Hostname: page-component-586b7cd67f-dsjbd Total loading time: 0 Render date: 2024-11-23T19:23:33.374Z Has data issue: false hasContentIssue false
  • English
  • Français

Authentication Studies (辨偽學) Methodology and the Polymorphous Text Paradigm

Published online by Cambridge University Press:  08 August 2014

Translated by
Paul Fischer 方破
Get access Rights & Permissions [Opens in a new window]

Abstract

The foundation of Chinese intellectual history is a group of texts known as “masters texts” (子書). Many masters texts were authored in the Han dynasty or earlier and many of these have as their title the name of a master who was generally regarded as the author. The inclination to treat a given book as the product of a single writer is apparently a strong one. Nevertheless, from the very beginning there were Chinese scholars who doubted the veracity of the putative authorship of some of these works and suggested that they may in fact have been the product of several authors. Over time, such scholars developed criteria by which to judge the authenticity of ancient masters texts. But as such textual criticism grew more penetrating, the object of its scrutiny began to come apart at the seams. In the last two decades there has been a growing consensus that most early Chinese masters texts were originally quite permeable and that only later were their received forms settled upon.

The branch of textual criticism that deals with authenticating early Chinese texts is called “Authentication studies.” This paper is a survey of the methodological advances made in the field of Authentication studies over the last two millennia. It is not a history of the field, as such a history would be a much longer project. The survey concludes with the idea of the “polymorphous text paradigm,” a paradigm that paradoxically obviates much of the preceding scholarship in its own field. Simply put, if authentication relies largely on anachronism, and anachronism relies largely on the dates of the putative author, then a multi-author work with no known “last author” will be impossible to authenticate. Furthermore, the polymorphous text paradigm does not posit these texts as necessarily having earlier and later “layers,” but rather as having had no set structure over the course of their early redactional evolution.

This survey examines the contributions of seventeen scholars to Authentication studies methodology, and concludes with how the changes in this field have influenced the work of three modern, Western scholars.

諸子書是研究中國思想史的基礎。其中許多書成於漢代及其前代的， 而這些書往往以被視為書的作者之名為書名。人們也傾向將某一部子 書視為單一作者所著。儘管如此，很早就有中國學者對上述推定作者 的觀點表示懷疑，而他們就提出多數子書其實是由許多作者所共同完 成的。隨著時間過去，這些持懷疑論的學者發展並形成了一套判斷古 書真偽的標準。但是，這種文獻學的驗證越是敏銳，它所檢視對象的 可信度也就開始分崩離析。在最近二十年，有一種日益為人接受的看 法，即認為早期中國的子書原具有較大的滲透性，到後來才漸漸形成 固定的文本形式。

文獻學中處理古書真偽問題的學科稱之為 “辨偽學”。本文檢視兩千 年來辨偽學方法的進展，用意不是要計劃寫一部辨偽學史，後者需要 更長遠的研究才能完成。文章以 “多形文獻範式” 作為總結，這種範式 似是而非推翻了存在其前的辨偽論點。簡而言之，如果判定文獻真偽 的主要根據是時代的錯置，而著作時代錯置的推論又依據於所假定的 作者所處的年代，則一部多人創作而無法確知其 “最後作者” 的著作將 無從判定其真偽。此外，本文所舉出的 “多形文獻範式” 並不假定這些 文獻必有先後層次，而是認定其在早期傳抄過程中有無固定結構的樣 貌。

本文評述了歷來十七位學者對辨偽學方法的貢獻，最後也提及此一 領域的變遷如何影響了三位現代西方學者的研究工作。

Type
Articles
Information
Early China , Volume 32 , 2009 , pp. 1 - 43
Copyright
Copyright © Society for the Study of Early China 2009

Access options

Get access to the full version of this content by using one of the access options below. (Log in options will check for institutional or personal access. Content may require purchase if you do not have access.)

References

1. Bianwei xue 辨偽學 is often rendered as “forgery studies” and conceived as dealing with “forged texts,” but this translation is too circumscribed for the variety of inauthentic texts discussed in the works of Authentication studies. This will become clear below.

2. Shaughnessy, Edward, Rewriting Early Chinese Texts (Albany: State University of New York Press, 2006), 260–61Google Scholar.

3. Probably the most famous examples of this are the various attempts to categorize the multiplicitous “thought” of the Zhuangzi 莊子; cf. Feng, Guan 關鋒, Zhuangzi zhexue taolun ji 莊子哲學討論集 (Beijing: Zhonghua, 1961)Google Scholar; Graham, A.C., “How Much of Chuang Tzu did Chuang Tzu Write?,” in Studies in Classical Chinese Thought, ed. Rosemont, Henry and Schwartz, Benjamin I., Journal of the American Academy of Religion Thematic Studies (Chico: CA, 1979), 459–501Google Scholar; Xiaogan, Liu 劉笑敢, Classifying the Zhuangzi Chapters (Ann Arbor: Center for Chinese Studies, The University of Michigan, 1994)Google Scholar.

4. Harold Love distinguishes four kinds of authorship: precursory (a source upon which a later author draws), executive (the usual sense of the term “author”; whom Love calls “the deviser”), declarative (a “validator” or “sponsor”; this describes the presumed authorship of many early Chinese Masters texts), and revisionary. Love notes that “executive” and “revisionary” are not as categorically distinct as they might appear: “With some misgivings I offer the term revisionary authorship … the problem being that revision is not always clearly distinguishable from primary composition.” See Love, Harold, Attributing Authorship: An Introduction (Cambridge: Cambridge University Press, 2002), 46CrossRefGoogle Scholar.

