The Textual History of Sanguo Yanyi

The Mao Zonggang Recension

Introduction

The Mao Zonggang commentary edition of Sanguo Yanyi [MZG] represents the final stage in the novel's evolutionary history, and the text established by Mao Zonggang and his father Mao Lun in the early Qing has remained the standard text of Sanguo Yanyi down to the present day. The appearance of MZG and its utter domination of Sanguo Yanyi's transmission history during the Qing at the expense of all previous textual recensions is in large measure due to the success of the earlier commentary edition of Shuihu Zhuan 水滸傳 produced by Jin Shengtan 金聖歎 (1608-1661) during the late Ming (Jin Shengtan's prefaces to this commentary edition are dated at Chongzhen 14 [1641]).

The Jin Shengtan commentary edition of Shuihu Zhuan was itself a product of the late Ming literati interest in producing commentary editions of works of fiction and drama. However, Jin Shengtan's commentary goes far beyond any of the early commentaries produced under the names of Li Zhuowu, Zhong Bojing, Chen Meigong and others. Earlier commentary editions of vernacular literature generally restricted themselves to random comments on the text and the actions of the protagonists, without any attempt to produce a systematic analysis of the text as a whole. Jin Shengtan was the first commentator whose intention was to systematically analyze the underlying narrative structure of a novel: he prefaced the novel with a "reading methodology" 讀法 which gives general instructions on how to approach the novel as well as introducing some of the basic narrative techniques used by the author; every chapter is headed by a lengthy commentary which seeks to explain the individual narrative threads that were being laid in that chapter; the text of each chapter is copiously endowed with intratextual comments that highlight important narrative patterns or explicate narrative techniques and devices.

Although Jin Shengtan was forced to make radical textual revisions to his source text in order to ensure that the text corresponded to his commentary, the most drastic change being to cut the original 100 or 120 chapter text off at chapter 71, the lack of textual fidelity was more than compensated for in the eyes of literati readers by the complex reading apparatus supplied by Jin Shengtan.

The reading apparatus for Shuihu Zhuan revolutionized the reading of fiction, for newly imbued with knowledge of the theory of narrative composition, educated readers were no longer content merely to read novels as simple stories, but sought to apply the same sophisticated analytical techniques used by Jin Shengtan to other works of fiction. Sanguo Yanyi had long been closely linked with Shuihu Zhuan, and during the late Ming many sister commentary editions of the two novels had been produced, as well as the "Roster of Heroes" edition [YXP] in which the texts of the two novels were printed in parallel. Given the kinship felt to exist between these two novels it was almost inevitable that sooner or later someone would attempt to apply Jin Shengtan's methods to Sanguo Yanyi. The Mao Zonggang commentary to Sanguo Yanyi was completed no more than twenty-five years after the publication of Jin Shengtan's commentary to Shuihu Zhuan, and in both its form and content is clearly indebted to the earlier work.

Only fifteen years after the first publication of the Mao Zonggang commentary to Sanguo Yanyi, the third major critical commentary edition of one of the great Ming novels came out. This was the commentary to Jin Ping Mei 金瓶梅 by Zhang Zhupo 張竹坡 (1670-1698), who, highly influenced by both Jin Shengtan and Mao Zonggang, provided this novel with an initial reading methodology, chapter commentaries, and copious intra-textual comments. These three commentary editions not only provided a model for later fiction commentators (few of whom reached the brilliance of Jin, Mao and Zhang), but also established the theoretical principles on which narrative fiction was supposed to be constructed. The influence of these three commentators' theories of narrative structure can be clearly seen in later Qing novels such as Honglou Meng 紅樓夢.

Composition of the Mao Zonggang Text

The recension of Sanguo Yanyi that is conventionally called the Mao Zonggang commentary edition was in fact largely the work of Mao Zonggang's father, Mao Lun 毛綸 (courtesy name Shengshan 聲山). We know little about the life of Mao Lun, except for that he went blind in his middle age, and thereafter passed his time by writing commentaries for works of vernacular literature, with his son Mao Zonggang 毛宗崗 working as his amanuensis.[1] Aside from his commentary edition of Sanguo Yanyi, there is also extant a commentary edition by Mao Lun of the Yuan southern drama Pipa Ji 琵琶記 (The Lute), which he titled "The Seventh Book of Talent" (Diqi Caizi Shu 第七才子書). In the general commentary at the head of this work, Mao Lun gives us some insight into the composition of his commentary version of Sanguo Yanyi:

Long ago Master Luo Guanzhong composed Tongsu Sanguozhi (The Popular History of the Three Kingdoms) in 120 juan (i.e. 120 chapters). His recording of the events is in no way less brilliant than the style of the Historian [Sima] Qian, but the text has been corrupted by village scholars, which I find a great pity. A few years ago I was afforded the opportunity of reading the original version, and so I edited the text. Furthermore, not hiding my inadequacies, I made a detailed analysis of the text and provided several items of general discussion at the start of each juan (i.e. each chapter). Moreover, in order to help finish the work, I also permitted my son [Mao Zonggang] to add his contributions to the commentary. When the book was completed a good friend of mine from Baimen (i.e. Nanjing) who thought highly of it, was going to publish it, when unexpectedly a disloyal student of mine wanted to claim it as his own work, and so plans for its publication were abandoned, which was a great shame. For this reason I now proffer up my Pipa (The Lute) for the reader's comments, and will let my edition of Sanguo (The Three Kingdoms) follow at a later date.

