The Background of Stone Pagoda Construction in Ancient Japan

: In this study stone pagodas from ancient Japan (7th to 9th centuries) were analyzed. The ﬁndings show that there are some apparently inﬂuenced by the Korean Peninsula and two other types. While there are examples of the former type that are large and serve as temple buildings, the latter are located in mountain forest temples. I am of the opinion that stone pagodas were important mechanisms that made possible the existence of mountain forest temples as Mahayana precepts-based transgression repentance ( keka 悔 過 ) training sites that complemented ﬂatland temples. This use of stone pagodas is different than China and Korea, which treated both wooden and stone pagodas in the same way. Moreover, ideas regarding Mahayana precepts-based transgression repentance were introduced from China, and I hold that the increase in stone pagodas at mountain forest temples corresponds to the Sinicization of Japanese Buddhism.


Introduction
While there are many examples of stonework in Japan, the most familiar are pagodas and Buddhist figures. Stonework pieces from the latter half of the 13th century and later are particularly prevalent throughout Japan, and have drawn the attention of many researchers and lay historians. However, there are few stone pagodas from the Heian period  and earlier, in terms of both type and absolute number. Therefore, the background to their construction and historical characteristics were unclear, and research that brought together examples of these stone pagodas was necessary. While prominent Japanese stonework scholar Kawakatsu Masatarō published his famous work Nihon sekizai kōgeishi 日本石材 工芸史 (The History of Japanese Stone Craftwork), covering the background of 7th and 8th century stonework creation and providing an overview of existing materials from the time, he also noted that there was insufficient knowledge of their details due to lack of materials (Kawakatsu 1957). In recent years, Sagawa Shin'ichi has pointed out that many stone pagodas predating the 9th century exist in places that were primarily used for mountain forest religious training, offering a new perspective for the field (Sagawa 2021). However, there is still a gap in the literature concerning the background of stone pagoda construction during that time. While scholars have discussed ancient stone pagodas' genealogies and dates with a focus on their morphological similarities to stone pagodas in China and Korea, Japan's stone pagodas are different from those pagodas in size, basic form, and construction techniques, making such discussions not very meaningful. Therefore, it is necessary to discuss the places where they were located, as Sagawa does. In this paper, I will focus on the nature of such places and stone pagodas' functions from this perspective.

Ishidōji's Three-Story Pagoda
The three-story pagoda at Ishidōji 石塔寺 temple in the city of HigashiŌmi in Shiga Prefecture is one of Japan's ancient-style stone pagodas. It features a thin, tall body

Okamasu Ishindō
Further evidence that Japanese pagodas from the 7th century were built by settlers with roots in the Korean Peninsula can be found in the Okamasu Ishindō 岡益石堂. The Okamasu Ishindō is located in Kokufuchō, Tottori City, and was also strongly influenced by Korean stone pagodas. It is made entirely of green tuff breccia. The center of its cut stone danjōzumi kidan 壇上積基壇 style podium (6 m per side) has somewhat of a bulge ( Figure 3). It contains a circular pillar with a central stand, the bottom of which features an ornamental double lotus petal motif (some believe this is actually a palmette). The structure is surrounded by a stone wall measuring 4.3 m in width and 2.1 m in height (Kasano et al. 1999) and which is engraved with the seal of the Ikeda family, who were the Tottori domain lords during the Edo period . It therefore appears that the Ikeda family repaired the wall during that time. However, as Mizuno Masayoshi points out, these repairs were limited to the stone walls (Mizuno 2006).  