Critical Readings on Tang China [Volume 1] 9004380159, 9789004380158

Table of contents :
Contents
Tang Emperors’ Accession Dates and Reign Titles (nianhao 年號)
General Introduction
History—Political, Intellectual, and Military
Wên Ta-ya: The First Recorder of T’ang History
The Rise to Power of the T’ang Dynasty: A Reassessment
The T’ang Imperial Family
Canonical Scholarship
Neo-Confucianism and Neo-Legalism in T’ang Intellectual Life, 755–805
The Structure of Tang Selection
Decree Examinations in T’ang China
The Bureaucratic Apparatus [of T’ang Historians]
Bureaucracy and Cosmology: The Ritual Code of T’ang China
Wei Cheng’s Thought [esp. Regarding Government]
Imperial Power and the Ruling Class [under Empress Wu]
The Chou Dynasty [of Empress Wu]
The Career of Yang Kuei-fei
The Flight from the Capital and the Death of Precious Consort Yang
Foreign Policy
The An Lu-shan Rebellion and the Origins of Chronic Militarism in Late T’ang China

Citation preview

Critical Readings on Tang China Volume 1

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

Critical Readings on Tang China volume 1

Edited by

Paul W. Kroll

LEIDEN | BOSTON

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

The Library of Congress Cataloging-in-Publication Data is available online at http://catalog.loc.gov Library of Congress Control Number: 2018962592

Typeface for the Latin, Greek, and Cyrillic scripts: “Brill”. See and download: brill.com/brill-typeface isbn 978-90-04-28113-4 (hardback, set) isbn 978-90-04-28169-1 (hardback, vol. 1) isbn 978-90-04-28168-4 (hardback, vol. 2) isbn 978-90-04-28167-7 (hardback, vol. 3) isbn 978-90-04-28166-0 (hardback, vol. 4) isbn 978-90-04-38015-8 (e-book, vol. 1) isbn 978-90-04-38016-5 (e-book, vol. 2) isbn 978-90-04-38019-6 (e-book, vol. 3) isbn 978-90-04-38020-2 (e-book, vol. 4) Copyright 2019 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.

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

Contents Volume 1 Tang Emperors’ Accession Dates and Reign Titles (nianhao 年號) xi General Introduction 1 Paul W. Kroll

History—Political, Intellectual, and Military 1 Wen Ta-ya: The First Recorder of T’ang History 11 Woodbridge Bingham 2 The Rise to Power of the T’ang Dynasty: A Reassessment 17 Howard J. Wechsler 3 The T’ang Imperial Family 41 Denis Twitchett 4 Canonical Scholarship 100 David McMullen 5 Neo-Confucianism and Neo-Legalism in T’ang Intellectual Life, 755–805 163 Edwin G. Pulleyblank 6 The Structure of T’ang Selection 214 P. A. Herbert 7 Decree Examinations in T’ang China 237 P. A. Herbert

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

vi

CONTENTS

8 The Bureaucratic Apparatus [of T’ang Historians] 267 Denis Twitchett 9 Bureaucracy and Cosmology: The Ritual Code of T’ang China 295 David McMullen 10

Wei Cheng’s Thought [esp. Regarding Government] 346 Howard J. Wechsler

11

Imperial Power and the Ruling Class [under Empress Wu] 367 Richard W. L. Guisso

12

The Chou Dynasty [of Empress Wu] 404 Richard W. L. Guisso

13

The Career of Yang Kuei-fei 455 Howard S. Levy

14

The Flight from the Capital and the Death of Precious Consort Yang 482 Paul W. Kroll

15

Foreign Policy 503 P. A. Herbert

16

The An Lu-shan Rebellion and the Origins of Chronic Militarism in Late T’ang China 518 Edwin G. Pulleyblank

Volume 2 Literature and Cultural History 17

T’ang Literati: A Composite Biography 545 Hans H. Frankel

18

Transparencies: Reading the T’ang Lyric 566 Stephen Owen

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

vii

CONTENTS

19

The Significance of the fu in the History of T’ang Poetry 584 Paul W. Kroll

20 An Offering to the Prince: Wang Bo’s Apology for Poetry 597 Ding Xiang Warner 21

Tamed Kite and Stranded Fish: Interference and Apology in Lu Chao-lin’s fu 636 Paul W. Kroll

22

A Re-evaluation of Chen Ziang’s “Manifesto of a Poetic Reform” 666 Timothy Wai Keung Chan

23

On Li Po 694 Elling O. Eide

24 Li Po’s Letters in Pursuit of Political Patronage 731 Victor H. Mair 25 Li Bai’s “Rhapsody on the Hall of Light”: A Singular Vision of Cosmic Order 762 Nicholas Morrow Williams 26 Tu Fu 825 Stephen Owen 27

Tu Fu’s Social Conscience: Compassion and Topicality in his Poetry 894 Shan Chou

28 Poems in Their Place: Collections and Canons in Early Chinese Literature 936 Pauline Yu 29 Heyue yingling ji and the Attributes of High Tang Poetry 967 Paul W. Kroll 30 The Formation of the Tang Estate Poem 1001 Stephen Owen

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

viii

CONTENTS

Volume 3 31

Lexical Landscapes and Textual Mountains in the High T’ang 1021 Paul W. Kroll

32 Historical and Literary Theory in the Mid-Eighth Century 1060 David McMullen 33 The Manuscript Legacy of the Tang: The Case of Literature 1096 Stephen Owen 34 Literary Collections in Tang Dynasty China 1126 Christopher M. B. Nugent 35 A Study of the Jinglong Wenguan Ji 1171 Jia Jinhua 36 The Old-Style fu of Han Yu 1204 David R. Knechtges 37

Another Go at the Mao Ying chuan 1230 Elling O. Eide

38 The Old Men [of the Early Ninth Century] 1235 Stephen Owen 39 The Inscription of Emotion in Mid-Tang Collegial Letters 1281 Anna M. Shields 40 Yüan Chen and ‘The Story of Ying-ying’ 1326 James R. Hightower 41

Nostalgia and History in Mid-Ninth-Century Verse: Cheng Yü’s Poem on “The Chin-yang Gate” 1360 Paul W. Kroll

42 Remembering Kaiyuan and Tianbao: The Construction of Mosaic Memory in Medieval Historical Miscellanies 1441 Manling Luo

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

ix

CONTENTS

43 The Dancing Horses of T’ang 1476 Paul W. Kroll 44 Falconry in T’ang Times 1504 Edward H. Schafer 45 Public Values in Calligraphy and Orthography in the Tang Dynasty 1540 Amy McNair

Volume 4 Religion 46 The Role of Buddhist Monasteries in T’ang Society 1559 Kenneth K. S. Ch’en 47 Buddhism and Education in T’ang Times 1580 Erik Zürcher 48 Imperial Patronage in the Formation of T’ang Buddhism 1622 Stanley Weinstein 49 Stūpa, Sūtra, Śarīra in China, c. 656–706 CE 1663 T. H. Barrett 50 The Birth of a Patriarch: The Biography of Hui-neng 1714 Philip B. Yampolsky 51

Metropolitan Chan: Imperial Patronage and the Chan Style 1743 John McRae

52 Time after Time: Taoist Apocalyptic History and the Founding of the T’ang Dynasty 1776 Stephen R. Bokenkamp 53 Taoist Ordination Ranks in the Tun-huang Manuscripts 1805 Kristofer M. Schipper

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

x

CONTENTS

54 Taoism in the T’ien-pao Era, 742–56 1829 T. H. Barrett 55 Li Po’s Transcendent Diction 1839 Paul W. Kroll 56 Immortality Can be Studied 1875 Jan De Meyer 57

The Worshippers of Mount Hua 1915 Glen Dudbridge Index of Personal Names 1949

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

Tang Emperors’ Accession Dates and Reign Titles (nianhao 年號) Lefthand column is Western year in which a particular reign-title was adopted. Ruler’s accession date (in parenthesis) follows name. Note that sometimes a ruler’s first new reign-title was not adopted until months after his actual accession or in the following year; e.g. Tang Taizong’s reign actually began in 626, not 627 when for the first time he proclaimed a new reign-title. Column following a reign-title gives the Chinese month and day of that reign-title’s adoption; righthand column gives corresponding Western month and day. 618 627 650 656 661 664 666 668 670 674 676 679 680 681 682 683 684 684 684 685 689 689

Gaozu 高祖 (18 Jne, 618) Taizong 太宗 (4 Sep, 626) Gaozong 高宗 (15 Jly, 649)

Zhongzong 中宗 (3 Jan, 684) Ruizong 睿宗 (27 Feb, 684)

Wude 武德 Zhenguan 貞觀 Yonghui 永徽 Xianqing 顯慶 Longshuo 龍朔 Linde 麟德 Qianfeng 乾封 Zongzhang 總章 Xianheng 咸亨 Shangyuan 上元 Yifeng 儀鳳 Tiaolu 調露 Yonglong 永隆 Kaiyao 開耀 Yongchun 永淳 Hongdao 弘道 Sisheng 嗣聖 Wenming 文明 Guangzhai 光宅 Chuigong 垂供 Yongchang 永昌 Zaichu 載初

v.20 i.1 i.1 i.7 ii.30 i.1 i.5 iii.6 iii.1 viii.15 xi.8 vi.3 ix.23 ix.30 ii.19 xii.4 i.1 ii.6 ix.6 i.1 i.1 i.1*

Jne 18 Jan 23 Feb 7 Feb 7 Apr 4 Feb 2 Feb 14 Apr 22 Mar 27 Sep 20 Dec 18 Jly 15 Sep 22 Nov 15 Apr 2 Dec 27 Jan 23 Feb 27 Oct 19 Feb 9 Feb 27 Dec 18

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

xii

Tang Emperors ’ Accession Dates and Reign Titles

Zhou 周 interregnum (690-705) 690 692 692 694 694 695 696 696 697 697 700 701 701

Wu Zetian 武則天 (16 Oct)

ix.9*

Oct 16

Ruyi 如意 Changshou 長壽 Yanzai 延載 Zhengsheng 證聖 Tiance wansui 天冊萬歲 Wansui dengfeng 萬歲登封 Wansui tongtian 萬歲通天 Shengong 神功 Shengli 聖曆 Jiushi 久視 Dazu 大足 Chang’an 長安

iv.1* ix.9* v.11* i.1* ix.9* la.20* iii.16* ix.9* i.1* v.5* i.3 x.22

Apr 22 Oct 23 Jne 9 Nov 23 Oct 22 Jan 20 Apr 22 Sep 28 Dec 20 May 27 Feb 15 Nov 26

* From Zaichu through Jiushi, old Zhou calendar employed, with usual 1th month designated as zhengyue 正月 and 12th month as layue 臘月.)

705 707 710 710 712 712 712 713 742 756 758 760 762 763 765 766 780 784

Tianshou 天授

Restored Tang Zhongzong 中宗 (23 Feb, 705) Shaodi 少帝 (5 Jly, 710) Ruizong 睿宗 (25 Jly, 710) Xuanzong 玄宗 (8 Sep, 712) Suzong 肅宗 (12 Aug, 756)

Daizong 代宗 (18 May, 762) Dezong 德宗 (12 Jne, 779)

Shenlong 神龍 Jinglong 景龍 Tanglong 唐隆 Jingyun 景雲 Taiji 太極 Yanhe 延和 Xiantian 先天 Kaiyuan 開元 Tianbao 天寶 Zhide 至德 Qianyuan 乾元 Shangyuan 上元 Baoying 寶應 Guangde 廣德 Yongtai 永泰 Dali 大曆 Jianzhong 建中 Xingyuan 興元

i.1 ix.5 vi.4 vii.20 i.15 v.13 viii.7 xii.1 i.1 viii.12 ii.5 iv.19 iv.15 vii.11 i.1 xi.12 i.1 i.1

Jan 30 Oct 5 Jly 5 Aug 19 Mar 1 Jne 21 Sep 12 Dec 22 Feb 10 Aug 12 Mar 18 Jne 7 May 13 Aug 24 Jan 26 Dec 18 Feb 11 Jan 27

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

xiii

Tang Emperors ’ Accession Dates and Reign Titles

785 805 806 821 825 827 837 841 847 860 874 880 881 885 888 889 890 892 894 898 901 904 904 907

Shunzong 順宗 (28 Feb, 805) Xianzong 憲宗 (5 Sep, 805) Muzong 穆宗 (20 Feb, 820) Jingzong 敬宗 (29 Feb, 824) Wenzong 文宗 (13 Jan, 827) Wuzong 武宗 (20 Feb, 840) Xuanzong 宣宗 (25 Apr, 846) Yizong 懿宗 (13 Sep, 859) Xizong 僖宗 (15 Aug, 873)

Zhaozong 昭宗 (22 Apr, 888)

Aidi 哀帝 (26 Oct, 904)

Zhenyuan 貞元 Yongzhen 永貞 Yuanhe 元和 Changqing 長慶 Baoli 寶曆 Taihe 太和 Kaicheng 開成 Huichang 會昌 Dazhong 大中 Xiantong 咸通 Qianfu 乾符 Guangming 廣明 Zhonghe 中和 Guangqi 光啟 Wende 文德 Longji 隆紀 Dashun 大順 Jingfu 景福 Qianning 乾寧 Guanghua 光化 Tianfu 天復 Tianyou 天佑

i.1 viii.5 i.2 i.4 i.7 ii.13 i.1 i.9 i.17 xi.12 xi.5 i.1 vii.11 iii.14 ii.22 i.1 i.1 i.21 i.1 viii.27 iv.25 iv.11

Feb 14 Sep 1 Jan 25 Feb 9 Jan 29 Mar 14 Jan 22 Feb 4 Feb 6 Dec 17 Dec 17 Feb 14 Aug 9 Apr 2 Apr 7 Feb 4 Jan 25 Feb 22 Feb 10 Sep 16 May 16 May 28

end of Tang

iv.22

Jne 5

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

General Introduction The fifty-seven articles and book chapters reprinted in this four-volume set of Critical Readings on Tang China aim to give a broadly representative survey of English-language scholarship on the Tang from the past three generations. The Tang 唐 dynasty, formally founded on 18 June 618 and officially terminated on 5 June 907, was the second great imperium of Chinese history (the first being the Han 漢 dynasty of 202 BCE to 220 CE) and, in the opinion of many people, its three centuries were the cultural and political high point of traditional China. During this time Tang China was the most advanced civilization in the world, as well as being the most extensive territorial empire. At its fullest strength the Tang controlled all of what we think of today as “China proper” and also exercised administrative suzerainty westward over most of central Asia (excepting Tibet), including the Tarim Basin and Taklamakan Desert and across the Pamir range to the Hindu Kush in present-day Afghanistan— domains once known in the West by the romantic names of Dzungaria, Sogdiana, Ferghana, and Transoxiana. To the southeast the Tang empire reached far into what we now know as northern Vietnam, the former Annam. And eastward during these centuries, Tang China was the dominant cultural influence over the newly developing states of Korea and Japan. Immediately preceding the Tang, the Sui 隋 dynasty, established in north China in 581 as successor to the Northern Zhou 北周 (557–581), conquered eight years later the Chen 陳 dynasty (557–589) that was the last of six southern dynasties centered on Jiankang 建康 (present-day Nanjing) after the demise of the Han. The Sui thus became the first dynasty in nearly three centuries to reunite all of China under a single rule. But the Sui did not last, being replaced in 618 by the Tang. Thus the short-lived Sui stands in relationship to the Tang rather like the brief but preparatory Qin 秦 (221–206 BCE) in its relationship to the following and longer-lasting Han. The vast empire of the Tang was administratively organized as a collection of roughly 350 prefectures (zhou 州) or commanderies ( jun 郡) and, under those more than 1200 districts (xian 縣) of roughly county size. There were also higher-level officials, often called “commissioners” (shi 使) with various specific titles and duties of supervision coordinating designated matters of several prefectures or larger areas. On the margins of the empire were military protectorates holding more or less loose control over various “foreign” areas and populations. The chief officials of the far-flung local domains of the empire were centrally appointed by the court and sent out from the capital for

© koninklijke brill nv, leiden, ���9 | doi:��.��63/9789004380158_002

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

2

General Introduction

set terms, ideally to be evaluated periodically and then promoted or demoted, as deserved, to posts in other locations, thus diminishing the possibility that any individual official might build up a private sphere of power. Of course the most desirable postings were always those in the capital city of Chang’an 長安 (near present-day Xi’an 西安) or, failing that, in the secondary capital at Luoyang 洛陽. The Tang capital of Chang’an was a walled city comprising approximately thirty square miles that had been built from the ground up with the founding of the Sui in the early 580s. It was located southeast of the site of several previous capitals called Chang’an, going back to pre-imperial times. This new city, which the Sui named Daxingcheng 大興城, was laid out on a grid pattern, comprising one hundred and six separate walled neighborhoods as well as a large bureaucratic compound and an even larger palace enclave for the imperial family. When the Tang took it over, they applied to it the historic name Chang’an and would continue to develop it throughout the dynasty. At its height it accommodated a million people within its walls, with another million in the near suburbs—an urban-centered population that would only begin to be approached elsewhere in the world by Baghdad in the mid-ninth century and Córdoba in the tenth—and which included large representations of traders and shopkeepers, monks and divines, musicians and entertainers from all corners and regions of Asia. The “foreign” element, particularly from the widespread oasis cities and settlements of central Asia, contributed to Chang’an’s truly cosmopolitan character. The dynasty’s secondary capital of Luoyang, where the court was in residence at various times, was not as large or as exactingly laid out but was only slightly less grand than Chang’an. The other great cities of the realm included especially Jinling 金陵 in the Yangzi delta, which (as Jiankang) had been the capital of a succession of “southern” dynasties during the four centuries between the Han and the Tang, and Yangzhou 揚州 (also called, as of old, Guangling 廣陵), slightly downstream from Jinling and on the opposite shore of the Long River (Changjiang 長江, as the Yangzi was then called). There were also dozens of other cities and hundreds of small towns that constituted the functional map of Tang imperial governance. The greatness and legacy of the Tang can be seen in many areas. In terms of political and institutional history, the forms and practices of administration during the Tang remained (with some alterations) largely the norm for the succeeding millennium of imperial history. A few examples are the civil-service examination for prospective officials, the nationally applicable legal code, establishment of a government bureau for the drafting of an official state history, the system of bureaucratic review and transferral of centrally appointed

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

General Introduction

3

officials, an extensively organized and carefully maintained network of postal and staging stations for official communication and administrative travel. The literary achievements of Tang writers (most of whom were at least sometime members of or aspirants to the state bureaucracy), especially evident in verse but also in other genres, are well known and have become a significant part of the history of world literature and the record of humanity’s highest individual responses to the beauties and sufferings of life. In religious and intellectual history, too, the Tang left its mark, most notably with the sinicized form of Buddhism known as Chan (Japanese: Zen) which even today is the most pervasive tradition of Buddhism in East Asia. In this sphere one must take account also of religious Daoism, which was equally prevalent throughout the Tang empire, having extra import because of the dynastic family’s claimed descent from the ancient sage Laozi. We speak of the Tang as a successful dynasty that lasted for nearly three centuries, unlike the relatively brief Sui, but this picture needs some qualification. It took several years after the founding of Tang in 618 before other rival claimants for the fallen Sui mantle were completely defeated; and even at that point one could not be sure whether the Tang would last or for how long. In the event, it did endure. Yet even this is something of an intentional misreading. For the Tang was formally abolished on 16 October 690, replaced by the Zhou 周 dynasty of Empress Wu 武后 (also known as Wu Zetian 武則天), formerly the chief wife of Emperor Gaozong 高宗 (r. 649–683) and now the first and only woman in Chinese history to wield power in her own name, after previously exercising unofficial control over the court during the last dozen or so years of Gaozong’s life. In 690, with the supplanting of the Tang, it would have looked to contemporaries as though the Tang had been just one more in the line of brief dynasties lasting only two or three generations that had been the norm in the Nanbeichao period. However, Empress Wu refrained from putting to death her two sons by Gaozong, each of whom had briefly occupied the throne, while she pulled the strings, after Gaozong’s demise and before Empress Wu’s own usurpation (Zhongzong 中宗, r. Jan.–Feb. 684; and Ruizong 睿宗, r. 684–690). While these demoted scions lived there persisted some anamnesis, even a reservoir of loyalty, toward the dynastic Li 李 family among a portion of the officialdom and aristocracy, so that when the aged empress was deposed on 23 February 705, it was not only possible but seemingly apposite that the Li line of succession should be restored and with it the Tang dynasty once more. Later historians, even up to the present, have regarded the fifteen-year Zhou interregnum as an illegitimate interruption and thus usually treat the Tang as having continued unbroken throughout that period, though in actuality it did not.

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

4

General Introduction

One curious result of the Zhou interregnum was that the feasibility—if not the formal acceptance—of female rule had been proven, to the point that Zhongzong’s court (r. 705–710) in the restored Tang was essentially dominated by women, namely his Empress Wei 韋后, their daughter Princess Anle 安樂, and Shangguan Wan’er 上官婉兒, a woman of exceptional literary talent who had been Empress Wu’s secretary and drafter of decrees and who had resourcefully transferred her services to Empress Wei. The poisoning of Zhongzong in 710 by his empress and daughter was meant to secure the throne, effectively if not in name, for the latter. But it was another woman, Princess Taiping 太平, favored daughter of Empress Wu and Gaozong, who, along with her nephew Li Longji 李隆基, a son of Ruizong, thwarted this plan and ended the lives of all three of those women. Princess Taiping had her own plans for ascendancy which became apparent during the second reign of Ruizong (710–712), but these were eventually scotched by Li Longji (known to history by his posthumous temple-name as Xuanzong 玄宗) who put the dynasty back on a strong footing and ruled for more than forty years during the great age that came to be known as the “High Tang.” Those were the golden years, the famous Kaiyuan 開元 (713–742) and Tianbao 天寶 (742–756) eras that saw the fullest flowering of Tang power, prosperity, and cultural distinction. However, Xuanzong’s reign terminated in the catastrophe of the An Lushan 安祿山 rebellion, which lasted from late 755 to 763 and included the emperor’s ignominious abandonment of the capital and flight to Szechwan. In fits and starts the Tang regained control of the whole country, but for the next century and a half certain parts of the state were supervised more by local military commissioners than by bureaucrats from the central government, although the official administrative machinery of the empire continued to function and both the idea and the reality of the Tang state was preserved. There was a pervasive sense, though, that the world had changed, and this second half of the Tang dynasty was in some ways a different polity than the first half. Culturally, nevertheless, there was much continuity. Although some scholars see the great rebellion of the mid-eighth century as the beginning of a new period in Chinese history, the so-called “early modern” or the inaptly termed “Middle Period” extending on even to the late Ming, it is more accurate to regard the Tang as a single whole, representing the era of “late medieval” China. Within general continuities, there were of course, as would naturally be expected over time, certain developments and changes of emphasis in political, scholarly, literary, and to some degree social trends during the dynasty’s last century and a half.

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

General Introduction

5

Two important features of Tang government, which lasted with some variation till the twentieth century, might be underlined here. The first is the famous civil-service examination system, with the jinshi 進士 (“Presented Scholar”) exam as its most competitive and prestigious challenge. Some misconceptions exist about this annual exam. It did not attain the high status it is usually thought of as having till 681, when a separate section focusing on literary composition was added to it. Although it is often said that lyric poetry (shi 詩) was the focus of this section, that was so only occasionally and even then only partially; more often when verse composition was required it was the fu 賦 that was asked for, a more expansive and demanding genre of poetry. The focus of the exam questions for the jinshi and the genres of writing required for it (almost always including one or another form of parallel prose) were often in dispute and often changed. By the ninth century an increasingly elaborate ritual program surrounded the presentation of candidates at court, the exam itself, the posting of results, and celebrations for those who passed. But passing the jinshi exam was not a ticket to high office or indeed to any office, for successful graduates had then to sit for a higher-level xuan 選 (“selection”) exam, the passing of which only then qualified one as a possible candidate for official appointment but guaranteed nothing. The most desirable initial appointment came to be a one- or two-year assignment as a collator or editor in one of the imperial libraries in Chang’an, allowing one to remain in the capital, have access to texts that might otherwise be difficult to consult, and also take advantage of opportunities to meet and become known to higher officials who could help a budding career. A sense of the competitive nature of the exam comes with the realization that there were typically no more than thirty individuals who were certified as passing the jinshi in any year, out of nearly a thousand candidates. Also worth noting is the importance of the so-called decree examinations, several of which might be held in any one year, these being called at the behest of the emperor for specified purposes, and under a variety of names; candidates were fewer and those who made the grade here but a handful. The meaningful point is that the exam system, despite the ideal of meritocracy and aura of brilliance attached to it, contributed a quite small number of individuals to the very large official bureaucracy, most of whose members came in through different routes of family privilege. Another particularly noteworthy feature of Tang government, which had a long and important legacy, was the establishment of a dedicated historio­ graphy office. The responsibilities of this office included full-scale and wide-ranging archiving of records pertaining to places, people, and events of relevance to the dynasty, both current and past, and the periodic composition

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

6

General Introduction

of summaries, narratives, and monographs drawn from these materials. The accounts of contemporary matters were recorded and collected, ideally for the ultimate use of historians of later dynasties who would write the definitive history of the Tang. But there was constant discussion and controversy regarding the proper interpretation of earlier documents, especially as their significance affected the formulation and execution of present-day policies. Indeed, history and scholarship were hardly to be distinguished, and both were of political as well as practical consequence. Shifting our perspective only slightly, one can also locate them both in the broader field of “literature” conceived generally. Like the dynasties that came before it, the Tang was part of a civilization that valued and emphasized literature—the written word—to a degree hardly matched elsewhere in the world. It is no wonder in this regard that sinological studies should by tradition be so obsessively text-centered, even while additional inquiry is today allotted to other matters such as material culture. But even the focus of archaeology, which is annually revealing to us new objects of study, is most often directed to excavated texts, notably epigraphy. It is what the Chinese of Tang times, as of other eras, said for and about themselves that draws our continuing attention. Moreover, it is literature in its narrower sense, the expression of individual thought and feeling, that has always been recognized as the most illustrious aspect of Tang civilization. This explains its comparatively prominent place in Tang studies generally and in the extracts collected here. Scholars typically divide their focus in studying the past into large categories, usually dependent on the disciplinary boundaries of university departments. Hence there are specialists of history, or literature, or religion, etc. Since life is short and knowledge is infinite, these are necessary concessions. We realize nonetheless that lived existence in Tang China, as in our own individual experience today, did not proceed in terms of discrete divisible quanta. The men and women we study from past times were not themselves specialists but faced all sides of human experience. To begin to understand them properly, we must try to broaden our own perspectives accordingly. Hence the articles selected for reprinting here cover many different fields. Still, one cannot take in everything at once, and for convenience’ sake I have organized these items into three major categories: Political History; Literature and Cultural History; and Religion. In the past hundred years, as Western scholarship on premodern China has developed in increasingly sophisticated and inevitably more specialized ways, study of the “late medieval” period centered on the Tang has yielded manifold new discoveries and understandings of that great age. The four volumes here at hand gather extracts of some of the most important English-language

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

General Introduction

7

e­ xamples of these studies, with the intent of serving as a first (but hopefully not last) reference for interested students and scholars. If possible, I would have reprinted here the four most outstanding books on Tang China published in recent times, namely Edwin G. Pulleyblank’s The Background of the Rebellion of An Lu-shan (Oxford Univ. Press, 1955), Antonino Forte’s Political Propaganda and Ideology in China at the End of the Seventh Century (Naples: Istituto universitario orientale, 1976; much revised 2nd edn., Kyoto: Italian School of East Asian Studies, 2005), Stephen Owen’s The Great Age of Chinese Poetry: The High T’ang (Yale Univ. Press, 1981), and David McMullen’s State and Scholars in T’ang China (Cambridge Univ. Press, 1988). Pulleyblank’s and Forte’s books are so tightly structured that it was not feasible to extract a single chapter from either of them. From Owen’s and McMullen’s books I have drawn a chapter apiece but also included separately published articles on different topics by both authors. These four books should be at the top of the reading list for anyone interested in Tang China. Also in this category are the three great books of Edward H. Schafer, The Golden Peaches of Samarkand: A Study of T’ang Exotics (Univ. of California Press, 1963), The Vermilion Bird: T’ang Images of the South (California, 1967), and Pacing the Void: T’ang Approaches to the Stars (California, 1977), which are in the nature of literary encyclopedias on different aspects of Tang culture and imagination. I had originally drawn up a list of eighty-five significant articles and book chapters for inclusion, but limitations of space demanded a severe reduction of that list, with the result that many equally deserving publications are, regrettably, not included here. Those that are included, as is Brill’s practice with all of the volumes in its Critical Readings series, have been completely reformatted in uniform style for ease of reading, and no longer carry their original pagination. Obvious typographical and factual errors have been corrected, but no substantive revisions of content have been made; the material stands as originally published, although most living authors would have welcomed the possibility to add later thoughts or more recent reflections. It should be noted that the older articles and chapters use the Wade-Giles romanization system that was standard in English-language scholarship till about two or three decades ago, while the more recent items use the Hanyu Pinyin system that is now the norm. Paul W. Kroll

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

History—Political, Intellectual, and Military

∵

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

Wên Ta-ya: The First Recorder of T’ang History Woodbridge Bingham One of the turning points in Chinese history came in 617 AD with the Duke of T’ang’s campaign from T’ai-yüan in Shansi to Ch’ang-an, the modern Sianfu, in Shensi. The empire of the short-lived Sui dynasty, which had reunited the whole of China for the first time in many years and had started a period of great cultural development, was rapidly disintegrating. Bandits and rebels had already gained control of large sections of the country. But it was reserved for the Duke of T’ang, Li Yüan, and his second son Li Shih-min (T’ang T’ai Tsung) to take advantage of this situation and to establish a power which would gradually supersede the Sui authority and gain control over all sections of the country. Their rule was the beginning of the T’ang dynasty, the “Golden Age” of China. The Duke of T’ang and his sons were in a position to make the best of the waning power of Sui Yang Ti (604–618). Briefly stated, their distinguished ancestry, the current prophecy concerning the future success of one of the Li name, their protected geographical location in Shansi, and Li Yüan’s personal record of official service and recent military success, all aided in making him unusually prominent at this time. Among the former Sui officials who looked to Li Yüan for leadership and flocked to join him even before he began his successful campaign through Shansi to the Western Capital, Ch’ang-an, was a scholar named Wên Ta-ya 温大雅. He accompanied the future emperor on his march into Shensi and served him for several years in positions of great confidence and responsibility. He was with Li Yüan when Ch’ang-an was captured and he helped to arrange the ceremonies at the inauguration of the new dynasty in 618. The Chinese literary works which are most commonly used as source materials for the history of this change in dynasty are three official dynastic histories: the Sui shu, Sui History or Book of Sui, which was written by men who took part in the founding of the T’ang dynasty within fifty years after the events, and the Chiu T’ang shu and T’ang shu, Old and New T’ang Histories, dated 945 and 1060 respectively, which are based on the official records and earlier histories of parts of the T’ang dynasty. Of these the Chiu T’ang shu is the most important for the events of 617 and 618.

Source: “Wen Ta-ya: The First Recorder of T’ang History,” Journal of the American Oriental Society 57 (1937): 368–74.

© koninklijke brill nv, leiden, ���9 | doi:��.��63/9789004380158_003

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

12

Bingham

To these standard histories must be added the work of Wên Ta-ya. In his capacity as official recorder at Li Yüan’s Headquarters he wrote a journal which is generally entitled the Ta T’ang ch’uang yeh ch’i chü chu1 or Court Journal of the Founding of the Great T’ang. This is a chronological résumé of events concerning Li Yüan and his founding of the dynasty from 615 up to the proclaiming of the new emperor. It differs in many important details from the narratives of the standard dynastic histories. Of especial significance is the fact that it makes out Li Yüan to be the real leader and gives little to indicate that his son Li Shih-min was particularly important. Most later accounts of the founding of the T’ang dynasty whether in Chinese or in western languages are based on the dynastic histories and stress the importance of Li Shih-min. This is done in spite of the fact that the Court Journal of Wên Ta-ya is recognized by Chinese scholars as containing reliable historical material. Hence there is a need for a critical study of this book in order that we may get a true understanding of the events which took place when the Duke of T’ang rebelled and took over the central authority at Ch’ang-an. At the commencement of each of the three chapters of the Court Journal and immediately following the title itself, the author is referred to in this way: “Written by the T’ang minister, ta hsing t’ai of Shan-tung tao 陝東道大行臺, Great State Pillar, Duke of Lo-p’ing chün and of the State’s Beginning, Wên Ta-ya.” The titles “Great State Pillar” (shang chu kuo) and “Duke of the State’s Beginning” (k’ai kuo kung) were both honorary titles indicating his participation in the founding of the new dynasty. The official biography of Wên Ta-ya is to be found in both of the T’ang histories. The version of the New T’ang History is obviously based on what had previously been written in the Old T’ang History and hence the newer history is used here chiefly as a check in translating the older version. The following is a translation of the biography of Wên Ta-ya from the Chiu T’ang shu, ch. 61, pp. 1a–1b2 (parallel account in T’ang shu, ch. 91, pp. 1a–1b): Translation Wên Ta-ya, tzŭ Yen-hung 彥弘, was a man of Ch’i 祁 in T’ai-yüan. His father was [Wên] Chün-yu 君悠.3 Under the Northern Ch’i he was a Scholar-Official of the 1  大唐創業起居注. 2  T’ung wên shu chü photo-lithographic reprint of the Ch’ien Lung palace edition of 1739, Shanghai, 1884. 3  T’ang shu ch. 91, p. 1a, gives 君攸, same pronunciation. Chiu T’ang shu chiao k’an chi ch. 34, p. 15b, considers 悠 a mistake. Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

Wên Ta-Ya: The First Recorder of T ’ ang History

13

Wên lin Academy 文林舘. Under the Sui [dynasty] he was ssŭ-ma of Ssŭ chou 泗州 [in present Kiangsu]. At the end of Ta Yeh [reign-period of Sui Yang Ti 605–617] he was Secretary in the Board of Justice [ssŭ li ts’ung shih 司隸從事. When he saw that the Sui government was daily becoming more confused he excused himself on account of illness and retired from office. [Wên] Ta-ya by nature was most filial. In his youth he was fond of learning, and for able accomplishment his reputation was well known. [The T’ang shu version here adds: (Wên Ta-ya) with his younger brothers Yen-po 彥博4 and Ta-yu 大有 were all well known. When Hsüeh Tao-hêng 薛道衡 (an official under the Northern Ch’i, Northern Chou, and Sui dynasties; Sui shu, ch. 57) saw them he sighed in admiration and said: “All three men indeed have the abilities of high ministers.” (This passage appears to be adapted from the Chiu T’ang shu biography of Wên Yen-po; ch. 61, p. 2a.)] [Wên Ta-ya] held office under the Sui as Scholar-Official of the Eastern Palace [i. e. at the court of the Heir Apparent, who was at this time Yang Yung, a prince opposed and superseded by the Emperor Yang Ti] and as District Marshal of Ch’ang-an. On account of mourning for his father he retired from office. Later on because the empire was disturbed he did not [again] seek to hold office. When Kao Tsu [Li Yüan] was in charge of the garrison at T’ai-yüan he greatly honored [Wên Ta-ya]. When the “Righteous Troops” rose [in revolt], he brought in [Wên Ta-ya] to act as Staff Officer in charge of Records [chi shih ts’an chün 記室參軍] at the Generalissimo’s Headquarters. He had special control of official documents. At the time of the abdication and succession [i. e. of the last Sui emperor, Yang Yu (Sui Kung Ti 617–618), and Li Yüan (T’ang Kao Tsu) respectively] he and the Staff Officer in charge of Interdepartmental Affairs [ssŭ lu (ts’an chün) 司錄(參軍)]5 Tou Wei and the Chief Registrar [chu po 主簿] Ch’ên Shu-ta 陳叔達 together arranged the ceremonies. During the first year of Wu Tê [618] he was transferred to be Vice-President of the Department of the Imperial Chancellory [huang mên shih lang 黃門 侍郞]. His younger brother [Wên] Yen-po became Vice-President of the Department of the Grand Imperial Secretariat [chung shu shih lang 中書 侍郞]. The pair occupied positions of close intimacy [with the Emperor Kao Tsu] and in [Imperial] deliberations they were honored. Kao Tsu in a relaxed mood spoke [to them] and said: “When I raised the Righteous Army at Chinyang, it was only by means of you [two].”

4  His recorded tzŭ Ta-lin 大臨 may possibly have been his real given name. 5  The full title with the last two characters added is given in Tou Wei’s biography, Chiu T’ang shu, ch. 61, p. 6a. Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

14

Bingham

After this he was transferred to the Board of Public Works. He was promoted to be ta hsing t’ai of Shan-tung tao and President of the Board of Public Works. On account of the Yin Heir Apparent [Li Chien-ch’êng] and Prince Ch’aotz’ŭ [Li Yüan-chi] [these two were plotting against their brother Li Shih-min] T’ai Tsung [i. e Li Shih-min] ordered [Wên] Ta-ya to take charge of the garrison at Lo-yang so as to anticipate [their] rebellion. [Wên] Ta-ya several times arranged secret plans and [thereby] received much official approval and rewards. When T’ai Tsung came to the throne he was successively transferred to be President of the Board of Rites and invested as Duke of the state of Li 黎國公. When [Wên] Ta-ya was about to rebury his paternal grandfather, the diviners said: “Burial at this place harms the elder brother and brings happiness to the younger.” Ta-ya said: “If this brings about my younger brother’s everlasting prosperity, I shall die with a smile on my lips.” The burial was finished. After a little more than a year he died. His posthumous name was Hsiao 孝 [meaning “Filial”]. He wrote “The Court Journal of the Founding [of the Great T’ang]” in three chapters 卷. In Yung Hui 5th year [654, the 5th year of the T’ang emperor Kao Tsung] he was posthumously entitled: Right Vice-President of the Department of State Affairs [shang shu yu p’u yeh 尙書右僕射]. This concludes the Chiu T’ang shu biography. There are a few further but mostly unimportant references to Wên Ta-ya in the early literature. Other biographies in Chiu T’ang shu ch. 61 give statements containing no additional information and the section on bibliography merely states his authorship of the Court Journal as it appears in his biography.6 Other items concerning Wên Ta-ya are contained in various Sung dynasty works: the T’ang hui yao, ch. 79, p. 18a; the T’ai p’ing yü lan, ch. 221; the Tzŭ chih t’ung chien of Ssŭ-ma Kuang, ch. 184, p. 2b; and the Ts’ê fu yüan kuei, ch. 782, p. 15b. These four all appear to be citations from the Chiu T’ang shu biography and are important only inasmuch as they present variant readings of the original. The Ts’ê fu yüan kuei is a collection of early documents published in 1013. Among them, in the section on “National History, (List of) Selected Works,” is also included an item which throws a different light on one of the points in which we are most interested, the writing of the Court Journal. This is the statement: “Wên Ta-ya was President of the Board of Rites; he wrote The Annals of the Establishment of the Present Ruler in his Royal Estate (Chin Shang wang 6  Chiu T’ang shu, ch. 46A, p. 31b.

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

15

Wên Ta-Ya: The First Recorder of T ’ ang History

yeh chi 今上王業記) in six chapters.”7 This helps to confirm the authorship of Wên Ta-ya and from the unusual wording of the title one is led to believe that the Court Journal was written during the reign of T’ang Kao Tsu, i. e. Li Yüan, as emperor and before he abdicated the throne to his son Li Shih-min. Another reference also indicates the early writing of the Court Journal and gives us what perhaps may have been the manuscript title while the events were being recorded and even before Li Yüan ascended the throne. In the New Revised Gazetteer of the District of Ch’i (this Ch’i was the native home of Wên Ta-ya), our scholar-official’s biography is found in the section on “local worthies” and has been compared with the biographies in the dynastic histories.8 It is identical in every way with the version of the New T’ang History except that the latter does not contain any reference to the Court Journal while at the end of the biography in the gazetteer is added this sentence: “He wrote the Official Annals of the Great Chancellor, the Prince of T’ang (Chu Ta ch’êng hsiang T’ang wang kuan shu chi 著大丞相唐王官屬記).” These personal titles were those assumed by Li Yüan a week after he had taken Ch’ang-an and at the time when he set up the boy Yang Yu as a puppet emperor, that is six months before he actually became emperor himself. The various problems connected with the Court Journal cannot be fully discussed at this time. But it may be pertinent in connection with Wên Ta-ya’s biography to add a little of what later scholars have thought about the date of his book. In a post-face to a Ming edition of the Court Journal the Ming scholar Hu Chên-hêng 胡震享 says that it was probably written while Li Chien-ch’êng (Li Shih-min’s elder brother) was Heir Apparent, that is during the period 618– 626, the reign of Li Yüan as emperor.9 This also is the generally accepted opinion among Chinese scholars of the present day,—that Wên Ta-ya’s book in its present form is an authentic document of the reign of T’ang Kao Tsu. Wên Ta-ya could not have written much later than this time for he probably died within the next ten years, i. e. during the first part of the reign of T’ang T’ai Tsung. We have no dates either for his birth or his death and can only calculate them from what we know about his younger brother Wên Yen-po. The latter died in 637 AD at the age of sixty-four according to Chinese reckoning.10 Hence he was born in 574 and his older brother some time prior to that date. Wên Ta-ya is not mentioned in the T’ang annals for the reign of T’ai Tsung or his successor Kao Tsung and so any reckoning concerning his death must be 7  Ts’ê fu yüan kuei (1814 edition), ch. 556, p. 14a. 8  Hsin hsiu Ch’i hsien chih 新脩祁縣志 (1882), ch. 7, pp. 24a–25a. 9  Chin tai pi shu 津逮祕書, Collection 10, Volume cm, Postface 3. 10   Chiu T’ang shu, ch. 61, p. 3b, and ch. 3, p. 8a.

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

16

Bingham

based on the details of his biography and the dates of his brother. From the anecdote about the reburial of his grandfather and the prophecy of his brother’s prosperity followed by Wên Ta-ya’s death a year later it is probable that Wên Yen-po was still alive when the prophecy was made. The other brother, Ta-yu, had died in 618.11 Hence Wên Ta-ya probably died some time between his last appointments, which were made after T’ai Tsungs’s accession in 626, and the death of Yen-po in 637. Our scholar-official may then have died about the same time as his former master the first T’ang emperor whose death came in 635 nine years after his abdication of the throne. Their lives were bound together in many ways and to the extent that Li Yüan is an important figure in Chinese history so also must Wên Ta-ya’s career be understood. The latter’s importance and that of his Court Journal is enhanced by the fact that he himself was a participant in the founding of the new dynasty. He was among the Shansi group of Sui officials who were with Li Yüan at the very start of the T’ang rebellion. As Recorder and keeper of official documents at the Duke of T’ang’s headquarters Wên Ta-ya was on the inside of what was happening and Li Yüan himself later referred to him as an important member of those who shared in the responsibility for the success of the whole enterprise. At the time when he probably was finishing his account of the founding he was an official of high standing at the court and was personally trusted not only by the first emperor Kao Tsu but by his successor T’ang T’ai Tsung. Such was the man who officially recorded the beginnings of T’ang history. 11   Chiu T’ang shu, ch. 61, p. 4b.

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

The Rise to Power of the T’ang Dynasty: A Reassessment Howard J. Wechsler The traditional interpretation of the founding of the T’ang, which holds that T’ai-tsung rather than his father, Kao-tsu, was the genius behind the revolt that eventuated in the establishment of the dynasty, was given final form in the tenth and eleventh centuries with the compilation of two Standard, or canonical, dynastic histories (cheng-shih). The first of these, the Old T’ang History (Chiu T’ang-shu) of 945, states, “At this time, because Sui rule had already come to an end, T’ai-tsung secretly planned the righteous uprising.”1 The second, the New T’ang History (Hsin T’ang-shu), completed in 1060, notes, “When Kao-tsu first rose in Taiyuan, it was not his own idea; rather the affair originated with T’ai-tsung.”2 Perpetuating the view of the Standard Histories was the great Sung historian Ssu-ma Kuang, who, in his monumental chronicle, the Comprehensive Mirror for Aid in Government (Tzu-chih t’ung-chien) of 1084, concluded, “The emperor’s raising of troops in Chinyang was entirely planned by the Prince of Ch’in (T’ai-tsung)” and “Kao-tsu obtained the empire entirely because of T’ai-tsung’s merit.”3 Until a short time ago this traditional view of early T’ang history received almost universal credence. Recently, though, a small group of revisionists led by the Chinese scholars Lo Hsiang-lin and Li Shu-t’ung have reassessed the roles played in the T’ang founding by the first and second emperors.4 They have reinterpreted data in the T’ang Standard Histories, relying chiefly on information culled from their biographical sections (lieh-chuan), and have also made important use of an early seventh-century source almost totally ignored by the traditionalists, the Source: “The Rise to Power of the T’ang Dynasty: A Reassessment,” in Howard J. Wechsler, Mirror to the Son of Heaven: Wei Cheng at the Court of T’ang T’ai-tsung, New Haven: Yale University Press, 1974, pp. 8–32. 1  c ts 2.2. 2  h ts 2.4b. 3  t ctc ch. 190, p. 5957; ch. 191, p. 6012. 4  See Lo Hsiang-lin, T’ang-tai wen-hua shih [A Cultural History of the T’ang Dynasty] (Taipei, 1955), chap. 1; Li Shu-t’ung, T’ang-shih k’ao-pien [An Examination of T’ang History] (Taipei, 1965), pp. 1–42, 43–98, 276–309. The chapter by Lo was originally published as an article in 1936; those by Li were originally published as articles in the 1950s and early 1960s.

© koninklijke brill nv, leiden, ���9 | doi:��.��63/9789004380158_004

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

18

Wechsler

Diary of the Founding of the Great T’ang Dynasty (Ta-T’ang ch’uang-yeh ch’ichü-chu) by Wen Ta-ya. They have, in short, essentially rewritten the history of the period. The arguments presented in this chapter are to a substantial degree based on their impressively detailed researches as well as on my own more recent excursions through early T’ang history.

The Decline of the Sui and the Rise of Li Yüan

The Sui dynasty (589–618), predecessor of the T’ang, came to power when its founder, Wen-ti, usurped the throne of the Northern Chou, last of the regimes that had dominated North China during the Period of Disunion. Eight years later Wen-ti toppled the southern house of Ch’en, brought to an end the political and cultural fragmentation that had plagued China for close to four centuries, and launched its Second Empire. But Wen-ti’s triumph was ephemeral, and within two generations his dynasty also lay in ruins. At first there was little to portend that the Sui would be so short-lived. Wen-ti reestablished a highly centralized government marked by an impressive level of administrative rationality. Many of the emperor’s officials were recruited by means of the newly revived civil service examinations, which were designed to reward merit over hereditary privilege. A magnificant capital of unprecedented scale—almost six miles east to west and more than five miles north to south—and aptly named Ta-hsing-ch’eng, or “Great Revival City,” was constructed southeast of the old Han capital, Ch’ang-an, on the site of modern Sian. Reimposition of the “equal-field” land-tenure system of the Northern Dynasties, combined with the development of new and more efficient methods of tax collection, swelled the national revenues. Grain shipments to the food-poor capital region were facilitated by the construction of a canal (the Kuang-t’ung ch’ü) linking Ta-hsing-ch’eng with the fertile plain lying to the east, beyond the T’ung Pass. A variety of state, prefectural, and village granaries were built to combat famine. The border defense along the northern frontier was substantially reinforced by massively repairing, rebuilding, and extending the Great Wall. Neighboring tribes and states were weakened as a result of the vigorous foreign policies Wen-ti pursued on the diplomatic and military fronts. Lastly, Wen-ti began the difficult task of uniting the many disparate political, ethnic, and cultural groups that had developed during the Period of Disunion under successive alien regimes in the north and Chinese royal houses in the south.5 By the time of his death in 604, the first Sui emperor had wrought what 5  See Arthur F. Wright, “The Formation of Sui Ideology, 581–604,” in Chinese Thought and Institutions, John K. Fairbank, ed. (Chicago, 1957), pp. 71–104. Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

The Rise to Power of the T ’ ang Dynasty: A Reassessment

19

must have seemed a miracle to most of his subjects: China was united, peaceful, and prosperous for the first time in centuries.6 The reign of Wen-ti’s son, Yang-ti, also began promisingly enough.7 Yang-ti continued to improve Sui administration, promulgated a new code of laws less stringent than his father’s, expanded the state system of education and civil service examinations (it was during Yang-ti’s time that the famous chin-shih examination originated), and sponsored a revival of Confucian learning. On the foreign front, he briefly extended Sui suzerainty over the Eastern Turks to the north of China and manipulated the Western Turks, those in the region west of the Jade Gate and north of the Tarim Basin, to China’s advantage. He drove the T’u-yü-hun from their homeland in modern Tsinghai province and opened relations with Japan. Great amounts of tribute from the Sui’s Central Asian vassals poured into Ta-hsing-ch’eng along the silk routes, which were the links between China and West Asia and, beyond it, Europe. In the end, if we are to believe accounts in the Sui Standard History, the Suishu, Yang-ti’s program of public works, carried out on a gigantic scale, brought disaster to the dynasty. During the period 605–10 the second Sui emperor constructed an eastern capital (tung-tu) at Loyang in northern Honan, roughly half the size of Ta-hsing-ch’eng, and expanded upon his father’s work on the Great Wall. He had a series of canals hundreds of miles in length redredged or constructed anew, linking Loyang first with the Huai and Yangtze River valleys (via the T’ung-chi ch’ü) and then with the region further south terminating at modern Hangchow (via the Chiang-nan ho). Following the completion of the T’ung-chi canal, the emperor set sail aboard a “dragon boat” for Chiang-tu (modern Yangchow), his “river capital” near the Yangtze, leading a flotilla of vessels stretching in close file for more than sixty miles. The longest of all Yangti’s canals (the Yung-chi ch’ü) linked Loyang with the region around modern Peking. Hundreds of thousands of laborers were conscripted for these vast undertakings, great numbers of whom perished, and the economic resources of the newly united country were strained to the breaking point. For these “excesses” Yang-ti has earned the opprobrium of countless Chinese commentators, who have damned him with a stereotyped portrayal as a “bad last” ruler. Yet in so doing they have largely ignored his many positive 6  Two recent studies that stress the accomplishments of Sui Wen-ti are T’ang Ch’eng-yeh, Sui Wen-ti cheng-chih shih-kung chih yen-chiu [The Political Achievements of Sui Wen-ti] (Taipei, 1967), and Arthur F. Wright, “The Sui Dynasty,” in the forthcoming Cambridge History of China. 7  The following material relating to Sui Yang-ti is largely based on Woodbridge Bingham, The Founding of the T’ang Dynasty: The Fall of Sui and the Rise of T’ang (Baltimore, 1941; reprint 1970), pp. 1–59, and Wright, “The Sui Dynasty.” Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

20

Wechsler

contributions to the consolidation of the Sui. To cite just two examples, the canals he built brought the Sui into easy communication with most of the major food-producing regions of China; and the elaborately mounted imperial progresses he made over their waters were in many respects a continuation by other means of his father’s efforts to weld together the highly heterogeneous peoples of late sixth-century and early seventh-century China. Étienne Balazs has justly called Yang-ti the real founder of the Second Empire because he integrated the southeast into the rest of the country for the first time since the Han.8 Unfortunately, following upon the completion of many of his most ambitious public works, and before the peasantry had time to recover from oppressive levies of forced labor and taxes, Yang-ti began the first of three disastrous campaigns to conquer Koguryŏ, a kingdom located east of the Liao River on the northern half of the Korean peninsula. The first campaign of 611–12 proved unsuccessful for two major reasons: a great flood in the lower Yellow River valley that caused heavy desertions in the Sui ranks among men from the affected region, and the failure of Sui forces to reach the Koguryŏ capital before the onset of winter. Nevertheless, Yang-ti proceeded with his second and equally disappointing Koguryŏ campaign of 613, levying new taxes and conscripting more men. Widely scattered revolts now erupted across China, the most serious of which took place in Honan province led by the Sui president of the Board of Rites (li-pu shang-shu), Yang Hsüan-kan.9 It was quickly crushed by the government. Beginning in 614, as Yang-ti embarked upon yet another attempt to take Koguryŏ, rebellion began to engulf the entire country. Desertions in the Sui ranks became rife, supplies for the army failed to reach their destinations, and Sui military power rapidly ebbed. Late in 615 China’s “barbarian” neighbors to the north, the Eastern Turks, weakened the emperor’s prestige still further by surrounding him at the town of Yen-men in northern Shansi while he was inspecting fortifications along the Great Wall. Yang-ti was forced to endure the embarrassment of being detained at Yen-men for a full month before the Turkish siege was broken by Sui reinforcements. Following his return to Ta-hsing-ch’eng from Yen-men, the growing threat of civil disturbance and the continued inability of Sui armies to restore order 8  Étienne Balazs, “L’Oeuvre des Souei: L’Unification,” in Histoire et institutions de la Chine ancienne, by Henri Maspero and Etienne Balazs (Paris, 1967), p. 165. 9  For a study of Yang Hsüan-kan’s revolt, see Nunome Chōfū, “Yō Genkan no hanran” [The Revolt of Yang Hsüan-kan], Ritsumeikan bungaku [Ritsumeikan University Journal of Cultural Sciences], 236 (1965), 1–30.

