Siku tiyao entry for Tao Yuanming ji

Shuxiang You， Dewei Shen, Marc Nürnberger

(威斯康辛大學(xué)麥迪遜分校東亞系,美國威斯康辛麥迪遜,53705;德國慕尼黑大學(xué)東亞系,德國慕尼黑,D-80539)

SikutiyaoentryforTaoYuanmingji

Shuxiang You， Dewei Shen, Marc Nürnberger

(威斯康辛大學(xué)麥迪遜分校東亞系,美國威斯康辛麥迪遜,53705;德國慕尼黑大學(xué)東亞系,德國慕尼黑,D-80539)

TaoYuanmingji陶淵明集 (Collected works of Tao Yuanming ji), 8juan, Edition stored in the Imperial Treasury.

Compiled by Tao Qian陶潛(365-427 A.D.)of the Jin Dynasty[1]. Editors Comments: according to the “PerfactoryNote” by Yang Xiuzhi陽休之(509-582 A.D.)of the Northern Qi Dynasty(550-577 A.D.),there were altogether three versions of the Collected works of [Tao]Qian circulating in the world: one in eight juan, with on preface; one in six juan, with preface and contents, but the arrangement was in a mess and at the same time also lacking[materials][2];one compiled by Xiao Tong蕭統(tǒng)(501-531 A.D.),editors comment: books edited and collected by ancients, were also referred to as “written” [by them]. Therefore, all the old versions of Wen Xuan文選[Selections of Literature]state in[their]title “written by Heir Apparent Zhaoming of the Liang Dynasty” and Xu Ling’s徐陵(507-583 A.D.)“Preface to Yutai xin yong”[New Songs from the Jade Terrace]”also claims that “wrote and collected love songs and composed altogether ten volumes.”[Yang]Xiuzhi claiming that the Collected works of Tao Qian was written by [Xiao]Tong, probably followed the parlance of his days, and today [we]still follow his old text[3], and also in eight Juan[4], with Wuxiao Zhuan五孝傳(Five Biographical Writings of Filial Piety)and Siba mu四八目[5](Four Times Eight Entries) missing.SibamuisShengxianqunfulu圣賢群輔錄[6][Records of Sages and Group of Assistants].[Yang]Xiuzhi comprehensively collated and combined these three versions and established [one in ]ten juan, already not the old [version]of Zhaoming [i.e. Xiao Tong].[7]

In addition, Song Xiang’s宋庠(996-1066 A.D.)“Siji”私記[8](Private Note)claims that in“Jingjizhi”經(jīng)籍志(Records of Bibliography), inSuishu隋書[Book of Sui],the collectedworks of Tao Qian has nine juan[9].[“Jingjizhi”]also says in the Liang Dynasty there was [one in]five Juan, and one volume of contents[10].Tang“Zhi”唐志[11][The “Records”in the Tang]says [there were ]five juan[12]. The circulating[editions]of [Song]Xiang’s day [were the following]:one was Xiao Tong’s eight-volume version, with the essays arranged in front of poems; one was Yang Xiuzhi’s ten-volume version; otherwise, there were dozens of other versions[13],[so]at last [Song Xiang]did not know which was the right one[14]. Only in his later years did he obtain the old version of Jiang Zuo[15], of which the order of his arrangement seemed to have the most cohesion. Nowadays the circulating version is what [Song]Xiang called“Jiang Zuo version.”[16]

However, the Heir Apparent Zhaoming昭明, who was close to the generation of Qian 潛, already had lost [the chance to]access theWuxiaozhuanandSibamu, and thus did not include them in [Qian’s]works. How could Yang Xiuzhi 陽休之 have the two books appended [to Qian’s corpus]? Moreover, the quotations from theShangshu尚書 (Book of Documents) in theWuxiaozhuanandSibamuare contradictory to one another[17]. They were therefore decidedly not processed by the same hand. The two books must have been writings falsely attributed to [Qian], which Xiuzhi trusted erroneously and added [to Qian’s corpus]. Although in each of the various later editions there were differences with regard to the number of fascicles and the order of the arrangement, they moved along the same track in the sense that they all allowed the interpolation of fake materials. This [process]actually began with what Xiuzhi compiled. [Song]Xiang’s “Siji” 私記 (“Private Postscript”) merely doubts the authenticity of the two entries “Ba ru”八儒 (“The Eight Confucian Schools”) and “San mo” 三墨[18](“The Three Moist Schools”), which also resulted from imprudence in evidential criticism.