The aforementioned Nihon shoki entry from the eighth year of Emperor Tenchi's reign (669) states, "About 700 men and women move to the Ōmi Gamō District" (男女七百餘人 遷居近江國蒲生郡). This textual passage indicates that a group of people from the Korean Peninsula settled around Ishidōji, where researchers have unearthed a "Kotō-style" roof tile with the same pattern as one from the abandoned temple Kabataji 綺 田 寺 (Gamōmachi Kokusai Shinzen Kyōkai 2000). Ogawasara Yoshihiko and Hishida Tetsuo argue that this roof tile reflects a new style introduced by Baekje immigrants (Ogasawara   The aforementioned Nihon shoki entry from the eighth year of Emperor Tenchi's reign (669) states, "About 700 men and women move to the Ōmi Gamō District" (男女七百餘人 遷居近江國蒲生郡). This textual passage indicates that a group of people from the Korean Peninsula settled around Ishidōji, where researchers have unearthed a "Kotō-style" roof tile with the same pattern as one from the abandoned temple Kabataji 綺 田 寺 (Gamōmachi Kokusai Shinzen Kyōkai 2000). Ogawasara Yoshihiko and Hishida Tetsuo argue that this roof tile reflects a new style introduced by Baekje immigrants (Ogasawara ( Figure 3). It contains a circular pillar with a central stand, the bottom of which features an ornamental double lotus petal motif (some believe this is actually a palmette). The structure is surrounded by a stone wall measuring 4.3 m in width and 2.1 m in height (Kasano et al. 1999) and which is engraved with the seal of the Ikeda family, who were the Tottori domain lords during the Edo period . It therefore appears that the Ikeda family repaired the wall during that time. However, as Mizuno Masayoshi points out, these repairs were limited to the stone walls (Mizuno 2006).  (Kasano et al. 1999), with modifications by the author).
The top of the central circular pillar is a piece of a multistory pagoda. The same pieces are also found at the neighboring temples Daiun'in 大雲院 and Okamasuji 岡益寺. Sagawa Shin'ichi posits that these were originally part of a five-story pagoda (Sagawa 2006) that he believes can be dated to the 9th century, based on a comparison with ancient multistory pagodas extant in Japan. Mizuno holds that the multistory pagoda, stone wall, and circular column form a set, and has reconstructed them by placing the multistory pagoda on top of a base with a stone wall. Noting that a tuff fragment, probably material for the pagoda, was unearthed from a ground layer dated to the 7th century, Mizuno also argues that the stone wall and other features were built during that time. He believes that the top part of the pagoda was initially wooden, but was changed to stone during the 9th century.  (Kasano et al. 1999), with modifications by the author).
The top of the central circular pillar is a piece of a multistory pagoda. The same pieces are also found at the neighboring temples Daiun'in 大雲院 and Okamasuji 岡益寺. Sagawa Shin'ichi posits that these were originally part of a five-story pagoda (Sagawa 2006) that he believes can be dated to the 9th century, based on a comparison with ancient multistory pagodas extant in Japan. Mizuno holds that the multistory pagoda, stone wall, and circular column form a set, and has reconstructed them by placing the multistory pagoda on top of a base with a stone wall. Noting that a tuff fragment, probably material for the pagoda, was unearthed from a ground layer dated to the 7th century, Mizuno also argues that the stone wall and other features were built during that time. He believes that the top part of the pagoda was initially wooden, but was changed to stone during the 9th century. Here, I will just mention that currently these two theories of Sagawa and Mizuno exist regarding the time when this pagoda was built.
The Okamasu Ishindō pagoda is unique in Japan, with no similar examples anywhere in the country. Certain characteristics do not have genealogies in Japan and clearly demonstrate the influence of peninsular stonework, including the ornamental double lotus petals on the bottom of the circular pillar's flower-shaped support (ukebana 請花) and the entasis pillar structure. In this regard, we should look to the peninsula for similar examples. However, at present, I will refrain from pursuing its prototypes.
This pagoda was built where the eastern pagoda of the temple's golden hall was located, and we can assume it was created as one of the temple's buildings.
The western part of Tottori Prefecture, where Okamasu Ishindō is located, contains Korean Peninsula-style relics. It is likely that settler clans from the Korean Peninsula were involved in the construction of both Okamasuji and Ishindō.
