Few issues in studies of early China arouse so much controversy as the question of the dating of major pre-imperial (i.e. pre-221 b.c.e.) texts. Already during the imperial era, particularly during the Song 宋 (960–1279) and Qing 清 (1636/1644–1912) dynasties, scholars noticed that many texts traditionally attributed to specific thinkers contain clear anachronisms and may have been written generations after the death of their putative author. Back then, their ideas were bitterly contested by other scholars.Footnote 1 These debates could become politically sensitive, as happened in the final years of the Qing dynasty, when a reformer Kang Youwei 康有爲 (1858–1927) based his radically revisionist view of Confucianism on—highly dubious—claims that some of the canonical texts were forged by the associates of the notorious usurper Wang Mang 王莽 (45 b.c.e.–23 c.e.).Footnote 2 In the early Republican era (1911–1949) an assault on traditional dating of major texts was part and parcel of intellectual and cultural awakening and of the break with traditional culture in general, and this was the case for a number of Japanese scholars as well.Footnote 3 A further upsurge in the debate surrounding the dates of pre-imperial texts occurred in China in the 1980s, and publications on the topic continue, albeit with lesser frequency, well into the present day.Footnote 4 Yet despite an impressive harvest of scholarly manuscripts and articles, few if any controversies have been settled. It is fair to conclude that on this topic the scholarly community in China, Japan, and the West agrees primarily to disagree.
Even a cursory survey of current treatment of the texts attributed to pre-imperial Masters (or Philosophers, zi 子) will show a huge variety of approaches. While many scholars (especially those coming from the field of philosophy) habitually treat these texts as if they were really penned by the putative author, others insist that the Masters are creations of the texts no less than the texts are creations by the Masters.Footnote 5 Some believe that even a lengthy text—such as, e.g., the Zuo zhuan 左傳—can be dated with great precision, while others argue that even short texts like the Laozi 老子 are “composite” in their nature: they could be repeatedly rearranged and reassembled from smaller textual units, which makes the very notion of “dating” highly tenuous.Footnote 6 Nor is there an agreement on which factors should be given priority in determining the text's date. Should we focus primarily on the text's internal information (e.g. references to historical events and personages), on its vocabulary and grammar, on its relations with other texts, on references to it in later (e.g. Han 漢 [206/202 b.c.e.–220 c.e.]) literature, or on anything else?
In a recent insightful study Paul Fischer outlined the evolution of authentication debates from the Han dynasty to the present day. Fischer analyzed different criteria used in these debates, pointed out their weaknesses, and put forward what appears to be an increasingly popular “polymorphous texts” paradigm. Fischer considers the Masters' texts as “fluid entities variously constructed from related but disparate pericopes circulating independently or in diverse editions.” This understanding in his eyes undermines the very reasonability of authentication debates. Fischer concludes: “This paradigm does not, of course, abrogate the field of Authentication studies at one stroke. But it is the greatest challenge to the field in all the years since its inception.”Footnote 7
At first glance, Fischer's conclusion undermines not just broadly defined authentication studies but their sub-field, i.e. clarifying the texts' dates. Indeed, if every text and every textual unit was repeatedly reworked and redacted, dating it with any precision becomes pointless.Footnote 8 Yet I believe that this conclusion is premature and also grossly counterproductive. Their fluid nature aside, the Masters' texts and different units thereof—be they clusters of chapters (the forerunners of future books), individual chapters, or just smaller segments (zhang 章)—display clear temporal differences in terms of their content and language, as I hope to demonstrate below. Eschewing the question of the date of their composition severely diminishes our ability to contextualize these texts and to properly position them within pre-imperial and early imperial ideological controversies. Rather than giving up studies of the text's dates, we should incorporate new understandings outlined by Fischer and like-minded scholars and put forward a more cautious, more up-to-date, and more methodologically transparent way of tackling this question.
It is against this backdrop that in what follows I want to address anew problems involved in ascertaining the dates of composition of the Masters' texts. For my case study I have chosen the Book of Lord Shang attributed to Shang Yang 商鞅 (d. 338 b.c.e.), a text to which I have been attracted for some time due to its peculiar social and political stance. The choice is not accidental. Debates about the composition of the Book of Lord Shang and the dates of its component chapters have continued in China and Japan for almost a century; yet in Western scholarship such issues remain largely ignored.Footnote 9 The resultant lack of clarity about the date of the book's composition is one of the reasons for a regrettable neglect of its content in Western Sinology, which I aim to reverse. Besides, I hope that by presenting a methodologically transparent analysis of the text's dating I shall be contributing toward renewed interest in the question of the dating of other Masters' texts. In addition, resolving some of the controversies around the Book of Lord Shang's dates may be helpful to contextualize it within Qin 秦 history and contribute therewith to studies of Qin history in general.
Background: Textual History of the Book of Lord Shang
The Book of Lord Shang is commonly associated with Shang Yang (aka Gongsun Yang 公孫鞅 or Lord of Shang 商君),Footnote 10 the singularly famous Qin statesman, whose reforms propelled this state to eventual supremacy over the rival polities of the Warring States (Zhanguo 戰國, 453–221 b.c.e.) era. The thinker and the text associated with him have become hugely controversial ever since the Warring States period; in the eyes of many imperial literati they epitomized political amorality—or, worse, immorality—and did not merit scholarly attention. This trend of “being ashamed to speak about Shang Yang,”Footnote 11 reflects the attitudes of not just many traditional scholars but also many modern ones, particularly in the West.Footnote 12
The first references to writings associated with Shang Yang come from an ideologically close text, the Han Feizi 韓非子. On one occasion, when urging the ruler to be ready to alter existing norms and regulations and to dispense, when necessary, with public opinion, Han Fei 韓非 (d. 233 b.c.e.) says: “it is explained in the ‘Internal and External’ of Lord Shang” (shuo zai Shangjun zhi ‘nei wai’) 說在商君之〈內外〉.Footnote 13 The current Book of Lord Shang does not include a chapter “Internal and External” (nei wai 内外); judging from the content of Han Fei's argument, he refers to what is currently the first chapter of the Book of Lord Shang, “Revising the laws” (“Geng fa” 更法). Elsewhere, Han Fei cites Shang Yang directly. After explicating Shang Yang's principle of imposing heavy punishments on minor offenses, Han Fei says:
公孫鞅曰: 行刑, 重其輕者——輕者不至, 重者不來。是謂以刑去刑。
Gongsun Yang said: “When punishments are implemented, inflict heavy [punishments] on light [offences]: then light [offences] will not come, and heavy [ones] will not arrive. This is called: to eradicate punishments with punishments.”Footnote 14
This statement by Gongsun Yang (i.e. Shang Yang) appears almost verbatim in two of the chapters of the Book of Lord Shang: 4, “Eliminating the strong” (“Qu qiang” 去強), and 13, “Making orders strict” (“Jin ling” 靳令).Footnote 15 Possibly, then, Han Fei had access to the writings attributed to Shang Yang. Indeed, elsewhere he mentions that families within the boundaries possess “laws/methods of Shang [Yang] and Guan [Zhong]” (Shang Guan zhi fa 商管之法).Footnote 16 Judging from the context, these “laws” or “methods” (fa 法) refer not only to legal or administrative regulations bequeathed by these statesmen but to the texts attributed to them, i.e. the [proto-] Book of Lord Shang and [proto-] Guanzi 管子.Footnote 17 If Han Fei is right, then the Book of Lord Shang might have circulated broadly in the Chinese world on the eve of imperial unification.