Now, [since the inauthenticity]of theSibamuhas already been pointed out by His Sagacious Discernment[19], and known clearly as a forgery, it is separately recorded under the “Compendia” of the “Zi” 子 (“Various schools”) category and carefully distinguished as false. As for theWuxiaozhuan, the meaning behind its words is mediocre and superficial; it definitely was not composed by Qian. Since [theWuxiaozhuan]emerged contemporaneously with theSibamu, its forgery is thus needless to say. Now [in our edition]they have both been deleted. As for the compilation of Tao Qian’s poetry and prose, we still follow what the Heir Apparent Zhaoming determined, and arrange [his works]into eight fascicles. Although the original order of the Liang era is no longer available for scrutiny, by the exclusion of the forgeries and the preservation of the authentic, [this edition]could almost be counted as still close to [its appearance]in antiquity.

Original Chinese Text of the Siku tiyao entry

《陶淵明集》八卷內(nèi)府藏本

晉陶潛撰。案北齊陽休之《序錄》,潛集行世凡三本:一本八卷,無序；一本六卷,有序目,而編比顛亂,兼復(fù)闕少；一本為蕭統(tǒng)所撰(案古人編錄之書,亦謂之撰,故《文選》舊本皆題“梁昭明太子撰”,而徐陵《玉臺新詠序》亦稱“撰錄艷歌,凡為十卷”。休之稱潛集為統(tǒng)撰,蓋沿當(dāng)日之稱,今亦仍其舊文)，亦八卷,而少《五孝傳》及《四八目》。《四八目》即《圣賢群輔錄》也。休之參合三本,定為十卷,已非昭明之舊。

又宋庠《私記》稱《隋·經(jīng)籍志》潛集九卷,又云梁有五卷,錄一卷?！短浦尽纷魑寰?。庠時所行,一為蕭統(tǒng)八卷本,以文列詩前。一為陽休之十卷本。其他又數(shù)十本,終不知何者為是。晚乃得江左舊本,次第最若倫貫。今世所行,即庠稱江左本也。

然昭明太子去潛世近,已不見《五孝傳》、《四八目》,不以入集,陽休之何由續(xù)得?且《五孝傳》及《四八目》所引《尚書》自相矛盾,決不出于一手,當(dāng)必依托之文,休之誤信而增之。以后諸本,雖卷帙多少、次第先后,各有不同,其竄入偽作,則同一轍,實自休之所編始。庠《私記》但疑“八儒”、“三墨”二條之誤,亦考之不審矣。

今《四八目》已經(jīng)睿鑒指示,灼知其贗,別著錄于子部類書而詳辨之。其《五孝傳》文義庸淺,決非潛作。既與《四八目》一時同出,其贗亦不待言。今并刪除。惟編潛詩文仍從昭明太子為八卷。雖梁時舊第今不可考,而黜偽存真,庶幾猶為近古焉。

注釋：

[1]Tao Yuanming陶淵明also known as Tao Qian陶潛(365-427A.D.) was a Chinese poet of the Eastern Jin Dynasty(317-420A.D.).He is the foremost of the poets who have written their greatest work while in reclusion and/or those poets in whose poems the theme of countryside solitude particularly resonates. Approximately 130 of his works survive. See Yuan Xingpei袁行霈,TaoYuanmingjijianzhu陶淵明集箋注.[Notes and Commentaries on Collectedworks of Tao Yuanming](Beijing:Zhonghua shuju,2003);Meng Erdong 孟二冬,TaoYuanmingjiyizhujiyanjiu陶淵明集譯注及研究[Translation, Comments and Studies on Collectedworks of Tao Yuanming ](Beijing: Kunlun chubanshe,2008);David Hinton, trans,TheSelectedPoemsofT’aoCh’ien( Washionton: Copper Canyon Press,1993).

[2]According to Chen Shangjun陳尚君and Zhang Jinyao張金耀(SikutiyaoJingdu四庫提要精讀[Intensive Reading of Precise of the Four Treasuries ][Shanghai: Fudan Daxue chubanshe,2008],pp:336-337),it is clear to which of the before mentioned editions Yang Xiuzhi’s comment “the arrangement was in a mess and at the same time again lacking[materials]”編比顛亂,兼復(fù)闕少refers. Given the flow of the text, his comment is more likely to be only on the six-volume version.

[3]This comment by the Siku ministers elaborates on the meaning of the technical term zhuan撰“to write”. In ancient times, the meaning of zhuan撰was apparently much broader and could also refer to making notes and commentaries, or to compiling a book. Hence Yang Xiuzhi could state “One edition was written by Xiao Tong”一本為蕭統(tǒng)所撰,meaning Xiao Tong “compiled” or “edited” that edition. To strengthen their argument, the Siku ministers offer two further examples of this usage: one is Xiao Tong’sWenXuanand the other is Xu ling’sYutaixinyong.WenXuanis one of the earlist existing anthologics of Chinese poetry and literature, which was compiled by Xiao Tong and a group of scholars he had assembled. However, all the old versions ofWenXuanstate that it was written by Xiao Tong. Likewise,Yutaixinyongis a collection of Chinese poetry dating to the time of the Six Dynasties(222-589A.D.),but it was claimed to be “written”by Xu Ling.

