The Journal of Korean Art and Archaeology

Introduction

The National Museum of Korea (hereafter NMK)’s Central Asia collection features a reed mat (葦席) (Accession No. Bongwan 3978) that was excavated in Turpan, in China’s Xinjiang Uyghur Autonomous Region (Fig. 1). The object, made of woven reed stems, is presumed to have served either as a mat for laying the body of a deceased individual (mummy) to rest or as a funerary item with another specific purpose. Attached to both sides of this reed mat were parts of official documents dating back to the Tang Dynasty (唐, 618–907) identified as the “Report to the Throne from the Board of Revenue in the 3rd Year of Yifeng and Imperial Directive from the Treasury Bureau in the 4th Year” (唐儀鳳三年(678)度支奏抄·同四年(679)金部旨符) (hereafter jointly referred to as “Report from the Board of Revenue”), which belonged to the Japanese Ōtani Documents (大谷文書) and the Chinese Turpan Documents (吐魯番文書). These artifacts are among the most valuable in the NMK’s Central Asia collection since they are the only official Tang Dynasty documents to have been found in Korea.

Fig. 1. Side A of the reed mat in the collection of the National Museum of Korea to which the Tang documents were attached (before the documents were detached). Tang, 703. (left) 16.0 × 48.5 cm (right) 20.0 × 23.5 cm. National Museum of Korea

Fragments of these documents, which are now scattered all across Korea, China, and Japan, were originally discovered in Tomb 230 of the Astana Cemetery in Turpan. Some fragments were transported to Japan following the Ōtani expedition in 1912 and later dispersed throughout Japan and Korea, while the tomb’s remaining documents were only excavated in 1972 by Chinese researchers.

The document fragments that found their way into Chinese and Japanese collections were studied by researchers in each country early on. The extensive amount of research carried out over the last half century has facilitated the classification of document fragments and provided insights into their nature. Furthermore, collaborative efforts between Chinese and Japanese researchers were made to restore the documents to their original form and conduct more in-depth research on their contents. As a result, the Report from the Board of Revenue was revealed to be a surviving section of an imperial directive issued by the Treasury Bureau in 679 CE that the Tang Department of State Affairs (尙書省) had sent out to each prefecture and which had been subsequently received and processed by the Director of the Granaries Section (倉曹司) of the Xizhou Area Command (西州都督府). This document has proven to be a valuable resource in filling the void left behind by the lack of historical materials regarding the history of national finances in the early Tang Dynasty.

In 1995, the reed mat and attached documents from the NMK collection were introduced to the academic world for the first time, which revealed that fragments from the Report from the Board of Revenue were also being kept in Korea. As a result, restoration research on this corpus of documents, which had come to a standstill, arrived at a new turning point. However, since investigations at the time were conducted so as to not alter the original state of the artifacts, this hindered efforts to identify the entirety of the contents contained within the documents and attempts to uncover the methods by which they had been repurposed.

The document fragments in the Japanese and Chinese collections consist exclusively of parts that were peeled off from the reed mat. On the other hand, the reed mat from the NMK collection is particularly valuable in that its condition closely resembles the original form of the funerary item to which the documents had been attached before being removed. In order to conduct a closer examination, however, it was deemed essential to separate the documents from the reed mat. In a previous study, it had been noted that the documents from the NMK collection were closely related to the Chinese documents, raising the strong likelihood that the fragments in the two collections had originally been part of the same sheet. In contrast, such a clear link could not be established between the Chinese and Japanese documents even though they belonged to the same set of records. Hence, it was believed that a more comprehensive understanding of the Report from the Board of Revenue could be attained by examining the soon-to-be disassembled document fragments in conjunction with the fragments in other collections. In addition, relying on the insights gained from the disassembly process regarding the layout of the document fragments, it was thought that it might be possible to deduce where and to what objects the Chinese and Japanese document fragments had originally been attached. With these aims in mind, in 2020, the NMK made the decision to separate the document fragments from the reed mat.

