by An Liu
Early commentaries attest to the popularity of the work among Han intellectuals, including Ma Rong (77–166); Lu Zhi (d. 192); Xu Shen (58–148?), who wrote the etymological dictionary Shuowen jiezi ; and Gao You (fl. 160–220), who also wrote a commentary on the Lüshi chunqiu. The latter two commentaries are the only ones that survived and only in fragmentary form. The original title of the Xu recension was Huainan honglie jiangu (The Vast and Luminous [Book] of Huainan, with Inserted Explanations), suggesting that it contained commentaries interspersed throughout the text, either within or above the lines on the page.3 The original title of the Gao recension was Huainan honglie jiejing (The Vast and Luminous [Book] of Huainan, with Classical Explanations), indicating that Gao had explained the meaning of the text using ideas from the Confucian classics.4 In addition, each chapter title of the Gao recension was followed by the character xun (to explicate), as in, for example, , “Originating in the Way, Explicated.” With this, Gao was paying homage to his teacher Lu Zhi, from whom, he says, he received the explications of the text.5 This character is not, as some have thought, part of the chapter titles in Liu An’s original work or those of the other recensions of the text.
The two independent commentarial traditions (each with its own recension of the text and commentary) were conflated at some point, perhaps as early as the fourth century C.E. or even earlier.6 To the best of our knowledge, by the end of the Northern Song dynasty, all editions contained this conflated commentary and recension of the text. Thirteen chapters were from the Gao recension (which itself contained some comments by Xu), and eight chapters were from the Xu recension. In all the extant editions, the Gao commentary is found in chapters 1–9, 13, 16, 17, and 19, and the Xu commentary is now found in chapters 10–12, 14, 15, 18, 20, and 21. In some editions, seven of the longest chapters are divided in half (chapters 1–5, 9, and 13), making a redaction of twenty-eight rather than twenty-one chapters. All these are Gao recension chapters, whose additional length is partly caused by the comparative length and frequency of the Gao commentary. Some extant editions still follow this division, as discussed later. Only a few editions survive from these earlier periods (only one complete and one fragmentary). But by the early Ming dynasty, with its flourishing printing trade, the number of editions of the Huainanzi proliferated. Accordingly, the textual history of the work now is the history and filiation of its many editions.
Main Editions of the Text
The more than eighty-seven complete editions of the Huainanzi can be organized into six distinct lineages, with an “ancestral redaction” at the head of each one.7 These redactions, the oldest extant editions in each of these lineages, are the following.
The Northern Song Redaction of 1050
The Northern Song redaction is a twenty-one-chapter edition that was originally printed around 1050. It remained in private collections and thus had little or no influence on the textual transmission of the Huainanzi until the nineteenth century, during which several traced facsimile copies were made. Unfortunately, in the copying, random errors were introduced into the text that were not in the original.8 One of these copies, made in 1872 by Liu Maosheng , was the edition reprinted in the Sibu congkan , and it is known to us today in this edition and several facsimiles. The original exemplar was in the rare book collection of the Library of the Southern Manchuria Railroad Company when the Russian army entered Dalian at the close of World War II in 1945 and has not been seen since then. The Northern Song redaction is one of the most important editions of the text because it preserves so many early and accurate readings, yet it is not without errors.
The Daozang Redaction of 1445
The very long history of the inclusion of the Huainanzi among the canonical works of the Daozang (Daoist Patrology) goes back to the Southern Song recension of 1121, the Northern Song recension of 1019, and possibly earlier.9 In 1923, the Hanfenlou and Commercial Press in Shanghai began a three-year project to make a photolithographic reproduction of the one extant complete exemplar of the Daozang recension of 1445, the one that is still preserved in the White Cloud Temple in Beijing. A number of exemplars of the 1445 Daozang were given to Daoist monasteries and temples throughout the land, but their circulation among non-Daoist literati was limited. Thus this twenty-eight-chapter recension of the Huainanzi had only limited influence during the Ming dynasty and even less influence during the Qing. Four editions based completely on this Daozang Huainanzi were published during the Ming and only one during the Qing.10 Because of the Daozang Huainanzi redaction’s superior readings, probably attributable to its Song or even Tang ancestry, this is one of the most reliable of the ancestral redactions. Nonetheless, it too is not without errors.
