Confucius and the Analects Revisited: New Perspectives on Composition, Dating, and Authorship 9004382771, 9789004382770

Featuring contributions by preeminent scholars of early China, Confucius and the Analects Revisited: New Perspectives on

860 128 9MB

English Pages 314 [325] Year 2018

Report DMCA / Copyright

DOWNLOAD FILE

Polecaj historie

Confucius and the Analects Revisited: New Perspectives on Composition, Dating, and Authorship
 9004382771, 9789004382770

Table of contents :
Contents
List of Figures and Tables
Notes on Contributors
Introduction
Michael Hunter and Martin Kern
The Continuing Currency of the Lunyu
Our Position on the Lunyu
Of Rugs and Dominoes
The Contributions
Chapter 1
A Critical Overview of Some Contemporary Chinese Perspectives on the Composition and Date of the Lunyu
John Makeham
1 The Guodian Materials and the Dating of the Lunyu
2 The Shanghai Museum Strips, Intertextuality, and a Proto-Lunyu Corpus
Concluding Comments
Chapter 2
The Lunyu as an Accretion Text
Robert Eno
The Concept of an Accretion Text
Theories of the Lunyu as an Accretion Text
Japanese Sinology and the Accretion Theory
The Brookses’ Accretion Theory
The Significance of Order within Disorder in the Lunyu
The Historical Context for the Compilation of the Lunyu
The Emergence of Confucian Aphoristic Collections
The Likely Role of Qin Encyclopedism
The Ru Underground of the Early Han and the Canonization of Confucius’s Wisdom
Closing the Canon
Conclusion
Chapter 3
The Lunyu as a Western Han Text
Michael Hunter
The Title “Lunyu” 論語
Han 漢 Bibliography and the Limitations Thereof
The Evidence from Lunyu Intertextuality: Kongzi Quotations
Developments in Kongzi Quotation Practice
The Lunyu as a Layered Text
Reading the Lunyu as a Western Han Text
Conclusion
Chapter 4
Confucius and His Disciples in the Lunyu: The Basis for the Traditional View
Paul R. Goldin
Evidence from Intellectual History
The Evidence from Philosophical Vocabulary
Evidence from References to Other Philosophers
Chapter 5
The Lunyu, a Homeless Dog in Intellectual History: On the Dating of Discourses on Confucius’s Success and Failure
Joachim Gentz
Terms, Concepts, and Ideas in the Lunyu: Taking Ren 仁 as an Example
Portrayals of Confucius in the Lunyu: Taking Confucius’s Success and Failure as an Example
Success and Failure Outside of the Lunyu
Conclusion
Chapter 6
Confucius’s Sayings Entombed: On Two Han Dynasty Bamboo Lunyu Manuscripts
Paul van Els
The Dingzhou Analects
Tomb and Excavation
Tracings and Transcriptions
Chapters, Sections, and Graphs
Textual Differences
The P’yŏngyang Analects
Tomb and Discovery
Features of the Manuscript
Differences from the Received Analects
Provenance of the Manuscripts
When were the Manuscripts Copied?
Where were the Manuscripts Copied?
Why were the Manuscripts Copied?
Conclusion
Chapter 7
Manuscript Formats and Textual Structure in Early China
Matthias L. Richter
Hypotheses about the Influence of Manuscript Formats on Texts
The Extension of Texts
Correspondence between the Text and Its Carrier
Mise-en-Page
Conclusion
Chapter 8
Interlocutor Collections, the Lunyu, and Proto-Lunyu Texts
Mark Csikszentmihalyi
Two Early Versions of the “Interlocutor Text” Origin Story
“Interlocutor Texts” in Early China: The Case of Zengzi 曾子
The First “Interlocutor Collection” and the Case of the Missing Zengzi 曾子
Topical Consistency in the Lunyu
Chapter 9
Sima Qian’s Kongzi and the Western Han Lunyu
Esther Sunkyung Klein
Overview of Sima Qian’s Lunyu
The Shiji on Kongzi’s Disciples
The Shiji “Kongzi shijia”
Kongzi’s Multiple Roles in the Shiji
The Kongzi of Historical Texts
An Esoteric Kongzi
Conclusion
Chapter 10
Kongzi as Author in the Han
Martin Kern
Epilogue
Index

Citation preview



Confucius and the Analects Revisited

© koninklijke brill nv, leiden, 2018 | doi:10.1163/9789004382947_001

i

ii

Studies in the History of Chinese Texts Edited by Martin Kern (Princeton University) Robert E. Hegel (Washington University, St. Louis) Theodore Huters (University of California at Los Angeles) Ding Xiang Warner (Cornell University)

VOLUME 11

The titles published in this series are listed at brill.com/hct





Confucius and the Analects Revisited New Perspectives on Composition, Dating, and Authorship Edited by

Michael Hunter Martin Kern

LEIDEN | BOSTON

iii

iv



Cover illustration: Portrait of Confucius on p. [cxvi] in the following work: Confucius sinarum philosophus, sive Scientia sinensis latine exposita... Published/Created: Parisiis : Apud Danielem Horthemels, viâ Jacobæâ, sub Mæcenate, 1687. Princeton University Library, Call Number: (Ex) 2082.2687q. The Library of Congress Cataloging-in-Publication Data is available online at http://catalog.loc.gov LC record available at http://lccn.loc.gov/2018040916

Typeface for the Latin, Greek, and Cyrillic scripts: “Brill”. See and download: brill.com/brill-typeface. issn 1877-9425 isbn 978-90-04-38277-0 (hardback) isbn 978-90-04-38294-7 (e-book) Copyright 2018 by Koninklijke Brill nv, Leiden, The Netherlands. Koninklijke Brill NV incorporates the imprints Brill, Brill Hes & De Graaf, Brill Nijhoff, Brill Rodopi, Brill Sense, Hotei Publishing, mentis Verlag, Verlag Ferdinand Schöningh and Wilhelm Fink Verlag. All rights reserved. No part of this publication may be reproduced, translated, stored in a retrieval system, or transmitted in any form or by any means, electronic, mechanical, photocopying, recording or otherwise, without prior written permission from the publisher. Authorization to photocopy items for internal or personal use is granted by Koninklijke Brill nv provided that the appropriate fees are paid directly to The Copyright Clearance Center, 222 Rosewood Drive, Suite 910, Danvers, ma 01923, usa. Fees are subject to change. This book is printed on acid-free paper and produced in a sustainable manner.

Contents Contents

v

Contents

List of Figures and Tables vii Notes on Contributors viii x



Introduction 1 Michael Hunter and Martin Kern

1

A Critical Overview of Some Contemporary Chinese Perspectives on the Composition and Date of Lunyu 17 John Makeham

2 The Lunyu as an Accretion Text 39 Robert Eno 3 The Lunyu as a Western Han Text 67 Michael Hunter 4 Confucius and His Disciples in the Lunyu: The Basis for the Traditional View 92 Paul R. Goldin 5 The Lunyu, a Homeless Dog in Intellectual History: On the Dating of Discourses on Confucius’s Success and Failure 116 Joachim Gentz 6 Confucius’s Sayings Entombed: On Two Han Dynasty Bamboo Lunyu Manuscripts 152 Paul van Els 7 Manuscript Formats and Textual Structure in Early China 187 Matthias L. Richter 8 Interlocutor Collections, the Lunyu, and Proto-Lunyu Texts 218 Mark Csikszentmihalyi 9 Sima Qian’s Kongzi and the Western Han Lunyu 241 Esther Sunkyung Klein

vi 10 Kongzi as Author in the Han 268 Martin Kern Index 309 314

Contents

List of and Figures and Tables List of Figures Tables

vii

List of Figures and Tables 2.1 2.2 2.3 3.1 6.1 6.2 6.3 6.4 6.5 6.6 6.7 6.8 6.9 6.10 6.11 7.1 7.2 7.3.1 7.3.2 7.4.1 7.4.2 7.5.1 7.5.2

3.1 3.2 3.3 4.1 6.1 8.1 8.2

Figures Diagram of Lunyu book 1 51 Diagram of Lunyu book 8 51 Diagram of Lunyu book 19 52 The intertextual profile of the Lunyu in pre-Han texts 82 Bamboo fragment from the Dingzhou tomb 155 Tracings of the Dingzhou Analects 159 Photograph of the P’yŏngyang Analects 168 Photograph of the P’yŏngyang Analects 169 Select tracings of Sayings of the Scholars 175 Select tracings of Wenzi 176 Select tracings of Six Secret Teachings 177 The graph 之 in Dingzhou manuscripts 177 The graph 道 in Han dynasty manuscripts 178 The graphs 何 and 可 in Han dynasty manuscripts 178 The graphs 可 and 何 in the P’yŏngyang Analects 179 Varying height of characters in Guodian manuscripts 200 Varying density of writing in Guodian and Shanghai Museum manuscripts 202 Correction in Guodian *Zi yi 緇衣 205 Correction in Guodian *Yucong 語叢 4 205 Wuwei *Yili manuscript “Fu zhuan” 服傳, section of slips 1 and 2, verso 207 Wuwei *Yili manuscript “Fu zhuan” 服傳, bottom of slip 28, recto 207 Guodian *Zi yi 緇衣 manuscript, slip 47, recto: end of text and 二十又三 213 Layout of the Wangjiatai manuscript *Zheng shi zhi chang 政事之常 213

Tables Lunyu mentions in early sources 69 Kongzi attributions in texts after ca. 100 BCE 74 Kongzi attributions in texts prior to ca. 100 BCE 75 Mentions of Masters in “Masters” texts 106 Possible matching chapters in bamboo manuscript and received text 160 Order of key disciples in Lunyu 11/3 229 Order of key disciples in the Shiji 229

viii

Notes on Contributors

Notes On Contributors

Notes on Contributors Mark Csikszentmihalyi is the Marjorie Meyer Eliaser Chair of International Studies and Chair of the Department of East Asian Languages and Cultures at the University of California, Berkeley. He uses both excavated and transmitted texts to reconstruct the religions, philosophies, and cultures of early China. His books include Material Virtue: Ethics and the Body in Early China (2004), Readings in Han Chinese Thought (2006), and Religious and Philosophical Aspects of the Laozi (co-edited 1999). Paul van Els is University Lecturer of China Studies at Leiden University, where he teaches Classical Chinese language tutorials and courses on Chinese culture, history, philosophy, and religion. His publications include the monograph The Wenzi: Creativity and Intertextuality in Early Chinese Philosophy (2018), a volume coedited with Sarah A. Queen, Between History and Philosophy: Anecdotes in Early China (2017), and an illustrated, two-volume Dutch-language textbook of Classical Chinese, Van orakelbot tot weblog (2011, 2015). Robert Eno Prior to retirement, Robert Eno taught for many years in the Department of East Asian Languages and Cultures at Indiana University. His research has focused on early Chinese thought and history, and on philological scholarship, including work on oracle bone and bronze inscriptions. Joachim Gentz is Chair of Chinese Philosophy and Religion at the University of Edinburgh. His main research focus is on Chinese history of thought. His work crosses the disciplinary boundaries of Sinology, Religious Studies, Philosophy, and Cultural Studies. His publications include Das Gongyang zhuan (2001), Komposition und Konnotation (co-edited, 2005) Keywords Re-Oriented (2009), Understanding Chinese Religions (2012), Religious Diversity in Chinese Thought (co-edited, 2013), and Literary Forms of Argument in Early China (co-edited, 2015). Paul R. Goldin is Professor of East Asian Languages and Civilizations at the University of Pennsylvania. His research focuses on the cultural and intellectual history of

Notes on Contributors

ix

early China, roughly from the Bronze Age through the end of the Han dynasty. Other interests include gender and sexuality in Chinese history and the history of the Chinese language. He is the author of Rituals of the Way: The Philosophy of Xunzi (1999); The Culture of Sex in Ancient China (2002); After Confucius: Studies in Early Chinese Philosophy (2005); and Confucianism (2011). In addition, he has edited the revised edition of R.H. van Gulik’s classic study, Sexual Life in Ancient China (2003), and has edited or co-edited six other books on Chinese culture and political philosophy. Michael Hunter is an Associate Professor of East Asian Languages and Literatures at Yale University whose research interests are at the intersection of early Chinese thought and literature. His first book, Confucius Beyond the Analects, is a critique of the Analects-centric approach to the study of Confucius and a survey of the wealth of Confucius material outside of the Analects. Martin Kern is the Joanna and Greg ’84 P13 P18 Zeluck Professor in Asian Studies and Chair of the East Asian Studies Department at Princeton University. He also directs the International Center for the Study of Ancient Texts Cultures at Renmin University of China, Beijing, and serves as Co-editor of T’oung Pao. His publications on early Chinese literature, history, and textual culture include The Stele Inscriptions of Ch’in Shih-huang: Text and Ritual in Early Chinese Imperial Representation (2000), various edited volumes, and numerous articles and book chapters, such as the opening chapter of the Cambridge History of Chinese Literature (2010). Esther Sunkyung Klein is a Lecturer in the School of Languages and Cultures at the University of Sydney and a visiting researcher (2018) in the Department of Philosophy at the Australian National University. She is the author of Reading Sima Qian from Han to Song (2018). Her current research explores the intersection of early Chinese philosophy and medieval Chinese historiography. John Makeham is Chair and Director of the China Studies Research Centre at La Trobe University, Australia. He is a specialist in the intellectual history of Chinese philosophy. His principal publications include Lost Soul: “Confucianism” in Con­ temporary Chinese Academic Discourse (2008); Transmitters and Creators: Chinese Commentators and Commentaries on the Analects (2003); Name and

x

Notes On Contributors

Actuality in Early Chinese Thought (1994); and translations of Xiong Shili’s New Treatise on the Uniqueness of Consciousness (2015) and Xu Gan’s Balanced Discourses (2002). Matthias L. Richter is Associate Professor of Chinese and currently Chair of the Department of Asian Languages and Civilizations at the University of Colorado, Boulder. His research on early Chinese literature focuses on questions of rhetoric and redactional strategies, textual criticism, the formational history of texts, and the methodology of studying early Chinese manuscripts. Recent publications include the monograph The Embodied Text: Establishing Textual Identity in Early Chinese Manuscripts (2013).

Introduction

1

Introduction Michael Hunter and Martin Kern For more than two millennia, readers have turned to the Lunyu 論語 (Analects or the Selected Sayings) as an authoritative guide to the teachings of Kongzi 孔 子 (Confucius; trad. 551–479 BCE), the most important figure in the East Asian tradition. Insofar as Kongzi has stood for certain foundational values and practices, including learning (xue 學), reverence for the past (gu 古), ritual propriety (li 禮), the nobility of official service (shi 仕 / shi 士), and the interdependence of family virtues like filial piety (xiao 孝) with official virtues like loyalty (zhong 忠), the Lunyu has been a potent “initiation text” into that tradition, to borrow a phrase from Robert Eno—or, to quote one early witness, it is “the linchpin of the Five Classics and the mouthpiece of the Six Arts” (五經之錧鎋, 六藝之喉衿 也).1 To this day, the practice of introducing traditional China via the Lunyu continues in classrooms around the world. The Lunyu certainly lends itself to the role of gatekeeper text. As a guide to the quotable Kongzi, it is short (ca. 16,000 characters) and divided into five hundred or so bite-sized, easily memorized bons mots. Even its challenges are conducive to reader engagement. The text does not present Kongzi’s teachings in ways that a modern academic philosopher would recognize as rigorous. Logical connections between and across entries are implicit at best. Contra­ dictions abound. Entries of various formats (sayings, comments, dialogues, anecdotes, testimonia) are strung together indiscriminately with little or no context.2 The Lunyu is not disorganized so much as unconstructed, the overall effect of which is to invite, even demand, the active participation of readers in ways that few other classical texts do. The text also facilitates this process by tempting readers with the promise of “a single thread tying [Kongzi’s Way] together” (一以貫之; 4/15 and 15/3) or the challenge of reconstructing the “three [unexpressed] corners” (san yu 三隅) for every “single corner” (yi yu 一隅; 7/8) in the text itself. The Lunyu has played an especially important role in the development of early China studies in the modern era. Surveys of early Chinese thought or philosophy typically open with a chapter or section on the Lunyu as a foundational stage in the development of Warring States (453–221 BCE) thought. 1 Jiao 1987: 14. 2 On the challenges involved in recovering the “historical meaning” of the Lunyu, see Makeham 2002.

© Koninklijke Brill NV, Leiden, 2018 | doi:10.1163/9789004382947_002

2

Hunter And Kern

Likewise, a conventional algorithm for the intellectual historical analysis of early terms and concepts is to consider their use (or absence) in the Lunyu before turning to other, ostensibly later, sources. As a record of the teachings of the figure widely considered the earliest philosopher and “first teacher” (xianshi 先師), the Lunyu continues to anchor the contemporary imagination of the Warring States “Masters” (zhuzi 諸子). In the last century or so, many of the most important voices in the field of classical Chinese philosophy have done a great deal of “thinking through Confucius” and the Lunyu, to quote David Hall and Roger Ames.3 Within the history of philology, the Lunyu has provided fertile ground for the development of various critical methodologies. Whenever philologists set out to reorder a text’s juan 卷 (fascicles, chapters) or zhang 章 (paragraphs, entries) with the aim of sequencing its layers chronologically, wittingly or not they are following a path paved centuries earlier by Itō Jinsai 伊藤 仁斎 (1627–1705), Cui Shu 崔述 (1740–1816), and others who sought to identify Kongzi’s original teachings amid the Lunyu’s miscellanies.4

The Continuing Currency of the Lunyu

Nearly two decades into the twenty-first century, the field of Lunyu studies is as vibrant as ever. Since 2000, publications in English have included new translations by Edward Slingerland (2003), Pan Fu’en and Wen Shaoxia (2005), and Ann-ping Chin (2014); reprints of older translations by Arthur Waley (2000), Burton Watson (2007), Simon Leys (2014), and David Hinton (2014); collections of essays edited by Bryan W. Van Norden (Confucius and the Analects: New Essays; 2002), David Jones (Confucius Now: Contemporary Encounters with the Analects; 2008), Amy Olberding (Dao Companion to the Analects; 2014), and Michael Nylan (The Analects [Norton Critical Edition]; 2014); monographs by Daniel K. Gardner (Zhu Xi’s Reading of the Analects: Canon, Commentary, and the Classical Tradition; 2003), John Makeham (Transmitters and Creators: Chinese Commentators and Commentaries on the Analects; 2003), Amy Olberding (Moral Exemplars in the Analects; 2012), Henry Rosemont (A Reader’s Companion to the Confucian Analects; 2013), and Michael Hunter (Confucius Beyond the Analects; 2017); and many more journal articles besides. Publications in 3 Hall and Ames 1987. 4 On this history, see Eno’s contribution to this volume (chap. 2) and Kim and Csikszentmihalyi 2014. A spectacular demonstration of this approach is Brooks and Brooks 1998, which systematically reordered the text on the basis of certain (not universally shared) assumptions about the workings of an early Confucius school. Our own view of the Brookses’ project mirrors Eno’s.

Introduction

3

Chinese from the same period number in the thousands.5 The Lunyu has been introduced and reintroduced and re-reintroduced so many times now that one could write a dissertation just on the genre of the Lunyu introduction.6 Given this context, one might wonder whether the world needs yet another book on the Lunyu. The easy response is that there is clearly a market for such publications.7 In the People’s Republic of China, China Central Television’s broadcast of Yu Dan’s 于丹 lectures on the Lunyu in 2006 created a national sensation. The printed version of those lectures, Lunyu xinde 論語心得 (Confucius from the Heart: Ancient Wisdom for Today’s World), has sold millions of copies. Even more remarkably given the anti-Confucian campaigns of the Maoist era, the Communist Party has come to embrace Kongzi 孔子 and the Lunyu as a way of promoting its version of traditional Chinese values, now manifested in some five hundred “Confucius Institutes” all over the globe.8 In 2015 various speeches by Xi Jinping 習近平, the general secretary of the Communist Party of China and president of the People’s Republic of China, were collected and published with the title Xi Jinping: How to Read Confucius and Other Chinese Classical Thinkers.9 Quotations of the Lunyu also featured prominently in the opening ceremony of the 2008 Beijing Olympics, a landmark event in China’s emergence on the global stage. The vibrancy of the East Asian classical tradition, with the Lunyu as one of its crown jewels, is something even the most contrarian of Lunyu scholars can be grateful for.10

Our Position on the Lunyu

Nevertheless, the Lunyu’s canonicity is not the point of departure for our project. Instead, what motivates this volume is the editors’ belief that direct 5 6

7 8 9 10

In the CNKI Journal Translation Project database of Chinese academic journals, http:// gb.oversea.cnki.net (accessed January 2, 2017), a search for “論語” in paper titles since the year 2000 yields over six thousand results. Here one might contrast the Lunyu with the Zhouli 周禮 (Rites of Zhou) and the Shangshu 尚書 (Exalted Documents), the subjects of other volumes within the Studies in the History of Chinese Texts series; see Elman and Kern 2010; Kern and Meyer 2017. Despite their significance in the Chinese intellectual tradition, both texts have received embarrassingly little attention in Western Sinology. Schaberg 2001. On the resurrection of Confucius and Confucianism, see esp. Song 2003; Billioud and Thoraval 2015. Zhang 2015. The disadvantageous position of classical Indian studies versus classical Chinese studies is a cautionary tale in this regard; see Pollock 2009: 944–945.

4

Hunter And Kern

evidence of the Lunyu’s authority and even its existence is sorely lacking before the Western Han period (202 BCE–9 CE). As argued in Hunter’s contribution in this volume, the earliest sources to corroborate the existence and circulation of a Lunyu text date to the mid- to late Western Han, when the imperial dynasty began teaching it to princes and invoking it as an instrument of imperial legitimacy.11 The earliest sources to describe the Lunyu as a record of Kongzi’s teachings compiled by his students emerged decades after its adoption by the imperium.12 Moreover, representations and quotations of Kongzi prior to the Han period exhibit so few parallels with the Lunyu as to preclude the possibility of direct borrowing from a Lunyu text. To date, the only excavated or looted manuscripts to corroborate the existence of the Lunyu date no earlier than the mid-first century BCE. The wealth of non-Lunyu Kongzi material in earlier manuscript finds only makes the Lunyu’s invisibility in the Warring States period and the first decades of the Han dynasty that much more remarkable. Here we must acknowledge our debt to scholars whose skepticism regarding the traditional account of the Lunyu’s origins inspired our own. These include the entire tradition of critical Rongo 論語 (Analects) scholarship in Japan, from Takeuchi Yoshio 武内義雄 (1886–1966) and Tsuda Sōkichi 津田左 右吉 (1873–1961) to Kaneto Mamoru 金戸守; the works of Zhao Zhenxin 趙貞 信 (1902–1989) and Zhu Weizheng 朱維錚 (1936–2012) in China; and the writings of John Makeham, Mark Csikszentmihalyi, Christiane Haupt, and Oliver Weingarten.13 Of special note is Makeham’s 1996 article “The Formation of Lun­yu as a Book,” an enduring work of scholarship that established the contours of the revisionist position within English-language scholarship. When we step back to consider the available evidence in its totality, we conclude that the likeliest context for the creation of a Lunyu text is the Western Han dynasty. (This is the argument summarized in Hunter’s contribution and 11 12

13

See Csikszentmihalyi 2002; Hunter 2017: chap. 3. That view’s earliest expression is a fragment attributed to Liu Xiang 劉向 (79–8 BCE), the official charged by Emperor Cheng 成帝 (r. 33–7 BCE) in 26 BCE with organizing and cataloging the imperial library: “All twenty sections of the Lu Lunyu are fine sayings recorded by Kongzi’s disciples” (魯論語二十篇皆孔子弟子記諸善言也). This fragment appears in the preface to the Lunyu jijie (Sibu congkan ed., 1.1a), attributed to He Yan 何晏 (ca. 190–249). For the expanded “Yiwen zhi” 藝文志 (Treatise on Arts and Letters) account, see Hanshu 30.1716. See also Mark Csikszentmihalyi’s contribution to this volume (chap. 8). Takeuchi 1939; Tsuda 1946; Kaneto 1970, 1971, 1972, 1973, 1974, 1975, 1976, 1977, 1978, 1979, 1980; Zhao 1961; Zhu 1986, 1987; Makeham 1996; Csikszentmihalyi 2002, 2004; Haupt 2006; Weingarten 2010. We would also like to extend our thanks to Cameron Moore, whose survey of Japanese Rongo scholarship reminded us of the importance of Japanese scholarship in the twentieth century.

Introduction

5

in his recent monograph.)14 We believe that the Lunyu is both an imperial and a dynastic text. By that we do not just mean that the Lunyu is a product of the Han period. The Lunyu as a book came into existence under the intellectual, political, and social conditions of a unified imperial state governed by members of a dynastic lineage. Given the amazing breadth and dynamism of the “Kongzi” phenomenon in the Warring States period, the standardization of Kongzi quotation practice via the Lunyu would not have been possible prior to the Qin 秦 unification or the increasing centralization of the early Western Han, nor is there any evidence of such standardization prior to empire. This is not to say that all of the Lunyu was written from scratch in the Han; our inquiry is about the compilation date of the Lunyu as a book, not about the dating of its individual passages. With regard to the latter, we remain largely agnostic: given the paucity of parallels between the Lunyu and pre-imperial Kongzi quotations, evidence that a given saying circulated as a Kongzi saying prior to the Lunyu is not forthcoming in the vast majority of cases. We simply insist that inclusion in the Lunyu is not a marker of a saying’s antiquity or authenticity. Moreover, it was under the Western Han dynasty that the need for a quotable Kongzi canon became pressing, as emperors and subjects alike looked to justify their policies with reference to the sage behind the Five Classics, the newly established state-sponsored curriculum for the training of imperial officials. As noted by Hunter, echoes of Western Han recruitment edicts in the Lunyu further reveal its interest in the “selection” (lun 論) of talented and virtuous candidates, a key ingredient in the emerging imperial system. As a companion text to the Xiaojing 孝經 (Classic of Filial Piety), which established the broader social, political, and cosmological implications of xiao 孝 (filiality) for an early imperial audience, the Lunyu indoctrinated imperial scions in values conducive to dynastic continuity.15 In short, even if it existed in the Warring States period in some form, a possibility that can never be dismissed, such a text could not have had the authority it enjoyed under empire. The Western Han dynasty did not merely put its seal of approval on a preexisting text; in a very real sense, it “created” (zuo 作) the Kongzi canon it needed. This is a controversial position, one with which not all the contributors to this volume (or participants in the 2011 conference that precipitated it) would entirely agree. Despite our enthusiasm for revisionist Lunyu scholarship, we do not offer this volume as a definitive answer to the question of the Lunyu’s origins. To the contrary, we openly acknowledge the contested nature of our 14 15

Hunter 2017. It is not a coincidence that the two most important themes in Lunyu 1 are xue 學 (“learning”; 1/1, 1/6–8, 1/14) and xiao 孝 (“filiality”; 1/2, 1/6, 1/11).

6

Hunter And Kern

claims and the impossibility of proving once and for all a particular date for any part of the Lunyu, let alone for the collection as a whole. (As pointed out by Goldin and others in the present volume, the undeniable presence of preHan material in the Lunyu significantly muddies the distinction between a Han and pre-Han text.) No less for the Lunyu than for other early texts, early China scholars cannot presume to offer anything approaching definitive proof of its origins. Paradoxically, this is even truer today than it was a few decades ago thanks to the combination of newly available manuscripts, digital research tools (as in Hunter’s contribution), and broader comparative approaches to the study of ancient text cultures. To quote the introduction of another volume in the Studies in the History of Chinese Texts series: The certainty that the eminent philologist Bernhard Karlgren still felt, in the mid-twentieth century, when deciding on the interpretation of individual Chinese characters and words, is long gone: a wealth of new data from unearthed ancient manuscripts, together with more sophisticated conceptual approaches that are informed by neighboring disciplines and cross-cultural comparisons, especially the study of ancient Mediterranean texts, has made us far less sure of ourselves in evaluating the “right” choice for this or that Chinese character. … In this endeavor, we have been happy to trade false certainty for more interesting and productive questions and possibilities.16 What is true of individual Chinese characters is even more true for processes of textual formation in the ancient context. In the face of such uncertainties, the best one can do is to consider all the evidence at one’s disposal in the hopes of devising a provisionally workable hypothesis, all while explicitly acknowledging the tentative nature of one’s conclusions and welcoming new evidence as it becomes available. As a field, the worst thing we can do is to promise more certainty than our sources allow us. However much we would like to recover The Original Analects or The Authentic Confucius, that should not be the goal of (self-)critical textual scholarship.17 Given the multiplicity of voices on the topic, all of them making very strong cases one way or the other, there is a larger question here: how do we engage in controversy? Or, how does one collaborate in such a way that all perspectives are cherished? Irrespective of their intellectual, methodological, or disciplinary commitments, all our contributors started from the premise that the 16 17

Kern and Meyer 2017: 6. Brooks and Brooks 1998; Chin 2007.

Introduction

7

question of the Lunyu’s origins remains open and vital. Our goal is not to rebut other views or to establish a new orthodoxy but to create a space in which different hypotheses can be advanced, tested, and modified.

Of Rugs and Dominoes

The essays in this volume are further united by a willingness to consider the implications of rethinking the Lunyu’s origins. By virtue of its canonicity, the Lunyu is embedded in the modern imagination of early China in ways that few other texts are. Consequently, every hypothesis about the Lunyu’s nature, origin, context, and circumstances of composition and compilation implies wider assumptions about ancient Chinese textuality and its intellectual, social, material, and political contexts. To quote Mark Csikszentmihalyi, problematizing the Lunyu’s chronology and its relationship to the historical Kongzi “has the effect of pulling the rug out from under the usual narrative of what is often called the ‘history of thought’ (sixiang shi 思想史) of early China.”18 Or to use another metaphor: knock over the Lunyu and many other dominoes also fall. These include Destabilizing the traditional timeline of early Chinese thought anchored to the Kongzi of the Lunyu as the earliest master and/or the Lunyu as the earliest work of philosophy. Problematizing the Lunyu as a source of social, political, or linguistic realities in the early Warring States period.19 Complicating the study of Kongzi by removing the most convenient Kongzi canon. If not via the Lunyu, how does one go about reading and teaching the voluminous yet scattered corpus of early Kongzi literature?20 Dissolving the disciplinary divide between pre-Qin and early imperial thought. If the Lunyu did not exist or did not circulate widely prior to the Western Han period, then it can hardly be read as a foundational text of “pre-Qin philosophy” (xian Qin zhexue 先秦哲學). Whether any texts traditionally dated to the Warring States period can be read exclusively from a pre-Qin perspective is an open question. Problematizing the master-student model of intellectual and textual transmission that is so central to the modern imagination of the early Chinese

• • • •



18 19 20

Csikszentmihalyi 2004: 25. For a contrasting view, see Goldin’s essay in this volume (chap. 4): “[T]he new insights regarding the relatively late compilation of the Analects do not invalidate the traditional understanding of the text’s philosophical importance.” Behr 2011. On this problem, see the introduction to Hunter 2017.

8





Hunter And Kern

intellectual scene. Insofar as Kongzi is thought to have formed the first “school” or “intellectual lineage” (jia 家), and insofar as he and his dizi 弟子 (followers) exemplify master-student relationships in general, where else should we look for evidence of master-student transmission if not to the Lunyu? If master-student schools and lineages did not play as large a role in the formation of “Masters” literature, then what is the underlying mechanism of intellectual and textual transmission in the Warring States period?21 Exposing the anachronisms inherent in using post-Lunyu sources to make sense of pre-Lunyu sources. Sima Qian’s 司馬遷 (ca. 145–85 BCE) Shiji 史記 (Grand Scribe’s Records), most of our earliest extant commentaries to the classics, and resources like the Shuowen jiezi 說文解字 (Explanations of Characters Simple and Compound) character dictionary of ca. 100 CE all feed the impression that the Lunyu was integral to the pre-Han textual record, and all come to us from the early empire.22 Questioning the appropriateness of the accretion model of textual formation. Insofar as the field of Lunyu studies has taught us to think of early texts as successive layers of accretion, rethinking the Lunyu also leads us to reconsider our default assumptions about how early texts were formed. Canonization is by definition a process of selection and exclusion,23 and early descriptions of editors, including Kongzi, Liu Xiang 劉向 (79–8 BCE), and many others, provide overwhelming evidence for the compilation of ancient Chinese texts through processes of reduction, not expansion.24

From a traditional perspective, these points might read as rallying cries for a deconstructionist, if not outright nihilistic, program. To the contrary, for us the payoff of rethinking the Lunyu is the thrill of exploring new perspectives in the study of early Chinese texts. The contributions to this volume are offered in that spirit.

The Contributions

Appearing more than two decades since the publication of his seminal essay “The Formation of Lunyu as a Book,” John Makeham’s chapter (“A Critical 21 22 23 24

See also Weingarten 2015. See Durrant 1995 for the importance of the Kongzi persona for Sima Qian’s self-presentation, which, in turn, colors the entire reception of the Shiji and our understanding of the Warring States period. Assmann and Assmann 1987. See, e.g., Kern 2015; Van der Loon 1952.

Introduction

9

Overview of Some Contemporary Chinese Perspectives on the Composition and Date of the Lunyu”) surveys the current state of Chinese-language scholarship on the Lunyu, in particular in relation to recently unearthed manuscripts. Makeham spotlights “methodologically naïve” efforts to use these manuscripts to reaffirm traditional accounts of early ru 儒 intellectual history involving both the Mengzi 孟子 (Mencius) and a (re)constructed Zisizi 子思子 (Master Zisi) and to claim an early date for the Lunyu within this history. In a detailed rebuttal of some of the most prominent studies to this effect, Makeham demonstrates how their use of manuscript evidence from Guodian 郭店 and from the Shanghai Museum corpus involves multiple leaps of faith and logical fallacies. Makeham urges us to distinguish between an irrecoverable past and a reconstructed past and to be more critically aware of our own roles as interpreters. In chapter 2 (“The Lunyu as an Accretion Text”) Robert Eno defends the accretion model of the Lunyu, the most influential theory of the text’s origins in the modern era. After broadly surveying (and critiquing) previous iterations, including those of Kimura Eiichi and E. Bruce Brooks and A. Taeko Brooks, Eno presents the most cautious and persuasive version of the accretion theory to date. Without assuming a strong link between the Lunyu and the historical Kongzi, Eno tentatively dates the core layer of the Lunyu to the late fourth or early third century BCE, with additional stages of redaction taking place in the Qin and Han periods. Of particular note is Eno’s discussion (following Kanaya Osamu) of “the likely role of Qin encyclopedism” in the compilation of the Lunyu and other ru texts both in the years before the imperial unification and then further with the official erudites at the Qin imperial court. Their textual work, according to Eno, prepared the basis for the subsequent canonization by Western Han scholars. Adopting a different approach to the question of the Lunyu’s origins, in chapter 3 (“The Lunyu as Western Han Text”) Michael Hunter summarizes the argument for reading the Lunyu as a Western Han text. From a statistical analysis of thousands of Kongzi quotations across all early Chinese texts, Hunter shows that before mid–Western Han times, anything resembling the received Lunyu would be unreconstructable from Kongzi quotations in other texts. Hunter places the compilation and canonization of the Lunyu in the reign of Han emperor Wu 武 (141–87 BCE), that is, the time between the Huainanzi (139 BCE) and the Shiji (ca. 100 BCE), “with post-Shiji texts drawing heavily from the Lunyu for their Kongzi quotations and pre-Huainanzi texts drawing their Kongzi material from elsewhere.” While Hunter allows for the Lunyu’s inclusion of pre-imperial material, he removes the Lunyu from its exalted position as the fountainhead of Chinese philosophy and, furthermore, opens perspec-

10

Hunter And Kern

tives on what a philosophical reading of the Lunyu as a Western Han text might reveal. In chapter 4 (“Confucius and His Disciples in the Lunyu: The Basis for the Traditional View”) Paul R. Goldin mounts a vigorous defense of “the traditional understanding of the [Lunyu’s] philosophical importance” even as he accepts a Western Han date for the redaction of the received Lunyu. Surveying various philosophical issues from the fourth and third centuries BCE, Goldin argues that the Lunyu’s silence on these issues is in keeping with an early Warring States intellectual milieu. He further shows how references to other philosophers in early texts suggest a robust overall framework for the traditional chronology of Warring States texts. Without committing to the view of the Lunyu as an authentic record of Kongzi’s teachings, Goldin thus maintains the Lunyu’s status as a canonical, or at least early, source of Chinese philosophy and that “whoever was responsible for compiling” the text “included an overwhelming proportion” of genuinely early material within it. In contrast, Joachim Gentz, in chapter 5 (“The Lunyu, a Homeless Dog in Intellectual History: On the Dating of Discourses on Confucius’s Success and Failure”), expresses skepticism regarding the prospects of dating the Lunyu. While acknowledging its wealth of “concepts, ideas, thoughts, terms, metaphors, discourses, and problems,” he argues that the Lunyu does not contextualize or systematize these elements in ways that lend themselves to intellectual historical analysis, because “[t]he purpose of the book is obviously not to take part in intellectual debates.” Drawing broadly on a wealth of transmitted texts and recently unearthed manuscripts, Gentz presents two case studies: the Lunyu’s contradictory presentation of ren 仁 (often translated as “humaneness” or “benevolence”) and its treatment of the problem of Kongzi’s success and failure in comparison to pre-Han debates on the subject. Yet neither individual concepts nor the success/failure problem, Gentz concludes, allow us to date the Lunyu—a text that with its fundamental focus on Kongzi’s action “is homeless in early Chinese intellectual history.” In chapter 6 (“Confucius’s Sayings Entombed: On Two Han Dynasty Bamboo Lunyu Manuscripts”) Paul van Els judiciously surveys the evidence from two fragmentary Lunyu manuscripts, one from the tomb of Liu Xiu 劉脩 (d. 55 BCE), the king of Zhongshan 中山, unearthed in the 1970s near the modernday city of Dingzhou 定州 in Hubei 湖北 Province, and the other from the tomb of a high-ranking Han official in Lelang Commandery 樂浪郡 in modernday North Korea. After summarizing the circumstances of the manuscripts’ discovery and scholarly history, as well as their archaeological contexts, physical characteristics, and paleographic features, including a number of textual variants, van Els addresses more difficult questions regarding the dating,

Introduction

11

provenance, and purpose of the manuscripts. Through a detailed comparative analysis of their script styles against various other Han dynasty manuscripts, he argues that both manuscripts date to the first century BCE and not to the early Western Han, as some have contended. Although not focused primarily on the problem of the Lunyu’s chronology, Matthias L. Richter’s contribution in chapter 7 (“Manuscript Formats and Textual Structure in Early China”) bears directly on the question of how we should imagine the processes of textual formation and transmission that ultimately produced the received Lunyu. Richter targets the widely held assumption that the composite nature and fluidity of texts like the Lunyu are attributable to the format of bamboo manuscripts, which (so the assumption goes) invited the rearrangement, addition, and subtraction of individual bamboo slips—in Erik Maeder’s memorable phrase, the “loose-leaf ring binder” theory of textual formation. In contrast, Richter finds little evidence for such a theory, instead pointing to numerous instances in which scribes endeavored “to define textual identity and to prevent a potential confusion of textual order.” When looking for explanations for the Lunyu’s heterogeneity, Richter concludes, we must look to factors other than early manuscript formats. In chapter 8 (“Interlocutor Collections, the Lunyu, and Proto-Lunyu Texts”) Mark Csikszentmihalyi examines the role of “interlocutor texts” in the formation of the received Lunyu. The point of departure for his essay is early theories of the Lunyu’s composition, which assign a prominent role to Kongzi’s students as recorders, compilers, and transmitters of his teachings. However, in his survey of early texts (including whole chapters and shorter dialogues) attributed to interlocutors like Zengzi 曾子, Csikszentmihalyi finds little evidence for the stable attribution of such texts to specific interlocutor figures. To the contrary, there is evidence to suggest that extant interlocutor texts reached their final form only in the Han period, perhaps under the influence of the Lunyu. Csikszentmihalyi also devotes a substantial section to the relationship between the Lunyu and the Shiji chapter of collected biographies of Kongzi’s students, which “likely was not done by someone with something very much like a modern Lunyu in front of them.” Esther Sunkyung Klein, in chapter 9 (“Sima Qian’s Kongzi and the Western Han Lunyu”), takes a closer look at that crucially important witness to the Lunyu’s emergence in the Western Han period: Sima Qian’s Shiji. Klein surveys the use of Lunyu parallels in three layers of the Shiji: the numerous taishigong yue 太史公曰 (the honorable senior archivist says) passages in which the historian speaks in his own voice, the Kongzi biography of the “Kongzi shijia” 孔子世家 (Hereditary House of Kongzi), and the disciple biographies of the “Zhongni dizi liezhuan” 仲尼弟子列傳 (Arrayed Traditions of Zhongni’s

12

Hunter And Kern

Disciples). Finding that “Sima Qian had a Lunyu, or at least some proto-Lunyu source(s),” Klein catalogs various discrepancies and contrasting points of emphasis between the Lunyu and the Shiji. These include the Shiji’s amplification of certain Lunyu sayings, its condensation of others, its fascination with Zigong coupled with its relative disinterest in Zengzi, its interest in invoking a more “esoteric” Kongzi, and its emphasis on Kongzi’s authorship of the Chunqiu 春秋 (Spring and Autumn Annals). Finally, in chapter 10 (“Kongzi as Author in the Han”) Martin Kern surveys early sources for the notion of Kongzi’s authorship of the Chunqiu as a test to falsify the hypothesis of a Han compilation date for the Lunyu. According to Kern, there is little if any evidence of this notion in pre-imperial texts; and notably, the Chunqiu is not even mentioned in the Lunyu. Kern further shows that contrary to common assumptions, the idea that Kongzi authored the Chunqiu also remains limited to very few texts during the Western Han period. The most important exception is Sima Qian’s Shiji, which, in Kern’s reading, differs strikingly from the Lunyu in one particular respect: unlike the Kongzi in the Lunyu, the Shiji Kongzi, suffering and frustrated, is obsessed with being recognized by others, and for this reason created the Chunqiu. While these observations could be interpreted to mean that the Lunyu comes from an earlier period, Kern suggests that it was the very nature of the text as an imperial primer for the education of the crown prince (and others at court) that shaped its ideological outlook, including the absence of any mention of the Chunqiu as a text critical of failed rulership.



Our thanks to the Princeton Institute for International and Regional Studies (PIIRS) and the David A. Gardner ’69 Magic Project in the Humanities Council at Princeton University for funding the original conference (“The Analects: A Western Han Text?”) on November 4–5, 2011. We would also like to thank the original participants in that conference whose conference papers did not end up in the present volume: Wolfgang Behr, Anne Cheng, Cameron Moore, Wojciech Simson, Ken’ichi Takashima, and Oliver Weingarten. Finally, we thank Oliver Weingarten for co-organizing our original conference and for serving as co-editor during earlier stages in the preparation of the present volume.

Introduction



13

Bibliography

Assmann, Aleida, and Jan Assmann, eds. 1987. Kanon und Zensur. Paderborn: Wilhelm Fink Verlag. Behr, Wolfgang. 2011. “Linguistic Approaches to the Dating of the Lunyu: Methodological Notes and Future Prospects.” Paper presented at “The Analects: A Western Han Text?,” Princeton University, November 4–5, 2011. Billioud, Sébastien, and Joël Thoraval. 2015. The Sage and the People: The Confucian Revival in China. Oxford: Oxford University Press. Brooks, E. Bruce, and A. Taeko Brooks. 1998. The Original Analects: Sayings of Confucius and His Successors. New York: Columbia University Press. Chin, Ann-ping. 2007. The Authentic Confucius: A Life of Thought and Politics. New York: Scribner. Chin, Ann-ping. 2014. Confucius–The Analects. New York: Penguin Books. Csikszentmihalyi, Mark. 2002. “Confucius and the Analects in the Hàn.” In Van Norden 2002: 134–162. Csikszentmihalyi, Mark. 2004. Material Virtue: Ethics and the Body in Early China. Lei­ den: Brill. Durrant, Stephen W. 1995. The Cloudy Mirror: Tension and Conflict in the Writings of Sima Qian. Albany: State University of New York Press. Elman, Benjamin, and Martin Kern, eds. 2010. Statecraft and Classical Learning: The Rituals of Zhou in East Asian History. Leiden: Brill. Gardner, Daniel K. 2003. Zhu Xi’s Reading of the Analects: Canon, Commentary, and the Classical Tradition. New York: Columbia University Press. Hall, David, and Roger T. Ames. 1987. Thinking through Confucius. Albany: State University of New York Press. Hanshu 漢書. 1962. Beijing: Zhonghua shuju. Haupt, Christiane. 2006. “Und der Meister sprach … : Die Darstellung des Konfuzius in Texten der Zhanguo- und Frühen Han-zeit.” PhD diss., Ludwig-Maximilians University, Munich. Hinton, David. 2014. Analects: Confucius. Berkeley: Counterpoint. Hunter, Michael. 2017. Confucius Beyond the Analects. Leiden: Brill. Jiao Xun 焦循. 1987. Mengzi zhengyi 孟子正義. Beijing: Zhonghua shuju. Jones, David, ed. 2008. Confucius Now: Contemporary Encounters with the Analects. Chicago: Open Court. Kaneto Mamoru 金戸守. 1970. “Rongo buntei kō: Gen Rongo seiritsu o megutte” 論語文 体考: 原論語成立をめぐって. Shitennōji joshi daigaku kiyō̄ 四天王寺女子大学紀要 2:32–58.

14

Hunter And Kern

Kaneto Mamoru 金戸守. 1971. “Kōshi to so no jidai: Rongo hihan oboegaki” 孔子とその 時代: 論語批判覚. Shitennōji joshi daigaku kiyō 四天王寺女子大学紀要 4:101–115. Kaneto Mamoru 金戸守. 1972. “Kobun Rongo kō: Shiki deshi iu, Kansho Geibunshi” 古 文論語考:史記弟子云, 漢書藝文志. Shitennōji joshi daigaku kiyō 四天王寺女子大学 紀要 5:43–59. Kaneto Mamoru 金戸守. 1973. “Shiki Rongo kō: Kōshi seika, sono ‘Rongo’ seiritsu shiron” 史記「論語」考: 世家、その「論語」成立試論. Shitennōji joshi daigaku kiyō 四天王寺女子大学紀要 6:14–38. Kaneto Mamoru 金戸守. 1974. “Shiki Rongo kō, sono ni: Chūji deshi retsuden yori Rongo seiritsu ni tsuite” 史記「論語」考、その二:仲尼弟子列伝より論語成立につい て. Shitennōji joshi daigaku kiyō 四天王寺女子大学紀要 7:51–70. Kaneto Mamoru 金戸守. 1975. “Shiki Rongo kō, so no mi: Honji, Hassho no Rongo seiritsu shiryō ni tsuite” 史記「論語」考、その三: 本紀,八書の論語成立資料につ いて. Shitennōji joshi daigaku kiyō 四天王寺女子大学紀要 8:45–59. Kaneto Mamoru 金戸守. 1976. “Shiki Rongo kō, so no shi: Seika inyo Koshi-gen yori Rongo no seiritsu” 史記「論語」考、その四: 世家引用論語の成立. Shitennōji joshi daigaku kiyō 四天王寺女子大学紀要 9:101–118. Kaneto Mamoru 金戸守. 1977. “Shiki Rongo kō, so no go: Haku’i retsuden o ronjite Rongo Koshi-gen ni oyobu” 史記「論語」考、その五: 伯夷列伝を論じて論語孔子言 に及ぶ. Shitennōji joshi daigaku kiyō 四天王寺女子大学紀要 10:65–79. Kaneto Mamoru 金戸守. 1978. “Shiki Rongo kō, so no roku: Soketsu, Rongo no seiritsu ni tsuite” 史記「論語」考、その六: 総結、論語の成立について. Shitennōji joshi daigaku kiyō 四天王寺女子大学紀要 11:19–26. Kaneto Mamoru 金戸守. 1979. “Shiki Rongo kōyo” 史記論語考余. Shitennōji joshi daigaku kiyō 四天王寺女子大学紀要 12:21–32. Kaneto Mamoru 金戸守. 1980. “Shiki Rongo ko yoshoku: Kansho yori mita Rongo seiritsu kō” 史記論語考余続:漢書より見た論語成立考. Shitennōji joshi daigaku kiyō 四天王寺女子大学紀要 13:17–29. Kern, Martin. 2015. “The ‘Masters’ in the Shiji.” T’oung Pao 101:335–362. Kern, Martin, and Dirk Meyer. 2017. Origins of Chinese Political Philosophy: Studies in the Composition and Thought of the Shangshu (Classic of Documents). Leiden: Brill. Kim, Tae Hyun, and Mark Csikszentmihalyi. 2014. “History and Formation of the Analects.” In Olberding 2014: 21–36. Leys, Simon. 2014. The Analects. In Nylan 2014: 62-126. Makeham, John. 1996. “The Formation of Lunyu as a Book.” Monumenta Serica 44:1–24. Makeham, John. 2002. “A New Hermeneutical Approach to Early Chinese Texts: The Case of the Analects.” Sophia 41.1:55–69. Makeham, John. 2003. Transmitters and Creators: Chinese Commentators and Com­ mentaries on the Analects. Cambridge, MA: Harvard University Asia Center.

Introduction

15

Nylan, Michael, ed. 2014. The Analects. Norton Critical Editions. New York: W. W. Norton. Olberding, Amy. 2012. Moral Exemplars in the Analects: The Good Person Is That. New York: Routledge. Olberding, Amy, ed. 2014. Dao Companion to the Analects. Dordrecht: Springer. Pan Fu’en 潘富恩 and Wen Shaoxia 温少霞. 2005. Lunyu jinyi 論語今譯. Jinan: Qi Lu shushe. Pollock, Sheldon. 2009. “Future Philology? The Fate of a Soft Science in a Hard World.” Critical Inquiry 35:931–961. Rosemont, Henry. 2013. A Reader’s Companion to the Confucian Analects. New York: Palgrave Macmillan. Schaberg, David. 2001. “‘Sell It! Sell It!’: Recent Translations of Lunyu.” Chinese Literature: Essays, Articles, Reviews 23:115–139. Slingerland, Edward. 2003. Confucius: Analects. Indianapolis: Hackett. Song Xianlin. 2003. “Reconstructing the Confucian Ideal in 1980s China: The ‘Culture Craze’ and New Confucianism.” In New Confucianism: A Critical Examination, ed. John Makeham, 81–104. New York: Palgrave MacMillan. Takeuchi Yoshio 武內義雄. 1939. Rongo no kenkyū 論語之研究. Tokyo: Iwanami sho­ten. Tsuda Sōkichi 津田左右吉. 1946. Rongo to Kōshi no shisō 論語と孔子の思想. Tokyo: Iwanami shoten. Van der Loon, Piet. 1952. “On the Transmission of Kuan-tzŭ.” T’oung Pao 41:357–393. Van Norden, Bryan W., ed. 2002. Confucius and the Analects: New Essays. Oxford: Oxford University Press. Waley, Arthur. 2000. The Analects: Confucius. New York: Alfred A. Knopf. Watson, Burton. 2007. The Analects of Confucius. New York: Columbia University Press. Weingarten, Oliver. 2010. “Textual Representations of a Sage: Studies of Pre-Qin and Western Han Sources on Confucius (551–479 BC).” PhD diss., Cambridge University. Weingarten, Oliver. 2015. “What Did Disciples Do? Dizi 弟子 in Early Chinese Texts.” Harvard Journal of Asiatic Studies 75.1:29–75. Yu Dan 于丹. 2007. Lunyu xinde 論語心得. Taipei: Lianjing chubanshe. Yu Dan 于丹. 2009. Confucius from the Heart: Ancient Wisdom for Today’s World. Trans. Estber Tyldesley. Beijing: Zhonghua shuju. Zhang Fenzhi. 2015. Xi Jinping: How to Read Confucius and Other Chinese Classical Thinkers. New York: CN Times Books. Zhao Zhenxin 趙貞信. 1961. “Lunyu jiujing shi shei bianzuan de”《論語》究竟是誰編 纂的. Beijing Shifan daxue xuebao 北京師範大學學報 4:11–24. Zhu Weizheng 朱維錚. 1986. “Lunyu jieji cuoshuo” 論語結集脞說. Kongzi yanjiu 孔子研 究 1:40–52.

16

Hunter And Kern

Zhu Weizheng 朱維錚. 1987. “Lishi de Kongzi he Kongzi de lishi” 歷史的孔子和孔子的 歷史. In Kongizi yanjiu lunwen ji 孔子研究論文集, ed. Zhonghua Kongzi yanjiusuo 中華孔子研究所, 156–172. Beijing: Jiaoyu kexue chubanshe.

Chinese Perspectives On The Composition And Date Of Lunyu

17

Chapter 1

A Critical Overview of Some Contemporary Chinese Perspectives on the Composition and Date of the Lunyu John Makeham Over the past three decades, the discovery and publication of an important body of recovered texts—including many previously unattested texts—have stimulated debate about issues of authorship, transmission, and interpretation. The manuscripts excavated from tombs at Mawangdui 馬王堆, Dingxian 定縣, Guodian 郭店, and the collection of Chu bamboo-strip texts purchased by the Shanghai Museum are of especial importance. A number of Chinese scholars have also used some of these texts to advance claims about the composition and date of the Lunyu. Some of this scholarship has been influenced by a paradigm shift over the last twenty years, which Chinese scholars themselves have referred to as moving from “doubting the past” (yigu 疑古) to “explaining the past” (shigu 釋古).1 The “explaining the past” attitude has been enthusiastically adopted by many Chinese scholars—and not just those on the mainland—to reaffirm traditional accounts of early ru 儒 intellectual history. For example, many mainland scholars assume that the Guodian corpus is, by and large, a corpus of ru texts and have sought to reaffirm traditional/conventional accounts (which date from the time of Han Yu 韓愈 [768–824]) of a Confucius, Zengzi, Zisi, Mencius lineage: the so-called SiMeng (Zisi-Mencius) school (Si-Meng xuepai 思孟學派). With Mencius serving as the “spiritual” forebear of the New Confucians and those influenced by New Confucian writings on “the learning of the mind and the nature” (xinxing zhi xue 心性之學), it is understandable why many Chinese scholars should seek to find in the Guodian materials evidence to support the existence of a Si-Meng school and to identify certain texts with that school. Thus, although they would concur with Du Weiming that “with the excavation of the Guodian Chu tomb strips, the history of Chinese philosophy and the history of Chinese scholarship need to be completely rewritten,” attempts to date certain texts to a period

1 This reorientation was highlighted with the publication of Li Xueqin 1994. For a discussion of the significance of these terms, see Makeham 2008: 210.

© Koninklijke Brill NV, Leiden, 2018 | doi:10.1163/9789004382947_003

18

Makeham

later than their traditionally accepted dates of composition are not what they have in mind.2 This last observation also applies to the case of the Lunyu, a text seen as foundational to the tradition of learning associated with the teachings of the historical Confucius. Over the past decade or more, however, there has been a difference in how recovered texts have been used with reference to the Lunyu. Essentially two trends can be discerned. In the late 1990s and early 2000s, when the Guodian materials were the focus of scholarly inquiry, there was a concentration of interest in how these materials could be used to identify the date when the Lunyu became a book. From about 2003 and onward, with the ongoing publication of the Shanghai Museum strips, there has been a shift in focus to issues of intertextuality. A related trend over this same period has been the gradual development of a hypothesis about a large proto-Lunyu corpus, compiled by Confucius’s disciples, which underwent significant subsequent editing. In what follows, I introduce some of the scholarship associated with each of these two trends and, where appropriate, comment on some shortcomings in that scholarship. I conclude with some broader methodological reflections on the distinction between a past that was once experienced and a reconstructed past, and the role that the historian/interpreter plays in that reconstruction. 1

The Guodian Materials and the Dating of the Lunyu

Example 1.1: The Guodian manuscript *Zi yi 緇衣 (Jet-Black Robes) and the Liji chapter “Fang ji” 坊記 (Embankment Record) Liang Tao 梁濤 argues that the “Fang ji” passage in the Liji which cites a passage from the Lunyu (second half of 1/11 and again at 4/20) proves that the Lunyu existed sometime before 402 BCE. His argument is as follows: “Kongzi shijia” 孔 子世家 (Hereditary House of Confucius) in the Shiji 史記 records that Zisi wrote the “Zhong yong” 中庸 (Doctrine of the Mean); and the Suishu 隨書 “Yinyue zhi” 音樂志 (Treatise on Music) cites Shen Yue 沈約 (441–513), who states that the “Zhongyong,” “Biao ji” 表記 (Record of Models), “Fang ji,” and “Zi yi” were all pian 篇 (chapters) in the Zisizi 子思子. Because a version of “Zi yi” was discovered at Guodian together with the texts Lu Mu Gong wen Zisi 魯穆公 問子思 (Duke Mu of Lu Asks Zisi), which features passages attributed to Zisi, and Wuxing 五行 (Five Phases), which many modern Chinese scholars 2 Du Weiming 1999: 2.

Chinese Perspectives On The Composition And Date Of Lunyu

19

attribute to the so-called Si-Meng school, this serves to confirm Shen Yue’s account about chapters that originally belonged to the Zisizi.3 Since the Zisizi was written by Zisi and since Zisi referred to “Lunyu” in his writings (i.e., in “Fang ji”), Liang concludes: “then at the very least, during the lifetime of Zisi, the Lunyu already existed as a book. According to the verification of modern scholars, Zisi lived circa 483–402 BCE. Therefore, the Lunyu existed as a book no later than 402 BCE.”4 In responding to Liang’s account, it should first be reiterated that we do not know the extent to which the content of the Zisizi in twenty-three pian recorded in the Hanshu 漢書 “Yiwen zhi” 藝文志 (Treatise on Arts and Letters) matches the content of the version said to have been known to Shen Yue. Yet even if we were to assume that the four pian listed by Shen Yue were once part of the Zisizi, this in no way establishes that the received version of the “Fang ji” (i.e., the version which exists as a chapter in the Liji) matches the original version. Witness the discrepancies between the Guodian and Shanghai Museum versions of “Zi yi,” on the one hand, and the Liji version of “Zi yi,” on the other.5 We simply do not know what editorial modifications some original pre-Qin “Fang ji” may have undergone before the end of the Han dynasty. The fact that specific reference to the title “Lunyu” is found in only one passage in only one supposedly pre-Han text—the “Fang ji”—surely warrants caution. It should also be pointed out that many other passages which quote Confucius in “Fang ji,” as well as other chapters in the Liji, use formulas such as zi yun 子云 (the Master said) or Kongzi yue 孔子曰 (Kongzi said); Lunyu yue 論語曰 (the Lunyu says) is used nowhere else.6 Example 1.2: The Guodian manuscript *Zi yi, the Liji chapter “Fang ji,” and Lunyu 1/11 and 4/20 As with Liang Tao, Guo Yi 郭沂 accepts that the existence of the Guodian version of “Zi yi” confirms traditional accounts that “Fang ji” was also written by 3 4 5 6

Influential proponents of this hypothesis include Li Xueqin 1998 and Ding 2000: 160–168. Liang 2002. See, e.g., Kern 2005: 293–318. This supports Takeuchi Yoshio’s 武内義雄 (1943: 127) hypothesis that the “Fang ji” reference to the Lunyu is a case of a marginal annotation being copied into the text. Thus, even if some original version of “Fang ji” did predate the Han, it would remain an open possibility (and quite a strong one given that we have no record of a redaction of the Liji before the close of the first century BCE) that the marginal annotation was not pre-Han. Cf. Mark Csikszent­ mihalyi’s observation in chapter 8 in this volume: “The attribution to the Lunyu of Lunyu 4/20 in the ‘Fang ji’ 坊記 chapter of the Liji is a good example of later interpolation, since manuscript versions of the affiliated ‘Zi yi’ 緇衣 chapter demonstrate the way in which the contents, especially the quotations, have been revised in the received editions.”

20

Makeham

Zisi, and on that basis he proceeds to use “Fang ji” to date the compilation of the Lunyu. Guo asserts that because the latest historical event referred to in the Lunyu is Zengzi’s death, which Guo dates at 436 BCE but does not explain why, “therefore that year can be established as the earliest date that the Lunyu was compiled.” Guo evidently has no problem dating the whole text on the basis of one “section” (zhang 章) in the Lunyu. He identifies 402 BCE as the latest year that “Fang ji” could have been compiled because “‘Fang ji’ is Zisi’s record of Confucius’s sayings and Zisi died in 402 BCE.”7 Guo accepts that “Fang ji” (together with “Zhong yong,” “Biao ji,” and “Zi yi”) was compiled by Zisi and that its passages citing Confucius (based on the content of the received “Fang ji” in the Liji) were recorded by Zisi. His authority for these claims is a passage in the Kong congzi 孔叢子, a work quite possibly of late Eastern Han provenance.8 He insists that the Lunyu must have been in existence already during the lifetime of Zisi because “Fang ji” cites the Lunyu by name. Guo further explains that the general principle Zisi adopted in compiling these works was to provide a record of Confucius’s remarks and sayings that had not been recorded in the Lunyu. Because Zisi regarded Lunyu 1/11 and 4/20 (“If, for three years, he simply makes no changes to the ways of his father, then this is enough to be known as being filial” [三年無改於父之道,可謂孝 矣]) as such an important passage, however, he made an exception to his general rule in this one case. As such, Guo concludes, these records have a value just as great as the Lunyu in that they too preserve a veritable record of Confucius’s sayings.9 Example 1.3: The Guodian manuscript *Yucong 語叢 3 (Thicket of Sayings, 3) and Lunyu 7/6 and 9/4 Yang Chaoming 楊朝明 has also sought to establish an early date for the formation of the Lunyu as a book, arguing that because the Mencius cites passages from the Lunyu, therefore Lunyu must have been in existence before the Mencius. (He ignores the question of when the different strata of the Mencius were composed and when the Mencius was finally compiled as a book.) It is, of course, reasonable to identify the Mencius as a work most likely to cite the Lunyu should the Lunyu have already attained the status of a book or a unique collection of records before the Mencius was compiled. As I have argued elsewhere, however: 7 Guo 2002: 23. 8 See Ariel 1989. 9 Guo 2002: 334, 339.

Chinese Perspectives On The Composition And Date Of Lunyu

21

if Lunyu were already a book or unique collection of records when Mencius was compiled, one would expect it to be quoted extensively, yet of the twenty-eight passages in Mencius which begin with “Kong Zi yue” or “Zhongni yue,” a mere eight are found in the received Lunyu. Of these eight only one is identical with the received text; of the other seven, one is slightly different while of the other six, the wording is so different that even if it were granted that they were quotations from a proto-Lunyu corpus that was then in existence, clearly between then and the end of the Western Han the contents of that corpus underwent significant editing. Yet rather than postulating a proto-Lunyu corpus that underwent significant editing, it is more reasonable to attribute all twenty-eight passages to a collection or, more probably, a number of collections of Confucius’ sayings that were already in existence when Mencius was written.10 It might be further noted that nowhere does the Mencius refer to a text called “Lunyu.” As it happens, in a paper delivered at the 2011 conference “The Analects: A Western Han Text?” Wojciech Simson also addressed the topic of sayings attributed to Confucius in the Mencius. The most significant example he cites is an extended passage in Mencius 7B/37 featuring sayings attributed to Confucius which correspond closely to four different zhang in the received Lunyu (9/22, 9/25, 17/18, 17/13). The Mencius passage is as follows (translation by Simson): Kongzi said, “I hate things that appear to be something but are not. I hate the false millet, because I fear it could confound the sprouts. I hate the flatterers, because I fear they could confound duty. I hate the sharptongued, because I fear they could confound trustworthiness. I hate the tunes of Zheng, because I fear they could confound the music. I hate the lilac, because I fear it could confound the vermillion. I hate the honest countryman, because I fear he could confound virtue.” 孔子曰:惡似而非者:惡莠,恐其亂苗也;惡佞,恐其亂義 也;惡利口,恐其亂信也;惡鄭聲,恐其亂樂也;惡紫,恐其 亂硃也;惡鄉原,恐其亂德也。 Simson takes this as evidence for the high probability that something like the received version of the Lunyu was already in existence in the late fourth or 10

Makeham 1996: 15–16.

22

Makeham

early third century BCE. He argues that this probability is further increased because the Confucius sayings in the Mencius consist of roughly a dozen passages which have no counterparts in the received Lunyu and roughly another dozen passages which do have discernible counterparts in the received Lunyu, therefore “suggesting that the whole stock of Confucius sayings at [the disposal of the Mencius authors] had a similar distribution. When, therefore, the authors of the Mencius chose a Confucius saying for the above passage from their stock indiscriminately, they had a one in two chance that it would be a nonLunyu saying and an equal one in two chance that it would be a Lunyu saying.” This is a very large leap of faith. We simply have no idea what sort of Confucius sayings the editors/authors of the Mencius had at their disposable; what was the provenance and format of those sayings; or even if they were written or orally transmitted. In addition to the possibility that Mencius 7B/37 is a case of the Mencius citing the Lunyu, two other possibilities seem just as plausible: that the received Lunyu drew on the Mencius or that both the received Lunyu and Mencius drew from a third source of Confucius sayings.11 Returning to Yang Chaoming, he accepts that Mencius was born in 372 BCE and that Confucius’s reputedly youngest disciple, Zengzi, died in 428. Because Lunyu 8/3 describes Zengzi on his deathbed instructing his own disciples, Yang insists that the Lunyu must have been compiled between 428 and 372 BCE (Yang, of course, assumes that Lunyu 8/3 was an integral part of an “original” Lunyu.) He further argues that with the discovery of the Guodian corpus, the date for the compilation of the Lunyu can be pushed toward the beginning of this sixty-year period. Yang takes it as an established fact that several of the rujia 儒家 (“Confucian”) texts in that corpus belong to the now-lost Zisizi. Because one of these texts, “Fang ji,” cites the Lunyu by name, the Lunyu must already have been in existence before the Zisizi was compiled. Further confirmation of the connection between the Guodian materials and the Lunyu is that the group of bamboo strips named *Yucong 語叢 3 (Thicket of Sayings, 3) by the modern editors contains certain passages that have been identified as having matching counterparts in the Lunyu. The first (strip 51) has been reconstructed as follows: “Fix one’s purpose on the Way, rely upon virtue, stand firm in humaneness, and associate with those of cultural attainments” (志於道。據於德。立於仁。游於藝).12 Yang concurs with the view that this is Lunyu 7/6. The second (from strips 64 and 65) has been recon11 12

Hunter (2014: 38–42) argues for a third possibility: that some Lunyu parallels were added to the Mengzi at some point in the Han after the compilation and circulation of the Lunyu. See Jingmen shi bowuguan 1998: 211; Li Xueqin 2002: 4–6.

Chinese Perspectives On The Composition And Date Of Lunyu

23

structed as follows: “Be without capricious thoughts, stubbornness, self- centeredness, and inflexibility” (亡(毋)意。亡(毋)固。亡(毋)我。亡(毋)必).13 Yang concurs with the view that this is Lunyu 9/4. On the basis of these three arguments he concludes that prior to 300 BCE (the approximate date of the tomb in which the Guodian strips were found), the Lunyu was circulated in the state of Chu; that Zisi cited the Lunyu in his writings; and that because Zisi lived sometime between 491 and 400 BCE, the Lunyu could not have been compiled later than 400 BCE or earlier than 428 BCE. In the second half of the article, Yang adduces a range of speculations and questionable arguments to claim that it was, in fact, Zisi who compiled the Lunyu.14 The problems with Yang’s argument are multiple. In regard to the first point, the conviction that the “rujia” texts in the Guodian corpus belong to the nowlost Zisizi seems to be widely accepted only within the Chinese scholarly community. Although there are reasonable grounds for speculating that texts such as the “Zhong yong” and “Da xue” 大學 (Great Learning) may once have belonged to a body of writings called Zisizi, the extent to which the received versions of those texts resemble their original versions is unknown. There are no similarly reasonable grounds for extending this speculation to *Yucong 3. Yang’s second point assumes that the received version of “Fang ji” (transmitted as a chapter in the Liji), which does cite the Lunyu by name, is faithful to some hypothesized fourth-century BCE text, yet ignores the fact that no version of “Fang ji” was recovered at Guodian.15 No mention is made of the scholarship that argues that the Liji was edited during the Han and how this may account for the anomalous reference to “Lunyu” in the “Fang ji” chapter of Liji. As for the third point, the text of Lunyu 7/6 reads, “The Master says, ‘Fix one’s purpose on the Way, rely upon virtue, stand firm in humaneness, and associate with those of cultural attainments’” (子曰:志於道,據於德,立於 仁,游於藝), and Lunyu 9/4 reads, “The Master forbade four things: one must not have capricious thoughts, one must not be inflexible, one must not be stubborn, and one must not be self-centered” (子絕四:毋意,毋必,毋固, 毋我). In both cases, the Lunyu version attributes the passages to Confucius

13 14 15

Jingmen shi bowuguan 1998: 212. Yang 2003: 34. Li Xueqin (2002: 6–7) cites a passage from the Guodian text *Yucong 1 (禮,作於情), which he claims has a close match with a line in the received “Fang ji” (禮者,因人之情 而為之節文). Even if we ignore the differences between the two passages (the *Yucong passage is considerably briefer than the “Fang ji” passage), the occurrence of this passage in a miscellaneous assortment of passages with no indication of its provenance cannot be used to sustain Li’s bald assertion that the *Yucong passage was extracted from “Fang ji.”

24

Makeham

(whereas in the *Yucong 3 version, they are attributed to no one, much less are they identified as passages from the Lunyu). In the case of Lunyu 9/4 there are several other noticeable textual inconsistencies. First, the reconstruction is made on the basis of passages taken from two different strips, 64 and 65. Strip 64 consists of four characters, transcribed as 亡意亡古. Strip 65 consists of seven characters transcribed as 亡義亡必皆至 焉. The differences with the received 9/4 are evident. Yang ignores these issues and does not even contemplate the possibilities that the Lunyu may not have existed prior to 300 BCE, that the apparent textual borrowing is actually from *Yucong 3 to the Lunyu rather than vice versa, or that both passages ultimately derive from a third source that predated both of them. Furthermore, that this particular Lunyu passage may have been borrowed from a text that was excavated at Guodian, or that both passages ultimately derive from a third source that predated both of them, applies equally to such cases as the apparent textual parallels between Lunyu 15/2 and a passage in the Guodian text *Qiongda yi shi 窮達以時 (Failure and Success Depend on Timeliness) and between Lunyu 15/6 and a passage in the *Zhongxin zhi dao 忠信之 道 (The Way of Loyalty and Trust). Despite this, and with no evidence to support the claim, scholars such as Zhou Fengwu 周鳳五 and Chen Liangwu 陳良 武 insist that the *Zhongxin zhi dao passage must have derived from the Lunyu passage.16 The above discussion has focused on scholarly developments in late-1990s and early-2000s China, when the Liji chapter “Fang ji” became pivotal to a range of arguments that linked a number of the Guodian texts with arguments that sought to prove that the Lunyu was compiled in the fifth century BCE. The above analysis highlights how prominent scholars employed questionable methodology and drew conclusions that were not adequately supported by the weight of evidence or by critical argument. The rise of the so-called “explaining the past” paradigm over the course of the 1990s and the related push to confirm the existence of a Si-Meng school were influential factors in the scholarly appropriation of the Guodian corpus. Both these factors were in turn buoyed by a strengthening cultural nationalism, the power of which has continued to grow, largely unchallenged. Of course, the impetus to preserve the sanctity of origins has much deeper cultural roots. My inquiry now turns to the question of whether matters began to improve from 2003 onward, as the scholarly gaze shifted to the Shanghai Museum strips.

16

Zhou 1998; Chen Liangwu 2008: 102.

Chinese Perspectives On The Composition And Date Of Lunyu

2

25

The Shanghai Museum Strips, Intertextuality, and a Proto-Lunyu Corpus

Example 2.1: Shanghai Museum manuscript *Cong zheng 從政 (Serving in Office) and Lunyu 20/2 Do not be oppressive; do not be cruel; do not steal; and do not be corrupt. If you neither nurture nor exhort [people], yet say to them that they must succeed, this is oppression. If you kill people without first having instructed them, this is cruelty. If your orders are not timely yet you expect prompt completion, this is theft. If you benefit from crooked affairs, this is corruption. I have heard it said that one who serves in government must esteem the five virtues; keep stable the relations between those above, those in the middle, and those below;17 and get rid of the ten enmities. The five virtues are as follows: leniency, respectfulness, generosity, humaneness, and vigilance. If the gentleman is not lenient, he will be unable to secure the support of the one hundred families; if he is not respectful, he will be unable to uproot shame; if he is not generous, he will be unable to hold the people; if he is not humane, he will be unable to conduct government; and if he is not vigilant, his affairs will come to nought. 毋暴、毋虐、毋賊、毋念(貪)。不修不武((戒/誡),謂之必成,則 暴。不教而殺,則虐。命無時,事必有期,則賊。為利枉事,則貪。 聞之曰:從政敦五德、固三制、除十怨。五德:一曰緩,二曰恭,三 曰惠,四曰仁,五曰敬 (警)。君子不緩則無以容百姓;不恭則無以除 辱;不惠則無以聚民;不仁則無以行政;不敬則事無成。

Zizhang asked Confucius, “What must someone be like to be able to serve in government?” Confucius said, “If he honors the five excellences, rejects the four evils, then he may serve in government.” Zizhang asked, “What is meant by the ‘five excellences’?” The Master replied, “The gentleman is generous without being extravagant; makes people work hard without bringing cause for complaint; has desires without being corrupt; maintains a dignified ease without being arrogant; is awe-inspiring without being severe.” Zizhang asked, “What is meant by ‘generous but not extravagant?’” The Master replied, “Benefit the people 17

Tentatively following the interpretation of Zhu 2003.

26

Makeham

by according with that which truly benefits them—is this not generosity?18 By selecting tasks worth working hard for, who would complain? Having desired to be humane and so achieving humaneness, what is there left to covet? Whether dealing with the many or the few, the mighty or the humble, the gentlemen dares not treat them indifferently—is this not maintaining a dignified ease without being arrogant? The gentleman keeps his robe and cap straight and maintains a dignified gaze, and people respectfully look upon him with awe. Is this not being awe-inspiring without being severe? Zizhang asked, “What is meant by the ‘four evils’?” The Master said, “To execute people without first having instructed them is called cruelty; to expect success without first having exhorted is called oppression; to be tardy in giving orders yet to expect prompt completion is called stealing; to dole out stingily what must be given is called clerkishness.” 子張問於孔子曰:何如斯可以從政矣?子曰:尊五美,屏四惡,斯可 以從政矣。子張曰:何謂五美?子曰:君子惠而不費;勞而不怨;欲 而不貪;泰而不驕;威而不猛。子張曰:何謂惠而不費?子曰:因民 之所利而利之,斯不亦惠而不費乎?擇可勞而勞之,又誰怨!欲仁而 得仁,又焉貪!君子無眾寡,無小大,無敢慢,斯不亦泰而不驕乎! 君子正其衣冠,尊其瞻視,儼然人望而畏之,斯不亦威而不猛乎!子 張曰:何謂四惡?子曰:不教而殺謂之虐;不戒視成謂之暴;慢令致 期謂之賊;猶之與人也,出納之吝,謂之有司。

Chen Jian 陳劍 notes the similarity of the “four evils” listed in each of the two passages and assumes that it must be a case of the *Cong zheng having been based on the Lunyu. He suggests that the fourth item in the respective lists of “four evils” differs because it was not clear even to the authors of the *Cong zheng just what the phrase “to dole out stingily what must be given is called clerkishness” (猶之與人也,出納之吝,謂之有司) in the Lunyu passage actually meant.19 In addition to the “four evils,” Zhou Fengwu 周鳳五 draws attention to the “five virtues” (wu de 五德) in the *Cong zheng and the “five excellences” (wu shan 五善) in the Lunyu passage, suggesting that although the *Cong zheng was based on the Lunyu, the content of the two lists differs because the *Cong

18 19

Beginning with the second answer, this series of answers appears without the matching question. This “truncated” format may reflect the influence of oral transmission. Chen Jian 2003.

Chinese Perspectives On The Composition And Date Of Lunyu

27

zheng provides an elaboration of the five virtues. Unfortunately, Zhou does not explain how the elaboration works.20 Example 2.2: Shanghai Museum manuscript *Cong zheng and Lunyu 13/25 If the gentleman is ahead of others [in his capacities], then he will lead them; if he lags behind others [in his capacities], then he will serve them. Accordingly, the gentleman is difficult to obtain and easy to employ. When he employs people, he values their [abilities]. When the petty person is ahead of others [in his capacities], he … them. When he lags behind others [in his capacities], he denigrates (?) them. Accordingly, it is said that the petty person is easy to obtain but difficult to employ. When he employs others, he is fully demanding of them. [君子先]人則啓道之,後人則奉相之,是以曰君子難得而易史(使) 也,其史(使)人,器之;小人先人則 敔之,[後人] 則 毀之,是 以曰小人易得而難史(使)也,其史(使)人,必求備焉。 The Master said, “The gentleman is easy to serve but difficult to please. If one seeks to please him by [doing something] that is not in accord with the Way, he will not be pleased. When it comes to employing people, he values their [abilities]. The petty person is difficult to serve but easy to please. If one seeks to please him by [doing something] even though it is not in accord with the Way, he will be pleased. When it comes to employing people, he is fully demanding of them.” 子曰:君子易事而難說也:說之不以道,不說也;及其使人也,器 之。小人難事而易說也:說之雖不以道,說也;及其使人也,求備 焉。

Again, Chen Jian maintains that the passage from the *Cong zheng was based on the Lunyu passage.21 The two passages are clearly related, yet in the *Cong zheng passage, the gentleman is presented as serving some ruler, whereas in the Lunyu passage, the gentleman is the ruler. Following Chen Wei 陳偉,22 Chen Jian speculates that the *Cong zheng was based either on an early protoLunyu or on a different edition of the Lunyu circulating at that time. 20 21 22

Zhou 2003. Chen Jian 2004: 4. Chen Wei 2003.

28

Makeham

Example 2.3: Shanghai Museum manuscript *Junzi wei li 君子為禮 (The Noble Man’s Practice of Ritual) and Lunyu 7/6 and 12/1 Yan Yuan was waiting in attendance on Confucius when Confucius said, “Hui, the gentleman carries out ritual by cleaving to humaneness.” Yan Yuan got up and responded, “I am not clever—I am unable to grasp things quickly.”23 The Master said, “Sit down. I have something to tell you. If it is not right to say something, then do not say it. If it is not right to look at something, then do not look at it. If it is not right to listen to something, then do not listen to it. If it is not right to move, then do not move.” 顏淵侍於夫子,夫子曰: 韋 (回),君子為禮,以依於仁。顏淵作而答 曰:韋(回)不敏,弗能少居也。夫子曰:坐,吾語汝。言之而不義, 口勿言也;視之而不義,目勿視也;聽而不義,耳勿聽也;動而不 義,身勿動焉。 

The Master said, “Fix one’s purpose on the Way; rely upon virtue; cleave to humaneness; and associate with those of cultural attainments.” 子曰:志於道 ,據於德,依於仁,游於藝。

Yan Yuan asked about humaneness, and the Master answered, “Restrain oneself and return to ritual. If for a whole day one can restrain oneself and return to ritual, then the whole world will ascribe humaneness to him. Does the practice of humaneness issue from oneself or from others?” Yan Yuan answered, “May I ask about the steps involved?” The Master answered, “If it is contrary to ritual, do not look; if it is contrary to ritual, do not listen; if it is contrary to ritual, do not speak; and if it is contrary to ritual, do not move.” 顏淵問仁。子曰:克己復禮為仁。一日克己復禮,天下歸仁 焉。為仁由己,而由人乎哉?顏淵曰:請問其目?子曰:非禮 勿視,非禮勿聽,非禮勿言,非禮勿動。顏淵曰:回雖不敏, 請事斯語矣!

23

On the term shao ju 少居, see Liao 2006.

Chinese Perspectives On The Composition And Date Of Lunyu

29

Xu Shaohua 徐少華 proposes that the whole of the *Junzi wei li—several paragraphs in all—is a development of Lunyu 7/6, with the cited lines evidencing a recognizable correspondence with the two transmitted Lunyu passages, 12/1 and 7/6. Again, it is assumed that the direction of the borrowing is from some original Lunyu of which our received text is a distant heir. 24 Huang Ren’er 黃人二 concurs with the view that the above-cited passage from the *Junzi wei li closely resembles the passage from Lunyu book 12.25 He then develops an argument to the effect that during the Western Han, the Lun­ yu was known by various titles and that the title “Lunyu” was but one of them, citing the following examples: (a) A passage from the “Zheng shuo” 正說 (Correcting Erroneous Statements) section of the Lunheng 論衡, which states that Fu Qing of Lu was the first to use the name “Lunyu” and that before this, during the reign of Emperor Xuan 宣帝 (r. 74–49 BCE), the academicians simply referred to it as “zhuan/ chuan” 傳 (tradition),26 thus indicating that the name “Lunyu” had yet to become fixed. (Huang also cites several passages from the Shiji, Hanshu, and Fa­ yan 法言 where the term zhuan/chuan is used to refer to passages that occur in the received text of the Lunyu.) (b) Examples from the Hanshu and the Lunheng where the term lun 論 is used either by itself or in such titles as Qilun 齊論 (Qi Lunyu), Lulun 魯論 (Lu Lunyu), Gulun 古論 (Ancient Lunyu), and Zhang Hou Lun 張侯論 (Marquis Zhang Lunyu). (c) The titles Lun yan dizi ji 論言弟子籍 (Lun yan—Register of Disciples) and Lunyu dizi wen 論語弟子問 (Lunyu—Disciples’ Questions)27 that Sima Qian refers to in Shiji “Zhongni dizi liezhuan” 仲尼弟子列傳 (Arrayed Traditions of Zhongni’s Disciples).28 Huang identifies Lun yan dizi ji as the so-called Gulun purportedly recovered from the wall of Confucius’s family house in 155 BCE and speculates that the Lunyu dizi wen was an early version of the Lunyu that was compiled by Confucius’s disciples and included more details about the names of his disciples than the material we find in the received Lunyu. 24

25 26 27 28

Xu 2007: 23–24. In his essay in this volume (chap. 8), Mark Csikszentmihalyi instead notes that even though the cited passage from *Junzi wei li is a parallel to Lunyu 12/1, there are multiple differences, maintaining “it is clear that basic elements of different passages were still changing in the Warring States period.” Huang Ren’er 2011. Huang Hui 1990: 2.1133. The translation of these two “titles” is necessarily speculative, as it is unclear just how Huang understood them. Shiji 67.2226. See Esther Klein’s discussion of these texts in chapter 9 in the present volume.

30

Makeham

Although he does not explicitly make the point, as these examples suggest, for Huang the Lunyu must have been in a fluid state and transmitted in more than one format well into Han times. Huang also appeals to internal evidence in the received Lunyu to explain why it was (purportedly) referred to by a variety of titles in the Warring States and Western Han periods, before the title “Lunyu” gradually became fixed. Specifically, he follows a particularly creative reading of Lunyu 5/22:29 When the Master was in Chen, he said, “Let’s go back! Let’s go back! Our young students are wildly writing on bamboo strips. They have brilliant talents but do not know how to edit what they have written.” 子在陳,曰:歸與!歸與!吾黨之小子狂簡,斐然成章,不知所以裁 之。

According to Huang, the younger students who had remained in Lu were each frantically compiling exceedingly lengthy records of the Master’s sayings and conversations, and it was not until Confucius returned to Lu that he personally abridged this material and reduced it to a more manageable size. We are told that this was in 484 BCE, five years before the Master’s death. The material that became edited down was subsequently referred to by a variety of names, such as Lun (and its variants such as Gulun, etc.), Zhuan, Lun yan dizi ji, Lunyu dizi wen, and eventually Lunyu. (As stated above, implicit in Huang’s thesis is the claim that the Lunyu must have been in a fluid state and transmitted in more than one format well into Han times.) Huang further claims that other texts recovered in recent years—such as the Guodian and the Shanghai Museum strips—that are similar to material found in the received Lunyu are derived from the large unedited body of material that Confucius’s students had compiled, a proto-Lunyu. Nor does Huang stop there. He further proposes that even though the material in a range of recovered texts was ultimately derived from some preedited Lunyu corpus, much of this material was subsequently transmitted independently of the Lunyu, in the form of commentaries or “traditions,” and served to elaborate on the broader meaning of particular sections and sentences of the Lunyu. As typical examples, Huang lists *Zhongxin zhi dao as an interpretation of Lunyu 15/6, *Qiongda yi shi as an interpretation of 15/2, *Cong zheng as an interpretation of 17/6 and 20/2, Zhonggong 仲弓 (Shanghai Museum corpus) as 29

Huang follows Mou 1987: 221–229.

Chinese Perspectives On The Composition And Date Of Lunyu

31

an interpretation of 13/2, and Jun zheng 君正 (The Ruler’s Government; Mawangdui) as an interpretation of 13/29. Example 2.4: Shanghai Museum manuscript Zhonggong 仲弓 and Lunyu 13/2 Zhonggong told Confucius that Ji Huanzi had appointed him as head official of a hamlet, 30 saying, “The Ji family … has appointed me as an official and I am worried about it.”… Zhonggong said, “May I dare to inquire, what should come first in governing?”… [Confucius] said, “In ancient times those who served their rulers did so with loyalty and respect. Even though the task of governing is difficult, just [practice] what you have learned: [treat the elderly as they should be treated; be kind to the young; guide your serving officials; promote the talented and worthy; and reform those who have been found guilty of] committing offenses. This is where governing starts.” Zhonggong said, “I am not smart. Even if there were talented and worthy persons, I would not be able to recognize them to promote. May I ask, how does one go about promoting the talented?” Confucius said, “Talent should not be kept concealed. Promote those you recognize. Are others likely to ignore those you did not recognize?”31 季桓子使仲弓為宰,仲弓以告孔子曰: 季氏 … 使雍也從於宰夫之後, 雍也童(忡) … 仲弓曰:敢問為政何先?… 曰:雍,古之事君者,以忠 與敬,唯(雖)其難也,汝惟以聞,[老老慈幼, 先有司, 舉賢才, 改過舉] 罪,政之始也。仲弓曰:若夫老老慈幼,既聞命矣。夫先有司為之如 何?仲尼曰:夫民安舊重舉 … 老老,慈幼,先有司,舉賢才,改過舉 罪…. 上下相復以忠,則民懽承教,曷為者不有成是故有司不可不先 也。仲弓曰:雍也不敏,唯 (雖)有賢才弗知舉也,敢問舉才如之何? 仲尼:夫才不可掩也。舉爾所知,爾所不知,人其舍之者。

Zhonggong was the head official of a hamlet [appointed] by the Ji family and asked about governing. Confucius said, “Guide your serving officials, forgive minor transgressions, and promote the worthy and talented.” Zhonggong said, “How does one recognize the worthy and talented in order to promote them?” Confucius said, “Promote those you recognize. Are others likely to ignore those you did not recognize?” 30 31

On whether zai 宰 should be understood as a senior official within the Ji family or as a hamlet chief, see Liao 2004. Following Li Ling’s (2004: 299) interpretation of this last sentence.

32

Makeham 仲弓為季氏宰,問政。子曰:先有司,赦小過,舉賢才。曰: 焉知賢 才而舉之?子曰:舉爾所知,爾所不知,人其舍諸?

On the basis of the similarity between the Shanghai Museum manuscript Zhonggong and Lunyu 13/2, Chao Fulin 晁福林 infers that the Zhonggong is one of the “several tens to one hundred pian” which, according to Wang Chong 王 充 (27–100),32 had made up the original, preedited Lunyu as recorded by Confucius’s disciples.33 Chen Tongsheng 陳桐生 advances a variation of this view, proposing that the Zhonggong text was taken from an original record of conversations between Confucius and Zhonggong, of which only a selection made its way into the received Lunyu.34

Concluding Comments

This hypothesis about a large proto-Lunyu corpus that was compiled by Confucius’s disciples and that underwent significant subsequent editing was developed in the wake of the publication of the Shanghai Museum strips, from about 2003 onward. Compared with the methodological assumption shared by Guo Yi and Yang Chaoming that the whole text of the Lunyu can be dated on the basis of an individual passage, the rehabilitation of Wang Chong’s thesis— that there once existed a large body of proto-Lunyu materials—is far less methodologically naïve. Of course, not all Chinese scholars insist that apparent textual borrowing must be from the Lunyu or from some presumed corpus of material from which the Lunyu itself was also compiled. In referring to the similarities between the Zhonggong text and Lunyu 13/2, Li Ling 李零, for example, argues for the possibility that our received version of the Lunyu, in part or in whole, consists of passages extracted or digested from texts that are themselves assorted collections of sayings and records of events and people.35 Unfortunately, Li’s openness on this issue is not widely shared. Yet even Li is not consistent. When the Shanghai Museum manuscript *Kongzi shilun 詩論 (Kongzi’s Discussions of the Odes) was first published, Li was quick to conjecture that in part this manuscript may have recorded the views of Confucius’s disciple Zigao 子羔. Because some of these discussions (lun 論) contain the phrase Kongzi yue 孔子曰 32 33 34 35

Huang Hui 1990: 2.1132. Chao 2005: 14. In chapter 3 in the present volume Michael Hunter also discusses the Zhonggong and Lunyu 13/2, arguing that 13/2 was composed after the Zhonggong. Chen Tongsheng 2008: 914. Li Ling 2004: 298–299, 324.

Chinese Perspectives On The Composition And Date Of Lunyu

33

(Confucius said), Li accepted that most of the text preserves the views of the historical Confucius as recorded by his disciples. Elsewhere Li also maintained that every single occurrence of the phrase Kongzi yue in the Liji marks the words of the historical Confucius.36 My own view is that at some point after the Master’s death, if not already before, there were probably various different corpora of material focusing on the Master’s sayings, transmitted either orally or in writing.37 (We can only conjecture who might have produced and circulated this material.) Our received Lunyu was drawn from some of that material and was edited into its more or less current form during the Western Han. This thesis, or rather hypothesis, remains necessarily speculative. In his essay in this volume (chap. 4), Paul R. Goldin quite rightly reminds us that “[a] work that was compiled in a certain century does not necessarily consist of material dating from that same century.” His own position is that even though the Lunyu may well have been redacted in the Western Han, it consists largely, if not almost exclusively, of sayings and discussions from the time of Confucius and his disciples. Insisting that “any theory of the compilation of the Analects is incomplete without an account of the purpose,” Goldin remains open to the traditional view that “the Analects was recorded to prevent Confucius’s teachings from disappearing at a time when thoroughgoing social and political changes threatened old-fashioned methods of transmission, such as oral transmission from master to disciple. The survival of ideas in this new context could not be guaranteed if they were not written down.” This is all very well, but the traditional view still fails to explain why “no pre–Han dynasty texts and few Western Han dynasty texts mention the Analects and [why] statements attributed to Confucius in pre-Han and Western Han texts differ markedly from those attributed to Confucius in the Analects.”38 In his paper presented at the conference, Wojciech Simson maintains that we should not expect to find the Lunyu cited by title in early China for two reasons. The first is because book titles were uncommon in pre-Qin literature. Second, even in the Western Han period, the Lunyu was rarely cited by title. Even if we accept these claims as unproblematic, this still does not address the more important fact that “statements attributed to Confucius in pre-Han and Western Han texts differ markedly from those attributed to Confucius in the 36 37 38

See Li Ling 2002: 67–68. This view is consistent with the thesis advanced by Mark Csikszentmihalyi in chapter 8 about the circulation of pericopes during the Warring States period and their varied appropriation. Borrowing a statement from Paul van Els’s essay in this volume (chap. 6).

34

Makeham

Analects.” In other words, there is little evidence to show that the Lunyu was cited in pre-Han texts.39 As for Goldin’s observation that the Lunyu fails to bear evidence of Han “philosophical” concerns, this does not strike me as a particularly compelling reason to assume that the Lunyu must therefore have been compiled before the Western Han. For the sake of argument, let us hypothetically assume that the Lunyu was compiled during the Western Han. If we lacked a good understanding of the purpose for which it was compiled at the time, why should this lead us to expect to find philosophical content in it? I also find Goldin’s related claim that “the new insights regarding the relatively late compilation of the Analects do not invalidate the traditional understanding of the text’s philosophical importance” to be questionable. Some years ago, in a review of Roger T. Ames and Henry Rosemont Jr., The Analects of Confucius: A Philosophical Translation (1998), I made the following point about the “philosophical” content of the Analects: “Philosophical significance is a matter of interpretation and perspective. Take, for example, the character shu 恕. Even though it only occurs twice in the Lunyu (and still only once in Mencius), in his Lunyu jizhu 論語集注 commentary Zhu Xi accorded it prime philosophical significance. By the same token, the character qi 氣 might ordinarily be seen to have no philosophical significance in the Lunyu, yet in Huang Kan’s Lunyu yishu 論語義疏 sub-commentary, it is clearly of philosophical significance.”40 Claims about the philosophical significance of the Lunyu are, of course, long-standing and seem to have acquired the status of a self-evident truth when the question of how the text should be approached is broached. To cite but one recent example, in the introduction to the Dao Companion to the Analects, Amy Olberding states that “although scholars agree that the text is effectively a pastiche, what this bodes hermeneutically for interpreting the text philosophically is an open question. . . . The wider issue in play and the 39

40

A different sort of argument from absence is developed by Martin Kern in chapter 10 of the present volume. Having demonstrated that a dominant perspective shared by certain Han authors was that Confucius created Chunqiu (Springs and Autumns), Kern further notes the anomalous absence of even a single reference to the Springs and Autumns in the Lunyu. What, he asks us, does this tell us about the Lunyu and its hypothetical Western Han compiler? In reply, he proposes an intriguing possibility: this is evidence that the Lunyu was compiled by those whose interests were aligned with the imperial voice rather than with the new class of career classicists. As enticing as this possibility is to those of us who seek to defend a Western Han date for the compilation of the Lunyu, if we apply the principle of Occam’s Razor, a more parsimonious explanation would simply be that the absence of any reference to the Springs and Autumns in the Lunyu points to a pre-Han date for the Lunyu’s compilation. Makeham 1999: 19.