5. Charles Gardner, remarking on the authorship of early Chinese texts, said: “Few writers of the pre-Han epoch (prior to the second century B.C.E.) are really known to us. Even when a name has been preserved, this is an illusory label, which fails to inform us seriously concerning the personality and motives of the author or of his qualifications for authorship. It is then almost a matter of indifference whether we do or do not attach a personal name to an early text.” See Gardner, Charles, Chinese Traditional Historiography (Cambridge: Harvard University Press, 1938; rpt. 1961), 23Google Scholar.

6. These categories of arguments for inauthenticity are my own, and therefore are not accompanied by technical terms in Chinese. Categorization schemes devised by other scholars are discussed later in the paper.

7. An “extrinsic, incongruous” text would simply be anachronistic.

8. The argument from precedence is usually, but not always, an argument from anachronism. I separate them because it is possible for two authors to be contemporaneous yet for a later scholar to insist on the possibility of discerning who came up with the idea or phrase first.

9. “Plagiarism” has strong negative connotations that are almost certainly inappropriate to the time when these texts were written. However, I nevertheless use “plagiarism” rather than “borrowing” because most of the scholars in this survey would have intended some negative connotations, as their writings below will make clear.

10. Huainanzi 漢淮字 (Sibu beiyao 四部備要 ed.), 19.11b. Doubt in the veracity of a text appears even earlier: the Mengzi 孟子 “Jin xin xia 盡心下” chapter quotes Meng Ke 孟軻 (390–305) as saying: “It would be better to be without the Documents than to entirely believe the Documents. (For example), in the ‘Martial Completion’ (chapter) I use two or three passages only. The good man has no enemies in the world; (therefore) when the best invaded the worst, how could it (say) ‘the (spilt) blood floated pestles’?” 盡信《書》, ，則不如無《書》 。 吾於 ‘武成’ 取二三策而已矣 。 仁人無敵於天下；以至 仁伐至不仁，而何其 ‘血之流杵’ 也. See Mengzi (Sibu beiyao ed.), 14.2a, or Lau, D.C., trans., Mencius (New York: Penguin, 1970), 194Google Scholar. Meng Ke's doubt, however, seems to have sprung more from a desire to rewrite history to fit his didactic enterprise than from detecting philological inconsistencies.

11. Han shu 漢書 (Beijing: Zhonghua, 1962), 30.1731Google ScholarPubMed. See also Xiang's, Liu notes for the Kong Jia pan yu 孔甲盤盂 at 30.1740Google Scholar, Shen Nong 神農 at 30. 1742Google ScholarPubMed, and Yi Yin shuo 伊 尹說, Shi Kuang 師曠, Tian yi 天乙, and Huang Di shuo 黃帝說, all at 30.1744. The “time of the six states” is c. 250 B.C.E., while Huang Di is said to have lived c. 2500 B.C.E. Ban Gu 班固 (32–92) is the author of the Han shu, however the content of the “Yiwen zhi” 藝文志—the imperial library catalog—within it derives almost entirely from the Bie lu 別錄 (8 B.C.E.) by Liu Xiang 劉向 (79–8 B.C.E.), written a century earlier. Liu Xiang made this collection of abstracts during his twenty-year tenure as imperial librarian, although he died before finishing it. His son Liu Xin 劉歆 (45–23 C.E.) took over his official post after his death, and two years later presented his Qi lüe 七略 (6 B.C.E.). The Qi lüe is largely an abridgement of the Bie lu, and the “Yiwen zhi” is an abridgement of that abridgement. Because the Bie lu and Qi lüe are both lost, it is impossible to know who wrote precisely what in the extant “Yiwen zhi,” but because Liu Xiang worked on the catalog for twenty years, while Liu Xin worked on it for only two, and Ban Gu did little more than edit it, I generally give credit for the contents to Liu Xiang. At the end of his preface to the “Yiwen zhi” (p. 1701), Ban Gu says of Liu Xin's Qi lüe: 今刪其 要，以備篇籍。 “Now (I have) selected its essentials in order to complete (the entries for) text-length and text-name.” For more on the origin and structure of this catalog, see Luming, Gao 高路明, Guji mulu yu Zhongguo gudai xueshu yanjiu 古籍目錄與中國 古代學術研究 (Nanjing: Jiangsu guji, 1997), 72–77Google Scholar.

12. Anonymous Buddhist sutras were routinely ascribed to the historical Buddha. Therefore, when, for example, Dao'an 道安 (312–385) made his Zongli zhongjing mulu 綜 理眾經目錄 and Daoxuan 道宣 (596–667) his Fojing mulu 佛經目錄, they felt compelled to set such texts aside in an “Yijing lu 疑經錄” appendix.

13. Han shu, 30.1740, 30.1744, 30.1744, and 30.1742.

14. It might be noticed that Liu Xiang had no such criticisms for any of the texts in the Classics or Ruist categories. Or perhaps he did, but Ban Gu chose not to include them.

15. Han shu, 30.1744. It is unclear whether Liu Xiang is referring to style or content here, but it is worth noting that both of these texts are listed in the “Xiao shuo” 小說 sub-section of the Various Masters section.

16. Han shu, 30.1729 and 30.1744. Zhang Shunhui 張舜徽 suggests that the last 15 graphs of the first note are garbled, and should instead read: “或又以近世有為太公術 者所增加也.” The meaning as I have translated it, however, would remain the same. See Shunhui, Zhang 張舜徽 (1911–1992), Zhang Shunhui ji: Zhongguo wenxian xue 張舜 徽集: 中國文獻學 (1982; rpt. Wuhan: Huazhong shifan daxue, 2004), 145Google Scholar.

17. Han shu, 30.1730. Other early texts write his given name as “Yukou” 禦寇.

18. Liu Xiang, Liezixinshu mulu 列子新書目錄, in Kejun, Yan 嚴可均 (1762–1843)Google Scholar, Quan shanggu sandai Qin Han Sanguo Liuchao wen 全上古三代秦漢三國六朝文 (1808; Beijing: Zhonghua, 1958, 1999), 333 下-334 上Google Scholar.