The aim of this paper is to paint a comprehensive picture of the Report from the Board of Revenue by accessing and restoring the document fragments from the NMK collection and combining them with those from the Chinese and Japanese collections. In addition, this paper also seeks to reconstruct a visual representation of the reed mat to which the document fragments in the Japanese collection had been attached by drawing on insights from the restoration of the reed mat in the NMK collection, which will ultimately shed light into the practice of repurposing documents for the production of unique funerary items that was characteristic of the Turpan region at the time. To that end, the paper will first investigate the process by which the reed mat and attached documents came to be scattered across the three countries and offer a historical overview of the research conducted on the document fragments in order to provide all the necessary background information. Next, the fragments from four document sheets, comprising two different document sets (including the Report from the Board of Revenue) which were detached from the reed mat, will be combined with the document fragments in the Chinese collection, and the fragments of the newly identified document set will be analyzed in order to identify the nature and date of this document. Finally, an attempt will be made to reconstruct the reed mat from the NMK collection which is currently divided into two pieces to gain additional insights on the original form and production method of similar reed mats. Relying on this information, it will then be proposed that the document fragments in the Japanese collection had originally been attached to a separate reed mat, which will allow for a modification of the pre-existing reconstruction plan of the reed mat in the Japanese collection.

Dispersal of the Reed Mat and Attached Documents to Korea, China, and Japan

Dispersal Process

Located in the southeastern section of the Astana cemetery in Turpan are the ancient tombs of the Zhang family, the region’s most powerful family from the late period of the Qu Clan Gaochang Kingdom (麴氏高昌國) to the Flourishing Tang (盛唐, 713–765) period. This is where Tomb 230, which yielded the reed mat and attached documents, was located. According to the epitaph tablet, this tomb is where Zhang Lichen (張禮臣, 655–702) and his wife were buried together. Zhang died in the second year of the Changan (長安) reign (702) and was buried the following year (703).

It is said that the first person to enter this tomb was Yoshikawa Koichiro (吉川小一郞, 1885–1978), a member of the 3rd Ōtani Expedition, in late March 1912. He took the reed mat and documents with him to Japan, providing details about the structure and condition of the tomb at the time of investigation, along with an illustration (Fig. 2), in his exploration diary titled Travelogue of China (支那紀行). In the “Condition of the Tomb and Coffin” (墳墓および槨中の状態) section, Yoshikawa recounted that “the mummies were mostly laid out on reed mats to which scrap paper was attached,” noting that “the scrap paper contained numerous records of contemporary events,” which prompted him to “bring them back for reference” (Yoshikawa Koichiro 1937, 606–607). It is highly probable that this description was made with the reed mat and documents that had been brought to Japan in mind, suggesting that some of the documents in Tomb 230 were intentionally left behind while the reed mat was retrieved and taken away.

Fig. 2. Illustration of the ‘condition of the tomb and coffin’ illustration in “Travelogue of China”. “Travelogue of China” by Yoshikawa Koichiro. (Tokyo: Yukosha, 1937, p.607)

The expedition team had been organized and dispatched by Ōtani Kozui (大谷光瑞, 1876–1948), who was the twenty-second abbot of Nishi Hongan-ji Temple (西本願寺) at the time. Due to the deteriorating finances of the Nishi Hongan-ji Temple sect, Otani resigned from his position as abbot in April 1914 to take responsibility for the actions of Buddhist priests who had been imprisoned for embezzlement. Yoshikawa, who was the sole member of the expedition team at the time, was ordered to return home, and the expedition came to an abrupt end. The collection, including the reed mat, left Ürümqi (烏魯木齊) in January 1914 and brought to Nirakuso (二樂莊), the Ōtani’s villa in Kobe on July 10 of the same year.

On November 27, 1914, Ōtani went into exile in Lushun, China. Although he took some artifacts from the Western Regions (西域) with him, some of them are believed to have been returned to the Nishi Hongan-ji Temple for some unknown reason since they were later discovered in Ōtani’s villa in Beppu in 1948 after his death. This collection, which was immediately transferred to Ryukoku University (龍谷大學), included some of the reed mat documents.

The reed mat which had been kept with the attached document remained at the villa, but immediately following Ōtani’s resignation, the Nishi Hongan-ji Temple sect stopped paying the maintenance fees for Nirakuso, and in January 1916, Ōtani sold Nirakuso and most of the collection stored there to entrepreneur Kuhara Fusanosuke (久原房之助, 1869–1965). At the end of April of the same year, Kuhara donated the collection to Terauchi Masatake (寺內正毅, 1852–1919), the first Governor General of Joseon, who hailed from the same hometown. After being acquired by the Japanese Government-General Museum, the reed mat was displayed at Gyeongbokgung Palace’s Sujeongjeon Hall, which served as a storage and exhibition facility for the artifacts from the Western Regions during the Japanese occupation period. It later became part of the NMK collection after Korea’s liberation, where it remains to this day.

