Chen Songchang 陳松長, Mawangdui boshu yanjiu 馬王堆帛書研究 (Studies on the Mawangdui Silk Manuscripts), Beijing 北京: Shangwu yinshuguan 商務印書館 2021. ISBN 9787100196345; 6, 440 pages; 198 RMB.

Reviewed by Michael Friedrich (Universität Hamburg)

Almost fifty years ago, tomb 3 of the early Han Mawangdui 馬王堆 site (hereafter MWD) in Changsha was excavated in December 1973 and could be dated to 168 BCE. The archaeologists unearthed a wealth of artefacts similar to those found in tomb MWD 1 which had already been excavated in 1972. However, a lacquered chest containing two bamboo manuscripts and more than two dozen silk manuscripts had a much longer-lasting impact on the study of ancient China than the mummy of the tomb owner found in MWD 1.[1]

Due to the political struggles of the time, so it is said, the excavation campaign and the treatment of the artefacts immediately following it were a far cry from international standards and caused, in addition to a lack of documentation, further damage.[2] All silk artefacts from the site, including the two paintings hanging on the eastern and western walls of the coffin chamber and the T-shaped banners from MWD 1 and MWD 3 were transferred to the Palace Museum in Beijing, because the Hunan Provincial Museum 湖南省博物館 (officially now Hunan Museum) did not have any previous experience with this type of artefacts. The Palace Museum treated the most spectacular ones first, among them the two scrolls containing the Laozi 老子 text. After conservation and mounting on paper, the fragments of the artefacts were taken back by conservators of the Hunan Provincial Museum.[3]

The history of research on the MWD written artefacts has yet to be written, a first and apparently incomplete overview is provided in the latest edition of the manuscripts’ contents from 2014.[4] According to this summary, three main stages may be discerned:

1974–1985: Shortly after the excavation was finished, the State Bureau of Cultural Relics 國家文物局 (officially now National Cultural Heritage Administration) took control and established a ‘collation group’ (zhengli xiaozu 整理小組) at the Cultural Relics Publishing House 文物出版社 in Beijing, composed of scholars from various institutions nation-wide who worked with black-and-white images of the manuscripts. After publishing a short excavation report and preliminary editions of some texts such as the Laozi, in 1978 volume 1 of the series Mawangdui Hanmu boshu 馬王堆漢墓帛書 appeared, followed by volumes 3 and 4 in 1983 and 1985 respectively, also containing black-and-white images of the manuscripts;[5] this six-volume edition was never completed. After 1978, the group dissolved because the scholars coming from other places were called back to their units.

1986–2007: For an international conference on the findings from MWD in 1992, the Museum in Changsha edited a sumptuous volume containing editions and colour images of some of the items so far unpublished.[6] Especially the Zhouyi 周易 and its commentaries led to a second wave of publications after the Laozi and ‘Yellow Emperor text’ (Huangdi shu 黃帝書) ‘fever’ immediately following the excavation and the first publications. In 1995 the ‘collation group’ was revived and worked for more than a year in the offices of the Cultural Relics Publishing House again. When this became too expensive, the scholars left after a year and continued their work at their home institutions. Just as before, this resulted in the de facto dissolution of the group. The next major task was the report on the excavation of MWD 2 and 3. Thirty years after the event had taken place, its first volume came out in 2004 and included black-and-white images and editions of the texts on bamboo slips, including the ‘inventory’ (qiance 遣冊), but the second volume on the silk manuscripts never did, thus leaving a second publication ruin.

2008 until today: It took another ten years, until finally in 2014, forty years after the excavation, a complete edition of all contents of the manuscripts from the MWD site including colour images was published in seven very big and very expensive volumes. This edition was the result of a joint venture of Hunan Museum, Fudan University’s ‘Center for Research on Chinese Excavated Classics and Palaeography’ 復旦大學出土文獻與古文字研究中心 and the renowned Zhonghua Book Company 中華書局. It was deemed needing revision even before it came out, and already in June 2015 a conference was held to collect addenda et corrigenda.[7] One may add the proceedings of a conference held on the occasion of the fortieth anniversary of the excavation of the MWD site. It appeared in 2016 and contained more than thirty contributions to the study of the manuscripts.[8]