Chinese Perspectives On The Composition And Date Of Lunyu

35

governing question that informs all interpretation is whether and to what extent the text [the Analects], despite its complex compositional history, evinces a conceptual and thematic unity that renders it available as a piece of philosophy rather than many fragmentary pieces of what may be multiple philosophies [emphasis added].”41 Again, I confess to being perplexed as to why we should expect to find philosophical content in a “pastiche” text of obscure provenance. This is not to deny that the application of philosophical concepts and frameworks can be fertile ground for the development of philosophical insights that may exist in potentia in a text but that require hermeneutical reflection and dialogue to bring to fruition. Yet equally, we should be vigilant about claims that might suggest the recovery of philosophical meaning in a text such as the Lunyu. I would urge greater critical awareness of (1) the difference between a past that was once experienced and a reconstructed past and (2) the role that the historian/interpreter plays in that reconstruction. In the case of the Lunyu, this historical consciousness should recognize that our knowledge of the historical contexts of the text’s genesis promises to provide little in the way of hermeneutic purchase and leverage and that the construction of historical contexts in which to situate the text tells us more about the present than it does about the past. This last point is amply demonstrated in this chapter. Elsewhere I have distinguished between a text’s “historical meaning” and its “scriptural meaning.”42 By historical meaning I refer to the meaning of a text as composed by its original author(s) and/or its original audience. By scriptural meaning I point to the meaning realized in the subsequent historical trajectory of that text. I have argued that the scriptural meaning of historical texts such as the Lunyu cannot be fixed within the range of a limited series of historical accretions because no interpreter is immune from the ongoing process of the text’s reception. Although this is not a reason to abandon historical context as a criterion to be used in adjudicating between competing interpretive claims, we should bear in mind the crucial role that the reader/interpreter plays in creating historical context. One consequence of this is that even if it were possible somehow to “recover” the historical meaning of a text, there would be no independent criteria by which such an understanding could ever be distinguished from scriptural meaning. This being so, I believe greater emphasis should be given to understanding the scriptural meanings of the text, as preserved in the commentarial literature. This is not to say that scriptural meanings have to be consulted 41 42

Olberding 2014: 3, 4. Makeham 2003: 9–10, 15–16.

36

Makeham

simply because they exist or were the fashion at some time, nor is it to claim that a correct interpretation has to be the sum of all other interpretations. Rather, understanding how prominent representatives of a tradition interpreted the text is to reflect on the preconditions (and preconceptions) of our own understanding.

Bibliography

Ariel, Yoav. 1989. K’ung-ts’ung-tzu: The K’ung Family Masters’ Anthology—a Study and Translation of Chapters 1–10, 12–14. Princeton, NJ: Princeton University Press. Chao Fulin 晁福林. 2005. “ShangBo jian Zhonggong shuzheng”上博簡《仲弓》疏證. Kongzi yanjiu 孔子研究 2:4–16. Chen Jian 陳劍. 2003. “ShangBo jian Zigao, Cong zheng pian de pinhe yu bianlian wenti xiaoyi” 上博簡《子羔》、《從政》〉篇的拼合與編連問題小議. http://www. jianbo.org/Wssf/2003/chenjian01.htm, accessed October 9, 2011. Chen Jian 陳劍. 2004. “Shanghai bowuguan cang Zhanguo Chu zhushu Cong zheng pian yanjiu (san ti)” 上海博物館藏戰國楚竹書《從政》篇研究(三題). Paper presented at the Third International Conference on Excavated Chinese Manuscripts, Mount Holyoke College, South Hadley, MA, 2004. Unpublished article but available in different formats on the web. I have used the version at http://www. gwz.fudan.edu.cn, accessed October 23, 2011. Chen Liangwu 陳良武. 2008. “Chutu wenxian yu Lunyu yanjiu” 出土文獻與《論語》研 究. Zhangzhou shifan Da xue xuebao 漳州師范學院學報 (Philosophy and Social Sciences) 69.3:99–105. Chen Tongsheng 陳桐生. 2008. “Cong ShangBo zhujian kan Lunyu de bianzuan tedian” 從上博竹簡看《論語》的編纂特點. Wuhan ligong Da xue xuebao 武漢理工大學學 報 (social sciences edition) 21.6:914–917. Chen Wei 陳偉. 2003. “Shanghai bowuguan cang Chu zhushu Cong zheng jiaodu”上海 博物館藏楚竹書《從政》校讀. http://www.bamboosilk.org/Wssf/2003/chenwei01. htm, accessed October 11, 2011. Ding Sixin 丁四新. 2000. Guodian Chu mu zhujian sixiang yanjiu 郭店楚墓竹簡思想研 究. Beijing: Dongfang chubanshe. Du Weiming 杜維明. 1999. “Guodian Chujian yu XianQin ru dao sixiang de chongxin diwei” 郭店楚簡與先秦儒道思想的重新地位. Zhongguo zhexue 中國哲學 20:1–6. Guo Yi 郭沂. 2002. Guodian zhujian yu XianQin xueshu sixiang 郭店竹簡與先秦學術思 想. Shanghai: Shanghai jiaoyu chubanshe. Huang Hui 黃暉. 1990. Lunheng jiaoshi 論衡校釋. Beijing: Zhonghua shuju. Huang Ren’er 黃人二. 2011. “ShangBo zhujian (wu) Junzi wei li yu Dizi wen shishi—jian lun ben pian pianming wei Lunyu dizi wen yu Lunyu zhi xingcheng he zhuyao bi-

Chinese Perspectives On The Composition And Date Of Lunyu

37

anji shijian” 上博藏簡(五)《君子為禮》與《弟子問》試釋——兼論本篇篇名 為《論語弟子問》與《論語》之形成和主要編輯時間. Zhongguo guojia bowuguan guankan 中國國家博物館館刊 2011.6. Unpaginated copyedited copy kindly supplied by the author. Hunter, Michael. 2014. “Did Mencius Know the Analects?” T’oung Pao 100:1–47. Jingmen shi bowuguan 荊門市博物館. 1998. Guodian Chu mu zhujian 郭店楚墓竹簡. Beijing: Wenwu chubanshe. Kern, Martin. 2005. “Quotation and the Confucian Canon in Early Chinese Manuscripts: The Case of ‘Zi yi’ (Black Robes).” Asiatische Studien / Études asiatiques 59.1:293–332. Li Ling 李零. 2002. Guodian Chujian jiaodu ji 郭店楚簡校讀記. Enl. ed. Beijing: Beijing Da xue chubanshe. Li Ling 李零. 2004. Jianbo gushu yu xueshu yuanliu 簡帛古書與學術源流. Beijing: Sanlian shudian. Li Xueqin 李學勤. 1994. Zouchu yigu shidai 走出疑古時代. Shenyang: Liaoning Da xue chubanshe. Li Xueqin 李學勤. 1998. “Cong jianbo yiji ‘Wuxing’ tan dao ‘Da xue’” 從簡帛佚籍《五 行》談到《大學》. Kongzi yanjiu 孔子研究 3:47–51. Li Xueqin 李學勤. 2002. “Yucong yu Lunyu”《語叢》與《論語》. In Qinghua Da xue sixiang wenhua yanjiusuo jikan 清華大學思想文化研究集刊, ed. Liao Mingchun 廖 名春, 2nd ser., 3–7. Beijing: Qinghua Da xue chubanshe. Liang Tao 梁濤. 2002. “Dingxian zhujian Lunyu yu Lunyu de chengshu wenti” 定縣竹簡 《論語》與《論語》的成書問題. http://www.jianbo.org/Wssf/2002/liangtao04. htm, accessed October 9, 2011. Liao Mingchun 廖名春. 2004. “Ji Huanzi shi Zhonggong wei zai: Chu jian Zhonggong pian zaji zhi yi” 季桓子使仲弓為宰——楚簡《仲弓》篇劄記 一. http://www.confucius2000.com/qhjb/jhzszgwzcjzgzj1.htm>, accessed November 16, 2011. Liao Mingchun 廖名春. 2006. “ShangBo wu: Junzi wei li pian jiaoshi zhaji”《上博五·君 子為禮》篇校釋箚記 . http://www.confucius2000.com/admin/list.asp?id=2276, accessed October 15, 2011. Makeham, John. 1996. “The Formation of Lunyu as a Book.” Monumenta Serica 44:1–24. Makeham, John. 1999. Feature review, “The Original Analects: Sayings of Confucius and His Successors, and: The Analects of Confucius: A Philosophical Translation.” China Review International 6.1:1–33. Makeham, John. 2003. Transmitters and Creators: Chinese Commentators and Com­ mentaries on the Analects. Cambridge, MA: Harvard University Asia Center. Makeham, John. 2008. Lost Soul: “Confucianism” in Contemporary Chinese Academic Discourse. Cambridge, MA: Harvard University Asia Center. Mou Runsun 牟潤孫. 1987. Zhu shi zhai conggao 注史齋叢稿. Beijing: Zhonghua shuju.

38

Makeham

Olberding, Amy. 2014. Introduction to Dao Companion to the Analects, ed. Amy Olberding, 1–17. Springer: Dordrecht. Shiji 史記. 1959. Beijing: Zhonghua shuju. Takeuchi Yoshio 武内義雄. 1943. Eki to Chūyō no kenkyū 易と中庸の研究. Tokyo: Iwanami shoten. Xu Shaohua 徐少華. 2007. “Lun zhushu Junzi wei li de sixiang neihan yu tezheng” 論竹 書 《君子為禮》的思想內涵與特徵. Zhongguo zhexueshi 中國哲學史 2:22–31. Yang Chaoming 楊朝明. 2003. “Xin chutu zhushu yu Lunyu chengshu wenti zai renshi” 新出竹書與《論語》成書問題再認識. Zhongguo zhexueshi 3:32–39. Zhou Fengwu 周鳳五. 1998. “Guodian Chujian Zhongxin zhi dao kaoshi” 郭店楚簡《忠 信之道》考試.” Zhongguo wenzi 中國文字 24:121–128. Zhou Fengwu 周鳳五. 2003. “Du Shangbo zhushu Cong zheng (jian pian) zhaji” 讀上博 楚竹書《從政(甲篇)》札記. http://www.jianbo.org/Wssf/2003/zhoufengwu01. htm#_ftnref19, accessed October 12, 2011. Zhu Yuanqing 朱淵清. 2003. “‘San zhi’ jie”《三制》解. http://www.jianbo.org/Wssf/ 2003/zhuyuanqing02.htm, accessed October 11, 2011.

The Lunyu As An Accretion Text

39

Chapter 2

The Lunyu as an Accretion Text Robert Eno In their impressive reinterpretation of the structure and history of the Lunyu, The Original Analects, E. Bruce Brooks and A. Taeko Brooks explain that their model of the Lunyu’s composition is based on an “accretion theory” of the text. They discuss briefly the history of such theories in connection with the Lunyu and apply in great detail their version of the theory.1 However, the Brookses do not define the term “accretion” or describe the range of textual phenomena and analytic tactics that an accretion theory may entail. This is not a defect in their book; their project does not depend on this type of definition. But the term can carry a range of meanings, and there are many ways that accretion theory may be applied. I am a proponent of applying accretion theory to the Lunyu, and although I do not accept the Brookses’ specific portrait of the way the Lunyu came to be a book, I think that any viable model of the etiology of the text must accommodate an accretion approach. My task here is to show the usefulness of an accretion approach, and I will pursue it in four stages. I will begin by attempting to clarify the concept of accretion and its usefulness in interpreting texts characterized by editorial disorder. A second section will discuss different ways in which the concept has been applied to the Lunyu by reviewing a number of past approaches, including the Brookses’. I will then analyze a particular feature of the Lunyu to argue that it must, indeed, be an accretion text. Finally, I will suggest how the historical conditions of the Qin-Han transition may have generated the accretion processes that yielded the Lunyu as a unified, canonical text.

The Concept of an Accretion Text

The phrase “accretion text” is ambiguous and can be used to denote a range of textual phenomena. Any text that has taken its current form through an additive process over time could be characterized as the product of accretion, even 1 Brooks and Brooks 1998, particularly appendix 1, “The Accretion Theory of the Analects,” 201–248.

© Koninklijke Brill NV, Leiden, 2018 | doi:10.1163/9789004382947_004

40

Eno

a single-author work that exists in multiple published or manuscript versions. However, in practice, the term is used to describe texts that are the products both of multiple authors and of multistage redaction processes. Thus, for example, a curated anthology of small texts, such as the Han Dynasty Shuiyuan 說苑 of Liu Xiang 劉向, would not usually be considered an accretion text; although its many component passages were authored by different people at different times, as a discrete text the Shuiyuan is the product of a single redaction process, guided by a single editorial agent.2 On one level, most traditional theories of the Lunyu have implicitly been accretion theories of a sort, because the approximately five hundred discrete passages of the Lunyu have historically been understood to have been the product of multiple authors—the original disciples and their followers— whose “records” of Confucius’s words and deeds, some recorded during the Master’s life or soon thereafter, were drawn together in a single book over some period of time, extending at least past the death of the disciple Zeng Shen 曾 參, depicted in 8/3–4 and datable to ca. 436 BCE, over forty years after Confucius’s death.3 However, “accretion” is generally used to denote more extensive disjuncture in the process of text creation. The term is used metaphorically, suggesting processes of natural growth, biological or geological, where gradual change occurs through undesigned but rationally intelligible processes. We can convey the force of this metaphor through a general definition: an accretion text is one that has been created through a series of dispersed processes of authorship and redaction resulting in a fragmentation of editorial agency, signaled by apparent disorder in the design of the text. This definition contrasts observed textual disorder with expectations that authors and editors design texts with intelligible organizing principles. Because the concatenation of multiple editorial agencies over time is an undesigned process, the multiplicity of intentions reflected in the ultimate form of an accretion text is heterogeneous enough that there is an appearance of disorder—a lack of structural or stylistic coherence, inconsistency of diction or of facts and assertions, repetition of passages, and so forth, all of which we would expect to have been resolved were a single, guiding editorial agency at work.

2 Variora in such texts may indicate some type of multiple redaction process, but unless discrepancies are major in scale, they would not affect this profile. 3 Recognition of the dating significance of Zeng Shen’s profile is not recorded before the Tang writer Liu Zongyuan 柳宗元 (773–819; quoted in Zhang 1954: 451–452), but it is hard to imagine that its general implications were unrecognized earlier. This traditional notion continues to have strong adherents, particularly in China. For example, in the introduction to his edition of the Lunyu, Yang Bojun (1984: 29–30) adopts precisely such a view.

The Lunyu As An Accretion Text

41

Moreover, to extend the geological metaphor, in an accretion model the lack of homogeneity or editorial coherence may be ascribed to a process of hardening, or resistance to alteration, that accrues to layers of text laid down earlier; the addition of subsequent material tends to leave earlier layers relatively untouched.4 In general, this tendency of redactors to avoid reordering and rationalizing existing text is explained by ascribing to early stages of the text some quality of authority, or what I will call here “canonicity,” which discourages later redactors from making radical alterations, constraining them instead to a limited range of strategies that include preposing or appending new text to old, interpolating new text without altering older text, or making minor adjustments in the wording of old text to make it say something new.5 The reluctance to revise and reorder existing text, even while adding new material, is the factor that creates apparent disorder. In the case of the Lunyu, an accretion theory serves as an alternative to two accounts of the text’s etiology. The first of these is traditional: that the text was the product of a single editorial community of disciples, stretching over no more than two or three generations, and was largely complete within the life spans of the first. The second of these accounts emerged only in the twentieth century and pictures the Lunyu as entirely the product of a brief redaction period just at or after the close of the Warring States period. Accretion requires both deep fragmentation in editorial agency and the time necessary to provide canonicity to early layers of text, which neither of these two profiles can provide. In Western text-critical traditions, the classic model of an accretion text is the Bible. Late nineteenth-century scholars of biblical hermeneutics, drawing on a century of precursors, outlined the “Documentary Hypothesis,” which posited that the Pentateuch was not a single-author text but a network of interwoven passages from a variety of sources, thus explaining a variety of inconsistencies and redundancies that had long puzzled close readers of the Hebrew Bible. Subsequent studies have performed similar analyses on the biblical chronicles and prophetic books, and New Testament scholars have pursued 4 Edward Slingerland (2000: 138–139), critiquing the Brookses’ specific accretion proposal, challenges the appropriateness of the geological metaphor, which I am extending. I hope the more flexible applications of the metaphor here will avoid the objections he raises. 5 My use of the terms “canon” and “canonicity” in this essay treats this imperative to preserve the words and form of a text, even if altering it by additions, as an index of a text’s authority. Of course, such a functional concept of canonicity involves matters of degree in the assessment of texts as authoritative—full canonicity would exclude alterations altogether. This usage is obviously different from, though related to, essential qualities ascribed to a text or its putative author, or criteria associated with church or government sanction of texts.

42

Eno

these strategies as well, particularly with regard to the synoptic Gospels. It is these practical studies that underlie the general notion of an accretion text and from which I have derived the general definition above.6 The Brookses wrote with an awareness of this tradition of biblical scholarship, which supplies scholars with many text-critical tools that we apply to the Lunyu today. However, in describing their accretion theory of the Lunyu, they based their model on an older, primarily East Asian scholarly lineage. If we examine the development of this lineage, extending it through the Brookses’ own work, we will see how the Lunyu fits the profile of an accretion text and the variety of ways in which its etiology has been understood.

Theories of the Lunyu as an Accretion Text

The list of earlier scholars the Brookses cite as advocates of the accretion approach is a short one, though it begins early, with the Song commentator Hu Yin 胡寅 (1098–1156). Hu was the first to note significant stylistic differences between Lunyu books 1–10 and 11–20, which he labeled the Shanglun 上論 and Xialun 下論 (Upper and Lower Lunyu), implying that the Lunyu might, in some sense, be a composite of multiple source texts.7 After Hu Yin, the Brookses list only Itō Jinsai 伊藤仁齋 (1627–1705) in Japan, Cui Shu 崔述 (1740–1816) in China, and Arthur Waley (1889–1966) in the West. Itō drew from Hu Yin’s division of the Lunyu the conclusion that books 1–10 represented an original (sei 正) and complete text, with book 10’s unusual form, recording Confucius’s habitual actions rather than his speech, signaling the coda of the work. Books 11–20 constituted a supplement (zoku 續), assembled at a later time.8 Cui Shu’s somewhat later but independent analysis was far more detailed, part of a systematic skepticism that Cui applied broadly in his text-critical work. Whereas Hu and Itō had focused on distinguishing between two moieties of the Lunyu, Cui focused at the level of the text’s twenty books. He claimed that features of the last five books marked them as later additions to an original and complete text of fifteen books.9 Like many who pursue 6 7 8 9

On the development of the Documentary Hypothesis and its extensions to later biblical books, see Nicholson 1998: 3–11; Campbell and O’Brien 2000: 11–12. For New Testament text criticism, Sanders and Davies 1989 is a comprehensive methodological guide. Brooks and Brooks 1998: 201. Itō’s comments appear in the preface to his Rongo kogi 論語古義. Zhang (1954: 453–459) records relevant comments extracted from several of Cui’s works. Cui had an optimistic view of the very early date of the books he regarded as original to the text, since he assumed that they were accurate life records of Confucius and his

The Lunyu As An Accretion Text

43

accretion analysis, Cui was concerned with the issue of authenticity (zhenwei 真偽).10 He wished to delineate an original core text that could be viewed as a faithful report of Confucius’s words and deeds, disentangling it from elements he felt were inferior or in conflict with his own view of Confucius.11 By contrast, in remarks that Arthur Waley included in the introduction to his 1938 translation of the Lunyu, Waley made clear that he was skeptical that any part of the text should be understood as a true report of Confucius’s own words. Waley did, however, follow Cui Shu in proposing an original core of the Lunyu text: he specified books 3–9, although he did not explain his reasoning, other than to point out that some other books, including those excluded by Cui, showed differences in style and substance.12

Japanese Sinology and the Accretion Theory

Waley’s speculations mark the end of the Brookses’ lineage of previous accretion hypotheses for the Lunyu. However, over the decades following Waley’s comments, three Japanese scholars—Takeuchi Yoshio 武內義雄 (1886–1966), Tsuda Sōkichi 津田左右吉 (1873–1961), and Kimura Eiichi 木村英一 (1906– 1981)—wrote pioneering monographic studies on the sources and composition of the Lunyu, using distinct methodologies and reaching different conclusions, which led to far more robust models of the Lunyu as an accretion text than anything Itō, Cui, or Waley envisioned.13 While their work is too

10

11

12 13

disciples. Thus, Cui regarded book 19, which records sayings of the disciples after Confucius’s death, as “late” but not inauthentic: it is late only with regard to books 1–15, which record the statements of the living Confucius. By contrast, books 16–18 and 20 are, for Cui, both late and of unreliable provenance (454). Issues of authenticity and accretion are often confused. While accretion analysis may identify certain sections or passages as late additions to a text, late addition does not mean late provenance: source texts added late may themselves have originated early. The absolute dating of text elements is a related but distinct issue. Cui is the earliest scholar I am aware of to identify an individual Lunyu passage as a late interpolation within an early book. He regarded the report of Lunyu 7/23, which states that Confucius met with the immoral consort of the ruler in the state of Wei, as inconsistent with Confucius’s character (Zhang 1954: 454). Waley (1938) 1989: 21–26. Waley’s analysis was perfunctory—he referred to it as a “guess”— but was influential because his translation was, for many years, treated as a standard English version. My thanks to Cameron Moore for his 2011 conference presentation, which provided broader and more fully contextualized portraits of the three monographic Japanese works discussed in this section. It is unclear why the Brookses do not list the contributions of these scholars in their account of the development of the accretion approach;

44

Eno

detailed to explore adequately here, I want to focus on those aspects that bear directly on the issue of how we may conceive of the Lunyu as an accretion text. In his 1939 monograph, Rongo no kenkyū 論語之研究, Takeuchi Yoshio extended the approaches of Itō Jinsai and Cui Shu to full text-critical scale. Relying heavily on records of variant Lunyu editions preserved in early sources, particularly in Wang Chong’s 王充 (27–100) Lunheng 論衡 (Balanced Discussions), Takeuchi built a model of the Lunyu as a composite text, comprised of an early core of two editions—a “Qi-Lu” 齊魯text in two juan 卷 (comprising books 1 and 10) and a “Hejian” 河間text in seven juan (books 2–8), associated with the school of Zeng Shen—as well as an alternative group of seven books (11–15, 19–20) derived from a Qi tradition with ties to the disciple Zigong 子貢, and four additional books incorporated individually (books 9, 16–18). For Takeuchi, the Hejian books were the earliest, linked directly to Confucius’s original disciples; the remainder likely postdated the Mengzi 孟子.14 While Takeuchi’s specific model can be criticized on a number of grounds, his general approach marks a major methodological shift. Unlike Cui and Waley, Takeuchi was not merely looking for an original core text: he sought to identify the Lunyu’s component source texts. While he treated one of the three main sources, the Hejian text, as the “original” Lunyu, he saw the Qi-Lu text and the group of books he assigned to the Qi tradition as being developed independently before being compounded with the Hejian text to create the received Lunyu. He saw the remaining chapters as appended or interpolated, most of them after the tripartite Lunyu had been compounded.15 Although Takeuchi pursued detailed interpretations of individual passages, like Cui Shu he focused his text-critical work on the Lunyu on book-level divisions, and his central motivating project was to disentangle early, more “authentic” reports of Confucius’s words and deeds from ones that were more suspect because of their late date. This latter project became the target of Takeuchi’s academic counterpart, Tsuda Sōkichi.

14

15

occasional citation of all three in the course of their discussion indicates awareness of these authors’ contributions. Takeuchi 1939: 106. Among accretion theorists, Takeuchi is perhaps most reliant on the external evidence of early reports of variant Lunyu. His use of the Lunheng’s somewhat garbled account of the Lunyu’s text history, which Takeuchi amended in critical respects (most significantly by amending on slender grounds Wang Chong’s count of nine juan in the Hejian edition to a count of seven), facilitates a model of source texts that Takeuchi may have conceived initially on the basis of internal evidence. According to Takeuchi (1939: 95–101), book 9 was probably an addendum to the Hejian text, implying that its addition may have preceded the editing of the compound Lunyu text, while books 16–18 were added last.

The Lunyu As An Accretion Text

45

Tsuda was not an advocate of the accretion approach: though he granted that the Lunyu was in some sense a composite, he challenged the accretion approach and in doing so contributed to its further development. Writing with an awareness of the comprehensive techniques of biblical higher criticism, as well as the critical methods that had been applied to the analysis of the Buddhist canon, Tsuda believed that the goals and methods of previous Lunyu analysts were fundamentally misconceived, and that more sophisticated analysis would undermine models of long-gestating source texts, such as Takeuchi’s.16 In his Rongo to Kōshi no shisō 論語と孔子の思想, published in 1946, Tsuda repeatedly used external evidence to question whether key passages in the Lunyu were associated with Confucius even as late as the Han era. For example, he argued that virtually every book in the Lunyu was pervaded by passages that reflected the intellectual environment of the late Warring States era, involving ideas drawn from Daoism, mantic traditions, and forms of canon exegesis that were unknown during Confucius’s time.17 He identified many Lunyu sayings ascribed to Confucius that appear unattributed in late Warring States texts, including Confucian texts such as the Xunzi 荀子, suggesting that the Lunyu appropriated these passages to attribute them to Confucius, treating later texts as sources, rather than the other way round.18 Going further, Tsuda found Lun­ yu material quoted by Confucians without expected attribution to Confucius even in the late Western Han.19 Using criteria such as these, Tsuda was able to point to so many Lunyu passages that seemed anachronistic or to have been independent of the text at a late date as to undermine the notion that any sustained portion of the text could have been established before the last years of the Warring States era or that there was any way to date the Lunyu other than to assess passages on an 16 17 18

19

Tsuda 1946: 514–515. Tsuda 1946: 157–218. Tsuda 1946: 158–160. It is an anomalous feature of the Xunzi that, although Confucius is cited as authority for many passages in the text, every apparent quotation of the Lunyu is unattributed to Confucius (altogether six instances, each from a different Lunyu book; see Eno 1990: 239n2). Tsuda (1946: 6–22) related this to broad patterns of variation in attribution and wording of maxims across the pre-Qin and early Han corpuses. For example, Tsuda adduced a memorial submitted in 74 BCE by the Confucian minister Wang Ji 王吉, which invoked without attribution phrasing found in Lunyu 17/19, despite the fact that elsewhere Wang is conscientious in citing Confucius by name (Tsuda 1946: 284, citing Hanshu 10.3061). (We now know that 17/19 was included in the Dingzhou version of the Lunyu, datable prior to 55 BCE, considerably lowering the odds of Tsuda’s conjecture in this case and increasing the likelihood that the same Hanshu text could be evidence that familiar Lunyu passages might routinely be cited without attribution, at least during the Han.)

46

Eno

individual basis.20 Although Tsuda did not specifically frame his study as a response to Takeuchi, his conclusions undermined Takeuchi’s premise that the twenty-book structure of the Lunyu could provide a basis for reconstructing its etiology and that the clues to the Lunyu’s textual history reported in Han sources could provide a blueprint.21 Tsuda believed that the great majority of the Lunyu’s books, sixteen of twenty, were the products of a single milieu and that their order in the text was not significant.22 Most important for us, in Tsuda’s view, whether those sixteen books were assembled by a single editor or by several—either of which he deemed possible—the intellectual and formal similarities so outweigh differences that there was no need to suppose that composition extended over any long period of time.23 Tsuda’s critique motivated Kimura Eiichi to develop a more refined model of the accretion approach in his 1971 monograph, Kōshi to Rongo 孔子と論語. Kimura articulates with great clarity the basic argument for applying an accretion approach to the Lunyu: Now, suppose the Lunyu had been compiled by one person or at one time: what results would we expect? Surely we would expect to find some significant kind of pervasive editorial plan. For example, we might find that passages were sorted by topic or theme. Or perhaps entries would be arranged in chronological order or grouped together on the basis of similar keywords. Even if we found the arrangement to be totally random, revealing that there had been no consideration of order whatever, that too would reflect an editorial principle. However, while we see in portions of the received Lunyu text examples of many such types of meaningful passage ordering, we see no overall editorial principle at work. 今かりに論語が、或る一人による或一時の編纂に成ったとすれば、 どういう結果になるであろうか。思うにそこには、必ずや何等かの 意味で、全書を一貫する何等かの編纂意圖が見られる筈である。例 えば事項別とか論題別とかに篇目章次が定められているとか、年代 順に排列が考慮されているとか、相似た文辭を集めて前後の聯關が つけられているとか、或は全く雑纂の體によって、如何なる排列順 序も考慮しないという態度を示しているとかは、いずれも全書を一 貫した編纂方針の表現である。ところが今の論語には、部分的に種 20 21 22 23

Tsuda 1946: 500. Tsuda 1946: 238–239. Tsuda 1946: 278. The exceptions are books 16, 18–20, which he acknowledged as distinct in origin; he also believed that books 5 and 6 were compiled together. Tsuda 1946: 272–274.

The Lunyu As An Accretion Text

47

々の意味での章次の聯關が見られながら、全體を一貫した編纂方針 が見當らず.24

In other words, the problem the Lunyu presents is one of editorial disorder. Kimura’s claim was not that the text had no order at all—as he notes, that too would serve as a consistent expression of editorial intent. The problem is that, while there are portions of the text that have no perceivable order, many portions, of variable size, are ordered quite clearly, though the ordering principles of these sections are varied. The greater part of Kimura’s extended analysis is, in the end, a discovery of the many kinds of order in the text. The order that he elucidates, however, is fragmented; different sections of the text exhibit different ordering principles in a way that cannot be cogently interpreted as the work of a single editorial authority. Perhaps Kimura’s most distinctive contribution to the theory of the Lunyu as an accretion text was his close analysis of microstructures within the Lunyu that he called “passage clusters” (shōgun 章群). Clusters of passages, some brief, others extended, do seem linked by topic or theme, by keywords, by passage structure, and so forth, even in the less coherent books. Kimura articulated five general editorial strategies governing these clusters, and his specific analyses illustrate great variety in the way these strategies are deployed.25 It is through these pockets of coherence that the component elements of the Lunyu as an edited collection, operating on a level beyond the individual passage, can be identified, and these are the basis for critical analysis of the Lunyu as a text and for explorations of its history. Kimura devotes most of his analytic energy to these clusters, which, on occasion, bridge book divisions.26 In this way, Kimura’s analysis mediates between Takeuchi and Tsuda, focusing neither on the level of the book nor on the level of the individual passage but on the level of passage clusters, searching for and interpreting “joins” between clusters, which can provide clues for both demarcating source texts and identifying which Confucian factions may have been responsible for both the sources and their redaction into a larger framework. On this basis, Kimura 24 25

26

Kimura 1971: 171–172. The five editorial strategies include: stringing together passages alike in content; sequencing strings with different content through connecting passages sharing features of both strings; establishing an initial string on one theme and adding strings that are variations of it; placing resonant passages as framing markers around a string to bring out an implicit theme; appending a string through late redaction to inflect the meaning of prior text (Kimura 1971: 219–228). For example, Kimura (1971: 269–270, 480) sees the final passages of book 2, which concern issues of ritual (li 禮), as a bridge to book 3, which is devoted almost entirely to ritual concerns.

48

Eno

develops a complex model of the etiology of the books, assigning relative dates and places of origin in terms of disciple generations and faction location.27 Despite differences in key assumptions, specific interpretations, and ultimate conclusions, Kimura’s methodology appears to be the closest antecedent of the Brookses’ methodology.

The Brookses’ Accretion Theory

To date, by far the most detailed attempt to chart the accretion process for the Lunyu is represented by the work of the Brookses. The goal of their analysis is a complete map of the development of the Lunyu, passage by passage, discovering within each book individual principles of order and dating each book and the timing of each later interpolation into an existing book within a narrow range.28 Their project is a form of “redaction criticism”: rather than focusing on the question of source texts, the Brookses’ address the question of the date and perspective of successive generations of redactors, explaining their contributions to the text in terms of the unique contexts of the eras in which they lived.29 To accomplish this goal, the Brookses adopt bold hypotheses: they posit that the earliest portions of the text were compiled under the guidance of Confucius’s original disciples in the state of Lu 魯: first under the leadership of Zigong, followed by, tentatively, one or two other original disciples, then shifting to Zeng Shen, who became the acknowledged master of a single branch of Lu Ruism that composed, preserved, and enlarged the Lunyu. Curatorial control of the text is pictured as being solely in the hands of the group’s designated leader. The Brookses argue that this leadership role remained with Zeng Shen’s lineal heirs for one further generation and then passed to Confucius’s grand27

28

29

For Kimura (1971: 473–475), books 1, 8, and 4, emerging in that order, are the earliest, derived from Zeng Shen’s school in Lu, which also produced books 14 and 17 somewhat later. Other Ru communities in Lu produced books 2–3, 5–7, and 9–10 relatively early, and books 11, 13, and 15 later. Communities in Qi produced the late books 16 and 18–20, with book 11 a hybrid, begun in Lu and completed in Qi. Kimura’s analysis of “passage clusters” is paralleled in the Brookses’ work by their analysis of each book into thematic subsections, treating as late interpolations passages that break continuity within subsection confines. Like Kimura, the Brookses are exceptionally conscientious readers of the Lunyu, and their structural analyses of its books are full of critical insights. Traditional redaction criticism refers to these issues of editorial context by the German phrase Sitz im Leben, or “life setting.” The Brookses support their critical conclusions through extensive engagement with Warring States history and texts.

The Lunyu As An Accretion Text

49

son, Zisi 子思, whose heirs controlled it thereafter.30 These hypotheses permit the Brookses to reconstruct the text’s history with remarkable specificity, deploying a range of text-critical methods well established in biblical studies. However, the validity of their key assumptions of authorial lineage and the insularity of the growing Lunyu text is lightly argued and supported by neither internal nor external evidence, other than the Brookses’ interpretations of internal evidence dictated by their governing theory itself. For the Brookses, the authoritative character of the Lunyu is established from the outset. They imagine book 4 as the product of a specific context, Confucius’s death, which they picture as the occasion for the creation of a “memorial” text, composed by the disciples, most likely under the direction of Zigong. This textual genesis ties book 4 to immediate, living memories of the Master, selected by the original inner circle at a sanctified point of heightened care. After this authoritative genesis, succeeding books derive their authority from the lineally sanctioned editorial chain of school leaders in control of the text.31 The text is pictured as growing in a linear fashion, with school leaders adding a book about every ten years over the course of nearly two centuries, until, with the absorption of the state of Lu by the state of Chu 楚 in 249 BCE, the school leader, Zishen 子慎, flees west, sealing the text away in the wall of his home and ending its life as a growing repository.32 It should be clear from this description that although the Brookses label their model an accretion theory of the Lunyu, it is a most unusual type of accretion, fundamentally different from the models developed in biblical studies. Although editorial agency is fragmented, it is also unified by an order of licensed authority, passed, for the most part, from father to son. There is also 30

31

32

The Kong congzi 孔叢子 and the “Kongzi shijia” 孔子世家 (Hereditary House of Kongzi) chapter of the Shiji 史記 provide lists of the Kong family heirs, including their names and life spans, which permits the Brookses to correlate individual leaders with what they interpret to be an appropriate Sitz im Leben for Lunyu books and interpolations. Every dated passage is assigned to a unique, identified redactor. However, apart from name and life span (and in two cases an added phrase on an office held), existing sources provide no documented personal characteristics for the figures after Zisi, and it is the Brookses’ model, rather than external sources, that supplies specific qualities to these featureless lists of heirs. Perhaps weakened to some degree when the school lineage leadership did not follow kinship lineage. The Brookses’ model allows no interpolations to disturb the earliest books until school leadership passes from Zeng family members to Kong family members. Thereafter, interpolation became an ongoing editorial strategy. Brooks and Brooks 1998: 196. The Brookses see Zishen’s hypothesized act as the origin of the Gulun 古論 (Ancient Lunyu) text, ancestral to today’s received Lunyu, which was reported to have been accidentally recovered from a wall of the Kong family compound in 157 BCE.

50

Eno

no compounding of distinct urtexts: accretion is entirely a matter of text expansion according to set rules for appending or preposing books and interpolating passages. The Brookses do not engage in source criticism because, for them, there is only one source text. The peculiarity of the Brookses’ approach to the accretion model lies in a particular feature of their philological project. Whereas the general text-critical function of an accretion theory is to explain textual disorder, the Brookses’ basic goal is to date the component books and passages of the Lunyu. These are distinct issues, and so long as a passage can be fitted into their dating scheme, the Brookses do not treat it as a sign of editorial disorder. For example, verbatim passage repetition is one type of editorial disorder that an accretion theory can help explain, since compounding source texts may include duplications that a conservative editor will leave intact, respecting the authority of each text. When the Brookses encounter verbatim duplication in the Lunyu, no disorder is recognized. For example, in the case of identical passages such as 1/13 and 17/17, the author of 17/17, identified as the son of the author of 1/13, is simply seen as repeating the passage because its expression was germane to him in the same way it was to his father.33 For all its technical virtuosity, the Brookses’ project has the significant problem that there exists no evidence in support of the a priori assumptions on which its success depends. There is no record to indicate that the anchor date for book 4 is accurate, that redaction authority was limited to a single editorial line, that Confucius’s lineage played a continuing doctrinal leadership role, or that the text was kept under confidential control for nearly two centuries. While these issues are problematic for the Brookses’ particular version of the accretion approach, they do not bear on the validity of the approach in general. I want now to draw on some elements of the accretion theories reviewed here to comment on features of the Lunyu that I think make the accretion approach persuasive and then to outline one model of the general historical processes that would have yielded the text we have today.

The Significance of Order within Disorder in the Lunyu

Earlier, I quoted at length Kimura’s insight into the central issue for the accretion approach: the editorial disorder of the Lunyu. Tsuda, arguing against an accretion approach, did not feel the force of this disorder, and the Brookses make it a subordinate issue where it conflicts with their premises. But I believe 33

Brooks and Brooks 1998: 164. See also the Brookses’ treatment of repetition in 6/27 and 12/15 (1998: 93).

51

The Lunyu As An Accretion Text 1

1

2

2

3

3

4

4

5

5

6

6

7

7

8

8

9

9

10

10

11

11

12

12

13

13

14

14

15

15

16

16

Figure 2.1 Diagram of Lunyu book 1. (Shaded passages quote disciples as authoritative voices.)  Note: The resonance between Lunyu 1/1 and 1/16 (noted early by Takeuchi [1939: 90]) depends on selecting one of two commentarial interpretations for 1/1; scholars have long been divided on which is to be preferred. It would be plausible to argue that the final editors of book 1 read 1/1 according to the sense implied by their placement of the resonant 1/16, but that an alternative that ignores this was transmitted prior to the incorporation 1 2 resonance3exists - 7 precisely because8 it- 17 18 - 21 of 1/1 in the Lunyu.

1

2

3-7

8 - 17

18 - 21

Figure 2.2 Diagram of Lunyu book 8. (Zeng Shen passages are more darkly shaded; quotations of Confucius are lightly shaded; comments on ancient sages are unshaded.)  Note: This analysis differs slightly from Kimura’s (1971: 322–329); see Eno 2010.

Kimura is right: editorial inconsistency is a pervasive feature of the Lunyu, and it is not conceivable that a single editor or a closely coordinated editorial team would have constituted 1 - 19the text in this way. 20 - 25 Kimura’s point about the Lunyu’s combination of orderliness on the microlevel and lack of overall editorial coherence can be illustrated by considering the shifting forms of organization that characterize individual books of the 1 - 19 20 - 25 text. For example, book 1, among the most clearly ordered books in the Lunyu, is organized as an anthology of school sayings, alternating teachings attributed to Confucius and those attributed to his disciples in sequences of one to two passages, all framed within the “bookends” of thematically resonant initial and final passages (see fig. 2.1). Book 8, on the other hand, is not organized as an anthology, but is composed of three clearly separate component texts, bound together by a nested structure that places a block of the disciple Zeng Shen’s teachings in old age within a selection of disparate teachings attributed to Confucius, all set again within a body of texts celebrating heroes of remote antiquity (see fig. 2.2).

52

Eno

1 - 19

20 - 25

Figure 2.3 Diagram of Lunyu book 19. (Passages quoting Zigong are shaded.)

Book 19, which is devoted entirely to the teachings and sayings of disciples after the death of Confucius, is organized as a compound text. It combines a relatively orderly string of passages involving the junior disciples, in competition with one another and in conversation with their followers, with an appended section that is devoted to the senior disciple Zigong in conversation with various grandees of the Lu court and in which independent origin is clearly signaled by consistent use of Confucius’s style name, Zhongni 仲尼, unique within the Lunyu (see fig. 2.3). Some books are models of thematic consistency. For example, almost all passages of book 3 are explicitly devoted to the theme of ritual and music, excepting only 3/5, a comment on the superiority of the Zhou states over nomadic and forest peoples, which appears to address the theme of ritual order implicitly, and 3/24, an isolated narrative concerning Confucius that does not invoke the theme and is thus, perhaps, a late interpolation. Other books are so disordered as to defy mapping. For example, the Brookses’ analyses of books 14 and 15 identify over 40 percent of the passages in each chapter as late interpolations in order to tease some type of editorial unity out of the remainder. 34 If we move from the level of the individual book to the text as a whole, most scholars (Tsuda seems to be the only exception) see the stylistic distinction that first prompted Hu Yin to divide the text into the moieties of the Shanglun and Xialun; and books 16, 18, and 20, at least, do seem to belong to separate intellectual discourses, as Cui Shu and all the Japanese scholars agree. But within these broad divisions, I believe one group of books is particularly suggestive of a core component text: books 3–5 and book 7. This group of books is distinctive because each book is internally homogeneous in form and theme, and they are mutually complementary in content. The uniform content of book 3, on the theme of ritual and music, has been noted above. Book 4 is a collection of brief ethical maxims, with two of the twenty-six passages, 4/15

34

For a very different analysis of the text’s editorial diversity, see Brooks and Brooks 1998: 248.

The Lunyu As An Accretion Text

53

and 4/26, often identified as interpolations.35 Book 5 collects comments on individuals.36 Book 7 collects comments on Confucius and descriptions of his manner.37 The fact that book 6 intervenes in this sorted collection of books appears to be a simple case of interpolation: the first half of the book (6/1–16) is uniform and duplicates book 5 in theme, explaining its point of insertion, while the second half (6/17–30) is a short miscellany of maxims. On the basis of this reasoning, I will treat books 3–5 and 7 as originally conjoined. A group of four books sorted into complementary themes, in the midst of sixteen other books that are, to greater or lesser degrees, heterogeneous, makes sense if those four books were at some point an independent collection. Thematic sorting is the sort of editorial principle that Kimura found lacking in the Lunyu as a whole; he sought it at the level of the passage cluster, but does not appear to have considered the possibility of a “book cluster.”38 On the theory that book 6 was probably interpolated before this cluster was incorporated with other source texts for the Lunyu, I am going to refer to books 3–7 together as the “Shanglun Core Source.”39 My minimal claim is that the redaction of the materials that composed the Shanglun Core Source was clearly performed at a different time and on different principles from other episodes of redaction that shaped the received Lunyu. When additional materials that shared themes with the chapters of the Shanglun Core Source were incorporated in the text, the Core Source books, apart from scattered interpolations, were not revised to accommodate them, 35 36

37 38

39

See, e.g., Kimura 1971, 290–291; Brooks and Brooks 1998, 149. Additionally, passages 4/5 and 4/6 seem to have been expanded by preposing and appending elements to an original maxim. Passage 5/13 appears to be a late interpolation, and 5/26–28 are not as straightforwardly on topic as the main body of passages. Among the four books discussed here, book 5 is the least homogeneous, and even so, 85 percent of its entries conform to the main theme of the book. Five of thirty-eight passages diverge from the theme: 7/6, 7/36, and 7/37 are brief maxims, and 7/26 and 7/29 are more complex diversions. Kimura’s final proposal for the accretion process assigns book 4 to a different Confucian faction from the others in this group, and book 7 to disciples one generation later than the rest. The Brookses treat books 4 and 5 as serially composed, with book 6 intervening before the addition of book 7, all over a period of about forty years, with the other member of this set, book 3, composed about a century later, after several chronologically intervening chapters (the late date for book 3 reflects their theory that the Confucian school focus shifted from ren 仁 to li 禮 when leadership in Lu moved from the Zeng lineage to the Kong lineage; see Brooks and Brooks 1998: 59). The notion of this four-book cluster is more closely compatible with Waley’s perception that books 3–9 form a core source and Takeuchi’s theory that books 2–8 represent a preserved Hejian edition. An argument can be made for viewing books 11–13 as a similar, sorted core source in the Xialun portion of the text, but I will not pursue that issue here.

54

Eno

and no further general sorting by theme occurred at the level of the full text.40 This claim is inconsistent with the processes of redaction pictured by Tsuda, and it is strong evidence that the received text is the product of an accretion process. I believe that strong arguments, incorporating aspects of the proposals of Takeuchi, Kimura, and the Brookses, can be presented for a more extensive accretion model, but I will not pursue them here, since my argument requires only that an accretion process be established in principle. There is a corollary to this claim that will form a bridge to the next step of this analysis. It is that, regardless of whether the date was early or late, we can say that at the time the Shanglun Core Source took its present form, the component sources for that redaction were not regarded as “canonically” sacrosanct, in the functional sense of the term “canon” used here.41 Respect for the canonical authority of a text is, in this analysis, signaled by an unwillingness to so radically alter the form of a text that the new redaction loses lineal continuity with its source. In such a canonical context, redaction and preservation are companion imperatives.

The Historical Context for the Compilation of the Lunyu

In this section I will consider how the accretion process may have occurred in the case of the Lunyu, covering a period that may have extended from the close of Confucius’s life until the date of the earliest recovered manuscript copy, placed in a grave closed in 55 BCE.42 I believe that the basic conditions for the gestation of an accretion text such as the Lunyu would involve the following succession of five stages: 1. The generation of passages that will, intact or altered, find their way into the received text; 2. The collection of these passages in manuscripts that will become the source texts for the Lunyu; 3. The assembly of these collections as components of a larger text; 40

41 42

Kimura (1971: 172–173) makes this point as well: single-theme chapters in the Lunyu highlight the disorder implied by the presence of similarly themed passages in other chapters that show no thematic unity. It is possible that some thematically consistent passages are late interpolations in chapters without thematic unity, but the fact that many thematically relevant passages were instead located elsewhere indicates that thematic unity was not a consistent redaction principle. Although I am claiming that the Shanglun Core Source predates the Lunyu’s final redaction, I am not claiming that it was necessarily earlier than other components of the received text. Hebei sheng wenwu yanjiusuo 1997: 1.

The Lunyu As An Accretion Text

4. 5.

55

Concurrent or sequential iterative processes of canonization that endow components and the growing composite text with such authority as to restrain any editorial initiative to reorder the text as a whole; Continued conservative editorial alteration through strategies of appendage and interpolation on the levels of the book, the passage, or wording within passages, until canonization is formalized or reaches a point where reverence for the disseminated text makes it inalterable.

I will devote most attention here to the key third and fourth stages, which seem to me the most challenging to explain concretely. I will argue that the third stage was most likely a by-product of larger-scale bibliographic activities of Ru 儒 and others under the sponsorship of the Qin state and empire, and that the fourth stage was most likely the product of a period of Ru persecution during the early decades of the Han. First, however, I want to comment more briefly on what we know concerning the initial two stages.

The Emergence of Confucian Aphoristic Collections

I do not believe it is possible, on the basis of current evidence, to pinpoint the time at which Confucius lore began to be recorded in writing. The Brookses speculate that Lunyu passages were first composed at the time of Confucius’s death. While this is certainly possible, no positive evidence exists to support it, and the speculation assumes a type of text-based culture that may not yet have existed.43 Writing was long established in certain contexts, but the notion of Confucian disciples recording the Master’s words—though pictured once in Lunyu 15/6—may be anachronistic, a better fit for the following century, when a variety of masters traditions competed for patrons and followers by circulating school texts. This does not mean that component passages of the Lunyu may not have begun with or before Confucius’s death; however, the likelihood is that any such aphoristic teachings would have been memorized and orally transmitted individually, not as an ordered collection with fixed content and sequence.44 43

44

The Brookses (1998: 204) claim a resemblance between the language of book 4 and contemporary bronze inscriptions and bamboo chronicle entries, but I believe there is little in common with bronze-inscription rhetoric and we have recovered no pre–Warring States court chronicles to assess the form they took at the time. For reasons of space, I will not consider further here the role of oral transmission in the formation of the Lunyu. I believe it is likely that during the Warring States era there was a widespread shift in authority from the spoken to the written word, and that this must be

56

Eno

Recently recovered bamboo grave manuscripts, dating from about 300 BCE, do testify to a robust culture of sectarian and other private writing by that date, and they provide a terminus ad quem for the first Confucian aphoristic collections.45 While no apparent source texts for the Lunyu have yet appeared among these Warring States era manuscripts, two of these documents consist, like the Lunyu, of strings of relatively brief aphorisms and discussions, quoting Confucius and his disciples. Contemporary editors have titled these documents *Junzi wei li 君子為禮 (The Noble Man’s Practice of Ritual) and *Dizi wen 弟子問 (The Disciples Ask). 46 Although these specific texts differ in content from the Lunyu and are not direct source texts, each includes passages that show clear filiation to the received text, either as near duplication or recognizable variant, or in terms of characteristic form and rhetoric.47

45 46

47

a central consideration in dating the Lunyu. In biblical contexts, issues of oral sources are usually pursued through analyses of formulaic language, or “form criticism,” and there are ways in which this can be applied to the Lunyu. I am considering here bamboo texts recovered from Baoshan 包山 and Guodian 郭店 in Hubei, as well as the Shanghai Museum and Qinghua University collections, presumably from the same era and region but of undocumented provenance. Huang Ren’er (2011) and others argue that these two texts are, in fact, a single text that intermingled the strips published separately, and they substantially rearrange passages on that theory. My own analysis of the calligraphy of the strips, going beyond the comments provided by the original commentator, Zhang Guangyu 張光裕, confirms that there are clearly two different, but similar, calligraphic hands at work and that the division proposed by Zhang and the Shanghai Museum staff does indeed, with one exception (*Dizi wen strip 3), track that distinction. Although this does not rule out the possibility of a single text with scribes alternating strips, no rearrangement I have seen is adequately compelling to reverse the presumption that these are distinct texts. The similarity in calligraphic styles between the two texts suggests that they were the product of a single workshop (on the basis of the model of calligraphy traditions and scribal workshops in Bin 2014: 4–9), and it may be that they were sequential parts of a single text, prepared by scribal colleagues, or even two books of one larger compendium. The clearest specific overlaps are *Dizi wen, Appended Strip, and Lunyu 1/3 and 17/17, and *Junzi wei li, strips 1–3, and Lunyu 12/1, the latter involving considerable variation from the received Lunyu. However, on the level of form and rhetoric, there are at least nine more passages that mark these texts as drawing from the same homiletic tradition as the Lunyu (see Yi [Eno] 2011: 536). In addition, the *Yucong 語叢 3 text, a string of maxims recovered at Guodian, includes two unattributed Lunyu quotations—7/6 (likely a late interpolation in book 7, see n. 37) and 9/4—and the *Zun deyi 尊德義 text, a sustained exposition, appears to embed Lunyu 8/9. These Confucian texts are not explicitly collections of Confucius’s aphorisms, like the two Shanghai Museum examples; they remind us that some maxims in the received Lunyu circulated outside Confucius-lore contexts.

The Lunyu As An Accretion Text



57

The Likely Role of Qin Encyclopedism

With certainty established for the existence by about 300 BCE of passages that appear in the Lunyu and of Lunyu-like aphoristic texts concerning Confucius and his disciples, though not the existence of specific source texts for the Lunyu, we now turn to the question of how and when the third stage of the process outlined above took place: the assembly of source texts into the embryonic Lunyu (bearing in mind that a smaller-scale redaction of sources for the independent Shanglun Core Source would have come earlier). As is well known, the sole text of possible pre-Han provenance that refers to the Lunyu by name is the “Fang ji” 坊記 (Embankment Record) chapter of the received Liji 禮記 (Records of Ritual), and the date of that text is a factor in this issue. Traditionally viewed as a pre-Qin text, for most of the past century the dominant opinion has been that it dates from the early Han.48 However, the “Fang ji” has also been seen as a member of a linked four-chapter corpus within the Liji, all texts associated with Confucius’s grandson, Zisi. Since a pre-Qin manuscript of the associated “Zi yi” 緇衣 (Jet-Black Robes) chapter emerged in both the Guodian find and the Shanghai Museum collection of recovered Warring States texts, the likelihood that we will discover that the “Fang ji” too was a pre-Qin product has increased.49 Models of a Lunyu accretion process can accommodate either the assumption that the text had not reached a stage of completion warranting a name by the rise of the Qin or that it had. Here, I will make the conservative assumption that the “Fang ji” is a Han product and then suggest the alternative scenario. The Qin-Han transition period is often viewed as a watershed for intellectual history in general and Confucianism in particular.50 I believe it was, but not in the way it is usually construed. Over half a century ago, Kanaya Osamu 金谷治, a student of Takeuchi’s, published a radical rethinking of the role of Confucianism in the Qin period, which portrayed the Qin as an era when Con­ fucianism was the beneficiary of important state support for purposes focused on the encyclopedic assemblage of the knowledge of the past. In what follows,

48 49 50

For a survey of views, see Lin 2008. Lin (2008: 34) notes that Takeuchi regarded the “Fang ji” as pre-Qin but the reference to the Lunyu as a late commentarial addition. Li Xueqin 李學勤 believes that because the Guodian *Yucong 1 text includes overlaps with the “Fang ji,” a pre-Qin date for the “Fang ji” has been proven, but as Scott Cook (2012: 886–887) argues, Li has overstated the evidence. John Henderson (1991: 39–40) notes a general relation of historical disjunctures on the scale of that occurring between pre- and post-Qin China to the treatment of texts from the prior era as canonically authoritative.

58

Eno

I will adapt Kanaya’s insights to apply them to the question of the Lunyu’s etiology.51 Traditionally, the Qin period has been understood as an era of Confucian persecution: the Qin is said to have persecuted Ru and excluded them from governance; the Qin ban on books is seen as destroying Ru texts; the Qin is understood to have pursued a policy of eradicating the learning of the past.52 None of these claims stands up to scrutiny. All appear to be products of a rewriting of history to serve parallel interests of Ru and the Han court during the period of Ru ascendance to orthodoxy during the mid- to late second century BCE. The records of the Shiji indicate that after the Qin conquest, a substantial number of Ru were appointed to the rank of “erudite,” or boshi 博士. Boshi, including the Ru, maintained retinues of disciples at the capital, were consulted by the emperor, and were charged to assemble as an advisory body on important matters at court.53 Despite significant tensions between Ru boshi and the government, recorded in the Shiji, Ru boshi remained part of the Qin court as late as 209 BCE, just as the Qin entered its final stage of dissolution.54 In addition to advisory duties, Qin boshi were charged to “comprehend the past and present” (tong gujin 通古今).55 The practical meaning of this phrase may be reflected in the terms of the book proscription of 213 BCE. While the Qin instituted a ban on the ownership of unapproved books and burned those in private possession, the government did not, in fact, destroy the texts held at court: boshi, as appointees in charge of intellectual knowledge, were specifically exempted from the ban as a consequence of their official duties (zhi 職).56 This exemption indicates that boshi were charged to perform text-related services for the dynasty. 51

52 53

54 55 56

See Kanaya 1960: 230–257. I became aware of Kanaya’s work through Martin Kern’s excellent précis (Kern 2000: 191–194). Kern took to task Western writers who had been offering revisionist accounts of the Qin without awareness that Kanaya had anticipated such work with superior arguments; I was among their number. I am omitting from this list the traditional tale of the murder of the Ru, which has been widely debunked (see Kanaya 1960: 234–235; Bodde 1987: 95–96). The appointment of Ru as boshi is confirmed in the Shiji “Fengshan shu” 封禪書 (28.1366), and some well-known Ru are identified as Qin boshi, including those discussed below. Attention is often given to the tension between the First Emperor and the Ru that ensued after the emperor summoned them to advise him on the fengshan rite. More significant is his convening them in the first place. Shiji 99.2720–2721. Note that six years after the tensions of the fengshan incident we also find the boshi Shunyu Yue 淳于越, plainly a Ru, offering counsel at court (Shiji 6.254). Hanshu 19A.726. Shiji 6.255.

The Lunyu As An Accretion Text

59

What those services may have been is suggested by the composition of the twenty-nine-chapter Shangshu 尚書, a book explicitly banned by the Qin. Perhaps the single greatest anomaly in the Shangshu is the fact that concluding an assemblage of chapters ascribed to the legendary sage-kings and the leading figures of the Xia, Shang, and Zhou dynasties, the book closes with the “Qin shi” 秦誓 (Harangue of Qin), a speech by Duke Mu 穆公, ruler of the Spring and Autumn era state of Qin. While the inclusion of this coda in a canonical text has been explained as a Qin addition to an existing compilation,57 there is, in fact, no reason to believe that any single canon of shu 書 had been fixed prior to the Qin conquest. Mohist and Confucian references to “shu” differ, and the term was used flexibly to refer to authoritative texts of antiquity rather than a fixed canon.58 It is far more cogent to conclude that the Shangshu was a compilation made by Qin boshi as part of their mandate to “comprehend the past and present.”59 After all, the existence of the received text is due to its preservation by a Qin era Ru boshi, Fu Sheng 伏勝, about whom I will say more below.60 A second example would be the Lüshi chunqiu 呂氏春秋 (Annals of Lü Buwei), ordinarily viewed as a preconquest product of Qin, assembled under the patronage of Prime Minister Lü Buwei 呂不韋, who, it is said, hung the completed text on the city wall of the state of Qin in 239 BCE, challenging anyone to improve it by a single word. There are serious problems with this dating of the text. Many of these have been identified by D. C. Lau, and they include the puzzling location of the text’s brief afterword, titled “Xuyi” 序意, at the close of the first of the book’s three major components, the ji 紀 division, indicating that the following lan 覽 and lun 論 divisions were added at a later date.61 Moreover, the “Xuyi” chapter discusses only the twelve subdivisions of the ji division, and it does so in terms of a cosmological framework that is most reasonably understood not as a comment on the subdivisions as we have them today but as a description solely of the lead chapters of the twelve subdivi57 58 59 60

61

Matsumoto 1966: 581–582. Thus, the explicit Qin ban on “shu” would have referred to all versions of such authoritative historical texts rather than a single version ancestral to the received Shangshu. Only four of the twenty books in the “Zhou shu” 周書 section of the Shangshu are cited in pre-Qin texts and manuscripts recovered to date, which may further indicate that the canonical bounds of the text were not set prior to the Qin. On this model, the “Da xue” 大學 (Great Learning), which quotes the “Qin shi” at length, may also be seen to be a Qin boshi product; see Matsumoto 1966: 572–573, 581. A second philosophical summa, the “Zhong yong” 中庸, may also date from the Qin (Kanaya 1960: 355). Lau 1991: esp. 46–53. For a brief summary of approaches to this issue, see Knoblock and Riegel 2000: 27–28.

60

Eno

sions, the calendrical yueling 月令 chapters, which appear as an independent text in the Liji. If the notion of a perfect text of 239 BCE is understood to be limited to the yueling chapters, the Shiji account is far more cogent, as it is difficult to credit the supposition that larger sections of the text could be accommodated in public display or deemed perfect to the last character. If we adopt this view, it is cogent to view the Lüshi chunqiu compendium as almost entirely the later product of dynastic boshi; its encyclopedic digest of pre-Qin learning, collected in thematically organized groups of succinct chapters, would fit their mandate to “comprehend the past and present.”62 Taken together, this evidence suggests that during the Qin, senior Ru masters were among those elevated to scholarly positions as encyclopedists, charged with assembling and editing literary works that would, far from destroying the learning of the classical era, preserve it in forms useful to the Qin court, which would alone have access to it. Viewing the Qin dynasty and the role of its Ru masters in this light, it is possible to see how a limited number of Ru may have gained access to a substantial range of source texts containing the sententious wisdom of Confucius and his disciples, whether attributed to them directly, as in the *Junzi wei li and *Dizi wen texts, or in an unattributed Confucian context, as in the case of *Yucong 語叢 3 (Thicket of Sayings, 3). However, on the model I am proposing here, because the state-sponsored goal of Qin boshi projects was primarily pansectarian and curatorial, and not intended to nurture the growth of sectarian schools, it is unlikely that the Lunyu itself would have been an official Qin product. Moreover, judging from the highly organized editorial structure of the Lüshi chunqiu, the disordered editorial form of the Lunyu does not resemble the type of text that the Qin court would have authorized.

62

There is actually no reason to mechanically separate the components of the Lüshi chunqiu into pre-imperial and imperial eras. The process whereby a yueling-based prototext, complete by 239 BCE, grew into the enormous received compendium could have been ongoing over a thirty-year period. There is no report that Lü Buwei’s stable of scholar-retainers left Qin after Lü’s disgrace and his death in 236 BCE, and it is possible that the Qin state continued to support them and may even have appointed some of these or other masters as boshi prior to the establishment of the imperial dynasty. Royal patronage of scholarly masters was a strategy of state legitimation from the time of Wei Wenhou 魏文 侯 in the mid-fifth century BCE and was a major reason the Qi kings sustained the academy at Jixia 稷下. The Qin imperial patronage of boshi described here was very likely an extension of Qin state practices during the prior era of conquest.

The Lunyu As An Accretion Text



61

The Ru Underground of the Early Han and the Canonization of Confucius’s Wisdom

While the Qin appears to have been a period when major Ru masters received state patronage and participated in state-sponsored bibliographic and editorial work, with the end of the Qin, the boshi corps lost its standing and, with it, their sanctioned possession of texts. The book ban of 213 BCE remained in place after the dynasty’s dissolution in 208 BCE, and it was not abrogated until 191 BCE, four years after the end of Liu Bang’s 劉邦 reign as the founding ruler of the Han dynasty. Apart from those items the former boshi retained, under ban and in hiding, the literary remains of the past were largely extinguished, not because of the Qin prohibition but likely because of a bibliocaust that resulted from Xiang Yu’s 項羽 invasion of the Wei River valley in 208 BCE, when his troops burned to the ground the Qin palace compounds, where the dynasty’s archive of banned texts would have been located.63 Until the end of the Qin, the leading Ru and their followers would not have experienced a sharp separation from the writings of the past, but the context of the early Han would have been well suited to the fourth stage of the Lunyu’s accretion process, involving heightened reverence by Ru for surviving Confucian texts and a growing propensity to treat them as canonical. The early Han proscription of unapproved texts would have had a particular impact on Confucianism because of Liu Bang’s recorded distaste for Ru. Liu Bang’s earliest Ru follower, the eccentric Li Yiji 酈食其, was, according to the Shiji, warned about this by being told that Liu Bang’s habit when confronted with Ru scholars was to grab their hats and urinate into them.64 The earliest Ru known to have worked his way into Liu Bang’s graces after the conquest, the former Qin boshi Shusun Tong 叔孫通, carefully discarded his distinctive Ru apparel in favor of a Chu 楚–style work uniform when appearing at court.65 Liu Bang’s attitude toward Ru may have been a response to the conduct of Ru during the Qin-Han interregnum, when Confucians flocked to banners other than his. At the time of Chen She’s 陳涉 initial rebellion against the Qin, the scion of Confucius’s own clan, Kong Jia 孔甲, led Ru from the state of Lu down to Chu to throw in their lot with the rebels.66 After the Xiang clan had taken control of 63 64 65 66

Shiji 7.315. Shiji 97.2692. Shiji 99.2721. A second Ru, Lu Jia 陸賈, was admitted to Liu Bang’s good graces as a consequence of diplomatic services rendered at the Han’s southern borders; his Shiji biography describes Liu Bang’s distaste for Ru and the classical canon (97.2699). The Shiji tells us he became a boshi for Chen She (which suggests Chen’s own imperial ambitions; Shiji 121.3116). The affinity of Ru for Chu was natural: their home region, Lu,

62

Eno

the rebellion, Xiang Yu arranged for the puppet king of Chu to appoint him as the Duke of Lu. Ultimately, the Ru clustered there, joining the forces of Lu as the last holdouts against the conquering armies of Liu Bang.67 This background helps account for the hostility of the early Han court toward the Ru, but its implication for the history of Confucian texts—particularly in relation to the process of canonization—is probably best conveyed by the experience of the Qin boshi Fu Sheng, from whom the received Shangshu text is derived. Fu had preserved his copy of the Shangshu by hiding it in a wall; he recovered it only after so many years had passed that the greater part of the text had been destroyed. He taught the chapters that remained privately to Ru in the Shandong region for years, until word of his knowledge reached the Han court in the time of Wendi 文帝, when the emperor was taking early steps to loosen the dynasty’s exclusion of Ru.68 As a boshi Fu Sheng would not have needed to hide his Shangshu text during the Qin period: he was exempt from the book ban. However, for seventeen years after the fall of the Qin, his private possession of the text would have been a legal violation, which would account for the extended period in which he kept it hidden. And although Liu Bang’s son lifted the ban in 191 BCE, after his death three years later, with power consolidated in the hands of Empress Lü 呂太后 until 180 BCE, Confucians would likely have remained a suspect group, still distrusted for their resistance to the Han founding and standing in stark opposition to the dominant ideologies of the state, Legalism and HuangLao 黃老. It is under these conditions that some Confucian texts retained by former Qin boshi like Fu Sheng and the students of the following generation who received them in inheritance would have come to be treated as sacred remnants of an authoritative past, to be edited only under conservative principles of interpolation and appendage, rather than by reediting and rationalizing the whole as components were added. Two features of this historical model of the fourth stage of the Lunyu accretion process make it attractive. First, while it is not hard to find within the QinHan era a period of Ru retrenchment that involved a sense of disjuncture from the past and a magnified valuation of surviving texts, under the common

67

68

had been under Chu control for several decades prior to the Qin conquest, and Chu patronage of Ru was reflected in its elevation of the last great pre-Qin Ru, Xunzi 荀子, to official position (Eno 1990: 134). Shiji 7.337–338. Shusun Tong, the former Qin boshi who ultimately defected to Liu Bang, was distinguished from fellow Ru only by the timing of his jump from the sinking Xiang family ship: after his flight from Qin he first served Xiang Liang 項梁 and Xiang Yu before concluding that it would be more prudent to discard his Confucian insignia and serve the Han house on the eve of its victory. Shiji 121.3124–3125.

The Lunyu As An Accretion Text

63

assumption that with the rise of the Qin Confucianism was suppressed and its texts were destroyed it is more difficult to identify a period when collection of disparate texts, a necessary stage of the Lunyu accretion process, would have occurred. Second, the model explains the late emergence of any reference to the title “Lunyu,” as the text assembled during the Qin-Han period, regardless of the dates of its component sources, would have been a new construct. However, a case can also be made for an earlier collection process, as might be required if the date of the “Fang ji,” with its reference to the Lunyu, were found to predate the Qin. For example, although the masters assembled at Jixia 稷下 in Qi during the late fourth and third centuries BCE are said not to have had official duties, most scholars agree that a group of them was responsible for the generation or collection of texts that grew into the Guanzi 管子 compendium.69 Ru such as Xunzi 荀子 were masters at Jixia, and if text collocation was part of the Jixia tradition, he or other Ru masters could have been part of that effort.70 Moreover, according to the Shiji, the office of boshi was first instituted in the state of Lu, perhaps during the time of Duke Mu 穆公 (r. 415–383 BCE), when Confucian masters such as Zisi and Zengzi’s son Zeng Shen 曾申 were receiving some form of court patronage.71 Those who consider it likely that the collocation of Lunyu source texts began early may take the account of this first formalization of scholarly bureaucracy as an anchor point. It is also valid to note that if the “Tianxia” 天下 chapter of the Zhuangzi 莊子, which claims that the scholar Hui Shi 惠施 possessed five cartloads of texts, was plausible to early readers, it signifies the possibility of private text collection on a scale that might lead to redacted compilations like the Lunyu.72 While I regard these earlier periods as less likely for the major curatorial work that produced the Lunyu, I do think it reasonable to see the largest of the Lunyu source texts, the Shanglun Core Source, as edited significantly earlier than the Qin, perhaps at Jixia or at the courts of the great third-century BCE warlords famous for patronizing vast scholar entourages prior to Lü Buwei. We can, at least, say that it is unlikely to have been a post-Qin compendium, since its topical reorganization of the materials of which it is comprised suggests that the 69 70 71

72

See Rickett 1985: 15–19, and the nuanced analysis in Kanaya 1987: 320–324. The Xunzi includes chapters that bear resemblance to the Lunyu, for example, “Faxing” 法 行 and portions of “Zidao” 子道; scholars generally view these as late chapters, but they could be based on school materials assembled by Xunzi’s Jixia circle. Shiji 119.3101; see Qian (1956: 158) for the ascription to Duke Mu’s reign. The Shiji account is vague and linked to anecdotal material in a way that may suggest less historical reliability than Qian Mu attributed to it. The Hanshu also speaks of boshi established in the state of Wei 魏 (51.2327). However, five cartloads of bamboo texts might amount to no more than a single bookcase of modern books. State-sponsored archives may have been a necessity for broad text collection.

64

Eno

Shanglun Core Source’s own sources were not yet regarded as canonical inheritances from a bygone world.

Closing the Canon

As for the final stage of Lunyu redaction, the reported multiplicity of Lunyu editions during the Han that so fascinated Takeuchi Yoshio indicates that certain aspects of the text continued to be contested for some time. However, apart from Wang Chong’s account of texts with widely disparate numbers of juan 卷, which Takeuchi took to be source texts that continued to be circulated independently, accounts seem to suggest that there were basically three Han variants (the Lu, Qi, and Gu editions), which differed in only minor respects. The partial Lunyu manuscript recovered from a grave in Dingzhou 定州, Hebei, sealed in 55 BCE, is not far removed from the received text, with sections of all twenty books among the recovered strips. Although the total number of strips accounts for only about 45 percent of the present text, and the original order of passages in the bamboo manuscript cannot be determined, it suggests that the received text was essentially fixed by the mid-first century BCE.73 Half a century earlier, some circulating versions may not yet have incorporated all twenty current books. The Shiji cites over 120 passages found in the Lunyu, most virtually verbatim, but no passages are cited from books 8, 16, or 20. This may be because the text Sima Qian relied on did not include them— books 16 and 20 are often regarded as among the latest to be added to the Lun­ yu, for reasons of form and content.74 However, it may also simply be a matter of chance. The Shiji most frequently cites passages that characterize Confucius or the disciples or that record conversations between them, and the three books include only a small handful of such passages.

Conclusion

When addressing issues of interpretation and dating for the Lunyu, whether of the entire text or component elements of it, an accretion approach is neces73 74

Hebei sheng wenwu yanjiusuo 1997: preface. Book 8 features a string of sayings by the disciple Zeng Shen, and it may be that components of the Lunyu celebrating Zeng Shen were absent from Sima Qian’s copy of the text. Despite his apparent influence among prominent Ru groups, Zeng Shen is almost completely ignored by the Shiji, and his words are nowhere quoted. (These points were suggested to me by Michael Hunter.)

The Lunyu As An Accretion Text

65

sary to accommodate the overriding problem of editorial disorder. If issues of “authenticity” are set aside and questions of the text’s connection to the historical Confucius are acknowledged to be insoluble on present evidence, an accretion approach can accommodate ranges of dating solutions that fit available evidence while addressing the critical issue of textual disorder in the Lunyu. I have tried here to clarify the nature of textual accretion, illustrate varieties of accretion theories of the Lunyu, and demonstrate the flexibility of the accretion approach by providing alternative historical applications of a five-stage template that I believe is necessary for any successful model of the Lunyu’s development. I hope that my proposal of how the Qin-Han transition was critical to the Lunyu’s development will be persuasive and fit future new evidence. But if it does not prevail in its particulars, an analytic structure involving multiple stages of redaction accompanied by a growing attitude of textual canonicity will still be needed to account for the Lunyu’s mix of order and disorder.

Bibliography

Bin Dongchoel. 2014. “Calligraphy and Scribal Tradition in Early China.” PhD diss., Indiana University. Bodde, Derk. 1987. “The State and Empire of the Ch’in.” In The Cambridge History of China, vol. 1, ed. Denis Twitchett and Michael Loewe, 20–102. Cambridge: Cambridge University Press. Brooks, E. Bruce, and A. Taeko Brooks. 1998. The Original Analects: Sayings of Confucius and His Successors. New York: Columbia University Press. Campbell, Anthony, and Mark O’Brien. 2000. Unfolding the Deuteronomistic History. Minneapolis, MN: Fortress Press. Cook, Scott. 2012. The Bamboo Texts of Guodian: A Study and Complete Translation. Ithaca, NY: Cornell University East Asia Program. Eno, Robert. 1990. The Confucian Creation of Heaven. Albany: State University of New York Press. Eno, Robert. 2010. “Sources for the Analects 8 Layers.” Warring States Papers 1:93–99. Hanshu 漢書. 1962. Beijing: Zhonghua shuju. Hebei sheng wenwu yanjiusuo 河北省文物研究所. 1997. Dingzhou Han mu zhujian Lunyu 定州漢墓竹簡《論語》. Beijing: Wenwu chubanshe. Henderson, John B. 1991. Scripture, Canon, and Commentary: A Comparison of Confucian and Western Exegesis. Princeton, NJ: Princeton University Press. Huang Ren’er 黃人二. 2011. “Shangbo cangjian (wu) Junzi wei li yu Dizi wen shishi” 上博 藏簡 (五) 《君子為禮》與《弟子問》試釋. Zhongguo guojia bowuguan guankan中 國國家博物館館刊 2011.6:65–80.

66

Eno

Kanaya Osamu 金谷治. 1960. Shin-Kan shisōshi kenkyū 秦漢思想史研究. Tokyo: Nihon gakujutsu shinkō. Kanaya Osamu 金谷治. 1987. Kanshi no kenkyū 管子の研究. Tokyo: Iwanami shoten. Kern, Martin. 2000. The Stele Inscriptions of Ch’in Shih-Huang: Text and Ritual in Early Chinese Imperial Representation. New Haven, CT: American Oriental Society. Kimura Eiichi 木村英一. 1971. Kōshi to Rongo 孔子と論語. Tokyo: Sōbunsha. Knoblock, John, and Jeffrey Riegel. 2000. The Annals of Lü Buwei. Stanford, CA: Stanford University Press. Lau, D. C. 1991. “A Study of Some Textual Problems in the Lü-shih ch’un-ch’iu and Their Bearing on Its Composition.” Zhongguo wenzhe yanjiu jikan 中國文哲研究集刊 1:45–87. Lin Baoquan 林保全. 2008. “‘Fang ji,’ ‘Qishi’erzi jie’ chengyin Lunyu shuti zaiyi” 〈坊 記〉、〈七十二子解〉稱引《論語》書題再議. Zhongguo xueshu niankan 中國學 術年刊 (Guoli Taiwan Shifan daxue guowenxi 國立台灣師範大學國文系) 30:31–58. Matsumoto Masaaki 松本雅明. 1966. Shunjū sengoku ni okeru Shōsho no tenkai 春秋戰 國における尚書の展開. Tokyo: Kazama shobō. Nicholson, Ernest. 1998. The Pentateuch in the Twentieth Century. Oxford: Clarendon Press. Qian Mu 錢穆. 1956. Xian-Qin zhuzi xinian 先秦諸子繫年. 2nd ed. Hong Kong: Xianggang daxue chubanshe. Rickett, W. Allyn. 1985. Guanzi: Political, Economic, and Philosophical Essays from Early China. Princeton, NJ: Princeton University Press. Sanders, E. P., and Margaret Davies. 1989. Studying the Synoptic Gospels. Philadelphia: Trinity Press International. Shiji 史記. 1959. Beijing: Zhonghua shuju. Slingerland, Edward. 2000. “Why Philosophy Is Not ‘Extra’ in Understanding the Analects.” Philosophy East and West 50.1:137–141. Takeuchi Yoshio 武內義雄. 1939. Rongo no kenkyū 論語之研究. Tokyo: Iwanami shoten. Tsuda Sōkichi 津田左右吉. 1946. Rongo to Kōshi no shisō 論語と孔子の思想. Tokyo: Iwanami shoten. Waley, Arthur. (1938) 1989. The Analects of Confucius. New York: Vintage Books. Yang Bojun 楊伯峻. 1984. Lunyu yizhu 論語譯注. Hong Kong: Zhonghua shuju. Yi Ruobo 伊若泊 [Robert Eno]. 2011. “Shangbo wu suojian Zhongni dizi Zigong de yanyu yu zaoqi Ruxue shi” 《上博。五》所見仲尼弟子子貢的言語與早期儒學史. In Zhongguo jianboxue guoji luntan lunwenji 2007 中國簡帛學 國際論壇論文集, 523–542. Taipei: Guoli Taiwan daxue zhongguo wenxuexi. Zhang Xincheng 張心澂. 1954. Weishu tongkao 偽書通考. Shanghai: Shangwu yin­ shuguan.

The Lunyu As A Western Han Text

67

Chapter 3

The Lunyu as a Western Han Text Michael Hunter This chapter summarizes the argument for reading the Lunyu 論語 (Analects) as a product of the early Western Han period (202 BCE–9 CE). As such, it is deeply indebted to a number of scholars who have critiqued the assumption that the Lunyu or some core section therein predates the Han period. These include Zhao Zhenxin and Zhu Weizheng in China, Tsuda Sōkichi and Kaneto Mamoru in Japan, and especially John Makeham, Mark Csikszentmihalyi, Christiane Haupt, and Oliver Weingarten.1 The version of the “Lunyu as a Western Han text” thesis proposed here is defended at greater length in my fulllength monograph on the subject.2 There are precious few certainties when it comes to the dating of early Chinese texts, so it is important to ask at the outset what degree of certainty we might reasonably hope to achieve when assessing the chronology of the Lunyu (or any other early source, for that matter). The arguments presented below do not prove the Lunyu’s Western Han origins beyond the shadow of a doubt, nor do they completely disprove the Lunyu’s traditional chronology. Whether or not the Lunyu originated in the Western Han period, as I contend, I hope to demonstrate at the very least that efforts to date the Lunyu prior to the Han do a poor job of explaining the available evidence, such as it is.

The Title “Lunyu” 論語

The crudest index of the Lunyu’s existence in a given period is to survey mentions of its title, of which there are many. Emperors listed the Lunyu as one of the texts they had studied,3 bibliographers ranked it alongside the most canonical traditions of the day,4 and numerous authors cited it as a source. However, explicit references to the Lunyu do not appear in any extant text 1 See Zhao 1961; Zhu 1986, 1987; Tsuda 1946; Kaneto 1970, 1971, 1972, 1973, 1974, 1975, 1976, 1977, 1978, 1979, 1980; Makeham 1996; Csikszentmihalyi 2001, 2002; Haupt 2006; Weingarten 2010. 2 Hunter 2017. 3 See Emperor Zhao’s 昭帝 82 BCE edict at Hanshu 7.223, discussed below. 4 See, e.g., the “Liu yi” 六藝 (Six Arts) division of the “Yiwen zhi” 藝文志 bibliography (Hanshu 30.1701–1723) and Wang Chong’s 王充 (27–100) discussion of the Lunyu at Lunheng 81, “Zheng shuo 正說” (Sibu congkan [hereafter SBCK] ed., 28.6b–7a), both of which rank the Lunyu im-

© Koninklijke Brill NV, Leiden, 2018 | doi:10.1163/9789004382947_005

68

Hunter

reliably dated to the pre-Han period. On current evidence, its title appears only sporadically in Western Han sources but much more frequently in the Eastern Han period (table 3.1). There are several candidates for the earliest mention of a Lunyu text in the extant record; of these, only one or two can be dated with any degree of accuracy. The single citation in the “Fang ji” 坊記 (Embankment Record) chapter of the Liji 禮記 (Records of Ritual) not only appears in a text whose own dating is uncertain but is the only such citation in the whole of the Liji.5 Another citation appears alongside more than a dozen verbatim quotations in a set of memorials attributed to Dong Zhongshu 董仲舒 (ca. 179–ca. 104 BCE) in the early part of Emperor Wu’s 武帝 reign. However, references in these documents to Kangju 康居 (Bactria) and Yelang 夜郎 (modern-day Guizhou 貴州), neither of which was incorporated into the Han empire until the Eastern Han period, caution against taking these documents as authentic representations of the early to mid–Western Han context.6 Another three citations appear in chapters 2, 5, and 6 of the Hanshi waizhuan 韓詩外傳 (Outer Tradition of the Han Poetry),7 a text traditionally dated to ca. 150 BCE on the basis of its attribution to Han Ying 韓嬰, a court erudite (boshi 博士) whose career spanned the reigns of Emperors Wen 文帝 (r. 180–157 BCE), Jing 景帝 (r. 156–141 BCE), and Wu 武 帝 (r. 141–87 BCE).8 Given that Han Ying also inspired an independent tradition

5

6

7 8

mediately after the classics. The term “Seven Classics” (qi jing 七經), probably the Six Classics plus the Lunyu, first appears in Eastern Han sources (see, e.g., Hou Hanshu 35.1196). For this argument, see Makeham 1996: 10–11; see also Makeham’s discussion (in chapter 1 of the present volume) of efforts to date the Lunyu on the basis of the “Fang ji.” The line in question—“Not revising the father’s way for three years; this can be called ‘filiality’” (三年無 改於父之道,可謂孝矣)—appears at Lunyu 1/11 and Lunyu 4/20. The quotation in the “Fang ji” (SBCK ed., 15.14b–15a) follows the line “The Master said, ‘A noble man eases his parents’ transgressions and respects their fine points’” (子云:君子弛其親之過,而敬其美). Loewe 2011. However, Loewe (2011: 336) ultimately concludes that “[t]he three essays that Dong Zhongshu presented in response to imperial rescripts are probably to be taken as some of the most authentic expressions of his views,” despite the “signs of editorial work that was presumably accomplished by Ban Gu.” Whether or not the memorials authentically express the historical Dong Zhongshu’s ideas, Loewe’s analysis leads me to suspect that their language and citations reflect an Eastern Han milieu. For the Lunyu citation, see Hanshu 56.2514. Not only does this particular citation break the pattern of Kongzi yue 孔子曰 quotations across the memorials, but its quotation marker—“Your subject has heard that the Lunyu says” (臣聞論 語曰)—has no parallel in the extant literature that I am aware of. SBCK ed., 2.13a, 5.18a, 6.4a. For a summary of the available biographical information on Han Ying, see Loewe 2000: 151. On the Hanshi waizhuan in particular, see Hightower 1993: 125–126. For Makeham (1996: 12–13), the Lunyu citations in the Hanshi waizhuan indicate that the Lunyu existed as a book by ca. 140 BCE. In my view, however, the evidence from the Hanshi waizhuan as a whole is decidedly mixed, with the bulk of its Kongzi sayings and stories deriving from non-Lunyu sources. For a discussion, see Hunter 2017: chap. 3.

69

The Lunyu As A Western Han Text Table 3.1

Lunyu mentions in early sources

Hanshi waizhuan

3

Shiji

1

Yantie lun

2

Liji

1

Xinxu

4

Xinlun

1

Hanshu

36

Lunheng

29

Baihu tongyi

47

Shuowen jiezi

35

Zheng Xuan’s  Li 禮 and Shi 詩 commentaries

30

Wang Yi’s Chuci commentary

11

Xu Shen / Gao You’s Huainanzi commentary

14

Gao You’s  Lüshi chunqiu com- 22 mentary Fengsu tongyi

22

Hou Hanshu

16

Total

274

of Shi 詩 (Odes) scholarship that was carried on by his students and descendants, the possibility that these citations were added after Han Ying’s death cannot be discounted.9 Other early mentions include a concluding comment to chapter 67 of Sima Qian’s 司馬遷 (d. ca. 86 BCE) Shiji 史記, the “Zhongni dizi liezhuan” 仲尼弟子 列傳 (Arrayed Traditions of Zhongni’s Disciples),10 and an edict issued in the 9 10

Hanshu 88.3613–3614. Shiji 67.2226. The second occurrence of “Lunyu” in the Shiji appears within material added after Sima Qian’s death. See Shiji 96.2688.

70

Hunter

name of the underaged Emperor Zhao 昭帝 (r. 87–74) in 82 BCE.11 The Hanshu 漢書 also records that Liu Qu 劉去, a great-grandson of Emperor Jing who was appointed king of Guangchuan 廣川 in 91 BCE, had received training in the Lunyu (together with the Yi 易 [Changes] and Xiaojing 孝經 [Classic of Filial Piety]).12 Writing in the Eastern Han period, Wang Chong 王充 (27–100 CE) credited a certain Fu Qing of Lu 魯人扶卿, who served as inspector (cishi 刺史) of Jingzhou 荊州 during the latter part of Emperor Wu’s reign, with having named the text.13 Although little is known of this figure, provincial inspectors were first appointed in 106 BCE,14 thus suggesting the latter part of Emperor Wu’s reign as a rough terminus ante quem for the emergence of a named Lunyu text. Sima Qian and the author(s) of the Hanshi waizhuan aside, many of the earliest recorded uses and mentions of the Lunyu are in the context of its use by the imperial Liu 劉 clan for the education and nomination of its scions. In addition to Liu Qu’s nomination and Emperor Zhao’s edict, the Lunyu was cited as a qualification of Emperor Xuan 宣帝 (r. 74–49 BCE) upon his nomination,15 it was invoked by Shi Dan 史丹 to press Emperor Yuan 元帝 (r. 48–33 BCE) to keep Liu Ao劉驁, the future Emperor Cheng 成帝 (r. 33–7 BCE), as his crown prince,16 and it featured prominently in Wang Mang’s 王莽 proclamations before and after his establishment of the Xin 新 dynasty (9–23 CE).17 The circulation of the Lunyu within the Liu clan has also been confirmed by the discovery of Lunyu manuscripts in the tombs of Liu Xiu 劉脩 and Liu He 劉賀 from the late Western Han (see Paul van Els’s contribution in this volume). And as Mark Csikszentmihalyi has observed, the list of Lunyu experts in the “Yiwen zhi” 藝 11 12

13

14 15 16 17

Hanshu 7.223. The circulation of a “Lunyu” during or before the reign of Emperor Wu is attested in two additional sources. The first is Zhao Qi’s 趙岐 (d. 201 CE) preface to his Mengzi 孟子 commentary, in which he states that Emperor Wen had established erudites for the Lunyu (as well as for the Mengzi and Xiaojing). See the preface to Mengzi (SBCK ed., 3a). The second is the legend of the “ancient-script” (guwen 古文) Lunyu dealt with below. However, Zhao Qi’s claim concerns an event reported to have taken place more than 250 years earlier and is completely unsubstantiated by earlier sources. Moreover, Zhao Qi’s reason for citing this precedent is only to criticize the lack of a Mengzi or Lunyu boshi in his own era. Lunheng 81, “Zheng shuo” 正說 (SBCK ed., 28.6b–7a). Whereas the “Yiwen zhi” refers to this figure ambiguously as “Lu Fuqing,” Wang Chong identifies him as “Fu Qing of Lu.” See Hanshu 30.1716–1717 and Lunheng 81 (SBCK ed., 28.6b–7a). The Jingdian shiwen 經典釋文 (SBCK 1/31b) quotes Zheng Xuan 鄭玄 to the effect that Fu Qing was also known as Fu Xian 扶先 or Fu xiansheng 扶先生 (Master Fu). On this point, see Makeham 1996: 11n42. Bielenstein 1980: 10. Hanshu 8.238. Hanshu 82.3376. See, e.g., Hanshu 99A.4053–4063, 99A.4089.

The Lunyu As A Western Han Text

71

文志 (Treatise on Arts and Letters) bibliography includes a number of figures who served as imperial tutors in the same period.18 This was the context within which the earliest extant sources of Lunyu bibliography were composed.19



Han 漢 Bibliography and the Limitations Thereof

It bears repeating that the traditional view of the Lunyu as a record of Kongzi’s teachings compiled by his closest students cannot be found in the Lunyu itself, nor does it appear in any pre- or early Han text. The earliest known sources of the traditional view are Liu Xiang’s 劉向 (79–8 BCE) catalog of the imperial library, the Bie lu 別錄 (Separate Listings), followed by the thirtieth chapter of the Hanshu, the “Yiwen zhi” bibliography.20 On current evidence, the Lunyu’s traditional dating was first articulated decades after the text’s adoption by the imperial clan as a qualification for rulership. Against this backdrop, it should come as no surprise that bibliographers from the late Western Han onward advanced a theory of the text that confirmed the Lunyu’s importance. To suggest otherwise would have been to undermine the legitimacy of the Han (or Xin) imperium itself. That aside, early bibliographic sources also highlight Emperor Wu’s reign as a pivotal moment in the history of the Lunyu. Wang Chong informs us that the Lunyu was “lost when the Han arose” (漢興失亡) only to be rediscovered in the walls of Kongzi’s ancestral home during the reign of Emperor Wu.21 The “Yiwen zhi” confirms the story of the “ancient-script” Lunyu’s discovery and goes on to explain—rather incredulously—that it occurred after Liu Yu 劉餘, or King Gong of Lu 魯恭王 (r. 154–128), set out to demolish Kongzi’s house in order to clear space for his own palace. Liu Yu supposedly called off the demolition after entering Kongzi’s house and “hearing tones of metal and stone and silk and bamboo” (聞金石絲竹之音)—that is, the sound of ancient music.22 However, not only does the roughly contemporaneous Shiji make no mention of an ancient-script Lunyu, but Liu Xiang’s son Liu Xin 劉歆 (46 BCE–23 CE), in a memorial dated to ca. 25 BCE, also neglected to mention the Lunyu in his seem-

18 19 20 21 22

Csikszentmihalyi 2002: 146. See also Martin Kern’s contribution (chap. 10) in the present volume. Lunyu jijie preface (SBCK ed., 1.1a); Hanshu 30.1716. The “Yiwen zhi” was based on the Qilüe 七略 (Seven Surveys) by Liu Xin. Lunheng 81 (SBCK ed., 28.6b–7a). Hanshu 53.2414. See also the pseudo–Kong Anguo 孔安國 preface to the Shangshu (SBCK preface, 3a).

72

Hunter

ingly exhaustive list of texts discovered in the walls of Kongzi’s house.23 Another inconsistency is that the “Yiwen zhi” dates the discovery of the text to the “latter part of Emperor Wu’s reign” (Wudi mo 武帝末) even though the Hanshu biography of Liu Yu records that he was appointed king of Lu in 154 BCE during the reign of Emperor Jing 景帝 (r. 156–141) and died twenty-six years later in 128 BCE, a mere thirteen years into Emperor Wu’s fifty-four year reign.24 This jumble of detail does not convince as a reliable account of the Lunyu’s history. However, Han era Lunyu bibliography makes a little more sense if understood as a retrospective attempt to explain certain uncomfortable truths about the Lunyu’s origins. If the Lunyu first came into existence during the reign of Emperor Wu or thereabouts, thus leaving no mark on earlier textual traditions (see below), then Han bibliographers who were already invested in the Lunyu as the preeminent source of Kongzi sayings would have had a strong incentive to invent a plausible backstory. John Makeham’s observation that the list of Lunyu experts in the “Yiwen zhi” extends no earlier than Emperor Wu’s reign is also relevant in this context.25

The Evidence from Lunyu Intertextuality: Kongzi Quotations

Needless to say, the coining of the term “Lunyu” need not have coincided with the creation of an as-yet-untitled Lunyu text or proto-Lunyu collection.26 In the absence of any pre-Han manuscript witnesses or some other direct evidence of a pre-Han Lunyu, the only way to detect such a text is indirectly, by looking for signs of its influence on other extant corpora. For instance, even if Eastern Han texts had not preserved a single mention of its title, we might still infer the 23

24

25 26

Hanshu 36.1969. In this memorial Liu Xin urges Emperor Cheng 成帝 (33–7 BCE) to establish an official chair for the study of the Zuozhuan 左傳. The only guwen texts mentioned are “remnant ritual texts” (yi li 逸禮) in thirty-nine sections and Documents (Shu 書) texts in sixteen sections. For the argument that the legend of the guwen Lunyu is a later fabrication, see Tsuda 1946: 97–101; Kaneto 1973. See esp. Nylan 1994 for a skeptical view of guwen legends more generally. Hanshu 30.1706. For Liu Yu’s Hanshu biography, see Hanshu 53.2414. In chapter 81 of the Lunheng, the section on the Lunyu dates the incident to Emperor Wu’s reign, but the section on the Shangshu dates it to Emperor Jing’s. On this point, see also Makeham 1996: 13–14. Makeham 1996: 19–20. According to Anne Cheng (1993: 315), “[o]riginally the work had been referred to as the Kongzi 孔子, in the same way as writings of other masters of the Warring States period.” However, I have found no evidence that any early author ever used “Kongzi” to refer to a text rather than to the Master himself.

The Lunyu As A Western Han Text

73

Lunyu’s existence from the observation that roughly 70 percent of the hundreds of Kongzi yue quotations from the period exhibit word-for-word parallels with the received Lunyu (see table 3.2).27 Eastern Han authors did not just pay lip service to the Lunyu as one of the most important texts in the tradition.28 They also used the text as if it was the preeminent source of Kongzi sayings. Contrast Kongzi quotations in the Eastern Han with those of the pre-Han era (table 3.3). Although a wealth of texts in both periods quote Kongzi and/or the anonymous “master” (zi 子),29 only 9 percent of Kongzi sayings in pre-Han sources exhibit Lunyu parallels. If the Warring States and Eastern Han periods represent two extremes, then the Western Han period falls somewhere in between. Some texts (e.g., the Shiji) look more like Eastern Han sources insofar as they include a relatively high percentage of Kongzi sayings with Lunyu parallels. On the other hand, texts like the Huainanzi 淮南子 and Liji do not appear to privilege Lunyu Kongzi sayings over non-Lunyu sayings. The Shiji/Huainanzi comparison is especially striking. Although both texts are large compendia (just under 600,000 characters for the Shiji and over 130,000 characters for the Huainanzi) compiled (at least in part) during the reign of Emperor Wu (141–87 BCE), the Shiji includes a much higher percentage of Lunyu sayings among its 240 Kongzi quotations (62 vs. 14 percent). Thus, we might infer that the Huainanzi/Shiji period marks a shift in the influence of the Lunyu as a source of Kongzi material, with post-Shiji texts drawing heavily from the Lunyu for their Kongzi quotations and pre-Huainanzi texts drawing their Kongzi material from elsewhere. There are also significant differences between how authors before and after Emperor Wu’s reign handled Kongzi quotations with Lunyu parallels. Post–Emperor Wu quotations not only exhibit very few variants but also are routinely invoked as authoritative proof texts or subjected to commentary. In some texts, the use of Lunyu sayings to critique or comment on others also indicates that post–Emperor Wu authors conceived of the Lunyu as a single, internally consistent source of Kongzi’s teachings.30 In contrast, even when pre–Emperor Wu sources do contain Kongzi sayings with Lunyu parallels, those parallels tend to 27 28 29 30

In my scheme, a “Lunyu parallel” is a passage that shares at least three characters in common with the Lunyu in a similar syntactic arrangement, allowing for semantic, phonological, and graphical variants. See, e.g., Zhao Qi’s description of the Lunyu as “the linchpin of the Five Classics and the mouthpiece of the Six Arts” (五經之錧鎋,六藝之喉衿也) in the preface to his Mengzi commentary (SBCK Mengzi preface, 3a–b). My numbers include explicit Kongzi quotations together with zi yue 子曰 (the Master said) quotations despite the indeterminacy of the latter in many texts. See, e.g., the discussion of Lunyu 6/2 at Shuiyuan 19 (19.20a). See also Weingarten 2010: 151–154 for a translation and analysis of this passage’s “commentarial structures.”