The Book of Lord Shang was relatively well known throughout the Former Han dynasty 前漢 (206/202 b.c.e.–9 c.e.). Several eminent Han statesmen reportedly studied Shang Yang's teachings, and his ideas were openly endorsed by a few leading courtiers.Footnote 18 Sima Qian 司馬遷 (c. 145–90 b.c.e.) claims to have read at least two chapters of the Book of Lord Shang, viz. “Opening the barred” (“Kai sai” 開塞) and “Agriculture and warfare” (“Nong zhan” 農戰). The first of these chapters is mentioned also in a slightly earlier treatise, Huainanzi 淮南子 (c. 140 b.c.e.).Footnote 19 By the end of the Former Han dynasty, the first catalog of the imperial library was compiled by Liu Xiang 劉向 (79–8 b.c.e.) and his associates. The catalog mentions the book named Lord Shang 商君 in twenty-nine chapters in the “Legalists” (fajia 法家) sub-section of the “Masters” (or “Philosophers,” zi 子) books. Separately, it lists a book named Gongsun Yang 公孫鞅 in twenty-seven chapters in the “Strategists” (bing quanmou 兵權謀) sub-section of “Military books” (bing shu 兵書). This latter may well be another version of the Book of Lord Shang.Footnote 20
The Book of Lord Shang continued to circulate in the aftermath of the Han Dynasty's fall in 220 c.e.; it is recorded (under the name Master Shang 商子 or alternately Book of Lord Shang) in the imperial catalogues of the Sui 隋 (581–618) and Tang 唐 (618–907) dynasties; both mention the text as comprising five scrolls (juan 卷) and do not refer to the number of chapters (pian 篇).Footnote 21 Under the Tang, the Book of Lord Shang was still relatively well known; sections of it were incorporated by Wei Zheng 魏徴 (580–643) in his Essentials of Orderly Rule from Multiple Books (Qunshu zhiyao 群書治要, compiled in 631), and references to it appear in the Comprehensive Institutions (Tong dian 通典) by Du You 杜佑 (735–812).Footnote 22 However, by the time of the Song Dynasty 宋 (960–1279) we have first indications of textual corruption in the Book of Lord Shang. Zheng Qiao 鄭樵 (1104–1162) mentions in the bibliographic section of his Comprehensive Treatises (Tong zhi 通志) that three of the twenty-nine chapters mentioned in the Han catalog are lost; a similar statement is made in the Bibliographic Treatise from the Jun Studio (Jun zhai dushu zhi 郡齋讀書志) by Chao Gongwu 晁公武 (1105–1180).Footnote 23 By the end of the Song dynasty, one more chapter is reported to have been lost, and by the end of the Yuan dynasty 元 (1271–1368) another had followed.Footnote 24 In the early Ming Dynasty 明 (1368–1644) two versions of the text existed, one with twenty-five and another with twenty-four chapters, but it was only the latter, shorter, one that continued in circulation.Footnote 25 Currently available recensions all miss chapters 16 and 21 of the twenty-six-chapter Song version. Of the three chapters that had been lost between the Han and the Song dynasties, only one fragment survived in Wei Zheng's compilation.Footnote 26
The loss of chapters is only the most visible aspect of the text's deterioration over the centuries. To this, one should add many instances of textual corruption within the surviving chapters. This corruption is most obvious in the chapters that deal with military matters (especially chapters 10–11, of which only fragments remained intact) and specific administrative and military regulations, such as chapter 19. It seems that early scribes and later literati who were in charge of reproducing the text were less acquainted with matters peculiar to the book's immediate military and administrative context, which resulted in several lacunae, misplaced slips, miswritten characters or later typos, and the like.
Aside from direct corruption, the Book of Lord Shang suffered from centuries of neglect by the imperial literati, which resulted in minimal editorial efforts on their part. The book was never divided into paragraphs and sentences (zhangju 章句); it had almost no annotations (only a few glosses by anonymous commentators are found in a single early recension);Footnote 27 and we know of no serious scholarly work on it prior to the very end of the eighteenth century. This neglect made the text notoriously difficult for later readers, and may explain why they often eschewed it, preferring to learn about Shang Yang's legacy from a much better preserved “Biography of Lord Shang” (“Shangjun liezhuan” 商君列傳) in Sima Qian's Records of the Historian (Shi ji 史記). This tendency regrettably persists well into the present day, although it is clear that Sima Qian's account is not accurate and should be read cum grano salis insofar as Shang Yang's career and ideas are concerned.Footnote 28
Nowadays, the earliest surviving recensions of the Book of Lord Shang come from the sixteenth and early seventeenth centuries, i.e. the latter half of the Ming dynasty.Footnote 29 Yet all these are superseded by the first known critical edition of the text prepared by the great Qing man of letters, Yan Wanli 嚴萬里, better known by his later name, Yan Kejun 嚴可均 (1762–1843).Footnote 30 Yan had in his possession a recension of the Book of Lord Shang dating back to the Yuan dynasty; he collated it with two Ming recensions and published a new text in 1793. Later, Yan (now under the name of Yan Kejun) prepared a second improved edition (1811), but it was the first one which became more popular, being republished in 1876 by Zhejiang Publishing House as part of the compilation Twenty-two Masters (Ershier zi 二十二子). This later version—which corrected some of Yan's typographical errors but unfortunately made a few new ones—became the foundation for most subsequent publications of the Book of Lord Shang throughout the twentieth century.Footnote 31
The publication of Yan Wanli's critical edition, as well as renewed interest in Shang Yang and his legacy from the second half of the nineteeth century, for the first time prompted Qing literati to engage in collating and glossing the Book of Lord Shang. By the late nineteenth century, the book had benefited from efforts of such eminent scholars as Sun Xingyan 孫星衍 (1753–1818), Yu Yue 俞樾 (1821–1907), and Sun Yirang 孫詒讓 (1848–1908).Footnote 32 Their efforts in turn influenced many new annotated editions published through the Republican Period (1912–1949). Of these the most notable are those by Wang Shirun (1915), which served as the basic text for Duyvendak's English translation; Zhu Shizhe 朱師轍 (1921, revised in 1948 [republished in 1956]), which served as the basic text for Perelomov's Russian translation); and that of Jiang Lihong 將禮鴻 (1944, published in 1996), which eventually became the standard version in the Mainland, being incorporated into the New Edition of the Masters Collection (Xinbian zhuzi jicheng 新編諸子集成).
After the establishment of the People's Republic of China (1949), textual work on the Book of Lord Shang slowed, but all of a sudden accelerated in the early 1970s in the context of the anti-Confucian campaign, when Shang Yang was lionized as a “progressive” thinker whose legacy deserved the utmost attention.Footnote 33 Many publications on Shang Yang and the Book of Lord Shang appeared then, but most have little, if any, scholarly value. One stands apart though, Gao Heng's 高亨 (1900–1986) The Book of Lord Shang, Commented and Translated.Footnote 34 Gao, who studied the Book of Lord Shang long before the Cultural Revolution (1966–1976), managed to avoid blatant politicization and presented balanced and academically solid work, which by far surpassed that of his predecessors. The only significant defect of his publication is the use of simplified characters, which makes his edition less appropriate for a critical study of the text.
In recent decades, many editions of the Book of Lord Shang have been published, but few merit mention; in most cases scholars simply reproduced earlier recensions and Gao Heng's translation into colloquial Chinese (baihua 白話). The major exceptions to this rule are several publications by Zhang Jue 張覺, who became deeply engaged in studying the Book of Lord Shang in the early 1990s. His efforts culminated in a new critical edition (2012), which fully benefited from modern library facilities, and which by far supersedes anything published on the Book of Lord Shang heretofore.Footnote 35 Zhang Jue consulted ten different recensions from the Ming and Qing period and, according to his testimony, “selected the best” of each without exclusively following any of them. His work records all cases of textual discrepancies among various recensions, collects most of the earlier commentaries, and thus is immensely helpful for researchers. Moreover, it contains a rich apparatus of auxiliary materials, including detailed introductions to each of the extant early recensions of the Book of Lord Shang.