History of Research

Research on the reed mat documents began after the discovery of documents that had previously been attached to the reed mat. The Western Region Cultural Research Association (西域文化硏究會) was formed in 1953 to study the Western Regions artifact collection found among Ōtani’s belongings. In 1959, Ogasawara Senshu (小笠原宣秀) first introduced findings concerning sixty-two documents he had found with reed mat traces from the Ōtani Documents, which he had identified as fragments of an official document dating back to the third and fourth years of the Yifeng (儀鳳) reign of the Tang Dynasty (Ogasawara 1959, 387–419). In the following year, Naito Genkichi (內藤乾吉) interpreted it as a document containing not only an order issued by the Treasury Bureau as an Imperial Directive of the Crown Prince in the third year of the Yifeng reign (titled “Notes on the Allocation of Yong (庸) and Diao (調) Taxes for All States in the 4th Year of the Yifeng Reign”), which had been sent out by the Department of State Affairs (尙書省), but also a record of the fact that this order had been received and processed by the Xizhou Area Command (Naito Kenkichi 1960, 87–88).

Meanwhile, in 1972, China’s Xinjiang Museum discovered eight reed mat document fragments that had not been recovered from Tomb 230 during the excavation work conducted by the Ōtani expedition team (Figs. 3 and 4). In 1983, Xu Fuqian (許福謙) annotated the text and made the document fragments known to the world for the first time. However, he was not aware that document fragments belonging to the same records were also present in the Ōtani collection. Based on the place names, and duration of the official names, and contents of the text, he interpreted the document fragments as having come from the records of the Board of Revenue and been produced between 670 and 679 (Xu Fuqian 1983, 543–580). Regarding the Chinese document fragments retrieved later from Tomb 230, it was Ikeda On (池田溫) who first suggested that they and the Japanese reed mat document fragments from the Ōtani collection may have come from the same set of records. He also argued that the document should be viewed as a decree containing financial-related accounts due to the inclusion of matters under the jurisdiction of the Treasury Bureau (Ikeda On 1984, 62–74).

Fig. 3. Fragments of the Report from the Board of Revenue in the Chinese collection. 72TAM230:46/1. Documents Unearthed in Turpan by China Institute of Cultural Relics et al. (Beijing: Cultural Relics Publishing House, 1996, pp.65–66)

Fig. 4. Fragments of the Report from the Board of Revenue in the Chinese collection. 72TAM230:46/2. Documents Unearthed in Turpan by China Institute of Cultural Relics et al. (Beijing: Cultural Relics Publishing House, 1996, p.67)

Based on the above research results, Otsu Toru (大津透) conducted an in-depth study of the document fragments in the Chinese and Japanese collections, coming up with the titles “Tang Yifeng Sannian Duzhi Zouchao Tong Sinian Jinbu Zhifu” (唐儀鳳三年度支奏抄·同四年金部旨符) for the documents. This title points to the third Year of Yifeng of the Tang Dynasty (678), when the Board of Revenue of the Department of State Affairs compiled a report on the national budget for the following year including the income and expenditure (i.e. yong and diao taxes) of the superior prefectures (府) and prefectures (州), and submitted it to the crown prince with a draft document for approval. In the fourth Year of Yifeng of the Tang Dynasty (679), an Imperial Directive issued by the Treasury Bureau for nationwide implementation was received and processed by the Xizhou Area Command, which produced copies to be sent out to the five counties (縣) under its jurisdiction. The document fragments in the Japanese and Chinese collections were seen as having come from one of these copies (Otsu Toru 1986, 1831–1880), allowing Otsu Toru and Enomoto Junichi (榎本淳一) to piece together the fragments in the Ōtani collection and successfully reconstruct the entire records (Fig. 5). In addition, based on their newfound understanding of the structure of the reassembled document, they were able to propose a reconstruction scheme for the reed mat as well (Otsu Toru and Enomoto Junichi 1987, 47–48).