昔羅關中先生作《通俗三國志》，共一百二十卷。其紀事之妙，不讓史遷，卻被村學究改懷，予甚惜之。前歲得讀其原本，因為校正，復不揣愚陋，為之條分節解，而每卷之前，又各綴以總評數段。且許兒輩亦得參附末論，共贊其成。書既成，有白門快友見而稱善，將取以付梓，不意忽遭背師之徒，欲竊冒此書為己有，遂致刻事中閣，殊為可恨。今特先 以《琵琶》呈教，其《三國》一書容當嗣出。[2]

From this account it seems that it was Mao Lun who initiated the project to revise and commentate the text of Sanguo Yanyi, and that his son, Mao Zonggang, was only a minor partner. It is not certain exactly how much Mao Zonggang contributed to the project, but from Mao Lun's words it would seem likely that Mao Lun was solely responsible for the textual revision of Sanguo Yanyi, the initial reading methodology, and the chapter commentaries, and that his son only helped with the intratextual comments. If so, this would explain the often disparate opinions expressed in various places within the commentary.

That Mao Lun was the major commentator of Sanguo Yanyi is also testified to by Li Yu, who in the preface quoted from below refers to Mao Lun as responsible for the commentary of Sanguo Yanyi, but makes no mention of Mao Zonggang. Nevertheless, when the Mao version of Sanguo Yanyi came to be published, Mao Lun's name was not prominently mentioned. In the earliest extant edition (MZG-A), the publication credits at the head of the text ascribe the commentary to Mao Zonggang and the textual editing to a certain Hang Yongnian 杭永年 from Suzhou, and Mao Lun is only obliquely referred to on the title-page. In later editions, Mao Lun's name remains inconspicuous (only in MZG-D and MZG-G is Mao Shengshan mentioned as the commentator on the title-page), and as the credit to Hang Yongnian is omitted from the text-initial credit (although retained at the head of the table of contents), this recension has become known as the Mao Zonggang edition. It is not known who Hang Yongnian was, nor the reason why Mao Lun's name was not prominently advertised in the published text, but it may well be that Hang Yongnian was the wayward disciple mentioned by Mao Lun as having upset his original publication plans, and that as a compromise solution to the dispute between the two of them, Hang Yongnian's name replaced Mao Lun's in the publication credits.[3]

The Mao commentary edition of Pipa Ji includes a preface dated at Kangxi 5 [1666], and so Mao Lun must have completed his commentary edition of Sanguo Yanyi prior to this date. Although the Mao text was not published immediately, it does not seem to have lain entirely dormant, but apparently circulated in manuscript form amongst the Nanjing literati.

Editions of the Mao Zonggang Text

Li Yu Preface Edition

The Li Yu Preface edition [MZG-A] is the earliest extant edition of the Mao Zonggang commentary text of Sanguo Yanyi. It was published by the Zuigeng Tang 醉畊堂 publishing house, and has a preface written by Li Yu at Hangzhou during the twelfth month of Kangxi 18 [1679] (early in 1680 by the Western calendar). Many works of fiction include forged prefaces attributed to famous literati figures, and there are forged prefaces attributed to Li Yu.[4] However, the two Li Yu prefaces to Sanguo Yanyi – the the one for the Mao commentary edition (MZG-A) and the one for his own commentary edition (LLW) – do appear to be genuine. In these prefaces Li Yu refers to his personal life (e.g. mention of his son-in-law returning from Nanjing) in a way that we would not expect see in a forgery, and the two prefaces closely accord with each other, suggesting that they were written by the same person: they are both signed by Li Yu at his villa, the Layered Gardens 層園 in Hangzhou where Li Yu resided at that time; they both have the same two seals at the end of the preface; and they both commence with a discussion of the "four great books of wonder".

In the preface to the Zuigeng Tang edition Li Yu says:

《水滸》之奇，聖歎嘗批之矣，而《三國》之評，獨未之及。予嘗欲探索其奇以正諸世，乃酬應日煩，又多出游少暇，年來欲踐其志，會病未果。適予婿沈因伯歸自金陵，出聲山所評書示予。…因伯索序，聲山既已先我而評矣，而予又為之序，不亦贅乎。…今聲山又布其錦心，出其繡口，條分句析，揭造物者秘藏，宣古人之義蘊。開卷井井，實獲我心。且使讀是書者知第一奇書之目果在《三國》也。因以證予說之不謬，則又何可以無言。[5]

Li Yu had moved to Hangzhou from Nanjing in Kangxi 16 [1677], and died there in Kangxi 19 [1680] at the age of seventy.[6] Shen Yinbo, who was married to Li Yu's eldest daughter, had remained in Nanjing to take care of Li Yu's household and business affairs there, and it is probable that after obtaining a preface from his father-in-law, Shen Yinbo took the manuscript back to Nanjing, and had it published by the Zuigeng Tang 醉畊堂 publishing house sometime in Kangxi 19 [1680].

This first edition is different to all later editions of the Mao recension in several important respects, notably the presence of the Li Yu preface, which is not found in any other extant edition. Li Yu opens his preface with a discussion of the "four great books of wonder" (sida qishu 四大奇書, a term which he says was originally used by Wang Shizhen (1526-1590) for Sima Qian's Shi Ji 史記 (Records of the Historian), the Nanhua Jing 南華經 (The Nanhua Sutra [i.e. The Book of Zhuangzi]), Shuihu Zhuan, and Xixiang Ji 西廂記 (The Romance of the Western Chamber), and then later applied by Feng Menglong (1574-1646) to the four great Ming novels Sanguo Yanyi, Shuihu Zhuan, Xiyou Ji and Jin Ping Mei. Li Yu's preface is largely a justification for his assertion that Sanguo Yanyi is the "foremost book of wonder" (diyi qishu 第一奇書) amongst these four great works of fiction, and in accord with this claim, the Zuigeng Tang edition is entitled "The Foremost of the Four Great Books of Wonder" (sida qishu diyi zhong 四大奇書第一種) on both the volume-initial position and the central margin of each folio (where the short title of the book, as well as volume or chapter number and page number were normally printed).