19. Liu, Liezixinshu mulu, in Yan, , Quan shanggu sandai Qin Han Sanguo Liuchao wen, 333 下Google Scholar.

20. See Boltz, William, “Textual Criticism More Sinico,” Early China 20 (1995), 393–406CrossRefGoogle Scholar.

21. But if Liu Xiang thought parts of the text were problematic, why did he nevertheless include them? Perhaps he acted out of deference to the texts that had come down to him, even though they were clearly in radically different editions. Perhaps the actual redaction was done by his employees and was therefore not his doing and he only read and commented on the final edition. If he knowingly put sections that he perceived to be from different authors together in one edition—perhaps because he thought they belonged to a single tradition or lineage—then why did he specifically comment that some texts had “that which later generations added”? Likewise, if jia 家 should be understood in his comments on the Liezi as “lineage” rather than “author,” why would he put writings from different lineages together in one text?

22. That is, his contention that it was “obtained later.”

23. Rong, Ma, shu, Shangzhuan xu 尚書傳序, in Guohan, Ma 馬國翰 (1794–1857), Yuhan shanfang jiyi shu 玉函山房輯佚書 (c. 1855; Yangzhou: Guangling, 2004), 393Google Scholar.

24. One explanation of Ma's description is simply that the received text chapter of the “Tai shi” is somewhat inferior in style or content, implying that it was written and added later, and that later still it lost some of its content. But most scholars think he is saying that the “Tai shi” chapter was originally extant in the pre-Qin period, was subsequently lost, and then redacted or forged by an author who was unaware of the prior quotes.

25. A text that has been proven merely to be missing five lines might give us pause to wonder at the logic of thereby describing it as inauthentic. I presume that Ma was struck by the apparent implausibility of a text losing so many lines that just happened to be quoted elsewhere. In any case, this example was taken up by many later Authentication scholars as a powerful method of authentication criticism; cf. Zhang, , Zhongguo wenxian xue, 145–46Google Scholar.

26. Sui shu 隋書 (656; Beijing: Zhonghua, 1973), 49.1298–99Google Scholar.

27. Ma, In, Yuhan shanfang, 1511Google Scholar.

28. Lu Zhi 陸質 (d. 806), Chunqiu, ji zhuan zuanli 春秋集傳纂例 (Siku quanshu 四庫 全書 ed.), 146.386Google Scholar.

29. It might be objected that this “record” lies outside the scope of my inquiry into Classics and Masters texts, but as it was quite likely a preface or postface to the Zuo zhuan, I consider it within my range of concern.

30. Zongyuan, Liu 柳宗元 (773–819), Liu Zongyuan ji 柳宗元集 (Beijing: Zhonghua, 1979), 109Google Scholar.

31. Liu Zongyuan claims that Wenzi stole from Mengzi and Guanzi. By using this argument he is also implying a date for the Wenzi much later than the Han shu account (30.1729) would allow. The Han shu says Wenzi was a student of Laozi and a contemporary of Kongzi (551–479). But Mengzi (390–305) lived a century and a half later, so how could Wenzi have plagiarized him? Saying Wenzi took from Guanzi (d. 645) would not upset the traditional dating of the former, but at least two scholars prior to Liu Zongyuan had already said that the Guanzi 管子 was written later, wholly or in part. See Fu Xuan 傅玄 (217–278), quoted by Liu Shu 劉恕 (1032–1078) in his Tongjian waiji 通鑑外紀, and Kong Yingda 孔穎達 (574–648) in his Zuo zhuan zhengyi 左傳正義. This may be why Liu Zongyuan listed Master Guan after Master Meng. That is, he may have thought all three texts were roughly contemporaneous at c. 300 B.C.E.

32. By “actually belonged to the prior author,” I mean that what was “stolen” was not in fact “common property,” as with a “cultural commonplace.” As mentioned above, for the English term “plagiarism” to be justified, the author of the text who appropriated material must be passing that material off as his own, a claim that is difficult to prove with many of these early texts. I discuss these problems further in my “Intertextuality in Early Chinese Masters-Texts: Shared Narratives in Shi Zi,” Asia Major 22.2 (2009), 1–34Google Scholar.

33. Of the thirteen classics, the Er ya 爾雅 and Zhou li 周禮, both attributed to Zhou Gong 周公 (r. 1042–1036), as well as the Zuo zhuan, attributed to Zuo Qiuming, had already come under suspicion as being authored later than traditionally thought. As we saw above, the Shang shu, or at least parts of it, had also been subject to unflattering scrutiny. But in the Song, scholars greatly increased their doubts regarding the provenance of the venerable classics.

34. These early commentaries are collectively called the Shi yi 十翼: the Tuan zhuan (Shang xia) 彖傳 (上下), Xiang zhuan (Shang xia) 象傳 (上下), Xici zhuan (Shang xia) 繫 辭傳 (上下), Wen yan 文言, Shuo gua 說掛, Xu gua 序掛, and Za gua 雜掛.

35. Xiu, Ouyang 歐陽修 (1007–1072), Ouyang Xiu quanji 歐陽修全集 (Beijing: Zhongguo, 1986), 568Google Scholar. Similar descriptions of heterogeneity are on pp. 430, 563, and 571. I translated “shengren” 聖人 as “sage” in the singular, in which case it certainly refers to Kongzi; it could just as well be translated in the plural, but given the following it is clear that Ouyang is arguing against Kongzi's authorship.

36. He notes that the opening lines of the Wen yan 文言 are very similar to lines spoken by Mu Jiang 穆姜 in 564 B.C.E., the ninth reign year of her grandson Lu Xiang Gong 魯襄公 (r. 572–542), thirteen years before Kongzi was even born; cf. the Zhou Yi and Chunqiu Zuo zhuan, both in Shisan jing zhushu 十三經注疏, ed. Yuan, Ruan 阮元 (1764–1849) (1816; Beijing: Zhonghua, 1980, 2003), 15 and 1942Google Scholar.