74

Hunter

Table 3.2 Kongzi attributions in texts after ca. 100 BCE

Attributions with Lunyu parallels / total Kongzi attributions Shiji (ca. 100)

148 / 240

62%

Yantie lun

35/43

81%

Chunqiu fanlu

23/30

77%

Liji

29 / 320

 9%

Da Dai Liji

8 / 141

 6%

Shuiyuan (late W. Han)

32 / 220

17%

Xinxu

13/22

59%

Lienü zhuan

0/7

 0%

Xinlun

4/5

80%

↓ Eastern Han ↓ Hanshu

146 / 174

84%

Lunheng

72 / 115

63%

Baihu tongyi

5 / 20

25%

Shuowen jiezi

2 / 13

15%

Qianfu lun

27 / 43

63%

Zheng Xuan’s Li 禮 and Shi 詩 commentaries

25/32

78%

He Xiu’s Gongyang commentary

27/36

75%

Xu Shen / Gao You’s Huainanzi commentary

4/5

80%

Gao You’s  Lüshi chunqiu commentary

7/8

87%

Fengsu tongyi

9 / 22

41%

Zhonglun

23 / 37

62%

Hou Hanshu

78 / 115

68%

Total

717/1,648

44%

The Lunyu As A Western Han Text

75

Table 3.3 Kongzi attributions in texts prior to ca. 100 BCE

Attributions with Lunyu parallels / total Kongzi attributions Zuozhuan

3 / 42

7%

Shanghai Museum manuscripts

~11 / ~100

Guodian (ca. 300 BCE)

3 / 23

Mengzi

9 / 31

Xunzi

3 / 38

Mozi

2/4

Han Feizi

1 / 33

8%

Guoyu

0 / 11

0%

Lüshi chunqiu (ca. 239)

4 / 34

Yanzi chunqiu

1 / 15

7%

Zhuangzi

5 / 93

5%

Zhouyi

0 / 30

0%

Gongyang

1 / 8

11% 9% 29% 8% 50%

12%

13%

↓ Western Han ↓ Xiaojing

1 / 15

7%

Mawangdui (168)

0 / 121

0%

Shuanggudui (165)

0 / ~33

0%

Xinshu

0 / 6

0%

Guliang

0 / 7

0%

Xinyu

5 / 9

56%

Shangshu dazhuan

3 / 26

12%

Hanshi waizhuan

14 / 112

11%

Huainanzi (139)

4 / 21

14%

Total

73

/812

9%

76

Hunter

vary significantly from their Lunyu counterparts. If we were to count only verbatim parallels in the pre-Han period, the percentage of Kongzi quotations with Lunyu parallels would fall from 9 percent to 4 percent.31 Moreover, there is hardly any evidence to suggest that pre–Emperor Wu authors treated such sayings as more authoritative than non-Lunyu Kongzi sayings. In the aggregate, then, the early record reveals a marked shift in the authority and influence of the Lunyu as a source of Kongzi quotations during the reign of Emperor Wu or thereabouts. However, a handful of passages in ostensibly earlier sources complicate this picture somewhat. One of these is Mengzi 7B/37, in which Mengzi comments on a variant of Lunyu 5/22 by invoking variants of Lunyu 13/2, 17/13, and 17/18.32 The Shanghai Museum bamboo-manuscript corpus, a looted collection dated to ca. 300 BCE, also contains a number of passages with striking parallels with the Lunyu, including the following dialogue from *Junzi wei li 君子為禮 (The Noble Man’s Practice of Ritual): Yan Yuan was waiting on the Master. The Master said, “Hui! A noble man in his conduct of ritual complies with the humane.” Yan Yuan rose to reply, “I am not intelligent. I am completely unable to make even a little sense of this.” The Master said, “Sit and I will tell you. If saying something is improper, then your mouth will not say it; if seeing something is improper, then your eyes will not see it; if hearing something is improper, then your ears will not hear it; if moving is improper, then your body will not move in that way.” 顏淵侍於夫子。夫子曰:回!君子為禮,以依於仁。顏淵作而答曰: 回不敏,弗能少居。夫子曰:坐吾語女。言之而不義,口勿言也;視 之而不義,目勿視也;聽之而不義,而勿聽也;動而不義,身毋動 焉。33 31

32 33

For example, consider the Lunyu 13/18 parallels at Lüshi chunqiu 11/4.3 (SBCK ed., 11.8a–b), Hanshi waizhuan (SBCK ed., 2.9b), and Baihu tongyi (SBCK ed., 4.12b). In Lunyu 13/18, the Duke of She 葉公 asks Kongzi about a certain Upright Self (Zhi gong 直躬) who had testified against his sheep-rustler father. Whereas the Hanshi waizhuan and Baihu tongyi versions quote Kongzi with close parallels of the line “fathers cover for their sons and sons cover for their fathers; therein lies uprightness” (父為子隱,子為父隱。直在其中矣) from Lunyu 13/18, Kongzi’s criticism of Upright Self in the Lüshi chunqiu is altogether different: “How different is Upright Self’s faithfulness! He made a name for himself for [the price of] his only father” (異哉直躬之為信也,一父而載取名焉). To cast as wide a net as possible, I have counted variants like the Lüshi chunqiu version as a “Lunyu parallel.” SBCK ed., 14.16a–19b. Shanghai bowuguan cang Zhanguo Chu zhushu 2001–2011: 5.254–256, strips 1–3.

The Lunyu As A Western Han Text

77

Yan Yuan asked about humaneness. The Master said, “To conquer oneself and abide by ritual is humane. If in one day one can conquer oneself and abide by ritual, then the whole world will return to humaneness. One becomes humane through oneself, so how could it be through another?” Yan Yuan said, “May I ask about the particulars?” The Master said, “If it is not ritually proper, then do not look at it; if it is not ritually proper, then do not listen to it; if it is not ritually proper, then do not speak of it; if it is not ritually proper, then do not act on it.” Yan Yuan said, “Although I am not intelligent, I beg to put your words into practice.” 顏淵問仁。子曰:克己復禮為仁。 一日克己復禮,天下歸仁焉。為仁 由己,而由人乎哉?顏淵曰:請問其目。 子曰:非禮勿視,非禮勿 聽,非禮勿言,非禮勿動。顏淵曰:回雖不敏,請事斯語矣。

Variants aside, the parallels between the two—the role of Yan Yuan, the question of the relationship between li 禮 (ritual) and ren 仁 (humaneness), the list of ritual prohibitions—are undeniable and present a strong prima facie case for some sort of underlying connection. How we understand the relationship between the Lunyu and passages like these is the critical question. At the very least, it seems clear that Lunyu 12/1 and a smattering of other Lunyu sayings and dialogues had a history that extended back to the fourth century BCE. On the other hand, the relative scarcity of these parallels both in recently discovered manuscripts and in the pre-Han textual record at large, their variability, and the absence of other signs of the Lunyu’s influence severely undermine their identification as Lunyu quotations. Simply put, it would be a mistake to treat the exceptions as anything other than exceptional. Moreover, those who would insist on reading these parallels as evidence of the Lunyu must also deal with a few other competing hypotheses, namely: (1) that the compilers of the Lunyu drew from texts like Mengzi 7B/37 and *Junzi wei li; (2) that the authors of *Junzi wei li and the compilers of the Lunyu drew from a common repertoire of Kongzi texts and traditions; or (3) that in the case of a transmitted text like Mengzi 7B/37, its Lunyu parallels reflect a post-Lunyu milieu, perhaps having been added by an editor like Zhao Qi who was keen to demonstrate the affinities between Mengzi and Kongzi.34

34

This is precisely the tenor of Zhao Qi’s preface to his Mengzi commentary, in which he calls attention to a number of parallels between the Lunyu and Mengzi. For further discussion, see Hunter 2014: 73–74.

78

Hunter

Developments in Kongzi Quotation Practice

The most common entry in the Lunyu is the standalone zi yue 子曰 (the Master said) saying, which accounts for 219 of the 513 (43 percent) entries in the text. Another 110 entries (21 percent) feature Kongzi responding to questions from students or other interlocutors; of these, forty-three follow a simple format in which the interlocutor asks a brief indirect question (e.g., Lunyu 2/6: “Meng Wubo asked about filial piety” [孟武伯問孝]), Kongzi responds, and the episode ends. Little effort is made to represent such exchanges as realistic conversations or scenes of instruction. Generally speaking, this was not how pre-Han authors made use of Kongzi. Pre-Han Kongzi sayings tend to be richly contextualized as comments on anecdotes (e.g., the Kongzi wen zhi yue 孔子聞之曰 [Kongzi heard of it and said] sayings found throughout the Zuozhuan 左傳 and Lüshi chunqiu 呂氏春秋 [Annals of Lü Buwei]), as commentaries (e.g., the Shanghai Museum *Kongzi shilun 孔子詩論 [Kongzi’s Discussion of the Odes] manuscript or the various Zhouyi 周易 [Zhou Changes] commentaries found at Mawangdui 馬王堆), or as elements of longer anecdotes and dialogues (e.g., the Shanghai Museum Zhonggong 仲弓 and *Min zhi fumu 民之父母 [Parent to the People] manuscripts). Prior to the rise of the Lunyu, Kongzi sayings derived their value primarily from the contexts to which they were attached; they were less often valued as independent teachings worthy of study in their own right. (Perhaps the most notable exceptions are the *Junzi wei li and *Dizi wen 弟子問 (The Disciples Ask) manuscripts from the Shanghai Museum corpus.)35 Also consider the various extant accounts of Kongzi’s travails when besieged “between Chen and Cai” (Chen Cai zhi jian 陳蔡之間), which constituted by far the most well-attested Kongzi tradition from the early period, with versions in the Lüshi chunqiu, Zhuangzi 莊子, Mozi 墨子, and elsewhere. In version after version, authors used the same narrative structure (the opening line “When Kongzi was in dire straits between Chen and Cai” [孔子窮於陳、蔡之間]) while varying the substance of the dialogue between Kongzi and his students, as if they cared less about mining the story for apposite sayings than exploiting its rhetorical potential for their own purposes.36 35 36

For a translation, see Eno 2015, appendix 4. See, e.g., Kong congzi 孔叢子 9 (SBCK ed., 3.5a; trans. Ariel 1989: 117) for Zisi’s 子思 admission that “among the sayings of my grandfather [i.e., Kongzi] recorded in my writings are some which I have personally heard and some which were brought to my attention by others. So even though my writings consist of words which are not precisely the Master’s, they do not fall short of the Master’s ideas” (臣所記臣祖之言,或親聞之者,有聞之 於人者。雖非其正辭,然猶不失其意焉).

The Lunyu As A Western Han Text

79

Certain early sources even call attention to the dynamism of Kongzi yue as a discursive space, including this example from the “Gui gong” 貴公 (Valuing Impartiality) section of the Lüshi chunqiu: A man of Chu lost his bow and had no desire to look for it, saying, “A man of Chu lost it and a man of Chu will find it, so why look for it?” Kongzi heard this and said, “Lose the ‘Chu’ and it’ll be fine.” Lao Dan heard this and said, “Lose ‘a man’ and it’ll be fine.” Thus it was Lao Dan who was perfectly impartial. 荊人有遺弓者,而不肯索,曰:荊人遺之,荊人得之,又何索焉?孔 子聞之曰:去其荊而可矣。老聃聞之曰:去其人而可矣。故老聃則至 公矣。37

The anecdote presents itself as a statement on the difference between Kongzi and Laozi, or perhaps between Kongzi-style and Laozi-style thought. While both figures are more “impartial” (gong 公) than the average person, Laozi is deemed more impartial than Kongzi because his perspective transcends Kongzi’s strictly human concerns. But the author of the anecdote frames that difference primarily in terms of rhetorical strategies, treating the Kongzi and Laozi sayings as two different approaches to creating memorable sayings out of preexisting material. The Kongzi version eliminates regional cues to create a shorter and more abstract saying, while Laozi shows himself to be “perfectly impartial” (zhi gong 至公), that is, to operate on the highest level of generality, by excising the human element. (The Lüshi chunqiu’s praise of Laozi was not tolerated by Liu Xiang, however, who omitted the Laozi saying altogether and instead praised Kongzi as “greatly broad-minded” [da gong 大公] in the Shuiyuan version of the anecdote.)38 How the intended audience of the Lüshi chunqiu received this anecdote is anyone’s guess. But its author seems to have conceived of Kongzi sayings not as venerable sources of “the Master’s” teachings but as a medium for abstracting wisdom from other sources. Kongzi quotation as revealed in examples like this one was an unabashedly dynamic and creative practice, in stark contrast to the Lunyu’s staid presentation of Kongzi dicta.

37 38

Lüshi chunqiu 1/4 (SBCK ed., 1.9b–10a). Shuiyuan 14 (SBCK ed., 14.6b).

80

Hunter

The Lunyu as a Layered Text

Arguably the most influential approach to dating the Lunyu to the Warring States period is the accretion theory, which is defended at length by Robert Eno in this volume (chap. 2). As articulated by Kimura Eiichi, the accretionist argument begins with an intuitive inference: one does not end up with a text as haphazardly organized as the Lunyu without a complex textual history consisting of multiple stages of accretion and redaction over an extended period of time. Had the Lunyu been compiled by a single person or group at a single moment (as Tsuda argued), we would expect the text to exhibit a consistent structure or organizing principle.39 Instead, what we find is a mishmash of editorial principles both across and within chapters, evidence that points to the intervention of multiple editors at various stages in the text’s early history. If we assume a terminus ante quem of ca. 55 BCE for the emergence of a text closely resembling the received Lunyu (on the basis of the Dingzhou 定州 Lunyu manuscript), then dating its initial compilation to the early Western Han period affords too little time for the growth of such an editorially chaotic text. Given the wealth of detail one finds in the most exhaustive accretionist studies, especially those of Kimura Eiichi and E. Bruce Brooks and A. Taeko Brooks, a sufficiently comprehensive critique of any one accretion model is beyond the scope of this essay.40 Instead, let us stipulate at the outset that the core intuition of the accretionist approach—that the Lunyu’s disorganization betrays a complicated redaction history—is correct, and that the early history of the Lunyu is a story about multiple individuals shaping the text’s form and content through compilation, interpolation, composition, redaction, etc. The critical question is whether the evidence adduced for an accretion model also compels us to imagine a lengthy textual history stretching back from the Western Han into the Warring States period. In other words, can an accretion model provide reliable information about the Lunyu’s chronology in addition to its composition? Specific applications of the accretion model can be criticized on various fronts, from the ways scholars have identified layers to the assumptions made (often on the basis of the Lunyu and post-Lunyu sources!) about the workings of an early Kongzi school.41 From a methodological perspective, however, 39 40 41

Tsuda 1946: 271–275. Kimura 1971; Brooks and Brooks 1998. With regard to the first point, see Weingarten 2010: 29–57 for the argument that “[t]he relevance of most of the criteria that scholars have so far used to distinguish textual layers in the Lunyu and to assign relative dates to them is debatable” (56). See also Schaberg’s (2001) critical review of the Brookses’ (1998) accretion model.

The Lunyu As A Western Han Text

81

using a long-term accretion model to date the Lunyu faces a few insurmountable obstacles. First, in documented examples of textual accretion, what allows one to speak confidently about accretion processes is the existence of multiple versions of a text at different stages in its transmission history.42 That is precisely what we lack in the case of the Lunyu, at least until the latter half of the Western Han period. For instance, the likelihood that Lunyu 20/3 (the final entry in the received text) was a later addition was confirmed only after the discovery of the Dingzhou Lunyu manuscript, in which 20/3 is written in smaller characters within the leftover space on the final bamboo strip of 20/2.43 In the absence of earlier manuscript witnesses, determining the form and content of any precursor texts is necessarily a matter of inference or conjecture. The second obstacle is the Lunyu’s minimal intertextual profile in the preHan textual record. Were it the case that pre-Han Lunyu parallels clustered in meaningful ways, it might be possible to read those clusters as evidence of Lunyu precursor texts. However, as seen in figure 3.1, such clustering is exceedingly rare in texts composed prior to Emperor Wu’s reign, a handful of examples excepted.44 Consequently, there is no way to corroborate applications of the accretion model to a pre-Han Lunyu tradition. The third major obstacle is that the Lunyu lacks the very thing that might ground inferences about its redaction history: a regular structure or organizing principle against which any variations or accretions stand out in sharp relief. Against the overarching narrative of the book of Genesis, for example, the inclusion of two different creation myths and the logical inconsistencies in the story of Joseph are among the anomalies explained by the “Documentary Hypothesis,” according to which the Pentateuch is a conglomeration of four different source texts.45 A cathechistic dialogue like the *Min zhi fumu manuscript 42 43 44

45

See, e.g., Matthias Richter’s (2013) analysis of the *Min zhi fumu manuscript and its relationship to the received version. For a discussion, see van Els’s contribution (chap. 6) in the present volume. See, e.g., the Lunyu 8/18–19 parallel at Mengzi 3A/4 (SBCK ed., 5.12b–13a) and the Lunyu 10/8 and 10/12 pairing at Mozi 39 (SBCK ed., 9.22a; discussed at length in Weingarten 2009). Of the ten Lunyu entries with parallels in more than one pre–Emperor Wu text (the entries in the center of fig. 1), Lunyu 7/2 and 7/34 and Lunyu 1/3, 5/25, 15/27, and 17/17 are duplicates or close intratextual parallels. For an introduction, see Friedman 1997. On the story of Joseph as an example of the explanatory power of the Documentary Hypothesis, see Baden 2012: 1–12. See also Baden 2012: 30: “The literary analysis of the Pentateuch is grounded in the basic inability to read the text as a whole, and that inability is not manifested in the variety of themes or style. Our reading of the Pentateuch is not undermined by the collocation of disparate themes or by the use of different words for ‘female servant.’ Instead, what makes the reading of the Pentateuch problematic is its lack of narrative flow, and only by addressing this problem first and foremost can we be responding authentically to the text before us.”

82

Hunter

Figure 3.1 The intertextual profile of the Lunyu in pre-Han texts

provides another kind of overarching structure, as do the canonical quotationcapped zi yue sayings of “Zi yi” 緇衣 (Jet-Black Robes).46 But the Lunyu provides no such help. On the one hand, the lack of a consistent organizing principle is cited as evidence of the accretion model’s general applicability; on the other, the Lunyu’s very disorganization frustrates efforts to parse its homologies and heterologies, let alone sequence such groupings as successive layers of accretion. In short, accretion models have to jump through too many methodological hoops to furnish reliable evidence of the Lunyu’s chronology. First, one must identify the textual units from which the Lunyu was constituted, all without the benefit of external evidence or some overarching editorial principle for determining the sequence, coherence, or completeness of such units. Second, 46

For example, it is the dialogue structure of *Min zhi fumu which grounds Richter’s (2013: 126ff.) argument that the text originally consisted of Kongzi’s single reply to Zixia’s initial question, with additional follow-up questions added at a later date.

The Lunyu As A Western Han Text

83

one must justify the interpretation of those units not just as source texts but as layers—that is, as texts added in sequence to a growing Lunyu collection.47 Third, one must place those layers in a particular sequence anchored to an initial core text whose existence cannot be corroborated. Finally, one must situate that relative chronology within an absolute chronological framework extending back into the pre-Han period. In their attempts to date the Lunyu according to internal criteria, accretion theorists from Itō Jinsai 伊藤仁齋 (1627–1705) and Cui Shu 崔述 (1740–1816) to Kimura Eiichi and E. Bruce Brooks and A. Taeko Brooks have made indispensable contributions to critical Lunyu studies. To date, however, no accretion theorist has satisfactorily answered the more basic question of whether the Lunyu is the kind of text that lends itself to such analysis.

Reading the Lunyu as a Western Han Text

The traditional view enjoys a significant advantage over the revisionist view in at least one respect. Generations of early China scholars (myself included) learned classical Chinese and first studied the masterworks of early Chinese thought under the assumption that the Lunyu preceded and thus influenced the Warring States masters and later authors. Not surprisingly, the vast bulk of early China scholarship has largely reinforced this approach by interpreting Warring States and Han texts against the Lunyu, thereby giving the traditional view an imposing aura of plausibility. When we encounter a Lunyu parallel in, say, the Mengzi, it simply makes sense to read it as a Lunyu quotation rather than as a Lunyu precursor. To argue that the Lunyu is a Western Han text, then, it is not enough to point out the evidence in favor of that view. One must also demonstrate the plausibility of reading the Lunyu as a product of the Western Han period, a task I have taken up elsewhere.48 Consider the brief dialogue between Kongzi and his disciple Zhonggong at Lunyu 13/2 together with its parallel in the much longer Shanghai Museum Zhonggong manuscript, both of which feature Kongzi dispensing advice to Zhonggong upon his appointment as chief steward of the Ji 季 clan of the state

47

48

Again to draw an analogy with biblical studies, this is the difference between the Documentary Hypothesis, which posits a single act of compilation from four originally independent source texts, and various supplementary hypotheses, according to which material was added to an original core text in stages over an extended period of time. That is the challenge taken up in Hunter 2017: chap. 5.

84

Hunter

of Lu 魯.49 In the manuscript version, Kongzi’s advice consists of four elements—“treating the elderly as elderly and being kind to the young, putting officeholders first, promoting the accomplished and talented, and forgiving transgressions and crimes” (老老慈幼,先有司,舉賢才,宥過與罪)—which are each elaborated in the course of the dialogue. But Lunyu 13/2 features three elements in a different order—“putting officeholders first, forgiving lesser transgressions, and promoting the accomplished and talented” (先有司,赦小 過,舉賢才)—of which only “promoting the accomplished and talented” is expanded upon. Judging from the close parallels in the wording and structure of these two texts, Lunyu 13/2 has a history which extends at least as far back as the late fourth century BCE (assuming that estimates of the dating of the Shanghai Museum corpus are correct). At some point someone, quite possibly the Lunyu compiler, made a conscious decision to emphasize “promoting the accomplished and talented” by rearranging the core Kongzi saying and excising the explanations of the other elements.50 Of course, it is impossible to know for sure whether the Lunyu compilers had access to a text like the Shanghai Museum Zhonggong manuscript when they included Lunyu 13/2 or whether Zhonggong had already been condensed into its “abridged” version (to borrow Chen Tongsheng’s term) prior to the creation of the Lunyu.51 Nevertheless, the preservation of those other elements as vestiges of the longer, unabridged version is one hint that the Lunyu compilers might have been the culprits. If so, then in Lunyu 13/2 we have an example of the Lunyu compilers having expressed a clear preference in the course of selecting material from an earlier tradition. Why would the Lunyu compilers value “promoting the accomplished and talented” over these other elements? Lunyu 13/2 is certainly not the only entry to indicate that “promoting” (ju 舉) and “recognizing men of accomplishment and talent” (zhi xian cai 知賢才) were among their central concerns. The term “knowing” or “recognizing others” (zhi ren 知人) appears in four sayings across four different books (Lunyu 1/16, 4/7, 12/22, and 20/3), and Lunyu 12/22 in particular argues that the most important kind of “knowing” (zhi 知) is “knowing others” (zhi ren 知人). A number of other Lunyu sayings feature Kongzi 49 50

51

For the text of the Zhonggong manuscript, see Shanghai bowuguan cang Zhanguo Chu zhushu 2001–2011: 3.261–283. The line in question appears on strip 7 (p. 268). As noted by Martin Kern in chapter 10 of the present volume, early accounts of textual redaction overwhelmingly represent it as a process of reducing corpora via selection or excision. Thus, the notion that a Lunyu grew over time is highly counterintuitive from an early perspective. Chen 2006.

The Lunyu As A Western Han Text

85

commenting on the virtues, vices, strengths, and weaknesses of his disciples and various historical figures. Such comments especially predominate in Lunyu book 5, the first half of book 6, book 11, and the middle section of book 14 and are often prefaced by questions like “what is so-and-so like?” (X也何如) (Lunyu 5/4, 5/8), “is so-and-so humane?” (ren 仁) (Lunyu 5/19, 6/30, 12/3, 14/1, 14/16, 14/17, 17/1), and “can so-and-so be charged with the conduct of government?” (可使從政也與) (Lunyu 6/8). Still other passages deal with the problem of identifying “teachers” (shi 師) (Lunyu 2/11), “stewards” (zai 宰) (Lunyu 5/8), “noble men” (junzi 君子) (e.g., Lunyu 8/6 and 14/5), and those who “understand ritual” (zhi li 知禮) (Lunyu 3/15, 3/22, 7/31). It is probably no accident that some Lunyu parallels (e.g., qiao yan ling se 巧言令色 and Lunyu 2/10) appear in a range of sources identified by Matthias Richter as “characterological” texts— that is, texts having to do with the evaluation of candidates for office.52 Character evaluation in the Lunyu is as much a matter of being known as it is of knowing others. Despite the Master’s instruction “not to worry that others do not know you” (不患人之不己知) at Lunyu 1/16, 4/14, and 14/30, arguably the most persistent anxiety in the text is that a talented and virtuous man will go “unrecognized” (bu zhi 不知) in his own time: for example, when Kongzi laments at Lunyu 14/35 that “there is no one who recognizes me” (莫我知也夫).53 Kongzi also teases his disciples for constantly complaining, “I am not known” (bu wu zhi 不吾知).54 Lunyu 1/1, the very first Kongzi saying in the text, likewise praises the junzi 君子 (noble man) who “does not resent being unrecognized by others” (人不知而不慍). In and of itself, the Lunyu’s interest in character evaluation does not necessarily point to a Western Han context, especially given that Kongzi’s assessments of the virtues and vices of his students and various (quasi-)historical figures constitute one of the most common subgenres of Kongzi sayings from the pre-Han period. However, resonances with an edict issued by Emperor Wen 文帝 (r. 180–157 BCE) in 165 BCE hint that the representation of character evaluation in the Lunyu owes something to Western Han political rhetoric surrounding the recommendation of talented individuals to the imperium.55 In the edict, Emperor Wen ordered his officials to “select men of worth and integrity” (xuan xian liang 選賢良) who were then expected to compose essays in 52 53 54 55

Richter 2005. For qiao yan ling se 巧言令色 parallels, see, e.g., Da Dai Liji 49, “Zengzi lishi” 曾子立事 (SBCK ed., 4.4a), Da Dai Liji 72, “Wen wang guan ren” 文王官人 (SBCK ed., 10.3b), and Yi Zhoushu 58, “Guan ren” 官人 (Yi Zhoushu jiaobu zhuyi 58.331). On this theme, see also Lunyu 1/1, 1/16, 4/14, 9/28, 11/26, 13/2, 14/30, 14/39, 15/4, and 15/19. See also Lunyu 14/39 and 15/4. Hanshu 49.2290. The text of the edict is contained within Chao Cuo’s 鼂錯 (d. 154) memorial.

86

Hunter

which they would criticize the emperor’s government and “fully display their ambitions without hiding anything” (悉陳其志,毋有所隱). Emperor Wen promised to “personally survey [the responses] in order to discern how the grandees will assist Us and whether or not they meet [Our criteria]” (朕親覽 焉,觀大夫所以佐朕,至與不至). The language of the 165 BCE call for recommendations was echoed in similar edicts issued by Emperor Wu in 134 and 130 and Emperor Yuan in 47.56 Compare the language of Emperor Wen’s edict with Lunyu 5/26 and 11/26, both of which feature Kongzi enjoining disciples (Yan Yuan 顏淵 and Zilu 子路 57 in 5/26 and Zilu 子路, Zeng Xi 曾皙, Ran You 冉有, and Gongxi Hua 公西華 in 11/26) to “each speak of [his] ambitions” (各言其志).58 In Lunyu 11/26, the disciples’ speeches are followed by Kongzi’s assessments of their qualifications for high office. Kongzi states that Zilu 子路 is not up to the challenge of governing a small state, that Ran You could indeed govern a small territory, and that Gongxi Hua’s seemingly modest desire to serve as a minor official in charge of ritual matters actually qualified him for higher office. Gongxi Hua’s interest in serving “the ancestral temple” (zong miao 宗廟) is also noteworthy given the three mentions of the ancestral temple in Emperor Wen’s edict and the frequent invocation of ancestral sacrifices in edicts throughout the Western Han period. The appearance of similar anecdotes in two other Western Han compilations, the Hanshi waizhuan and Shuiyuan,59 indicates that such “interview scenes” (to borrow Oliver Weingarten’s term)60 were an important subgenre of Kongzi dialogue in the period. These other versions also happen to exhibit more definite parallels with Western Han edicts. Just as in Lunyu 11/26, disciples relate their “ambitions” (zhi 志) so that Kongzi can then pronounce on their qualifications. In the Hanshi waizhuan book 9 parallel, Zilu is deemed “a brave man of service” (yong shi 勇士), Zigong “an eloquent man of service” 56

57 58 59 60

For Emperor Wu’s 134 BCE edict, see Hanshu 6.160–161; for the 130 edict, see Hanshu 58.2613–2614; for Emperor Yuan’s 47 BCE edict, see Hanshu 9.281–282 and 75.3171–3172. All these edicts feature the claim that the emperor will “personally survey the responses” (qin lan yan 親覽焉) of recommended grandees. See also Kern’s contribution (chap. 10) for the observation that one key difference between Sima Qian’s Kongzi and the Kongzi of the Lunyu is that the former repeatedly laments his nonrecognition. Lunyu 5/26 actually references a Jilu 季路, which seems to be an alternate name for Zilu. This is the formulation at Lunyu 11/26. Lunyu 5/26 has a slightly different wording: “Why don’t you each speak of your ambitions?” (盍各言爾志). See Hanshi waizhuan 7 (SBCK ed., 7.15b–16a), Hanshi waizhuan 9 (SBCK ed., 9.7b–8b), and Shuiyuan 15 (SBCK ed., 15.8a–9b). Weingarten 2010: 156: “Interview scenes read like sequences of conversations specifically devoted to plumbing the character and capabilities of the respective disciples.”

The Lunyu As A Western Han Text

87

(bian shi 辯士), and Yan Yuan “a great man of service” (da shi 大士). The overlap with the Emperor Wen edict is most evident in the same passage from Hanshi waizhuan book 9, in which Kongzi demands, “let each of you speak of your ambitions, and I shall survey them” (二三子各言爾志,予將覽焉). There the use of the verb lan 覽, a term normally reserved for kings, emperors, sages, or others in a position of political or intellectual authority, is striking. While lan 覽 does not appear in the relevant Lunyu episodes, its use in the Hanshi waizhuan points to the subgenre’s contemporary appeal: Kongzi models the role of emperor; his disciples, would-be officials. The setting of the single Shuiyuan and two Hanshi waizhuan versions also lends support to this reading. When the Qin First Emperor 秦始皇 embarked on a tour of his newly unified empire in 219, he ascended Mount Tai 泰山 to “completely survey the eastern reaches [of his realm]” (周覽東極), according to the stele inscription that marked the occasion.61 Situating these interview scenes on mountaintops was a way to emphasize Kongzi’s authority as a supreme judge of character, much as emperors sat atop the Han bureaucracy and personally judged (or so they claimed) the abilities of recommended officials.62 Taken together, the comparison between Lunyu 13/2 and the Zhonggong manuscript, the large number of Lunyu sayings dealing with character evaluation, the parallels with characterological texts, and resonances with Western Han edicts calling for the recommendation of talented officials are compelling evidence for taking “recognizing men of worth and talent” as one of the core concerns of the Lunyu compilers. The Lunyu’s perceived value for the imperial Liu clan may have even been as a manual of character evaluation, a skill that emperors and kings were expected to exercise once in office. On the flip side, aspiring officials might have read such sayings as advice on becoming, or at least presenting themselves as, “accomplished” and “talented” candidates. Given the Lunyu’s overwhelming interest in such matters, in a Western Han context the Lunyu’s title might even have been understood as Sayings on the Selection [of Men for Office] (*Lunren zhi yu 論人之語).

Conclusion

To reiterate a point from the introduction, there are few certainties in the dating of early Chinese texts. I believe that the available evidence on balance 61 62

For a translation and discussion of the Mount Tai stele, see Kern 2000: 17–23. However, whether any Western Han emperor personally oversaw the selection of capable officials is an open question.

88

Hunter

points to a Western Han date for the compilation of the Lunyu, perhaps toward the beginning of Emperor Wu’s reign in 141 BCE. However, it must be admitted that the arguments assembled here are circumstantial at best. I have examined the evidence from mentions, Lunyu intertextuality, Han bibliography, and Kongzi quotation practice without making any effort to date the language or ideas of the Lunyu. For arguments from linguistics I defer to others with more expertise than I possess.63 My reluctance to engage arguments from ideas or philosophy such as those advanced by Paul Goldin in this volume (chap. 4) stems from a genuine methodological concern. Until now, the identification of earlier versus later ideas has depended on a traditional timeline that places the Lunyu at the head of a series of Warring States master texts. To question the traditional dating of the Lunyu is thus to question the legitimacy of models built on that timeline. Future research and manuscript finds may eventually corroborate much of the traditional chronology and standard narrative of early Chinese thought. In the meantime, before using ideas to date the Lunyu, we must first ask how our understanding of those ideas would change if we were to abandon the traditional chronology and everything that follows from it. Whatever the problems with the Lunyu chronology proposed here, in my view these problems pale in comparison to those facing the traditional view. Those who would insist on the Lunyu’s pre-imperial origins are obliged to acknowledge (1) the Lunyu’s striking lack of influence on the pre-Han textual record, (2) the abundance of non-Lunyu Kongzi material in those very same texts,64 (3) the crucial importance of the Han dynasty in the canonization of the Lunyu, and (4) the biases of the Han bibliographers ultimately responsible for the traditional view. Whether or not the Lunyu existed as a book prior to the Han, to grant it a privileged place in pre-Han studies is to perpetuate a Han anachronism.

Bibliography

Ariel, Yoav. 1989. K’ung-ts’ung-tzu: The K’ung Family Masters’ Anthology—a Study and Translation of Chapters 1–10, 12–14. Princeton, NJ: Princeton University Press. Baden, Joel S. 2012. The Composition of the Pentateuch: Renewing the Documentary Hypothesis. New Haven, CT: Yale University Press. 63 64

My neglect of linguistic arguments is somewhat justified in light of Wolfgang Behr’s (2011) overview of attempts to date the language of the Lunyu, in which he concluded that such efforts have proved inconclusive. An excellent source for dealing with this material is the Kongzi jiyu jiaobu 孔子集語校 補, an updated version of Sun Xingyan’s 孫星衍 (1753–1818) Collected Sayings of Kongzi.

The Lunyu As A Western Han Text

89

Behr, Wolfgang. 2011. “Linguistic Approaches to the Dating of the Lunyu: Methodological Notes and Future Prospects.” Paper presented at “The Analects: A Western Han Text?,” Princeton University, November 4–5, 2011. Bielenstein, Hans. 1980. The Bureaucracy of Han Times. Cambridge: Cambridge University Press. Brooks, E. Bruce, and A. Taeko Brooks. 1998. The Original Analects: Sayings of Confucius and His Successors. New York: Columbia University Press. Chen Tongsheng 陳桐生. 2006. “Kongzi yulu de jieben he fanti” 孔子語錄的節本和繁 本. Kongzi yanjiu 孔子研究 2006.1:16–122. Cheng, Anne. 1993. “Lun yü 論語.” In Early Chinese Texts: A Bibliographical Guide, ed. Michael Loewe, 313–323. Berkeley: Society for the Study of Early China and Institute of East Asian Studies, University of California. Csikszentmihalyi, Mark. 2001. “Confucius.” In The Rivers of Paradise: Moses, Buddha, Confucius, Jesus, and Muhammad as Religious Founders, ed. David Noel Freedman and Michael James Mcclymond, 233–308. Cambridge: W. B. Eerdmans. Csikszentmihalyi, Mark. 2002. “Confucius and the Analects in the Hàn.” In Confucius and the Analects: New Essays, ed. Bryan W. Van Norden, 134–162. Oxford: Oxford University Press. Eno, Robert. 2015. The Analects of Confucius: An Online Teaching Translation (Version 2.2). http://www.indiana.edu/~p374/Analects_of_Confucius_(Eno-2015).pdf, accessed May 1, 2018. Friedman, Richard Elliott. 1997. Who Wrote the Bible? 2nd ed. San Francisco: Har­per­ Collins. Hanshu 漢書. 1962. Beijing: Zhonghua shuju. Haupt, Christiane. 2006. “Und der Meister sprach … : Die Darstellung des Konfuzius in Texten der Zhanguo- und Frühen Han-zeit.” PhD diss., Ludwig-Maximilians Uni­ versity, Munich. Hightower, James R. 1993. “Han shi wai chuan.” In Early Chinese Texts: A Bibliographical Guide, ed. Michael Loewe, 125–128. Berkeley: Society for the Study of Early China and Institute of East Asian Studies, University of California. Hou Hanshu 後漢書. 1965. Beijing: Zhonghua shuju. Hunter, Michael. 2014. “Did Mencius Know the Analects?” T’oung Pao 100:33–79. Hunter, Michael. 2017. Confucius Beyond the Analects. Leiden: Brill. Kaneto Mamoru 金戸守. 1970. “Rongo buntei kō: Gen Rongo seiritsu o megutte” 論語文 体考: 原論語成立をめぐって. Shitennōji joshi daigaku kiy 四天王寺女子大学紀要 2:32–58. Kaneto Mamoru 金戸守. 1971. “Kōshi to so no jidai: Rongo hihan oboegaki” 孔子とその 時代: 論語批判覚.” Shitennōji joshi daigaku kiy 四天王寺女子大学紀要 4:101–115.

90

Hunter

Kaneto Mamoru 金戸守. 1972. “Kobun Rongo kō: Shiki deshi iu, Kansho Geibunshi” 古 文論語考:史記弟子云, 漢書藝文志. Shitennōji joshi daigaku kiy 四天王寺女子大学 紀要 5:43–59. Kaneto Mamoru 金戸守. 1973. “Shiki Rongo kō: Kōshi seika, sono ‘Rongo’ seiritsu shiron” 史記「論語」考: 世家、その「論語」成立試論. Shitennōji joshi daigaku kiy 四天王寺女子大学紀要 6:14–38. Kaneto Mamoru 金戸守. 1974. “Shiki Rongo kō, sono ni: Chūji deshi retsuden yori Rongo seiritsu ni tsuite” 史記「論語」考、その二:仲尼弟子列伝より論語成立に ついて. Shitennōji joshi daigaku kiy 四天王寺女子大学紀要 7:51–70. Kaneto Mamoru 金戸守. 1975. “Shiki Rongo kō, so no mi: Honji, Hassho no Rongo seiritsu shiryō ni tsuite” 史記「論語」考、その三: 本紀,八書の論語成立資料につ いて. Shitennōji joshi daigaku kiy 四天王寺女子大学紀要 8:45–59. Kaneto Mamoru 金戸守. 1976. “Shiki Rongo kō, so no shi: Seika inyo Koshi-gen yori Rongo no seiritsu” 史記「論語」考、その四: 世家引用論語の成立. Shitennōji joshi daigaku kiy 四天王寺女子大学紀要9:101–118. Kaneto Mamoru 金戸守. 1977. “Shiki Rongo kō, so no go: Haku’i retsuden o ronjite Rongo Koshi-gen ni oyobu” 史記「論語」考、その五: 伯夷列伝を論じて論語孔子 言に及ぶ. Shitennōji joshi daigaku kiy 四天王寺女子大学紀要 10:65–79. Kaneto Mamoru 金戸守. 1978. “Shiki Rongo kō, so no roku: Soketsu, Rongo no seiritsu ni tsuite” 史記「論語」考、その六: 総結、論語の成立について. Shitennōji joshi daigaku kiy 四天王寺女子大学紀要 11:19–26. Kaneto Mamoru 金戸守. 1979. “Shiki Rongo kōyo” 史記論語考余. Shitennōji joshi daigaku kiy 四天王寺女子大学紀要 12:21–32. Kaneto Mamoru 金戸守. 1980. “Shiki Rongo ko yoshoku: Kansho yori mita Rongo seiritsu kō” 史記論語考余続:漢書より見た論語成立考. Shitennōji joshi daigaku kiy 四天王寺女子大学紀要 13:17–29. Kern, Martin. 2000. The Stele Inscriptions of Ch’in Shih-Huang: Text and Ritual in Early Chinese Imperial Representation. New Haven, CT: American Oriental Society. Kimura Eiichi 木村英一. 1971. Kōshi to Rongo 孔子と論語. Tokyo: Sōbunsha. Kongzi jiyu jiaobu 孔子集語校補. 1998. Compiled by Sun Xingyan 孫星衍 et al. Jinan: Qi Lu shushe. Loewe, Michael. 2000. A Biographical Dictionary of the Qin, Former Han and Xin Periods (221 BC–AD 24). Leiden: Brill. Loewe, Michael. 2011. Dong Zhongshu, a “Confucian” Heritage and the Chunqiu fanlu. Leiden: Brill. Makeham, John. 1996. “The Formation of Lunyu as a Book.” Monumenta Serica 44: 1–24. Nylan, Michael. 1994. “The Chin Wen / Ku Wen Controversy in Han Times.” T’oung Pao 80:83–145.

The Lunyu As A Western Han Text

91

Richter, Matthias. 2005. Guan Ren: Texte der altchinesischen Literatur zur Charakterkunde und Beamtenrekrutierung. Bern: Lang. Richter, Matthias. 2013. The Embodied Text: Establishing Textual Identity in Early Chinese Manuscripts. Leiden: Brill. Schaberg, David. 2001. “‘Sell It! Sell It!’: Recent Translations of Lunyu.” Chinese Literature: Essays, Articles, Reviews 23:115–139. Shanghai bowuguan cang Zhanguo Chu zhushu 上海博物館藏戰國楚竹書. 2001–2011. Ed. Ma Chengyuan 馬承源. 8 vols. Shanghai: Shanghai guji chubanshe. Shiji 史記. 1959. Beijing: Zhonghua shuju. Sibu congkan (chubian) 四部叢刊 (初編). 1922. Shanghai: Shangwu yinshuguan. Digit­ ized by Unihan Digital Technology Co. in 2001. Version 1.0. Suishu 隨書. 1973. Beijing: Zhonghua shuju. Tsuda Sōkichi 津田左右吉. 1946. Rongo to Kōshi no shisō 論語と孔子の思想. Tokyo: Iwanami shoten. Weingarten, Oliver. 2009. “Confucius and Pregnant Women: An Investigation into the Intertextuality of the Lunyu.” Journal of the Americal Oriental Society 129:1–22. Weingarten, Oliver. 2010. “Textual Representations of a Sage: Studies of Pre-Qin and Western Han Sources on Confucius (551–479 BC).” PhD diss., Cambridge University. Zhao Zhenxin 趙貞信. 1961. “Lunyu jiujing shi shei bianzuan de”《論語》究竟是誰編 纂的. Beijing Shifan daxue xuebao 北京師範大學學報 4:11–24. Zhu Weizheng 朱維錚. 1986. “Lunyu jieji cuoshuo” 論語結集脞說. Kongzi yanjiu 孔子研 究 1:40–52. Zhu Weizheng 朱維錚. 1987. “Lishi de Kongzi he Kongzi de lishi” 歷史的孔子和孔子的 歷史. In Kongzi yanjiu lunwen ji 孔子研究論文集, ed. Zhonghua Kongzi yanjiusuo 中華孔子研究所, 156–172. Beijing: Jiaoyu kexue chubanshe.

92

Goldin

Chapter 4

Confucius and His Disciples in the Lunyu: The Basis for the Traditional View Paul R. Goldin There is an emerging consensus that the received text of the Analects (Lunyu 論語), though regarded throughout Chinese history as the best single source for the life and philosophy of Confucius,1 did not exist before the Han dynasty. The work of scholars such as Zhu Weizheng 朱維錚, John Makeham, and Mark Csikszentmihalyi has left little doubt that the text was redacted sometime during the Western Han.2 This does not necessarily mean, however, that the contents must date to a period later than Confucius and his disciples.3 A work that was compiled in a certain century does not necessarily consist of material dating from that same century.4 Thus, the new insights regarding the relatively late compilation of the Analects do not invalidate the traditional understanding of the text’s philosophical importance. In this chapter, I shall present several examples suggesting that the Analects reflects an intellectual environment from long before the Han dynasty. These distinctive features of the text would have to be explained by any theory of its origin. The same evidence will also support the traditional chronology, which postulates the sequence Analects– Mozi–Mencius–Laozi–Xunzi. 1 E.g., Creel 1949: 291: “All scholars seem to be agreed that, while some parts of the Analects are subject to question, the book in general is our best single source for Confucius.” Creel was already aware, it should be noted, that there are no secure references to the received text of the Analects from before the Han. 2 Zhu Weizheng 2002: 97–123; Makeham 1996, 2006; Csikszentmihalyi 2002, 2004: 28ff. 3 Despite, e.g., Weingarten 2009: 598: “the textual material found in the received Lunyu may have originated anytime between Confucius’s death in the early fifth century BCE and the second century BCE, or even later in the case of interpolations.” See also Weingarten 2010: 199. Stumpfeldt 2010: 24: “In the past two decades … it has been recognised that the material in the [Analects] is highly heterogeneous in nature and of varying authenticity” (without any references). Close parallels between two lines from *Yucong 3 語叢三 (a manuscript from Guodian 郭店) and Analects 7/6 and 9/4 show that at least some material in the Analects has truly early origins: Cook 2012: 2.866–868; Li Xueqin 2009: 293–297. (Note that there is an earlier edition of Li with different pagination.) 4 This is similar to Li Ling’s (2008: 213) analogy of an ancient Chinese text as a glass of wine, with the wine (the textual contents) and the glass (the edition in which the textual contents have been transmitted) to be distinguished analytically.

© Koninklijke Brill NV, Leiden, 2018 | doi:10.1163/9789004382947_006

Confucius And His Disciples In The Lunyu



93

Evidence from Intellectual History

The Analects make no mention of philosophical developments that took place after the fifth century BCE. Most strikingly, the text never refers to the concept of physical self-cultivation.5 This point is well illustrated by a comparison of the Confucian Analects with the Mohist Analects (sometimes called “Moyu” 墨 語 in Chinese), a set of documents with a very similar structure and thus, though profoundly neglected, a natural comparandum.6 First, a word about the two texts. When I refer to the Confucian Analects, I shall restrict myself to chapters 1–15, as I accept the argument of D. C. Lau and others, on the basis of observations by Cui Shu 崔述 (1740–1816), that the last five chapters of the received text are written in a different style and belong to a different era.7 The informal title “Mohist Analects” is commonly used to refer to chapters 46–50 of the Mozi 墨子. These sections relate discussions between Mo Di 墨翟 (d. ca. 390 BCE) and an array of interlocutors consisting of disciples, political figures such as territorial lords, and representatives of various dissenting schools and points of view. The most famous passage in the series appears in chapter 50 (“Gongshu” 公輸), which tells of Mo Di’s role in dissuading the state of Chu 楚 from attacking Song 宋 with siege machinery invented by one Gongshu Ban 公輸般.8 But this is the least useful portion of the text for a number of reasons, in addition to its suspicious narrative mode. The chapter includes virtually no information on Mo Di’s philosophy and makes only passing reference to his disciples.9 For chapters 46–49, however, the epithet “Analects” fits very well, as the language and composition indeed remind the reader of the Confucian Analects. The text presents one brief episode after the other, with a minimum of details, always culminating in a teaching from the mouth of the Master. Some scholars accept the Mohist Analects as the Mohist equivalent of the Lunyu—that is, as an authoritative record of Mo Di’s life and sayings compiled after his death by his disciples.10 There are some thorny issues of dating, however, and it is clear, 5 6 7 8 9 10

Slingerland 2000: 139. See, e.g., Ding 2011. Lau 1992: 265–275. This is not to suggest that chapters 16–20 can be treated as a block from a single extraneous source. For the controversy regarding Gongshu Ban’s given name, see the commentary in Wu 2006: 13.50.749n1; Goldin 2005: 158n45. Luan 1957: 119. E.g., Zhang (1936) 1975: 11.46.412, at the beginning of the “Gengzhu” 耕柱 chapter. Zhang also cites the opinion of Wang Kaiyun 王闓運 (1833–1916). Wu Yujiang (2006: 1027) cites the opinions of Hu Shi 胡適 (1891–1962) and Liang Qichao 梁啟超 (1873–1929), both of whom compare these chapters with the Lunyu. See also Zheng Jiewen 2006: 1.1–4; Jiang

94

Goldin

at any rate, that the received version has been standardized by some editor.11 (Mo Di’s disciple Qin Guli 禽滑釐, for example, is referred to as a master in his own right.)12 Nevertheless, those scholars who treat the Mohist Analects as a genuine document from the period just after Mo Di’s death must consider it one of the most precious texts in the history of Chinese philosophy. Not only would it qualify as one of the oldest surviving works, but we would have to regard it as a Mohist yulu 語錄, a record of the Master’s spoken teachings.13 In other cases—notably that of Zhu Xi 朱熹 (1130–1200)—where a thinker’s yulu is extant, it generally comes to be one of the most carefully studied texts in his oeuvre. In the case of the Mohist Analects, however, I know of only one such intensive investigation: Zheng Jiewen’s 鄭傑文 Zhongguo Moxue tongshi 中國 墨學通史, published in 2006.14 This lack of attention may explain why the most important difference between the Mohist Analects and their Confucian analogue has eluded critics: the Mohist Analects discuss a greater variety of philosophical positions. For

11

12 13 14

1985: 477–478; Wang Dongzhen 1981: 51–52. Japanese scholars tend to take an opposing view; e.g., Yoshinaga 2004: 34–69; Watanabe 1973: 538–539. Both argue that these chapters are among the latest in the entire Mozi. (See also Ding 2011.) Durrant (1977–1978: esp. 265–266) observes that the grammar of the Mohist Analects is distinct in certain respects from that of other portions of the Mozi and suggests that differences in the intended audience (i.e., not necessarily differences in date) may account for these “textual contrasts.” There are two notorious problems of dating. The first involves Mo Di’s interview with Qi Taiwang 齊大王, which ostensibly means “King Tai of Qi,” in “Luwen” 魯問 (Wu 2006: 13.49.718). Su Shixue 蘇時學 (fl. 1865) and others identified this as a reference to Tian He 田和 (d. 385 BCE), otherwise known as Lord Tai 太公, who usurped the throne of Qi in 391 BCE. See Shiji 46.1886–1887. Some scholars have suggested that because of the title “King Tai,” the received text of the “Luwen” must date from after 357 BCE, since it was only in that year that the Tian family began calling themselves kings. However, the issue is not so simple. First, the Japanese edition of Hōryaku 寶曆 7 (= 1757) notes that one version reads Qi dafu 齊大夫—in other words, “a grandee of Qi” rather than “King Tai of Qi.” In addition, Bi Yuan 畢沅 (1730–1797) pointed out that the parallel passage in Taiping yulan 346.4a does not even contain the character da/tai at all, yielding simply “the king of Qi.” There is no conclusive evidence, therefore, that the phrase refers to Tian He. Note that Forke (1922: 579n7) is thoroughly confused regarding the dates; the best summary is still to be found in Qian Mu 1956: §§64–65, 70. The second problem is more complicated and has to do with the figure of Lord Wen of Luyang 魯陽文君, who appears in several anecdotes. For an overview of the issues (if not a perfect solution), see He Hao 1994. See Luan 1957: 118; Luo Genze 1958: 194. Tan 1995: 21. Zheng Jiewen (2006: 1.4) writes at the beginning of his study that the Mohist Analects are “the most original material of the Mohist school, and therefore in what follows I shall rely chiefly on what is recorded in these five chapters” (更爲原始的墨家資料。所以,以 下主要依據此5篇所記).

Confucius And His Disciples In The Lunyu

95

example, A. C. Graham suggested that one of the passages addresses what he called “Yangism”:15 Master Wuma said to Master Mo, “I am different from you. I cannot love universally. I love people from Zou more than people from Yue, people from Lu more than people from Zou, people from my district more than people from Lu, people in my family more than people from my district, my parents more than [other] people in my family, and myself more than my parents. I go by what is closer to me. If you strike me, it hurts; if you strike someone else, it does not hurt me. For what reason should I not resist what hurts me and resist what does not hurt me? Because I exist, there are [cases] where I [might] kill someone else in order to benefit myself, but there are no [cases] where I [might] kill myself in order to benefit someone else.”  Master Mo said, “Will you conceal your principle? Will you tell people your opinion?”  Master Wuma said, “Why should I conceal my principle? I would tell people.”  Master Mo said, “Then if one person is persuaded by you, one person will want to kill you in order to benefit himself. If ten people are persuaded by you, ten people will want to kill you in order to benefit themselves. If the whole world is persuaded by you, the whole world will want to kill you in order to benefit itself.” 巫馬子謂子墨子曰:我與子異,我不能兼愛。我愛鄒人於越人,愛魯 人於鄒人,愛我鄉人於魯人,愛我家人於鄉人,愛吾親於我家人,愛 我身於吾親,以為近我也。擊我則疾,擊彼則不疾於我,我何故疾者 之不拂,而不疾者之拂?故有我,有殺彼以利我,無殺我以利彼。   子墨子曰:子之義將匿邪?意將以告人乎?   巫馬子曰:我何故匿我義?吾將以告人。   子墨子曰:然則一人說子,一人欲殺子以利己;十人說子,十人欲 殺子以利己;天下說子,天下欲殺子以利己。16

What is wrong with Master Wuma’s “principle” (yi 義), Mo Di argues, is that if everyone in the world were to act on it, the result would be death and hardship.17 To judge from his own examples, Master Wuma is from Lu; some 15 16 17

The association of Master Wuma with Yangism goes back to Sun Daosheng 1982. “Gengzhu” 耕柱 (Wu 2006: 11.46.645). Cf. the translation in Mei 1929: 219–220. Graham 1989: 61–62.

96

Goldin

commentators equate him with a disciple of Confucius named Wuma Shi 巫馬 施 (b. 521 BCE) or with one of his descendants.18 Whether he is really a “Yangist,” that is to say, a follower of Yang Zhu 楊朱, is impossible to determine,

because little is known about Yang Zhu’s teachings, and the term “Yangism” (like the corresponding Chinese term weiwo 為我) has been used promiscuously in the scholarly literature to cover many different ideas.19 Still, it is evident that the Mohist Analects are aware of an important new movement in Chinese thought: concern for the self and the body. Elsewhere, Mo Di says: Suppose you said to someone, “I shall grant you a cap and sandals but cut off your hands and feet. Would you do it?” He certainly would not. For what reason? Because a cap and sandals are not as valuable as hands and feet. Or you might say, “I shall grant you the world but kill your body. Would you do it?” He certainly would not. For what reason? Because the world is not as valuable as one’s body.20 今謂人曰:予子冠履,而斷子之手足,子為之乎?必不為。何故?則 冠履不若手足之貴也。又曰:予子天下,而殺子之身,子為之乎?必 不為。何故?則天下不若身之貴也。

This was a very common idea in Warring States China: worldly possessions are not as valuable as one’s life and limbs. (It is close, incidentally, to the kind of position that some later testimony associates with Yang Zhu himself.)21 The Annals of Lü Buwei (Lüshi chunqiu 呂氏春秋), for example, appears to borrow Mozi’s image:

18

19 20 21

See the comments by Su Shixue and Wu Yujiang (Wu 2006: 11.46.647n9). Wuma Shi’s name is sometimes read Wuma Qi 旗/期; see the commentary of Chen Qiyou in Chen 2002: 18.1240n18. However, it is possible that some other Master Wuma is intended; earlier in the same chapter (Wu 2006: 11.46.643), Master Wuma is quoted as saying, “To set aside men of the present and praise the former kings is to praise decomposed bones; it is as though a carpenter were to know decomposed wood but not to know fresh wood” (舍今 之人而譽先王,是譽槁骨也。譬若匠人然,知槁木而不知生木). This does not sound like the opinion of a disciple of Confucius. Goldin 2003: 204ff. “Guiyi” 貴義 (Wu 2006: 12.47.670). Cf. the translation in Mei 1929: 222. E.g., Liu 1989: 13.436 (“Fanlun” 氾論): “Keeping one’s nature whole and protecting one’s purity, not tying down one’s body with [material] objects—this is what Master Yang proposed” (全性保真,不以物累形,楊子之所立也).

Confucius And His Disciples In The Lunyu

97

Suppose there is a man here. He cuts off his head in order to exchange his cap [i.e., for a better one?] and kills his body in order to replace his clothes. The world would certainly think him deluded. Why is this? A cap is what adorns the head; clothes are what adorns the body. If one kills that which is adorned but treats as essential that which adorns, then one does not know what things are for. 今有人於此。斷首以易冠,殺身以易衣。世必惑之。是何也?冠所飾 首也,衣所飾身也。殺所飾而要所以飾,則不知所為矣。 22

Other passages in the Mohist Analects reinforce the reader’s sense that Mo Di has known about such arguments for some time. Consider the following: Master Mozi said, “The scholars of today handle their bodies less carefully than a merchant handles a roll of cloth. When a merchant handles a roll of cloth,23 he dares not sell it recklessly; he must choose the best ones. Scholars of today do not handle their bodies like this. Whatever their wills desire, they do. In the worst cases, they are punished; in the best cases, they suffer the stain of defamation. Thus, the scholars of today handle their bodies less carefully than a merchant handles a roll of cloth.” 子墨子曰:今士之用身,不若商人之用一布之慎也。 商人用一 布,不敢繼苟而讐 [=售 ]24 焉,必擇良者。今士之用身則不然,意之 所欲則為之,厚者入刑罰,薄者被毀醜。則士之用身,不若商人之用 一布之慎也。25

For all their moralistic pretensions, “the scholars of today” are reckless and end up endangering themselves. Mo Di himself, of course, is not sympathetic to the idea of saving one’s skin at all costs,26 and as we have seen, he cannot accept the arguments of Master Wuma uncritically. Nevertheless, he is portrayed as 22

23 24 25 26

“Shenwei” 審為 (Chen 2002: 21.1463). Cf. the translation in Knoblock and Riegel 2000: 557. The same chapter attributes a similar argument to a philosopher named Zihuazi 子華子 (Chen 2002: 21.1464). The latter argument also appears almost verbatim in the Zhuangzi (Guo 1961: 9B.28.969ff.; “Rangwang” 讓王). Following Wu Yujiang’s commentary: the second bu 布 in the original text is excrescent. Following Bi Yuan’s commentary. “Guiyi” (Wu 2006: 12.47.672). Cf. the translation in Mei 1929: 225. Consider the oft-cited story of Meng Sheng 孟勝 (d. 381 BCE), a Mohist master who preferred to die rather than betray the honor of the school (“Shangde” 上德; Chen 2002: 19.1266). Cf. Zheng Jiewen 2006: 1.67–69; Kanaya 1997: 1.312–313, 404; Graham 1989: 44–45.

98

Goldin

being so comfortable with such rhetoric that he is willing to adapt it to his own purposes. Whether or not we identify the theme specifically as “Yangist,” it is well known that this concern for one’s own health and well-being came to play a significant role in Chinese intellectual history from about the fourth century BCE onward. And yet the whole idea is absent from the Confucian Analects. Other terms familiar from the materialistic paradigm of most self-cultivation discourse, such as yinyang 陰陽, are likewise unattested—in striking contrast to a text like Mencius 2A.2, which must be read against the background of physical self-cultivation theories that Mencius wished to discredit.27 [The disciple Gongsun Chou 公孫丑 said], “I venture to ask: wherein lie your strengths, Master?”  [Mencius] said, “I know words. I am good at nourishing my flood- like qi.”  I venture to ask, “What do you mean by ‘flood-like qi’?”  “It is difficult to say. It is the kind of qi that is greatest and firmest. If it is nourished with uprightness and is not damaged, it fills the space between Heaven and Earth. It is the kind of qi that accompanies righteousness and the Way. Without it, [the body] starves. It is engendered by the accumulation of righteousness and is not obtained through sporadic righteousness. If there is something in one’s actions that does not satisfy the heart, then [the flood-like qi] starves.” 敢問夫子惡乎長?   曰:我知言,我善養吾浩然之氣。   敢問何謂浩然之氣?   曰:難言也。其為氣也,至大至剛;以直養而無害,則塞于天地之 間。其為氣也,配義與道;無是,餒矣。是集義所生者,非義襲而取 之也。行有不慊於心,則餒矣。28

Mencius’s theory, in a nutshell, is that habitual good conduct, not meditation or gymnastics, is what produces a healthy physique.29 It is hard to imagine how this dialogue could have proceeded without an awareness of traditions like that of The Internal Enterprise (Neiye 内業), which explains how nourishing one’s qi can be used to attain a heightened state of intelligence and well- 27 28 29

For the observation about yinyang, see Pines 2002: 701. Mengzi zhengyi 6.199–202. Cf. the translation in Lau 2003: 61ff. Goldin 2011a: 42.

Confucius And His Disciples In The Lunyu

99

being.30 In sum, both Mo Di and Mencius seem to be well informed about contemporary practices of physical self-cultivation; the Confucius of the Analects does not. To move now to another philosophical innovation that seems to have taken place after Confucius, but before Mo Di and Mencius: the Mohist Analects display an awareness of the school of Shennong 神農, the Divine Farmer.31 This is especially noteworthy in that arguments attributed to followers of Shennong are discussed in the Mencius, as we shall see presently. “Among the rustic people of the south of Lu,” we read in the Mohist Analects, “there was one Wu Lü.32 In the winter he made pottery and in the summer he plowed; he compared himself to Shun” (魯之南鄙人有吳慮者,冬陶夏 耕,自比於舜). Wu Lü believes that there is no use in talking so much about ethics: all one needs to do is attend to the daily labors necessary for survival. Mo Di has a devastating response: Master Mozi said, “I have calculated this. I considered plowing and feeding the people of the world. If I am successful, I shall do the plowing of one farmer. Divide this among the world, and one will be unable to let each person get a single sheng of grain. Even if one could get a sheng of grain, it is obvious that that is not enough to satisfy those who are starving in the world. I considered weaving and clothing the people of the world. If I am successful, I shall do the weaving of one woman. Divide this among the world, and one will be unable to let each person get a single chi of cloth. Even if one could get a chi of cloth, it is obvious that that is not enough to warm those who are cold in the world.” 子墨子曰:翟嘗計之。翟慮耕而食天下之人矣,盛,然後當一農之 耕,分諸天下,不能人得一升粟。籍 [=藉] 33 而以為得一升粟,其不 能飽天下之飢者,既可睹矣。翟慮織而衣天下之人矣,盛,然後當一 婦人之織,分諸天下,不能人得尺布。籍而以為尺布,其不能煖天下 之寒者,既可睹矣。34

Mo Di goes on to claim that his value as a teacher is far greater than it would be as a farmer, since a teacher can teach everyone to plow, whereas a farmer plows 30 31 32 33 34

Roth 1999. For the seminal study of the subject, see Graham 1986: 67–110; also Graham 1989: 61–74. Bi Yuan notes that in the parallel passage in Taiping yulan 822.8a, Wu Lü is called Wu Xian 吳憲. Following Bi Yuan’s commentary. “Luwen” 魯問 (Wu 2006: 13.49.720). Cf. the translation in Mei 1929: 248–249.

100

Goldin

singly. It is impossible to say whether this Wu Lü was in fact an adherent of the Shennong group, since the text does not tell us anything more about him. But the whole episode, including Mo Di’s rebuttal, is comparable to Mencius’s encounter with Xu Xing 許行 (3A.4), who is identified explicitly as a practitioner of the words of Shennong.35 Xu Xing’s tenet is: “The worthy [ruler] plows and eats together with his people; he governs while he prepares his morning and evening meals” (賢者與民並耕而食,饔飧而治).36 When Mencius hears about Xu Xing from his disciple, he proceeds to dismantle his simplistic position: Mencius said, “Must Master Xu sow grain before he can eat?” [His disciple] said, “Yes. ” … “Does Master Xu wear a cap?” “Yes.” “What kind of cap?” “His cap is made of plain [silk].” “Did he weave it himself?” “No, he traded grain for it.” “Why does Master Xu not weave himself?” “It would interfere with his plowing.” 孟子曰:許子必種粟而後食乎?曰:然 … 許子冠乎?曰:冠。 曰:奚冠?曰:冠素。 曰:自織之與?曰:否。以粟易制。 曰:許子奚為不自織?曰:害於耕。37

Mencius goes on to show that even the metal in Xu Xing’s plow must have been acquired from someone else. A harmonious society, Mencius demonstrates, must have various professionals devoting their energies to different tasks. For Mencius, who believes that human beings are moral animals, such professionals include specialists in ethics like himself. But the key to his refutation of Xu Xing is echoed, if not anticipated, by Mo Di’s argument in the Mohist Analects: it is simply not efficient for everyone to farm. Mencius stresses the importance

35

36 37

Cf. Mei 1934: 132–133; Hsiao 1979: 220n21; Yang Junguang 1992: 228–229, 304–305. Qian Mu (1956: §113) identified Xu Xing with one Xu Fan 許犯, who is said in “Dangran” 當染 (Chen 2002: 2.98) to have studied with Qin Guli; see also Qian 1930: 56–57. This idea has been refuted several times; see, in addition to the above, Tong 2005: 2.658–661; Fang 1937: 143– 144. Mengzi zhengyi 11.367. See Zhao Qi’s 趙歧 (d. 201) commentary for the terms yong 饔 and sun 飧; Jiao Xun (1987) also cites the concurring opinion of Wang Niansun. Jiao 1987: 11.369–370. Cf. the translation in Lau 2003: 113.

Confucius And His Disciples In The Lunyu

101

of overseers, Mo Di that of teachers; both are really talking about the division of labor.38 In the Confucian Analects, Confucius and his disciples are silent on this issue. There are various hermits who appear here and there, and who criticize Confucius for reasons that are often inscrutable. But no one makes a case remotely resembling those of Wu Lü and Xu Xing, whether as an argument to be advanced or to be refuted. Last, the absence of any probing discussions relating to xing 性 and qing 情 must be mentioned too.39 As A. C. Graham has shown, qing (often written 請) is a widely attested philosophical concept from the Mozi onward,40 but the term appears only twice in the Analects (13/4 and 19/19), and in neither case is it a matter of any controversy. As for xing, we are told pointedly by Zigong 子貢 (i.e., Duanmu Ci 端木賜, 520–456 BCE) that it is one of the concepts never discussed by Confucius (Analects 5/12). The only other appearance of xing is in Analects 17/2, where Confucius states, “By their xing, people are close to one another; they grow distant from one another through practice” (性相近也,習 相遠也). Not only is the quotation doubtful because chapter 17 is spurious in its entirety; the fact that Confucius and his disciples would have contented themselves with such a thin account of xing suggests that they were uninterested in, and perhaps uninformed of, the raging debates over xing that characterized philosophy in the ensuing centuries. (As examples, consider the positions taken not only by Mencius and Xunzi but also in such texts as the recently excavated *Xing zi ming chu 性自命出 [Natural Disposition Emerges from the Mandate].)41 Zigong’s comment would seem to convey that the term was known to Confucius but that he did not think it demanded extensive investigation.

38 39 40 41

For other references to this concept in the Mozi, see, e.g., Wang Dongzhen 1981: 203–204; Okamoto 1992: 21ff. One might also note the absence of su 俗 (vulgar), a word that is not common in texts from before the fourth century BCE (Pines 2005–2006: 185). For a general discussion of the language of the Analects, see Charles N. Li 1996. Graham 1978: 179–182; also Graham 1986: 59–65. A recent conference volume contains two helpful overviews of qing: Puett 2004 and Harbsmeier 2004. See, e.g., Goldin 2005: 36–57. The conversation in “Tangong” 檀弓 (Sun Xidan 1989: 10.270–272) between Youzi 有子 and Ziyou 子游 about the reasons for leaping (yong 踴) at a funeral leads to an analysis of qing that is of a piece with such texts as Xing zi ming chu and even the “Great Preface” (Daxu 大序) to the Odes. The fact that such discourse is entirely absent from the Analects undermines the commonplace assumption that “Tangong” can be considered alongside the Analects as a record of Confucius’s speech and actions. I am indebted to Kenneth W. Holloway for this reference.

102

Goldin

In Analects 9/14, Confucius states a viewpoint that might be construed as a forerunner of the theories of Xunzi and the Guodian manuscripts: “If a noble man were to dwell among [the Nine Barbarians], what crudeness would there be?” (君子居之,何陋之有). The implication is that people without the benefit of instruction by a noble man will be as crude as savages (and, by the same token, that savages can be civilized with the right training).42 Yet, unlike most Han writers,43 he seems to have felt no need to associate these ideas with the keyword xing.

Evidence from Philosophical Vocabulary

The philosophical vocabulary of the Confucian Analects does not reflect changes in the intellectual world that took place in the third and fourth centuries BCE. The linguistic tests pioneered by Bernhard Karlgren (1889–1978) have proven to be almost totally invalid; they were questionable even in their own day, and the plethora of Warring States and early imperial manuscripts discovered since the 1970s have shown, once and for all, that texts were not composed in a manner that would support Karlgren’s assumptions.44 But these developments do not disqualify the study of philosophical terminology, which inevitably reflects the intellectual orientations of an author’s expected audience. For example, a knowledgeable reader will know that a text discussing “qualia” in connection with consciousness and perception cannot date from before the twentieth century, and not for any arcane linguistic reasons but because the word was coined—in its specific philosophical sense—by Clarence Irving Lewis in 1929.45 Or, to take an example closer to the phenomenon that I shall describe in the Analects: when a text employs a technical term whose meaning changed over time, its usage can be helpful to historians who wish to place it in its intellectual context. Thomas S. Kuhn has pointed out that in the history of science, one such telling word is “planet,” which can be indicative of an entire underlying system of thought. If a work of astronomy refers to the moon as a “planet,” a reader has strong reason to presume that it was written in ignorance of Copernicus.46 Another term that can be used as a benchmark is “element”; if a 42 43 44 45 46

See Goldin 2011a: 220–224. Goldin 2011b: 228–232. See the trenchant comments by Rudolf Wagner in Allan and Williams 2000: 130; also Shaughnessy 2006: 40–41; Luo Shaodan 2003. C. Lewis 1929: 121. Kuhn 1957: 45–77.

Confucius And His Disciples In The Lunyu

103

text refers to Earth, Air, Fire, and Water as “elements,” it is clearly operating according to an ontological system other than that of modern chemistry. Naturally, not all words are equally valuable for this sort of analysis. “Mastiff” has had essentially the same relatively narrow array of connotations for centuries. “Perversion” has not.47 The first of two lexical items that have odd senses in the Analects—or, more precisely, antique-looking senses—is zhong 忠. In most philosophical writing from the Eastern Zhou and Han, zhong means something very close to “loyalty,” but this is obviously not what the disciple Zeng Can 曾參 (505–436 BCE?) means when he identifies zhong 忠 and shu 恕 as the keystone of Confucius’s ethics (e.g., Analects 4/15). As I have argued elsewhere,48 the zhong of the Analects is related to a word that is more typically written with the graph 中 in bronze inscriptions and the Exalted Documents (Shangshu 尚書), where it bears the sense of “being impartial,” “being unbiased,” sometimes specifically “hearing both sides of a case.” Only from the Shenzi 慎子 (attributed to Shen Dao 慎到, b. ca. 360 BCE) onward, it turns out, is the sense of “loyalty” attested. The older meaning did not die out immediately. There are other texts, such as Mozi and Zuozhuan 左傳, where zhong is used to mean either “loyalty” or “impartiality,” depending on the context. But by the time we get to the world of Sunzi 孫子 and Xunzi 荀子, zhong has scarcely any meaning other than “loyalty.” (And this detail would also seem to throw into doubt the theory that the Sunzi is the oldest Chinese philosophical text—older, supposedly, than even Confucius.)49 The use of zhong in the Analects is thus markedly conservative, if not archaic. For that matter, the whole “Golden Rule” approach to moral reasoning, which informs the concepts of zhong and shu, does not seem to have remained popular after the time of Confucius’s disciples.50 Confucians and non-Confucians alike—whether Mo Di, Mencius, the philosophers of the Laozi 老子 and Zhuangzi 莊子 traditions, or Xunzi—soon preferred to base their ethics on other foundations. Shu is taken seriously in Shizi 尸子;51 this text, now extant

47 48 49

50 51

Davidson 2001. Goldin 2008b. For the sake of economy, I shall not repeat all the examples cited in that study. E.g., He Bingdi 1999; Chang 2007: 1.412n38; esp. M. Lewis 1990: 285n5. In Mark Edward Lewis’s view, military manuals were the first philosophical texts in China and indeed launched the philosophical burgeoning of the Warring States period. Another problem is that the Sunzi refers to crossbows and triggers (Pines 2002: 703). Goldin 2011a: 15ff. Qunshu zhiyao 36.625.

104

Goldin

only in fragments, is attributed to Shi Jiao 尸佼 (alive in 338 BCE),52 but not enough is known about it to permit strong inferences.53 The second telltale philosophical term is dao 道. When it appears as a noun in the Analects (in other words, not in the verbal sense of “to lead, to guide”), dao means something like “right or exemplary conduct.” It is not the cosmological concept known from texts such as Laozi and Xunzi. The main difference between dao in Laozi and Xunzi is that dao generates Heaven in the former and is dictated by Heaven in the latter; a related difference is that dao in Laozi has no inherent moral orientation, whereas for Xunzi, dao is the bedrock of morality itself. But otherwise, the usage is similar: dao is the inalterable structure of the universe, which we should learn to perceive and apply in our daily lives, rather than struggling bootlessly against it.54 When Neo-Confucians read the Analects in the Middle Ages, they assumed that the word dao had the same connotations in Confucius’s idiom as it did in their own. Thus, when Confucius says, “If one hears dao in the morning, it is acceptable to die in the evening” (朝聞道,夕死可矣; Analects 4/8), Zhu Xi’s commentary reads, “Dao is the principle by which things are as they should be” (道者,事物當然之理).55 That may fit Zhu Xi’s naturalistic metaphysics, but it is anachronistic. There is no attempt in the Analects (or in Mencius, it should be noted) to relate dao to natural processes; indeed, the Confucius of the Analects espouses no cosmological theory of any complexity. The closest he comes to making a statement about the cosmos is the claim—again in chapter 17— that Heaven speaks through the regular progression of the seasons (Analects 17/19). Michael J. Puett, among others, has argued that systematic cosmologies emerged only in the late Warring States period;56 thus, it stands to reason that Confucius’s discourse is free of them.

Evidence from References to Other Philosophers

References to other philosophers within various philosophical texts provide the framework for an overall chronology. Examining who responds to whom, and who cites which precedents, is generally an undervalued technique in the study of early Chinese intellectual history. For example, I think it can put to 52 53 54 55 56

Qian 1956: §90. Fischer (2012: e.g., 5, 12, 42) takes it as an authentic work by Shi Jiao. Goldin 2011a: 81ff. This account is terser, but perhaps more precise, than Goldin 1999. See also the analysis of “foundational naturalism” in Peerenboom 1993. Zhu Xi 1983: 2.71. Puett 2002: 145–200.

Confucius And His Disciples In The Lunyu

105

rest the nagging question “What is Chinese philosophy?” over which much ink has been spilled in recent years. In most twentieth-century scholarship, “Chinese philosophy” corresponded more or less to the Chinese bibliographical category of zi 子 or zhuzi 諸子, “masters.” (Sometimes texts more strictly classified as jing 經, “canons,” were included as well, inasmuch as the Analects and Mencius are recognized as two of the Thirteen Canons, and other jing, such as the Zuozhuan, also contain considerable philosophical material.) More recently, some scholars have questioned the practice of delimiting such inquiry according to traditional Chinese classificatory schemes. Donald Harper has used the phrase “natural philosophy” to refer to traditions involving medicine, divination, exorcism, and other practices that he terms “occult,” arguing that “philosophy” need not be restricted to the elevated “Masters” literature.57 Others have asked whether “Chinese philosophy” refers to any coherent subject, and whether the very concept is hopelessly polluted by Western philosophical presuppositions.58 One glance at the personages whom Xunzi mentioned by name (usually in the process of refuting them) will reveal the contours of what he considered his field of study: most prominently, Confucius, Mo Di, Mencius, Laozi, Zhuangzi, Sunzi, and Shen Dao—in other words, all the names that would appear in a textbook of “Chinese philosophy.” He has hardly anything to say about medicine, divination, or exorcism and does not, as far as I can recall, name a single “occult” specialist. Thus, even if one is reluctant, for whatever circumspect methodological reason, to call Xunzi’s enterprise “philosophy,” there can be little doubt that Masters literature was regarded in its day as constituting a coherent subject. Such references can also imply a sequence of debate: Xunzi responded to Mencius; Mencius did not respond to Xunzi. (There are similar phenomena in Western philosophy: Aristotle responded to Plato; Plato did not respond to Aristotle.)59 And here we find impressive evidence supporting the traditional chronology. Consider table 4.1: It must be borne in mind that the texts in this chart were not composed at a single moment in time, and thus, there is considerable chronological overlap among them. One aspect of the traditional view that must be revised is its ten57 58 59

Harper 1999: 813ff. Defoort 2001 and the response in Raud 2006; see also Denecke 2010: 1–31, 326–346. Van Norden (1996) has defended a conception of Chinese philosophy, along with a consideration of its value. Taylor 1969: 100–105, 116–138.

106

Goldin

Table 4.1 Mentions of Masters in “Masters” texts

Text

Figures cited

Analects Mozi Mencius Zhuangzi Xunzi

(none below) Confuciusa (and none below) Confucius,b Mo Di,c Song Kengd 宋牼 (and none below) Confucius, Mo Di,e Song Keng,f Laozi,g Shen Daoh (and none below)i Confucius,j Mo Di,k Mencius,l Song Keng,m Laozi,n Shen Dao,o Zhuangzi,p Sunziq

a Most famously in “Fei Ru xia” 非儒下 (Wu 2006: 9.39.428–433), but see also “Gengzhu” (Wu 2006: 11.46.643; alluding to the dialogue known from Analects 13/16) and “Gongmeng” 公 孟 (Wu 2006: 12.48.692). b Too many instances to cite singly. c Mencius 3A.5, 3B.9, 7A.26, 7B.26. d Mencius 6B.4. e The compound Ru-Mo 儒墨 appears as many as seven times in the Zhuangzi; for specific references to Confucius, see Littlejohn 2010. f “Tianxia” 天下 (Guo 1961: 10B.33.1082). g Again, too many instances to cite singly. h “Tianxia” (Guo 1961: 10B.33.1086–1088). i Mencius is not named in Zhuangzi, but there are allusions to phrases in the received Mencius, e.g., “the endpoints of humanity and righteousness” (renyi zhi duan 仁義之端), which is famous from Mencius 2A.6 (“Qiwu lun” 齊物論; Guo 1961: 1B.2.93). j Many instances, e.g., “Jiebi” 解蔽 (Wang Xianqian 1988: 15.21.393–394). k Again, many instances; one of the harshest is “Yuelun” 樂論 (Wang Xianqian 1988: 14.20.379ff.). l Primarily in “Xing’e” 性惡 (e.g., Wang Xianqian 1988: 17.23.439). m “Fei shi’er zi” 非十二子 (Wang Xianqian 1988: 3.6.92); “Tianlun” 天論 (Wang Xianqian 1988: 11.17.319); “Zhenglun” 正論 (Wang Xianqian 1988: 12.18.340–345); and “Jiebi” (Wang Xianqian 1988: 15.21.392). n “Tianlun” (Wang Xianqian 1988: 11.17.319). o “Fei shi’er zi” (Wang Xianqian 1988: 3.6.93); “Tianlun” (Wang Xianqian 1988: 11.17.319); “Jiebi” (Wang Xianqian 1988: 15.21.392). p “Jiebi” (Wang Xianqian 1988: 15.21.392). q Obliquely in “Yibing” 議兵 (Wang Xianqian 1988: 10.15.265–266); see Knoblock 1988–1994: 2.331n4.

dency to associate single texts with single authors;60 recent work has shown that this is an unlikely scenario for any text from before the time of Xunzi and 60

A tendency that goes back for centuries, incidentally; witness Sima Qian’s (Shiji 74.2343) assertion that Mencius wrote the Mencius, a claim that few Western scholars would take seriously (e.g., Hunter 2014: 58–74, in contrast to Dong 1997: 151–154).

Confucius And His Disciples In The Lunyu

107

Han Fei 韓非 (d. 233 BCE)—and even they did not produce the editions of the texts that bear their names today.61 Zhuangzi, in particular, contains material that must range across a couple of centuries.62 Nevertheless, the overall pattern is not easily reconciled with the notion that the Analects includes material from long after the time of Confucius and his disciples.63 If that were the case, we would expect to find references to the same later personages mentioned in the other texts. In this vein, it is also worth repeating Arthur Waley’s (1889–1966) observation that “the picture of Confucius given in the Analects … differs from that of all other books in that it contains no elements that bear patently and obviously the stamp of folk-lore or hagiography.”64 Han dynasty representations of Confucius have him performing all sorts of superhuman feats (such as lifting portcullises with his bare hands), but such fantasies are nowhere to be found in the Analects.65



There are two general hypotheses that would account for the evidence presented in this study. The first is what might be considered an updated version of the traditional understanding: the Analects consists largely, if not almost exclusively, of sayings and discussions from the time of Confucius and his disciples. The second would be that the text derives from a later period and was craftily composed so as to avoid embarrassing anachronisms. Readers must decide for themselves which hypothesis they find more plausible, but I shall conclude with some reasons for doubting the second. The notion that the redactors of the Analects would have avoided anachronisms is anachronistic in itself. Early Chinese literature is filled with anachronisms; they are concomitants of a historiography that prized moral accuracy 61 62 63

64 65

Fischer 2008–2009; see also M. Lewis 1999: 57ff. For an authoritative textual history of the Xunzi, see Gao 2010. For Han Feizi, the best studies remain Zheng Liangshu 1993 and Lundahl 1992. For a recent criticism of the commonplace understanding that the so-called “Inner Chapters” (neipian 内篇) represent an authentic core of the Zhuangzi, see Klein 2010. It is equally problematic for the so-called “accretion theory” advanced in Brooks and Brooks 1998, where the methodology behind their novel dates is laid out in appendix 1 (201–248). Their reasoning is fallacious: while they are correct that different chapters display different themes and styles, Brooks and Brooks seem to assume that the only way to explain such phenomena is to postulate a unique date of composition for each chapter. Waley 1938: 14. “Zhushu” 主術 (Liu 1989: 9.312). The classic statement of this problem is Gu 1982. For a recent overview of the various images of Confucius in later eras, see Nylan and Wilson 2010.

108

Goldin

over factual accuracy.66 More important, craftily compiled texts that purport to be relics from much older times still betray their engagement with the intellectual world into which they were released. As Alexander Campbell (1788– 1866) wrote of The Book of Mormon: This prophet Smith, through his stone spectacles, wrote on the plates of Nephi, in his book of Mormon, every error and almost every truth discussed in New York for the last ten years. He decides all the great controversies;—infant baptism, ordination, the trinity, regeneration, repentance, justification, the fall of man, the atonement, transubstantiation, fasting, penance, church government, religious experience, the call to the ministry, the general resurrection, eternal punishment, who may baptize, and even the question of freemasonry, republican government, and the rights of man. All these topics are repeatedly alluded to.67 Examples can be readily drawn from Chinese history as well. Most modern scholars consider the Liezi 列子 to be a spurious product of the Six Dynasties and not the text by the same name listed in the Hanshu 漢書 bibliography.68 Not surprisingly, recent work has been able to show how the Liezi would have resonated with literati of that period.69 Similarly, Wang Su 王肅 (195–256) claimed in his preface to The Family Sayings of Confucius (Kongzi jiayu 孔子家 語)—a work now widely thought to have been doctored by Wang himself— that it was presented to him by one of Confucius’s descendants, and that he considered it urgently necessary to publish the text, as it confirmed many of the positions that he had taken against the followers of Zheng Xuan 鄭玄 (127– 200).70 Thus, if it did not speak to the intellectual concerns of the early third century, School Sayings of Confucius would never have appeared; a less charitable judgment would be that Wang simply cobbled it together to provide doctrinal cover for his opinions. This subtle inability of questionable texts to escape their milieu is discernible even in works that contain genuine older 66 67 68

69 70

Goldin 2008a. Campbell 1831: 93. The presence of Greek names such as Timothy (3 Nephi 19:4) is sufficient for most nonbelievers to doubt the orthodox account of The Book of Mormon. For more on Campbell’s critique of The Book of Mormon, see, e.g., Gutjahr 2012: 42ff. The two most influential opinions have been those of Ma Xulun 1925 (a summary of which appears in Yang Bojun 1979: 301–305) and Graham 1986: 216–282. In the West, the case for the authenticity of Liezi is now usually taken to be hopeless, but several Chinese scholars still defend it (e.g., Ma Da 2000). Yuan 2005; Dippmann 2011. Kramers 1950: 91ff.; Goldin 1999: 138n53.

Confucius And His Disciples In The Lunyu

109

material within them. After all, much of the Liezi, and indeed almost the entirety of School Sayings, consist of passages with parallels in less controversial texts. When we turn to the Confucian Analects, by contrast, we see very little that would have been germane to the primary philosophical questions of the early Han. There is no discussion of the structure of the universe, the imperial mandate, the science of gauging Heaven’s intention on the basis of signs in the natural world, or the proper ritual and sumptuary distinctions for the various ranks in the social hierarchy—which were, to judge from the works of Lu Jia 陸 賈 (ca. 228–ca. 140 BCE), Jia Yi 賈誼 (201–169 BCE), and Dong Zhongshu 董仲舒 (ca. 179–ca. 104 BCE), the issues that excited philosophers at the highest level.71 By contrast, one of the most overt concerns of the Analects is to discourage the adoration of ghosts and spirits because they provide no useful moral guidance.72 This viewpoint would surely have aroused controversy during the period when Confucius lived, that is, the transition from Bronze Age religion to the more secular worldviews suited to the teeming society of the Eastern Zhou. But in the Han, it might almost have seemed quaint. This difficulty should remind us that any theory of the compilation of the Analects is incomplete without an account of the purpose. People do not compose, record, and then preserve texts without some reason that, in their minds, justifies the effort. My suggestion is that the Analects was recorded to prevent Confucius’s teachings from disappearing at a time when thoroughgoing social and political changes threatened old-fashioned methods of transmission, such as oral transmission from master to disciple. The survival of ideas in this new context could not be guaranteed if they were not written down. Mark Csikszentmihalyi has gathered compelling evidence to show that the Analects was “closely associated with the office of grand tutor to the Heir Apparent for much of the Western Hàn,” and it is a plausible hypothesis that the book was produced at that moment for use in the education of imperial princes.73 But the weight of the evidence suggests that whoever was responsible for compiling this textbook included an overwhelming proportion of genuine material within it.

71 72 73

Goldin 2007. Goldin 2011a: 13–14. Csikszentmihalyi 2002: 149.

110

Goldin

Epilogue

Readers will have noticed that the foregoing is predicated on certain convictions about the nature of ideas: they have distinct identities; they are not interchangeable; and, above all, they matter to people who hold or oppose them. I object to the notion, occasionally expressed in this volume and abundantly beyond, that because our knowledge of the history of this or that ancient text remains imperfect, we cannot permit ourselves to say anything definite about the philosophy that it advances, or that attempting to take its ideas seriously can only result in a naïve reformulation of traditionalist prejudices. To be sure, texts are continually interpreted and reinterpreted, and no text that gains an appreciable audience can have the same meaning for every reader. (“Jamais deux personnes n’ont lu le même livre, ni regardé le même tableau,” according to Mme. Swetchine.) But I will always be fundamentally unpersuaded by an argument about a philosophical text that does not account for the philosophy.

Bibliography

Allan, Sarah, and Crispin Williams, eds. 2000. The Guodian Laozi: Proceedings of the International Conference, Dartmouth College, May 1998. Berkeley: Society for the Study of Ancient China and Institute of East Asian Studies, University of California. Brooks, E. Bruce, and A. Taeko Brooks. 1998. The Original Analects: Sayings of Confucius and His Successors. New York: Columbia University Press. Campbell, Alexander. 1831. “Delusions.” Millennial Harbinger 2.2 (February 7). Chang, Chun-shu. 2007. The Rise of the Chinese Empire. Ann Arbor: University of Michigan Press. Chen Qiyou 陳奇猷. 2002. Lüshi chunqiu xin jiaoshi 呂氏春秋新校釋. Shanghai: Shang­ hai guji chubanshe. Cook, Scott. 2012. The Bamboo Texts of Guodian: A Study and Complete Translation. Ithaca, NY: Cornell University East Asia Program. Creel, H. G. 1949. Confucius: The Man and the Myth. New York: John Day. Csikszentmihalyi, Mark. 2002. “Confucius and the Analects in the Hàn.” In Confucius and the Analects: New Essays, ed. Bryan W. Van Norden, 134–162. Oxford: Oxford University Press. Csikszentmihalyi, Mark. 2004. Material Virtue: Ethics and the Body in Early China. Leiden: Brill. Davidson, Arnold I. 2001. The Emergence of Sexuality: Historical Epistemology and the Formation of Concepts. Cambridge, MA: Harvard University Press.

Confucius And His Disciples In The Lunyu

111

Defoort, Carine. 2001. “Is There Such a Thing as Chinese Philosophy? Arguments of an Implicit Debate.” Philosophy East and West 51.3:393–413. Denecke, Wiebke. 2010. The Dynamics of Masters Literature: Early Chinese Thought from Confucius to Han Feizi. Cambridge, MA: Harvard University Asia Center. Ding Sixin 丁四新. 2011. “Lun Mozi ‘Moyu’ Mojia houxue zhi guishenguan” 論《墨子・ 墨語》墨家後學之鬼神觀. Anhui Da xue xuebao (Zhexue shehui kexue ban) 安徽大 學學報(哲學社會科學版)2011.2:1–12. Dippmann, Jeffrey. 2011. “Reading the Zhuangzi in Liezi: Redefining Xianship.” In Riding the Wind with Liezi: New Perspectives on the Daoist Classic, ed. Ronnie Littlejohn and Jeffrey Dippmann, 151–164. Albany: State University of New York Press. Dong Hongli 董洪利. 1997. Mengzi yanjiu 孟子研究. Nanjing: Jiangsu guji. Durrant, Stephen W. 1977–1978. “A Consideration of Differences in the Grammar of the Mo tzu ‘Essays’ and ‘Dialogues.’” Monumenta Serica 33:248–267. Fang Shouchu 方授礎. 1937. Moxue yuanliu 墨學源流. Shanghai: Zhonghua shuju. Fischer, Paul. 2008–2009. “Authentication Studies (辨偽學) as Methodology and the Polymorphous Text Paradigm.” Early China 32:1–43. Fischer, Paul, trans. and ed. 2012. Shizi: China’s First Syncretist. New York: Columbia University Press. Forke, Alfred. 1922. Mê Ti des Sozialethikers und seiner Schüler philosophischen Werke. Berlin: Mitteilungen des Seminars für Orientalische Sprachen. Gao Zheng 高正. 2010. Xunzi banben yuanliu kao 荀子版本源流考. 2nd ed. Beijing: Zhonghua shuju. Goldin, Paul Rakita. 1999. Rituals of the Way: The Philosophy of Xunzi. La Salle, IL: Open Court. Goldin, Paul R. 2003. Review of Chuang-tzŭ: The Inner Chapters, trans. A. C. Graham. Early China 28:201–214. Goldin, Paul R. 2005. After Confucius: Studies in Early Chinese Philosophy. Honolulu: University of Hawai‘i Press. Goldin, Paul R. 2007. “Xunzi and Early Han Philosophy.” Harvard Journal of Asiatic Studies 67.1:135–166. Goldin, Paul R. 2008a. “Appeals to History in Early Chinese Philosophy and Rhetoric.” Journal of Chinese Philosophy 35.1:79–96. Goldin, Paul R. 2008b. “When zhong Does Not Mean ‘Loyalty.’” Dao 7.2:165–174. Goldin, Paul R. 2011a. Confucianism. Berkeley: University of California Press. Goldin, Paul R. 2011b. “Steppe Nomads as a Philosophical Problem in Classical China.” In Mapping Mongolia: Situating Mongolia in the World from Geologic Time to the Present, ed. Paula L. W. Sabloff, 220–246. Philadelphia: University of Pennsylvania Museum of Archaeology and Anthropology. Graham, A. C. 1978. Later Mohist Logic, Ethics and Science. Hong Kong: Chinese Uni­ versity Press.