The only extant edition of the Li Yu Preface Edition is that issued by the Zuigeng Tang 醉畊堂 publishing house. This edition has a textual division of 60 juan, and a lineation of 8 lines of 24 characters per half-folio. It has 40 portrait illustrations of the novel's protagonists, and the text includes the initial reading methodology, as well as chapter-initial commentaries and intertextual commentary ascribed to Mao Zonggang.

Jin Shengtan Preface Editions

The Li Yu preface edition apparently only enjoyed a very brief period of circulation, and was quickly supplanted by a new edition of the Mao Zonggang commentary recension, from which all other extant editions are descended. In this new edition, the author of the highly successful commentary edition of Shuihu Zhuan, Jin Shengtan, was falsely associated with the commentary, presumably in order to increase the sales potential of the book:

• The Li Yu preface of the Zuigeng Tang edition was replaced by a preface ascribed to Jin Shengtan and dated the twelfth month of Shunzhi 1 [1644] (early in 1645 by the Western calendar).
• Jin Shengtan's name was appended to the volume-initial credits.
• Jin Shengtan's name was given as the commentator on the title-page.

Although the designation of Sanguo Yanyi as "the foremost book of talent" is attributed to Jin Shengtan in the false preface, it was in fact a term first used by Li Yu in his commentary edition of Sanguo Yanyi (LLW) that he issued shortly after he had written the preface for the Mao commentary edition. In this preface Li Yu says:

I have already made so bold as to write a preface for [Mao] Shengshan's commentary version, but I recall that long ago [Jin] Shengtan wanted to produce a commentary version of the historian [Sima] Qian's Shi Ji (Records of the Historian) which would be "the foremost work of talent". However that never came to fruition, and now my commentary version of the text … can truly be described as "the foremost book of talent" (diyi caizi shu). For this reason I am republishing it in order to make it available to those who appreciate things of old.

余於聲山所評傳首，已僭為之序矣，復憶曩者聖歎擬欲評定史遷《史記》為第一才子書，既而不果。余茲閱評是傳之文…誠哉第一才子書也。因再梓以公諸好古者。

The title "Foremost Book of Talent" is also present on the title-page of the Li Liweng commentary edition, and it must have been from this edition that the editor of the second edition of the Mao commentary recension came up with the idea of having Jin Shengtan reevaluate his list of "books of talents", and place Sanguo Yanyi in the position of honour.

Why then did Li Yu designate Sanguo Yanyi as the "foremost of the four great books of wonder" in his earlier preface to the Mao commentary recension, and as the "foremost book of talent" in his later preface to his own commentary edition? It may simply have been that he needed a new title to differentiate his edition from the Mao text, and so he could not reuse the term "book of wonder". Although the publication of Jin Shengtan's commentary editions of Shuihu Zhuan and Xixiang Ji under the respective titles of "The Fifth Book of Talent" (diwu caizi shu) and "The Sixth Book of Talent" (diliu caizi shu) was in the context of a specific set of six books (see above), Mao Lun established a precedent when he published his commentary edition of Pipa Ji in Kangxi 5 [1666] under the title of "The Seventh Book of Talent" (diqi caizi shu). With a fifth, sixth, and seventh book of talent already known to the public, it may have seemed natural to Li Yu to transform Jin Shengtan's list of six works of Chinese literature into an open series of fiction and drama, with Sanguo Yanyi placed at the head as the "foremost book of talent".

The designation "book of talent" was also used in a somewhat different meaning for the "talent and beauty" 才子佳人 romances Haoqiu Zhuan 好逑傳 (The Fortunate Union), Yu Jiao Li 玉嬌梨 (The Two Fair Cousins) and Ping Shan Leng Yan 平山冷燕 (The Two Gifted Maidens), which had two, three, and four talented heroes and heroines respectively, and so were known by the alternative titles of "The Book of Two Talents" (liang caizi shu), "The Book of Three Talents" (san caizi shu), and "The Book of Four Talents" (si caizi shu).[8] By the addition of the ordinal prefix, Haoqiu Zhuan, Yu Jiao Li, and Ping Shan Leng Yan became known as the third, fourth, and sixth books of talent, and were integrated into the series of "books of talent" established by Jin Shengtan and Mao Lun. This series was eventually expanded to a total of "ten books of talent" (shi caizi shu) by the designation of Huajian Ji 花箋記 (The Flowery Scroll), Pinggui Zhuan 平鬼傳 (The Pacification of the Demons) and Baigui Zhi 白圭志 (The White Sceptre) as the eighth, ninth, and tenth books of talent respectively.[9]

The Jin Shengtan preface edition is found in three editions, which vary by format but not by content. As with the Li Yu preface edition, they are all entitled Sida Qishu Diyizhong 四大奇書第一種, and all include the initial reading methodology, as well as chapter-initial commentaries and intertextual commentary ascribed to Mao Zonggang.

MZG-B

This is a small format edition with a 60 juan textual division, and a lineation of 11 lines of 25 characters per half-folio (rarely 12 lines of 25 characters). The prefatory matter includes 40 portrait illustrations of the novel's protagonists.

This is the most common woodblock edition of the Mao Zonggang text, and is preserved in a wide range of printings issued by various publishing houses, mostly located in Guangzhou. Publication dates for those printings that I have examined range from Jiaqing 19 [1814] through to Daoguang 25 [1845].