37. Ouyang, , Ouyang Xiu quanji, 571 and 129Google Scholar.

38. See the “Yi huowen sanshou 易或問三首” section of the Jushi ji 居士集 in Ouyang, Ouyang Xiu quanji, 1.129. Kongzi's attitude toward divination has historically been a contested issue, but to Ouyang Xiu and many Authentication scholars after him, this was a clear case of inauthentication based on a description of the purported author that made his authorship impossible.

39. This is one instance of a broad, but relatively obscure, category. Sometimes the “description” will be of the text rather than the author; cf. Karlgren's seventh type, described below.

40. Most of Hu's brief descriptions can only be understood by looking at the examples he provides, therefore what I have written in this list are not translations, but are rather brief explanations based on further reading.

41. Yinglin, Hu 胡應麟 (1551–1602), Shaoshi shanfang bicong 少室山房筆叢 (1586; rpt. Shanghai: Shanghai, 2001), 290–91Google Scholar.

42. Hu, , Shaoshi, 322Google Scholar.

43. Elman, Benjamin, From Philosophy to Philology: Intellectual and Social Aspects of Change in Late Imperial China (Cambridge: Council on East Asian Studies, Harvard University, 1982; 2nd edition, Los Angeles: Asia Pacific Institute, UCLA, 2001)Google Scholar.

44. “That which cannot be added to” is “perfection.”

45. That is, no one claimed that what was said or written was their personal property.

46. The title of the first Guanzi chapter is problematic; see Rickett, W. Allyn, Guanzi: Political, Economic, and Philosophical Essays from Early China (Princeton: Princeton University Press, 1985, 1998), 1.148Google Scholar. I presume the “Di zi” 弟子 cited by Zhang is the “Dizi zhi” 弟子職 chapter in the received text. I do not know the provenance of the “Tu fang” 土方 text, but it seems to be a problematic ascription.

47. This writing is not in the extant Han Feizi.

48. That is, the “Gao Zi” and the “Wan Zhang” are writings named after people, but were not written by those people. Likewise, the Yanzi chunqiu was written by Mohists, but Yanzi himself was not a Mohist.

49. Now usually called the Huainanzi 淮南子.

50. Ye Ying 葉瑛, tongyi, Wenshijiaozhu 文史通義校注 (Beijing: Zhonghua, 1985), 170Google Scholar.

51. Shi, Su 蘇軾 (hao 號: Dongpo 東坡; 1037–1101), Dongpo quanji 東坡全集 (Shanghai: Shanghai guji, 1987), 15–16Google Scholar.

52. It appears that Zhang is being unfair to Liu here. Liu actually said that he suspected the Yanzi chunqiu to have been written by Mohists—just as Zhang thought—but that it was possible that Yanzi himself was a Mohist. Simply reading this passage by Zhang one might get the impression that Liu stated categorically that Yanzi was a Mohist. Cf. Liu, Liu Zongyuan ji, 113.

53. David Nivison explains Zhang's perception of the masters' authorship thus: “The ancient writer['s] … concern was with the preservation and development of what is true and right, not with his own originality.” Nivison, David, The Life and Thought of Chang Hsüeh-ch'eng (1738–1801) (Stanford: Stanford University Press, 1966), 129Google Scholar. Nivison implies in this sentence and elsewhere that Zhang thought all Masters texts were not written by their titular “authors.” I am not ready to credit Zhang with such a theory. It seems to me, rather, that while he thought all Chunqiu-era texts did not have single authors, for Zhanguo texts, he thought some had single authors and some did not.

54. Liang Qichao lunzhu xuancui 梁啓超論著選粹, ed. Qitai, Chen 陳其泰, Shuqing, Lu 陸樹慶, and Shu, Xu 徐蜀 (Guangzhou: Guangdong renmin, 1996), 790–93Google Scholar.

55. For the sixth category, Liang cites the Guanzi and the Shangjun shu 商君書 as containing anachronisms; for the seventh category he advises caution with the Shi ji 史 記 because it has later additions; and for the eighth category he pronounces the Liexian zhuan 列仙傳 inauthentic for mentioning Buddhist scriptures, which Liang asserts did not enter China until two hundred years later. The situation for all three are the same, although that of the Shi ji is much clearer than the others, yet Liang's advice differs, and he is particularly harsh in the third case.

56. Although for the sixth category Liang prudently says anachronism may indicate the text in question is only partially inauthentic, in the eighth and ninth categories he condemns entire texts for only one anachronism or inaccuracy. The anachronism of the example text in category eight is in the previous footnote, while the example of contradiction for category nine is the claim of two Buddhist sutras, the Niepan jing 涅 盤經 and Rulengqie jing 入楞伽經, to both be the Buddha's deathbed sayings. Anti-Buddhist prejudice may have played a role here.

57. Chen, , et. al., Liang Qichao lunzhu xuancui, 1076Google Scholar.

58. As an example of the first, he argues the Xiao jing 孝經 could not have been written by Kongzi or his student Zengzi 曾子 (505–436) because the term “jing 經 (classic)” was not used in book titles in ancient times. For literary style, he points out the Da Yu 大禹 was described in the Han shu “Yiwen zhi” as being of a later style 其文 似後世語, and that Zhao Qi 趙岐 (108–201) edited out several chapters of the Mengzi based on style, as Guo Xiang 郭象 (d. 312 C.E.) did with the Zhuangzi. For grammar, he cites Cui Shu's 崔述 (1740–1816) argument that the Zhong yong 中庸, said to be by Zisi 子思 (483–402), must come after the Mengzi based on its use of xuzi 虛字 (particles). And for rhyme, he argues the Laozi 老子 Chapter Ten must be late because it does not use an ancient rhyme scheme.

