminating condition within a dark room. If that is the case, the first image they will see is the huge meditating monk on the bottom right corner of the rear wall mural (see Figure 3).

The pictorial program of the cave is not symmetrical either. The head of the meditating monk on the rear wall has been ruined. However, it seems that all the meditating monks depicted in the cave are turning their heads to their left sides. Based on this observation, I argue that the meditating monk under the huge tree should be facing toward the right wall. A similar mural on the rear wall of Cave 42 seems to further support that idea (Figure 22). That is to say, the two main figures on the rear wall mural are not forming a symmetric pair flanking the center, but both turning their heads to their left, which would visually guide the viewers to turn their sight to the right wall.

All monks performing the contemplation of the impure painting on the right wall are facing the east, the direction of the entrance, whereas those performing the Pure Land visualizations depicted on the left wall are facing the west, the direction of the rear wall. In addition, the apsaras depicted on the top of the murals on all three walls are flying in a direction consistent with the orientation of the meditating monks, with the only exception of the one above the bodhisattva figure on the rear wall. This seems to further indicate the order of the meditation practices performed in the cave. More specifically, the meditators started with the contemplation of the impure on the right wall and gradually turned around to the left wall via the entrance, which seems to further support our previous analysis on the purpose of the contemplation of the impure. The apsaras on the bottom-left of the rear wall mural appear to be interacting with the bodhisattva in a face-to-face manner. The opposition between their body orientations and movement tendencies makes this scene a stabilized, static image rather than an in-motion one. This tension functions as a tightened knot marking the bodhisattva figure as the ending point of the meditation practices aided by the visual clues painted in the cave.

To reiterate, the strong visual guidance of movement direction in this cave marks the meditating monk on the rear wall as the starting point of the meditation, whereas the bodhisattva in the opposite corner marks the ending point. Thus, the meditator starts as a meditating monk, embodied in the bottom-right corner, before turning to the right wall to practice the contemplation of the impure. After concluding the contemplation of the impure, he prepares himself with adequate karmic merit and wisdom to turn to the left wall via the front wall and practice the Pure Land visualizations. Eventually, he ends as a bodhisattva successfully reborn in Sukhāvāti, embodied in the bottom-left corner of the rear wall mural, as promised in the Visualization Sūtra.

5.2. The Issue of Portable Icons

In addition, I suspect that portable images of the Amitāyus triad were once placed inside Cave 20 when meditations were performed, most likely placed in the center of the front chamber. Though no immovable, monumental icons were found in pre-Tang caves at Toyok, unlike those in Kizil and Dunhuang Grottoes, I believe portable images must have been once used in the rituals and meditations performed in these caves. Due to their high cultural significance as well as monetary value, those portable icons were unlikely to remain in situ after hundreds of years. However, there are still some clues showing they once existed. Grünwedel and Stein had each found some small wooden sculptures of bodhisattvas in the west precinct of the Toyok site during the early twentieth century (see Von Le Coq 1926, p. 8; Stein 1928, vol. 3, pp. 131–36). In the archaeological report of the 2010s excavation, fragments of small wooden and clay icons were discovered (Zhongguo Shehui Kexueyuan Kaogu Yanjiusuo Bianjiang Minzu Kaogu Yanjiushi and Tulufan Yanjiuyuan 2020, pp. 448–53). In Cave 33, a cave contemporaneous to Cave 20, a wooden head of bodhisattva was found. The head is only 15 cm in height, suggesting that the size of the image should be modest, probably 50–80 cm in total. In other words, the icons used at Toyok Caves were likely to be small portable images, which were easy to be carried back and forth by monks.