[4]Chen Shangjun and Zhang Jinyao(p.337)notes that Yang Xiuzhi praised Xiao Tong’s version, saying that “the edition and extraction have a good style, and his arrangement can be followed”編錄有體,次第可尋.

[5]There is a controversy about the meaning of Siba mu. Hu Xiangyun胡祥云(“Siba mu tiyi xiyi”四八目題意析疑wenshi 3[2006],pp.79-86)argues that Siba四八means“four times eight”,which equals to thirty-two, and mu目means entries. Thus, Siba mu means thiry-two entries, and refers to thirty-two entries of Pre-Qin figures which were finished earlier on. Chen Shangjun 陳尚君and Zhang Jinyao張金耀 (p.343)provide another hypothesis about the meaning of Siba mu. They claim that mu was an evaluation system of people during the period from the Eastern Han(25-220A.D.) to the Southern and Northern Dynasties(420-589A.D.). They also claim that si and ba refer to the entries whose titles contain “four” and “eight” respectively, since the number of these two types of entries is the largest. This translation follows Hu’s opinion and thus translates Siba mu as “Four Times Eight Entries.”

[6]Siba mu provides brief introductions of ancient sages, men of virtue, great ministers as well as other famous political assistants. Therefore, it is also calledShengxianqunfulu.

[7]Chen Shangjun 陳尚君and Zhang Jinyao張金耀(p.337)claim that Yang Xiuzhi’s ten-volume version was based on Xiao’s version and also referred to the other two versions. According to Yang Xiuzhi’s “Prefactory Note”, he quite appreciated Tao Qian’s works and was afraid that Tao Qian’s works would be lost since the existing three versions were different. Therefore, he addedWuxiaozhuanandSibamuas well as the preface and contents, and finally compiled a ten-volume version余頗賞潛文,以為三本不同,恐終致忘失。今錄統(tǒng)所闕,并序目等,合為一帙十卷,以遺好事君子.

[8]See Song Xiang宋庠,“Si ji”私記(Private Note),in Yuan Xingpei, Tao Yuanming ji jianzhu,p.615.

[9]According to Chen Shangjun陳尚君 and Zhang Jinyao張金耀(p.338),Tao Shu陶澍(1779-1839A.D.)considered this nine-volume version to be Xiao Tong’s eight-volume version including one volume of contents, one volume of the “Preface” and “Eulogy”, and seven volumes of the main text..

[10]Chen Shangjun陳尚君and Zhang Jinyao張金耀 (p.338)claimed that the Liang version refers to the six-volume version mentioned in the first paragraph.

[11]It is unclear whether here Tang“Zhi”refers to “Jingji zhi”經(jīng)籍志[Records of Bibliography], inJiuTangShu舊唐書[Old Book of Tang],or refers to “Yiwen zhi”藝文志[Records of Art and Literature], inXinTangShu新唐書[New Book of Tang].Both mentioned that the Collectedworks of Tao Qian had five Juan. In “Jingji zhi”經(jīng)籍志inJiuTangShu舊唐書(Taipei: Dingwen shuju,1981),p.2067,it says 陶淵明集五卷；in“Yiwen zhi”藝文志,inXinTangShu新唐書(Taipei: Dingwen shuju,1981),p1590,it says陶潛集二十卷又集五卷。

[12]As mentioned in last note, “Yiwen zhi”藝文志[Records of Art and Literature], inXinTangShu新唐書[New Book of Tang](p.1590)claims that the Collectedworks of Tao Qian also had a twenty-volume version. However, Chen Hanzhang陳漢章(1864-1938A.D.)(ChongwenZongmujishibuzheng崇文總目輯釋補正,vol.4[Beijing:Zhonghua shuju,2006])considers this claim unconvincing.

[13]According to Cheng Shangjun 陳尚君and Zhang Jinyao張金耀(p.338),these dozens of versions might all be handwritten versions.

[14]In Song Xiang’s “Si ji”, he claims that he collected dozens of versions and at last did not know which one is right.余前后所得本僅數(shù)十家,卒不知何者為是。

[15]It is unclear what this “Jiang Zuo” refers to. In “Si ji”,Song Xiang claims that he obtained this version in his later years, which was said to be from the old version of Jiang Zuo. Song Xiang considered the order of the arrangement of this version to have the most cohesion晚獲此本,云出于江左舊書,其次第最若倫貫。Chen Shangjun陳尚君 and Zhang Jinyao 張金耀claim that it should refer to a version of Southern Tang. However, Chen and Zhang also mentioned that the editor and the number of volumes of this southern Tang version were not known clearly. Moreover, “Jiang Zuo” can refer to the area east to the Yangzi River, or a version of the Eastern Jin Dynasty.