As we have seen, stone pagodas influenced by the Korean Peninsula tend to be large and form part of a temple complex. This is similar to how large stone pagodas were, like wooden pagodas, built as temple buildings on the Korean Peninsula. However, the only stone pagodas in Japan with such peninsular influences are the two discussed above. Other stone pagodas are clearly different. Below, I will review the characteristics of such stone pagodas from the eighth and ninth centuries.  Table 1 shows a list of multistory pagodas from the 8th to 9th centuries in Japan (Table 1, Figures 4 and 5). In Japan, there are no examples of stone pagodas with inscriptions or unearthed artifacts from the 900s. Stone pagodas began to recede in the 10th century and then were absent up through the 12th century. Included in the table's stone pagodas are ones that, like wooden pagodas, have a deeply contoured shape (5-10, 13-17) and ones with wooden pagoda structures, such as hip rafters (sumigi 隅木) on eave undersides and descending ridges (kudarimune 降棟) on eaves (3,4,5,7,8,10,(13)(14)(15)(16). A ceramic pagoda held by Shōsōin 正倉院 from the 8th century is similar. Also, at the Enichiji 恵 日寺 temple pagoda was a pot apparently from the 9th century. Furthermore, Iwanaga Shōzō points out that ancient-style pagodas' roofs are slanted the same amount and that this slant is different from stone pagodas from the 12th century onwards (Iwanaga 2001). While I ask the readers to refer to my below reference list for information on the dating of individual ones, we can see these pagodas as all being from the 8th or 9th centuries. During this time, the number of stone pagodas in Japan greatly increased. They are notable for being relatively small, around 2 m in height. Regarding location, several were found at temples deep in the forest mountains, includingŌganji 応願寺 temple's multistory pagoda fragment (6), Tōnomori's 塔ノ森 hexagonal 30-story pagoda (7), Murōji 室生寺 temple's wish-granting jewel pagoda (10), Rokutanji temple ruins' 13-story pagoda (11), Iwaya's 岩 屋 three-story pagoda (12), Iwatakiyama's 岩滝山 multistory pagoda (13), and Shusshakuji 出釈迦寺 temple's multistory pagoda (15). On the other hand, Kannonji temple's threestory pagoda (5) and Ryūfukuji 龍福寺 temple's five-story pagoda (9) were found in valleys at the bases of mountains, and can be seen as parts of mountain forest temples. Excluding that of Rokutanji, these stone pagodas are small, around 2 m in height. Further, many of the temple buildings connected to them do not exhibit clear monastery style (garan 伽藍; Skt. *sam . ghârāma) structures.Ōganji's multistory pagoda (6) and Tōnomori's hexagonal 13-story pagoda (7) are found at mountain forest training sites east of Tōdaiji 東大寺 temple and Kōfukuji 興福寺 temple. As Sagawa Shin'ichi points out, they appear to have been created as pagodas for mountain forest temples that formed sets with flatland temples (Sagawa 2021). Moreover, their shapes and locations are clearly different from those of the many wooden pagodas built during this time.  Inscribed on Ryūfukuji's five-story pagoda in Nara Prefecture's village of Asuka, we find the year Tenpyō-shōhō 3 (751) and the name of Princess Takeno (Takeno-ō 竹野王; female royalty connected to Prince Nagaya or Nagaya-ō) (Figures 6 and 7). It states, "South of Asakaze and North at Tōnomine . . . " (朝風南葬談武之峯北). A hill approximately 1 km to the northwest is still referred to as Asakaze. It therefore appears that this stone pagoda was originally located on top of this hill. One of Prince Nagaya's wooden tablets (mokkan 木簡) contains the following: "Prince Takeno hired people for 3.6 L of rice and dispatched them to a mountain temple" (竹野王子山寺遣雇人米二升γγγ; Nagaya-ō Mokkan No.1829). Apparently, Prince Takeno built this stone pagoda for his own memorialization and to, as parishioner of this mountain temple, support the overall development of its sangha. While the primary aim of this pagoda's construction was the repose of Prince Takeno after death, secondarily it functioned as the sangha's core, symbolizing Ryūfukuji's status as a mountain temple. The next section expands on this issue with a discussion of Rokutanji (Taishi Town, Osaka Prefecture).   