59. His examples are: the Zihuazi 子華子 must have been written in the Northern Song 北宋 (960–1127) because it often copies from Wang Anshi's 王安石 (1021–1086) Zi shuo 字說, and the Shen Pei Shi shuo 申培詩說 copies from Zhu Xi's 朱熹 Mao Shi ji zhuan 毛詩集傳.

60. His example for using this method is the use of the term “wu hou” 無厚 in the Deng Xizi 鄧析子. Both the Mozi 墨子 and the Zhuangzi had the specific definition of “without thickness” for this term, but because the Deng Xizi uses it to mean “without favor,” Liang says it is later. This important category is rather ill-served by Liang's example. He cites one line each from the Mozi (actually, he misquotes Mozi, mistaking xu 序 as hou 厚) and the Zhuangzi where the meaning of hou 厚 is clearly “thickness” and then cites the Deng Xizi where the meaning is “favor.” To prove his point, however, Liang needs to do much more than merely cite counter-examples; he would need to show that hou 厚 as “favor” was not used until a date later than the attributed authorship of the Deng Xizi. I also do not see how this term qualifies as “technical terminology.”

61. Cf. Pines, Yuri, “Lexical Changes in Zhanguo Texts,” Journal of the American Oriental Society 122.4 (2002), 691–705CrossRefGoogle Scholar.

62. As examples of inauthenticity via “thought-systems and school transmission” he argues that Kongzi could not have authored both the Lun yu and the Xici zhuan because the former is practical and the latter “mysterious” 有很深的玄學氣味. Similarly, he says the Yanzi chunqiu uses sayings of Mozi and therefore must be later than him. Likewise, the Laozi described in the Li ji 禮記 “Zengzi wen 曾子問” chapter emphasizes the importance of ritual, but the Laozi in the Laozi text denigrates it, therefore they could not have been the same person. As examples of inauthenticity from the “relation between thought and time period,” he argues that the Liezi uses Buddhist thought, which came later, and that yin-yang 陰陽 thought does not appear in early texts: therefore, its use in the Xi ci and the Wen yan commentaries to the Yi jing demonstrate that they are later. He again argues for a later date for the Laozi because it criticizes ren 仁 (goodness) and yi 義 (propriety), which traditional Ruism held were the didactic inventions of Kongzi and Mengzi respectively.

63. Liang explicitly cites Hegel's theory approvingly in Chen, , et. al., Liang Qichao lunzhu xuancui, 1044Google Scholar.

64. This doubt was arguably first planted in him by his young Peking University teacher, Hu Shi 胡適 (1891–1962), and the means for resolving the doubt was often to be found in the “double-proof method” 二重證據法 of Wang Guowei 王國維 (1877–1927), a teacher from Tsinghua University. This method proposed the mutual use of both received texts (zhishang cailiao 紙上材料) and excavated texts (dixia cailiao 地下材料). Wang first proposed this method in 1924. See Dunheng, Sun 孙敦恒, Wang Guowei nianpu xinbian 王國維年譜新編 (Beijing: Zhongguo wenshi, 1991), 146Google Scholar. Wang and Gu, however, often seemed to be pursuing opposite agendas. Wang was interested in using archeology to support the traditional account while Gu was primarily engaged in undermining it. See Bonner, Joey, Wang Kuo-wei: An Intellectual Biography (Cambridge: Harvard University Press, 1986), 190Google Scholar.

65. See Jiegang, Gu in his preface to Gushi bian 4 (1933), 11Google Scholar. The idea of a slow evolution and stratification of historical legend was probably first proposed by Cui Shu 崔述 (1740–1816).

66. This situation led directly to another theoretical shift in Gu Jiegang's method. Rather than having a fixed, traditional date for a text, then upon discovering an anachronism having to pronounce all or part of it inauthentic and casting it aside, Gu wanted to subsume Authentication studies under the task of dating: “That which is called ‘dating’ is simply discerning its attributed time period and placing it in its correct time period. Dating is the end, discerning authenticity is the means.” 所謂考 年代，也就是辨去其僞託之時代而置之於其真時代中。考年代是目的，辨真僞是手 段 Gushi bian 4 (1933), 18Google Scholar.

67. Gushi bian 4 (1933), 21–22Google Scholar. The difference here between Song and Gushi bian scholars is that while Song scholars wanted to directly re-connect with the intentions of the sages, Gushi bian scholars wanted to discern the historicity of events in the time of the so-called sages.

68. Gushi bian 4 (1933), 21Google Scholar.

69. Karlgren, Bernhard (1889–1978), “On the Authenticity and Nature of the Tso Chuan,” Gøteborgs högskolas årsskrift 32.3 (1926)Google Scholar; and “The Authenticity of Ancient Chinese Texts,” Bulletin of the Museum of Far Eastern Antiquities 1 (1929), 165–83Google Scholar.

70. He accepted the inauthenticity of the “Tai shi” chapter of the Shang shu because quotes from it in later texts are not found in the received edition, but noted that some people did not, on the same grounds, find the Chunqiu fanlu 春秋繁露 inauthentic.

71. Karlgren, , “Authenticity,” 167Google Scholar.

72. See n. 69 above.

73. Shi, Hu 胡適 (1891–1962), “Pinglun jinren kaoju Laozi niandai de fangfa” 評論 近人考據老子年代的方法 in Zhexue luncong 哲學論叢 1 (1933)Google Scholar; translated as “A Criticism of Some Recent Methods Used in Dating Lao Tzu” in Harvard Journal of Asiatic Studies 2.3/4 (1937), 373–97CrossRefGoogle Scholar; rpt. in Gushi bian 6 (1938), 387–410Google Scholar. It is followed in Gushi bian 6 by Feng Youlan's reply, pp. 410–17. This dialogue between Hu and Feng is contextualized by Shaughnessy, Edward in “The Guodian Manuscripts and Their Place in Twentieth-Century Historiography on the Laozi,” Harvard Journal of Asiatic Studies 65.2 (2005), 417–57Google Scholar.