The Liu Ji Redaction of 1501
The Liu Ji redaction in twenty-eight chapters, published in 1501, is in some ways the most interesting edition of the Huainanzi. Liu, a scholar and official from Jiangxia in Hubei Province who passed the metropolitan exam in 1490, was a brilliant textual scholar who was interested in Chinese science and produced a new edition of the Guanzi.11 The significance of his Huainanzi redaction comes in large part from his subcommentary in which he discussed both the meaning of passages and the text-critical decisions he was making. His redaction has caused considerable scholarly controversy about its provenance.12 Liu mentions that he used three earlier editions to make his new one: an “old edition,” an “other edition,” and a “different edition.” The first shows affinities with the 1445 Daozang redaction but probably was closer to the redaction in the Daoist Patrology recension of 1019. The second is distinctly different from both the Northern Song and Daozang redactions but otherwise cannot be identified. The third edition varies significantly from the other two editions and shares meanings with the quotations from the Xu recension of the Huainanzi preserved in the encyclopedia Taiping yulan (published in 983); it thus appears to be quite old.13 This redaction has had a regrettably minor influence on later editions, with only seven descendants, none of which is later than 1670. It now is available only on microfilm from an exemplar in the National Central Library in Taibei. Its lack of popularity may well be due to the sheer size and complexity of the Liu subcommentary.
The Zhongli siziji Redaction of 1579
The Zhongli siziji, a twenty-eight-chapter redaction of the Huainanzi, is one of four works published in this collection, along with the Laozi, Zhuangzi, and Guanzi. This collection was compiled after a change in districting under the first Ming emperor in 1373, in which the ancestral homes of these four famous philosophers were included in the same prefecture of Anhui Province, whose capital city had briefly been named Zhongli at that time. Although construction of a Ming “central capital” was begun there in 1370, it was abandoned five years later, and its name was changed to Fengyang in 1375. The supervising editor for the project was a scholar and official named Zhu Dongguang (fl. ca. 1540–1585). Careful research has demonstrated that this redaction, which has no descendants, was created by conflating readings in two earlier editions: the Wang Ying edition of 1536 from the Liu Ji lineage and the Anzheng tang edition of 1533 from the Daozang lineage.14 Because it is completely derived from two earlier ancestral redactions, it does not have to be consulted when creating a modern critical edition.
The Mao Yigui Redaction of 1580
The Mao Yigui redaction was extremely prolific, with thirty descendants, and is thus one of the most influential editions of the Huainanzi.15 Its descendants include the Mao Kun edition of 1590, the beautifully executed Wang Yiluan edition of 1591, and the Siku quanshu edition of 1781. It is the ultimate source of all the extant twenty-one-chapter editions of the text that are not reprints of the Northern Song redaction, which means most of them. Its own provenance is rather complicated. Mao (fl. ca. 1555–1615) was the nephew of the famous statesman and philosopher Mao Kun (1512–1601); both hailed from Gui’an in northern Zhejiang Province. Mao Yigui passed the provincial exams in 1588 but failed to rise any higher and held a succession of minor official posts in far-off d
istricts during his career. Along with his senior collaborator, Wen Bo from Wucheng in Zhejiang Province, Mao established an edition with a unique arrangement of text and abridged commentary that was a conflation of three earlier editions.16 One of these was the Daozang redaction, and another was the Wang Ying edition of the Liu Ji redaction, both of which have twenty-eight chapters. The identity of the third edition is significant because it must have had the twenty-one-chapter arrangement that Mao adopted and must have been his basic text. There is reason to believe that this third edition was the actual exemplar of the Northern Song redaction that was transmitted into the twentieth century, which was likely in a private collection in Wucheng at that time.17 If it was not, then it must be from an edition closely related to the Northern Song redaction, given its distinct textual variants. Because the Mao Yigui redaction is likely to have been derived completely from older extant ancestral redactions, its independent testimony to the text of the Huainanzi is minimal.