Inscribed on Ryūfukuji's five-story pagoda in Nara Prefecture's village of Asuka, we find the year Tenpyō-shōhō 3 (751) and the name of Princess Takeno (Takeno-ō 竹野王; female royalty connected to Prince Nagaya or Nagaya-ō) (Figures 6 and 7). It states, "South of Asakaze and North at Tōnomine… "(朝風南葬談武之峯北). A hill approximately 1 km to the northwest is still referred to as Asakaze. It therefore appears that this stone pagoda was originally located on top of this hill. One of Prince Nagaya's wooden tablets (mokkan 木簡) contains the following: "Prince Takeno hired people for 3.6 L of rice and dispatched them to a mountain temple" (竹野王子山寺遣雇人米二升 γγγ; Nagaya-ō Prince Takeno after death, secondarily it functioned as the sangha's core, symbolizing Ryūfukuji's status as a mountain temple. The next section expands on this issue with a discussion of Rokutanji (Taishi Town, Osaka Prefecture).

The 13-Story Pagoda at the Rokutanji Ruins and the Background of Its Creation
As discussed in the previous section, clear patterns and tendencies in construction and usage emerge by inspecting numerous 8th century stone pagodas in Japan. Close consideration of one specific 8th century pagoda, located at the Rokutanji ruins, brings even further to light important aspects of the 8th century pagodas and their background influences.
The Rokutanji ruins are located at an altitude of about 280 m to the west of Mt. Nijō, which separates Yamato and Kawachi (Figure 8). Its ground consists of a 400 m 2 flat area created by quarrying. Compared to other mountain forest temples from its time, Rokutanji  Prince Takeno after death, secondarily it functioned as the sangha's core, symbolizing Ryūfukuji's status as a mountain temple. The next section expands on this issue with a discussion of Rokutanji (Taishi Town, Osaka Prefecture).

The 13-Story Pagoda at the Rokutanji Ruins and the Background of Its Creation
As discussed in the previous section, clear patterns and tendencies in construction and usage emerge by inspecting numerous 8th century stone pagodas in Japan. Close consideration of one specific 8th century pagoda, located at the Rokutanji ruins, brings even further to light important aspects of the 8th century pagodas and their background influences.
The Rokutanji ruins are located at an altitude of about 280 m to the west of Mt. Nijō, which separates Yamato and Kawachi (Figure 8). Its ground consists of a 400 m 2 flat area created by quarrying. Compared to other mountain forest temples from its time, Rokutanji

The 13-Story Pagoda at the Rokutanji Ruins and the Background of Its Creation
As discussed in the previous section, clear patterns and tendencies in construction and usage emerge by inspecting numerous 8th century stone pagodas in Japan. Close consideration of one specific 8th century pagoda, located at the Rokutanji ruins, brings even further to light important aspects of the 8th century pagodas and their background influences.
The Rokutanji ruins are located at an altitude of about 280 m to the west of Mt. Nijō, which separates Yamato and Kawachi (Figure 8). Its ground consists of a 400 m 2 flat area created by quarrying. Compared to other mountain forest temples from its time, Rokutanji is unusual due to its stone 13-story pagoda, which is located in front of a cave containing three engraved nyorai 如来 (Skt. tathāgata). This stone pagoda is 5.1 m in height, with a base width of 1.5 m (Figure 9). It was created by shaping a tuff outcrop into a stone pillar, then adding features (Yamamoto 1993). There is a hole at the top of the base. The relative positioning of the stone pagoda and stone cave indicate that it was probably a single pagoda/golden hall (金堂) temple with nyorai as its main object of veneration (Taketani 1989). The closely connected eaves, wide body, and the shape of the hole of this 13-story pagoda are not found elsewhere in Japan. The stone pagoda was certainly built in the mid-eighth century; it is made from an outcrop carved into a stone pillar, and only relics from the mid-eighth century exist around it.