The author of the book under review was one of the key figures during the second phase. Chen Songchang 陳松長 (b. 1957), since 2006 professor at the Yuelu Academy 嶽麓書院 of Hunan University, entered Hunan Provincial Museum in 1988 and became its Vice-director in 1995. He was one of the co-editors of the 1992 volume and edited a volume on the ‘art of the Mawangdui silk manuscripts’ in 1996.[9] Just as the previous publication, it included images of still unpublished manuscripts and partial transcriptions, grouped according to ‘calligraphic’ styles. He became a member of the revived ‘collation group’ and was in charge of the edition of the ill-fated second volume of the excavation report which apparently failed, because one or more of his colleagues were not able to surmount the obstacles posed by some of the manuscripts.[10] In his years with the museum, Chen acquired an unsurpassed familiarity with the artefacts, allowing him to draw his conclusions from direct engagement with them, whereas most scholars had to rely on images. Since the early 1990s, he has published dozens of articles and some monographic studies on the calligraphy, codicology, palaeography and philology of the MWD silk manuscripts. In 2008, a first collection of his earlier articles dealing with ancient manuscripts appeared, thirty of them were devoted to those from MWD.[11]

According to the author’s 2019 Foreword (qianyan 前言, pp. 3–6), he was invited by The Commercial Press商務印書館 to publish a collection of his articles solely devoted to the MWD silk manuscripts and gladly accepted, as this would allow him to conclude his research on the MWD corpus and facilitate access to his dispersed works. He further states that he only corrected a few characters and adjusted related content, in order to faithfully present his earlier work even if marred by ignorance. Unfortunately, the author (or the publisher) did not only forget to provide the references for the original publications but also to mention the existence of the 2008 collective volume, where the original places of publication were duly noted. Twenty-four of its thirty articles on the MWD silk manuscripts re-appear in the present volume.

Following the author’s short foreword, there are seven chapters: one on codicology (馬王堆帛書形制研究, pp. 1–36; two articles), one on the nature and classification of the contents of the manuscripts (馬王堆帛書性質與分類, pp. 37–49; two articles), chapters on three types of related texts and the manuscripts containing them: commentaries on the Zhouyi (馬王堆帛書《易傳》整理與研究, pp. 50–73; three articles); the Xingde texts (馬王堆帛書《形德》研究, pp. 74–193; nine articles); and the Yinyang wuxing texts (馬王堆帛書《陰陽五行》研究, pp. 194–231; four articles); followed by one on the ‘silk diagrams’ (馬王堆帛圖研究, pp. 232–312; seven articles), finally one on ‘calligraphy’ (馬王堆帛書書法藝術研究, pp. 313–345; three articles). A classified bibliography of almost one hundred pages continuing the one in the 2008 collection closes the volume (pp. 346–440). It covers Chinese, Japanese and a few English publications and includes translations from other languages, but excludes archaeological and material studies as well as any on the bamboo manuscripts; it reaches up to September 2018.

The chapters are not only of uneven length, but also differ in quality and cohesion. Chapter Two is probably the weakest one. Its first article (pp. 37–44) states that the manuscript collection despite its seemingly heterogenous nature must represent an ‘organic whole’ (youjide zhengti 有機的整體). The author then tries to prove its Daoist ‘tendency’ (qingxiang 傾向) by quoting Sima Tan’s 司馬談 (d. 110 BCE) statement that the Daoists integrated the best of all other ‘schools’ (jia 家). If taken literally, it would be difficult to find criteria for non-Daoist ‘tendencies’. The second article (pp. 44–49) classifies the content of the written artefacts according to the bibliographic chapter (Yiwenzhi 藝文志) of the Hanshu 漢書 (History of the Han), claiming that in this way one would use Han-time categories to look at the texts.[12] This claim is more than doubtful, because the ‘Confucian’ system developed for the imperial collection and taken over into the Hanshu was probably without precedent and would not have corresponded to realities more than one and a half centuries earlier when the tomb owner of MWD 3 was active, and the imperial family had a pronounced interest in ‘Daoism’.

Chapters Three through Five contain more or less copious philological notes, still relevant for studying the contents and the dates of the Zhouyi commentaries and the divinatory texts, although now to be read in conjunction with the 2014 editions. The texts belonging to various types of military divination were unknown before, which posed an even greater challenge to their reconstruction from the surviving fragments than the piecing-together of texts with transmitted parallels. This puzzle led to ever-changing arrangements of the fragments with consequences for understanding their content. Chen was one of the first scholars to study these texts in depth, and even if not all of his suggestions were generally accepted, they have contributed to shaping the field as it presents itself today.