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

The Rise to Power of the T ’ ang Dynasty: A Reassessment

21

prompted the emperor to sail south to Chiang-tu. Thereafter he remained in seclusion, isolated from the news of his crumbling empire, immobilized by self-doubt. The resulting political vacuum was increasingly filled by other men: Sui local officials, commanders of local military elites (hao-chieh), “bandit” (tsei) chiefs, and rebel leaders. Early in 618 Yang-ti was assassinated at his Yangtze capital by one of his own officials, Yü-wen Hua-chi. By this time, however, he had already been demoted to the status of “retired emperor” (t’aishang huang) by the rebel conqueror of Ta-hsing-ch’eng, the Duke of T’ang, Li Yüan (the given name of T’ang Kao-tsu), who would shortly establish a dynasty that came to rival the Han in extent and splendor. Like the majority of dynastic founders before and after him, Li Yüan was of noble lineage. His ancestors, members of the northwestern aristocracy centered in Shensi and Kansu, occupied prominent offices in successive semi“barbarian” regimes in North China from the time of the Northern Wei dynasty; some scholars have even hypothesized that the Li-T’ang house was itself originally “barbarian.”10 What is clear, at any rate, is that the first T’ang emperor was at least half non-Chinese, for his mother, née Tu-ku, was of noble Hsien-pei stock. At the beginning of the Northern Chou dynasty, Li Yüan’s grandfather was posthumously ennobled Duke of T’ang principality, a title that was first passed on to Li Yüan’s father and then on to Li Yüan at the age of six in the year 574. Since his mother was the sister of Sui Wen-ti’s consort, in his youth Li Yüan was a frequent guest at the Sui imperial palace. While still in his teens, Li Yüan embarked upon his official career as one of the elite palace guard (ch’ien-niu pei-shen). His relationship by marriage to Emperor Wen brought him rapid promotion, and soon he was enjoying high office in the provincial bureaucracy. Under Emperor Yang, Li Yüan was summoned back to the imperial palace first as the assistant director of the Department of Imperial Domestic Service (tien-nei shao-chien) and later as the vice-president of the Court of Imperial Insignia (wei-wei shao-ch’ing). In 613, during Yang-ti’s second campaign against Koguryŏ, he was put in charge of transporting provisions at the Huai-yüan garrison (chen) in modern Liaoning province. A turning point in Li Yüan’s career came this same year with the eruption of Yang Hsüan-kan’s revolt at Li-yang in northern Honan. According to one 10  While it is clear that Li Yüan’s immediate paternal forebears came from Lung-hsi, modern Kansu province, the more remote origins—both geographical and racial—of the Li-T’ang house have been hotly argued since the 1930s by Liu P’an-sui, Ch’en Yin-k’o, Kanei Yukitada, Chu Hsi-tsu, Wang T’ung-ling, and Ts’en Chung-mien, among others. The debate continues.

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

22

Wechsler

source, when Yang was preparing to revolt, his brothers fled to Honan from the ranks of the Sui expeditionary army in Liaoning; it was Li Yüan who first discovered and reported these defections to Yang-ti.11 The emperor then appointed Li Yüan garrison commander (liu-shou) of Hung-hua commandery (chün) at modern Ch’ing-yang in the easternmost part of Kansu province, and ordered that all armies “to the right of the Pass” (kuan-yu, the region to the west of the T’ung Pass, comprising modern Shensi and eastern Kansu provinces) were to be placed under his command in order to resist the rebels. As a result of steadily deteriorating conditions throughout the country, from this time forward Li Yüan held posts exclusively in the Sui military hierarchy. During the period 615 to 616 he destroyed two “bandit” organizations in southern Shansi and successfully opposed Turkish incursions into the northern portion of the province. By 616 Sui control of the central Shansi region had become seriously weakened as a result of the defeat of imperial forces in Taiyuan by the “bandit” Chen Ti-erh and a force of followers said to have numbered one hundred thousand men. When Li Yüan crushed Chen less than a year later, he was awarded the post of garrison commander (liu-shou) of Taiyuan (T’ai-yüan fu), with headquarters at the town of Chin-yang. As Bingham has observed, this new appointment, made some months after Yang-ti had sailed into splendid isolation at Chiang-tu, may merely have reflected the considerable de facto power Li Yüan already wielded in the region.12 Despite this promotion, Li Yüan’s position under the Sui was actually quite insecure. About the year 614 a ballad of enigmatic wording had gained currency among the people predicting that the next person to occupy the throne would be surnamed Li, a name as common in China as Smith and Jones in our country. The ballad helped to launch or advance precipitately the careers of other rebels surnamed Li, such as Li Mi and Li Kuei.13 At the same time, it made Yang-ti paranoiacally suspicious of all those in his employ bearing the tainted surname, a condition most notably illustrated by his execution of thirty-two members of the clan of Li Hun, including that hapless official. Li Yüan well knew that there was a distinct possibility he would soon suffer a similar fate.

11   h ts 1.1b. 12   The Founding of the T’ang Dynasty, pp. 79–80. 13  Woodbridge Bingham, “The Rise of Li in a Ballad Prophecy,” Journal of the American Oriental Society, 61 (1941), 272–73; tctc ch. 183, p. 5709; ch. 184, pp. 5745–46.

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

The Rise to Power of the T ’ ang Dynasty: A Reassessment



23

The T’ang Uprising—Variations on a Theme

It is at this point that the traditional and revisionist accounts of the T’ang founding part company. According to the traditional and widely accepted version of events based on the T’ang Standard Histories and the Comprehensive Mirror,14 when Li Yüan became the garrison commander of Taiyuan, his second son, Li Shih-min (the later T’ai-tsung), by himself conceived of a revolt intended to sweep away the already moribund Sui and establish the Li-T’ang house in its place. Shih-min, the traditional accounts say, first discussed plans for such an undertaking with the former prefect (ling) of Taiyuan, Liu Wenching, who had been clapped into prison at Chin-yang because he was related by marriage to the anti-Sui rebel Li Mi. Liu energetically encouraged Shih-min in his plans. But the latter, fearing that the elder Li would be too timid to second the venture and would thus pose an obstacle to the revolt, decided he would have to force his father’s hand. He therefore engaged the aid of Li Yüan’s old friend, P’ei Chi, then the assistant superintendent ( fu-chien) of the Sui palace at Chin-yang. P’ei secretly sent ladies from the palace harem for Li Yüan’s private enjoyment without informing him of their origin. When Li Yüan belatedly discovered that he had unknowingly been compromised, he realized that it would only be a matter of time before Yang-ti learned of his crime and ordered his execution. Therefore, early in the fifth month of 617 he reluctantly assented to raising the banner of revolt. Li Yüan’s first and fourth sons, Li Chien-ch’eng and Li Yüan-chi,15 were then summoned from southern Shansi to Chin-yang and the creation of a “righteous army” of revolt was begun. The revolt, however, could not get under way until a serious obstacle had been removed. At the time Yang-ti appointed Li Yüan garrison commander of Taiyuan, he had also sent two deputies, Wang Wei and Kao Chün-ya, to serve under him. Li Yüan now became fearful that they would report the troop buildup to Chiang-tu. Thus, according to the traditional accounts, he summoned both deputies to audience, where they were seized by troops led by Li Shih-min, imprisoned, and executed shortly thereafter. In the subsequent T’ang campaign to conquer the Sui capital, we are also told, Li Yüan relied most heavily on the talents of Shih-min, who more than anyone else contributed the leadership and strategy which led to the T’ang triumph. 14  The major accounts of the T’ang uprising and the campaign to take the Sui capital are in cts 1.2b–4b, 2.1b–3b, 57.6–8b; hts 1.2b–5, 2.1b–2b, 88.1–3; tctc ch. 183, pp. 5728–35; ch. 184, pp. 5737–61. 15  Li Yüan’s third son, Li Hsüan-pa, had died earlier.

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

24

Wechsler

The traditional accounts portray Shih-min, who at the time of the revolt was about seventeen years of age,16 as a brilliant military commander, physically powerful and intensely charismatic. On the other hand, they depict Li Yüan, then about fifty-one, as a doddering and spineless old man, buffeted about by events over which he had no control, an unwilling pawn in the hand of his wily son, Shih-min. The following, from C. P. Fitzgerald’s 1933 monograph, Son of Heaven: A Biography of Li Shih-min, The Founder of the T’ang Dynasty, admirably sums up the view of Li Yüan’s qualities that has persisted for more than a millennium: “Li Yüan, duke of T’ang, was an easygoing aristocrat, not remarkably intelligent, a weak character. He lacked tenacity, foresight and resolution. Had he not been the father of Shih-min there was no man living in China less likely to win his way to the throne.”17 There is, however, another portrait of Li Yüan contained in Wen Ta-ya’s Diary of the Founding of the Great T’ang Dynasty, compiled sometime during the period 617 to 626, that sharply contrasts with the above. Wen Ta-ya’s Li Yüan is a great archer: “Whenever he sighted a running animal or a flying bird he shot without missing the mark.”18 He overwhelms his adversaries with dazzling displays of his martial power and swaggering bravado in the face of danger.19 He is so adept at employing military strategies against his enemies that even the fierce Eastern Turks are loath to face him on the battlefield.20 Wen Ta-ya’s Li Yüan—a man of powerful ambition, inexhaustible energy, and indomitable

16  The large amount of conflicting data regarding Shih-min’s birthdate and his age at the time of the Taiyuan revolt is discussed by Bingham, The Founding of the T’ang Dynasty, p. 49, n. 79; Nunome Chōfū, “Tensaku jōshō, Sentōdō daikōdai shōshorei, Shin Ō Seimin— sokuimae no To no Taisō” [Supreme Commander of Heavenly Strategy, President of the Department of Affairs of State of the Shan-tung Circuit Grand Field Office, the Prince of Ch’in, Shih-min—Prior to His Accession as T’ang T’ai-tsung], Ritsumeikan bungaku [Ritsumeikan University Journal of Cultural Sciences], 255 (1966), 5–7; and Ma Ch’i-hua, “Chen-kuan cheng-lun” [Commentary on the Politics of the Chen-kuan Period], Kuo-li cheng-chih ta-hsüeh hsüeh-pao [National Chengchi University Journal], 1 (1960), 270–71. Nunome arrives at a birthdate of the first month of solar 599 (lunar 598), Ma at a birthdate of 600. This would make Shih-min eighteen years of age and seventeen years of age, respectively, at the time of the Taiyuan revolt. Because Shih-min himself once observed (hts 102.6b) that he was seventeen years old (eighteen sui) when the revolt took place, I am inclined to accept a birth-date of 600. 17  P. 32. 18   c yccc 1.2b. 19  Ibid. 1.2b, 4–4b. 20  Ibid. 1.3.

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

The Rise to Power of the T ’ ang Dynasty: A Reassessment

25

will—emerges as the real leader of the Taiyuan uprising and the victorious T’ang campaign to conquer the Sui capital. Wen shows that long before the T’ang uprising in the fifth month of 617, Li Yüan had already begun to dream of replacing the Sui. As we have seen, in 616 Li Yüan, Duke of T’ang, was assigned to combat “banditry” in Taiyuan, which was the site of his nominal fief. Now, Wen notes, Li Yüan came to view the coincidence of his fief and place of assignment as a propitious sign from Heaven.21 Early in 617, still some months before the Taiyuan uprising, Wen records the following comment made by Li Yüan to his son Shih-min: “The allotted time of the Sui is about over, [but] our house will continue to respond to auspicious omens from Heaven. If we have not raised troops at this early time, it is because you and your brothers have not yet assembled [your forces].”22 It is thus apparent that sometime before the T’ang revolt Li Yüan had already become convinced that he was destined to inherit the Mandate, a conviction reinforced by the “Li Ballad,” which, Wen informs us, was sung by young and old on the streets of Chin-yang. On one occasion Li Yüan even remarked, “I ought to rise up [and march] one thousand li to fulfill that prophecy!”23 The Diary also casts doubt on Shih-min’s contributions to the Taiyuan revolt. First, it makes no mention of his direct participation in the seizure of Wang Wei and Kao Chün-ya, noting instead that the troops which seized Li Yüan’s deputies came from an army nominally under Shih-min’s command but actually led by other generals.24 Second, whereas the Old T’ang History describes only Shih-min’s role in the pacification of Hsi-ho prefecture (located just to the southwest of Taiyuan) shortly following the T’ang revolt, the Diary shows that Shih-min’s elder brother Chien-ch’eng joined with him in the venture and deserved equal credit.25 The Diary serves to deflate Shih-min’s reputation still further. According to the T’ang Standard Histories, it was Shih-min who in 615 devised the strategy at Yen-men that freed Yang-ti from the Turkish siege. Shih-min’s activities at this frontier town have traditionally marked his 21  Ibid. 1.1–1b. The passage in question appears to contain an allusion to the “Li Ballad.” The text says: “Thinking that the common people of Taiyuan were the ‘old folk’ of T’ao T’ang, and [that the area for which] he had received his pacification commission did not exceed his own [nominal] fief, the emperor [Kao-tsu] was secretly joyous.” The way in which the “Li Ballad” alludes to T’ao T’ang (the culture-hero Yao) is discussed by Bingham, “The Rise of Li,” 277. 22   c yccc 1.5. At this time Kao-tsu’s sons, Li Chien-ch’eng and Li Yüan-chi, were still in southern Shansi. 23  Ibid. 1.15b. 24  Ibid. 1.9b. This episode is corroborated in hts 90.1. 25   c ts 1.3; cyccc 1.16b.

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

26

Wechsler

debut in Chinese historical records. Yet the Diary makes no mention of his role at Yen-men, recording only that Li Yüan led the army that helped free the Sui emperor.26 Significantly, the revisionists’ contention that long before the Taiyuan uprising Li Yüan had decided to turn against the Sui is supported by various biographies in the T’ang Standard Histories. The biographies of Li Ching in the Old and New T’ang History record that in 616, when Li Yüan was made assistant to the deputy prefect (ch’eng) of Ma-i commandery, Li Ching noticed that Li Yüan “had ambitions to conquer the empire (yu ssu-fang chih chih).” He even set out to report this information to Yang-ti but found progress impossible because the roads were blocked.27 The Old T’ang History biography of Liu Wen-ching, the official with whom Shih-min is said to have first plotted the T’ang revolt, records that the following year, when Li Yüan became garrison commander of Taiyuan, Liu similarly observed that the former had designs on the empire.28 Moreover, it is clear that about the same time no fewer than four of Li Yüan’s subordinates—Hsü Shih-hsü, T’ang Chien, Wu Shih-huo, and Ts’ui Shan-wei— were all exhorting him to revolt.29 The Old T’ang History biography of Wu Shihhuo notes that when Li Yüan became garrison commander and the empire was daily falling into greater disorder, Wu secretly advised him to raise troops and presented him with lucky charms and treatises on military strategy. “Please do not say anything more,” Li Yüan replied. “Books on military strategy are forbidden, yet you still bring them to me, so I well understand your meaning. [In the future] we will grow rich together.”30 Yet another piece of evidence comes from the Old T’ang History biography of Yü-wen Shih-chi, younger brother of Yü-wen Hua-chi, Yang-ti’s assassin. Shih-chi had once served under Li Yüan when the latter was in the Sui employ at Ta-hsing-ch’eng. When Shih-chi surrendered to the T’ang early in 619 following his brother’s death, Li Yüan, who had already ascended the throne, turned to P’ei Chi and observed, “It has already been six or seven years since this person and I discussed the taking of the empire. Those like you all came after him.”31 26   c ts 2.1b; hts 2.1–1b; cyccc 1.1b; Bingham, The Founding of the T’ang Dynasty, p. 49, n. 82. Actually, both these versions of the Yen-men incident may be incorrect. The biographies of Hsiao Yü, cts 63.5b and hts 101.1b, note that the Sui I-ch’eng Princess, wife of Shih-pi Qaghan, leader of the Turks, persuaded her husband to lift the siege. 27   c ts 67.1b; hts 93.1–1b. 28   c ts 57.6. 29  See, respectively, cts 57.14, 58.1b, 58.12b; hts 91.14b. 30   c ts 58.12b. This story also appears in the P’an-lung-t’ai pei of Li Ch’iao (644–713); ctw 249. 7b–8. Wu Shih-huo was the father of the later Empress Wu. 31   c ts 63.16.

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

The Rise to Power of the T ’ ang Dynasty: A Reassessment

27

It thus seems apparent that long before Shih-min allegedly precipitated the Taiyuan uprising, Li Yüan was already envisioning the succession of the Li-T’ang house to the Mandate and was under great pressure from numerous quarters to raise a standard of revolt. It is hardly likely, then, that Shih-min single-handedly forced his father to take up arms against the Sui or even that he was the most significant contributor to this decision. Moreover, other bits of evidence widely scattered throughout the T’ang Standard Histories and the Comprehensive Mirror suggest, like Wen Ta-ya’s narrative, that once the decision to revolt had been made Li Yüan had more than ample ability and strength to lead the T’ang forces to victory. It is said, for example, that the first T’ang emperor won his wife’s hand in an archery contest by hitting the eyes of two peacocks painted on two gates with just two shots.32 In a furious battle against bandit forces in Shansi, while loosening a quiver of seventy arrows, he is reported to have killed seventy men, then to have built a mound of their bodies and prudently retrieved the arrows for reuse.33 Even if these stories are partly apocryphal, we know that prior to his appointment as garrison commander, Li Yüan had amassed a splendid record of successes on the battlefields of Shansi against both “bandit” marauders and the Eastern Turks. Furthermore, at the time of the T’ang revolt, Li Yüan seems still to have been quite vigorous of body, for although he was already past fifty, he is said to have fathered at least seventeen of his twenty-two sons after he became emperor!34

The Origins of the Conflicting Views

The conflicting narratives we have examined surrounding the launching of the T’ang dynasty give rise to two obvious questions: (1) why are the versions contained in the Standard Histories and Wen Ta-ya’s Diary so much at variance? and (2) why are there so many internal contradictions within the Standard Histories themselves? The answers to these questions are to a large extent related to the manner in which these records were compiled. At the very outset, it should be mentioned that Wen’s Diary is a very special source for the Sui-T’ang transition period, owing primarily to the author’s intimate association with the T’ang rebel movement. Wen’s native place 32  Ibid. 51.3. 33   h ts 1.2. Tuan Ch’eng-shih, Yu-yang tsa-tsu (853; Ts’ung-shu chi-ch’eng ed., 1937), ch. 1, p. 1, records the number of dead as eighty. 34   t ctc ch. 190, pp. 5957–58, and note of Hu San-hsing.

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

28

Wechsler

(pen-kuan) was Taiyuan. Sometime prior to Li Yüan’s assignment there as garrison commander, Wen resigned his post as a Sui official and returned home. Subsequently, he joined Li Yüan’s staff as a secretary (chi-shih ts’an-chün) and accompanied the T’ang leader on the campaign from Chin-yang to Ta-hsingch’eng.35 His Diary, covering a period of 357 days from the Taiyuan uprising to Li Yüan’s enthronement in the Sui capital, thus represents nothing less than an eyewitness account of the T’ang founding. Overwhelmingly, the evidence suggests that Wen’s work is genuine and that it was written shortly after the events it narrates. Judging by the official titles which precede Wen’s name in various notices of the Diary still extant, it was first compiled during 617–18; a later version was probably made sometime during Kao-tsu’s reign, 618–26. This view is reinforced by the existence of the titles of what may have been earlier versions of the work: the Record of the Great Chancellor Prince of T’ang and His Officials (Ta-ch’eng-hsiang T’angwang kuan-shu chi) in two chüan, and Record of the Kingly Enterprise of the Present Ruler (Chin-shang wang-yeh chi) in six chüan. Great Chancellor Prince of T’ang is a title Li Yüan held during the last Sui reign, I-ning (617–18), presided over by a Sui puppet emperor he had placed on the throne. The Diary’s avoidance of T’ang taboo names, references to the second Sui emperor by his posthumous name, Yang-ti, and to Shih-min as Prince of Ch’in, a title he held after the beginning of 618, demonstrate that Wen’s final version of his text was not completed until sometime after the founding of the T’ang. The work is mentioned in the Conspectus of History (Shih-t’ung) of Liu Chih-chi, written in 710, and is included in the “Monograph on Literature” (ching-chi chih and i-wen chih) of the Old and New T’ang History, each of which was based on catalogues of books held in the Imperial Library dating from the K’ai-yüan period (713–42).36 From all this we may conclude that the Diary was not a forgery of later date. It has been pointed out, moreover, that the Diary may well contain an objective treatment of Shih-min’s role in the establishment of his house. First, it was probably written prior to Shih-min’s assassination of his elder brother, the crown prince Chien-ch’eng, and his usurpation of the throne in 626,37 thus before any need arose for him to alter the historical records in such a way as to 35   c ts 61.1–1b. 36  Liu Chih-chi, Shih-t’ung [Conspectus of History] (710; Shanghai, 1928), p. 47; Woodbridge Bingham, “Wen Ta-ya: The First Recorder of T’ang History,” Journal of the American Oriental Society, 57 (1937), 372–73; Fukui Shigemasa, “Ō Tō sōgyō kikyochū kō” [A Study of the Diary of the Founding of the T’ang Dynasty], Shikan, 63–64 (1961), 83–88; Lo Hsianglin, T’ang-tai wen-hua shih, pp. 1–3. 37  See below, pp. 67–77.

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

The Rise to Power of the T ’ ang Dynasty: A Reassessment

29

lend legitimacy to his rule. Second, on the eve of the assassination plot against Chien-ch’eng, Shih-min ordered the author of the Diary, Wen Ta-ya, then his own subordinate, to garrison his base at Loyang on the northeastern plain, and sent him secret messages there.38 Had Shih-min read Wen’s Diary (a point to be raised again shortly) and had Wen’s judgment of him therein been unduly harsh, it is hardly likely that he would have demonstrated such confidence in Wen. The present title appended to Wen Ta-ya’s work, Ta-T’ang ch’uang-yeh ch’ichü-chu, makes it the only example surviving in its entirety of the T’ang “diary of activity and repose” (ch’i-chü-chu) genre. During the T’ang such diaries were compiled four times a year (once each season) from daily records of the emperor’s actions and words,39 and served as raw materials for two other kinds of historical compilations, the “veritable records” (shih-lu) and “dynastic history” (kuo-shih). Yet Wen’s Diary differs from orthodox examples of ch’i-chü-chu in that it was privately rather than officially compiled and covers a period of time completely anteceding the inauguration of the dynasty. Perhaps because of this, it was not kept at the History Office (shih-kuan) with all the other records of its kind but was included as a separate title in the Imperial Library. This is probably how it was able to survive the conflagration during the An Lu-shan rebellion of 755–56, which destroyed 3,682 chüan of diaries and numerous other records stored in the History Office.40 The complex process by which the T’ang Standard Histories were compiled has been traced in some detail by Pulleyblank and Rotours, among others.41 Simply speaking, it is apparent that the Old T’ang History narration of the T’ang founding was ultimately based on the Kao-tsu Veritable Records (Kao-tsu shihlu) and T’ai-tsung Veritable Records (T’ai-tsung shih-lu).42 The New T’ang History 38   c ts 61.1b; hts 91.1b. 39  Chu Hsi-tsu, “Han-T’ang-Sung ch’i-chü-chu k’ao” [A Study of the Diaries of Activity and Repose of the Han, T’ang, and Sung Dynasties], Kuo-hsüeh chi-k’an [Journal of Sinological Studies, Peking National University], 2 (1930), 634. 40   t hy ch. 63, p. 1095; letter from Denis Twitchett to Robert Somers, 29 September 1970, used by permission of the recipient. 41  See E. G. Pulleyblank, “The Tzyjyh Tongjiann Kaoyih and the Sources for the History of the Period 730–763,” Bulletin of the School of Oriental and African Studies, 13 (1950–51), 448–57; Robert des Rotours, Le traité des examens (Paris, 1932), pp. 56–71. 42  Chao I, Nien-erh-shih cha-chi [Detailed Notes on the Twenty-two Histories], 2 vols. (1795; Taipei, 1965), vol. 1, ch. 16, pp. 214–16. The veritable records were, of course, utilized by Wei Shu, compiler of the T’ang-shu, a dynastic history covering the period 618 to the mid-eighth century that was for the most part incorporated verbatim into the Old T’ang History; see Pulleyblank, “The Tzyjyh Tongjiann Kaoyih.” Ssu-ma Kuang quotes from both

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

30

Wechsler

represents a revision of the Old, making use of many materials unavailable at the time its predecessor was compiled. Yet these new materials relate chiefly to the period after the An Lu-shan Rebellion, not to the early T’ang. The Kaotsu and T’ai-tsung Veritable Records thus provided both of the T’ang Standard Histories with the bulk of their materials on the establishment of the dynasty. Compilation of the Kao-tsu Veritable Records, which covered the period 618–26, and the first half of the T’ai-tsung Veritable Records, covering the period 626–40, began sometime in the late 630s or early 640s. While the compilation of these materials was proceeding, the ruler by custom did not interfere in any way with the work of the historians lest he impair their veracity. By the late 630s T’ai-tsung had been on the throne for more than a decade and had achieved a remarkable record of accomplishment. Yet his murder of Crown Prince Chien-ch’eng and his usurpation of the throne had sullied an otherwise exemplary reputation. It seems, therefore, that he decided to influence the historical materials surrounding his rise to power to make it appear that he, even more than his father or elder brother, deserved to inherit the Mandate. T’ai-tsung amply demonstrated his anxiety over his historical image on various occasions. At such times he made it clear to the historians who were compiling the records of his father’s reign and of his own that he was more than just routinely interested in the outcome of their labors. About the year 641 he said to Ch’u Sui-liang: “Since you recently have been an official in charge of recording the deeds and actions of the emperor (chih ch’i-chü),43 what kinds of affairs have you recorded? Generally, is the ruler allowed to examine [the records] or not? We wish to read these records so that We may take as a warning what they consider to be Our successes and failures.” Sui-liang replied: “The present recording officials [correspond to] the ancient historians of the left and right. In recording the ruler’s words and actions, good and evil must be written down so that the ruler will not act improperly. I have never heard that rulers could themselves examine the histories [of their reigns].” T’ai-tsung said: “If We have bad points, must you record them?” Sui-liang replied: “I have heard that ‘it is better to fulfill the duty of one’s the Kao-tsu and T’ai-tsung Veritable Records in his k’ao-i, or “investigations of discrepancies” sections, now usually interspersed throughout the text of the Comprehensive Mirror. 43  Robert des Rotours, Traité des fonctionnaires et traité de l’armée, 2 vols. (Leiden, 1947), 1: 153, notes that the chih ch’i-chü designation meant that the holder occupied the posts either of grand secretary of the Chancellery or remonstrating counselor. Ch’u was appointed remonstrating counselor in 641; cts 80.1b.

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

The Rise to Power of the T ’ ang Dynasty: A Reassessment

31

office than an obligation towards one’s ruler.’44 My duty in office is to uphold the brush, so how could I not record them?”45 On yet another occasion T’ai-tsung approached Fang Hsüan-ling, an official in charge of supervising the compilation of dynastic history, and said: “When We read the histories of former dynasties, [the way in which] they ‘distinguish the good so as to make it bad for the evil’46 is sufficient to provide Us with a warning for the future. [Thus] We do not know why since ancient times rulers have not been allowed personally to read the dynastic history (kuo-shih) of their reigns.” [Fang] replied: “Since the dynastic history must record good and evil so that the ruler will not act improperly, [the compilers] fear that they will offend him. Therefore, he may not see them.” T’ai-tsung said: “Our reasoning is quite different than that of the men of old. If We now wish to read the dynastic history, it is because if there are good deeds [recorded therein] they need not be discussed, but if there are faults, We wish to use them as a mirror-warning by which to improve Ourself. You are to compile the records and present them.”47 At this time the remonstrating counselor (chien-i ta-fu) Chu Tzu-she vehemently protested T’ai-tsung’s attempt to inspect the records. Ssu-ma Kuang laconically records only that “the emperor did not pay heed,” and that he ordered Fang Hsüan-ling and his fellow historians, Hsü Ching-tsung and others, to edit the dynastic history into chronicle (pien-nien) form, the results of which were the Kao-tsu and T’ai-tsung Veritable Records.48 It is apparent, then, that during the period of their compilation both the dynastic history and veritable records received T’ai-tsung’s strong editorial influence. Yet when the veritable records were finally completed and presented to the emperor in the seventh 44  From the Tso-chuan, Duke Chao, 20th year; see James Legge, The Chinese Classics, 2nd ed. rev., 5 vols. (Hong Kong, 1961), vol. 5, The Ch’un Ts’ew with the Tso Chuen, p. 684. 45   c kcy 7.7b–8. The ckcy gives the year of this episode as 639. thy ch. 63, p. 1102 and tfyk 554.25b both place it in 642. Ch’u Sui-liang’s biography in cts 80.1b–2 gives the year 641. For a related episode, see ckcy 6.18b–19. The ckcy has been the subject of a master’s thesis; George Winston Lewis, “The Cheng-kuan Cheng-yao: A Source for the Study of Early T’ang Government” (University of Hong Kong, 1962). 46  Legge, The Chinese Classics, vol. 3, The Shoo King, p. 573. 47   c kcy 7.8–8b; see also tctc ch. 197, p. 6203; thy ch. 63, p. 1103. 48   t ctc ch. 197, p. 6203; ckcy 7.8b; thy ch. 63, p. 1102. The sources do not agree on the year in which this episode occurred. The biography of Chu Tzu-she, cts 189A.10b, lists the year of his death as 641.

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

32

Wechsler

month of 643, they evidently still were not written in a manner calculated to please him. On the contrary, he claimed that he was unsatisfied with the way in which events surrounding his assassination of the crown prince were narrated: Long ago, when the Duke of Chou killed [his brothers] Kuan[-shu] and Ts’ai[-shu] the House of Chou was made peaceful, and when Chi-yu poisoned [his elder brother] Shu-ya the state of Lu was made tranquil. My action was as righteous as theirs because it has brought security to the state and benefit to all the people. Why then do the historians obscure [this fact] with their brushes? They should delete their embellishments and write a true account of the affair.49 Although the emperor ordered that the historians rectify their narrative concerning the circumstances surrounding Chien-ch’eng’s death, probably far more than just these sections were altered at the time. Ironically, Wei Cheng’s concern that history record the truth may have prompted him to become an unwitting accomplice in T’ai-tsung’s efforts to tamper with the record. Wei, who by this time had achieved a substantial reputation in his own right as a historian, sent a memorial to the throne applauding T’ai-tsung’s emendation order. At first glance this may seem a rather strange mode of behavior for an official renowned for his fearless and forthright opposition to all irregularities in his prince’s conduct. However, the reason Wei gave in his memorial for supporting T’ai-tsung’s action—that the purpose of history was to condemn evil and to encourage good and that if the narratives of the period were “not written truthfully” (shu pu i shih), then posterity would have no way of learning from the past50—suggests that he may not have been in full possession of all the facts regarding the case. Certainly, it is difficult to believe that he was consciously advocating that his fellow historians falsify their accounts simply to conceal the darker aspects of T’ai-tsung’s rise to power and thereby to enhance his place in history. One historian, however, who may have been persuaded by the emperor to do exactly this was a chief compiler of dynastic history and veritable records for the reigns of Kao-tsu and T’ai-tsung, Hsü Ching-tsung (592–672).51 Hsü had 49   c kcy 7.8b; see also tctc ch. 197, p. 6203. 50   c kcy 7.9. For a discussion of Wei’s historiographical beliefs, see below, pp. 138–39. 51  The exact nature of Hsü Ching-tsung’s contributions to the Kao-tsu and T’ai-tsung Veritable Records remains unclear, but there is no doubt that they were quite extensive. The New T’ang History “Monograph on Literature” (hts 58.11b) notes that Ching Po compiled, Fang Hsüan-ling supervised, and Hsü Ching-tsung revised the Kao-tsu Veritable

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

The Rise to Power of the T ’ ang Dynasty: A Reassessment

33

been a colleague of Wei Cheng’s at the camp of the rebel Li Mi during the last years of the Sui, and would, long after Wei’s death, gain a dubious reputation as the chief hatchet man of Empress Wu. Hsü was notorious for his avarice, even betrothing, it is said, one of his own daughters to the son of the Man “barbarian” chief, Feng Ang, in return for a considerable amount of gold and other valuables.52 But even more important for the present discussion, Hsü was an unscrupulous historian, numerous instances in which he distorted his records having been well documented.53 At T’ai-tsung’s behest, and provided with sufficient means to make it worth his while, it seems possible that Hsü might have made his accounts appear as if T’ai-tsung alone had masterminded the Taiyuan revolt. When work was begun on compiling the dynastic history of the early T’ang, of the four key people involved in the revolt according to the traditional narratives—Kao-tsu, T’ai-tsung, Liu Wen-ching, and P’ei Chi—only T’ai-tsung was still alive.54 Moreover, Hsü’s penchant for dissembling was apparent at least to the third T’ang emperor, Kao-tsung, who in 673 read accounts of his predecessors’ reigns that had been written in large part by Hsü, found several passages he knew to be false, and ordered them corrected.55 Kao-tsung was evidently unaware of the real story of the Taiyuan uprising and of the exaggerated accounts of T’ai-tsung’s contributions to the founding of the dynasty. So, although certain offending sections in the historical records were changed at his order, probably only the most blatant errors were corrected, since it was not known precisely which passages had been emended by Hsü, then dead. Moreover, it is apparent that Hsü Ching-tsung, or whoever else lent a hand in falsifying the dynastic history and veritable records, could not deal effectively with all the material that eventually comprised the narration of events Records. The veritable records for T’ai-tsung’s reign up to 640, known as the Chin-shang shih-lu, are listed as having been compiled by Ching Po and Ku Yin and supervised by Fang Hsüan-ling; hts 58. 11b–12. But there are several other places where Hsü is listed as a contributor to the work; see thy ch. 63, p. 1092 and Li Shu-t’ung, T’ang shih k’ao-pien, p. 30. Even more intriguing is the information provided by the Sung scholar Ch’en Chen-sun. He notes that title pages of his own editions of the Kao-tsu and T’ai-tsung Veritable Records list Hsü Ching-tsung alone as having received the order to compile them, although he concludes that this information is probably false; Chih-chai shu-lu chieh-t’i [Annotated Catalogue of the Chih Library] (ca. 1235; Ts’ung-shu chi-ch’eng ed., 1937), ch. 4, p. 117. 52   c ts 82.2b. 53  A long list of historical falsifications perpetrated by Hsü appears in his biographies, cts 82.3b–4 and hts 223A.3b–4, and also in thy ch. 63, pp. 1093–94. See also Chao I, Nienerh-shih cha-chi, vol. 1, ch. 16, p. 211, and Pulleyblank, “The Tzyjyh Tongjiann Kaoyih,” 451. 54  Li Shu-t’ung, T’ang-shih k’ao-pien, p. 31. 55   t hy ch. 63, p. 1093; tctc ch. 202, p. 6371.

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

34

Wechsler

surrounding the T’ang uprising and its aftermath in the two T’ang Standard Histories and their intermediate compilations, thus giving rise to numerous internal contradictions in the texts. Already by the eleventh century Ssu-ma Kuang was noting in his work entitled Investigations of Discrepancies (K’ao-i) inconsistencies and outright contradictions he had encountered in his sources for the early T’ang period while compiling the Comprehensive Mirror. He perceived, for example, that T’ai-tsung’s contributions to the Taiyuan uprising and Ta-hsing-ch’eng campaign had been exaggerated at the expense of the crown prince and partially took account of this fact in compiling his own narrative.56 But although he occasionally made use of information culled from Wen Ta-ya’s Diary, he apparently failed to be persuaded by its central thesis and chose to perpetuate the traditional interpretation of the T’ang founding. Despite evidence to the effect that T’ai-tsung altered the historical records in his own favor, a few nagging questions remain. Because Wen’s Diary is biased in favor of Kao-tsu,57 who is portrayed virtually without fault, it may be that the version of events recorded in it are as distorted as those in the Standard Histories. If the Diary was written during Kao-tsu’s reign, what would have been more natural than to glorify the record of a “good first” emperor? Since Li Chien-ch’eng was the T’ang heir at the time the Diary was compiled, is it possible that Wen Ta-ya exaggerated his role in the Taiyuan uprising and its aftermath so that he would compare favorably with Shih-min? It seems strange that T’ai-tsung failed to suppress the Diary even after he began to take a supervisory role in the compilation of the dynastic history and veritable records. Did he underestimate the Diary’s threat to his historical image because only the dynastic history and veritable records were intended to be transmitted to posterity? Did he never himself read the Diary since it was a short, private account of the founding of his house, and so remain completely ignorant of its content and the necessity for its destruction? All these questions, unhappily, remain unresolved.

Li Yüan Rehabilitated

The above imponderables cannot, of course, negate the impressive body of evidence in the T’ang Standard Histories we have already reviewed supporting 56  See tctc ch. 184, p. 5738. 57  This bias has been noted by several scholars. See Ssu-ma Kuang’s note in tctc ch. 184, p. 5737; Bingham, The Founding of the T’ang Dynasty, p. 120; Fukui, “Ō Tō sōgyō kikyochū kō,” 86–87; Nunome, “Tensaku jōshō,” 3, 13.

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

The Rise to Power of the T ’ ang Dynasty: A Reassessment

35

Wen Ta-ya’s contention that Li Yüan himself conceived and led the Taiyuan revolt. Nor did Li Yüan’s role in the founding of the T’ang dynasty simply end here. On the contrary, the strong qualities of leadership he evinced at the very outset of his revolt continued greatly to benefit the T’ang cause even afterwards. A strong case, I believe, can be made that Li Yüan made a substantial contribution first to the capture of the Sui capital at Ta-hsing-ch’eng and later to the pacification and unification of the entire country under T’ang rule. Particularly impressive, it seems to me, was Li Yüan’s ability to parry successive challenges to his power from various quarters. Initially, the gravest of these challenges came from the Eastern Turks, the effective rulers of northern Asia, who controlled territory north of the Great Wall stretching from modern Liaoning province to western Mongolia. Seeing profit in a weak and divided China, the Turks had compelled many anti-Sui rebels in North China to declare themselves Turkish vassals and had provided them with soldiers, arms, and precious horses for their campaigns.58 Shortly after Li Yüan executed his deputies Wang Wei and Kao Chün-ya in the fifth month of 617, the Turks invaded the Taiyuan region and jolted the T’ang camp by advancing as far as the walls of Chin-yang before retreating. Clearly, it would have been arrant folly for Li Yüan to quit Taiyuan—a major bastion of defense against the Turks—on a campaign to conquer China without first reaching an accommodation with them. He therefore sent a letter to the Turkish qaghan, Shih-pi, professing a desire to save the Sui and, to that end, offering to restore harmonious relations with the Turks. The proposal was sweetened with the promise that if Shih-pi allied with Li Yüan, all the booty to be gained from the campaigns against the rebels would be turned over to the qaghan.59 The letter was very respectful in tone, and when it was completed Li Yüan affixed to it the character ch’i, “communication from inferior to superior,” a move ethnocentric Chinese commentators have taken to mean that Li, like his rivals for the Mandate, had for the sake of expediency decided to become a Turkish vassal. Recently, some controversy has arisen as to whether this was indeed the result.60 What is significant, however, is that the maneuver succeeded in 58  According to hts 215A.6 and Tu Yu, T’ung-tien [Comprehensive Statutes] (801; Taipei, 1966), ch. 197, p. 1069a, numbering among these rebels were Hsüeh Chü, Li Kuei, Liu Wu-chou, Liang Shih-tu, Wang Shih-ch’ung, Tou Chien-te, and Kao K’ai-tao. 59   c yccc 1.11b–12. 60  Opposing views on this question are found in Ch’en Yin-k’o, “Lun T’ang Kao-tsu ch’engch’en yü T’u-chüeh shih” [On T’ang Kao-tsu Calling Himself a Subject of the Turks], Lingnan hsüeh-pao, 11 (1951), 1–9; Li Shu-t’ung, T’ang-shih k’ao-pien, pp. 214–46, and idem, “Tsai-pien T’ang Kao-tsu ch’eng-ch’en yü T’u-chüeh shih” [A Further Examination of T’ang

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

36

Wechsler

its objective. It created a rapprochement between the T’ang and the Turks that worked to Li Yüan’s advantage and allowed him a crucial breathing spell during which he was able to strengthen his forces and plan the strategy of his advance. With the Turkish threat out of the way for the moment at least, Li Yüan could turn to the task of military organization, which he completed in the sixth month. The executive and administrative arm of this organization was known as the Administration of the Grand General (ta-chiang-chün fu) and was staffed largely by incumbent or former Sui military and civil officials working or residing in the Taiyuan vicinity. The T’ang army was initially recruited and supplied by the efforts of these men. In the building of his army, Li Yüan appears to have made excellent use of the Sui militia organization (ying-yang fu) in Taiyuan. Since several members of the Administration of the Grand General were officers in the Sui militia, they simply transferred militia troops to Li Yüan’s command, thereby swelling his ranks.61 P’ei Chi, the assistant superintendent of the Sui palace at Chin-yang, provided the T’ang forces with abundant supplies of grain and armor from the palace storehouses. The first major military objective of Li Yüan was the Sui capital, about three hundred and fifteen miles as the crow flies to the southwest of Taiyuan, for which he set out in the seventh month, marching down the Fen River valley toward the T’ung Pass, the gateway to Shensi. The capture of Ta-hsing-ch’eng was important for two reasons. Politically, it represented a symbol of dynastic legitimacy—the passing of the Mandate from the Sui to the T’ang. Strategically, it commanded the Shensi plain, which Li Yüan, like other dynastic founders before him—Ch’in Shih-huang-ti, Han Kao-tsu, and Sui Wen-ti—planned to use as a springboard for the conquest of all China. At Huo-i, a town in the southern Fen River valley, Li Yüan met with yet another major threat to his advance. He had called a halt at Huo-i because of strong Sui resistance and torrential late summer rains. While waiting for the weather to clear, he received a message from Li Mi, the powerful Honan rebel leader and fellow beneficiary of the “Li Ballad,” proposing an alliance with the T’ang. Li Yüan suspected that Li Mi would pose a formidable obstacle if he

Kao-tsu Calling Himself a Subject of the Turks], Ta-lu tsa-chih [The Continent], 37 (1968), 248–66. 61  Nunome Chōfū, “Ri En shūdan no kōzō” [The Structure of Li Yüan’s Organization], Ritsumeikan bungaku [Ritsumeikan University Journal of Cultural Sciences], 243 (1965), 27–29. The rebels Liang Shih-tu and Li Kuei probably profited in similar fashion from their connections with the Sui militia; see tctc ch. 183, p. 5718, and ch. 184, p. 5745.

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

The Rise to Power of the T ’ ang Dynasty: A Reassessment

37

b­ ecame aware of his plans to take Ta-hsing-ch’eng.62 He therefore wrote a polite but deceptive letter to his rival in which he portrayed himself as a loyal subject of the Sui who had raised an army only to reestablish order and disclaimed any ambition other than that of being left with his noble title, Duke of T’ang. At the same time he encouraged Li Mi in his own efforts to topple the Sui: “Heaven has created the common people, who need a shepherd. Now, who else is that shepherd if not you!”63 Whether Li Mi was reassured by the letter or preoccupied by troubles in his own camp,64 we do not know. In any event, he allowed Li Yüan to march on Ta-hsing-ch’eng without hindrance—a move he subsequently came to regret. When the rains let up in the eighth month, the T’ang army was able to overcome Sui resistance at Huo-i and move on. During the battle of Huo-i, Li Yüan, it is said, contributed to the defeat of the Sui forces by circulating the rumor that their commander had been killed, thus throwing them into full rout.65 P’u-chou fu, a town in southern Shansi that controlled access to the T’ung Pass and, beyond it, the Shensi plain, was taken by T’ang forces in the ninth month of 617, largely, it appears, by the planning of Li Yüan.66 Perhaps one of the reasons for these early T’ang successes was the high morale among the T’ang troops, for Li Yüan speedily rewarded all those meritorious in battle regardless of their rank.67 The T’ang campaign to conquer Ta-hsing-ch’eng has been described in some detail elsewhere and need not be elaborated here.68 It is worth mentioning, however, that, judging from Wen Ta-ya’s Diary, Shih-min appears to have contributed no more to its capture than his elder brother Chien-ch’eng, perhaps even less. All during the march on the capital the two brothers were almost invariably appointed to military duties with comparable responsibilities. For example, both shared in the initial planning of strategy for the battle of Huo-i,

62  Li Mi certainly realized the strategic importance of capturing the capital region in any campaign to win the empire. He had so advised Yang Hsüan-kan after the former had revolted against the Sui, and had long been encouraged to capture the capital by his own officials; see cts 53.2 and tctc ch. 183, p. 5735. 63   t ctc ch. 184, p. 5743. 64  See below, p. 51. 65   c yccc 2.15. 66  Ibid. 2.21–21b. 67  Ibid. 2.15b–16. 68  See, especially, Bingham’s account in The Founding of the T’ang Dynasty, pp. 95–104 and the author’s treatment, “Kao-tsu the Founder,” in the forthcoming Cambridge History of China volume on the T’ang dynasty.

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

38

Wechsler

and both led troops to surround Ta-hsing-ch’eng.69 Even more important, when the city finally fell to the T’ang, troops under the command of Chien-ch’eng’s subordinate, Lei Yung-chi, were first to breach its walls.70 The notion, sedulously propagated by the T’ang Standard Histories and Comprehensive Mirror, that Shih-min almost single-handedly brought about a victorious conclusion to the Ta-hsing-ch’eng campaign, must now be laid to rest. Once inside the Sui capital, Li Yüan established a grandson of Yang-ti as a puppet emperor. Yet never for a moment was there any doubt that he would himself shortly assume the imperial mantle: he had already begun acting and speaking like Han Kao-tsu after that dynastic founder marched through the T’ung Pass,71 and was assuming the highest official and noble titles in the land preparatory to his accession. On the twentieth day of the fifth month, 618 (June 18, 618), the first anniversary by the Chinese calendar of the seizure of the Sui deputies Wang Wei and Kao Chün-ya and the effective beginning of the Taiyuan revolt, Li Yüan mounted the throne amid great pomp and ceremony. Immediately following his accession, Kao-tsu (as we will hereafter refer to him) appointed his eldest son, Chien-ch’eng, as the T’ang heir, named his second son, Shih-min, Prince of Ch’in, and his fourth son, Yüan-chi, Prince of Ch’i. The T’ang capital was established at Ta-hsing-ch’eng and renamed Ch’ang-an, a self-conscious act linking the T’ang to the last great Chinese empire of Han. If important questions have arisen concerning the nature of Shih-min’s role in launching the T’ang, there is nevertheless wide agreement that after the dynasty was formally proclaimed, Shih-min made major contributions toward the strengthening of T’ang military power and the destruction of rival contenders for the empire. Between 618 and 622, the Prince of Ch’in was almost constantly in the field as commander-in-chief (yüan-shuai) of numerous T’ang armies-on-campaign (hsing-chün) that were raised piecemeal to combat “rebel” groups. His defeat of Hsüeh Jen-kuo in the northwest and, especially, Tou Chien-te and Wang Shih-ch’ung on the northeastern plain,72 virtually assured the reunification of China by the T’ang. Because Kao-tsu led no troops in the field after the capture of the Sui capital, Shih-min’s role in the T’ang pacification effort during the next several years greatly overshadowed his father’s. It has thus not been sufficiently ­appreciated 69  See cyccc 2.13–15, 21–25b; Lo Hsiang-lin, T’ang-tai wen-hua shih, p. 17; Li Shu-t’ung, T’angshih k’ao-pien, p. 283. 70   c yccc 2.27; Li Shu-t’ung, T’ang-shih k’ao-pien, pp. 279–80; Lo Hsiang-lin, T’ang-tai wenhua shih, p. 19. 71  Fukui, “Ō To sōgyō kikyochū kō,” 88–89. 72  See below, pp. 64–66.