MZG-C

This is a large format edition with a 60 juan textual division, and a lineation of 10 lines of 23 characters per half-folio. The prefatory matter includes 40 portrait illustrations of the novel's protagonists, which in one printing are credited as having been engraved by Feng Yunlong 馮雲龍 of Yangcheng 羊城 (i.e. Guangzhou).

MZG-D

This is a large format edition with a 19 juan textual division, and a lineation of 12 lines of 26 characters per half-folio. The prefatory matter includes 40 portrait illustrations of the novel's protagonists, although in one printing that I have examined, these have been replaced by 120 full-page illustrations based on those found in LLW.

Most printings of this edition seem to emanate from the mid-Chinese cities of Nanjing and Suzhou.

Dual Sanguo Yanyi – Shuihu Zhuan Editions

These editions have the texts of Sanguo Yanyi and Shuihu Zhuan printed in parallel after the format of the Ming "Roster of Heroes" edition (YXP), each page being divided into two registers, with the text of Sanguo Yanyi in the lower register, and the text of Shuihu Zhuan (a 115 chapter text, not the Jin Shengtan commentary version) in the upper register. That these editions are directly based on the earlier YXP edition, but with the Mao text replacing the original Sanguo Yanyi text, is indicated by the fact that they reproduce the title-page and/or Xiong Fei preface from YXP, as well as preserving the original 20 juan textual division of that edition.

There are two editions of this dual Sanguo Yanyi – Shuihu Zhuan text, both entitled Sida Qishu Diyizhong 四大奇書第一種, and having a textual division of 20 juan.

MZG-E

This edition is subtitled Yingxiong Pu 英雄譜 ("Roster of Heroes"), and has a lineation of 15 lines of 23 characters per half-folio for Sanguo Yanyi, and a lineation of 16 lines of 12 characters per half-folio for Shuihu Zhuan. It includes the chapter-initial commentaries and intertextual commentary ascribed to Mao Zonggang, but omits the initial Reading Methodology, and is not illustrated.

MZG-F

This edition is subtitled Han-Song Qishu 漢宋奇書 ("Han and Song Works of Wonder"), and has a lineation of 12 lines of 20 characters per half-folio for Sanguo Yanyi, and a lineation of 13 lines of 10 characters per half-folio for Shuihu Zhuan. It includes the chapter-initial commentaries and intertextual commentary ascribed to Mao Zonggang, as well as the initial Reading Methodology, and is illustrated with 80 portrait illustrations (1-40 for Sanguo Yanyi, 41-80 for Shuihu Zhuan).

Li Yu – Mao Zonggang Joint Commentary Edition

This edition [MZG-G] represent a hybrid between the Mao Zonggang recension and the Li Liweng commentary edition (LLW). In addition to the Mao Zonggang text and commentary, the marginal comments from the Li Liweng commentary edition are appended in the margin of the text. Moreover, this version of the Mao Zonggang text preserves features of LLW such as the initial table of dramatis personae, and the 24 juan textual division. Also, unlike all other editions of the Mao Zonggang recension, this edition is entitled Guanban Dazi Quanxiang Piping Sanguozhi 官板大字全像批評三國志 ("Official edition of The History of the Three Kingdoms with large characters, illustrations and commentary"). There is a preface by Huang Shuying 黃叔瑛 dated at Yongzheng 12 [1734], which was presumably written for the first printing of this edition.

This edition has a 24 juan textual division, and a lineation of 11 lines of 22 characters per half-folio. It includes the initial reading methodology, chapter-initial commentaries and intertextual commentary ascribed to Mao Zonggang, as well as marginal comments derived from LLW. It also has 40 portrait illustrations of the novel's protagonists.

Abridged Edition

There is also an abridged version of the Mao recension [MZG-H], with a textual division of 12 juan, in which the narrative text is abridged, all poetry and intra-textual commentary are omitted, and the chapter commentaries are collected together at the head of each juan (with abridgement where necessary, so that each chapter commentary does not exceed one folio).

The only example of this edition that I have encountered is a fragmentary volume (covering chapters 61-70) of what appears to be an early Qing edition held at the British Library in London. As the short title on the central fold is given as "The Foremost of the Books of Wonder" (qishu diyi zhong), which accords with the Li Yu preface edition (MZG-A), but not with the standard Jin Shengtan preface editions (MZG-B, MZG-C, MZG-D), it is possible that this is an early abridgement of the Li Yu preface edition.

Features of the Mao Zonggang Text

At the hands of Mao Lun and Mao Zonggang, Sanguo Yanyi underwent its most extreme textual revision. The commentary provided by Mao Lun was the most comprehensive that had ever been given to Sanguo Yanyi, comprising detailed chapter-initial commentaries and line-by-line intratextual comments. Furthermore, following in the footsteps of Jin Shengtan, Mao Lun prefaced the text with an initial "Reading Methodology" which sets out the basic parameters for understanding how to approach the text (according to Mao Lun).[10] However, unlike the earlier commentary editions, such as those ascribed to Li Zhuowu and Zhong Bojing, this was not merely the old text provided with a new commentary, but was a systematic critical commentary combined with a thorough textual revision. For Mao Lun, as for Jin Shengtan with his earlier commentary on Shuihu Zhuan, the two processes of commentary and textual revision went hand-in-hand. Earlier commentators on Sanguo Yanyi may have been critical of the text in places, but they preferred to remain as outside observers, merely pointing out to the reader those passages which they agreed with and those which they disagreed with (passages which "Li Zhuowu" and other earlier commentators thought were fit to be excised were marked by sidelining). However, Mao Lun had a specific agenda to promote, and he could not allow inconsistencies between the text and his commentary to weaken his position. This necessitated the rectification of the text in those places which departed from Mao Lun's concept of what the original authorial intent should have been.