Fig. 5. Photo of the document fragments of the Report from the Board of Revenue in the collection of Ryukoku University in their connected state. Otani Documents Collection by Oda Yoshihisa. (Kyoto: Hozokan, 1990, p.10)

In 1995, it was announced to the academic world that associated document fragments also existed in Korea. After interpreting the text featured on the back of the reed mat document fragments, Min Byung-hoon proposed that these document fragments, as with those in the Japanese and Chinese collections, also formed part of the Report from the Board of Revenue. It was also revealed that there was another document layer hiding beneath the uncovered document fragments, and was suggested that their contents were closely related with those from the Chinese collection (Min Byeong-hoon and Ahn Byeongchan 1995, 156–180). Upon learning about the new data and interpretations, Otsu compared all of the translations that had been produced up to that date and compiled a restoration plan for the entire records by synthesizing the document fragments from the three countries (Otsu Toru 2006, 27–112). In particular, it was maintained that, judging from the content and form of the document fragments, the fragments from two documents in the NMK collection were likely to be connected to the fragments from two documents in the Chinese collection (Otsu Toru 2000, 239–244).

Separation and Investigation of Documents and Prototype of NMK Collection Reed Mat

Before presenting the outcomes of the document separation, it is essential to conduct a brief examination of how these documents were initially arranged on the reed mat. Side A of the larger reed mat piece was entirely covered with the Report from the Board of Revenue, and the remaining parts of the documents were folded along three edges to cover the outer edges of Side B. On Side B, an additional two sheets of documents were glued to the surface, aligned at a right angle to the document that covered Side A and extended to parts of Side B (Fig. 6). The separation process revealed the existence of four document sheets in total, comprising two sets of documents. It was revealed that two layers of documents belonged to the Report from the Board of Revenue, while the two documents attached to Side B were part of another set of documents. The NMK assigned serial numbers to the documents according to the document set and the order in which they were processed. The two documents belonging to the Report from the Board of Revenue were numbered “2020NMK1:1” and “2020NMK1:2,” while the documents from Side B were numbered “2020NMK2:1” and “2020NMK2:2.”

Fig. 6. Scheme of how document fragments were attached to Side B of the reed mat in the NMK’s collection. National Museum of Korea. Author

Report from the Board of Revenue

The two documents from the Report from the Board of Revenue were placed in an overlapping state and were of nearly identical shape and size (Figs. 7 and 8). While the edges on the right side of both sheets are smooth, the edges on the left side are irregularly shaped, which is likely the result of external force having been applied to tear off this section of the overlapping sheets. There are also traces of a purple fabric called qi (綺), which appears to have been attached to the outer edges of the reed mat, which is especially visible on the backside of the upper right section of 2020NMK1:1. The pattern of the fabric is not clear, but it is presumed to be a geometric pattern with repeated diamonds (Fig. 9).

Fig. 7. Document sheet from Report from the Board of Revenue numbered 2020NMK1:1 (forming the outer layer). Tang, 679. 28.5 × 52.4 cm. National Museum of Korea. (Red box denotes the location (on the opposite side) of a remaining piece of the purple fabric that had been attached to the back side of the reed mat.)

Fig. 8. Document sheet from Report from the Board of Revenue numbered 2020NMK1:2 (forming the inner layer). Tang, 679. 28.5 × 53.0 cm. National Museum of Korea

Fig. 9. Purple qi (綺) fabric attached to the back side of the reed mat. 2.2 × 0.9 cm. National Museum of Korea

The remaining number of lines on each document is twentieth and eighteenth respectively, with the number of characters ranging from fifteen to seventeen per line. Letters were written in a font similar to the regular script (楷書), leaving an empty space of around 4.5 centimeters at the top of the manuscript. Above each line where a new entry (條目) begins, there is a space with the character ‘一’ serving as a marker. Each of the documents contains three complete entries, as well as incomplete records that would have continued on to the left and right. The documents also feature a paper seam, with the character ‘儉’ written at the base on the reverse side of each document, which was made by Yuan Huaijian (元懷儉), the Military Advisor of the Granaries Section (倉曹參軍事) of the Xizhou Area Command, in his capacity as judge to confirm the binding of the document.

As expected, it was confirmed that two documents belonging to the Report from the Board of Revenue in the NMK collection could be matched to the two documents from the Report from the Board of Revenue in the Chinese collection (72TAM230:46/1, 72TAM230:46/2) (Fig. 10). Since the two document sheets in the Chinese collection are similar in size and shape, it was presumed that these two sheets had also been layered upon each other when attached to a reed mat. As a result of combining and reading the separated documents together, four new entries were identified, and six previously incomplete entries could be completely restored. These entries contained a wealth of information that could not have been confirmed through other sources (Table 1).