Needless to say, my research on the Book of Lord Shang has benefited enormously from Zhang Jue's efforts, and in what comes next I normally follow his edition, unless indicated otherwise. It should be noticed, however, that even this excellent edition suffers from several flaws. Zhang Jue's parsing of the text and his attempts to divide it into paragraphs and sentences are not always convincing. Some of his personal glosses are of lower quality than could have been expected; and his marked lack of interest in recent studies is self-defeating. Zhang neither incorporates insights from excavated materials, nor appears to be interested in these in general. Moreover, his repeated attacks on Gao Heng result in the rejection of many sensible suggestions made by Gao, and sometimes in glossing over Gao's views. These negative aspects notwithstanding, Zhang's work remains an impressive scholarly achievement, to which I am immensely indebted.
Debates over the Text's Dating
Because of the lackluster interest in the Book of Lord Shang on the part of imperial literati, there were very few discussions of its dating throughout the imperial period. Traditional scholars focused—if they did at all—not on the text's dates but on its authenticity, i.e. whether or not it was composed by Shang Yang. The first known instances of questioning Shang Yang's authorship date from the Southern Song dynasty 南宋 (1127–1279). Zhou Duanchao 周端朝 (1172–1234) observed that the book refers to “a lot of later affairs” (duo fuhui houshi 多附會後事) and contains “redundant and excessive words” (fanlan yinci 汎濫淫辭).Footnote 36 He concluded that the book was inadequate as an introduction to Shang Yang's ideas; instead, it was better that one read the biography in the Records of the Historian. The poor literary quality of the text was also observed by Huang Zhen 黃震 (1213–1280), who claimed that the book is too “disordered” (fanluan 煩亂) to have been produced by a “gifted law official” such as Shang Yang: “its authenticity is doubtful and cannot be verified” (zhenwei dai wei kezhi 真僞殆未可知).Footnote 37 These comments did not inspire further studies of the text. It was only under the Qing dynasty that more substantial claims were made about the book's authenticity. Some scholars, such as Ma Su 馬驌 (1621–1673) and Wang Zhong 汪中 (1745–1794) remarked that chapter 15, “Attracting the People” (“Lai min” 徠民), is obviously late.Footnote 38 The compilers of the great imperial encyclopedic project, the Complete Books in the Four Treasuries (Siku quanshu 四庫全書), opined, in turn, that the entire Book of Lord Shang is late because the first chapter uses the posthumous title of Lord Xiao of Qin 秦孝公 (r. 361–338 b.c.e.). If Shang Yang was executed immediately after the lord's death, when could he have prepared this book?Footnote 39 These claims laid the ground for modern scholars doubting the book's authenticity.
During the Republican period, interest in the authenticity of early texts in general, and the Book of Lord Shang in particular, intensified, and the topic was addressed, even if briefly, by almost every eminent scholar of that age. The dominant tendency then was to point out obvious or perceived anachronisms in the text (i.e. references to events that happened after Shang Yang's execution) and dismiss the book in its entirety as a later forgery, unworthy of further study. In the heyday of the atmosphere of “doubting antiquity” (yigu 疑古), a cavalier attitude toward many early texts, including the Book of Lord Shang, prevailed. While a notable minority of scholars were engaged in detailed study of the book, they were clearly outnumbered by those who—echoing Hu Shi 胡適 (1891–1962)—considered Shang Yang an important reformer but not a thinker, and surely not an author of the Book of Lord Shang.Footnote 40
The dismissive attitude toward the Book of Lord Shang displayed in the majority of Republican-period studies reflected several common methodological problems. First, scholars were quite often eager to trumpet the discovery of yet another “anachronism,” which, under closer scrutiny, could not be accepted as evidence for the book's lateness. For instance, the fact that chapter 1, “Revising the laws,” refers to Lord Xiao of Qin by his posthumous appellation does not necessarily suggest that the chapter was penned after Lord Xiao's death; the ruler's name could have been updated by later editors or transmitters of the text. Similarly, the close similarity between chapter 13 of the Book of Lord Shang and chapter 53, “Chi ling” 飭令, of the Han Feizi does not presuppose, pace Qian Mu 錢穆 (1895–1990), Fu Sinian 傅斯年 (1896–1950), and others that the former is later than the latter; actually, the opposite is much more probable (see below).Footnote 41 But even when there are obvious anachronisms in the Book of Lord Shang (and these are undeniable), does this mean that the entire book is late, or just that a few chapters were composed long after Shang Yang's death? This simple question somehow escaped many of the grand masters of the Republican period, even though the fallacy of their methodology was pointed out quite early, for instance by Chen Qitian 陳啓天 (1893–1984).Footnote 42
The Republican-period legacy of dismissing the Book of Lord Shang remained influential in the West up to the end of the twentieth century, resulting in a minuscule number of studies of the Book of Lord Shang there.Footnote 43 In China and Japan, however, it yielded to a much more sophisticated approach. The dominant scholarly opinion nowadays is that the Book of Lord Shang—like many other pre-imperial texts—came into existence after a long period of accretion, during which chapters or paragraphs were added, altered, and possibly edited out.Footnote 44 With this understanding in mind, scholars are no longer engaged in the methodologically untenable attempt to date the entire book as a single entity, but rather try to date its individual chapters and determine their authorship. The roots of this approach are discernible already in the Republican period itself (e.g. works of Chen Qitian and Rong Zhaozu); later it was endorsed by Gao Heng.Footnote 45 Yet it was only with Zheng Liangshu's 鄭良樹 magnum opus (1989) that the concept of accretion can be said to have been fully absorbed into the study of the Book of Lord Shang.Footnote 46
Zheng Liangshu presented an insightful—albeit at times overtly speculative—theory of the formation of the Book of Lord Shang. He reads the Book of Lord Shang as a testimony to the evolution of “Shang Yang's school” (Shang Yang xuepai 商鞅學派). Zheng's highly detailed study outlined what he considers traces of ideological evolution among Shang Yang's followers as reflected in changing emphases in different chapters of the treatise. Zheng's publication was almost immediately followed by another magnum opus by Yoshinami Takashi 好并隆司, who proposed an alternative scenario for the book's formation.Footnote 47 Notably, both scholars asserted that the vast majority of the chapters were produced after Shang Yang's death, although they differed with regard to certain details.
In recent years, attempts to date individual chapters of the Book of Lord Shang and to assert their authorship continue. The current tendency, reflecting the general mood of “trusting antiquity” (xin gu 信古) which is now prominent among Mainland scholars, is to assert that most of the chapters were penned by Shang Yang, with only a few added after his death. The most radical representative of this trend is Zhang Jue,Footnote 48 who considers just one single chapter (15, “Attracting the people”), as spurious, and regards all the rest as authentic. Other scholars who recently addressed the dating question are less radical, but in general the dating they propose for most of the chapters is markedly earlier than that put forward by Zheng Liangshu and Yoshinami.Footnote 49
In general, I think that the accretion hypothesis proposed by Zheng Liangshu, Yoshinami Takashi, and other scholars can advance our knowledge of the Book of Lord Shang, were we to avoid certain methodological pitfalls embedded in its simplistic utilization. To begin with, recall that most of the currently extant Warring States-period texts bear the imprint of at least four contributors: the original author, his disciples and followers, later editors, such as Liu Xiang and his associates, and also manifold copyists and transmitters, especially those who, under the Han dynasty, transcribed the pre-Qin texts into a “modern” Han script. The degree of intervention of each of the two latter actors in the text's content is debatable. I accept Sarah Allan's view that the Han copyists' alterations of the Warring States-period texts were “probably not intentional forgeries but editorial changes and misguided attempts to restore an original text or meaning,” even though in some cases even a minor change could become ideologically consequential.Footnote 50 As for later editors, their impact could in certain cases be considerable, ranking from a new arrangement of the text, to excision of supposedly “duplicate” or “spurious” chapters and paragraphs, to substitution of barely legible or allegedly “wrong” characters with new ones, to—most annoyingly—adding new textual segments. In a few cases, later editors identify themselves and their rearrangements;Footnote 51 but most of them remain silent, such as in the case of the Book of Lord Shang. This silence does not rule out, however, the possibility of later editors and transmitters' tampering with the text's content.