Religions 2021, 12, x FOR PEER REVIEW 10 of 17 is unusual due to its stone 13-story pagoda, which is located in front of a cave containing three engraved nyorai 如来 (Skt. tathāgata). This stone pagoda is 5.1 m in height, with a base width of 1.5 m (Figure 9). It was created by shaping a tuff outcrop into a stone pillar, then adding features (Yamamoto 1993). There is a hole at the top of the base. The relative positioning of the stone pagoda and stone cave indicate that it was probably a single pagoda/golden hall (金堂) temple with nyorai as its main object of veneration (Taketani 1989). The closely connected eaves, wide body, and the shape of the hole of this 13-story pagoda are not found elsewhere in Japan. The stone pagoda was certainly built in the mid-eighth century; it is made from an outcrop carved into a stone pillar, and only relics from the mid-eighth century exist around it.  Rokutanji's main object of veneration was likely the nyorai triad that is line-carved into the stone cave ( Figure 10). These nyorai figures are thought to be either a Miroku 弥 勒 (Skt. Maitreya) triad (Fujisawa 1985), or Amida 阿弥陀 (Skt. Amitâbha), Shakamuni 釈 迦牟尼 (Skt. Śākyamuni), and Yakushi 薬師 (Skt. Bhaiṣajyaguru) (Nishimura 1942). Taketani previously published a detailed measured drawing of these nyorai (Figure 2) (Taketani 1989). Comparing it with the figures on site, I found that the north figure clearly forms a meditation mudra (jōin 定印), and that the central figure probably forms a wish-granting mudra (yogan'in 与願印). By contrast, the southern figure's mudra is unclear; having observed it in person, I believe it is a Miroku. These three figures are the buddhas of the three dimensions of time (sanzebutsu 三世仏); that is, the past, present and future. It is Rokutanji's main object of veneration was likely the nyorai triad that is line-carved into the stone cave ( Figure 10). These nyorai figures are thought to be either a Miroku 弥勒 (Skt. Maitreya) triad (Fujisawa 1985), or Amida 阿弥陀 (Skt. Amitâbha), Shakamuni 釈迦牟尼 (Skt.Śākyamuni), and Yakushi 薬師 (Skt. Bhais . ajyaguru) (Nishimura 1942).
Taketani previously published a detailed measured drawing of these nyorai (Figure 2) (Taketani 1989). Comparing it with the figures on site, I found that the north figure clearly forms a meditation mudra (jōin 定印), and that the central figure probably forms a wishgranting mudra (yogan'in 与願印). By contrast, the southern figure's mudra is unclear; having observed it in person, I believe it is a Miroku. These three figures are the buddhas of the three dimensions of time (sanzebutsu 三世仏); that is, the past, present and future. It is highly likely that transgression repentance at this site entailed the recitation of buddhas' names based on the Sangōsanzenbutsumyōkyō 三劫三千仏名経. Based on earthenware found around the ruins, it should be noted that Rokutanji appears to have been built during the second quarter of the 8th century, around 745. Generally, rituals that involve recitations of the names of buddhas and transgression repentance are referred to as butsumyōe. Based on the Nihon kiryaku 日本紀略, such a ritual is thought to have first appeared in Tenchō 7 (830) (Takei 1980). Therefore, the Buddhist service conducted here would not have been a full-fledged butsumyōe 仏名会. However, from the time that Buddhism was transmitted to China, sutras on the names of the Buddhas existed there, and gained wide popularity during the Tang Dynasty (at the latest) (Yamaguchi 2018). In addition, Shōsōin sutra copying records indicate that such sutras existed in Japan during the first half of the 8th century. There are also many popular-level sutras on the names of the buddhas that are not found in the Kaiyuan lu 開元録 (Jp. Kaigen roku). It is entirely possible that primitive buddha name recitation-based transgression repentance was conducted at Rokutanji. In addition, the figures comprising Rokutanji's main object of veneration (Amida, Shakamuni, and