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

The Rise to Power of the T ’ ang Dynasty: A Reassessment

39

that Kao-tsu’s political and military strategies designed to secure the allegiance of various power groups throughout the country effectively complemented Shih-min’s military campaigns and thus greatly accelerated the T’ang reunification. China in the year 618 was a patchwork of contending authorities, great and small. Hundreds of “rebel” and “bandit” organizations had occupied territory of varying size and had established at least rudimentary governments to administer them. Many areas continued to remain under the jurisdiction of Sui civil and military officials who hoped to be confirmed in their power by a new dynasty. In still other localities, leaders among the gentry had raised private armies to oppose civil strife and were filling with varying degrees of success the administrative vacuum left by the fall of the Sui. All of these areas had to be brought under T’ang control, by military means if necessary, but preferably by persuasion. To this end, Kao-tsu promoted a three-pronged policy of amnesty, appointment, and reward to persuade his adversaries to capitulate. Those who voluntarily surrendered with their armies and territories as well as many of those defeated on the battlefield were spared their lives. The followers of those who for one reason or another were executed were usually pardoned. Important “rebel” leaders were allowed to incorporate their forces into the T’ang military as whole units remaining under their personal command,73 thus increasing their willingness to fight for the T’ang. Incumbent Sui local officials were often reconfirmed in their posts, and “bandit” or “rebel” leaders in control of given localities were often appointed to govern the identical areas as T’ang prefects.74 At court, the surname Li, which Kao-tsu liberally bestowed on important rebels, conferred high prestige on them, and granted them such privileges as being honored above officials of equal rank.75 Frequently, the emperor generously rewarded former foes and granted them high noble titles. The considerable number of voluntary surrenders to the T’ang and the few instances of revolt by those who had surrendered suggest that, beyond purely military considerations, Kao-tsu’s policies were effective in rallying support for his regime over much of the country.

73  See cyccc 2.16; 2.23. 74  See tctc ch. 185, pp. 5796, 5805; ch. 186, pp. 5807, 5808, 5827; ch. 187, pp. 5852, 5863, 5869; ch. 189, pp. 5921, 5929; ch. 190, p. 5944; cts 57.17, 69.13b; cyccc 2.19b. 75  See cts 1.9, 1.10, 1.10b, 1.11, for the cases of Hsü Shih-chi, Lo I, Liu Hsiao-chen, Tu Fu-wei, Kao K’ai-tao, and Hu Ta-en. tctc ch. 187, p. 5840, lists special privileges accorded those with the Li surname.

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

40

Wechsler

By early 624, Kao-tsu’s pacification policies, combined with the military campaigns led by Shih-min and other T’ang generals, brought disorder within the Great Wall to an end for the first time in a decade. The emperor now proclaimed a great amnesty and joyfully declared that China was once more at peace. Although the evidence presented in this chapter has undercut some of the foundations upon which Li Shih-min’s historical reputation rests, the view that he was one of China’s more extraordinary rulers remains unassailable. Yet he did not, it seems, possess the prescience and almost superhuman qualities the traditional narratives of the founding of his house have attributed to him. In toppling him from the “founding ruler” pedestal he has occupied for so long, we place him in a more accurate perspective not only in relation to his father, Kao-tsu, but also in relation to his courtiers, among whom stood Wei Cheng.

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

The T’ang Imperial Family Denis Twitchett The single most important political objective for the T’ang ruling house was ensuring the orderly succession to the throne and the continuity of their dynasty. In a system where so many powers were vested in the sovereign, it was vitally important that there should always be a capable man on the throne, able to provide leadership for his court and stability for the dynastic regime, as well as to mediate among his ministers and make essential decisions. The T’ang house, which came from the northwestern aristocracy, at first operated a style of government common among the northern peoples in which the emperor ruled through achieving whenever possible consensus among his council of minister-advisers. To do this effectively he also needed to exercise unquestioned authority over them, and when necessary to enforce his own decisions with or without their concurrence.

The Succession

The succession of an unsuitable ruler could be a disaster: a violent or tyrannical emperor could easily alienate not only the people as a whole but also the ministers and officials through whom he had to act; a feckless and inattentive ruler would allow his proper authority to be usurped by his ministers or worse by self-seeking and sycophantic favorites. The succession of a child-emperor raised especially difficult problems; it automatically required a regency, which would place the control of the court in the hands either of a dowager-empress, or of a dominant chief minister. Such a succession carried the threat of factional dissention, or worse, a threat to the dynastic family’s continued dominance. Moreover, the succession of a very young emperor could set off a chain reaction (on emperors’ lifespans, see table 1, appended). If he died young, and this was not unlikely, for an immature young emperor barely in his teens was presented with every sort of temptation to excess and self-indulgence, this could set the stage either for a further succession by an infant son or, if the young emperor died without male issue, for the selection of a prince out of the direct line of succession. In either case this could lead to serious factional disputes, and become the prelude to a series of successions outside the normal pattern. Source: “The T’ang Imperial Family,” Asia Major 3rd ser. 7.2 (1994): 1–61.

© koninklijke brill nv, leiden, ���9 | doi:��.��63/9789004380158_005

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

42

Twitchett

This happened, for example, between 820 and 846, when five emperors succeeded to the throne in a quarter century, at a time when the court and empire urgently needed stability and strong leadership. The throne was the only patrimony in the empire where the succession did not automatically pass from father to eldest son by the principal consort, or, if the eldest son had predeceased his father, either to his eldest grandson, or, if there were no grandsons, to a younger son. To avert the possibility of the succession to the throne of an inept or unsuitable descendant, and to avoid any chance of a disputed succession, it was usual to select and appoint an heirapparent (huang t’ai-tzu 皇太子; t’ai-tzu 太子) early in each reign (for a chronological list of heirs-apparent, and their rank by birth order among siblings, see table 2, appended). The emperor and his advisers could, in theory, designate any of the royal princes, and since many T’ang emperors had large numbers of sons, the choice was wide. But seniority was important: the older the succeeding emperor, the less the chance of political instability. Thus the eldest son was the normal first choice. Hsüan-tsung 玄宗 (r. 712–56), for example, was appointed heir-apparent by his father, Jui-tsung 睿宗 (r. 684–90, 710–12), in 710 only after his elder brother Li Ch’eng-ch’i 李成器, who had a better claim to the succession and had served as heir-apparent during Jui-tsung’s earlier reign under the tutelage of empress Wu 武后, had declined in his favor.1 At times, when there seemed a particular need to settle the succession beyond the next generation, the emperor could select one of the heir-apparent’s sons as the imperial grandson (huang t’ai-sun 皇太孫), although there was disagreement about the propriety of doing this while the heir-apparent was himself still alive.2 In such cases the infant prince was assigned preceptors and given 1  Liu Hsü 劉晦 et al., Chiu T’ang shu 舊唐書 (Peking: Chung-hua shu-chü, 1975; hereafter, cts) 95, p. 3010; Ssu-ma Kuang 司馬光, Tzu-chih t’ung-chien 資治通鑑 (Peking: Ku-chi ch’u-panshe, 1956; hereafter, tctc) 209, p. 6650. 2  Wang P’u 王薄, T’ang hui-yao 唐會要 (Shanghai: Shang-wu yin-shu kuan, 1935; hereafter, thy) 4, pp. 49–50. This first happened in 642, when T’ai-tsung was worried about public concern over the heir-apparent Ch’eng-ch’ien’s conduct and fitness to rule. To attempt to stabilize the situation he issued an edict ensuring that if the heir-apparent were to die, succession would pass to his five-year-old son, not to another of his sons by a concubine. Again, at the end of Kao-tsung’s reign in 682, after Li Hsien, the future emperor Chung-tsung, was installed as heir-apparent, his eldest son Li Chung-jun was appointed huang t’ai-sun, once again as a measure to stabilize the succession. When Chung-tsung was dethroned by empress Wu, and banished to Fang-chou in 684 this appointment was canceled. See cts 6, p. 116; 86, p. 2835; and Ou-yang Hsiu 歐陽修, Sung Ch’i 宋祁 et al., Hsin T’ang-shu 新唐書(Peking: Chung-hua, 1975; hereafter, hts) 81, p. 3593. Another case occurred when empress Wu finally seized the throne in 690. The former puppet-emperor Jui-tsung was reduced to the rank of imperial heir

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

The T ’ ang Imperial Family

43

an establishment of officials in the same way, though on a lesser scale, as the heir-apparent. The Heir-Apparent The installation of an heir-apparent was a major political act. It was conducted with elaborate ritual, by the emperor in full court,3 and the occasion was marked by a major Act of Grace (ta she 大赦) including an empire-wide amnesty, and often some days of “public license” (p’u 酺).4 (huang-ssu), and his eldest son, who had been heir-apparent, was given the title imperial grandson (huang-sun); see cts 95, p. 3009. 3  This ritual is specified in Hsiao Sung 蕭嵩 et al., [Ta T’ang] K’ai-yüan li 大唐開元禮 (Tokyo: Koten kenkyū kai, 1972; hereafter, kyl) installing heirs-apparent, see Sung Min-ch’iu 宋敏求, T’ang ta chao ling chi 唐大詔令集 (Peking: Shang-wu yin-shu kuan, 1959; hereafter, ttclc 27, pp. 93–96; and the brevets of appointment (ts’e), ttclc 28, pp. 98–101. thy 4, pp. 47–48, has a description of the ritual, “on a reduced scale because of the emperor’s being unwell,” observed at the installation of Mu-tsung’s heir-apparent in the 12th mo., 822 (translated following; a fuller account is given in Wang Ch’in-jo 王欽若 et al., Ts’e-fu yüan-kuei 冊府元龜 [Peking: Chung-hua, 1960; hereafter, tfyk] 257, pp. 15a–16b).  The emperor presided “beneath the eaves”; the ceremony was accompanied by the full court orchestra (kung-hsien) and the entire imperial bodyguard were deployed. The courtiers assembled by rank outside Hsüan-cheng Gate, took some food in the covered galleries, and then dressed in their full court robes and caps, wearing their swords and sandals. Then they entered the main courtyard (cheng-ya) through Yüeh-hua Gate. A quarter-hour after dawn they entered the court; the emperor took his place beneath the eaves attended by the ranks of eunuchs. The heir-apparent approached on foot from Ch’ung-ming Gate followed and protected by the officials of the palace, two of the consorts of royal princesses ( fu-ma) supporting his robes and cap, and the ritual commissioner (li-i shih) conducted him toward the emperor. When he arrived the orchestra played, the fan insignia were opened, and the assembled courtiers made their obeisance. The heir-apparent advanced as far as the southeastern Dragon’s Head (these ch’ih-t’ou, were the sculptured heads of dragons ornamenting the stone balustrade of the great stairway leading up to the audience hall), made a double obeisance and received the brevet of investiture. The provisional president of the Secretariat Tu Yüan-ying knelt and read the text of the brevet, and when he was finished gave it to the heir-apparent, who again made a double obeisance and performed a dance (wu-tao), before retiring to his royal pavilion (mu-tien) at Ch’ung-ming Gate. When the assembled officials had finished offering their congratulations to the emperor, they went to Ch’ung-ming Gate to pay their respects to the heir-apparent. The heir ordered the curtains of his pavilion be raised and returned their obeisance holding his tablet of office. The palace officials bowed and received the tablet from him. 4  On major Acts of Grace, see Brian McKnight, The Quality of Mercy (Honolulu: U. of Hawaii P., 1981). For a sample of Acts of Grace issued to commemorate the installation of a new heirapparent, see ttclc 29, pp. 103–7.

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

44

Twitchett

Subsequently, his “capping” (kuan 冠), the ceremony marking his coming to maturity, and his assuming “adult dress” (yüan-fu 元服),5 were also occasions of great state ritual. His marriage was also a major dynastic decision, for it involved the selection of a principal consort ( fei 妃) who would probably become the future empress, and it too was accompanied by very elaborate ceremonies.6 Once the sovereign-in-waiting was selected, he was placed in the care of experienced statesmen who would act as his mentors, and took up residence in his own palace, the Eastern Palace (Tung-kung 東宮; sometimes called Ch’ing-kung 青宮), where he presided over a large staff of officials, deliberately designed as a microcosm of the central government, to accustom him to the routine of governing with royal dignity. (This organization is described in detail, below.) From time to time, if the emperor was temporarily incapacitated or was absent from the capital on a military campaign or a royal progress (hsün-shou 巡守, “tour of inspection”),7 the heir-apparent would be appointed temporary regent (chien-kuo 監國), in order to give him experience of actually governing the empire, of course under the careful supervision of the highest ranking ministers at court.8 At other times he might be deputed the responsibility of reviewing criminal sentences or deciding minor legal disputes with the same end in mind.9 It was important that the heir-apparent’s position of unquestioned superiority over the rest of the imperial sons be maintained. Early in the T’ang, as we see below, every prince of the blood (ch’in wang 親王) had his own personal household establishment comprising a princely administration (wang-fu 王府), a group of tutors and adviser-companions, and his personal guardsunits. The princes and royal princesses (kung-chu 公主) also had offices to administer the fiefs of maintenance that provided their revenues. Great pains were taken to prevent other princes and princesses from flaunting their power in excess in an attempt to rival or overshadow the 5  t hy 26, pp. 495–96; kyl 110; ttclc 29, pp. 107–8. 6  k yl 111. For edicts appointing consorts for heirs-apparent, see ttclc 31, pp. 120–21. 7  On royal progresses, see Howard J. Wechsler, Offerings of Jade and Silk: Ritual and Symbol in the Legitimation of the T’ang Dynasty (New Haven: Yale U. P., 1985), pp. 161–69. 8  For collections of edicts appointing the heir-apparent as temporary regent, see ttclc 30, pp. 110–15; tfyk 259, pp. 3b–9b. Until the 8th c. the term used for such a temporary regency was chien-kuo. The heir might also sometimes be deputed with “the duty of deciding all important matters, military and civil” (ch’u-chüeh chün-kuo chi-wu). In the 9th c. the common term used was kou-tang chün-kuo cheng-shih, “manage the civil and military affairs of the empire.” 9  E. g., tfyk 259, pp. 4a–b, 6a–7a.

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

The T ’ ang Imperial Family

45

heir-apparent.10 A careful balance needed to be struck between allowing the princes their proper dignity, and ensuring that they and their entourages did not constitute a political faction, or in any way threaten the heir’s succeeding to the throne. In the eighth century the personal establishments of princes and princesses were gradually whittled away, the ranks of their advisers reduced almost to vanishing point, and their revenues cut to the bone. The heir-apparent alone among the royal children kept his large establishment and this problem of potential rivalry receded. Even after the formal installation of an heir-apparent had been carried out, the emperor could change his mind, and replace him.11 Removing an heirapparent was an extremely serious act. Although the empress Wu did this twice, and also arbitrarily dethroned two young emperors to secure her own power, she was, as in so many areas of politics, an exception. In normal circumstances an established heir-apparent could not be replaced arbitrarily, but the act had to be justified by accusations either that the heir-apparent was plotting to usurp his father’s throne by armed force (T’ai-tsung 太宗, r. 626–49, having come to power in just this way); or that he was building his own political faction with a view to usurpation; or that he was irretrievably delinquent or mentally incompetent. Such accusations cannot, in most cases, be tested for lack of surviving evidence. But the fact that they were made at all shows that there was a common perception that any decision changing the succession needed powerful moral or political justification to make it generally acceptable. The emperor usually consulted with his ministers about such cases, but most officials were unwilling to accept responsibility for such a potentially dangerous decision, and deferred to the emperor’s opinion, claiming that the succession was a personal or family matter. Immediately before an emperor’s death, an imperial testamentary edict (i-chao 遺詔, or i-shu 遺書) was promulgated, usually confirming the choice that had been made when the current heir-apparent had been installed.12 The 10  See the memorials of Kao Chi-fu, cited in thy 26, pp. 497–98; others quoted in Wu Ching 吳棘, Chen-kuan cheng-yao 貞觀政要, ed. Harada Taneshige 原田種成, Jōgan seiyō teihon 貞觀政要定本 (Tokyo: Tōyōbunka kenkyūjo kiyō, 1962; hereafter, ckcy) 4, pp. 113–31; those by Ch’u Sui-liang in Li Fang 李昉 et al., Wen-yüan ying-hua 文苑英華 (Peking: Chung-hua, 1966; hereafter, wyyh) 623, pp. 10a–11b; cts 76, p. 2654. 11  For convenient collections of documentation on the succession, see thy 4, pp. 39–44, under “Ch’u-chün 儲君,” and tfyk 227, pp. 1a–19a. 12  The testamentary edicts of many of the T’ang emperors are collected in ttclc 11–12, pp. 66–73. The testaments of emperors who abdicated (i-kao) are in ttclc 12, pp. 73–74.

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

46

Twitchett

testament was read by a senior member of government in the presence of the deceased emperor’s encoffined body in one of the great audience halls, and the new emperor succeeded to the throne “before his father’s coffin,” as a solemn public sign that the deceased emperor’s will was being done. In the ninth century, as we shall see, a succession of emperors came to the throne only days after having been appointed heir-apparent, and there were well-founded suspicions that the whole process, the selection and installation of the heir, and the composition of a testamentary edict had been hastily cooked up by his palace eunuchs after the late emperor’s death or while he was incapacitated by terminal illness. Before his succession, in normal times, the heir would have spent some years in preparation for the responsibilities of his new role. From the very beginning of the dynasty there was a clear perception of the necessity for this, clearly expressed for example, in a long memorial presented to T’ai-tsung by Liu Chi 劉洎 (d. 645) shortly after the future Kao-tsung 高宗 had been appointed as heir-apparent.13 A very complex administrative organization was established to give the future emperor training and personal experience of officialdom and of the routines of administering the empire. The Heir-Apparent’s Administration The heir-apparent to the throne headed a personal establishment known as the Eastern Palace (Tung-kung), which was a copy in miniature of the government of the emperor.14 The idea behind this was not simply that he should live in a grand style that set him apart from the other royal princes, but that by superintending the activities of a large and varied household he would be familiarized with the solemn routines and procedures that would face him every day after his accession. Like the emperor, he had provision for the following posts:

13  For this important memorial see thy 4, pp. 39–40; cts 74, pp. 2609–11; ckcy 4, pp. 120– 24; Ma Tuan-lin 馬端臨, Wen-hsien t’ung-k’ao 文獻通考, in Shih t’ung 十通 (Shanghai: Shang-wu, 1935; hereafter, whtk) 623, pp. 5a–6b; and Tung Kao 董詰 et al., [Ch’in-ting] Ch’üan T’ang wen [欽定]全唐文 (Taipei: Hua-wen shu-chü, 1965; hereafter, ctw) 151, pp. 2a–4b. 14  See Li Lin-fu 李林甫 et al., Ta T’ang liu-tien 大唐六典 (Taipei: Wen-hai ch’u-pan she, 1962; hereafter, tlt) 26, pp. 27–28; thy 67, pp. 1167–72; Tu Yu 杜佑, T’ung tien 通典 (hereafter, tt), printed in Shih t’ung (see prev. n.) 30, pp. 171–76; cts 44, pp. 1906–13; hts 49A, pp. 1291–1303 (trans. Robert des Rotours, Traité des fonctionnaires et traité de l’armée: Traduits de la Nouvelle histoire des T’ang [chaps. XLVI–L] [Leiden: E. J. Brill, 1947–48; hereafter, Fonct.] 2, pp. 570–628).

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

The T ’ ang Imperial Family

47

1. three preceptors (san-shih): grand preceptor to the heir-apparent (t’ai-tzu t’ai-shih 太子太師), grand adviser to the heir-apparent (t’aitzu t’ai-fu 太傅), and grand guardian of the heir-apparent (t’ai-tzu t’ai-pao 太保); 2. their deputies: lesser preceptor to the heir-apparent (t’ai-tzu shaoshih 太子少師), lesser adviser to the heir-apparent (t’ai-tzu shao-fu 少傅), and lesser guardian of the heir-apparent (t’ai-tzu shao-pao 少保). All these posts were extremely high-ranking positions of honor; the three preceptors each held the lower-first rank (or, 1.2),15 their deputies the lowersecond rank (2.2). These preceptors were normally elderly statesmen of great distinction and unblemished reputation, to whom the heir-apparent was expected to render deep respect and reverence. On formal occasions they acted as his escort of honor (lu-pu 齒簿). Some or all of the posts were frequently left vacant, since they were only to be filled if a suitable candidate was available. When they were filled they provided sinecures of great dignity and moral prestige for retired holders of the highest office.16 Supervision of the everyday conduct of the heir-apparent was largely left in the hands of lesser, though still very high-ranking, officials—his four monitors (t’ai-tzu pin-k’o 太子賓客), who held rank 3.1. They too were elderly officials whose duties were to accompany, assist, and advise him in his duties, and to help him in his ritual observances.17 There were also his readers-in-waiting (shih-tu 侍讀), scholarly officials holding substantive posts elsewhere in the bureaucracy who were appointed, as a concurrent responsibility, to lecture 15  T’ang ranks followed the basic chiu-p’in 九品 (“nine ranks”) system. Below, I use two numerals separated by a period, the first referring to the “nine ranks.” The second numeral shows further subdivisions into two classes: cheng 正 (upper) and ts’ung 從 (lower). Ranks 1–3 were subdivided only to that level; thus, e. g., 1.1 means upper-first, or first rank, upper class 正一品, while 1.2 would be lower-first, or first rank, lower class 從一品. Ranks 4–9, however, contained a further division of both upper and lower, the resulting subdivisions being called shang 上 and hsia 下. Thus 4–9 each had four grades in all. For these lower ranks, the second numeral refers to the subdivisions: 6.1 would thus be upper grade of the upper class of the sixth rank 正六品上; 6.2 is lower grade/upper class 正六品下; 6.3 is upper grade/lower class 從六品上; and 6.4 is lower grade/lower class 從六品下. 16   t lt 26, pp. 10b–13a; thy 67, p. 1167; tt 30, pp. 171b–72a–c; cts 44, p. 1906; hts 49A, pp. 1291–92; Fonct. 2, pp. 570–72. For the formalities surrounding the heir’s reception of his three preceptors, see thy 26, pp. 496–97. 17   t lt 26, pp. 13a–b; thy 67, p. 11678; tt 30, p. 172a; cts 44, p. 1906; hts 49A, p. 1292; Fonct. 2, p. 572.

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

48

Twitchett

him on the canonical literature.18 Monitors were first appointed in 656 and at first often seem to have been chief ministers performing this function as a concurrent duty. The posts of both monitors and readers-in-waiting were often left vacant. All these admonitory officials had a relationship with the prince not unlike that of the chief ministers with the emperor. In a well-established Confucian tradition, they were to accustom him to paying heed to the advice, both moral and practical, of elder statesmen. The central administration of the heir-apparent’s establishment comprised three offices paralleling the three major ministries (san-sheng 三省) of the central government. The Intendance-General of the heir-apparent (Chan-shih fu 詹事府) was responsible for the control and administration of the Eastern Palace, paralleling the Department of State Affairs as its central executive department. The two Secretariats of the heir-apparent, the Left 左 and Right 右 Ch’un-fang 春坊 (also known collectively as the Erh fang 二坊), paralleled the Secretariat (in the case of the Left Secretariat, Tso ch’un-fang) and the Chancellery (in the case of the Right Secretariat, Yu ch’un-fang). Like the Chancellery and Secretariat at court, the heir-apparent’s two secretariats handled his correspondence, drafted documents and memorials, and drew up his orders (ling-shu 令書), which had the force of law like imperial edicts. The acts of the heir-apparent, like those of the court itself, were formally recorded, and these documents became matters of record, and were sent to the Historiographer’s Office (Shih-kuan 史館) annually.19 When the heir-apparent was placed in charge of government affairs because of the emperor’s temporary absence or incapacity, the heads of these three offices were expected to provide the prince with especially careful guidance, and under such circumstances they were known as the “commissioners of the three offices” (san-ssu shih 三司使).20 For a brief period, from 662 to 670, the heir-apparent’s court also had its own internal Censorate, the Kuei-fang 桂坊, responsible for the control and discipline of the officials of his staff.21 After its abolition, the functions of maintaining discipline and investigating malpractice among the heir-apparent’s officials were vested in two judicial inspectors (ssu-chih 司直), on the staff of the Intendance-General.22 18  See thy 26, pp. 510–11; tt 30, p. 172a; Fonct. 2, pp. 572–73. 19  Denis Twitchett, The Writing of Official History under the T’ang (Cambridge: Cambridge u.p., 1992), p. 23. 20   h ts 49A, p. 1293. 21   t lt 26, pp. 17a–18a; thy 67, p. 1168; tt 30, p. 172b; Fonct. 2, p. 576. 22   t lt 26, pp. 17a–18a; cts 44, p. 1907; hts 49A, p. 1293.

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

The T ’ ang Imperial Family

49

The Intendance-General (chan-shih fu) of the Heir-Apparent The Intendance-General was the executive of the heir-apparent’s establishment.23 Its intendant-general (t’ai-tzu chan-shih 太子詹事 held rank 3.1 and was thus nominally the equal of the presidents of the Chancellery and Secretariat, while the deputy intendant-general (shao chan-shih 少詹事) held rank 4.1, making him the equal of their vice-presidents. Under their control were the household and the guards of the heir-apparent. The household comprised the Three Courts of the heir-apparent (T’ai-tzu san-ssu 太子三寺): the Court of Domestic Services (Chia-ling ssu 家令寺), the Court of Ceremonial Affairs (T’ai-tzu lei-keng ssu 太子率更寺), and the Court of Equipages (T’ai-tzu p’u-ssu 太子僕寺). The Court of Domestic Services was responsible for the provision of food, for the granaries and storehouses of the Eastern Palace, and apparently for the heir-apparent’s estates.24 It controlled three subordinate offices, the Office of Foodstuffs (Shih-kuan shu 食官署),25 the Office of Granaries (Tien-ts’ang shu 典倉署)26 and the Office of Storehouses (Ssu-tsang shu 司藏署).27 The second of these offices was responsible for managing the heir-apparent’s farms, and for the maintenance of their slaves, and of the families of corvée laborers designated for service in the Eastern Palace. The Court of Ceremonial Affairs had broad responsibilities.28 Its head (leikeng ling 率更令) was responsible for the various ritual observances that had to be performed by the heir-apparent, for his attendance at the shih-tien 釋奠 ceremony at the Confucian temple, and for the lectures on the canonical texts provided for his instruction. The office also was responsible for the clepsydra (water clock) in the Eastern Palace, and for the maintenance of such other symbols of regular and correct normative behavior as the pitch pipes for its musicians. The office was lastly responsible for the discipline and punishment of officials in the heir-apparent’s establishment, although many cases seem to have been referred to the Supreme Court of Justice (大理寺).

23   t lt 26, pp. 13b–17b; thy 67, pp. 1167–68; tt 30, p. 172a–c; cts 44, pp. 1907–9; hts 49A, pp. 1292–93; Fonct. 2, pp. 574–77. 24   t lt 27, pp. 5b–8b; thy 67, p. 1171; tt 30, p. 174a–c; cts 44, pp. 1910–11; hts 49A, pp. 1297– 98; Fonct. 2, pp. 598–603. 25   t lt 27, pp. 8b–10a; tt 30, p. 174b; cts 44, p. 1910; hts 49A, p. 1297; Fonct. 2, pp. 600–1. 26   t lt 27, pp. 10a–b; tt 30, p. 174b; cts 44, p. 1911; hts 49A, p. 1297; Fonct. 2, pp. 601–2. 27   t lt 27, pp. 11a–b; tt 30, p. 174b; cts 44, p. 1911; hts 49A, p. 1298; Fonct. 2, pp. 602–3. 28   t lt 27, pp. 12a–15a; thy 67, pp. 1171–72; tt 30, p. 174; cts 44, p. 1911; hts 49a, p. 1298; Fonct. 2, pp. 603–5.

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

50

Twitchett

The Court of Equipages was responsible for the carriages, palanquins, sedan chairs in the Eastern Palace, the upkeep of the heir-apparent’s stables and pastures and the provision of fodder for them, and also for the keeping of the insignia and ceremonial arms for the heir-apparent’s court, and for funeral ceremonies.29 It controlled a special subordinate Office for the Stables and Pastures of the heir-apparent (Chiu-mu shu 廄牧署),30 which was responsible for the provision and preparation of horses and carriages, and also supervised the many pastures that were reserved for the heir-apparent’s needs. Lastly, the Intendance-General was in control of the Ten Guards of the heirapparent, described below. The Left Secretariat of the Heir-Apparent (t’ai-tzu tso ch’un-fang) The Left Secretariat was the larger and more important of the two parallel secretariats. It looked after the official documents issued by the heir-apparent and the memorials received by him.31 The Left Secretariat was headed by two mentors (t’ai-tzu tso shu-tzu 太子左 庶子) rank 4.1 and two companions (chung-yün 中允) rank 5.1. Beneath them were the following: 1. two remonstrance secretaries (ssu-i lang 司儀郎), rank 6.1, who kept the record of the heir-apparent’s daily acts (tung-kung chi-chu 東宮 記注) and prepared material for the Historiographer’s Office; 2. the grand counsellor of the left (tso yü-te 左諭德), rank 4.2, who was charged with providing moral guidance to the prince; 3. five censor-counsellors (tso tsan-shan ta-fu 左贊善大夫), rank 5.1, who transmitted the prince’s orders and proffered criticism of his actions in the same way as the remembrancers and omissioners at court; and 4. necessary clerical officials and staff. The Secretariat of the Left also supervised a number of specialized offices. Most important of these, and again paralleling the organization of the Chancellery and Secretariat, were the Ch’ung-wen Academy 崇文館, a small college with a

29   t lt 27, pp. 15a–17b; thy 67, p. 1172; tt 30, pp. 174c–75a; cts 44, p. 1911; hts 49A, pp. 1298– 99; Fonct. 2, pp. 606–7. 30   t lt 27, pp. 18a–19a; tt 30, pp. 174c–175a; cts 44, p. 1911; hts 49A, p. 1911; Fonct. 2, pp. 607–8. 31   t lt 26, pp. 18a–32b; thy 67, pp. 1168–71; tt 30, pp. 172c–74a; cts 44, pp. 1907–9; hts 49A, pp. 1293–96; Fonct. 2, pp. 577–94.

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

The T ’ ang Imperial Family

51

staff of scholars and collators and fifteen students,32 and the heir-apparent’s Library Service (Ssu-ching chü 司經局) headed by two director-librarian (hsien-ma 先馬), rank 5.4.33 This was not only a library, but also an archive for all documents, maps, and pictures submitted to the heir-apparent. Five other subordinate offices had charge of the domestic arrangements of the Eastern Palace and the heir-apparent: 1. the Service of Food (Tien-shan chü 典瞎局);34 2. Service of Medicines (Yao-tsang chü 藥藏局);35 3. Service of Furnishings (Nei-chih chü 內直局);36 4. Service for Palace Internal Arrangements (Tien-she chü 典設局).37 Each of these had two (four in the case of Internal Arrangements) chiefs of service (lang 郎) and two assistants (ch’eng 丞) controlling large numbers of palace servants and attendants; and 5. Service of the Palace Gatekeepers (Kung-men chü 宮門局),38 whose chiefs and assistants commanded the 100 gate-porters and other employees charged with the security of the Eastern Palace gates. The Right Secretariat of the Heir-Apparent (t’ai-tzu yu ch’un-fang) This office shared the handling of the heir-apparent’s business and paperwork with the Left Secretariat.39 Its two mentors (t’ai-tzu yu 右 shu-tzu), rank 4.2, and two companion-secretaries (chung she-jen 中舍人), rank 5.2, were responsible for transmitting the heir-apparent’s orders and decisions. There were also four secretaries to the heir-apparent (t’ai-tzu she-jen 太子舍人), rank 6.1, who drafted the prince’s orders and his formal communications with the emperor. The Secretariat also had eight protocol officers (t’ung-shih she-jen 通事舍人), who were responsible for the reception of visitors and for handling their official communications with the prince. The Right Secretariat was thus the main channel of communication to the heir-apparent. 32   t lt 26, pp. 22b–23a; thy 67, pp. 1117–18; tt 30, p. 173b–c; cts 44, p. 1908; hts 49A, p. 1294; Fonct. 2, pp. 584–85. 33   t lt 26, pp. 23a–25a; thy 67, pp. 1170–71; tt 30, p. 173c; cts 44, p. 1908; hts 49A, pp. 1908– 9; Fonct. 2, pp. 585–88. 34   t lt 26, pp. 25a–b; cts 44, p. 1908; hts 49A, p. 1295; Fonct. 2, pp. 588–89. 35   t lt 26, pp. 26a–b; cts 44, p. 1908; hts 49A, p. 1295; Fonct. 2, pp. 589–90. 36   t lt 26, pp. 26b–30a; cts 44, p. 1908; hts 49A, p. 1295; Fonct. 2, pp. 590–91. 37   t lt 26, pp. 30a–31a; thy 67, p. 1171; cts 44, p. 1908; hts 49A, pp. 1295–96; Fonct. 2, pp. 591–93. 38   t lt 26, pp. 31a–32b; cts 44, pp. 1908–9; hts 49A, p. 1296; Fonct. 2, pp. 593–94. 39   t lt 26, pp. 32b–37a; thy 67, p. 1171; cts 44, p. 1909; hts 49A, p. 1296; Fonct. 2, pp. 594–98.

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

52

Twitchett

Also forming a part of its staff were a grand counsellor of the right (yu yü-te) and five censor-counsellors (yu 右 tsan-shan tai-fu) corresponding to those in the Left Secretariat, and their clerical staff. All officials of the Right Secretariat held equal rank with their opposite numbers in the Left Secretariat. The Heir-Apparent’s Guards The heir-apparent, like the emperor, had control of his own personal military forces. These comprised two separate groups.40 First was the personal bodyguard of the heir-apparent, the so-called Three Personal Guards (San-fu 參府), comprising the Personal Guard (Ch’in-wei 親衛), the Merit Guard (Hsün-wei 動衛) and the Standby Guard (I-wei 翔衛). These served as personal bodyguards and acted as ceremonial attendants on the heir-apparent at state functions. They paralleled the Five Personal Guards (Wu-fu 五府) of the emperor. Their members were “guards-officials” (wei-kuan 衛官) and were recruited from among the sons of officials and high-ranking military men. Second were the Ten Guards of the Heir-Apparent (T’ai-tzu shih shuai-fu 太子十率府),41 which paralleled the Sixteen Guards of the Emperor, and were organized in much the same way. Each guard had a permanently appointed group of commanders and officers; some included detachments from among the Three Personal Guards, but the main body of troops was provided from militia (che-chung fu 折衝府) units who came to the capital and served for short periods of duty with the Guard in regular rotation. The heir-apparent’s guards were much smaller than those of the emperor. Des Rotours has calculated that each guard probably had between 400 and 700 men at full strength.42 Like the Guards of the Emperor, each of the Guards of the Heir-Apparent had its permanent headquarters staff. A general (shuai 率) with two deputies ( fu-shuai 副率) held overall command. These were high-ranking officials, holding ranks 4.1 and 4.3, respectively. Beneath them was an adjutant (changshih 長史) responsible for the day-to-day operation of the guard and for making executive decisions. He had a chief administrative clerk (lu-shih ts’an-chün shih 錄事參軍事) and a group of clerks, to whom were responsible the administrators (ts’an-chün shih 參軍事) of the guard’s various special administrative services (ts’ao 曹). These were each responsible for specific duties:

40   t lt 28; thy 71, p. 1289; tt 30, p. 175a–c; cts 44, pp. 1911–13; hts 49A. pp. 1299–1303; Fonct. 2, pp. 608–28. 41   t lt 28; cts 44, pp. 1911–13; hts 49A, pp. 1299–1303. 42  For details of the complicated changes in nomenclature, see Fonct. 2, appendix 1.

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

The T ’ ang Imperial Family

53

1. the Service of Granaries (Ts’ang-ts’ao 倉曹) for rations and stores, building materials, and the like; 2. Service of Troops (Ping-ts’ao 兵曹) for the men; 3. Service of Helmets (Chou-ts’ao 冑曹) for the armory, weapons, and defensive armor; and 4. in some guards there was also a Service of Cavalry (Chi-ts’ao 騎曹) responsible for the guard’s mounts.43 The Service of Granaries was responsible for keeping the service records of the Guard’s civil officials, the Service of Troops for those of the military officers and soldiers. In most of the Guards the Service of Troops was also responsible for its horses. The Chien-men Shuai-fu 監門率府 and Nei Shuai-fu 內率府 guards had no Service of Granaries, and the Service of Troops took over these duties as well. The Guards fell into two groups. The two Chien-men Shuai-fu and two Nei Shuai-fu, although their commanding officers held equal rank with those of the other Guards, had a reduced number of services, and their subordinate officials all held slightly lower ranks and duties somewhat different from those in the other Guards. Their military personnel were also somewhat fewer. Both groups were custodial guards. The former (Chien-men) strictly controlled and recorded the passage of both people and goods in and out of the Eastern Palace. The latter (Nei) formed the heir-apparent’s personal bodyguard, mounted guard in the audience hall when he received his court, when he practised ceremonial archery, and so forth. Each of these pairs of Guards included in their number some permanent-duty Guards-officers, who should not be confused with the ordinary guardsmen-soldiers (wei-shih 衛士) who came from the militias (che-chung fu) to the palace to perform their regular turns of duty, or with the “guards-officials” (wei-kuan) belonging to the Three Militias (san fu), mentioned above. These permanent-duty gate-guards (chien-men chih-chang 監門 首長) and permanent-duty swordsman bodyguards (ch’ien-niu pei-shen 千牛 備身) were counted as regular officials of the seventh rank. The other Guards, the two Wei shuai-fu 衛率府, the two Ssu-yü shuai-fu 司禁率府, and the two Ch’ing-tao shuai-fu 清道率府, had no such large cadres of soldiers holding guards-official ranks on their permanent establishment. Instead each had a group of low ranking officers with specific military duties:

43  The various sources disagree about which guards had which administrative services, a problem nearly impossible to reconcile; see Fonct. 2, pp. 612–26, and nn. given there.

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

54

Twitchett

1. 2. 3. 4.

an officer of the stairways (ssu-chieh 司階); watch officers (chung-hou 中候); officers of javelins (ssu-ko 司戈); and officers of the two-pronged halberds (ssu-chi 司戟).

In the two Wei shuai-fu there were also ten “guards permanently on duty” (san ch’ang-shang 散長上) holding the lowest official rank. These officers may have held only ceremonial duties; but it is also possible that they were originally low-ranking professional soldiers permanently assigned to a military unit most of whose personnel were constantly being rotated. Perhaps their role was not unlike that of the regular army officers or non-commissioned officers assigned to Territorial Army units in the old British system. The two Wei shuai-fu, the two Ssu-yü shuai-fu and the two Ch’ing-tao shuaifu also included among their ranks some of the “guards-officials” belonging to the Three Militias, mentioned above. These sons of high-ranking officials, whose rank derived from hereditary privilege, were employed only in these Guards, whose duties were largely ceremonial, not in the custodial guards. Control of the Heir-Apparent’s Establishment The entire household establishment of the heir-apparent was thus large and complex. A description in the important source-compendium T’ang liutien 唐六典 lists 95 ranking officials with posts “within the current,” and 122 Guards-officers holding official rank. The household services employed over 1,100 servants and specialists. Some 240 of these held established posts “outside the current,” the rest were classed as “miscellaneous employees” (tsa-jen). The household services employed over 500 cooks, storekeepers, porters, and so forth; 95 employees in the stables; over 100 in connection with the clepsydra and the drum, which marked the hours. Finally, there would have been at least 5,000 guardsmen performing regular turns of duty in the heir-apparent’s Guards. In addition to these civilian and military employees the heir-apparent, if he was no longer a child, would have his own harem (Nei-fang 內坊). In addition to his princess-consort ( fei 妃), his concubines were divided into five ranks: 1. 2. 3. 4. 5.

two ladies of excellence (liang-ti 良娣, rank 3.1); six ladies of excellent beauty (liang-yüan 良媛, rank 4.1); ten ladies of inherent excellence (ch’eng-hui 承徽, rank 5.1); sixteen ladies of clear instruction (chao-hsün 昭訓, rank 7.1); and twenty-four ladies of decorous service ( feng-i 奉儀, rank 9.1).

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

The T ’ ang Imperial Family

55

There was also a large staff of female palace attendants headed by the lady directors of his Inner Quarters (ssu-kuei 司閨).44 The heir-apparent’s Inner Palace establishment was not entirely autonomous. It was controlled by a special office, the T’ai-tzu nei-fang chü 太子內坊局,45 which was subordinated to the emperor’s Department of the Inner Palace Intendancy (Nei-shih sheng 內使省), while the Harem of the Heir-Apparent (T’ai-tzu nei-kuan),46 with its various subordinate offices, seems to have come under the general control of the emperor’s harem service (Nei-kuan 內官). All these offices were, of course, filled by eunuchs in the Nei-fang chü, and by female officials in the Nei-kuan. The Nine Horse Pastures of the Heir-Apparent (Tung-kung chiu mu-chien 東宮廢牧監) were likewise not under the heir-apparent’s direct control but were administered as a part of the Court of Imperial Equipages (T’ai-p’u ssu 太僕寺).47 It is not clear how these pasture directorates related to the Court of Equipages (P’u-ssu 僕寺) under the heir-apparent, or what relations existed between the latter and the Court of Imperial Equipages. The Plan of the Eastern Palace and the Heir-Apparent’s Offices The heir-apparent’s Eastern Palace was a large walled compound, 4,894 feet from north to south and about 984 feet from east to west,48 an area of just less than two-tenths of a square mile (or, just under half a square kilometer), which formed the easternmost section of the old Palace City (kung-ch’eng 宮城, hsi-nei 西內) in Ch’ang-an 長安.49 It was entered from the south through the main Chia-fu Gate 嘉福門 beyond which the visitor passed through two further inner gates, Ch’ung-ming Gate 重明門 and Chia-te Gate 嘉德門, to reach the large main courtyard in front of Chia-te Hall 嘉德殿, the great audience hall. On the west and east of this great courtyard were gates leading to inner courtyards containing the offices of the heir-apparent’s two secretariats, 44   t lt 26, pp. 38a–40a; cts 44, pp. 1909–10; hts 47, pp. 1231–32; Fonct. 1, pp. 276–80. 45   c ts 44, p. 1909; hts 47, p. 1224; Fonct. 1, pp. 254–56. 46   c ts 44, pp. 1909–10; hts 47, pp. 1231–32; Fonct. 1, pp. 276–80. 47   h ts 44, p. 1256; Fonct, 1, pp. 403–4. 48  The east-west dimension is nowhere given precisely. My estimate is based on the figures in Hsü Sung 徐松, T’ang liang-ching ch’eng-fang k’ao 唐兩京城坊考, ed. Chang Mu 張穆 (1848; rpt. in Hiraoka Takeo 平岡武夫, ed., Chōan to Rakuyō 長安と洛陽, vol. 6 of Tōdai kenkyū no shiori 唐代研究のしおり [Kyoto: Kyoto daigaku jimbun kagaku kenkyū sho, 1956]), ch. 1, p. 6a. 49  On the Eastern Palace, see Ch’ang-an chih 長安志, in Ching-hsün t’ang ts’ung-shu 經訓堂 叢書 (1784 edn.; rpt. in part, Hiraoka, Chōan to Rakuyō, pp. 89–121) 6, pp. 4a–b; Hsü, T’ang liang-ching 1, pp. 6a–7a.

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

56

Twitchett

the Tso ch’un-fang and Yu ch’un-fang. The Left Secretariat together with the Ch’ung-wen Academy and the heir-apparent’s Library lay on the east, through the Feng-i Gate 奉義門; the Right Secretariat on the west through the Fenghua Gate 奉化門. Their position on either side of the central axis was exactly parallel to the layout of the Secretariat and Chancellery in the “public area” of the emperor’s palace. Behind Chia-te Hall was a range of other halls set one behind the other on the central axis, each dominating its own walled enclosure: the Ch’ung-chiao 重教殿, Li-cheng 麗正殿, Kuang-t’ien 光天殿, and Ch’eng-en 承恩殿 Halls. Behind the latter was the northern gate of the Eastern Palace, the Yüan-te Gate 元德門, which opened into the Imperial Park (Chin yüan 禁苑). In the northern section of the Eastern Palace was the heir-apparent’s own Northern Garden (Pei yüan 北苑). There were various other halls and buildings in courtyards on either side of the main range of great halls; their precise positions and functions are not certainly known. These buildings in the rear of the palace were mostly the heir-apparent’s private quarters. They included a Buddhist shrine (Fo-t’ang yüan 佛堂院), a hall for ritual archery (She-tien 射殿), the quarters for the heir-apparent’s harem, and the palace kitchens, which in the latter part of the dynasty also serviced the nearby Residence of the Ten Princes.50 The business premises of most of the administrative offices pertaining to the heir-apparent, and those of his Guards, were not in the Eastern Palace itself but occupied the northeastern section of the Imperial City (Huang-ch’eng 皇城), the administrative quarter lying immediately to the south of the main palace. The heir-apparent’s offices lay immediately opposite Chia-fu Gate, the main gate of the Eastern Palace. In this compact area were the heirapparent’s Audience Hall (Tung-kung ch’ao-t’ang 東宮朝堂), the outer offices of his Intendancy-General and his two Secretariats, the offices of the various guards, and the stables and storehouses of the heir’s Court of Equipages. Lü Tafang’s 呂大防 stone map of 1080 also shows the Office of Works for the Eastern

50  The site of the Eastern Palace, which lies under the Ming-Ch’ing city of Hsi-an, has never been excavated. It is shown, only in general outline, on the preserved fragments of Lü Ta-fang’s stone engraved map of 1080. The modern architectural historian Fu Hsi-nien 傅喜年, “T’ang-tai sui-tao hsing-mu te hsing-chih kou-tsao ho suo fan-ying te ti-shang kung-shih” 唐代隧道型墓的形制构造和所反映的地土宮室, in Wen-wu yü k’ao-ku lun-chi 文物與考古論集 (Peking: Wen-wu ch’u-pan she, 1986), pp. 323–43, makes a most imaginative study comparing what can be gleaned about the palace from textual evidence with the plan of the well-preserved tomb of the I-te heir-apparent, which was constructed as an analogue of the prince’s earthly residence.

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

The T ’ ang Imperial Family

57

Palace (Tung-kung nei tso-fang 東宮內作坊), which is unknown in the written sources. Was the Eastern Palace Permanently Staffed? One major problem in trying to understand this huge and costly organization is its degree of permanence. There was not always an heir-apparent officially installed, and after 838 there was never an incumbent heir-apparent. When there was no appointed heir, obviously the Eastern Palace had to be kept in repair, since it was always to be presumed that an heir-apparent would eventually be named. In addition, the Eastern Palace began to provide some services for the imperial princes living in the Sixteen Princely Residences. But it is not at all clear whether the Eastern Palace staff was kept up to strength, or whether the whole establishment was put on a “care and maintenance” footing with a skeleton staff. Moreover, there must have been considerable differences between the total years that mature princes served and those of infant heirsapparent.51 Very probably a large number of the posts described above were not always filled. However, posts in the Eastern Palace, like those on the staffs of royal princes, always provided the court with a large pool of available “grace and favor” appointments, comfortable sinecures that might be used to reward deserving officials and keep them employed at the capital. Whatever their actual utility, these posts remained a considerable source of moral prestige for their incumbents. In 785, for example, the mentors of the heir-apparent’s Left and Right Secretariats were promoted in rank to hold precedence over the vice-presidents (shao-ch’ing 少卿) of the Nine Courts.52 These posts were “pure” appointments (ch’ing-kuan 清官), although sometimes the persons appointed might well have been thought unsuitable for their responsibilities. In 813, for example the former grand master of works (ssuk’ung 司空) Yü Ti 于頃 was appointed adviser (pin-k’o 賓客) to the newly installed heir-apparent (the future Mu-tsung 穆宗; r. 820–24). Earlier in the year he had been disgraced after a spectacular scandal involving his sons, and in 817 was demoted for failing to bring up his own sons properly. Later, when he asked to retire from office in 818 the chief ministers even suggested that he be made the heir-apparent’s grand guardian, although the emperor refused.53 In 830, president of the Left Secretariat Sun Ke 孫革, presented a passionately worded memorial protesting that whereas posts in the various 51  E.g., Su-tsung was heir-apparent for 18 years, Te-tsung for 15, and Shun-tsung 26. 52   t hy 67, p. 1169. 53  On this case, see Denis Twitchett, “The Seamy Side of Late T’ang Political Life: Yü Ti and His Family,” am 3d ser. 1.2 (1988), pp. 29–63.