There are insights of a more general nature. Concerning the MWD Xici 繫辭 that is shorter than the received text and has a partially different order of the sections, the author concludes that one of the main differences to the transmitted text is the absence of divinatory practice, such as the section on the yarrow stalks in the transmitted Xici, and of arcane language. In addition, passages which have troubled later commentators such as xian zhu ren, cang zhu yong 顯諸仁, 藏諸用 are also absent as well as statements on the line positions (yaowei 爻位) that connect the two trigrams of a hexagram, thus resulting in a more intelligible text. Chen concludes that the MWD Xici represents an earlier and more homogenous stage of the text which was later augmented by Han specialists (56-59). Following the principle of lectio difficilior, however, one may argue just the other way around and consider the MWD version as a redaction intentionally removing the opaque parts. The author observes similar features in the Ersanzi wen 二三子問. Confucius’s replies to his disciples’ questions are based on textual exegesis only, without regard to line configuration or to the imagery of the trigrams and hexagrams. The author not only draws the hardly surprising conclusion that this text is Confucian, but goes one step further: Sima Qian 司馬遷 (d.  c. 86 BCE) was right, Confucius did study the Zhouyi (pp. 72-73).[13] Finally, if both commentaries interpret the text, but do not refer to divinatory practices, then the scroll could be understood as a witness to Confucian appropriation of the ‘classic’. It is a pity that he did not further elaborate on the codicological units.

Zhouyi and Ersanzi wen were written on one scroll, while the other five texts related to the Zhouyi, namely Xici, Zhong 衷, Yao 要, Miu He 缪和 and Zhao Li 昭力, were contained in a second scroll.[14] Like most MWD silk scrolls, these two are multiple-text manuscripts and need to be studied as a whole; otherwise important evidence may simply be overlooked, such as scribal habits pertaining not only to one text but to all texts on a scroll. Even more important, however, is the overall intention guiding the composition of such manuscripts which will not be understood if the textual units are studied in isolation.[15]

With nine articles, Chapter Four on the Xingde texts is by far the longest.[16] There are three versions of these manuals addressing various kinds of divination, according to Chen A (jia 甲) is in ‘old clerical’, B (yi 乙) in ‘Han clerical’ and C (bing 丙) in ‘seal clerical’. Aside from a general introduction (pp. 74–87), philological commentaries (pp. 134–193) and studies of terminological questions (pp. 93–98, 120–134), there is a detailed discussion of Xingde C which, according to the author, is closer in content to the Yinyang wuxing texts than to A and B (pp. 87–93). Another topic dealt with in two articles is the relation between the A and B versions (pp. 98–112, 112–120). With some significant exceptions, both manuscripts in principle contain the same content units, but use a mirror-inverted layout. In more than one content unit, A in ‘old clerical’ script mentions ‘[year] 11 of the present emperor’ (jin huangdi shiyi今皇帝十一) which must correspond to the first Han ruler and to the year 196 BCE, while B in ‘Han clerical’ script only once has the date of ‘[year] 1 of Hui, the Filial’ (Xiao Hui yuan 孝惠元), corresponding to 194 BCE. In version A there are more hints of concrete circumstances, while these are absent from B. Following Marc Kalinowksi, the author considers B to be a revised copy or more precisely a work based on A (p. 113), thus overcoming the simplistic concept of all manuscripts merely being ‘copies’ and differences between two copies resulting from different models (see p. 10). He also observes that the text of A is sometimes better than that of B (pp. 108–109), without drawing any conclusions.

The studies on the Yinyang wuxing texts in Chapter Five mainly focus on version A which due to its heavy damage was fully published only in 2000, and is still described as being written in ‘seal script’ in the first article on terminology (p. 195). Besides a contribution to its palaeography (pp. 202–212) and an introduction to its content, now under the title Shifa 式法 because of some similarities to the diviner’s board shi 式 (pp. 212–226),[17] there is an interesting piece on terms for the periods of the day in the Chuxing zhan 出行占, which may be of interest to those studying more mundane topics (pp. 223–231).