Nor should we treat every chapter a priori as an organic unit to be analyzed in the context of the text's accretion. Actually, chapters differ considerably in their nature. Some were produced as separate essays and circulated independently for a considerable period of time before being put together with other essays to form a “book” such as we have.Footnote 52 These “organic” chapters have a distinct style and are normally well organized, with an identifiable beginning and conclusion; in the case of the Book of Lord Shang many of these might have originated from Shang Yang's and his followers' court memorials (see, e.g., chapters 6, 9, 15, 17, 23, 24, and 25). Yet in other cases, a chapter is just a patchwork prepared by later editors from independent paragraphs (zhang 章) associated with a putative author. Chapters 13 and 18 of the Book of Lord Shang may well represent this “composite” type. It makes perfect sense to try to date the organic chapters, but surely not the composite ones.
It is with these understandings in mind that I shall proceed now to present my hypothesis about the dating of some of the book's chapters. I shall try to avoid inevitably speculative attempts to date every single chapter or to guess who its author was. Yet I believe we have enough data to determine with relative precision the dating of at least some of the book's chapters, and to propose tentative dating for many more. At the current stage of our knowledge, any definitive answer regarding the process of the book's accretion is impossible, but even a partial answer will advance us considerably toward a better understanding of the nature and the content of the Book of Lord Shang.
The Text's Dating: A Tentative Reconstruction
Eighty years ago, Chen Qitian criticized scholars engaged in debates over the dating of the Book of Lord Shang for their lack of methodological rigor.Footnote 53 In the generations that have passed since then, the situation has improved, but a single commonly accepted methodology of assessing the dating of pre-imperial texts and of their component chapters is still not in sight. Creating such a methodology goes beyond the ambitions of the present study; as a preliminary step I shall try carefully to outline different criteria employed and their relative weight in assessing the dates of individual chapters. In particular I shall try to explain why “internal” criteria (e.g. historical information contained in the text, its language, and, whenever appropriate, interrelations among its chapters) are by far superior to “external” criteria (e.g. references or citations of the text elsewhere).Footnote 54 I hope that the methodological transparency adopted in what follows will encourage further research that will fine-tune my findings.
A. Historical Information
The first and by far the most convincing criterion for dating a chapter is that of relating its content to what we know of the history of the Warring States period. For instance, it is clear that if a chapter invokes events that occurred after Shang Yang's death, mentions personalities who lived after Shang Yang, or applies terms that were not in existence during his lifetime, it could not have been be authored by him. Yet historical data can not only prove a late date, but also suggest a relatively early one. For instance, since many chapters discuss the ways to unify All-under-Heaven, but not a single one (with the possible exception of chapter 26, “Fixing divisions” [“Ding fen” 定分]) evinces any knowledge of the eventual imperial unification of 221 b.c.e., it is plausible that most or all of the chapters were composed before the establishment of the Qin Empire. Similarly, insofar as no military chapter refers to cavalry, it is likely that they were produced before the third century b.c.e., by which time cavalry had gained importance in the world of the Warring States, and in Qin in particular.Footnote 55 In such cases, the argumentum ex silentio can be meaningfully employed.
Of all the chapters of the Book of Lord Shang, chapter 15, “Attracting the people,” is one of the most easily datable in the entire corpus of pre-imperial texts. Judging from the events it mentions (e.g. the Changping 長平 campaign against the state of Zhao 趙, 262–260 b.c.e.), and from what it does not mention, i.e. Qin's unstoppable expansion from the early 230s b.c.e., it must have been produced in between these dates. There are other indicators of its dating around 250 b.c.e., which I shall mention below; but here we can move to other chapters.
Another clearly anachronistic (i.e. post-Shang Yang) chapter is the last one, 26, “Fixing divisions.” It depicts an administrative system and employs nomenclature which came into existence only on the eve of the imperial unification. For instance, it mentions the offices of chengxiang 丞相 (prime minister, established in 309 b.c.e.) and yushi 御史 in the meaning of “chief censor” (a meaning acquired on the eve of unification).Footnote 56 In marked distinction to the rest of the book, chapter 26 focuses not on the means to control a single state as a preliminary step before the future unification, but rather on how to rule the entire sub-celestial realm. Since this realm is depicted as comprising both autonomous territories of “regional lords” (zhuhou 諸侯) and centrally ruled commanderies and counties (junxian 郡縣), many scholars have proposed an early Han date for this chapter. Yet this need not be the case: as has been observed by Ōkushi Atsuhiro among others,Footnote 57 on the eve of unification Qin treated some of the neighboring polities ruled by regional lords as its dependencies, which means that the chapter could have been composed back then. In any case, the chapter could not have been produced long before 221 b.c.e., and is probably the latest chapter in the entire Book of Lord Shang.
Two other chapters contain obvious anachronisms. Chapter 9, “Implementing laws” (“Cuo fa” 錯法) mentions a Qin strongman, Wu Huo 烏獲, who was active three decades after Shang Yang's death (Wu Huo was executed in 306 b.c.e.).Footnote 58 The final section of chapter 20, “Weakening the people” (“Ruo min” 弱民), again mentions Wu Huo, and, in addition, refers to a series of events in the state of Chu 楚 that occurred prior to 278 b.c.e. Moreover, inaccuracies in the depiction of the events show that the section under discussion (20.11) was probably composed much later than 278 b.c.e. Footnote 59 Yet, as has been noticed by most observers, this section was in all likelihood misplaced in chapter 20, because it is not related either topically or stylistically to the rest of the chapter. Putting chapter 20 aside for the time being, we may conclude that at least three chapters (9, 15, and 26) were obviously composed long after Shang Yang's death.
On several other occasions, historical data allow us to conclude that certain chapters were penned relatively early, i.e. during Shang Yang's lifetime. One of the clearest examples is chapter 2, “Order to cultivate wastelands” (“Ken ling” 墾令). This chapter appears to reflect the situation of an old aristocratic system that was current in the state of Qin prior to Shang Yang's reforms. It remains conspicuously silent about ranks of merit, which were Shang Yang's single most significant innovation;Footnote 60 instead, when it refers to elite members, it cautions against the power of hereditary nobles (dafu 大夫) (2.14), and of the heads of noble lineages (jia zhang 家長) (2.7), i.e. groups that disappeared from Qin's social landscape in the aftermath of Shang Yang's reforms and are never mentioned again in the Book of Lord Shang. Elsewhere, the chapter warns against exemptions from taxation and labor services granted to minor sons of the nobles and to the merchants' servicemen (2.13, 2.18), which again hints at the situation in pre-reform Qin society. Besides, some of the chapter's statements, e.g. “when the emoluments are bountiful, taxes are abundant” (2.4), seem likewise to hint at the pre-reform situation, in which the nobles' income derived directly from the subordinate population, whose tax quotas could be adjusted by the master.Footnote 61 If this is the case, then, again, the chapter refers to a dispersed mode of rule that characterized the state of Qin before the reform era.Footnote 62 All these suggest that chapter 2 is indeed one of the earliest in the Book of Lord Shang.Footnote 63
Another possibly early chapter is 12, “Military defense” (“Bing shou” 兵守). It starts with enumerating the difficulties faced by a state that has to combat enemies on the four frontiers; this state has to engage in defensive warfare and compensate for its strategic vulnerability with the ability to utilize its human resources fully. The situation of the state surrounded by enemies on the four sides, and its comparison with another state which backs the sea (presumably Qi 齊), is strongly indicative of the chapter having been written in the state of Wei 魏, where Shang Yang's career started.Footnote 64 If this chapter was indeed produced by Shang Yang or a member of his circle, then it should be a very early product, preceding Shang Yang's departure to Qin in 360 b.c.e. It will also explain the markedly defensive focus of the chapter, which is at odds with other military chapters of the Book of Lord Shang and with the known facts from Shang Yang's biography.Footnote 65
Moreover, comparing the “Military defense” chapter with the so-called defensive chapters of the Mozi composed probably in the latter half of the Warring States period reflects clear differences in the date of their composition. This system of complete mobilization of women, elderly, and the infirm in the Mozi 墨子 appears incomparably more sophisticated than in the Book of Lord Shang.Footnote 66 The “Military defense” chapter does envision mobilizing everybody, but its recommendations remain very rudimentary (for instance, it emphasizes primarily the need to prevent contacts among the three armies staffed by adult men, by adult women, and by the “elderly and the infirm”: but would it be possible to prevent these contacts in the cramped confines of a defended fortress?). It is likely then that the chapter reflects early experiments with complete mobilization of the entire populace to defend the fortress, and that it is datable to the early Warring States period. All these allow the placement of chapter 12 among the earliest in the book.