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

58

Twitchett

services subordinate to his office had been traditionally filled by sons of the highest ministers with a scholarly training, who coveted such offices in spite of their low rank because they were “pure” appointments, in recent years men transferred from clerical posts “outside the current” had been appointed to them, thus confusing and defiling their moral standing. His protests were heeded.54 Such appointments were forbidden in future; the privileged elite had won another small battle to retain their hold on these desirable and undemanding outposts of privilege. The provision of prestigious sinecures was not, however, the primary function of the heir-apparent’s organization. Behind it lay a very serious attempt to provide the future ruler with the experience of presiding over an independent establishment. He was given the guidance and the advice of experienced and trusted officials, and was familiarized with something of the constant formality and routine ceremonial that would surround his every action after he came to the throne. It also gave an opportunity for at least some members of the bureaucracy to get the feel of the young man who would one day become their master. Some future emperors formed long-lasting associations with the officials who had been their preceptors or members of their household establishment while they were heirs-apparent, and such advisors could exert a lasting influence on the politics of their “pupil’s” reign.55 These well-intentioned and carefully considered measures to provide imperial heirs with a suitable training to fit them for their future responsibilities were, however, to a large measure thwarted by the remarkable instability of the succession from Kao-tsu’s 高祖 reign (618–26) to that of Hsüan-tsung, and again in the ninth century by the manipulation of the succession by eunuch factions within the palace.

Succession Problems in the Early-T’ang Period

In spite of all these precautions, for a century or more the succession remained far from secure. Kao-tsu’s own successor had been fixed even before the foundation of the dynasty, when at the very end of 617 his eldest son Li Chiench’eng 李建成 was appointed heir to his fief of T’ang by the Sui emperor Yang-ti

54   t hy 67, p. 1169. 55  In 642 T’ai-tsung already foresaw that such deep personal associations could become a political problem, and put a limit of four years on the term for which an official might serve on the staff of any of the princes; see ckcy 4, p. 116.

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

The T ’ ang Imperial Family

59

煬帝.56 Kao-tsu confirmed this choice of heir after assuming the throne in 618,

when he formally installed Chien-ch’eng as heir-apparent.57 But this arrangement was shattered when bitter rivalry between Chien-ch’eng and his younger brother Shih-min 世民 (the future T’ai-tsung) culminated in 626 in the Hsüanwu Gate incident. Shih-min murdered the heir-apparent and another of his brothers, and after first being declared Kao-tsu’s heir-apparent subsequently usurped the throne, accepting Kao-tsu’s abdication in his favor, and leaving his father as retired emperor.58 Two months after his own enthronement in 626, T’ai-tsung installed his eldest son Li Ch’eng-ch’ien 承乾 as heir-apparent.59 Still an eight-year-old child, at first he gave a very good impression, and whenever T’ai-tsung was traveling away from the capital he was left in charge, of course under the close supervision of well-chosen ministers, and conducted business seriously. But as he grew older the prince gave cause for grave doubts over his health and about his fitness to rule.60 For all his serious demeanor in public, in his private life he was devoted to sensual pleasures, and this became public knowledge. T’aitsung began to show unusual favors to his son Li T’ai, prince of Wei 魏王李泰, who formed the ambition of replacing Li Ch’eng-ch’ien as the imperial heir. The extraordinary favors shown by the emperor to Li T’ai met with protests from his senior advisers, who foresaw trouble between the two princes. But the heir-apparent’s indiscretions continued. He developed a homosexual passion for a young palace entertainer. The emperor heard of this liaison and was 56   c ts 1, p. 4, which gives the cyclical date chia-tzu; tctc 184, p. 5765 gives the date as chi-ssu. 57   c ts 1, p. 7; tfyk 257, p. 1a; tctc 185, p. 5795, dates it July 4, 618. 58  Andrew Eisenberg, “Kingship, Power and the Hsüan-wu Men Incident of the T’ang,” tp 80 (1994), pp. 223–59, puts forward the novel idea that Kao-tsu stayed aloof and deliberately left his sons to compete for the throne so that he himself could become retired emperor with the comfortable knowledge that the best, or at least the most ruthless, among his sons would succeed him. Although he adduces interesting evidence about the politics of Kao-tsu’s reign, there is, unfortunately, no positive proof that Kao-tsu ever entertained such an idea. 59  The chronology of these events was as follows: the Hsüan-wu Gate coup occurred July 2, 626; T’ai-tsung was appointed heir-apparent July 5; Kao-tsu abdicated in his favor on Sept. 3; and T’ai-tsung ascended the throne the day after. His heir-apparent Li Ch’eng-ch’ien, prince of Chung-shan, then eight years of age, was installed sixty days later—on Nov. 2; see cts 2, pp. 30–31; tctc 191, p. 6017; 192, p. 6024; thy 1, p. 2. Throughout the T’ang period it remained a normal practice, when a succession was in some degree irregular or unexpected, that the succeeding emperor be briefly appointed heir-apparent before actually ascending the throne. 60   t hy 50, p. 869, contains an edict dated 631 ordering a Taoist adept called Ch’in Ying to offer prayers for the cure of the heir-apparent’s illness.

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

60

Twitchett

infuriated, had the prince’s lover killed, and several members of the prince’s entourage were executed. The heir-apparent believed that Li T’ai had denounced him and this further embittered the relationship between the brothers. The heir-apparent now began to encourage members of his household to indulge in violent behavior and became besotted with Turkish customs, dressing in Turkish costume and speaking Turkish. In 643 he recruited a group of desperados and plotted to murder Li T’ai. But one of his fellow plotters was arrested for involvement in another abortive rising led by the heir-apparent’s younger brother, Li Yu, prince of Ch’i 齊王李祐, and under interrogation revealed the plot to kill Li T’ai. The heir-apparent was stripped of his titles and exiled to the wild frontier district of Ch’ien-chou 黔州 (now in Kweichow),61 where he died at the end of 644. His erstwhile rival Li T’ai was secluded in the Northern Park,62 and then demoted.63 After Ch’eng-ch’ien’s dismissal as heir a bitter court dispute broke out over which prince should replace him, with T’ai-tsung and his ministers backing rival candidates. T’ai-tsung’s ninth son Li Chih, prince of Chin 晉王李治 (the future Kao-tsung), was finally agreed upon and installed as heir-apparent, and the emperor began to devote serious personal attention to his preparation for ruling the empire. But the factional alignments that had arisen from the dispute over the succession remained, and continued to divide the highest levels of the bureaucracy for some years after Kao-tsung’s accession in 649. During Kao-tsung’s long reign (649–683) there were far more serious problems. Kao-tsung’s original empress née Wang was barren, and his first heirapparent, Li Chung 李忠, installed in 652, was his son by a secondary consort. When the emperor’s new favorite Wu Chao 武曌 (the future empress Wu) brought about the dethronement and death of empress Wang and was herself installed as empress in 656, she was determined that the succession should go to one of her own children, so that the next emperor, like her own sickly husband, should remain firmly under her thumb. As soon as she became empress, Li Chung was demoted to be prince of Liang 梁王 and sent away from the capital, first to be governor-general of Liang-chou 梁州 and then to be prefect of Fang-chou 房州 (in modern Hupei). He lived in fear of his life, sometimes wearing women’s clothes to escape imaginary assassins. Later, in 660, he was accused of making political prophecies based on his dreams, and further reduced to commoner status. He was banished to Ch’ien-chou, where he was imprisoned in the former residence of T’ai-tsung’s demoted 61   t fyk 257, p. 2a; ttclc 31, p. 121; tctc 197, p. 6193. 62  See edict in tfyk 257, p. 4b; tctc 197, p. 6196. 63   c ts 3, p. 55; cts 76, p. 2655; tfyk 297, p. 27a; ttclc 39, p. 179.

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

The T ’ ang Imperial Family

61

heir-apparent Li Ch’eng-ch’ien. In 664 he was falsely accused of plotting sedition with Shang-kuan I 上官儀 and the eunuch Wang Fu-sheng 王伏勝, and was ordered to commit suicide.64 Kao-tsung’s remaining sons by other palace ladies were also removed from the scene. His second son, Li Hsiao, prince of Yüan 原王李孝, was banished to provincial posts. In 652 he was made governor-general of T’ai-yüan 太原, Ping-chou 并州, and prefect of Sui-chou 遂州 in 658. He too died in 664.65 The third son, Li Shang-chin, prince of Tse 澤王李上金, and the fourth son, Li Suchieh, prince of Hsü 徐王李素節, also spent the years after empress Wu’s rise to power banished to provincial posts well away from the capital. Finally, in 690 Wu Ch’eng-ssu 武承嗣 had the inquisitor Chou Hsing 周興 accuse them of plotting rebellion. They were brought to Lo-yang 洛陽, where Su-chieh was immediately murdered and Shang-chin hanged himself.66 With the other princes safely out of the way, in 656 Chung was replaced as heir-apparent by the empress Wu’s own three-year-old eldest son Li Hung 李弘, who, as he grew to manhood, promised to become a great emperor. But in 675 he died suddenly of poison during a visit to the Ho-pi Palace in Lo-yang. He was posthumously given the imperial title of Hsiao-ching Huang-ti 孝敬皇帝 and a mausoleum was built for him, as if he had actually reigned as emperor. Many people believed that he was poisoned at empress Wu’s orders, because prior to his death he had been showing opposition to her and to her policies.67 He was replaced as heir-apparent by her second son Li Hsien 賢. He was the most talented of Kao-tsung’s sons, a considerable scholar and a great favorite of his father. When the emperor fell seriously ill in 679 he was made temporary regent and carried out his duties forcefully and decisively. Empress Wu feared that he too might become a potent threat to her own authority, and in 680 he was demoted to commoner status and banished on a blatantly trumped-up charge of treason,68 whereupon late in 680 her third son Li Che 哲 (originally named Li Hsien 顯, the future Chung-tsung 中宗; r. 684, 705–10) became heir. On Kao-tsung’s death Li Hsien, the fourth designated heir-apparent, succeeded to the throne, with the newly titled empress-dowager (huang t’ai-hou 皇太后) Wu making all governmental decisions.69 Kao-tsung died on December 28, 64   c ts 4, p. 86; cts ¡86, pp. 2823–25; hts 81, p. 3586. 65   c tS 86, p. 2825; hts 81, p. 3586. 66   c ts 86, pp. 2825–27; hts 81, pp. 3586–88. 67  He was Kao-tsung’s fifth son, his eldest by empress Wu; see biogs. cts 86, pp. 2828–31; hts 81, pp. 3588–90; also tctc 202, p. 6377. 68   t ctc 202, pp. 6397–98; cts 86, pp. 2831–32; hts 81, pp. 3590–91. 69   t ctc 203, pp. 6416–17.

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

62

Twitchett

683, and Chung-tsung’s enthronement was delayed until January 3, 684. His consort Wei was installed as empress on January 23. On February 26, alarmed by Chung-tsung’s wishing to appoint the new empress’s father as president of the Chancellery, empress Wu had some of her ministers and generals lead troops into the audience hall, proclaim the emperor’s deposition, and drag him off to imprisonment and banishment.70 His reign had lasted fifty-five days. On the next day he was replaced on the throne by his twenty-two-yearold younger brother Li Tan, prince of Yü 預王李旦, the empress’s fourth and youngest son, as Jui-tsung.71 Although the empress-dowager was completely in control, and the young emperor was settled in a separate hall in the palace, the outward forms of sovereignty were followed. The emperor’s consort Lady Liu 柳妃 was enthroned as empress, and his eldest son Li Ch’eng-ch’i 成器 was installed as heir-apparent.72 Still unsatisfied by her indirect authority, empress Wu dethroned Jui-tsung in 690, arranged her own enthronement as sovereign, and founded a new dynasty, the Chou 周. The Chinese imperial system had never envisaged a female sovereign, and because all the empress’s children were by birth members of the T’ang royal family of Li it proved impossible to devise a way for the succession to remain within the Wu family so as to perpetuate a new dynasty. Jui-tsung had never abdicated when the empress seized the throne, but was appointed the “imperial successor” (huang-ssu 皇嗣), although his surname was changed to Wu, and his childhood personal name of Lun 輪 was restored. He resided in the heir-apparent’s Eastern Palace, and was treated with all the ceremony due to an heir-apparent.73 Eventually in 698, deteriorating political circumstances forced the empress to make sweeping changes. The deposed former emperor Chung-tsung was 70  See tctc 203, pp. 6417–18. The former emperor Chung-tsung was demoted to prince of Lu-ling on Feb. 26, 684. His empress was demoted to consort ( fei). His eldest son, Li Ch’ung-jun, who had been appointed imperial grandson in 682 (see cts 5, p. 109), was reduced to commoner status on March 10 (cts 6, p. 116). After Jui-tsung’s accession, the former emperor was banished, first to Chün-chou and subsequently to Fang-ling, in northeast Hupei (cts 7, p. 135). 71   t ctc 203, p. 6418; cts 6, p. 116; cts 7, p. 152. 72  See n. 1, above. Chung-tsung was dethroned on Feb. 26, 684. Jui-tsung was made emperor and his consort made empress on the 27th; tctc 203, p. 6418; cts 6, p. 116. His five-yearold eldest son was probably appointed heir-apparent on April 20. tctc 203 gives the cyclical date jen-tzu of the second lunar mo., but there was no such day. The next jen-tzu day was in the fourth month. cts 6 does not record the event, and Li Ch’eng-ch’i’s biog. at cts 95, p. 3009, gives only the year 684. 73   c ts 7, p. 152.

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

The T ’ ang Imperial Family

63

summoned back to the court in Lo-yang from banishment on the pretext that he required medical treatment. His brother Jui-tsung ceded his position as imperial successor to him,74 and he was once again appointed heir-apparent, thus ensuring that the throne would revert to the royal house of Li after the empress’s death. When in 705 she was finally dethroned by a coup d’etat, Chung-tsung automatically resumed the throne.75 In 706 Chung-tsung’s third son, Li Ch’ung-chün 重俊, was appointed heirapparent.76 However, although the old empress was dead, the Wu family had not been eliminated from power. Her nephew Wu San-ssu 武三思, rumoured to be the new empress Wei’s 韋皇后 lover, quickly engineered the downfall of the leaders of the coup that had overthrown the empress Wu, and in 705 married his own son to empress Wei’s daughter, princess An-lo 安樂公主. He then began scheming to have the princess appointed as heir-apparent. Although this totally unprecedented plan to settle the succession on a woman was temporarily blocked, the empress Wei and Wu San-ssu revived it in the next year. Ch’ung-chün, in self-preservation, then attempted an armed coup, in which 74  He was given the title prince of Hsiang 相王 and given command of the new heir’s Right Guard. In 702 he was appointed minister of education, but resigned in 703. Later that year he was made governor of Yung-chou (Ch’ang-an); cts 6, pp. 130–31; cts 7, p. 152. 75  On these issues, see R. W. L. Guisso, Wu Tse-t’ien and the Politics of Legitimation in T’ang China (Bellingham, Wash.: Western Washington U., 1978), esp. pp. 126–28. 76  Li Ch’ung-chün’s biogs. are cts 86, pp. 2837–39; hts 81, p. 3595. Chung-tsung had four sons in all. The oldest, and the only one bom to empress Wei, was Li Ch’ung-jun 重潤 (biogs. cts 86, p. 2835; hts 81, p. 3593), a particular favorite of Kao-tsung, who created for him the post of imperial grandson. When Chung-tsung was dethroned by empress Wu in 683, he was reduced to commoner status and separately imprisoned. When his father Chung-tsung was brought back to Lo-yang and reinstated as heir-apparent, he was re-enfeoffed, but in 701 he and his sister the princess Yung-t’ai 永泰公主 were accused of criticizing the empress’s favorites, the Chang brothers, and were immediately executed.  The second son, by an unnamed palace lady, was Li Ch’ung-fu, prince of Ch’iao 譙王 李重福, in 704 (biogs. cts 86, pp. 2835–37; hts 81, p. 3594). After Chung-tsung’s accession the empress Wei accused him of having colluded with the Chang brothers to frame her son Ch’ung-jun and thus cause his death. He was banished to a nominal post in Chünchou 均州, in what is now northeast Hupei, and detained there even though imperial amnesties permitted other banished officials to return to the capital. On the grounds of seniority he still had the best claim to the succession, and when in 710 empress Wei seized power, one of her first acts was to send 500 troops to Chün-chou to apprehend him. But she herself fell from power a few days later, and the newly enthroned Jui-tsung conferred the post of prefect of Chi-chou 集州 upon Ch’ung-fu. But he was already embroiled in a rebellion of his own, and had gone to Lo-yang to try to seize the throne. The plot was a fiasco, and ended in Ch’ung-fu’s suicide and the public exposure of his corpse.

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

64

Twitchett

Wu San-ssu and his son the consort of princess An-lo were killed, but which ultimately failed and led to his own death. Chung-tsung himself, empress Wei, and the princess An-lo survived, and there was renewed speculation that empress Wei intended to appoint An-lo as imperial heir. However, the post of heir-apparent was allowed to remain unfilled. Only one of Chung-tsung’s sons remained at court, his fourth son Li Ch’ung-mao, prince of Wen 溫王李重茂.77 In 710 Chung-tsung’s brief reign ended in his murder by empress Wei, who with her daughter princess An-lo had presided over five years of misrule and utter corruption and who now tried to seize power for herself, as empress Wu had done. She enthroned the sixteen-year-old Li Ch’ung-mao as “Wen-wang 溫王” and had herself appointed dowager-empress, planning to overthrow the T’ang and set up a new dynasty. Her regime was doomed from the outset, and lasted only two weeks before a counter-coup was mounted, led by the third son of Jui-tsung, Li Lung-chi 隆基 (the future Hsüan-tsung). The empress, with many of her own family and partisans, and princess An-lo were killed, and the child-emperor was forced to abdicate in favor of his uncle, Chung-tsung’s younger brother, the former emperor Jui-tsung, who ascended the throne once again.78 Jui-tsung was a reluctant ruler, who had survived under empress Wu’s dominance by keeping a low profile and withdrawing from involvement in public affairs. The stability of the dynasty was still seriously threatened, however, by the bitter rivalry between the new emperor’s ambitious sister the princess T’ai-p’ing 太平公主 and his third son Li Lung-chi, who had been designated heir-apparent only three days after Jui-tsung ascended the throne on July 25, 710. After two years of dissention, Jui-tsung had had enough, and in September of 712 he abdicated in favor of Li Lung-chi, but under an unwieldy compromise arrangement that left himself as retired emperor (t’ai-shang-huang 太上皇), holding court every fifth day and retaining powers over senior appointments and the most important administrative decisions. The emperor, moreover, still constantly deferred to the opinions of his formidable sister the T’ai-p’ing princess, who continued her intrigues unabated and consolidated support among the court officials. Eventually, in the summer of 713 events reached a crisis. The princess prepared a conspiracy to murder Li Lung-chi and seize power for herself. Li 77  Biog. at cts 86, p. 2839. 78   t ctc 209, p. 6643; cts 51, pp. 2171–75; hts 76, pp. 3486–87. On these events, see Denis Twitchett, ed., Sui and T’ang China (589–906), Part 1, vol. 3 of The Cambridge History of China (Cambridge: Cambridge u.p., 1979), pp. 326–28, 335–341.

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

The T ’ ang Imperial Family

65

L­ ung-chi learned of her plans and carried out a successful preemptive strike against the conspirators; the princess was ordered to commit suicide on July 29, 713. The next day Jui-tsung formally divested himself of his remaining powers, and Hsüan-tsung became emperor in reality.79 Until Hsüan-tsung became full emperor and dynastic stability was finally restored, the succession had thus been repeatedly threatened. In the preceding decade alone the empress Wu had been dethroned and her dynasty brought to an end; Chung-tsung had been murdered by his empress Wei, who had aimed to set up a dynasty of her own; the same empress had previously tried to have a daughter appointed heir-apparent; there had been attempts to seize the throne first by Chung-tsung’s heir-apparent, and then by another royal prince; there had been a plot to seize power by a royal princess; and Hsüan-tsung had himself led two armed interventions to restore stability. Until this time no eldest son, duly installed as heir-apparent at the beginning of his father’s reign, had yet succeeded to the T’ang throne. These disturbances of the system laid down in the dynasty’s statutes had mostly been the result of intrigues by a succession of powerful and ambitious women of the imperial household attempting to entrench themselves and their own influential families in power, taking advantage of a succession of weak-willed emperors. Changes under Hsüan-tsung Hsüan-tsung lost little time in installing his own heir-apparent. In 715 his second son Li Ssu-ch’ien, prince of Ying 郢王李嗣謙, was installed as heirapparent, the event being marked by an amnesty and three days of public celebration. In 719 the court celebrated both his coming to maturity (yüan-fu 元服) and the ch’ih-chou 齒胄 ceremony, when he first attended lectures in the State Academy Directorate (Kuo-tzu chien 國子監). In 725 he was married to a consort of the surname Hsüeh 薛氏, a ceremony accompanied by a local Act of Grace 曲赦 in the capital. At last the arrangements for the succession appeared to be proceeding according to the rules. However, palace politics soon interfered yet again. In 720 the emperor’s younger brother Li Fan 範 was involved in a palace plot together with a brother-in-law who was the consort of a royal princess. As a result strict limits were placed on the social contacts of royal princes and princesses, and the emperor’s two elder brothers were recalled to the capital from their posts

79  On these complex events, see Howard S. Levy, “How a Prince Became Emperor: The Accession of Hsüan-tsung (713–755),” Sinologica 6 (1958), pp. 101–21; Twitchett, ed., Cambridge History 3, pp. 335–45, and the sources given there, esp. tctc 210, pp. 6681–85.

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

66

Twitchett

in the provinces. In 725 a similar plot involving the emperor’s youngest brother Li Yeh 業 and his in-laws was uncovered. More serious was the position of the empress Wang 王皇后, who had been installed in 712, but who was still childless. Before his accession to the throne Hsüan-tsung had already had sons by a number of consorts. His designated heir-apparent was the son of his secondary consort Chao Li-fei 趙麗妃, a talented musician and dancer who had become the future emperor’s favorite while he was vice-prefect of Lu-chou 潞州 (708–10). The son was probably bom in 709 and first enfeoffed as a commandery prince in 710. After Hsüan-tsung’s first years Chao Li-fei, like several of his other consorts, gradually lost the emperor’s affections. She died in 726. In Hsüan-tsung’s early years on the throne his two principal favorites were the kuei-pin lady Yang 楊貴嬪, the mother of the future Su-tsung (see below), who also died early in his reign, and Wu Huifei 武惠妃, who bore him four sons and three daughters. In 724 the empress Wang, desperate to conceive a son, was found to have consulted a monk to perform magic ceremonies to ensure a child, was degraded to commoner status, and died shortly after. The emperor wished to appoint lady Wu as empress, but the court objected that she was a descendant of the family of empress Wu, and that since she was not the mother of the heir-apparent but had sons of her own, her promotion might imperil the succession. She was thus not enthroned empress, but installed as hui-fei, although with the sort of ceremonies appropriate to an empress.80 The objectors’ arguments proved all too accurate. In 736, amid much court intrigue, Wu Hui-fei claimed that the heir-apparent and other royal princes were planning to kill both herself and her son Li Mao, prince of Shou 壽王李瑁. Hsüan-tsung dismissed these charges. But next year the claims were renewed by Yang Hui 楊洄, the consort of one of Wu Hui-fei’s daughters. The heirapparent, together with his brother-in-law Hsüeh Hsiu 薛鏽, the consort of the princess T’ang-ch’ang 唐昌公主, Li Yao, prince of E 鄂王李瑶, and Li Chü, prince of Kuang 光王李据, the emperor’s fifth and eighth sons, respectively, were reduced to commoner status and then ordered to commit suicide. Wu Hui-fei herself died at the end of 737, and was posthumously granted the title of empress. There was now neither empress nor heir-apparent, and the position of heir remained vacant for more than a year. The emperor refused to appoint Wu 80  Wu Hui-fei’s biogs. are cts 51, pp. 2177–78; hts 76, pp. 3491–92 See also Howard S. Levy, “Wu Hui-fei, A Favoured Consort of T’ang Hsüan-tsung,” tp 46.1–2 (1958), pp. 49–80, which includes an excellent geneaological chart of the Wu clan.

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

The T ’ ang Imperial Family

67

Hui-fei’s son Li Mao, as she had hoped, and eventually decided that the succession should go to Li Yü, prince of Chung 忠王李璵 (the future Su-tsung 肅宗; r. 756–62), his third son by the deceased lady Yang, whose claim to seniority was secure. (His personal name was changed to Heng 亨 in 744.) Hsüan-tsung’s eldest son, Li Ts’ung 琮, was not considered because he was childless. After this, in spite of the dominance of the aging emperor by his favorite consort Yang Kuei-fei 楊貴妃, no change was made in the succession, and Li Yü/Heng eventually succeeded to the throne. Li Heng came to the throne as a man of twenty-nine, better prepared to rule than any of his predecesors, after having served uninterruptedly as heirapparent for eighteen years. He did not, however, succeed to the throne in the normal fashion, on the death of his father. In 756 An Lu-shan’s rebel army had taken Ch’ang-an, and Hsüan-tsung decided to flee and take refuge in Szechwan. On the road the heir-apparent parted company with him, deciding to retreat to the northwest, rally the imperial forces, and continue to resist the rebels. Within a few days Li Heng had been persuaded to usurp the throne. To satisfy propriety the old emperor went through the motions of abdicating in his favor, and was appointed retired emperor (shang-huang t’ien-ti 上皇天帝). In 757, after Su-tsung’s forces had recaptured the two capitals from the rebels, the former Hsüan-tsung was returned to Ch’ang-an and allowed to reside in Hsing-ch’ing Palace, which he had had constructed more than thirty years before. Later, in 760, Su-tsung began to fear that the old emperor still retained the loyalty of many of his subjects, and might become the focus of dissidence. He was moved into Ta-ming Palace, where he lived until his death on May 3, 762. Su-tsung was himself already mortally ill, and survived his father by only eleven days. His son Li Yü 預 (the future Tai-tsung 代宗; r. 762–79) had been designated heir-apparent in 758, and he was appointed regent immediately after Hsüan-tsung’s death because Su-tsung was already incapacitated. Sutsung’s empress Chang had opposed his appointment as heir-apparent and had wished to have one of her own sons installed to replace him as heir. Now she attempted to have one of her sons murder the heir-apparent in the palace. However, the plot was thwarted by Su-tsung’s chief eunuchs Li Fu-kuo 輔國 and Ch’eng Yüan-chen 程元振. The empress and two of the dissident princes were executed, and Li Yü was enthroned on May 18, 762. The succession subsequently became more a matter of routine. Only one T’ang emperor after Su-tsung ever appointed an empress, and she died on the same day as her installation.81 Much of the former instability had been the result 81  Te-tsung did arrange to install his favorite consort, Wang Shu-fei 王淑妃, the mother of his heir-apparent (the future Shun-tsung), as empress in 786. The brevet of appointment

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

68

Twitchett

of palace intrigues led by a series of ambitious and unscrupulous women, all of whom had close ties both with their own politically dominant families and with other influential aristocratic groups at the capital. Henceforth the emperor’s principal wife was appointed only as a consort ( fei). However, after an emperor died the most favored of his consorts would be appointed empressdowager. Chief among these would be the mother of the new emperor, whose elevation helped to firmly establish his claims to the legitimate succession. During the ninth century, when several short-lived emperors followed one another, there were at times as many as three empresses-dowager or grand empresses-dowager living simultaneously in different palaces or halls. They were not politically neutral but interfered in succession disputes and meddled in other aspects of politics. But, formidable old ladies as they undoubtedly were, they no longer had the unquestioned authority of an active empress or the capacity to do mischief of their predecessors. The other royal princes too ceased to be dangerous rivals; confined since 725 to the princely residences (see below, under “The Royal Princes”) and denied the opportunity to hold office, they no longer had either the status, the independence, or the wealth that their predecessors had enjoyed in the seventh and early-eighth centuries. For almost half a century the throne passed uneventfully to the eldest son of each deceased emperor, after he had served an apprenticeship as heir-apparent. This period in waiting had lasted for fifteen years in the case of Tetsung 德宗, twenty-five years in the case of Shun-tsung 順宗. The accession of Shun-tsung, February 28, 805, however, broke this pattern. Shun-tsung came to the throne already a very sick man. Since the preceding autumn he had lost the power of speech. His condition deteriorated, and, incapable of ruling, he was forced to abdicate in favor of Li Ch’un 純 (r. Hsien-tsung 憲宗, 805–20), the eldest of his twenty-three sons, who had been installed as heir-apparent on May 8. On August 26 the heir was appointed regent, and Shun-tsung’s formal abdication followed on September 4. He retained the title of retired emperor, and his residence was moved to Hsing-ch’ing Palace, where he died February 11, 806.

was issued Dec. 3. However, the empress had already been sick for a long period, and died on the fifth, just as the rites conferring the title of empress were completed; cts 12, p. 355; tctc 232, p. 7474; cts 52, pp. 2193–94; hts 77, p. 3502 (hts incorrectly dates this event to 787).

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

The T ’ ang Imperial Family



69

The Succession and the Eunuchs after 805

The persons who engineered Shun-tsung’s abdication were a group of highranking eunuchs, allied with a number of powerful provincial governors. The eunuchs, whose power was in the ascendant in every field of government from this time onwards, controlled the succession, just as they controlled the royal palaces and the residences of the royal princes, whose household establishments survived in name only. Hsien-tsung died in mysterious circumstances in 820, said to have been the result of poison administered by a eunuch, but more likely killed by the highly toxic “elixirs” prepared by his alchemists. The heir-apparent,82 his third son Li Heng 恒 (Mu-tsung), succeeded him.83 From this time on eunuchs were involved in every T’ang succession, and in every dispute about the succession. Every successful candidate for the throne was enthroned with the backing of the palace consuls (hu-chün chung-wei 護軍中尉), the supreme commanders of the eunuch-led Palace Armies of Inspired Strategy (Shen-ts’e chün 神策軍).84 Mu-tsung was incapacitated in a polo match in 822,85 after which he was completely dominated by the eunuchs and died prematurely in 824.86 He was succeeded by his eldest son Li Chan, prince of Ching 景王李湛 (r. Ching-tsung 敬宗, 824–27), who had been hastily created heir-apparent after the accident.87 82  Hsien-tsung had appointed his eldest son Li Ning, prince of Teng 鄧王李寧, as heirapparent in 809, but he died in 811. His third son, prince of Sui Li Heng 遂王李恒 (the future Mu-tsung), was appointed heir-apparent in his place in 812. 83   c ts 15, pp. 471–72; cts 16, p. 475; tctc 241, pp. 7776–77. 84  See Liu Yat-wing “The Shen-ts’e Armies and the Palace Commissions in China, 755–875 AD” (Ph.D. diss., London University, 1970). The evidence on the various late-T’ang successions is tabulated on pp. 273–76. I adopt Dr. Liu’s translation of “hu-chün chung-wei” as “Palace Consul,” in preference to Hucker’s “Chief Palace Commandant-Protector.” 85   c ts 16, p. 501; tctc 242, pp. 7822–23. On this incident, see James T. C. Liu, “Polo and Cultural Change: From T’ang to Sung China,” hjas 45.1 (1985), pp. 203–24. 86   c ts 16, p. 504; cts 17A, p. 507; tctc 243, pp. 7830–31. The emperor’s death was not due to his injuries. By the New Year of 824, he was fit enough to preside over the court according to the normal ceremonial. But the emperor had begun ingesting mineral drugs. A recluse named Chang Kao 張皋 sent in a memorial strongly criticizing him for repeating his father’s fatal error. Mu-tsung appreciated his advice and tried to summon him to the court, but he could not be found. The emperor’s illness recurred on Feb. 23, and he died two days later. See cts 171, pp. 4449–50, for Chang Kao’s memorial. See also Joseph Needham and Ho Ping-yü, “Elixir Poisoning in Medieval China,” Janus 48 (1959), pp. 221–51; rpt. in Joseph Needham, Clerks and Craftsmen in China and the West (Cambridge, Cambridge u.p., 1970), pp. 318–19. 87   t ctc 242, pp.. 7822–23; tfyk 257, pp. 15a–17a.

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

70

Twitchett

Ching-tsung’s accession at the age of fifteen was a dynastic disaster. He was quite unfit to rule, and his irresponsible and unpredictable behavior led to his murder by a group of eunuchs in 827 during a wild night of debauchery and carousal. He was still only eighteen, and, although he left five infant sons,88 they could not be enthroned. Until Ching-tsung, the succession had remained in the direct line from father to eldest son since Su-tsung’s time. There now followed two decades of confused and disputed successions. The most powerful eunuchs of the time, Wang Shou-ch’eng 王守澄 and Liang Shou-chien 梁守謙, enthroned his younger brother Li Ang 昂 (r. Wentsung 文宗, 827–40), after a bitter struggle between three rival cliques of eunuchs backing the claims of different royal princes.89 Wen-tsung was a far more serious ruler than his two pleasure-loving predecessors. But he proved weak and indecisive, and his reign was marred by constant and bitter factional wrangling both among factions of courtiers and ministers and among rival groups of eunuchs. Things were made worse by two ill-conceived and incompetently executed attempts to curb or eliminate the power of the eunuchs, which had disastrous results for the civil bureaucracy. In 832 he appointed the elder of his two sons, Li Yung, prince of Lu 魯王李永, as heir-apparent, but he proved to be frivolous and dissolute.90 In 838 his mother, Wang Tefei 王德妃, who had fallen out of favor with Wen-tsung and quarelled with Yang Hsien-fei 楊賢妃, the influential former consort of Mu-tsung, died. The emperor called a conference of his ministers, which advocated removing the heir-apparent. The indecisive emperor took no action, but shortly afterwards the heir-apparent met a violent death in circumstances that were carefully hushed up, although the case led to the execution or banishment of several dozen eunuchs and officials.91

88   c ts 175, pp. 4538–40; hts 82, p. 3632; hts 70B, pp. 2170–71; and thy 5, p. 54. thy omits the eldest son Li P’u, prince of Chin 晉王李普, a special favorite of his uncle Wen-tsung, who intended to make him his own heir. But he died in 828, aged five, after which he was posthumously entitled the Tao-huai heir-apparent 悼懷太子 (“The Lamented Heir-Apparent”). 89   c ts 17A, pp. 522–23; tctc 243, pp. 7851–53. Wen-tsung was Mu-tsung’s second son, and Ching-tsung’s younger brother. 90   t fyk 257, pp. 16b–17a; tctc 244, p. 7880. The long delay in appointing an heir was the result of Wen-tsung’s disppointment over the death in infancy of Li P’u, prince of Chin, the eldest son of Ching-tsung whom he had loved as if he were his own son and wished to designate as his successor. 91   c ts 17B, p. 575; cts 175, pp. 4540–42; hts 82, p. 3634; tctc 246, pp. 7935–36.

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

The T ’ ang Imperial Family

71

In the winter of 839 the emperor was pressed to appoint a new heirapparent. Mu-tsung’s former consort Yang Hsien-fei pushed the claims of her own son Li Jung, prince of An 安王李溶,92 but she was a relative of the powerful minister Yang Ssu-fu 楊嗣復, and the prince’s appointment was strongly opposed for political reasons. Wen-tsung therefore issued an edict announcing that the new heir should be Li Ch’eng-mei, prince of Ch’en 陳王李成美, the fifth son of Ching-tsung.93 Ch’eng-mei was never, however, formally installed, and on February 8, 840, another edict appointed yet another of Mu-tsung’s sons, Li Ch’an, prince of Ying 穎王李瀍 (the future Wu-tsung 武宗; r. 840–46), who was secretly summoned from obscurity in the Residence of the Sixteen Princes to be heir-apparent and regent. Two days later Wen-tsung was dead, and the prince of Ying ascended the throne. The succession had been stage-managed and the testamentary edict appointing the prince of Ying had been forged by the all-powerful eunuch consuls Ch’iu Shih-liang 仇士良 and Yü Hung-chih 魚弘志.94 No emperor after Wen-tsung ever appointed an heir-apparent before he was himself on the brink of death, and it is probable that in some cases they had no hand in the testamentary edicts in which their successors were named, but that these were written while the emperor was on his death bed, or even written after his death, by the eunuchs who manipulated the succession. After Wu-tsung’s death, on April 22, 846, the unexpected appointment at the age of thirty-five of Li I, prince of Kuang 光王李怡 (renamed Ch’en 忧; r. Hsiuan-tsung 宣宗, 846–59), which resulted from a conference of eunuchs held in the Imperial Park, made a complete break with the pattern of succession, for he was a younger brother of Mu-tsung, the thirteenth son of Hsientsung, and thus a member of the generation senior to the last three emperors. He was named as imperial uncle (huang t’ai-shu 皇太叔) and provisional regent (ch’üan kou-tang chün-kuo cheng-shih 權勾當軍國政事) in a testamentary edict issued the day before Wu-tsung’s death.95 92   c is 175, p. 4538; hts 82, p. 3632; and tctc 246, p. 7941. cts 175 says that he was Mu-tsung’s eighth son, but this is inconsistent with the fact that earlier (p. 4536) the same text says that Mu-tsung had only five sons, a number confirmed by hts 82, p. 3630. 93   c ts 175, pp. 4539–40; hts 82, p. 3632; tfyk 257, pp. 17b–18a. 94  The testamentary edict is at tfyk 11, p. 15b; ttclc 12, pp. 70–71; wyyh 73, p. 15a; and cts 18A, p. 584. 95   c ts 18A, p. 610; cts 118B, p. 613; tfyk 257, p. 18b; and tctc 248, p. 8022. For Wu-tsung’s testamentary edict, see ttclc 12, p. 71; and wyyh 76, p. 13a. The choice of Hsiuan-tsung was surprising not only because he was virtually unknown and had always kept out of the limelight, acting modestly and unassumingly, but also because his mother, Lady Cheng

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

72

Twitchett

His eldest son Li Ts’ui 漼 (r. I-tsung 懿宗, 859–73) was appointed heirapparent and regent only six days before his succession at the age of twentyfive in 859.96 I-tsung’s fifth son and successor Li Yen 儼 (now renamed Hsüan 儇; r. Hsi-tsung 僖宗, 873–88) succeeded at the age of eleven in 873 just two days after a testamentary edict named him heir-apparent and provisional regent.97 After his death in 888 his younger brother Li Chieh 傑 (now renamed Yeh 曄; r. Chao-tsung 昭宗, 888–904) was appointed as imperial younger brother (huang ti 皇弟) in a testamentary edict issued two days before he came to the throne aged twenty-one.98 This meant that after Ching-tsung all T’ang emperors came to the throne without having spent any time as the formally designated heir-apparent, and thus lacked the institutionalized training for power provided by the statutes. Instead of spending some years at least presiding over his own court in miniature in the Eastern Palace, under the tutelage of carefully selected tutors and mentors, and perhaps having acted for short periods as temporary regent, after Mu-tsung’s reign future rulers grew up along with the other imperial princes in the restrictive conditions of the Sixteen Princely Residences, which were administered by the eunuchs. Moreover the policy imposed by Hsüan-tsung, prohibiting the employment in office of royal princes of the blood and debarring them from participation in public affairs, meant that no royal prince unless he was designated heir-apparent could acquire any practical experience of administration. It is not surprising that some were irresponsible and led frivolous lives.

鄭氏, who was still living, had a most unusual background. She had been a concubine of the rebel Li Ch’i 李錡, who had taken her into his harem after a fortune teller had prophesized that she would bear a future emperor. After Li Ch’i’s defeat and execution in 807 she had been officially enslaved and assigned to Hsien-tsung’s inner palace, where she became an attendant of his favorite consort Kuo Kuei-fei. She bore the future Hsiuan-tsung in 810. hts 77, p. 3505, gives this information, and also says that some people believed her original surname had been Erh-chu 爾朱, a double surname associated with the northern frontier peoples of Northern Wei times; cts 52, p. 2198, says that the “old histories lack any information about where her family came from and how she entered the palace service.” 96   c ts 18B, p. 6454; cts 19A, p. 649; tfyk 257, p. 18b; and tctc 249, pp. 8075–76. Hsiuantsung’s testamentary edict is at ttclc 12, pp. 71–72; wyyh 80, p. 17b. 97   c ts 19A, p. 683; cts 19B, p. 689; tfyk 257, p. 18b; and tctc 249, pp. 8166–67. The testamentary edict is at ttclc 21, p. 72; wyyh 84, p. 7a. 98   c ts 20A, p. 735; tftk 257, pp. 18b–19a; and tctc 257, p. 8376. The testamentary edict is at ttclc 12, p. 72; wyyh 88, p. 6a.

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

The T ’ ang Imperial Family

73

Several of them also suffered chronic ill-health, and a number of them, beginning with Hsien-tsung, died of poisoning as the result of ingesting highly toxic “elixirs” prepared by their alchemists.99 The life-expectancy of emperors declined sharply in the ninth century. Every T’ang ruler down to Te-tsung survived into his fifties; Hsüan-tsung lived to be seventy-seven, the empress Wu died at eighty-two. By contrast, no T’ang emperor after Te-tsung reached the age of fifty. Of the eight emperors who followed Hsien-tsung only one, Hsiuantsung, even survived into his forties (see table 1). The careful provisions for grooming and providing appropriate experience for future emperors and for protecting the succession had clearly broken down completely by the 830s.

The Royal Princes

In addition to supporting and preparing the heir-apparent, the T’ang royal house also needed to make adequate provision for the large numbers of imperial princes. It was essential to enable them to live in a style and dignity befitting their ranks, for it was always conceivable that any one of them might become the next candidate for the succession, and it was equally important to keep them content and loyal and at the same time to limit their opportunities and capabilities of threatening the throne. T’ai-tsung placed great stress on the importance of the members of the royal family as a reliable body of support for the emperor and for the dynasty. He believed, as did some senior officials at his court, that an administrative system in which hereditary local power was devolved on enfeoffed members of the royal family and other proven allies (the so-called feng-chien 封建 system) would be inherently more stable and long-lasting than the hierarchical bureaucratic system of counties and prefectures administered by career appointees of the central power. The latter system, the so-called chün-hsien 郡縣 system, had been established by the Sui dynasty and retained by his father Kao-tsu. Advocacy of the feng-chien ideal was not an irrational espousal of an ideal system deriving from antiquity. At the time of T’ai-tsung’s coup against his brothers in 626, the T’ang regime was far from securely established and the country was still recovering from the destruction and bloodshed resulting from several years of large-scale warfare; the last major rivals to the T’ang had 99  See Needham and Ho, “Elixir Poisoning,” pp. 316–39. The imperial victims were Hsientsung, Mu-tsung, Wu-tsung, and probably Hsiuan-tsung.

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

74

Twitchett

been suppressed only two years before. The empire was still governed very loosely. At the local level a huge number of small splinter counties and prefectures (roughly three times as many as in the heyday of the Sui), had been set up ad hoc as local forces surrendered to T’ang control or were pacified. At a higher level, a network of regional branches of government (hsing-t’ai 行台)— essentially strategic—commanded different theatres of operations; and regional military commands (tsung-kuan fu 總管府) also controlled the local civil administration. It took many years before T’ai-tsung enjoyed complete control over the local administrations of his empire, and while the civilian hierarchy was steadily being extended, in some cases in areas still not fully reconciled to the T’ang victory, it was quite understandable that T’ai-tsung’s preference was to command at least the strategic backbone of his empire through the indirect agency of administrations under family members. Debate on the subject began in 627 when the reestablishment of a fengchien system was suggested by Hsiao Yü 蕭瑀 (575–648), who had long experience of politics under the Sui and during Kao-tsu’s reign. It was strongly opposed by the other chief ministers, but T’ai-tsung persisted with his advocacy of this form of government in the face of all opposition. In 631 several of the emperor’s younger brothers were re-enfeoffed with new titles, as were seven of his sons. Later in that year an edict ordered that detailed regulations be set up for hereditary territorial fief administrations for members of the royal family and also for meritorious and sage ministers.100 Beginning in 632 his surviving younger brothers were given appointments as prefects, mostly in important regional centers and in strategic cities defending Ch’ang-an. Then, early in 636, twenty-one imperial princes were reenfeoffed and appointed governors-general (tu-tu 都督) of various major strategic prefectures, and sent out to govern them as if they were fiefs.101 In some cases the administrative center of their governorship was historically the same as the newly endowed fief name.102 In 637 T’ai-tsung attempted to enlarge this 100   t ctc 193, p. 6089. 101   t ctc 194, p. 6118. 102  On this policy, see Howard J. Wechsler, Mirror to the Son of Heaven: Wei Cheng at the Court of T’ang T’ai-tsung (New Haven: Yale u.p., 1974), pp. 175–76; and Cambridge History 3, pp. 210–11. See also Robert M. Somers, “Time, Space and Structure in the Consolidation of the T’ang Dynasty (AD 617–700),” jas 45.5 (1986), p. 976. Somers, who clearly grasped the significance of this policy, was incorrect, however, in his assertion that the locations assigned to royal princes are unknown. On March 4, 636, probably as a result of the death of retired emperor Kao-tsu a few months earlier, 17 princes, 10 brothers of the emperor, and 7 of his sons were promoted to grander fiefs. On March 16 six of these royal brothers and six of the emperor’s sons were appointed governors-general of named major centers;

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

The T ’ ang Imperial Family

75

scheme by sending fourteen of his senior courtiers to similar regional appointments, and to make all these provincial posts, together with those held by the royal princes, hereditary. This provoked bitter opposition. The new appointees included some of the emperor’s closest supporters, who looked on such provincial appointments as banishment from the center of power, and resisted strongly and vociferously. The policy was abandoned and the hereditary appointments terminated in 639. But the policy of using royal princes as governors and prefects of a protective ring of strategic centers continued, and remained in T’ai-tsung’s mind until the end of his life, when he raised it as an issue of profound importance in his Ti-fan 帝範, the political testament written for his heir in 648.103 The appointment of royal princes to prefectural posts, where they could gain experience of administration, while remaining under close supervision and at a safe distance from the court and its intrigues, continued under his successor Kao-tsung, one of whose tutors as heir-apparent had been Hsiao Yü. It continued during the nominal reigns of Chung-tsung and Jui-tsung, under the dominance of empress Wu before she usurped the throne in 690. There was a group of important prefectures that seems to have been reserved as postings for royal princes, and their positions clearly show that T’ai-tsung’s vision of enfeoffed family members ( fan-ch’i 藩戚) providing a protective hedge around the capital still remained alive. Several of T’ai-tsung’s brothers and those of his sons who survived him remained in service in one prefecture after another until their deaths.104 Under Kao-tsung his own brothers and uncles as well as some of his sons continued to be appointed to posts as prefects or governorsgeneral, as were the sons of Chung-tsung and Jui-tsung. see the imperial order in cts 64, p. 2424. Li T’ai, the emperor’s scholarly fourth son was among those appointed, but was allowed to stay at the capital and pursue his scholarly interests, while Chang Liang was deputed to perform his actual duties. The princes left for their places of duty on May 2, after an emotional send-off by T’ai-tsung; see tctc 194, pp. 6118–19; ttclc 35, p. 150. On Feb. 5, 637, three more of the sons of the emperor were appointed governors-general; see tctc 194, p. 6125. The appointment of 21 princes including the above as hereditary prefects of their districts, and the appointment of 14 senior court statesmen as the hereditary prefects of other places, came in July 637; tctc 195, p. 6130; ttclc 35, pp. 149–50. The affair dragged on. The senior ministers protested bitterly to the emperor that these provincial appointments amounted to banishment, and on April 5, 639, the scheme for hereditary prefects was terminated; cts 3, p. 50; tctc 195, pp. 6145–46. 103  See Cambridge History 3, pp. 210–12. 104  Of Kao-tsu’s 22 sons, 16 survived his own death in 636; 15 survived the death of T’ai-tsung in 649, 5 survived the death of Kao-tsung in 683.

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

76

Twitchett

There is no question that they did actually participate in government. Some proved model administrators, others were reprimanded by T’ai-tsung or Kaotsung for their shortcomings. One of them, Li Yüan-hsiang, prince of Chiang 江王李元祥, Kao-tsu’s twentieth son, had such a fearsome reputation for strictness with his underlings that they claimed that to be appointed to serve on his staff was worse than banishment to the outer prefectures of Ling-nan. There is evidence, too, that the princes were accompanied to their postings by their princely households, including their fief administrations. Most of the surviving princes and their immediate heirs holding provincial posts were killed or sent into penal banishment in the 680s and early 690s during the purges of empress Wu. But although princes continued to hold prefectural posts almost to the end of the seventh century there was no further talk of making the princes’ prefectural appointments hereditary. That such a plan was even contemplated, much less that an attempt was made to implement it, shows how highly T’ai-tsung valued the royal princes, and their potential importance as supporters of the dynastic regime. T’ai-tsung saw enfeoffed family members as a powerful source of protection for the throne, “as fences and screens to the dynasty.” Others, perhaps more realistically, thought that they were more likely to become serious potential rivals to imperial authority entrenched hereditarily in their own semiautonomous power-bases away from the capital. Later in the dynasty, feng-chien government was from time to time discussed, but the term is then usually used more as a coded reference to the devolution of political authority, rather than as the literal granting of hereditary territorial powers to members of the royal family and other great persons that T’ai-tsung had envisaged. On the one occasion, during the An Lu-shan rising, when the policy of appointing members of the royal family to command hereditary fiefdoms was attempted, the only royal prince who actually proceeded to his fief promptly rebelled and attempted to establish an independent regime.105 T’ai-tsung may have failed to establish hereditary territorial fiefs for his family members, but in his time and during the following reigns the royal princes nevertheless enjoyed far more powers and privileges than their successors later in the T’ang.

105  This was Li Lin, prince of Yung 永王李璘, Hsüan-tsung’s 16th son; cts 107, pp. 3264–66; hts 82, pp. 3611–12; tctc 219, pp. 7007, 7009, 7019.