Miroku) are different from the buddhas of the three dimensions of time generally found in Japan (Amida, Shakamuni, and Yakushi). This combination appeared in China during the Sui dynasty, and was popular during the Tang dynasty (Hagiwara 2007). Here, as well as in the above-discussed existence of buddha name-based transgression repentance, we can identify a strong influence from Chinese Buddhism. In this context, it is important to note a point made by Inoue Kaoru regarding the appearance of Rokutanji's 13-story pagoda. With the strong influence of Tang culture during the Nara period in mind, Inoue notes that Rokutanji's pagoda resembles the Small Wild Goose Pagoda (Xiaoyan ta 小雁塔) at Jianfu 薦福寺 temple in China's Xi'an (Figure 11: Inoue 1982). In light of my above examination, we should fully consider this possibility. With that said, there are no technical similarities; only a mental image of a pagoda was introduced to Japan. I also noted this with regard to Ishidōji's three-layer pagoda. Almost all other stone pagodas in Japan were, in the same way, built only based on an image and with completely different techniques from the peninsula. These pagodas were not supported by the full-fledged transfer of techniques using blueprints. In addition, the figures comprising Rokutanji's main object of veneration (Amida, Shakamuni, and Miroku) are different from the buddhas of the three dimensions of time generally found in Japan (Amida, Shakamuni, and Yakushi). This combination appeared in China during the Sui dynasty, and was popular during the Tang dynasty (Hagiwara 2007). Here, as well as in the above-discussed existence of buddha name-based transgression repentance, we can identify a strong influence from Chinese Buddhism. In this context, it is important to note a point made by Inoue Kaoru regarding the appearance of Rokutanji's 13-story pagoda. With the strong influence of Tang culture during the Nara period in mind, Inoue notes that Rokutanji's pagoda resembles the Small Wild Goose Pagoda (Xiaoyan ta 小雁塔) at Jianfu 薦福寺 temple in China's Xi'an ( Figure 11: Inoue 1982). In light of my above examination, we should fully consider this possibility. With that said, there are no technical similarities; only a mental image of a pagoda was introduced to Japan. I also noted this with regard to Ishidōji's three-layer pagoda. Almost all other stone pagodas in Japan were, in the same way, built only based on an image and with completely different techniques from the peninsula. These pagodas were not supported by the full-fledged transfer of techniques using blueprints.

The Adoption of Stone Pagodas in Ancient Japan
The preceding sections have examined the development of stone pagodas in Japan. Some were strongly influenced by the Korean Peninsula, and some were not. The Ishidōji three-story pagoda is a stonework made by settlers from Baekje or their descendants. Due to its distinctive scale, it appears to have been made in imitation of pagodas that were built as temple buildings. Similarly, excavations have shown that Okamasu Ishindō overlaps with the building of the abandoned Okamasuji (Tottori-ken Maizō Bunkazai Sentā 2000); it appears to have functioned as a temple building (Mizuno 2006). These Korean Peninsula-type stone pagodas are different from other stone pagodas in terms of their large size and status as temple buildings, and have strong settler clan monument characteristics.
By contrast, stone pagodas not strongly influenced by the Korean Peninsula include many measuring approximately 2 m in height. The majority of these are found in mountain forests, like Ryūfukuji's pagoda, and some were built as mountain temple equipment, so to speak. Around Ryūfukuji, we can reconstruct an early butsumyōe ritual space in which transgressions were repented to the buddhas three times. It can be assumed that many of these mountain forest stone pagodas were used for these transgression repentance rituals.