112

Goldin

Graham, A. C. 1986. Studies in Chinese Philosophy and Philosophical Literature. Singa­ pore: Institute of East Asian Philosophies. Graham, A. C. 1989. Disputers of the Tao: Philosophical Argument in Ancient China. La Salle, IL: Open Court. Gu Jiegang 顧頡剛. 1982. “Chunqiu shi de Kongzi he Handai de Kongzi” 春秋時的孔子 和漢代的孔子. Reprint. In Gushi bian 古史辨, vol. 2, ed. Gu Jiegang 顧頡剛 et al., 130–139. Shanghai: Shanghai guji chubanshe. Guo Qingfan 郭慶藩. 1961. Zhuangzi jishi 莊子集釋. Ed. Wang Xiaoyu 王孝魚. Beijing: Zhonghua shuju. Gutjahr, Paul C. 2012. The Book of Mormon: A Biography. Princeton, NJ: Princeton Uni­ versity Press. Harbsmeier, Christoph. 2004. “The Semantics of qíng in Pre-Buddhist Chinese.” In Love and Emotions in Traditional Chinese Literature, ed. Halvor Eifring, 69–148. Leiden: Brill. Harper, Donald. 1999. “Warring States Natural Philosophy and Occult Thought.” In The Cambridge History of Ancient China, ed. Michael Loewe and Edward L. Shaughnessy, 813–884. Cambridge: Cambridge University Press. He Bingdi 何炳棣. 1999. “Zhongguo xiancun zui gu de sijia zhushu Sunzi bingfa” 中國 現存最古的私家著述《孫子兵法》. Lishi yanjiu 歷史研究 1999.5:72–89. He Hao 何浩. 1994. “Luyang jun, Luyang gong ji Luyang shexian de wenti” 魯陽君、魯 陽公及魯陽設縣的問題. Zhongyuan wenwu 中原文物 1994.4:46–51. Hsiao, Kung-chuan. 1979. A History of Chinese Political Thought, vol. 1, From the Be­ ginnings to the Sixth Century A.D. Trans. F. W. Mote. Princeton, NJ: Princeton University Press. Hunter, Michael. 2014. “Did Mencius Know the Analects?” T’oung Pao 100.1–3:33–79. Jiang Boqian 蔣伯潛. 1985. Zhuzi tongkao 諸子通考. Hangzhou: Zhejiang guji chubanshe. Jiao Xun 焦循. 1987. Mengzi zhengyi 孟子正義. Ed. Shen Wenzhuo 沈文倬. Beijing: Zhonghua shuju. Kanaya Osamu 金谷治. 1997. Kanaya Osamu Chūgoku shisō ronshū 金谷治中國思想論 集. Tokyo: Hirakawa. Klein, Esther. 2010. “Were There ‘Inner Chapters’ in the Warring States? A New Examination of Evidence about the Zhuangzi.” T’oung Pao 96.4–5:299–369. Knoblock, John. 1988–1994. Xunzi: A Translation and Study of the Complete Works. Stanford, CA: Stanford University Press. Knoblock, John, and Jeffrey Riegel. 2000. The Annals of Lü Buwei. Stanford, CA: Stanford University Press. Kramers, R. P. 1950. K’ung tzu chia yü: The School Sayings of Confucius. Leiden: E. J. Brill. Kuhn, Thomas S. 1957. The Copernican Revolution: Planetary Astronomy in the Development of Western Thought. Cambridge, MA: Harvard University Press.

Confucius And His Disciples In The Lunyu

113

Lau, D. C., trans. 1992. Confucius: The Analects. 2nd ed. Hong Kong: Chinese University Press. Lau, D. C., trans. 2003. Mencius: A Bilingual Edition. Revised ed. Hong Kong: Chinese University Press. Lewis, Clarence Irving. 1929. Mind and the World Order: Outline of a Theory of Knowledge. New York: Charles Scriber’s Sons. Lewis, Mark Edward. 1990. Sanctioned Violence in Early China. Albany: State University of New York Press. Lewis, Mark Edward. 1999. Writing and Authority in Early China. Albany: State University of New York Press. Li, Charles N. 1996. “A Cryptic Language with a Minimal Grammar: The Confucian Analects of Late Archaic Chinese.” In Lexical Structures and Language Use: Proceedings of the International Conference on Lexicology and Lexical Semantics, Münster, September 13–15, 1994, ed. Edda Weigand and Franz Hundsnurscher, 1:53– 118. Beiträge zur Dialogforschung 9. Tübingen: Niemeyer. Li Ling 李零. 2008. Jianbo gushu yu xueshu yuanliu 簡帛古書與學術源流. Revised ed. Beijing: Sanlian. Li Xueqin 李學勤. 2009. Zhongguo gudai wenming yanjiu 中國古代文明研究. Shanghai: Huadong Shifan Da xue chubanshe. Littlejohn, Ronnie. 2010. “Kongzi in the Zhuangzi.” In Experimental Essays on Zhuangzi, 2nd ed., ed. Victor H. Mair, 177–194. Dunedin, FL: Three Pines. Liu Wendian 劉文典. 1989. Huainan Honglie jijie 淮南鴻烈集解. Ed. Feng Yi 馮逸 and Qiao Hua 喬華. Beijing: Zhonghua shuju. Luan Tiaofu 欒調甫. 1957. Mozi yanjiu lunwenji 墨子研究論文集. Beijing: Renmin chubanshe. Lundahl, Bertil. 1992. Han Fei Zi: The Man and the Work. Stockholm: Institute of Oriental Languages, Stockholm University. Luo Genze 羅根澤. 1958. Zhuzi kaosuo 諸子考索. Beijing: Renmin chubanshe. Luo Shaodan. 2003. “Inadequacy of Karlgren’s Linguistic Method as Seen in Rune Svarverud’s Study of the Xinshu.” Journal of Chinese Linguistics 31.2:270–299. Ma Da 馬達. 2000. Liezi zhenwei kaobian 《列子》真僞考辨. Beijing: Beijing chubanshe. Ma Xulun 馬敍倫. 1925. Tianma shanfang congzhu 天馬山房叢著. N.p. Makeham, John. 1996. “The Formation of Lunyu as a Book.” Monumenta Serica 44:1–24. Makeham, John. 2006. “A New Hermeneutical Approach to Early Chinese Texts: The Case of the Analects.” In Hermeneutical Thinking in Chinese Philosophy, ed. Lauren Pfister, 95–108. Journal Supplement Series to the Journal of Chinese Philosophy. Malden, MA: Blackwell. Mei, Yi-pao, trans. 1929. The Ethical and Political Works of Motse. London: Probsthain.

114

Goldin

Mei, Yi-pao. 1934. Motse: The Neglected Rival of Confucius. London: Probsthain. Nylan, Michael, and Thomas Wilson. 2010. Lives of Confucius: Civilization’s Greatest Sage through the Ages. New York: Doubleday. Okamoto Mitsuo 岡本光生. 1992. “Bokushi ni okeru ‘zai’ no kōkan” 「墨子」における 「財」の交換. Tōyō no shisō to shūkyō 東洋の思想と宗教 9:17–30. Peerenboom, R. P. 1993. Law and Morality in Ancient China: The Silk Manuscripts of Huang-Lao. Albany: State University of New York Press. Pines, Yuri. 2002. “Lexical Changes in Zhanguo Texts.” Journal of the American Oriental Society 122.4:691–705. Pines, Yuri. 2005–2006. “Bodies, Lineages, Citizens, and Regions: A Review of Mark Edward Lewis’ The Construction of Space in Early China.” Early China 30:155–188. Puett, Michael J. 2002. To Become a God: Cosmology, Sacrifice, and Self-Divinization in Early China. Cambridge, MA: Harvard University Asia Center. Puett, Michael. 2004. “The Ethics of Responding Properly: The Notion of qíng in Early Chinese Thought.” In Love and Emotions in Traditional Chinese Literature, ed. Halvor Eifring, 37–68. Leiden: Brill. Qian Mu 錢穆. 1930. Mozi 墨子. Wanyou wenku. Shanghai: Shangwu yinshuguan. Qian Mu 錢穆. 1956. Xian-Qin zhuzi xinian 先秦諸子繫年. 2nd ed. Hong Kong: Hong Kong University Press. Qunshu zhiyao 群書治要. 1936. By Wei Zheng 魏徵 (580–643). Shanghai: Shangwu yinshuguan. Raud, Rein. 2006. “Philosophies versus Philosophy: In Defense of a Flexible Definition.” Philosophy East and West 56.4:618–625. Roth, Harold D., trans. 1999. Original Tao: Inward Training and the Foundations of Taoist Mysticism. New York: Columbia University Press. Shaughnessy, Edward L. 2006. Rewriting Early Chinese Texts. Albany: State University of New York Press. Shiji 史記. 1959. Beijing: Zhonghua shuju. Slingerland, Edward. 2000. “Why Philosophy Is Not ‘Extra’ in Understanding the Analects.” Philosophy East and West 50.1:137–141. Stumpfeldt, Hans. 2010. “Thinking beyond the ‘Sayings’: Comments about Sources concerning the Life and Teachings of Confucius (551–479).” Oriens Extremus 49:3–27. Sun Daosheng 孫道昇. 1982. “Wumazi yu Yang Zhu xuepai de guanxi” 巫馬子與楊朱學 派的關係. In Gushi bian 古史辨, vol. 6, ed. Gu Jiegang 顧頡剛 et al., 193–197. Shanghai: Shanghai guji chubanshe. Sun Xidan 孫希旦. 1989. Liji jijie 禮記集解. Ed. Shen Xiaohuan 沈嘯寰 and Wang Xingxian 王星賢. Beijing: Zhonghua shuju. Taiping yulan 太平御覽. Sibu congkan 四部叢刊 ed. Tan Jiajian 譚家健. 1995. Mozi yanjiu 墨子研究. Guiyang: Guizhou jiaoyu chubanshe. Taylor, A. E. 1969. Aristotle on His Predecessors. 2nd ed. La Salle, IL: Open Court.

Confucius And His Disciples In The Lunyu

115

Tong Shuye 童書業. 2005. Tong Shuye shiji kaozheng lunji 童書業史籍考證論集. Ed. Tong Jiaoying 童教英. Beijing: Zhonghua shuju. Van Norden, Bryan W. 1996. “What Should Western Philosophy Learn from Chinese Philosophy?” In Chinese Language, Thought, and Culture: Nivison and His Critics, ed. Philip J. Ivanhoe, 224–249. La Salle, IL: Open Court. Waley, Arthur, trans. 1938. The Analects of Confucius. London: George Allen and Unwin. Wang Dongzhen 王冬珍. 1981. Moxue xintan 墨學新探. Taipei: Shijie shuju. Wang Xianqian 王先謙. 1988. Xunzi jijie 荀子集解. Ed. Shen Xiaohuan 沈嘯寰 and Wang Xingxian 王星賢. Beijing: Zhonghua shuju. Watanabe Takashi 渡邊卓. 1973. Kodai Chūgoku shisō no kenkyū—“Kōshiden no keisei” to Ju Boku shūdan no shisō kōdō 古代中國思想の研究—〈孔子傳の形成〉と儒墨 集團の思想行動. Tokyo: Sōbunsha. Weingarten, Oliver. 2009. “Confucius and Pregnant Women: An Investigation into the Intertextuality of the Lunyu.” Journal of the American Oriental Society 129.4:597– 618. Weingarten, Oliver. 2010. “Textual Representations of a Sage: Studies of Pre-Qin and Western Han Sources on Confucius (551–479 BCE).” PhD diss., Cambridge University. Wu Yujiang 吳毓江. 2006. Mozi jiaozhu 墨子校注. Ed. Sun Qizhi 孫啓治. 2nd edition. Beijing: Zhonghua shuju. Yang Bojun 楊伯峻. 1979. Liezi jishi 列子集釋. Beijing: Zhonghua shuju. Yang Junguang 楊俊光. 1992. Mozi xinlun 墨子新論. Nanjing: Jiangsu jiaoyu chu­banshe. Yoshinaga Shinjirō 吉永慎二郎. 2004. Sengoku shisōshi kenkyū: Juka to Bokuka no shisōshiteki kōshō 戰國思想史研究:儒家と墨家の思想史的交涉. Kyoto: Hōyū. Yuan Jixi 袁濟喜. 2005. “Liezi yu Liuchao wenshi de yansheng” 《列子》與六朝文士的 演生. Zhongguo Renmin Da xue xuebao 中國人民大學學報 2005.6:121–128. Zhang Chunyi 張純一. (1936) 1975. Mozi jijie 墨子集解. In Wuqiubeizhai Mozi jicheng 無 求備齋墨子集成, ed. Yan Lingfeng 嚴靈峰. Taipei: Chengwen chubanshe. Zheng Jiewen 鄭傑文. 2006. Zhongguo Moxue tongshi 中國墨學通史. Beijing: Renmin chubanshe. Zheng Liangshu鄭良樹. 1993. Han Fei zhi zhushu ji sixiang 韓非之著書及思想. Taipei: Xuesheng shuju. Zhu Weizheng 朱維錚. 2002. Zhongguo jingxueshi shijiang 中國經學史十講. Shanghai: Fudan Da xue chubanshe. Zhu Xi 朱熹. 1983. Sishu zhangju jizhu 四書章句集注. Beijing: Zhonghua shuju.

116

Gentz

Chapter 5

The Lunyu, a Homeless Dog in Intellectual History: On the Dating of Discourses on Confucius’s Success and Failure Joachim Gentz Who quotes the Lunyu in academic discourse and why? Intellectual historians regard Confucius as “the first known ideologically active member of the shi stratum”1 or even as “the most influential thinker in Chinese civilization and the first whose philosophy can be reconstructed to any significant degree,”2 and they view Confucianism accordingly as “the earliest of the competing tendencies in the thought of ancient China”3 and “China’s oldest and most revered philosophy.”4 The Lunyu is still one of the major sources used to reconstruct Confucius’s thought, to interpret early Confucianism, and to construct early Chinese intellectual history.5 Using the Lunyu as a source in intellectual history means to assume that the Lunyu is part of an early intellectual discourse and even an important reference point for defining the beginning of Chinese philosophy. This assumption has been called into question as the dating of the Lunyu has become increasingly controversial.6 In an article that summarizes some of the main arguments of this controversy, Csikszentmihalyi and Kim conclude that in the light of recent archaeological finds “the Analects changes from a text about Confucius to a text that preserves the early textual practices of generations of writers that may include Confucius.” “It is hardly a reflection of Confucius’s life and times, but instead represents a complex space occupied by a number of people who claimed to be his spiritual or ethical followers— people who possessed divergent, contrasting, and continually retouched portraits of him.”7

1 2 3 4 5 6

Pines 2012: 78. Goldin 2011: 7. Graham 1989: 31. Goldin 2011: 1. Hunter 2012: 1n4. See the critical analyses in Makeham 1996; Simson 2006; Hunter 2012. For further references, see Hunter 2014: 42n41. 7 Csikszentmihalyi and Kim 2014: 163–164.

© Koninklijke Brill NV, Leiden, 2018 | doi:10.1163/9789004382947_007

The Lunyu, A Homeless Dog In Intellectual History

117

To critically reflect upon the practice of taking the Lunyu as a source for constructing intellectual history, this chapter will follow a hypothesis that reverses this relationship: if the Lunyu can be taken as a source that relates to intellectual history, then intellectual history (as reliably as it can be reconstructed in the light of the constant new excavated sources) in turn can be taken to date the Lunyu. This chapter will thus ask what happens if we continue to follow the general assumption of the Lunyu as an intellectual locus and try to locate it in intellectual history. To use intellectual history to date texts, the following five conditions have to be fulfilled: 1. An intellectual history will focus on one or more elements that might include simple concepts, ideas, thoughts, terms, metaphors, or more complex discourses or problems. According to one’s analytical focus, one might pursue the “history of concepts,” “history of ideas,” “history of thoughts,” “history of discourses,” “history of problems,” “history of terms,” or “history of metaphors.” 2. These elements must be consistent and related to a coherent position or perspective at a given time. 3. These elements must change over time. 4. These elements must be datable and amenable to historical contextualization so that their shifts or changes can be conceptualized in a historical framework. 5. Something of the elements must remain identical amid these changes so that they can be related to and compared with similar elements in a different historical time such that we can write a “history of X.” In a “history of terms,” for example, the terms will remain the same even as their meanings change. The Lunyu is obviously full of concepts, ideas, thoughts, terms, metaphors, discourses, and problems that would seem to be amenable to intellectual historical analysis, thus fulfilling condition 1. Yet we find hardly any argumentation or discursive exposition of these elements, and references to intellectual discourses (such as Lunyu 2/3)8 remain so vague that no particular historical positions or arguments can be identified. All we find are short authoritative statements of the Master, evaluations and definitions that are not even 8 “The Master said, ‘If you guide them by governmental rules and order them by means of penal law, the people will seek to avoid these and become shameless; if you guide them by virtues and order them by means of rituals, they will know shame and be correct’” (子曰:道之以 政,齊之以刑,民免而無恥;道之以德,齊之以禮,有恥且格).

118

Gentz

consistent throughout the book. The purpose of the book is obviously not to take part in intellectual debates.9 Condition 2 thus seems not to be fulfilled. This will be further proven by looking at one concrete example of a central term, concept, and idea in the Lunyu.

Terms, Concepts, and Ideas in the Lunyu: Taking Ren 仁 as an Example

Historians of Chinese thought consider benevolence or humanity (ren 仁) as one of Confucius’s central ideas. To explain its meaning, they often cite a selection of Lunyu passages in which the word is defined and then provide their own interpretation.10 I would like to proceed similarly and present a few passages on ren to demonstrate, in contrast, that such an approach is meaningless when applied to the Lunyu. I will pick just a handful of central passages on ren from the first few books of the text: Youzi said, “There are very few who in their actions are filial toward their parents and respectful toward their elder brothers and yet like to offend their superiors. That someone who doesn’t like to offend his superiors would be fond of creating disorder has never happened. The gentleman regards the roots of things as fundamental. When the right root is established, then the right way grows out of it. Filial piety and respect toward elder brothers, isn’t this the root of ren?”11 有子曰:其為人也孝弟,而好犯上者,鮮矣;不好犯上,而好作亂 者,未之有也。君子務本,本立而道生。孝弟也者,其為仁之本與?

Let us reconstruct the line of thought here: filial and respectful people rarely offend their superiors and thus rarely cause disorder. This attitude is the basis of ren. Ren in this passage is therefore presented as rooted in a spirit of respect and obedience to superiors.

9 10 11

The principle in Lunyu 3/7 that “a gentleman has nothing to compete for” (君子無所爭) might reflect this. Even Hirase Takao (2016: 307–320), who in his recent book on ren interprets each Lunyu passage that uses the term, focuses on particular sociopolitical aspects of ren only. Note the wordplay here with the formulation 其為仁 in the last sentence echoing the first sentence’s 其為人.

The Lunyu, A Homeless Dog In Intellectual History

119

The Master said, “Skillful speech and insinuating appearance rarely go together with ren.”12 子曰:巧言令色,鮮矣仁!

Here, ren is opposed to artificial performance, so it means something like genuineness. It is unclear to me how this can be rooted in a spirit of obedience. The Master said, “Those who are not ren cannot endure hardship for long and cannot enjoy pleasure for long.13 Those who are ren have their home in ren; those who are wise have their benefit from ren.” 子曰:不仁者不可以久處約,不可以長處樂。仁者安仁,知者利仁。

In this passage ren is defined as something that is independent from outer circumstances such as hardship and pleasure. It can be positively used by wise people. This aspect of ren could be related (by the reader of the Lunyu) to the two aspects defined above, especially genuineness. Such a relation, however, is not established by the text. The Master said, “Only those who are ren are able to like people and to detest people.” 子曰:唯仁者能好人,能惡人。

 Ren is here related to the fair emotional judgment of others. The passage does not give any reason why only those who possess ren can judge others, but it seems that ren means something like impartiality, detachedness, neutrality, fair-mindedness, being unbiased, and the like. The Master said, “If your attention is directed to ren, then there is no wickedness.” 子曰:苟志於仁矣,無惡也。

Again, ren is related to e 惡 (wickedness), and this might be the reason why these two passages have been arranged next to each other. Yet the character 惡 12 13

The same saying (巧言令色,未可謂仁也) can be found in the *Dizi wen manuscript from the Shanghai Museum find (Zhang 2005: 274, strip 11). Note the similar sound of 約 (*[q](r)ewk) and 樂 (*[r]ˤ awk).

120

Gentz

(pronounced nominal “e” and not verbal “wu” as in the preceding passage) has a different meaning here. Ren means something like goodness (as opposed to wickedness). It does not connect easily to 4/3, as goodness is not the same as impartiality, and neither is it a sufficient condition for impartiality. The Master said, “Wealth and social status are what people desire. Yet one shouldn’t settle in them if this cannot be achieved by means of the right way. Poverty and low status are what people despise. Yet one should not do away with them if this cannot be achieved by means of the right way. If a gentleman gives up ren, how can he establish a name for himself? Not even for the time it takes to finish a meal will a gentleman turn against ren, because getting into unfavorable situations and falling into difficulties will certainly result from this.” 子曰: 富與貴是人之所欲也,不以其道得之,不處也;貧與賤是人之 所惡也,不以其道得之,不去也。君子去仁,惡乎成名?君子無終食 之間違仁,造次必於是,顛沛必於是。

 Ren is introduced in this passage as a true means to establish a name. This seems to contradict passages like 1/3, where ren is dissociated from attempts to please others, or 4/2, where ren is exactly the means to be independent from outer circumstances. In an entirely unconnected last sentence, the absence of ren is then associated with getting into unfavorable situations and falling into difficulties. Ren is thus a sense of keeping out of trouble and turning situations into favorable circumstances. This recalls the wise men of 4/2 who know how to use ren for their benefit. Looking at this sample of passages on ren in the Lunyu, it is quite obvious that it is an assemblage of unrelated and miscellaneous predications about ren. Ren is presented as a spirit of respect and obedience, genuineness, independence from outer circumstances, impartiality, goodness, or a sense of keeping out of trouble and turning situations into favorable circumstances. Is it possible to find a common ground among these passages that would provide us with some very general and basic meaning of ren, one that does equal justice to all six passages without being a mere accumulation of definitions? Within the scope of just two pages, A. C. Graham defined ren variously as “an unselfish concern for the welfare of others,” covering “like English ‘noble’ the whole range of superior qualities distinctive of the man of breeding,”14 “the orientation which makes right action effortless, following attainment of just 14

Graham 1989: 19.

The Lunyu, A Homeless Dog In Intellectual History

121

the right balance between self and other, a precarious balance which hardly anyone is able to sustain,” “the perfectly and permanently disinterested person,” “a matter of attuning the desires on behalf of self and others,” and, finally, “the instant in which you conquer self to see self and others in perfect proportion,” that being “an instant in which accord with conventions becomes effortless and the exercise of style within fixed forms is an uninterrupted flow.”15 To summarize, ren for Graham is an unselfish, disinterested attitude toward the self that, by attuning one’s own desires and concerning oneself with the welfare of others, creates a proper balance between self and others and thus makes right and conventional actions effortless. We see Graham’s struggle to pin down this complex notion not only in his use of an open-ended phrase like “the whole range of superior qualities” but also in his neglect of obedience (1/2), genuineness (1/3), favorable action (4/5), and other associations with ren in the Lunyu. To include these aspects within Graham’s definition would probably reduce the gentleman to a butler. Paul R. Goldin gives a similar definition: “Humanity (ren JG) … is the virtue based on the method of shu [恕]. … Shu is placing oneself in the position of others, and acting towards them as one imagines they would desire … by taking oneself as analogy.”16 Again, a number of elements highlighted above are missing in this definition. Roger Ames and Henry Rosemont, who translate ren as “authoritative conduct” and “authoritative person,” define it more openly as “one’s entire person: one’s cognitive, aesthetic, moral, and religious sensibilities as they are expressed in one’s ritualized roles and relationships.”17 They choose to define the term so broadly that it includes almost all aspects of a person, with the result that it cannot be used anymore with analytical precision. Likewise, Arthur Waley defines ren as “‘good’ in an extremely wide and general sense,”18 illustrating it with a selection of definitions of ren from the Lunyu. It is a typical and, given the nature of the Lunyu, necessary feature of intellectual historical analyses of the Lunyu that they never use all the available features of a concept or an idea, instead offering a selective account.19 That is because the many diverse definitions given in the Lunyu cannot be, and are 15 16 17 18 19

Graham 1989: 22. Goldin 2011: 19. Ames and Rosemont 1998: 49. Waley 1989: 28. The definitions of ren in the cited Lunyu passages above (and in thirty to forty other passages besides) occur in isolated form or within small clusters of two to seven thematically related passages. However, as shown by the examples above, this does not mean that the grouped passages are conceptually more related to one another than to passages on ren elsewhere in the Lunyu.

122

Gentz

therefore probably not meant to be, reduced to a less complex common ground even though a number of passages claim repeatedly, and perhaps quite desparately, that at least in the Master’s teaching there was one coherent line of thought (yi yan 一言 or yi yi guan zhi 一以貫之).20 The explanations of ren given in my small sample above show that we find no consistent meaning of the term ren, no common concept, idea, or even problem that would link these different propositions. This also applies to other terms, concepts, or ideas in the Lunyu. The book resembles a collection of tesserae, some of which have been loosely grouped together like the ren-clusters in 1/2–3, 4/1–7, 12/1–3, and 15/9–10 or the xiao 孝-cluster in 2/5–8. Intellectual historians have not only shaped these tesserae with their diverging interpretations but also used them highly selectively to reconstruct very different images of Confucius and his philosophy.21 One does not have to take a postmodern perspective to claim that the Lunyu is an assemblage of mutable tesserae that in their entirety do not (and were never meant to) amount to a full mosaic.22 Instead, they can only be (and have always been) used selectively to form an incomplete mosaic of fragments of diverse discourses. The Lunyu can thus be compared to a mosaic construction kit or a box of Lego bricks containing individual textual units that according to different interpretations can even change color and shape and have to be (and have always been) used as convertible modular building blocks to construct a variety of Confucianisms. The systematic incoherence of the book thus makes it difficult if not impossible to approach the Lunyu by means of an intellectual history analysis. The reason for this lack of coherence is that the primary purpose of the book is not to formulate a consistent position within an intellectual landscape but to portray the Master in as many face(t)s as were available at the time of its compilation. This raises the question of whether the Lunyu might have constructed a particular and consistent image of Confucius. If the Lunyu takes part in an intellectual discourse on the nature, character, personality, importance, or impact of Confucius and depicts him in a coherent way that is relatable to other texts, then conditions 1, 2, and 5 of the introductory catalog would be fulfilled. An intellectual history approach might then be used to date the Lunyu by locating its portrayal of Confucius in a history of portrayals of Confucius, a phenomenon that Hunter calls “Kongzigraphy.”23 20 21 22 23

Lunyu 2/2, 4/15, 13/15, 15/3. See, e.g., Harbsmeier 1990 and, for a different reconstruction, Wagner 1991, 2004. This reluctance to provide a clear image of Confucius is somehow mirrored in the later iconoclast critique against the production of Confucius images. See Sommer 2002; Murray 2001, 2009. Hunter 2012: 130.

The Lunyu, A Homeless Dog In Intellectual History



123

Portrayals of Confucius in the Lunyu: Taking Confucius’s Success and Failure as an Example

It has been noted by many earlier authors that the Lunyu Confucius is quite distinct from later depictions of the sage. Gu Jiegang suggested that we should study “one Confucius at a time.”24 Arthur Waley has noted that “the picture of Confucius given in the Analects … differs from that of all other books in that it contains no elements that bear patently and obviously the stamp of folklore or hagiography.”25 The manufacturing of the apotheosis of Confucius as a quasidivine being has been the object of many studies, and a historical development of Confucius portrayals is generally assumed in literature on Confucius.26 Yet this process is difficult to date because there is no clear line of development from a Confucius bare of any hagiographic elements to the divine sage of later ages. We rather see the appropriation, transformation, and invention of Confucius lore in Warring States and Han texts alongside constant constructions and reconstructions of Confucius tesserae.27 These pieces aim in the first place to support the arguments of the respective texts and certainly do not contribute to a complete and coherent image of Confucius. They gain their authority only by being recognizable as part of a discourse which I will call “Confucian” because it gains its identity solely by being related to Confucius. This discourse is intertextual, as it is not based on, or shaped by, one single text but constituted by a network of interrelated texts.28 The discourse is neither systematic nor unified but consists of single themes that are first loosely and later (by Sima Qian) more coherently linked to a biographical narrative. These themes can be regarded as individual and quite independent subdiscourses. Despite their changes over the course of Chinese intellectual history, these subdiscourses gain their identity and historical continuity by being related to specific problems that are held to be identical and continuous by participants 24

25 26

27 28

Gu 1988: 2.130: “The Confucius of the Chunqiu period was a gentleman, the Confucius of the Zhanguo period was a sage, the Confucius of the Western Han period was a pope, the Confucius of the Eastern Han period was again a sage, and now he is just about to turn into a gentleman again” (春秋时的孔子是君子,战国的孔子是圣人,西汉时的孔 子是教主,东汉后的孔子又成了圣人,到现在又快要成君子了). Waley 1989: 14. See, e.g., Nylan and Wilson 2010. As we know, this process is still going on. In 2006 an author with the (quite pretentious) pseudonym Sanren 三人edited a book titled Liushige Kongzi 六十个孔子 (Sixty Confuciuses) that provides an overview of over sixty different approaches to Confucius and includes essays by Hu Shi, Lu Xun, Lin Yutang, Gu Jiegang, Feng Youlan, Chen Duxiu, and others. See the very meticulous studies by Weingarten (2010) and Hunter (2012). Weingarten 2010.

124

Gentz

in the discourse.29 The identity of these problem-related discourses is marked by intertextual references to particular text passages, terms,30 utterances, and stories (or the historical protagonists who synecdochically represent those stories), all of which constitute—and are thus associated with—particular problems. The most adequate method to date these problem-related discourses is via a “history of problems.”31 I will therefore use this approach to date one of the central subdiscourses in the Lunyu, that of the problem of Confucius’s success and failure, which explores the stinging issue of why Confucius never held high office.32 It can be identified as a continuous topos in early Chinese intellectual history and appears so frequently in portrayals of Confucius that historical shifts can be determined and dated (which satisfies conditions 3 and 4). This will be attempted in the following section to prove whether the fourth condition can be fulfilled in our intellectual history analysis of the Lunyu. 29

30 31

32

I will define “discourse” here, following the approach by Busse and Teubert (1994), as a virtual text corpus consisting of Chinese texts that were probably written in the third century BCE by members of the Chinese intellectual elite and that deal with a particular problem (which I am going to define), refer to each other, and thus create an intertextual relationship. The unity of the discourse is delineated by my own definition of the particular problem(s) to which the texts relate. An early Chinese attempt to assemble the central key terms of Neo-Confucianist discourses in one book was Chen Chun’s 陳淳 (1159–1223) Beixi Zi yi 北溪字義 (translated in Chan 1986). I do not assume that the problems discussed in these subdiscourses were de facto relating to identical problems based in human reason as Windelband (1892) and Hartmann (1910: 469), the inventors of Problemgeschichte, believed, but rather, I follow Collingwood’s (1946: 283) and Gadamer’s (1990: 375–384) views that they change and are merely reconstructed within the historical awareness of the readers and their own investigative horizon. Following Gadamer’s harsh critique of Kant’s and Neo-Kantian scholars’ (such as Windelband and Hartmann) concept of problem history, which he saw as a “bastard of historism” (1990: 382), later scholars like Sgarbi have in their reconstruction of Problemgeschichte unduly simplified the Neo-Kantian positions and downplayed the important historical dimension found in the works of both Windelband and Hartmann. See Sgarbi 2010 and his very similar English article (2011). A more differentiated analysis can be found in Oexle 2001. A detailed analysis of the development of the main subdiscourses regarding Confucius is still outstanding, and such an analysis has never been used to date the Lunyu. One of the few works to analyze one of these subdiscourses has been Elstein 2006, where he analyzes depictions of Confucius’s authority and teacher-disciple relationship in the Lunyu (see also Elstein 2009). In an earlier article (Gentz 2012) I analyzed another central subdiscourse of the Lunyu, the discourse on Confucius’s knowledge, and compared it with other early texts. The only text that shows similar attitudes (not on a discursive level but in its exegetical practice) is the Gongyang zhuan 公羊傳. The evidence, however, does not suffice to date the Lunyu. See Weingarten 2014 for an analysis of the same subdiscourse and Weingarten 2010 for an analysis of further early textual traditions about Confucius from a literary point of view.

The Lunyu, A Homeless Dog In Intellectual History

125

The Confucian subdiscourse turns into a wider theoretical problem when it is broadened to relate to sages in a more general sense and can be used to discuss abstract questions about the relation between sageliness and taking office. One of the central messages of early historiographical and hagiographical accounts, perhaps starting with those of the Shangshu 尚書 (Exalted Documents) and the Zuozhuan 左傳 (Zuo Traditions), is the claim of empirical proof of the early Zhou ideology regarding the relationship between moral conduct and success in life. That virtuous behavior is the most powerful means to transform politics and society is an assumption that can be found in all texts connected to the Confucian tradition. Yet the biographies of the three most eminent founders of that tradition, Kong Qiu 孔丘 (Kongzi), Meng Ke 孟軻 (Mengzi; 371–289), and Xun Kuang 荀況 (Xunzi; ca. 310–ca. 210), did not match this ideal pattern of a successful life. Despite following their moral principles and professing knowledge of how to rule a state, none of them held a longterm or high government post or was able to bring about change in their own deplorable times. The Xunzi even concludes with a eulogy explaining why Master Xun had no official post, followers, or fame yet still had to be considered a worthy with the mind of a sage (sheng zhi xin 聖之心), who just did not encounter the right times (bu yu shi 不遇時).33 The problem of this fundamental contradiction between sagehood and successful life fostered a number of discussions of the life of Confucius, considered by some to have been the last of the great sages in the world. The main works that I have found to reflect the general problem of the relation between virtuous action and success in the pre-Qin era are the *Qiongda yi shi 窮達以時 (Failure and Success Depend on Timeliness) manuscript from Guodian 郭店, the *Guishen zhi ming 鬼神之明 (Evidence of Ghosts and Spirits) from the Shanghai Museum corpus, the Mengzi 孟子, the Xunzi 荀子, and the Lüshi chunqiu 呂氏春秋 (Annals of Lü Buwei). In the following section I will reconstruct the main argumentative lines of these discourses, analyze their historical development, and then try to relate relevant passages in the Lunyu to these discourses in order to date them.

Success and Failure Outside of the Lunyu

The *Qiongda yi shi opens with a general statement supported by a range of historical examples:

33

Wang 1988: 553.

126

Gentz

There is Heaven and there is Man. Heaven and Man are distinct. When investigating Heaven and Man’s distinction, one understands what makes things go. When there is the right man, but there are not the right times, then even though he be worthy, he will not be able to put his worthiness into action. However, if there are the right times, what difficulties will there be? … [In all the above cases,] whether or not they encountered favorable circumstance lay with Heaven. Their actions were not executed for the sake of success; therefore, when failing they were not distressed. Their pains were not suffered for the sake of achieving a reputation; therefore, when nobody knew them, they did not feel disgraceful. … [S]uccess and failure correspond to timeliness. … It is for this reason that the gentleman prizes self-examination.34 有天有人,天人有分。察天人之分,而知所行矣。有其人,亡其世, 雖賢弗行矣。茍有其世,何難之有哉?… 遇不遇,天也。動非為達 也,故窮而不[怨,隱]非為名也,故莫之智而不吝。… 窮達以時 … 故 君子敦於反己。35

This short text summarizes the theme of our discussion in an extremely dense and stylistically sophisticated form. It shows that the main arguments and terms of a discussion on the problem of the relationship between a good life and a good fate were already set up around 300 BCE. They were embedded in a discourse on timeliness that referred to a number of well-known historical precedents, particularly the sage-emperor Shun 舜 and Wu Zixu 伍子胥 (d. 484 BCE). The text further shows that this discourse at an early stage was not necessarily related to Confucius, whose predicament “between Chen and Cai” (Chen Cai zhi jian 陳蔡之間) was central to later discussions of success and failure.36 The main point established by the *Qiongda yi shi is the strict separation between Heaven and Man. It clearly defines and explains the limits of Man’s effort in regard to a successful good life. Sober historical and empirical observations had prompted the conclusion that the ideological assumptions of a correspondence between Man’s virtue and his success so strongly propagated by the early Zhou were no longer tenable. Neither virtue nor merit is sufficient to generate success in life; thus, the efficacy of any human action depends on 34 35 36

Several English translations of this text have been made: D. Meyer 2012: 57–67; Cook 2012: 453–464; my unpublished manuscript (2014), from which this translation is taken. Text from Jingmen shi bowuguan 1998: 145. This story has been analyzed in detail by Makeham (1998) and A. Meyer (2013). A similar analysis of Confucius stories that appear across early Chinese texts has been presented by Weingarten (2010).

The Lunyu, A Homeless Dog In Intellectual History

127

the right moment in time, which lies entirely with Heaven, not with Man. The last paragraph of the *Qiongda yi shi formulates the consequences of this insight (proven empirically in the text with historical examples) for human aspirations. As humans have no influence on the efficacy of their actions, the only thing they can do is generate the potential for virtuous action (de xing 德行) and concentrate on their own perfection (fan ji 反己) without any consideration for its effects. The *Qiongda yi shi does not discuss the extent to which the virtues cultivated by men provide a condition for success, a claim put forward in later discussions. The *Guishen zhi ming takes the same problem as a starting point in its discussion of whether ghosts and spirits interfere in the human world to ensure that humans are rewarded or punished according to their virtue. After presenting some of the historical examples commonly connected to this discourse (Shun, Wu Zixu, etc.), it concludes in a fashion similar to the *Qiongda yi shi, albeit within a different conceptual framework: If one examines this, then some among the good people were not re­ warded and some among the brutes were not punished. Therefore, when I accordingly propose that “[the power of] ghosts and spirits is not evident,” then this must have a [further] reason. Is it that their power suffices to reach out to them and yet they won’t do it? I do not know. Or is it that their power in fact does not suffice to reach out to them? I also do not know. As these two are different, I [just] say: there are cases in which [the power of] ghosts and spirits is evident and cases in which it is not.37 如以此詰之,則善者或不賞,而暴 [者或不罰。古(故)] 吾因加“鬼 神不明”,則必 有故。其力能致焉而弗爲乎? 吾弗知也;意其力固不 能致焉乎?吾又弗知也。此兩者歧。吾故 [曰:鬼神有] 所明,有所不 明。此之謂乎!38

The same problem of the disconnect between virtue and reward leads to a very different explanation. That Shun became Son of Heaven was not because he encountered Yao but because the power of ghosts and spirits in his case became evident with regard to his good actions. That the worthy Wu Zixu died was because the ghosts’ and spirits’ power did not become evident with regard 37 38

Several English translations of this text have been made: Ding 2006; Brindley 2009; ­Sterckx 2013: 122–125; an unpublished manuscript by Marco Caboara (2013); my unpublished manuscript (2014), from which this translation is taken. Text from Cao 2005 as quoted in Ding 2006.

128

Gentz

to his worthiness. In contrast to the *Qiongda yi shi’s timely encounter approach, the *Guishen zhi ming’s demonological approach to explaining humans’ failure or success is unique in early Chinese literature and is nowhere taken up, or even referred to. The lack of conditions 3 and 5 from the introductory catalog thus makes it difficult to include this text in an intellectual history of our topic. The Mengzi discusses the same problem but with a different set of assumptions. Focusing on Confucius as the new model of a sage, it defines a successful life in moral terms as a life in which inner virtues have been cultivated. Mengzi is probably the first text that makes an explicit terminological distinction between inner and outer factors of morality, especially in some of the Gaozi dialogues.39 Fame, wealth, and official positions are viewed as signs of external success of secondary importance. Confucian virtues such as benevolence and righteousness are opposed to material profit, prosperity, and animal appetites, even in existential situations. A beggar starves to death rather than taking food that is given abusively as long as he has not lost his true heart,40 which is now considered the most basic criterion for a successful life. When the Mengzi discusses taking office, it criticizes Bo Yi 伯夷,41 the sage who refused to serve immoral governments, for eschewing his responsibilities and being too concerned with moral purity. It is equally critical of Yi Yin 伊尹, who would take any office under any circumstances. In contrast, Confucius is praised at Mengzi 5B/1 as the “sage of timeliness” (孔子,聖之時者也), who would assess the circumstances before deciding whether to serve.42 Timeliness (shi 時) here is not related to a Heavenly moment that would have allowed Confucius to realize his sagely potential in office and bring peace and order to the world, as in the *Qiongda yi shi, but to the weighing of whether to take office in times of disorder.43 As we read in Mengzi 3B/3, Confucius, like other gentlemen, was eager to take office; he even became agitated when he was not in service for three months and is criticized for showing unseemly impatience. Asked why Confucius, being so eager in seeking office, found it so hard to take one, Mengzi responds that he would never seek it by dishonorable means.44 The whole of Mengzi 5B also deals with 39 40 41 42 43 44

Mengzi 6A/6: “Benevolence, righteousness, ritual sense, and wisdom were not infused into me from outside; I had them originally” (仁義禮智,非由外鑠我也,我固有之 也). For the use of nei 內 and wai 外, see also Mengzi 6A/4, 6A/5. Mengzi 6A/10. Mengzi 2A/2, 3B/10, and 5B/1 criticize Bo Yi much in the same manner as Lunyu 18/8. See also Mengzi 2A/2. Weighing up according to the circumstances (quan 權) is a central concept of Mencian philosophy. See also Vankeerberghen 2005–2006. According to Mengzi 5B/4, Confucius’s motives when seeking office were threefold: “Confucius took office sometimes because he could realize his Way, sometimes because he was

The Lunyu, A Homeless Dog In Intellectual History

129

the question of the proper relationship between gentlemen and rulers, what kind of positions gentlemen should take, and what salaries or gifts they should accept. Mengzi 5B/5 suggests poverty as another motive for Confucius’s acceptance of (low) office. Poverty, however, plays no central role in the general discussion about taking office. As John Makeham notes, “the root problem was not poverty but powerlessness, a lack of position,”45 resulting in constant “job hunting,”46 as reflected in Mengzi 3B/3, 5B/4, and elsewhere. The Mengzi never mentions Heaven in its discussion of office seeking despite elsewhere emphasizing its crucial role in determining fate and in conferring the right to rule over the empire. In the final passage of 1B/16 we find the following passage on fate: [Mengzi] said, “When one moves, something causes it; when one halts, one is hindered by something. Moving and halting are nothing that lies in man’s ability. That I did not encounter the Marquis of Lu is due to Heaven.” 曰:行或使之,止或尼之,行止非人所能也。吾之不遇魯侯,天也。

The passage strikingly recalls the *Qiongda yi shi. Heaven, not Man, is responsible for whether or not humans encounter favorable circumstances (yu 遇). And destiny decides Man’s fate.47 On the right to rule, Mengzi 5A/5 is illustrative: Wan Zhang said, “That Yao gave the world to Shun, did that really happen?” Mengzi said, “No. The Son of Heaven cannot give the world to somebody else.” “In that case, who then gave the world to Shun?” [Mengzi] said, “Heaven gave it to him.” 萬章曰:堯以天下與舜,有諸?孟子曰:否。天子不能以天下與人。 然則舜有天下也,孰與之?曰:天與之 。

Elsewhere, the Mengzi takes a long-term view of human history, introducing a general law of alternating order and disorder (yi zhi yi luan 一治一亂) as an

45 46 47

treated with decency, and sometimes because a prince supported good people at court” (孔子有見行可之仕,有際可之仕,有公養之仕). Makeham 1998: 78. The Mozi’s account of the story of Confucius between Chen and Cai also takes up the theme of poverty to illustrate Confucius’s hypocrisy; see Makeham 1998: 81. Makeham 1998: 89. See also the first few passages of book 7A, also 2B/13.

130

Gentz

external condition for the appearance of sages (3B/9), who appear in the world every five hundred years (2B/13, 7B/38).48 However, in contrast to the *Qiongda yi shi, these passages stand alone and are nowhere developed into any conceptual discourse.49 Instead, the argumentative focus in the Mengzi with respect to Confucius is on the lack of support from men in power, not Heaven or any other external laws. Mengzi 5A/6 begins with the continuation of the argument in 5A/5 that Heaven decides who rules the world. The reason Confucius never became a ruler, however, is not, as in the other cases, claimed to be Heaven’s decision but, rather, is because the Son of Heaven did not recommend him: To rule the world, a common man must have not only the virtue of a Shun or a Yu [which Confucius had, of course] but also the support of an emperor. Therefore, Confucius never ruled the world.50 匹夫而有天下者,德必若舜、禹,而又有天子荐之者。故仲尼不有天 下。

In 7B/18 the Mengzi further refers to the famous story of Confucius’s stranding, along with his disciples, in the region of Chen and Cai, one of the frequently cited illustrations of Confucius’s failed life. The Mengzi explains this situation again by the lack of support shown by the authorities in power: Mengzi said, “That the gentleman [Confucius] was in danger between Chen and Cai was because he had no connections to superiors at court.” 孟子曰:君子之厄於陳蔡之閒,無上下之交也。

On the surface, these statements seem to contradict other statements in the Mengzi which claim that benevolence (ren) is matchless. In 4A/7 we read: Confucius said, “Benevolence cannot be quantified. As a matter of fact, if the ruler of a state loves benevolence, he will be matchless in the world.”51 48 49 50 51

See also Mengzi 2B/13, 7B/38. Hans van Ess (2009) argues that 7B/38 must be later, probably second century BCE. For a more detailed analysis of the Mencian view of ming (“fate,” “destiny”), see Eno 2005: esp. 8–9. Based on the translation in Lau 1970: 145. Trans. Lau 1970: 121.

The Lunyu, A Homeless Dog In Intellectual History

131

孔子曰:仁不可為眾也。夫國君好仁,天下無敵。

Or, using another Confucius quotation in 2A/1: Confucius said, “[Governmental] virtue spreads faster than an order transmitted through posting stations. At the present time, if a state of ten thousand chariots were to practice benevolent government, the people would rejoice as if they had been released from hanging by the heels. Just now is the time therefore when one will surely, with half the effort, achieve twice as much as the ancients.”52 孔子曰:德之流行,速於置郵而傳命。 當今之時,萬乘之國行仁政, 民之悅之,猶解倒懸也。故事半古之人,功必倍之,惟此時為然。

 Ren 仁 and other virtues appear like magic weapons, particularly in times of disorder, yet this power seems to be effective only if exerted by a ruler. The Mengzi thus insists on a strict separation between power and responsibility on the one side and Confucius’s virtues on the other, with the result that his failure appears entirely as the failure of the authorities in power. The main question in the Mengzi is, therefore, where and how do we seek office at rulers’ courts? The case of Confucius, the greatest sage of all, is nowhere connected to Heaven, timeliness, or any other external or nonhuman factors. The Xunzi is the first text to combine the discourses of the *Qiongda yi shi and the Mengzi. In “Rong ru” 榮辱 (Of Honor and Disgrace), the Xunzi seems to propagate a straightforward correlation between virtue and success. Honorable men will enjoy benefits; disgraceful men will suffer harm. In “Fei shier zi” 非十二子 (Contra Twelve Philosophers), the example of Confucius and other sages demonstrates that “even this kind of sage does not always gain a position of power” (是聖人之不得埶者也).53 Counterexamples—that is, sages who attained power—are, again, Shun and Yu. Although a great Ru will influence the whole world with his greatness, he is able to do so only if he is in office. If he fails to get an office (qiong 窮), he will be ridiculed by ordinary Ru but he will obtain a great reputation because he never parts from the true Way although he lives as a poor man in reclusion.54 The Xunzi thus shares the Mengzi’s assumption of a division between a virtuous life and great wealth, fame, and official posts. At the same time, however, it maintains that virtuous conduct 52 53 54

Trans. Lau 1970: 75–76 with amendments. Wang 1988: 97. See Xunzi, chap. 8, “Ru xiao” 儒效 (The Teachings of the Ru).

132

Gentz

leads to respect and fame. The Xunzi’s explanation of Confucius’s failure toward the end of “Zhongni” 仲尼 (On Confucius) reveals striking parallels to the *Qiongda yi shi: “Hence the gentleman bends in times that require bending and straightens out in times that require straightening” (故君子時詘則詘,時 伸則伸也).55 In this and other chapters we find a reflection of the Mencian idea of proper weighing and Confucius as the sage of timeliness, which, however, is still an insufficient condition of success. Elsewhere in the Xunzi, “Cheng xiang” 成相 (Working Songs) contains a number of laments that, as in the Mengzi, place the responsibility for success or failure of great Ru on the times of good and bad government: Yao conferred [the empire] on an able man, and Shun encountered the right times. He elevated the worthy and promoted the virtuous, so the world was well ordered. But even if a man is a worthy or a sage, if he does not encounter the right times, who will recognize him?56 堯授能,舜遇時,尚賢推德天下治。雖有聖賢,適不遇世,孰知之?

Alas! Who am I that I alone do not encounter the right times in this disordered age!57 嗟!我何人,獨不遇時當亂世!

However, this still does not explain why sages such as Shun and the great Ru encountered propitious circumstances whereas Xunzi encountered unlucky ones. In “Ru xiao” 儒效 (Teachings of the Ru), the Xunzi gives an explanation that refers to a discourse on the charioteer Zaofu 造父 that also appears in the *Qiongda yi shi and echoes the Mencian idea that a ruling position is pivotal for a sage to exert his influence and to change the world.58 In contrast to the *Qiongda yi shi, the Xunzi does not claim that success depends in any way on Heaven. However, we find an identical explanation in “You zuo” 宥坐 (The Warning Vessel on the Right) in the story of Confucius between Chen and Cai, a key element in all discourses about Confucius’s failure:

55 56 57 58

Wang 1988: 113. Wang 1988: 462. Wang 1988: 467. Wang 1988: 137. Xunzi uses the same criterion as Mengzi 2A/2 of a ruler needing only a hundred square li (百里之地) for his perfect rule.

The Lunyu, A Homeless Dog In Intellectual History

133

When Confucius once traveled southward toward Chu, he ran into difficulties in the territory between Chen and Cai. For seven days he and his disciples had not eaten cooked food, only herb soup without a single grain of rice, so that the disciples all had a famished look. Zilu stepped forward and asked, “I have heard that Heaven responds with good fortune to those who do good and with disasters to those who do what is bad. As to our situation here and now, you, Master, have accrued your virtue, accumulated acts of righteousness, and dwelt on the good, and have done so for a long time. Why, then, do you live in obscurity?” Confucius replied, “Yóu, you don’t understand, I will tell you. Do you think that the wise are certain to be employed? But did not Prince Bigan have his heart cut out?!59 Do you think that the loyal are sure to be employed? But did not Guan Longfeng endure punishment?! Do you think that those who remon­strate are always followed? But was not Wu Zixu slashed apart and exposed outside the eastern gate of Gusu?! “As a matter of fact, whether one encounters [the right opportunity] or not depends on the right time; whether one becomes a worthy or not depends on innate ability. Gentlemen who broaden their studies and make profound plans and yet do not meet with the right time are numerous. From this can be seen that those who have not met with the right time are legion. How should I be the only one? And, indeed, consider the orchid and angelica that grow deep in the forest: that there is no one [to smell them] does not mean that they are not fragrant. The learning of the gentleman is not undertaken for the sake of a success [in his career], but so that if he fails, he will not be beset with hardship, so that in times of grief his sense of purpose will not diminish, and so that in regard to knowing fortune and misfortune, ends and beginnings, his heart will not suffer illusions. In fact, whether one is worthy or not depends on innate ability, whether one acts or not depends on the man, whether one en­ counters [favorable circumstances] or not depends on the right time; whether one lives or dies depends on fate. Now, if there is the right person but he does not meet with the right time, then even though he is worthy, would he be able to put his worthiness into practice? Yet if he should chance to meet with the right time, what difficulty would there

59

The Xunzi refers to this event again in one of its fu-poems on the disordered world, in which it states that since antiquity it has been a constant rule that things turn around after one thousand years (千歲必反,古之常也). Cf. chap. 26.6 (Wang 1988: 482).

134

Gentz

be! Therefore, the gentleman studies broadly, develops profound plans, cultivates himself, and gives his utmost to await his right time.”60 孔子南適楚,厄於陳蔡之間,七日不火食,藜羹不糝,弟子皆有飢 色。子路進而問之曰: 由聞之:為善者天報之以福,為不善者天報之 以禍。今夫子累德積義懷美,行之日久矣,奚居之隱也?孔子曰:由 不識,吾語女。女以知者為必用邪?王子比干不見剖心乎!女以忠者 為必用邪?關龍逢不見刑乎!女以諫者為必用邪?吳子胥不磔姑蘇東 門外乎!夫遇不遇者,時也;賢不肖者,材也;君子博學深謀,不遇 時者多矣!由是觀之,不遇世者眾矣,何獨丘也哉!且夫芷蘭生於深 林,非以無人而不芳。君子之學,非為通也,為窮而不困,憂而意不 衰也,知禍福終始而心不惑也。夫賢不肖者,材也;為不為者,人 也;遇不遇者,時也;死生者,命也。今有其人,不遇其時,雖賢, 其能行乎?苟遇其時,何難之有!故君子博學深謀,修身端行,以俟 其時。61

The numerous verbatim parallels with the *Qiongda yi shi are striking. The Xunzi propounds three arguments in the story. First, it continues the earlier explanations of the Mengzi that the gentleman studies not for worldly comfort but for inner virtues and that he remains firm and steady in times of hardship. Second, in the final passage it takes the experience of hardship as a condition for a gentleman to broaden his horizons and deepen his thoughts. It thereby concludes with a redefinition of what real success means for a gentleman. Third, the Xunzi agrees (almost verbatim) with the *Qiongda yi shi that meeting with the right time determines one’s worldly success, official position, fame, and wealth (although the topic of official posts is not explicitly mentioned here). In the Xunzi, however, time is not connected to Heaven. This becomes quite explicit when we look at the sentence “whether one encounters [the right opportunity] or not depends on Heaven” (遇不遇者,天也) of the *Qiongda yi shi, which is rendered “whether one encounters [the right opportunity] or not depends on the right time” (遇不遇者,時也) in the Xunzi. Time is one of four factors involved in producing a good life, the others being innate ability (賢不肖者,材也), personal engagement (為不為者,人也), and existential fate (死生者,命也). As timeliness trumps worthiness, it is the basic condition for any successful action. The Xunzi, although sharing the Mengzi’s veneration for Confucius as a sage, concedes that Confucius was not able to 60 61

Based on the translation by Knoblock (1994: 249–250). See also Makeham 1998: 79–80; Wang 1988: 526–527; A. Meyer (2013). Wang 1988: 526–527.

The Lunyu, A Homeless Dog In Intellectual History

135

put his sageliness into practice. It introduces the notion of timeliness into the Confucian discourse on success, which is controlled neither by Man nor by Heaven. The sage has to turn back to himself and await the right time to be able to change the world. The most complex work on timing in the pre-Qin era, the Lüshi chunqiu, continues the combined discourse of the Xunzi while also asking how far humans are able to control the success of their actions, how much of their success is attributable to their own wisdom or efforts as opposed to Heaven or other external factors.62 In line with the *Qiongda yi shi and the Xunzi, chapter 14.3 of the Lüshi chunqiu, “Xu shi” 胥時 (Awaiting the Right Time), argues that great success depends on both the right time and the worthiness of those who accomplish the action. It is Heaven that gives the opportunity, but it is Man who seizes it: Heaven will not give him an opportunity twice; time will not remain right for long; someone who is skilled does not do his work twice; handling affairs lies in being appropriate in time.63 天不再與,時不久留,能不兩工,事在當之。

Sages have an insight into when the accomplishment of a certain action will be realistic or most effective and therefore await the right time for their respective actions. Chapter 14.6, “Shen ren” 慎人 (Being Mindful of the Human), explains in detail what the Heavenly and the human contributions to success are. With numerous intertextual links, the chapter connects directly back to our previous discussion of the *Qiongda yi shi and the Xunzi, including the story of Confucius in the border region of Chen and Cai, but it supplies a conclusion that goes one step further, claiming that the happiness of those who have achieved the Way lies beyond failure (qiong 窮) and success (da 達/tong 通).64 As its title “Yuhe” 遇合 indicates, chapter 14.7 deals with “encounter” and “concord.” Confucius is presented as an example of someone who met more than eighty lords and yet was able to reach only the position of minister of crime in Lu. The chapter thus reflects a new problem in the discourse of success. There is no longer a consensus on what exactly a successful “encounter” 62 63 64

Sellman 2002. Chen 1990: 769; Knoblock and Riegel 2000: 315; Wilhelm 1979: 186–187. Instead of the contrastive pair qiong 窮—tong 通, as used in the same story in the Xunzi and the Zhuangzi, the text uses the pair qiong 窮—da 達, like the *Qiongda yi shi.

136

Gentz

is. Obviously, not every opportunity that provides a gentleman with the chance to realize his own virtues counts as a successful “encounter.” An encounter needs to “concord” (he 合) with the favorable circumstance of a good ruler in order to be successful. The chapter emphasizes that, because of the lack of wise rulers, the gentleman has to critically examine his own capabilities before he accepts the huge responsibility that comes with an office, a theme seen in the Mengzi. However, the Lüshi chunqiu here shifts attention from the ruler’s morality to whether he possesses the basic intellectual qualification to choose the right people. The (unreliable) morality of a ruler no longer serves as a benchmark for the definition of a real encounter, nor is it relevant for the discussion of whether, when, and why to take office: This is the reason why it is said, “There are no constant principles in ‘encounter’ and ‘concord.’ Pleasing/persuading others is merely a matter of behaving compliantly.” 故曰:遇合也無常。說,適然也。65

Chapter 14.8, “Bi ji” 必己 (Certainty in Oneself), moves this discourse further into a broader philosophical field: “External things cannot be dealt with in any certain way” (外物不可必).66 The outer world is utterly unreliable, and humans have no way of controlling the outcomes of their actions. There is no general rule or solution for one’s own personal situation in the external world, no guarantee that any theory is applicable or that any method assuredly works. Everything depends on contexts that cannot be known or controlled. Whether an encounter will lead to success or not is unpredictable, as a number of the stories in this chapter illustrate with a touch of cynical black humor.67 The chapter ends with the recommendation that one should turn to and rely on oneself.68 This expression of extreme mistrust and anxiety in regard to the outer world might reflect some aspects of a late Warring States period psychology of utmost insecurity in an unpredictable world of constant warfare. It is also the ending point of a discourse that had started quite confidently with assumptions about, and the hopes for a return to, a golden era when powerful inner virtues had a transformative effect on the outer world. Later discussions, in 65 66 67 68

Chen 1990: 816; Knoblock and Riegel 2000: 329; Wilhelm 1979: 198. Chen 1990: 828; Knoblock and Riegel 2000: 331; Wilhelm 1979: 200. Chen 1990: 829–830; Knoblock and Riegel 2000: 334; Wilhelm 1979: 203. See also Chen 1990: 830; Knoblock and Riegel 2000: 335; Wilhelm 1979: 203–204. Chen 1990: 830; Knoblock and Riegel 2000: 335; Wilhelm 1979: 204.

The Lunyu, A Homeless Dog In Intellectual History

137

contrast, reflect the loss of any confidence in the outer world and in the efficacy of morality. They suggest instead an inward turn to the only reliable point that can be trusted, the self. What started as a strict separation of Heaven and Man and the advice “to turn back to oneself” (fan ji 反己) in the *Qiongda yi shi developed into a strict separation between self and others together with the quite similar advice to focus on that which is “certain in oneself” (bi zai ji 必在 己). It is difficult to draw straight lines between the stages of the discourse that have been presented so far, but it seems that the analyzed texts allow us to construct a narrative of the development of this discourse in stages that follow quite neatly upon each other even though we do not know how many, and what kind of, discursive elements come in between these stages and how exactly they relate to one another.69 It seems that for these texts all five conditions for a reconstruction of an intellectual history as outlined above are given so far, and a narrative summary could run as follows: As in so many other ancient cultures,70 the observation that moral worthiness is not the only decisive factor that leads to a successful life stimulated a debate in early China on the question of which other factors play a role in achieving success (and which factors other than moral action could play a role in avoiding harm). We have looked at some early textual witnesses of this discourse and have reconstructed some stages of a discourse that starts from two different approaches to the problem. On the basis of a number of historical precedents regarding the success of famous worthies, the *Qiongda yi shi introduced timely encounters with rulers as a second decisive factor besides the worthiness of the historical individuals. The two instances of Heaven and Man were identified as the key factors that, in strict separation and entirely independently of each other, both equally contributed to success in life. Taking Confucius as the historical example by reference to which the relation between moral worthiness and successful life had to be defined, the Mengzi identified worthiness with success and therefore had to distinguish inner moral success as the true and ultimate success from an inferior outer success of fame, wealth, and official posts. It used neither Heaven nor timely encounters as models to explain Confucius’s failure. His not taking 69

70

The *San de 三徳 manuscript from the Shanghai Museum collection (Li Ling 2005) could, for example, combine some aspects of the concept of timeliness as employed in the *Qiongda yi shi and some aspects of the power of ghosts and spirits as envisioned in the *Guishen zhi ming, a text that otherwise cannot be connected to any of the analyzed discourses. See, e.g., the Ludlul bēl nēmeqi (also called the Babylonian Job; Müller 1978) or The Dispute between a Man and His Ba (Assmann 1990).

138

Gentz

office in bad governments was interpreted as an act of sageliness. The Xunzi, much more pessimistic about the prospect of changing the bad times, combined these two approaches and introduced timeliness into the Confucian discourse as the decisive factor explaining why Confucius was not successful in implementing his sageliness to restore order by means of an official post. The Lüshi chunqiu developed the discourse several steps further by questioning and relativizing basic notions of the discourse such as “timeliness” and “encounter.” It ended by entirely giving up any belief even in the notions of good and bad times or anything outside the inner self that could be relied upon to guarantee success. Turning to the Lunyu’s participation in this discourse, we must analyze how a successful life is defined in the Lunyu and how Confucius’s life is portrayed, evaluated, and explained. The questions of why Confucius was not successful in his political career and how an unrecognized gentleman should conduct himself occupy Confucius and his students throughout the Lunyu.71 I take this as a clear indication that a correlation between worthiness and success in a political career was expected in the Lunyu context(s). Despite its interest in the problem, however, the Lunyu does not develop a coherent solution. Lunyu 14/5 comes closest to articulating the problem of virtue and success but without providing an answer: Nangong Kuo asked Confucius, “Yi was good at archery and Ao was dragging boats, yet neither met with a natural death. Yu and [Hou] Ji bent down to sow grain, and yet they came to rule over the world.” Confucius did not respond. After Nangong Kuo had left, he said, “A gentleman indeed is this man! High virtue indeed has this man!” 南宮适問於孔子曰: 羿善射,奡盪舟,俱不得其死然;禹稷躬稼,而 有天下。夫子不答,南宮适出。子曰:君子哉若人!尚德哉若人!

Overall, the Lunyu addresses the relationship among official service (wei 位, shi 試, shi 士, shi 仕, wei zheng 為政), recognition (zhi 知), and fame (ming 名, wen 聞) with respect to the worthy (xianzhe 賢者) or gentleman (junzi 君子) from five different perspectives, the first of which explores the relation between a successful gentleman and recognition by contemporaries.72 We find this already in the *Dizi wen 弟子問 (The Disciples Ask) manuscript of the Shanghai 71 72

Lunyu 1/1, 4/14, 8/1, 14/30, 14/35, 15/19. Even though Confucius appears to be rather indifferent to this problem, he nevertheless constantly addresses this topic throughout the Lunyu.

The Lunyu, A Homeless Dog In Intellectual History

139

Museum collection, which is dated to around 300 BCE: “The Master sighed and said, ‘Oh! Nobody recognizes me!’” (子嘆曰: 烏!莫我知也夫!).73 The Lunyu opens (1/1) with a passage on this very topic: The Master said, “To study and in due time to put it into practice, is this not a pleasure? To have friends coming from afar, is this not a joy? Not being indignant when people do not recognize oneself, is this not a gentleman?” 子曰:學而時習之,不亦說乎?有朋自遠方來,不亦樂乎?人不知而 不慍,不亦君子乎?

Lunyu 4/14 also suggests that a successful gentleman does not strive for the recognition of others but aims at what makes others recognize him: “The Master said, ‘Don’t be troubled that you have no official position; be troubled about the means by which you establish yourself. Don’t be troubled that nobody recognizes you; seek what is worth being recognized for’” (子曰:不患無位,患 所以立;不患莫己知,求為可知也). A second perspective arises in the Lunyu’s discussion of the relationship between success and fame, which assumes that true successful gentlemen, even though not recognized by contemporaries, will eventually be praised, as at Lunyu 15/20: “The Master said, ‘A gentleman is concerned that when he departs from the world his name is not praised’” (子曰:君子疾沒世而名不稱焉). Lunyu 16/12 could be read as clarifying the difference between recognition (perspective 1) and fame (perspective 2): Lord Jing from Qi possessed one thousand teams of horses, yet on the day he died, the people knew no virtue for which to praise him. Bo Yi and Shu Qi starved at Mount Shouyang, yet the people praise them down to the present day. Isn’t this what it is saying? 齊景公有馬千駟,死之日,民無德而稱焉。伯夷叔齊餓于首陽之下, 民到于今稱之。其斯之謂與?

The passage makes clear that a person’s true success and fame rests not on material goods but on his moral merits. We find a similar argument in the 73

Zhang 2005: 279, 269, strips 20, 4. Hunter (2012: 230 and the table in appendix 3:H, 399– 400) notes that this precise phrase is also found in other early texts (such as the Laozi and the “Li sao” 離騷).

140

Gentz

discussion of success (da 達) at Lunyu 12/20, which is also a central concept in the *Qiongda yi shi: Zizhang asked, “How should a shi-scholar be so that he can be called successful?” The Master said, “What then would you consider successful?” Zizhang replied, “One who is certain to be heard no matter whether he is employed at the royal court or at a great official’s house.” The Master said, “This is being heard, not being successful. As a matter of fact, someone who is successful presents himself straight and is fond of righteousness, he examines what others say and observes their appearance, and he is thoughtful of deferring to others. Then he is certain to be successful no matter whether he is employed at the royal court or at a great official’s house. In general contrast, someone who is being heard just puts on an air of benevolence while acting contrary to it, and he just abides by this without any doubt. He then is certain to be heard no matter whether he is employed at the royal court or at a great official’s house.” 子張問:士何如斯可謂之達矣? 子曰:何哉,爾所謂達者?子張對 曰:在邦必聞,在家必聞。子曰:是聞也,非達也。夫達也者,質直 而好義,察言而觀色,慮以下人。在邦必達,在家必達。夫聞也者, 色取仁而行違,居之不疑。在邦必聞,在家必聞。

Confucius in this passage is eager to point out the difference between being heard and being truly successful. In Lunyu 15/2, the famous story of Confucius’s distress between Chen and Cai is used to discuss the other side of the failure/ success (qiong 窮 / da 達) opposition. In this passage, Zilu asks the Master the interesting question of whether it is possible for a gentleman to fail (qiong 窮): When they were in difficulties between Chen and Cai, they ran out of grain and the disciples became so weak that none of them was able to stand up. Zilu indignantly came to see the Master and said, “Does failure actually ever exist for a gentleman?” The Master said, “A gentleman is steadfast when he fails; a petty man in such circumstances lets himself go.” 在陳絕糧,從者病,莫能興。子路慍見曰:君子亦有窮乎?子曰:君 子固窮,小人窮斯濫矣。

The Master responds to Zilu’s desperate question in the negative. A gentleman might fail, but so long as he is steadfast, then he does not fail at being a

The Lunyu, A Homeless Dog In Intellectual History

141

gentleman. Failures of the sort encountered between Chen and Cai remain outer problems. The basic argument in all these passages follows the distinction between self versus others and inner versus outer that we also find in the Mengzi. Even the famous story of Confucius’s distress between Chen and Cai is explained within the oppositional pattern of challenging external circumstances and the inner virtues of a gentleman. Material wealth, skills, fame, and position are the main issues that are (and obviously need to be) minimized in contrast with the real values that a gentleman strives for: dao 道, de 德, ren 仁, yi 義, gu 固, etc. This division between material wealth, worldly success, office, power, etc. and Ruist values in the Lunyu becomes even more apparent in the second ten chapters of the Lunyu, where Confucius does not intend to seek office.74 Yet there is no coherent terminology or clear conceptual framework. Reputation is regarded as positive when it is based on moral achievements, and this is what one should strive for. It is not worth striving for when it is based merely on wealth, power, or performance. The third perspective addresses the question of whether an official position is necessary for exerting a moral impact on the world.75 On the one hand, Lunyu 2/21 implies that one can have the same impact on the world when following moral standards as when occupying an official position: Someone asked Kongzi, “Why don’t you take an active role in governance?” The Master said, “The Book of Documents says, ‘Filial piety, just filial piety [toward your parents], and friendliness toward your brothers, extend this to the realm of governance.’ This is also taking an active role in governance. Why should this relate only to officially playing an active role in governance?” 或謂孔子曰:子奚不為政?子曰:《書》云:孝乎惟孝、友于兄弟, 施於有政。是亦為政,奚其為為政?

Lunyu 5/6 and 15/7 also endorse the position found in the Mengzi and developed in the Lüshi chunqiu that an official position should be accepted only under the appropriate circumstances, and Lunyu 8/14 indicates that there are 74

75

Makeham 1998: 92–93. This is somehow closer to the way Confucius is depicted in the Zhuangzi than to his representations in the Mengzi and Xunzi, where he appears eager to seek office (Makeham 1998: 90–91). However, there is constant frustration and lament in the Lunyu, whereas the “Rang wang” 讓王 chapter of the Zhuangzi has him singing and being joyful so that even his disciples misunderstand him. Lunyu 2/21, 3/24, 4/14, 4/26, 5/6, 15/7, 17/1, 18/5, 18/7.

142

Gentz

certain political actions that should be executed only when holding the appropriate post in government.76 On the other hand, passages like Lunyu 17/1 insist on the value of official service: … “Can someone be called benevolent who holds his moral treasure hidden and lets his state go astray?” Confucius said, “One can’t.” “Can someone be called wise who is fond of regulating affairs but constantly misses the right time to do so?” Confucius said, “One can’t.” “Days and months are passing by; the years are not with us.” Confucius said, “Right, I will get employed!” […]懷其寶而迷其邦,可謂仁乎?曰:不可。好從事而亟失時,可謂 知乎?曰:不可。日月逝矣,歲不我與。孔子曰:諾。吾將仕矣。

Similarly, at Lunyu 18/7 Zilu goes so far as to assert that “not to serve is to lack righteousness” (不仕無義). The fourth set of passages depicts Confucius as a servant not of a worldly ruler but of Heaven, which bestowed upon him a task beyond that of any common officeholders: When the Master was surrounded in Kuang, he said, “King Wen has long passed away, but aren’t the accomplished patterns of order still here with us? If Heaven intended to destroy these patterns, then those remaining after his death would not have gained access to these patterns. If, however, Heaven has not yet destroyed these patterns, how would the people of Kuang then deal with me?”77 子畏於匡。曰:文王既沒,文不在茲乎?天之將喪斯文也,後死者不 得與於斯文也;天之未喪斯文也,匡人其如予何?

Despite all the doubts and sorrows displayed by Confucius in the Lunyu, the self-confidence expressed here renders the problem of worldly success and 76

77

“The Master said, ‘If one does not occupy the appropriate official position, then one should not devise its governance’” (子曰:不在其位,不謀其政). The parallel passage in Lunyu 14/26 has one more sentence: “Zengzi said, ‘The gentleman’s thinking does not go beyond his own particular position’” (曾子曰:君子思不出其位). This sentence in turn has a parallel in the xiang 象 commentary to the Zhouyi hexagram liang 艮. Lunyu 9/5.

The Lunyu, A Homeless Dog In Intellectual History

143

recognition irrelevant. Even so, one wonders why Confucius in other passages, including 14/35, continues to mention his lack of recognition: The Master said, “Nobody recognizes me!” Zigong said, “What do you do when nobody recognizes you, Master?” The Master said, “I don’t bear a grudge against Heaven and do not blame other people; instead, I study human affairs below and thereby reach up to Heaven. The one who knows me, it is Heaven!”78 子曰:莫我知也夫!子貢曰:何為其莫知子也?子曰:不怨天,不尤 人。下學而上達。知我者,其天乎!

Other passages express the same view from the perspective of other figures, as at Lunyu 3/24: A border official from Yi asked to meet Confucius and said, “I managed to meet every single gentleman who arrived at this place.” Confucius’s followers thereupon arranged a meeting with him. When he left he said, “Why are you guys so distressed about the general decline?79 The world has been without the right Way for a long while, but Heaven is about to use Confucius as his wooden clapper [for making announcements].” 儀封人請見。曰:君子之至於斯也,吾未嘗不得見也。從者見之。出 曰:二三子,何患於喪乎?天下之無道也久矣,天將以夫子為木鐸。

Like 9/5, this passage envisions Confucius in times of disorder not as a government official but as an “official” employed by Heaven and fulfilling a special kind of duty. Confucius’s teaching is presented as being of much greater importance and exerting a much greater impact than a common official’s worldly officeholding. Crucially, this last perspective does not appear in any of the other early pre-Qin texts surveyed above. A fifth way to present the theme is through the voice of people outside Confucius’s circle (like the preceding passage from Lunyu 3/24). In later chapters the theme has a slightly different tinge. It appears as a topic that is so much associated with Confucius that all kinds of people address it by either urging 78 79

Lunyu 14/35. Most commentators and translators interpret sang 喪 as a loss of position. Although this reading would support my analysis much better, I think it makes more sense to read it in the context of Lunyu 9/5 as a state of general decline of the patterns of order (喪斯文).

144

Gentz

Confucius to take office (see 17/1 above), warning him to do so (18/5), or condemning him and his disciples for doing no work (18/7). Where do we place the Lunyu in the historical development of early Chinese discourses on success and failure? The Lunyu addresses the general problem of the correspondence between moral accomplishments and successful life that is discussed throughout the texts that I have analyzed. It thus can be regarded as a player in the history of this problem (fulfilling condition 1). Continuous and changing elements in this history could be established for the *Qiongda yi shi, the Mengzi, the Xunzi, and the Lüshi chunqiu: the *Qiongda yi shi and the Mengzi start two different discourses on the same problem that are then combined in the Xunzi and further developed in the Lüshi chunqiu. However, not only does the Lunyu present a diversity of perspectives on the problem of success and virtue (thus not fulfilling condition 2), but it very rarely echoes discourses found in other early texts (weakening condition 4). There are no references to the impact of ghosts and spirits on human fate as in the *Guishen zhi ming, Confucius is not depicted as someone who acted in a timely manner or who awaited the right time as in the *Qiongda yi shi,80 and there are no references to determined good or bad times.81 We do find elements of fatalism in the Lunyu where Confucius expresses his belief in the dependence of his own fate on the actions of Heaven. A fatalistic notion of timeliness occurs sometimes in the sense that his time has come or has not yet come or, as in 9/22, with a slight doubt: “There are, are there not, sprouts that do not produce blossoms and blossoms that do not produce fruit?”82 It is perhaps for this reason that the Mozi charges Confucianism with being “fatalistic.”83 Never do we find, however, any more systematic reflection on this issue. Moreover, the Lunyu does not refer to the historical precedents that are central to this particular discourse. Wu Zixu is not mentioned once in the Lunyu, nor is Shun ever associated with the question of success or fate. Yet, the terms qiong and da are used in a similar way as in the *Qiongda yi shi and in the Xunzi as expressions of success and failure in a human’s life. The Mencian ideas that incapable rulers were responsible for Confucius’s failure to take office and that Confucius appeared according to a five-hundredyear sagely cycle are also absent from the Lunyu. Only two Lunyu concepts of success have parallels in the Mengzi. The first is the idea, perhaps the most 80 81 82 83

This idea seems to start with the Gongyang commentary to the fourteenth year of Lord Ai’s 哀公 reign, the last entry of the text. The term ming 命 in Lunyu 14/36 is sometimes interpreted in this way but could equally well, and in my view more convincingly, refer to human orders. See also Lunyu 14/36. Eno 2005.

The Lunyu, A Homeless Dog In Intellectual History

145

fundamental and coherent in the Lunyu, that success and failure have to be measured by inner moral standards and not by external criteria such as recognition, wealth, or high office. The second is the idea that Confucius cautiously assessed the conditions before taking office. Lunyu 8/13 is particularly trenchant in this matter. It states not only that no post should be taken in times of disorder but that one should have a position and be successful in times when the Way prevails.84 Timeliness is brought into this discussion in the Lunyu only to argue that one should take office in times of good government and hide (or even resist) in bad times.85 Unlike in the Lüshi chunqiu we also do not find any doubt in the belief that humans are able (at least partly) to control success. Instead, and in clear opposition to the Lüshi chunqiu and the other early discourses on the topic, the Lunyu expresses full confidence that a gentleman’s actions do have a reliable impact on the outer world and that certain methods can be applied to secure social and moral order. The Lunyu discussions are not complex enough to allow us to identify more detailed features indicative of historical shifts in order to contextualize these more precisely (as demanded by condition 4). On this basis alone, the Lunyu could be positioned before, between, or after the Mengzi and the Xunzi.86 The one element that is Lunyu specific, Confucius’s higher calling conferred by Heaven, cannot be found in any of the other texts and thus cannot be dated. We also find very little abstract or theoretical reasoning on the problem of (Confucius’s) success or failure as such. There is no general question about reasons or responsibilities for a successful life that could provide answers that lie beyond the behavior of the persons concerned. Heaven, ghosts, fate, timeliness, time cycles, bad rulers, or a combination of several of these had been proposed as explanations in other early texts reflecting upon this problem. The Lunyu, in contrast, is mainly concerned with how Confucius acted, or thought a gentleman should act, in such circumstances and thereby constructs a unique 84

85 86

“The Master said, ‘Be sincere in your trustworthiness and be passionate in your learning, and abide to the death in the good Way. Do not enter a state that is in peril; don’t stay in a disordered state. Show yourself when the Way prevails in the world, but hide yourself when it does not. It is shameful to be poor and humble when the Way prevails in the state; it is shameful to be rich and noble when it does not’” (子曰:篤信好學, 守死善道。危 邦不入,亂邦不居,天下有道則見,無道則隱。邦有道,貧且賤焉,恥也; 邦無道,富且貴焉,恥也). If we relate these statements to the numerous expressions of grievance against the declining world in the Lunyu, they appear as yet another explanation of why the very fact that Confucius never gained an official post is not an indication of his failure but rather a proof of his true sageliness. Lunyu 5/2. For a critical discussion on the relationship between the Lunyu and the Mengzi, see Hunter 2014.

146

Gentz

approach to the problem, one that provides casuistic models of action without addressing the problem as such. Thus, it is an approach that is homeless in early Chinese intellectual history.

Conclusion

In my methodological reflections at the beginning of this analysis, I postulated that in order to be dated by means of intellectual history, a text first and foremost needs to contain intellectual historical elements. Such elements can be detected in the Lunyu. Terms, concepts, ideas, and problems such as the two problems of Confucius’s knowledge and of his success relate to discourses that can be found in other early texts as well. These, however, are not consistent on the level of intellectual discourses within the Lunyu, as the example of ren showed in a first analysis. The Lunyu also does not present intellectual discourses that relate to a common thread of an argument or that relate to other identifiable discourses. This prompted my model of Confucius sayings in the Lunyu as unrelated tesserae. It is also in line with Hunter’s thesis that “Confucius quotation before the advent of the Analects was a dynamic, creative practice in which authors treated Confucius sayings as venues for the re-performance of inherited wisdom.”87 It also accords with Csikszentmihalyi and Kim’s conclusion quoted above. As the main focus of the Lunyu is not on the individual discourses but on a larger discourse on Confucius, however, this is where I hoped to find its contribution to intellectual history and where I tried in a second step of analysis to find coherent views that run through the whole book. It turned out, however, that on this level of Confucius portrayals that represent one of the continuous, vivid, and central discourses in Chinese intellectual history, no consistent elements of intellectual discourse were found either. Second, I claimed that for any dating, these elements of intellectual discourses need to show features of change that make them datable in a history of X, in our case a history of problems. Yet the construction of Confucius in the Lunyu is so unique among the early pre-Qin texts that it is not possible, at least 87

Hunter 2012: i. Hunter concedes however, that although “Kongzi quotation practice in the early period was a decidedly messy phenomenon,” Kongzi was more than “a cipher who parroted whatever ideas an author wished to attribute to him. … Kongzi yue sayings were … not infinitely flexible. Close attention to the functions of Kongzi sayings reveals a handful of common threads and interests, which, taken together, illuminate their distinctive role within early intellectual discourse” (120). Hunter identifies eight such common threads (120–124).

The Lunyu, A Homeless Dog In Intellectual History

147

on the basis of the subdiscourse on Confucius’s success and failure that I analyzed in this chapter (and the example on Confucius’s knowledge analyzed in Gentz 2012), to link it to any datable elements in other texts. We do find similar and overlapping representations of a Master who fails in exterior aspects of his life such as official position and wealth but who has great accomplishments on the inner moral side, and a Master who carefully weighs whether and when to take office. But no development of these aspects can be found in other texts that would allow us to reconstruct a history of these aspects within which the Lunyu could be located. My attempt to use an intellectual history approach to date the Lunyu by analyzing the discursive field of success did not fully meet conditions 2, 4, and 5 set out at the beginning of the chapter and was thus unsuccessful.88 This does not mean that a dating of the Lunyu by means of intellectual history is impossible in principle. But it is not possible, I would argue, on the level of internal intellectual discourses on individual terms and concepts such as ren 仁, yi 義, xiao 孝, and li 禮, nor is it possible on the basis of discourses surrounding Confucius’s success and failure or Confucius’s knowledge.

Bibliography

Ames, Roger, and Henry Rosemont Jr. 1998. The Analects of Confucius: A Philosophical Translation. New York: Ballantine. Assmann, Jan. 1990. “Der ‘leidende Gerechte’ im alten Ä gypten: Zum Konfliktpotential der ägyptischen Religion.” In Loyalitätskonflikte in der Religionsgeschichte: Festschrift für Carsten Colpe, ed. Christoph Elsas and Hans G. Kippenberg, 203–224. Würzburg: Königshausen und Neumann. Brindley, Erica. 2009. “‘The Perspicuity of Ghosts and Spirits’ and the Problem of Intellectual Affiliations in Early China.” Journal of the American Oriental Society 129.2:215– 236. Busse, Dietrich, and Wolfgang Teubert. 1994. “Ist Diskurs ein sprachwissenschaftliches Objekt? Zur Methodenfrage der historischen Semantik.” In Begriffsgeschichte und Diskursgeschichte: Methodenfragen und Forschungsergebnisse der historischen Se­ man­tik, ed. Dietrich Busse, Fritz Hermanns, and Wolfgang Teubert, 10–28. Opladen: Westdeutscher Verlag.

88

As referenced in n. 32, an earlier effort to date the Lunyu by analyzing the discursive field of knowledge was similarly unsuccessful.

148

Gentz

Cao Jinyan 曹錦炎. 2005. “‘Guishen zhi ming’ shiwen zhushi” 《鬼神之明》釋文註釋. In Shanghai bowuguan cang Zhanguo Chu zhushu 上海博物館藏戰國楚竹書, ed. Ma Chengyuan 馬承源, vol. 5, 307–320. Shanghai: Shanghai guji chubanshe. Chan, Wing-tsit. 1986. Neo-Confucian Terms Explained: The “Pei-hsi tzu-i” by Ch’en Ch’un, 1159–1223. New York: Columbia University Press. Chen Qiyou 陳奇猷. 1990. Lüshi chunqiu jiaoshi 呂氏春秋校釋. Shanghai: Xuelin chubanshe. Collingwood, Robin George. 1946. The Idea of History. Oxford: Clarendon. Cook, Scott. 2012. The Bamboo Texts of Guodian: A Study and Complete Translation. Ithaca, NY: Cornell University East Asia Program. Csikszentmihalyi, Mark, and Tae Hyun Kim. 2014. “The Formation of the Analects.” In The Analects of Confucius, trans. Simon Leys, ed. Michael Nylan, 152–165. New York: Norton. Ding Sixin丁四新. 2006. “Shangbo Chujian ‘Guishen’ pian zhushi” 上博楚简《鬼神》 篇注释. http://www.bsm.org.cn/show_article.php?id=337, accessed August 18, 2014. Elstein, David. 2006. “Friend or Father? Competing Visions of Master-Student Relations in Early China.” PhD diss., University of Michigan. Elstein, David. 2009. “The Authority of the Master in the Analects.” Philosophy East and West 59.2:142–172. Eno, Robert. 2005. “Mencius, Readings 5: Timeliness and ‘Fate’ (Ming).” http://www. indiana.edu/~p374/Mengzi5.pdf, accessed March 29, 2017. Ess, Hans van. 2009. “Die Idee des Zyklus von fünfhundert Jahren bei Mengzi, Sima Qian und einigen anderen Denkern der Han.” In Dem Text ein Freund: Erkundungen des chinesischen Altertums; Robert H. Gassmann gewidmet, ed. Roland Altenburger, Martin Lehnert, and Andrea Riemenschnitter, 57–76. Bern: Peter Lang. Gadamer, Hans-Georg. 1990. Wahrheit und Methode. Tübingen: Mohr. Gentz, Joachim. 2012. “Confucius Confronting Contingency in the Lunyu and the Gongyang zhuan.” Journal of Chinese Philosophy 39.1:60–70. Goldin, Paul R. 2011. Confucianism. Berkeley: University of California Press. Graham, Angus C. 1989. Disputers of the Tao: Philosophical Argument in Ancient China. La Salle, IL: Open Court. Gu Jiegang 顧頡剛. 1988. “Chunqiu shidai de Kongzi he Handai de Kongzi” 春秋時代 的孔子和漢代的孔子. In Gu Jiegang gushi lunwen ji 顧頡剛古史論文集. Beijing: Zhong­hua shuju. Guoyu 國語. 1988. Shanghai: Shanghai guji chubanshe. Harbsmeier, Christoph. 1990. “Confucius Ridens: Humor in The Analects.” Harvard Journal of Asiatic Studies 50.1:131–161. Hartmann, Nicolai. 1910. “Zur Methode der Philosophiegeschichte.” Kantstudien 15: 459–485.

The Lunyu, A Homeless Dog In Intellectual History

149

Hirase Takao 平勢隆郎. 2016. “Jin” no gengi to kodai no suri: Nijushishi no “jin” hyōka “tenri” kan o kiso toshite 「仁」の原義と古代の數理:二十四史の「仁」評價「天 理」觀を基礎として. Tokyo: Tōkyō Daigaku Tōyō bunka kenkyūjo hōkoku. Hunter, Michael. 2012. “Sayings of Confucius, Deselected.” PhD diss., Princeton University. Hunter, Michael. 2014. “Did Mencius Know the Analects?” T’oung Pao 100:1–47. Jingmen shi bowuguan 荊門市博物館. 1998. Guodian Chu mu zhujian 郭店楚墓竹簡. Beijing: Wenwu chubanshe. Knoblock, John. 1994. Xunzi, a Translation and Study of the Complete Works. Stanford: Stanford University Press. Knoblock, John, and Jeffrey Riegel. 2000. The Annals of Lü Buwei. Stanford, CA: Stanford University Press. Lau, D. C. 1970. Mencius. London: Penguin. Li Ling 李零. 2005. .“‘San de’ shiwen zhushi” 《三德》釋文註釋. In Shanghai bowuguan cang Zhanguo Chu zhushu 上海博物館藏戰國楚竹書, ed. Ma Chengyuan 馬承源, vol. 5, 286–303. Shanghai: Shanghai guji chubanshe 125–148 Makeham, John. 1996. “The Formation of Lunyu as a Book.” Monumenta Serica 44:1–24. Makeham, John. 1998. “Between Chen and Cai: Zhuangzi and the Analects.” In Wandering at Ease in the Zhuangzi, ed. Roger T. Ames, 75–100. New York: State University of New York Press. Meyer, Andrew. 2013. “The Frontier between Chen and Cai: Anecdote, Logic and Argumentation in Early China.” Paper presented at the Anecdotes Workshop, Leiden University, May 31. Meyer, Dirk. 2012. Philosophy on Bamboo: Text and the Production of Meaning in Early China. Leiden: Brill. Müller, Hans-Peter. 1978. Das Hiobproblem: Seine Stellung und Entstehung im Alten Orient und im Alten Testament. Darmstadt: Wissenschaftliche Buchgesellschaft. Murray, Julia K. 2001. “Portraits of Confucius: Icons and Iconoclasm.” Oriental Art 47.3:17–28. Murray, Julia K. 2009. “‘Idols’ in the Temple: Icons and the Cult of Confucius.” Journal of Asian Studies 68:371–411. Nylan, Michael, and Thomas Wilson. 2010. Lives of Confucius: Civilization’s Greatest Sage through the Ages. New York: Doubleday. Oexle, Otto Gerhard, ed. 2001. Das Problem der Problemgeschichte, 1880–1932. Göttingen: Wallstein. Pines, Yuri. 2012. The Everlasting Empire: The Political Culture of Ancient China and Its Imperial Legacy. Princeton, NJ: Princeton University Press. Sanren 三人, ed. 2006. Liushige Kongzi 六十个孔子. Changsha: Hunan wenyi chubanshe.

150

Gentz

Sellman, James D. 2002. Timing and Rulership in Master Lü’s Spring and Autumn Annals. Albany: State University of New York Press. Sgarbi, Marco. 2010. “Umriß der Theorie der Problemgeschichte.” In Eine Typologie der Formen der Begriffsgeschichte, ed. Riccardo Pozzo and Marco Sgarbi, 7.185–199. Hamburg: Meiner. Sgarbi, Marco. 2011. “Concepts vs. Ideas vs. Problems: Historiographical Strategies in Writing History of Philosophy.” In Begriffs-, Ideen- und Problemgeschichte im 21. Jahrhundert, ed. Riccardo Pozzo and Marco Sgarbi, 69–80. Wiesbaden: Harrassowitz. Shanghai bowuguan cang Zhanguo Chu zhushu 上海博物館藏戰國楚竹書, vol. 5. 2005. Ed. Ma Chengyuan 馬承源. Shanghai: Shanghai guji chubanshe. Simson, Wojciech Jan. 2006. Die Geschichte der Aussprüche des Konfuzius (Lunyu). Bern: Peter Lang. Sommer, Deborah. 2002. “Destroying Confucius: Iconoclasm in the Confucian Temple.” In On Sacred Grounds: Culture, Society, Politics, and the Formation of the Temple of Confucius, ed. Thomas A. Wilson, 95–133. Cambridge, MA: Harvard University Asia Center. Sterckx, Roel. 2013. “Mozi 31: Explaining Ghosts, Again.” In The Mozi as an Evolving Text: Different Voices in Early Chinese Thought, ed. Carine Defoort and Nicolas Standaert, 95–141. Leiden: Brill. Vankeerberghen, Griet. 2005–2006. “Choosing Balance: Weighing (quan 權) as a Metaphor for Action in Early Chinese Texts.” Early China 30:47–89. Wagner, Rudolf G. 1991. “Die Unhandlichkeit des Konfuzius.” In Weisheit: Archäologie der literarischen Kommunikation II, ed. Aleida Assmann and Jan Assmann, 455–464. Munich: Finck. Wagner, Rudolf G. 2004. “Making Sense: Wang Bi’s Commentary on the Lunyu.” In Zhong Ri “Sishu” quanshi chuantong chutan 中日《四書》詮釋傳統初探, ed. Huang Junjie 黃俊傑, 109–158. Taipei: Taiwan Daxue chuban zhongxin. Waley, Arthur. 1989. The Analects of Confucius. New York: Random Vintage. Wang Xianqian 王先謙. 1988. Xunzi jijie 荀子集解. Beijing: Zhonghua shuju. Weingarten, Oliver. 2010. “Textual Representations of a Sage: Studies of Pre-Qin and Western Han Sources on Confucius (551–479 BCE).” PhD diss., University of Cambridge. Weingarten, Oliver. 2014. “The Sage as Teacher and Source of Knowledge: Editorial Strategies and Formulaic Utterances in Confucius Dialogues.” Asiatische Studien / Études asiatiques 68.4:1175–1223. Wilhelm, Richard. 1979. Frühling und Herbst des Lü Bu We. Düsseldorf: Diederichs. Windelband, Wilhelm. 1892. Geschichte der Philosophie: Geschichte der Probleme und der zu ihrer Lösung erzeugten Begriffe. Freiburg i.B.: Mohr.

The Lunyu, A Homeless Dog In Intellectual History

151

Zhang Guangyu 張光裕. 2005. “‘Dizi wen’ shiwen zhushi” 《弟子問》釋文註釋. In Shanghai bowuguan cang Zhanguo Chu zhushu 上海博物館藏戰國楚竹書, ed. Ma Chengyuan 馬承源, vol. 5, 265–283. Shanghai: Shanghai guji chubanshe.

152

Van Els

Chapter 6

Confucius’s Sayings Entombed: On Two Han Dynasty Bamboo Lunyu Manuscripts Paul van Els

As one of the foundational texts of Chinese culture, the Analects (Lunyu 論語) was copied throughout the centuries and in all corners of the Chinese empire, from the capital city to the very edges of the Sinosphere. The text was inked on bamboo, silk, paper, and wood and durably engraved in stone. Some age-old manuscripts have made it to modern times. Until the final quarter of the twentieth century, the earliest extant version of the Analects was one of the “Stone Classics of the Xiping era” (Xiping shijing 熹平石經). These canonical Confucian texts were engraved in stone in Luoyang, the seat of government during the Eastern Han dynasty, around the year 175 CE. Several decades ago, however, archaeologists discovered two handwritten copies of the Analects in tombs that had been closed around 50 BCE. Well over two centuries older than the stone carvings, these handwritings on strips of bamboo now rank as the earliest Analects ever found.1 The two bamboo manuscripts have come to be known as the Dingzhou Analects 定州論語 and P’yŏngyang Analects 平壤論語, after the respective locations of the tombs in which they were found. Although unearthed decades ago, news about their spectacular discoveries has only gradually trickled out into

1 In 2011, when I wrote the present essay, archaeologists in China started excavating a tomb complex near Nanchang 南昌 in Jiangxi Province 江西省. The person buried in the main tomb is said to have been Liu He 劉賀, the Marquis of Haihun 海昏侯, who died in 59 BCE. The archaeological excavations yielded numerous grave goods related to Confucius, including a lacquer screen containing the earliest known portrait of the master and bamboo strips on which reportedly the chapter “Zhi dao” 知道 (Knowing the Way) was inked. As this chapter does not exist in the received Analects, scholars speculate that the Haihun bamboo strips may be part of a Qi 齊 version of the Analects (for more on this, see below). This remains speculative, though, given the absence of academic publications on the Haihun bamboo strips to date. For the same reason, the Haihun bamboo strips are not included in the present discussion of Han dynasty Analects manuscripts, even though the date of the Marquis of Haihun’s tomb roughly corresponds to the dates of the two tombs under discussion in the present essay.

© Koninklijke Brill NV, Leiden, 2018 | doi:10.1163/9789004382947_008

Confucius’s Sayings Entombed

153

the academic world. As a result, studies of the excavated manuscripts are few and far between.2 The present essay is intended as a gateway to both two-thousand-year-old manuscripts. The first two sections discuss the archaeological context of the discoveries and analyze the manuscripts themselves, including characteristic features of the bamboo strips and the texts inked thereon and notable differences between these and other Analects versions. In these sections, I also critically evaluate present-day Analects studies and offer alternative hypotheses where there is room for debate. The third and final section of the essay discusses what I consider the most fascinating (and most complex) issue regarding the manuscripts: their provenance. In that section, I examine when, where, and why the Analects was copied onto the bamboo strips. The ultimate goal is to present a nuanced understanding of the two bamboo manuscripts that conveys the fascinating insights they offer while also exploring the limitations of what these manuscripts can actually tell us.