The Zhuang Kuiji Redaction of 1788
The twenty-one-chapter Zhuang Kuiji redaction greatly influenced twentieth-century editions and studies of the Huainanzi because of its relatively late creation during the Qing period and its eighteen descendants. Among these were several widely circulated collections and editions, including the Zhejiang Publishing Company’s 1876 Ershierzi edition, the 1923 critical edition of Liu Wendian , and the 1935 steel-movable-type edition in the collection Sibu beiyao . The editor of the Zhuang Kuiji redaction, Zhuang Kuiji (1760–1813), was a young scholar who was born into an important family in the district of Wujin in Jiangsu Province. He was a student of the famous textual scholar Bi Yuan (1730–1797) and served under him in the Shenxi provincial capital of Xi’an when Bi was governor.18 While governor, Bi assembled in his office some of the luminaries of the Han Learning textual movement, and Zhuang undoubtedly benefited from his position as a junior scholar among them.19 In addition to establishing a new edition of the text, Zhuang wrote a subcommentary in which he cited variant readings, gave textual emendations, and explained some difficult passages in the text.20
The provenance of Zhuang’s edition is complex, but it appears to have been based directly on a hand-collated exemplar of the Mao Kun edition of 1590 produced by his mentor, the Han Learning scholar Qian Dian (1744–1806). Qian added emendations from a number of other Ming editions: the Mao Yigui redaction, the Zhongli redaction, the Yejinshan edition of 1582 (a poor-quality reprint of the Daozang redaction), and perhaps the Daozang redaction itself. Using this method, Qian supplemented the abridged commentary in the Mao lineage until it was almost complete.21 Unfortunately, Zhuang’s redaction, although influential, is, in the last analysis, rather poor compared with the early Ming editions. It is based on a mediocre edition that contains textual errors and adds many hidden emendations and conflations.22 Since Zhuang’s redaction seems wholly derived from still extant editions and since it introduces many new errors into them, the possibility that it contains any accurate readings not in the other ancestral redactions is almost nonexistent. Thus despite its wide circulation, it need not be consulted to make a modern critical edition of the Huainanzi.
Because the Zhongli and the extremely prolific Mao Yigui and Zhuang Kuiji redactions and their descendants are derived from other extant redactions, none of these more than fifty editions are likely to contain possibly accurate variant readings not already present in the Northern Song, Daozang, and Liu Ji redactions. Thus only these three need be consulted for a new critical edition.
Modern Critical Editions
In addition to using the three oldest ancestral redactions of the Huainanzi, a modern critical edition should include the emendations of the major textual scholars of the nineteenth and twentieth centuries.23 Using a variety of scholarly techniques but often working on the inferior later editions of the Huainanzi —those in the Mao Yigui and Zhuang Kuiji lineages—Qing and Republican scholars nonetheless provided invaluable emendations that go well beyond corrections based on collations with the superior editions that were rare during their time but that are now well known, such as the Northern Song and Daozang redactions. The first published attempts to collect nineteenth-century textual critics of the Huainanzi were made in Japan in 1914 by Hattori Unokichi . Less than a decade later, more complete critical editions were compiled almost simultaneously in China by two scholars, Liu Wendian and Liu Jiali . The Hattori edition was based on the 1798 Kubo Chikusui edition from the Mao Yigui lineage, and it was combined with the Daozang redaction and one of its later editions in the Daozang jiyao collection of 1796.24 The Hattori edition includes the textual notes of Tokugawa scholars from the Kubo edition but not the Kubo edition’s detailed comparisons with other editions. The textual emendations of the major Qing critics Wang Niansun , and Yu Yue are included in this edition.
Liu Wendian’s Huainan honglie jijie was completed in 1921 and published in 1923. It includes textual emendations by twenty-two of the most important Qing-dynasty textual critics, including Wang Niansun and his son Wang Yinzhi , Gu Guangqi , and Yu Yue.25 Liu Jiali’s Huainan jizheng was completed in 1921 and published in 1924 and seems to have been written independently. It cites sixteen Qing critics, many of the most important of whom are also found in Liu Wendian’s work.26 Both critical editions were likely based on the Zhejiang Publishing Company’s edition of the Zhuang Kuiji redaction published in the Ershierzi (Twenty-two Philosophers) collection in 1876. Because Liu Wendian’s work included more Qing textual scholars, it became much more widely circulated than Liu Jiali’s, and it has been reprinted a number of times in different collections. It has also been quite popular with Western translators, serving as the basic text for the translations by Evan Morgan, Roger Ames, John Major, Claude Larre, Isabelle Robinet, and most other modern Western scholars, including Charles Le Blanc, Rémi Mathieu, and colleagues. Unfortunately, the original Zhuang Kuiji redaction had many textual errors, only some of which were corrected in later editions. Despite the convenience of its inclusion of so many Qing textual critics, since it was based on a derivative and flawed edition, there are better editions to use as the basis for translation.