In the past, scholars believed 8th-century mountain forest temples were private Buddhist hubs "of the people" that stood in opposition to state Buddhism and, in light of ritsuryō 律令 rules, sometimes even illegal (Futaba 1957). However, it eventually became clear that mountain forest temples and government temples actually formed sets, as pointed out by Sonoda Kōyū, who argues that government temple priests traveled between government temples and mountain forest temples in order to acquire jinenchi 自然 智, or knowledge with which individuals are originally equipped, yet currently unaware of (Sonoda 1981). In recent years, Kikuchi Hiroki has also highlighted cases in which lay male practitioners engaged in jōgyō (浄行), or religious training, to become monks, arguing that this practice originated in Chinese scripture and was transgression repentance (Skt. poṣadha, Jp. fusatsu 布薩/keka 悔過) in essence based on the Mahayana precepts. He holds that while at flatland temples monks had to observe the complete precepts (based

The Adoption of Stone Pagodas in Ancient Japan
The preceding sections have examined the development of stone pagodas in Japan. Some were strongly influenced by the Korean Peninsula, and some were not. The Ishidōji three-story pagoda is a stonework made by settlers from Baekje or their descendants. Due to its distinctive scale, it appears to have been made in imitation of pagodas that were built as temple buildings. Similarly, excavations have shown that Okamasu Ishindō overlaps with the building of the abandoned Okamasuji (Tottori-ken Maizō Bunkazai Sentā 2000); it appears to have functioned as a temple building (Mizuno 2006). These Korean Peninsulatype stone pagodas are different from other stone pagodas in terms of their large size and status as temple buildings, and have strong settler clan monument characteristics.
By contrast, stone pagodas not strongly influenced by the Korean Peninsula include many measuring approximately 2 m in height. The majority of these are found in mountain forests, like Ryūfukuji's pagoda, and some were built as mountain temple equipment, so to speak. Around Ryūfukuji, we can reconstruct an early butsumyōe ritual space in which transgressions were repented to the buddhas three times. It can be assumed that many of these mountain forest stone pagodas were used for these transgression repentance rituals.
In the past, scholars believed 8th-century mountain forest temples were private Buddhist hubs "of the people" that stood in opposition to state Buddhism and, in light of ritsuryō 律令 rules, sometimes even illegal (Futaba 1957). However, it eventually became clear that mountain forest temples and government temples actually formed sets, as pointed out by Sonoda Kōyū, who argues that government temple priests traveled between government temples and mountain forest temples in order to acquire jinenchi 自 然智, or knowledge with which individuals are originally equipped, yet currently unaware of (Sonoda 1981). In recent years, Kikuchi Hiroki has also highlighted cases in which lay male practitioners engaged in jōgyō (浄行), or religious training, to become monks, arguing that this practice originated in Chinese scripture and was transgression repentance (Skt. pos . adha, Jp. fusatsu 布薩/keka 悔過) in essence based on the Mahayana precepts. He holds that while at flatland temples monks had to observe the complete precepts (based on the Theravāda four-part Vinaya), at mountain forest temples they maintained Mahayana precepts based on the likes of the much simpler "three categories of pure precepts" (Skt. trividhāniśīlāni, Jp sanju jōkai 三聚浄戒), and engaged in active mountain forest religious training characterized by more freedom in daily life (Kikuchi 2020). In this way, mountain forest temples in Japan were religious training sites that belonged to flatland temples, and at them transgressions were repented and diverse activities engaged in. Why, then, were stone pagodas installed in such places?