The Dingzhou Analects

This section discusses the Dingzhou Analects from its entombment to its discovery in modern times, its analysis by archaeologists and paleographers, and finally its assessment by other scholars in the field. Tomb and Excavation In the Western Han dynasty (202 BCE–9 CE), the area south of present-day Beijing 北京 was known as Zhongshan 中山. In the year 55 BCE, the ruler of 2 The Dingzhou Analects was discovered in 1973, but its transcription was not published until 1997. Since then, the bamboo manuscript has been mentioned in a dozen or so publications. Ames and Rosemont (1998: 271–277) in their Analects translation devote an appendix to the manuscript; Csikszentmihalyi (2002: 146–147) discusses it in his book chapter on Confucius and the Analects in the Han dynasty; and Makeham (2003: 367–368) offers a brief description in his essay on the development of the early commentarial tradition on the Analects. Other scholars mention the Dingzhou Analects in passing, while translators note differences between the bamboo manuscript and the received text in their translations. Specialized studies focusing entirely on the Dingzhou Analects have appeared only in Chinese. For details, see the bibliography at the end of this chapter, or see Tang 2007 for a helpful overview of Chinese post–Dingzhou Analects research. The P’yŏngyang Analects was discovered in 1992. Since access to the manuscript was (and still remains) highly restricted, no academic study on it was published until 2009. The only relevant publications to date are by Kim Kyŏng-ho, Shan Chengbin, and Yi Sŏng-shi et al. (see the bibliography).

154

Van Els

Zhongshan, Liu Xiu 劉脩, passed away.3 Posthumously known as King Huai of Zhongshan 中山懷王, Liu Xiu was buried near Lunu 盧奴, the capital city of Zhongshan, in a wooden tomb under a large burial mound surrounded by an earthen wall. Clothed in a precious suit made of jade tesserae sewn together with gold thread, he was laid to rest in a nested coffin and accompanied by a wealth of funerary objects (including jade, gold, bronze, and lacquerware) and a number of texts written on strips of bamboo.4 While his tomb had been fitted with every conceivable posthumous comfort, the king’s afterlife was far from peaceful. Not long after his burial, robbers entered the wooden-tomb construction. Yet before they could plunder many of the valuable objects within, they inadvertently sparked a fire with their torches, setting the place ablaze, and were thus forced to make a quick escape.5 While the flames saved numerous objects from the bandits’ hands, an unknown number of artifacts and manuscripts went up in smoke, and many of those remaining were scorched and scattered (fig. 6.1). After the fire, no one is known to have entered the tomb for another two millennia, until 1973. In 1973 a team of Chinese archaeologists excavated the tomb, located in what is now the village of Bajiaolang 八角廊 near Dingzhou (a city built on the soil that once was Lunu) in Hebei Province. Eight months of work were required to complete the excavation. The excavated materials were sent to the National Cultural Relics Bureau 國家文物局 in Beijing, where specialists analyzed the bamboo strips, which had been severely damaged by the tomb fire.6 With the help of the renowned paleographer and historian Li Xueqin 李學勤 and other scholars, they assigned consecutive numbers to the bamboo strips and transcribed legible graphs onto notecards, one strip per card. Sadly, in 1976 the devastating Tangshan 唐山 earthquake toppled the storage boxes in which the ancient manuscripts were contained, causing further damage to the strips. The tomb fire, the earthquake, and spectacular manuscript discoveries elsewhere (Mawangdui 馬王堆, Zhangjiashan 張家山, Guodian 郭店, to give a few examples) delayed further analysis of the excavated objects. In 1981 the research team published a brief excavation report and a short introduction to

3 Van Els (2009: 916–919) reflects on the exact year of the king’s death and on the likelihood of Liu Xiu being the person buried in the Dingzhou tomb. 4 Hebei sheng bowuguan et al. 1976 and Hebei sheng wenwu yanjiusuo 1981 provide detailed information about the jade suit and the excavated grave goods. 5 Hebei sheng bowuguan et al. 1976: 57. 6 Hebei sheng wenwu yanjiusuo 1995: 38–39 is a detailed report of the analysis of the bamboo strips.

Confucius’s Sayings Entombed

155

Figure 6.1 Bamboo fragment from the Dingzhou tomb (photograph by the author)

the bamboo strips.7 This is when the world first learned that eight distinct texts had been found in the king’s tomb: Rujia zhe yan 儒家者言 (Sayings of the Scholars) Wenzi 文子 Lunyu 論語 Taigong liu tao 太公六韜 (The Grand Duke’s Six Secret Teachings) Ai gong wen wuyi 哀公問五義 (Duke Ai Inquires about the Five Ways of Righteousness) Baohu zhuan 保傅傳 (Biography of the Grand Tutor) rishu, zhanbu 日書占卜 (hemerological and divinatory texts) Liu’an wang chao wufeng ernian zhengyue qiju ji 六安王朝五鳳二年正月起 居記 (Record of the King of Lu’an’s Visit to the Imperial Court in the First Month of the Second Year of the Five Phoenixes Reign)

• • • • • • • •

7 Guojia wenwu ju et al. 1981a.

156

Van Els

In a later publication the research team mentioned en passant that one more text, a memorial by the prominent Han dynasty statesman Xiao Wangzhi 蕭望之 (ca. 114–46 BCE), had also been discovered in the same tomb.8 To date, four of the texts have been published in transcription: Sayings of the Scholars (1981), Wenzi (1995), Analects (1997), and The Grand Duke’s Six Secret Teachings (2001). The long intervals between these publications and the apparent dormancy of the project since 2001 have done little to enhance academic awareness of the tomb and its discovery. This is regrettable, because the tomb yielded impressive objects and texts aside from the oldest handwritten copy of the Analects ever found, and their function and significance will not be fully understood without an in-depth study of all the tomb’s contents.9 Tracings and Transcriptions As mentioned, the Dingzhou Analects was discovered in 1973, but a description of the manuscript was not published until eight years later, in the August 1981 issue of the Chinese academic journal Wenwu 文物.10 Another sixteen years later, a transcription of select bamboo strips was published in the May 1997 issue of Wenwu, accompanied by tracings, notes on the transcription, and an explanatory essay by the research team responsible for arranging the Dingzhou bamboo strips, written by team leader Liu Laicheng 劉來成.11 The same year also witnessed the publication of the full transcription of the Dingzhou Analects as a separate monograph.12 Compared with other manuscripts, published in sumptuous books replete with magnificent pictures, such as those unearthed in Guodian or those purchased by the Shanghai Museum, the publication of the Dingzhou Analects (and other manuscripts from the same tomb) leaves much to be desired. The few tracings are accompanied neither by photographs nor by explanations as to why these few bamboo fragments were selected for tracing. Without photographs or a complete set of accurate tracings, those who wish to study the bamboo Analects must rely solely on its transcription. Sadly, the transcription is not flawless. Because of the fire that once raged in the tomb, the surviving bamboo fragments were found in disorder. Since it is impossible to know their original or8 9 10 11 12

Hebei sheng wenwu yanjiusuo 1997b: 61. Van Els 2009 provides more information about the Dingzhou tomb and its unfortunate fate. Guojia wenwu ju et al. 1981a. Hebei sheng wenwu yanjiusuo 1997b. Hebei sheng wenwu yanjiusuo 1997a.

Confucius’s Sayings Entombed

157

der, the transcription presents them in the order of corresponding passages in the received text, which does not necessarily reflect the original order. Owing to the Tangshan earthquake, numerous bamboo fragments were destroyed or damaged to the point that the graphs on them are no longer legible. The graphs from these fragments survive only as transcriptions on cards made prior to the natural disaster. Because these graphs can no longer be verified against the original manuscript, they appear within square brackets in the published transcription. The transcription was published in modern regular script (kaishu 楷書) in simplified characters. This is a methodological flaw. As William G. Boltz has written, manuscripts “should be transcribed so as to reveal as precisely and unambiguously as possible the exact form of what is written, without introducing any interpolations, alterations, or other extraneous material based on assumptions, biases, or subjective decisions of the scholar-transcriber or of anyone else. In a nutshell, this means that the transcription should reflect exactly what is written and nothing more.”13 Boltz’s argument also applies here: the change to regular script is an alteration of the manuscript, and even more so is the change to simplified characters. This violates the principle of structural consistency, which, Boltz explains, entails that the transcription of a graph “should not deviate from the actual structural form of the graph in the manuscript.”14 To facilitate reading, the transcription also contains modern punctuation marks. While helpful, this “extraneous material” (Boltz’s terminology) is uncalled for in a methodologically correct transcription because it forces an interpretation of the text that may limit the possibilities offered by the unpunctuated transcription. The reader should have the opportunity to see exactly what the ancient scribe wrote, not what the modern editor thinks the scribe intended to write.15 These are just a few issues with the transcription of the bamboo Analects. Other issues are outlined in a four-page article by Sun Qinshan.16 I emphatically note that the purpose of pointing out these problematic aspects of the transcription is not to criticize Chinese colleagues who faced the unenviable task of making sense of the unpromising heap of charred pieces of bamboo (see fig. 6.1), and whose professional facilities may not have met international standards. However, these problems do highlight the need for especially 13 14 15 16

Boltz 1999: 596. Boltz 1999: 597. Richter (2003) and Xing Wen (2005) discuss methodological issues concerning the transcription of excavated early Chinese manuscripts. Sun 2007.

158

Van Els

careful treatment of ancient manuscripts, including the publication of transcriptions. Moreover, the problems emphasize that the transcription must be used with caution. When using the transcription in research, scholars rely heavily on choices made by editors several decades ago, with no opportunity to verify the accuracy of their work. Any study involving the Dingzhou Analects should ideally contain a disclaimer stating that its results are tentative. Chapters, Sections, and Graphs The Dingzhou Analects consists of 620 bamboo strips, most of which are fragments with one or both ends broken off. Only a handful of strips are complete. When they were placed in the tomb two millennia ago, the strips were probably 16.2 centimeters long and 0.7 centimeters wide, with 19–21 graphs per strip. Three binding threads joined these bamboo strips at the top, middle, and bottom. The threads did not survive, but their imprints are still visible on the excavated bamboo fragments. On those fragments, 7,576 graphs have been discerned, which amounts to just under half the length of the received Analects. The graphs were written in a mature, highly rectilinear Han dynasty clerical script (lishu 隸書), in which the graphs are square to wide in shape, with wavelike flaring of major strokes, as shown on figure 6.2. There are notable differences between the bamboo manuscript and the received text in terms of their (1) chapters, (2) sections, and (3) graphs. 1. The bamboo manuscript appears to have been a complete version of the Analects when it was placed in the tomb, as corresponding bamboo fragments have been found for all twenty chapters in the received Analects. However, the degree of survival differs markedly per chapter, from as few as 20 graphs (about 4 percent of the chapter total) for chapter 1 to as many as 694 graphs (about 77 percent) for chapter 15. I think the different degrees of survival are coincidental and not due to an inherent feature of the manuscript, as the opening chapter was probably positioned at the outer edge of the roll of bamboo strips and therefore most susceptible to destruction by the tomb fire. The manuscript does not mention chapter titles or chapter numbers. It does, however, mention the number of sections and graphs within coherent textual units that we would probably call chapters. Ten excavated bamboo fragments list such information, and Chen Dong explored likely counterparts in the received text for each of them (table 6.1). Five of these transcribed bamboo fragments display the total number of graphs in the textual unit to which they belong. In each case, the number is lower than that of the corresponding chapter in the received text. This could potentially indicate that the handwritten copy is a condensed version of the Analects. More likely, in view of what we know from other manuscripts, it shows that the growth of the Analects at that

Confucius’s Sayings Entombed

159

Figure 6.2 Tracings of the Dingzhou Analects (Hebei sheng wenwu yanjiusuo 1997b: 50)

time was still in progress. The Former Han dynasty was a time of textual fluidity, when texts were susceptible to change.17 It seems that despite occasional 17

One famous case of (relative) textual fluidity in the Han dynasty is the Laozi 老子, or Daodejing 道德經, in which the text’s two constitutive parts occur in reverse order in Han dynasty manuscripts when compared with the received text. For more on this, see Henricks 1990: xvi.

160

Van Els

Table 6.1 Possible matching chapters in bamboo manuscript and received texta

Strip

Manuscript text and translation

Received text

0612

• 凡二章 [凡三百廿二字] • In all, 2 sections. A total of 322 graphs. • 凡卅七章 …… • In all, 37 sections ……

Chapter 20 堯曰 3 sections, 341 graphsb Chapter 7 述而 37 sections

0614

…… [章] …… 五百七十五字 ……sections …… 575 graphs.

Chapter 2 為政 581 graphs

0615

凡[卅六]章 • 凡九百九十字

Chapter 17 陽貨 26 sections, 1,020 graphs

0616

• 凡卅章 • 凡七百九十字 • In all, 30 sections. • A total of 790 graphs. • 凡[卌]四章 …… • In all, 44 sections …… [• 凡卌七章] [□□百八十一字] • In all, 47 sections. {XX}81 graphs. • 凡十三章 …… • In all, 13 sections …… [凡十]三章 • …… In all, 13 sections. • …… • 凡廿八章 [• 凡八百五十一字] • In all, 28 sections. • A total of 851 graphs.

Chapter 9 子罕 30 sections, 812 graphs

0613

0617 0618 0619 0620 0621

In all, 36 sections. • A total of 990 graphs.

Chapter 14 憲問 44 sections Chapter 15 衛靈公 49 sections, 900 graphs Chapter 16 季氏 14 sections Chapter 11 先進 23 sectionsc Chapter 5 公冶長 28 sections, 871 graphs

a This table is based on Chen Dong 2003: 8. The transcribed text of the Dingzhou Analects, here and elsewhere in this chapter, corresponds to Hebei sheng wenwu yanjiusuo 1997a. b Chen Dong (2003: 8) explains that the number of graphs here refers exclusively to the first two sections of chapter 20 in the received text. The third section is often considered spurious. See also below. c Chen Dong (2003: 8) suspects that the bamboo strip actually mentions 23 sections, which the modern editors of the transcription mistakenly transcribed as 13. Legend: symbols used in the transcription: □ An illegible graph on the bamboo strip. Rendered as {X} in the English translation. …… A sequence of illegible graphs on the bamboo strip. • A black dot on the bamboo strip. [ ] Graphs on the bamboo strip that were transcribed onto notecards before the Tangshan earthquake but have become illegible after the forces of nature destroyed the part of the bamboo strip on which they were written. These graphs now exist only on the notecards. The accuracy of their transcription can no longer be confirmed. For aesthetic reasons, square brackets are omitted in the English translation.

Confucius’s Sayings Entombed

161

attempts to secure the number of sections and graphs, the Analects slightly increased in size after this particular manuscript was placed in the tomb. 2. Specialists managed to determine the division of sections in the manuscript because at the end of each section the ancient scribe left the remainder of the bamboo strip empty, starting the next section on a new bamboo strip.18 There are some differences in the division of sections in the manuscript and the received text. A single section in the Dingzhou Analects may correspond to two or more sections in the received Analects. Conversely, one section in the received Analects may appear as two sections in the Dingzhou Analects. Most differences are fairly inconsequential. For instance, chapter 10 in the Analects describes how Confucius acted in different situations, such as “when there was a sudden clap of thunder or a violent wind, he invariably assumed a solemn attitude” (迅雷風烈必變) or “when climbing into a carriage, he invariably stood squarely and grasped the mounting-cord” (升車必正立執綏).19 In the received text, these sentences occur in consecutive sections. In the manuscript, the second sentence immediately follows the first, suggesting that they belong to one section. There is one noteworthy variation in sections. In the received text, the concluding chapter of the Analects contains three sections. In the bamboo manuscript, it contains two sections written in regular-sized graphs, followed by two small dots, followed by the third section, which is written in two columns and in half-sized graphs. In other words, the third section is squeezed onto the very same bamboo strip that contains the last sentence of the second section. Liu Laicheng suggests that the small graphs were added to the manuscript after it had already been completed.20 His evidence is bamboo fragment 0612, which mentions “2 sections” and “322 graphs” (see table 1) and likely refers to the concluding chapter. The third section then must have been added after the total number of sections in this textual unit had been written down. It seems that someone in the old days was apparently aware of at least one other Analects version and felt the need to tally the division of sections and chapters with that other version.21 Why did this person squeeze the additional section onto the last bamboo strip? One practical explanation is that this may have been considered less cumbersome than adding additional bamboo strips to the manuscript. A more likely explanation is that the third section was already known at the time when the text was copied onto the bamboo strips but was not 18 19 20 21

Hebei sheng wenwu yanjiusuo 1997b: 49. Translations by D. C. Lau (1979: 105). Hebei sheng wenwu yanjiusuo 1997b: 59. Csikszentmihalyi (2002: 147, 157) also makes this point.

162

Van Els

considered canonical. In that scenario, it was included with the text for the sake of completeness, in smaller graphs to mark its inferior status. This is where a photograph or an accurate transcription of the bamboo strip would come in handy, for it could possibly reveal whether the third section was written by the same hand as the other sections. 3. Many graphs on the bamboo strips are written differently from the graphs in modern regular script to which they are said to correspond. The report of the Dingzhou research team mentions no fewer than seven hundred variants, which amounts to 10 percent of the received text.22 In a meticulous analysis of these variants, Ma Yumeng groups them into various categories and shows that the manuscript contains, among others, graphs that are now written with an added semantic element (such as 立, now written with an additional 亻 “man” element on the left: wei 位 “place, location”); graphs that are now written with the same phonetic element but a different semantic element (such as 功, now written with a 攵 “beat, strike” element on the right: gong 攻 “to attack”); and graphs with a similar pronunciation but no shared structural components (such as 葆, now written bao 寶 “treasure”).23 Most of these variants are what we have come to expect from a Han dynasty manuscript: loanwords, alternative writings, or mere slips of the brush. Such variants are frequently found in other manuscripts of that period. They show that the text was inked on bamboo before the gradual process of standardization of Chinese script had come to completion.24 Finally, Ma Yumeng also notes that the manuscript contains mistakes, such as yue 曰, which should have been you 由 “You [name of a disciple],” and jun 君, which should have been ju 居 “to reside.”25 Textual Differences The Analects was transmitted over a period of two thousand years. In the course of its transmission, the text naturally underwent changes, whether by accident or on purpose. The Dingzhou manuscript, by contrast, spent all this time under layers of soil, unaffected by changes above the ground. One major theme in post-Dingzhou Analects scholarship is the comparison of the bamboo manuscript and the received text, often to clarify contested passages in the latter. A primary difference, as noted by Ma Yumeng, concerns the use of grammatical particles.26 Generally speaking, the manuscript is much less inclined 22 23 24 25 26

Hebei sheng wenwu yanjiusuo 1997b: 59–60. Ma 2010. For an equally meticulous analysis of the variants, see Xu 2006. Galambos (2006) describes this process. Ma 2010: 70. Ma 2010: 73ff.

Confucius’s Sayings Entombed

163

to use function words than the received text. For example, there are over sixty instances where modal particles (e.g., hu 乎, yi 矣), auxiliary words (e.g., zhi 之, zhe 者), and sentence connectives (e.g., ze 則, er 而) are used in the received text but not in the bamboo version. Conversely, the manuscript version contains over two dozen function words, mostly sentence-final modal particles, that are absent in the received text. On the basis of the latter observation, Ma Yumeng suggests that the bamboo manuscript has a rather colloquial flavor and may have been copied from oral recitation, as opposed to being copied from a written version of the Analects. While tempting, this does not harmonize with the former observation of the many cases where the manuscript has fewer modal particles than the received text. Moreover, some of the mistakes in the manuscript appear to be visual rather than aural.27 Thus, it may be difficult to devise one coherent explanation to account for all the differences in the use of grammatical particles. On the one hand, words without a lexical meaning can sometimes be left out without changing the meaning of the text; on the other, grammatical particles can be added to sentences to clarify grammatical relationships between words. As such, the variation in the use of grammatical particles simply means that this manuscript instantiates a unique reading of the text just as other early Analects manuscripts presumably instantiated other unique readings. In addition to function words, there are other noteworthy differences between the bamboo manuscript and the received text. Here is an amusing example from section 7/1 in the received Analects noted by Zhao Jing:28 The Master said, “I transmit but do not innovate. I trust in and love the ancient. One could, perhaps, compare this to our Old Peng.”29 子曰:述而不作,信而好古,竊比於我老彭。

No one knows who this Old Peng was. Some commentators suggest that he was an intimate of Confucius’s given the atypical grammatical construction with the word wo 我 (I, my, we, our) placed immediately before Old Peng’s name, which suggests familiarity and forces the translation “our Old Peng.” Interestingly, bamboo fragment 0138 has a different word order:

27 28 29

Compare: 曰 *gwat vs. 由 *lu or 君 *C.qur vs. 居 *ka. Old Chinese reconstructions, marked by an asterisk, are by Baxter and Sagart (2011). Zhao Jing 2005: 176. Cf. the translation by D. C. Lau (1979: 86).

164

Van Els

but do not innovate. I trust in and love the ancient. One could, perhaps, compare me to Old Peng. [而不作,信而好古,竊比]我於老彭。 The bamboo manuscript leaves the question of Old Peng’s identity unresolved, but the different word order opens up possibilities for someone outside the group of Confucius’s intimates, and it makes a lot more sense than “our Old Peng.” One final textual difference between the bamboo manuscript and the received text concerns the designations “Master Kong” 孔子 and “the Master” 子, both of which introduce statements ascribed to Confucius. Scholars have used these designations for centuries to date different portions of the text, the underlying rationale being that “the Master” displays greater intimacy (and hence an earlier date) than the more distant “Master Kong.” However, the validity of these designations as a dating criterion is questionable, as Weingarten persuasively shows.30 Moreover, as several scholars point out, the usage of these designations in the bamboo manuscript differs from that in the received text.31 There are two sections where the received text reads “Master Kong” and the manuscript has the shorter “the Master.” Conversely, there are five sections where the received text reads “the Master” and the manuscript has the fuller “Master Kong.” Interestingly, four of these five sections occur in chapter 11 of the Analects. It therefore seems that this chapter ascribed more statements to “Master Kong” in the Han dynasty than it does now (see also below). Apart from this conspicuous chapter, there does not appear to be a clear trend or a strict system in the Analects for using these designations. They obviously differ in different versions of the text and cannot be meaningfully used as a dating criterion. Thus, other criteria must be used in dating the Analects, as I shall explain below.

The P’yŏngyang Analects

The P’yŏngyang Analects, also known as Lelang Analects 樂浪論語, is a bamboo manuscript unearthed in the early 1990s. Quite astonishingly, its discovery is still shrouded in mystery. No official report of the discovery has been released, nor have official tracings or transcriptions of the manuscript been 30 31

Weingarten 2009: 37–48. Yang 2003; Liang 2005; Weingarten 2009.

Confucius’s Sayings Entombed

165

published. All we have are a few photographs, with partly overlapping content, which were made public through the relentless efforts of a small number of scholars. This section discusses the P’yŏngyang Analects on the basis of the information that has been made available. Tomb and Discovery About a decade after the Dingzhou tomb was closed, a high official passed away in Lelang Commandery 樂浪郡, an administrative unit in the far northeastern corner of the Han empire. The official was buried with a number of texts, including a copy of the Analects written on bamboo strips and a household register (hukou bu 戶口簿) inked on wooden tablets. The register lists increases and decreases in the number of households and inhabitants of the prefectures that constituted Lelang Commandery.32 The text states that it was drawn up in the fourth year of the Chuyuan 初元 period, or 45 BCE. In all likelihood this was done under the auspices of the high official, who presumably died not long after the population census was completed.33 Two thousand years later, the location of the official’s posthumous abode has become known as the Chŏngbaek-tong 貞柏洞 neighborhood, which is part of the Nakrang-kuyŏk 樂浪區域 district in P’yŏngyang, the capital of North Korea. In the early 1990s, most probably in 1992, the tomb was excavated by North Korean archaeologists, who today preserve its contents at the North Korean Academy of Social Science. The excavated texts probably would have remained unknown to the rest of the world if not for the efforts of Japanese and South Korean scholars. From the 1990s to the early 2000s, teams of Japanese scholars repeatedly visited North Korea to gather material for their study of the ancient Koguryŏ 高句麗 culture. The teams included the renowned archaeologist Egami Namio 江上波夫, the equally famous painter Hirayama Ikuo 平山郁夫, and the acting director of the Koguryŏ Society 高句麗会, Itō Toshimitsu 伊藤利光. During one of their visits, they participated in the celebrations for Kim Il-sŏng’s 金日成 birthday and were even granted an audience with the North Korean leader. In 2003, perhaps in part due to their acquaintance with Kim Il-sŏng, they received a large batch of photographs depicting excavations of ancient tombs in North Korea, including 152 color photos and 3,400 black-and-white photos. Itō Toshimitsu, as the head of the Koguryŏ Society, preserved the photographs. He kept the color photos, gave two albums of black-and-white photos to Tsuruma Kazuyuki 鶴間和幸, professor at Gaku­ shūin University 学習院大学, and donated the remaining black-and-white 32 33

Kim 2011: 61–63. Yi, Yun, and Kim 2011: 163.

166

Van Els

photos to an unspecified research institute in Ōsaka 大阪. One of the blackand-white photos shows bamboo strips on which parts of the Analects are written. In 2003 Yi Sŏng-shi 李成市 of Waseda University 早稲田大学 learned of the P’yŏngyang Analects after its discovery had been mentioned at a conference. He applied for permission to visit North Korea, which he received in 2005. Regrettably, the person responsible for the Analects manuscript was not in P’yŏngyang at the time, so Yi Sŏng-shi returned from North Korea having been unable to view it. After a three-year lull, he happened to visit Tsuruma Kazuyuki, who showed him a photo of the Analects bamboo strips. Together with Yun Yong-gu 尹龍九 and Kim Kyŏng-ho 金慶浩, Yi Sŏng-shi then set out to study the P’yŏngyang Analects based on the photo he saw in Tsuruma Kazuyuki’s office. In the process, the three scholars discovered that a similar photo had already been published as early as in 2001 in the Bulletin of the Koguryŏ Society 高句麗会会報 (issue 63), so they used both images for their analysis. In 2009 the North Korean Academy of Social Science gave Yi Sŏng-shi, Yun Yonggu, and Kim Kyŏng-ho permission to show Tsuruma Kazuyuki’s photo to the wider academic world. Hence, thanks to the combined efforts of North Korean, South Korean, and Japanese scholars and institutions, the world now finally has a chance to learn about the P’yŏngyang Analects. Given this remarkable background, research on the P’yŏngyang Analects is still in its infancy. Yi Sŏng-shi, Yun Yong-gu, and Kim Kyŏng-ho coauthored an article in Korean, published in 2009, in which Yi Sŏng-shi details how they came to study the photographs of the manuscript, Kim Kyŏng-ho describes typical features of the manuscript and provides an annotated transcription, and Yun Yong-gu discusses the historical value of the Analects manuscript.34 Their article has since been updated and translated into Japanese (2010), and again updated and translated into Chinese (2011).35 Kim Kyŏng-ho also discusses the manuscript in his English article on the spread of Confucianism and Chinese script, which was published in 2011.36 One year later, in 2012, Kim Kyŏng-ho and Yi Yŏng-ho published a comprehensive collection of papers under the catchy title Chiha ŭi Nonŏ, chisang ŭi Non 지하의논어, 지상의논어 (Analects Underground, Analects on Paper), which contains a number of  papers on the P’yŏngyang Analects.37 In addition to these mostly Korean publications, the Chinese scholar Shan Chengbin, who has also worked on the Dingzhou Analects, has an unpublished conference paper on the P’yŏngyang 34 35 36 37

Yi, Yun, and Kim 2009. Yi , Yun, and Kim 2010, 2011. Kim 2011. Kim and Yi 2012.

Confucius’s Sayings Entombed

167

Analects.38 Given that no official report or transcription has been published by the North Korean Academy of Social Science, the writings of scholars mentioned in this paragraph are currently the only sources of information regarding the P’yŏngyang Analects. Particulars about the bamboo manuscript in the present essay are drawn from their work, to which I am deeply indebted. Features of the Manuscript The photograph from Tsuruma Kazuyuki’s collection (fig. 6.3) shows thirtynine bamboo strips. Of these, thirty-one strips correspond to chapter 11 in the received Analects, and eight strips to chapter 12. The photo that was published in the Bulletin of the Koguryŏ Society (fig. 6.4) shows the same bamboo strips as the Tsuruma Kazuyuki photo, plus an additional bunch of strips on the right, fourteen of which are legible. The two photos show that three binding threads originally joined the bamboo strips at the top, middle, and bottom. The threads did not survive, but they did create discoloration across the width of the strips and indentations of uneven size at the right-hand side of each strip. If the manuscript was created specifically for the burial, Kim Kyŏng-ho reasons, one would expect indentations of fairly similar shape.39 Their unequal shapes may suggest that the manuscript was repeatedly rolled and unrolled for reading, allowing the binding threads to cut deeper into some strips than others. If Kim Kyŏng-ho’s hypothesis holds, this could potentially make the P’yŏngyang manuscript the earliest Analects copy, depending on how long it had been in use prior to its entombment. The graphs are neatly written between the binding threads. One way of explaining this, as Kim Kyŏng-ho does, is that the bamboo strips were joined into a bundle before the text was copied on them.40 Another possibility, I would add, is that the scribe marked the position of the binding threads before copying the text onto the bamboo strips and bundled the strips only after the text was fully copied on them. At present, our knowledge of the process of text copying in early China is insufficient to make conclusive statements in this regard. Moreover, for the P’yŏngyang Analects we would require more and higher-quality photographs. The medial binding thread divides the text on the strips into two halves. Bamboo strips corresponding to chapter 11 of the Analects contain ten graphs above the medial thread and ten graphs below, totaling twenty graphs per 38 39 40

Shan 2011. Yi, Yun, and Kim 2011: 171. Yi, Yun, and Kim 2011: 170.

Figure 6.3 Photograph of the P’yŏngyang Analects (Tsuruma Kazuyuki Collection; Yi Sŏng-shi, Yun Yong-gu, and Kim Kyŏng-ho 2009: 131)

168 Van Els

Figure 6.4 Photograph of the P’yŏngyang Analects (Bulletin of the Koguryŏ Society; Yi Sŏng-shi, Yun Yong-gu, and Kim Kyŏng-ho 2009: 131)

Confucius’s Sayings Entombed

169

170

Van Els

strip. Bamboo strips corresponding to chapter 12 of the Analects contain nine graphs above and nine graphs below, totaling eighteen graphs per strip. Each section in the P’yŏngyang Analects starts on a new bamboo strip and is preceded by a black dot to mark the beginning of the section. Similar to the Dingzhou Analects, when a section ends before the end of a strip is reached, the remainder of the bamboo is left uninscribed. The manuscript contains some amusing textual peculiarities which clearly reveal a scribe at work. For instance, some graphs are written small and squeezed in between other graphs. Kim Kyŏng-ho plausibly suggests that they were initially forgotten and inserted later, as is common in manuscripts of that period.41 On some bamboo strips the scribe also decreased the spacing between graphs or omitted words so as to fit an entire section on a strip and avoid wasting an extra strip on the last couple of graphs of the section. The omitted words were mostly grammatical particles. This feature is noted by Kim Kyŏngho, who observes that some bamboo strips contain fewer particles than the equivalent text in the received Analects, particularly when a section covers the entire length of the strip.42 Thus, it seems that the arrangement of the text on the writing material somewhat influences the content of the text.43 Differences from the Received Analects One notable textual difference between the P’yŏngyang Analects and the received text is that the manuscript fairly consistently attributes statements to “Master Kong” rather than to “the Master,” particularly on bamboo strips corresponding to chapter 11. Take, for instance, strip 27:44 Master Kong said, “What is You’s zither doing inside my gate?” His other disciples ceased to treat Zilu [= You] with respect. Master Kong 孔子曰由之瑟奚爲於丘之門門人不敬子路孔子

In 11/15 of the received Analects we find the same passage, with identical wording, but with quotations ascribed to “the Master.” As we saw earlier, the Dingzhou Analects likewise prefers “Master Kong” on bamboo strips related to chapter 11. Judging by the two manuscripts, then, chapter 11 contained more attributions to “Master Kong” in the Han dynasty than it does now. The impli41 42 43 44

Yi, Yun, and Kim 2011: 168. Yi, Yun, and Kim 2011: 168. See Richter’s essay in the present volume (chap. 7) on the influence of book format on text structure. Yi, Yun, and Kim 2011: 177.

Confucius’s Sayings Entombed

171

cations of this observation are unclear, for there are no apparent differences between “Master Kong” and “the Master” attributions with respect to their content. Apart from these conspicuous differences, major trends are hard to observe when comparing the P’yŏngyang Analects with other Analects. Detailed analyses by Kim Kyŏng-ho and Shan Chengbin show that the P’yŏngyang and Dingzhou manuscripts sometimes share a textual variant that is not found in other Analects versions, while at other times the P’yŏngyang manuscript resembles the received text where the Dingzhou manuscript has a textual variant, and then there are instances where the two manuscripts and received editions are all different.45

Provenance of the Manuscripts

One important issue in present-day Analects scholarship is the position of the Han dynasty bamboo manuscripts among the various Analects versions. The issue is normally discussed in a conceptual framework that dates back to the Han dynasty. Historiographical sources of that period, such as the Hanshu 漢書, mention three Analects versions circulating in the Western Han dynasty: (1) the so-called “Lu Analects” 魯論, a version in twenty chapters from an exegetical tradition in the ancient state of Lu 魯; (2) the so-called “Qi Analects” 齊論, a version in twenty-two chapters from an exegetical tradition in the ancient state of Qi 齊; and (3) the so-called “Ancient-Script Analects” (guwen Lun­ yu 古文論語), or “Ancient Analects” (gu Lun 古論) for short, a version in twenty-one chapters that was allegedly copied in the Warring States era and hidden in a wall of Confucius’s former mansion, where it was discovered in the Western Han dynasty (or so the story goes), by which time its script had become outdated (hence “ancient”).46 To which of these three versions are the two excavated manuscripts affiliated? Liu Laicheng suggests that the Dingzhou Analects is probably a copy of the Lu Analects.47 Shan Chengbin concurs and provides further support for this claim.48 Li Xueqin, however, notes the manuscript’s chapter division and considers the likelihood of a connection with the Qi Analects somewhat higher.49 Sun Qinshan, on the other hand, suggests that 45 46 47 48 49

Yi, Yun, and Kim 2011; Shan 2011. Makeham (2003: 363–377) offers a detailed description of these Analects. Hebei sheng wenwu yanjiusuo 1997b: 61. Shan 2002: 124. Li Xueqin 2001: 422.

172

Van Els

it resembles the Ancient Analects.50 The problem with these hypotheses, as Li Ruohui perceptively points out, is that all modern Analects editions are conflations.51 Since no one has ever seen an actual Qi, Lu, or Ancient Analects, how are we to associate the excavated manuscripts with any one of these lineages? Accordingly, a growing number of scholars nowadays subscribe to the idea of the bamboo manuscripts as independent copies of the Analects that existed alongside the three main lineages and that may have been related in one way or another to one or more of those lineages, although the exact nature of the relationship can no longer be ascertained.52 The differences between the two manuscripts and other versions of the Analects suggest that a discussion of a Han dynasty manuscript within the Lu/Qi/Ancient Analects framework is likely to be ineffective and that excavated materials ought to be studied in their own right. To gain a better understanding of the two excavated manuscripts, I propose to study them from the perspective of three interrelated questions: when, where, and why were the Analects copied onto the bamboo strips? When were the Manuscripts Copied? The inscribed bamboo strips were found in tombs dating from the mid-first century BCE, but how old were the two manuscripts when they were placed in their tombs? Let us start with the Dingzhou Analects. Scholars nowadays identify the “early Han dynasty” (Han chu 漢初) as the date of the manuscript.53 In fact, this date is so firmly accepted in present-day Analects studies that only one scholar felt the need to support it with evidence. In his study of taboo characters in the Dingzhou Analects, Chen Dong observes that the manuscript avoids mentioning the personal name of Liu Bang 劉邦 (r. 202–195 BCE), founder of the Han dynasty. There are over a dozen instances where the received text contains the word bang 邦 (state), which is written as guo 國 in the bamboo manuscript.54 Chen Dong also notes that the manuscript does not avoid the personal names of later emperors. This leads him to conclude that the text was copied before the names of the later emperors were tabooed—in other words, before these men became emperors. Whether explicitly support50 51 52 53 54

Sun 1998: 4. Li Ruohui 2006: 20. These scholars include Chen Dong (2003), Li Ruohui (2006), Tang Minggui (2007), and Ma Yumeng (2010). In a similar vein, Makeham (2003: 368) suggests it may be a hybrid text. These scholars include Chen Dong (2003), Tang Minggui (2007), Zheng Chunxun (2007), and Ma Yumeng (2010). Chen Dong 2003. In fact, the graph bang 邦 is used only once on the surviving bamboo fragments. Chen Dong explains this single occurrence as a mistake by the modern editors of the transcription.

Confucius’s Sayings Entombed

173

ing Chen Dong’s hypothesis or not, many scholars now maintain that the Dingzhou Analects dates from the very beginning of the Han dynasty. What does this date imply? How persuasive is the evidence? One implication is that at the beginning of the Han dynasty, the Analects already existed as a text that closely resembles its current form. The vast majority of present-day Analects scholars will have no problem with this.55 They concur that the Analects was, as Tang Minggui puts it, “basically formed as a book within 100 years following the death of Confucius” (在孔子去世後100年之內已 基本成書)—in other words, long before the founding of the Han dynasty.56 The scholarly consensus is perplexing, as is this date. Let us look at the basic facts. The two excavated manuscripts, the earliest representations of the Analects we have, reveal that the text had by and large acquired its current form when the bamboo strips were placed in their respective tombs, around 50 BCE. To postulate that the Analects had already acquired this form a full three centuries earlier is quite a stretch, one that requires solid evidence.57 Surprisingly, claims that the Analects was created within a hundred years after Confucius’s death are scarcely ever supported by evidence, let alone evidence from archaeological finds. It seems to me that the bamboo may have added an air of ancientness and authenticity to the Analects, which reinforced preconceived notions about the text’s date. In actual fact, the excavated manuscripts in no way prove or even remotely hint at a date close to the passing of the Master. If, hypothetically, the Analects did exist that early, then given Confucius’s renown we would reasonably expect to find references to the Analects, or to the Master’s sayings contained therein, in texts reliably dated between the mid-fifth century BCE (when Confucius crossed the great divide) and the mid-first century BCE (when the two tombs were closed). Scholars who have scrutinized those writings, such as Makeham and Hunter, astonishingly conclude that no pre–Han dynasty texts and few Western Han dynasty texts mention the Analects and that statements attributed to Confucius in pre-Han and Western Han texts differ markedly from those attributed to Confucius in the Analects. They therefore situate the formation of the Analects well into the Han dynasty, 55 56

57

There are, as usual, scholars with alternative views, such as Zhao Zhenxin (1961) and Zhu Weizheng (1986). Tang 2007: 50. In his critical overview of various contemporary Chinese perspectives on the composition and date of the Analects, Makeham’s essay in the present volume (chap. 1) discusses the gradual development of a Chinese hypothesis about a large proto-Analects corpus, compiled by Confucius’s disciples. Confucius reportedly died in 479 BCE. If the Analects was created within a centrury after his death, this brings us to 379 BCE, which is over three centuries before the closing of the two tombs around 50 BCE.

174

Van Els

perhaps even as late as 140 BCE, even though it may consist in part of pre–Han dynasty material.58 Another implication of the Dingzhou Analects’ supposed “early Han” date is that the manuscript was transmitted for a century and a half between its inception (around 200 BCE under Liu Bang) and its interment (around 50 BCE). This scenario, (implicitly) supported by all those who favor the “early Han” date, leads to many other fascinating questions. Was it common practice in those days to hand down bamboo manuscripts over such an extensive period? What would be the underlying rationale? Did people have a penchant for antiques? Did they treasure their books and pass them on to their offspring as heirlooms? Did people actually read these antique books? Would it not have been more practical to create new copies from time to time? If the bamboo manuscript was indeed kept in the family of the king of Zhongshan for all this time, why then was it taken out of circulation by putting it in his tomb? And why did the manuscript not accompany an earlier holder into his grave? Such questions may not be answerable, but in my opinion they should at least be explicitly reflected upon when proposing an early Han date. One possibility, proposed by Ho Yung-chin, is that the bamboo manuscript is a copy (of a copy of a copy) of an Analects version that dates from the foundational years of the Han dynasty. In other words, not the actual bamboo strips but the text on them was transmitted from the time of Liu Bang.59 While tempting, this view is flawed. If the archetype of this Analects lineage was copied and recopied since the early Han dynasty, copyists faithfully observed the taboo on Liu Bang’s name but ignored taboos introduced after his reign, which essentially invalidates taboo observance as a criterion for dating texts. The early Han dynasty date for the bamboo Analects manuscript rests solely on one piece of evidence: taboo observance. How persuasive is this? It is indeed clear that the scribe makes a conscious attempt to avoid the tabooed name of the dynasty’s founder, but does that necessarily mean the manuscript, or its archetype, dates from his reign? It is also clear that the manuscript does not avoid the personal names of later emperors, but does that mean the manuscript could not have been copied during or after their reigns? What if, for instance, the manuscript was copied in a time or place where taboos for emperors other than the dynasty’s founder were not strictly observed? Of course, this is mere speculation, but it does indicate that the taboo theory may not be watertight and that prudence is in order when applying the taboo criterion in the dating of texts. In this context, Lundahl aptly notes that taboo practices 58 59

Makeham 1996; see also Hunter’s essay in the present volume (chap. 3). Ho 2007: ii.

Confucius’s Sayings Entombed

175

Figure 6.5 Select tracings of Sayings of the Scholars (Hebei sheng wenwu yanjiusuo 1981: 7)

“differed not only between different dynasties, but even between different periods of a single dynasty.”60 In his analysis of name tabooing in the Han dynasty, Adamek notes “many instances of not avoiding taboo in inscriptions and writings,” which can be explained, among other reasons, “by a lax attitude toward tabooing at the time.”61 In view of this ambiguity in taboo practices, I propose to approach the manuscript’s date from a different angle: namely, by looking at the structural form of graphs. The structural form of Dingzhou Analects graphs can be seen on the tracings of bamboo fragments that were published with their transcription. Admittedly, the quantity and quality of the tracings and photographs may not be optimal, but let us see what the materials at hand tell us. First, let us compare the handwriting of the Dingzhou Analects (fig. 6.2) with the handwriting of other manuscripts discovered in the same tomb (figs. 6.5–6.7). The Dingzhou research team published tracings of three other manuscripts: Sayings of the Scholars, Wenzi, and Six Secret Teachings.

60 61

Lundahl 1994: 181. Adamek 2012: 131.

176

Van Els

Figure 6.6 Select tracings of Wenzi (Hebei sheng wenwu yanjiusuo 1995: 28)

The handwriting on all these manuscripts is remarkably similar. Still, a few differences can be observed. For instance, the graphs of the Six Secret Teachings appear to be slightly thicker than the graphs of the other manuscripts, which could be due to a thicker brush or more pressure on the brush in the process of copying. If we zoom in, minute differences can be observed between individual graphs. In the manuscript Sayings of the Scholars (nos. 1 and 2 in fig. 6.8), the upperleft stroke is a dot with a little horizontal “tail” to the right where the brush was lifted from the bamboo. In the Analects (nos. 3 and 4) and the Wenzi (no. 5) manuscripts, the upper-left stroke is written downward. If the tracings are accurate and the small sample is representative, these subtle differences may reveal different hands at work. Apart from minute differences, what is most striking about the tracings is that the handwriting on all tracings of Dingzhou strips is remarkably similar—especially when compared with other manuscripts.

Confucius’s Sayings Entombed

177

Figure 6.7 Select tracings of Six Secret Teachings (Hebei sheng wenwu yanjiusuo 2001: 79)

Figure 6.8 The graph 之 in Dingzhou manuscripts

If the Dingzhou Analects dates from the beginning of the Han dynasty, as is now commonly assumed, one would expect it to resemble other manuscripts from that period. Let us have a look at images and tracings of manuscripts that were placed in tombs in the first century of the Han dynasty, at Mawangdui 馬王堆 (tomb date: 168 BCE), Fuyang 阜陽 (tomb date: 165 BCE), Fenghuangshan 鳳凰山 (tomb date: 156–141 BCE), and Yinqueshan 銀雀山 (tomb date: 140–118 BCE).

178

Van Els

Figure 6.9 The graph 道 in Han dynasty manuscripts

Figure 6.10 The graphs 何 and 可 in Han dynasty manuscripts

There are major differences between manuscripts from these tombs and manuscripts from the Dingzhou tomb. Note, for instance, how 道 dao (the Way) is written in the various manuscripts (fig. 6.9).62 In manuscripts from Mawangdui (no. 1), Fuyang (no. 2), and Yinqueshan (no. 3), the graph is written with two distinct elements: a chuo 辵 (go) component on the left and a shou 首 (head) component on the right. In the Dingzhou manuscripts (nos. 4 and 5), by contrast, the 辵 component is simplified to 辶 and occupies the left and bottom parts of the graph, with the 首 component resting on top of its final stroke. Figure 6.10 provides another example. In manuscripts from Mawangdui (no. 1), Fenghuangshan (no. 2), and Yinqueshan (no. 3), the graphs he 何 (what) and ke 可 (possible) are normally written with a long, elongated final stroke that gradually curves from the upper-right corner to the lower-left corner. In all Dingzhou manuscripts (nos. 4 and 5), by contrast, the final stroke of these graphs is written with a sharp hook to the left. Of course, this may reflect regional variation or scribal preference, but the Dingzhou manuscripts’ rectilinear style, which demands the final stroke to bend to the left rather than to the bottom, more likely signals a development over time, and hence a later date.63 62

63

In figure 9 the graphs from Mawangdui, Fuyang, Fenghuangshan, and Yinqueshan are taken from, respectively, Chen Songchang et al. 2001; Hu and Han 1988; Jilin Da xue 1976; and Pian 2001. Dingzhou graphs are taken from the tracings provided with the transcriptions of the various manuscripts. My use of the word “style” is based on Richter’s (2006) typology of handwriting, which distinguishes three levels: types, styles, and hands. Style, the middle level, refers to the

Confucius’s Sayings Entombed

179

Figure 6.11 The graphs 可 and 何 in the P’yŏngyang Analects

Now let us have a look at the P’yŏngyang Analects, the manuscript that was placed in a tomb in 45 BCE. On those bamboo strips, the graphs 何 he (what) and 可 ke (possible) are written as shown in figure 6.11. Regrettably, high-resolution photographs have not yet been made available. That said, even these blurry images make it clear that, much like with the Dingzhou handwriting, the final stroke ends horizontally to the left. To be sure, these select comparisons are not ironclad proof. They do, however, highlight how similar the Dingzhou and P’yŏngyang manuscripts are to each other, and how distinct they are from manuscripts dating from the first century of the Han dynasty. If the calligraphic style of the two Analects manuscripts is unlike any manuscript from the first century of the Han dynasty, scholars who argue that they were copied under Liu Bang would have to account for this discrepancy. In sum, while evidence is scarce, I would argue that, contrary to the prevailing sentiment in Analects studies, the two bamboo manuscripts date from the mid-first century BCE. Where were the Manuscripts Copied? The two Analects manuscripts were discovered in tombs located in presentday Dingzhou and P’yŏngyang, but that does not necessarily mean the texts were copied onto the bamboo strips at these locations. If we examine where the manuscripts could have been produced, the two most likely possibilities are at places that were either central (at the imperial court in Chang’an) or local (at the seats of power in Zhongshan and Lelang). There is something to be said in favor of both possibilities but none of the arguments are particularly persuasive. In favor of a local manuscript production, one could point to the differences between the two Analects, such as the different ways in which they ascribe quotations to “Master Kong” or “the Master,” as discussed earlier. A centralized reproduction would likely yield more homogeneous results, so a local reprofashion in which a certain type of script is executed. In Richter’s understanding, a style can be typical of a certain school of scribes or even of an entire region or period.

180

Van Els

duction could explain the heterogeneity. That said, the differences between the manuscripts could also be explained as temporal (copied in different periods) or even personal (preferences of different scribes). In further favor of a local manuscript production, one could point to the fact that the tomb in P’yŏngyang yielded a household register in addition to the Analects. The data for such a document must have been gathered in Lelang Commandery, and it would therefore be odd if the text was not composed at the local level. However, the argument that both the Analects and the household register were produced locally would hold only if their manuscripts had similar physical features. This is not the case, if only because the household register was inked on wooden tablets whereas the Analects was written on bamboo. The different writing materials could reflect different values attached to these documents, cheaper wood used for practical administrative documents and bamboo reserved for venerated canonical texts. Yet it could equally reflect different provenances, the wooden document produced locally and the bamboo document sent from elsewhere, perhaps from the capital city. In favor of a central manuscript production, one could point to two documents found in the Dingzhou tomb that must have come from elsewhere. The first document, Record of the King of Lu’an’s Visit to the Imperial Court, is said to tell of the journey made by Liu Ding 劉定, King Miu of Lu’an 六安繆王 (r. 73–50 BCE), to Emperor Xuan’s court in 56 BCE. In this travelogue, King Miu mentions the places he passed through and the distances between them, and he describes the court activities he witnessed or participated in. One could imagine that King Miu sent copies of the travelogue to his peers for them to enjoy. One could also imagine that the document was copied in the capital city and distributed to all kings as a model for their dealings with the emperor. Regrettably, a transcription of the document has not yet been published, so its precise contents are unknown. The second document, also unpublished, is a memorial written by the aforementioned Xiao Wangzhi, the tutor of the imperial crown prince and a known transmitter of the Analects.64 Xiao Wangzhi was a senior scholar at the imperial court who spent most of his adult life in the capital city, and since his memorial was directed to the imperial throne, the document was in all likelihood drawn up in Chang’an and, for whatever reason, taken to Zhongshan as a copy. If these two documents were sent to Zhongshan from the capital city, one could speculate as to whether the same might hold true for the Analects.

64

Hebei sheng wenwu yanjiusuo 1997b: 61. See Loewe 2000: 606–608 for a biography of Xiao Wangzhi.

Confucius’s Sayings Entombed

181

Given the lack of materials available to us, all we can do, unfortunately, is speculate. We have only a limited number of manuscripts from the Dingzhou and P’yŏngyang tombs, and only two of them are Analects. Furthermore, what is true for one manuscript need not necessarily be true for another, as we saw in the case of the household register and the Analects of P’yŏngyang. Even if the handwriting is strikingly similar, as is the case with the Dingzhou manuscripts, that still does not mean that all the manuscripts have the same provenance, for it could simply mean that the various scribes who copied the manuscripts were trained in the same place.65 Why were the Manuscripts Copied? In his article on Confucius and the Analects in the Han dynasty, Csikszentmihalyi mentions renewed interest in the canonical text in the 50s and 40s of the first century BCE.66 This vogue, Csikszentmihalyi explains, may have something to do with the shift that Loewe observed from a modernist to a reformist ideology, that is, from an expansionist and extravagant rule to a more sober, humane, and inward-looking style of government.67 The Analects provided the moral foundation for this new government and was actively promoted for this purpose by Emperors Xuan 漢宣帝 (r. 74–49 BCE) and Yuan 漢元帝 (r. 49–33 BCE). In his article on the spread of Confucianism and Chinese script, Kim Kyŏngho formulates it a little more strongly. He argues that the two documents discovered in the P’yŏngyang tomb testify to the growing influence of a Chinese-script-based bureaucracy and culture. The one document, the household register, uses formulaic expressions (such as “more than the previous [year]” 多前) that were also used in similar inventories of households and populations elsewhere in the Han empire, thus showing a standardized bureaucratic language throughout the empire.68 The other document, the Analects, resembles the canonical text discovered in the Dingzhou tomb. Since both tombs are located far away from the Han dynasty capital city of Chang’an, Kim Kyŏng-ho argues that the two Analects “should be understood in the milieu of Chinese rulers’ heightened efforts to spread Confucianism throughout the empire.”69 It is a well-attested fact, Nylan shows, “that certain emperors, empresses, and ministers were anxious to promote Confucian values,” but that does not 65 66 67 68 69

I thank Ken-ichi Takashima (personal communication) for this observation. Csikszentmihalyi 2002: 146. Loewe 1986: 198. Kim 2011: 63. Kim 2011: 67.

182

Van Els

mean Confucianism existed as a well-defined, uniform state ideology.70 It also does not necessarily mean, I would add, that copies of the Analects were distributed from Chang’an to be read as a vessel of that ideology. Two tomb manuscripts simply do not provide enough evidence for that. The limitations of the evidence obtained from the two tombs demand that we be cautious in our formulations. Yet it is probably safe to say that the two unearthed Analects copies bear witness to the newfound interest in the canonical text in the mid-first century BCE. The documents may have informed the political views of the occupants of the tombs and hence played a part in their political lives. Then again, the deceased may also have read the Analects for personal moral selfcultivation. They may have cherished the text for being part of the education they had received in the early years of their lives. Or the trendy Analects might have been placed in the tombs simply to show that the tomb owners were au courant with their social echelon.

Conclusion

We are fortunate to have two Han dynasty Analects manuscripts at our disposal.71 They offer a fascinating glimpse into the manuscript culture of the Western Han dynasty, and they bear witness to the popularity of the Analects in that period. While the manuscripts provide valuable insights, the current state of the manuscripts, the shortcomings of official publications, and the limitations of the field of early Chinese manuscripts at large force us to be careful in our analyses. Consequently, this present essay explores many different possibilities and offers more questions than positive conclusions. We can only hope that as the study of early Chinese manuscripts advances and more high-quality publications of the two Analects manuscripts appear (with color photos and methodologically accurate transcriptions), we can come to more definitive conclusions. Meanwhile, rather than building shaky hypotheses on scanty premises, it may be preferable to clearly delineate the boundaries of our present state of knowledge. While this may be somewhat disappointing, it is consonant with the views espoused by Confucius, who is believed to have said, “to recognize what you know as what you know, and recognize what you do not know as what you do not know, this is true knowledge” (知之為知之,不知為 不知,是知也).72 70 71 72

Nylan 1999: 22. As mentioned above, a recent discovery may have yielded a fragmentary third bamboostrip Analects manuscript from the same period (mid-first century BCE). Analects 2/17. Translation based on Slingerland 2003: 13.

Confucius’s Sayings Entombed



183

Bibliography

Adamek, Piotr. 2012. “A Good Son Is Sad if He Hears the Name of His Father: The Tabooing of Names in China as a Way of Implementing Social Values.” PhD diss., Leiden University. Ames, Roger, and Henry Rosemont Jr. 1998. The Analects of Confucius: A Philosophical Translation. New York: Ballantine Books. Baxter, William H, and Laurent Sagart. 2011. “Baxter-Sagart Old Chinese Reconstruction.” Version of February 20, 2011. http://crlao.ehess.fr/docannexe.php?id=1202, accessed August 14, 2012. Boltz, William G. 1999. “The Fourth-Century B.C. Guodiann Manuscripts from Chuu and the Composition of the Laotzyy.” Journal of the American Oriental Society 119.4:590–608. Chen Dong 陳東. 2003. “Guanyu Dingzhou Han mu zhujian Lunyu de jige wenti” 關於 定州漢墓竹簡《論語》的幾個問題. Kongzi yanjiu 孔子研究 2003.2:4–13. Chen Sipeng 陳斯鵬. 2001. “Dingzhou Han mu Lunyu jiaodu juli” 定州漢墓《論語》校 讀舉例. Jianbo yanjiu 簡帛硏究 2001.1:385–391. Chen Songchang 陳松長 et al. 2001. Mawangdui jianbo wenzi bian 馬王堆簡帛文字編. Beijing: Wenwu chubanshe. Csikszentmihalyi, Mark. 2002. “Confucius and the Analects in the Hàn.” In Confucius and the Analects: New Essays, ed. Bryan W. Van Norden, 134–162. Oxford: Oxford University Press. Feng Shengjun 馮勝君. 2007. Guodian jian yu Shangbo jian duibi yanjiu 郭店簡與上博 簡對比研究. Beijing: Xianzhuang shuju. Galambos, Imre. 2006. Orthography of Early Chinese Writing: Evidence from Newly Excavated Manuscripts. Budapest: Department of East Asian Studies, Eötvös Loránd University. Guojia wenwu ju 國家文物局 et al. 1981a. “Dingxian 40 hao Han mu chutu zhujian jianjie” 定縣 40 號漢墓出土竹簡簡介. Wenwu 文物 1981.8:11–13. Guojia wenwu ju 國家文物局 et al. 1981b. “‘Rujia zhe yan’ shiwen” 《儒家者言》釋文. Wenwu 文物 1981.8:13–19 (with an outline of the manuscript on 20–22). Hebei sheng bowuguan 河北省博物館 et al. 1976. “Dingxian 40 hao Han mu chutu de jinlü yuyi” 定縣 40 號漢墓出土的金縷玉衣. Wenwu 文物 1976.7:57–59. Hebei sheng wenwu yanjiusuo 河北省文物研究所. 1981. “Hebei Dingxian 40 hao Han mu fajue jianbao” 河北定縣 40 號漢墓發掘簡報. Wenwu 文物 1981.8:1–10. Hebei sheng wenwu yanjiusuo 河北省文物研究所. 1995. “Dingzhou Xi-Han Zhongshan Huai wang mu zhujian Wenzi shiwen” 定州西漢中山懷王墓竹簡《文子》釋文. Wenwu 文物 1995.12:27–34 (with collation notes on 35–37 and 40 and an essay on the arrangement and significance of the bamboo strips on 38–40).

184

Van Els

Hebei sheng wenwu yanjiusuo 河北省文物研究所. 1997a. Dingzhou Han mu zhujian Lunyu 定州漢墓竹簡《論語》. Beijing: Wenwu chubanshe. Hebei sheng wenwu yanjiusuo 河北省文物研究所. 1997b. “Dingzhou Xi-Han Zhong­ shan Huai wang mu zhujian Lunyu shiwen xuan” 定州西漢中山懷王墓竹簡《論 語》釋文選. Wenwu 文物 1997.5:49–54 (with collation notes on 55–58 and an introduction on 59–61). Hebei sheng wenwu yanjiusuo 河北省文物研究所. 2001. “Dingzhou Xi-Han Zhongshan Huai wang mu zhujian Liu Tao shiwen ji jiaozhu” 定州西漢中山懷王墓竹簡《六 韜》釋文及校注. Wenwu 文物 2001.5:77–83 (with an essay on the arrangement and significance of the bamboo strips on 84–86). Henricks, Robert G. 1990. Lao-Tzu Te-Tao Ching: A New Translation Based on the Recently Discovered Ma-wang-tui Texts. London: Bodley Head. Ho Yung-chin 何永欽. 2007. “Dingzhou Han mu zhujian Lunyu yanjiu” 定州漢墓竹簡 《論語》研究. MA thesis, National Taiwan University. Hu Pingsheng 胡平生 and Han Ziqiang 韓自強. 1988. Fuyang Han jian Shijing yanjiu 阜陽漢簡《詩經》研究. Shanghai: Shanghai guji chubanshe. Jilin Da xue lishixi kaogu zhuanye fu Jinancheng kaimen banxue xiao fendui 吉林大學 歷史系考古專業赴紀南城開門辦學小分隊. 1976. “Fenghuangshan 167 hao Han mu qian ce kaoshi” 鳳凰山一六七號漢墓遣冊考釋. Wenwu 文物 1976.10:38–46. Kim Kyŏng-ho 金慶浩. 2011. “A Study of Excavated Bamboo and Wooden-Strip Analects: The Spread of Confucianism and Chinese Script.” Sungkyun Journal of East Asian Studies 11.1:59–88. Kim Kyŏng-ho 김경호 and Yi Yŏng-ho 이영호. 2012. Chiha ŭi Nonŏ, chisang ŭi Non 지 하의논어, 지상의논어. Seoul: Sŏnggyun’gwan Taehakkyo Ch’ulp’anbu. Lau, D. C. 1979. The Analects. London: Penguin Books. Li Ruohui 李若暉. 2006. “Dingzhou Lunyu fenzhang kao” 定州《論語》分章考. Qi Lu xuekan 齊魯學刊 2006.2:20–23. Li Xueqin 李學勤. 2001. “Bajiaolang Han jian rushu xiaoyi” 八角廊漢簡儒書小議. In Jian bo yiji yu xueshushi 簡帛佚籍與學術史, 418–429. Nanchang: Jiangxi jiaoyu chubanshe. Liang Tao 梁濤. 2005. “Dingxian zhujian Lunyu yu Lunyu de chengshu wenti” 定縣竹簡 《論語》與《論語》的成書問題. Guanzi xuekan 管子學刊 2005.1:98–102. Loewe, Michael. 1986. “The Former Han Dynasty.” In Cambridge History of China, vol. 1, The Ch’in and Han Empires, 221 B.C.–A.D. 220, ed. Dennis Twitchett and Michael Loewe, 103–222. Cambridge: Cambridge University Press. Loewe, Michael. 2000. A Biographical Dictionary of the Qin, Former Han and Xin Periods (221 BC–AD 24). Leiden: Brill. Lundahl, Bertil. 1994. “Tehui 特諱: A Special Kind of Chinese Language Taboo.” In Outstretched Leaves on His Bamboo Staff: Studies in Honour of Göram Malmqvist on

Confucius’s Sayings Entombed

185

His 70th Birthday, ed. Joakim Enwall, 177–188. Stockholm: Association of Oriental Studies. Luo Jiyong 羅積勇. 2006. “Dingzhou Han mu zhujian Lunyu wuzi kaoding yize” 定州漢 墓竹簡《論語》誤字考訂一則. In Jianbo yuyan wenzi yanjiu 簡帛語言文字研究, ed. Zhang Xiancheng 張顯成, 210–214. Chengdu: Bashu shushe. Ma Yumeng 馬玉萌. 2010. “Dingxian Han mu zhujian Lunyu yiwen yanjiu” 定縣漢墓竹 簡《論語》異文研究. MA thesis, East China Normal University. Makeham, John. 1996. “The Formation of Lunyu as a Book.” Monumenta Serica 44: 1–24. Makeham, John. 2003. Transmitters and Creators: Chinese Commentators and Com­ mentaries on the Analects. Cambridge, MA: Harvard University Asia Center. Makeham, John. 2006. “A New Hermeneutical Approach to Early Chinese Texts: The Case of the Analects.” Journal of Chinese Philosophy 33, Issue Supplement s1:95–108. Originally published in Sophia 41.1 (2002): 55–69. Mansvelt Beck, Burchard. 1987. “The First Emperor’s Taboo Character and the Three Day Reign of King Xiaowen.” T’oung Pao 73:68–85. Nylan, Michael. 1999. “A Problematic Model: The Han ‘Orthodox Synthesis,’ Then and Now.” In Imagining Boundaries: Changing Confucian Doctrines, Texts, and Herme­ neutics, ed. Kai-wing Chow, On-cho Ng, and John B. Henderson, 17–56. Albany: State University of New York Press. Pian Yuqian 駢宇騫. 2001. Yinqueshan Han jian wenzi bian 銀雀山漢簡文字編. Beijing: Wenwu chubanshe. Richter, Matthias. 2003. “Suggestions concerning the Transcription of Chinese Manuscript Texts: A Research Note.” International Research on Bamboo and Silk Documents: Newsletter 3.1:1–12. Richter, Matthias. 2006. “Tentative Criteria for Discerning Individual Hands in the Guodian Manuscripts.” In Rethinking Confucianism: Selected Papers from the Third International Conference on Excavated Chinese Manuscripts, Mount Holyoke College, April 2004; 儒學的再思考:第三屆國際簡帛研討會論文集, ed. Xing Wen 邢文, 132– 147. San Antonio, TX: Trinity University. Shan Chengbin 單承彬. 2002. “Dingzhou Han mu zhujian ben Lunyu xingzhi kaobian” 定州漢墓竹簡本《論語》性質考辨. Kongzi yanjiu 孔子研究 2002.2:29–38, 124. Shan Chengbin 單承彬. 2011. “Pingrang chutu Xi-Han Lunyu zhujian jiaokanji” 平壤出 土西漢《論語》竹簡校勘記. Paper presented at the “Di Si Jie Zhongguo Jingxue Guoji Xueshu Yantaohui” 第四屆中國經學國際學術研討會, Taiwan Daxue Wen­ xueyuan 臺灣大學文學院, Zhongyang Yanjiuyuan Zhongguo Wenxue Yan­jiusuo Heban 中央研究院中國文學研究所合辦, March 18–19, 2011. Slingerland, Edward. 2003. Confucius: Analects. Indianapolis: Hackett. Sun Qinshan 孫欽善. 1998. “Qianyan” 前言. In Sibu yaoji zhushu congkan Lunyu 四部要 籍註疏叢刊《論語》. Beijing: Zhonghua shuju.

186

Van Els

Sun Qinshan 孫欽善. 2007. “Dingzhou Han mu zhujian Lunyu jiaokan zhixia” 定州漢墓 竹簡《論語》校勘指瑕. Wenxian jikan 文獻季刊 2007.2:149–152. Tang Minggui 唐明貴. 2007. “Dingzhou Han mu zhujian Lunyu yanjiu gaishu” 定州漢墓 竹簡《論語》研究概述. Guji zhengli yanjiu xuekan 古籍整理研究學刊 2007.2: 48–50. Van Els, Paul. 2009. “Dingzhou: The Story of an Unfortunate Tomb.” Asiatische Studien / Études asiatiques 63:909–941. Wang Su 王素. 1998. “Hebei Dingzhou chutu Han jianben Lunyu xingzhi xintan” 河北 定州出土西漢簡本《論語》性質新探. Jianbo yanjiu 簡帛硏究 3:459–470. Weingarten, Oliver. 2009. “Textual Representations of a Sage: Studies of Pre-Qin and Western Han Sources on Confucius (551–479 BCE).” PhD diss., University of Cam­bridge. Xing Wen 邢文. 2005. “Towards a Transparent Transcription.” Asiatische Studien / Études asiatiques 59.1:31–60. Xu Fuchang 徐富昌. 2006. Jianbo dianji yiwen cetan 簡帛典籍異文側探. Taipei: Guojia chubanshe. Yang Chaoming 楊朝明. 2003. “Xinchu zhushu yu Lunyu chengshu wenti zai renshi” 新 出竹書與《論語》成書問題再認識. Zhongguo zhexue shi 中國哲學史 2003.3:32– 39. Yi Sŏng-shi 李成市, Yun Yong-gu 尹龍九, and Kim Kyŏng-ho 金慶浩. 2009. “P’yŏngyang Chŏngbaek-tong 364 kobun ch’ult’o chukkan ‘Nonŏ’ e taehayŏ” 平壤貞柏洞 364 號 墳출토竹簡『論語』에대하여. Mokkan gwa munja 木簡과文字 2009.4:127–166. Yi Sŏng-shi 李成市, Yun Yong-gu 尹龍九, and Kim Kyŏng-ho 金慶浩. 2010. “Pyongyan Jon-Go dō san roku yon gō fun shutsudo chikkan ‘Rongo’ ni tsuite” 平壤貞柏洞三六 四號墳出土竹簡『論語』について. Chūgoku shutsudo shiryō kenkyū 中国出土資料 研究 14:110–149. Yi Sŏng-shi 李成市, Yun Yong-gu 尹龍九, and Kim Kyŏng-ho 金慶浩. 2011. “Pingrang Zhenbodong 364 hao mu chutu zhujian Lunyu” 平壤貞柏洞 364 號墓出土竹簡《論 語》. Chutu wenxian yanjiu 出土文献研究 10:158–194. Zhao Jing 趙晶. 2005. “Qianxi Dingzhou Han jianben Lunyu de wenxian jiazhi” 淺析定 州漢簡本《論語》的文獻價值. Zhejiang shehui kexue 浙江社會科學 2005.3:150–151, 176. Zhao Zhenxin 趙貞信. 1961. “Lunyu jiujing shi shei bianzuan de” 《論語》究竟是誰編 纂的. Beijing shifan Da xue xuebao 北京師範大學學報 1961.4:1–24. Zheng Chunxun 鄭春汛. 2007. “Cong Dingzhou Han mu zhujian Lunyu de xingzhi kan Han chu Lunyu mianmao” 從《定州漢墓竹簡論語》的性質 看漢初《論語》面貌. Chongqing shehui kexue 重慶社會科學 2007.5:38–40. Zhu Weizheng 朱維錚. 1986. “Lunyu jieji cuoshuo” 《論語》結集脞說. Kongzi yanjiu 孔 子研究 1:40–52.

Manuscript Formats And Textual Structure In Early China

187

Chapter 7

Manuscript Formats and Textual Structure in Early China Matthias L. Richter

In recent years, along with a stronger emphasis on the historical criticism of early Chinese literature, the fluidity and composite nature of most texts, particularly from the pre-imperial period, have garnered increased attention.1 At the same time, the study of the material conditions of text production and reception in early China has been greatly stimulated by substantial manuscript finds since the early 1900s, especially the larger number of manuscripts with literary texts found since the 1970s.2 Inevitably, this inspired discussions about the relatedness of the material aspects of manuscript production and the structure of texts. Since the Lunyu is undoubtedly a composite text par excellence, it is not surprising that early Chinese manuscript formats have been cited as arguments to explain the fact that the Lunyu is a compilation of short, mostly disconnected textual units. It is the purpose of the present chapter to examine the underlying logic and validity of such arguments. Undeniably, material conditions of book production may have an influence on the composition of texts. I am using the term “book” here in the broad sense of the material carrier of one or several texts, regardless of whether in manuscript or print form or in a particular format, such as a codex (i.e., stacked pages, bound together at one edge to form the spine of the book) or a scroll or bound bamboo slips from early China. It would be ill-advised, however, to draw a connection between book production and composition of texts in a generalizing fashion. A failure to indicate in what particular way and to what extent a certain text was influenced by the material conditions of book culture at the time of its composition would invite the speculative application of such an argument as a general truth. For example, the assertion that the develop1 Among others, Li Ling (2008) emphasizes textual fluidity most prominently, and the composite nature of early Chinese texts is discussed in some detail by William G. Boltz (2005). 2 For an overview of the manuscript finds, see Giele 1998–1999, as well as Pian and Duan 2006. On the basis of these newly discovered sources, numerous studies of manuscript culture in early China continued the pioneering earlier works of Wang Guowei (1914), Tsien Tsuen-hsuin (1962), and Chen Mengjia (1980). Among these works, Zhang 2004, Tsien 2004, as well as Hu Pingsheng and Ma Yuehua’s new edition of Wang Guowei 2004, deserve special mention.

© Koninklijke Brill NV, Leiden, 2018 | doi:10.1163/9789004382947_009

188

Richter

ment of the novel has benefited from certain advantages the codex had over the scroll should not lead us to assume that the composition or reception of a novel depended on the codex format.3 After all, most of Charles Dickens’s novels, which can certainly be considered milestones in the development of this genre, were first published in journals in serial mode and reached their readers in installments.4 Moreover, a substantial portion of the audience for Dickens’s novels were not readers at all but listeners.5 In the case of early China it is infinitely more difficult to decide which social and material conditions actually determined the production and reception of a particular text. We have reason to examine in detail the various arguments that attribute the fluidity of early Chinese texts and their composite nature to certain manuscript formats, because such arguments might divert our attention from the actual reasons why some texts not only were composed of short units but also continued to be transmitted in this form, while in other texts such short units were increasingly forged into sustained, logically coherent treatments of a topic. A particularly important source of error that may lead us to overestimate the influence of manuscript formats on the structure of some early Chinese texts lies in the attribution of textual fluidity to the rearrangement of bamboo slips. Theories of this kind are clearly influenced by modern notions of what a text is, how and to what end texts are composed, and how they are used. We should heed John Makeham’s cautioning that “we should bear in mind the crucial role that the reader/interpreter plays in reconstituting historical context,” with regard not just to overconfident claims as to the historical meaning of a text but also to premature interpretations of the materiality of texts.6 3 “The codex is built for nonlinear reading … the way a deep reader does it, navigating the network of internal connections that exists within a single rich document like a novel. Indeed, the codex isn’t just another format, it’s the one for which the novel is optimized. The contemporary novel’s dense, layered language took root and grew in the codex, and it demands the kind of navigation that only the codex provides” (Grossman 2011: 13). 4 For a discussion of how the serial mode of publication and episodic structure of Dickens’s novels interacted, see Allen 2012. Patten (1978) gives a comprehensive and detailed account of the publication history of Dickens’s novels; and Stone (1987) provides details of the process of their composition. 5 “Dickens’s novels were criticized as being ‘literature for semi-illiterates’ because the monthly parts were enjoyed by hundreds of thousands of working-class men and women who only read with difficulty and whose horizons of understanding were limited. Indeed many thousands of ordinary people who vastly enjoyed Dickens could not read at all; they met in groups in a room or the village hall and the parts—sometimes rented and sometimes bought between them—were read aloud by the schoolmaster or a literate neighbour” (Muir 1992: 193). For the role of public readings in the widening of Dickens’s audience, see also Patten 1978: 333–341. 6 See Makeham 2006: 96.

Manuscript Formats And Textual Structure In Early China

189

One of the perceptions involved in our problem is the predominant habit of thinking of texts as entirely dependent on their written form. However, even our modern textual culture abounds with texts that are not necessarily written down or at least not primarily communicated in written form. The rise of aural media such as telephone, radio, and television has made oral communication hardly less prominent in our day than written media, whose importance has also increased with the rise of new technologies such as e-mail and text messaging—the mobile phone exemplifies how closely interconnected the written and oral hemispheres of our textual culture have become. Nevertheless, the types of modern texts that tend to capture the most scholarly attention— namely, authored books or articles in journals or newspapers—are all written texts. Even performative texts such as academic talks and public speeches and theater plays are usually first composed in writing, and they are frequently also published in writing after their performance. So it seems natural to say that a text was written whenever we refer to the composition of this text. In modern English, even in academic writing, the word “textual” has largely become a synonym for “written.” Also in early China studies, many articles and books juxtapose orality with textuality, as if an oral text were not textual. For the study of early Chinese literature, however, in which it is crucial to explore the interdependent, complementary functions of the oral and written modes of text production and transmission, we cannot narrow the concept of textuality to the sphere of writing and thus assume an absolute dependence of a text on its written instantiation. Owing to the fortunate circumstance that a sizable number of original manuscripts from early China have become available to students of that period, we finally have a chance to examine the actual potential and limitations of written texts without projecting our modern notions of textual culture back onto earlier periods. In what follows I shall discuss the question of whether the currently available manuscript evidence corroborates ideas that have been proposed about the role that manuscript formats played in the composition, use, and transmission of texts. I shall first look at some hypotheses about the influence of manuscript formats on texts, then name the assumptions that such hypotheses necessarily imply, and finally examine the question of whether the archaeological evidence supports these assumptions.

Hypotheses about the Influence of Manuscript Formats on Texts

In his insightful discussion of the formational history and authorship of the Lunyu, Maurizio Scarpari remarks that “in the Zhou period … texts paralleled a

190

Richter

rich oral tradition”—a way of phrasing which shows that by “texts” he understands written texts.7 Yet, despite his important observation, Scarpari then discusses the variability of early Chinese texts as if the manuscript culture of that time not only determined the written form of these texts but also influenced how texts were composed—a process that includes the gradual compilation of shorter textual units into larger texts: A variable number of characters were written . . . on each strip (jian). The written strips were then fastened together to form units of different size, from ce (small sets of bamboo strips bound together) to the more authoritative and partially “definitive” pian or juan. . . . This way of compiling the written text allowed for the strips, the sets of strips, and the bundles to be assembled and reassembled in different ways, explaining [emphasis added] why the various received versions sometimes display differences in the order of the pian and of the work as a whole, with “books” downgraded as “sections,” or “paragraphs,” or “sentences,” or “passages” within a received text.8 Since in the early Chinese manuscripts now in our possession none of the codicological terms Scarpari names here—ce 冊/策, pian 篇, and juan 卷—occurs with a clear reference to a particular codicological form or unit, it is impossible to determine the meaning of these terms at the time.9 Scarpari shows his awareness of this problem by adding a cautionary footnote but maintains his argument nevertheless.10 Given the lack of evidence for early China, the 7 8 9

10

Scarpari 2007: 448. Scarpari 2007: 448–450. By codicological units I refer to units not of the text but of the material it is written on, defined either by the boundaries of an entire manuscript or by parts of it that are delineated by layout, spacing, or marks and thus intentionally treated by the producers of the manuscript as distinct physical units of this text carrier. Only rarely are the terms ce, pian, and juan used in archaeologically retrieved texts to describe the physical object rather than the text it carried. Examples where the word juan 卷 is clearly used to describe the specific format of a written text occur on the recto of the wooden board YM6D13 found in tomb 6 of Yinwan 尹灣, e.g., Lienü zhuan yi juan 列女傳一卷, Liujia yinyang shu yi juan 六甲陰陽書一卷, and ji yi juan 記一卷. See Lianyungang shi bowuguan et al. 1997: 24, 131. I thank Olivier Venture for bringing these examples to my attention. Erik Maeder (1992: 28) has posited the same hierarchy of ce and pian with more assurance: “Han bibliographers … set out to reassemble and copy afresh writings which had been transmitted to them in the partially ‘finished’ form of pian 篇 (‘bound roll of bamboo slips,’ a technical term which has generally been translated into English as ‘chapter’), that is, as already relatively closed ‘wholes’”; “isolated slips and relatively short ce 冊 (‘bound sets of bamboo slips’ of a size corresponding to internally consistent ‘paragraphs’

Manuscript Formats And Textual Structure In Early China

191

modern use of terms like ce, pian, and juan in discussions of ancient manuscripts is necessarily retrospective; that is, it is guided by what these terms refer to in transmitted literature. Hence, the notions that ce were combined to form pian and that pian were more authoritative are mere conjectures that seem to assume a uniform terminology and production standard of manuscripts. The excavated artifacts do not confirm such uniformity at all. Moreover, the absence of unambiguous indications as to the codicological reference of terms like ce, pian, and juan in the archaeological evidence also means that they are of no consequence for interpreting this archaeological evidence. This codicological terminology becomes important only when we interpret transmitted accounts of texts that actually use such terminology, for example, the bibliographic treatise of the Hanshu 漢書 or Liu Xiang’s 劉向 (79–8 BCE) and Liu Xin’s 劉歆 (46 BCE–23 CE) accounts of their reconstruction of the textual heritage in the imperial library.11 It is the second part of Scarpari’s statement that is interesting for our inquiry. Scarpari is doubtless right in observing that textual units that once used to be understood as complete and independent texts were over time gathered together in more extensive compilations and consequently “downgraded” to dependent parts of a text. What requires closer scrutiny, however, is Scarpari’s assertion that the fluidity of text in early China goes back to peculiarities of manuscript formats, particularly the possibility of rearranging individual bamboo slips or whole sets of these. This theory supposes that a new version of a text was produced by reordering an existing written version of it rather than by writing the text afresh in its changed sequence. This point is consequential, since it implies that the potential independence and movability of smaller units of text are significantly, if not mainly, caused by the material conditions of early Chinese manuscripts rather than by considerations of content and social uses of the texts. An intentional use of the bamboo- or wood-slip type of manuscript as a means of ensuring textual flexibility might even imply that texts were composed by their authors in short, distinct units with the purpose in mind to make them match the smaller codicological units of the manuscripts, in order to facilitate their future rearrangement. An alternative—and somewhat more realistic—explanation ascribes the fact that early Chinese texts mostly consist of short units to the weight and unwieldy

11

within a ‘chapter’) are allowed to roam about more or less freely within the limits of an already constituted pian.” See Hanshu 30.1701–1784. For Liu Xiang’s Bie lu 別錄 (Separate Listings) and Liu Xin’s Qi lue 七略 (Seven Summaries), see the reconstruction of these texts by Yan Kejun 嚴可均 (1762–1843) in his Quan shanggu sandai Qin Han sanguo liuchao wen (Bie lu 38.336a–339b; Qi lue 41.351b–353a).

192

Richter

nature of bamboo manuscripts.12 Here, too, the influence of writing material on text production is often overestimated. For example, a bamboo manuscript with 5,000 characters of text weighed, even by a generous estimation, no more than the latest issue of the journal Early China (33–34, 2010–2011), that is, a little over 600 grams. This can only be a rough estimate, since the weight of a slip must have differed considerably depending on not just its exact dimensions but also the humidity of the area where it was used. My own measurements of slips of the typical dimensions of the 300 BCE manuscripts from Guodian (i.e., 30 cm × 5 mm × 1.5 mm) indicate an approximate weight of 3.5 grams. Assuming an average of 27 characters per slip, a manuscript bearing 5,000 characters would have weighed ca. 650 grams and thus would not have been difficult to transport or carry on one’s person. It would still seem, however, that a bamboo manuscript bearing text of several thousand characters in length would have been impractical for the purpose of frequent perusal. If we assume slips of ca. 30 centimeters in length, a comfortable length for reading purposes, and hence 25–30 characters per slip, a text of 2,000 characters would have required at least 67 slips and the unrolled manuscript would have been almost 50 centimeters wide (calculating 5–6 mm width per slip and 1–2 mm for space between slips to accommodate binding). But we need to keep in mind that reading was not the sole or even primary purpose of a manuscript. Written texts may in some instances have had a primarily representative function.13 Scarpari’s idea that bamboo manuscripts facilitated textual fluidity is obviously influenced by an earlier article by Erik Maeder, from which he quotes extensively. In this article, Maeder compares the bamboo-slip book format with “a loose-leaf ring binder into which miscellaneous material, including both class notes by different hands and documentary handouts, can be entered, only later to be rearranged, shortened or expanded as new material is found which is deemed pertinent, and as the compilers’ concerns change.”14 This idea of Maeder’s has become remarkably popular.15 And Scarpari adds 12

13 14 15

Michael Nylan (2009: 722n7) confirms the currency of this view when, among the working hypotheses held by most scholars in the field, she names the following: “Texts circulated in much smaller units than today, something on the order of a chapter or an essay in today’s books, presumably because of the sheer bulkiness and weight of the bamboo slips.” For a discussion of this phenomenon, see Richter 2011. Maeder 1992: 28. See also Scarpari’s citation of this passage in Scarpari 2007: 450. To name just a few examples, aside from Scarpari’s article, Mark Lewis (1999: 54–55) asserts, “A chapter or even a subsection was an independent unit. New units could be added, old ones removed or the order re-arranged without doing physical violence to the work. As Erik Maeder has described it, Warring States texts worked like a loose-leaf binder into which one inserted essays or notes by different hands, and added, removed, or

Manuscript Formats And Textual Structure In Early China

193

more color to the picture: “Whoever was in possession of the text could make direct changes, if they so desired or considered it necessary, personally inserting their own observations and annotations. The strips could be removed (even if only as a result of damage or loss), added or moved according to personal conviction and present necessity, creating a sense of permanent textual fluidity.”16 Scarpari then combines the idea of texts being written on material that allows for a rearrangement of textual sequence, already implying shortness of textual units, with Mark Lewis’s statement that “authority appeared in the guise of quotation, with the quoted words rendered authoritative by the implicit presence of disciples as audience and scribes.”17 Lewis, too, describes the fluidity of texts, as texts were “invariably social creations passing through numerous hands. The notion of authorship was weak or absent. By contrast, the role of reader or transmitter involved a more active role than that assigned to someone who picks up a modern book. To create any text involved a group of people gathered together, often in a teaching situation, and to preserve or expand it across time required that this initial situation be constantly repeated.”18 The flaw in this otherwise apt characterization of textual culture in early China, especially the connection between quotation and authority, is that Lewis invariably speaks about texts as written texts and does not seem to consider the possibility that many of these early Chinese didactic texts (which he calls “scholarly texts”) were composed, communicated, and transmitted to a large extent either without any reference to a written form or in a predominantly oral textual culture, in which written texts played at best an auxiliary role. It does not seem helpful for the purpose of the present discussion to reenter a fundamental debate about written versus oral textual culture in early China. Critical arguments aiming to balance Lewis’s overemphasis on writing have long been put forth in great detail.19

16 17 18 19

rearranged the material to suit the evolving interests of the compiler.” Kenneth Brashier (2011: 48, 362n6) uses the same analogy (acknowledging Maeder as its originator) to describe the formational history of the Liji: “To borrow an apt analogy, texts such as the Ritual Records were like messy loose-leaf ring binders into which chapters and their commentarial notes were inserted, shuffled and removed; they were not permanently bound, tidy textbooks handed forward from antiquity.” Scarpari 2007: 450. Lewis 1999: 95; cited in Scarpari 2007: 451. Lewis 1999: 55. Michael Nylan (2000: 210–211) has raised this point in her critical appraisal of Lewis’s monumental work: “Lewis collapses verbal speeches, recitations, complex modes of behavior, the technical arts, cosmic regularities … and historical events, seeing all from the single vantage point of ‘writing.’” In a similar vein, Martin Kern (2000: 350–351), referring more specifically to the argument cited above, asks, “when the ‘masters’ were invented in

194

Richter

Of course, it lies in the very nature of oral textual culture that we have no immediate evidence of it. We can only infer from features of transmitted literature and manuscripts of the time, as well as from our knowledge of their historical context, that orality played a major role in the creation and transmission of texts in didactic contexts. What is important for our discussion is the fact that, despite the absence of firsthand evidence of orality, there is also no positive indication that texts were necessarily composed and primarily transmitted in writing. What Lewis says about the implicit presence of disciples in the text as not just audience but also scribes seems to imply live transcriptions of the teacher’s pronouncements during instruction. He is cautious in interpreting literary statements as direct evidence of an actual practice and concedes that “the ‘enunciatory scene’ portrayed in the text … may or may not correspond to the actual ‘communication situation’ in which the recorded phrases were produced” when he asserts that “the writer casts himself in the role of a secretary transcribing the speech of another.”20 Scarpari takes such literary representations of Confucius’s instructions more literally when he cites as a “clear example of this practice” a narrative in Lunyu 15/6, where Zizhang is said to have written his master’s instruction on his sash.21 Scarpari adds a note, explaining the phrase “wrote it on his sash” (shu zhu shen 書諸紳) as referring “to the custom dating back to the earliest times of suspending a bamboo strip document from the sash securing one’s own robe.”22 This sounds more plausible than assuming that Zizhang literally wrote the teacher’s words on his sash. Confucius’s answer is forty-eight characters long and would fit on two average-size bamboo slips. However, the Western Zhou Song-gui 頌簋 bronze inscription, which Scarpari adduces as evidence of this old custom, differs significantly in several points from the Lunyu passage we are trying to understand. First, the inscription refers to the third year of King Xuan’s 宣 reign, that is, 825 BCE, which is over three centuries before the narrated time in the Lunyu.23 Second, the social setting—an

20 21 22 23

their texts, did this necessarily imply the notion of writing these texts, or did the oral teaching in fact gradually generate the image of the master, recreating his original mode of instruction? When the texts began to become transmitted through individual teachers, how important was the role of their written version for matters of instruction … and transmission. … was it indeed the written version that created and sustained the textual lineage …?” Lewis 1999: 57. Scarpari 2007: 452. Scarpari 2007: 452n28. Shaughnessy dates the inscription to 825 BCE, whereas Zizhang’s assumed year of birth is 503 BCE (see Qian 1990: 616); so the narrative in Lunyu 15/6 probably refers to a time no earlier than 490 BCE.

Manuscript Formats And Textual Structure In Early China

195

appointment ritual at the royal court of the Zhou king—is vastly different from that of the Lunyu episode; for the conversations recorded in the Lunyu we have to assume much more modest settings than “a form of transcription, like that performed by mythical historians of the right and left who recorded every word and deed of the king.”24 Finally and most importantly, the bronze inscription describes this practice in completely different words: The appointee “Song knelt in obeisance, touching his head to the ground. He received the [spoken] command and the slips [with a record of it]; and, having suspended these from his sash, went out” (頌拜稽首,受命冊,佩以出).25 It seems implausible that if the Lunyu passage were to describe the same practice, it would be in the words “wrote it on his sash.” So should we then take the Lunyu passage as evidence for Confucius’s disciples noting his teachings on their sashes, even if only in exceptional cases? After all, this is what the text literally says. Such a scenario, however, is unlikely for various reasons. The circumstance alone that Lunyu 15/6 is the only instance where “writing on a sash” is attested in early Chinese literature should give us pause. But there are also technical aspects to consider. The scenario would assume that Zizhang took off his sash during or after instruction to record his teacher’s words on it. Further, the material of the sash must have been of a kind that allowed Zizhang to write at least as many as forty-eight characters on it. To be sure, very large characters could most probably be written on almost any surface, but for the purpose of writing characters of the size typical for early Chinese manuscripts (i.e., ca. 5–8 mm high and 4–5 mm wide), one needed a special fabric made of tightly woven, thin threads; for the writing to stay legible, the surface of the silk would have to have been prepared to ensure that the fine brushstrokes adhered to the surface without being absorbed by the threads.26 It appears most unlikely that Zizhang or anyone else would have worn such delicate, expensive fabric as a sash. Moreover, Zizhang would have needed a brush, ink, some water, an ink slab, and somewhere to place the ink slab while writing. He would also have needed a perfectly even surface on which to spread out the sash. A table might have accommodated both the sash 24 25

26

Lewis 1999: 95. For a translation of the complete inscription, see Shaughnessy in Loewe and Shaughnessy 1999: 298–299. The inscription is published in Ma et al. 1988, where the passage in question is transcribed as 頌拜𩒨首,受令,冊佩㠯出. My reading follows that of Guo Moruo 郭沫若 in Liang Zhou jinwenci daxi 兩周金文辭大系, as cited in Zhong 2010. I thank Wolfgang Behr for bringing this article to my attention. The type of silk used for manuscripts and paintings was typically the thin-threaded juan 絹 or its equally densely woven, double-threaded, and hence somewhat thicker and more expensive variety, jian 縑. (See Chen Songchang 2000: 15.) If Zizhang’s sash was made of silk at all, it would probably have been the coarser zeng 繒.

196

Richter

and the ink. He would have also had to fix the textile on the surface of this table to keep it from moving while writing on it. Even if we could assume that it was possible for Zizhang to record Confucius’s teachings on his sash or that the Lunyu passage uses the words “Zizhang wrote it on his sash” as an ellipsis for “Zizhang wrote it on bamboo slips and suspended these from his sash,” we still need to consider the rhetorical function of this sentence in the Lunyu passage. While the information given in the Song-gui inscription about Song carrying a written record of the king’s decree out of the place of his appointment ceremony is factually credible and appears as one organic narrative element among others in the middle of the text, the Lunyu sentence “Zizhang wrote it on his sash” concludes the short scene of instruction. The narrative function of this conclusion of a brief exchange between master and disciple is clearly not to mention a common practice. If it had been common, it would probably not have been mentioned at all. Rather, Zizhang’s act of writing the words on his sash is narrated as an unusually strong reaction to his teacher’s words. It dramatizes the narrative and thus lends particular emphasis to the teaching. Emphatic conclusions are frequently used in literary representations of Confucius’s teaching in the Lunyu and other texts. They may take the form of an evaluative statement by Confucius or that of the students’ humbly proclaiming their inadequacy and promising to lend their best efforts to enacting Confucius’s teachings.27 For example, in a Kongzi jiayu 孔子家語 (Family Sayings of Kongzi) episode, Zixia responds to Confucius’s instruction with the assurance: “Allow me to remember this and abide by it as long as I live.”28 Emphatic conclusions may also be couched in the narrator’s language or a combination of the narrator’s language and direct speech, for example, “Repeatedly bowing, Zilu said, ‘Reverently have I received your instruction’”; or “When the three disciples had heard these words from their master, they were enlightened, as if awakened from blindness”; or, even more dramatically, “Zixia agitatedly jumped to his feet, stood with his back against

27

28

Lunyu 6/2 concludes with Confucius’s approval of Zhonggong: “Yong’s [i.e., Zhonggong’s] words are to the point” (雍之言然). In Lunyu 5/12, however, the Master chides Zigong for an ambitious pronouncement: “Si, this is not something within your reach” (賜也非爾所 及也). At the end of Lunyu 12/1, Yan Yuan responds to Confucius’s instruction: “Although I am not smart, please allow me to commit myself to these precepts” (回雖不敏請事斯 語矣). Lunyu 12/2 ends in precisely the same words, spoken by Zhonggong. 商請志之而終身奉行焉 (Kongzi jiayu zhuzi suoyin 29.8–9). Alternatively, we could understand this passage as a combination of direct speech and narrator’s speech: “[Zixia said], ‘Allow me to remember this.’ And he abided by it as long as he lived.”

Manuscript Formats And Textual Structure In Early China

197

the wall, and said, ‘Would your disciple dare not to respectfully receive this (instruction)?’”29 Considering these literary conventions in similar accounts of Confucius’s teaching, it seems clear that the conclusion of Lunyu 15/6 with the words “Zizhang wrote it on his sash” does not describe an actual notation, as in the English “noting on one’s shirt-cuff,” which was an attested practice. Rather, it is a rhetorical gesture signifying that Zizhang intended to remember his teacher’s pronouncement at all times—a requirement already included in Confucius’s words and likewise clearly not intended to be taken literally: “when you stand you should have this ideal there in front of you, and when you are in your carriage you should see it leaning against the handle-bar.”30 For the particular passage in the Lunyu or a similar text it may seem unnecessary to examine at such length whether or not a described action is to be understood literally or as a figure of speech. However, whenever we interpret such information as historical reality, this calls for a broader examination of all aspects involved in such an assumption. In the following I shall examine in a similar way whether the manuscript evidence we now have corroborates the idea that bamboo manuscripts were a cause of or determining influence on the structure of the Lunyu or similar early Chinese texts.

The Extension of Texts

The first assumption concerns the extension of texts. Did early Chinese manuscript formats make it necessary for texts to be short or to be composed as a rearrangeable sequence of several short units and then to circulate in such smaller units, because a longer text would have been too unwieldy? Indeed, the conditions set by bamboo and wood manuscripts may have favored shorter texts, not so much for reasons of weight (as explained above) but to keep the size of the manuscript within a range that would allow a reader to unroll and read it while holding it comfortably. When we encounter manuscripts larger than ca. forty slips and of a greater slip length than ca. thirty centimeters, we have reason to doubt that the manuscript was the primary means of access to an otherwise unknown text or that it was used much for reciting and possibly memorizing a text. Manuscripts of such large sizes may have been produced 29 30

子路再拜曰敬而受教 (Kongzi jiayu zhuzi suoyin 38.2–3); 三子者既得聞此論於夫子 也煥若發矇焉 (Liji zhuzi suoyin 137.32); and 子夏蹶然而起負墻而立曰弟子敢不承 乎 (Liji zhuzi suoyin 139.6–7). Lunyu 15/6: 立則見其參於前也,在輿則見其倚於衡也. Translation by D. C. Lau (1992: 151).