[16]According to Chen Shangjun and Zhang Jinyao (P.338), Song Xiang based on the Jiang Zuo version and proofread it by using other versions, finally compiling a new version of the Collected works of Tao Qian in ten juan. This version is called “Prime Minister Song version”宋丞相本or “Song version”宋本for short. After Song Xiang, a monk named Si Yue思悅compiled a ten-volume version.宋庠之后,有僧人思悅編訂的十卷本。 But both versions have been lost(See Chen and Zhang,p.339).

Besides these two versions, Chen and Zhang mention several other known versions of the Northern Song(960-1127A.D.)for textual research, including Chao Wenyuan’s晁文元(951-1034A.D.)version, Chen Shugu’s陳述古(1017-1080A.D.)version, Prime minister Zhang’s version(according to Chen and Zhang p.339, this “Prime Minister Zhang”may refers to Zhang Shangying張商英(1043-1121A.D.)),the version of Donglin Temple東林寺in Mount Lu廬山,Cen Rang’s 岑穰(?)collected version, Chao Yongzhi’s晁詠之(c.1055-1106A.D.)collected version, Wang Zhongliang’s王仲良(?)block-printed version. These above seven versions have all been lost(See Chen and Zhang,p.339).

[17]According to the Siku catalogue summary of theShengxianqunfulu圣賢群輔錄 (i.e.Wuxiaozhuan), the editorial board displays the following evidence: In theWuxiaozhuan, the punctuation of the sentence “xiao hu wei xiao you yu xiongdi” [孝乎惟孝友于兄弟]cited fromShangshu尚書 follows Bao Xian’s 包咸 commentary that belongs to the New Character School. However, in theShengxianqunfuluunder the entry of “Siyue” 四岳, the citation of Shangshu adopts Kong An’guo’s 孔安國 commentary which belongs to the Old Character School. This adequately demonstrates that the two books should not have been compiled by the same person. SeeSikuquanshu, juan 137.

[18]In Song Xiang’s 宋庠 (996-1066A.D.) preserved postscript for Tao Qian’s works, “Benchao Song Chengxiang siji” 本朝宋丞相私記, it mentions that although he collected several dozen of the editions of Tao Qian’s works, he could not tell which one should be the best. However, the presentation of the chapter order of the Jiangzuo 江左 edition seemed most consistent. In this Jiangzuo edition, the annotation ofWuxiaozhuanandSibamuis more elaborate than in any other editions he saw. Nevertheless, Song Xiang was especially aware of two entries inSibamutitled “Ba ru” and “San mo”, which seem to have been randomly inserted by later scholars. He doubts the two entries could really reflect the intention of Tao Qian. In addition, right before the two entries, there is one sentence that obviously attempts to end the book: “All that could be recorded by books or transmitted by the wise elders, and all those whose virtue or disrepute are widely heard, should have been enumerated here to exhaustion.” 凡書籍所載及故老所傳, 善惡聞于世者, 蓋盡于此矣。Song Xiang considered it another evidence for the adulteration of the two entries. In contrast, Yuan Xingpei 袁行霈 offers a detailed evidential criticism to defend the two entries. See Yuan XingpeiTaoYuanmingjijianzhu陶淵明集箋注 (Beijing: Zhonghua shuju, 2003),pp.597-600,p.615.

[19]It looks that since the Qianlong 乾隆Emperor (r. 1736-1795A.D.), by his “sagacious discernment,” pointed out in person that theSibamu(a.k.aShengxianqunfulu圣賢群輔錄) was a book falsely attributed to Tao Qian, the editorial board of the Siku imperial library attacked its inauthenticity vehemently on many occasions in theSikuquanshu. For instance, in the “Front Matter” (fanli 凡例) of theSikuquanshu, Entry 18 used the Siba mu as a bad exemplar of forged transmitted books of little value. On the contrary, the modern scholars Chen Shangjun 陳尚君 and Zhang Jinyao 張金耀 criticize the editorial board as too impressionable towards the emperor’s opinion to make the argument tenable. They come up with several rational antitheses and conclude that it is better to append Wuxiao zhuan andSibamuafter Tao Qian’s works than to completely discard them, since neither the pros nor the cons are convincing. See Chen Shangjun and Zhang JinyaoSikutiyaojingdu四庫提要精讀 (Shanghai: Fudan daxue chubanshe, 2008), pp.344-345.