74. That is, as distinct from government-sponsored.

75. That is, as distinct from commentary. “Classics” here has a wider application than the thirteen classics.

76. Hu, , “Pinglun jinren,” in Gushi bian 6 (1938), 390Google Scholar.

77. As above, where Liang Qichao argued that Laozi must be later than Mengzi because both used ren yi 仁義 (goodness and propriety) pedagogically and Mengzi had invented the topic, so too do Hu Shi and Gu Jiegang use the same method. They both noted that the Lun yu advocated wuwei er zhi 無爲而治 (ruling without contrivance) and since this is a well known Laozi trope, Hu Shi concluded that this “proves” Laozi's priority while Gu Jiegang concluded precisely the opposite.

78. Hu Shi's specific answer to Liang Qichao's assertion that Mengzi was the first to discuss “ren yi” 仁義 (goodness and propriety) was to point out that this phrase appears more than once in the Zuo zhuan. See Hu, , “Pinglun jinren,” in Gushi bian 6 (1938), 400Google Scholar.

79. He lists four reasons why one should discern inauthenticity; nine types of inauthentic texts; nine methods of making inauthentic texts; twelve reasons for making inauthentic texts; six ways that inauthentic texts are discovered; four reasons why inauthentic texts may have been produced; six rules that the Authentication scholar must abide by; four lists of eight, two, eight, and ten methods for authentication; and six conditions under which authentication must be carried out. The list that I am most interested in here, methodology, is in fact four lists: the one given by Hu Yinglin, the two given by Liang Qichao, and the one given by Bernhard Karlgren. All of these have already been mentioned above.

80. Xincheng, Zhang, Weishu tongkao 偽書通考 (1939; rev. ed., Shanghai: Shangwu, 1957), 781–801Google Scholar. The “personal name” mentioned here is “Qi daifu 七大夫” (p. 784).

81. The other main reference work for Authentication studies is, of course, that by Yun, Ji 紀昀 (1724–1805), quanshu, Sikuzongmu 四庫全書總目 (1782; rpt. ed., Beijing: Zhonghua, 1965, 1981, 1995)Google Scholar.

82. Jiaxi, Yu 余嘉錫 (1883–1955), Siku tiyao bianzheng 四庫提要辨證 (Beijing: Kexue, 1954; rpt. Beijing: Zhonghua, 1980; rpt., Kunming: Yunnan renmin, 2004), 500Google Scholar.

83. Jiaxi, Yu, Gushu tongli 古書通例 (Shanghai: Shanghai guji, 1985Google Scholar; rpt. in Mulu xue fawei han Gushu tongli 目錄學發微含《古書通例》, Beijing: Zhongguo renmin daxue, 2004)Google Scholar. It was not published in his lifetime.

84. Yu, , Gushu tongli, 174Google Scholar.

85. Yu, , Gushu tongli, 211–13Google Scholar.

86. Although the history of recently excavated Warring States, Qin, and Han texts is far too long to describe here, the 1972 excavations of two tombs at Yinqueshan 銀 雀山 uncovered some early texts with transmitted counterparts and was thus particularly significant for Authentication studies. These included parts of the Sunzi bingfa 孫子兵法, Sun Bin bingfa 孫臏兵法, Weiliaozi 尉繚子, Liu tao, Yanzi, and Guanzi. The archeological report was published in Wenwu 文物 1974.2. See also Yates, Robin D.S., “The Yin-Yang Texts from Yinqueshan,” Early China 19 (1994), 75–144CrossRefGoogle Scholar. Several more were soon to come, including finds at Dingzhou 定州 in 1973, Mawangdui 馬王堆 also in 1973, Shuanggudui 雙古堆 in 1977, Wangjiatai 王家台 in 1993, Guodian 郭店 in 1993, and the acquisitions made by the Shanghai museum in 1994. A brief introduction to excavated texts can be found in Pian Yuqian 駢宇騫 and Shuan, Duan 段書安, Ben shiji yilai chutu jianbo gaishu 本世紀以來出土簡帛概述 (Taibei: Wanjuanlou, 1999)Google Scholar. Some excavations discovered texts without received counterparts, and these are also of great use to Authentication studies. Some may in fact be parts of works recorded in the catalogs but since lost. See Li Xueqin 李學勤, “Jingmen Guodian Chu jian zhong de Zisizi” 荊門郭店楚簡中的《子思子》, Wenwu tiandi 文物天地, 1998.2; rpt. in Chongxie xueshu shi 重寫學術史 (Shijiazhuang: Hebei jiaoyu, 2001), 7–12Google Scholar.

87. Xueqin, Li, “Chongxin gujia Zhongguo gudai wenming” 重新估價中國古代文 明, Renwen zazhi 人文雜誌 supplementary publication Xian-Qin shi lunwen ji 先秦史 論文集, 1982Google Scholar; rpt. in Xueqin, Li, Li Xueqin ji 李學勤集 (Harbin: Heilongjiang jiaoyu, 1989), 15–27Google Scholar, and Xueqin, Li, Dangdai xuezhe zixuan wenku: Li Xueqin juan 當代學者自 選文庫：李學勤卷 (Hefei: Anhui jiaoyu, 1999), 1–14Google Scholar. Hereafter, the Li Xueqin juan pagination is used.

88. Li, , “Chongxin gujia Zhongguo gudai wenming,” 11Google Scholar.

89. Youwei, Kang, Xinxue weijing kao 新學偽經考 (1891; rpt. ed., Beijing: Wenhua xueshe, 1931; rpt. edited and with notes by Qian Zhongshu 錢鍾書, Beijing: Sanlian, 1998)Google Scholar.