The third case in which the chapter's content suggests an early date of composition is chapter 6, “Calculating the land” (“Suan di” 算地), one of the ideological centerpieces of the text. The chapter depicts the composition of lands under the Qin's control as follows:
故為國任地者, 山林居什 一, 藪澤居什一, 谿谷流水居什一, 都邑蹊道居什{一, 惡田居什二, 良田居什}Footnote 67 四。此先王之正律也。
Hence, in ruling the state and making use ofFootnote 68 the land, the correct standard of the former kings was: mountains and forests occupy one-tenth; swamps and marshlands occupy one-tenth; valleys, dales and running rivers occupy one-tenth; towns, settlements, paths and roads occupy {one}-tenth; {infertile fields occupy two-tenths, fertile fields occupy} four-{tenths}. (6.2)
The text continues:
小畝五百, 足待一役, 此地不任也。方土百里, 出戰卒萬人者, 數小也。
The territory of five hundred small mu is enough to provide for one serviceman; [yet] this means that the land is not properly utilized. A territory of one hundred li squared can provide for ten thousand soldiers: [yet] the number is [still] small. (6.2)
The “correct standard of the former kings” referred to in this text reflects what might have been a standard view of land division in the Loess Plateau during the Warring States period; similar calculations appear in several other texts.Footnote 69 The relatively high proportion of cultivated areas in comparison to mountains and rivers clearly suggests that the text was authored before the Qin conquest of Sichuan c. 316 b.c.e., which radically altered the nature of the terrain under Qin's control (see more below).Footnote 70 Yet what is more important for the text's dating is the author's dissatisfaction with the current situation in which a household possesses 500 small (100 paces long) mu 亩, while 100 li squared (i.e. approximately 1,600 km2) could provide only for 10,000 soldiers.Footnote 71 This dissatisfaction is related to one of the important agricultural reforms allegedly initiated by Shang Yang: replacing a small mu with a large one (240 paces long), and the subsequent determination that the standard allotment to one household should be 100 large mu (c. 100 × 461 m2 = 4.61 ha) rather than 500 small mu (c. 500 × 192 m2 = 9.6 ha) under the previous system.Footnote 72 Since we know from the Qin Statute on Land, unearthed from Tomb 50 at Haojiaping, Qingchuan 青川郝家坪 (Sichuan), that by 309 b.c.e. the large mu was the standard measurement of agricultural fields, it is clear that the text of chapter 6 must be considerably earlier.Footnote 73 It may not be too bold to suggest that chapter 6 reflects the situation during the period of the promulgation of Shang Yang's reforms, when a smaller mu was still current in the state of Qin.
B. The Chapter's Language
Back in 1926, Bernhard Karlgren published his seminal study “On the Authenticity and Nature of the Tso Chuan,” in which he explored linguistic differences among several major pre-imperial texts as a possible indicator of their dating and authenticity. This pioneering study—despite its manifold flaws, which were inevitable in an age before concordances and digital resources—became hugely influential, spurring heated debates, and encouraging many follow-ups.Footnote 74 Duyvendak, in particular, attempted to employ Karlgren's method in dealing with the chapters of the Book of Lord Shang, yet his effort did not yield convincing results. Now, almost a century after Karlgren's study, the advance in the field remains minuscule, especially insofar as the Warring States-period literature is concerned. It seems that there are simply no unequivocal grammatical parameters that would allow us distinguish between, e.g., fourth- and third-century-b.c.e. texts. So frustrating is the situation that some scholars came to a simple conclusion: “it is impossible to date pre-Han texts with any degree of accuracy.”Footnote 75
In an earlier study, I tried to find another parameter for temporary changes in the language of pre-imperial texts, namely their specific lexicon. For instance, there are a few terms and compounds which are non-existent in the fifth-century-b.c.e. texts but become ubiquitous thereafter (e.g. “benevolence and righteousness” [renyi 仁義], “myriad things” [wanwu 萬物], “ten-thousand-chariot-[strong state]” [wansheng 萬乘], or crossbow [nu 弩], and crossbow-related terms such as ji 機, “trigger”). Other terms are not attested to before the late fourth century b.c.e., but dominate the third-century-b.c.e. texts: e.g. “pattern” or “principle” (li 理), the pair yin-yang 陰陽 in its meaning of basic cosmic forces or complementary opposites, or the term “plain-clothed” (buyi 布衣, referring to poor members of the shi 士 stratum). Insofar as we speak of terms that became fairly widespread in political and philosophical texts of the late Warring States period, their appearance or the lack thereof may serve as a useful indicator of the text's date.Footnote 76
Back then, I analyzed the language of the Book of Lord Shang, showing that, overall, it corresponds to the fourth-century-b.c.e. layer.Footnote 77 The text never employs such terms as buyi or yin-yang as cosmic forces or binary opposites. The term li 理 appears only four times, thrice in a single passage in chapter 18, “Charting the policies” (“Hua ce” 畫策) (18.7); overall it seems that, at the time of the compilation of the bulk of the Book of Lord Shang, this term had not yet become common in intellectual discourse.Footnote 78 This allows the tentative conclusion that most of the chapters of the book were composed before c. 300 b.c.e. But can we fine-tune this method to discern the date of individual chapters?
Lexical changes may be a fairly efficient means of distinguishing the date of the composition of the bulk of an ancient text (or its Ur-text), yet their efficiency diminishes when we deal with smaller textual units, such as individual chapters. The very brevity of a chapter (usually around one thousand, and never more than two thousand, characters in the Book of Lord Shang) makes the sample too small to be reliable. The smaller the sample, the greater is the danger that an inadvertent “correction” of an illegible character by, e.g., a Han-dynasty scribeFootnote 79 will add a third-century-b.c.e. character into a fourth-century-b.c.e. text, skewing my results. Nonetheless, with due caution some meaningful lexical differences among the chapters can still be discerned.
One of the most significant differences among the chapters of the Book of Lord Shang is the difference in the ruler's appellations. While most chapters apply the neutral “ruler” (jun 君) and “sovereign” (zhu 主), a few others employ “king” (wang 王), which suggests that they were produced after the appropriation of the royal title by Lord (later King) Huiwen of Qin 秦惠文王 in 325 b.c.e. In addition to chapter 15, the lateness of which has been clarified already, these include chapter 17, “Rewards and punishments” (“Shang xing” 賞刑), and possibly chapter 19, “Within the borders” (“Jing nei” 境内), although in this latter case the appearance of the term wang is disputed.Footnote 80 Chapter 26 speaks of the ruler as a “Son of Heaven” (tianzi 天子), which adds yet another indication of its provenance around the age of the imperial unification of 221 b.c.e.Footnote 81
Variations in the ruler's appellations bring us to another interesting difference among the chapters of the Book of Lord Shang: their preoccupation with the concept of the True Monarch. As I have argued elsewhere, from the Middle Warring States period on, many texts focus on this figure of an ideal ruler who will be able to unify All under Heaven under his aegis and bring about universal peace and stability.Footnote 82 The Book of Lord Shang is one of the earliest texts in which this notion figures prominently; yet the intensity of discussion differs considerably among various chapters. Some chapters, most notably 3, “Agriculture and warfare” (“Nong zhan” 農戰), and 4, “Eliminating the strong,” as well as those related to chapter 4 (see below), abound with promises for a ruler who would heed the authors' advice to become the True Monarch; in others, including all the identifiably late chapters, these pronouncements disappear. It is likely that the difference again revolves around 325 b.c.e.: after the Qin rulers adopted the title king/monarch (wang 王), it was no longer politically advisable to emphasize the ruler's need to become a True Monarch, which would imply that his current title was fraudulent. This suggests that chapters 3 and 4 and a few others, such as 7, “Opening the barred,” represent ideas and knowledge of the mid-fourth century b.c.e.