The inference about the existence of the icons of Amitāyus Triad is not ungrounded. First, based on the mural paintings on the rear wall and left wall, the core theme of this cave appears to be the entry into Sukhāvāti. As discussed previously, the visualizations missing on the left wall are all related to the Amitāyus Triad. The eighth visualization teaches the meditators to visualize the statues of the Amitāyus Triad. The ninth, tenth and eleventh visualizations instruct the meditator to visualize Amitāyus, Avalokiteśvara and Mahāsthāmaprāpta, respectively. Finally, the thirteenth visualization asks the meditator to imagine these three spiritually advanced beings manifested in their various forms in the Western Pure Land. Particularly, the thirteenth visualization instructs the visualizer to first imagine an image of Amitāyus in the size of zhangliu suspended above a water surface.52

The main legitimacy behind this assumption is that, in most of the illustrations of the Visualization Sūtra produced before the eighth century, the visualizations related to the Amitāyus/Amitābha Triad were usually not represented. Instead, icons of the triad are present, either in the form of main sculptures or as central figures dwelling in the paradisical setting of Sukhāvāti represented on the wall painting. As shown by the only other sixth-century illustration of the Visualization Sutra, a stone relief on the top of the west wall of the Xiaonanhai Central Cave (Figure 26), the five visualizations related to the Amitāyus Triad were aided by the stone sculptures below the relief featuring representations of the Nine Grades of Rebirth identified by inscriptions amid randomly arranged motifs representing the remaining parts of Thirteen Visualizations, probably due to ontological concerns (see Li 1999, pp. 168–81).

Finally, the circular structure of the pictorial program of Cave 20 strongly alludes that there should be a visual focus at the center of the cave. Cave 26 at Ajanta and the Kanjingsi Cave at Longmen Grottoes both feature such circular pictorial programs. Noteworthily, there is a main icon in the center of each cave. In addition, the murals in Cave 20 were painted in the upper part of the walls, about 1.8 m from the floor level. When the monks sit down to start meditation, there is no image at their eye level. In that case, I propose that portable wooden images of Amitāyus Triad might be once set up in the center of the chamber, probably slightly leaning backward to the rear wall, to become the main images facilitating the meditations performed in this Pure Land cave.

6. Methodological Note: On the Function of Murals in Cave 20

The nature and function of Buddhist images in the cave shrines of medieval China and Central Asia have triggered some methodological debates among art historians and scholars of Buddhism (see Sharf 2013, pp. 38–64, and Howard 2015, pp. 20–39). Whether these images were meant to be used in Buddhist rituals and practices lies at the center of discussion. This is such a big issue, and I do not believe that the existence of a universal answer is applicable to the thousands of caves spanning such a wide geographical range and time scope. Nevertheless, I would like to argue that the various elements constituting Cave 20 at Toyok, especially the architectural structure, pictorial program, inscriptions and recovered objects, seem to suggest the function of the murals as meditation aids.

Robert Sharf argues that many caves at Mogao Grottoes were intended as merit projects or mortuary shrines, which were barely visited after their completion (Sharf 2013, pp. 38–64). In Cave 20, there is a wind-blocking wall in the front of the entrance, standard equipment for meditation caves, suggesting that people were expected to stay in the cave for a long duration. The floor behind the wind-blocking wall was artificially raised 13 cm and flattened, seemingly to be purposefully prepared as a meditation zone. The key point made by Sharf, the limited visibility of images due to the darkness in the cave, does not exist in Cave 20 as there are two huge lighting windows on the front and rear wall of the front chamber and a diagonal lighting hole on the ceiling of the back chamber (Zhongguo Shehui Kexueyuan Kaogu Yanjiusuo Bianjiang Minzu Kaogu Yanjiushi and Tulufan Yanjiuyuan 2020, pp. 440–42) There are grooves on the entrance to the back chamber and the lighting windows, showing traces of wooden doors/mullions once installed there. It seems unnecessary to spend resources and efforts on these environment-control facilities if the cave was not intended to be regularly visited after completion.

All figures depicted in the murals are meditating monks.53 Even Sharf himself acknowledges that these may be depictions of practices that took place in the cave (Sharf 2013, p. 51). Below the image of the monk meditating on a half-skeleton body in Cave 42 (Figure 19), there is an inscription that reads “Zhikong, a monk from the Kaijue Temple” (Xia 2020, p. 19), suggesting the existence of monks active in the cave. In the back chamber, two narrow bars of clay platforms were built at the bottom of the side walls. A few pits were opened on the platforms to accommodate utensils. A clay jar with a sealing flake stone was discovered at such a pit. The back chamber is lower in floor level with no murals and more irregular surfaces, appearing to be the living quarters of the meditating monks. The clay jar is probably a water container used by monks for body cleaning and purifying rituals: such vessels were frequently depicted below the meditating couch of the monks in the murals (e.g., see Figure 18). Again, it seems these murals represent the meditators in the cave and aid the meditations performed herein.