The major textual changes made in his version of Sanguo Yanyi are enumerated by Mao Lun in a section entitled "General Principles" (fanli 凡例) at the head of the text:

1. The literary particles of the vulgar edition are mostly incomprehensible, and the language long-winded and repetitive. I now change it all according to the ancient edition, which I really think makes [the language] clear and direct.
   一，俗本之乎者也等字，大半齟齬不通。又詞語冗長，每多複沓處。今悉依古本改正，頗覺直捷痛快。
2. The vulgar edition is largely corrupt in its recording of events. For example the episodes "Zhaolie (Liu Bei) drops his chopsticks at the clash of thunder" (Item 41),[11] "Ma Teng is killed on entering the capital" (Item 114),[12] and "Lord Guan is enfeoffed as the Han Marquis of Shouting" (Item 51)[13] are all in disagreement with the ancient edition. In addition, the episode "Empress Cao berates Cao Pi" is detailed in Fan Ye's Hou-Han Shu (The History of the Latter Han), yet the vulgar editions actually write erroneously that she took the side of evil (Item 159);[14] the episode "Lady Sun throws herself in the river and drowns" is detailed in Xiaoji Zhuan (Biographies of Ruthless Women), and yet the vulgar editions merely record that she returns to Wu (Item 121).[15] I now identify the truth of these events according to the ancient edition.
   一，俗本紀事多訛。如昭烈聞雷失箸及馬騰入京遇害，關公封漢壽亭侯之類，皆與古本不合。又曹后罵曹丕詳于范曄《後漢書》中，而俗本反誤書其黨惡，孫夫人投江而死詳于《梟姬傳》中，而俗本但紀其歸吳。今悉依古本辨定。
3. There are some episodes that cannot be omitted, such as "Lord Guan holds up a torch through the night",[16] "Guan Ning cuts the mat and sits apart [from his friend Hua Xin]",[17] "Cao Cao shares out incense and [tells his concubines to] sell shoes [for a living]",[18] "Yu Jin sees the painting at the temple of [Cao Cao's] tomb",[19] including descriptions of the talent of the wife of Martial Marquis (Zhuge Liang),[20] the cleverness of Kangcheng's serving girl,[21] Deng Ai's reply of "Phoenix, oh phoenix",[22] Zhong Hui's answer about not sweating,[23] Du Yu's obsession with the Zuo Zhuan (The Zuo Commentary [to the Spring and Autumn Annals]).[24] All these episodes are deleted from the vulgar edition and so are left unrecorded. I now preserve these episodes according to the ancient edition, so that the reader may see the full picture.
   一，事有不可闕者，如關公秉燭達旦，管寧割席分坐，曹操分香賣履，于禁陵廟見畫，以至武侯夫人之才，康成侍兒之慧，鄧艾鳳兮之對，鍾會不汗之答，杜預《左傳》之癖，俗本皆刪而不錄。今悉依古本存之，使讀者得窺全豹。
4. Some fine literary pieces from the Three Kingdoms period that are recorded in Wen Xuan (Literary Selections) such as Kong Rong's "Memorial recommending Mi Heng" and Chen Lin's "Denouncement of Cao Cao" are as worthy for transmission as [Zhuge Liang's] former and latter "Memorial to launch a campaign", but the vulgar edition omits them both, and so they are not recorded. I now insert these pieces according to the ancient edition so that lovers of the past may read them.[25]
   一，三國文字之佳，其錄于《文選》中者，如孔融《薦禰衡表》，陳琳《討曹操檄》實可與《前後出師表》並傳。俗本皆闕而不載。今悉依古本增入，以備好古者之覽觀焉。
5. In the vulgar edition the outline of contents (i.e. the chapter titles) is uneven and without antithesis, confused and without order. Moreover each chapter is divided into two sections. Putting myself in the mind of the author, I now connect up [the two halves of each chapter], and for each chapter ensure that there is an antithetical couplet for its title, insisting upon clever use of language [for the composition of the titles] in order to delight the reader.
   一，俗本題綱參差不對，錯亂無章。又于一回之中分上下兩截。今悉體作者之意而聯貫之。每回必以二語對偶為題，務取精工以快悅（閱）者之目。
6. The vulgar edition falsely attributes its commentary to Master Li Zhuowu, but actually it is not known who is really responsible for the commentary. In the commentary there are often remarks which are offensive to Zhao Lie (Liu Bei) or derogatory to the Martial Marquis (Zhuge Liang). I now delete them all, and replace them with a new commentary to rectify [the erroneous opinions of the earlier commentary].
   一，俗本謬托李卓吾先生批閱，而究竟不知出自何人之手。其評中多有唐突昭烈、謾罵武侯之語。今俱削去，而以新評校正之。
7. What is even more laughable about the vulgar edition is that it marks narrative with which it agrees with circles and dots at the side, and it marks narrative with which it disagrees with deletion marks (side-lining). If one does not judge the language but only the events, then what about the Chun Qiu (The Spring and Autumn Annals) [which was reputedly edited by Confucius] where there are thirty-six instances of regicide and fifty-two instances of the overthrowal of states. Should we then take the sage's classic and cover it with marks of deletion? Now in this edition, at the points of commentary there are side-circles and dots but no side-lines, thus removing completely the earlier vulgarity.
   一，俗本之尤可笑者，于事之是者則圈點之，于事之非者則塗抹之。不論其文而論其事，則《春秋》弒君三十六、亡國五十二，將盡取聖人之經而塗之抹之耶。今斯編評閱處有圈點而無塗抹，一洗從前之陋。
8. Within the narration of events there are inserted poems, and this should be an exquisite part of the literary style. Yet in the vulgar edition time after time we come across the words "Later someone wrote a poem in admiration saying" or everywhere else we find Master Zhou Jingxuan, whose poems are laughably vulgar in the extreme. Now in this edition I select poems by famous Tang and Song authors to replace them. These are completely different from those in the vulgar edition.
   一，敘事之中夾帶詩詞，本是文章極妙處，而俗本每至「後人有詩嘆曰」，便處處是周靜軒先生，而其詩又甚俚鄙可笑。今此編悉取唐宋名人作以實之，與俗本大不相同。
9. Heptasyllabic regulated verse originated in the Tang dynasty, and during the Han dynasty this verse form had not yet come into being. The vulgar edition often falsifies poems by men of old, for example the poems by Zhong You and Wang Lang in praise of the Bronze Phoenix Terrace (Item 111)[26] and the poem that Cai Mao inscribes on the walls of (Liu Bei's) residence (Item 67)[27] are all examples of heptasyllabic regulated verse, which really is thought laughable by men of knowledge. Now according to the ancient edition I delete them all in order to preserve the authenticity [of the text].
   一，七言律詩起于唐人，若漢則未聞有七言律也。俗本往往捏造古人詩句，如鍾繇、王郎頌銅雀臺，蔡瑁題館驛壁，皆偽作七言律體，殊為識者所笑。今悉依古本削去以存其真。
10. There are some episodes which were invented by later authors and are absent in the vulgar edition of the popular elaboration (Sanguo Yanyi) but which can be found today in drama, for example such episodes as Lord Guan executing Diaochan,[28] and Zhang Fei capturing Zhou Yu.[29] Although these episodes are false, yet people today know them to be so. There are some episodes which are absent in the ancient edition of Sanguo Yanyi yet present in the vulgar edition of the popular elaboration, such as those episodes in which Zhuge Liang wishes to burn Wei Yan alive in Shangfang Valley (Item 205),[30] and Zhuge Zhan receiving Deng Ai's letter and being unable to come to a decision (Item 234).[31] Although these episodes are false, yet people today do not know them to be so. If they do not know them to be false is this not too great an injustice for the men of old to bear? Now I delete them all so that the reader will not be misled by [the false words of] the people of Eastern Qi.
    一，後人捏造之事，有俗本演義所無而今日傳奇所有者，如關公斬貂蟬，張飛捉周瑜之類。此其誣也，則今人之所知也。有古本《三國志》所無而俗本演義所有者，如諸葛亮欲燒魏延于上方谷，諸葛瞻得鄧艾書而猶豫未決之類。此其誣也，則非今人之所知也。不知其誣，毋乃冤古人太甚。今皆削去，使讀者不為齊東所誤。