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

77

The T ’ ang Imperial Family

The Household Establishments of Royal Princes There were several grades of prince in the T’ang system of nobility: Title

Rank

Nominal fief

prince of the blood (ch’in-wang 親王) successor prince (ssu-wang 嗣王) prince of commandery (chün-wang 郡王)

1.1

10,000 households

1.2

5,000 households

1.2

5,000 households

Only the princes of the blood, that is the brothers and the sons of the reigning emperor, enjoyed special household establishments that were similar in general structure to that of the heir-apparent, but on a much reduced scale.106 This entitlement continued on the statute book until the ninth century, even though in reality these great princely administrations did not outlast the first decade of Hsüan-tsung’s reign. In the very early years of the dynasty the other royal princes, the hereditary princes (ssu-wang), that is the eldest sons of princes of the blood by their principal wives, and the princes of commanderies (chün-wang), the sons of the heir-apparent, although they had no princely administrations, were provided with subordinate staff to administer their fiefs of maintenance.107 It is 106   t lt 29, pp. 6a–8a; cts 44, pp. 1914–15; thy 67, pp. 1172–73; hts 49B, pp. 1305–8; Fonct. 2, pp. 629–44, for formal descriptions of the princes’ household organizations. 107  Such an organization is not included in any of the main descriptive sources, all of which are based on the K’ai-yüan Statutes of 719 or of 737. However, the ms. fragments of ch. 6 of the 651 Statutes (Yung-hui ling 永徽令) discovered at Tun-huang include such a clause. See Yamamoto Tatsurō, Ikeda On, and Okano Makoto, Tunhuang and Turfan Documents concerning Social and Economic History (Tokyo: Tōyō bunko, 1978, 1980), vol.1A, p. 103 (ll. 51–54). The organization they describe is a somewhat reduced version of the fief administration of a prince of the blood, with a director (ling), a supervisor of agriculture (ta-nung), two marshals (wei), a chief and deputy of Offices of Instruction (hsüeh-kuan chang, ch’eng) and of Stables (chiu-mu chang, ch’eng), a chief of another office that is illegible, and five secretaries (she-jen). No ranks are attributed to these posts. The same ms. describes similar staff provided for dukes of states (kuo-kung) and commandery dukes (chün-kung); 11. 55–64.

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

78

Twitchett

unknown when this provision was ended. It no longer existed in the 730s, when T’ang liu-tien was compiled, and is never mentioned, as far as I know, in any contemporary document. Establishments of Princes of the Blood Each prince of the blood had his personal official establishment, known as his Princely Administration. The highest ranking officials included in this were a group of preceptors and advisers, forming his personal entourage and known collectively as the prince’s officers (wang kuan 王官): 1. a mentor ( fu 傅, rank 3.2);108 2. an administrative counsellor (tzu-i ts’an-chün shih 諮議參軍事, rank 5.1); 3. a companion (yu 右, rank 5.4); 4. an indeterminate number of readers-in-waiting (shih-tu 侍讀); 5. two masters of letters (wen-hsüeh 文學, rank 6,3);109 6. a master of ceremonies in the Eastern Hall (tung-ko chi-chiu 東閣祭 酒, rank 7.3); 7. a master of ceremonies in the Western Hall (hsi-ko chi-chiu 西閣祭 酒, rank 7.3). These officials counseled and advised the prince and saw to his education and practical training.110 Most of the cases on record show these posts filled as honorary concurrent appointments (chien-kuan 兼官) for high ranking members of the central bureaucracy. The princely administration also included his routine administrative staff. This consisted of:

108  In the Statutes of 624, the mentor held the title preceptor (shih). This was changed to mentor in 711. The post was abolished in 714, but reestablished shortly after; thy 67, p. 1172; tlt 29, p. 6b. 109  Two according to tlt 29, p. 7b; cts 44, p. 1914. One according to hts 49B, p. 1305. 110  The Statutes of 624 included an additional set of officials, two attendants-in-ordinary (ch’ang-shih) and four gentlemen attendants-in-ordinary (ch’ang-shih lang), four secretaries (she-jen) and receptionists (yeh-che). According to thy 67, p. 1172; hts 49B, p. 1305, these posts were all eliminated in 715, in the first set of K’ai-yüan Statutes. However, they were probably abolished before 651, since ch. 6 of the Tunhuang ms. Yung-hui ling (see n. 107, above) lists only the same posts as those in tlt 29, p. 1a; cts 44, p. 1914; and hts 49B, p. 1305. See Tun-huang and Turfan Documents, p. 105, ll. 42–46.

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

79

The T ’ ang Imperial Family

1. 2. 3.

a chief administrator (chang-shih 長史, rank 4.3); an assistant administrator (ssu-ma 司馬, rank 4.4); and two directors of services (yüan 掾, rank 6.1).

These officials presided over a central clerical staff, consisting of: a. a registrar (chu-pu 主簿, rank 6.4); b.  two secretarial aides (chi-shih ts’an-chün shih 記室參軍事, rank 6.4); c. an administrative supervisor (lu-shih ts’an-chün shih 錄事參軍事, rank 6.4); d. an office overseer (lu-shih, rank 9.4); e. two administrators (ts’an-chün shih 參軍事, rank 8.3); f. four unattached administrators (hsing ts’an-chün shih 行參軍事, rank 8.4); and g. two document clerks (tien-ch’ien 典籤, rank 8.4). The executive functions of the prince’s household were carried out by seven specialized services (ts’ao 曹), each headed by an administrator (ts’an-chünshih, rank 7.1): 1. Service of Personnel (Kung-ts’ao 功曹); 2. Service of Granaries (Ts’ang-ts’ao 倉曹); 3. Service of Finance (Hu-ts’ao 戶曹, these first three under the oversight of one of the directors of services (yüan 掾); 4. Service of Troops (Ping-ts’ao 兵曹); 5. Service of Cavalry (Chi-ts’ao 騎曹); 6. Service of Law (Fa-ts’ao 法曹); and 7. Service of Works (Shih-ts’ao 士曹), these latter controlled by the second director of services (yüan).111 Each prince also had his personal military establishment, comprising his Personal Guard (Ch’in-shih fu 親事府), and his Palace Escort Brigade (Changnei fu 帳內府). Each of these was headed by:

111  At the very beginning of the dynasty there had been three additional services, the Service of Armor (K’ai-ts’ao), the Service of Lands (T’ien-ts’ao), and the Service of Waterways (Shui-ts’ao).

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

80

Twitchett

1. two brigade colonels (tien-chün 典軍, rank 5.1); 2. two lieutenant-colonels ( fu tien-chün 副典軍, rank 5.2); and 3. five majors (hsiao-wei 校尉, rank 6.3), with a cadre of junior officers, captains (lü-shuai 旅帥, rank 7.4), lieutenants (tui-cheng 隊正, rank 8.4), and second lieutenants ( fu tui-cheng 副隊正, rank 9.4) that varied in number according to the importance of the prince. The nominal strength of these units in T’ai-tsung’s reign was 450 men in the Personal Guard, and 667 men in the Palace Escort Brigade.112 Finally, the prince had a separate administration for his fief (ch’in wang kuo 親王國). The fiefs of royal princes and other nobles as actually established were quite different from those envisaged by T’ai-tsung. Although their titles, being mostly the names of ancient kingdoms or commanderies, usually had regional connotations, they were not territorial jurisdictions, nor did the prince reside in or have any direct special relationship with his fief. The fief consisted merely of a specified number of fief households ( feng-hu 封戶) in a designated prefecture (usually unconnected with the region implied in the prince’s title) the revenues collected from which were paid to the prince’s fief administration and formed his main source of income.113 The fief was usually described simply as of so many hundred households. But this figure was in almost all cases only the number of fief households to which the prince was nominally entitled, and he actually received the revenues from a much smaller fief of maintenance (shih-feng 食封, or shih shih-feng 實食封). Thus the fief organization in effect was the office that supervised the collection of the prince’s revenues and controlled his expenditures. Its personnel all held low ranks compared with the members of his household administration. The fief administration was headed by the director of the fief (ling 今, rank 7.4), assisted by a supervisor of agricultural works (ta-nung 大農, rank 8.4), a marshal (wei 尉, rank 9.2), and an assistant (ch’eng 丞, rank 9.4) with a non-ranking office staff of twenty-five. The director also controlled four subordinate bureaus, each consisting of a bureau chief (chang, 長 rank 9.2) and his assistant (ch’eng, rank 9.4). These bureaus were: 1. Office of Instruction (Hsüeh-kuan 學官) responsible for the instruction of the women of the prince’s household; 112   t lt 29, pp. 3b–4b; hts 49B, p. 1307. 113  It was common in the seventh century for the fief administrator to collect the regular taxes but to exchange these for local commodities in order to make a profit; e. g., cts 64, pp. 2430–31.

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

The T ’ ang Imperial Family

81

2. Office of Provisions (Shih-kuan 食官), which was also in charge of building works; 3. Office of Stables (Chiu-mu 廄牧); and 4. Office of Storehouses (Tien-fu 典府). The offices comprising a prince’s household according to the Statutes were thus very numerous and relatively high ranking: every prince’s mentors and household administration comprised 30 ranking civil officials, 5 of them holding the fifth and higher ranks; his bodyguards included somewhat more numerous ranking military officers; and his fief administration numbered another 12 ranking civil officials. Therefore, each princely household, if it had been staffed at full strength, would have had more ranking official posts than any one of the central government’s Six Boards (the staff of which could number up to twenty-seven) or the Administration of the Metropolitan District (thirty). Moreover, there were considerable numbers of princes of the blood. Until 682 a household administration was not established for a prince until he was old enough to leave the palace, normally at about the age of fourteen or fifteen, when he began to participate in the major court assemblies. Even so, twenty of Kao-tsu’s sons and twelve of T’ai-tsung’s were enfeoffed as princes. Of course, not all these princes were endowed with a fief or provided with a household administration. Some died in childhood, many others were reduced in rank, cashiered, or executed for involvement in various abortive conspiracies and for other political reasons. Moreover, those who were enfeoffed received only a tiny fraction of the fief households to which they were entitled.114 In the 680s and 690s the repeated purges under the empress Wu drastically thinned the ranks of the T’ang royalty, but after the restoration of Chung-tsung 114  We can trace the fiefs of 13 of Kao-tsu’s sons and 6 of T’ai-tsung’s. All received an original fief of maintenance of from 600 to 800 households. On the accession of Kao-tsung these were increased to 1,000, or in two exceptional cases to 1,500; for these data, see cts 64, pp. 2425–36; cts 76, pp. 2659–66; Ch’ien I 錢易, Nan-pu hsin-shu 南部新書, ch. chia, p. 6. On the basis of the regulations in the Statutes they were each entitled to 10,000 households. For the difference between paper entitlements (hsü-i 虛邑; hsü-feng 虛封) and actual fiefs of maintenance (shih-i 實邑; chen-i 貞邑; shih-feng 食封; shih shih-feng 實食 封), see the classic study of Niida Noboru 仁井田陞, “Tōdai no hōshaku oyobi shokuhō sei” 唐代の封爵及び食封, thgh 10 (1939), pp. 1–63. See also Tonami Mamoru 礪波護, “Zui no bōetsu to Tōsho no shokujippō” 隋の貌閲と唐初の食實封, thgh 37 (1966), pp. 153–82; rpt. idem, Tōdai seiji shakai shi kenkyū 唐代政治社會史研究 (Kyoto, 1986), pp. 249–91. This is translated with additional material in Jens Østergård Petersen, “The Sui Dynasty Inspection of Countenances and the Early T’ang Fiefs of Maintenance,” East Asian Institute Occasional Papers 2 (East Asian Institute, U. of Copenhagen 1988), pp. 5–59.

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

82

Twitchett

a systematic attempt was made to restore many princely lines that had been extinguished. Their attendant fiefs were restored, and the sizes of the fiefs granted to current princes and princesses were greatly increased. Beginning with Hsüan-tsung the royal family began to proliferate wildly. Nineteen of Hsüan-tsung’s thirty sons, 1 uncle, and his 4 brothers were ennobled as princes of the blood, while 34 of his grandsons were created princes of commanderies and another 24 dukes of states. The relentless growth of the imperial family continued after Hsüan-tsung’s abdication and death. Eleven of Su-tsung’s sons were ennobled as princes of the blood, as were 18 of Tai-tsung’s, 9 of Te-tsung’s, 22 of Shun-tsung’s, and 17 of Hsien-tsung’s sons. The drain on the treasury represented by so many inflated princely household establishments would have been incalculable, had the household posts all been filled and had the fiefs of maintenance to which the princes were formally entitled (the tax revenues from 10,000 fief households for a prince of the blood) been granted in full. At the end of T’ai-tsung’s reign, when the registered tax-paying households in the empire had still recovered only to 3,800,000, from a peak of 9,075,791 in 609, such a level of fief households would have been completely unsustainable. Even in the 720s, when the taxable population had almost doubled, if the regulations had been followed to the letter the revenues of the princes ennobled in Hsüan-tsung’s reign alone would have consumed the taxes from more than a quarter of a million households, and the salaries and allowances of their household officers an enormous additional sum. However, a crisis had already been reached by the second decade of Hsüantsung’s reign, a crisis probably provoked as much by the desire to restrict the princes’ political activities as by the costs of their maintenance. Beginning in 720, a series of problems beset the imperial family. First, the emperor’s younger brother Li Fan, prince of Ch’i 岐王李範, was involved together with highranking palace officials, one of them the consort of one of his sisters, in some sort of alleged conspiracy.115 Then in 724 the childless empress Wang was degraded after allegations that she had been dabbling in magic to ensure a child. 115   c ts 95, p. 3016. The offenses involved seem to have been minor. At this time the emperor had forbidden the princes and dukes from close intimacy with persons outside the court. P’ei Hsü-chi 裴虛己, consort of the Huo-kuo princess 霍國公主, Jui-tsung’s youngest daughter, was found guilty of making excursions and enjoying banquets with Li Fan, and of privately holding prognosticatory apocrypha—politically dangerous material. He was exiled to Ling-nan. In a linked incident, Liu T’ing-ch’i 劉庭琦, marshal of Wan-nien county, and Chang E 張諤, grand invocator (t’ai-chu 太祝) of the Court of Imperial Sacrifices, were demoted and banished on the quite extraordinary charge of drinking and composing poetry with the prince, who himself went unpunished. Hsüan-tsung’s explicit point was to remove bad influences from him.

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

83

The T ’ ang Imperial Family

She was never replaced. In 725 Li Yeh, prince of Hsüeh 薛王李業, the emperor’s other younger brother, was apparently the center of another plot that involved his brother-in-law and the director of the Imperial Household Department.116 Neither prince was himself punished in these incidents, but it was becoming clear that the lack of control over members of the royal family living independently outside the palace not only made them a potential danger in themselves but also made them vulnerable targets for political conspirators looking for a figurehead around whom they might form a political faction or plot a coup.

The Establishment of the Residences of the Ten Royal Princes

In 725, after the completion of the vastly expensive feng-shan 封禪 sacrifices on Mount T’ai, a radical change was made to the system.117 Since Hsüan-tsung’s accession, his sons had remained resident in the Inner Palace. Now many of the young princes were approaching maturity, and instead of setting up each of them in his own mansion in the city outside the palace, as had been the custom, a special great common residence (ta-chai 大宅) was constructed for them in the walled park east of An-kuo Temple 安國寺, in the northeast corner of Ch’ang-an. This site was divided up into separate residential courts (yüan 院) for the individual princes, and was known as the Residence of the Ten princes (Shih-wang chai 十王宅),118 after the ten sons of the emperor who were first settled therein. The compound was placed under the management of the palace eunuchs (chung-kuan 中官). The princes and their households were fed 116   c ts 95, pp. 3018–19. The various participants were charged with “privately discussing political good fortune and disaster” 私議休咎, i.e., political portents. Private prophecy was forbidden, and was punishable by death by strangulation under article 268 of the Code if it touched on predictions concerning the fortunes of the dynasty. His coaccused were Wei Pin 韋賓, director of palace attendants in the heir-apparent’s Left Secretariat, who was the younger brother of his consort, and Huang-fu Hsün 皇甫恂, director of the Imperial Household Service. Hsüan-tsung ordered Wei flogged to death and Huang-fu banished as prefect of Chin-chou 錦州 in Szechwan. The prince’s consort surrendered herself awaiting punishment, and Li Yeh himself dared not visit the emperor. But when they were summoned to court the emperor pardoned them completely. 117   c ts 107, pp. 3171–72; thy 5, pp. 51–52; hts 82, pp. 3615–16; tctc 213, pp. 6777–78; Hsü, T’ang liang ching 3, p. 27a; Ch’ang-an chih 9, p. 1a. 118  It was also known colloquially as the Ten Residences (Shih chai 十宅), the Sixteen Residences (Shih-liu 十六 chai), Ten Palaces (Shih-kung 十宮), or sometimes the Courts of Cordial Relatives (Mu-ch’in yuan 睦親院); see tctc 213, pp. 6777–78; cts 107, pp. 3171– 72; thy 5, pp. 51–52; hts 82, pp. 3615–16.

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

84

Twitchett

daily from the kitchens of the heir-apparent’s Court of Domestic Services, and some court officials were appointed as concurrent tutors-in-waiting, and came to the compound to instruct them in literature and calligraphy. The princes’ compound communicated with the emperor’s court in the Ta-ming Palace by an enclosed passageway. Later seven more of Hsüan-tsung’s sons were moved to the residence when they grew old enough to be enfeoffed. Some of these princes were killed both during the succession crisis in 737 and as a result of various alleged conspiracies in the 740s. Irrespective of the actual number of resident princes, however, the compound remained known as the Residence of the Ten princes. As their numbers expanded, the Residence spilled over into the nearby Eastern Imperial Park. Later a similar residence called the Court for the Hundred Grandsons (Pai-sun yüan 百孫院) was built to house the rapidly proliferating next generation. The princes seem to have lived in rather cramped quarters and under irksome restrictions. The residences housed the princes’ consorts and children and great numbers of palace women servants (kung-jen 宮人), nominally 400 for each of the courts of the Ten Princes, 30 or 40 for each court of the Hundred Grandsons. There was a special treasury in the residence, the Treasury of the Royal Brothers (Wei-ch’eng k’u 維城庫) from which the princes’ stipends were disbursed and their expenses paid. Similar but smaller compounds for the royal sons and grandsons were also built at the emperor’s separate summer palace, the Chiu-ch’eng kung 九成宮, and at the hot-spring palace east of Ch’ang-an, the Hua-ch’ing kung 華清宮, which he visited annually with many of his court entourage. This change put an end to the lavish independent life style that the imperial princes had previously enjoyed. Their households of officials and guards became superfluous, and although these organizations and provision for their offices remained in the statutes of 737, in practise they had already been drastically reduced to a mere handful of officials. The vestigial household administrations of the various imperial princes were all accommodated together in cramped common premises in Hsüan-p’ing ward 宣平坊, south of the Eastern Market 東巾 and far from the princely residences, and their officials ceased to have any real functions apart from making their formal appearance at the great seasonal audiences.119 Probably most of the posts simply remained vacant to be sometimes used as dignified sinecures.

119  See cts 107, pp. 3271–22; thy 5, p. 51; hts 82, p. 3616; also Hsü, T’ang liang ching 3, pp. 27a–b.

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

The T ’ ang Imperial Family



85

Ending of Provincial Posts for Royal Princes

For the time being the royal princes themselves continued to be given provincial appointments, for example, as military governors. But these were now usually clearly designated as nominal (ling 領) or in absentia (yao-ling 遙領) appointments. The prince remained in the palace, while the actual duties of his post were performed by a proxy holding the title of deputy-governor ( fu-shih 副使). Discontinuance of Imperial Branch Lineages ( fang) Hsüan-tsung was also responsible, at about the same time, for a further change in the status of the royal princes. Until his reign, on the death of every royal prince his descendants had formed a separate, royal branch-lineage ( fang 房). By the 730s there were forty-one such formal branches of the imperial family recognized, ten of which were collateral lines from predynastic generations. From Hsüan-tsung’s time no new branches were set up.120 The careful and detailed genealogy of family members previously kept up regularly by the Court for Imperial Clan Affairs (Tsung-cheng ssu 宗正寺) also seems to have fallen into disarray after the 760s. A new imperial genealogy was commissioned in 766, and a continuation of this was completed in 839.121 But at the end of 852 the Court for Imperial Clan Affairs was again complaining of widespread confusion in their records, and requesting that every household of the royal clan should compile a record of the date when every prince had been first enfeoffed, noting also his ancestors for five generations and listing his sons and grandsons. This was to be sent to the Court of Genealogy (T’u-p’u yüan 圖譜院) of the Court for Imperial Clan Affairs, so that they could unravel the confusion in the records and differentiate senior and junior lines.122 We are not told what resulted from this cry of desperation. The historians who wrote up the history of the later reigns certainly seem to have lacked even the most basic information about the imperial family. Whereas we are 120   h ts 70A, pp. 955–57; 70B, pp. 2147, 2176. 121  The formal imperial geneaology Yung-t’ai hsin p’u 永泰新譜, or Huang-shih Yung-t’ai p’u 皇室永泰譜 (20 ch.) was compiled in 766 by the official historian Liu Fang 柳芳 and presented to Tai-tsung in that year by Li Ch’i, prince of Wu 吳王李祈, then president of the Court of Imperial Clan Affairs. In 839 Wen-tsung ordered the vice-president of the Supreme Court of Justice Li Ch’ü 李衢 to bring this genealogy up to date, and Liu Fang’s grandson, the Han-lin academician Liu Ching 柳璟 was commissioned to write a continuation (10 additional ch.) covering the reigns from Te-tsung onward; cts 149, p. 4033, thy 36, p. 666. 122   t hy 36, p. 666.

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

86

Twitchett

reasonably well informed about the imperial clan members before the An Lushan rebellion, the level of information progressively declines thereafter. For the reigns after Mu-tsung we have only the barest listing of the names and fiefs of princes and princesses, with no indication even of the princes’ mothers, and with confusion over the number of imperial sons and daughters and the order of their birth for some reigns.123

The Royal Princes in the Late-T’ang Period

In 827 the chief administrator of the household of prince Ch’iung 瓊王 petitioned the throne complaining that the common office in Hsüan-p’ing ward that he and his colleagues in the households of other princes had once shared, and which had long been allowed to fall derelict, had been repossessed by the commissioners for estates (chuang-chai shih 莊宅使) in 818, and the site sold to a wealthy military governor. Since then, they had had no place whatever in which to perform even the formalities of their offices.124 As a result of this memorial they were given the use of another recently confiscated mansion in the Yen-k’ang ward 延康坊, near the Western Market and even farther from the princely residences, to use as a common office.125 Replacement of Fiefs by Money Stipends The financial independence of the princes, ensured by the system of fiefs of maintenance, was also greatly reduced. The numbers of fief households to which princes and other nobles were entitled as of right were very large, and although these entitlements bore little relationship with the much smaller fiefs of maintenance that were actually granted, the costs of the system were 123   t hy 5, pp. 52–54; hts 79–82. Chiu T’ang-shu’s coverage is as follows: ch. 64 (Sons of Kaotsu); 75 (T’ai-tsung); 86 (Kao-tsung, Chung-tsung); 95 (Jui-tsung); 107 (Hsüan-tsung); 116 (Su-tsung, Tai-tsung); 150 (Te-tsung, Shun-tsung); 175 (Hsien-tsung, Mu-tsung, Chingtsung, Wen-tsung, Wu-tsung, Hsiuan-tsung, I-tsung, Hsi-tsung, Chao-tsung). For many of the sons of Wu-tsung and his successors we lack information on their mothers, date of death, posthumous titles, and details of their heirs; we mostly receive the date of their enfeoffment. The genealogical tables of the imperial family in hts 70A–70B also give no details beyond the names and fief-titles of the sons of emperors from Mu-tsung onwards, rarely information beyond the grandsons’ generation for any emperor from Hsüan-tsung to Hsien-tsung. 124  It is significant of the loss of any real function for these offices that it took them nine years to draw the court’s attention to the fact that they no longer had any premises. 125   t hy 67, p. 1172; Hsü, T’ang liang-ching 3, pp. 23b–24a; 4, p. 13b; Ch’ang-an chih 10, p. 3b.

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

The T ’ ang Imperial Family

87

a problem on a national scale.126 In 765 the princes’ entitlements to fiefs of maintenance were reduced by a third “pending the end of the current military crisis.”127 The reduction was still in force twenty years later. In 797 the revenues from fief households were commuted to a standard payment of 800 lengths of silk in respect of each hundred nominal fief households.128 There was also a special Court for Ritual Assemblies (Li-hui yüan 禮會院), in the Ch’ung-jen ward 崇仁坊, which was used for the marriage ceremonies of the heir-apparent, royal princes, and princesses. After the rebellion this court became derelict, and was eventually sold off by the commissioners for estates.129 New premises in the Ch’ang-hsing ward 長興坊 were acquired for the same purpose in 814, but within a decade these too had fallen into disuse and disrepair.130 The only posts in the princes’ administrations that sometimes continued to be filled were those of their various preceptors, the mentors and companions. These offices, however, were now sinecures, titles merely allocated to highranking officials in the central government as honorary concurrent duties.131

The Royal Princesses

The daughters of an emperor presented a lesser, but still considerable, expense to the government. Imperial princesses, like their brothers, were bestowed 126  See the lengthy edict dated 709, included in thy 90, pp. 1642–44, which says that by that date the twenty or thirty enfeoffed households of Kao-tsu’s time had now grown to more than 140, and that over 600,000 taxpayers in 54 prefectures, all in the richest parts of the empire, were required to pay part of their revenues to various fief-holding nobles. The total expense was computed at 1,200,000 lengths of silk. 127   t hy 90, p. 1646. 128   t hy 90, p. 1647. 129   c ts 107, p. 3172; thy 5, p. 3616; hts 82, p. 3616. According to Ch’ien, Nan-pu hsin-shu, ch. ting, p. 37; Hsü, T’ang liang-ching 3, p. 4a, this court was established in 731, when the government bought the site, formerly the western half of the mansion of the Ch’ang-ning princess 長寧公主 and her consort Yang Shen-chiao 楊慎交. It was administered under the Court of Agriculture (Ssu-nung ssu 司農寺). It may have been abandoned because noble ladies were massacred there during the rebellion. 130  Hsü, T’ang liang-ching 2, p. 8b. One of the omissioners (shih-i) was dismissed at this time for requesting the use of it for his own family wedding; Ch’ien, Nan-pu hsin-shu, ting, p. 37. 131  E.g., in 832 the emperor Wen-tsung selected three well-respected Confucian scholars to serve the prince of Lu, one of whom became his monitor when he was installed as heirapparent later that year; thy 67, p. 1173.

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

88

Twitchett

with noble ranks and titles, and were granted fiefs of maintenance as a source of income.132 Title

Relationship

Rank

Fief

senior grand princess t’ai chang kung-chu 大長宮主 grand princess chang-kung chu 長宮主 imperial princess kung-chu 宮主 commandery princess chün-chu 郡主 county princess hsien-chu 縣主

paternal aunt of emperor

1.1

2,000 families

sister of reigning emperor daughter of reigning emperor daughter of heir-apparent daughter of prince of the blood

1.1

up to 1,000 families 300 families

1.1 1.2 2.2

All five grades of princess were considered as titled ladies of the outer court (wai ming-fu 外命婦) and as such enjoyed many privileges. However, whereas the royal princes remained members of the imperial clan, the princesses, like all women, ceased to belong to their family of birth as soon as they were married, and became part of their husbands’ households. From that time the government continued to provide the emoluments due from the princess’s fief, and of course provided a substantial dowry, but otherwise she was treated as the wife of her consort (who was given the title fu-ma tu-wei 駙馬都尉, commonly abbreviated to fu-ma 駙馬, or 副馬), not as a part of the royal family. Although the consort and his family were placed on a special register of relatives by marriage (chu-ch’in 諸親) kept by the Court for Imperial Clan Affairs, 132  In the table below, princess’s ranks are “equivalent ranks 視品,” the ceremonial nature of which is treated in Niida Noburu, Tōryō shūi 唐令拾遺 (Tokyo, 1933), pp. 316–17. See also tlt 2, p. 37b; hts 46, p. 1188. To confuse the issue, imperial princesses (kung-chu) were commonly referred to as the monarch’s princesses (chün-chu 君主). Only the first three categories, below, all of upper-first rank, were allowed fief administrations, although all grades of princess were granted fiefs of maintenance of a specified number of households contributing to their maintenance.

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

The T ’ ang Imperial Family

89

their children were not considered as royalty. A princess’s sons were, however, eligible to use her rank to claim entry into the roll of officials by hereditary privilege (yin 蔭).133 A royal princess had no formal household administration, except for a brief period from 706 to 710 when, while the court was under the dominance of Chung-tsung’s empress Wei, her daughter princess An-lo, and empress Wu’s powerful daughter, the princess T’ai-p’ing, princesses were provided with administrations and household officials similar to those for royal princes. The princess T’ai-p’ing was granted a household administration and entitlement to dignities equal to those of a prince of the blood with, in addition, a real fief of 10,000 households. The empress Wei’s own two daughters, princess Ch’angning 長寧 and princess An-lo were granted the same household administration but without a chief administrator. They were also provided with their own guards and fiefs of maintenance of 2,500 and 3,000 households, respectively. This did not satisfy princess An-lo, who demanded unsuccessfully, as we have already seen, to be created crown princess-apparent (huang t’ai-nü 皇太女). The remaining royal princesses, Chung-tsung’s daughters by other royal consorts, were granted establishments one-half this size. These dignities for royal princesses were ended with the fall of empress Wei in 710 and the accession of Jui-tsung, except in the case of his sister the princess T’ai-p’ing, who was still too powerful politically for him to be able to reduce her entitlements. After her fall and death in 713 no other princess was ever again provided with a full household administration.134 The only exception was made for royal princesses married off to nonChinese rulers in dynastic marriages (so-called ho-ch’in 和親).135 In such cases it was considered appropriate to “specially establish an administration” (t’e chih 133   t hy 6, p. 72. An edict of 799 denied this privilege to adopted sons of princesses. 134  See thy 6, p. 69; cts 51, p. 2172; hts 83, p. 3650. 135  This seems to have applied only to imperial princesses, and not even to all of them. Most of the “princesses” married to foreign rulers under the T’ang were daughters of more distant collateral relatives, or daughters of imperial clan members. This was admitted in a speech made by the T’ang delegate Li Yü, commandery prince of Han-chung, who escorted Su-tsung’s daughter, the Ning-kuo princess 寧國公主 (previously twice married), to wed the Uighur khaghan in 758:  Up to this time, when China has made a marriage alliance with neighboring countries the brides have been daughters of imperial clan members, given the name “princesses.” Now, however, the princess Ning-kuo is a real daughter of the emperor. Moreover she has talent and good looks, and has made a marriage with the khaghan ten-thousand li from home. You, Khaghan, are now the T’ang emperor’s son-in-law, and appropriate ritual

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

90

Twitchett

fu 特置府) for her, equivalent to the household of a royal prince of the blood. This household administration accompanied her to her new country as a personal entourage.136 In normal times a princess had no personal administration, only a small office (i-ssu 邑司) to administer her fief.137 This was headed by a director (rank 7.4) and an assistant (rank 8.4), who had control of the princess’s revenues, her personal wealth,138 her accumulated goods, and her lands and gardens.139 Two low-ranking registrars, rank 9.2, and an overseer had control of the tax revenue from her fief, and managed her household’s incomings and behavior is to be expected. How could she be made to sit on the couch and receive your commands?  See cts 195, pp. 5200–1; cts 10, p. 253; hts 83, p. 3360; 217A, p. 6116–17. For a list of hoch’in marriages, see thy 6, p. 75. Other material is included in tfyk 978 and 979. 136  In the case of Ning-kuo (see previous n.), accounts of her marriage do not mention establishment of an administration for er, but when she returned to Ch’ang-an in 789 her Princess’s Administration (kung-chu fu) was abolished and replaced by a normal fief office (i-ssu); thy 6, p. 75. Elsewhere she is said to have voluntarily given up her establishment; hts 83, p. 3660. Te-tsung’s daughter, the Hsien-an princess 咸安公主 (later Hsiang-mu, of the state of Yen), was married to the Uighur khaghan in 788. Before the marriage Te-tsung established an administration with officials and subordinates equal to those attributed to a prince of the blood; cts 195, p. 5208; hts 217A, p. 6124; tfyk 979, pp. 16b–17a. A similar special arrangement (parity with the household of a royal prince) was made when the T’ai-ho princess 太和公主, Mu-tsung’s tenth younger sister, was married to the Uighur khaghan in 821; cts 195, p. 5211; hts 217B, p. 6129. That this was a specific arrangement relating to her marriage was clearly understood, since when in 843 she returned to Ch’ang-an after the collapse of the Uighur empire she was first promoted as senior princess, after which her administration was abolished; hts 83, pp. 3668–69. 137  Except for the brief period of Chung-tsung’s reign, the fiefs of maintenance actually provided for princesses were comparatively small. Until 682 these fiefs did not exceed 300 households (princes of the blood then received a standard grant of 800, in exceptional cases up to 1,000). Princess T’ai-p’ing, as empress Wu’s only daughter, was exceptionally, granted 350. The granting of extravagant fiefs of maintenance began in 698, when T’aip’ing’s fief was increased to 3,000, and reached new excesses after the accession of Chungtsung in 705, when her fief was increased to 10,000 and those of Ch’ang-ning and An-lo were increased to 2,500 and 3,000. (thy 90, p. 1638, gives these figures as 3,500 and 4,000.) Newly acceded, Hsüan-tsung attempted to reduce the fiefs of princesses to 300 households, but met with strong opposition and eventually established a standard of 1,000. 138  The most important item was presumably her dowry, and since a good many princesses married more than once, the fief office would ensure that this remained the property of the princess, and did not remain with the family of her husband. 139  Princesses sometimes held lands of their own. The Han-yang princess 漢陽公主, eldest of Shun-tsung’s daughters, once repudiated her husband Kuo Ts’ung 郭鏦. Later in life

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

The T ’ ang Imperial Family

91

expenditures.140 The fief administrations of the various princesses came under the general control of the Court for Imperial Clan Affairs. After the An Lu-shan rebellion the fief revenues of princes and princesses alike were, as we have seen, reduced.141 Following the “liang-shui fa” tax reform of 780 it was no longer practical to earmark the tax revenues from designated wealthy households in specific prefectures to pay for the support of individual princesses, and in 791 their fief revenues were replaced by a tariff of set payments, either in cloth or cash.142 But this rule must have been rescinded at some point or simply disregarded, as in 851 they were forbidden to have their fief administrations approach local prefectures and counties directly, except in respect of the revenues due from their fief households.143 The Consorts of Royal Princesses The consorts of royal princesses were carefully chosen.144 Marriage into the royal family always had a political dimension, and almost all consorts came from well-established aristocratic families, and were often the sons of highranking officials or military officers. Unlike the royal princes, who after Hsüantsung’s time were deliberately kept out of public life and never granted high substantive offices, the consorts of princesses were almost always related to men holding high positions, and often themselves held substantive offices,

she became a by-word for her frugality, said to have scratched a record of her rent receipts on the wall with an iron hairpin; hts 83, p. 3665. 140   t lt 29, pp. 20a–b; cts 44, p. 1915; hts 49B, p. 1308. hts lists only a single registrar. There was also a small office staff of eight clerks (shih), two receptionists (yeh-che), two secretaries (she-jen), aand two domestic attendants (chia-li). 141   t hy 90, p. 1645. Under the new scale, imperial princesses were allocated 500 fief households, and lesser princesses had theirs reduced by one-third. All their revenues were to be collected in the prefectures of Hsüan-chou (southern Anhwei), Ming-chou, Yüeh-chou, Chü-chou, and Mu-chou in Chekiang. 142   t hy 90, p. 1647. Henceforth imperial princesses were to receive as fief payments ( fengwu) 700 lengths of fine hemp cloth payable in two annual installments corresponding to the two annual collections of taxes under the “liang-shui” system. Commandery princesses were to receive 100 strings of cash quarterly, and county princesses 70 per quarter. 143   t hy 6, p. 74. 144  Lists of the daughters of successive emperors, with details of their marriages, are given in thy 6, pp. 63–66; hts 83, pp. 3641–76. These lists are, however, incomplete, particularly after about 850. On these marriages, see Wang Shou-nan 王壽南, “T’ang-tai kung-chu chih hun-yin” 唐代公主之婚姻, in Li Yu-ning 李佑寧 and Chang Yü-fa 張玉法, eds., Chung-kuo fu-nü shih lun-wen chi 中國婦女史論文集 (Taipei, 1988) 2, pp. 90–144.

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

92

Twitchett

sometimes becoming very important officials.145 Many of their sons too became officials, sometimes rising to high rank. Thus these relatives by marriage of royal princesses were often deeply involved in the politics of the day, and their royal wives provided them with an important personal link to the emperor and the palace. Some of the princesses in this way played an indirect but nonetheless important role in politics. The royal princesses lived outside the palace in their own private mansions, often in great splendor. When the emperor wished to issue an order to them he would employ palace eunuch messengers (chung-shih 中使) to deliver it, but in general the princesses enjoyed considerable independence. Not all of the princesses married; some entered the Taoist religious life and became nuns or abbesses of Taoist temples. Others, even though they were married, lived openly promiscuous lives, in some cases being sequestered in the palace to avoid further public scandal.146

145  This problem is very complex and requires a careful study. In the early-9th c. it seems to have been the practise that a princess’s husband would first serve two full tours of duty in an acting (chien-chiao 檢校) post, and would then be eligible for an appointment to a regularly established court office (cheng-yüan kuan 正員官). After 808 their appointments were restricted to the capital and princesses were no longer allowed to accompany their husbands to provincial postings. Their postings and promotions were, in theory at least, supposed to come under the overview of the Court of Imperial Clan Affairs; thy 6, p. 72. 146  E. g., the Yen-kuang princess 延光公主 (later princess of the state of Kao 郜國), the seventh daughter of Su-tsung. Originally she was married to P’ei Hui 裴徽, and then remarried to Hsiao Sheng 蕭升. After the death of her second husband she had a string of lovers. When this became known Te-tsung was furious and secluded her in her residence, while her principal lover was flogged to death and three others banished to Ling-nan. In 788 she was stripped of her title for practising ku 蠱 black magic, while five of her sons were imprisoned in distant parts of the empire for their involvement. Her daughter was the consort of the heir-apparent, and the emperor, fearing the power of the princess’s hatred, was going to have her killed, but she died in 790. The heir-apparent just then fell sick, and to prevent further disasters from the magic spells his consort too was killed; hts 83, p. 3662.  One of Te-tsung’s daughters, the princess I-yang 義陽公主, was secluded in the palace, and her consort confined in fetters in their mansion because of their scandalous impropriety; hts 83, p. 3664. Shun-tsung’s sixth daughter, the Hsiang-yang princess 襄陽 公主, who had been married to Chang K’o-li 張克禮, the son of a Ho-pei military governor, also led a scandalous life, going into the city and the marketplace incognito. She had several paramours, and paid officials heavily to hush the matter up, but eventually was discovered by her husband, and sequestered in the palace by Mu-tsung; hts 83, p. 3666.

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

The T ’ ang Imperial Family

93

Like other high-born ladies of their era, they remarried freely without this being considered an outrage to morality, as it would have been in later centuries.147 Some were remarried more than once. In law they were forbidden to remarry if they had already born sons, but this rule was not strictly enforced.148 In the later part of the ninth century, remarriage seems to have become less common, and in 851 the remarriage of princesses with living children was forbidden by edict. In case a princess without children wished to remarry, the Court for Imperial Clan Affairs was responsible for determining whether the marriage should be permitted. If a princess falsely claimed to have no children in order to facilitate her remarriage her case was to be referred to the competent authorities for punishment.149 Summary For the early-T’ang state, engaged in reestablishing a single dynastic power after the failure of the Sui empire, it was essential not only that the regime establish temporal stability, but that it should also be seen as preparing for permanence. This permanence could be ensured only by measures to protect the succession to the throne, and to ensure that it should descend generation after generation to an acceptable successor, the choice of the dying emperor and his principal advisors, and that he should be prepared to face the daunting demands of his office. As an institutional foundation for this, the state made generous provision for the support and maintenance of the large pool of imperial children. To ensure that the designated successor, the heir-apparent, who would be chosen from their number, would be given a suitable training under chosen tutors and administrative experience to fit him for becoming emperor, complex arrangements were made and large numbers of officials were involved.

147  See the entries on royal princesses in thy 6, pp. 63–66; hts 82. The details on their spouses are by no means complete, particularly for later emperors’ daughters, but there are a good many cases of remarriage. Three of Chung-tsung’s eight daughters remarried, 3 of Jui-tsung’s, 7 of Hsüan-tsung’s (plus a daughter who was married three times), and 2 of Su-tsung’s seven daughters. Remarriage seems to have become less common in the ninth c., but this reflects the fact that most of the data were lost, rather than a change in mores. 148   h ts 46, p. 1188. 149   t hy 6, p. 74. These new rules were to be conveyed to the Court for Noble Ladies (Ming-fu yüan 命婦院) and to become a permanently applicable ordinance (ch’ang-shih 常式).

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

94

Twitchett

In the first quarter-century of the dynasty, when grip on the provinces was still far from secure, T’ai-tsung viewed the other imperial princes as potential allies who could be enfeoffed to rule strategic regions of the empire on behalf of the emperor. Although the hereditary succession among these princes was almost immediately abandoned, the idea persisted until the early-eighth century. But it also became clear that in reality large numbers of imperial princes, removed from strict court control, were liable to become potential rivals rather than allies. The succession and the arrangements for the royal family were destabilized in the reign of Kao-tsung by his powerful empress Wu, who wished to ensure the succession for one of her own sons, deposing two heirs-apparent in the process. After her husband’s death in 683 she dominated the court as regent and decimated the imperial family in repeated purges. Dismissing two young emperors, eventually she set up a dynasty of her own from 690 to 705, during which period she attempted various ways of transferring the succession to her own Wu clan. Under her successor, Chung-tsung, real power was once again wielded by his empress, who even attempted to appoint a daughter as the imperial heir. The institutional arrangements designed in the early-T’ang proved resilient, however, although the succession had rarely gone as originally intended. Under Hsüan-tsung, who had suffered a number of threats to his own position before coming to the throne, it became apparent that the great number of imperial princes living independent lives, with some of them serving in influential offices, was a source of political instability. The princes were resettled in a special palace complex, their social contacts limited, and they were forbidden to be appointed to any but nominal offices. By degrees these restrictions were retained and made even more severe under Hsüan-tsung’s successors. The princes’ “fiefs” were replaced by money salaries, which were repeatedly diminshed, their administrations and numerous staff were gradually reduced almost to nil, and by the early-ninth century the princes, living in the cramped quarters of the Princely Residences, were an object of pity to some of the court, who described them as a pool of frustrated talent with no outlet for their abilities. The arrangements for selecting and training a dynastic heir, on the other hand, were restored and strengthened, and probably functioned better during the latter half of the eighth century than at any other period under the T’ang. This was important to the dynasty’s survival, since this was a period of severe pressures from growing provincial autonomy. Not until the unexpected succession of Hsien-tusng in 805 did an emperor succeed to the throne without many years of preparation as heir-apparent.

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

95

The T ’ ang Imperial Family

But this system too was undermined in the ninth century. The succession, which up to this time in almost every case had been determined by the will of the deceased emperor and his chief ministers, now fell into the hands of the palace eunuchs, who arranged every succession after 805, sometimes with scant attention either to the principle of generational seniority or to the dying emperor’s intentions. A series of young emperors, from 820 to 845, who died leaving only infant sons, broke the generational pattern of the succession, and disturbed the institutional arrangements for the appointment of an heirapparent. After 838 there was never a serving heir-apparent, the eunuchs’ chosen successor being appointed heir only days before succeeding. The monarchy was fatally undermined, having lost control of its own destiny. All emperors after Mu-tsung were chosen from a large pool of princes, living in their residences, totally lacking any experience of government and with very limited contacts beyond the palace. Many of them came to the throne very young, sometimes as little more than children. Only one of them lived to the age of fifty. From the end of Hsiuan-tsung’s reign in 859, during which some degree of administrative regularity was restored but no successor had been designated, the T’ang royal house declined into little more than a symbol of continuity, a continuity that they could no longer themselves influence. The careful provisiuons put in place by the founding emperors had come to nothing.

List of Abbreviations

ckcy Chen-kuan cheng-yao 貞觀政要 cts Chiu T’ang-shu 舊唐書 ctw [Ch’in-ting] Ch’üan T’ang wen [欽定]全唐文 Fonct. des Rotours, Traité des fonctionnaires et traité de l’armée hts Hsin T’ang-shu 新唐書 kyl K’ai-yüan li 開元禮 tctc Tzu-chih t’ung-chien 資治通鑑 tfyk Ts’e-fu yüan-kui 冊府元龜 thy T’ang hui-yao 唐會要 tlt T’ang liu-tien 大唐六典 tt T’ung-tien 通典 ttclc T’ang ta chao-ling chi 唐大詔令集 whtk Wen-hsien t’ung-k’ao 文獻通考 wyyh Wen-yüan ying-hua 文苑英華

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

96 Table 1

Twitchett Life-Spans of T’ang Emperors

Emperor

Service as heir-apparent

Age at accession

Age at death

Kao-tsu T’ai-tsung Kao-tsung Chung-tsung (i) Jui-tsung (i) Empress Wu Chung-tsung (ii) Jui-tsung (ii) Hsüan-tsung Su-tsung Tai-tsung Te-tsung Shun-tsung Hsien-tsung Mu-tsung Ching-tsung Wen-tsung Wu-tsung Hsiuan-tsung I-tsung Hsi-tsung Chao-tsung Ai-tsung

none none 6 yrs. 3 yrs. none

52 27 21 27 22 67 49 48 27 45 35 37 44 27 24 14 17 25 35 25 11 21 12

69 50 55 54 54 82

none 2 yrs. 18 yrs. 4 yrs. 15 yrs. 26 yrs. 3 mos. 8 yrs. 13 mos. nonea none none none none none none

76 50 52 53 45 42 28 17 30 31 49 39 25 47 15

note: This table is based on the rulers’ actual western-style ages, calculated on the basis of information in A. C. Moule, The Rulers of China (London, 1957). a  Emperors from Wen-tsung onward were nominally appointed heir-apparent, but only immediately before their accession. They never actually served as heir.

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

97

The T ’ ang Imperial Family Table 2

T’ang Emperors and Heirs-Apparent

Emperor (No. of sons)

Li 李 Heirs-apparent (Frat. birth order)

H-A outcome / other remarks

Kao-tsu; r. 618–26 (22 sons) T’ai-tsung; r. 626–49 (14 sons)

Chien-ch’eng (1st) h-a. 618–26 Ch’eng-ch’ien (1st) h-a. 635–43 Chih (9th) h-a. 643–49 Chung (1st) h-a. 652–56 Hung (5th) h-a. 656–75 Hsien (6th) h-a. 675–80 Che (7th) h-a. 680–83 none

murdered, 626

Ch’eng-ch’i (1st) h-a. 684–90

ceased h-a. when Wu dethroned Jui-tsung, 690.

Kao-tsung; r. 649–683 (8 sons)

Chung-tsung; r. 694 (4 sons) Jui-tsung; r. 684–90 Chung-tsung jr. bro.; Kao-tsung 8th son. (6 sons) Wu-Chou Dynasty Empress Wu; r. 690–705

Restored T’ang Chung-tsung; r. 705–10 (4 sons) [Wen-wang Chung-tsung 4th son.]

deg. 643; d. 644 suc. 649 deg. 656 d. 675 deg. 680 suc. 684, under empress Wu regency. dethroned by Wu, 684

former Jui-tsung “imp. ceded position, 698 successor” & renamed Wu Lun, 690–98. former Chung-tsung h-a. 698 suc. 705

Ch’ung-chün (3d) h-a. 706–7 no h-a. 707–10

rebelled & killed, 707

installed briefly by Wu, 710.

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

98 Table 2

Twitchett T’ang Emperors and Heirs-Apparent (cont.)

Emperor (No. of sons)

Li 李 Heirs-apparent (Frat. birth order)

H-A outcome / other remarks

Jui-tsung; r. 710–12 abdicated, 712; “retired emperor,” 712–13. (6 sons) Hsüan-tsung; r. 712–56 abdicated, 756; “retired emperor,” 756–62. (30 sons)

Lung-chi (2d) h-a. 710–12

suc. 712; with full powers, 713

none, 712–15 Ssu-ch’ien (3d) h-a. 715–37 Yü (3d) h-a. 738–56 Yü (1st) h-a. 758–62 K’uo (1st) h-a. 764–79 Sung (1st) h-a. 779–805 Ch’un (1st)

condemned to suicide, 737 renamed Heng, 744; usurped throne, 756.

none, 805–09 Ning (1st) h-a. 809–11 Heng (3d) h-a. 812–20 Chan (1st) h-a. 822–24 none

d. 811 suc. 820

Su-tsung; r. 756–62 (14 sons) Tai-tsung; r. 762–79 (20 sons) Te-tsung; r. 779–805 (11 sons) Shun-tsung; r. 805 abdicated, 805 (23 sons) Hsien-tsung; r. 805–20 (20 sons)

Mu-tsung; r. 820–24 (5 sons) Ching-tsung; r. 824–27 (5 sons) Yung (1st) Wen-tsung; r. 827–40 h-a. 832–38 Ching-tsung jr. bro.; none, 838–40 Mu-tsung 2d son. (2 sons) none Wu-tsung; r. 840–46 Ching-tsung and Wen-tsung jr. bro.; Mu-tsung 5th son. (5 sons)

suc. 762 suc. 779 suc. 805 suc. 805

suc. 824

killed in palace, 838

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

99

The T ’ ang Imperial Family

Emperor (No. of sons)

Li 李 Heirs-apparent (Frat. birth order)

Hsiuan-tsung; r. 846–59 Mu-tsung jr. bro.; Hsien-tsung 13th son. (12 sons) I-tsung; r. 859–73 Hsiuan-tsung 1st son (8 sons) Hsi-tsung; r. 873–88 I-tsung 5th son (2 sons) Chao-tsung; r. 888–904 Hsi-tsung jr. bro. I-tsung 7th son. (10 sons) Ai-ti; r. 904–7 Chao-tsung 9th son

none

H-A outcome / other remarks

none

none

Yü (1st) h-a. 897–904

murdered, 904

none

note: h-a. = heir-apparent / served as heir-apparent deg. = degraded d. = died suc. = succeeded to imp. throne

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

Canonical Scholarship David McMullen The Confucian canons occupied a central place in the intellectual life of T’ang China. They provided the basis of education for successive generations of the intellectual elite, and a part of the ‘memorization corpus’ that scholars took through their serving careers.1 Mastery of the canons was assumed, to be mentioned in biographical accounts only if it was achieved very young, very quickly, or, just occasionally, very late.2 There was no question of an alternative curriculum to any who aspired to official service. The canons pointed those who studied them in the direction of an official career. They sanctioned almost all the important activities in which the state was involved. The very activity of governing was traced back to canonical sources, and the institutions that formed the T’ang state were likewise considered to have had canonical origins. T’ang scholars justified all the administrative operations with which this book is concerned by appeal to the canons: education, the ritual programme, the study and documentation of the past, both recent and remote, and the composition of literature in the service of the state. The influence of the canons, however, like the political ambition of those who interpreted them, was greater even than this. Almost every policy or measure that the dynasty formulated might be introduced by, or might later elicit, Source: “Canonical Scholarship,” in David McMullen, State and Scholars in T’ang China, Cambridge: Cambridge University Press, 1988, 67–112. © Cambridge University Press, reproduced with permission. 1  Hartman (1986), p. 236. 2  References to early mastery of the canons are very numerous in biog. accounts at all levels of documentation. For the cases of some well known scholars, see stts 10/11a, specifying the Tso chuan, quoted by Pulleyblank (1961), p. 137; wyyh 924/5b–6a, Ts’ui Yu-fu, of Tu-ku Chi, specifying the Hsiao ching; wyyh 678/5a–7a, Hsiao Ying-shih, specifying the Analects and Shang shu; wyyh 690/11b, Lu Kuei-meng, specifying the Six canons, with Mencius and Yang Hsiung. For some examples of late, but successful acquisition of learning, Yüan Tz’u-shan chi, Appendix p. 166, Yüan Chieh at 17 sui; Ch’en Tzu-ang chi, Appendix p. 252, Ch’en Tzu-ang at 17 or 18 sui; Yen Lu kung wen chi 9/10a–b, of Hsien-yü Hsiang, 693–752, who learnt to read only at twenty, and succeeded at the chin shih at ‘close on forty’. For an instance of notably rapid learning, see Yen Lu kung wen chi 9/5a, spirit path stele for Yen Yüan-sun, who mastered the Shang shu ‘in six days, including the commentary’.