Fig. 10. Connected state of the fragments of the Report from the Board of Revenue in the Korean and Chinese collections. Documents Unearthed in Turpan by China Institute of Cultural Relics et al. (Beijing: Cultural Relics Publishing House, 1996, pp.65-67)

Additionally, it was possible to determine the size of the document sheets that had been cut out from the Report from the Board of Revenue to be repurposed into funerary items. As indicated by the combined restoration diagram, both edges of 2020NMK1:1 and 72TAM230:46/2 were cut in a neat fashion, suggesting that the length of the sheet that was cut out would have been around 92 centimeters, making it identical to the combined length of 2020NMK1:2 and 72TAM230:46/1 at around 93 centimeters. In other words, segments around 93 centimeters in length were cut out from a long paper roll report, comprised of paper sheets around 40 centimeters wide and 28.5 centimeters high, that were stitched together when making the reed mat. Meanwhile, in a previous study, it was observed that the Report from the Board of Revenue documents in the Japanese collection consisted of six document sheets that would have followed the Chinese collection document sheets within the original report, meaning that the documents from the NMK collection constitute the very first part of the extant Report from the Board of Revenue. Although the entire length of the original report remains uncertain, considering the nature of the document as a budget directive concerning all of the administrative districts of the Tang Dynasty, it is likely to have been of considerable length. The arrangement of documents from the Korean, Chinese, and Japanese collections in the order in which they originally appear in the Report from the Board of Revenue can be illustrated as follows (Fig. 11).

Fig. 11. Reconstruction of the Report from the Board of Revenue and the sequence of the document sheets from the Korean, Chinese, and Japanese collections. Author

Official Document Regarding Fugitive Imperial Guards

The two documents covering Side B of the reed mat (Fig. 12) were confirmed to have been part of the same set as documents 72TAM230:62 and 72TAM230:63 from the Chinese collection, both excavated from the Astana Tomb 230. Most notably, it was verified that 2020NMK2:1 could be combined with 72TAM230:62 and three other “document fragments” (文書殘片) numbered 72TAM230:80/4, 72TAM230:80/9, and 72TAM230:80/10 (Fig. 13).

Fig. 12. Document fragments of the Official Document Regarding Fugitive Imperial Guards in the in the collection of the NMK 2020NMK2:1 (left), 2020NMK2:2 (right). Tang, 674–676. (left) 17.9 × 8 cm (right) 29 × 16 cm. National Museum of Korea

Fig. 13. Document Sheet (3) – Connection of document fragments in the Korean and Chinese collections (solid line: cutting line; dotted line: seam; [元]: signature written on the seam; band-shaped outline: range of the reed mat impression). Documents Unearthed in Turpan by China Institute of Cultural Relics et al. (Beijing: Cultural Relics Publishing House, 1996, pp.85–88)

The document fragments in the Korean and Chinese collections, numbering seven in total, came from three document sheets that had been attached to Side B of the reed mat in the NMK collection. For convenience, according to the order of the contents, the three document sheets will be referred to as Document Sheets (1), (2) and (3). Document Sheet (3) (Fig. 13) consisted of 2020NMK2:1 from the Korean collection and four document fragments from the Chinese collection, making up the uppermost layer among the three overlapping document layers of Side B. Document Sheet (2), which corresponds to 2020NMK2:2 (Fig. 14), was covered by Document Sheet (3) and attached to the right section of the reed mat, with the left and right ends of the sheet tailored to match the width of the mat. The document sheet features a band-shaped impression mark from the reed mat in the center, which can also be observed on Document Sheet (1) (72TAM230:63) (Fig. 15). It is presumed to have been attached to the left section of the reed mat which does not have any documents attached to it in its current state.

Fig. 14. Document Sheet (2) – 2020NMK2:2 (solid line: cutting line; band-shaped outline: range of the reed mat impression). National Museum of Korea

Fig. 15. Document Sheet (1) – 72TAM230:63 (solid line: cutting line; dotted line: seam; [大]: band-shaped outline: range of the reed mat impression). Documents Unearthed in Turpan by China Institute of Cultural Relics et al. (Beijing: Cultural Relics Publishing House, 1996, p.85)

This set of documents addresses the measures taken by the Xizhou Area Command in response to a Gaochang Prefecture (高昌縣) report on the loss of taxes in the form of cloth, known as yongxie (庸緤), due to the desertion of Imperial Guards (衛士), which includes an estimation of the incurred losses to be collected from the Xizhou Area Command in the form of silver coins. As such, this set of documents was given the title “Official Document Issued by the Granaries Section of the Xizhou Area Command in the Tang Shangyuan Era Regarding Fugitive Imperial Guards of Gaochang Prefecture” 唐上元年間(674–676) 西州倉曹司案卷爲高昌縣申送逃走衛士庸緤價錢事) (hereafter “Official Document Regarding Fugitive Imperial Guards”).