198

Richter

primarily for archival or representative purposes and may have played an ancillary role in the actual communication of their content. Despite the fact that bamboo and wood manuscripts could carry only a limited amount of text and still stay within a size practical for convenient reference and reading, we have no reason to assume that these limits restricted the length of texts that were composed in early China. Rather, the length of text depended on content and genre. The reason many of the texts in ancient manuscripts consist of short, rearrangeable units does not lie in their material carrier but rather in their didactic content and in the fact that short textual units are easier to memorize and can be used flexibly in teaching and in persuasive political speech. We have no positive indications that these texts were composed in writing, nor can we know when and to what extent they were laid down in writing and used with reference to such a written record. What we do know for certain is that bamboo manuscripts did not pose an obstacle to composing and recording long texts. We have manuscript evidence of long texts whose constituent parts cannot be rearranged because their narrative content or sustained systematic argument does not allow for any other textual sequence. To name two Chu 楚 manuscripts from ca. 300 BCE as examples, the narrative historical text Rong Cheng shi 容成氏 ([The Minister] Rong Cheng) is over 2,300 characters long, and the text *Xing zi ming chu 性自命出 (Natural Disposition Emerges from the Mandate) extends over 1,550 characters, at least 60 percent of which contain a continuous systematic treatise, the logic of which precludes any rearrangement of its text.31

Correspondence between the Text and Its Carrier

If indeed manuscript formats had facilitated or even motivated the composition of texts in short units and a general fluidity of textual sequence, we would expect certain correspondences between typical lengths of text and codicological units. However, the vastly different slip lengths of the manuscripts 31

The unprovenanced manuscript Rong Cheng shi of the Shanghai Museum collection comprises fifty-three slips, 44 cm in length, each with forty-two to forty-five characters. The title of the text is written on the verso of the last slip. See Shanghai bowuguan cang Zhanguo Chu zhushu, vol. 2. The manuscript *Xing zi ming chu stems from the controlled excavation of tomb 1 of the Guodian 郭店 site. (The asterisk indicates that the title is not present in the manuscript itself but was devised by the editors.) The text is written on sixty-seven slips, 32 cm in length, each bearing twenty-three to twenty-five characters (in exceptional cases up to thirty). See Jingmen shi bowuguan 1998. For further information on the Guodian find, see the articles by Li Boqian and Liu Zuxin in Allan and Williams 2000: 9–32.

Manuscript Formats And Textual Structure In Early China

199

discovered so far do not encourage us to assume any such general correspondence. The dimensions of the slips probably depended on various factors, such as the type of document, the social status of its owner, or different local production standards.32 For example, comparing the manuscripts from the Guodian find with the unprovenanced Shanghai Museum manuscripts, we can establish three size ranges in the Guodian corpus (15.1–17.4, 26.4–28.2, and 30.6–32.2 cm). With one single exception (the 24 cm long slips of *Jian da wang po han 柬大王泊旱 [Great King Jian Stops the Drought]), the manuscripts in the Shanghai Museum collection are all larger than the largest in the Guodian find, ranging from 572–517, 475–426, 400–390, to 339–320 millimeters. Whether the Shanghai Museum manuscripts come from one or several sites, it cannot be coincidental that their sizes differ so markedly from the Guodian find. Since both collections contain the same type of literary and didactic texts with politico-philosophical content, even some instances of different versions of the same text occurring in both collections, the reason for the different size ranges cannot lie in the type of document but must be sought elsewhere. It may be that the Shanghai Museum manuscripts were all produced for people of higher social status than the Guodian tomb occupant, or they may be the products of different groups of stationary producers and scribes whose conventions in manuscript formats differed. It is only in the early empire that more standardized manuscript sizes begin to appear. The preferred sizes of the Mawangdui silk manuscripts, ca. 24 and ca. 50 centimeters, approximate the size standards described by Wang Chong 王充 (27–ca. 100 CE).33 Even then, it seems unlikely that manuscript production across the empire adhered to such standards. Even if we could assume a certain regular slip length, this would still not allow us to determine a regular amount of text that each slip accommodated. The tendency to use relatively uniform character sizes, each taking up the same (approximately square) space, developed only in the early empire and was apparently restricted to manuscripts of a particularly high quality.34 In 32 33

34

Hu Pingsheng (2000) studies the formats of tomb inventories (qiance 遣策), documents (wenshu 文書), literary texts (shuji 書籍), and statutes and decrees (lüling 律令) separately, understanding their formats to be governed by different factors. In Lunheng 論衡, chap. 36, a size of two chi and four cun (二尺四寸, i.e., ca. 55 cm) is specified as the size for manuscripts bearing the “patterned speech of the sages” (sheng­ ren wenyu 聖人文語), which are “propagated and recited by the Classicists” (ruzhe tui du 儒者推讀), while documents of lesser status are said to have had the smaller format of one chi (ca. 23 cm). See Huang 1990: 557 (as well as 551 and 821 for further references to manuscript sizes in chaps. 35 and 57). The late Western Han *Yili 儀禮 manuscripts from Wuwei 武威, Gansu, are examples of such elaborately produced manuscripts. See Gansu sheng bowuguan and Zhongguo ke­ xueyuan kaogu yanjiusuo 2005.

200

Richter

*Laozi A, slip 2, character for si {私}

*Laozi C, slip 8, character for sang {喪}

Figure 7.1 Varying height of characters in Guodian manuscripts

earlier manuscripts, character sizes vary greatly, depending not just on style of script but, more importantly, the content of the text, as different words were written with characters of different sizes. While the narrowness of slips (typically 0.5–0.7 cm) restricted the width of characters, their height varied vastly, depending on their general shape and structural complexity, which led to a vertical rather than horizontal extension. Classifiers placed to the left of the phonophoric in modern script are in Chu script often placed at the bottom. While simple characters like 一, 厶, 已, or the more frequent 亡, 曰, or 之 could require relatively little space, taller ones like 譽, 清, 然, 喪, 爲, 唯, and 事 take up a greater portion of the slip. Within a group of Guodian manuscripts written in the same style of script (i.e., *Laozi 老子 A–C, *Taiyi sheng shui 太一生水 [The Great One Generates Water], *Yucong 語叢 4 [Thicket of Sayings, 4],35 character heights vary from 2.3 millimeters for the character for si {私} on slip 2 of *Laozi A to more than five times higher at 12.2 millimeters for the character for sang {喪} on slip 8 of *Laozi C (see fig. 7.1).36 The space needed for a certain amount of text further depends on character spacing and the amount of punctuation inserted between characters. The number of characters fitted into a certain amount of space is very different in different Warring States manuscripts. For example, the typical ratio in the Guodian manuscripts is ca. 22 characters per 30 centimeters. Yet, within the same corpus of manuscripts, we find the most extreme difference in character density even within four manuscripts of comparable size titled by the editors *Yucong 1–4. While the 15-centimeter-long slips of *Yucong 4 accommodate 15 characters each, the 15- to17-centimeter-long *Yucong 1–3 manuscripts place fewer than 8 characters on each slip. Frequently, a slip of 17 centimeters has no 35 36

The translation of this title follows that of Scott Cook. For an explanation, see Cook 2012: 800–801. This is not counting the exceptional cases of characters added as corrections on the verso of slips in a smaller size or the character 一, which consists of a single, less-than-onemillimeter-thick horizontal stroke but which, due to its slant, occupies up to 1.7 mm of space on the slip.

Manuscript Formats And Textual Structure In Early China

201

more than 2 characters on it (see fig. 7.2). One might be tempted to dismiss the extraordinarily low density of writing in the manuscripts *Yucong 1–3 as an exception, since their very tall characters are written in a special ornamental style reminiscent of weapons inscriptions and are spaced very generously. On the other hand, these manuscripts are exactly the type of document Maeder and Scarpari refer to when they discuss manuscript formats as facilitating textual fluidity. The *Yucong are a collection of apothegms with a low degree of coherence, just like the Lunyu. Striking differences in character density also occur in manuscripts written in a less ornamental style. While most of the Guodian manuscripts on average fit ca. 22 characters in a space of 30 centimeters, the Shanghai Museum manuscript Wu ming 吳命 ([King] Wu’s Command) has as many as 65 characters in a space of 52 centimeters; if written in the average density of the Guodian manuscripts, almost 90 centimeters would have been required. See also Figure 7.2, showing a segment of the manuscript Heng xian 恆先 (Before Permanence) that fits only six characters on a section of slip slightly longer than that one of Wu ming that accommodates as many as ten characters. Looking at the other side of the assumed correspondence between texts and bamboo slips, it is likewise impossible to determine a typical length of textual unit that would regularly match one or several bamboo slips. Taking the Lunyu as an example, textual units are of extremely different lengths, ranging from 5 or 6 characters (10/12: 席不正不坐; 2/12: 子曰君子不器; 10/19: 食不語寢不言) to as many as 273 (16/1) or even 315 (11/26). One could still argue that most textual units of the Lunyu stay within a range of one dozen to ca. three dozen characters, and hence most of them would fit on a single bamboo slip. But even so, they would be rearrangeable only under the condition that every new unit begins on a new slip and that after the end of this unit no text of a different unit follows on the same slip.

Mise-en-Page

If indeed it was a practice in early China to use bamboo manuscripts as receptacles for “miscellaneous material” which could later “be rearranged, shortened or expanded as new material is found,” as Maeder surmises, and if “whoever was in possession of the text could make direct changes, … personally inserting their own observations and annotations,” and if indeed “strips could be removed …, added or moved according to personal conviction and present necessity,” as Scarpari asserts, we would expect the manuscripts to reflect a practice of arranging the text on the slips in a manner that facilitated such

202 *Yucong 4, slip 1 (complete)

Richter *Yucong 1, slip 1 (complete)

Wu ming, slip 1 (section)

Heng xian, slip 2 (section)

Figure 7.2 Varying density of writing in Guodian and Shanghai Museum manuscripts

Manuscript Formats And Textual Structure In Early China

203

rearrangements, and we would also expect to find evidence of insertions or indications that unbinding and rebinding a manuscript was a common practice used to make changes to the text.37 A manuscript of bound bamboo or wood slips that was intended to be rolled up, to be repeatedly unrolled for the purpose of reading, and possibly even to be transported surely needed a secure and stable binding. The binding strings had to be wound around each slip very tightly and fitted as closely as possible into the V-shaped notch cut into the right edge of each slip, in order to hold the binding string in the correct position and to reduce the space between individual slips.38 The tight and accurate binding required to produce a functional bamboo manuscript was probably not a technique that everyone mastered who might have desired to make a change to a text. The high calligraphic quality of most manuscripts with literary texts indicates that manuscript production was, as a rule, in the hands of specially trained craftsmen. Hence, anyone who wanted to extract a slip from a bound bamboo manuscript and insert a new one needed to command the skill of properly binding a manuscript or needed a qualified person to do this for him. To exchange a slip, one needed to unbind the strips, beginning from the end, at least as far as the slip that was to be exchanged—a decidedly more cumbersome process than exchanging a page in a modern ring binder. This of course does not rule out the possibility that slips were exchanged or inserted in a manuscript. Indications of such an occurrence would be either differences in the writing itself or in its carrier. The fact that uniformity on both levels is among the main criteria by which archaeologists reconstruct manuscripts of course reduces the chance of finding examples of inserted or exchanged slips. If any slips of a particular manuscript had looked somewhat different from the adjacent slips at the time the manuscript was interred, this would not be easy to recognize after the excavation two millennia later, and even if we could recognize such a difference, it is still possible that slips of different quality were used when the text was originally written. Nor do differences of handwriting allow us to recognize inserted or exchanged slips. Judging from the few existing detailed studies of scribal hands in Warring States manuscripts, the change of scribal hand does not usually coincide with beginnings and ends of the individual slips and thus cannot possibly reflect exchanged or inserted slips.39 Rather, it was common practice for 37 38

39

See Maeder 1992: 28; Scarpari 2007: 450. Achieving a continuous writing surface with gaps between slips made as narrow as possible was especially important for manuscripts with charts or any sort of lines drawn across several slips, as, for example, in the hemerological Qin manuscripts (*Ri shu 日書) from Shuihudi 睡虎地. See Yunmeng Shuihudi Qin mu bianxie zu 1981: 89–140. See Li Songru 2006, 2012; as well as Richter 2006.

204

Richter

more than one scribe to be involved in writing one manuscript. Consequently, even if the different hands do occur on different slips, this is no reliable indication of inserted or exchanged slips. The absence of positive indications of exchanged or inserted slips in early Chinese manuscripts does of course not exclude such a possibility. But several indications to the contrary make it less probable that such rearrangements were common practice or even a preferred method of making changes to texts. If inserting slips had been a feasible method of making additions to a text, we would expect this method to be used to insert omitted text, instead of writing it on the verso, which, unless on a slip close to one of the ends of the manuscript, required a reader to turn over the entire document to read the missing characters. For example, the scribe of the Guodian *Zi yi 緇衣 (Jet-Black Robes) manuscript omitted, possibly because of an eye-skip, the first of two parallel rhyming sentences from the passage 人[苟有言必聞其聲]苟有行必見其成 (When people speak, be sure to listen to their voices; when they act, be sure to look at the outcome) on slip 40 of forty-seven.40 The seven missing characters are written in smaller size on the verso of the same slip, beginning exactly at the point of the slip where the omission on the recto occurred: that is, between 人 and 苟 (see fig. 7.3.1). The writing on the verso could come from a different hand, or the different character shape may just be an effect of the same scribe writing in a smaller size. In the Guodian manuscript *Yucong 4, an eye-skip error occurred on the very last slip. In the passage 聖君視朝而入[入之或入之 至之或至之]之至而亡及也已, nine of the ten characters added in smaller size on the verso were omitted.41 Interestingly, the last four characters of the text (亡及也已) are also written on the verso, but at the top of the slip (see fig. 7.3.2). At the very least, both examples indicate that inserting a new slip or even just adding one at the end of a manuscript was not a convenient matter of routine. Nevertheless, we cannot exclude the possibility that the owner of a manuscript might want to rearrange its text or insert new text at the cost of having to rebind the slips, in order to save the effort of writing the text anew. The need for rebinding would also arise when the bindings of a manuscript broke. The issue of broken bindings brings to mind the anecdote that portrays Confucius as so exceedingly fond of perusing the Classic of Changes that its

40 41

Cf. the transcription and annotation in Jingmen shi bowuguan 1998: 131, 136. Cf. the transcription and annotation in Jingmen shi bowuguan 1998: 218. Despite numerous and widely diverging attempts to explain this passage of the text, it still remains unclear. Hence, I refrain from attempting a translation.

Manuscript Formats And Textual Structure In Early China Section of slip 40 (recto and verso)

Figure 7.3.1 Correction in Guodian *Zi yi 緇衣

Slip 27 (recto and verso)

Figure 7.3.2 Correction in Guodian *Yucong 4 語叢

205

206

Richter

leather bindings broke three times (讀易韋編三絕).42 Just like the statement about Zizhang writing on his sash, this anecdote should not be taken literally. It makes a rhetorical point, impressing us with its depiction of Confucius as an avid reader and student of the Changes. At the same time, it confirms that users of written texts in early China were familiar with the problem of broken manuscript bindings. This is also reflected in the fact that the producers of manuscripts took measures to define textual identity explicitly and to prevent the confusion of textual sequence. The most effective and unambiguous way of achieving this goal is to number the slips of a bound manuscript. If the currently known manuscript finds are representative, this practice occurred only rarely in the Warring States period but became more common in the early empire. Examples of numbered Warring States slips (dated to around 300 BCE) surfaced in the Qinghua University collection. The unprovenanced manuscripts, *Yin zhi 尹至 ([Yi] Yin’s Arrival), *Yin gao 尹誥 ([Yi] Yin’s Announcement), Qi ye 耆夜 (Feast upon the Victory over Qi), Zhou Wu wang you ji Zhougong suo zi yi dai wang zhi zhi 周武王 有疾周公所自以代王之志 (Zhou Gong’s Resolution to Offer Himself Up in the King’s Stead, when King Wu of Zhou Fell Ill; for convenient reference renamed by the editors *Jin teng 金縢 [The Metal-Bound Coffer], after the corresponding Shangshu chapter), *Huang men 皇門 (The Resplendent Gate), and Ji gong zhi gu ming 祭公之顧命 (Ji Gong’s Testament), as well as *Xi nian 繫年 (Chronicle), all have slip numbers in the middle of the verso.43 Of course, this did not exclude the possibility of intentionally rearranging the slips. One could even erase the slip numbers and write new ones. But again, we have no evidence for such a practice, and at the very least the presence of slip numbers shows that the manuscript cannot have been designed to facilitate rearrangements of its text. Aside from slip numbers, other features restricted rearrangements of slips as well. *Xi nian 繫年 consists of twenty-three separate textual units. The end of each unit is indicated unambiguously by a hook-shaped mark directly after its last character and by leaving the rest of the slip blank, so that each new textual unit begins at the top of a slip. This would still allow for a rearrange42 43

Shiji 47.1937; Hanshu 88.3589. As binding strings were usually made of silk or hemp, the soft leather (wei韋) indicated in this anecdote may serve the rhetorical aim of emphasizing the intensity of Confucius’s engagement with the manuscript. See Li Xueqin 2010–2011. Recently, a number of Qinghua University manuscripts have been published in which the sequence of slips is indicated by a continuous diagonal line incised across the verso of the slips. Apparently, this feature was overlooked in manuscripts published previously. Further studies of this phenomenon will probably confirm that the practice of securing slip sequence was already prevalent in the Warring States period.

Manuscript Formats And Textual Structure In Early China

Figure 7.4.1 Wuwei *Yili manuscript “Fu zhuan” 服傳, section of slips 1 and 2, verso

207

Figure 7.4.2 Wuwei *Yili manuscript “Fu zhuan” 服傳, bottom of slip 28, recto

ment of the textual units, had not the producers of this manuscript numbered the 138 slips of all twenty-three units continuously, so that even if the units were bound separately, their correct sequence remained clear. At the same time, the continuous numbering of the slips of several textual units indicates a textual hierarchy in the sense that these several units are not independent but have to be understood as constituent parts of a larger text. The Wuwei *Yili 儀禮 manuscripts achieve the same effect by numbering the slips of the several textual units separately but in addition assigning each textual unit an ordinal number, which not only indicates that these units are parts of a larger text but also specifies their correct sequence. For example, the sixty slips of the Wuwei *Yili manuscript titled “Fu zhuan” 服傳 are marked as “eighth” (di ba 第八) on the verso of slip 1 and so assigned a certain position in a sequence of chapters of the *Yili manuscripts that were all bound separately; the title “Fu zhuan” 服傳 is written on the verso of slip 2 (see fig. 7.4.1). “Fu zhuan” is divided into ten codicologically separate parts, each beginning at the top of a new slip, usually marked by an initiator dot, and ending with a blank after the last character of the unit. This type of layout would allow an easy transposition of these ten units. However, the manuscript numbers all sixty slips consecutively and thus ensures that the order of the ten units remains unchanged (e.g., the slip number at the bottom of slip 28, recto; see fig. 7.4.2).44 To be sure, such a degree of explicitness in defining textual identity is typical for the early empire and rarely to be found in the Warring States period. However, even at that earlier time levels of textual hierarchy were frequently indicated by means of punctuation. For example, the twenty-three textual units of the Guodian *Zi yi manuscript are all explicitly marked by section terminator marks: small black squares at the right edge of the slip immediately 44

See Gansu sheng bowuguan and Zhongguo kexueyuan kaogu yanjiusuo 2005: table 2.

208

Richter

below the last character of the unit. The characters 二十又三 in the middle of the last slip are divided by some extra space from the last of these section terminator marks (see fig. 7.5.1). These numbers, together with the spacing, fulfill two functions: they serve as a text terminator and at the same time they explicitly indicate the number of sections, as a means to control completeness of the text but also to prevent any possible addition or excision of textual units. Although—unlike the examples from Wuwei or the Qinghua University collection—this Guodian manuscript has no slip numbers, it still precludes any rearrangements of its twenty-three textual units by means of layout. The ends of these units never coincide with the ends of slips. This alone would show us only that the producers of the Guodian manuscript *Zi yi had no intention to facilitate future changes to the text. But their interest in defining textual integrity went further than that: by specifying the number of sections on the last slip, they made it impossible for textual units to be added or subtracted unnoticed. Of course, such measures did not prevent texts from getting changed. The transmitted counterpart of this text, the Liji 禮記 (Record of Rituals) chapter “Zi yi,” has two sections that are absent in the manuscript, it combines two sections of the manuscript into one, and the sections it has in common with the manuscript text are in a radically different sequence.45 Edward L. Shaughnessy makes a strong argument to the effect that the changed sequence is a result of a physical reordering of a bamboo manuscript version of the text. Conscious of the fact that neither the Guodian nor the Shanghai manuscript versions would allow such a procedure, because of the layout features discussed above, he posits that “the source text of the Zi yi used by the editor of the Li ji text was written in such a way that a new pericope always began at the top of a strip.”46 This is certainly a possibility. However, fully to explain the differences between the Guodian and Liji versions, we would need to make a number of further assumptions.47 With regard to the different number of sections, we would need to assume one of three scenarios: (a) the Liji editor used a manuscript version that already contained the sections that are present in the Liji but absent from the Guodian manuscript; or (b) after having decided on an arrangement of a manuscript version with broken bindings that he used to edit the Liji text, the editor felt free to add further sections to the text he edited; or (c) the text of these sections was on misplaced bamboo slips. 45 46 47

For a convenient comparison of textual order, see Boltz 2005: 57; Kern 2005: 302; Shaughnessy 2006: 67. Shaughnessy 2006: 78. Some of these have already been noted by Kern 2005: 304–305n22.

Manuscript Formats And Textual Structure In Early China

209

At this point it is worth considering some implications of the assumption that the Liji sections absent from the manuscript versions are misplaced bamboo slips. If the Liji editor had indeed added these slips by mistake (otherwise we would not call them “misplaced”), he not only had to have worked from a bamboo-strip *Zi yi manuscript that began each textual unit at the top of a new slip and had broken binding strings. He would also have needed to have had available at least one other bamboo manuscript with broken bindings and with slips of identical size and shape and written in a similar hand in order to mistake one slip of that second manuscript for part of the *Zi yi manuscript. Even if all this had been the case, it still would not explain the orthographic and lexical differences between the manuscript and Liji versions. To explain these, we would need to assume that either the manuscript version that the Liji editor reconstructed already contained these details that differ from the Guodian text or that the Liji editor took the liberty to make these changes. If we were to make the first assumption, we could claim that all the Liji editor did was to reconstruct texts by arranging the bamboo slips of several manuscripts with broken bindings to the best of his ability and bind the original slips anew. The second scenario, however, would involve making a new copy of the text from these slips, which allowed all the changes in phrasing and orthography. Once we entertain the possibility that the Liji editor copied the text anew and made changes to it, this again opens up two possible scenarios. First, if we wanted to maintain the assumption that the changed textual sequence was caused by the broken bindings of the original manuscript that the editor reproduced, we would need to assume that the editor, who otherwise freely rephrased the text, considered changing the textual sequence to be inappropriate or at least had no intention of changing it. The second scenario would include the possibility that the editor copied a new version of the text *Zi yi, for which he used one or several manuscripts bearing this text. He was aware that this was a time-honored text, but this did not preclude changing and adapting it. These adaptations could be restricted to details of phrasing, but they could also include changing the textual sequence. If that had been the case, there would not have been any need to physically rearrange parts of the text he copied. One could just as well copy an intact manuscript and change multiple aspects of the text. Other cases of early manuscripts with transmitted counterparts, as well as other transmitted parallel texts, show that the variation between these texts does not solely—indeed, not even primarily—concern the sequence of textual units but that the actual variation also occurs within the textual units, the so-called “building blocks” of a text. It ranges from orthographic variants to changes of particular words, differences in the use of function words or other syntactic features, increasing ideological specification of

210

Richter

statements, modernization of certain concepts, and transpositions not just of whole sections of text but sometimes of just a phrase or a sentence. All these different types of variation potentially interact to achieve the desired textual change.48 This is not to deny the fact that some of the variants in textual history are unintentional. But the multifaceted nature of textual variation in early China certainly requires us to consider the possibility that transpositions of textual units of any size are among the intentional changes made to texts in the course of their social use and their transmission. Thus, the physical rearrangement of parts of a written version of a text is not a necessary condition for such variants. Indeed, it seems that the prevalent understanding of positional variants in traditional Chinese texts as unintentional is influenced by the concept and term cuo jian 錯簡, which broadly refers to positional variants but—because of its literal meaning “misplaced slip(s)”—potentially carries a strong connotation of “mistaken’” or “wrong.” Once we accept that users of texts in early China freely varied texts as needed and that loose bamboo slips are not a necessary condition for positional variants, the assumption that the bamboo manuscript format was intentionally used as a tool to ensure creative flexibility of texts becomes much less likely. One might object that a text like *Zi yi has more extensive logical coherence and is thus too different from the Lunyu to challenge the idea that the particular textual structure of the Lunyu can be explained as resulting from the use of bamboo slips to take class notes, which were later collected in rearrangeable bamboo manuscripts. Indeed, the Guodian *Yucong manuscripts would seem to provide a confirmation of the “ring binder theory.” They are collections of short textual units, comparable to the Lunyu. *Yucong 1–3 in particular are written in a layout that makes very short textual units of a few (often just one or two) slips freely rearrangeable. Even *Yucong 4 can be divided into three transposable parts. But judging from the current archaeological record, such manuscripts were the exception rather than the rule in early China. Even in the Guodian *Laozi manuscripts the matches to the transmitted Laozi chapters are not freely transposable. Although the three manuscripts divide their texts into five, three, and four transposable units, respectively, the sequence of the vast majority of chapters is not changeable.49 48 49

For a case study demonstrating the interaction of different types of variation, see Richter 2013. For a convenient overview of the matches of these manuscripts with transmitted Laozi chapters, as well as for an indication of the transposable units in each manuscript, see Boltz 2005: 55–56.

Manuscript Formats And Textual Structure In Early China

211

Other manuscripts with similarly structured didactic texts, in which we would most expect the “ring binder” functionality to occur, are produced in a way that not only does not facilitate later rearrangements but makes them impossible. The Shanghai Museum texts *Junzi wei li 君子爲禮 (The Noble Man’s Practice of Ritual) and *Dizi wen 弟子問 (The Disciples Ask) are very similar in genre and content to the Lunyu, and they are written continuously from one slip to another, leaving no option to rearrange the text.50 Even if the layout of a manuscript ensures that each new textual unit begins at the top of a new slip, so that a rearrangement of these units would become possible, this need not indicate that such a rearrangement was ever intended. Just as in modern books, paragraph breaks are used to divide the text into manageable sections and guide the reader. The recently discovered P’yŏngyang Lun­yu manuscript was apparently produced in such a manner.51 Other types of layout emphasize the composite nature of texts, breaking the text down into manageable units, without allowing for any rearrangement. The Shuihudi Qin manuscript *Wei li zhi dao 為吏之道 (The Way of Acting as a Petty Official) bears a didactic text of little coherence and hence was ideally open to rearrangement if this had been the goal. However, this text is written in five registers over fiftyone bamboo slips, with each register typically accommodating five to eight characters.52 The text is written starting at the top of slip 1 in the first register, continuing in the same register and moving to the left to slip 51; only then is the second register started under the first on slip 1 and again read leftward to the last of the 51 slips, and so forth. Consequently, even without slip numbers there is only one possible arrangement of the bamboo slips. The same is true for a different version of this text in the unprovenanced Qin manuscript Wei li zhi guan ji qianshou 為吏治官及黔首 (Acting as a Petty Official, Managing Offices and the Populace), held by the Yuelu shuyuan 嶽麓書院 in Changsha.53 In this manuscript of eighty-seven slips, the text is written in four registers, but the last three slips are written continuously.54 Hence, the user of the manuscript read each register separately from slip 1 to slip 84, then moved on to the next register below. After finishing the lowest, fourth register at slip 84, one continued at the top of slip 85 and began to read the last three slips from 50 51 52 53 54

See Shanghai bowuguan cang Zhanguo Chu zhushu, vol. 5. See Paul van Els’s essay (chap. 6) in this volume. See Yunmeng Shuihudi Qin mu bianxie zu 1981: 81–85. For an overview of the manuscript layout, see table 2 in Richter 2013. See Zhu and Chen 2010: 109–149. After publication of this manuscript, however, the study of continuous diagonal lines across the verso of the slips led to a revision of the sequence. For a new reconstruction, see Staack 2014, with rich illustrations.

212

Richter

top to bottom, as in any other manuscript. These two manuscripts with largely parallel texts differ significantly not only in the sequence of textual units but also in the phrasing of sentences within these units and in how these units spread over the slips and registers of the different manuscripts. They are a good example to demonstrate how texts that were used in didactic contexts in particular—something we can surely assume for the Lunyu as well—could be freely varied in sequence and phrasing without any need to physically rearrange parts of the manuscript. We should not be deceived by the format of bamboo manuscripts into thinking that their individual slips were strongly perceived as independent objects. As mentioned earlier, slips were bound together as closely as possible to create the impression of a continuous surface. The Qin manuscript *Zheng shi zhi chang 政事之常 (Principles of Government Service) from Wangjiatai 王家臺 tomb 15, closely related in genre and subject matter to the ones from Shuihudi and the Yuelu shuyuan, is an impressive example of how the bound manuscript could be treated as a continuous surface. This surface is divided by horizontal and vertical lines into four concentric squares, which are further divided into sections by diagonal lines. The resulting delineated spaces are filled with the text, which has to be read clockwise, beginning in the middle (see fig. 7.5.2).55 Despite the separate slips, the bound manuscript was used as a continuous surface, and the layout was motivated by breaking up the text into manageable small units, which were most probably memorized, rather than dividing the text in a way that would allow a textual rearrangement by means of repositioning individual slips of the manuscript.

Conclusion

In sum, the manuscript evidence from early China does not lend any support to theories that ascribe the composite nature and fluidity of early Chinese texts to the fact that most manuscripts of the time consisted of bound bamboo or wood slips bearing one column of text each. There is neither a general preferred size of textual units nor a degree of regularity in manuscript formats or number of characters per slip that would establish a general correspondence between the two and thus justify the assumption that manuscript formats influenced textual structure. I cannot see any confirmation of the theory that this format was utilized to allow future additions to or rearrangements of the text through a manipulation of the physical parts of the manuscripts. Rather, 55

Wang Mingqin 2004: 39–42.

Manuscript Formats And Textual Structure In Early China

213

Figure 7.5.2 Figure 7.5.1 Guodian *Zi yi 緇衣 Layout of the Wangjiatai manuscript *Zheng shi zhi chang 政事之常 manuscript, slip 47, recto: end of text and 二十又三

there are indications to the contrary: manuscripts frequently show efforts on the part of their producers to define textual identity and to prevent a potential confusion of textual order. A manuscript was bound as tightly as possible and largely treated as a continuous surface on which text was written in singlecharacter columns (corresponding to the lines of text in our horizontal writing system). Unbinding and rebinding such a manuscript was not a convenient enough process to make it a preferred method of effecting changes to a text. The actual reasons for the composite nature and fluidity of early Chinese texts have to be sought elsewhere. One reason why many texts consist of relatively distinct short units—especially in the didactic texts where they most typically occur—lies in the use of these texts. Both in contexts of instruction

214

Richter

and in those of political persuasion, the most effective text is not a long treatise but concise units that make a pointed argument. Much of early Chinese literature is composed of such short, originally independent textual units. While many texts to varying extents forge these textual units into a longer, more or less coherent text, no such attempt is visible in the Lunyu, whose individual passages (zhang 章) remain separate. Many of them also occur verbatim or have close parallels in other texts.56 As studies of intertextuality in early China show, such textual units could be combined in vastly different ways in the composition of larger texts. There is no evidence to show that the particular arrangement of textual units in the Lunyu was in any way determined by the material conditions of manuscript culture at any time during its compilation.

Bibliography

Allan, Sarah, and Crispin Williams, eds. 2000. The Guodian Laozi: Proceedings of the International Conference, Dartmouth College, May 1998. Berkeley: Society for the Study of Early China and Institute of East Asian Studies, University of California. Allen, Rob. 2012. “‘Boz versus Dickens’: Paratext, Pseudonyms and Serialization in the Victorian Literary Marketplace.” In From Compositors to Collectors: Essays on BookTrade History, ed. John Hinks and Matthew Day, 155–180. New Castle, DE: Oak Knoll Press, British Library. Boltz, William G. 2005. “The Composite Nature of Early Chinese Texts.” In Text and Ritual in Early China, ed. Martin Kern, 50–78. Seattle: University of Washington Press. Brashier, K. E. 2011. Ancestral Memory in Early China. Cambridge, MA: Harvard Uni­ versity Asia Center. Chen Mengjia 陳夢家. 1980. Han jian zhuishu 漢簡綴述. Beijing: Zhonghua shuju. Chen Songchang 陳松長. 2000. Boshu shihua 帛書史話. Beijing: Zhongguo da baike quanshu. Cook, Scott. 2012. The Bamboo Texts of Guodian: A Study and Complete Translation. Ithaca, NY: Cornell University East Asia Program. Gansu sheng bowuguan 甘肅省博物館 and Zhongguo kexueyuan kaogu yanjiusuo 中國科學院考古研究所, eds. 2005. Wuwei Han jian 武威漢簡. Beijing: Zhonghua shuju.

56

See the chapters by Mark Csikszentmihalyi (chap. 8) and Michael Hunter (chap. 3) in this volume.

Manuscript Formats And Textual Structure In Early China

215

Giele, Enno. 1998–1999. “Early Chinese Manuscripts: Including Addenda and Corrigenda to New Sources of Early Chinese History: An Introduction to the Reading of Inscriptions and Manuscripts.” Early China 23–24:247–337. Grossman, Lev. 2011. “The Mechanic Muse: From Scroll to Screen.” New York Times Sun­ day Book Review, September 4. Hanshu 漢書. 1962. Beijing: Zhonghua shuju. Hu Pingsheng 胡平生. 2000. “Jiandu zhidu xintan” 簡牘制度新談. Wenwu 文物 2000.3:66–74. Huang Hui 黃暉. 1990. Lunheng jiaoshi 論衡校釋. Beijing: Zhonghua shuju. Jingmen shi bowuguan 荊門市博物館. 1998. Guodian Chu mu zhujian 郭店楚墓竹簡. Beijing: Wenwu chubanshe. Kern, Martin. 2000. “Feature: Writing and Authority in Early China, by Mark Edward Lewis.” China Review International 7.2:336–376. Kern, Martin. 2005. “Quotation and the Confucian Canon in Early Chinese Manuscripts: The Case of ‘Zi yi’ (Black Robes).” Asiatische Studien / Études asiatiques 59.1:293– 332. Kongzi jiayu zhuzi suoyin 孔子家語逐字索引. 1992. Chinese University of Hong Kong, Institute of Chinese Studies Ancient Chinese Concordance Series. Hong Kong: Commercial Press. Lau, D. C., trans. 1992. Confucius: The Analects (Lun-yü). Hong Kong: Chinese University Press. Lewis, Mark Edward. 1999. Writing and Authority in Early China. Albany: State Uni­ versity of New York Press. Li Ling 李零. 2008. Jianbo gushu yu xueshu yuanliu 簡帛古書與學術源流. Rev. ed. Beijing: Sanlian chubanshe. Li Songru 李松儒. 2006. “Guodian Chu mu zhujian ziji yanjiu” 郭店楚墓竹簡字跡研究. Liaoning sheng bowuguan guankan 遼寧省博物館館刊 2006.1:149–167. Li Songru 李松儒. 2012. “Zhanguo jian bo ziji yanjiu: Yi Shangbo jian wei zhongxin” 戰國簡帛字跡研究—以上博簡爲中心. PhD diss., Jilin University. Li Xueqin 李學勤, ed. 2010–2011. Qinghua Da xue cang Zhanguo zhujian 清華大學藏戰 國竹簡. 2 vols. Shanghai: Zhongxi shuju. Lianyungang shi bowuguan 連雲港市博物館 et al. 1997. Yinwan Han mu jian du 尹灣漢 墓簡牘. Beijing: Zhonghua shuju. Liji zhuzi suoyin 禮記逐字索引. 1992. Chinese University of Hong Kong, Institute of Chinese Studies Ancient Chinese Concordance Series. Hong Kong: Commercial Press. Loewe, Michael, and Edward L. Shaughnessy, eds. 1999. The Cambridge History of Ancient China: From the Origins of Civilization to 221 B.C. Cambridge: Cambridge University Press.

216

Richter

Ma Chengyuan 馬承源 et al., eds. 1988. Shang Zhou qingtongqi mingwen xuan 商周青 銅器銘文選, vol. 3. Beijing: Wenwu chubanshe. Maeder, Erik W. 1992. “Some Observations on the Composition of the ‘Core Chapters’ of the Mozi.” Early China 17:27–82. Makeham, John. 2006. “A New Hermeneutical Approach to Early Chinese Texts: The Case of the Analects.” Journal of Chinese Philosophy 33.1:95–108. Muir, Frank. 1992. The Oxford Book of Humorous Prose: From William Caxton to P. G. Wodehouse. Oxford: Oxford University Press. Nylan, Michael. 2000. “Textual Authority in Pre-Han and Han.” Early China 25:205– 258. Nylan, Michael. 2009. “Classics Without Canonization: Learning and Authority in Qin and Han.” In Early Chinese Religion, pt. 1, Shang through Han (1250 BC–220 AD), ed. John Lagerwey and Marc Kalinowski, 721–776. Leiden: Brill. Patten, Robert L. 1978. Charles Dickens and His Publishers. Oxford: Clarendon Press. Pian Yuqian 駢宇騫 and Duan Shu’an 段書安. 2006. Ershi shiji chutu jian bo zongshu 二 十世紀出土簡帛綜述. Beijing: Wenwu chubanshe. Qian Mu 錢穆. 1990. Xian Qin zhuzi xinian 先秦諸子繫年. Taipei: Dongda. Quan shanggu sandai Qin Han sanguo liuchao wen 全上古三代秦漢三國六朝文. 1951. Comp. Yan Kejun 嚴可均 (1762–1843). Beijing: Zhonghua shuju. Richter, Matthias. 2006. “Tentative Criteria for Discerning Individual Hands in the Guodian Manuscripts.” In Rethinking Confucianism: Selected Papers from the Third International Conference on Excavated Manuscripts, Mount Holyoke College, April 2004; 儒學的再思考:第三屆國際簡帛研討會論文集, ed. Xing Wen 邢文, 132–147. San Antonio, TX: Trinity University. Richter, Matthias. 2011. “Textual Identity and the Role of Literacy in the Transmission of Early Chinese Literature.” In Writing and Literacy in Early China: Papers from the Columbia Early China Seminar, ed. Li Feng and David Branner, 206–236. Seattle: University of Washington Press. Richter, Matthias. 2013. The Embodied Text: Establishing Textual Identity in Early Chinese Manuscripts. Leiden: Brill. Scarpari, Maurizio. 2007. “Zi Yue, ‘The Master said …,’ or Didn’t He?” In Guru, the Spiritual Master in Eastern and Western Traditions: Authority and Charisma, ed. Antonio Rigopoulos, 437–469. Venice: Venetian Academy of Indian Studies. Shaughnessy, Edward L. 2006. Rewriting Early Chinese Texts. Albany: State University of New York Press. Shanghai bowuguan cang Zhanguo Chu zhushu 上海博物館藏戰國楚竹書. 2001–2011. Ed. Ma Chengyuan 馬承源. 8 vols. Shanghai: Shanghai guji chubanshe. Shiji 史記. 1959. Beijing: Zhonghua shuju. Staack, Thies. 2014. “The Wei li zhi guan ji qianshou Manuscript from the Yuelu Academy Collection: A New Reconstruction Based on Verso Lines and Verso Imprints of

Manuscript Formats And Textual Structure In Early China

217

Writing.” In Documents and Writing Materials in East Asia, ed. Tomoyuki Nagata, 1–27. Kyoto: Institute for Research in Humanities, Kyoto University. Stone, Harry, ed. 1987. Dickens’ Working Notes for His Novels. Chicago: University of Chicago Press. Tsien Tsuen-hsuin. 2004. Written on Bamboo and Silk: The Beginnings of Chinese Books and Inscriptions. 2nd ed. Chicago: University of Chicago Press. Wang Guowei 王國維. 2004. Jian du jianshu kao jiaozhu 簡牘檢署考校注. Ed. Hu Pingsheng 胡平生 and Ma Yuehua 馬月華. Shanghai: Shanghai guji chubanshe. Wang Mingqin 王明欽. 2004. “Wangjiatai Qin mu zhujian gaishu” 王家臺秦墓竹簡概 述. In Xin chu jianbo yanjiu: Xinchu jianbo guoji xueshu yantaohui wenji 2000 nian 8 yue, Beijing 新出簡帛研究:新出簡帛國際學術研討會文集2000年8月, 北京, ed. Ai Lan 艾蘭 [Sarah Allan] and Xing Wen 邢文, 26–49. Beijing: Wenwu chubanshe. Yunmeng Shuihudi Qin mu bianxie zu 雲夢睡虎地秦墓編寫組. 1981. Yunmeng Shuihudi Qin mu 雲夢睡虎地秦墓. Beijing: Wenwu chubanshe. Zhang Xiancheng 張顯成. 2004. Jianbo wenxian xue tonglun 簡帛文獻學通論. Beijing: Zhonghua shuju. Zhong Ning 鍾寧. 2010. “Zhou chao ceming dianli de zhenshi zaixian: Qingtong Song gui” 周朝冊命典禮的真實再現—青銅頌簋. Zhongguo bowuguan 中國博物館 2010.2:72–74. Zhu Hanmin 朱漢民 and Chen Songchang 陳松長, eds. 2010. Yuelu shuyuan cang Qin jian 嶽麓書院藏秦簡. Shanghai: Shanghai cishu chubanshe.

218

Csikszentmihalyi

Chapter 8

Interlocutor Collections, the Lunyu, and ProtoLunyu Texts Mark Csikszentmihalyi

Why is the Lunyu organized in the way that it is? It is a text that has been divided into about twenty “chapters” at least since the Han dynasty, but only chapter 10, with its focus on rituals, has a consistent topic. Within the chapters are many short passages—dialogues and occasional short narratives—that scholars today call pericopes (from the Greek for “cut around”), borrowing a term used for the smallest units of text in biblical form criticism, or zhang 章, which derives from breaks in musical pieces and connotes something like “paragraph” in the context of written texts.1 Within each chapter, more often than not, one pericope follows the next with little or no sign of connection between them beyond a common involvement with Kongzi. This does not mean that the Lunyu is a “random” collection of passages. Beyond the thematic consistency of chapter 10, within chapters there are a number of stretches of consistency across pericopes. Looking at other loosely organized works, it becomes clear that such stretches may contain information about the way the Lunyu was formed. In the Bible, for example, two chapters of the book of Matthew (8:1–34 and 9:1–34) contain a number of stories about healing. These same stories are not grouped but are instead sprinkled throughout Mark (chaps. 1, 2, 4, and 5) and Luke (chaps. 4, 5, 7, 8, and 9). In other words, Matthew arranges the pericopes by topic; the other two books, by other principles such as the chronology of the life of Jesus. Marcus Aurelius’s (121–180) Meditations consists of diary-like jottings arranged into chapters that reflect his thinking at different stages in his life.2 Blaise Pascal’s (1623–1662) Pensées is a posthumously published collection of notes whose loose organization is largely the product of rearrangements by later editors.3 1 An early definition of zhang as “up to when the music finishes” (樂竟為一章) is found in Duan Yucai’s 段玉裁 (1735–1815) Shuowen jiezi zhu 說文解字注 3B.33a. Zhang is also the term used in the Dingzhou 定州 Lunyu exemplar from the first century BCE to subdivide chapters. That is, chapters were followed by the number of zhang and then the number of characters in each chapter. 2 See, e.g., Marcus Aurelius 2006. 3 E.g., Blaise Pascal 1995.

© Koninklijke Brill NV, Leiden, 2018 | doi:10.1163/9789004382947_010

Interlocutor Collections, The Lunyu, And Proto-lunyu Texts

219

These examples represent several different types of consistency that a set of dialogues such as those found in the Lunyu might exhibit. They might be (1) temporally connected, that is, arranged chronologically; (2) topically connected, that is, all touching on the same subject; or (3) have taken place between Kongzi and a single interlocutor. When we find instances of consistency, a further question might be asked: what accounts for this consistency? For example, if the members of a set of pericopes are chronologically arranged, is it because they were recorded in that order, or is it the result of a later editor’s attempt to rearrange them to make them accord with a biography? If they share the same topic, is this the function of a Matthew-like disciple’s arrangement, or could they perhaps reflect a compilation of versions of the same event from different sources? Finally, if we find consecutive dialogues featuring the same disciple, does that mean that they come from notes of that disciple or his “school,” or is the consistency due to the hand of a later editor? In this chapter, I will begin exploring these questions by recalling the two earliest accounts of the formation of the Lunyu. Although these Han period explanations suggest rather different things about the process by which the text formed, they do share a basic assumption that the text was the product of a conscious decision to assemble a record—either comprehensively or in an exemplary fashion—of the teachings of a particular person, and that this decision was made during the lifetime of first- or second-generation disciples who contributed the content of this record. Both accounts dismiss the possibility that materials unconnected with the direct master-disciple transmission are in the text. As the editors of the present volume have pointed out, recent work has called many elements of these accounts into question, but the next step is to begin to explain the phenomena on which these early accounts are based. If the Lunyu does not come from disciple records, what explains this Han impression that it did? To attempt to begin to answer this question, this chapter will look specifically at multiple recensions of several dialogues that are found in the Lunyu, both transmitted and excavated, from the pre-imperial and Qin-Han periods. I will ask whether there is evidence that the formal consistencies seen in parts of the Lunyu might be interpreted as traces of the process whereby interlocutor texts were combined and incorporated into a single text. (By the term “interlocutor text,” used in the context of the Lunyu, I mean, at the very least, a set of pericopes that contain exchanges featuring a single interlocutor, usually a disciple or a ruler.) The most telling examples I have found call into question the assumption that the Lunyu was based on disciple-specific texts that were combined into a single record. By contrast, I will argue that even in the Han, questions, statements, and entire pericopes were still being put into the

220

Csikszentmihalyi

mouths of different interlocutors in a fluid way that does not fit very well with the picture of a fixed transmission of dialogues that had come down from previous centuries. Finally, I will conclude that the more likely scenario is actually much more complicated—rather than a simple movement from disciplebased collections to a single synthetic record, the period saw texts being reattributed and reformulated according to multiple principles of arrangement.

Two Early Versions of the “Interlocutor Text” Origin Story

The traditional story that the Lunyu was composed from the notes of disciples derives in part from the two earliest surviving accounts of its origins. The bibliographical listings in the Hanshu 漢書 compiled by Ban Gu 班固 (32–92 CE) are the outgrowth of an imperially sponsored inventory of books commissioned originally during the reign of Emperor Cheng 成 (33–7 BCE) and revised during the subsequent reign of Emperor Ai 哀 (7–1 BCE). The Lunyu subsection of the “Liu yi” 六藝 (Six Arts) section of the inventory lists twelve titles in 229 chapters and begins with three versions of the Lunyu itself, an “ancient” (gu 古) version in twenty-one chapters, a state of Qi 齊 version in twenty-two chapters, and a state of Lu 魯 version in twenty chapters. The Lu version has the same number of chapters, twenty, as the modern one. The bibliography also discusses the way the text was composed: The Lunyu is composed of discussions in which Confucius responded to his disciples and his other contemporaries, and of his disciples’ conversations with each other and the sayings they each heard directly from the Master. At the time, each of the disciples kept their own personal records. Then, after the Master died, his disciples gathered them all together, judging them and recompiling them,4 and so they were called his “selected sayings” [i.e., Lunyu]. 4 The phrase lunzhuan 論篹 is glossed by the Tang commentator Yan Shigu 顏師固 (581–645) as lunzhuan 論撰. The latter phrase appears in the “Jitong” 祭統 (Comprehensive Account of Sacrifices) chapter of the Liji 禮記 (Records of Ritual) in discussions of inscriptions, where it refers to the commemoration of the deeds of one’s ancestors: “inscriptions discuss and commemorate the virtue and goodness of a person’s ancestors” (銘者論譔其先祖之有德 善) and “The gentlemen of antiquity discussed and commemorated the good points of their ancestors in order to clearly exhibit them to people of later generations” (古之君子論譔其 先祖之美而明著之後世者也). Sun Xidan 1989: 47.1250–1252. In both phrases Legge translates lunzhuan as “panegyrise.” Neither of these interpretations really gets at the sense of “selection” that lun often conveys and that suits a context where multiple records are consolidated. The same word zuan is used elsewhere in the Hanshu bibliography to talk about

Interlocutor Collections, The Lunyu, And Proto-lunyu Texts

221

論語者,孔子應答弟子時人及弟子相與言而接聞於夫子之語也。當時 弟子各有所記,夫子既卒,門人相與輯而論篹。故謂之論語。5

The catalog explains the title of the work as the sayings (yu 語) of the Master that had been judged or selected (lun 論) and then recompiled (zhuan 篹) for inclusion in a standard version of Kongzi’s words. The picture this description paints is of a work that was composed soon after Kongzi’s death from the notes and recollections of his disciples. On the basis of such a description, there is little room for doubt about the internal consistency or the ultimate authenticity of sections of the text. The Lunyu is described as the product of a decisionmaking process that renders it “orthodox” in its own way. Wang Chong’s 王充 (27–ca. 100 CE) “Zheng shuo” 正說 chapter from the Lunheng 論衡 (Balanced Discussions), however, takes the first real text-critical approach to the Lunyu by using a more complex model of the formation of the text. In Lunheng, Wang Chong challenges the notion that the Lunyu originally had twenty-one chapters. According to Wang Chong, the original Lunyu had several dozen or hundreds of chapters of Kongzi’s words and acts recorded by his followers, but that version of the text was lost soon after the rise of the Han. Then, the twenty-one-chapter version of the Lunyu found in Kongzi’s home supplanted the thirty chapters that were available in other editions. Wang’s description of the first stage of the work’s transmission paints a picture not of Kongzi’s compilation of Documents texts, contrasting their origin in the distant past (“the origins of the Documents text were already in the distant past” [書之所起遠矣]) with a later moment “when Kongzi recompiled them together” [至孔子篹焉]). Here I use the word “recompile” to translate zhuan, and the sense in which it is used is clearly that this editorial act is in part to commemorate the Master. See Hanshu 30.1962. 5 Hanshu 30.1962. This implies that the two characters in the title are a result of the fact that the text is composed of types of speech, including “the speech they heard directly from the Master” (接聞於夫子之語), hence “speech” (yu) preceded by “judge” or “select” (lun) in the sense of eulogize. There are other occurrences of the title in pre-Han and early Han works. For example, in his work on Qin dynasty classical studies, Uchino Kumaichirō (1939: 71) cites the occurrence of the title Lunyu in the Liji 禮記, Hanshi waizhuan 韓詩外傳 (Outer Tradition of the Han Poetry), and Shiji in his discussion of its textual history during the Qin period. The attribution to the Lunyu of Lunyu 4/20 in the “Fang ji” 坊記 chapter of the Liji is a good example of later interpolation, since manuscript versions of the affiliated “Zi yi” 緇衣 chapter demonstrate the way in which the contents, especially the quotations, have been revised in the received editions. For advocates of a first- or second-century date for the Lunyu, the three citations by title in the Hanshi waizhuan are a phenomenon that is ostensibly harder to explain away. The beginning of Lunyu 10/18 is cited in Hanshi waizhuan 2.21, a short Kongzi quotation from 13/3 is cited in Hanshi waizhuan 5.33, and a longer quotation from the end of the same Lunyu passage is cited in Hanshi waizhuan 6.6, each just prior to quotations of the Shijing 詩經. It is worth noting, from a disciple text perspective, that all these Lunyu quotations are in passages that involve the disciple Zilu 子路.

222

Csikszentmihalyi

a single creation of a text but of numerous attempts to record the words and actions of the sage on short slips as befits a lesser work: Now, the Lunyu was the collective records of Kongzi’s words and actions by his disciples, and since the occasions when they took down and recorded them were very many, it came to several dozen or hundreds of chapters of slips measuring eight cun. The reduction of the records was for the convenience of portability. As what was left was not a “classic,” they feared the loss of the transmitted texts and entries and so used only eight-cun slips, not two-chi and four-cun slips. These were lost at the rise of the Han. 夫論語者,弟子共紀孔子之言行,勑記之時甚多,數十百篇,以八寸 為尺,紀之約省,懷持之便也。以其遺非經,傳文紀識恐忘,故以但 八寸尺,不二尺四寸也。漢興失亡。6

The loss of the many-chapter Lunyu at the start of the Han was followed in that period by the replacement of a surviving thirty-chapter version with the twenty-one-chapter “ancient-text” version by the time of Emperor Zhao 昭帝 (r. 87–74 BCE): When in Emperor Wu’s time the “ancient” version was discovered in Kongzi’s wall, it had twenty-one chapters. The two versions from Qi and Lu, and the Hejian version in nine chapters, made thirty chapters. … What people today call the Lunyu in twenty chapters is missing the nine chapters of the Qi, Lu, and Hejian versions. Parts of the original thirty chapters were scattered and lost, so some versions have twenty-one chapters. The contents are sometimes more and sometimes less; the text is sometimes right and sometimes wrong. 至武帝發取孔子壁中古文,得二十一篇,齊、魯二,河間九篇,三十 篇. . .今時稱論語二十篇,又失齊、魯、河間九篇。本三十篇,分布亡 失,或二十一篇。目或多或少,文讚或是或誤。7

6 For Wang Chong’s entire discussion of the Lunyu, see Huang 1990: 4.1135–1139. Here I am grateful for John Makeham’s suggestion to read 數十百 “several dozen to a hundred,” and for many discussions with Kim Tae Hyun about early descriptions of Lunyu formation. 7 Here I follow Yamada Katsumi (1984: 3.1745) in reading 讚 as 辭.

Interlocutor Collections, The Lunyu, And Proto-lunyu Texts

223

Wang Chong makes it clear that the work was shortened twice. A number of versions were first lost in the compression from several dozen or hundreds of chapters down to thirty, and then somehow from thirty to an abridged version like the “ancient” text version with 21 chapters. One example of such an abridgment is the “Hejian” version of the Lunyu from the library of Prince Xian 獻 of Hejian (fl. 155 BCE), who, among other things, edited Shusun Tong’s 叔孫通 anthology of ritual texts after Shusun’s death and was connected with the compilation and transmission of Han classicist works like the Yueji 樂記 (Records of Music), Zhou zhi 周制 (Zhou Regulations), and Hejian Xian wang dui shangxia san yonggong 河間獻王對上下三雍宮 (Prince Xian of Hejian’s Responses concerning Superior and Inferior in the Three Yong Palaces).8 No other contemporary sources mention this work, which presumably had seven chapters. Wang Chong’s more general point in the chapter is that Han scholars are ignorant of both the true meanings and circumstances of transmission of the classics and focus instead on deriving omens and portents from them at the expense of actually understanding them. Even the title Lunyu, according to Wang Chong, was a relatively late addition: At the beginning, when Kongzi’s grandson Kong An’guo started teaching Fu Qing from the state of Lu, who in his official career reached the rank of inspector of Jingzhou, it began to be called Lunyu. 初,孔子孫孔安國以教魯人扶卿,官至荊州剌史,始曰論語。

Wang’s understanding of the history of the text differs from that found in the Hanshu, reflecting a critical awareness that calls into question the inventory’s assumption that editions circulating in the Han were pristine versions of the notes and recollections of Confucius’s disciples. Wang identifies different factors behind the abridgement of the text, such as the ease of transporting the text and an obsession with nonclassical applications of the classics. He describes the Lunyu as a set of fixed records of Kongzi’s words and actions that were divided and somewhat randomly abridged. Nevertheless, as does the Hanshu, Wang still presumes that the text originated from the notes of multiple disciples. The idea of the accretion or gradual compilation of these key texts began to be explored in later imperial China, but it was not a facet of Han writing on text formation. Since both accounts argued that the work originated 8 Hanshu 22.1042, 30.1725, 30.1726. Michael Hunter’s (2012: 268) dissertation advances a hypothesis that Prince Xian had a significant role in the collection of materials that ended up forming the Lunyu.

224

Csikszentmihalyi

through an editorial process of reconciling disciple accounts, it is worth looking critically at the types of consistency around sections of the Lunyu and other early texts that concern Kongzi’s disciples.

“Interlocutor Texts” in Early China: The Case of Zengzi 曾子

Since both these accounts assume that the text was formed from the records of direct or second-generation interlocutors of Kongzi, it is worth asking if such “interlocutor” or “disciple” texts existed or if they still exist. The Hanshu bibliography lists the following four texts with titles compatible with disciple names in the Shiji chapter “Zhongni dizi liezhuan” 仲尼弟子列傳 (Arrayed Traditions of Zhongni’s [i.e., Kongzi’s] Disciples): Zisi 子思 in 23 pian 篇 (Shiji: cognomen Yuan Xian 原憲) Zengzi 曾子 in 18 pian (Shiji: Zeng Shen 曾參, cognomen Zi Yu 子輿) Qidiaozi 漆雕子 in 13 pian (Shiji: Qidiao Duo 漆雕哆, cognomen Zi Lian 子斂) Fuzi 宓子 in 16 pian (Shiji: Fu Buqi 宓不齊, cognomen Zijian 子賤)9 It is worth noting that these interlocutor texts for the most part do not overlap with the interlocutor texts that are among the Warring States slips held by the Shanghai Museum, which include dialogues with Zigao 子羔, Zhong Gong 中 弓 (i.e., Ran Yong 冉雍), Ji Kangzi 季康子, Ji Huanzi 季桓子, and Yan Yuan.10 The lack of overlap between these lists suggests that a large number of interlocutor texts were in circulation. In addition to tomb texts likely from the Warring States period and works identified in the late Western Han inventories, some received sources contain multiple dialogues that include Kongzi, such as the Xunzi 荀子, Da Dai Liji 大戴禮記 (Elder Dai’s Records of Ritual), Liji 禮記, Shiji 史記, Kongcongzi 孔叢子, 9 10

Hanshu 30.1724. Shanghai bowuguan cang Zhanguo Chu zhushu 2001–2011. Among the eight volumes of this series, key interlocutor texts are in vol. 2, Zigao 子羔 (2002); vol. 3, Zhonggong 中弓 (2003); vol. 5, *Ji Kangzi wen yu Kongzi 季康子問於孔子 (Ji Kangzi Questions Kongzi), *Dizi wen 弟子問 (The Disciples Ask), and *Junzi wei li 君子為禮 (The Noble Man’s Practice of Ritual) (2005); vol. 6, *Kongzi jian Ji Huanzi 孔子見季桓子 (Kongzi Visits Ji Huanzi) (2007); and vol. 8, *Yan Yuan wen yu Kongzi 顏淵問於孔子 (Yan Yuan Questions Kongzi) (2011). Other relevant finds are the 1973 *“Rujia zhe yan” 儒家者言 (Sayings of the Ru) found at Dingxian 定縣 in Hebei and the 1977 Fuyang Shuanggudui 阜陽雙古堆 Han tomb 1 finds from Anhui. All of these finds support the claim of Wang Chong that a tremendous number of Kongzi-related texts were circulating in the Han.

Interlocutor Collections, The Lunyu, And Proto-lunyu Texts

225

and Kongzi jiayu 孔子家語 (Family Sayings of Kongzi). These texts are effectively repositories that include interlocutor texts, and it is an open question whether their sets of pericopes featuring Kongzi and a particular interlocutor preserve aspects of the Hanshu’s titles cited above. Because not all texts that mention a single interlocutor explicitly connect that figure to Kongzi, either through dialogue or otherwise, I will refer to those that do not as “independent interlocutor texts.” To explore these works, this section will focus on the Zengzi texts incorporated into the Da Dai Liji. These texts constitute a “cache” of material related to a single disciple, but on closer examination contain several formally distinct types of work related to interlocutors. The Da Dai Liji contains multiple interlocutor texts: four chapters that we would call interlocutor texts and more that are independent interlocutor texts. Those four chapters are chapter 39, “Zhuyan” 主言 (Ruler’s Words), a Zengzi chapter with close parallels in the “Wangyan jie” 王言解 (Ruler’s Words, Explained) chapter of the Kongzi jiayu; chapter 40, “Aigong wen wuyi” 哀公問五義 (Duke Ai Asks about Five Kinds of Righteousness), a Duke Ai 哀公 chapter with close parallels in the “Ai gong” 哀公 chapter of the Xunzi; chapter 41, “Aigong wen yu Kongzi” 哀公問於孔子 (Duke Ai Questions Kongzi), another Duke Ai chapter with close parallels in the “Aigong wen” 哀公問 (Duke Ai’s Questions) chapter of the Liji; and chapter 62, “Wudi de” 五帝德 (Virtues of the Five Emperors), which is mostly conversations between Zaiwo 宰我 and Kongzi, with parallels in the “Wudi benji” 五帝本紀 (Basic Annals of the Five Emperors) chapter of the Shiji. There are, in addition, ten chapters of Zengzi “independent interlocutor” texts, including three chapters of dialogue between Zengzi and Shan Juli 單居離 as interlocutor.11 Pi Xirui 皮錫瑞 has argued that these chapters are part of the 18-juan disciple text Zengzi listed in the Hanshu.12 We do not know much about the antiquity of these interlocutor texts, but there is little question that they contain elements of an earlier Zengzi tradition and preserve a topical organization within the independent-interlocutor texts. A rare opportunity to compare transmitted exemplars with tomb texts presents itself with the Shanghai Museum manuscript “Neili” 內禮 (Household Rituals), a work that has significant parallels with chapters of the Da Dai Liji 11

12

The Zengzi chapters are 49, “Zengzi li shi” 曾子立事; 50, “Zengzi ben xiao” 曾子本孝; 51, “Zengzi li xiao” 曾子立孝; 52, “Zengzi da xiao” 曾子大孝 (close at times to “Jiyi” 祭義 in the Liji); 53, “Zengzi shi fumu” 曾子事父母; 54, “Zengzi zhi yan shang” 上曾子制 言上; 55, “Zengzi zhi yan zhong” 中曾子制言中; 56, “Zengzi zhi yan xia” 曾子制言下; 57, “Zengzi ji bing” 曾子疾病; and 58, “Zengzi tian yuan” 曾子天圓. The chapters with Shan Juli are 53, 57, and 58. Pi 1954: §2.2. Jiang Boqian (1985: 344–345) summarizes a number of Qing perspectives on the Zengzi in his Zhuzi tongkao.

226

Csikszentmihalyi

that have long been identified with the lost Han work Zengzi: chapters 51, “Zengzi li xiao” 曾子立孝 (Zengzi Establishes Filial Piety), and chapter 53, “Zengzi shi fumu” 曾子事父母 (Zengzi Serves His Parents). Both are among the four of the ten chapters that have only one reference of any kind to Zengzi, in the opening phrase “Zengzi yue” 曾子曰.13 The first section of the “Neili” begins with the phrase “Junzi zhi li xiao” 君子之立孝 (A gentleman’s establishing filial piety), where chapter 51 begins with “Zengzi yue: Junzi li xiao” 曾子曰君子立 孝 (Zengzi said, “A gentleman establishes filial piety”). The second section of the “Neili” begins with the phrase “Junzi shi fumu” 君子事父母 (A gentleman serves his parents) where the parallel in chapter 53 begins “Xiaozi” 孝子 (A filial child). While the better part of the “Neili” is made up of these two sections with extensive parallels to the pseudo-Zengzi, there are neither explicit nor implicit references to Zengzi anywhere in the text, and the title “Neili” is written on the back of the first slip.14 Given that many of the pericopes in four (barely) Zengzi chapters of the Da Dai Liji begin with the word junzi 君子, it appears likely that the thin skein of affiliation with Zengzi in the titles was not original, and that the original titles were indeed “Junzi li xiao” 君子立孝 and “Junzi shi fumu” 君子事父母.15 Here we might consider Jiang Boqian’s argument that the Xiaojing 孝經, a text affiliated with Zengzi, might have been created during one of the several periods of the Han when xiao 孝 (filial piety) was actively promoted by the state.16 Perhaps Zengzi’s name was affixed to these texts in the same period. The repositories of early Zengzi material in places like the Zengzi sections of the Da Dai Liji and the “Xiaoxing” 孝行 (Acting Out Filial Piety) chapter of the Lüshi chunqiu contain some characteristics of a corpus that is still forming. Consider the important sentiment that “the body is the form left to one by one’s parents” (身者親之遺體也). It is a part of Zengzi’s answer to a question by Gongming Yi 公明儀 (“What does the Master mean by filial piety?” [夫子可 (何) 謂孝乎]) in Da Dai Liji chapter 52, “Zengzi da xiao” 曾子大孝 (Zengzi’s Great Filial Piety), but it is simply a freestanding quotation of Zengzi in the “Jiyi” 祭義 (Meaning of Sacrifice) chapter of the Liji.17 There is no way to know if the phrase moved from a quotation to a dialogue or vice versa, but both versions are preserved. The Lüshi chunqiu preserves a dialogue between Yuezheng 13 14 15 16 17

Chapters 50–52 and 55 fit this description. Ji 2007: 106–107. Matthias Richter (2002) examines one Zengzi chapter and shows how it is a repurposed character diagnosis text, and so it is actually not a part of the genre that it is widely described as being and is only thinly connected to the character Zengzi. Jiang 1985: 357–358. Cf. Wang 1983: 82–85 with Sun Xidan 1989: 46.1228.

Interlocutor Collections, The Lunyu, And Proto-lunyu Texts

227

Zichun 樂正子春 and his disciple where the former says that Zengzi relayed to him a quotation from Kongzi: “What one’s father and mother put together and give birth to, their child keeps together when he returns; such a child can be said to be filially pious” (父母全而生之,子全而歸之, 可謂孝矣).18 This sort of double attribution might be an editorial response to resolve the difficulty of the quotation’s being attributed to both Zengzi and Kongzi in collections of quotations that were in circulation at the time. Yet the relative dearth of such instances of parallels to the early Zengzi texts in other works like the Shiji, and the fact that there is no overlap between the many Zengzi filial piety stories in Da Dai Liji and Zengzi materials in the Lunyu, combined with the suggestive reattributions of passages, point to a weakness in the evidence that Zengzi texts predated later collections with more variegated formats. This certainly goes against the accepted narrative, but it is also worth considering other hypotheses when evaluating the reliability of the influential “interlocutor text” story that is first told in the Eastern Han period. Wang Chong’s account and that of the Hanshu appeared during something of a dizi feng 弟子風 (disciple vogue) in the Han that fetishized the completeness of the set of the Master’s disciples. One can chart this trend from the biographies of the seventy-odd disciples in the Western Han Shiji, through lost Eastern Han works like the Kongzi turen tufa 孔子徒人圖法 (Kongzi and His Followers’ Methods for Making Charts),19 Zheng Xuan’s 鄭玄 (127–200) Kongzi dizi mulu 孔子弟子目 錄 (Catalog of Kongzi’s Disciples),20 and Eastern Han teacher Wen Weng’s 文翁 Lidian tu 禮殿圖 (Chart of the Ritual Hall).21 These works were likely part of the material that was used to compose the “Dizi xing” 弟子行 (Acts of the Disciples) and “Qishi’er dizi jie” 七十二弟子解 (The Seventy-Two Disciples, Explained) chapters of Wang Su’s 王肅 (195–256) Kongzi jiayu (Family Sayings of 18 19 20

21

Chen 2002: 736. Cf. Knoblock and Riegel 2000: 305. Listed in the Hanshu (30.1717) bibliography as having two juan 卷. According to the reconstruction by Yuan Jun 袁鈞 (Zhengshi yishu 鄭氏佚書 9.1b–9a), the text consisted of a table of names, cognomens, and birthplaces. A number of Qing scholars reconstructed portions of this lost text, which was listed in the Sui bibliography as Lunyu Kongzi dizi mulu 論語孔子弟子目錄 and in Tang bibliographies as Lunyu pianmu dizi 論語篇目弟子. For a comparison of their various editions, see Sun Qizhi and Chen Jianhua 1997: 165. Hu Lanjiang has argued that the Eastern Han administrator and teacher Wen Weng was responsible for a stone “reconstruction” of a ritual chamber of the Duke of Zhou’s, which he used as a study hall. According to a Jin 晉 dynasty account preserved in the Yiwen leiju 藝文類聚, this hall had paintings of ancient sages and worthies, who probably included Kongzi and his disciples. See Yiwen leiju 38.15a. The paintings may or may not be related to another title attributed to Wen Weng (and cited three times in Sima Zhen’s 司馬貞 [679–732] subcommentary to the Shiji), the Kongmiao tu 孔廟圖. See Hu 2002.

228

Csikszentmihalyi

Kongzi). This impulse was likely related to what Martin Kern has called “the late Western and then Eastern Han imperial needs to agree on a unified text and its more or less normative interpretation.”22 This presents us with a cultural moment when the character, images, and accounts of the disciples were of great cultural interest, and raises the possibility that some extant interlocutor texts actually were in part a product of this historical moment. Likewise, accounts of the formation of the Lunyu may have been a retroprojection into the past of the values of that moment. While none of this is decisive, it is worth considering the hypothesis that at some point topical texts about the filial piety of “the gentleman” were repurposed into Zengzi interlocutor texts because Zengzi became the paradigmatic filial son. In other words, it is possible that during the Han topical texts were attributed to a disciple associated with that topic, with the result that they were later considered to be part of a lost Han Zengzi. One place to test this hypothesis might be the earliest inventory of Kongzi’s disciples, the disciple biography chapter in the Shiji.

The First “Interlocutor Collection” and the Case of the Missing Zengzi 曾子

E. Bruce Brooks and A. Taeko Brooks have argued that the “Zhongni dizi liezhuan” chapter in the Shiji and the “Qishi’er dizi jie” chapter of the Kongzi jiayu are both based on a lost “register” of Kongzi’s disciples.23 The “Zhongni dizi liezhuan” could tell us something more concrete about how disciple stories circulated in the Han. The text is a compendium of biographical information concerning seventy-two or seventy-seven of the Master’s disciples and contains quotations concerning the first thirty-five of them; it was likely compiled at the end of the second century BCE.24 The chapter begins by explaining what these disciples had in common: “Kongzi said, ‘Seventy-seven people learned their vocation from me through personal communication. All of them had special abilities’” (孔子曰:受業身通者七十有七人。皆異能之士也). It then 22 23 24

Kern 2007: 776. Brooks and Brooks 1998. This is one of two chapters about Kongzi in the Shiji, the other being the “Kongzi shijia” 孔子世家 (Hereditary House of Kongzi). The latter has a line similar to the one that begins the “Zhongni dizi liezhuan”: “Kongzi taught the [classics of] Poetry, Documents, Ritual, and Music to some three thousand disciples. Those who personally learned the Six Attainments from him numbered seventy-two” (孔子以詩書禮樂教,弟子蓋三千 焉,身通六藝者七十有二人). See Shiji 47.1938.

Interlocutor Collections, The Lunyu, And Proto-lunyu Texts

229

introduces four sets of disciples grouped by ability: those skilled in “virtuous conduct” (dexing 德行), “matters of governing” (zhengshi 政事), “occasional language” (yanyu 言語), and “literary studies” (wenxue 文學). Of the seventyodd disciples mentioned by name, only the first ten are listed under these four headings.25 What becomes fairly clear is that the compiler of this chapter was not familiar with a couple of key passages in the Lunyu, or at least not exactly as they appear in the Lunyu. One difference is that the order of the “four specialties” is off compared with an otherwise parallel list in Lunyu 11/3 (as well as in the Kongzi jiayu).26 To be precise, the two categories and their associated disciples in the “Matters of governing” and “Occasional language” sections are reversed. Table 8.1 Order of key disciples in Lunyu 11/3

德行:

言語:

政事:

文學:

1. 顏淵 2. 閔子騫 3. 冉伯牛 4. 仲弓

5. 宰我 6. 子貢

7. 冉有 8. 季路

 9. 子游 10. 子夏

Table 8.2 Order of key disciples in the Shiji

德行:

政事:

言語:

文學:

1. 顏淵 2. 閔子騫 3. 冉伯牛 4. 仲弓

5. 冉有 6. 季路

7. 宰我 8. 子貢

 9. 子游 10. 子夏

Another difference between the Lunyu passage and the start of the Shiji chapter is how the list of remaining disciples was filled out. Beginning in the Song, 25 26

By contrast, the “Qishi’er dizi jie” lacks introductory matter, and while it also begins with Yan Hui, it contains few quotations. See Yang Chaoming 2005: 430–453. That passage organizes them as follows: 德行: 顏淵, 閔子騫, 冉伯牛, 仲弓. 言語: 宰我, 子貢. 政事: 冉有, 季路. 文學: 子游, 子夏. This matches the presentation of the disciples in Yang Chaoming 2005: 430–435. References to passages in the Lunyu follow the Lunyu zhuzi suoyin.

230

Csikszentmihalyi

scholars have noticed that in order to include Gongbo Liao 公伯寮, the author of the Shiji chapter must have frightfully misread Lunyu 14/36, which paints Gongbo Liao as a schemer who tries to slander Zilu: 27 Gongbo Liao slandered Zilu to Ji Sun. Zifu Jingbo reported this, saying, “Our master is certainly being given the wrong impression by Gongbo Liao, but I still have it in my power to execute him and expose his corpse in the public square.” The Master said, “If my Way is put into practice, it is a matter of allotment. If my Way is abandoned, it is a matter of allotment. How can Gongbo Liao do anything when it comes to allotment?” 公伯寮愬子路於季孫。子服景伯以告,曰:夫子固有惑志於公伯寮 , 吾力猶能肆諸市朝。子曰:道之將行也與?命也。道之將廢也與?命 也。公伯寮其如命何。28

The passage makes it clear that Gongbo Liao is more of a figure like Huan Tui 桓魋, who is a threat to Kongzi’s life in Lunyu 7/23, yet the compiler of the Shiji chapter lists him among Kongzi’s disciples. There are also disciples not listed in the Lunyu, and Lunyu disciples left out. Finally, not all the stories come from the Lunyu. For example, in the very long section about the disciple Zilu 子路 there are two consecutive sections that are not in the Lunyu. The first, in which Kongzi assures his disciple that reverence and respectfulness can beat strength and numbers, has a parallel only in the Shuiyuan 說苑 (Garden of Persuasions), while the second is an extended section from the Zuozhuan 左傳 entry for year 15 (480 BCE) of Duke Ai’s 哀公 reign.29 Earlier we saw evidence that suggested that some Zengzi interlocutor texts were not originally associated with Zengzi. Looking at the Shiji chapter, the fact that there are no Zengzi quotations is interesting, and perhaps consistent with the idea that the conversion of filial piety texts into Zengzi interlocutor texts was a late event. The Shiji chapter’s entire reference to Zengzi is limited to the following formulaic note: Zeng Shen was from Southern Wucheng. His cognomen was Ziyu, and he was forty-six years younger than Kongzi. Kongzi thought that he understood the Way of filial piety, and so he transmitted this specialty to him. He wrote the Classic of Filial Piety. He died in the state of Lu. 27 28 29

Jiang 1985: 140. Liu 1990: 14.593–596. Xiang 1987: 7.163; Yang 1993: 1590–1597.

Interlocutor Collections, The Lunyu, And Proto-lunyu Texts

231

曾參南武城人,字子輿。少孔子四十六歲。孔子以為能通孝道,故授 之業。作孝經,死於魯。30

It is conceivable that these sorts of notes were added later. In such a case, the original chapter may well have contained no references to Zengzi, something that was remedied by a later systematizing hand intent on representing an interlocutor whose filial piety texts had recently come into fashion. Just as Zengzi’s omission in the Shiji chapter and the presence of independently circulating Zengzi collections preserved in other Han works suggest that there may have been periods of vogue for particular disciples or topics with which they were connected, it is also evident that some anecdotes were repurposed so that some disciples were added to or left out of them. In Lunyu 5/8 a figure named Meng Wubo 孟武伯 asks Kongzi whether Zilu is benevolent. Kongzi answers that while he does not know whether Zilu is benevolent, the latter may be relied upon for collecting taxes in a state of a certain size. Two similar anecdotes appear in the Shiji. One appears in the Zilu section: Ji Kangzi asked, “Is Zhong You benevolent?” Kongzi said, “In a state of a thousand chariots he could be employed to manage the fu taxes, but I do not know whether he is benevolent.” Meng Wubo asked, “Is Zilu benevolent?” The Master said, “I do not know.” He asked again. The Master said, “In a state of a thousand chariots, [Zhong] You could be employed to manage the fu taxes, but I do not know whether he is benevolent.”31 季康子問:仲由仁乎? 孔子曰:千乘之國可使治其賦,不知其仁。 孟武伯問:子路仁乎? 子曰:不知也。 又問。 子曰:由也,千乘之國,可使治其賦也,不知其仁也。

There are several idiosyncrasies in this version of the passage. First, the same question and answer occur twice (or, actually, the question is asked three times), but the form of address used for Zilu changes twice across the passage. 30 31

Shiji 47.1938. Shiji 67.2192.

232

Csikszentmihalyi

The Lunyu is parallel to the second part of this passage only. It almost seems as if this passage is a less coherent version of a set of questions about two disciples that appears in the Ran Qiu section of the same Shiji chapter: Ji Kangzi asked Kongzi, “Is Ran Qiu benevolent?” He said, “In a city of a thousand households, or a house of a hundred chariots, [Ran] Qiu could be employed to manage the fu taxes, but as for his benevolence, I do not know.” He asked again, “Is Zilu benevolent?” Kongzi responded, “He is the same as [Ran] Qiu.”32 季康子問孔子曰:冉求仁乎? 曰:千室之邑,百乘之家,求也可使治其賦。仁則吾不知也。 復問:子路仁乎? 孔子對曰:如求。

Here, Ran Qiu takes center stage, and Zilu plays second fiddle. The Lunyu 5/8 and Shiji Zilu section parallels examined above do not mention Ran Qiu. It is possible that they reflect the combination of two separate source passages featuring two different interlocutors (i.e., a question about Ran Qiu asked by Ji Kangzi combined with a similar question about Zilu asked by Meng Wubo). There are two reasons that a more likely hypothesis is that these versions reflect a proxy battle being fought in the trenches of textual reformulation between the two traditions or practitioners of the method of particular disciples known for their management of estates. First, comparisons between Zilu and Ran Qiu are found elsewhere. Lunyu 11/22 and 11/24 are both examples of such comparisons, and parallels to both are included in the Shiji chapter. Second, the parallels surrounding 11/24 show a similarity to Lunyu 5/8 above. In this case, Lunyu 11/24 begins with a question about whether Zhong You (i.e., Zilu) and Ran Qiu are “great ministers” (dachen 大臣). Kongzi describes the conduct of great ministers and says that Zhong You and Ran Qiu are simply “serviceable ministers” (juchen 具臣). In the Shiji, however, the question is just about Zilu, and prose that compares Zilu unfavorably with great ministers is missing, even though Kongzi’s reply is still that he is a serviceable minister. In short, there appears to have been a time when the two disciples identified in the Shiji chapters as being experts in matters of governing were grouped and another point at which the same anecdotes were only about one of them. Interestingly, Lunyu 11/24 pairs the two disciples, while the Shiji omits Ran Qiu; and Lunyu 5/8 32

Shiji 67.2190.

Interlocutor Collections, The Lunyu, And Proto-lunyu Texts

233

omits Ran Qiu, while the Shiji incorporates two parallels that seem to reflect two different stages in the development of a passage that omits Ran Qiu. Rather than follow this interesting thread further, a preliminary conclusion would be that there was a porous boundary between passages comparing and contrasting interlocutors and versions of those passages where the spotlight was limited to a single interlocutor. With these observations in mind, it seems that the Shiji chapter’s own explanation of where it came from is false, or it reflects a relationship between this Shiji chapter and a very different Lunyu. On the surface, at least, the relation between the two texts is made explicit in the final note on the chapter itself: “completely taking from the Lunyu’s disciples’ questions, which I have put together and sequenced to make this chapter. Where anything was doubtful, I have omitted it” (悉取論語弟子問并次為篇,疑者闕焉).33 However, most of the quotations are not complete but heavily and systematically abridged. As seen above, the order of the disciples does not reflect the order of Lunyu 11/3. These different points indicate that at the very least, the construction of the chapter likely was not done by someone with something very much like a modern Lunyu in front of them. Some contrasts between the Shiji and the Lunyu support the idea that interlocutors stood in as proxies for deeper conflict. These texts demonstrate significant variation between the presentations of Han period versions of Kongzi dialogues. Foremost, of course, is the fact that the Shiji arranges quotations by disciples while the Lunyu has stretches of consistency only occasionally. In the first section of the Shiji, devoted to Yan Yuan, we find “Yan Yuan asked about benevolence. Kongzi replied, ‘By controlling oneself and returning to the rites, the people of the world will return “benevolence” to one’” (顏淵問仁,孔子 曰:克己復禮,天下歸仁焉), which appears to be a shorter version of Lunyu 12/1. However, in the fourth section, devoted to Zhong Gong, we find significant differences from Lunyu 12/2: Zhong Gong asked about benevolence. The Master said, “When you go out, behave as if you were receiving a great guest. When you employ the people, behave as if you were assisting at a great sacrifice. What you yourself do not desire, do not impose on others. Have no 33

Shiji 67.2226. Elsewhere in this volume, John Makeham (chap. 1) points out Huang Ren’er’s 人二 view that Lunyu dizi wen 論語弟子問 was the title of a work that “was an early version of the Lunyu that was compiled by Confucius’s disciples and included more details about the names of his disciples than the material we find in the received Lunyu.” Esther Klein’s translation (see chap. 9) does not take this phrase as the title of a different text, a view I find more convincing.

234

Csikszentmihalyi

resentment in your state; have no resentments in your family.” Zhong Gong said, “Although I am not clever, I will try to put these words into practice.” 仲弓問仁。子曰:出門如見大賓,使民如承大祭。己所不欲,勿施於 人。在邦無怨,在家無怨。仲弓曰:雍雖不敏,請事斯語矣。34

Zhong Gong asked about good government. The Master said, “When you go out, behave as if you were receiving a great guest. When you employ the people, behave as if you were assisting at a great sacrifice. Have no resentment in your state; have no resentments in your family.” 仲弓問政。孔子曰:出門如見大賓,使民如承大祭,在邦無怨,在家 無怨。35

These are not small differences between the Shiji and the received Lunyu. Putting zheng 政 (good government) in place of ren 仁 (benevolence) is consistent with the omission of a version of the Golden Rule (“Do not impose what one does not desire on others” 己所不欲勿施於人).36 It is worth looking into why, in these two texts, Kongzi gives two versions of the same answer to the same disciple, despite the fact that the disciple asks different questions in each text. Earlier we saw that certain passages on filial piety and other topics were likely labeled as Zengzi texts at some point in the Han. Zhong Gong’s situation was more complex, as there were two points of view in the Han about how to classify Zhong Gong. One version identifies Zhong Gong’s key ability as being in the area of “matters of governing.” In the Shuiyuan, Kongzi praises Zhong Gong as “one who can face south” (可使南面). That passage cites Kongzi as one who “clearly understands the way of the king” (明於王道),” and states that Zhong Gong is Kongzi’s equal in this area. The above version of 12/2 in the Shiji in which Zhong Gong asks about governing makes sense in that context. Yet Zhong Gong is also identified as the fourth of four disciples who excel at “virtuous action” in the Shiji and Kongzi jiayu typologies (which include all three disciples who ask Kongzi about benevolence in Lunyu 12). That label is internalized in the Lunyu version of Kongzi’s evaluation of Zhong Gong, as also seen in the Shuiyuan: “Kongzi took Zhong Gong to 34 35 36

Liu 1990: 12.485–486. Shiji 67.2190. Kaneto Mamoru (1974: 55–56) was the first to argue that the Shiji version was more original and that the subject was changed in the Lunyu version. For a contextualization of the Golden Rule in the early Chinese ritual-ethical context, see Csikszentmihalyi 2008.

Interlocutor Collections, The Lunyu, And Proto-lunyu Texts

235

have ‘virtuous action,’ saying, ‘[He] can face south’” (孔子以仲弓為有德行, 曰:…可使南面).37 While these two categories are not mutually exclusive, the

Zengzi case suggests that sometimes the proper identification of disciples was seen as important in the Han and so might have influenced the reconstitution of dialogic texts. Such considerations may have affected the organization of the Lunyu. While the parallel passages examined in this section only begin to hint at the relationship between the compilations of the two works Lunyu and Shiji, there is no doubt that they both were available in the Han, and the significant variations indicate a tolerance for different editions that the origin stories examined at the beginning of the chapter do not describe. The examples in this section demonstrate that even the Shiji chapter itself preserves evidence of jockeying between interlocutors and topics in the Lunyu and texts that preserve parallel passages to it. Again, this does not support the idea that various interlocutor texts were boiled down to an authoritative edition by the second generation of Kongzi’s disciples, as the Hanshu avers. Rather, the Zengzi texts and the Shiji/Lunyu parallels both indicate that attributions to different interlocutors were frequently manipulated in Han times.

Topical Consistency in the Lunyu

A hybrid in almost every way, the Lunyu contains sections of dialogue between Kongzi and a single disciple, as well as sections linked by a common topic.38 Topical texts united by common questions include the four pericopes 2/5–8, which have Meng Yizi 孟懿子, Meng Wu 孟武, Zi You 子游, and Zi Xia 子夏 “asking about filial piety” (問孝), as well as 12/1–3, which have Yan Yuan 顏淵, Zhong Gong 仲弓, and Sima Niu 司馬牛 “asking about benevolence” (問仁). Other “asking about benevolence” sections include 6/22 and 13/19, in which 37

38

For a complete translation of this passage from Shuiyuan 19, see Hunter 2012: 316. Note also that in Lunyu 13/1, Zilu asks about good government. The Zilu/Zhong Gong confusion or competition, then, also seems to apply to their interest in asking questions about good government. For the sake of precision, I define an interlocutor sequence as a run of three or more exchanges that grow out of different questions—as in a press conference, questions that are clearly not “follow-ups” to previous questions—between a single disciple and Kongzi. Its counterpart is a topical sequence, which is a set of three or more exchanges with different disciples on the same topic. I have not explored the chronological sequence possibility in this chapter but refer the reader to Brooks and Brooks 1998, which pays close attention to reconstructing chronology. The two categories of disciple and topical texts are not mutually exclusive, and some sections seem to feature both types of consistency.

236

Csikszentmihalyi

Kongzi provides two different answers to Fan Chi’s 樊遲 question, and 17/6, in which Zi Zhang 子張 “asks Kongzi about benevolence” (問仁於孔子). The five questions or scenes in 11/7–11, which end with references to Yan Yuan’s death, are examples of a topical text connected by common answers. The Lunyu also contains strictly defined interlocutor sequences. Pericopes 12/3–5 are dialogues between Kongzi and Sima Niu, and 12/17–19 are dialogues between Kongzi and Ji Kangzi 季康子. More common than that, however, are runs of statements by interlocutors who are not in dialogue with the Master. These include statements by Zengzi 曾子 (8/5–7 and 19/16–19), Zi Zhang 子張 (19/1–3), Zi Xia 子夏 (19/3–11), and Zigong 子貢 (19/20–25). These are similar to what we are calling interlocutor texts but may be distinguished from them by calling them independent-interlocutor sequences because they do not establish a personal relationship between the disciple and Kongzi.39 The consistency of topical and interlocutor sequences in the Lunyu suggests a strong authorial or editorial hand, making the appearance of parallel passages in other texts outside of such sequences particularly interesting from the perspective of text formation. For example, we have already seen that Sima Qian reproduces the middle pericope from the Lunyu sequence 12/1–3 that treats questions about benevolence from the three disciples Yan Yuan, Zhong Gong, and Sima Niu. In the Shiji, however, it appears as a question about government. Let’s consider two possible scenarios in light of the Shiji’s own story of Sima Qian’s cutting and pasting from the Lunyu to compose this chapter. Since the Shiji uses 12/3 without editorial changes elsewhere, a compiler could have started from the topical sequence 12/1–3. For this to have been the case, however, the Shiji compiler would have had to: (1) ignore the fact that this was the middle of three topically uniform pericopes in the text, (2) change the topic of 12/2 from benevolence to governing, and (3) remove the one clause of 12/2 associated with benevolence (or, at least, shu 恕) to render the resulting passage consistent with the understanding of Zhong Gong’s special skill with governance . The reverse process, a Lunyu compiler adding a version of the Golden Rule in order to fit it into a topical sequence about benevolence, is arguably

39

Categories are tricky. In the category of “almost interlocutor sequences” are 11/24–26, which all mention Zilu but are formally very different passages, and the nearby passages that concern Zigong, 5/9, 5/12, 5/13, and 5/15. “Nearly topical sequences” include 5/1–5 (Kongzi’s opinions of different disciples) and 8/18–21 and 10/1–26 (his views on the ancient sage-kings and on ritual matters, respectively). Note also that 13/1, 13/2, 13/16, and 13/17 are “asking about good government” (問政) passages (as are 12/7, 12/11, 12/14, 12/17, and 12/19).

Interlocutor Collections, The Lunyu, And Proto-lunyu Texts

237

easier to envision, although this by no means exhausts the possible editorial scenarios. The continuous reworking of individual passages is apparent when one compares three exemplars of Kongzi’s famous line from Lunyu 12/1: “Overcome oneself and return to ritual to act with benevolence” (克己復禮為仁). By contrast, the *Junzi wei li 君子為禮 (The Noble Man’s Practice of Ritual) manuscript from the Shanghai Museum corpus reads, “A gentleman performs the rites by relying on benevolence” (君子為禮以依於仁).40 These two formulations are different formally, and as they are followed by a similar exchange about benevolence, they might have been based on two different written formulations of the same teaching. So in the *Junzi wei li the discussion of benevolence and ritual propriety is the first of several consecutive pericopes that form a disciple text, while in the Lunyu it is the first of several pericopes that form a topically consistent text. A third parallel text, from the Zuozhuan 左傳 entry for year 12 of Duke Zhao 昭 (530 BCE), has the Master quoting the above Lunyu phrase (“Overcome oneself and return to ritual to act with benevolence”) but prefacing it with the phrase “In ancient times, there was a principle…” (古也有志).41 In this example, there is no sign of a disciple conversation. With respect to the question of the topically consistent passages in the Lunyu, the 12/1 and 12/2 parallels suggest that the topical consistency in 12/1–3 may have been a relatively late phenomenon.42 The interesting thing about this case is what it says about the fluidity of texts and the fact that individual pericopes flowed in and out of texts that exhibited topical or interlocutorial consistency. 40

41 42

The entire line reads, “The Master said, ‘Hui, a gentleman performs the rites by relying on benevolence’” (夫子曰:回!君子為禮以依於仁). Shanghai bowuguan cang Zhanguo Chu zhushu 2001–2011: 5.253–265. The manuscript version contains an expression found elsewhere in the Lunyu. In Lunyu 7/6, the Master says, “Be intent on the Way, occupy virtue, rely on benevolence, and wander in attainments” (志於道,據於德,依於仁,游 於藝). While there were likely more versions than these two, the text itself suggests that an editor added the three-character phrase “rely on benevolence” to a new context, because it either originated in the 12/1 passage (if the manuscript version is older) and was pasted into the 7/6 passage, or it originated in the 7/6 passage (if the Lunyu is older) and was inserted into the bamboo-slip version of the 12/1 passage. See Yang 1993: 1338–1342. One suggestion for an alternative reading for 12/1 would fit well with this reading. Kaneto Mamoru (1974: 52) notes the interesting possibility that yiri 一日 is actually a version of a note of a textual variant that was originally yi yue 一曰, implying that the Lunyu version incorporates alternative formulations such as the one described below. Then, we have the following variations of the first line of 12/1: Junzi wei li: 君子為禮以依於仁。/ Zuozhuan: 克己復禮,仁也。/ Lunyu: 克己復禮為仁。一曰:克己復禮,天下歸仁 焉。

238

Csikszentmihalyi

Turning back to the issue of the origin of the Lunyu, the fact that some “interlocutor texts” appear to have been formed at a relatively late date certainly disproves the notion that the Lunyu was static over the centuries leading to the Han. Does it say anything about the hypothesis that the Lunyu was formed out of the records of the disciples after the death of Kongzi? Evidence from the Han will likely never prove or disprove a hypothesis about events in the fourth century BCE. However, looked at another way, it does prove something about the origin stories that were told in the Eastern Han. We have seen how Kongzi’s interlocutors changed, and both the Zengzi texts and certain dialogues about Zilu and Zigong were still very fluid in the Western Han. Given this fluidity, it is hard to imagine that Eastern Han sources like the Hanshu and Lunheng could really know that the interlocutor texts that they posited as the origins of the Lunyu truly dated to antiquity. Having shown that texts about particular Kongzi interlocutors date only from the Han, we cannot rule out that others were from the pre-Qin period. But it does suggest that later writers who thought the interlocutor texts came from the pre-Qin periods were at least in part basing their views on an assumption that the Lunyu was ancient. While none of this is ironclad, the idea that some of the topically consistent stretches of the Lunyu were cobbled together from earlier works, and that some of the interlocutor-text stretches may have been reformatted from earlier works, calls into question the early accounts of the Lunyu being culled from disciple texts and suggests a much more fluid and complex model of text formation. Albeit by using somewhat arbitrary definitions, we have seen how pericopes circulated and were repurposed across several porous membranes. One is attribution, with “ancient principles” being attributed to Kongzi, and Kongzi sayings to his interlocutors. Another is type of text, with topical texts featuring sections elsewhere found in interlocutor texts, and interlocutor texts being constructed from topical texts. Indeed, the obvious thematic selections that show up in excavated texts show there is a centrifugal tendency exerted on disciple texts, while the “disciple vogue” would be a centripetal tendency. The Lunyu is a mixture of both tendencies, but by looking at its affiliated bundles of pericopes, it is possible that we can better understand the process by which individual sections were formed. While this is just a hypothesis, it is the product of a methodology that does not make assumptions about the relative ages of any of these texts and so is arguably better than simply finding some parallels in ancient texts and concluding that they are quoting the Lunyu and thus that the Lunyu is ancient.43 43

An example of this latter approach is Ding Yuanzhi 2007: 115–116. Ding argues that the exactness of parallels in works like the Guodian *Yucong 語叢 (Thicket of Sayings) manuscripts and other early texts associated with Zisi shows that the Lunyu was formed between 428 and 400 BCE.