© koninklijke brill nv, leiden, ���9 | doi:��.��63/9789004380158_006

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

Canonical Scholarship

101

relevant comment from them. They were quoted in the context of symbolically remote yet politically vital questions such as the management of the barbarians, and for issues central to the prestige of the dynastic house, like the maintenance of the imperial ancestral temple. They contained the principles of political morality, to which the conduct of all, from the sovereign to the individual serving official, should conform. The very comprehensiveness of the canonical corpus, the fact that it could provide guidance on almost the full range of issues that, in the political arena, scholars were confronted with, was one of the reasons for its unassailable position. There was a basic assumption among the intellectual elite that in any government context administrative competence was not enough. Their ideal was to ‘embellish administration with Confucian [learning]’ (i ju shih li).3 To fail to do this would lead to an unacceptable lack of moral awareness, to a purely administrative mode of government that would degenerate into disorder and tyranny. Confronted with the prestigious inheritance of the canons and their subsequent exegetical literature, the state undertook certain practical operations. The motive behind these was to ensure that the texts of the canons were widely disseminated, and that they were understood in ways compatible with the interests of the dynasty. The first of these operations was to determine, preserve and make widely available standard versions of the texts, and to provide, in the form of commentaries and sub-commentaries, a definitive understanding of them. High authority discharged this responsibility intermittently throughout the dynasty. When it was overlooked, individual scholar officials initiated attempts to attend to it. It resulted in a series of practical measures. The most important of these involved the display of the standard text of the canons at the state academy directorate, first written on wood and then, from the mid ninth century, engraved on stone. A single interpretation of the 3  For the term li or li tao in this negative sense, see wyyh 884/9a, Chang Chiu-ling, spirit path stele for P’ei Kuang-t’ing; T’ang Lu Hsüan kung han-yüan chi 6/8a and ttclc 106/542, Lu Hsüan kung, decree examination question of 785; wyyh 689/1b, Liu Mien, letter to Ch’üan Te-yü, attributing li tao to the Sui; also Yüan Tz’u-shan chi 9/135. For li tao in the careers of officials, see e.g. cts 77/2673, biog. of Yang Tsuan, d. c. 652; cts 113/3353, biog. of P’ei Mien, d. late 769 or early 770; cts 129/3599, biog. of Han Huang; cts 146/3965, biog. of Li Jo-ch’u, d. 799. Confucian (ju) values were also opposed to administrative ones (li) in e.g. cts 102/3164, biog. of Ma Huai-su; cts 157/4156, biog. of Wang Yen-wei. Cf. also Yu yu chi 5/3a, Lu Chao-lin, characterizing Kao tsung’s reign as esteeming li against ju and Mo. The term was not, however, necessarily negative, when the official concerned also had learning and integrity; see e.g. P’i-ling chi 6/9a canonization discussion for Lü Yin; cts 96/3025, biog. of Yao Ch’ung. For the phrase shih li i ju, see P’i-ling chi 20/3a–b, sacrificial graveside prayer for Wei Yüan-fu and cf. P’i-ling chi 6/9a, canonization discussion for Lü Yin.

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

102

McMullen

canonical texts, however, was never rigidly enforced, despite what appear to have been official initiatives in this direction. Successive emperors and ministers had a pluralistic attitude towards interpretation that was consistent with T’ang religious attitudes and with T’ang modes of thinking in other fields. It was also the government’s policy throughout the dynasty to make detailed knowledge of the canons and their commentaries a requisite of the examination system, through which the scholarly elite within the official hierarchy was recruited. The use of the canons in this way was never seriously questioned; but it proved difficult to administer meaningfully. The mechanical and unthinking approach to the texts and their commentaries that inevitably resulted among candidates provoked indignation among the minority of scholars for whom the canons provided authority for deeply held political and moral views. For the great majority of serving officials, knowledge of the canons meant learning the texts sufficiently to acquire official status and to run successful careers. Their experience of intensive study did not extend much beyond the symbolic ‘three winters’ (san tung) in the much used phrase from the Han shu, deemed necessary to master them.4 Their aims, like those that an official scholar enjoined on Hsüan tsung, the dynasty’s greatest patron of scholarship, in 711 or 712, were practical, and their approach was sometimes even explicitly limited to what they called the ‘general meaning’ (ta lüeh) of the canonical texts, rather than their very detailed exegesis.5 A minority within the scholarly elite, however, hoped that their special knowledge and skill in interpreting the canons would be of more specific benefit to the dynasty and use in their careers. They hoped to participate in the scholarly commissions set up by the dynasty to further its control over the canonical tradition or in debates about the value of commentaries and sub-commentaries. Such participation increased their own standing, and when successfully completed, brought promotion and substantial monetary reward. If they were not selected for commissions, individual scholars with expertise in specific canons sometimes submitted their own commentaries, and this in turn could bring them acknowledgement and reward. Whether undertaken on official commission or privately, canonical scholarship was understood to be a continuing process. Throughout the dynasty, there was a tacit assumption that a more accurate text and a better understanding were possible.

4  Han shu 65/2841, biog. of Tung-fang Shuo. For T’ang use of this phrase, see e.g. ctcc epitaph of Ts’ui Chih-tao, d. 682, aged 72 sui, an examination graduate; ctcc epitaph for Ts’ui Kuangssu, d. 732 aged 71, a decree examination graduate in the three teachings. 5  c ts 190B/5016–7, biog. of Liu Hsien; see also below, note 97.

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

Canonical Scholarship

103

The pattern of the state’s interest in the canons, combined with accidents of transmission, has resulted in an uneven picture of the tradition over the dynasty. Voluminous official sub-commentaries produced under T’ai tsung survive intact; but alongside them have to be set the laconic accounts of later attempts to re-establish the standard texts of the canons, or protests about their use in the examinations. Many unofficial works of commentary have been lost, so that an overview of the way in which attitudes to the canons evolved, though highly detailed at certain points, is indistinct at others. Nonetheless, over nearly three centuries, great changes are evident in the attitudes of scholars towards the canons. The early reigns saw the state’s interest in controlling the tradition at its keenest and most successful. Official operations predominated. By the mid eighth century, however, the state’s statutory provision of definitive texts, commentaries and sub-commentaries provoked a succession of scholarly controversies within the official community. As the examination system developed, the role of the canons in the examinations caused academic officials increasing dissatisfaction. In the post-rebellion period, despite its lack of resources, the state continued to discharge its traditional responsibility of preserving and displaying definitive texts. But the scholarly world of the late eighth and early ninth centuries was much less effectively centralized on the court and no longer dominated by imperial commissions. State academic institutions no longer provided the main institutional framework in which the canons were reinterpreted, and the most important intellectual developments took place unofficially. Some scholars now reinterpreted the canons primarily to justify their own ideas on political reform and on questions of religious belief. Their writings were no longer necessarily intended for submission to high authority and use in the official education system. They represented the response of individual scholars to a more devolved and less court-centred political order. This post-rebellion development within the scholar community towards an independent critical tradition in canonical learning may be seen as moving in two directions. These corresponded to the opposing extremes in the wellknown Confucian polarity between governing the state on the one hand and the cultivation of the self on the other.6 There was, first, a new interest in political theory, institutional history and technical administrative problems. This interest is to be linked with the utilitarian re-appraisal of the origins, institutions and functions of the dynastic state that took place among some scholars in the late eighth century. Secondly, at the opposite extreme, there took place the beginnings of the ‘deep interiorization’ of the tradition which was to 6  Schwartz (1959), pp. 52–3; Li chi chu shu 60/1a.

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

104

McMullen

culminate in the idealistic and introspective Confucian philosophy of the early T’ang phase of the Neo-Confucian movement. The divergence of these two intellectual emphases did not threaten the social cohesion of the scholar community. Both represented aspects of the change in outlook away from immediate focus on the emperor and the court towards a more generalized concept of the state and the individual. Some scholars at the turn of the eighth century recognized the value of both approaches and even tried to promote both. Both claimed sanction from the full range of Confucian canonical texts. At the risk, however, of adopting an over-schematic approach, a group of texts or single text may be seen as especially important for each. Interest in institutions and in political theory often correlated with specialization in the Ch’un-ch’iu and its Three traditions, the Tso chuan, the Kung-yang chuan and the Ku-liang chuan. For the movement towards interiorization, it was the group later to be known as the Ssu shu (Four books), the Analects, the Ta hsüeh (Great learning), the Chung yung (Doctrine of the mean), and above all the Mencius, that registered particularly well the changing attitudes of T’ang scholars. The emergence of these two emphases indicate how radically, by the second half of the eighth century, the scholarly climate had changed. Their evolution forms a counterpoint to the main subject of this chapter, the interest of the state in canonical scholarship.

The Reigns of Kao tsu and T’ai tsung

From the beginning of the T’ang, the Confucian canons were often quoted at court discussions and in a wide range of official contexts.7 The dynasty also made knowledge of the canons central to the curricula of the schools and to the examinations. But they also inherited from the re-unifying Sui dynasty the more specific concern to impose order and unity on the great range of divergent commentarial literature to which by the early seventh century the scholarly community was heir. They supervised in this undertaking a relatively small but productive community of official scholars. The majority of these held office in the Ch’ang-an state academy directorate, the institution responsible for empire-wide education. These scholars produced a survey of the tradition of canonical learning, that, fully extant, is not only a major source for early T’ang

7  A rough indication of how often, in the surviving record of court discussions, the canons were quoted is provided by Harada Tanishige’s index to ckcy; see p. 69 for the Mao shih and p. 80 for the Ch’un-ch’iu. See also below, notes 48–51.

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

Canonical Scholarship

105

scholarly attitudes, but also the most comprehensive account of the whole tradition from the medieval period. The Textual and Exegetical Traditions In surveying existing canonical scholarship, early T’ang scholars were concerned with three distinct stages in the development of the tradition. The first was that of the texts themselves of the Confucian canons. Secondly, there were primary commentaries, usually written between the Han and the Chin dynasties, of which, in most cases, two or more existed for each canon.8 Finally, there was a considerable body of exegetical or sub-commentarial literature that scholars had produced nearer T’ang times. The programme for canonical scholarship on which high authority embarked in the Chen-kuan period involved all these three stages. It demonstrates particularly well some of the features of early T’ang scholarship that recur in all the learned disciplines considered in this book. The emperor at first made use of the scholarship of individual scholars who had served under the Sui. Then, in the second half of the reign of T’ai tsung, a large ad hoc commission was appointed to produce the dynasty’s own definitive exegesis. The work of this commission, like that in other disciplines in this period, again exploited material inherited from the Sui dynasty. It also expressed certain early T’ang political priorities, among which one that recurred in all fields of learning was the unification of the northern and southern traditions after the period of disunion. For all the ambition involved, the commission ran into short-term difficulties and failed to produce long-term solutions to the problems it set out to resolve. In the early years of the dynasty, the T’ang almost incidentally managed to extend approval to the compilations of a major lexical and phonological scholar of the Sui. This was Lu Te-ming, a southerner by origin, who had represented the Confucian side in the three-teachings debate of 624 (Ch. 2, p. 33). Lu Te-ming had produced a series of phonological glossaries for a group of seven canons that included the Taoist texts Lao tzu and Chuang tzu, and in its Neo-Taoist emphasis was typical of the pre-T’ang southern dynasties. In this series, the Ching tien shih wen (Explanations for canonical texts), he drew in total from 230 works, in this breadth being characteristic of early seventh-century exegetes. He also supplied short surveys of both the primary commentaries and the sub-commentaries on each of the texts he glossed. Although he may have completed his compilations under Sui rule, he was honoured and given office by the T’ang. His Neo-Taoist outlook contrasts with the less speculative, 8  For the period of primary commentaries, see Honda (1935), pp. 156–206; P’ei P’u-hsien (1969), pp. 237–40. For sub-commentaries, see Mou Jun-sun (1960), pp. 356–9.

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

106

McMullen

more cautious attitudes of official canonical scholars of the second half of the Chen-kuan period.9 T’ai tsung’s approval of his compilations provides an early indication of the generally pluralistic outlook of the early T’ang state towards different interpretations of the canons. Lu Te-ming’s glossaries enjoyed official favour until well into the eighth century, and circulated empire-wide.10 In 631, however, T’ang scholars took the first step in a programme intended to produce for the dynasty its own definitive text and exegesis of the canons. In that year, T’ai tsung commissioned Yen Shih-ku (581–645, canonized Tai), a member of a scholarly clan that claimed descent from Yen Hui, and one of the most prominent official scholars of the first two reigns, to ‘determine’ (ting) the text of the Five canons.11 The choice of canons was significant, because it provides a first hint of the extent to which early T’ang official scholars followed their Sui predecessors in matters of scholarly policy. The group of five did not represent the complete range of Confucian texts, but rather those that had been particularly important under the Sui. They were: the Chou i; Shang shu; Mao shih; Li chi; and Ch’un-ch’iu with the Tso chuan. Yen Shih-ku therefore set aside the four texts that made up the long series of Nine canons, the Chou li and I li, and the Kung-yang chuan and Ku-liang chuan. The study of these last

9  The juxtaposition of Confucian ( ju) and classical Taoist (hsüan) was common in the period of disunion, as the numbers of scholars commended for mastery of both traditions suggests; e.g. Nan Ch’i shu 54/929, biog. of Ku Huan; Ch’en shu 33/447, biog. of Shen Pu-hai; Liang shu 3/96, of the emperor Liang Wu ti himself. kfspwk epitaph for the Sui official Chang Po, tzu Fang-chin, shows that this juxtaposition still occurred late in the period of disunion or in the Sui. tt 27/161.1 indicates that in the Liu Sung dynasty (420–79) ju and hsüan with literature (wen) and history (shih), constituted the four categories of learning, and were therefore the antecedents of the four categories (ssu pu) of the T’ang. 10  Lu Te-ming, biog. cts 189A/4944–45. For the Ching tien shih wen, see preface by Lu Teming; cts 46/1983; hts 57/1446 and Ho chih p. 51. Thompson (1979), pp. 56–61 and notes, reviews in detail the evidence concerning the chronology of Lu Te-ming’s life. Evidence for the wide circulation of the Ching tien shih wen derives from the survival of fragments at Tun-huang and at Nara, for which see Thompson, p. 58, note 31. That this work remained highly respected by official scholars in the T’ang may be seen by the remarks of Chang Shen, who was required to re-determine the text of the Five canons in 776. Chang Shen called the work ‘uniquely thorough’, and used it in the version of the canons he erected in the state academy directorate. See below, note 147. Lu Te-ming’s account of the transmission of the Tso chuan also figured in the writings of the Ch’un-ch’iu school in the post-rebellion period; see below note 165, and Ch’un-ch’iu tsuan li 1/12b, Chao K’uang. For the three-teachings debate of 624, see Lo Hsiang-lin (1954), pp. 85–7. 11  For Yen Shih-ku’s commission, see ckcy 7/215–16, dating 630; cts 73/2594, biog. of Yen Shih-ku; cts 189A/4941; cts 3/43, dating promulgation to 633.

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

Canonical Scholarship

107

two was said to have been so neglected in the Sui period that ‘they were almost without teachers to explain them’.12 The determining of a definitive text for the Confucian canons was not an undertaking that was original to the T’ang. In the period of disunion and the Sui, emperors had commissioned ‘definitive versions’ (ting pen), and these had therefore existed well before the T’ang.13 Yen Shih-ku, pursuing the project in the imperial library, of which he became deputy director from 633, used versions of the canons dating from no earlier than the Chin (ad 265–420) and Liu Sung (420–79). In discussions arranged by T’ai tsung, he successfully defended his decisions. In 633, these ‘definitive versions’ were distributed throughout the empire. A list of standard-form characters that Yen Shih-ku compiled with these standard texts, the Yen shih tzu yang (Mr Yen’s model characters), circulated widely.14 The fact that in several accounts Yen’s project immediately precedes the second stage in the early T’ang programme for canonical scholarship has been taken to imply that his ‘definitive versions’ were used as its basis.15 This was not, however, the case. This second stage began in 638, when T’ai tsung appointed a commission to compile sub-commentaries for the Five canons.16 Its brief, differing considerably from that of Yen Shih-ku, was not only to provide exegesis of the texts of the canons, but also to explain the subsequent strata: the primary commentaries and the most recent, sub-commentarial layer. The commission was directed by the leading official scholar K’ung Ying-ta, the ‘Confucius of Kuan-hsi’, who was appointed president of the state academy directorate that year. It was made up of thirteen or more scholars, most of whom, 12   s s 32/933. 13  For pre-T’ang ‘definitive versions’ (ting pen), see Tso chuan chiu shu k’ao cheng, by Liu Wen-ch’i (1789–1856), preface 3a–4b. Liu Wen-ch’i shows that K’ung Ying-ta’s subcommentary cited ‘definitive versions’ in its exegesis for the Chou i appendices, the Shang shu, and the Li chi, while in the Mao shih and Tso chuan sub-commentaries ‘definitive version’ citations were particularly numerous. In some cases, e.g. Tso chuan chu shu 49/3b, the sub-commentary rejected ‘definitive version’ readings. For references in pre-T’ang histories to determining the text of the canons, see ss 66/1554, biog. of Lang Mao, under the Northern Ch’i; ss 75/1715, biog. of Hsiao Kai. 14  For the Yen shih tzu yang, see Yen Lu kung wen chi (sptk ed.), Pu i 2a–3a. Yen Chen-ch’ing also mentions, critically, a later work in the same tradition, by Tu Yen-yeh. 15  P’ei P’u-hsien (1969), p. 242. 16   t hy 77/1404 implies that the work was submitted in 638; but this is the earliest date at which the project is mentioned, and is therefore likely to be the date of its commission. For other accounts of the project, see ckcy 7/216; cts 189A/4971; cts 73/2602–3, biog. of K’ung Ying-ta; tfyk 606/14b; cts 4/71.

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

108

McMullen

appropriately, held basic office in one or other of the directorate’s constituent schools.17 It differed from the major historical and literary commissions of the period, therefore, in not being directed by a scholar minister of the seniority of Wei Cheng or Fang Hsüan-ling, and this is possibly an indication of its lower standing. This commission completed a version of the sub-commentary series four years later, in 642, when T’ai tsung gave it his approval.18 When this version was discussed in the state academy directorate, however, for reasons that are not recorded, it proved unacceptable. It is possible that at this stage T’ai tsung himself encouraged debate between members of the commission and other canonical scholars.19 A new commission, of eleven, with basic posts similar to those of its predecessor, was appointed. In 643, K’ung Ying-ta left the commission, apparently because of his age, and over the final years of T’ai tsung’s reign momentum was lost.20 Then, a year after T’ai tsung’s death, a new commission, to include among others K’ung Ying-ta’s son, K’ung Chih-yüeh, was established. In 651 or 652, the project came under the direction of the senior scholar minister Chang-sun Wu-chi. In 653, Chang-sun Wu-chi, with a number of other high ranking ministers, submitted the series, which had taken fifteen years to complete, to Kao tsung. An edict the following year ordered the distribution of copies throughout the empire.21 The series became known as the Wu ching cheng i (The true meaning of the Five canons). 17  See the lists of names given in the prefaces to each of the sub-commentaries, before the mention of imperially ordered revision. Su Ying-hui (1968), pp. 182–4 notes discrepancies between the Tun-huang version and transmitted versions over the names of co-compilers. 18   Chin shih ts’ui pien 47/9a and cts 73/2603, biog. of K’ung Ying-ta mention that K’ung’s reward for completing the commission was 300 bolts of silk, about standard for a senior academic official for a large project. 19  The objections came from Ma Chia-yün; see cts 73/2602–3, biog. of K’ung Ying-ta and cts 73/2603, biog. of Ma Chia-yün. It is conspicuous that 642 was the year in which T’ai tsung ordered a discussion in connection with another, more radical submission concerning canonical exegesis; see cts 74/2620, biog. of Ts’ui Jen-shih and tfyk 606/14a–b. The new director of the project, Chao Hung-chih, 572–653, canonized Hsüan, had been a member of the I wen lei chü commission (see below, Ch. 6 at note 9) and had lectured on the Hsiao ching; see cts 188/4921–22, biog., which does not, however, mention his service on the sub-commentary commission. 20   c ts 73/2602, biog. of K’ung Ying-ta. Yü Chih-ning’s stele text for K’ung’s grave at Chao-ling does not mention the project following K’ung’s submission in 642; see Chin shih ts’ui pien 47/9a–b and Lo chen-yü (1909) 1/17b. 21   t kck 2/2b–4b, citing a Northern Sung edition of the sub-commentary. A fragment preserved at Tun-huang gives a list of the names of scholars responsible for the final revision, with names of collators, in one case of a student of the Ssu-men hsüeh; see Su Ying-hui

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

Canonical Scholarship

109

The sub-commentary series, at a total length of 180 chüan,22 was, like other seventh-century compilations, a cumulative work. Both in its basic features and in its detail, it was also the result of the need to compromise and to honour political priorities. Its editors sometimes sacrificed intellectual consistency to their loyalty to their chosen authorities, and had a compendious approach to the controversies that marked the pre-T’ang history of canonical scholarship. Like all early T’ang official scholars, the compilers owed a considerable debt to their predecessors under the Sui. Two major features of the sub-commentary series, in particular, derived from Sui scholarship. The first was that, like Yen Shih-ku, the commission restricted themselves to the inner series of Five canons. Secondly, in selecting a single primary commentary for each of the Five canons, as seventh century practice demanded that they should, K’ung Ying-ta’s commission adopted precisely those that had been in favour in the Sui.23 For the Chou i, the commission selected the commentary by the NeoTaoist philosopher Wang Pi (ad 226–49), and, for the Hsi-tz’u (Appendices) that of Han K’ang-po (fl. c. 385); for the Shang shu, the commentary attributed to K’ung An-kuo (fl. c. 156–74 bc); for the Mao shih, the commentaries first by Mao Heng of the Former Han, and then by the great eclectic Later Han commentator Cheng Hsüan (ad 127–200), whose approach to the canons was marked by the ‘sense of factual enquiry’ of the Old Text (Ku-wen) school. For the Li chi, the (1968), pp. 192–3. Su, pp. 186–7, refutes the suggestion of Ts’en Chung-mien that this belongs to a copy of the Shang shu cheng i. There is some discrepancy among the sources that mention the dates of the revision that took place after T’ai tsung’s death. K’ung Yingta’s hts biog., 198/5645, states that the revision was ordered in 651, but makes no mention of Chang-sun Wu-chi. thy 77/1405 states that a revision was ordered in 651, but under Chang-sun Wu-chi. A date of 652, rather than 651, for Chang-sun Wu-chi’s direction is put forward by Su, p. 185, on the basis of a citation of the Hui yao in Yü hai 42/32a–b. This leads Su to suggest that a revision in 651 under Yü Chih-ning was followed by the appointment of Chang-sun Wu-chi as director in 652. All sources except Chung-hsing kuan ko shu mu, quoted in Yü hai 37/10a, Su, p. 184, agree that the final completion was in 653. Cf. Thompson (1979), pp. 69–71 and notes. 22  Lo Chen-yü (1909), 1/18b, stele by Yü Chih-ning, referring to the first submission of the work, reads 170 chüan, cts 73/2602, biog. of K’ung has 180 chüan, cts 46/1968, 1970, 1971, 1974, 1978, and Ho chih, pp. 13, 18, 21, 16 and 38 have a total of 180 chüan. By the Sung, the series is entered with a larger number: see hts 57/1426, 1428, 1430, 1433, 1440, and Ho chih, pp. 13, 18, 21, 26, and 38; also Yang Hsiang-k’uei (1958), p. 8, note 1. 23  For indication that the same primary commentators were in favour under the Sui, see the short accounts of the history of the exegesis of each canon: ss 32/912–13 for the Chou i; ss 32/914–15 for the Shang shu; ss 32/918 for the Mao shih; ss 32/924–26 for the Li chi; ss 32/932–33 for the Ch’un-ch’iu. The fact that K’ung Ying-ta participated in the compilation of both ss and the sub-commentary helps explain this consistency.

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

110

McMullen

chosen commentary was also by Cheng Hsüan. Finally, for the Ch’un-ch’iu they selected the Tso chuan rather than either of its rivals, the Kung-yang or the Ku-liang, and for the Tso, the commentary by the scholar and soldier of the Three Kingdoms and early Chin Tu Yü (ad 222–84).24 This range of primary commentaries, endorsing Sui decisions as it did, was said to have provided a balance between what was characterized as the northern commentarial tradition, more practical, less speculative and typified by Cheng Hsüan’s acceptance in the north, and the southern, more philosophical and less concerned to gloss historical details, represented by the Neo-Taoist philosopher Wang Pi.25 K’ung Ying-ta’s commission probably also owed much of the fabric of their sub-commentaries to their Sui predecessors. The exegesis of two Sui scholars, Liu Cho (544–610), who had actually taught K’ung Ying-ta, and Liu Hsüan (d. c. 606 aged 68 sui) was particularly important. Both were northerners and official scholars; both were recorded as unifying the northern and southern traditions, and both produced sub-commentaries for the canons in series.26 It is possible that at least in the cases of the Shang shu, Mao shih and Tso chuan, K’ung’s commission took much of their material from these scholars. K’ung’s sub-commentary to the Shang shu sometimes used the term Great Sui (Ta Sui), implying that the Sui was the current dynasty.27 In the sub-commentary to the Tso, the text attributes certain comments to Liu Hsüan; exactly the same comments appear, without attribution, in the Mao shih sub-commentary.28 Though 24  For Wang Pi, see A. F. Wright (1947), pp. 124–61. An invaluable concise history of the Shang shu and K’ung An-kuo’s commentary is given by Hung (1957), note 5 on pp. 99–100 and p. 79 and note 103 on pp. 124–5. For Tu Yü’s commentary, see Yeh Cheng-hsin (1966). 25  For the relationship between northern and southern traditions, see Mou (1960), p. 393; also Liang shu 48/678, biog. of Lu Kuang; Pei shih 81/2709; ss 75/1705–06. 26  Liu Cho’s biogs. are at ss 75/1718–19 and Pei shih 82/2762–63. Liu Hsüan’s at ss 75/1719–23 and Pei shih 82/2763–67. See also ss 75/1726–27 and ss 75/1707. For K’ung Ying-ta’s debt to both, see tfyk 768/13b. For the debt of other early T’ang official scholars to Liu Cho, see tfyk 768/13b and cts 198A/4951, biog. of Kai Wen-ta, and tfyk 768/14b and cts 198A/4949, biog. of Chang Shih-heng. 27  For instances of Ta Sui occurring in the text of the sub-commentaries, see Shang shu chu shu 3/9b and 19/17a. The latter reference supplies a date of ‘early K’ai-huang’ for the abolition of castration as a punishment. 28  See Chien Po-hsien (1970), quoting the Ch’ing scholar Liu Wen-ch’i. In Tso chuan chu shu 39/8b and Mao shih chu shu 1/10a, the two sub-commentaries have almost identical texts, except that the Tso chuan passage includes four additional words, ‘Liu Hsüan also says’, before the phrase, ‘When shield and spear have been put down’, which both passages contain. The passages then continue to run almost exactly parallel for several columns. The omission of the phrase ‘Liu Hsüan also says’ from the Mao shih sub-commentary justifies

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

Canonical Scholarship

111

in neither case does this prove that more than the short passages concerned were of Sui origin, it is suggestive of a larger debt. Whether as a direct result of the compilers’ indebtedness to Sui predecessors or on their own initiative, the commission left in the sub-commentary series a work that clearly registers the intention to produce a single, standard interpretation of their texts. Each sub-commentary provides elaborate and comprehensive refutation of alternative viewpoints that its exclusive endorsement of a single primary commentary required it to reject.29 In doing so, it analyzes the views of a wide range of authorities. The scale of its survey may be inferred by the number of sources that the series drew on. For example, Cheng Hsüan’s primary commentary to the Li chi had cited 45 different sources by name. In K’ung’s sub-commentary, this number was increased by nearly 200, the majority of the titles being of works from the period of disunion.30 Wang Pi’s philosophical commentary to the Chou i had not cited any authority by name, but K’ung’s sub-commentary drew explicitly from 26 works of Chou i exegesis, besides other canons, apocrypha, lexical works, histories and philosophical texts.31 Even these figures do not give an adequate impression of the compendious character of the series, or its capacity to record and sometimes to emphasize the controversies of the preceding period.32 This impressive breadth reflects another purpose underlying the series, that of refuting views that might carry political risks for the dynasty. Perhaps the theme in the exegetical tradition over which early T’ang scholars, and probably T’ai tsung himself, felt a special sense of threat concerned prognostication. They were not radical sceptics over divination as it was practised in the early seventh century, and they certainly did not deny that the future might be foretold. But they wanted to ensure that the activity should be controlled, the speculation that much of the sub-commentary series may have originally contained attributions of this kind, which were removed in the course of editing. 29  K’ung’s method was to follow the primary commentary wherever possible. This led him, however, to refute Liu Hsüan at certain points. Liu Hsüan had written two treatises identifying and refuting errors in the primary commentary to the Tso chuan by Tu Yü, which K’ung Ying-ta endorsed. For these works, the Ch’un-ch’iu kuei kuo (Reproving the errors in [Tu Yü’s] Ch’un-ch’iu commentary) and Ch’un-ch’iu kung mei (Attacking the blind spots in [Tu Yü’s] Ch’un-ch’iu commentary) see the recensions by Ma Kuo-han (1794–1857), in Yühan shan fang chi i shu. K’ung therefore attacked Liu for faulting the primary commentary he otherwise endorsed, for ‘giving exposition to Tu and yet attacking Tu’. For instances of K’ung refuting Liu Hsüan’s attack on Tu Yü, see e.g. Tso chuan chu shu 49/4b; 49/10b. 30  Yeh Ch’eng-i (1970), pp. 313–322, at p. 314; Ho Hsi-ch’un (1966), pp. 1–2. 31  Wang Chung-lin (1959), pp. 1–111. 32  For example, in the preface to the Chou i sub-commentary, 2a–7a.

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

112

McMullen

and should not, as it had done in the pre-T’ang period, become a threat to dynastic power. In the late Former Han period, the later phase of the New Text (Chin-wen) movement had developed a tradition of divination and prognostication to a level that was politically unacceptable to the early T’ang. K’ung’s sub-commentary series therefore often attacked the texts that represented this tradition, dismissing them as shallow and unreliable.33 K’ung probably again reflected the political priorities of the Chen-kuan court when he countered some of the more mystical readings of the Confucian canons that had been current in the period of disunion. He indicated in the preface to the Chou i sub-commentary that he had set aside Buddhist explanations of the text.34 Inevitably, though, he was influenced to some extent by Buddhism.35 The Confucian-Neo-Taoist synthesis that had developed in the south, as Lu Te-ming had shown, remained an intellectual force in the early seventh century, and its influence is to be seen in the sub-commentary series. K’ung Ying-ta’s exposition of certain Chou i passages indicates that he held that the tao was ‘without form’ (wu hsing).36 He defined both the tao and the original ‘oneness’ of which the Chou i, the Lao tzu and the Li chi spoke as ‘void’ (k’ung).37 On the other hand, he also posited a physical process whereby matter came into being, through the agency of ‘particles’ (chi), which ‘left the void and entered the realm of being’.38 This belief in a physical transition from 33  For the background to New Text (Chin wen) prognosticatory books (ch’an shu), see Fung Yu-lan tr. Derke Bodde (1953), pp. 88–132; also Tjan Tjoe Som (1949), pp. 100–20. For another condemnation of the apocryphal and prognosticatory texts of the late Former Han period by K’ung Ying-ta and others, see ss 32/940–41. Bodde (1953), p. 89, translates part of this important passage. See also ss 32/948. ss 78/1743–64 constitutes the early T’ang official world’s cautious endorsement of the idea of prognostication. For criticism of apocryphal and divinatory texts in the sub-commentary series, see Shang shu chu shu preface 2a; 8/14b; Chou i chu shu preface 7a; Yang Hsiang-k’uei (1958), p. 11. Early T’ang attitudes to divination were also embodied in the revision of the Yin-yang shu (Book on yin and yang) undertaken on imperial commission by the polymath Lü Ts’ai, for which see thy 36/651–56; tt 105/558.2–559.3; cts 79/2719–27, biog.; cts 47/2044; hts 59/1557 and Ho chih p. 252. Also, Hou Wai-lu and Chao Chi-pin (1959), pp. 1–21; Needham (1956), p. 387. For later T’ang references to Lü Ts’ai’s Yin-yang shu, see Feng shih wen chien chi 5/21; this and the inclusion of excerpts in tt and Su Mien’s comment on the passage of work he excerpted for the Hui yao (thy 36/656) suggest that its pre-suppositions were supported by the scholar community in the eighth and early ninth centuries. 34   Chou i chu shu pref. 1a. 35  Honda (1935), pp. 232–3; Mou Jun-sun (1960). 36   Chou i chu shu 7/18b–19a; Yang Hsiang-k’uei p. 9. 37   Chou i chu shu 7/7a; 3/11b; Yang Hsiang-k’uei pp. 8–9. 38   Chou i chu shu 1/8b; cf. also Chou i chu shu 7/15b for further exposition of chi.

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

113

Canonical Scholarship

being to non-being has been seen as more indebted to Han cosmology than to the Neo-Taoism of the period of disunion. K’ung’s detailed formulations of the shape of the universe may derive from a similar intention to replace metaphysical speculation with a physically precise cosmology.39 Perhaps, however, the feature of the sub-commentary series that best reflects the early T’ang scholarly milieu was its exhaustive treatment of issues relating to ritual. The dynasty’s ritual programme (Chapter 4) was of enormous importance to the seventh- and eighth-century scholar community. K’ung Y’ing-ta’s commission repeatedly testifies to this by the thoroughness of their examination of the canonical origins of the ritual observances that made up the state ritual programme. It was an irony that, despite the exhaustive treatment K’ung accorded ritual issues, his position on many of these issues was to be set aside by the authorities in the decades that followed. The Policy of Inclusiveness K’ung’s Wu ching cheng i series was an attempt, in part politically motivated, to survey and delimit the vast canonical and exegetical traditions. But the series did not treat the whole range of Confucian canonical literature. Nor did its promulgation prevent further exegesis of the Five canons that it had glossed. In the Chen-kuan period, there were already indications that high authority adopted a more pluralistic attitude to the interpretation of the canonical corpus than that implicit in K’ung’s series. T’ai tsung himself encouraged controversy among Confucian exegetical scholars. His early approval of Lu Te-ming’s Ching tien shih wen series, with its southern commitment to classical Taoism, suggests the breadth of his interest. The list of sub-commentators of Liang, Northern Chou and Sui dynasties whom he honoured by edict in 640 (Ch. 2, p. 34) was more inclusive than might have been expected if a narrow concept of orthodoxy, based on K’ung’s sub-commentaries, had dictated their selection. The same is true of the list of primary commentators represented in the temple by edict in 647 (Ch. 2, p. 35), for this included figures whose commentaries K’ung had expressly criticized. Another incident of T’ai tsung’s reign points even more clearly to a willingness to tolerate mutually exclusive interpretations of the canons. In 642, a collator in the imperial library submitted commentaries on the Shang shu and Mao shih which refuted the primary commentaries of K’ung An-kuo and Cheng Hsüan. In the context of seventh-century official scholarship, this was a particularly radical approach, first because it entailed rejecting the primary commentaries that two of K’ung’s sub-commentaries had amplified. Secondly, 39   Shang shu chu shu 3/3b and ss 19/505 ff. See also Cullen (1976), pp. 107–9.

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

114

McMullen

it meant an exegetical approach that dismissed altogether the second, primary commentarial layer of the tradition and directly interpreted the text of the canons. Nonetheless, T’ai tsung ordered a discussion, and it was proposed that these new commentaries be taken into the imperial library, and that they be allowed to ‘circulate together with’ (ping hsing) the K’ung An-kuo and Cheng Hsüan commentaries. Only a detailed refutation, by an official from outside the academic establishment, holding financial office, ultimately brought about their rejection.40 The surviving evidence for the place of the canons in the education and examination system of the first two reigns also makes it plain that even at this early stage the system did not simply involve the Five canons ‘determined’ by Yen Shih-ku and glossed by K’ung’s commission. Under T’ai tsung, examiners also used the Chou li and the I li,41 and in 644 the crown prince demanded knowledge of the Hsiao ching as well as the Li chi.42 More indirectly, a Japanese code that is held to have been based on the T’ang statutes of 651 listed in its official school curricula two commentaries for each of the canons. It also gave the long series of nine canons, not just the short series of five that Yen Shih-ku and K’ung Ying-ta had used.43 It may have been, therefore, that by 651 the state had established two commentaries for each of the main series of canons, not one only, as K’ung Ying-ta’s position implied it should have done. In the course of the reigns of T’ai tsung and Kao tsung, moreover, official scholars compiled, probably on their own initiative, sub-commentaries to all the remaining canons in the long series of nine, with the possible exception only of the Kung-yang chuan. The most important scholar involved here 40   c ts 74/2620, biog. of Ts’ui Jen-shih. T’ai tsung’s encouragement of debate among canonical scholars is also suggested by cts 73/2602–03, biog. of K’ung Ying-ta; cts 73/2603, biog. of Ma Chia-yün. 41  For the place of the Chou li and the I li, see the decree of 635, thy 75/1375; tfyk 639/19a; tkck 1/15a. 42  For the place of the Hsiao ching in early T’ang promotion of education, see tkck 1/5b–6a, edict of seventh month of 624; cts 73/2602, biog. of K’ung Ying-ta, indicating that he lectured on the canon at the state academy directorate in 640. K’ung also compiled a sub-commentary to the Hsiao ching for the crown prince, often the focus for educational ideals; see cts 73/2602; hts 57/1443; Ho chih p. 45. See also thy 35/640, showing that Chao Hung-chih lectured on this canon again in 648, after a Shih-tien observance at the directorate; Chao’s biog., cts 188/4922, indicates that he lectured again on the canon, to Kao tsung, with scholars, academic officials and the students of the Hung-wen kuan attending. For later examples, see cts 22/864, Hsing Wen-wei in 690 in the Ming-t’ang; thy 35/642, Ch’u Wu-liang in 719. For such lectures, see tlt 21/6a. 43  Kao Ming-shih (1977), p. 11B.

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

Canonical Scholarship

115

was Chia Kung-yen, who compiled sub-commentaries for the Chou li and the I li that are extant. The fact that Chia also wrote sub-commentaries for the Analects and the Hsiao ching, and is recorded as having compiled a Li chi subcommentary that must, in effect, have duplicated K’ung’s own sub-commentary underlines the freedom of scholars to produce their own exegesis.44 Chia’s Chou li and I li sub-commentaries used the primary commentaries of the great Cheng Hsüan, and also relied on pre-T’ang exegesis. Though they were probably less detailed on philosophical questions, they were otherwise similar in perspective to K’ung’s own Li chi and Mao shih sub-commentaries and may have been seen as complementing them.45 One at least of the other sub-commentaries official scholars compiled in this period, however, is likely to have related only uneasily to K’ung’s series. This was the sub-commentary to the Ku-liang chuan compiled by Yang Shih-hsün, another academic official and a member of the commission that had worked under K’ung on the Tso chuan sub-commentary. Yang Shih-hsün used the primary commentary of Fan Ning (ad 339–401), a scholar in the same Ku-wen tradition as Tu Yü, the primary commentator to the Tso chuan whom K’ung had endorsed.46 But in his sub-commentary to the Tso, K’ung had included emphatic attacks on the Ku-liang tradition. If the dynasty had 44  Chia Kung-yen had worked on the revision of the Wu ching cheng i directed by Changsun Wu-chi; see tkck 2/4a, memorial of submission. He has a very brief biog. notice at cts 1289A/4949–50, which enters both these sub-commentaries. For these works, see also cts 46/1972; hts 57/1433; cf. Ho chih, p. 22, p. 24; also skcstmty, p. 364–5 and p. 387–8. hts 57/1433 also enters a Li chi cheng i in 80 chüan by Chia Kung-yen; cf. Ho chih p. 26. In addition, the bibliographies list a Hsiao ching sub-commentary, cts 46/1981; hts 57/1442; Ho chih, p. 45; and an Analects sub-commentary, cts 46/1982; hts 57/1444; Ho chih, p. 48. 45  Chien Po-hsien (1975), p. 75 cites Wen hsien t’ung k’ao 181/1a–b quoting a remark by Tung Yu, fl. c. 1130, that Chia Kung-yen’s Chou li sub-commentary owed to the exegesis of Ch’en Shao, biog. Chin shu 91/2348, and Shen Chung, biogs. Chou shu 45/808–11 and Pei shih 82/2741–2. skcstmty, p. 387, quoting Chia’s own preface to the I li, indicates his debt to the sub-commentaries of Huang Ch’ing of the Ch’i and Li Meng-che of the Sui. Yang Hsiang-k’uei (1958), pp. 12–13, suggests that his ritual scholarship was much stronger than his cosmology, and notes his reliance on apocryphal works. Cf. the remark of Chu Hsi, quoted in skcstmty, p. 365, that the Chou li sub-commentary was the best of all the canonical sub-commentaries. 46  For Yang Shih-hsün, see Tso chuan chu shu, pref. 2a by K’ung Ying-ta, mentioning him as an erudit of the Ssu-men hsüeh. Cf. also the remark of the skcstmty editors, p. 519, on him and on the integrity of the present text of the Ku-liang sub-commentary. cts 46/1979 and hts 57/1440, Ho chih p. 40 enter this sub-commentary. Yang Hsiang-k’uei, pp. 13–4, shows that Fan Ning quoted Tu Yü to refute Kung-yang concepts. Yang Shih-hsün in turn refuted Ho Hsiu’s primary commentary to the Kung-yang, in order to endorse Fan Ning’s

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

116

McMullen

envisaged enforcing, or even promoting, an intellectual orthodoxy, based on K’ung’s series, it would hardly have accepted a sub-commentary that took the part of a canon that K’ung had tended to condemn. The place in seventh-century exegesis of the remaining canon in the long series of nine, the Kung-yang chuan, is more difficult to determine. The Kungyang was a particularly important vehicle for the New Text school of the Han and was the authority behind much of the Han tradition of prognostication that the early seventh-century scholars rejected. A sub-commentary to the Kung-yang chuan and the primary commentary of Ho Hsiu (ad 129–82) is extant; but its date is uncertain.47 Despite K’ung’s efforts at producing a detailed and coherent survey of the canonical tradition, therefore, T’ang emperors, the educational agencies and examiners recognized a wider range of primary and commentarial literature than was consistent with his outlook. Some of them may also have retained a measure of traditional disdain for the detailed commentarial approach (chang-chü hsüeh) that K’ung Ying-ta’s great project exemplified. In the context of political discussion and policy recommendation, so important to the T’ang official community, it was usual to cite the text of the canons direct, without reference to intervening exegesis. A relaxed sense of what was orthodox, and freedom to appeal directly to the canons, were to remain very important characteristics of the Confucian canonical tradition throughout the T’ang. In due course they led to major developments in attitude within the tradition. Ch’un-ch’iu Scholarship in the Early T’ang Of all the canons in the short or long series, the Ch’un-ch’iu, demonstrates particularly clearly the politically determined, state-centred outlook of the canonical scholars of the early T’ang. The early T’ang scholarly world believed that Confucius had edited a profound message into the text of the Ch’un-ch’iu, and they associated the canon particularly closely with his teaching. Scholar ministers under T’ai tsung cited the Ch’un-ch’iu in order to justify a range of

use of Tu Yü. The sub-commentary was therefore broadly in the same Old Text (Ku-wen) tradition as K’ung Ying-ta’s sub-commentary to the Tso chuan. 47  Neither cts nor hts bibliography enter this sub-commentary. See Ch’ung-wen tsung mu 1/24–5; Chih chai shu lu chieh t’i 3/4a. The skcstmty editors, pp. 517–8, cite Tung Yu in identifying Hsü Yen as of late T’ang date. Yang Hsiang-k’uei, p. 14, accepts a ninth-century origin. However P’an Chung-kuei (1955), p. 11, believes it to be of Northern Ch’i date and probably by Kao Yün, whose biogs. are at Pei shih 31/1117–32 and Wei shu 48/1067–96.

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

Canonical Scholarship

117

policies. It was used to sanction the ideals invested in official history writing.48 Officials quoted it to justify the attitude of radical separation that they consistently enjoined on successive emperors in managing the barbarians.49 It also played a part in the maintenance of the imperial ancestral temple,50 and in the interpretation of omens.51 In the scholarly community, preference for the Tso chuan, the least speculative of the Three traditions to the Ch’un-ch’iu, and by far the most copious in its concrete historical detail, was general. T’ai tsung himself had been taught the Ch’un-ch’iu and the Tso, and quoted from them in conversation.52 The Tso, moreover, suited the T’ang temperament, with its love of action. It can be no accident that throughout the dynasty, when military men were commended for Confucian learning, this was the text that they were said to have known.53 But the Ch’un-ch’iu and the Tso were also important for more philosophical reasons. The ‘moderate scepticism’ towards such activities as prognostication 48  For the role of the Ch’un-ch’iu and the Tso chuan in the seventh-century outlook on history compilation, see Tso chuan chu shu 10/1b quoted in ss 33/966, to sanction court diaries. Wei Cheng’s admonition to T’ai tsung in 640, for which see ckcy 7/220 and Ch. 5, note 47, is an adaptation of this passage. The Ch’un-ch’iu and the Tso chuan were the primary sources for both Tung Hu and Nan Shih, much cited archetypes for the upright historian; see below, Ch. 5, note 8. Cf. also thy 46/815, Ch’u Sui-liang, discussion of 652. 49   c kcy 9/285. 50   c kcy 8/263. 51   c kcy 10/297 and cts 72/2566–67, Yü Shih-nan, interpreting a mountainslide, the appearance of snakes and of floods, to T’ai tsung. thy 36/651–56, passim, Lü Ts’ai, Yin-yang shu, used evidence from the Ch’un-ch’iu to refute contemporary superstition relating to life expectancy and auspicious dates for burial. 52  For a clear statement that the Tso predominated, see ss 32/933; cf. Pei shih 81/2709. See also Hung (1957), p. 79, p. 124 note 101. For examples of early T’ang specialization in the Tso chuan, see cts 189A/4944, biog. of Hsü Wen-yüan; p. 4948, biog. of Chu Tzu-she; p. 4950, biog. of Li Hsüan-chih; p. 4950, biog. of Chang Hou-yin; p. 4952, biog. of Hsiao Te-yen. A more eclectic approach had, however, been current in the period of disunion, and was even encouraged by the Liang royal house; see Liang shu 40/574, biog. of Liu Chih-lien; also Pei shih 81/2726, biog. of Li Hsüan; Pei shih 82/2760, biog. of Fang Hui-yüan. Cf. also the Ch’un-ch’iu exegetical titles that include the San chuan in ss 32/932. For T’ai tsung’s own instruction in the Tso chuan, see cts 189A/4950, biog. of Chang Hou-yin. In ckcy 7/221, he cited the Ch’un-ch’iu, duke Chuang 16th year, in discussing the tabooing of his own given name; see Tso chuan chu shu 47/7b. 53  For examples of generals who knew the Tso chuan, see cts 104/3212, biog. of Ko-shu Han, d. 756; cts 141/3850, biog. of T’ien Hung-cheng, d. 821; cts 151/4061, biog. of Wang O, 760– 815; Yen Lu kung wen chi 8/12a, Li Kuang-pi, 708–64; Ch’üan Tsai-chih wen chi 19/1b, Ma Sui, 726–95; Ch’üan Tsai-chih wen chi 13/4, Hun Chen, 737–800.