A word of caution is needed here. Lexical changes can rarely be used with precision, and they leave much room for doubt. A good case of such doubt is chapter 18 of the Book of Lord Shang. On the one hand, its fondness for the term li (pattern, principle), which appears there thrice (18.8), is indicative of a relatively late date: no earlier than the late fourth century b.c.e., which is the dividing line for the sudden outburst in the usage of li in philosophical texts. On the other hand, this chapter employs the term wang as a verb, in a mode which is reminiscent of earlier chapters of the Book of Lord Shang (18.3). Possibly, the composite nature of this chapter explains why passages composed at different periods appear side-by-side here. In any case this chapter, and several other chapters, lack clear lexical indicators that would help to determine their dating. More studies need to be done to advance our understanding of linguistic or lexical changes in Warring States-period texts, and to allow meaningful employment of this method.
C. Interrelations among the Chapters
The Book of Lord Shang is not a haphazard collection of disparate treatises; its chapters share not only fundamental ideas but even whole chunks of text, from one or two sentences to lengthy passages. These interrelations are often assumed to indicate the proximity in the chapters' date of composition,Footnote 83 but this is not necessarily the case. In some cases it is conceivable indeed that two similar or identical chunks of texts were put by a single person into two more-or-less simultaneously composed essays (just like a modern scholar can repeat the same passage in two or more articles). This may serve as a plausible explanation for similarities among a few sections in the Book of Lord Shang (e.g. almost verbatim repetitions of 3.5 in 4.3, of 3.7 in 4.5, or what appears a direct citation of a sentence from 6.6 in 7.2 preceded by “therefore it is said” 故曰). Yet it is equally possible that similar paragraphs were assembled into composite chapters by later editors (which may be the case of parallels between chapter 4 and 13, for instance). Moreover, as I shall demonstrate below, at times an almost identical textual segment can serve quite different ideological needs in two different chapters. Yet on other occasions an apparent similarity among a few chapters may reflect more complex patterns of intellectual development within “Shang Yang's school.”
One of the most interesting features of the Book of Lord Shang is the existence of several clusters of chapters which are related to each other. In each of these clusters we may discern earlier and later chapters. The clearest instance is the cluster of three chapters: 4, “Eliminating the strong”; 20, “Weakening the people”; and 5, “Explaining the people” (“Shuo min” 說民). As was demonstrated by Meng Jifu 蒙季甫, the two latter chapters provide a line-by-line exegesis of the former.Footnote 84 It is not clear why the exegesis was divided into two chapters, or why the final sections of chapter 4 remain without exegesis, but the “canon-commentary” (sometimes called jingshuo 經說) relation among the three chapters is undeniable. It seems that chapter 4 was conceived as a canonical text, probably a summa of Shang Yang's wisdom, which deserved special commentarial treatment.Footnote 85 Clearly, then, chapter 4 is earlier than its two exegetical chapters, and it is likely that it belongs to an earlier layer of the Book of Lord Shang.
To the cluster centered on chapter 4, we should add chapter 13, which is closely related to chapter 4, and which is reproduced to a large extent in chapter 53 of Han Feizi. The relations between the three have been analyzed by Zheng Liangshu, who demonstrated convincingly that chapter 4 is earlier, chapter 13 is later, and chapter 53 of the Han Feizi is derivative of chapter 13 of the Book of Lord Shang. I shall not repeat Zheng's detailed analysis, the basic parameters of which I accept; but my conclusion differs slightly from his.Footnote 86 I think that chapter 13 may be an alternative version of chapter 4; a kind of a draft that is less organized than chapter 4, less forceful in its style, and also is less abusive in its rhetoric. Possibly it is a composite chapter, which contains segments of chapter 4 and later additions. What matters for our discussion of the chapters' dating is that chapter 4 clearly belongs to an earlier stratum than three others, while each of these three reflects attempts by anonymous members of “Shang Yang's school” to explicate some of the enigmatic statements of chapter 4, but also, at times, to moderate its harsh messages.Footnote 87
Another, less evident, cluster comprises the three chapters that deal with land utilization: 2, 6, and 15. Rong Zhaozu, who noticed their similarities, opined that they were authored simultaneously, and dated all of them to post-250 b.c.e., because this is the unequivocal terminus ante quem non for chapter 15.Footnote 88 A more careful reading of these chapters not only undermines this uniform dating, as I have demonstrated above, but also calls for more caution when comparing ostensibly similar accounts. For instance, section 15.1 repeats almost verbatim the depiction of a supposedly standard land division from chapter 6.2 (cited above). Yet while in 6.2 the division of land in which 60 percent of the territory is arable is considered a norm, the authors of chapter 15 complain bitterly that “arable lands are just two-tenths” of the territory under Qin's control.Footnote 89 In all likelihood, this reflects the new realities in the aftermath of Qin's expansion into Sichuan and the Han River valley between 316 and 278 b.c.e. In the mountainous terrain of these areas, the old divisions based on Loess Plateau realities no longer worked. Moreover, the two accounts also reflect sharply different densities of population. In 6.2, a unit of 100 li squared is supposed to provide for 10,000 soldiers (zu 卒) or servicemen (yi 役), which refers to 10,000 households or about fifty thousand people. In chapter 15.1, in contrast, the same territory is occupied by 50,000 laborers (zuo fu 作夫), the term which refers to every able-bodied man or woman (aged 15 to 60), who constituted 60–70 percent of household members. In this case, we can speak of 70–80,000 expected inhabitants of a unit of 100 li squared: a 40–60 percent rise in the population in the wake of increased agricultural productivity in the second half of the Warring States period.
These differences are revealing. They not only provide additional proof for the relative earliness of chapter 6 and a post-260 b.c.e. composition of chapter 15, but also show how the same paragraph in two chapters may be utilized to convey different ideas and can reflect different realities. This warns against simplistic attempts to consider similarity between the two chapters as immediately reflecting their closeness in terms of authorship and dates.Footnote 90
The third, and less obvious, cluster consists of military chapters (10–12 and 19), which are dedicated to practical issues of combat. As these chapters (especially 10 and 11) are very badly damaged, it is difficult to discuss the nature of their interconnectedness. However, a comparison between chapters 12 and 19 is revealing. Both discuss siege warfare, but the differences are huge. Chapter 12, ”Military defense,” focuses on a purely defensive strategy. It shows how the full mobilization of the besieged will nullify the advantages of the assaulting army, and doom this army to inevitable failure. In marked distinction, chapter 19 is concerned with a purely offensive strategy. The attackers will overpower the defenders thanks to the proper use of positive incentives for valiant head-cutters and sappers, who will undermine the enemy's walls. The differences could not be greater. They strongly suggest that chapter 19 could not have been produced at a very early stage of the composition of the Book of Lord Shang, as is often proposed,Footnote 91 though it is still possible that it was authored by Shang Yang. Examining other chapters in the same treatise from a comparative perspective may thus allow us to correct many arbitrary conclusions.