Finally, as previously analyzed, the composition of images and curation of inscriptions in the cave were based on the actual needs of meditation practices performed in the cave rather than literal illustration and transcription of sutra texts. More complicated visualizations entailing multiple steps of vision construction and transformation were given more images and compiled inscriptions hinting at the key thresholds of the mental processes. That might also be why the square units aiding the Thirteen Visualizations were not arranged in an ordered, boustrophedonic way, like an illustrated comic book. The front chamber is about three meters in width and four meters in depth, a space large enough to accommodate multiple monks to meditate at the same time. Based on the mural images, the monks were likely to meditate on couches in a static manner. A linear or boustrophedonic arrangement would not make the images assisting different visualizations equally accessible for monks randomly seated within the cave. In addition, it would be quite easy for the adepts to locate images relevant to appropriate visualization while facing the mural, especially considering that they must have looked around the cave before starting to meditate. In the only contemporaneous image of the Visualization Sūtra in Xiaonanhai Central Cave (Figure 26), which is also considered a meditation cave, the images representing different visualizations were randomly placed amid the nine stems of lotuses depicting the Nine Grades of Rebirth (see Li 1999; Yan 1998).

7. Conclusions

The mural painting on the left wall of Cave 20 at the Toyok Grottoes in Turfan, created in the late sixth century, is one of the two earliest illustrations of Sukhāvatī made based on the apocryphal Visualization Sūtra. I argue that this wall painting was designed to assist the meditating monks in situ with visual and verbal cues to separately visualize the different individual visions constituting the “Thirteen Visualizations” taught in the sutra, excluding those five visualizations directly related to the Amitāyus Triad. These five visualizations were likely to be facilitated by portable icons of the Amitāyus Triad that were likely once placed in the center of the front chamber. The meditators started from the contemplations of the impure represented on the right wall as preparational practices to eliminate their sins, improve their karma and thereby enhance their spiritual purity to a level appropriate for performing the Pure Land visualizations. Then, they turned to the left wall to perform the Pure Land visualizations represented there, which ended with the twelfth visualization: imagining oneself being reborn in Sukhāvatī. Continuing from the twelfth visualization, the meditators, in a state of meditational concentration, turned their faces to the rear wall to enter the Western Pure Land via the various lotus ponds represented on the rear wall. In general, Cave 20 was constructed as a Pure Land cave for believers of Sukhāvatī to perform a set of meditation practices that grant admission to this paradise.

The Toyok mural and the relief at Xiaonanhai Central Cave are the only existing Sukhāvatī illustrations created based on the Visualization Sūtra, dated before the Tang dynasty. Their distinct look from later Sukhāvatī illustrations in Mogao Grottoes at Dunhuang might shed light on the practices adopted by sixth-century believers to achieve rebirth in Sukhāvatī. Sixth-century illustrations display no intention to depict an environmental setting of the Western Paradise. On the contrary, it focuses on depicting the different objects of visualization taught in the sutra, one by one. According to the sutra, the practitioners visualize the “Thirteen Visualizations” separately rather than piling up these individual visions together to create a holistic, eidetic image of the Western Pure Land. The “Thirteen Visualizations” were like the laksanas of the Western Pure Land: the meditators gained different merits, bliss and even the direct promise of entering the Western Pure Land after completing the different individual visualizations. As summarized by Cuong T. Mai, a rebirth in Sukhāvatī is already assured after performing any one of the visualizations on Ground (V3), Jeweled Tower (V6), Lotus Seat (V7), One’s Own Rebirth in Sukhāvatī (V12), or Amitāyus Triad in Sukhāvatī (V13) (Mai 2009, p. 349). I suspect that the “Thirteen Visualizations” might not have to be performed in a successive and comprehensive manner, which might also explain why images representing different visualizations both in Toyok mural and in Xiaonanhai relief were organized in a random order.