To determine the date of this document set, research was carried out on an individual by the name of Yuanli (元利), who served as the Military Advisor of the Granaries Section (倉曹參軍) of the Xizhou Area Command. According to document 2004TBM207:1-12 excavated from Tomb 207 in the Turpan cemetery of Badamu (巴達木), the individual’s full name was Zhang Yuanli (張元利), and he was serving as the Military Advisor of the Granaries Section of Xizhou Area Command in the twenty-ninth day of the seventh month of 675, resigning from his position not long after in the third month of 676 (Rong Xinjiang, Li Xiao, and Meng Xianshi 2008, 72–73). If we leave open the possibility that his tenure may have started earlier than the recorded dates, then it may be suggested that he served as Military Advisor of the Granaries Section from around 674 to early in the third month of 676. The Official Document Regarding Fugitive Imperial Guards could then be regarded as having been written around this time. In summary, the making of the reed mat in the NMK collection was done with sheets from two sets of documents that had been produced by the Granaries Section of Xizhou Area Command around the 670s. In addition, the document fragments from the Chinese collection would have originally been attached to the reed mat which is now part of the NMK collection.

Prototype and Production of NMK Collection Reed Mat

Lastly, in order to determine how these documents were arranged, it is necessary to understand what the reed mat, currently divided into a larger and a smaller piece, originally looked like. Both pieces feature an unraveled side, suggesting they were originally part of a single item. Adjusting the shapes of the reed mat pieces by rearranging loose reeds and virtually filling the space in between, the restored reed mat measures approximately 16×78 centimeters, which corresponds to the length of the document sheet cut out from the Report from the Board of Revenue (92 cm) when the lengths of the folded parts at the left and right ends (14 cm in total) are taken out.

Based on the above, and comprehensively taking into account the nature of the connection between the document types in the Korean and Chinese collections, the size of the connected document pieces, as well as the layout and layer sequence of the document sheets attached to the reed mat including the appearance of the currently attached document fragments, a restoration diagram of the extant document pieces can be produced with an understanding of how they were attached to the reed mat, as illustrated in Fig. 16. In addition, the traces of purple fabric which covered the documents along the edges of the reed mat on both sides suggest that it was used as a finishing material (Fig. 17). Based on this information, the production process of the original reed mat from the NMK collection can be restored as follows (Fig. 18).

Fig. 16. Scheme of how the document fragments in the Korean and Chinese collections were attached to Side B of the reed mat in the collection of the National Museum of Korea. National Museum of Korea. Author

Fig. 17. Purple qi (綺) fabric attached to the reed mat in the collection of the National Museum of Korea, Side A (left): Glass Plate No. 030231; Side B (right): Glass Plate No. 030234. Images in the National Museum of Korea’s collection of dry glass plate negatives

Fig. 18. Prototype and production process of the reed mat in the collection of the National Museum of Korea. Author

New Restoration Plan for the Reed Mat Document Fragments in the Ryukoku University Collection

The text featured in the eighty-six pieces of “reed mat document fragments” in the Ryukoku University collection is approximately three times the length of the text featured in the combined document fragments in the Korean and Chinese collections. Based on the restoration case of the reed mat in the NMK collection, a review will be undertaken on the previous restoration plan for the document fragments in the Japanese collection.

According to Otsu and Enomoto’s research, the eighty-six document fragments were originally composed of two document sheets glued together which were torn to pieces as they were separated from the reed mat. By connecting the document fragments in their current state (combined in the two document layers), it was possible to make a restoration consisting of six sheets that had been cut out from the Report from the Board of Revenue, with two layers of document sheets positioned to form three rows (Fig. 5). The first version of the conceptual diagram illustrating the connected state of the document fragments (Fig. 19) is composed of the following three rows: D-E-F (F´-E´-D´), G-A (G´-A´) and B-C (B´-C´). However, it can be observed that fragment F (F´), which should be part of the first row, is actually attached to the backside of fragment A´, on the left side. In addition, G (G´), which should be connected to the right of fragment A (A´) in the second row, is attached to the right of D (D´) in the first row. This demonstrates that some of the document fragments in the original restoration plan were misplaced. Based on the fold marks on the documents and the traces of purple fabric on the back, Otsu and Enomoto presented a second version of the conceptual diagram in which the connected state of the documents was closer to the original (Fig. 20).