Two of the seven contributions in Chapter Six deal with the diagram that features the deity Taiyi 太一 (pp. 232–252), two with the meteorological divination in Tianwen qixiang zazhan 天文氣象雑占 (pp. 252–281), and one each with Wu ze you xing 物則有形 (pp. 282–291), Jiu zhu 九主 (pp. 292–304) and Yinyang zhaiwei zhaixing jixiong tu 陰陽宅位宅形吉凶圖 (pp. 304–312). In addition to advanced reconstructions of the mostly heavily damaged artefacts and further palaeographic, philological and terminological studies, Chen connects some of their contents not only to other texts in the MWD corpus, but also to transmitted literature, archaeological evidence and manuscripts from Dunhuang. He demonstrates the omnipresence of Chu religion and mythology in these texts and images, which corresponds to the tomb architecture. The last article on the divination of favourable spots for buildings shows how ancient traditions made it into medieval times. It is perhaps not by accident that this article was first published in 2016 and explicitly criticises the 2014 edition (p. 306).[18]

The chapters on codicology and calligraphy address issues pertaining not only to the MWD corpus but also to the study of ancient Chinese manuscripts in general. Contrary to what the common translation of shufa 書法 as ‘calligraphy’ may suggest, Chapter Seven does not only deal with the artistic quality of the brushwork, although in all three articles it is appreciated. In addition, it provides descriptions of various writing styles found in the manuscripts, thereby implying criteria for an assessment of the corpus as a whole. Correcting earlier classifications into ‘seal script’ (zhuanshu 篆書) and ‘clerical script’ (lishu 隸書) being the two main types, in the first article Chen assigns all specimens to ‘clerical script’ and distinguishes ‘seal clerical’ (zhuanli 篆隸, such as the medical recipes in Wushier bingfang 五十二病方), ‘old clerical’ (guli 古隸, such as Laozi A or Yinyang wuxing B) and ‘Han clerical’ (Hanli 漢隸, such as Laozi B or Zhouyi) (pp. 315–325).[19] Although these terms still imply a development resulting in a unified and orderly appearance of the script, it is evident from the author’s own work that at least the latter two styles including cursive forms were both (still) in use during the lifetime of the owner of tomb MWD 3. Dating scroll Xingde A (in ‘old clerical’) to 196 BCE and scroll Xingde B (in ‘Han clerical’) to approximately 193 BCE, he speculates that both were written for Li Cang 利蒼 (d. 186 BCE), Chancellor of the Kingdom of Changsha 長沙國相, since 193 BCE Marquis of Dai 軑侯, and owner of tomb MWD 2, but produced for different purposes, and inherited by his son, owner of tomb MWD 3 (p. 116). The choice of script must have had other reasons.

After characterisation of styles in a general manner, the second article focuses on individual characters and provides many examples on which readers may train their eyes. Against traditional genealogies of the script styles, the silk manuscripts demonstrate that the ‘clerical’ style was already fully developed in early Han (pp. 338–339). It is obvious that these observations from the silk manuscripts are important for the study of other written artefacts from late pre-imperial and early imperial times as well, just as the evidence from the bamboo manuscripts, including the inventory, is relevant for the study of the silk ones. The classical distinction of ‘book-hand’ and ‘cursive writing’ may be more helpful here than general developments.[20] The last chapter is devoted to a single manuscript, titled Yinyang wuxing A, belonging to Chen’s first type of script. This should be read in conjunction with the second article in Chapter Five, which also discusses palaeographic issues, now under the new title Shifa 式法.[21]

A comprehensive overview of the codicology of the silk manuscripts and some scribal practices is provided in Chapter One.[22] Ten silk sheets found in a bamboo basket in MWD 3 had a width of about 50 cm and a length of about 100 cm, with the highest density of 100 warp threads by 70 weft threads. 50 cm is considered to be in the range of a ‘full panel’ (zhengfu 整幅, two feet two inches), considering that the width of 48 cm found in many of the silk manuscripts was caused by shrinking due to their having been soaked in water for centuries. According to chemical analysis, a colloid similar to today’s peach resin was used for sizing, then a hammer for polishing. Ruling of columns and upper and lower borders of the text field was done in red (zhusilan 朱絲欄) or black (wusilan 烏絲欄), but apparently there has not been a scientific study yet, as the author discusses various possibilities of achieving straight lines, including weaving in dyed threads. In most cases, the layout was carefully planned ahead of writing and drawing. If there were diagrams to be added, the space for them was marked first, then the ruling for the text was done. While in the beginning usually one column was left blank, the situation at the end differs. In most cases, there is some blank space left, perhaps meant to protect the scroll when rolled up. A special case is the scroll containing Xingde C. This manuscript is exceptional in more ways: Its content units (diagrams and texts) are separated by thick dark-brown frames which were woven into the silk panel, and written and drawn in red, with dark-brown ruling lines woven into the silk panel too (pp. 3–4, also see the image on p. 88). The author refrains from speculating about the reasons for this opulent outfit.