The political and social upheavals in China and Japan undoubtedly influenced the fact that no new critical editions of note were published for half a century after the works of Liu Wendian and Liu Jiali. In 1974, Tokyo University professors Togawa Yoshiro , Kiyama Hideo , and Sawaya Harutsugu produced a critical edition derived from the Northern Song and Daozang redactions and the Liu Wendian edition. Major Qing textual scholars from Liu are included. Waseda University professor Kusuyama Haruki compiled a three-volume critical edition published between 1979 and 1988. It is based on Liu Wendian’s edition with emendations from the Northern Song and Daozang redactions and Liu’s assembled textual scholars. Both editions contain modern Japanese translations.27
Three modern critical editions were published in China during the next decade:
1. Chen Yiping’s Huainanzi jiao zhu yi , published in 1994, is based on Liu Wendian’s edition and collated with nine other editions, including the Northern Song redaction. Chen also includes text critical comments by twenty-two Qing, Republican, and modern scholars, including Wang Niansun, Yu Yue, and Tao Fangqi.
2. Zhang Shuangdi’s Huainanzi jiaoshi was published in 1997. Zhang’s base edition is the Daozang redaction, and it is collated with twelve other editions, including all the other ancestral redactions and the Liu Wendian critical edition. Zhang also made emendations drawing on the work of eighty-four textual scholars, including Wang Shumin , the dean of twentieth-century textual critics, and his younger followers Zheng Liangshu and Yu Dacheng .
3. He Ning’s Huainanzi jishi , published in 1998, uses the Zhejiang reprint of the Zhuang Kuiji redaction as the base text and corrects its readings by comparing them with the Northern Song, Daozang, Zhongli, and Mao Yigui redactions. He made further emendations based on a long list of Ming, Qing, and twentieth-century scholars given in a very useful bibliographical appendix. He Ning als
o includes the same major modern scholars as Zhang Shuangdi does.
Of the three Chinese editions, Zhang’s is the most reliable because it is based on the high-quality Daozang redaction and presents its material in a clear and reader-friendly format. He’s and Zhang’s critical editions are superb because of the breadth of the editions cited and the wide variety of textual scholars included. Along with Kusuyama’s work, these were the editions we most often consulted for our translation when we had difficulty understanding our basic edition.
The modern critical edition we used as the basis of our translation was D. C. Lau’s Huainanzi zhuzi suoyin of 1992, included in the Chinese University of Hong Kong’s Institute of Chinese Studies Ancient Chinese Text Concordance Series. This Lau edition has all the basic elements of a modern critical text: it is based on the Sibu congkan edition of the Northern Song redaction but is collated with the other major ancestral redactions, the Daozang and the Liu Ji. Lau also consulted all the major nineteenth- and twentieth-century textual critics and a wide variety of other works. Added to that, of course, is the fact that Lau and his research team created a concordance to every character in the text—133,827, to be exact. The main drawbacks to Lau’s work are the absence of the Gao and Xu commentaries and its sometimes questionable punctuation and sectioning. But this invaluable work will be the standard for the foreseeable future.
The following bibliography includes all translations of the Huainanzi into Western languages and many of the recent modern Chinese translations. These are followed by a list of the modern critical editions just discussed. Next comes a bibliography of publications on the philosophy of the Huainanzi, whose number has increased dramatically in the past several decades as more scholars have turned their attention to this too-long-overlooked masterpiece. The bibliography concludes with a list of the major pre-1960 textual and historical studies of the Huainanzi and a bibliography of such works since 1960. Studies published in the past fifty years are likely to be the most accessible for scholars wishing to pursue further research. These bibliographies are fairly complete but by no means exhaustive. For a full bibliography of textual and historical works as of 1990, see Harold D. Roth, The Textual History of the Huai-nan Tzu.