Fig. 19. Conceptual diagram of how the document fragments of the Report from the Board of Revenue in the Ryukoku University collection were connected (first version). (Otsu Toru and Junichi Enomoto 1987, p.66)

Fig. 20. Conceptual diagram of the reconstruction of the document fragments of the Report from the Board of Revenue in the Ryukoku University collection (second version). (Otsu Toru and Junichi Enomoto 1987, p.68)

Meanwhile, Otsu and Enomoto interpreted the state of the document fragments illustrated in the second version of the conceptual diagram as the result of the detached document fragments having been reassembled into the shape of an envelope for ease of transportation by the expedition team during the recovery process, rather than representing the original layout of the documents as they had been attached to the reed mat (Otsu Toru and Enomoto Junichi 1987, 68–70). Originally, document sheets of the same length had been attached in three rows and two layers to the reed mat. In the process of making an envelope, however, the expedition team removed several centimeters from both ends of the first row, folded the document sheets along the dotted lines (3), (4) and (2) marked on the second version of the conceptual diagram, and folded the sheets by tucking the first row inwards along dotted line (1). In addition, the two small document fragments attached to the backside of the sheets forming the third row were believed to have fallen off the upper section of sheet D-E-F of the first row. It was also proposed that the sheets forming the first row (D-E-F (F´-E´-D´)) would have originally been attached below the sheets currently forming the third row (B-C (B´-C´)). In other words, the connection between the sheets of the first and second rows was seen as having been artificially made by the expedition team who reassembled the documents. As a result, a third version of the conceptual diagram was produced in which the three rows of documents looked as though they had been spread out and attached to one side of the reed mat, rather than having been folded and attached to both sides of the reed mat (Fig. 21). Otsu and Enomoto argued that the reed mat, over which a deceased individual would have been laid to rest, should have been longer than the length suggested in the proposed restoration plan for the three rows of documents (the dimension of which was 94.0 × 103.0 cm) in order to be closer to the estimated height of the deceased. In other words, it was suggested that there could have been additional documents attached above or below the three rows of documents, in which case some of them could have been documents from the Chinese collection.

Fig. 21. Conceptual diagram of the final reconstruction of the document fragments of the Report from the Board of Revenue in the Ryukoku University collection (third version). (Otsu Toru and Junichi Enomoto 1987, p.69)

If we re-examine the three restoration plans based on the way the documents were laid out on the reed mat from the NMK collection, it appears more likely that documents had been laid out in a way closer to the second version of the conceptual diagram, rather than the third version. The traces of the purple fabric located along the fold lines on the backside of the document can be regarded as clear evidence of the outline of the reed mat and its original form. Therefore, the documents of the first and second row should be viewed as having been combined at the time of the reed mat’s production rather than as having been artificially put together at the time that the documents were taken out to Japan. In addition, the reason that the length of the first row (D-E-F (F´-E´-D´)) is shorter than that of the second and third rows is not because the expedition team cut off both ends to form an envelope, but because the sheet had been tailored to fit the dimensions of the reed mat during the production stage. A similar phenomenon could be observed in the case of the reed mat from the NMK collection, in which the front side of the reed mat was covered using the Report from the Board of Revenue, which was folded at the edges to cover parts of the back side, and the remaining exposed surface of the back side were covered using sheets from the Official Document Regarding Fugitive Imperial Guards, which were tailored according to the dimensions of the reed mat. One plausible explanation for the slight offset (about 10 cm) between document sheets D-E-F and F´-E´-D´ is that sheet D-E-F was initially used during the process of covering the exposed section of the reed mat before sheet F´-E´-D´ was subsequently added to cover the remaining exposed areas. In light of the arguments presented above concerning the restoration process, the original reed mat to which the Ryukoku University collection documents were attached can be reproduced using the following method (Fig. 22).

Fig. 22. New reconstruction scheme and production process of the reed mat to which the document fragments in the Ryukoku University collection would have been attached. Author

Conclusion

The Official Document Regarding Fugitive Imperial Guards and Report from the Board of Revenue documents were each produced between 674–676 and 679 respectively by the Granaries Section of Xizhou Area Command. After the completion of relevant events, the corresponding documents would have been stored in the Granaries Section for a certain period of time, after which they were distributed to the private sector once their original purpose was fulfilled. The document sheets were repurposed into materials used to make funerary items, eventually ending up in the burial site currently known as Astana Tomb 203 the year following Zhang Lichen’s death in 702.