In order to justify his textual changes to the reader, Mao Lun invents the deceit that his text was based on some newly-discovered "ancient edition" (guben 古本), and that it was the "vulgar edition" (suben 俗本) in current circulation which had corrupted the original text, and violated the author's intentions. This claim does not stand up to scrutiny. If the textual changes implemented by Mao Lun did actually have some textual basis, we would expect to find some evidence of this in at least some Ming editions. In fact no other extant edition has any of the features which Mao Lun specifically records as being present in his "ancient edition". On the contrary, in all other respects the Mao recension is a typical member of the textual group which includes LZW and the other 120 chapter commentary editions (B1b branch).

If the Mao text is not based on an "ancient edition", then what was Mao Lun's source text? Mao Lun describes the "vulgar edition" of Sanguo Yanyi as comprising 120 bipartite chapters, containing poems ascribed to Zhou Jingxuan, having side-marking of the text, and possessing a commentary ascribed to Li Zhuowu which is often critical of Liu Bei and Zhuge Liang. This profile exactly accords with the Li Zhuowu commentary edition (LZW), and as the narrative text of the Mao recension basically follows that of LZW (both share the same textual affiliation), it seems quite probable that the "vulgar edition" (LZW) that Mao Lun so despised was actually the source edition for his text, and that there is no textual basis for any of the changes made by Mao Lun.

This suspicion is confirmed by a comparison between the Mao commentary and the Li Zhuowu commentary. Mao Lun claims that he has replaced the Li Zhuowu commentary of the "vulgar edition" with his own commentary, but a close examination shows that this is not entirely true, and that there are in fact quite a few examples of intratextual or chapter-initial comments in the Mao recension which are derived from the marginal commentary or chapter-final commentary of the Li Zhuowu commentary edition (LZW). For example, in MZG ch.107 an intratextual comment on Xin Chang's sister Xianying is almost word-for-word identical to an interlinear comment in LZW ch.107a; in MZG ch.110 an intratextual comment on Wen Qin and his son Wen Yang is almost word-for-word identical to an interlinear comment in LZW ch.110a; and in MZG ch.5 an item of the chapter-initial commentary comparing Yuan Shu (unfavourably) and Cao Cao (favourably) is almost identical to one item of the LZW ch.5 chapter-final commentary.

In summary, the changes effected to his source text (LZW) by Mao Lun were as follows.

Textual Redivision

The earliest editions of Sanguo Yanyi comprise 240 narrative items divided into a number of juan 卷 volumes (10, 12, 20, or 24), but in LZW and other late Ming and early Qing commentary editions the 240 items were redivided into 120 hui 回 chapters, each chapter comprising two discrete narrative items. Mao Lun converted the 120 bipartite chapters of his source edition into 120 unitary chapters by eliminating the division between the two items constituting each chapter. Furthermore, as each item was originally an independent narrative unit, the two narrative titles of the earlier bipartite chapters did not form a matching couplet, and so Mao Lun also rewrote the original item titles as antithetical couplets, with each couplet forming the chapter title of a single unitary chapter.