Interlocutor Collections, The Lunyu, And Proto-lunyu Texts



239

Bibliography

Aurelius, Marcus. 2006. Meditations Trans. Martin Hammond. London: Penguin Books. Brooks, E. Bruce, and A. Taeko Brooks. 1998. The Original Analects: Sayings of Confucius and His Successors. New York: Columbia University Press. Chen Qiyou 陳奇猷. 2002. Lüshi chunqiu xin jiaoshi 呂氏春秋新校釋. Shanghai: Shang­ hai guji chubanshe. Csikszentmihalyi, Mark. 2008. “The Golden Rule in Confucianism.” In The Golden Rule in World Religions, ed. Jacob Neusner and Bruce Chilton, 157–169. London: Con­tin­ uum. Ding Yuanzhi 丁原植. 2007. Chutu wenxian yu Rujia xueshu yanjiu 出土文獻與儒家學 術研究. Taipei: Wanjuanlou. Duan Yucai 段玉裁. 1981. Shuowen jiezi zhu 說文解字注. Shanghai: Shanghai guji chubanshe. Hanshu 漢書. 1962. Beijing: Zhonghua shuju. Hu Lanjiang 胡蘭江. 2002. “Wen Weng Lidian tu xiaokao” 文翁禮殿圖小考. Zhongguo dianji yu wenhua 中國典籍與文化 2002.3:31–34. Huang Hui 黃暉. 1990. Lunheng jiaoshi 論衡校釋. Beijing: Zhonghua shuju. Hunter, Michael. 2012. “Sayings of Confucius, Deselected.” PhD diss., Princeton Uni­ versity. Ji Xusheng 季旭昇, ed. 2007. Shanghai bowuguan cang Zhanguo Chu zhushu duben 上海博 物館藏戰國楚竹書讀本, vol. 4. Taipei: Wanjuanlou. Jiang Boqian 蔣伯潛. 1985. Zhuzi tongkao 諸子通考. Hangzhou: Zhejiang guji chubanshe.Analects,” Kaneto Mamoru 金戸守. 1974. “Shiki ‘Rongo’ kō, sono ni: Chūji deshi retsuden yori Rongo seiritsu ni tsuite” 史記「論語」考、その二: 仲尼弟子列伝より論語成立に ついて. Shitennōji joshi daigaku kiyō 四天王寺女子大学紀要 7:51–70. Kern, Martin. 2007. “Excavated Manuscripts and Their Socratic Pleasures: Newly Discovered Challenges in Reading the ‘Airs of the States.’” Asiatische Studien / Études asiatiques 61.3:775–793. Knoblock, John, and Jeffrey Riegel. 2000. The Annals of Lü Buwei. Stanford, CA: Stanford University Press. Liu Baonan 劉寶楠. 1990. Lunyu zhengyi 論語正義. Beijing: Zhonghua shuju. Lunyu zhuzi suoyin 論語逐字索引. 1995. Chinese University of Hong Kong, Institute of Chinese Studies Ancient Chinese Concordance Series. Hong Kong: Commercial Press Pascal, Blaise. 1995. Pensées. Rev. ed. Trans. A. J. Krailsheimer. London: Penguin Books. Pi Xirui 皮錫瑞. 1954. Jingxue lishi 經學歷史. Beijing: Zhonghua shuju. Richter, Matthias. 2002. “Self-Cultivation or Evaluation of Others? A Form Critical Approach to Zengzi Li Shi.” Asiatische Studien / Études asiatiques 56.4:879–917.

240

Csikszentmihalyi

Shanghai bowuguan cang Zhanguo Chu zhushu 上海博物館藏戰國楚竹書. 2001–2011. Ed. Ma Chengyuan 馬承源. 8 vols. Shanghai: Shanghai guji chubanshe. Shiji 史記. 1959. Beijing: Zhonghua shuju. Sun Qizhi 孫啟治 and Chen Jianhua 陳建華. 1997. Guyishu jiben mulu: Fu kaozheng 古佚書輯本目錄: 附考證. Beijing: Zhonghua shuju. Sun Xidan 孫希旦. 1989. Liji jijie 禮記集解. Ed. Shen Xiaohuan 沈嘯寰 and Wang Xingxian 王星賢. Beijing: Zhonghua shuju. Uchino Kumaichirō 內野熊一郎. 1939. Shindai ni okeru keisho keisetsu no kenkyū 秦代 に於ける經書經說の硏究. Tokyo: Tōhō Bunka Gakuin. Wang Pinzhen 王聘珍. 1983. Da Dai Liji jiegu 大戴禮解詁. Beijing: Zhonghua shuju. Xiang Zonglu 向宗魯. 1987. Shuiyuan jiaozheng 說苑校證. Beijing: Zhonghua shuju. Yamada Katsumi 山田勝美. 1984. Ronkō 論衡. Tokyo: Meiji Shoin. Yang Bojun 楊伯峻. 1993. Chunqiu Zuozhuan zhu 春秋左傳注. Rev. ed. Beijing: Zhong­ hua shuju. Yang Chaoming 楊朝明. 2005. Kongzi jiayu tongjie 孔子家語通解. Taipei: Wanjuanlou. Yiwen leiju 藝文類聚. 1999. Shanghai: Shanghai guji chubanshe. Zhengshi yishu 鄭氏佚書. 1888. Ed. Yuan Jun 袁鈞. Hangzhou: Zhejiang shuju.

Sima Qian’s Kongzi And The Western Han Lunyu

241

Chapter 9

Sima Qian’s Kongzi and the Western Han Lunyu Esther Sunkyung Klein Around 100 BCE, Sima Qian 司馬遷 (b. 145 BCE) wrote the first extant biographies of Kongzi 孔子 and his disciples, including them as chapters in the Shiji 史記. One of his many sources for Kongzi material was an earlier version of the text now known as the Lunyu 論語 (Analects). Many scholars today consider the Lunyu to be the most important source for Kongzi’s thought, the standard against which all other sources are judged. Yet as other chapters in the present volume show, the Lunyu acquired its authoritative status only in the early Eastern Han, a few generations after Sima Qian. Sima Qian’s work was therefore written at what may have been a crucial moment in the Lunyu’s textual history and, as such, demands close attention as a source of evidence. First, I will consider the nature of Sima Qian’s Lunyu source. Most work on Kongzi in the Shiji assumes that Sima Qian had something like the full text of the received Lunyu. Recent scholarship questioning the Lunyu’s traditional chronology calls this into question, however. As Michael Hunter has pointed out,1 the Shiji is the first extant text to contain extensive Lunyu parallels that look like Lunyu quotations. If we let go of the assumption that Sima Qian’s Lunyu source was identical to the one we have today, the Shiji can perhaps provide interesting evidence about what an earlier version of that source might have looked like. My second goal in this chapter is to consider how Sima Qian uses his Lunyu material and how he views the Lunyu as a source. There is some evidence that in Sima Qian’s eyes the Lunyu was neither canonical nor especially well known. He relies on it heavily in his chapter on Kongzi’s disciples, but it is of merely secondary importance as a source for the chapter on Kongzi himself. Furthermore, he seems anxious to preserve its contents by including them in his work. This is not an attitude he takes toward texts that circulated widely in his time. Finally, I will consider an aspect of Sima Qian’s Kongzi that is less often studied: Kongzi as he appears outside his own biography. Shiji chapter 47, “Kongzi shijia” 孔子世家 (Hereditary House of Kongzi) is often read as a very human portrait, giving warmth to the scattered sayings of the Lunyu by placing them 1 See chapter 3 of the present volume.

© Koninklijke Brill NV, Leiden, 2018 | doi:10.1163/9789004382947_011

242

Klein

in a psychologizing narrative framework.2 This may or may not be an accurate characterization of the “Kongzi shijia.” However, elsewhere in the Shiji Sima Qian uses sayings of Kongzi (including Lunyu sayings) for quite different purposes. These scattered Lunyu quotations are often overlooked but are worth considering because of what they reveal about Sima Qian’s reading of the Lun­ yu as he knew it.

Overview of Sima Qian’s Lunyu

Those who question the Lunyu’s traditional chronology tend to argue that the Lunyu was compiled much later than traditionally believed, probably sometime in the Western Han. A particularly extreme version of this argument was put forth by Kaneto Mamoru, who argued that the Shiji was a source for the Lunyu rather than the other way around.3 However, Kaneto’s hypothesis is problematic. There is a high probability that Sima Qian had some source or sources for his Lunyu material. Furthermore, he does mention Lunyu by name as a source, in a passage discussed below. Kaneto does address the passage but suggests that it may be a later interpolation.4 Arguments about that passage aside, Shiji parallels provide witness for about a quarter of the full received Lunyu text. Some chapters are not attested at all (e.g., chaps. 16 and 20, which are particularly anomalous in form), while other chapters are very well represented (e.g., chaps. 5 and 11, each of which has more than half of its entries attested in the Shiji). Unattributed parallels by themselves cannot, of course, reliably indicate the direction of borrowing. However, Sima Qian for all his virtues was not a great innovator of Confucian thought: one would not expect to see him inventing Kongzi sayings ex nihilo. He was a historian, and his talent lay in selecting from and organizing his source documents so that they would tell an engaging narrative. We see this clearly for Shiji chapters where the direction of borrowing is far more secure. It is possible to say with a high degree of confidence, then, that Sima Qian did not compose all the roughly 130 Lunyu-parallel passages that are attested in the Shiji, that he had some source or sources of Kongzi materials. It is certainly possible that Sima Qian did have a “major source” for Kongzi materials, one from which the majority of the parallel passages might have been 2 Michael Nylan’s first chapter in Lives of Confucius (Nylan and Wilson 2010), Annping Chin’s Confucius: A Life of Thought and Politics (2008), and the film Kongzi (dir. Hu Mei, 2010) all use it for this purpose—to name only a few examples. 3 Kaneto 1978: 19. 4 Kaneto 1973: 17–19.

Sima Qian’s Kongzi And The Western Han Lunyu

243

drawn. Such a source, which might have included as much as a quarter of the received Lunyu, could reasonably be called a proto-Lunyu, whether it bore the title Lunyu or not. The pattern of Sima Qian’s Lunyu citations might then reveal something about what his proto-Lunyu looked like. For example, it is interesting (if a little speculative) to consider the holes in the Shiji-attested Lunyu—what is missing from it. One of Kaneto Mamoru’s more interesting observations is that Lunyu material associated with the disciple Zengzi is completely absent from the Shiji.5 Zengzi is the disciple of Kongzi who is most closely associated with the virtue of xiao 孝 (filial piety). Perhaps not by accident, Lunyu discussions of filial piety are also almost entirely absent from the set of Shiji/Lunyu parallels.6 Were Zengzi and filiality materials not yet part of Sima Qian’s Lunyu, or did he deliberately exclude them for some reason? The virtue of xiao is not particularly emphasized in the Shiji, with the character mostly appearing as part of various rulers’ posthumous names. But there is no striking pattern of avoidance either. Various figures are occasionally described as filial, and in the case of Lord Wanshi’s 萬石君 (i.e., Shi Fen 石奮, d. 124 BCE) descendants, their filiality is remarked upon no fewer than five times.7 Sima Qian also appears to be aware of the association between Zengzi and xiao, since Zengzi’s biography in Shiji chapter 67, “Zhongni dizi liezhuan” 仲尼 弟子列傳 (Arrayed Traditions of Zhongni’s Disciples; hereafter “Disciples”), contains the remark that “Kongzi considered that [Zeng Shen] was able to fully comprehend the way of filiality and thus transmitted it to him. He [Zeng Shen] made the Xiaojing [Classic of Filial Piety]” (孔子以為能通孝道,故授之業,作 孝經).8 But the “Kongzi shijia” contains no references to filial piety and none of its Lunyu sayings discuss it. This kind of thematic pattern is striking. Factors other than Sima Qian’s source documents may well be involved, however: filial piety was a political issue in the Han in general and of great personal

5 Kaneto 1978: 21. 6 Discussions of xiao or, in general, the proper attitude toward one’s parents, appear in Lunyu 1/2, 1/6–7, 1/11, 2/5–8, 2/20–21, 4/18–21, 8/2, 8/21, 9/16, 11/5, 13/18–20, 17/9, 17/21, and 19/17 (twenty-four items in total). Only one of these, Lunyu 11/5, has a Shiji parallel (Shiji 67.2189). A characterization of the disciple Min Ziqian 閔子騫, it is not a substantive discussion of fil­iality in theory or practice but rather a comment that Min Ziqian’s parents and brothers had nothing objectionable to say about him. 7 Shiji 103.2764–2768. 8 Shiji 67.2205.

244

Klein

importance to Sima Qian in particular. Thus, it would be reasonable to suspect the Shiji’s treatment of this value to be somewhat complicated.9 Perhaps a clearer case is Lunyu items which consist of a disciple’s own pronouncements in the absence of the Master. Some of the biggest holes in the Shiji-attested Lunyu are these types of items, most notably in chapters 1, 8, and 19.10 If these pericopes were part of Sima Qian’s Lunyu at all, he did not consider them worth preserving—with two notable exceptions. The Shiji does contain Lunyu parallels for one independent saying by Yan Yuan 顏淵 (Lunyu 9/11) and three independent sayings by Zigong 子貢 (Lunyu 1/10, 5/13, 19/22). The last of these is the only saying from Lunyu chapter 19 with a Shiji parallel. No doubt it is significant that all four sayings actually describe the Master himself rather than being truly independent expansions on his ideas. Nonetheless, the compiler(s) of the received Lunyu placed the Zigong passages from chapter 1 and chapter 19 within long sequences of independent disciple sayings, suggesting that they were considered comparable to the others. Yet for Sima Qian, they seem to have been different—assuming, again, that the unattested sayings were there in his source(s) at all. In this brief overview, I have argued that Sima Qian had a Lunyu, or at least some proto-Lunyu source(s). On the other hand, it may be that Sima Qian’s Lunyu looked very different from the one we have today—it might have been much shorter, for example. Just because one pericope from a chapter seems to have been part of Sima Qian’s Lunyu does not imply that the entire chapter was.11 His Lunyu source(s) could easily have been organized differently, especially at the beginning and end. In the remainder of this study, I will continue to entertain the question of what Sima Qian’s Lunyu might have looked like. However, I turn now to a more focused study of how Sima Qian used his protoLunyu as a source in various parts of the Shiji. Overall, the Shiji uses the Lunyu-like materials in three distinct ways. We see Lunyu parallels in places where the historian reflects on his own work, the socalled taishigong yue 太史公曰 (the honorable senior archivist says) passages. 9 10 11

For a discussion of the political implications of xiao in the Western Han, see, e.g., Knapp 1995: 219–221. It may be worth noting that the final chapter of the Shiji explicitly connects Sima Qian’s work on that text to his father’s deathbed lament (Shiji 130.3295). Disciple-only pericopes are found in Lunyu 1/2, 1/4, 1/7, 1/9–10, 1/12–13, 4/26, 5/13, 6/9, 8/3–7, 12/24, and 19/1–25 (i.e., the whole of chap. 19). The unattested sayings in chapter 8 are a long sequence of Zengzi sayings, already mentioned above. Mark Csikszentmihalyi’s (2004: 33n47) reflection on Erik Maeder’s work on the Mozi (1992) is particularly useful in this regard. He writes, “Dating [a] text by existing chapter titles is problematic, because the oldest stratum of the text appears as a kernel of many chapters, surrounded by later argument and elaboration.”

Sima Qian’s Kongzi And The Western Han Lunyu

245

These evaluative comments are often embroidered with quotations, many (although certainly not all) of which have Lunyu parallels. The second way the Shiji uses material with Lunyu parallels is to build up the portrait of Kongzi in the chapter devoted to him. There Lunyu-parallel quotations are liberally mixed with other sources, as I will discuss in a moment, and show some intriguing variants. Finally, the Shiji uses the Lunyu parallels extensively in a separate chapter, chapter 67, devoted to Kongzi’s disciples.

The Shiji on Kongzi’s Disciples

I start with the chapter on the disciples, because the Lunyu plays a surprisingly major role here. It is also in this chapter that the Lunyu is actually mentioned by name. The historian’s comment at the end of the “Disciples” chapter is clearly important, yet exceedingly difficult to interpret: 太史公曰學者多稱七十子之徒譽者或過其實毀者或損其真鈞之未睹厥 容貌則論言弟子籍出孔氏古文近是余以弟子名姓文字悉取論語弟子問 并次為篇疑者闕焉12

I have left out modern punctuation because editions tend to disagree on how to punctuate and interpret the passage. This is an awkward and difficult passage, but it is possible to make sense of it: there is no need to interpret it as a clumsy interpolation. The beginning of the passage is not seriously problematic. “Many scholars talk about the seventy disciples of the Master. In praising them, some exagge­ rate the reality. In defaming them, some diminish the truth.” Afterward, the interpretation becomes much more difficult. The crucial questions are, how many sources are there, what are they called, and how are they being used? If one wants a balanced appraisal without being able to personally see the disciples’ faces or demeanors, then one could discuss their words. The “Register of Disciples” that comes out of the Kong family ancient script [texts] approaches correctness. I have used [this list of]13 the disciples’ names and surnames, and their patterned phrases, and entirely extracted

12 13

Shiji 67.2226. I.e., the “Register of Disciples”?

246

Klein

the questions asked by disciples in the Lunyu. I combined and arranged [these materials] to make a chapter. What was doubtful I left blank.14 I should emphasize again that the parsing of this passage is highly controversial, but I take there to be two texts, a “Register of Disciples” (which lists names and particulars) and the Lunyu. The title of the former text might alternatively be parsed as “Discussing and Talking about the Register of Disciples” (論言弟 子籍).15 One or both of these is probably also being referred to as the “Kong family ancient script” (孔氏古文). From the grammar, as Burton Watson’s translation suggests, it seems as if it is the “Register of Disciples” that “comes out of” the Kong family ancient-script texts. But John Makeham has argued that the Lunyu was among the texts that were supposedly discovered in the walls of the Kong family home, thus making it also a plausible choice of referent for this descriptor.16 Whatever the case, the comment raises the possibility that there were at least two different source texts for the chapter. The structure of the chapter itself gives some indication of how to interpret the passage. The chapter does seem to take as its base a list of disciples’ names and, after each entry, optionally includes extra information about them, often (but not only) in the form of selections from the Lunyu. It would make sense to identify the base list as the “Register of Disciples” mentioned in the historian’s comment. An instructive parallel can be found in chapter 38 of the Kongzi jiayu 孔子家語 (Family Sayings of Kongzi), the “Qishi’er dizi jie” 七十二弟子解 (The Seventy-Two Disciples, Explained).17 Although the Kongzi jiayu list appears to be later (as we would expect),18 both it and the Shiji “Disciples” chapter give the disciple’s name, courtesy name, place of origin (where available), and age relative to Kongzi. The “Qishi’er dizi jie” often includes a single brief anecdote of biographical relevance after the disciple’s name and particulars. 14

15 16 17 18

Csikszentmihalyi (2004: 29) translates the last part of this passage as “I have made this chapter by gathering information about the disciples’ full names, their speech and their writing, from the disciples’ questions in the Analects, and when anything was doubtful I have omitted it.” By comparison, Burton Watson (1958: 190) translates: “There is a register of disciples preserved by the Kong family and written in archaic characters which I believe to be generally accurate. I have used this list of names and surnames of the disciples, combining it with the questions and answers of these men found in the Analects, and arranged the material in this chapter.” Jin Dejian (1963: 205–208), in his study of Sima Qian’s sources, understands it in this manner, punctuating it 論言弟子籍. Makeham 1996: 14. Kongzi jiayu zhuzi suoyin 38.65–72. The Kongzi jiayu is generally accepted to be a later compilation, perhaps owing its present form to Wang Su 王肅 (195–256); see discussion in Kramers 1993.

Sima Qian’s Kongzi And The Western Han Lunyu

247

The Shiji contains many of the same anecdotes but in a more expansive form; they often seem to derive at least partially from multiple sources. Interestingly, however, the “Qishi’er dizi jie” includes more biographical information on the disciples than was apparently available to (or at least utilized by) Sima Qian. Thus, the Shiji could not have been its sole source.19 This pattern of variation suggests what the Shiji evaluation also tells us: that there was an independent “Register of Disciples” that both the Shiji and the Kongzi jiayu made use of. One can remain agnostic as to the origins and compilation of this allegedly ancient “Register of Disciples”: parts of it might go all the way back to the time of Kongzi’s death, but it could also have been a product of the Han dynasty “disciple vogue” discussed by Mark Csikszentmihalyi, an attempt to fill out the quasi-magical complement of seventy disciples (give or take).20 Most names later in the list lack all but the most basic details. In the Shiji “Disciples” chapter, selections drawn from (some form of) the Lunyu (and apparently other sources) then function as a kind of commentary on the disciple list. One difficult issue here is how to take xi 悉, which I have translated as “entirely.” Commentators tend to understand it in two different ways: either as saying that all the material in the “Disciples” chapter is taken from the Lunyu text, or that all the material in the Lunyu text is included in the chapter. Neither of these appears to be true, at least as far as the received Lunyu is concerned.21 Of course, Sima Qian doesn’t simply say “Lunyu” but rather “questions asked by disciples in the Lunyu.” Even so, there are many questions asked by disciples in today’s Lunyu that Sima Qian does not include here. Interestingly, however, if we look at the roughly 25 percent of the Lunyu that is attested in the Shiji, the statement is strikingly accurate. With only a very few exceptions, all the questions asked by disciples that are attested in Sima Qian’s

19 20 21

To give just one of many examples, the disciple Zai Yu 宰予 is not assigned a place of  origin in the Shiji (67.2194), whereas in the Kongzi jiayu he is said to be a person of Lu (Kongzi jiayu zhuzi suoyin 38.66). See chapter 8 in this volume. Han Zhaoqi 韓兆琦 (2004: 67.3920) cites Wang Juntu’s 王駿圖 comment: “If you look at what it says in [Sima Qian’s] evaluation, everything that does not appear in the Lunyu he basically does not dare to record” (觀其自贊所言凡不見於論語者概不敢錄). However, even leaving aside the “Register of Disciples” framework, there is a great deal of other material in the “Disciples” chapter that cannot be found in today’s Lunyu. Wang Ruoxu 王若虛 (1174–1243), on the other hand, is fully aware of the non-Lunyu material: “Among the eclectic remarks and vulgar affairs cited by [Sima] Qian, there are many that lack credibility. How could they all be recorded in the Lunyu?” (遷所引雜說鄙事,有不 足信者矣,又豈皆論語所載也) (Han 2004: 67.3920).

248

Klein

Lunyu appear in the “Disciples” chapter.22 Furthermore, the number of passages is large enough to make this significant. It would be unwise to place too much weight on this argument, since it relies on one word in one passage. However, it deserves to be considered as evidence that Sima Qian’s Lunyu was indeed different from the received text and perhaps much smaller. It might even be worth considering whether passages found in the “Disciples” chapter and not in today’s Lunyu might not have originally been part of Sima Qian’s Lunyu. Turning to the distribution of Lunyu passages used in the “Disciples” chapter, we see a rather interesting pattern: the majority of them (thirty-seven of the fifty-two) are clustered in what are now received Lunyu chapters 5, 6, 11, and 12. Again, this seems potentially significant from the point of view of the compilation of the Lunyu. It cannot be taken as conclusive, of course, but it might provide at least some hints about the structure of Sima Qian’s source. As we might expect from the comment at the end of the “Disciples” chapter, the Shiji itself shows signs of marked disciple factionalism, a fascination with the disciples as a set and with the fixing of their respective status in some kind of moral or philosophical hierarchy. The evaluation from the “Disciples” chapter mentions exaggerations of both praise and blame. Sima Qian himself appears to engage in some of this as well, both in this chapter and elsewhere in the Shiji. The discussion in Shiji chapter 124, “Youxia liezhuan” 游俠列傳 (Arrayed Traditions of Roving Warriors), implies that Ji Ci 季次 and Zisi 子思 were especially admired by many in Sima Qian’s time but not in their own: “If we 22

Fifty-two disciple dialogue passages in the Lunyu are also attested in the Shiji “Disciples” chapter. Six additional disciple dialogues are found both in the Lunyu and elsewhere in the Shiji (Lunyu 7/15, 13/3, 15/2, 15/11, 17/5, 17/7). Of these, Lunyu 7/15 is paralleled in the “Bo Yi” 伯夷 chapter (Shiji 61.2122), but we find there only the Master’s pronouncement—that Bo Yi and Shu Qi 叔齊 “sought benevolence and received it” (求仁而得仁)—not the disciple dialogue that frames the saying. Lunyu 13/3, the “rectification of names” passage, is again an understandable exception. It is paralleled (in much abbreviated form) in two places. In Shiji 23, “Li shu” 禮書 (Treatise on the Rites), we find only Kongzi’s pronouncement that “it is necessary to rectify names” (必也正名) (Shiji 23.1159). In the “Kongzi shijia,” it is a dialogue with the disciple Zilu, but it is so much more crucial to the portrait of Kongzi than to that of Zilu that it is understandable why it should be included in the “Kongzi shijia” (Shiji 47.1933) instead of the “Disciples” chapter. Similar reasoning applies to Lunyu 15/2, which is part of the Chen-Cai story and is also found in the “Kongzi shijia” (Shiji 47.1930, discussed below). As with Lunyu 7/15, the Shiji parallel for Lunyu 15/11 is not a dialogue (Shiji 3.109). Finally, Lunyu 17/5 and 17/7 both concern Kongzi’s temptation to join rebellions. He is dissuaded by Zilu 子路, but again, the episodes show more about Kongzi’s character than they do about Zilu’s and are thus included in the “Kongzi shijia” (Shiji 47.1914, 1924).

Sima Qian’s Kongzi And The Western Han Lunyu

249

sincerely compare the warriors of the hamlets and villages to men like Ji Ci and Yuan Xian [= Zisi] in terms of actual authority and the power and effect of their actions in their own times, then the former so far surpass the latter that they can hardly be discussed on the same day” (誠使鄉曲之俠,予季次﹑原憲比權 量力,效功於當世,不同日而論矣). Sima Qian’s remark, that the roving warriors were actually more influential than the impoverished disciples, would draw the ire of many generations of readers until it was reinterpreted as a covert autobiographical statement during the Northern Song.23 However, in the context of the Shiji as a whole, the statement should perhaps be interpreted not as anti-Confucian but (for reasons discussed below) as pro-Zigong. In chapters other than the “Roving Warriors,” Zisi in particular is contrasted unfavorably with Zigong. A passage in chapter 129, “Huozhi liezhuan” 貨殖列傳 (Arrayed Traditions of Merchants), defends the pursuit of profit on the grounds that it gives one the power to do more good than would a pure but impoverished lifestyle: Zigong, having studied with Zhongni [= Kongzi], [later] withdrew and held office in Wei. He bought cheap, stored goods and sold them at a profit in the regions between Cao and Lu. Of the seventy disciples, Si [= Zigong] was the wealthiest. Yuan Xian [= Zisi] could not even get his fill of coarse grain and concealed himself in an impoverished alley. Zigong had a fine team of four matched horses and was attended by mounted followers. With gifts of bundles of silk, he paid visits to the feudal lords. Wherever he went, the ruler of the state would always divide the courtyard and treat him as an equal. The one who was responsible for spreading Kongzi’s fame throughout the realm was none other than Zigong from beginning to end. Is this not what is called one who “gets power and increases his brilliance”? 子贛24既學於仲尼,退而仕於衛,廢著鬻財於曹﹑魯之閒,七十子之 徒,賜最為饒益。原憲不厭糟糠,匿於窮巷。子貢結駟連騎,束帛之 幣以聘享諸侯,所至,國君無不分庭與之抗禮。夫使孔子名布揚於天 下者,子貢先後之也。此所謂得埶而益彰者乎?25

23 24

25

See discussion in Klein 2010: 277–283. Variant of 貢, which occurs several times in the Shiji. As in this passage, sometimes both variants occur very close together. It is unclear why the text sometimes uses this variant and sometimes the other. One could speculate that it indicates different sources, but it could just as well reflect inconsistent substitution by later editors. Shiji 129.3258.

250

Klein

Returning to the “Disciples” chapter, the contrast there between the biographies of Zigong and Zisi is also telling. Zisi’s biography consists only of a version of the dialogue now found in Lunyu 14/1, about how it is shameful to act only for monetary gain. We also find a story about the wealthy Zigong going to visit the impoverished Zisi and being verbally (and morally) bested by the former.26 The Shiji biography of Zigong, on the other hand, is much longer than Zisi’s. It begins with five dialogues that loosely parallel Lunyu entries (Lunyu 5/9, 5/4, 19/22, 1/10, and 1/15). It then gives an extremely involved account of how Zigong was able to save Lu through a series of manipulative political persuasions. The summary reads, Thus, in one expedition, Zigong preserved Lu, formented chaos in Qi, broke Wu, strengthened Jin, and made Yue into a hegemon. Zigong in one mission pitted the strategic powers [of the states] against one another. Within ten years, these five states all had some upheaval. 故子貢一出,存魯,亂齊,破吳,彊晉而霸越。子貢一使,使勢相 破,十年之中,五國各有變。27

The final statement in Zigong’s biography sums up his career by emphasizing his expertise in turning a profit, his agreeable but honest character, his service in both Lu and Wei, and the wealth accrued by his family. In short, it seems that in Sima Qian’s view Zigong was an active and powerful player in historical events and the true hero of Confucian tradition. Zisi, on the other hand, was genuinely virtuous but insignificant and, in Sima Qian’s time, vastly overrated. Portraying the two disciples in this way does not seem to be merely a shallow admiration for wealth and power. Instead, for Sima Qian these two seem to have represented a split in what it meant to be a follower of Kongzi. Sima Qian’s portraits bring out tensions that are at most implicit (or perhaps absent) in the Lunyu. Did virtue and failure necessarily—tragically— go hand in hand? Or could one be a good follower of Kongzi and also prosper like Zigong? In seeming to answer in the affirmative, Sima Qian does betray a certain bias toward Zigong. On the other hand, perhaps he was just trying for a balanced appraisal, correcting what he saw as an excessive bias in favor of Zisi.

26 27

For a discussion of problems relating to the identification and biography of Zisi, see Csikszentmihalyi 2004: 86–100. Shiji 67.2301.

Sima Qian’s Kongzi And The Western Han Lunyu



251

The Shiji “Kongzi shijia”

At this point, I turn to the Shiji chapter on Kongzi himself. I am not going to present a full analysis of this chapter, because it has already been heavily studied.28 Instead, I will again focus on its relationship with the Lunyu, mentioning just a few of the more interesting features. As with the “Disciples” chapter, I begin at the end. Sima Qian’s comment at the end of this chapter does not mention the Lunyu. Instead, he mentions reading Kongzi’s own writings, by which he appears to mean commentaries on the classics (as those are the writings he attributes to Kongzi in the chapter devoted to him).29 He also describes a trip he made to visit the Kong family ancestral home. It is interesting that what Sima Qian emphasizes is the living ritual performance being done there, not textual records. But did he obtain special Kongzi lore during his trip? Did he give it any credence? If the Lunyu really was the historically important text that the Hanshu and popular belief make it out to have been, it seems as if Sima Qian ought to have been able to get a good look at it during his trip; however, he says nothing about it either way. Sources for the “Kongzi shijia” do appear to be much more heterogeneous than those for the “Disciples” chapter. As so often in the Shiji, we see Sima Qian’s much-criticized tendency to be inclusive rather than selective. Anecdotes from the Guoyu, where Kongzi displays arcane knowledge of weird phenomena,30 are placed alongside the ethical pronouncements for which the Lunyu Kongzi is known. Most of all, the star of the “Kongzi shijia” is an active participant in the events of his time and a successful custodian of the textual tradition. 28 29

30

For some useful discussions in English, see Durrant 1995: 29–69; Hardy 1999: 153–168. In Chinese, a thorough study of the chapter in comparison with other sources is Qian 2002. The specific phrase that he uses is Kong shi shu 孔氏書, and he says that reading this writing allows him to “see in his thoughts the person Kongzi was” (想見其為人; Shiji 47.1947). It is tempting to connect this to chu Kong shi guwen 出孔氏古文 (that which comes out of the Kong family ancient script [texts]) in the “Disciples” chapter discussed above. However, when the Shiji elsewhere uses the appellation Kong shi in connection with Kongzi’s writings, it tends to be clearly related to the classics: “Kong shi made the Spring and Autumn [Annals]” (孔氏著春秋; Shiji 110.2919); “Kong shi had an ancient-script Exalted Documents” (孔氏有古文尚書; Shiji 121.3125); and “thus, the Documents tradition and the Record of Rituals came from Kong shi” (故書傳、禮記自孔氏; Shiji 47.1936). In the latter two cases, it is unclear whether Kong shi is meant to refer to Kongzi or to his clan, and elsewhere in the Shiji, it is often the latter. Given the Shiji’s complicated notion of jia 家 and collective family involvement in authorship, the slippage here is unsurprising. E.g., the identification of a mysterious buried animal or a giant bone (see Shiji 47.1912– 1913).

252

Klein

Focusing on the chapter’s use of Lunyu parallels, though, we see some interesting tendencies. Shiji passages with Lunyu parallels tend to get assigned a date and a historical context. This aspect of the “Kongzi shijia” is traditionally explained by saying that Sima Qian had the Lunyu sayings and made guesses as to their contexts. A later date for the received Lunyu opens up other possibilities. Could it have been that both Sima Qian and the Lunyu compilers had a common source which contained these quotations embedded in stories? Did Sima Qian anthologize (or rework) the full stories while the Lunyu compilers only stripped out the quotations? Interesting as these suggestions are, I would argue that the traditional picture of the Shiji “Kongzi shijia” chapter’s compilation is closer to correct. The evidence either way is somewhat circumstantial, but a reading of the Shiji’s use of the Lunyu in this chapter does yield a few curious observations. First, there is some evidence that Sima Qian’s Lunyu source was its own text, not just drawn from scattered parallels in other texts. It happens that one of the few Lunyu items paralleled in the Mengzi (7B/37) is Lunyu 5/22: Kongzi was in Chen and said, “Perhaps we should return home. The men of my faction are wild and rustic, but in being promoted they will not forget their origins.” 孔子在陳曰:盍歸乎來!吾黨之士狂簡,進取,不忘其初。

The Master was in Chen and said, “Let us return! Let us return! The young men of my faction are wild and rustic, their accomplishments are becoming known, but they do not know how to be restrained.” 子在陳曰:歸與!歸與!吾黨之小子狂簡,斐然成章,不知所以裁 之。

Sima Qian has Kongzi pronounce the Mengzi version just before leaving Chen.31 From Chen he goes on to Wei for a short time, and then returns to Chen. There he makes what is essentially the same remark, this time in the Lunyu version.32 This would seem to reflect a certain carelessness in compilation practice. As the Shiji commentator Sima Zhen points out, these are two versions of the

31 32

Shiji 47.1923. Note that the Mengzi version also appears in Shiji 121.3117 as part of a discussion of the strong tradition of rites and music in the state of Lu. Shiji 47.1927.

Sima Qian’s Kongzi And The Western Han Lunyu

253

same saying, not two different sayings.33 But it does tell us something really valuable: that Sima Qian apparently had two different sources for the saying, sources that he understood to be distinct. Another curious item hints that at least part of Sima Qian’s Lunyu may have looked similar to the received version—that is, that the sayings there lacked context and were similarly (dis)organized. One of the major events in Kongzi’s life is always understood to be the occasion on which he and his disciples ran short of provisions between Chen and Cai. There are a great many accounts of this event in different Warring States texts, and it has been used for widely varying purposes.34 In the Shiji, the story includes a passage that is a close parallel to Lunyu 15/2, which is explicitly set during the Chen/Cai incident:35 Zilu, showing his vexation, said, “Does the princely man, too, get into such dire straits!?” Kongzi said, “The princely man holds firm when in such dire straits. The petty man in dire straits is the one who loses control like that.” 子路慍見曰:君子亦有窮乎?孔子曰:君子固窮,小人窮斯濫矣。

In the Lunyu, however, nothing in the subsequent passage (15/3) indicates that it has anything to do with Chen and Cai: Kongzi said, “Si, do you take me for one who has studied much and understood it?” [Zigong] said, “Yes. Is it not so?” Kongzi said, “It is not so. I use one [thread] to string it all together.” 孔子曰:賜,爾以予為多學而識之者與?曰:然。非與?孔子曰:非 也。予一以貫之。

 Lunyu 15/3 is usually taken to be related to Lunyu 4/15, where Kongzi also mentions the “single thread,” not to any version of the Chen/Cai story. But in the Shiji “Kongzi shijia,” Sima Qian adds a line of narrative to connect the two passages, making it clear that he understood this dialogue with Zigong as having taken place during the Chen/Cai crisis.36 The Lunyu 15/2 parallel is followed by 33 34 35 36

Shiji 47.1927. See Chen 2003; Makeham 1998. The Lunyu version begins “In Chen, their provisions were cut off …” (在陳絕糧). The Shiji closely parallels this, though also including a line about Kongzi singing unconcernedly. Shiji 47.1930.

254

Klein

the line “Zigong looked upset” (子貢色作), after which comes the Lunyu 15/3 parallel and then additional Chen/Cai narratives. The Lunyu 15/3 saying does not add much to the Chen/Cai narrative, nor does it relate to the issues at hand. However, it does suggest that in Sima Qian’s Lunyu source, 15/2 and 15/3 were sequential as they are today. This at least is a plausible explanation for why Sima Qian would take them as being chronologically related. The above two examples suggest that Sima Qian’s source, at least in some sections of the text, was organized in a manner similar to the received Lunyu. The third type of evidence points in the opposite direction, suggesting that there may also have been real differences between Sima Qian’s source and the received Lunyu. Although the Shiji often adds to Lunyu parallels by supplying them with a context, an opposite tendency can also be detected in the Shiji’s use of the Lunyu: many of the Shiji’s Lunyu parallels are shorter than their counterparts in the received Lunyu. This is especially apparent and revealing in a sequence of meetings between Kongzi and mysterious men who criticize him. For example, a stranger comes by and hears Kongzi playing the stone chimes (Lunyu 14/39), after which he criticizes Kongzi for stubbornly trying to do the impossible and refusing to compromise: “How stubborn is that kengkeng sound. If no one understands him, he should just quit. If it is deep, cut through. If it is shallow, skim over” (鄙哉!硜硜乎!莫己知也,斯已而已矣。深則厲,淺則揭). In the Lunyu version, Kongzi has a snappy response, criticizing the stranger in turn: “That would get results all right. There is no arguing with [such people]” (果哉!末之難矣). In the Shiji, he has no rejoinder. But on the other hand, the stranger’s remark is also less critical: “Keng! It’s just that no one understands him” (硜乎,莫己知也夫而已矣).37 It is as if each text chooses a different way to make Kongzi look better. In Lunyu 18/7, a hermit speaks with Zilu and implicitly criticizes Kongzi’s teachings. The Lunyu version is much more elaborate and ends with a long comment by Zilu on why Kongzi is right and the hermit is wrong. In the Shiji version, there is no Zilu speech. They simply go back to look for the hermit, and he has disappeared.38 Such variants suggest that the Shiji and the present Lun­ yu versions of these passages may both derive from a third source text. Perhaps the present Lunyu version was edited or added to by a Confucian apologist, and Sima Qian’s version was not, or was edited differently. Anecdotes that bear traces of anti-Confucius attitudes might be a special case. Perhaps they were not part of the text I am calling Sima Qian’s Lunyu 37 38

Shiji 47.1925. Shiji 47.1929.

Sima Qian’s Kongzi And The Western Han Lunyu

255

source. Perhaps they represented a separate bundle of bamboo strips. On present evidence, it is impossible to say. But the point is worth noting: even where the Shiji contains close parallels with Lunyu passages, one should not assume that it is quoting from a text identical to the received Lunyu. A last feature of the “Kongzi shijia” worth mentioning here is some strikingly messy passages near the end. As mentioned above, the Shiji tends to be inclusive. Remarks within the Shiji itself support this characterization, such as part of a dialogue in chapter 130, “Taishi gong zixu” 太史公自序 (Honorable Senior Archivist’s Self-Narration), where Sima Qian promises his dying father that he will faithfully preserve the texts their family has collected.39 Although Sima Qian did not take everything from all of his sources, he might have included as much of his Lunyu or proto-Lunyu as he could manage, perhaps even out of anxiety that it might otherwise be lost. Given the scope of his project, however, it may have been inevitable that Sima Qian did not get around to discussing or contextualizing some sections which did not fit clearly into any of the Shiji chapters. For example, near the end of the “Kongzi shijia,” after having discussed Kongzi’s work on each of the Six Classics, the Shiji narrative devolves into a series of miscellaneous sayings, all with Lunyu parallels.40 This seems significant. The chapter thus far has freely mixed together all kinds of different source material. But here the scope narrows to just the Lunyu. The parallels are also often abbreviated, variant, or switched around, and they are almost all drawn from chapters 7, 9, and 10 of the received Lunyu.41 First, I do not think it is reasonable to say that this section is an interpolation just because it is not smoothly integrated into the narrative flow. In writing about the Shiji, it is fashionable to focus on chapters that meet one’s own contemporary standards of literary artfulness and ignore chapters that do not. However, various parts of the Shiji show that such aesthetic standards were not the only ones to influence the Shiji’s compilation (the ten tables [Shiji chaps. 13–22] serve as the most striking examples of non-narrative content). Add to that the inclusive tendency mentioned earlier, and the presence of a “miscel39

40 41

Sima Tan says, “I am deeply afraid that the archives and patterned writings of the realm will be lost” (廢天下之史文,余甚懼焉). Sima Qian replies, “Although your little son is not very intelligent, he requests to discuss all the ancient knowledge that the one who came before has put in order. He will not dare to be remiss” (小子不敏,請悉論先人所 次舊聞,弗敢闕) (Shiji 130.3295). Shiji 47.1938–1941. Specifically, eight of the sayings have parallels in the received Lunyu chapter 7, six have parallels in chapter 9, and seven have parallels in chapter 10. There is also one saying that has a parallel in chapter 5.

256

Klein

laneous” section like this one should not necessarily be taken as evidence of textual corruption. If one considers some of the actual content, the section begins in a fairly orderly manner, though perhaps not by present notions of good style and organization. It starts with “fours”: Kongzi had four [types of] teachings: culture, conduct, loyal striving, and trustworthiness. 孔子以四教:文,行,忠,信。

It goes on: [He] refused four things: be without speculation, without insistence, without inflexibility, and without egotism. 絕四:毋意,毋必,毋固,毋我。

It then moves to two sets of “threes”: What he was cautious about: fasting, war, and sickness. 所慎:齊,戰,疾。

The Master seldom spoke about what was profitable or fated or benevolent. 子罕言利與命與仁。

The number game continues, with “taking one corner and returning three” (舉一隅以三隅反) followed by two sayings that are heavy on reduplicative binomes. But then we get a series of descriptions that just detail Kongzi’s behavior in various situations. After that we have a few sayings where he talks about himself. In short, this passage is not merely a disorganized “junk drawer” for Confucius-related textual material. The compiler has made an attempt to organize these passages, as is especially clear in the numerical section at the beginning. Though this is not necessarily evidence that the compiler was Sima Qian, I would argue that it is not necessarily evidence for the opposite conclusion either.

Sima Qian’s Kongzi And The Western Han Lunyu

257

Passages like this one support middle-of-the-road conclusions about Sima Qian’s Lunyu. On the one hand, he likely had a Lunyu-like source. The miscellaneous section in the “Kongzi shijia” seems to consist of items that Sima Qian could not fit in elsewhere but considered worthy of preservation. Also, some of the structure of the received Lunyu seems to be visible in the Shiji. There is a clustering of parallels from chapters 5, 6, 11, and 12 (regarding disciples, discussed above), and in the “Kongzi shijia” there is a clustering of items from chapters 7, 9, and 10 (regarding the Master). On the other hand, it seems that Sima Qian’s Lunyu differed from today’s Lunyu in some ways. First, it probably lacked some whole chapters (like chap. 16)42 and perhaps some sequences of sayings (such as the filial piety materials and independent disciple sayings discussed above). Sima Qian’s Lunyu seems to have had shorter sayings than the received Lunyu. This suggests that Sima Qian’s source material was more likely a textual ancestor to, rather than descendant of, the received Lunyu. A final point on this issue: Sima Qian talks about at least some texts in a way that makes it clear they are widely known in his day. Describing the Guanzi 管子 and the Yanzi chunqiu 晏子春秋, he says, “When it comes to their writings, many of our generation possess them. Thus, I do not discuss them but discuss instead anecdotes about them” (至其書,世多有之,是以不論,論其軼事).43 There is no such statement about his Lunyu. One could even read the “Disciples” chapter evaluation as saying that scholars of Sima Qian’s day lacked the moderating influence of the Lunyu and tended to overpraise or overcriticize certain disciples as a result. Finally, various textual oddities in the “Kongzi shijia” could be explained by the hypothesis that Sima Qian was laboring to preserve a messy text that he found valuable and potentially in danger of being discarded.

Kongzi’s Multiple Roles in the Shiji

This study has so far focused mostly on textual issues: what Sima Qian’s Lunyu might have looked like and how he used it. One could also broaden the scope and talk about Kongzi’s roles within the rest of the Shiji. The figure of Kongzi is ubiquitous in the Shiji, not only inhabiting his own chapter (47) and that of his disciples (67) but also being mentioned in more than fifty other chapters. One interesting general feature of the Shiji’s use of Kongzi in those fifty chapters, which Michael Hunter has pointed out, is the fact that the Lunyu Kongzi 42 43

See Goldin’s chapter (chap. 4) in this volume. Shiji 62.2136.

258

Klein

appears only in what we might call “authored material” and not in “compiled material.”44 That is, Sima Qian includes Lunyu materials but his sources do not. As various scholars have discussed, Sima Qian’s Kongzi is first and foremost the editor of the classics.45 A point worth noting is that Kongzi’s association with different classics leads him to play different roles within the Shiji. Some of these are more extensively reflected in the received Lunyu than others. There are many potential ways of carving up this subject, but two interesting aspects of it involve the use of the Lunyu to build up two different but perhaps overlapping roles for Kongzi within the Shiji: the Spring and Autumn Kongzi and the esoteric Kongzi.46

The Kongzi of Historical Texts

One aspect of the Shiji Kongzi that seems distinct from what we find in the Lunyu is what I call the Spring and Autumn Kongzi: Kongzi in relation to the political events of his own time, as a historical figure and as a commentator. As 44

45 46

See Hunter’s chapter (chap. 3) in this volume. There is one exception, which Hunter (2017: 168–170) also acknowledges, at Shiji 79.2422. It is found in a speech by Cai Ze 蔡澤, a minister in the state of Qin. The relevant portion of the speech is: “To advance and retreat, grow full and shrink, this is the constant Way of the sage. Thus [it is said], ‘When the state has the Way, then one serves in office; when the state is without the Way, one goes into reclusion.’ The sage says, ‘The flying dragon is in the heavens: it is beneficial to see the great man’; ‘wealth and honor [gained] without rightness are to me like floating clouds’” (進退盈縮,與時變化,聖人之常道也。故國有道則仕,國無道則隱。聖人曰 飛龍在天,利見大人。不義而富且貴,於我如浮雲). The first quotation is a slight parallel with Lunyu 8/13: “When the realm has the Way, then one is seen. When it is without the Way, then one goes into reclusion” (天下有道則見,無道則隱). This is clearly not attributed to Kongzi and seems to have been a common saying, though it is interesting that it does appear in the Lunyu. The second quotation is now found in the first hexagram of the Zhouyi 周易 (Qian 乾, “Heaven”) under nine in the fifth place. A long discussion of this line, attributed to Kongzi, is found in the “Wen Yan” 文言 commentary (Commentary on the Words of the Text). (For a translation, see Lynn 1994: 137.) The third quotation is an exact parallel to the latter part of Lunyu 7/16. Although one could doubt that “the sage” refers to Kongzi, the coincidence of the Zhouyi and Lunyu citations makes it seem very likely. That there is only this one exception does not seriously undermine the point, however. Clearly, the Lunyu was compiled from at least some earlier materials, and this saying could well have been one of them. Alternatively, Sima Qian or someone else could easily have rewritten the speech and added the quotation from the Lunyu; it would not be foreign to editorial practices at the time. See Nylan 2001: 16–23, 253–306. Csikszentmihalyi (2002) addresses similar aspects of the Han Confucius. The current study focuses more narrowly on Kongzi and the Lunyu in the Shiji.

Sima Qian’s Kongzi And The Western Han Lunyu

259

a historical figure, Kongzi has a slightly better career in the Shiji than he seems to have in the Lunyu, although still not exactly a political success. From the Chunqiu 春秋 (Spring and Autumn Annals) and its three commentarial traditions, Sima Qian takes Kongzi’s triumph through ritual at the Jiagu 夾谷 meeting,47 his daring, though unsuccessful, campaign against the three great families of Lu,48 and Duke Ai’s flattering, if rejected, elegy.49 Sima Qian operated within the constraints of Kongzi’s biography as he knew it from his sources. He could emphasize traditions in which Kongzi was efficacious and spin the story so that Kongzi’s failures were due to those who envied or feared him. But he could not make Kongzi—as an active participant in history—a true success. Contrasting him with Zou Yan 鄒衍, who was fawned on by the feudal lords, he writes that Kongzi (as well as Bo Yi 伯夷 and Mengzi) “had no thought of merely pandering to the customs of the time just to get by. But if you grasp a square tenon and want to insert it into a round hole, can it go in?” (此豈有意阿世俗苟合而已哉!持方枘欲內圜鑿,其能入乎 ).50 This again highlights the unresolved tension in the Shiji between holding to what is moral and having a successful life. In Kongzi’s case, Sima Qian attempts to resolve the tension by showing his posthumous success. One of the many ways in which Sima Qian achieves this is by weaving Kongzi’s life story into the very structure of the Shiji. Sima Qian inserts the events of Kongzi’s life into the annalistic frameworks of nearly all the chapters that overlap with his lifetime. These insertions deliberately amplify Kongzi’s importance in his own time. The events include his birth, death, comings and goings, and supposed service as minister of Lu. Kongzi may have been a political loser, but he is pictured in the Shiji as an astute historical commentator and judge of character. This is one interesting place where the Lunyu Kongzi reenters the picture. Sima Qian quotes Kongzi’s comments on events of his own time, borrowing his authority (via words from the Lunyu) to discuss the subjects of Shiji biographies. He does this using the 47

48

49 50

Shiji 47.1915–1916. Compare the entries for Ding 定 10 in the Zuozhuan (Yang 1993: 1577– 1579) and Guliang zhuan (Sibu beiyao ed., 11.7a–8a). In this incident, Kongzi is depicted as scoring a diplomatic victory over the aggressively posturing delegation from Qi merely by shaming them for ritually inappropriate behavior. Shiji 47.1916–1917. Compare the entries for Ding 定 12 in the Zuozhuan (Yang Bojun 1993: 1586–1587) and Gongyang zhuan (Chunqiu Gongyang zhuan zhushu 578–579). Concerned that the three families were becoming too powerful, Kongzi was said to have forced the issue by urging the ruler of Lu to move against them. Shiji 47.1945. Compare Zuozhuan Ai 哀 16 (Yang Bojun 1993: 1698–1699). Zigong responds to Duke Ai’s elegy for Kongzi by suggesting that it was inappropriate to speak so highly of Kongzi when during his lifetime Duke Ai was not willing to employ him. Shiji 74.2345.

260

Klein

formal conventions developed in historical texts. For example, the evaluation at the end of Shiji chapter 103 says, “Zhongni has a saying, ‘The princely man wishes to be slow in his speech and earnest in his conduct.’ Could not this refer to Wan Shi, Jian Ling, and Zhang Shu?” (仲尼有言曰:君子欲訥於言而敏於 行,其萬石﹑建陵﹑張叔之謂邪).51 There are various examples of this type in the Shiji: Lunyu sayings are matched to figures who postdate Kongzi. This is an extension of Kongzi’s role as a historical commentator. An Esoteric Kongzi Another aspect of Kongzi that is reflected in both the Shiji and the Lunyu is his circumspection. Both passages about what the Master did not talk of are found in Sima Qian’s Lunyu,52 with the second of them being quoted or alluded to in four different chapters.53 Passages about being cautious in speech and writing are also shared between the two texts. The above-quoted evaluation at the end of the Shiji “Disciples” chapter concludes with the words “what was doubtful I left blank” (疑者闕焉), which may be an allusion to the last part of Lunyu 2/18: “Listen much and leave blank what is doubtful; cautiously speak of what remains” (多聞闕疑,慎言其餘). In Lunyu 13/3, Kongzi states, “The princely man leaves blank that which he does not know” (君子於其所不知,蓋闕如也). Although Sima Qian does not quote that specific saying or the passage about archivists leaving blanks (Lunyu 15/26), he does write elsewhere about how Kongzi did not attempt to put dates into the Shangshu 尚書 (Exalted Documents): “Some places have [dates] to an extent, but there are many blanks for what could not be recorded. Thus, where there is doubt he transmitted the doubt—such was his caution” (或頗有,然多闕,不可錄。故疑則傳疑,蓋 其慎也).54 This does suggest that the practice of leaving blanks is associated with Kongzi’s editorial practices, at least for Sima Qian. Yet this is another tension within the Shiji, for Sima Qian seems eager to talk about exactly those things that Kongzi was too cautious to discuss. There is even some implication that Kongzi refused to discuss such things, not because they were fundamentally dubious, but because they were intended for only a small, select audience. This aspect of the Shiji Kongzi, an “esoteric” Kongzi so to speak, surfaces in several ways. One, curiously, has to do with the muchmaligned disciple Zai Yu. Zai Yu is portrayed very badly in the received Lunyu. In Lunyu 6/26 he seems to sneer at benevolent people, implying that they are 51 52 53 54

Shiji 103.2773; Lunyu 4/24. I.e., Lunyu 7/21 and 9/1. Shiji 27.1343, 47.1938, 49.1967, 74.2343. See discussion below. Shiji 13.487.

Sima Qian’s Kongzi And The Western Han Lunyu

261

gullible. (He is corrected by the Master.) He is listed, along with Zigong, as being notable for his ability as a talker (Lunyu 11/3), though elsewhere the Lunyu is suspicious of good talkers.55 In Lunyu 17/21, Zai Yu incurs Kongzi’s scorn for questioning the necessity of three years’ mourning. The Master tells him that if he can enjoy pleasures after only one year he should do so, but afterward pronounces that he is truly not benevolent. Most damning is Lunyu 5/10, where the Master excoriates him for sleeping in the daytime, even going so far as to imply that Zai Yu is responsible for the Master’s loss of faith in using people’s words as an indication of their character. The Shiji portrayal of Zai Yu is not especially favorable either, with several of the more unflattering Lunyu passages quoted in his biography.56 Interestingly, however, he also appears in the very first chapter of the Shiji: “What was transmitted by Kongzi when Zai Yu asked about the Virtues of the Five Thearchs and the Genealogies of the Thearchs, some Ru do not transmit” (孔子所傳宰予問 五帝德及帝繫姓,儒者或不傳).57 Sima Qian explicitly refers to this tradition as being limited to only some Ru. One might initially read this as a mark of its dubiousness. Yet another way to see it is as being an esoteric tradition, not for everyone. The “Virtues of the Five Thearchs” and the “Genealogies of the Thearchs” now appear as chapters in the Da Dai Liji 大戴禮記 (Elder Dai’s Records of Ritual). The “Virtues of the Five Thearchs” is indeed a dialogue between Kongzi and Zai Yu, in which Kongzi tells Zai Yu about the Five Thearchs but at the end also tells him that he is the wrong person to make inquiries about it. Afterward, Zai Yu speaks of it to someone else and Kongzi disowns him.58 The last part of this dialogue also appears in Sima Qian’s biography of Zai Yu.59 Zai Yu was the wrong person to receive the transmission, but Kongzi told him certain things anyway. It was when he revealed it to someone else that it became a real problem. This “esoteric” Kongzi is silent not out of ignorance but out of concern that his knowledge not be transmitted to the wrong people. Another example of Kongzi’s association with esoteric matters involves the feng 封 and shan 禪 sacrifices, ignorance of which brought the Ru great disgrace in Sima Qian’s time. The Shiji treatise describes some “abbreviated words” (lueyan 略言) passed down by Kongzi about these sacrifices and then adds:

55 56 57 58 59

E.g., when the Master says at Lunyu 1/3, “Clever words and a pleasing appearance: rarely indeed is [such a person] benevolent” (子曰:巧言令色,鮮矣仁!). Shiji 67.2194–2195. Shiji 1.46. Wang 1983: 125. Shiji 67.2195.

262

Klein

But he did not make clear the specific details of the rite—probably it was difficult to discuss. Someone asked for an explanation of the great sacrifice. Kongzi said, “I do not know. One who knows the explanation of the great sacrifice would be in relation to the realm as one who looks down at the palm of his hand.” 其俎豆之禮不章,蓋難言之。或問禘之說,孔子曰:不知。知禘之 說,其於天下也視其掌。60

An interesting aspect of this passage is the way in which the Lunyu parallel at the end fits in. It stands as a summation and distillation of the larger point being made: namely, that the details of this most powerful sacrifice seem to have been unclear even to Kongzi, though he does know exactly how powerful it could potentially be. Curiously, another quotation associated with Zai Yu comes up in the same chapter and is connected with the same issue: A tradition says, “If for three years a rite is not done, then the rite will certainly be discarded. If for three years a piece of music is not played, the music will certainly be ruined.” If an age was at its height, the feng and shan responded to it. But when there was decline, [the feng and shan] were in abeyance. The distance between them can be over a thousand years. Since the most recent [performance] it has been several hundred years. Thus, the rite is incomplete, even irrecoverable. Its details cannot be obtained, written down, heard, or spoken. 傳曰:三年不為禮,禮必廢;三年不為樂,樂必壞。每世之隆,則封 禪答焉,及衰而息。厥曠遠者千有餘載,近者數百載,故其儀闕然堙 滅,其詳不可得而記聞云。61

The connection here with Zai Yu is indirect—in Lunyu 17/21, he quotes the saying about the three years in such a way that it sounds like a proverb. So perhaps here it is a proverb not meant to be explicitly associated with Zai Yu. Still, Zai Yu’s association with the transmission of secret or possibly lost ancient knowledge makes this coincidence seem potentially interesting. Besides the ancient sacrifices, another subject that the Lunyu Kongzi does not talk much about is heaven or fate. Omens and anomalies were perhaps a 60 61

Shiji 28.1363–1364. The quotation parallels Lunyu 3/6. Shiji 28.1355. The “tradition” parallels Lunyu 17/21.

Sima Qian’s Kongzi And The Western Han Lunyu

263

sensitive subject in Sima Qian’s time, given the number of fake omens detailed in the “Fengshan shu” 封禪書 (Treatise on the Feng and Shan Rites) chapter of the Shiji. In the “Tianguan shu” 天官書 (Treatise on the Office of Astronomy), Sima Qian delivers a much longer version of the famous “the Master seldom spoke of” or “did not speak of” passages.62 He writes: The time before You and Li is high antiquity indeed. The changes in heaven that were observed were all in different regions of the country. The [different] experts prognosticated aberrations in order to meet the needs of the time, and their writings, charts, records, and prayers for good fortune did not follow any model. This is why when Kongzi discussed the Six Classics, he chronicled anomalies but wrote no explanations. When it came to the command of the heavenly way [i.e., fate], he did not transmit [anything about it]. Should one transmit it to people [of the right sort], they would have no need to be told [because they understand it already]. To people of the wrong sort, even if one did talk about it, it could not be made clear. 幽厲以往,尚矣。所見天變,皆國殊窟穴,家占物怪,以合時應,其 文圖籍禨祥不法。是以孔子論六經,紀異而說不書。至天道命,不 傳;傳其人,不待告;告非其人,雖言不著。63

A similar passage is found in chapter 49, the “Waiqi shijia” 外戚世家 (Hereditary House of Imperial Relatives by Marriage). This chapter is much concerned with the strange twists of fate that lead to women becoming empresses. The chapter begins with a meditation on how imperial marriages are of the highest importance, and yet the factors that lead one woman (and her family) to succeed while another woman fails seem unfathomable and dependent on some inextricable tangle of inborn nature and mysterious fate. After musing on this issue, the prefatory comment says: How can one go against fate!? Kongzi rarely spoke of fate, probably because it is problematic to talk about. If one is not able to master the transformations of the hidden and the illuminated, how could one be able to know about inborn nature and fate?

62 63

Lunyu 9/1, 7/21. Shiji 27.1343.

264

Klein 豈非命也哉?孔子罕稱命,蓋難言之也。非通幽明之變,惡能識乎性 命哉?64

The implication is that Kongzi does understand these “transformations of the hidden and the illuminated.” His listeners, however, would not understand, so there was no point in his talking much about it. The mention of “mastering the transformations of the hidden and the illuminated” recalls another, perhaps-related aspect of Kongzi, namely, the Kongzi who studied the Yi 易 (Changes). Three times in the Shiji, Kongzi’s knowledge of the Changes is emphasized. First, in the “Tian Jingzhong Wan shijia” 田敬仲 完世家 (Hereditary House of Tian Jingzhong Wan), the evaluation mentions Kongzi’s work on the Changes to add authority and gravity to the Changes tradition: It seems that Kongzi in his later years delighted in the Changes. The technique of the Changes is obscure, illuminating, and far-reaching indeed. If one is not a person of profound understanding and penetrating talent, how could one truly attend to it? Thus, when the senior archivist of Zhou divined by hexagrams about Tian [Jingzhong] Wan, his prognostication reached down to the tenth generation afterward. 蓋孔子晚而喜易。易之為術,幽明遠矣,非通人達才孰能注意焉!故 周太史之卦田敬仲完,占至十世之後。65

The kind of divination being done by the senior archivist of Zhou here is similar to that found in the Zuozhuan, featuring long-range political predictions. The more direct reference to Kongzi’s study of the Changes is of course found in Kongzi’s own biography. Interestingly, it begins with exactly the same words: Kongzi in his later years delighted in the Changes. He put in order the Tuan, the Attached [Verbalizations], the Images, the “Discussion of the Hexagrams,” and the “Commentary on the Words of the Text.” In reading the Changes, he three times wore out the leather bindings of his text. He

64 65

Shiji 49.1967. Shiji 46.1903. Note that the mention of ten generations also echoes the saying attributed to Kongzi that it is possible to know how things will be ten generations hence (Lunyu 2/23).

Sima Qian’s Kongzi And The Western Han Lunyu

265

said, “Suppose I could be given a few more years. If so, then I would be evenly balanced with regard to the Changes.” 孔子晚而喜易,序彖、繫、象、說卦、文言。讀易,韋編三絕。曰: 假我數年,若是,我於易則彬彬矣。66

This last saying bears a certain resemblance to Lunyu 7/17, the only passage in the Lunyu that also supports this Changes Kongzi. It is a very loose parallel but recognizable: The Master said, “If a few years could be added to my life, I would take fifty to study the Changes. I could thereby be without major errors.” 子曰:加我數年,五十以學易,可以無大過矣。

In both cases, full understanding of the Changes appears to require long study—it is not a type of insight that is available to just anyone. The final reference to Kongzi’s mastery of the Changes occurs in the “Disciples” chapter, where we find a complete line of transmission for the Changes, stretching from Kongzi to Sima Qian’s own generation. Significantly, the penultimate name is Yang He 楊何 (fl. 134 BCE). Shiji 130 mentions that Sima Tan studied the Changes from none other than this master. Here, as with other examples of the things Kongzi is reluctant to speak of, we find something hinted at in the Lunyu that is more fully elaborated in the Shiji.

Conclusion

As mentioned above, Sima Qian’s Kongzi is primarily associated with the classics. It appears that something like a Lunyu was available to Sima Qian, but as a source for the portrait of Kongzi himself it was of secondary importance. Lunyu quotations elsewhere in the Shiji are used just as one might expect: they are picked out and put into new contexts to lend authority. Sayings from other sources are used this way as well. Sima Qian tends to use Lunyu sayings to reflect Han concerns, especially the ones he particularly cared about: How should one judge and write history? What happened in the ancient past? And 66

Shiji 47.1937.

266

Klein

how should one perform powerful rituals? Sima Qian places almost no emphasis on the virtue of filial piety in association with Kongzi. The development of Confucian texts and traditions in early China was complicated, far more complicated than myths of a single core canonical Lunyu tradition can capture. In trying to get a sense of it, we are fortunate to have the Shiji, which just barely precedes the late Western Han dominance of the Lunyu Kongzi. Sima Qian seems to have had some of the same materials as his successors would have, but he did not hold quite the same reverent attitudes toward those materials. Trying to get a sense of Sima Qian’s Lunyu and of his Kongzi begins to reveal just how complicated the whole picture must have been in the early Western Han and earlier.

Bibliography

Chen Ning. 2003. “Mohist, Daoist, and Confucian Explanations of Confucius’s Suffering in Chen-Cai.” Monumenta Serica 51:37–54. Chin, Annping. 2008. Confucius: A Life of Thought and Politics. New Haven, CT: Yale University Press. Chunqiu Gongyang zhuan zhushu 春秋公羊傳注疏. 2000. Beijing: Beijing Da xue chu­ banshe. Csikszentmihalyi, Mark. 2002. “Confucius and the Analects in the Hàn.” In Confucius and the Analects: New Essays, ed. Bryan W. Van Norden, 134–162. Oxford: Oxford University Press. Csikszentmihalyi, Mark. 2004. Material Virtue: Ethics and the Body in Early China. Leiden: Brill. Durrant, Stephen W. 1995. The Cloudy Mirror: Tension and Conflict in the Writings of Sima Qian. Albany: State University of New York Press. Han Zhaoqi 韓兆琦. 2004. Shiji qianzheng 史記箋證. Nanchang: Jiangxi renmin chubanshe. Hardy, Grant. 1999. Worlds of Bronze and Bamboo. New York: Columbia University Press. Hunter, Michael. 2017. Confucius Beyond the Analects. Leiden: Brill. Jin Dejian 金德建. 1963. Sima Qian suo jian shu kao 司馬遷所見書考. Shanghai: Renmin chubanshe. Kaneto Mamoru 金戸守. 1973. “Shiki Rongo kō: Kōshi seika, sono ‘Rongo’ seiritsu shiron” 史記「論語」考: 世家、その「論語」成立試論. Shitennōji joshi daigaku kiyō 四天王寺女子大学紀要 6:14–38.

Sima Qian’s Kongzi And The Western Han Lunyu

267

Kaneto Mamoru 金戸守. 1978. “Shiki ‘Rongo’ kō, so no roku: Soketsu, Rongo no seiritsu ni tsuite” 史記「論語」考、その六: 総結、論語の成立について. Shitennōji joshi daigaku kiyō 四天王寺女子大学紀要 11:19–26. Klein, Esther Sunkyung. 2010. “The History of a Historian.” PhD diss., Princeton University. Knapp, Keith. 1995. “The Ru Reinterpretation of Xiao.” Early China 20:195–222. Kongzi jiayu zhuzi suoyin 孔子家語逐字索引. 1992. Chinese University of Hong Kong, Institute of Chinese Studies Ancient Chinese Concordance Series. Hong Kong: Commercial Press. Kramers, R. P. 1993. “K’ung tzu chia yu.” In Early Chinese Texts: A Bibliographical Guide, ed. Michael Loewe, 258–262. Berkeley: Society for the Study of Early China and Institute of East Asian Studies, University of California. Lynn, Richard. 1994. The Classic of Changes. New York: Columbia University Press. Maeder, Erik W. 1992. “Some Observations on the Composition of the ‘Core Chapters’ of the Mozi.” Early China 17:27–82. Makeham, John. 1996. “The Formation of Lunyu as a Book.” Monumenta Serica 44:1–24. Makeham, John. 1998. “Between Chen and Cai: Zhuangzi and the Analects.” In Wandering at Ease in the Zhuangzi, ed. Roger T. Ames, 75–100. New York: State University of New York Press. Nylan, Michael. 2001. The Five “Confucian” Classics. New Haven, CT: Yale University Press. Nylan, Michael, and Thomas Wilson. 2010. Lives of Confucius: Civilization’s Greatest Sage through the Ages. New York: Doubleday. Qian Mu 錢穆. 2002. Kongzi zhuan 孔子傳. Repr., Beijing: Sanlian. Shiji 史記. 1959. Beijing: Zhonghua shuju. Wang Pinzhen 王聘珍. 1983. Da Dai Liji jiegu 大戴禮解詁. Beijing: Zhonghua shuju. Watson, Burton. 1958. Ssu-ma Ch’ien, Grand Historian of China. New York: Columbia University Press. Yang Bojun 楊伯峻. 1993. Chunqiu Zuozhuan zhu 春秋左傳注. Rev. ed. Beijing: Zhonghua shuju.

268

Kern

Chapter 10

Kongzi as Author in the Han Martin Kern Thinking about the possible compilation date of the Lunyu 論語, one cannot but be struck by its unusual format and title. Every other pre-Qin text that purports to express the thought or reveal the persona of a particular thinker has had that person’s name (or some identifying phrase) in its title at least since the Han; the Lunyu never does. Whoever titled the text was not concerned with a reader who needed guidance as to whom or what the book might be about;1 instead, he presupposed that anyone encountering a text titled Lunyu would know its affiliation with Kongzi. The text that lacks Kongzi’s name in the title was taken for granted to stage, first and foremost, Kongzi the person. If the early interpretation of Lunyu as “Selected (or “Ordered”)2 Sayings” is to be trusted, the title was understood as some sort of digest, offering the distilled essence of a much larger textual repertoire together with an idealized account of Kongzi as a person. In other words, a title such as Lunyu makes sense only (a) esoterically or (b) as a designation given at a late stage when both Kongzi the sage person and the idea that he could be represented by sayings and anecdotes had already become widely recognized. It does not make sense to use the impersonal title Lunyu in order to introduce a corpus of pithy sayings, short dialogues, and barely contextualized anecdotes related to an otherwise unfamiliar master. Instead, the title “Selected/Ordered Sayings” signals to its potential readers that the text is the thoughtfully arranged compilation of materials related to a familiar persona or discourse. Unlike the “Masters” (zhuzi 諸子) texts of early China, the title is not eponymous with (or at least contains the name of) its purported author or protagonist but, precisely by withholding such information, stages the text’s authority and, by extension, that of its protagonist. In this, the text called Lunyu is also a meta-text that refers back to the nature of its own textuality as well as to the community of its implied (i.e., well-informed) audience. It signals that it has little need to join the argumenta1 Note, however, Huang Ren’er’s argument that the Lunyu was known by various titles, as noted in John Makeham’s chapter (chap. 1) in the present volume. Cheng (1993: 315) mentions that the text was originally known as Kongzi, “in the same way as writings of other masters of the Warring States period.” Finally, note also Michael Hunter’s tentative suggestion (in chap. 3 of the present volume) that the title Lunyu was related to the “selection” of official candidates. 2 For a discussion of lun as “ordered,” see Graham 1978: 194.

© Koninklijke Brill NV, Leiden, 2018 | doi:10.1163/9789004382947_012

Kongzi As Author In The Han

269

tive fray. And that is indeed the case: the Lunyu does not argue or explain why Kongzi is right on this or that; it does not even let Kongzi argue for himself. Instead, it shows him as being Kongzi. I find it very difficult to imagine that such a text existed very early or grew somehow organically over time. Michael Hunter’s massive evidence of citation data has further convinced me to reject the “accretion model” claimed by Brooks and Brooks3 nearly thirty years after it had been forcefully advanced by Kimura Eiichi 木村英一.4 At the same time, considering the meta-textual nature of the Lunyu, I am curious as to its representation of Kongzi himself as a master of texts. More specifically, I am interested in his image as the “author” of the Chunqiu 春秋 (Springs and Autumns [Annals])—or at least as the person responsible for the editing of the text5—that since antiquity has been so firmly inscribed in the common imagination of Kongzi. Can this representation help us contextualize the Lunyu in a particular time and intellectual environment? Traditional (including recent) scholarship suggests that Kongzi’s association with the “making” of the Chunqiu was of paramount significance for how Kongzi was imagined during the Western Han. If this is true, the thesis for a compilation date of the Lunyu around or after ca. 140 BCE faces an enormous conundrum: if Han political and philosophical thinkers imagined Kongzi first and foremeost as the author of the Chunqiu, how could they possibly avoid any mention of the latter in the Lunyu? If anything, a Han dynasty compiler of the Lunyu would have been highly interested in the idea of Kongzi as the author of the Chunqiu; or at the very least, he would not have chosen to avoid its mention entirely. This question alone suffices to throw doubt on the idea of a Han dynasty compilation date for the Lunyu.6 Therefore, the present essay explores the subject of “Kongzi made the Springs and Autumns” from a Han perspective. 3 Brooks and Brooks 1998. 4 Kimura 1971. For a critique of Brooks and Brooks 1998, see Schaberg 2001b. As noted by Schaberg (133), “To be shown what Brooks and Brooks have promised to show, one must accept the following [six] premises, all of them faulty.” 5 To call Kongzi the “author” of the Chunqiu is misguided. Since antiquity, different readers have imagined Kongzi’s engagement with the text in different ways, but nobody seems to have claimed that he composed it out of nothing. The question is, rather, to what extent Kongzi’s rephrasing with “subtle words” (wei yan 微言) may have changed the original annalistic records he had inherited from the scribes of the state of Lu 魯. In the present essay, I use “author” not in the modern sense of original creator, but, instead, in the Latin meaning of the Latin verb augere (to “augment,” to “increase” something that is already in existence) that is the origin of the modern word “author.” More on this question below. 6 In other words, the issue would not just require us to admit that the Lunyu contains pre-Han textual material; it would question how the text, even if including such material, could have been compiled in the Western Han.

270

Kern

My question is not whether or not there was a Lunyu text in pre-Han times. Instead I ask, is it plausible, or even possible, that the text was compiled during the Han? Can we falsify the hypothesis of a compilation date post-140 BCE? In what follows, I do not aim, or claim, to “prove” the matter one way or the other; the available data do not allow for that.7 Somewhat to my own surprise, I show that the notion of “Kongzi made the Springs and Autumns” does not suffice to reject the Han compilation hypothesis. What I have found, and will present below, is that the evidence is at best tenuous, and that scholarly consensus on the Han nexus between Kongzi and the Chunqiu does not hold, at least not for the time before Sima Qian 司馬遷 (ca. 145–ca. 85 BCE). As it turns out, the idea of Kongzi as the person responsible for the Chunqiu can and must be situated more specifically in a particular, and very limited, set of sources.8



Before I begin to examine Warring States sources with regard to Kongzi’s involvement with texts, a caveat lector: the term “Warring States sources” itself is extremely problematic. When looking at a text that is traditionally dated to pre-imperial times, it is impossible to separate its original core from the shape and organization it was given by its Han editors. To some extent, all received pre-imperial texts are Han texts.9 This is not to say they were newly created (let alone “forged”) in the Han. But it is to insist that the Han “editing”—which is closely associated with the Han imperial bibliographer Liu Xiang 劉向 (79–8 BCE) and his son Liu Xin 劉歆 (46 BCE–23 CE)—must in each case be understood as a reconstitution of the text in a new form. In fact, according to the little we know about Liu Xiang’s work in response to an imperial edict of 26 BCE, he was tasked to collate and rearrange the writings within the imperial library and to create new physical versions of the ancient texts, written in Han clerical script on bamboo slips of standardized length, to be stored in the imperial palace. This monumental effort entailed countless decisions on archaic, regional, or otherwise obscure graphs, on the assembly of texts of great variety under the headings of particular titles, and on placing texts into a new bibliographic format that constituted nothing less than a late Western Han intellectual history of the entire textual heritage, as far as it was available. Many of 7 To some extent, this echoes the results of Wolfgang Behr’s linguistic analysis of the Lunyu. As Behr (2011) has demonstrated, the available linguistic data do not allow any conclusion on the dating of the Lunyu. 8 For a fine study that reaches a similar conclusion but altogether has a focus that differs from that of the present essay, see Cai 2010. 9 See the seminal study by Van der Loon 1952.

Kongzi As Author In The Han

271

these decisions intervened in the texts in ways we would normally reserve for an author, including the arrangement of textual material into chapters and possibly “books” that in most cases had never existed in this form before. Thus, our modern and culture-specific conceptualizations of, and distinctions among, “author,” “compiler,” “editor,” and even “commentator” are quite anachronistic when applied to Chinese antiquity. With this in mind, I am using the term “Warring States sources” in the most charitable way, as if the texts traditionally attributed to that period (such as the Mengzi 孟子, quoted below) could be accepted just as such. This is a useful but purely heuristic maneuver, for I do still believe that these sources express some ideas operative in a pre-imperial intellectual milieu, which set them apart from truly imperial texts—that is, texts first created under the political, social, intellectual, and administrative conditions of the Qin-Han empire. That said, if the preponderance of evidence speaks forcefully against accepting a particular passage from such a “Warring States text” as indeed predating the empire, one is obviously still obliged to take note of this fact. Even if we accept the Mengzi in general as dating from the late fourth or early third century BCE, we cannot presume that everything in it does so as well. Furthermore, when reading the pertinent passages in the Mengzi and in the Gongyang zhuan 公羊傳, it is necessary to be aware of how much the established understanding of these texts is guided by subsequent elaborations. (The same logic extends to the reading of the Lunyu: we must be open to the possibility of later interference with an earlier text.) It is therefore useful to recall briefly the traditional view and its tenuous relationship with the actual textual evidence. In his influential book Writing and Authority in Early China, Mark Edward Lewis states the following: The Gongyang zhuan, which treats Confucius as the author of the Chun qiu, traced its teacher-disciple transmission back to the Xunzi. Finally, several late Warring States texts in other traditions, including the Han Feizi and the Zhuangzi, refer to Confucius’ composition of the Chun qiu. Since Han Fei was also a student of Xun Kuang’s it is likely that the latter espoused the theory of Confucius’ authorship of the Chun qiu. At any rate, the idea was widely held by the late Warring States period.10 These assertions regarding the idea of Kongzi’s authorship of the Chunqiu are dubious. First, there is but a single explicit reference in the Gongyang zhuan where Kongzi is quoted as taking responsibility for the phrasing—though 10

Lewis 1999: 234.

272

Kern

explicitly not for the content—of the Chunqiu. In addition, the epilogue to the Gongyang zhuan, of unknown origin, refers to a “noble man” (junzi 君子) as the person who “made” (wei 為, though not the stronger zuo 作) the Chunqiu.11 Second, neither the Han Feizi 韓非子 nor the Zhuangzi 莊子 makes any such claim.12 Third, as Lewis himself notes elsewhere, there is no mention of Kongzi’s authorship of the Chunqiu in the Xunzi 荀子, despite the insinuation that the purported author of the Xunzi somehow “espoused the theory of Confucius’ authorship.”13 Fourth, nothing suggests that “the idea was widely held by the late Warring States period.” To the contrary, very few people appear to have been aware of it. It seems to me that Lewis’s sweeping claims, and the underlying misreading of the textual evidence involved, can come from only one source: the powerful tradition that included both the hagiography of Kongzi and the anachronistic projection of later textual practices and properties into pre-imperial times.14 When thinking about Kongzi as author, it remains important not to lose sight of this tradition and the extent to which it still holds sway in most quarters of contemporary scholarship. The picture I draw throughout the following pages departs decisively from such views.



Across all Warring States sources, Kongzi as author appears only with respect to a single text, the Chunqiu, and very rarely so. The first source, possibly, is Mengzi 3B/9:15 When the world declined and the Way fell into obscurity, heresies and violence arose. There were instances of regicides and patricides. Kongzi 11 12

13 14 15

See below for both passages. I thank Christoph Harbsmeier, Jens Østergaard Petersen, Paul R. Goldin, and Joachim Gentz for discussing these passages in detail with me. The three passages Lewis (1999: 454n187) cites to support this claim are a Han Feizi passage, discussed below, that he misreads (and which, to the contrary, seems to indicate that, here, Kongzi is exactly not seen as the author); a Han Feizi passage (Wang Xianshen 1998: 34.314) that has Zixia 子夏 commenting on the Chunqiu but has nothing to say about Kongzi’s authorship either; and a fragment of dubious origin that only in the seventh-century anthology Yiwen leiju 藝文類聚 is attributed to the Zhuangzi 莊子. One may also note that the traditional commonplace that Han Fei 韓非 was “a student” of Xun Kuang 荀況 (i.e., Xunzi) has been forcefully challenged by Sato 2013. For two critiques of Lewis’s (over)emphasis on the status of writing in early China, see Nylan 2000; Kern 2000. Note, however, that Van Ess (2015), in a detailed and thoughtful discussion, suggests that Mengzi 3B/9 is a later interpolation that follows the account of Kongzi in the Shiji.

Kongzi As Author In The Han

273

was apprehensive and made the Springs and Autumns. The Springs and Autumns is the business of the Son of Heaven. Thus, Kongzi said, “Those who recognize me will do so for the Springs and Autumns; those who condemn me will do so for the Springs and Autumns.” 世衰道微,邪說暴行有作。臣弒其君者有之,子弒其父者有之。孔子 懼,作《春秋》。《春秋》、天子之事也;是故孔子曰:知我者其惟 《春秋》乎!罪我者其惟《春秋》乎!16

A few lines later, the Mengzi concludes: After Kongzi had completed the Springs and Autumns, rebellious ministers and murderous sons lived in fear. 孔子成《春秋》而亂臣賊子懼。17

Note what the Mengzi does not say: it does not relate Kongzi’s authorship to any specific event, nor does it integrate the composition of the Chunqiu with the Master’s biography. By contrast, in Kongzi’s biography in the Shiji 史記 (“Kongzi shijia” 孔子世家), the Mengzi passages are paralleled in reverse order and strikingly expanded, including with a parallel from Lunyu 15/20: The Master said, “Alas, alas! The noble man resents leaving the world without having his name recognized. My Way is not put into practice, so what can I use to show myself to later generations?” Thus, relying on archival records, he made the Springs and Autumns. ... [His] principles of criticizing and diminishing [the rulers of the past] were upheld and applied by true kings of later times. When the principles of the Springs and Autumns are put into practice, rebellious ministers and murderous sons from all across the realm will live in fear of them. ... When the disciples received the Springs and Autumns, Kongzi said, “Those who in later generations will recognize me will do so for the Springs and Autumns, and those who will condemn me will also do so for the Springs and Autumns.” 子曰:弗乎弗乎,君子病沒世而名不稱焉。吾道不行矣,吾何以自見 於後世哉?乃因史記作《春秋》 …貶損之義,後有王者舉而開之。 16 17

Jiao 1987: 452. Jiao 1987: 459.

274

Kern 《春秋》之義行,則天下亂臣賊子懼焉…弟子受《春秋》,孔子曰: 後世知丘者以《春秋》,而罪丘者亦以《春秋》。18

Here, the Shiji itself speaks universally of the “true kings of later times” (後有王 者), who include not only Kongzi’s immediate posterity but the rulers of all times, presumably including Sima Qian’s own Emperor Wu 漢武帝 (r. 141–87 BCE), insofar as these were “true kings” (wangzhe 王者). In addition, Kongzi is made not once but twice to voice his concern for readers of “later generations” (houshi 後世), to whom he “shows himself” (zi xian 自見) and who will “recognize” (zhi 知) him because of the Chunqiu. His final exclamation is positioned right before the concluding narrative of his death, marking it as his testament to posterity and sealing a narrative that, altogether, emphasizes his failure in life.19 In this, he inscribes himself into the very history he is chronicling: the text of the Chunqiu is radically reinterpreted and transformed into an act of dramatic self-expression. It marks the end of his life, and it marks the end of the historical period his text has chronicled. The way the Shiji presents Kongzi’s quest for posterity as compensation for this failure has guided the traditional reading of the Mengzi passage ever since. The possible second instance of a text’s mentioning Kongzi as involved with the Chunqiu may be found at the end of the Gongyang zhuan. Here, the Chunqiu text proper closes on a laconic note for the year 481 BCE: In the fourteenth year [of Duke Ai], in the spring, at the hunt in the western regions they caught a unicorn. 十有四年春,西狩獲麟。20

The Gongyang zhuan explains: The unicorn is a beast of benevolence. When there is one who acts as a true king, it arrives; when there is none who acts as king, it does not 18 19

20

Shiji 47.1943–1944. Cf. Lunyu 15/20 (“Wei Ling gong” 衛靈公), Shiji 61.2127 (“Boyi liezhuan” 伯夷列傳), and Mengzi 3B/9. Note that, here, the statement “When the principles of the Springs and Autumns are put into practice, rebellious ministers and murderous sons from all across the realm will live in fear” (《春秋》之義行,則天下亂臣賊子懼焉) is explicitly directed at the time of “future kings” (hou wang 後王). Chunqiu Gongyang zhuan zhushu 709. In the Gongyang zhuan and Guliang zhuan 榖梁傳 versions, the Chunqiu ends with the year 481 BCE, while the Zuozhuan 左傳 version continues the text until 463 BCE.

Kongzi As Author In The Han

275

arrive. … Kongzi said, “For whom did it come! For whom did it come!” As he turned his sleeve and wiped his face, tears soaked his gown. When Yan Yuan died, the Master said, “Ah! Heaven has bereft me!” When Zilu died, the Master said, “Ah! Heaven has cut me off!” When at the hunt in the western regions they caught the unicorn, Kongzi said, “My Way has reached its end.” 麟者,仁獸也。有王者則至,無王者則不至 …孔子曰:孰為來哉!孰 為來哉!反袂拭面,涕沾袍。顏淵死,子曰:噫!天喪予!子路死, 子曰:噫!天祝予!西狩獲麟。孔子曰:吾道窮矣。21

The structure is similar to that of the Mengzi passage quoted above: a brief factual statement followed, in this case, by the series of Kongzi’s emphatic exclamations—and, in addition, his emotional collapse. On both the textual and the meta-textual level, “My Way has reached its end” is the perfect, if somewhat melodramatic, ending of the text and ending of Kongzi’s life. In one crucial respect, however, the passage is ambiguous.22 The reading of wu dao qiong yi 吾道窮矣 (“My way has reached its end”) in the sense of “I am spent; I am desperate” is the one familiar from later tradition. However, in Han times, perhaps first with Sima Qian’s purported “teacher” Dong Zhongshu 董仲舒 (ca. 179–ca. 104 BCE), the Gongyang zhuan entry also became the fountainhead of the theory of Kongzi as the “uncrowned king” (su wang 素王). When the lin appears, Kongzi is first distressed over its arrival (“For whom did it come! For whom did it come!”) because it should not have arrived in the absence of a true king. Yet some Han readers, Dong Zhongshu among them, had an answer to Kongzi’s seeming incredulity: Heaven, in recognizing Kongzi as the “uncrowned king,” had sent the unicorn for him, as an omen of his kingship that overruled the worldly kings, and as the mandate (ming 命) to create the Chunqiu in order to overwrite, and indeed rectify, history. This reading appears at the beginning of the fragmentary chapter 16 (“Fu rui” 符瑞 [Auspicious Signs and Omens]) of the Chunqiu fanlu 春秋繁露 (Luxuriant Dew of the Springs and Autumns):

21 22

Chunqiu Gongyang zhuan zhushu 711–716. I am indebted to Paul R. Goldin for having brought this ambiguity to my attention and for initiating and sustaining a thorough discussion of it that further involved Christoph Harbs­meier and Jens Østergaard Petersen. (All three of these friends also offered detailed bibliographic and editorial help throughout the essay, not to mention their learned corrections.)

276

Kern

Regarding what human effort cannot bring about but is instead brought about just by itself: when at the hunt in the western regions they caught the unicorn, this was an auspicious sign marking the reception of the mandate [of Heaven]. This being the case, thereafter [Kongzi] engaged himself with the Springs and Autumns to rectify what was not right, and to illuminate the principle of change in dynastic stipulations. 有非力之所能致而自致者,西狩獲麟,受命之符是也。然後託乎《春 秋》正不正之間,而明改制之義。23

While the chapter is clearly fragmentary, and while in general it is impossible to authenticate with certainty any particular part of the Chunqiu fanlu as coming directly from Dong Zhongshu or his inner circle,24 the fragments of chapter 16 are a good candidate to belong to the early layers of the text, perhaps dating to the latter part of the second century BCE. They certainly fit a Western Han intellectual context (as does chapter 17, mentioned below) and belong to a line of thought that inspired, for example, Liu Xiang’s explicit statement in Shuiyuan 說苑 (Garden of Persuasions) that the arrival of the unicorn showed “how Heaven recognized the Master” (此天之知夫子也; see below). This reading of the Gongyang zhuan passage raises the question of the meaning of Kongzi’s final words, wu dao qiong yi: are they a sigh of despair or, rather, one of relief? While the former is favored by the tradition, the latter, advocated by Paul R. Goldin,25 may be closer to the understanding of at least some Han exegetes. In that reading, Kongzi does not simply despair at the absence of a true king; he also realizes that Heaven recognizes him as a sage. Furthermore—and this is missing in the Gongyang zhuan passage—Heaven gives him the mandate to compose the Chunqiu. In my view, the two readings can be combined: Kongzi receives the mandate only in the absence of a true worldly king; thus, he at once despairs at the world and is relieved, albeit with a heavy heart, that Heaven has recognized him as a sage. This, in fact, can be found in Sima Qian’s Shiji. The Shiji biography of Kongzi refers twice to Kongzi’s response to the appearance of the unicorn. The biography shortens the Gongyang zhuan account

23 24 25

Su 1992: 157–158; Zhong 2005: 352–354. For the seminal study on the authenticity of the Chunqiu fanlu, see Arbuckle 1991. For a proposed stratification of the text, see Queen and Major 2016: 20–29. Personal communication. Early texts use qiong 窮 in both senses, “to exhaust” or “to reach the ultimate.”

Kongzi As Author In The Han

277

while combining it with passages that have verbatim parallels in the Zuozhuan 左傳, the Mengzi, and three separate entries in the Lunyu:26 In the fourteenth year of Duke Ai of Lu, in spring, there was a hunt in Daye. One of Mr. Shusun’s chariot drivers captured a beast at Chushang and considered it inauspicious. Zhongni looked at it and said, “It is a unicorn.” It was seized. He said, “The [Yellow] River does not bring forth the Diagram; the Luo River does not bring forth the Writing. I am at the end!” When Yan Yuan died, Kongzi said, “Heaven has bereft me!” When after the hunt in the western region he saw the unicorn, he said, “My Way has reached its end!” Sighing deeply, he said, “Nobody recognizes me!” Zigong said, “Why is it that nobody recognizes you?” The Master said, “I do not complain against Heaven, nor do I blame other people. I study from below and reach up above. The one who recognizes me may be just Heaven!” 魯哀公十四年春,狩大野。叔孫氏車子鉏商獲獸,以為不祥。仲尼視 之,曰:麟也。取之。曰:河不出圖,雒不出書,吾已矣夫!顏淵 死,孔子曰:天喪予!及西狩見麟,曰:吾道窮矣!喟然歎曰:莫知 我夫!子貢曰:何為莫知子?子曰:不怨天,不尤人,下學而上達, 知我者其天乎!27

The passage suggests the richness, fluidity, and flexible applicability of Kongzi lore in Han times. Sima Qian’s account is a patchwork from at least four different sources, where Kongzi also appears under three different designations (Zhongni 仲尼, Kongzi 孔子, and zi 子). It jumps abruptly from point to point: first, Kongzi is shown as the person who correctly identifies the unknown beast; second, a passage on the absence of auspicious omens (the Yellow River Diagram and the Luo River Writing) is inserted in which Kongzi claims to be “at the end” (吾已矣夫); third, as in the Gongyang zhuan, he is quoted as feeling “bereft” after the death of his student; fourth, now returning to the capture of the unicorn, he claims, “My Way has reached its end”; fifth, he declares that only Heaven may recognize him (as in Lunyu 14/35; see below), but not without “sighing deeply” (喟然歎) and lamenting that, otherwise, “Nobody recognizes me!” (莫我知也夫). 26

27

Lunyu 9/9 (“Zi han” 子罕): 鳳鳥不至,河不出圖,吾已矣夫!Lunyu 11/9 (“Xian jin” 先進): 顏淵死。子曰:噫!天喪予!天喪予!Lunyu 14/35 (“Xian wen” 憲問): 子 曰:莫我知也夫!子貢曰:何為其莫知子也?子曰:不怨天,不尤人,下學 而上達。知我者其天乎!Mengzi 2B/13: 君子不怨天,不尤人。; Zuozhuan Ai 14. Shiji 47.1942.

278

Kern

This patchwork sequence emphasizes different aspects of Kongzi’s personality: his perspicacious mind in identifying supernatural portents and judging the state of the polity;28 his deep emotionality, paired with his concern for posterity, in this case of his student; and his sense of not being recognized. Even though the passage does not relate these traits explicitly to Kongzi’s authorship of the Chunqiu, they reappear in many depictions of other authors in the Shiji. Furthermore, the passage may well imply Sima Qian’s knowledge of the connection between the unicorn as an omen of Kongzi’s recognition by Heaven and of Kongzi’s “mandate” to compose the Chunqiu as mentioned in the Chunqiu fanlu fragment. In fact, these various aspects of Kongzi’s personality are also attributed to Sima Qian as he presents himself, or is presented, in the taishigong yue 太史公曰 (“the Honorable Lord Archivist says”) comments found throughout the Shiji, and they further relate to his own authorship of the latter.29 In short, the Shiji develops the image of Kongzi as both ideal and prototypical. By contrast, the passage in the Shiji’s “Rulin liezhuan 儒林列傳” (Arrayed Traditions of the Forest of Ru Scholars) is drastically shortened while making the connection with the Chunqiu explicit: When at the hunt in the western region they captured the unicorn, [Kongzi] said, “My Way has reached its end!” Thus, relying on archival records he made the Springs and Autumns so as to conform to the kingly law, with his phrasing subtle and his guidance broad. In later generations, many were the scholars who quoted from it. 西狩獲麟,曰:吾道窮矣!故因史記作《春秋》,以當王法,其辭微 而指博,後世學者多錄焉。30

Here, a third element is added: relating the making of the Chunqiu to the experience of “My Way has reached its end,” the passage claims that now, because of the Chunqiu, Kongzi’s influence continues through subsequent generations—an idea that resonates deeply and repeatedly elsewhere in the Shiji (see below), including in the parallel to the Mengzi passage cited above. While the concluding Gongyang zhuan entry says nothing about the quest for posterior recognition (or Kongzi’s authorship), it is followed by an epilogue 28 29 30

In early texts, Kongzi is celebrated as being particularly perspicacious in understanding and judging others (just as he was the one to correctly identify the unicorn); see Hunter 2017: chap. 2. For his ability to read portents, see also Nylan and Wilson 2010: 14, 20, 90. See Kern 2015, 2016. Shiji 121.3115.