The reconstruction scheme for the reed mat in the Korean collection, along with the lost reed mat to which the document fragments in the Japanese collection came from, contains information on how each document sheet was sequentially cut out from the original report, which is reflected in the order in which the document sheets were organized. In the restoration scheme presented above, the circled numbers in Figs. 18 and 22 indicate the order of the documents in the original report. Comparing these numbers with the order of production of the two reed mats, it becomes apparent that sheets were cut out from the Report from the Board of Revenue in a reverse order, beginning with part C and ending with sheet 2020NMK1:1. This fact indicates that the reed mat onto which the documents in the Japanese collection had been attached was made before the reed mat from the NMK collection. It is also apparent that the document sheets were cut out from the report with the dimensions of the reed mat in mind from the outset, rather than using materials cut in advance. Since there does not appear to be any missing parts between the cut lines of the documents in the Japanese collection, it is unlikely that the document sheet was cut anew after having been cut out from the original report. Therefore, it can be proposed that, after six document sheets were cut out in reverse order from the Report from the Board of Revenue (now part of the Ryukoku University collection) to be used to make a reed mat (which no longer exists), an additional two document sheets were cut out from the Report from the Board of Revenue (now part of the Chinese and NMK collections) and three document sheets from the Official Document Regarding Fugitive Imperial Guards (now part of the Chinese and NMK collections) to be used to make a smaller reed mat which currently finds itself in the NMK collection.

The above results naturally lead to the question of how the reed mat was used, which remains an unresolved issue. Was this reed mat actually used as a type of covering upon which to lay the body of a deceased individual to rest? Yoshikawa’s mention of reed mats and scrap, quoted above, is most likely to have been referring to the reed mat from the NMK collection and the document fragments in the Japanese collection, given that they represent the only extant examples of a reed mat with documents still attached. Although both his description and illustration (Fig. 2) indicate that the size of the reed mat had been close to the height of the bodies, the reed mat from the NMK collection and the one that can be reconstructed from the document fragments in the Ryukoku University collection are actually much smaller.

Nevertheless, although the two reed mats are different in size, it is clear that they were both made into the shape of a rectangular sheet using the same materials and manufacturing method. In other words, it is clear that they had been originally planned and produced as a set, consisting of at least two items that served the same purpose. In addition, the entire surface of the reed mat had been covered with paper to completely cover the exposed surface, and a strip of purple fabric had been attached along the square border of the mat on both sides, embellishments that cannot be interpreted simply as having served a purpose for maintaining the original shape of the reed mat.

In relation to this, it is worth nothing that the purple fabric attached to the reed mat (Fig. 9) is identical to the fabric attached to the border of the six-panel folding screen painting entitled “Woman Figure with a Musical Instrument” (樂伎圖) also excavated from Tomb 230, as well as the fabric used in the “dark purple diamond lattice cicada pattern twill” (絳紫色菱格蟬紋綺) (72TAM 188:B) excavated from Tomb 188, which is presumed to be the tomb of Zhang Lichen’s brother judging from the color, size, and pattern (Tokyo National Museum 2002, 60, Figure 43; Xinjiang Institute of Cultural Relics and Archaeology 2000, 118, Figure 24-5, Plate 6-2). Since the qi excavated from Tomb 188 is only a piece of fabric, and the dimensions of a single folding screen panel from Tomb 230 are different from those of the reed mat, it is obviously difficult to establish a correlation between the three. However, given that the same type of fabric was used, that the method of using a band of fabric to wrap the edge as a finishing technique is similar, and that the purpose of the folding screen painting was to cover the back wall of the location where the deceased would have been laid to rest, it is entirely possible that the reed mat and the folding screen were made by the same group of individuals. As such, the interrelationship between the groups who produced these funerary items is a topic that requires consideration from various angles.

This paper expanded on previous research efforts aimed at elucidating the connection between the document fragments of the Report from the Board of Revenue and their content through an investigation of the process involved in the production of reed mats used as funerary items, which included the repurposing of official documents. The reed mat along with the attached Tang documents that are part of the NMK collection serve as important artifacts with the potential to generate new topics of research in the future.