D. Intellectual and Stylistic Consistency
Descending the ladder of reliable dating criteria, we come now to the issue of intellectual and stylistic differences among the chapters. Scholars have long noticed that while the Book of Lord Shang is fundamentally consistent in its outlook and, to a lesser extent, in its style, there are still observable differences among the chapters that can be employed to fine-tune their dating. In particular, Zheng Liangshu and Yoshinami Takashi used these differences to outline the evolution of the views of “Shang Yang's school.”Footnote 92 Yet while I do accept the validity of much of their analysis, I think that many of their conclusions were premature.
There are several problems with focusing excessively on the book's intellectual and stylistic consistency or the lack thereof. First, what exactly constitutes “consistency” is hard to measure. Some parameters that one scholar may consider important may be rejected by another. For instance, Zheng Liangshu observed what he took to be an essential divide in the Book of Lord Shang between chapters (e.g. 14, 17) that emphasize the need to combine harsh punishments for wrongdoers with lavish rewards for meritorious soldiers and tillers, and other chapters (e.g. 4 and those of its cluster, as well as chapters 7, 8, and 18) that advocate harsh punishments but minimal rewards. For Zheng, these differences suggest that the second group of chapters was not produced by Shang Yang and does not reflect his views.Footnote 93 This is a questionable assumption, not only because it is based on a highly tendentious interpretation of the “real Shang Yang's thought,” as supposedly reflected in Sima Qian's biography of Shang Yang, but also because the very issue of a proper balance between punishments and rewards was not necessarily a major point of discord among Shang Yang's followers. Authors could modify their recommendations in different circumstances and when facing different audiences, without feeling that they deviated from fundamental views which focus on rewards and punishments as the two handles through which the populace could be manipulated.
Second, even when differences among the chapters are undeniable, these do not necessarily mean that the chapters were produced by different authors. Recall that books associated with the leading thinkers of the Warring States period were not produced as coherent treatises. Rather, they typically represent something akin to the collected oeuvres of a thinker: a combination of different essays composed in different circumstances at different stages of a thinker's life, and to which works by a thinker's disciples and followers were usually added. Even among the chapters penned by a single thinker, differences can be expected. An author could improve his style and modify his views during his career; or he could adapt the style and the content of his proposals to meet specific demands of the chapter's addressees. Paul R. Goldin's insightful observation that different emphases in different chapters of the Han Feizi may reflect different audiences targeted by each of the chapters is equally valid for the Book of Lord Shang.Footnote 94 Demanding complete ideological and stylistic uniformity in these chapters is then incongruous. It becomes counterproductive when scholars first presuppose how the text evolved ideologically and then date individual chapters. The result is inevitably speculative, and at times fairly misleading.Footnote 95
Speaking of the text's literary qualities, scholars such as Zheng Liangshu presuppose a kind of neat stylistic progress from earlier unpolished chapters of the Book of Lord Shang, such as 2 and 19, to more literally engaging chapters such as 17, 15, or 24. The difference between the two groups is undeniable: for instance, while chapter 2 is dull and repetitious,Footnote 96 the latter chapters are better structured and employ a much broader repertoire of argumentative techniques. Yet while it is plausible that chapters composed at a later stage of the Book of Lord Shang's formation would develop more articulate argumentation,Footnote 97 it is equally possible that differences in style reflect individual preferences and skills of the chapters' authors. Any attempt to date the chapters purely on the basis of stylistic criteria will remain highly speculative.
It is equally difficult to use differences in the chapters' content as a tool for determining the date of their composition. For instance, no fewer than three chapters (7, “Opening the barred,” 18, “Planning policies,” and 23, “Ruler and minister” [“Jun chen” 君臣]) narrate how the state was formed at the dawn of human history. The first two chapters presuppose the existence of maintainable social order in a primeval stateless society; it is just that accelerating demographic growth made this order unsustainable. Chapter 23, in contrast, associates stateless society with intrinsic violence and turmoil, which can be overcome only through the creation of political and social institutions. This latter version is closer to the mainstream views of Warring States-period thinkers. But does it represent an earlier approach of the book's authors, as argued by Yoshinami, or a later one, as assumed by Zheng Liangshu?Footnote 98 Any answer will remain merely a scholar's educated guess. We simply lack sufficient data to decide.
Elsewhere, however, differences in the chapters' content may indeed presuppose evolution of the authors' views, as is the case with administrative thought in the Book of Lord Shang. The issues of proper maintenance of the bureaucratic apparatus and of ruler–minister relations are generally of little interest to the authors of all the identifiably early chapters. These chapters do discuss the policy of appointment to government positions, but this is done exclusively through the prism of its impact on the population's motivation to engage in agriculture and warfare rather than its importance for preserving the ruler's authority vis-à-vis scheming ministers, as would be the case, e.g., in Han Feizi. Administrative issues are raised primarily in chapters 24 and 26 and to a lesser extent in chapters 14, “Cultivation of authority” (“Xiu quan” 修權), and 25, “Attention to law” (“Shen fa” 慎法). My personal feeling is that these chapters may belong to the later stratum of the Book of Lord Shang, but I do not have sufficiently compelling arguments to make this assessment convincing.
E. Relations to Other Texts
One final possible criterion employed to date the book and its chapters is identifying potential influences by other thinkers. Many Republican-period scholars, and some of the later ones (e.g. Zheng Liangshu and Yoshinami Takashi), argued that certain chapters bear an imprint of, e.g., Mengzi 孟子 (c. 380–304 b.c.e.), Laozi 老子, Xunzi 荀子 (c. 310–230 b.c.e.), Shen Buhai 申不害 (d. 337 b.c.e.), Shen Dao 慎到, and Han Fei. Accordingly, they postulated that the chapters under discussion post-date those thinkers' lifetimes. What exactly constitutes a proof of “external influence” is not well defined though. For some, it means fundamental ideological impact; thus, Yoshinami considers Shang Yang's ideal of “eradicating punishments with punishments” as a reflection of Laozi's notion of “non-action” (wu wei 無爲).Footnote 99 For others, discovering a certain term in a chapter of the Book of Lord Shang is sufficient to claim its association with a putative inventor or promoter of this term. Thus, shi 勢 (positional power, or power of authority) is associated with Shen Dao; shǔ 數 (method), and shù 術 (technique) are associated with Shen Buhai; the term yi 義 (righteousness, dutifulness, justice) is considered to reflect the impact of Mengzi, while the character cheng 誠 (sincerity) is seen as related to the Confucian text “Doctrine of the Mean” (“Zhong yong” 中庸).Footnote 100
None of these assertions appears convincing to me. For instance, the idea of “eradicating punishments with punishments” is broadly shared by many texts, and is traceable even to the canonical Classic of Documents (Shujing 書經);Footnote 101 surely it cannot be reduced to Laozi's influence alone. Even less tenable is Zheng Liangshu's (and his predecessors') identification of key terms in the Book of Lord Shang as reflective of other thinkers' influence. Terms of political and ethical discourse were normally utilized by more than one thinker (for instance, yi is an important term in the Mozi, i.e. it gained prominence long before the Mengzi); hence the very postulation that term X is reflective of an impact by thinker Y is simplistic. Besides, even if the impact of a certain thinker or text on the Book of Lord Shang can plausibly be assumed, this does not necessarily help in dating chapters of the Book of Lord Shang because the dating of the oeuvres of other pre-imperial thinkers remains equally contestable.
We do have parallels, however, between the Book of Lord Shang and a few other texts, such as Laozi (possibly referred to in 17.4 and 18.7) or Guanzi (paralleled, e.g., in 17.4, 23.3), but these parallels cannot help much in determining the chapters' dates simply because the dates of Laozi and Guanzi themselves are difficult to verify. On two occasions, the Book of Lord Shang apparently cites Xunzi (20.11) and Shenzi (i.e. Shen Dao, 26.5); both cases, though, are from demonstrably late textual segments and the citations themselves help little in fine-tuning these chapters' date. Of greater interest is chapter 10, “Methods of war” (“Zhan fa” 戰法), which appears to be engaged in intensive dialogue with the Sunzi 孫子.Footnote 102 Yet given the ongoing debate on the date of the Sunzi's composition, we cannot use this information to assert when chapter 10 was penned. Finally, we may ask whether or not chapter 24 reflects the impact of Shen Dao and Shen Buhai. On the one hand, this chapter is the only one in the Book of Lord Shang to place the terms shi and shǔ 數 at the center of its discussions. On the other hand, as noticed above, the very supposition that these terms are derivative from Shen Dao and Shen Buhai respectively is debatable; it is equally plausible that the prominence of both terms in chapter 24 reflects their general importance in the administrative thought of the Warring States period (recall that chapter 24 is one of the very few to focus exclusively on administrative issues). None of the cases above allows us to improve the dating of chapters and sections involved.