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

118

McMullen

that K’ung Ying-ta’s sub-commentary commission had urged is especially well exemplified in the outlook of early T’ang scholars on the Ch’un-ch’iu and its exegetical literature. The main feature of K’ung’s critique of the Ch’un-ch’iu and its Three traditions was its polemical advocacy of a cautious attitude to them.54 An important reason for this caution lay in the pre-T’ang history of the Ch’un-ch’iu tradition, and the association it had come to have by the seventh century with the semipopular tradition of prognostication and omen lore that derived ultimately from Han interpretations of the Kung-yang chuan. To scholars in the early T’ang, if it were not controlled, or if it were taken up by unwise emperors, this tradition might become a threat to the state. Scholarly exposition of the Ch’unch’iu provided a highly respected means for them to emphasize their views. It was this cautious attitude to the Ch’un-ch’iu that led K’ung Ying-ta, obeying Sui precedent, not only to select the Tso chuan, the least speculative of the Three traditions, but also to use the primary commentary of the Ku-wen scholar Tu Yü. To K’ung, following Tu Yü, the canon, far from being a text-book for prophecy as the New Text tradition had had it, was simply a chronicle history. Compiled by ‘more than one’ archivist in the state of Lu, it had contained inconsistencies of style and format even before Confucius had edited it. Confucius had, by his use of select terminology, concealed moral judgements on events in the text. But he alone was not responsible for its profundity. There were also cases where later scholars had erroneously detected his hand in what were no more than accidental phrases in a composite text. For K’ung, the Tso chuan, compiled by Tso Ch’iu-ming, an archivist in Lu and a contemporary of Confucius, was a separate work. Tso Ch’iu-ming ‘had not necessarily received instruction face to face from [Confucius]’. Moreover the Tso, composed much closer in time to the canon itself than the Kung-yang or Ku-liang, was much more reliable as a record of events. K’ung also followed Tu Yü in refuting some of the beliefs that the New Text tradition had invested in the person of Confucius. It was not true, he argued, that Confucius had had an ambivalent attitude to the Chou dynasty, or that in editing the Ch’un-ch’iu he covertly ‘demoted the Chou and made Lu the royal house’. K’ung argued rather, on the basis of Analects evidence and Confucius’s use of the Chou calendar, that he was a loyal subject of the Chou. Nor was it acceptable to K’ung that Confucius should be known as the ‘uncrowned king’

54  This summary of K’ung Ying-ta’s view of the Ch’un-ch’iu and the Tso chuan is drawn from Tso chuan chu shu, preface by Tu Yü with sub-commentary by K’ung Ying-ta, passim.

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

Canonical Scholarship

119

(su wang).55 Passages in the Analects in which he condemned usurpation again indicated that he was a loyal subject of the Chou. To call him the ‘uncrowned king’ was a calumny. ‘Alas! Confucius has long been slandered. Only through Tu [Yü] is his innocence established.’ Elsewhere in the sub-commentary K’ung Ying-ta attempted to discredit the Kung-yang and Ku-liang traditions, charging them, in an Analects phrase, with ‘the scholarship of wayside hearsay and roadside gossip’.56 In the subcommentaries to other canons also, he criticized both these texts.57 His purpose was to rid the Ch’un-ch’iu tradition as a whole of the superstitions that had become attached to it, and to divest Confucius of the royal and semi-divine status he had acquired. K’ung did not pursue this aim rigorously or consistently in the sub-commentary, for, when it suited his purpose to do so, he accepted evidence supplied by texts he elsewhere condemned as unreliable. But his basic position was fully consonant with the political priorities that, in court discussions, the Chen-kuan scholar ministers formulated for the emperor. Early T’ang attitudes to the Mencius The practical temper of early seventh-century canonical scholarship becomes particularly clear in the attitude of official scholars to the Mencius. The ancient philosopher who, from the ninth century on, was to provide one of the main sanctions for a revived, more introspective Confucianism, had a very different status in the early T’ang. Mencius himself had less standing than he had had in Han times, and a far lower position that he was to enjoy as the ‘second sage’ (ya sheng) of the Neo-Confucian era. Since he was not a commentator, he was not represented in the Confucian temple. There was no record of a local temple or cult to him, though his tomb may well have survived.58 The preeminent Meng clan of the T’ang, from P’ing-ch’ang in modern Shantung, 55  Despite K’ung’s condemnation of the title su wang, it remained current throughout the T’ang; see e.g. Ch’üan T’ang wen 175/5b, Ts’ui Hsing-kung, commemorative text of 666; stts 20/14a–b and Hung (1969), p. 6 and note 19 on p. 15; Ta T’ang chiao ssu lu 10/13a, Li Hsü in 788, attempting to prove the superiority of Confucius over T’ai kung; twt 26A/7a, Li Kuan, asking for repair of the T’ai hsüeh; Ch’üan Tsai-chih wen chi 23/1b, of Wei Chümou’s commitment to Confucianism. 56   Analects 17/12. 57  For K’ung’s criticism of the Kung-yang and Ku-liang commentaries, see e.g. Shang shu chu shu 4/7b–8a, quoted by Chang Hsi-t’ang (1935), p. 8; Li chi chu shu 53/29a; Tso chuan chu shu 1/16a. 58  A tomb to Mencius is mentioned in T’ai-p’ing huan yü chi 21/8b. The Yü-ti kuang chi of c. 1111–7, 7/73, also mentions a tomb. A local temple to Mencius was established in 1083, when Mencius was enfieffed as duke of the state of Tsou. Mencius’s official introduction

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

120

McMullen

claimed descent from a related forbear who had been mentioned in the Analects, and references to Mencius as an ancestor are rare.59 Mencius himself was not considered a canonical teacher, even though he was, in K’ung Ying-ta’s words, ‘a wise man of antiquity, in all the disciple of Tzu-ssu, not far distant from the sage Confucius’.60 Official scholars entered the book that bore his name in the philosophical rather than the canonical division in the official bibliographical scheme, after books ascribed to two disciples of Confucius, Tseng Shen and Tzu-ssu.61 In their attitude to the content of the Mencius, however, seventh-century scholars provide an even greater contrast with their ninth-century successors. On the one hand, the Mencius contained much prescriptive information on such topics as ritual usage, taxation and land tenure, besides short, chance remarks on a wide range of practical subjects. On the other, it also had long comments on doctrinal questions, particularly on the interior emphasis within Confucianism, the emotions, fate and the moral quality of the nature (hsing). Early T’ang scholars found the Mencius important above all for the practical information the text contained. In the Mao shih sub-commentary, by far the greatest number of K’ung Ying-ta’s citations of the Mencius concern remarks on ritual, social or administrative questions. Only one, on man’s inner capacity as a recipient of offerings in the Confucian temple followed in 1088; see Shan-tung t’ung chih, p. 13; T’ao Hsi-sheng (1972b), p. 85. 59   Yüan-ho hsing tsuan 9/15b. The Meng clan of P’ing-ch’ang in modern Shantung claimed descent from duke Huan of Lu, whose second son Ch’ing-fu had adopted the surname Chung-sun, later changed to Meng-sun. In the T’ang period, the Meng clan of P’ingch’ang claimed descent from Meng Ching-tzu, a member of this family who figured in the Analects, 8/4. See the ctcc epitaphs for Meng Chun, d. 714 aged 56 sui and for Meng Hui, d. 733 aged 65 sui. Mencius himself was a great grandson of Meng Ching-tzu. The kfspwk epitaph for Meng Shih, d. 704 aged 70 sui, would seem uncommon in claiming for him direct descent from Mencius. Meng Ching-su, prioress of a Taoist monastery, for whom Ts’en Wen-pen wrote a stele in 638, was traced, in imprecise terms, by descent from the circle of Confucius and Mencius; see Chin shih hsü pien 4/5a. I am grateful to Professor van der Loon for this reference. 60   Li chi chu shu 16E/13a. 61   s s 34/997; see also ss 34/999 for description of Mencius, Tzu-ssu and Hsün tzu as true disciples of Confucius in the Chan-kuo period. cts 47/2024, Ho chih, pp. 168–9. Mencius’s association with Hsün tzu, deriving in part from the juxtaposition of their biogs. in Shih chi ch. 74, was as characteristic of the medieval period as the juxtaposition of K’ung and Mo. For examples of the former, see tkck 2/3a, memorial of 653 by Chang-sun Wu-chi; Yang Ying-ch’uan chi 3/2a; stts 3/14a; twt 84/2a, Liu Mien; wyyh 680/10a, P’ei Tu. The Sung pi-chi writer Hung Mai (1123–1202) commented critically on the latter juxtaposition; see Jung chai hsü pi 14/135–6.

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

Canonical Scholarship

121

to distinguish good and evil, bears on the interior life.62 In the sub-commentary to the Chou i, the canon that of the inner series of five provided most detailed sanction for speculation on fate and the individual, K’ung did not draw once from the Mencius.63 In Chia K’ung-yen’s sub-commentaries for the Chou li and the I li,64 and in the early T’ang encyclopaedias Ch’ün-shu chih yao (Essentials of good government from many books), and I wen lei chü (Literary material arranged by category),65 the pattern is similar; the citations are generally restricted to concrete matters, rather than to psychological or interior questions. This neglect of the psychological theme in the Mencius does not mean that early T’ang scholars never considered interior ideas in Confucian terms. They certainly recognized as Confucian a process of self-cultivation, of, in the expression from the Chou i, ‘making an exhaustive discrimination of truth and effecting the complete development of their natures’ (ch’iung li chin hsing).66 K’ung Ying-ta, defending Confucius from an implied charge that he was afraid of death, stated that he had ‘effected the complete development of his nature’ before dying,67 and Yü Shih-nan (558–638, canonized Wen-i), in a commemorative text for the Confucian temple, stated that Confucius had no equal in the past, or the future, for self-cultivation as the Chou i defined it.68 When seventh-century Confucian exegetical scholars needed to cite Confucian formulations on interior questions, they probably relied on the commentaries on passages in the Ta hsüeh, Chung yung, Li yün (Cycle of ritual) and Yüeh chi (Record of music) chapters of the Li chi, and on the Chou i

62   Li chi chu shu yin shu yin-te, Peking: Harvard-Yenching, 1937, p. 8; Mao shih chu shu yin shu yin-te, p. 12. For the one exception, see Li chi chu shu 2/6a. The passage quoted here is the same passage that the scholar minister Wei Cheng anthologized in the Ch’ün-shu chih yao, 37/2b–3a. K’ung Ying-ta simply followed Cheng Hsüan in citing the Mencius, and did not use the passage to amplify a view of the nature. 63  For the Chou i sub-commentary, see Wang Chung-lin (1959), pp. 1–111. 64  For Chia Kung-yen’s citations from the Mencius, see Chou li yin-te fu chu shu yin shu yin-te, p. 161, and I li yin-te fu Cheng chu Chia shu yin shu yin-te, p. 61. 65  See especially I wen lei chü 21/384–87, entry for hsing ming. Though the views of Hsün tzu are represented, in a fu by Chung Chang-ao of the Chin, no mention is made of the Mencius in the entry. 66   Chou i chu shu 9/2a. 67   Tso chuan chu shu 1/16a–b, referring to Li chi chu shu 7/9b–10a. In his gloss on the term meng tien, K’ung Ying-ta argues that Confucius had feelings like ordinary men and dreamt like ordinary men. 68   Yü Pi-chien chi 1/14a; cf. also 15b.

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

122

McMullen

and Analects.69 What may be inferred from their relative lack of interest in the Mencius is that, in this period, Confucian ideas on introspective doctrinal questions were not reviewed with the degree of precision or sense of enquiry that was focussed on problems of a political, institutional or ritual kind. The great majority of scholars and others in the China of Kao tsu and T’ai tsung turned to Buddhism or Taoism for answers to questions of ultimate value. Only from the late eighth century, in the more devolved intellectual milieu brought about by the political decline of the dynasty, did radical hostility to these religions develop among a minority of scholars, and a revival of interest in treating such problems in exclusively Confucian terms take place.

From 653 to 755

When compared to its history in the first two reigns, the official canonical exegesis that scholars produced over the century from the promulgation of K’ung Ying-ta’s Wu ching cheng i until the outbreak of rebellion of 755 is sparsely documented. Probably few compilations of the length of the great subcommentary series, or of the historical importance of late eighth- and early ninth-century re-interpretations of the canons were written, and none has survived. Although in other areas of learning, in state ritual for example and in official genealogy and in administrative fields such as statutory and criminal law, second or even third versions of early T’ang official compilations were commissioned, K’ung’s sub-commentaries remained the only officially commissioned series in canonical exegesis. Scholars in academic office over this century, nonetheless, continued on their own initiative to compile numbers of sub-commentaries or glossaries to the canons. In a succession of controversies that took place in the academic agencies, they also raised questions about the value of the officially established primary commentaries. Scepticism towards the established texts reached an extreme point when the historian Liu Chih-chi cast doubt on the value of certain features of the Ch’un-ch’iu and Shang shu. The problem of how his and other spirited opinions affected the standing of K’ung Ying-ta’s official subcommentaries is one that dominates an account of canonical scholarship for the hundred years from T’ai tsung’s death.

69  For formulations of the nature of the sub-commentary series, see Chou i chi shu 1/4a–b; Li chi chu shu 52/1b (Chung yung); Mao shih chu shu 18C/7b; Li chi chu shu 27/6a–b (Yüeh chi). See also Yang Hsiang-k’uei (1958), pp. 8–11.

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

Canonical Scholarship

123

In the school system and in the examinations, the authorities continued to use the established texts and approved commentaries. But, in the increasingly harsh competition for success, idealistic scholars considered the attitudes of candidates to the canons less than satisfactory. The tradition of complaint that grew up over candidates’ outlook on their curricula was another feature of the canonical tradition in this century. Finally, towards the end of the period, under the changed conditions of the late K’ai-yüan (713–42) and T’ienpao (742–56) periods, among the ‘Confucian opposition’ the more distanced attitude towards the court and the political centre was also expressed in the context of the study of the canons. Criticism of Officially Established Texts Much evidence suggests that, from Kao tsung’s reign on, K’ung Ying-ta’s official sub-commentaries and the primary commentaries that they endorsed enjoyed general acceptance in the scholarly world and in the school system. Not only did high authority commission no second sub-commentary series, but no complete work of Confucian canonical scholarship survives from the second half of the seventh century, and relatively few from the first decades of the eighth. At Tun-huang, fragments of the primary commentaries endorsed by K’ung Ying-ta far outnumber those other commentaries, and fragments of K’ung’s own sub-commentaries were preserved in some numbers there.70 The impression is therefore given that the series, following its promulgation in 653, effectively displaced existing works, while new, unofficial exegesis posed no threat to its pre-eminence. In view of the immense prestige of official service and of the examinations in the canons as a means of gaining official status, it is not surprising to see the officially approved commentaries and sub-commentaries prevail in this way. Despite the impression given by this lack of surviving material, however, over the century from 653 to 755, individual scholars initiated a series of attempts, supported in some cases by detailed scholarly argument, to add to or modify the existing range of commentaries. These attempts, coming from academic officials and resting more on scholarly grounds than those of administrative convenience, form a distinct body of material, peculiar to the second half of the seventh and the early eighth century. In Kao tsung’s reign, examples of compilations that might have threatened the place of the official exegesis are not well documented. But very brief accounts survive suggesting that the court considered the work of other 70  Ch’en T’ieh-fan (1969), pp. 149–81, reviewing the Chou i; Shang shu; and Mao shih; Su Yinghui (1968b).

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

124

McMullen

canonical scholars. Early in the reign, the emperor ordered that the erudits and others should discuss the correctness of emendations and new glosses by Ts’ui I-hsüan, a commentarial and phonological scholar of this period.71 Again, not long after the completion of K’ung’s series and its extension, a certain Li Hsüan-chih, a disciple in ritual scholarship of the ritual exegete Chia Kung-yen and a representative on the Confucian side in three-teachings debates at court, wrote a dictionary for the three ritual canons that was said to have circulated widely. He was also an expert on the Ch’un-ch’iu with the Tso chuan, and on the Mao shih, Han histories and Lao tzu and Chuang tzu.72 There is no evidence that the dynasty tried to suppress his activities; indeed, the fact that the titles of his compilations were recorded in his official biography indicates approval of them. But again there is not enough evidence to make detailed comment on his scholarship in relation to official exegesis possible. Several decades later, in 703, Wang Yüan-kan (d. after 705), an erudit of the Ssu-men hsüeh and concurrently scholar of the Hung-wen kuan, submitted what were evidently controversial writings on the Shang shu, Ch’un-ch’iu, Li chi, Hsiao ching and Shih chi. An edict ordered official scholars to discuss the advisability of his interpretations. Some senior scholars, who ‘adhered exclusively to the commentaries (chang-chü) of earlier exegetes, deeply ridiculed him for tampering with former interpretations’. But others in more junior posts, among them the historian critic Liu Chih-chi and his colleague Hsü Chien (659–729, canonized Wen), came to his defence, and he was finally given an official commendation that praised his conduct and scholarship. Again, however, beyond the fact that they directly challenged established views, and that in one debate he advocated a mourning period of 36 rather than 25 months, nothing is known of his controversial opinions.73 In the long reign of Hsüan tsung, however, dissatisfaction among official scholars over the established commentaries is documented in some detail, very possibly because the strength of the emperor’s own interest ensured the preservation of the records concerned. From the start of his reign, the emperor 71   c ts 77/2689, biog. of Ts’ui I-hsüan; tfyk 606/14a–b, undated edict, probably in Kao tsung’s reign. The biog. states that ‘in the end the matter did not proceed’. 72   c ts 189A/4950, biog. of Li Hsüan-chih. His glossary, the San li yin i (Sounds and meanings in the Three ritual canons), is mentioned in the biog., but not in the cts bibliography. 73  For Wang Yüan-kan, see thy 77/1405; cts 189B/4963, biog. His compilations are not entered in the cts bibliography; but see hts 57/1428, 1434, 1441, 1443, 1457, 1500, and Ho chih, pp. 19, 29, 42, 45, 68, and 152. Wang Yüan-kan’s hts biog., 199/5666–68, includes an argument by Chang Chien-chih, c. 625-after 705, canonized Wen-chen in 808 (thy 80/1488), refuting his contention that the mourning period should last for 36 months.

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

Canonical Scholarship

125

seems likely to have encouraged a climate in which discussion took place. Even before his accession, Chang Yüeh, later one of the great scholar ministers of his reign, proposed that Hsüan tsung convene a commission to re-establish the texts of the canons.74 There then followed two episodes in which official scholars and former colleagues of Chang Yüeh tried to alter the statutory establishment of texts. The first of these started in 719, when an edict was issued ordering scholars to discuss the respective merits of primary commentaries for the Shang shu and the Hsiao ching by Cheng Hsüan and K’ung An-kuo, neither of which was established. In a second edict a few days later, the Shang shu was removed, and the discussion was limited to the Hsiao ching commentaries, to primary commentaries on the Lao tzu by Ho-shang Kung and Wang Pi and to a section of the Chou i. In this second edict, the possibility of a plurality of approved commentaries was made explicit: ‘If more than one commentary is found to be about equal in doctrinal exposition, may they not all be made current?’ The debate that followed was narrowed in effect to argument over the value of one official and one unofficial commentary each for the Lao tzu and Hsiao ching. It is dominated in surviving accounts by the submissions of the great historian critic Liu Chih-chi. Liu, in advocating in both cases that the unofficial commentaries be adopted, was opposed by ten other official scholars led by Ssu-ma Chen, an erudit of the Kuo-tzu hsüeh and author of an important commentary on the Shih chi (Ch. 5, p. 174). Both sides argued over the transmission of the primary commentaries involved. They cited their occurrence as titles and their length in chüan as these were recorded in earlier sources, and discussed the evidence for their authorship and the value of their content. Liu Chih-chi, in arguing for flexibility and change, quoted the precedent of the Tso chuan and its establishment, after controversy, as late as the Han period.75 Invaluable as this debate is in showing the sophistication of early eighthcentury views of textual transmission, its outcome itself was no less significant. The emperor ordered that the official commentaries remain the established ones, but that the study of the alternatives should be encouraged, so that their transmission would be ensured. Then, not long after the debate, in 722, Hsüan tsung produced his own commentary to the Hsiao ching.76 Some years later, 74   Chang Yen kung chi 13/1a; wyyh 652/2a–b. 75  For Liu Chih-chi’s debate, see thy 77/1405–10; ttclc 81/467–68; tfyk 50/6b–7a; tkck 6/1a–8a; Ta T’ang hsin yü 9/143–44; Hung (1957), pp. 74–134. 76   t hy 36/658; cts 102/3178, biog. of Yüan Hsing-ch’ung; cts 8/183. For Hsüan tsung’s preface and its date, see tkck 7/13b, note by Hsü Sung; for a rubbing of the text, see Tō Gensō Sekidai Kokyō.

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

126

McMullen

in 735, he wrote a Lao tzu commentary.77 He also continued to be interested in the Shang shu, the other canon involved in the first stage of the debate.78 By the completion of the Liu tien in 738, both his Hsiao ching and his Lao tzu commentaries had been officially established.79 The emperor’s compilation of these commentaries, against the background of his encouragement of discussion, confirmed his position as head of the scholar community. It placed him, as he surely intended, in the tradition of the scholar emperors of the pre-T’ang period, including even those of the scholarly Liang house in the south. A second expression of dissatisfaction with officially established texts in the K’ai-yüan period was initiated by Yüan Hsing-ch’ung (653–729, canonized Hsien), another leading official scholar, a bibliophile, president of the state academy directorate and major figure in the Chi-hsien yüan, Hsüan tsung’s own advisory college. The controversy he caused illustrates the other face authority might turn, when confronted with the demand for administratively inconvenient change. Yüan Hsing-ch’ung, who had compiled a sub-commentary to Hsüan tsung’s commentary on the Hsiao ching, also undertook a subcommentary for a reorganized version of the Li chi that Wei Cheng, the scholar minister of T’ai tsung’s reign, had edited and that a descendant now promoted (Ch. 4, pp. 120–21). This reorganized version probably involved arranging the prescriptive parts of the text as a code for contemporary ritual practice. In 726, however, having presented the compilation, Yüan Hsing-ch’ung encountered the resistance of Chang Yüeh, who was now chief minister and president of the Chi-hsien yüan. Chang claimed that the Li chi was ‘an inerasable authority’ (pu k’an chih tien), and that it should not be changed. The outcome was that Yüan’s sub-commentary, instead of being ‘established in the schools, transmitted and taught’, was ‘stored in an inner depository and not in the end established in the schools’. Yüan consoled himself by writing a polemical essay, in which he bitterly condemned the exaggerated and inflexible respect paid to established primary commentaries. Like Liu Chih-chi, he cited the precedent of the Tso chuan, as a text whose history demonstrated the benefits of more flexible attitudes.80 The scholarly life of the second half of Hsüan tsung’s reign was dominated by Li Lin-fu’s autocratic control of administration, and by the emperor’s growing commitment to Taoism. The freer atmosphere that had permitted 77  For the Lao tzu commentary, see thy 36/658; tfyk 53/16a; also thy 77/1410–11. 78  For the Shang shu, see thy 77/1410–11; tfyk 50/11a–b; tkck 9/6a; hts 57/1428. 79   t lt 21/6b. 80  For Yüan Hsing-ch’ung’s argument, see cts 102/3178–81, biog. of Yüan Hsing-ch’ung; thy 77/1410; Ta T’ang hsin yü 2/117; Kramers (1955).

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

Canonical Scholarship

127

c­ ontroversies in the academic agencies, like those involving Liu Chih-chi and Yüan Hsing-ch’ung, was no longer possible. The two developments in official canonical scholarship of the T’ien-pao period were, rather, prescriptive. Both may be seen as expressing the interests of the emperor or of high authority. Neither is represented in surviving documentation by any record of debate in the community of official scholars. The first concerned the Yüeh ling (Monthly commands). This was a chapter of the Li chi which had from pre-T’ang times been recognized as being distinct in its origins from the rest of the canon, and which contained a calendar of ritual and administrative prescriptions, given in terms of the yin-yang theory of the late pre-Ch’in period.81 As one of the principal canonical sanctions for the cosmic dimension to the dynastic state, the Yüeh ling was of special importance to successive T’ang sovereigns. Early in the dynasty, in 640, T’ai tsung had raised the question of how literally the Yüeh ling was to be followed.82 The text had also played a prominent part in controversies about the Ming-t’ang.83 A ritual that involved the reading out of the monthly or seasonal commands that it contained was associated with the building, and was the subject of debate in the late seventh and early eighth centuries (Ch. 4, p. 127).84 Then, in about 725, Hsüan tsung appears to have had the text of the Yüeh ling separately re-determined.85 From 738 for a brief period, he required the president of the court of sacrifices to introduce and lecture on one section of the text each month.86 In 743, by imperial order, the Yüeh ling was placed at the head of the Li chi.87 In 746, its name was changed by official 81  For seventh-century views of the origin of the Yüeh ling, see K’ung Ying-ta, Li chi chu shu 14/1a; ss 32/925; Lu Te-ming, Ching tien shih wen 1/22a. 82   c kcy 2/29–30. 83  For the place of the Yüeh ling in Ming-t’ang controversies, see e.g. thy 11/272; 11/277; cts 22/857; 22/869 etc. Cf. also the role of the Yüeh ling under the empress Wu in 699, in determining the scheduling of a military rehearsal ceremony; thy 26/502–03 and cts 89/2900, biog. of Wang Fang-ch’ing. Also the memorials on the scheduling of executions, which according to Yüeh ling theory had to be in the autumn; cts 102/3175 and wyyh 617/6b–8a, Hsü Chien; Ch’ü-chiang Chang hsien-sheng wen chi 7/8b. 84  For the ceremony of reading out seasonal commands, an observance that was associated with the text, see thy 26/491–92; tt 70/385. 1–386.1, with disapproving comment by Tu Yu. 85  For arguments that the text of the Yüeh ling was separately determined in about 725, see ‘T’ang Yüeh ling chu po’, by the Ch’ing scholar Ch’eng Jung-ching, in T’ang Yüeh ling hsü k’ao. 86   t hy 26/491–2; stts 24/914. 87   t hy 75/1374.

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

128

McMullen

edict to Shih ling (Seasonal commands).88 At about this time, the chief minister Li Lin-fu directed a commission of eight, including the Taoist Ch’en Hsi-lieh (d. 757) and the prolific official scholar Lu Shan-ching, to compile a new commentary for it.89 This commentary, which used the text re-determined earlier in the reign, survives. From this time on, the Yüeh ling was brought to the fore in the state ritual programme. Its portrayal of the emperor as ensuring the harmony of the cosmos by the quality of his government and by his ritual acts commended it not only to Hsüan tsung but also to post-rebellion rulers. But the Yüeh ling was also to register particularly clearly the difference between the public outlook of scholars, especially as this concerned the emperor, and their unofficial attitudes. It was to be the subject of incisive criticism by scholars writing unofficially about state ritual in the sceptical climate of the late eighth and ninth centuries (Ch. 4, pp. 155–57). The second development in official canonical scholarship of the second half of Hsüan tsung’s reign concerned the Hsiao ching. Much favoured for court lectures, and as a concise summary of Confucian teaching, the Hsiao ching was held to encapsulate Confucian teaching, just as the Lao tzu stood for Taoism and the Chin-kang ching (Diamond sutra) for Buddhism. In 743, Hsüan tsung ordered that a revised version of his own Hsiao ching commentary be distributed throughout the empire.90 In an act of grace the following year, he required that all households have a copy.91 In 745, the canon and his commentary were engraved in stone and set up in the state academy directorate. This, the first datable instance in the dynasty of the stone engraving of a Confucian canon, still stands at Sian. Alongside the text is an engraved list of the official scholars commissioned to prepare it.92

88   t hy 77/1410; cts 9/219; tfyk 50/11b. 89  For the commentary, see T’ang Yüeh ling chu. This version of the text displaced earlier officially determined versions; see Chang Shen, Wu ching wen tzu hsü (Preface to model characters of the Five canons), quoted in tkck 11/5b, double column entry); see also below at note 153. 90   t hy 77/1411; thy 36/658. For the Hsiao ching as a text representing Confucianism, see Lo Hsiang-lin (1954), pp. 91–2, showing that Hsüan tsung produced commentaries to all three. Before this, Wang Te-piao, d. 699 aged 80 sui, had done the same; see his ctcc epitaph. Later, the south-eastern scholar and Te tsung favourite Wei Chü-mou, 749–801, canonized Chung, did likewise; see Ch’üan Tsai-chih wen chi 23/2b–3a. 91  For the order to distribute empire wide, see ttclc 74/417, act of grace of 12th month of 744 (corresponding to 1st day of 2nd month of 745 on the Julian calendar); cts 9/218; tkck 9/6a–b, noting variant date. 92  For the Shih t’ai Hsiao ching, see above, note 76.

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

Canonical Scholarship

129



Complaints against the Standards of Canonical Learning in the Examinations Hsüan tsung’s promotion of his own commentaries on the Lao tzu and the Hsiao ching was ultimately addressed to the operation that played a large part in shaping T’ang attitudes to the canons, the examinations. Here in the K’ai-yüan and T’ien-pao periods the authorities ran into a number of practical problems. The increase in competition for office that took place over the K’ai-yüan period93 affected for the worse the attitudes of candidates to examination curricula. Aiming for success, their approach was one of practical expediency, and they all opted for the shortest texts that the curriculum regulations allowed. The response of academic officials to this problem took the form of attempts to ensure that all the canons in the series of nine were studied by examination candidates, and that none was neglected. Scholars were also led to criticize more generally the place canonical learning had assumed in the examinations, and to condemn as particularly meaningless the rote learning of the canons that they involved. The wish to see all nine of the canons figure in candidates’ choice of text, and perhaps also to have more candidates from the state academy directorate successful in the competition for office, is evident from two mid K’ai-yüan memorials from the directorate’s senior officials. In 720, a vice-president memorialized indicting that students were neglecting the Chou li and the I li, two medium canons, and the Kung-yang chuan and Ku-liang chuan, two minor canons, while, ‘with their minds bent on passing, they all rushed to the Li chi, on the grounds that [for a major canon] its text was short’. He requested that lower pass standards for the context questions on the four neglected canons be approved.94 In 728, Yang Ch’ang, the president, protested that the Tso chuan was also neglected, and asked that those who selected it or the other four should be given concessions in the competition for office.95 Dissatisfaction with the level of knowledge expected of candidates in the examinations was most eloquently expressed in this period by Liu Chih, a son of the great critic Liu Chih-chi, and later a reformer of the directorate (Ch. 2, p. 52). In a memorial submitted in about 735, Liu Chih criticized the basis of 93  For the increase in the numbers of candidates, see Yang Ch’ang, memorial of 729, thy 75/1376; tt 17/96.1; tfyk 639/22b–23a; also the memorial of Liu Chih, tt 17/98.3 on the increase in the numbers of those with official status. See also Guisso (1978), pp. 87–106 and Herbert (1986a), pp. 205–8, for the seventh-century background. 94   t t 15/83.2; thy 75/1376; tfyk 639/22a–b. 95   t hy 75/1376; tfyk 639/22b–23a; cts 185B/4820 and hts 130/4496, biogs. of Yang Ch’ang; Herbert (1986), p. 208.

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

130

McMullen

selection, especially the increasingly literary emphasis of the chin shih and the superficiality of the approach that required only rote learning of the canons.96 Another way in which scholars expressed disapproval of the approach that examinations required was more indirect: they commended, in biographical tributes, those who did not study detailed exegesis of the canons (chang chü), but rather, in their concern for the world and its problems, sought for the deeper and more general meaning of the texts. Even those who went on to be successful in the chin shih might be praised in this way, in effect for an approach that transcended examination requirements.97 Perhaps in response to the complaints that Liu Chih and others submitted, there was some attempt in 737 to make both the chin shih and the ming ching a more substantial test.98 In the T’ien-pao period, there were other, minor adjustments to the place of the canons in both the chin shih and the ming ching syllabuses. But, under the restrictive control of Li Lin-fu,99 the basic problems remained unsolved. Nor did the brief three-year period from Li’s death in 752 until the outbreak of the rebellion, during which chin shih quotas were relaxed, have any lasting effect on the system.100 Liu Chih-chi’s views on the Ch’un-ch’iu and the Tso chuan Despite the disdain that, on the grounds of its length, examination candidates showed the Tso chuan, this was still one of the canonical texts over which scholars of the canons expressed lively opinions. The Ch’un-ch’iu, the canon that was integrated with the Tso, moreover, retained its position as the cryptic but authoritative source for a wide range of principles, policies and ritual 96   t t 17/96.1–97.1; Herbert (1986a), pp. 95–8. 97  For some commendations of scholars who were educated in the reign of Hsüan tsung, and whose attitude to the canons transcended the limitations of mere exegesis, see wyyh 955/7b, Sun Ti, epitaph for his father; Yen Lu kung wen chi 9/10a, of Hsien-yü Hsiang; Yen Lu kung wen chi 8/1b, of Kuo K’uei; Ch’üan Tsai-chih wen chi 19/1b, of Ma Sui; Ch’üan Tsaichih wen chi 23/3b, of Wei Yü; Liu Ho-tung chi 8/114–5, of Liu Hun; wyyh 972/9b, Liang Su, of Tu-ku Chi. 98   t lt 4/5a–b; thy 75/1377; tfyk 639/24a–25a; Herbert (1985), p. 86. tkck 8/16b, note by Hsü Sung, cites a comment in tfyk stating that this reform was in response to a memorial by Yao I. thy 76/1379 mentions a reform proposal by Yao I in 736. 99  In 740, only 15 chin shih were passed; in 741, only 13; see tkck 8/27a; 8/30a; 9/25a; 9/27b. Li Lin-fu also affected decree examinations, as the unusually direct indictment by Yüan Chieh indicates; see Yüan Tz’u-shan chi 4/52, and Sun Wang (1957), pp. 16–7. See also Feng shih wen chien chi 3/2. 100   Feng shih wen chien chi 3/2; T’ang chih yen 14/154; tfyk 640/2b; des Rotours (1932), p. 141, note 2.

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

Canonical Scholarship

131

i­ssues. It remained a much quoted authority in court and in the course of administration. Hsüan tsung cited it to encourage agriculture and to discourage the submission of omens. Individual scholars continued to refer to it in connection with history writing; it provided sanction for the management of the ancestral temple, for the problem of the length of the mourning period, and for the canonization system. In 731, Yü Hsiu-lieh, then a corrector at the imperial library, and later a major official scholar, even tried to prevent it from passing into Tibetan hands. His grounds were that it testified to a deteriorated political situation, in which the feudal lords were over-powerful in relation to the Chou emperor. It was not in the dynasty’s interests, he argued, that the barbarians should know of such a development.101 For most of this century, official scholars continued to endorse the outlook on the Ch’un-ch’iu promoted in the first two reigns, namely that the Tso chuan was the most important of the Three traditions to the canon. Most of the works written over the period now exist only as titles. But the provocative critique of the whole tradition that Liu Chih-chi incorporated in his Shih t’ung (Generalities on history) of 710 does much to make good this loss. Liu’s perspective on the Ch’un-ch’iu and the Tso represents an extreme point in the questioning of the authority of Confucius that is perhaps to be related indirectly to the success and confidence of seventh- and early eighth-century government. His great love for the Tso points again to the T’ang preference for detailed and vivid historical narrative. Much of what Liu said about the Ch’un-ch’iu and the Tso chuan was not controversial in the context of seventh- and early eighth-century opinion. Just as K’ung Yung-ta had done, he argued that the Ch’un-ch’iu was one annal among many extant when Confucius, adhering strictly to the Chou calendar and paying particular attention to ritual and music, had edited the text.102 Confucius’s 101  For Hsüan tsung’s mentions of the Ch’un-ch’iu, see thy 28/534, in 725; cts 8/201, in 734. For Yü Hsiu-lieh’s request of 731, see thy 36/667; wyyh 694/12a; cts 196A/5232. After a discussion, his proposal was rejected. The influence of the canon in history writing is to be seen for example in the projected extension of the Ch’un-ch’iu and the Tso chuan by P’ei Kuang-t’ing and others; see below, note 108. For a more modest extension of the Ch’un-ch’iu, see cts 190B/5013, biog. of Liu Yün-chi. For the Ch’un-ch’iu in the management of the ancestral temple, see thy 17/353; 17/355 etc.; in policy towards the barbarians, Ch’ü-chiang Chang hsien-sheng wen chi 16/3b; in argument over the length of the mourning period, hts 199/5666–68, memorial by Chang Chien-chih. 102   s tts 1/5b; 3/3b; Tso chuan chu shu 1/5a, both quoting Mencius 4B/21. Liu Chih-chi also quoted Mo tzu, who suggested the figure of 100, which in turn was supported by the Kungyang tradition; see Kung-yang chu shu 1/1a–b and stts 1/5b and commentary at 1/7a. Cf. also ss 42/1197, biog. of Li Te-lin, for the same Mo tzu citation.

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

132

McMullen

main aim, also, had been to show how retribution followed evil action and goodness brought success. What made Liu’s critique original was that he characterized as shortcomings precisely those features of the Ch’un-ch’iu that others had accepted as essential to its sanctity as a canon. To Liu, the fact that Confucius, with copious material available to him, had limited the coverage of the Ch’un-ch’iu to the state of Lu was a demerit.103 The way in which, by distorting the account, Confucius had edited his own judgements into the text was also unsatisfactory.104 Liu’s main attack on the Ch’un-ch’iu analyzed these grounds for dissatisfaction in twelve instances. His critique as a whole, despite formal deference to the canon, includes references to the ‘straight brush’, the ‘clear mirror’, the ‘faithful echo’; also to clarity of structure, and ‘making the good and evil plain’. In this way Liu indicated his impatience with the way in which Confucius was held to have concealed his judgements in the canon, and his own preference for a dispassionate and objective narrative. A recurrent feature of Liu’s criticism of the Ch’un-ch’iu was his appeal, sometimes explicit, sometimes implied, to other historical texts of the Chou period. Of those none was more important to him than the Tso chuan, the text he had loved since early youth.105 Liu believed that the Tso had been edited into its final form by Tso Ch’iu-ming, who, as an archivist in Lu, had a status, he seems to imply, close to that of Confucius himself. Tso Ch’iu-ming had had access to a wide range of documents, and to the disciples of Confucius from all the states, thus greatly increasing the range of his evidence. The result was that the Tso was much more comprehensive than its two rival traditions, the Kung-yang and the Ku-liang.106 Liu defended the Tso against the charge that it continued material not directly relevant to the Ch’un-ch’iu, for, ‘… supposing people down the ages had when studying the Ch’un-ch’iu merely drawn from the other two traditions, then our knowledge of the events of [the] 240 years [spanned by the canon] would be vague and imperfect, so that subsequent scholars would be like the blind and the deaf …’107 Liu’s trenchant advocacy of the Tso as superior to the Kung-yang and Kuliang represents a polemical formulation of the conventional seventh-century position. But his implication that the Tso was in some respects better than the Ch’un-ch’iu itself, and his attack on Confucius as editor of the canon exceeded the bounds of acceptable criticism. Later in the dynasty, it provoked indignant 103   s tts 14/8b–9a. 104   s tts 14/7b–8a. 105   s tts 10/11a, quoted by Pulleyblank (1961), p. 137. 106   s tts 14/14a–20a. 107   s tts 14/18a–b.

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

Canonical Scholarship

133

attempts to restore to the Ch’un-ch’iu and to Confucius himself the authority and solemnity that Liu had impugned. These attempts were an aspect of the general trend, from the middle of the eighth century, towards developing the comprehensiveness and independence from the rival religious systems of Confucian ideas and attitudes. Liu’s criticism of Confucius, far from being followed up, was to be emphatically rejected. The Shih t’ung is the only extant work of the pre-rebellion century to embody important opinions on the Ch’un-ch’iu. But in the reign of Hsüan tsung, academic interest in the canon continued. Probably in the mid K’ai-yüan period, official scholars planned, but never completed, a continuation of the canon, to cover the period from the Warring States to the Sui, for which Hsüan tsung was to write the ‘canon’, or main text, while a small commission appointed to the Hung-wen kuan composed the ‘tradition’ on the model of the Tso.108 Soon after 735, also, a standard set of pronunciations and character readings was compiled for the Ch’un-ch’iu in the Chi-hsien yüan, by an otherwise unknown scholar, Lü Cheng.109 By the second half of the reign, there were signs of a shift away from exclusive endorsement of the Tso that had been conventional in the seventh and early eighth centuries. One of Liu Chih-chi’s sons, Liu K’uang, himself an official historian, may possibly have tried to downgrade its status; he held that the use of full canonization titles in the Chi chung Chou shu (Chou books from the tomb in Chi), a Chou dynasty chronicle discovered in ad 281, and in the Tso, indicated that they ‘were retrospectively prepared by later scholars and were not contemporary “orthodox form” histories’.110 The prolific official scholar of the late K’ai-yüan and T’ien-pao periods Lu Shan-ching compiled a work in 30 chüan on the Ch’un-ch’iu and all the Three traditions, suggesting that he saw value in each.111 But a more significant instance is provided again by Hsiao Ying-shih, the unofficial teacher whose students left the state academy directorate to receive his instruction. A disciple later described how, in 108   w yyh 884/10b, epitaph for P’ei Kuang-t’ing by Chang Chiu-ling; cts 84/2807, biog. of P’ei. 109  See note 143 below. 110  Liu K’uang’s opinion is preserved in Lu Ch’un, Ch’un-chiu Tan Chao chi chieh tsuan li (A compilation of the principles of the collected commentaries to the Ch’un-ch’iu by Tan [Chu] and Chao [K’uang]), 1/11b. Liu K’uang’s statement is also quoted by James Legge in Prolegomena to The Chinese Classics, Volume V, The Ch’un Ts’ew, with The Tso Chuen, Hong Kong: Lane Crawford and Co., and London: Trübner and Co., 1873, p. 31. 111  For Lu Shan-ching’s compilation, see Nihonkoku genzaisho mokuroku p. 4; Niimi Hiroshi (1937), p. 135. Lu was known posthumously as a vice-president of the state academy directorate, which suggests that this academic tenure was his highest substantive post; see cts 118/3415, biog. of Yüan Tsai.

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

134

McMullen

compiling a work of history of his own, he adopted an eclectic approach to what were considered the traditional shortcomings of each of the Three traditions.112 Hsiao taught away from the capital, in Shantung and possibly at Tan-yang, near modern Nanking. A member of the post-rebellion Ch’un-ch’iu school, Chao K’uang, was said to have been one of his followers.113 An eclectic attitude to all three of the Traditions was one of the main features of this school. Hsiao Ying-shih therefore provides an important link between early T’ang attitudes to the canon, dominated by the outlook of historians and by preference for the Tso chuan, and the school that was in the late eighth century to express some of the most striking textual and philosophical insights of the Confucian tradition in T’ang times. Disillusion with the Study of the Canons as a Means to Advancement For the majority of T’ang scholars and aspiring officials down to the K’ai-yüan period, the study of the Confucian canons was identified above all with entry into bureaucratic service. The few with specifically scholarly ambitions aspired to enter the official scholarly agencies over which the emperor presided, and in which spirited scholars like Liu Chih-chi, Yüan Hsing-ch’ung and others served. Towards the end of Hsüan tsung’s reign, however, the same complex political and social forces that had resulted in the evolution of a ‘Confucian opposition’, began also to bring about different attitudes to the study of the canons. This development took place mainly among relatively low-ranking, often provincial officials, aspirants to office and temporary or permanent recluses. Their attitudes were not of course new; but scholars now reasserted them in a new context and with stronger emphasis. By the T’ien-pao period, the largely provincial society which such writers addressed had developed its own values and literary styles. In this milieu there was never any question of surrendering the ideal of official service under an enlightened ruler. Rather verse writers like Kao Shih (704–65, canonized Chung) and others complained about the unjust hardship involved in protracted study of the Confucian canons, necessary before

112   T’ang shih chi shih 27/422; 28/426; CTShih 209/2174, Chia Yung. Hsiao’s remark was an adaptation of Fan Ning’s in his preface to the commentary on the Ku-liang chuan, see Ch’un-chiu Ku-liang chuan chu shu, pref. 5b–6a. This remark was also quoted by Chang Chiu-ling, though without reference to the Kung-yang chuan; see Ch’ü-chiang Chang hsien-sheng wen chi 16/4b. Later in the dynasty, it became a tenet of the eclectic Ch’unch’iu scholars; see twt 95/3b, Liu K’o; Ch’un-ch’iu shih li, pref. by Liu Fen. 113   h ts 202/5768, biog. of Hsiao Ying-shih.

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

Canonical Scholarship

135

a candidate could consider competing in the regular examinations.114 Tu Fu, likewise, contrasted the suffering that study involved with the idle and luxurious lives of the sons of the highly placed.115 Other scholars, like Sun Ti and Yen Chen-ch’ing, praised those who sought for deeper meaning in the canons and were not satisfied simply by detailed commentaries.116 Some writers contrasted the opportunism of successful officials at court with the poverty and integrity they found in students and aspiring scholars in the provinces. They condoned failure, whether it was long-term or relatively brief, in highly traditional terms, as the justified and dignified refusal to serve except in acceptable conditions.117 When the court was dominated by ‘close ministers’ (chin ch’en) who were morally unsuitable, the true scholar was right to withdraw and cultivate his own person until times changed.118 In this milieu, there was great respect for the religious recluse, the figure who in the words of one writer had, ‘traced back his footsteps, taken leave of humanity, shut the door and maintained his solitude’.119 As in the seventh century, for the majority of scholars the rival teachings of Buddhism and Taoism gave answers to the questions of ultimate value that their lives as contemplatives might raise. Interest in the supra-mundane faiths, moreover, in this period when Hsüan tsung was deeply committed to Taoism, did not conflict with attitudes at the apex of the official hierarchy. Some scholars and verse writers, however, drew on the Confucian learned tradition to justify their failure to gain access to high office. Individuals sometimes located a philosophy of selfcultivation in the Confucian canons, or described their withdrawal in terms drawn from the introspective strand within classical Confucianism.120 More typically, for numerous verse writers of this period, the Confucian symbols of their situation were the ‘old fisherman of Ts’ang-lang’, a figure representing composure in retirement,121 or the stronger image of Po I and Shu Ch’i, two 114   Ho yüeh ying ling chi 1/80, Kao Shih; 1/60, Wang Wei. 115   Chiu chia chi chu Tu shih 1/1, Hung (1952), pp. 56–7. 116  See above, note 97. 117  See above, Ch. 2, note 146. 118  For the term chin ch’en in this period, see Ch. 6, note 151. 119   Yen Lu kung wen chi 12/5a; Ho yüeh ying ling chi 3/114, Yen Fang; Yen Keng-wang (1959), p. 696; cf. Ts’en Chia-chou shih 1/171–18a and Yüan Tz’u-shan chi 8/116–17 for post-rebellion examples. 120   c tcc epitaph for the ming ching graduate Cheng Shen, prefect of Ch’ing-chou, d. 734 aged 83 sui; he maintained that his nature was united with the tao, and that ‘in all his actions, he did not violate goodness’ (Analects 6/7). 121  For the ‘old fisherman of Ts’ang-lang’ (Ts’ang-lang weng), see Mencius 4A/9, in which Confucius commends the fisherman, here referred to as a boy; also David Hawkes (1985),

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

136

McMullen

exemplars of the Shang dynasty who had starved to death on mount Shouyang rather than compromise their integrity by taking office under the Chou. Confucius had cited both these figures in his teaching. Mencius had referred to Po I alone, but had made him stand emphatically for the idea that service to an emperor was conditional on the emperor’s moral acceptability.122 To many of the politically unsuccessful scholars of the later part of Hsüan tsung’s reign, these images, drawn from the canons, dignified their exclusion from the political centre. In the post-rebellion period, it was to be the successors to these alienated writers, as well as official scholars in metropolitan posts, who would take up the Confucian canons with motives quite different from those of their seventh century forbears, and make existential demands of these same texts.

The Post-rebellion Period

In the first half of the T’ang, the dynasty had made two main demands of the canonical tradition: that it express a range of political values in the interests of the dynasty; and that it satisfy the practical, administrative needs of the examination system. In the post-rebellion period, despite the deterioration in political conditions, both the problem of the definitive text of the canons and the question of their function in the examinations remained important to the scholar community. Until well into the ninth century, officials submitted a succession of reformist initiatives aimed at restoring the health of the dynasty through education and learning in the Confucian canons. They also made a series of attempts at determining and preserving in permanent form the texts of the canons. From the late eighth century on, however, the most significant reformist ideas derived not from the academic agencies, but from scholars who formulated them in a society very different from the court-centred community of pp. 206–7. For examples of the Ts’ang-lang image in the verse of the K’ai-yüan and T’ienpao periods, see Ho yüeh ying ling chi 1/52, Ch’ang Chien; 1/73 and 74, Li Ch’i; 2/82, Ts’en Shen; 2/99, Wang Ch’ang-ling. 122  For Po I and Shu Ch’i, see Analects 18/8; 7/15; 16/12. For Mencius’s references to Po I, see Mencius 2A/2; 2A/9; 5B/1; 6B/6. A temple to Po I and Shu Ch’i stood in this period; see Ch’üan T’ang wen 357/1b, text by Liang Sheng-ch’ing. This is to be dated to 725, according to Ku Yen-wu, Chin shih wen-tzu chi 3/25a. See also Ho yüeh ying ling chi 1/73, Li Ch’i. For the image of ‘gathering ferns’ (ts’ai wei) associated with them, see Shih chi 61/3123, biogs. of Po I and Shu Ch’i; Ho yüeh ying ling chi 1/61, Wang Wei; 1/73, Li Ch’i; 2/95–6, Ch’u Kuang-hsi. For later comment on or commemoration of Po I and Shu Ch’i, see Huang-fu Ch’ih-cheng wen chi 2/1a–2a; Han Ch’ang-li chi 3/12/97–8.