Kongzi As Author In The Han

279

on the “making” of the Chunqiu that reads like an external insertion. Even if the Gongyang zhuan were to be accepted as a Warring States text,31 it remains difficult to decide how to date this epilogue and whether or not it even relates to Kongzi: Why did the noble man make the Springs and Autumns? Given that, in order to bring order to an age of chaos and to return it to correctness, nothing comes even close to the Springs and Autumns, would it be that he made it for this reason? Or was it because, as a noble man, he delighted in speaking of the Way of Yao and Shun? Or, finally, was it not because he was delighted that [future sages like] Yao and Shun would recognize the noble man?32 When establishing the right principle of the Springs and Autumns in order to await [his recognition by] later sages, this surely is what a noble man would delight in. 君子曷為為《春秋》? 撥亂世,反諸正,莫近諸《春秋》,則未知其 為是與?其諸君子樂道堯舜之道與?末不亦樂乎堯舜之知君子也?制 《春秋》之義,以俟後聖,以君子之為亦有樂乎此也。33

Who is the “noble man” (junzi 君子)? To a faithful reader of the Mengzi (and of a host of later texts), the answer is clear: Kongzi. But if one situates the Gongyang zhuan epilogue in pre-imperial times, there are serious arguments against that understanding. To begin with, we do not know whether or not the term junzi is referential at all—it may well be understood as “a noble man.” Both the Zuozhuan and the Guliang zhuan 榖梁傳 repeatedly invoke Kongzi as commentator; the former identifies him as either “Kongzi” or “Zhongni” 仲尼 (Kongzi’s courtesy name), and the latter invariably as “Kongzi.” At the same time, the Zuozhuan attributes yet another set of comments to “the [or “a”?] noble man.” As noted by Eric Henry, the ways in which the “noble man” and “Kongzi” express themselves on events in the Zuozhuan differ strikingly and consistently, and the “Kongzi” comments clearly postdate those of the “noble man.”34 Thus, whoever added the “Kongzi/Zhongni” comments seems to have 31 32 33 34

As argued by Gentz 2001: 345–403. The sentence is ambiguous; I agree with Malmqvist (1971: 218–219), Gentz (2001: 90), and Li (2007: 412) who take it to express the hope that future sages in the mold of Yao and Shun will recognize the author of the Springs and Autumns. Chunqiu Gongyang zhuan zhushu 719–721. Gentz (2001: 89–90, 384) has also identified this passage as a postface. In addition to Gentz’s analysis, see also the discussions in Schaberg (2001: 305–306) and, most detailed, in Li (2007: 411–421). Henry 1999; see also Schaberg 2005.

280

Kern

assumed that the “noble man” was not Kongzi/Zhongni.35 Likewise, the authoritative “noble man” who in eleven instances delivers pithy statements in the Guoyu 國語 is unrelated to Kongzi. On the other hand, in its dual references to an age in turmoil and to the theme of recognition (zhi 知), the Gongyang zhuan epilogue echoes the Mengzi passage quoted above. As will be shown below, the dual themes of “Kongzi made the Springs and Autumns” and “the noble man being recognized” together belong to the core of the Kongzi image nowhere more prominently than in the Shiji.36 While the Lunyu lacks any mention of the Chunqiu, it repeatedly touches on the question of recognition, as noted by Hunter in this volume (chap. 3): aside from 1/1 (“Not taking offense when not being recognized, is this not the mark of the noble man?” [人不知而不慍,不亦君子乎]), in 1/16 the Master says, “Do not worry about people not recognizing you; worry about you not recognizing others” (不患人之不己知,患不知人也); in 4/14 (“Li ren” 里仁), he declares, “Do not worry about not having a position; worry about what it takes to establish yourself. Do not worry about nobody recognizing you; strive for that for which you can be recognized” (不患無位,患所以立。不患莫己知,求為可 知也); in 11/26 (“Xian jin” 先進), the Master admonishes his disciples, “You constantly say, ‘People don’t recognize me!,’ but if someone recognized you, what would you do with that?” (居則曰: “不吾知也!”如或知爾, 則何以哉); in 14/30 (“Xian wen” 憲問), a variant of 1/16 is given (“Do not worry that people do not recognize you; worry that you yourself are incapable” [不患人之不己知,患其 不能也]), just as another one appears in 15/19 (“Wei ling gong” 衛靈公) with “The noble man is distressed by his lack of ability; he is not distressed that people do not recognize him” (君子病無能焉, 不病人之不己知也). While Hunter reads some of these passages as indicating that the Lunyu Kongzi is indeed much concerned with recognition, I think they must predominantly be taken to say that the noble man should not worry about not being recognized—as is clearly stated in the paradigmatic passage of Lunyu 1/1. This is the opposite of the Kongzi in the Mengzi, who exclaims that he will be recognized, or condemned, only for the Chunqiu, and even more so of the anonymous “noble man” in the Gongyang zhuan epilogue, who awaits posterity to 35

36

The relationship between the “noble man” and “Kongzi/Zhongni” comments may also be conceptualized differently, namely, that both were woven together by an editor. (By contrast, Henry 1999 takes the “noble man” as the Zuozhuan narrator.) In other words, while one stratum may be older, both may have entered the Zuozhuan text at the same time. (I thank Paul R. Goldin for this insight.) But even then it would appear that the editor distinguished “Kongzi/Zhongni” from the “noble man.” For a broader study on the problem of recognition, see Henry 1987.

Kongzi As Author In The Han

281

give him his rightful recognition. In my reading, only two passages in the Lunyu show a somewhat different take on recognition, but even these do not accord with the Gongyang zhuan or the Mengzi. One is 14/39, where someone observes that the Master’s playing of the chime stones seems to reveal his frustration over not being recognized; the person then remarks, “If nobody recognizes him, he should just stop it!” (莫己知也,斯已而已矣) and should instead adapt to the circumstances, an argument that the Master readily accepts. The other is 14/35, where the Master states, “Nobody recognizes me!” (莫我知也夫) and then concludes—at most with an implied sense of frustration—that “the one who recognizes me may be just Heaven!” (知我者其天乎). Here, Kongzi may be lamenting the ignorance of others in an imperfect world, but what truly matters to him is to be recognized by Heaven—which both Chunqiu fanlu (explicitly) and Shiji (implicitly) relate to his mandate for making the Chunqiu. There are several other passages in the Lunyu (e.g., 9/13 and 13/2) that dwell, directly or indirectly, on the theme of recognizing the worthy, but we do not find Kongzi advocating explicitly that one should worry about others’ recognition of oneself. In sum, while the dual connection of Kongzi with the theme of recognition and the creation of the Chunqiu attains a strong presence with Sima Qian (see below) and gains further traction in the last decades of the Western Han, it is weak in the Warring States and never directly advanced in the Lunyu. Mark Edward Lewis and Stephen W. Durrant both cite with appreciation Chen Renxi’s 陳仁錫 (1581–1636) statement that Kongzi’s entire biography in the Shiji hinges on the notion of Kongzi’s not being employed (篇中以 用不用二字為關鍵).37 Remarkably, the very part of the biography that emphasizes this element is also densely populated with lines from the Lunyu—as if the latter could be appropriated for a stance that it never takes. Finally, the third passage possibly of pre-Han origin that mentions Kongzi as the author of the Chunqiu is also found in the Gongyang zhuan, under the twelfth year of Duke Zhao 昭 (530 BCE): In the twelfth year, in spring, Gao Yan of Qi led an army and brought to power the Northern Yan Earl at Yang. What is meant by “Earl at Yang”? It is Prince Yang. The Master said, “I already knew this [miswriting of a personal name as a location].” A bystander said, “If you knew this, why did you not change it?” [The Master] said, “What about those [other instances] where one does not know [that something is wrong]? The Springs and Autumns is so faithful to history that its sequence [of lords is 37

Durrant 1995: 38; Lewis 1999: 228, citing Durrant. For Chen’s original remark, see Ling 1576: 47.24a.

282

Kern

the one established] by Duke Huan of Qi and Duke Wen of Jin, and its accounts of assemblies [match] how they were arranged by the hosts. [However,] as for its [choice of] words, I, Qiu, must bear the blame alone.” 十有二年。春。齊高偃帥師。納北燕伯于陽。伯于陽者何。公子陽生 也。子曰:我乃知之矣。在側者曰:子苟知之,何以不革。曰:如爾 所不知何。《春秋》之信史也。其序則齊桓晉文。其會則主會者為之 也。其詞則丘有罪焉耳。38

This is the only passage in the Gongyang zhuan where Kongzi, referring to himself in an intimate register by his first name, is mentioned explicitly as taking responsibility for the phrasing of the text. Together with Mengzi 3B/9 and, perhaps, the Gongyang zhuan epilogue, it constitutes a claim that is isolated among Warring States texts but fits tightly with Sima Qian’s account of Kongzi’s involvement with the Chunqiu. This observation does not constitute proof that the Mengzi and Gongyang zhuan passages are Han-dynasty interpolations, though it must be noted that no Han dynasty text cites Mengzi 3B/9 or, more generally, the Mengzi’s claim regarding Kongzi’s authorship. Perhaps the Gongyang zhuan, as Joachim Gentz and others hold,39 was indeed connected to Kongzi already in late Warring States times, and this connection is even implied within the text itself, including in its use of “the noble man” as a designation for Kongzi.40 But to judge from the evidence of our available sources, this idea would have been confined to a very small community of thinkers, if it had any influence at all. It had yet to gain prominence in the broader intellectual discourse of its time, and—most important for my present concerns—it had yet to become a defining feature in the conceptualization and representation of the Master.



The Hanshu 漢書 “Yiwen zhi” 藝文志 (Treatise on Arts and Letters) begins its account of the writings in the imperial library as follows: In the past, after Zhongni had perished, his subtle words were cut off; after his seventy disciples had died, the great meaning became perverted. 38 39 40

Chunqiu Gongyang zhuan zhushu 566–569. See also Gentz 2001: 96–98; Malmqvist 1978: 137; Malmqvist 1971: 203. Gentz 2001. For a critique of this last assumption, see Schaab-Hanke (2002: 283–291) who further refers to Pu 1995: 240–256.

Kongzi As Author In The Han

283

Therefore, the Springs and Autumns split into five [interpretive lineages], the Poetry into four, and the Changes accumulated traditions of multiple lineages. 昔仲尼沒而微言絕,七十子喪而大義乖。故《春秋》分為五,《詩》 分為四,《易》有數家之傳。41

While it remains inconclusive whether or not the “noble man” in the Gongyang zhuan epilogue originally referred to Kongzi, there is no question that readers and writers since Han times accepted this identification; no Han reader after Sima Qian would have failed to identify the “subtle words” (wei yan 微言) as those of the Chunqiu. If, for the sake of tracing the Han view of Kongzi, we read the Mengzi and Gongyang zhuan as mutually supportive statements on Kongzi’s authorship, we recognize in them a number of points, all of them relevant to the Kongzi figure in the Shiji: First is the choice of verbs. In both Mengzi 3B/9 and the Gongyang zhuan epilogue, Kongzi “makes” the Chunqiu as a textual response to the collapse of the moral and social order. The former uses the term zuo 作, while the latter uses wei 為 and then also notes that he “fashioned” (zhi 制) “the right principle” (yi 義) of the Chunqiu. Especially the use of zuo may be an oblique way of calling Kongzi a sage who “makes” at the level of the earlier sage-kings.42 What Kongzi “makes” is not just a text, but a new model of sovereignty that replaces the ways of earlier kingship with his own, as he appropriates “the business of the Son of Heaven.” Before the empire, nobody except Kongzi is ever credited with “making” a text meant to be read by an anonymous audience of readers, including those of posterity. Second, the Mengzi does not simply speak about Kongzi; it quotes him directly, infusing the account with the immediacy and authenticity of the Master’s own voice; the final entry of the Gongyang zhuan (before the epilogue) does the same, and so does the entry under Duke Zhao, which even uses the intimate self-designation Qiu. This voice is highly personal and resonates throughout the subsequent tradition—Kongzi consistently speaks not just his mind but also his heart. Needless to say, we do not hear Kongzi speak; we hear him as the Mengzi and the Gongyang zhuan, some centuries after his death, imagine and present him as speaking.

41 42

Hanshu 30.1701. For a discussion of zuo in early China, see Puett 2001.

284

Kern

Third, again in the Mengzi, Kongzi assumes true ownership of his work, and he is willing to bear the consequences. He is, in other words, a Foucauldian author who accepts punishment for his work.43 Fourth, the writing of history is conceived as a direct intervention into the political and social status quo. According to the Gongyang zhuan epilogue, the Chunqiu “brings order to an age of chaos and restores it to correctness” (撥亂 世,反諸正),44 while the Mengzi states, “After Kongzi had completed the Chunqiu, rebellious ministers and murderous sons lived in fear” (孔子成《春 秋》而亂臣賊子懼). While the former statement rectifies an imperfect past by means of a retrospective judgment that, in turn, is to be taken as guidance for the present and future, the comment in the Mengzi indicates an immediate and pervasive reception of Kongzi’s work—which raises very interesting questions about how it was “published,” that is, how it was transmitted to those “rebellious ministers and murderous sons.” In my reading, this is not merely pure fiction but an early step in the development of Kongzi’s hagiography. And fifth, Kongzi in both texts is portrayed as a self-conscious author who— while responding to his own time—writes for posterity, a motif that becomes central with Sima Qian. Thus, Kongzi’s writing provides us with his judgments on history, but more important, it tells us about his own moral stance. Kongzi inscribes himself into the text, assuming the role of the true author: the Chunqiu text attains a new meaning, and with it a new hermeneutical challenge, because it is now associated with Kongzi as its author. It is by no means clear what it means that Kongzi “made” the Chunqiu, considering that the text reflects (in a way that we do not really understand, considering Kongzi’s purported reworking of the text) the chronicles of his home state of Lu from 722 to 486 BCE (to 481 BCE in the Gongyang zhuan version). The Chunqiu is not a narrative; it is not even the skeleton of a possible narrative, presuming far more than it actually says, including the audience’s familiarity with numerous names, events, and their actual significance. Thus, it cannot possibly have been directed at any wider general audience (including “rebellious ministers and murderous sons”) because such an audience would not have been able to make any sense of it. Perhaps the early texts merely suggested that Kongzi “initiated” or “gave rise to” (other possible readings of the verb zuo, and a better match with Latin augere, “to augment”) its particular significance—in other words, that he transformed the chronicle into a discourse. The ambivalence over Kongzi’s role is apparent from the fact that in Han texts, he is also said to have “fashioned” (zhi 制), “made” (wei 為), 43 44

Foucault 1979. On the function of historiography to rectify history itself, see Schaberg 2001a: esp. chap. 8.

Kongzi As Author In The Han

285

“organized” (zhi 治), “arranged in sequence” (ci 次), “transmitted” (shu 述), “brought to completion (cheng 成), or “perfected” (xiu 修) the Chunqiu. Each of these terms still assigns to him the principal responsibility for the text in its final form. In light of these ambiguities, it is perhaps not surprising that there even seems to exist some early evidence against Kongzi’s authorship of the Chunqiu. In an anecdote in Han Feizi 韓非子 chapter 30, “Nei chushuo, Part One” 內儲說 上, Duke Ai of Lu 魯哀公 asks Kongzi why in the Chunqiu it is recorded that “[i]n winter, in the twelfth month, hoarfrost fell without killing the beans” (冬 十二月霣霜不殺菽).45 Kongzi responds that sometimes, someone who should be killed is not killed, and that “when Heaven loses the Way, even grasses and trees will go against it—how much more so if the ruler of men loses [the Way (that his people will go against him)]!” (天失道,草木猶犯干之,而況於人君 乎).46 Here, Kongzi is portrayed as a perspicacious conversational commentator on the Chunqiu—in other words, he would be explaining his own text. While some scholars take these comments to suggest Kongzi’s authorship of the Chunqiu, one may just as well conclude the opposite: as Kongzi used his “subtle words” to reveal the truth of history in coded ways, why would he then also go on record with an explicit commentary, and even autocommentary? Even more improbable is the setting of the anecdote: the entry supposedly raised by Duke Ai is the final (thirty-third) year of Duke Xi 僖公, the fifth duke chronicled in the Chunqiu, while Duke Ai is the twelfth and final duke chronicled there. Of course, Duke Ai could not possibly refer to the (already completed?) text of the Chunqiu—a text that includes his own reign—and engage the very author of the text in conversation.47 If anything, the anachronistic anecdote seems to suggest that whoever was responsible for including it in “Nei chushuo”—one of a mere handful of Han Feizi chapters named in the Shiji 48—considered Kongzi not to be the author of the Chunqiu. We can assume that the author of the anecdote was aware of these contradictions and expected the same from his audience. In having Kongzi explain the Chunqiu to Duke Ai, he granted him authority over the text. To some extent, this situation parallels—as noted above—Kongzi’s role as commentator on the Zuozhuan, where the voice of Kongzi/Zhongni is clearly external to the 45 46 47 48

The actual wording in the received Chunqiu is slightly different: “there fell hoarfrost without killing the grass” (隕霜不殺草). Wang Xianshen 1998: 30.223–224. Of course, Duke Ai may have had access to the earlier court annals created by the Lu court scribes; but he could not have referred to the Chunqiu as a text whose “subtle phrases” were fashioned by Kongzi. Shiji 63.2147.

286

Kern

text and even postdates that of the “noble man.” In both cases, Kongzi’s authority is not that of an author, but of a most perceptive reader. Yet at the same time, we must remain alive to the possibility that ancient readers were less troubled than we are today by textual and logical inconsistencies. Perhaps the author of the “Nei chushuo” could have it both ways and simply consider Kongzi the ultimate authority on all matters related to the Chunqiu, and hence present him as both author and commentator. Or perhaps he was playing with the expectations of an audience that already took Kongzi for granted as the author of the Chunqiu. We do not know; but if the latter was indeed the case, it remains curious that such an assumption was not voiced elsewhere as well. While modern scholars like Yang Bojun and others have long questioned any involvement of Kongzi with the Chunqiu,49 the idea of Kongzi as its “maker” became accepted over the long course of the Han dynasty and has been widely current since. Thus, Michael Nylan has stated that “The Han saw Kongzi, above all, as the author of the Spring and Autumn Annals. … [T]he story about his compilation of the Annals seems to drive all the other stories about Kongzi.”50 This is certainly the view one takes away from reading Sima Qian’s comments on Kongzi and from much of the literature from the late first century BCE through the end of the Eastern Han in the early third century CE. Yet it is not at all what we find before or even a generation after Sima Qian. There are, in fact, very few sources that attribute the Chunqiu to Kongzi.51 The first is Dong Zhongshu, who in two of his three responses to Emperor Wu’s policy questions—in the early years of the emperor’s reign—stated (or repeated the statement in the Mengzi) that “Kongzi made the Chunqiu” (孔子 作《春秋》).52 Likewise, in his proposal to ban all teachings “that are not within the curriculum of the Six Arts and Kongzi’s precepts” (不在六藝之科孔子之 術者),53 Dong connects the sage to the entire body of learning that gradually became distilled into the Five Classics (wu jing 五經). Finally, chapter 17 of the Chunqiu fanlu, “Yu xu” 俞序 (Summary Postface[?]), begins with the phrase “As for Zhongni’s making of the Springs and Autumns” (仲尼之作《春秋》也). However, not only is the attribution of the entirety of Chunqiu fanlu to Dong Zhongshu, and hence the date of any particular section, uncertain (though 49

50 51 52 53

Yang 1993: “Introduction,” 5–16. Yang cites not only a range of compelling reasons to question Kongzi’s involvement with the text but also a series of traditional thinkers from the seventh century onward who already doubted it. For further discussion, see also Gentz 2001: 38–40. In Nylan and Wilson 2010: 68, 74. For a similar argument, see Hunter 2017: 73n118. Hanshu 56.2509, 2515. Hanshu 56.2523.

287

Kongzi As Author In The Han

both chapters 16 and 17 are plausible as dating to mid-Western Han times); scholars have also called into question the reliability of Dong’s responses to Emperor Wu, raising the possibility that they were reshaped or even retrospectively created by Ban Gu 班固 (32–92), the author of the Hanshu.54 The second source to speak of Kongzi’s engagement with the Chunqiu is the Huainanzi 淮南子 of 139 BCE. In chapter 9 of this text, “Zhushu” 主術 (The Precepts of Rulership), Kongzi is called the “uncrowned king” (su wang 素王) for his “single-minded concentration on teaching the Way” (zhuanxing jiaodao 專行教道). In relating the 242 years of the Springs and Autumns period, “selecting what was good and weeding out what was shameful, he accomplished the kingly Way and was broad-minded in his deliberations” (采善鉏醜,以成 王道,論亦博矣); furthermore, “he created the Springs and Autumns without speaking of ghosts and spirits and without daring to concentrate on his own concerns” (作為《春秋》,不道鬼神,不敢專己).55 Here, the Master is commended for a “selfless” text that that does not advance his personal causes.



This, then, is the entire evidence for Kongzi’s authorship from the Warring States and early imperial periods up to Sima Qian. As Michael Loewe has noted, neither Lu Jia 陸賈 (ca. 228–ca. 140 BCE) nor Jia Yi 賈誼 (201–169) mentions Kongzi’s involvement with the Chunqiu anywhere in their received writings,56 nor does any early Han source other than the Huainanzi, including Sima Qian’s contemporaries—not in expository prose, not in memorials to the throne, not in imperial edicts. The earliest recorded instances following Sima Qian’s writings would come only a generation later, during the reign of Emperor Xuan 宣帝 (r. 74–49 BCE): when Huan Kuan 桓寬 retrospectively compiled the (idealized)57 Yantie lun 鹽鐵論 (Debate on Salt and Iron) of 81 BCE, he mentioned that Kongzi “made the Springs and Autumns”: Therefore, [Kongzi] traveled east and west, north and south, to persuade [the regional lords] but was not employed. Thereupon, he retired and perfected the kingly Way and made the Springs and Autumns, bequeathing it to the posterity of a myriad years.

54 55 56 57

See Loewe 2011: 118–121; Arbuckle 1991: 66–76; even more critical is Sun 2000. Liu 1997: 9.313. Loewe 2011: 149. See Loewe 1974: 81–112.

288

Kern 是以東西南北七十說而不用丙,然後退而修王道,作《春秋》,垂之 萬載之後。58

Finally, shortly after 68 BCE the then governor of Shanyang 山陽, Zhang Chang 張敞, noted in a memorial that “Zhongni made the Springs and Autumns” (仲尼作春秋).59

In light of the meager overall evidence to attribute such authorship to Kongzi, it is perhaps surprising that the few references available all appear in an almost offhand, matter-of-fact way. Nobody in early China felt the need to argue over Kongzi’s authorship of the Chunqiu—but then again, nobody in early China ever argued over the authorship of any text. “Kongzi made the Springs and Autumns” was never explicitly disputed, nor was there an attempt to explain in further detail how he did it, or what it was exactly that he did; the only characterization of that process is the repeated formula, first appearing in the Shiji, that he “relied on archival records” (因史記). We also know that the Shiji was not in general circulation during the first century BCE, and the surge of references to Kongzi’s authorship during the final three decades of that century is not sufficiently explained by way of reference to Sima Qian’s Kongzi biography.



The Chunqiu was prominent enough already in the fourth century BCE to give rise to a body of texts that directly or indirectly responded to it, including Zuozhuan and Gongyang zhuan. In the *Yucong 語叢 1 manuscript from the Guodian 郭店 corpus, probably dating from the late fourth century BCE, it is listed together with the other curricula of the Yi 易 (Changes), the Shi 詩 (Poetry), the Shu 書 (Documents), the Li 禮 (Rituals), and the Yue 樂 (Music)—that is, the Six Arts (liu yi 六藝)—and characterized as that “which brings together the affairs of the past and the present” (《春秋》所以會古今之事也).60 Moreover, the Shiji knows of an entire lineup of texts that were said to have drawn on the Chunqiu, and where it appears that the term is not (as elsewhere) used as the generic designation of historical annals but refers to a single text of elevated 58 59 60

Wang Liqi 1992: 20.253–254 (“Xiangci” 相刺). Hanshu 76.3217; Loewe 2011: 149n123. Jingmen shi bowuguan 1998: 195 (slips 40–41). I deliberately say “curricula” and not “books,” as I believe that the Six Arts, which in Han times became distilled into the Five Classics (by losing the Music), were not yet defined texts but rather repertoires (or discourses) of textual and ritual practices, forming the core of the Ru 儒 body of learning.

Kongzi As Author In The Han

289

status.61 It seems utterly unlikely that mere archival records, drawn up in the most basic format of annalistic notations, were as such considered important. What made them important was their moral and political interpretation, or the possibility of such interpretation.62 In other words, the Chunqiu constituted a curriculum of learning not as a body of factual records but as a repertoire of historical knowledge from which to extract lessons of precedence and edification. In this sense, it mirrored the songs from the Shi, which in *Yucong 1 are characterized as that “which brings together the aspirations of the past and the present” (《詩》所以會古今之志也);63 and it can be compared to the Yi, in the same manuscript said to be that “which brings together the Way of Heaven and the Way of Man” (《易》所以會天道人道也).64 Before the Han, no available source associates the Yi or the corpus of the Shi with an author or a compiler, which suggests there was no felt need for one. What both texts needed, however, were interpreters and teachers—masters who would turn these nonnarrative, nonargumentative, and hermeneutically wide-open (and therefore problematic) texts into repositories of cultural, political, and moral meaning. Leaving the Mengzi and Gongyang zhuan passages aside for a moment, this is precisely what can be said about the Chunqiu wherever else this text is mentioned. Before the Han, its prestige and authority were located not in the figure of an author but in its hermeneutic possibilities to reveal the past as meaningful for the present. To the extent that any one person mattered in relation to the text, it was not its originator but its interpreter and teacher, just as Kongzi is drawn upon as interpreter and teacher of the Poetry in the Shanghai Museum manuscript now titled, by its modern editors, *Kongzi shilun 孔子詩論 (Kongzi’s Discussion of the Poetry). In evaluating the historical reliability and significance of the Mengzi and Gongyang zhuan passages quoted above, we must consider this intellectual milieu of the Warring States, which on the whole had no need, and therefore likely no place, for a single authorial figure within the textual and ritual repertoire of the Six Arts. For both the Shi and the Yi, the authorial absence was not a deficiency but a quality of the text and its traditional authority. What mattered was the interpretability of the text, which was neither controlled nor constituted by its attribution to a historical persona, nor constrained by the 61 62

63 64

Shiji 14.509–510. Pines (2009: 318–332) has suggested that the Chunqiu originated from records for the ancestral sacrifice, and that at least some of its commentaries—especially the Gongyang zhuan—maintained a strong religious significance. Sima Qian, however, does not dwell on the religious meaning of the Chunqiu. Jingmen shi bowuguan 1998: 194 (slips 38–39). Jingmen shi bowuguan 1998: 194 (slips 36–37).

290

Kern

author function as “the principle of thrift in the proliferation of meaning”;65 it was inexhaustibly profound because it “has never finished saying what it has to say.”66 Discussions such as in *Yucong 1, but also in the Zuozhuan and Gongyang zhuan, suggest that the same was generally true for the Chunqiu. It was not before the Han that all the Five Classics became texts in search of their authors, and that the figure of the exemplary interpreter—Kongzi—became reconfigured as the exemplary author (for the Chunqiu), commentator (for the Yi), and compiler (for the Shi as well as for the Shu and the Li). Remarkably, while the figure of Kongzi as interpreter and teacher to some extent receded— note that there is no Han text comparable to the Kongzi shilun—it did not entirely disappear; the result is the self-contradictory image of the Master as both author of and (auto)commentator on the Chunqiu. Meanwhile, the Lunyu never once mentions the Chunqiu, but it does portray Kongzi as involved with the Shi and, to a lesser degree, also with the Yi, the Shu, and matters of ritual—yet nowhere as their author, compiler, or systematic commentator. It is fair to say that the Lunyu is largely uninterested in exploring Kongzi’s authorship or arrangement of any one of the classics, the Chunqiu included. In fact, compared with its emphasis on exemplary conduct, it is remarkably uninterested in texts—their existence, their production, their circulation and reception—altogether.



By far the single most important source for Kongzi’s authorship is the Shiji. In its concluding chapter 130, where the historian gives account of both his own life and his text, he creates a genealogy of suffering authors in which Kongzi’s authorship of the Chunqiu takes a decidedly different turn compared with the passages in the Mengzi and the Gongyang zhuan epilogue. Like so many others, Kongzi writes not merely out of political frustration and moral indignation but in direct response to personal suffering: In the past, the Earl of the West was incarcerated in Youli, and he expanded the Classic of Changes; Kongzi was in a desperate situation between Chen and Cai, and he made the Springs and Autumns; Qu Yuan was banished, and he composed “Encountering Sorrow”; Zuo Qiuming lost his eyesight, and there was the Discourses of the States; Sunzi got his 65 66

Foucault 1979: 159. Calvino 1986.

Kongzi As Author In The Han

291

feet chopped off, and he discoursed on the Art of War; Lü Buwei was banished to Shu, and his contemporaries transmitted Lü’s Survey; Han Fei was imprisoned in Qin, and [there were] the “Difficulties of Persuasion” and “Resentment about Solitude.” Most of the three hundred Odes [in the Poetry] were made by worthies who gave expression to their rage. All these men had something eating away at their hearts. They could not carry out the Way, and hence they wrote about the past while thinking of those to come. 昔西伯拘羑里,演《周易》;孔子厄陳蔡,作《春秋》;屈原放逐, 著《離騷》;左丘失明,厥有《國語》;孫子臏腳,而論《兵法》; 不韋遷蜀,世傳《呂覽》;韓非囚秦,《說難》、《孤憤》;《詩》 三百篇,大抵賢聖發憤之所為作也。此人皆意有所鬱結,不得通其道 也,故述往事,思來者。67

While echoing Kongzi’s quest for posterity in more general terms, this passage—repeated in Sima Qian’s famous letter to Ren An 任安68—in several of its details directly contradicts what the Shiji tells us about these authors elsewhere: Han Fei 韓非 (d. 233 BCE) was invited to the Qin court because the King of Qin (and future First Emperor) admired his already-existing essays “Gu fen” 孤憤 (Solitary Resentment) and “Wu du” 五蠹 (Five Vermin). He also had composed “Shui nan” 說難 (Difficulties of Persuasion) before arriving in Qin, where his imprisonment occurred later.69 Likewise, Lü Buwei 呂不韋 (d. 235 BCE) was the chancellor of Qin when he oversaw the composition of the Lüshi chunqiu 呂氏春秋 (Mr. Lü’s Springs and Autumns), which included a section of “surveys” (lan 覽); only several years later was he banished.70 Leaving aside the question of whether such openly contradictory statements should be attributed to the same author, the narrative in chapter 130 offers a genealogy of a particular ideological bent. Most important, the text constructs a direct relationship between personal suffering and authorship. In this account, Kongzi (like Han Fei) no longer writes in response to the political circumstances of his time (as he does in the Shiji passage from the “Arrayed Traditions of the Forest of Ru Scholars” where he composes the Chunqiu following the capture of the unicorn); instead, he becomes an author because of his personal fate. As the catalog of suffering 67 68 69 70

Shiji 130.3300. Hanshu 62.2735. The letter is also preserved as “Letter in Response to Ren Shaoqing” (Bao Ren Shaoqing shu 報任少卿書) in chapter 41 of the Wenxuan 文選. Shiji 63.2155. Shiji 85.2510–2512.

292

Kern

authors makes clear, they all respond to personal disaster in the same way: it is, in fact, the experience of suffering itself, and only this experience, that turns them into authors—none of them is writing by choice. In Kongzi’s case, the “making” of the Chunqiu arises from being in dire straits “between Chen and Cai,” which is “by far the most well-attested Kongzi tradition from the early period.”71 To be sure, the distinction between Kongzi’s personal suffering and his political frustration is not absolute. But there is still a striking difference between writing when “between Chen and Cai” and, as noted elsewhere in the Shiji, writing after the appearance of the unicorn. While the latter, close to the end of Kongzi’s life, portends both the futility of Kongzi’s political ambition and his imminent demise, the former aligns a significantly earlier moment of physical suffering (in this case, starvation) with the assaults on other heroes from the past. The difference between the two narratives is remarkable. The narrative about the unicorn shows Kongzi as a sage and historian of unique apprehension and clairvoyance, a solitary man who recognizes the unicorn for what it is, and who is being recognized by Heaven. The Kongzi “between Chen and Cai” is merely one among many who wrote out of suffering, one point in the line leading to Sima Qian himself. But both versions are meaningful for the author of the Shiji, as they express the dual motivations of his own authorship: writing out of personal suffering while also writing out of the desire to give testimony to a past that ends only in the time and person of the historian. For both the Chunqiu and the Shiji, these two motivations have generated two different meanings and readings of the text that cannot easily be reconciled. Both advance fundamental truth claims, but these claims—one by the historian, the other by the man speaking of his pain—are not entirely mutually supportive. At best they can be read as parallel or complementary; at worst, mutually undermining. Together they constitute the torn personas of both Sima Qian and his Kongzi.72



The Shiji’s conflicting statements regarding Kongzi’s authorship of the Chunqiu mark him as a composite persona, an image that could be invoked in different ways and for different rhetorical ends. Even though the Mengzi passage does not include the quest for posterity, it shows Kongzi as a high-minded author 71 72

Hunter’s chapter 3 in the present volume, p. 78. It may be argued that Kongzi’s personal suffering is somehow related to his desire to rectify the sociopolitical conditions of both past and present, but the texts do not actually say so.

Kongzi As Author In The Han

293

who, with fierce intent, fuses historical truth with personal emotion. Speaking its author’s heart and mind, the Chunqiu “makes sense” not as a mere chronicle but as the manifestation of both personal self-expression and political and historical critique. Sima Qian may well have inherited his understanding from recent tradition, that is, the Gongyang zhuan reading of the text for which his purported teacher Dong Zhongshu and, before Dong, Scholar Huwu 胡毋(母?)生 (early second century BCE) and Gongsun Hong 公孫弘 (ca. 200–121 BCE) were principal experts. But more than those of his predecessors, Sima’s account is repeatedly focused on the Master’s authorship. In the taishigong yue statement to Kongzi’s Shiji biography, Sima states, “When reading the writings of Master Kong, I see him before me as the person he was!” (余讀孔氏書,想見其爲人).73 The idea that through a person’s writing one gains direct access to his personality is repeated again verbatim (想見其爲人) in the statement on the other archetypal author of early China, Qu Yuan 屈原 (trad. ca. 343–278 BCE),74 who is likewise represented as having composed his writings out of the same fusion of political frustration and personal suffering. Several similar, though more indirect, statements to the same effect can be found in the Shiji’s evaluation of various philosophical “Masters.”75 In all such cases, it is the text that leads to the true nature of the author as a person—but only through its perceptive posthumous reader, who thus constitutes the posterity in which the author is finally recognized. The author becomes dependent on his reader: it is the latter who now imagines the former, who rescues both the text and the person. The principal difference in the accounts discussed so far concerns the authorial motivation: did Kongzi make the Chunqiu in response to a world without a true king (and for the mandate that he has therefore received from Heaven), and hence as a moral critique of the past and a guide for the future, or did he write from the experience of personal suffering and hence as a means of self-expression? The Shiji has it both ways, even though the second motivation has the potential to undermine and destabilize the first: how can Kongzi be an impartial critic of the past if he is driven by personal resentment? This tension is given voice already before Sima Qian—namely, in the above-quoted passage from the Huainanzi, “Zhu shu” (chap. 9): “He created the Springs and Autumns without speaking of ghosts and spirits and without daring to concentrate on his own concerns” (作為《春秋》,不道鬼神,不敢專己).76 In other 73 74 75 76

Shiji 47.1947. Shiji 84.2503. See Kern 2015. Liu 1997: 9.313.

294

Kern

words, this passage of the Huainanzi contradicts the notion that Kongzi wrote out of personal suffering. In “Fan lun” 氾論 (Boundless Discourses; chap. 13), the Huainanzi repeats the theme that the Chunqiu originated in response to the decline of order. Here, the text appears entirely impersonal—not “created” but “having arisen” in “an age of decline,” playing on the intransitive meaning of zuo 作 as “to arise”: When the kingly Way became deficient, the Poetry arose; when the House of Zhou failed and the principles of ritual collapsed, the Springs and Autumns arose. The Poetry and the Springs and Autumns are praised by the scholars, but each was produced by an age of decline. 王道缺而《詩》作,周室廢、禮義壞而《春秋》作。《詩》、《春 秋》,學之美者也,皆衰世之造也。77

A parallel can be found in Mengzi 4B/21, which also mentions the two classics most closely associated with Kongzi: Mengzi said, “When the footprints of the [ancient] kings disappeared, the Poetry vanished. After the Poetry had vanished, the Springs and Autumns arose.” 孟子曰:王者之迹熄而《詩》亡,《詩》亡然後《春秋》作。78

Late in the Western Han, Liu Xiang’s Shuiyuan goes so far as to quote Kongzi himself on this impersonal, quasi-natural rise of the Chunqiu: Kongzi said, “Had the Way of Xia not vanished, the Virtuous Power of Shang would not have arisen; had the Virtuous Power of Shang not vanished, the Virtuous Power of Zhou would not have arisen; had the Virtuous Power of Zhou not vanished, the Springs and Autumns would not have arisen. After the Springs and Autumns had arisen, the noble man [men?] understood that the Way of Zhou had vanished.”

77 78

Liu 1997: 13.427. Jiao 1987: 572. For a different reading, with 迹 (traces) as 䢋 (wooden clappers), see Yang 1988: 193: “When the wooden clappers of the [ancient] kings went silent, the Poetry vanished.”

Kongzi As Author In The Han

295

孔子曰:夏道不亡,商德不作;商德不亡,周德不作;周德不亡, 《春秋》不作;《春秋》作而後君子知周道亡也。79

In yet another passage, the Shuiyuan states, The Master traveled to persuade seventy regional lords, not having a stable residence. He was intent on making the common people across the realm obtain their proper station, yet his Way was not carried out. He retired and perfected the Springs and Autumns . . . 夫子行說七十諸侯無定處,意欲使天下之民各得其所,而道不行。退 而修《春秋》…80

Finally, the Shuiyuan takes a turn that connects it directly to the Chunqiu fanlu fragment noted above: [Kongzi] appreciated the finest of what was good and censured the tiniest of what was detestable. Human affairs were penetrated, the kingly Way was fulfilled; as [Kongzi] refined and harmonized the sage stipu­ lations, he communicated them to Heaven above, and the unicorn arrived—this shows how Heaven recognized the Master! Thereupon he sighed deeply and said, “Isn’t Heaven of utmost brightness so that it cannot be overshadowed? How then can there be a solar eclipse? Isn’t Earth of utmost security so that it cannot be in peril? How then can Earth be shaken?” Yet Heaven and Earth are still being overshadowed and shaken; thus, when the worthies and sages speak to the world but cannot have their Way carried out, disasters and natural anomalies arise together. The Master said, “I do not complain against Heaven, nor do I blame other people. I study from below and reach up above. The one who recognizes me might be just Heaven!” 采毫毛之善,貶纖介之惡,人事浹,王道備,精和聖制,上通於天而 麟至,此天之知夫子也。於是喟然而歎曰:天以至明為不可蔽乎?日 何為而食也?地以至安為不可危乎?地何為而動?天地尚有動蔽,是 故賢聖說於世而不得行其道,故災異並作也。夫子曰:不怨天,不尤 人,下學而上達,知我者其天乎!81 79 80 81

Xiang 1987: 1.31 (“Jun dao” 君道). Xiang 1987: 14.350 (“Zhi gong” 至公). Xiang 1987: 14.351 (“Zhi gong”).

296

Kern

Here, we witness an amalgam of various passages from the Shiji, the Mengzi, and the Lunyu, combining the various ideas already encountered above. Yet more explicitly than in any other Western Han source, the unicorn, sent by Heaven, appears in response to the presence of the sage because, indeed, only Heaven recognizes Kongzi.



In sum, by late Western Han times the image of Kongzi as an expert in the interpretation of the Chunqiu had fully developed into one of his authorship of the text itself. For pre-imperial times, however, their general (let alone widespread) conflation is questionable: if anything—as in the Han Feizi passage where Kongzi is consulted by Duke Ai—the two ideas appear as mutually exclusive. Among our available sources, by far the most influential account of Kongzi as author of the Chunqiu is that of the Shiji, even though it is riddled with contradictions about Kongzi’s motivation. Across its various chapters, the Shiji connects Kongzi repeatedly to the Chunqiu, for example, in its table on the feudal lords.82 Later Western and Eastern Han texts such as Shuiyuan, Yantielun, and Kongzi jiayu 孔子家語 (Family Sayings of Kongzi),83 but also public figures such as Yang Xiong 揚雄 (53 BCE–18 CE)84 and Wang Mang 王莽 (r. 9–23),85 drawing on a pool of earlier statements, could routinely refer to Kongzi as having created the Chunqiu. In addition, Xin 新, Eastern Han, and Wei 魏 dynasty thinkers would refer to Kongzi as having encoded the Chunqiu with predictions about the rise of Wang Mang’s Xin dynasty,86 the restoration of the (Eastern) Han,87 or the establishment of the Wei dynasty,88 respectively. In view of these various sources and the authority that the Mengzi, the Gongyang zhuan, and, most of all, the Shiji have exerted over the tradition, it is easy to lose sight of the fact that the great majority of Warring States and early Han texts never speak of Kongzi as the author of the Chunqiu. These include the “Masters” texts as well as Han texts such as Han Shi waizhuan 韓詩外傳 (Outer Tradition of the Han Poetry), Liji 禮記 (Records of Ritual), Da Dai Liji 大 戴禮記 (Elder Dai’s Records of Ritual), Jia Yi’s Xinshu 新書 (New Writings), of Lu Jia’s Xinyu 新語 (New Discourses), and Liu Xiang’s Xinxu 新序 ([Writings] Newly 82 83 84 85 86 87 88

Shiji 14.509–510. Kongzi jiayu 39/2 (“Benxing jie” 本性解); 42/1 (“Quli Zigong wen” 曲禮子貢問). Hanshu 87B.3578. Hanshu 99B.4109. Hanshu 99B.4109. Hou Hanshu 13.538. Sanguo zhi 3.108.

297

Kongzi As Author In The Han

Arranged). From this larger perspective, it is not surprising that the Chunqiu (let alone its authorship) is never once mentioned in the Lunyu. Some of the characteristics attributed to Kongzi there are also applied to him as author in the Shiji and later texts, in particular his sense of rightness and ritual order. Yet other aspects of the Kongzi persona seem to show a particular Han perspective: his subtle phrasing, his unique perspicaciousness in identifying the unicorn and other portents, his sensibilities as a reader, his sage nature as the “uncrowned king,” his intense emotion, his desire for recognition and quest for posterity. It would not have escaped an author like Sima Qian that all these traits were precisely the ones that he claimed for himself.89 We are well advised to consider how much the Kongzi persona in the Shiji depends on his biographer and therefore should not be projected back into earlier sources, including the Lun­yu. According to the Kongzi biography in the Shiji, Kongzi becomes involved with the classics only toward the end of his failed career.90 When after fourteen years of unsuccessful travel, he finally returns home to Lu 魯, he begins to talk about good government in terms that no reader can fail to relate to Sima Qian’s personal fate (and that of his father): “Good government,” Kongzi is made to pronounce, “lies in the selection of ministers” (政在選臣); and then he goes on to explain that in order to lead the people, one does not rely on rewards but on the good moral example of the ruler. Perhaps unsurprisingly, “thereafter Lu in the end could not make use of Kongzi, and Kongzi also did not seek office” (然魯終不能用孔子,孔子亦不求仕).91 Just like Mengzi, who retires after not having been employed,92 Kongzi advises his ruler in vain and moves to the fringes; and also like Mengzi, he then turns to writing as the logical consequence of having been rejected as an adviser.



89

90 91 92

Here, I paraphrase David Schaberg’s (2001a: 257) fine observation with respect to the Zuozhuan compilers, namely, that “the historiographers . . . could not have failed to recognize what they had in common with the men whose deeds they were commemorating.” For extensive discussion of how much of Kongzi Sima Qian saw in himself, see Durrant 1995. See also Kongzi’s involvement with all the classics in Sima Qian’s “Self-Narrative” (Shiji 130.3296–3297), where particular emphasis is placed on the Chunqiu. Shiji 47.1935. Shiji 74.2343. Zhao Qi 趙歧 (d. 201), the author of the earliest extant commentary of the Mengzi (and possible organizer of the Mengzi as a book), emphasizes this parallel between Mengzi and Kongzi. Hunter (2014: 73–74) suggests that Zhao Qi may have gone so far as to structure and title the Mengzi chapters in order to stress the connection with the Lunyu.

298

Kern

The Huainanzi—perhaps together with the writings attributed to Dong Zhong­ shu—provides evidence that the attribution of the Chunqiu to Kongzi predates the Shiji. However, the Huainanzi may slightly postdate a possible compilation window for the Lunyu toward the beginning of Emperor Wu’s reign, the approximate date proposed by Hunter, as do Dong Zhongshu’s compositions (which could even be significantly later). This would leave us with the Mengzi and Gongyang zhuan passages as our only available—and highly tenuous— references to the connection between Kongzi and the Chunqiu. In other words, the compilers of the Lunyu may not have been aware of this connection at all, which would be a fine explanation for their failure to mention the Chunqiu, let alone Kongzi’s authorship of it. Yet on the other hand, we must remain alive to the fact that so much of the textual world of Warring States, Qin, and Han China is lost to us. There is every possibility that tomorrow, some excavated manuscript will plainly state, “Kongzi made the Springs and Autumns.” And then what? Let us therefore assume, if only for the sake of the argument, that the Lunyu compilers of the mid-second century BCE were indeed aware of this very phrase. Why did they decide—and why were they free—not to include it in their representation of the Master? One explanation for this decision might be found in the way the Chunqiu was read in the Han—the very way that made it so potent in political discourse. While the text rarely served as an authoritative “proof text” as the Shi and the Shu were rhetorically invoked to cap philosophical and political arguments, it offered (a) a repository of applicable precedents together with (b) a model of political judgment and criticism available to the newly emerging class of scholar-officials. In this combination, the text extended the “oppositional stance”93 of the pre-Qin philosophers to the new class of career classicists—the salaried Ru scholars at the imperial court. If the Shu was frequently invoked in imperial edicts,94 it was because it represented the royal voice. By contrast, the Chunqiu, as read through the Gongyang, Guliang, and Zuo traditions, provided the model of the upright minister and noble adviser whose wisdom was as often ignored as it was heeded. In other words, if the voice of the Documents could be appropriated as the imperial voice, the voice of the Chunqiu and its three traditions was available to those speaking truth to power. After all, this was the role by which Kongzi himself, as the “uncrowned king,” was defined: the sage king who, in Heaven’s view, replaces the worldly kings. As scholars since Chen Renxi in the late Ming have repeatedly noted, the Shiji suggests that Kongzi’s entire involvement with all of the Six Arts (and Five Classics) derived from his 93 94

See Lewis 1999: chap. 2, esp. 70–73. As noted by Michael Hunter (personal communication).

Kongzi As Author In The Han

299

experience of “not being employed” (bu yong 不用)—neither abroad nor after his final return to his home state of Lu. Wang Chong 王充 (27–ca. 100), who in his Lunheng 論衡 (Balanced Discussions) refers to Kongzi’s authorship of the Chunqiu some twenty times (but barely to his engagement with the other classics), goes so far as to assert that “had Kongzi become king, the Springs and Autumns would not have arisen” (使孔子得王,《春秋》不作).95 In this, he echoes not only Sima Qian’s genealogy of suffering authors but also a taishigong yue comment in the Shiji on the writing of another Chunqiu text, this one by Excellency Yu 虞卿 (Yu Qing; third century BCE): like Kongzi, Excellency Yu is said to have written the Yushi chunqiu 虞氏春秋 (Mr. Yu’s Springs and Autumns) in order to “show himself to later generations” (自見於後世): However, had Excellency Yu not gone through hardship and grief, one might say that he also would not have been able to compose writings and show himself to later generations. 然虞卿非窮愁,亦不能著書以自見於後世云。96

But perhaps something even simpler and more obvious is behind Wang Chong’s comment: a text like the Chunqiu cannot be the work of a king. In its reading as a work of “praise and blame” (baobian 褒貶), it can only be a text speaking upward to power, not one speaking downward from there. The figure of Kongzi as a failed political adviser only enhanced its critical stance and trustworthiness. What further made the text both powerful and problematic was its association with an age of moral decline, as is emphasized in so many passages— starting with the Mengzi—quoted above; like none of the other classics, it was the very product of a corrupt time. When a Han official invoked the Chunqiu to speak, directly or indirectly, in the voice of Kongzi, he immediately assumed a position that was doubly fraught: as a morally superior yet at best marginally powerful person speaking toward his ruler, and as a man whose appropriation of the Chunqiu could carry the implication of criticizing his own times. Considering the constant factionalism that marked political debate throughout the Han, such thoughts would never have been far away, neither for the officials nor for their emperors. In fact, the issue is addressed head-on in Sima Qian’s “Self-Narrative” that concludes the Shiji. In an exchange with Grandee Hu Sui 壺遂, Sima defends his own composition of a work of history by saying 95 96

Huang 1990: 19/82.1152 (“Shu jie” 書解). Shiji 76.2376. For further discussion, see Kern 2015.

300

Kern

that it is precisely not like the Chunqiu, that is, a work of political criticism. To his praise of Kongzi’s work that “to bring order to an age of chaos and to return it to correctness, nothing comes even close to the Springs and Autumns” (撥亂 世反之正,莫近於《春秋》), Hu Sui retorts, Since in Kongzi’s time, above there was no enlightened lord, below [Kongzi] could not gain employment. Therefore, he made the Springs and Autumns. … Today, as you, sir, above have met an enlightened Son of Heaven, below you have secured a position; the ten thousand matters are complete, each having been placed in correct order. In what you discourse upon, what is it that you wish to explain? 孔子之時,上無明君,下不得任用,故作《春秋》…今夫子上遇明天 子,下得守職,萬事既具,咸各序其宜,夫子所論,欲以何明?

This prompts Sima Qian to offer his remarkable disclaimer, dripping with irony, self-effacement, and accusation: Yet if servicemen and worthies are capable but not employed, it is the shame of one who holds the state [i.e., his emperor]; and if the ruler above is enlightened and sagacious but the fame of his virtue is not spread, it is the fault of the officials. Moreover, I have enjoyed holding my office, but to discard and not record the flourishing virtue of the enlightened sage and to erase and not transmit the accomplishments of the meritorious ministers, hereditary houses, dignitaries, and grandees and to let the words of the former men [of virtue] fall away would be the greatest of all crimes. What I call [my task of] transmitting affairs from the past is [merely] to put what has come down from the earlier generations into good order and balance, and it is not what one calls “making”; thus, your comparing [my work] to the Springs and Autumns is mistaken. 且士賢能而不用,有國者之恥;主上明聖而德不布聞,有司之過也。 且余嘗掌其官,廢明聖盛德不載,滅功臣世家賢大夫之業不述,墮先 人所言,罪莫大焉。余所謂述故事,整齊其世傳,非所謂作也,而君 比之於《春秋》,謬矣。97

Here is what Sima Qian manages to say in a most eloquent act of affirmation by way of denial: First, Kongzi made the Chunqiu in response to an age of 97

Shiji 130.3297–3300.

Kongzi As Author In The Han

301

moral decline. Thus, his work not merely offers historical criticism but is the foremost way “to bring order to an age of chaos and return it to correctness.” Second, because Sima’s own time is blessed with an enlightened sage, there cannot be anything to criticize. Third, it would be a ruler’s shame not to employ (bu yong 不用, the very phrase Sima uses elsewhere for Kongzi’s fate) the worthies—which, as it happens, is precisely what Emperor Wu did to Sima Tan 司馬談 (who in 110 BCE died out of grief over the humiliation) and then to Sima Qian himself (who was forced to choose between suicide and castration). Fourth, it would then be the historian’s greatest crime not to record the worthies of his time (who, just like those from the past, suffer from neglect and abuse and whose virtue would otherwise be forgotten)—a direct echo of the description of Kongzi’s efforts in Shiji chapter 61, “The Biography of Boyi” (Boyi liezhuan 伯夷列傳), and another slight aimed at Emperor Wu. Fifth, in claiming that his work therefore cannot be compared to the Chunqiu, he appropriates Kongzi’s most famous disclaimer, “I transmit but do not make” (述而不作), pronounced in Lunyu 7/1.98 In short, Sima Qian artfully concurs with Hu Sui that indeed he is a second Confucius; and his work, a new Chunqiu. This is the climate of Western Han intellectual life, and we cannot think about the role of the Lunyu in the Han—and its relation to the Chunqiu—without thinking about Han political debates. There is very limited evidence that the Chunqiu was part of the upbringing of members of the imperial family. For the year 123/122 BCE, the Hanshu notes that Emperor Wu had his crown prince, Liu Ju 劉據 (Wei taizi 衛太子), tutored in the Gongyang teaching of the text; in addition, Liu Ju also received the Guliang teaching.99 I do not know what to make of this singular case other than to point out how exceptional it is. Fifty years later, shortly after 74 BCE, Emperor Xuan heard that Liu Ju had been “fond of” (hao 好; as opposed to “be instructed in”) the Guliang teaching; after some consultation with his scholars at court, the emperor deemed it appropriate to promote the Guliang zhuan.100 The very fact that the initial event was remembered two generations later suggests that it was highly exceptional. By contrast, the Hanshu repeatedly mentions the Lunyu as one of the central texts for the education of the crown prince and other members of the ruling family,101 making it an imperial text par excellence. All these mentions also include references to one of the other classics—that is, 98 99 100 101

As Michael Hunter has pointed out to me (personal communication), Sima Qian himself does not explicitly attribute the phrase shu er bu zuo to Kongzi. Hanshu 63.2741. Hanshu 88.3618. Hanshu 53.2428, 68.2947, 71.3039, 75.3159, 78.3282, and 81.3347. See also Hunter’s essay in the present volume (chap. 3).

302

Kern

either the Shi, the Shu, the Yi, or instructions in ritual—yet not once to the Chunqiu. Was it seen as incompatible with the Lunyu? The Kongzi of the Chunqiu is by definition a man of judgment and argument, most keenly interested—as is made explicit repeatedly when Han authors talk about the text—in future applications of the experiences from the past, including his subtle but nevertheless harsh criticism of rulers. The Kongzi of the Lunyu, by contrast, is overall a less explicitly political figure and, indeed, is notably deferential toward political authority.102 He does not speak for the ruler, nor does he speak up to or against the ruler. He speaks only for himself, advocating a set of social norms and moral ideals that could easily be accommodated to the education of the imperial crown prince. Given our limited evidence, it would be quite hazardous to state any of this in absolute terms. But the closer and longer one looks at the Han, the less surprising the absence of the Chunqiu in the Lunyu seems to become. The traditional view that sees Kongzi’s authorship of the Chunqiu as central to how early thinkers envisioned the Master has little support before Sima Qian and remains strikingly limited even for another century after him.

Epilogue

It should be clear from the foregoing pages that I am not primarily interested in the question of whether or not “Kongzi made the Springs and Autumns.” Whether Kongzi was regarded as its author (in whatever sense) or as its supreme interpreter does not change our understanding of the text. For the dating of the Lunyu, the issue cannot be used to falsify a Western Han compilation date. The more interesting question would be why Kongzi’s authorship was suddenly so important to Sima Qian while barely ever mentioned before—and how it then disappeared again from the textual record until several generations after Sima Qian. In light of the fact that the Shiji was not widely known or available during the first century BCE, there seems to have been some undercurrent of intellectual history where this claim was considered credible or even important; moreover, such an undercurrent, invisible to us today, may have existed already in the late Warring States. Our surviving sources are too fragmentary; we simply do not know. Whatever the case, it still remains unclear what “Kongzi zuo Chunqiu” 孔子 作春秋 may have meant initially—for example, in the Mengzi (if that passage should date from the Warring States). Before Sima Qian, one could accept 102

Hunter 2017: 287–288.

Kongzi As Author In The Han

303

Kongzi as the supreme authority on the Chunqiu without implying that he was the author of the text. For once, we could take his statement “I transmit but do not make” seriously. In this case, what Kongzi does is not passive transmission but active interpretation: he brings the text from the past to life. As such, shu is not less than zuo. While only the sages “create” or “make,” they depend on posterity to “transmit,” that is, to reactivate and reconstitute the meaning of antiquity for an ever-evolving present time. This—not creation as such—is the principal significance of Kongzi’s involvement with the Chunqiu in all early passages, including those (such as the one in Han Feizi) that cannot be constructed as direct statements on authorship. Kongzi is the person who makes the Chunqiu speak—not about himself but about the past that has given rise to the present.103 “Kongzi zuo Chunqiu,” at least initially, may thus best be understood as “Kongzi gave the Springs and Autumns its meaning,” and made that meaning matter. He did not “make” the text but “augmented” (augere) it. What is more, the title Chunqiu may then refer to more than just the text we have: namely, to the entire discourse and field of knowledge of the past as represented by the text and its interpretation.104 Or put in other terms: “Kongzi zuo Chunqiu” refers to giving rise to the interpretation of the recent past through textual means—which is what elevates the Chunqiu to the level of a classic.105 Without this interpretive appropriation, the court chronicle of Lu would have remained just that: an annalistic list of events, stored in the archive. While Sima Qian wants us to believe that Kongzi created the Chunqiu out of his personal urge and necessity, it may well be the other way around: the text needed Kongzi in order to become a classic that, to repeat Italo Calvino’s beautiful observation, “has never finished saying what it has to say.” But to assume the durability and authority of a classic, it did not really need Kongzi the author; it needed Kongzi the “augmenter” and interpreter (just as, from a Han perspective, the Shi, Shu, and Yi needed Kongzi’s editorial and commentarial involvement). Authorship without interpretable intent means literally nothing, while on the other side, interpretation does not require authorship. From this perspective—which matches how Kongzi is portrayed across early sources, including in the Lunyu—Kongzi becomes a function of the process 103 104 105

As Hunter (2017: 54–56 and chap. 2) notes, Kongzi is unique in his status as a universal commentator. I assume the same situation to explain the inflated numbers of characters that the Shiji notes for many early “Masters” (zi 子) texts; see Kern 2015. For a dense and sophisticated analysis of the mutually constitutive relationship between Kongzi as sage author and the Chunqiu as canonical writing, see Guo 2016. Guo’s discussion of the discursive practices of Chinese classicism since the Han touches on many of the issues raised in the present essay.

304

Kern

of the text’s canonization. Before that, he already is the established authority on the Shi, and it is easy to see how this authority is then transposed to the other classics, including the Chunqiu. But with the latter text, Kongzi also takes on a new role: more than being the transmitter and teacher of the glorious era of high antiquity, he is now also the interpreter of the recent past and even of his own present time, in which he intervenes by way of his text. Among the classics, the Chunqiu is the true bridge to antiquity, as it connects an earlier past all the way down to Kongzi’s own time and even to his death. It is therefore also the one text that can be claimed to have arisen only at a time of decline— which, finally, with the appearance of the unicorn, seems to collapse the moment of interpretation into that of creation, and the interpreter into a man of political and even personal despair. As noted above, the original prestige and authority of the text may have rested not with the figure of an author but with the text’s hermeneutic possibilities to reveal the past as meaningful for the present. The author figure—as “the principle of thrift in the proliferation of meaning”—was then both: a function that opened the text to a particular, unified interpretive perspective and one that restricted it to just that. It enabled the text to mean something while impeding it to mean something else. But if the original connection of Kongzi with the Chunqiu turned a corpus of annalistic records into a classic of hermeneutic needs so that it could continue “saying what it has to say,” Sima Qian’s account, together with late Western Han ideological and bibliographical purposes and practices, helped to stabilize further the significance of the Chunqiu as Kongzi’s “work” and by this also to limit its possible range of interpretation. Textual repertoires and traditional records do not have authors; when they acquire authors, they become “works.” With Sima Qian and then in the late Western Han, the attribution of the classics to Kongzi cannot be divorced from this textualization, definition, and stabilization of the traditional canon into a fixed set of books. But even by 140 BCE, this development had yet to occur, whether with the alleged establishment of the “erudites for the Five Classics” (五經博士) in 136 BCE or significantly later.106 In other words, the Kongzi of the Lunyu may have arrived just a bit too early to participate in this endeavor.

106

For the most systematic review—and rejection—of the claim in Hanshu 6.159 that Emperor Wu “established the erudites for the Five Classics” (zhi wujing boshi 置五經博士), see Fukui 2005: 109–258. Fukui argues for a much later date.

Kongzi As Author In The Han



305

Bibliography

Arbuckle, Gary. 1991. “Restoring Dong Zhongshu (BCE 195-115): An Experiment in Historical and Philosophical Reconstruction.” PhD diss., University of British Co­lumbia. Behr, Wolfgang. 2011. “Linguistic Approaches to the Dating of the Lunyu: Methodological Notes and Future Prospects.” Paper presented at “The Analects: A Western Han Text?,” Princeton University, November 4–5, 2011. Brooks, E. Bruce, and A. Taeko Brooks. 1998. The Original Analects: Sayings of Confucius and His Successors. New York: Columbia University Press. Cai, Liang. 2010. “Who Said, ‘Confucius Composed the Chunqiu’? The Genealogy of the ‘Chunqiu’ Canon in the Pre-Han and Han Periods.” Frontiers of History in China 5:363–385. Calvino, Italo. 1986. “Why Read the Classics?” New York Review of Books, October 9, 1986. Cheng, Anne. 1993. “Lun yü 論語.” In Early Chinese Texts: A Bibliographical Guide, ed. Michael Loewe, 313–323. Berkeley: Society for the Study of Early China and Institute of East Asian Studies, University of California. Chunqiu Gongyang zhuan zhushu 春秋公羊傳注疏. 2000. Beijing: Beijing daxue chubanshe. Durrant, Stephen W. 1995. The Cloudy Mirror: Tension and Conflict in the Writings of Sima Qian. Albany: State University of New York Press. Foucault, Michel. 1979. “What Is an Author?” In Textual Strategies: Perspectives in Poststructuralist Criticism, ed. Josué V. Harari, 141–160. Ithaca, NY: Cornell University Press. Fukui Shigemasa 福井重雅. 2005. Kandai Jukyō no shiteki kenkyū: Jukyō no kangakuka o meguru teisetsu no saikentō 漢代儒教の史的研究: 儒教の官學化をめぐる定說の再 檢討. Tokyo: Kyūko shoin. Gentz, Joachim. 2001. Das Gongyang zhuan: Auslegung und Kanonisierung der Früh­ lings- und Herbstannalen (Chunqiu). Wiesbaden: Harrassowitz Verlag. Graham, Angus C. 1978. Later Mohist Logic, Ethics and Science. Hong Kong: The Chinese University Press. Guo Xi’an 郭西安. 2016. “Shengren, shengyan yu shengxing: zuowei yizhong jingxue panduan de ‘Kongzi zuo Chunqiu’” 聖人、聖言與聖性:作為一種經學判斷的“孔 子作《春秋》”. Hunan daxue xuebao (shehuikexue ban) 湖南大學學報(社會科 學版)2016.1:47–55. Hanshu 漢書. 1962. Beijing: Zhonghua shuju. Henry, Eric. 1987. “The Motif of Recognition in Early China.” Harvard Journal of Asiatic Studies 47:5–30.

306

Kern

Henry, Eric. 1999. “‘Junzi Yue’ versus ‘Zhongni Yue’ in Zuozhuan.” Harvard Journal of Asiatic Studies 59:125–161. Hou Hanshu 後漢書. 1965. Beijing: Zhonghua shuju. Huang Hui 黃暉. 1990. Lunheng jiaoshi 論衡校釋. Beijing: Zhonghua shuju. Hunter, Michael. 2012. “Sayings of Confucius, Deselected.” PhD diss., Princeton University. Hunter, Michael. 2014. “Did Mencius Know the Analects?” T’oung Pao 100:33–79. Hunter, Michael. 2017. Confucius Beyond the Analects. Leiden: Brill. Jiao Xun 焦循. 1987. Mengzi zhengyi 孟子正義. Beijing: Zhonghua shuju. Jingmen shi bowuguan 荊門市博物館. 1998. Guodian Chu mu zhujian 郭店楚墓竹簡. Beijing: Wenwu chubanshe. Kern, Martin. 2000. “Feature: Writing and Authority in Early China, by Mark Edward Lewis.” China Review International 7:336–376. Kern, Martin. 2015. “The ‘Masters’ in the Shiji.” T’oung Pao 101:335–362. Kern, Martin. 2016. “Shiji li de ‘zuozhe’ gainian” 《史記》裡的‘作者’概念. In Shiji xue yu shijie hanxue lunji xubian 史記學與世界漢學論集續編, ed. Martin Kern and Lee Chi-hsiang 李紀祥, 23–61. Taipei: Tonsan Publications. Kimura Eiichi 木村英一. 1971. Kōshi to Rongo 孔子と論語. Tokyo: Sōbunsha. Lewis, Mark Edward. 1999. Writing and Authority in Early China. Albany: State University of New York Press. Li, Wai-yee. 2007. The Readability of the Past in Early Chinese Historiography. Cambridge, MA: Harvard University Asia Center. Ling Zhilong 凌稚隆. 1576. Shiji pinglin 史記評林. Liu Wendian 劉文典. 1997. Huainan honglie jijie 淮南鴻烈集解. Beijing: Zhonghua shuju. Loewe, Michael. 1974. Crisis and Conflict in Han China, 104 BC to AD 9. London: George Allen and Unwin. Loewe, Michael. 2011. Dong Zhongshu, a “Confucian” Heritage and the Chunqiu fanlu. Leiden: Brill. Malmqvist, Göran. 1971. “Studies on the Gongyang and Guuliang Commentaries.” Bulle­ tin of the Museum of Far Eastern Antiquities 43: 67–222. Malmqvist, Göran. 1978. “What Did the Master Say?” In Ancient China: Studies in Early Civilization, ed. David T. Roy and Tsuen-hsuin Tsien, 137–155. Hong Kong: The Chinese University Press. Nylan, Michael. 2000. “Textual Authority in Pre-Han and Han.” Early China 25: 205–258. Nylan, Michael, and Thomas Wilson. 2010. Lives of Confucius: Civilization’s Greatest Sage through the Ages. New York: Doubleday.

Kongzi As Author In The Han

307

Pines, Yuri. 2009. “Chinese History Writing between the Sacred and the Secular.” In Early Chinese Religion, pt. 1, Shang through Han (1250 BC–220 AD), ed. John Lagerwey and Marc Kalinowski, 315–340. Leiden: Brill. Pu Weizhong 浦衛忠. 1995. Chunqiu sanzhuan zonghe yanjiu 春秋三傳綜合研究. Taipei: Wenjin chubanshe. Puett, Michael J. 2001. The Ambivalence of Creation: Debates Concerning Innovation and Artifice in Early China. Stanford, CA: Stanford University Press. Queen, Sarah A., and John Major. 2016. Luxuriant Gems of the Spring and Autumn, Attributed to Dong Zhongshu. New York: Columbia University Press. Sanguo zhi 三國志. 1987. Beijing: Zhonghua shuju. Sato, Masayuki. 2013. “Did Xunzi’s Theory of Human Nature Provide the Foundation for the Political Thought of Han Fei?” In Dao Companion to the Philosophy of Han Fei, ed. Paul R. Goldin, 147–165. Dordrecht: Springer. Schaab-Hanke, Dorothee. 2002. “Review of Joachim Gentz, Das Gongyang zhuan: Auslegung und Kanonisierung der Frühlings- und Herbstannalen (Chunqiu).” Oriens Extremus 43:275–291. Schaberg, David. 2001a. A Patterned Past: Form and Thought in Early Chinese Histo­ riography. Cambridge, MA: Harvard University Asia Center. Schaberg, David. 2001b. “‘Sell It! Sell It!’: Recent Translations of Lunyu.” Chinese Literature: Essays, Articles, Reviews 23:115–139. Schaberg, David. 2005. “Platitude and Persona: Junzi Comments in the Zuozhuan and Beyond.” In Historical Truth, Historical Criticism, and Ideology: Chinese Historiography and Historical Culture from a New Comparative Perspective, ed. Helwig SchmidtGlintzer et al., 177–196. Leiden: Brill. Shiji 史記. 1959. Beijing: Zhonghua shuju. Su Yu 蘇輿. 1992. Chunqiu fanlu yizheng 春秋繁露義證. Beijing: Zhonghua shuju. Sun Jingtan 孫景壇. 2000. “Dong Zhongshu de Tianren san ce shi Ban Gu de weizuo” 董 仲舒的《天人三策》是班固的僞作. Nanjing shehui kexue 南京社會科學 2000.10: 29–35. Van der Loon, Piet. 1952. “On the Transmission of Kuan-tzu.” T’oung Pao 41: 357–393. Van Ess, Hans. 2015. “Mengzi 3B9 und die Unzufriedenheit mit dem Recht der Han.” Bochumer Jahrbuch zur Ostasienforschung 38:225–243. Wang Liqi 王利器. 1992. Yantie lun jiaozhu 鹽鐵論校注. Beijing: Zhonghua shuju. Wang Xianshen 王先慎. 1998. Han Feizi jijie 韓非子集解. Beijing: Zhonghua shuju. Xiang Zonglu 向宗魯. 1987. Shuiyuan jiaozheng 說苑校證. Beijing: Zhonghua shuju. Yang Bojun 楊伯峻. 1988. Mengzi yizhu 孟子譯註. Beijing: Zhonghua shuju. Yang Bojun 楊伯峻. 1993. Chunqiu Zuozhuan zhu 春秋左傳注. Rev. ed. Beijing: Zhonghua shuju. Zhong Zhaopeng 鐘肇鵬. 2005. Chunqiu fanlu jiaoshi ( jiaobu ben) 春秋繁露校釋(校 補本). Shijiazhuang: Hebei renmin chubanshe.

308

Kern

Index Index

309

Index accretion model of textual formation, 2, 8, 35, 39–55, 64–65, 80–83, 223–224, 269 auspicious signs (see omens) authority, textual, 4, 5, 41, 49–51, 54, 59, 73, 76, 117–118, 123, 191–193, 241, 264–266, 268, 285–286, 289–290, 296, 298, 302–304 Ames, Roger, 2, 34, 121 authorship, 17, 40, 189, 193, 271–303 see also Kongzi 孔子, as author of the Chunqiu 春秋 Bajiaolang 八角廊 (see Dingzhou 定州 tomb texts) Ban Gu 班固, 220, 287 baobian 褒貶, 299 Beijing Olympics, 3 bian shi 辯士, 87 “Biao ji” 表記, 18, 20 Biblical studies, parallels with, 41–42, 81, 206, 218–219 bibliography under the Han dynasty, 71–72, 88, 108, 220–224 see also “Yiwen zhi” 藝文志 Bie lu 別錄, 71, 191 Boltz, William G., 157 book formats, 187–193, 211 Book of Mormon, 108 boshi 博士, 58–63, 68 Bo Yi 伯夷, 301 bronze inscriptions, 103, 194–195 Brooks, E. Bruce and Taeko, 2n, 6, 39, 42, 48–50, 54–55, 80, 83, 228, 269 building blocks, textual, 122, 209 bu yong 不用, 299–301 Calvino, Italo, 290n, 303 Campbell, Alexander, 108 canonization, 2, 5, 7–9, 41, 45, 54–55, 59, 61–62, 64–65, 88, 303–304 ce 冊/策, 190–191 Chao Fulin 晁福林, 32 Chen Dong 陳東, 158, 160, 172–173 Chen Jian 陳劍, 26–27 Chen Liangwu 陳良武, 24 Chen Renxi 陳仁錫, 281, 298

Chen She 陳涉, 61 Chen Tongsheng 陳桐生, 32, 84 Chen Wei 陳偉, 27 Chu 楚 state, 23, 49, 61–62, 93, 200, Chunqiu fanlu 春秋繁露, 275–276, 278, 281, 286, 295 Chunqiu 春秋, 34n, 59, 258–259, 269–304 classicism, 35n, 199, 223, 298, 303n commentary, 8, 30, 34, 42, 73, 78, 96, 104, 163, 247, 251–252, 258–260, 264–265, 271, 279, 285–290, 303 composite nature of early texts, 42, 44–45, 55, 187–188, 211–213, 292 Confucius (see Kongzi 孔子) Confucius Institutes, 3 Copernicus, 102 cosmology, 5, 59, 104 Csikszentmihalyi, Mark, 2n, 4, 7, 67, 70, 92, 109, 116, 146, 181, 247 Cui Shu 崔述, 2, 42–44, 52, 83, 93 Da Dai Liji 大戴禮記, 85n, 224–227, 261, 296 da 達 (see success and failure, debates concerning) dao 道, 104, 141, 178 “Da xue” 大學, 23, 59n destiny (see ming 命) Dickens, Charles, 188 Dingzhou 定州 tomb texts, 154–155 see also Lunyu 論語, Dingzhou 定州 manuscript dizi 弟子, 8, 29–30, 48, 51–53, 87, 92–109, 193–197, 218–238, 271, 280, 282 as compilers of the Lunyu, 18, 29, 32–33, 40–44, 49, 55–60, 64 biographies, 224–235, 241–260 disciple texts, 218–238 Documentary Hypothesis, 41–42, 81 Dong Zhongshu 董仲舒, 68, 109, 275–276, 286, 293, 298 “doubting the past” yi gu 疑古, 17 Du Weiming, 17–18 Duke Ai of Lu 魯哀公, 155, 225, 230, 259, 274, 277, 285, 296 Duke Mu of Lu 魯穆公, 63

© Koninklijke Brill NV, Leiden, 2018 | doi:10.1163/9789004382947_013

310 Duke Mu of Qin 秦穆公, 59 Duke Zhao of Lu 魯昭公, 222, 237, 281–283 Durrant, Stephen W., 8n, 94n, 281 Egami Namio 江上波夫, 165 emotion, 119, 275, 278, 293, 297 Emperor Ai (Han) 漢哀帝, 220 Emperor Cheng (Han) 漢成帝, 4n, 70, 220 Emperor Jing (Han) 漢景帝, 68, 70, 72 Emperor Wen (Han) 漢文帝, 68, 70n, 85–87 Emperor Wu (Han) 漢武帝, 68, 70–73, 76, 81, 86, 88, 222, 274, 286–287, 298, 301 Emperor Xuan (Han) 漢宣帝, 29, 70, 180, 287, 301 Emperor Yuan (Han) 漢元帝, 70, 86 Emperor Zhao (Han) 漢昭帝, 70, 222 Empress Lü 呂太后 (Han), 62 encountering [the right circumstances], see yu 遇 encyclopedism, 9 Excellency Yu 虞卿, 299 “explaining the past” (shigu 釋古), 17 extension, textual, 197–198 “Fang ji” 坊記, 18–24, 57, 63, 68, 221n fate (see ming 命) Fayan 法言, 29 feng 封 and shan 禪 sacrifices, 261–262 Fenghuangshan 鳳凰山 tomb texts, 177–178 filial piety (see xiao 孝) Five Classics, 1, 5, 8, 68n, 223, 251, 255, 258, 263, 265, 286, 290, 294, 297–299, 301, 304 fluidity, textual, 77, 159, 187–188, 190–193, 198, 201, 212–213, 237–238, 277 frustration, 12, 141n, 281, 290–293 Foucault, Michel, 284, 290 Fu Buqi 宓不齊, 224 Fu Sheng 伏勝, 59, 62 Gentz, Joachim, 282 Golden Rule, 103, 234, 236 Goldin, Paul, 33–34, 88, 121, 276 Gongbo Liao 公伯寮, 230 Gongshu Ban 公輸般, 93 Gongsun Chou 公孫丑, 98 Gongsun Hong 公孫弘, 293 Gongxi Hua 公西華, 86

Index Gongyang zhuan 公羊傳, 124n, 271–284, 288–290, 293, 296, 298, 301 government, 25, 31, 58, 85–86, 108, 125, 128, 131–132, 138, 142–145, 152, 181, 212, 234, 236, 297 Graham, A.C., 95, 101, 120–121 Gu Jiegang 顧頡剛, 123 Guanzi 管子, 63, 257 Guliang zhuan 穀梁傳, 259n, 279, 298, 301 Guo Yi 郭沂, 19, 32 Guodian 郭店 tomb texts Lu Mu Gong wen Zisi 魯穆公問子思, 18 Taiyi sheng shui 太一生水, 200 Wuxing 五行, 18 Xing zi ming chu 性自命出, 101, 198 Yu cong 語叢, 20, 22–24, 60, 200–201, 204, 210, 288–290 Zhongxin zhi dao 忠信之道, 24, 30 Zi yi 緇衣 (see Zi yi 緇衣) Zun deyi 尊德義, 56n Guoyu 國語, 251, 280 Han Feizi 韓非子, 107, 271–272, 285, 291, 296, 303 Han Ying 韓嬰, 68–69 Han Yu 韓愈, 17, 68n Hanshi waizhuan 韓詩外傳, 68–70, 86–87 Hanshu 漢書, 19, 29, 70–72, 108, 171, 191, 220, 223–227, 235, 238, 251, 282, 287, 301 Haupt, Christiane, 4, 67 Heaven (see tian 天) Hejian Xian wang dui shangxia san yonggong 河間獻王對上下三雍宮, 223 Hirayama Ikuo 平山郁夫, 165 historian, 11, 18, 35, 154, 242–244, 290–292 see also Sima Qian 司馬遷 historical meaning, 1n, 35, 188 history, writing of, 265–266, 274–275, 281–285, 299–300 Huainanzi 淮南子, 73, 287, 293–294, 298 Hu Sui 壺遂, 299–301 Hu Yin 胡寅, 42, 52 Huang Ren’er 黃人二, 29 Huan Kuan 桓寬, 287 Huan Tui 桓魋, 230 Hunter, Michael, 122, 173, 241, 257, 280, 298 Huwu sheng 胡毋生, 293

Index intellectual history (sixiang shi 思想史), 2, 7, 17, 57, 98, 104, 116–117, 121–124, 128, 137, 146–147, 270, 302 interlocutor texts, 219, 224–238 intertextuality, 18–32, 72–79, 81–82, 123–135, 209–214, 225–238, 242–257, 277–278 Itō Jinsai 伊藤仁斎, 2, 42–44, 83 Itō Toshimitsu 伊藤利光, 165 Jia Yi 賈誼, 109, 287, 296 jing 經, 105 “Ji tong” 祭統, 220n Jixia 稷下, 60n, 63 “Ji yi” 祭義, 226–227 juan 卷, 2, 64, 190–191 junzi 君子, 25–28, 76, 85, 102, 118–123, 132–145, 226, 228, 272–273, 279–280, 282–283, 286, 294 Kanaya Osamu 金谷治, 57–58 Kaneto Mamoru金戸守, 4, 67, 242–243 Karlgren, Bernhard, 6, 102 Kern, Martin, 34n, 228 Kim Kyŏng-ho, 153n, 166–167, 170–171, 181 Kimura Eiichi 木村英一, 43, 46–47, 51–54, 80, 83, 269 Kong congzi 孔叢子, 20, 49n, 224 Kong Jia 孔甲, 61 Kongzi dizi mulu 孔子弟子目錄, 227 Kongzi jiayu 孔子家語, 108, 196, 225, 227–229, 234, 246–247, 296 Kongzi turen tufa 孔子徒人圖法, 227 Kongzi 孔子 as “uncrowned king” (su wang 素王), 275, 287, 297–298 as author of the Chunqiu 春秋, 34n, 269–304 “between Chen and Cai” (Chen Cai zhi jian 陳蔡之間), 78, 126, 129–135, 140–141, 253–254, 290–292 biography, 123, 241–264, 273–276, 281, 297, 301 designations for, 52, 164, 170–171, 179, 277, 279, 282–283 disciples (see dizi 弟子) interlocutors, 78, 219, 224–238 Kuhn, Thomas S., 102 Lau, D.C., 93

311 Laozi, 79, 92, 103–105, 200, 210 layouts, manuscript, 207–213 learning (see xue 學) Lelang 樂浪 Lunyu (see Lunyu 論語, P’yŏngyang Lunyu) Lewis, Clarence Irving, 102 Lewis, Mark Edward, 193–194, 271–272, 281 li 禮, 1, 47n, 52, 53n, 109, 121, 147, 218, 251, 259, 266, 288–290, 297, 301 Liang Tao 梁濤, 18–19 Lidian tu 禮殿圖, 227 Liezi 列子, 108–109 Liji 禮記, 18–20, 23–24, 33, 47n, 57, 60, 63, 68, 73, 208–209, 224–226, 261, 296 Li Ling 李零, 31n, 32 lin 麟, 274–278, 291–292, 295–297, 304 Li Ruohui 李若暉, 172 Li sao 離騷, 290 Liu Bang 劉邦, 61–62, 172, 174, 179 Liu 劉 clan, 87 Liu He 劉賀, 70, 152n Liu Ju 劉據, 301 Liu Laicheng 劉來成, 156, 161, 171 Liu Qu 劉去, 70 Liu Xiang 劉向, 4n, 8, 40, 71, 79, 191, 270, 276, 294, 297 Liu Xin 劉歆, 71, 191, 270 Liu Xiu 劉脩, 70, 154 Liu yi 六藝, 1, 67n, 220, 286–289, 298 Liu Yu 劉餘, 71–72 Li Xueqin 李學勤, 154, 171 Li Yiji 酈食其, 61 Loewe, Michael, 68n, 181, 287 Lü Buwei 呂不韋, 59–60, 63, 291 lun 論, 5, 29, 32, 59, 221, 294 Lundahl, Bertil, 107n, 174 Lunheng 論衡, 29, 44, 67n, 70, 221, 238, 299 Lunyu 論語 Dingzhou 定州 manuscript, 17, 64, 80–81, 152–181 gu Lun 古論, 29–30, 171–172 P’yŏngyang Lunyu manuscript, 152, 164–181, 211 proto-Lunyu, 18, 21, 30, 32, 72, 243–244, 255 Qi Lun 齊論, 171 Rongo 論語 studies, 4, 44, 45, 46 title, 19, 29–30, 63, 67–70, 220–224, 242, 245–246, 268, 303

312 Lunyu 論語 (cont.) Upper 上 and lower 下 Lunyu, 42, 52, 53, 54, 57, 63 Zhang Hou Lun 張侯論, 29 see also accretion model of textual formation Lüshi Chunqiu 呂氏春秋, 59, 78–79, 96, 125, 291 Ma Yumeng 馬玉萌, 162–163 Maeder, Erik, 190n, 192, 201 Makeham, John, 1n, 2, 4, 67, 92, 129, 173, 246 manuscript formats, 187–212 Marcus Aurelius, 218 masters (see zhuzi 諸子) Mawangdui 馬王堆 tomb texts, 178, 199 Jun zheng 君正, 31 Meng Ke 孟軻, 17, 98, 100–106, 125, 259 Mengzi 孟子, 98–105, 128–130, 252 and the Lunyu, 20–22, 44, 75–77, 83, 92 on success and failure, 128–137, 141–145 and Kongzi’s authorship of the Chunqiu, 271–299 ming 命, 126, 129, 133–134, 138, 144–145, 201, 206, 262–263, 275, 291, 297–298, 301 mise-en-page, 201–212 Mo Di 墨翟, 81n, 93–106 Mohist Analects, 93–100 Mozi 墨子, 59, 78, 81n, 92–106, 144 music (see yue 樂) Neiye 內業, 98 Neo-Confucianism, 104, 124n Nylan, Michael, 2, 72n, 181, 286 official service, 1, 73, 86–87, 128, 138, 142, 250, 259 omens, 23, 262–263, 275–278 Pascal, Blaise, 218 philosophy, 1–2, 7, 17, 34–35, 88, 92–94, 99, 101–105, 109, 116, 122, 136, 199, 248, 269, 293, 298 pian 篇, 18, 190–191, 233 Plato, 105 poverty, 129 Puett, Michael J., 104 Qi lüe 七略, 71n

Index Qin First Emperor 秦始皇, 87, 291 Qin 秦 Dynasty, 5, 7, 9, 19, 33, 39, 55, 57–65, 87, 94, 125, 135, 143, 146, 211–212, 219, 238, 268, 271, 291, 298 Qinghua University 清華大學 manuscript corpus, 56n, 206, 208 qing 情, 101 Qin Guli 禽滑釐, 94 qiong 窮 (see success and failure, debates concerning) qualia, 102 Qu Yuan 屈原, 290, 293 Ran You 冉有, 86 ren 仁, 2, 53n, 76–77, 85, 118–122, 130–131, 141–142, 147 Richter, Matthias, 85, 187 ritual (see li 禮) Rosemont, Henry, 2, 34, 121 Ru 儒, 17, 48, 55, 58–63, 131–132, 141, 155, 261, 298 Scarpari, Maurizio, 189–194, 201 scriptural meaning, 35 selection as a mode of text formation, 5, 8, 32, 51, 84n, 118, 121, 297 Shan Chengbin 單承彬, 166, 171 Shanghai Museum manuscript corpus, 17–19, 24–32, 57, 76, 78, 83–84, 125, 156, 199, 201, 211, 224–225, 237, 289 Cong zheng 從政, 25–27, 30 Dizi wen 弟子問, 56, 60, 78, 138, 211 Guishen zhi ming 鬼神之明, 125–128, 137n, 144 Heng xian 恆先, 201 Jian da wang po han 柬大王泊旱, 199 Junzi wei li 君子為禮, 28–29, 56, 60, 76–78, 211, 237 Kongzi shilun 孔子詩論, 32, 78, 289–290 Min zhi fumu 民之父母, 78, 81 Neili 內禮, 225–226 Qiongda yi shi 窮達以時, 24, 30, 125–137, 140, 144 Rong Cheng shi 容成氏, 198 San de 三徳, 137n Wu ming 吳命, 201 Zi yi 緇衣 (see Zi yi 緇衣) Zhonggong 仲弓, 30–32, 78, 83–84, 87

Index Shaughnessy, Edward L., 208 Shen Dao 慎到, 103, 105 Shenzi 慎子, 103 Shen Yue 沈約, 18–19 Shennong 神農, 99–100 Shenzi 慎子, 103 Shi Dan 史丹, 70 shi 時, 24, 125–128, 131–138, 144–145 Shi 詩, 69, 228n, 283, 288–291, 294 Shiji 史記 and Kongzi’s 孔子 interlocutors, 224–236 and the Lunyu 論語, 8, 29, 71–74, 241–266 and the Qin 秦 dynasty, 58–64 chap. 27, “Tianguan shu” 天官書, 263 chap. 28, “Fengshan shu” 封禪書, 58n, 263 chap. 46, “Tian Jingzhong Wan shijia” 田 敬仲完世家, 264 chap. 47, “Kongzi shijia” 孔子世家, 18, 49n, 228n, 241–243, 258n, 251–257, 273 chap. 49, “Waiqi shijia” 外戚世家” 263 chap. 61, “Bo Yi liezhuan” 伯夷列傳, 274n, 301 chap. 67, “Zhongni dizi liezhuan” 仲尼弟 子列傳, 29–30, 69, 224, 228, 233n, 243 chap. 121, “Rulin liezhuan” 儒林列傳, 278, 291 chap. 124, “Youxia liezhuan” 游俠列傳, 248 chap. 129, “Huozhi liezhuan” 貨殖列傳, 249 chap. 130, “Taishi gong zixu” 太史公自序, 255, 299, 297n Taishigong yue 太史公曰 comments, 244–245, 278, 293, 299 Shizi 尸子, 103 shu 恕, 103–104 Shu 書, 59, 62, 103, 125, 141, 206, 260, 288, 298 Shuihudi 睡虎地 tomb texts, 203n, 211–212 Wei li zhi dao 為吏之道, 211 Shuiyuan 說苑, 40, 73n, 79, 86–87, 230, 234, 276, 294–296 Shun 舜, 99, 126–132, 144, 279 Shuowen jiezi 說文解字, 8n 218n Shusun Tong 叔孫通, 61–62, 223 Sima Niu 司馬牛, 235–236 Shiji, 8 Sima Qian 司馬遷 and the Lunyu, 8, 64, 70, 123, 241–266,

313 and the Chunqiu 春秋, 274–278, 281–305 see also Shiji 史記 Sima Tan 司馬談, 255n, 265, 301 Sima Zhen 司馬貞, 227n, 252 Si-Meng school 思孟學派, 17–19, 24 Song 宋 state, 93 Song Keng 宋牼, 106 Stone Classics of the Xiping era 熹平石經, 152 structure, textual, 80–84, 187–214, 246–248 see also accretion model of textual formation success and failure, debates concerning, 123–146 suffering, 97, 126, 131–133, 266, 290–294, 299–301 Sunzi 孫子, 103, 105, 290 tabooed names, 172–175 Takeuchi Yoshio 武内義雄, 4, 43–44, 46–47, 54, 64 Tang Minggui 唐明貴, 172–173 Tangshan 唐山 earthquake, 154, 157, 160n tian 天, 104, 109, 127–137, 142–145, 262–263, 273–283 timeliness (see shi 時) tong 通, 58, 135, 165 true king, 273–276 Tsuda Sōkichi 津田左右吉, 4, 43–47, 50–54, 67, 80 Tsuruma Kazuyuki 鶴間和幸, 165–168 tutor, office of, 70–71, 109, 155, 180, 301 Waley, Arthur, 2, 42–44, 121, 123 Wang Chong 王充, 32, 44, 64, 70–71, 199, 221, 223, 227, 299 Wang Su 王肅, 108, 227, 246n Wangjiatai 王家臺 tomb texts, 212–213 Zheng shi zhi chang 政事之常, 212–213 wealth, 120, 128, 131, 134, 137, 141, 145, 147, 249–250 Weingarten, Oliver, 4, 67, 164 Wen Weng 文翁, 227 wu de 五德, 25–27 Wumazi 巫馬子, 95–97, wu shan 五善, 26 Wuwei 武威 tomb texts, 199n, 207–208 Wu Zixu 伍子胥, 126–127, 133, 144

314 Xi Jinping 習近平, 3 Xiang Yu 項羽, 61–62 Xiao Wangzhi 蕭望之, 156, 180 xiao 孝, 1, 5, 78, 141, 226–228, 230–231, 234–235, 243, 257, 266 Xiaojing 孝經, 5, 70, 226, 243 xing 性, 101–102 Xinshu 新書, 296 Xinyu 新語, 296 Xu Shaohua 徐少華, 29 Xu Xing 許行, 100–101 xue 學, 1, 5n, 17, 18, 58, 60, 133, 286, 289 Xun Kuang 荀況, 62n, 63, 125, 271–272 Xunzi 荀子, 45, 92, 101–106, 131–135, 138, 141, 144–145, 224–225, 271–272 Yan Hui 顏回, 28, 76–77, 86–87, 196n, 224, 233–236, 244, 275, 277 Yang Chaoming 楊朝明, 20, 22, 32, 229n Yang He 楊何, 265 Yangism, 95–96 Yang Xiong 揚雄, 296 Yang Zhu 楊朱, 96 Yantie lun 鹽鐵論, 287 Yanzi chunqiu 晏子春秋, 257 Yao 堯, 127, 129, 132, 279 yi 義, 141, 147, 283 Yi 易, 70, 78, 204, 206, 264–265, 283, 288, 290 Yili 儀禮, 199n, 207–208 Yinqueshan 銀雀山 tomb texts, 177–178 yinyang 陰陽, 98 Yi Sŏng-shi, 153n, 166 “Yiwen zhi” 藝文志 bibliography, 19, 70–72, 282 yu 遇, 50, 83, 100, 125–138, 197, 300 Yu Dan 于丹, 3 yue 樂 , 52, 218, 252n Yue 樂, 228n, 288 Yueji 樂記, 223 Yuelu Academy嶽麓書院 manuscript corpus, 211–212 Wei li zhi guan ji qianshou 為吏治官及 黔首, 211 Yuezheng Zichun 樂正子春, 226

Index Yun Yong-gu, 166 Yushi chunqiu 虞氏春秋, 299 Zai Yu 宰予, 247n, 260–262 Zeng Can 曾參, 103 Zeng Shen 曾參, 40, 44, 48, 51, 63, 224, 230, 243 Zengzi 曾子, 17, 22, 85n, 224–238, 243 zhang 章, 2, 20, 46–47, 214, 218 Zhang Chang 張敞, 288 Zhangjiashan 張家山 tomb texts, 154 Zhao Qi 趙岐, 70n, 77 Zhao Zhenxin 趙貞信, 4, 67, 173n Zheng Jiewen 鄭傑文, 93–94 Zheng Xuan 鄭玄, 70n, 108, 227 zhi 知 (recognition), 84–87, 125–146, 273–283, 292–297 zhong 忠, 1, 31, 103 Zhong Gong 仲弓, 30– 32, 78, 83–84, 87, 224, 233–236 Zhongshan 中山 kingdom, 153–154, 174, 179–180 “Zhong yong” 中庸, 18–20, 23, 59n Zhong You 仲由, 231–232 Zhou Fengwu 周鳳五, 24, 26 Zhou zhi 周制, 223 Zhu Weizheng 朱維錚, 4, 67, 92 Zhuangzi 莊子, 63, 78, 97n, 103, 105, 107, 271–272 zhuzi 諸子, 2, 55, 60–63, 73, 83, 105, 268, 289 Zigao 子羔, 32, 224 Zigong 子貢, 44, 48–49, 52, 86, 101, 143, 236, 238, 244, 249–250, 253–254, 261, 277 Zilu 子路, 86, 133, 140, 142, 170, 196, 230–232, 238, 253–254, 275 Zisi 子思, 17–20, 23, 49, 57, 63, 224, 248–250 Zisizi 子思子, 18–19, 22–23 Zixia 子夏, 196, 235–236 Zi yi 緇衣, 18–20, 57, 82, 124, 205–210, 213, 221n Ziyou 子游, 101n, 235 Zizhang 子張, 25–26, 140, 194–197, 206, 236 Zou Yan 鄒衍, 259 Zuo Qiuming左丘明, 290 Zuozhuan 左傳, 72n, 78, 103, 105, 125, 230, 237, 264, 277, 279, 285, 288, 290, 298