Preliminary Conclusion
Let us summarize our findings. First, it is clear that the Book of Lord Shang as a whole was produced before the imperial unification of 221 b.c.e. (the only possible exception being chapter 26). Second, the text's lexicon suggests that the bulk of it came from the fourth century b.c.e. Third, it is clear that the formation of the text was a result of a lengthy period of accretion which started in Shang Yang's lifetime and ended more than a century after his death. Let us now turn to individual chapters and their precise location within this time span.
We can date with relative precision two chapters (2, 12) to the earliest stage of Shang Yang's career, i.e. prior to and on the eve of his reforms in Qin (c. 360s b.c.e.). Other chapters which may come from Shang Yang's lifetime, and quite possibly from an early stage of his career, are 3 and 4. Chapters 6 and 7, the ideological centerpieces of the entire book, are probably later, but were in all likelihood penned during Shang Yang's lifetime or shortly thereafter. Chapters 13, 20, and 5 (all of which are related to chapter 4) were surely composed sometime after chapter 4, but we have no clear date of composition (aside from the separate section 20.11, which certainly came decades after 278 b.c.e.). The military chapters 10 and 11, as well as 19, definitely come from the fourth century b.c.e., as is suggested by their unawareness of the use of cavalry, but their precise dating is less clear.
Among the identifiably later chapters, the two latest ones (15, 26) are safely datable to the final decades before Qin's imperial unification in 221 b.c.e. Chapter 9 is likely to post-date 306 b.c.e., while chapter 17 surely dates from after 325 b.c.e. For no other chapter is decisive dating possible. Some chapters (e.g. 18) are composite in nature, and therefore speaking of a date for their composition is meaningless. Other chapters are either too short (8, 22, 23), or lack unequivocal indicators of the date for their composition (14, 24, 25). My tentative results are summarized in Table 1 in the Appendix below.
Epilogue: Hierarchy of the Dating Criteria
Above, I focused on the determinants of the Book of Lord Shang's dating proposed in earlier studies and tried to arrange them according to the degree of their importance. In particular, I argued that historical information contained in the text should be given a clear priority over other criteria in ascertaining the date of the text's composition. To demonstrate the potential usefulness of this claim to other Masters' texts, I want to focus briefly on an example that shares many similarities with the Book of Lord Shang, viz. the Mozi. Like the Book of Lord Shang, Mozi was neglected for almost two millennia before it was rediscovered by Qing philologists; it also suffers from considerable textual corruption; and most notably it is yet another instance of an “evolving text” in which different temporary layers can be discerned. Similarities aside, in recent decades Mozi has attracted an incomparably greater amount of attention from Western scholars than has the Book of Lord Shang; in particular, the relations among the triplet chapters that constitute its core were discussed repeatedly.Footnote 103 It is these debates that I want to address briefly below. Normally, the discussants point out distinctions between the chapters' vocabulary and in their usage of certain grammatical particles to postulate different authorship; then they turn to the chapters' philosophical content as the primary indicator of the sequence of their composition. Without ruling out the importance of this line of investigation, I want to point out its major weakness: namely, ignoring Mozi's historical information as the most reliable criterion for the date of the chapters' composition.
Take for instance chapter 19, the third in the “Contra aggression” (“Fei gong 非攻”) triplet. Not a few scholars have opined that it was likely to have been written later than the two other chapters of the triplet, dating it to the middle or end of the fourth century b.c.e. Footnote 104 Regrettably, the discussants ignored a singularly significant piece of historic information in this chapter. The chapter cites Mozi's anonymous rivals who argue that, contrary to Mozi's claims, the states of Chu, Yue 越, Jin, and Qi benefited from aggressive warfare: hence, despite their humble beginnings, they were able “to divide All-under-Heaven into four parts and possess it” (si fen tianxia er you zhi 四分天下而有之).Footnote 105 This list of four superpowers is puzzling. While Chu, Jin, and Qi were indeed the major powers from the Spring and Autumn Period (Chunqiu 春秋, 770–453 b.c.e.) on, inclusion of Yue, as well as exclusion of Qin, are exceptional to the Mozi.Footnote 106 The Qing commentator Su Shixue 蘇時學 (1814–1874) noted this peculiarity and directly related it to the text's dating, suggesting that Mozi did not witness the age of the Warring States when Yue declined, while Qin reappeared as the major power.Footnote 107
Newly available data confirm Su's insight. The recently published bamboo manuscript Xinian 繫年 from the Tsinghua (Qinghua 清華) University collection narrates the history of the late fifth century, showing that then four major powers indeed competed for supremacy under Heaven: the Jin–Yue axis fought (successfully) against the Qi–Chu axis. It was then that the state of Qin, plagued by internal struggles, reached the nadir of its power.Footnote 108 This peculiar situation lasted until the first decades of the fourth century b.c.e., by which time Jin was finally partitioned among its component states of Wei, Han 韓, and Zhao; Yue declined, while Qin restored its fortunes. This means that the situation narrated in the Mozi existed only during Mozi's lifetime (c. 460–390 b.c.e.) and in its immediate aftermath. Were a middle- or late fourth-century follower to have invented the chapter, or were it to have been heavily edited then, one might expect that the list of major powers would be altered, or at the very least Mozi would be attributed with prescience about Yue's imminent decline. That this is not the case suggests that chapter 19, and other chapters that speak of Yue's power, are closer to Mozi's lifetime and are more reliable in reproducing the Master's views than is often assumed.
A single example, of course, does not account for a full-fledged dating methodology, but it suffices to show that a discussion of the dates of a Masters' text should start with analyzing the text's content from the angles of political, military, administrative, economic, social, or legal history; only then can other criteria be considered. While it is always possible that a later forger would try to create an archaic-looking text, either by resorting to archaizing language,Footnote 109 or by utilizing historical facts from the past, the evidence suggests that it would be all but impossible for him to truthfully recreate the realities of the long bygone age.Footnote 110 Besides, as the richness of recently unearthed documents allows us to grasp details of sociopolitical developments during the Warring States period incomparably better than was previously possible, we can now evaluate with greater accuracy the dates of the texts' composition. Only after a text's historical content was taken into consideration can we move to its linguistic peculiarities, and then to differences among its chapters and their relations to other texts. Each of these criteria is, in turn, superior to the focus on the text's ideological content as a determinant of its dating.
To be sure, the sequence of the dating criteria outlined above has its limits. For instance, it fits well those texts that abound with historic information, but not those which focus on, e.g., ethical or metaphysical questions and do not contain verifiable historical data. Nor have I addressed a complex problem of the later transmitters' impact on the text's content. Some scholars stress the magnitude of this impact,Footnote 111 yet their arguments will remain insufficiently convincing unless they consider the text's historical information and the (im)possibility that it was forged by a later transmitter. More challenging are the instances of adding small but ideologically important segments to an earlier text, as is the case, for instance, with the Zuo zhuan:Footnote 112 these tiny additions can considerably skew my analysis. Historical information aside, much progress should be made before we are able to determine grammatical or lexical changes in pre-imperial texts with sufficient precision. Many problems related to the texts' dating will remain unresolved for the time being. Yet insofar as my discussion above may contribute toward renewed interest in this issue, then the major goal of the present article will be achieved.
Appendix