In ancient Japan, there were no stone pagodas without peninsular influences built as temple buildings. Wooden and stone pagodas had different roles. Mahayana scriptures often advocate building pagodas and making offerings. The Lotus Sutra is a major example of this. Its "Expedient Means" and "Spiritual Powers" chapters state that to build a pagoda is both a virtuous act that opens the path to buddhahood, as well as that where a pagoda is built-whether in forests, under trees, on mountains, or in valleys-will become a site for religious training and enlightenment. In East Asia, where Mahayana Buddhism spread, the influence of the Lotus Sutra has been consistently strong. In Japan, importance was particularly attached to the "Expedient Means" and "Spirituals Powers" chapters (they were considered one of the sutra's four essential chapters), and the sutra would widely spread amongst not only priests but also laypeople. Based on these facts, it appears that at mountain forest temples, which were the Mahayana precept-based training sites of flatland temples, first a pagoda was built to symbolize their status as such. In other words, we should see these stone pagodas as having been built on small mountains to establish places for Buddhist training. While, of course, in Japan large wooden pagodas were built at mountain forest temples that served as hubs, I believe that these small stone pagodas were installed at religious training sites where it was not possible to build large temple complexes. In this sense, it appears that in Japan stone pagodas were entirely replacements for wooden pagodas.
Looking at the background of such repentance systems, religious beliefs and practices related to the buddhas of the past, present, and future, and ideas regarding pagoda construction and the establishment of religious training sites, we can clearly see the influence of Chinese Buddhism. In the case of Rokutanji, it appears that a repentance ritual existed which predates the standardization of Buddhism that followed the establishment of a national precepts platform when Jianzhen 鑑真 arrived in Japan. This ritual was based on diverse groups of sutras continually introduced to Japan beginning in the 7th century (e.g., the group of sutras brought by Dōshō 道昭 1 ). Excluding the time after Buddhism's introduction to Japan, Sinification advanced greatly in the seventh century, and it appears that this is a manifestation of this phenomenon.
Yamamoto Jun's research has found that transgression repentance rituals, which were introduced with the transmission of Buddhism to Japan, transitioned from repentance by individuals in the 7th century to rituals directed toward specific objects of veneration in the 8th century. Then, in the 9th century, he argues, these prevailing practices were again replaced by new esoteric Buddhist rituals, including the butsumyōe, before becoming obsolete (Yamamoto 2018). Due to the old transgression repentance rituals held at stone pagoda-centered mountain forest temples shifting to esoteric Buddhist rituals from the 9th century onwards, the symbols of mountain forest temples changed from small stone pagodas to esoteric mandalas, buddhas, bodhisattvas, and deities. The decline in stone pagodas from the tenth century onwards was probably a result of this.

Conclusions
In this paper, I argued the following. First, I noted that ancient stone pagodas in Japan fall under three types, one of which is pagodas exhibiting the influence of the Korean Peninsula. The two types without Korean Peninsular influences are small or medium-sized multistory pagodas, and the small ones were primarily located in mountain forest temples. Second, I pointed out the possibility that all Korean Peninsula-influenced pagodas were built as temple buildings and the other types as symbols to establish mountain forest temples as sites of religious training, as well as that the decline in stone pagodas coincided with the shift to of esoteric Buddhist rituals. Japanese Buddhism was established by taking in Buddhist culture from the Korean Peninsula, and then would became increasingly inclined toward Chinese Buddhism as envoys were sent to China. While one cause of this was changes in foreign relations during the latter half of the 7th century that led to a confrontational relationship with the peninsula, this orientation toward China remained unchanged even after relations improved. This paper's findings basically match the process of Japanese Buddhism maturing as it incorporated new information on Chinese Buddhism. The development of stone pagodas matches the development of ancient Japanese Buddhism.
As noted at the beginning of this article, Kawakatsu Masatarō points out that there are few stone pagodas from this period, and of those that do exist, only several can be dated with certainty. In order to clarify the background of pagoda construction from this limited information, it is necessary to decipher the ideas regarding the spaces in which pagodas were placed. I hope this paper will serve as a springboard for further advancements, based on the relationship between mountain forest temples and ancient Buddhism, in our understanding of stone pagodas.