Stylistic Rewriting

Stylistic rewriting of his source text is the most pervasive type of change effected by Mao Lun, and there is hardly a line that has not been affected to some degree by Mao Lun's surgical knife. Stylistic changes range from the pruning of inappropriate classical particles and minor alteration in the phrasing of the text through to the wholesale deletion or rewriting of passages. The result is that the narrative text of the Mao recension is considerably shorter than that of its source edition: the total length of narrative (excluding commentary and extra-narrative poetry) for MZG is only about 475,000 characters compared with about 550,000 characters for LZW, which is a reduction in length of about 15%.[32] Although this line-by-line textual rewriting often obscures or obliterates the original narrative text, the stylistic alteration does not generally affect the overall meaning of the original text, and in some places Mao Lun's stylistic alterations can even be considered to be beneficial to the text. For example, his fine descriptive use of language is illustrated in the account of Zhang Fei's murder in Chapter 81. In LZW the assassination is described with the minimum of detail:

[Fan Qiang and Zhang Da] went straight up to [Zhang Fei's] bed. Zhang Fei's snoring was like thunder. The two assassins did their deed, and killed Zhang Fei. Concealing the severed head, they came out, and together with several score men they went over to the State of Wu by boat.

直至床前，飛鼻息如雷。二賊下手，將飛殺之，藏其首級而出。便下船來，引數十人，投東吳去了。(LZW 81:5a)

However in the Mao Zonggang text the event is described with a far greater vividness:

[Fan Qiang and Zhang Da] went straight up to [Zhang Fei's] bed. It happened that Zhang Fei never closed his eyes when he slept, so that night as he slumbered in his tent the two assassins saw him with eyebrows bristling and eyes staring wide, and so they dared not make a move [against him]. Only when they heard his thunderous snoring did they dare approach him, and with their daggers stab him in the midriff. Zhang Fei let out a bellowing roar and died.

直至床前。原來張飛每睡不合眼。當夜寢於帳中，二賊見他須豎目，本不敢動手，因聞鼻息如雷，方敢近前，以短刀刺入飛腹。飛大叫一聲而亡。(MZG-B 41:5b-6a)

Deletion or Substitution of Poetry

A number of the intra-narrative poems in the original text were written as regulated verse, which is a poetic form that was not invented until long after the end of the Three Kingdoms period. Mao Lun deleted or replaced all such anachronistic poems. Mao Lun was also unsatisfied with the extra-narrative poems in his source text, and both excised a large number of old poems, and added a great many new ones. Of the set of 391 extra-narrative poems present in LZW, only 151 of these poems are present in MZG, and of those original poems that do remain most have been altered, and many rewritten almost beyond recognition. Furthermore the original ascriptions of authorship for most of the poems have been eliminated. Mao Lun singles out those poems written by Zhou Jingxuan for special condemnation, but in fact some twenty-seven (out of the original seventy) Zhou Jingxuan poems are still preserved in the Mao recension, only the references to their authorship have been deleted.

Narrative Expansion

In a number of places Mao Lun expands upon the original narrative text by adding material from other sources. These include cases where he quotes historical documents (e.g. Chen Lin's denouncement of Cao Cao (ch.22), and Kong Rong's memorial (ch.23)), and where he inserts anecdotal material derived from other sources (e.g. anecdotes relating to Zheng Xuan and his serving girl (ch.22), Guan Yu standing guard over Liu Bei's two wives (ch.25), Guan Ning dissociating himself from Hua Xin (ch.66), Cao Pi shaming Yu Jin (ch.79), Lady Sun's suicide on hearing of the death of Liu Bei (ch.84), Deng Ai's stutter (ch.107), Zhong Hui's clever response (ch.107), Zhuge Liang's virtuous wife (ch.117) and Du Yu's love for the Zuo Zhuan (ch.120)).

Textual Rectification

Many of the textual changes made by Mao Lun reflect his pro-Shu (Liu Bei) anti-Wei (Cao Cao) position, and are insidious in intent, for they quietly realign the reader's interpretation of the central protagonists. Although the original author clearly comes out on the side of Liu Bei, and acknowledges the legitimacy of the state of Shu over the state of Wei, the pro-Shu anti-Wei bias in the text of the early editions is rarely overstated, and the attitude towards Cao Cao is often ambivalent. On the other hand, Mao Lun is unequivocal in his praise for Liu Bei and his condemnation for Cao Cao.

Several of the differences which Mao Lun mentions in the general principles between the text of his edition and that of the "vulgar editions" are examples of Mao Lun deliberately altering his source text in order to bring it in line with his pro-Shu sentiments. For example, the cynical attempt by Zhuge Liang to murder Wei Yan (Item 205) is emended in the Mao text so that Zhuge Liang is not shown in a negative light; and Cao Cao's final words to his ministers regarding the succession (Item 156) are supplemented in the Mao text by frivolous instructions to his concubines (thus showing Cao Cao in a negative light, and eroding his status as a good ruler in the eyes of the reader, especially when compared with Liu Bei's statesmanlike final words in Item 169). Elsewhere, characters who in the original text are critical of Shu or prevaricate in their support for Shu are shown to be steadfastly loyal to the Shu cause in the Mao text: Ma Teng's siding with Cao Cao (Item 114) is changed so that he only pretends to throw in with Cao Cao; Empress Cao's criticism of her husband for not abdicating in favour of her brother Cao Pi (Item 159) is changed so that she berates her brother for being disloyal to the Han dynasty (an abrupt change from a pro-Wei stance to a pro-Shu stance); Zhuge Zhan's hesitation as to whether he should defect to Wei (Item 234) is changed to an immediate rejection of the idea.