90. Li, , “Chongxin gujia Zhongguo gudai wenming,” 12Google Scholar. Li Xueqin goes on to cite the Wenzi as an example of a text previously considered inauthentic then proven to be authentic. The 1973 Dingzhou find turned up a text clearly related to, yet still quite different from, the transmitted text. See Wenwu 1981.8 and 1995.12, as well as Blanc, Charles Le, Le Wen zi: À la Lumière de l'Histoire et de l'Archéologie (Montréal: Université de Montréal, 2000)Google Scholar. Given that even an “authentic” Wenzi appears to many scholars to be completely derived from the Huainanzi, the question of authenticity for this particular Masters text is far from settled with the Han manuscript unearthed at Dingzhou.

91. Xueqin, Li, “Dui gushu de fansi 對古書的反思,” Fudan daxue lishi xi 复旦大 學歷史系, Zhongguo chuantong wenhua de zai guji 中國傳統文化的再估計 (Shanghai: Shanghai renmin, 1987)Google Scholar; rpt. in Xueqin, Li, Jianbo yiji yu xueshu shi 簡帛佚籍與學術史 (Taibei: Shibao wenhua, 1994; rpt. Nanchang: Jiangxi jiaoyu, 2001), 29–32Google Scholar.

92. For example, he suggested that bronze fish excavated from tombs datable to the Chunqiu period (770–481) support the authenticity of the Yi li 儀禮 insofar as it describes these artifacts as coming from that period.

93. Xueqin, Li, “Zouchu yigu shidai,” Zouchu yigu shidai 走出疑古时代 (1994; rev. 2nd ed., Shenyang: Liaoning daxue, 1997), 1–19Google Scholar.

94. Li, , “Chongxin gujia Zhongguo gudai wenming,” 12Google Scholar. If Li Xueqin is here implying Gu Jiegang and the Gushi bian debates with his use of the term “yigu” 疑古 (doubting antiquity), then his characterization of them as “blindly” doubting antiquity is more strongly worded than elsewhere.

95. Li, , “Dui gushu de fansi,” 32Google Scholar.

96. Li, , “Zouchu yigu shidai,” 9Google Scholar. On p. 16 however, his target is once again Qing “new text” factionalism.

97. On the one hand, Jiegang's, Gu student Qiyu, Liu 劉起釪 wrote an ardent defense of what he took to be an attack on his teacher in “Guanyu ‘Zouchu yigu shidai’ wenti” 關於 ‘走出疑古時代’ 問題 in Chuantong wenhua yu xiandaihua 傳統文化與現代化 1995.4Google Scholar. On the other hand, Li Xueqin's co-worker on the government project to date the Xia, Shang, and Zhou dynasties (夏商周斷代工程), Song Jian 宋健, fell only just short of calling anyone who might doubt traditional Chinese history a traitor to his country. See his “Chaoyue yigu, zouchu mimang” 超越疑古、走出迷茫, Hebei shifan daxue xuebao: Shehui kexue ban 河北師範大學學報: 社会科學版 19.4 (10 1996), 1–8Google Scholar. Both of these articles, plus Li Xueqin's 1994 article, as well as two others, have been translated into English in the “Doubting of the ‘Doubting of Antiquity’” issue of Contemporary Chinese Thought 34.2 (Winter 2002–2003)Google Scholar. A good article that correctly locates the “debate” in attitude rather than methodology is one by Shin'ichi, Yanaka 谷中信一, “Shin shutsudo shiryō no hatsugen to Gikoshugi no yuku-e” 新出土資料的發現與疑古主義的行方, Chūgoku shutsudo shiryō kenkyū 中國出土資料研究 2 (1998), 150–66Google Scholar.

98. Broadening his target yet again, Li Xueqin in 1997 redefined “doubting antiquity” as a historical phenomenon in three phases: first as Song scholars doubting Han scholars, second as Qing scholars doubting Song scholars, and third as the Gushi bian debates of the 1920's and 1930's; see Li, , Zouchu yigu shidai, 343Google Scholar. One problem with discerning whom Li is arguing against is simply that the phrase “doubting antiquity” can refer both generally to those who doubt antiquity as well as specifically to the Gushi bian debates of the 1920's and 1930's. Li means the former, while his detractors take him to mean the latter.

99. See n. 64 above.

100. Liangshu, Zheng 鄭良樹, Guji bianwei xue 古籍辨偽學 (Taibei: Taiwan xuesheng, 1986)Google Scholar. Chapter Eight of this work now serves as the preface for later reprints of his Xu Weishu tongkao 續僞書通考 (Taibei: Taiwan xuesheng, 1984).

101. Zheng, , Guji bianwei xue, 209Google Scholar.

102. Specifically, he highlights the work of Yang Bojun 楊伯峻 (1909–1992) in his Liezi, jishi 列子集釋 (Beijing: Zhonghua, 1979; rpt. 1997)Google Scholar, and several articles by Zhang Yiren 張以仁 on the Guo yu and Zuo zhuan that appeared in Lishi yuyan yanjiusuo jikan 歷史語言研究所集刊.

103. Zheng, , Guji bianwei xue, 224Google Scholar.

104. Ling, Li 李零, “Chutu faxian yu gushu niandai de zairenshi” 出土發現與古書 年代的再認識, Jiuzhou xuekan 九州學刊 3.1 (1988), 105–36Google Scholar; rpt. in Li Ling zixuan ji 李零 自選集 (Guilin: Guangxi shifan daxue, 1998), 27–31Google Scholar. Li Ling acknowledges Yu Jiaxi's work in Gushu tongli as foundational to this list on p.27.

105. Li, , “Chutu faxian,” 31Google Scholar. In a neat reminder of the fungibility of even modern texts, not only did the publishers of the 1998 reprint change the graphs from traditional to simplified, but the author made an unremarked revision of the word “dao 道 (practice)” in the eighth conclusion to the “yi 意 (meaning)” there now.