It would be futile to assess the volume by what it does not contain: an index, cross-references, updates and discussion of recent works, especially the 2014 editions. One may miss a treatment of the bamboo manuscripts or the maps and banners as well as of the chest in its archaeological context and its relationship to the tomb owner. However, it must not be forgotten that Chen Songchang was the first scholar to recognise the importance of codicology for the study of the MWD silk manuscripts, a field that simply did not exist before. In addition, his contributions mainly deal with those manuscripts deemed less important by historians and philologists, namely those containing various divination methods. Together with the medical texts, these make up the overwhelming majority of the corpus and are probably much closer to the realities of the times than the ‘philosophical’ texts modern scholars prefer to interpret.[23]

Studies of ancient Chinese ‘literary’ manuscripts should take into account the evidence from legal and administrative documents as well, as these have apparently shaped the appearance of the former. Adding experience with the Qin manuscripts held by the Yuelu Academy and the Western Han documents from Zoumalou 走馬樓 to his familiarity with the MWD corpus, it is to be hoped that Professor Chen will provide a comprehensive synthesis of his work in the future.

The Commercial Press has produced a fine volume with traditional characters and many colour images in high quality, some larger ones even on foldouts, fully justifying its price and the re-publication of articles first having appeared thirty years ago. It does not only belong in every sinological library but on the desk of everyone studying ancient Chinese manuscripts. The chapter on codicology deserves a full and updated translation into English.

---

[1] For a summary of previous research on the MWD site in Chinese see Chen Jianming 陳建明 (ed.) (2013), Mawangdui Hanmu yanjiu 馬王堆漢墓研究, Changsha: Yuelu shushe 嶽麓書社; approximately two thirds of the volume are devoted to the silk manuscripts.

[2] For some of the problems see Shan Yingwen 單穎文 (2014), ‘Zai tan “dixia tushuguan”—Mawangdui boshu zhengli sishi zai huimou’ 再探“地下圖書館”——馬王堆帛書整理四十載回眸, Zhongguo chuban jituan gongsi 中國出版集團公司 30 October 2014, http://www.cnpubg.com/book/2014/1030/22544.shtml, accessed 12 March 2023; originally from Wenhui bao 文匯報.

[3] For details see the list in Changsha Mawangdui Hanmu jianbo chutu yu zhengli qingkuang huigu 長沙馬王堆漢墓簡帛出土與整理情況回顧 (hereafter Huigu), in Qiu Xigui 裘錫圭 (ed.) (2014), Changsha Mawangdui Hanmu jianbo jicheng 馬王堆漢墓簡帛集成 (hereafter Jicheng), 7 vols, vol. 1, front matter with separate pagination, pp. 1–8, here pp. 4–6.

[4] Huigu, pp. 6–8; see also Shan Yingwen 2014 and the annals of ‘memorabilia’ in Yu Yanjiao 喻燕姣 and Wang Hui 王卉, Hunan sheng bowuguan 湖南省博物館 (ed.) (2016), Jinian Mawangdui Hanmu fajue sishizhounian guoji xueshu yantaohui lunwenji 紀念馬王堆漢墓發掘四十周年國際學術研討會論文集 (hereafter Lunwen ji), Changsha: Yuelu shushe 嶽麓書社, pp. 708–741.

[5] Published by Wenwu chubanshe 文物出版社 in Beijing.

[6] Fu Juyou 傅舉有 and Chen Songchang 陳松長 (eds) (1992), Mawangdu Hanmu wenwu 馬王堆漢墓文物, Changsha: Hunan chubanshe 湖南出版社.