Mao Lun's political realignment of the text goes far deeper than suggested by the few examples mentioned in the general principles, and throughout the text minor textual changes, reinforced by the Mao commentary, are designed to manipulate the reader's interpretation of the novel. A good example is Mao Lun's orchestration of the reader's initial impressions of the two central protagonists of the novel, Liu Bei and Cao Cao. In Item 1 of the pre-Mao text, when Liu Bei is first introduced it is narrated that as a child whilst once playing under a large mulberry tree with some other children he said "I am the emperor, and should ride upon this canopied chariot!" 我為天子，當乘此羽葆車蓋 (LZW 1:5a). On hearing these insurrectionary words his uncle scolds him: "Don't talk nonsense! You'll be the end of our family line!" 汝勿妄言，滅吾門也 (LZW 1:5a), thereby giving the reader a negative impression of Liu Bei's words. However in the Mao recension, the condemnation of Liu Bei by his uncle is deleted along with some of the following text with the result that Liu Bei's remark is now followed by the out of context words of his uncle: "This boy is no ordinary person!" 此兒非常人也 (MZG-B 1:5a), thus giving implicit approval of Liu Bei's words. A positive interpretation of Liu Bei's words is further strengthened by the Mao commentary at this point, which compares Liu Bei's childhood words with those reputedly made by Liu Bang, founder of the Han dynasty, when he was a child. When Cao Cao is first introduced in Item 2, it is explained in the pre-Mao text that he is the twenty-fourth generation descendant of the Han prime-minister Cao Shen, that his great-grandfather Cao Jie was a man of benevolence and generosity, that his grandfather Cao Teng was one of the ten eunuch attendants at court who was later enfeoffed as the Neighbourhood Marquis of Bi, and that his father Cao Song, originally a member of the Xiahou clan adopted by Cao Teng (a eunuch), was a loyal and filial, refined and elegant official who held a succession of high posts under Emperor Ling. This impressive genealogy for Cao Cao is omitted in the Mao recension, and replaced by the derogatory words "[Cao] Cao's father Cao Song originally had the family name Xiahou, but as he was the adopted son of the eunuch attendant Cao Teng he passed himself off as being a member of the Cao clan" 操父曹嵩本姓夏侯氏，因為中常侍曹騰之養子，故冒姓曹 (MZG-B 1:8b). This negative impression of Cao Cao is again heightened by the accompanying commentary at this point: "With a genealogy like this, how could Cao Cao ever be compared with the scion of the Prince of Jing, the distant descendant of Emperor Jing (i.e. Liu Bei)?" 曹操世系如此，豈得與靖王後裔景帝玄孫同日論哉.[33]

Notes

1. For studies on Mao Lun and Mao Zonggang, see the introduction to "Mao Tsung-kang on How to Read the San-kuo yen-i (The Romance of the Three Kingdoms)" in David Rolston (ed.), How to Read the Chinese Novel pp.146-151; Chen Xianghua, "Mao Zonggang de shengping yu Sanguozhi Yanyi Mao-pingben de Jin Shengtan xu wenti"; and Huang Lin, "Youguan Mao-ben Sanguo Yanyi de ruogan wenti". Chen Xianghua suggests that Mao Zonggang was born in Chongzhen 5 [1632] and died about Kangxi 48 [1709].

2. Diqi Caizi Shu 1:28a-28b.

3. This suggestion is also put forward by Huang Lin in "Youguan Mao-ben Sanguo Yanyi de ruogan wenti" p.328.

4. For example the "Li Yu" preface to a Jianyang edition of Shuihu Zhuan that is discussed by Y.W. Ma in "Shuihu shushou ziliao liuzhong (wai yizhong)" p.121.

5. Cited in Chen Xianghua, "Mao Zonggang de shengping yu Sanguozhi Yanyi Mao-pingben de Jin Shengtan xu wenti" p.79. Cf. the "Jin Shengtan" preface to MZG, in which it is rather feebly claimed that Jin Shengtan "suddenly saw a draft copy of Mao's commentary version of Sanguo Yanyi on a friend's desk" 忽於友人案頭見毛子所評《三國志》之稿.

6. See Sun Kaidi, "Li Liweng yu Shi'er Lou" p.158-162.

7. See Chen Xianghua, "Mao Zonggang de shengnian yu Sanguozhi Yanyi Maopingben de Jin Shengtan xu wenti" pp.76-83 for a discussion of the relationship between the "Jin Shengtan" preface and the Li Yu preface.

10. For a translation of Mao Zonggang's "Reading Methodology" see David Roy's "How to read The Romance of the Three Kingdoms" in David Rolston (ed.), How to Read the Chinese Novel pp.146-195.

32. These figures are based on my own calculations. I have excluded poetry from the calculations as the two editions have a largely dissimilar set of extra-narrative poems. Moss Roberts states that "[t]he revision [of Mao Zonggang] entailed reducing the length of the novel by about one-sixth, from some 900 thousand characters to 750 thousand" (The Three Kingdoms: A Historical Novel p.965). The figure of 900,000 seems to be derived from the number of print characters (891,000) given on the imprint page of the Shanghai Guji Chubanshe 1980 typeset edition of 1522, whilst the figure of 750,000 seems to be derived from the number of print characters (734,000) given on the imprint page of the Zuojia Chubanshe 1955 typeset edition of MZG. However, as these figures include punctuation as well as Chinese characters, they are wildly inaccurate (my own calculation for 1522 is about 588,000 characters). It is interesting to note, however, that the percentage of abridgement for the Mao Zonggang recension is also about 15% for these inaccurate figures.

Contents | Bibliography

BabelStone Home Page