106. Ling, Li, Jianbo gushu yu xueshu yuanliu 簡帛古書與學術源流 (Beijing: Sanlian, 2004), 198Google Scholar. This conception of early Chinese text formation is evident in Li Ling's examination of the Sunzi 孫子. Although he had at his disposal several editions, including an excavated text from the Han, a Dunhuang text from the Wei-Jin period, and lengthy Tang collectanea quotations, he does not present a collated edition in an attempt to recover an “original” text. Rather, he presents them consecutively, as an evolving subject, and, contrary to those who believed that an early excavated text vindicated the views of those who would date the text traditionally to the general Sun Wu 孫武 of Wu 吳 (c. 500 B.C.E.), he said: “The ‘personal authorship theory’ goes against (the facts of) ancient text formation” ‘親著說’ 是違反古書體例的. See Li Ling, Sunziguben yanjiu 《孫子》 古本研究 (Beijing: Beijing daxue, 1995), 205Google Scholar.

107. I use the metaphor of a “polymorphous text” not only to summarize Li Ling's description here, but also in conscious contradistinction to some other metaphors previously used; for example, Gu Jiegang's cenglei zaocheng 層累造成 “layered construction” as applied to textual formation, Bruce Brooks' “accreted” Lun yu text, and Carine Defoort's “core” Shizi text. These descriptions imply that once a part of the received text has been “established,” it can only be added to and not removed. My “polymorphous text” category, following Yu and Li, does not recognize any part of an early text as “settled.” See Jiegang, Gu, “Yu Qian Xuantong lun gushi shu” 與錢玄同論古史書, in Dushu zazhi 讀書雜誌 9 (1923)Google Scholar; Bruce, E. and Brooks, A. Taeko, The Original Analects: Sayings of Confucius and His Successors (New York: Columbia University Press, 1998)Google Scholar; and Defoort, Carine, “Ruling the World with Words: The Idea of Zhengming in the Shizi,” Bulletin of the Museum of Far Eastern Antiquities 73 (2001), 217–42Google Scholar.

108. Li, Jianbo gushu yu xueshu yuanliu, 204n3.

109. For example, Early Chinese Texts: A Bibliographical Guide, ed. Loewe, Michael (Berkeley: The Society for the Study of Early China, 1993)Google Scholar.

110. Some specialized Authentication studies texts have been published since 1988, when Li Ling published his “Chutu faxian” article, but while acknowledging the importance of excavated texts, they mostly seem not to have realized the implications of the work of Yu Jiaxi, Li Xueqin, and Li Ling on methodology. See, for example, Zhongguo weishu zongkao 中國僞書綜考, ed. Ruiquan, Deng 鄧瑞全 and Guanying, Wang 王冠英 (Hefei: Huangshan, 1998)Google Scholar, a reference book that simply adds another hundred texts to the 1104 of Xincheng, Zhang; Zhongguo weishu daguan 中國僞書大觀, ed. Zhaopeng, Yu 俞兆鵬 (Nanchang: Jiangxi jiaoyu, 1998)Google Scholar, a reference book that largely re-states the cases against 186 of Zhang Xincheng's 1104; and Jianguo, Liu 劉建國, Xian-Qin weishu bianzheng 先秦僞書辨正 (Xi'an: Shaanxi renmin, 2004)Google Scholar, a work in the vein of Zheng Liangshu's, which argues for the authenticity of forty-nine pre-Qin texts.

111. Thompson, Paul, The Shen Tzu Fragments (Oxford: Oxford University Press, 1979), xviiGoogle Scholar.

112. Thompson, Paul, “On the Formal Treatment of Textual Testimony,” in The Guodian Laozi: Proceedings of the International Conference, Dartmouth College, May 1998, ed. Allan, Sarah and Williams, Crispin (Berkeley: The Society for the Study of Early China, 2000), 89Google Scholar.

113. Roth, Harold, The Textual History of the Huai-nan Tzu (Ann Arbor: The Association for Asian Studies, 1992)Google Scholar. A critical edition has not yet been published. However, Roth has contributed to a complete English translation of the Huainanzi, The Huainanzi: A Guide to the Theory and Practice of Government in Early Han China/Liu An, King of Huainan, trans. and ed. Major, John S., Queen, Sarah A., Meyer, Andrew Seth, and Roth, Harold D., with contributions by Puett, Michael and Murray, Judson (New York: Columbia University Press, 2010)Google Scholar.

114. Roth, , The Textual History of the Huai-nan Tzu, 2Google Scholar.

115. Roth, Harold, “Some Methodological Issues in the Study of the Guodian Laozi Parallels,” in Allan, and Williams, , The Guodian Laozi, 81Google Scholar.

116. Roth, , “Some Methodological Issues,” 75 and 83Google Scholar; Roth credits LaFargue, Michael, The Tao of the Tao Te Ching (Albany: State University of New York Press, 1992), on p. 75Google Scholar.

117. Boltz, William, “Textual Criticism and the Ma wang tui Lao tzu,” Harvard Journal of Asiatic Studies 44.1 (1984), 185–224CrossRefGoogle Scholar.

118. Boltz, William, “The Study of Early Chinese Manuscripts: Methodological Preliminaries,” in The Guodian Laozi, ,, ed. Allan, and Williams, , 46Google Scholar.

119. Boltz, William, “The Composite Nature of Early Chinese Texts,” in Text and Ritual in Early China, ed. Kern, Martin (Seattle: University of Washington Press, 2005), 59Google Scholar.

120. Boltz, , “The Composite Nature of Early Chinese Texts,” 59–60Google Scholar.

121. Boltz, William, “Reading the Early Laotzyy 老子,” Asiatische Studien/Études Asiatiques 56.3 (2005), 213Google Scholar.

122. Boltz, , “Reading the Early Laotzyy 老子,” 213–14Google Scholar.

123. Boltz, , “Reading the Early Laotzyy 老子,” 230Google Scholar.