[7] See Jicheng, vol. 6, p. 66 for the Yinyang wuxing A edition where the editors slipped in a note saying that according to recent insights the reconstruction of the manuscript and thus the textual order was mistaken but could not be changed anymore, see also Chen Songchang 2021, pp. 357–358, nos. [125]–[130], 396, nos. [213–216] for some of the contributions to the conference; in addition, the colours do not always correspond to those of the original objects.

[8] See Lunwen ji, pp. 133–514.

[9] Mawangdui Hanmu yishu 馬王堆帛書藝術, Shanghai: Shanghai shudian 上海書店.

[10] Huigu, pp. 7–8.

[11] Chen Songchang (2008), Jianbo yanjiu wengao 簡帛研究文稿 (hereafter: Wengao), Changsha: Xianzhuang shuju 綫裝書店. It includes a bibliography on research on the MWD corpus originally published in 1998 and updated through 2007 (pp. 524–572).

[12] This is actually the brief introduction to the silk manuscripts in the excavation report of 2004 (pp. 87–91).

[13] The article was written at a time when politicians and some historians were trying to ascertain the reliability of traditional sources against critical scholarship dating back at least to the Song.

[14] In Chapter One, he mentions in passing that these two scrolls were actually parts of one and the same scroll (pp. 7-–9), because the empty column at the end of the lower part of the Zhouyi scroll tallied that in the upper part of the Xici scroll. Closer inspection shows that this is highly improbable: 1) on other scrolls of this group, texts are never separated by a full blank column; 2) the marginal ruling on the Zhouyi scroll is executed in red with the same fine lines as those drawn for the columns, while the Xici scroll has thick black lines instead; 3) the hands and the number of characters per column differ, see the images in Jicheng, vol. 1 and vol. 3, pp. 6-–10.

[15] For the concept of multiple-text manuscript see Alessandro Bausi (2010), ‘A Case for Multiple Text Manuscripts being “corpus organizers”’, manuscript cultures no. 3, pp. 34-–36 and Michael Friedrich and Cosima Schwarke (2016), ‘Introduction – Manuscripts as Evolving Entities’, in M. Friedrich and C. Schwarke (eds), One-Volume Libraries: Composite and Multiple-Text Manuscripts , pp. 1-–26.

[16] Even before the 2008 volume, these contributions had already been re-published in Chen Songchang (2001), Mawangdui boshu “Xingde” yanjiu lungao 馬王堆帛書《刑德》研究論稿, Taipei: Taiwan guji chuban youxian gongsi 臺灣古籍出版有限公司.

[17] This was not accepted by the editors of Jicheng, see n. 8.

[18] Lunwen ji, pp. 410–415.

[19] For this classification, Chen follows Qiu Xigui 裘錫圭 (1974), ‘Cong Mawangdui yihao Hanmu “qiance” tan guanyu guli de yixie wenti’ 從馬王堆一號漢墓“遣冊”談關於古隸的一些問題, Kaogu 考古 no. 1, pp. 46–55.

[20] See William G. Boltz (2012/2013), ‘Hand-writing Styles in Early Chinese Manuscripts’, manuscript cultures no. 5, pp. 11–19; the cursive elements were already noted on the inventory slips from MWD 1 (which was sealed after MWD 3) in Qiu Xigui 1974.

[21] This title was rejected by the editors of the 2014 editions, who decided, for clarity’s sake, to follow the established titles in spite of knowing that some definitely were wrong, see Jicheng, vol. 5, p. 66. They also raised the question of whether the two parts of Yinyang A originally were two independent panels and not one scroll, ibid.

[22] This sub-chapter is based on Chen Songchang (2007), ‘Mawangdui boshu de chaoben tezheng’ 馬王堆帛書的抄本特徵, Hunnan daxue xuebao (shehuikexue ban) 湖南大學學報（社會科學版） no. 9, pp. 20–29, but partially re-organised and expanded.

[23] According to the bibliography, the number of monographs devoted to one group of texts is headed by studies on the Laozi (28), closely followed by those on the medical texts (25) and the Zhouyi and its commentaries (23); for articles, the list is dominated by studies of medical manuscripts (559), followed by those on the Zhouyi (388) and on the Yellow Emperor texts (231), works on the Laozi are much less (156) and all others number well below one hundred.

---

---