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

Canonical Scholarship

137

the seventh and early eighth centuries, and sometimes in places remote from Ch’ang-an. Most of the individual scholars concerned had some experience of academic office, either in their own careers or through colleagues. But scholarly tenures now very seldom involved participation in academic commissions like those characteristic of the pre-rebellion period. In turn, the reformist ideas of both official scholars and of their colleagues in general service posts or on the periphery of official society reflected the more devolved character of scholarly life. Efforts to Reform Canonical Learning in the Examinations The government’s evacuation of Ch’ang-an in 756 meant that for a year the regular examinations were conducted on an improvised basis away from the capitals. The examinations returned to Ch’ang-an in 758, and from 765, in order to reduce the burden on the capital’s resources, were also established as a concurrent operation at Lo-yang.123 From 758 on, the chin shih and ming ching were held, with very few exceptions, every year until the end of the dynasty. They also retained their importance as the main official operation with which scholars associated canonical scholarship. Like the school system itself, therefore, they became the focus for a long succession of reformist initiatives. From Tai tsung’s accession and the court’s return late in 763 from its flight to Shan-chou, successive proposals for the reform of the schools and the examinations contained complaints at the ineffective role of canonical learning in both the ming ching and the chin shih. Typically, however, such memorials referred to the established texts by title only. They condemned mere rote learning, but did not go into more specific detail. Moreover, just as the dynasty’s response to the deterioration of the school system had been intermittently to implement reform suggestions, so the characteristic response of authority to the ineffectiveness of the examinations was to welcome and discuss reformist ideas, while making no permanently effective changes. 123   t kck 10/2b. For the unconventional performance as chief examiner of Li K’uei, in 759, see thy 76/1379 and cts 126/3559, biog. For the examinations at Lo-yang, see tfyk 640/11a and tkck 10/18a–19a. Probably the first Lo-yang examiner was Yang Wan (cts 11/276), while Chia Chih, holding the basic office of president of the right of the department of affairs of state (shang shu yu ch’eng) was chief examiner at Ch’ang-an. That Hsü Sung is probably correct in this emendation is indicated by the later accounts of a chin shih success in 765, that of Hsü Shen. Hsü was passed by Chia Chih (Ch’üan Tsai-chih wen chi 24/1a–b), and Li Ao specifically recorded of him that he was ‘entered on the register at Ching-ch’ao fu (i.e. Ch’ang-an) and was offered as a chin shih candidate’ (Li Wen kung chi 11/91a). Cf. Fu Hsüan-ts’ung (1980), p. 190, note 1.

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

138

McMullen

In 763–4, a number of officials submitted the education system as a whole to radically critical review. The vice-president of the board of rites, Yang Wan, asked for the abolition of the regular Taoist examinations introduced on Hsüan tsung’s order from 741, and suggested that the ming ching and chin shih be replaced by tests involving one of nine established canons or the Hsiao ching, Analects and Mencius combined as a single canon.124 A discussion followed. In this, Chia Chih denounced the standards by which candidates were selected. Contemporary examinees, he claimed, were ignorant of the basic concepts of moral self-analysis, including even the attainments for which Confucius had praised Yen Hui. As a result of this discussion, in 764, certain relatively unimportant categories of ‘annual tribute’ were stopped. But an edict announced that, ‘the chin shih and ming ching have been long established. Now suddenly to alter their curricula (yeh) would, we fear, cause difficulties to those [involved]’.125 Another, later detailed set of detailed proposals, from Chao K’uang, the pre-rebellion pupil of Hsiao Ying-shih and member of the Ch’un-ch’iu school, represents a first point of contact between a reform group whose ideas were developed in the provinces and the central academic system. Chao K’uang demonstrates the interest in institutional reform that often accompanied specialization in the Ch’un-ch’iu. He condemned rote learning of canons as ‘childish’, and advocated a series of new examinations, including one that would have involved collating the differences between the Three traditions to the Ch’un-ch’iu, and one on pre-Ch’in philosophers, including Mencius.126 Under Te tsung, intermittent and largely ineffectual attempts at reforming the role of official canonical scholarship continued. The issues involved were widely discussed in the official community and by prospective officials. Just as the state academy directorate officials had done under Hsüan tsung, reformers criticized the bias among candidates in favour of shorter or easier canons. Their criticisms led to the devising of several alternatives to the ming ching degree, intended to provide incentives for those selecting less popular canons. In 793, the emperor gave formal endorsement to a special examination on the 124   t hy 76/1395–96; tt 15/84.1; cts 119/3430–34, biog. of Yang Wan; cts 190B/5029–31, biog. of Chia Chih; tfyk 640/4b–10b; tkck 10/7a–13b. 125   t hy 76/1399; cts 11/275; tctc 223/7165, 764.19; tfyk 640/10b–11a; tkck 10/17a–18a. In this year, the annual tribute of candidates in the categories of child prodigies and ‘the filial and fraternal who give their strength to the fields’ was stopped. 126  For Chao K’uang’s submission, see tt 17/97.1–99.1; also wyyh 765/6a–8a, for the first part. The post of prefect of Yang-chou, by which he is referred, was probably his highest tenure. In Lü Ho-shu wen chi 4/5a, he is called ‘the late prefect of Yang-chou’; in Ch’üan Tsai-chih wen chi 41/6a, he is also referred to as ‘Chao of Yang-chou’. For his proposals, see also Herbert (1986a), pp. 108–11.

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

Canonical Scholarship

139

three ritual canons that had been established in 789, and one on the dynasty’s ritual code, to counteract what was described as the deterioration in the ritual scholarship.127 Towards the end of Te tsung’s reign, some three or four of the chief examiners were literary scholars with reformist sympathies. These scholars had social contact with some of those who, in unofficial circles, developed new approaches to the canons. One of the main trends was an increased interest in Confucian analysis of questions of ultimate value, in the ‘relation of heaven and man’ (t’ien jen chih chi), and in interior questions, or ‘nature and destiny’ (hsing ming). This interest was perhaps partly a response to the decline of the dynasty, and to the political pessimism and frustration of the second half of Te tsung’s reign. In the late Chen-yüan period, it became something of a vogue, not just for the alienated periphery of official society, but also among those who attained high office. The survival of a fuller record of questions and answers in this period, too, makes it clear that some examiners brought these ideas into the questions they set. Ku Shao-lien,128 Kao Ying (740–811, canonized Chen)129 and Ch’üan Te-yu130 all set topics that reflected renewed interest in the 127  For the San li examination in 789–93, see tt 15/84.1–84.2; thy 76/1396–97; tfyk 640/13b– 14b; cts 44/1892, commentary; hts 44/1159; des Rotours (1932), p. 130 and n. 1. For the names of successful candidates, see tkck 27/38a (Five Dynasties period); ctcc epitaph for Fan Yin, died 870 aged 42 sui, mentioning a relative, Wang Hsiu-fu, as a successful candidate. 128  Ku Shao-lien, himself a chin shih of 770 and a former grand secretary, was examiner in 793, 794, and 798. In 793, in the po hsüeh hung tz’u examination, he set the topic ‘Yen tzu does not repeat his faults’ (Analects 6/7; 6/11). Han Yü’s answer, Han Ch’ang-li chi 4/14/ 32–33, contained references to concepts of ‘sincerity and enlightenment’ drawn from the Chung yung. Tu Huang-shang said of Ku that he was ‘thoroughly versed in [the relations between] heaven and man, and exhausted fully the beginnings of nature and destiny’; see wyyh 918/3b–7b. For another important tribute to Ku as an examiner, see wyyh 988/ 4b–6b, Lü Wen. 129  Kao Ying, himself also a chin shih of 763, and a former grand secretary, was chief examiner in 799 and 800. In 800, Kao set the line ‘By nature men are close; by practice they grow apart’, from Analects 17/2. For a translation of Po Chü-i’s answer, see des Rotours (1932), pp. 335–42. In his preface for Po’s collected works, Yüan Chen named this fu as being celebrated among new chin shih candidates at the capital; see Yüan shih Ch’ang-ch’ing chi 51/1a. 130  Ch’üan Te-yü was recognized as an exception in not having taken a regular examination, a fact of which he seems to have been proud; see Kuo shih pu 2/33 and Yin hua lu 2/77. He was, however, more conventional as an examiner in having been a grand secretary; see Ch. 6, note 195. In his second chin shih dissertation question of 802, Ch’üan Tsai-chih wen chi 40/3a–b, he asked for comment on the Analects 6/7 statement that ‘for three months

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

140

McMullen

interior pole or questions relating to traditional Confucian understanding of psychology. In one of the very few ming ching dissertation questions that survived from the dynasty, set on the Li chi in 805, Ch’üan Te-yü even foreshadowed the later Neo-Confucian grouping of the Four books, when he asked about the attitude to self-cultivation of the Ta hsüeh and the Chung yung.131 From the start of the ninth century on, there were further intermittent attempts at improving the standard of canonical learning required in the examinations. The reformist submissions these involved typically cited the canons by name only and complained of the poor state of candidates’ knowledge of them.132 Rote knowledge of commentaries, in particular, continued to be frequently condemned, often in tandem with criticism of the use of literary composition as a means of testing candidates (Ch. 6, pp. 241–44). In 823, an examination of the Ch’un-ch’iu and all of the Three traditions was proposed, to redress the unpopularity of these texts among candidates.133 The establishment of this examination suggests that the eclectic approach that Chao K’uang and other members of the south-eastern Ch’un-ch’iu school had developed, and which was already reflected in examination questions in the late eighth and early ninth centuries,134 was now officially accepted. Later in the century, there were more attempts to reform the ming ching and chin shih, and these occasionally mentioned the place of canonical learning. The most radical proposal, operative for just one year after 833, involved Yen Hui did not contravene goodness’, and the Mencius 2A/2 remark that ‘at forty my mind was no longer moved’. Cf. Ch’üan Tsai-chih wen chi 37/10b–11a, in which he attributed importance to the latter phrase. It also figured in Li Wen kung chi 2/7b. 131  In 805, Ch’üan Te-yü’s ming ching Li chi dissertation question opened with the sentence, ‘The Ta hsüeh contains the way of illuminating virtue; the Chung yung has the techniques for exhausting fully the nature’; see Ch’üan Tsai-chih wen chi 40/17a. Ch’üan expected candidates to be ‘thoroughly versed in the principles of nature and destiny’; see 40/10a–b; but cf. his own understanding of hsing ming primarily as destiny, 30/8a–b. 132   Yüan shih Ch’ang-ch’ing chi 28/5a–b, decree examination answer of 806, containing one of the strongest denunciations of the knowledge of the canons required in the examinations. 133   t hy 76/1398; tfyk 640/17b–18b; cts 44/1892; cts 16/502; hts 44/1159 and 1161; des Rotours (1932), p. 29, p. 34, pp. 149–50, dates 822 or 823; tkck 19/28a–29a and note by Hsü Sung on p. 29a dates 823. 134   Lu Hsüan kung han-yüan chi 6/10a; wyyh 473/6b; ttclc 106/543; tkck 12/3b–4a. Ch’üan Te-yü’s extant ming ching dissertation questions on the Ch’un-ch’iu and the Three traditions, (for 802, see Ch’üan Tsai-chih wen chi 40/5a and 6b–7a; for 803, 40/11a and 13a; for 805, 40/16b and 18a–b), were for the Tso chuan and the Ku-liang chuan; the Kung-yang chuan is not represented. Ch’üan mentioned his own addiction to the Ch’un-ch’iu in Ch’üan Tsai-chih wen chi 30/1a–3a.

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

Canonical Scholarship

141

abolishing shih and fu in its tsa wen section (Ch. 6, p. 244).135 But the regular examinations ran on, much as before, until the very end of the dynasty; and it must be assumed that in the second half of the ninth century, in this area as in education more generally, the dynasty was no longer able to respond effectively to reformist suggestions. The Problem of the Definitive Text Successive attempts to require more meaningful knowledge of the Confucian canons in the examination system concentrated on adjusting curricula or on promoting less literary and more practical demands of the learning the canons contained. But the need, at an administrative level, to provide candidates with a definitive version of the text of the canons remained central, and continued to cause difficulties. The dynasty also recognized a related ideal, that the imperial library should house the definitive version of the canons. In contrast to so many of the problems that confronted T’ang official scholars, this found its enduring solution late in the post-rebellion period. Long before the T’ang, first in the Later Han and then under the San Kuo Wei, the text of the canons had been engraved in stone and displayed in the precincts of the T’ai hsüeh.136 The T’ang learned world considered that these early engravings constituted a precedent, and that their own dynasty should not only care for what remained of them, but also, in due time, erect their own monumental stone engravings of the texts of the canons. Under the Sui, in 586, the authorities had tried to bring the remains of the early engravings from Loyang to the capital. When T’ang scholars turned their attention to them, they expressed abhorrence because, in the disorders at the end of the Sui, parts of the engravings had been purloined for use as pillar bases.137 T’ai tsung’s great scholar minister Wei Cheng, as director of the imperial library, had collected what he could of the engraved stones. Rubbings taken from them were preserved in the imperial library. The dynasty also showed its reverence for the ideal of a stone engraving by including the Han engraving in the curriculum of the orthography school in the state academy directorate. The surviving stone

135   t kck 21/5a–b and note by Hsü Sung; cts 17B/551; cf. tfyk 641/4a–b; ttclc 29/106. 136  For the two early engravings, see Tsien Tsuen-hsuin (1962), pp. 73–9. 137   s s 75/1718, biog. of Liu Cho; ss 32/947; Feng shih wen chien chi 2/18–19. Sections of Ts’ai Yung’s engraving, however, remained buried at Lo-yang, and some were discovered later in the dynasty, to find their way into private possession; see Liu Pin-k’o chia hua lu, p. 9.

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

142

McMullen

fragments were considered to contain the authentic forms of characters for which variants had come to circulate.138 Early in the dynasty, T’ai tsung had attempted to solve the problem of a definitive version of the canons by commissioning Yen Shih-ku, whose ‘established texts’ had been completed in 633.139 Despite Yen’s separate list of orthodox-form characters, which circulated widely, this by no means ended the difficulties. Towards the end of the seventh century, there was a new attempt to produce an authoritative text of the canons, with the classical histories, the Shih chi and the Han shu included. In 696, a commission of scholars, including the long-serving official historian Wu Ching (c. 670–749), was convened in the imperial library. No result, however, is recorded.140 Then again, when Hsüan tsung was crown prince, Chang Yüeh had submitted a proposal to him asking that the text of the ‘nine canons and three histories’ be re-determined; but again there was no public outcome.141 In the K’ai-yüan and T’ien-pao periods, the questions again came to the fore. One official, an assistant instructor at the Kuo-tzu hsüeh and finally a vicepresident of the directorate, ‘regretted that the errors in the texts of the canons were so numerous and that successive generations had perpetuated them for so long’. He had been on the point of determining the text of the Three histories and the Five canons, ‘to have them displayed at the gate of the school’, when the death of his father intervened and forced him to leave office.142 Phonological glossaries compiled in the Chi-hsien yüan in the late K’ai-yüan period for the Mao shih, Ch’un-ch’iu and Taoist canon Chuang tzu may have been the result of the demand for definitive readings of these texts.143 In the T’ien-pao period, the writer of an important collection of reminiscences, Feng Yen, himself a directorate student, raised the issue of a T’ang engraving of the canons with the directorate’s erudits and students, again 138   Feng shih wen chien chi 2/18–9. For the place of the ‘canons in stone’ in the syllabus of the orthography school, see tlt 21/16a; also tkck 11/5a, Chang Shen. The entries in ss 32/945–6 and cts 46/1986–87 also indicate the importance the scholarly world attached to these engraved texts. 139  For Yen Shih-ku’s established versions’, see above at note 11. The language used to describe his commission deliberately echoes that of Ts’ai Yung’s biog. in Hou Han shu 60B/1990. 140   c ts 183/4728, biog. of Wu Min-chih; cts 191/5099, biog. of Li Ssu-chen confirms the dating. 141   Chang Yen kung chi 13/1a; wyyh 652/2a–b. 142   c tcc epitaph for K’ai Hsiu-yüan, died 733 aged 55 sui; Teng k’o chi k’ao pu p. 98 indicates that he was a ming ching of the year 700. 143   t hy 36/658; Ch’ü-chiang Chang hsien-sheng wen chi 15/1a–b; Yü hai 42/34a–b and 45/18b– 19a; Ikeda (1971), p. 67.

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

Canonical Scholarship

143

without success.144 In the event, under Hsüan tsung, only the text of the Hsiao ching was given a stone engraving, as a part of Hsüan tsung’s promotion of this concise summary of Confucian ethical norms. After the rebellion, both the need for a definitive version and the ideal of a monumental engraving to reflect the greatness of the dynasty became more acute. The famous seal script calligrapher Li Yang-ping (fl. c. 765–80) described how he yearned ‘to engrave stone in the seal script, to write out in full the Six canons and erect them in the Ming-t’ang, as an inerasable authority, to call them the T’ang canons in stone, so that a hundred ages hence no adjustments [in their texts] need be made’.145 At about this time, the text of the Ch’un-ch’iu, with exegesis by the military governor of Huai-nan, Han Huang, was engraved in stone and erected, probably at the Confucian temple at Jun-chou, modern Nanking. For reasons that are not clear, Han Huang’s Ch’un-ch’iu scholarship was ridiculed in the T’ang. But for the far south-east to have anticipated the capital, even on the small scale that this short canon involved, was an indication of the favourable conditions that obtained there.146 Han Huang was the political enemy of Chang Shen, a senior scholar at the Ch’ang-an state academy directorate. Perhaps because of this, his initiative in using stone at this time had no recorded parallel in the metropolis. It was Chang Shen, this same enemy of Han Huang, who in 776 led a small commission of scholars charged with producing definitive texts of the Five canons. On this occasion, the texts were written on wooden panels and set in the walls of the corridors of the east and west sides of the Kuo-tzu hsüeh. Chang Shen and another scholar also compiled a supplement giving the correct orthography for disputed characters. This supplement, displayed separately as

144   Feng shih wen chien chi 2/19; the account ends with a lament that an engraving in stone would have to wait ‘until the [Yellow] river flowed clear’. 145  For Li Yang-ping’s proposal, see twt 81/7a–b. This letter may be dated to the reign of Tai tsung, by Li Yang-ping’s reference to ‘eight generations’. Cf. wyyh 703/2b, Liang Su, referring to Te tsung as the ninth, and Liu Ho-tung chi 1/22, Liu Tsung-yüan referring to Shun tsung as the tenth emperor of the dynasty. In 780, Li Yang-ping was an assistant in the state academy directorate, see des Rotours (1975a), p. 92. 146   c ts 129/3602, biog. of Han Huang; wyyh 973/4b, report of conduct by Ku K’uang; wyyh 769/6a–b, Lu Kuei-meng; cf. T’ang yü lin 2/49. For the hostility between Han Huang and Chang Shen as a vice-president of the directorate, see cts 69/3445, biog. of Ch’ang Kun. For the suggestion that a work of Ku-liang exegesis be engraved in the south-east, see Ch’üan Tsai-chih wen chi 20/2b–3a.

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

144

McMullen

the Wu ching wen tzu (Model characters of the Five canons), was later engraved in stone and has been transmitted to modern times.147 Ten years later, in 786, another initiative, based on the imperial library and intended to produce a definitive text of the long series of nine canons was taken. But the director of the imperial library, Liu T’ai-chen (725–92), a former pupil of Hsiao Ying-shih, was unsuccessful in his request for a special commission to undertake the project.148 Only a few years after this, Han Hui, a brother of Han Huang, as director of the imperial library, made yet another attempt to redetermine the canonical texts.149 Under Hsien tsung, further efforts were made to arrive at definitive versions, and in the state academy directorate supplements giving correct orthography were produced, at least for the Li chi and the Ch’un-ch’iu.150 In the course of the reign, the Confucian reformer Cheng Yüch’ing drew attention to the deteriorating condition of the version displayed in the directorate. Over this period, the texts were probably replaced or restored for display there.151 It was not until the reign of the scholarly emperor Wen tsung, between 829 and 831, however, that proposals resulted in the complete renovation of

147   t kck 11/1a and 11/4b–6b, quoting a rubbing of the stone engraving of 837. This text praised Ts’ai Yung’s stone engraving version, and claimed to have used it in determining certain variants. The assistant was Yen Ch’uan-ching, a hsiao lien graduate, and possibly a member of the same clan as Yen Shih-ku and Yen Chen-ch’ing, both noted philologists. The list of standard characters that Chang Shen and he compiled was cited by Wang Ching in Ta T’ang chiao ssu lu 1/7a. An account of its transmission is given in skcstmty pp. 855–6. A fu on the theme of the version of the canons on the roofed wall in the state academy directorate is at wyyh 61/2b–3b, and would seem to describe the version done by Chang Shen’s commission. Another, wyyh 61/3b–4b, seems to refer to the K’ai-ch’eng engraving in stone. For Chang Shen, see also Kuo shih pu 3/54. 148   t hy 65/11246. Liu T’ai-chen’s spirit path stele by P’ei Tu, Ch’üan T’ang wen 538/14b, refers very briefly to this incident. It also shows that Liu held tenures typical of a post-rebellion academic official, in the court of sacrifices, as a court diarist, rescript writer and grand secretary and chief examiner, besides this tenure as director of the imperial library. 149   Ch’üan Tsai-chih wen chi 20/3a. 150  Two titles entered in the hts bibliography suggest an attempt to re-determine the text of the canons in 818; hts 57/1434, Ho chih p. 30, has a Li chi tzu li i-t’ung (Variants in the character standards in the Li chi), in one chüan, with an additional note, ‘determined on imperial order in 818’. hts 57/1441, Ho chih p. 43, enters a Ch’un-ch’iu chia chien (Additions and excisions from the Ch’un-ch’iu), again in one chüan, with the note, ‘prepared and determined by the state academy directorate in 818’. Chih chai shu lu chieh t’i 3/55 states that this was ‘in a similar category to the Wu ching wen tzu’. 151   t hy 66/1160, dated 12th month of 819.

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

Canonical Scholarship

145

these texts in wood.152 But the interest of the emperor and of Cheng T’an, his Confucian minister, went further than this. For it was in this period of relative weakness in T’ang power that the dynasty’s ambition to complete a monumental stone engraving was finally achieved. Cheng T’an asked that ‘the precedent of the Han be followed, and that stone be incised in the T’ai hsüeh, to display a model for ten thousand ages’. In 837, this stone engraving, of the long series of Nine canons, was completed and set up in the state academy directorate. The K’ai-ch’eng shih ching (K’ai-ch’eng canons in stone), as the engraving is known, are among the grandest examples of the highly developed art of stone engraving to have survived from the dynasty. But in some respects they caused disappointment. At the textual level, the drafting commission reverted to the choices made by K’ung Ying-ta at the start of the dynasty. The versions they chose were precisely those that, two centuries before, K’ung Ying-ta’s sub-commentary commission and their contemporary sub-commentators Chia Kung-yen and Yang Shih-hsün had endorsed. The only exceptions were the Hsiao ching, for which the text that Hsüan tsung had used for his commentary was used, and the Yüeh ling for which the text that Li Lin-fu’s commentary had glossed was adopted. Although in the stone engraving the names of these established commentaries were repeated at the head of each chüan, neither commentaries nor any other exegesis were included. Whether for this or for other reasons, the engraving was controversial, though few details of the ensuing debates survive. Yet this monumental achievement, as Wen tsung surely intended, preserved and displayed the texts of the canons for the state, until from the tenth century the widespread use of printing put the problems of circulation and transmission on a completely different footing.153 Post-rebellion Ch’un-ch’iu Scholarship These intermittent attempts at reform and renovation, however, took place in a scholarly climate very different from that of the early eighth century. Postrebellion re-interpretation of the canons was distinguished by the fact that 152  For the 831 renovation, see Liu Yü-hsi chi 8/73–74. Pien Hsiao-hsüan (1963), p. 70, dates this record to between 828 and 831. For the Wei Kung-su mentioned, see Ch. 4, note 193. 153  For the K’ai-ch’eng engraving in stone, see thy 66/1162; thy 77/1411; cts 17B/571; cts 173/4490–91, biog. of Cheng T’an, implying a date of 830 for the commission; tctc 245/7930, 837.12; ctcc epitaph for Chao ?-chih, d. early 835 aged 59 sui, states that as an erudit of the Li chi, he ‘concurrently directed the matter of the canons in stone’. Shih ching k’ao 2/1a–8a; skcstmty, pp. 856–7, describing the Chiu ching tzu-yang, by T’ang Yüan-tu, and its relation to the Wu ching wen-tzu of 776. Taga (1953), p. 263, called the K’ai-ch’eng engraving ‘an epitaph for the T’ang school system’.

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

146

McMullen

almost all of it was unofficial, by its dismissal of the compendious, commentarial approach to the text of the canons that had been officially in favour early in the dynasty, and by the keenness of the critical ideas that were now read out from the ancient Confucian texts. Post-rebellion Ch’un-ch’iu scholarship, particularly, contrasts with the official outlook of the early T’ang on the Ch’un-ch’iu and the Three traditions. In the central political arena, the Ch’unch’iu remained a sanction for the typical concerns of the T’ang scholar elite, for issues of ritual management, for a range of practical policy questions, and for principles of political morality. Emperors, ministers and officials continued to cite it with great frequency.154 But the major intellectual movement associated with this text was promoted by academic officials at Ch’ang-an only at the start of the ninth century, in the middle of the three phases of its history. Its members, moreover, went far beyond the disdain for the commentarial and subcommentarial approach to the canons that had become common in the eighth century. They saw the Ch’un-ch’iu not simply as an authority that would bring peace and stability to the world, but, in their metaphor, as a medicine that might effect an immediate cure on the body politic. They dismissed the official exegesis of the early T’ang, and rejected the traditional T’ang preference for the Tso chuan. They attempted to restore to Confucius the central authority that Liu Chih-chi’s attacks on the Ch’un-ch’iu had threatened to take from him. The Ch’un-ch’iu had always provided T’ang scholars with justification for the principles of government they advocated; but members of the post-rebellion school used the text to a new extent, to sanction a range of reformist ideas.155 For thirty years after the An Lu-shan rebellion, the far south-east, preserved from the worst effects of militarism, provided the locale for the Ch’un-ch’iu school. Three scholars, Tan Chu (d. 770), Chao K’uang (fl. c. 770) and Lu Ch’un (d. 805), all of whom served or retired in this region, followed one another as the main figures of the school.156 Tan Chu, originally a northerner who w ­ ithdrew

154  For post-rebellion emperors citing the Ch’un-ch’iu, see cts 10/262, Su tsung; ttclc 1/2 and thy 66/1156, Te tsung; thy 29/537 and tctc 236/7620, 805.28, Hsien tsung. Wen tsung had a scholarly interest in the Ch’un-ch’iu; see hts 200/5707, biog. of Shih Shih-kai; tfyk 40/28b, describing Wen tsung’s compilation Yü chi Ch’un-ch’iu Tso shih lieh kuo ching chuan (Imperially compiled canon and commentary for the various states [drawn from] the Ch’un-ch’iu and the Tso chuan). This work was referred to the history office in 835. 155  For the post-rebellion Ch’un-ch’iu school, see Pulleyblank (1960), pp. 88–91; and (1959), p. 147; Chang Ch’ün (1974), pp. 149–59; Yoshihara (1974), pp. 67–104; and (1976), pp. 633–53; Inaba Ichirō (1970), pp. 389–96. 156  For references to all three scholars, see cts 189B/4977, biog. of Lu Ch’un; Kuo shih pu 3/54; hts 168/5127–8; hts 200/5705–07. For some characterizations of the far south-east as an

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

Canonical Scholarship

147

to Tan-yang near modern Nanking, compiled two works on the canon.157 In these, he emphatically reasserted the belief that Confucius had edited the Ch’un-ch’iu with a far-reaching view of history and profound moral purpose. To Tan Chu, the canon could ‘remedy the decline of our age and reform the deterioration in ritual’.158 Discarding early T’ang official reverence for primary commentaries and sub-commentaries, he charged that too much attention had been paid to the Three traditions. Though its detail was useful, the Tso particularly had been over-rated, and had been understood as a literary and historical rather than a canonical text. Scholars of the Ch’un-ch’iu should, in the Analects phrase, ‘have no constant teacher’ from among the Three traditions. Still less should they adhere to later exegesis.159 The second member of the school, Chao K’uang, served as a staff officer in the south-east and met Tan Chu in 770, to be influenced by his ideas, probably through Tan’s son. Chao K’uang believed that Confucius, in editing the canon, had promoted ‘constant laws’, especially those to do with ‘ceremonial offerings, court visits, hunts and marriages’, and that these were true for all time.160 He also held, adopting a theme from early Kung-yang exegesis, that Confucius in the Ch’un-ch’iu had illustrated issues of conduct for which unchanging norms did not apply, and for which understanding of ‘expedient action’ (ch’üan) was necessary. This was a doctrine that Confucius had described in the Analects as particularly difficult.161 Expediency, however, like belief that institutions were always in a state of evolution, justified political toughness and interventionism. The reaffirmation of this idea may well be connected with the activist attitudes of some of the reformist scholars of Te tsung’s reign, including even Tu Yu, with whom the Ch’un-ch’iu scholars were probably connected. Chao’s insistence that certain rituals were specific to the son of heaven may be related to the reformers’ emphasis on centralized authority, and to their insistence that area of stability where, after the rebellion, scholars gathered, see twt 73/6b, Li Han; wyyh 988/5a–b, Lü Wen; Liu Yü-hsi chi 2/15. 157  For Lu Ch’un’s account of Tan Chu and Tan’s relationship with Chao K’uang and with himself, see Ch’un-ch’iu tsuan li 1/18a–19a. Tan Chu compiled two works, the Ch’un-ch’iu t’ung li (General principles of the Ch’un-ch’iu) in 3 chüan, and the Ch’un-ch’iu chi chuan (Collected commentaries to the Ch’un-ch’iu). According to Ch’un-ch’iu tsuan li 1/18b, his work was completed in 770. 158   Ch’un-ch’iu tsuan li 1/1b. 159   Ch’un-ch’iu tsuan li 1/5b–6a; for the idea of ‘no constant teacher’, Ch’un-ch’iu tsuan li 1/6a, see Analects 19/22. 160   Ch’un-ch’iu tsuan li 1/7a–b. 161   Ch’un-ch’iu tsuan li 1/7b; Analects 9/30. D. C. Lau, Confucius: the Analects, Harmondsworth, Middlesex: Penguin Books, 1979, p. 100, translates, ‘the exercise of moral discretion’.

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

148

McMullen

‘power should come from one outlet’.162 His belief also in the inevitability of institutional decline, justifying radical reform,163 and his downplaying of the place of omens in the Ch’un-ch’iu164 also coincided with the outlook of Tu Yu and the reformers. Chao K’uang continued Tan Chu’s efforts to demote the Tso chuan. He attacked as spurious the version of its transmission given by the Sui-T’ang exegete Lu Te-ming.165 In his boldest textual conclusion, he criticized Tan Chu for ‘following the old theory’ and identifying Tso Ch’iu-ming as its author. The Tso was a superficial and inaccurate work, and Confucius in the Analects had reserved praise only for those who had lived before him in time. The Tso Ch’iuming of the Analects must therefore have been a different, much earlier figure than the man who had given his name to the Tso chuan.166 Chao also argued that the Tso chuan and Kuo yü, both traditionally and officially ascribed to Tso Ch’iu-ming, were actually by different hands.167 Lu Ch’un, the third member of the school, like other scholars with reformist sympathies, came north to the capital, probably late in Tai tsung’s reign.168 After in 781 Yang Yen fell from power, he was banished to the south-east, mak162  See the two essays on the ti sacrifice by Chao K’uang, Ch’un-ch’iu tsuan li 2/13a–19a; also 2/12a. 163   Ch’un-ch’iu tsuan li 6/10b. 164  For the small space the Ch’un-ch’iu scholars gave to analysis of omens, see Ch’un-ch’iu tsuan li 6/10b–11a, 9/7a. 165   Ch’un-ch’iu tsuan li 1/12b, quoting Ching tien shih wen. 166   Ch’un-ch’iu tsuan li 1/10a–b. 167  Chao’s argument that the Tso chuan and the Kuo yü were by different hands was not new; see Tso chuan chu shu 59/5a and 37/9a, where it is given as the opinion of Fu Hsüan and of the Sui exegete Liu Hsüan. See Hung (1937), p. xlv. hts 200/5706, biog. of Tan Chu implies that it was Tan who argued that the Tso and the Kuo yü were by different authors; this is likely to be an error, in view of the above. 168  Lu Ch’un had an 11-year discipleship under Tan Chu; see Ch’un-ch’iu tsuan li 1/16b. Following Tan’s death, and with the help of his son, and in consultation with Chao K’uang, he copied out Tan’s work, completing an account of the school’s writings in 775. Lu Ch’un compiled three works relating to the Ch’un-ch’iu. The Ch’un-ch’iu chi chuan tsuan li (Compilation of the principles from collected commentaries to the Ch’un-ch’iu) in 10 chüan; see hts 57/1441, Ho chih p. 42; skcstmty p. 522; and Ku ching chieh hui han 18; the Ch’un-ch’iu wei chih (Concealed import of the Ch’un-ch’iu) in 3 chüan; cf. hts 57/1441, Ho chih p. 42, both of which read 2 chüan, and do not attribute the title to an author; skcstmty p. 523; and Ku ching chieh hui han 19; and Ch’un-ch’iu chi-chuan pien i (Resolution of doubts from collected commentaries to the Ch’un-ch’iu) in 10 chüan; see skcstmty p. 524 and Ku ching chieh hui han 20. hts 57/1441 and Ho chih p. 42 enter an unattributed Ch’un-ch’iu pien i in 7 chüan, and also a Lu Chih chi chu Ch’un-ch’iu in 20 chüan, which may derive from thy 36/660.

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

Canonical Scholarship

149

ing the journey there with the official historian Shen Chi-chi, a friend of Tu Yu. But he soon returned to the capital, to hold a variety of metropolitan scholarly posts, and to take part in some of the state ritual controversies of the Chenyüan period (Ch. 4, p. 153). Under Shun tsung, he was an elder scholar who combined commitment to reform with interest in the Ch’un-ch’iu.169 A number of younger scholars, Han T’ai, Han Yeh, Ling Chun, Lü Wen and P’ei Chin were in the group of which he was now the senior figure. But the most important younger member was Liu Tsung-yüan, who took up with the school probably in 803.170 Despite the fact that some of these scholars held academic office, the canonical scholarship of the school in this its metropolitan phase remained unofficial. In contrast to the debates in which Liu Chih-chi and other K’ai-yüan 169  The main sources for Lu Ch’un’s career are Liu Ho-tung chi 9/132–33 and 31/504–05; cts 189B/4977, hts 168/5127–28, biogs.; hts 200/5705–07. From these, from successive contributions he made to ritual controversies, and from other evidence, it appears he was recommended to the court and appointed to a junior post in the court of sacrifices probably between 775 and about 778 (Ch’un-ch’iu tsuan li 1/19a); then became an omissioner of the left (cts 189B/4977) in the period of Yang Yen’s financial reforms. He was banished to the south following Yang Yen’s downfall, making the journey with the official scholar Shen Chi-chi in 781 (T’ai-p’ing kuang chi 352/3697 and Dudbridge (1983), p. 61 and note 9). Later he became an erudit of the court of sacrifices, in 785 (Ta T’ang chiao ssu lu 4/8a, cf. cts 21/844); a supernumerary secretary to the board of punishments in 788 (Ta T’ang chiao ssu lu and hts 15/379); a secretary to the board of granaries, with concurrent ritual responsibilities in late 790 (cts 149/4032, biog. of Liu Mien); and a secretary of the left of the central department of the ministry of affairs of state (tso ssu lang chung) in 795 (thy 13/313; tt 50/290.3). He was demoted to be an erudit of the Kuo tzu hsüeh and held this post in 803 (see Liu Ho-tung chi 9/133, cts 189B/4977, Lü Ho-shu wen chi 4/4b–5b and thy 36/660 and note 171 below; also tctc 236/7603, 803.13). He then served as prefect in Hsinchou, modern Kiangsi, and T’ai-chou, modern Chekiang (cts 189B/4977 and Liu Ho-tung chi 9/133). When the reform movement gained power, he was recalled as grand secretary in the chancellery, and finally reader to the crown prince (Liu Ho-tung chi 9/133 and cts 189B/4977–8). Probably some time during the period 802–3, he may have corresponded with Ch’üan Te-yü on selection policy; see Ch. 6, note 213. 170   Liu Ho-tung chi 31/504–06; see also Liu Ho-tung chi 11/172–3, a slightly different grouping; cts 135/3736; tctc 236/7603, 803.13; tctc 231/7609–10, 805.6; also Pien Hsiao-hsüan, Liu Yü-hsi nien-p’u pp. 28–9. For Ling Shih-hsieh, see Liu Ho-tung chi 25/413; cf. 43/720–21 and cts 135/3736 for the reformer Ling Chun, who later wrote a Hou Han Ch’un-ch’iu. For Lü Wen as a pupil of Lu Ch’un, see Lü Ho-shu wen chi, preface by Liu Yü-hsi, 1b; Lü also wrote on the importance of the Ch’un-ch’iu to a relative; see Lü Ho-shu wen chi 3/1a–3a. For some other Ch’un-ch’iu specialists in this period, see cts 160/4209, biog. of Yü-wen Chi; cts 189B/4978, biog. of Feng K’ang and thy 36/660; hts 200/5708, biog. of Shih Shih-kai; Huang-fu Ch’ih-cheng wen chi 2/9a.

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

150

McMullen

scholars had taken part eight decades previously, the discussions of the school were not conducted at the emperor’s request or in an official institutional framework. But at some point towards the end of Te tsung’s reign, when Lu Ch’un had been demoted to be an erudit of the Kuo-tzu hsüeh, they presented his compilation of the school’s scholarship to the throne, with a memorial describing its merits.171 This point of contact between their unofficial scholarship and high authority, however, only underlines the differences between imperial direction of the learned world in the K’ai-yüan period and the early ninth century. Lü Wen, the author of the memorial of presentation, did not ask that the works of the group be established and taught in the schools; no discussion is recorded, and there is no record that the compilation was taken into the imperial library. At about this time, the dramatic political events of Shun tsung’s reign inter­ vened. For six months in 805, members of the movement briefly enjoyed access to political power. But there is little indication, in the records for 805, that its tenets were prominent in the programme that the reformers implemented. By the ninth month of 805, the reform movement had been broken up. Its senior members or sympathizers, like Tu Yu and even Lu Ch’un, avoided recrimination; but many of its junior followers were exiled, among them the Ch’un-ch’iu scholars Han Yeh, Han T’ai and Liu Tsung-yüan. The sequel to these reverses was that after 805 the Ch’un-ch’iu scholarship of the school was pursued among the exile group and their contacts, by letter and in the provinces. In the present picture too, it is Liu Tsung-yüan who was at the centre of this phase of the movement. Liu’s response to the three earlier leaders of the movement, Tan Chu, Chao K’uang and Lu Ch’un, even in his exile, was one of excitement and almost unqualified approval.172 Nearly a milennium and a half, he believed, had passed since Confucius had edited the 171  For the memorial of presentation, see Lü Ho-shu wen chi 4/4b–5b. This enters Lu Ch’un’s post as erudit of the Kuo-tzu hsüeh. thy 36/660, however, enters the work as by Lu Chih, secretary of the chancellery, probably a double error in that the character printed is the given name of Lu Hsüan kung and because Lu Ch’un was not yet promoted. For Lü Wen’s career, focussing on his T’u-fan mission of 804–5, on which he accompanied the official historian Chang Chien until Chang’s death, see cts 137/3769 and hts 160/4967, biogs. and Ogawa (1964), pp. 70–84. On this mission, Lü was nearly joined by another second generation disciple of Tan Chu, Tou Ch’ün; see cts 155/4120, biog. of Tou. 172  Liu quoted Lu Ch’un in his exile period: in Liu Ho-tung chi 31/505, commenting on Lu Ch’un’s gloss on the 11th year of duke Hsüan (Tso chuan chu shu 22/8a and 9b–10a), Liu remarked, ‘We see here the sage’s praising and blaming, bestowing and taking away, and wherein its appropriateness resides. This is what is called “defects and good points not being concealed”.’ Lu Ch’un, Ch’un-ch’iu wei chih 2/30b–32a, concludes with the same

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

Canonical Scholarship

151

canon, and ‘hundreds and thousands’ of scholars had wrangled over its meaning in essays, commentaries, sub-commentaries and treatises. But these three scholars had now at last realized its true meaning, so that ‘common men and small children could understand it’.173 Liu Tsung-yüan traditionally has a position as one of the leading agnostic intellectuals of the T’ang. But in his response to the new Ch’un-ch’iu scholarship, he perpetuated precisely the school’s paradoxical and basically conservative orientation. On the one hand, he adopted its critical approach, combining it with his own agnosticism in brief reviews of a number of early texts. On the other hand he held, no less than Tan, Chao and Lu had done, to the profundity and sanctity of the Ch’un-ch’iu. For all the critical boldness he showed, particularly in his extended analysis of the Kuo yü,174 he held back from a systematic application of his ideas to canonical literature itself. Elsewhere in his writing too, Liu expressed an attitude of profound respect for Confucius and for the integrity of the teaching contained in those texts, the Ch’un-ch’iu175 and phrase. For the problem of identifying Yüan of Jao-chou, the recipient of Liu’s letter, see Ts’en Chung-mien (1962b), pp. 409–10. 173   Liu Ho-tung chi 9/132. 174  For the Fei Kuo yü, see Liu Ho-tung chi 44–45/746–88; see also particularly Gentzler (1966) esp. pp. 184–6. At the close of the Fei Kuo yü, Liu Ho-tung chi 45/788, and in a letter to Lü Wen, Liu Ho-tung chi 31/506–7, Liu condemned the attitude that considered the Kuo yü close to the canons. Ironically, it was later classified as a work of Ch’un-ch’iu scholarship; see hts 57/1441 and Ho chih p. 43. 175   Liu Ho-tung chi 31/305; 44/767–68; Tso chuan chu shu 13/5a–8a (9th–10th year of duke Hsi). Hsün Hsi, adviser to duke Hsien of Chin, had allowed the duke to kill his heir apparent and set up another son, the child of his favourite concubine, in his place. Hsün Hsi had sworn to instruct his son and to remain true to him till death. After the duke’s death, however, this son was killed, and Hsün Hsi, far from dying with him, instituted yet another son as heir. Only when this son in turn was killed did Hsün Hsi die. In the Ch’un-ch’iu, the style used in narrating the death of Hsün Hsi was traditionally interpreted as meaning that he was a good man, rather than a bad one, for it conformed with the style used in two other episodes in which good advisers were murdered (Tso chuan chu shu 5/2a–b, 2nd year of duke Huan; Tso chuan chu shu 9/3a, 12th year of duke Chuang). To Liu Tsung-yüan, however, it was beyond doubt that Hsün Hsi was culpable. Confucius’s purpose in the Ch’un-ch’iu, in using a form of words that implied he was good, was rather ‘to stir up those who might be unable to die [in comparable circumstances]’. Confucius therefore drew attention, in isolation from the earlier facts of his career, to Hsün Hsi’s eventual death. In this, Confucius, in Liu’s view, was fulfilling the principle he had announced in Analects 7/29, that he ‘should not enquire into a man’s past’. Liu further suggested that Confucius used a similar pedagogic technique when in the Ch’un-ch’iu he charged that Chih, heir apparent of Hsü, had murdered his father, when all Chih had done was to fail to taste his

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

152

McMullen

Analects,176 closely associated with him. Despite his reputation therefore, he was less direct as a critic than Liu Chih-chi, whose trenchantly expressed ideas of nearly a century before had been such a catalyst to the Ch’un-ch’iu school. Liu Tsung-yüan, who died in the far south in 819, did not transmit the teaching of the Ch’un-ch’iu school to disciples of any importance.177 But other adherents of the school, including one or two in academic office, continued the tradition.178 Its eclectic approach now found general acceptance, to be endorsed by scholars such as the historian Liu K’o (d. c. 839),179 by the monitory official Yin Yu, who in 823 successfully proposed the examination in the Ch’unch’iu and all the Three traditions,180 and by the rescript writer Wei Piao-wei (c. 770–c. 829),181 the decree examination candidate Liu Fen (fl. 828)182 and

father’s medicine before him or prepared it badly (Tso chuan chu shu 48/12a–b, 19th year of duke Chao). 176   Liu Ho-tung chi 4/68–70. In the first part of a two section note on the Analects, Liu conceded that the text was written down finally probably only by Confucius’s disciples’ disciples. In the second part, he held that the anomalous opening to the final chapter (Analects 20/1), which gives the words of the emperor Yao, represented statements that Confucius constantly recited in his teachings and that his disciples, only partially understanding, had transmitted and reproduced when they compiled the text. Both parts of the note imply that the text of the Analects contained Confucius’s own teaching. For respect for the profundity of Confucius’s thought, see also Liu Ho-tung chi 16/297–98, and Ch. 4, note 243. 177  Even in the far south, however, Liu had met a Ch’un-ch’iu specialist; see Liu Ho-tung chi 5/76. 178  The most interesting instance is that of Tou Ch’ün, who was taught the Ch’un-ch’iu in the south-east by a disciple of Tan Chu. He compiled a work in 30 chüan which was twice presented on his behalf to Te tsung but not acknowledged. Liu Yü-hsi and Liu Tsung-yüan were said to have despised him (but cf. Liu Ho-tung chi 38/606–07), and he later avoided accompanying the official historian Chang Chien and the Ch’un-ch’iu scholar and Lu Ch’un disciple Lü Wen on their mission to the Tibetans; see cts 155/4120–1, biog. and Tou shih lien-chü chi p. 15a. For Ch’un-ch’iu scholarship in the middle and late Yüan-ho period, see also Han Ch’ang-li chi 6/34/85–86, Fan Shao-shu; Han Ch’ang-li chi 4/18/81–82, Yin Yu; Han Ch’ang-li chi 2/5/38–40, Lu T’ung. Inaba (1970), p. 383, notes that Lu T’ung was the author of a work called the Ch’un-ch’iu che wei (Selecting the subtle in the Ch’un-ch’iu) in 4 chüan. This is entered in Chün chai tu shu chih 1B/4a. 179   t wt 95/2b–3a, Pulleyblank (1959), pp. 147–9. 180  See above, note 133. 181  Wei Piao-wei was author of a work entitled Ch’un-ch’iu san chuan tsung li (Gathering of the principles of the Three traditions to the Ch’un-ch’iu); see hts 57/1441 and Ho chih p. 42. 182  For Liu Fen, see cts 190B/5064. A preface to Tu Yü’s Ch’un-ch’iu shih li is identified as by him in skcstmty pp. 520–2.

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

Canonical Scholarship

153

others. Even late in the dynasty, scholars such as Ch’en Yüeh,183 Ssu-k’ung T’u (837–908)184 and Lu Kuei-meng (d. c. 881)185 were in the same eclectic tradition. Yet another late T’ang scholar, P’i Jih-hsiu, a friend of Lu Kuei-meng and an ardent promoter of both Mencius and Han Yü, was important in Ch’un-ch’iu scholarship in a different way: he compiled a short collection of comments on the text that constitutes nothing less than a point by point refutation of the attack that Liu Chih-chi had made one hundred and fifty years before.186 It is no longer possible to relate the Ch’un-ch’iu scholarship of these late T’ang scholars to the political climate of their time. But they clearly promoted some of the same concerns that the post-rebellion school had formulated, an eclectic approach to the Three traditions, and a careful endorsement of the authority and sanctity of the canon and of Confucius as its editor. The hsing ming Scholars The change that took place in the intellectual milieu of post-rebellion T’ang China is particularly well demonstrated by the successive stages in the evolution of Confucian interest in interior or psychological questions. Like the renewed interest in the Ch’un-ch’iu, this development again took place largely in unofficial society; but almost all its leading proponents were examination graduates, and most held academic office at some point in their careers. The exponents of this hsing ming emphasis in its second, late eighthcentury, stage held a number of attitudes in common with the Ch’un-ch’iu school. Like the Ch’un-ch’iu scholars, they dismissed inherited exegesis and were impatient of mechanical commentarial approaches to the canons. Like them too, they focussed on a small number of canonical texts. Their underlying motive was also to restore the authority of the Confucian tradition, in this case in the face of Buddhist and Taoist intrusion. Their greatest contribution was their attempt to redefine answers to questions of ultimate value in exclusively 183  Ch’en Yüeh was the author of a work entitled Che chung Ch’un-ch’iu ( Judging appropriately between [the Three traditions to] the Ch’un-ch’iu), in 30 chüan, pref. in Ch’üan T’ang wen 829/3b–4b; see also hts 57/1441 and Ho chih p. 43 and Ssu-k’ung Piao-sheng wen chi 3/2b. 184   Ssu-k’ung Piao-sheng wen-chi 3/1a–2a. 185   w yyh 796/6a–b, endorsing a remark attributed to Wen-chung tzu, and attacking the Ch’un-ch’iu scholarship of Han Huang (cts 129/3603, biog. and hts 57/1441 and Ho chih p. 42) as it was preserved in a stone inscription at Jun-chou, probably in the Confucian temple there. Lu Kuei-meng and another late T’ang scholar, Ch’en Shang, discussed the idea of Liu Chih-chi that the Ch’un-ch’iu was an historical rather than a canonical text; see wyyh 690/10b–11b, and cf. T’ang yü lin 2/56–7. 186   P’i tzu wen sou 3/35; stts 14/2b–11a; Inaba (1970), pp. 389–90.

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

154

McMullen

Confucian terms. They brought an area of experience that scholars had hitherto considered largely a private matter briefly into the centre of the scholarly arena, and gave it new priority for the political and social world of their time. Despite the fact that they were, like most T’ang officials, ambitious for high office, they took their ideas to some extent out of the state-centred, official context that had shaped so much of the early T’ang outlook on the canons, and made them universal statements about man. At the level of the individual text, they contributed substantially to the long process the end result of which was that the Confucian canon was reorganized, and the Four books and their authors were given the official recognition that in their own day were given to the series of Nine canons. The early post-rebellion stage of this renewed interest among official scholars in the interior emphasis is not, however, as clearly documented as the corresponding stage in the Ch’un-ch’iu school. The background to the movement was the interest in seclusion, religious contemplation and the ‘relationship between heaven and man’ that had developed unofficially in the scholarly community before the rebellion. But again there was no sense in which this interest was restricted to those out of office or in low-level posts. Officials of any rank, whether in metropolitan or provincial posts, might pursue it, just as others, by involving traditional Confucian caution towards the supra-mundane, decline to do so. Interest in contemplation or self-cultivation therefore is likely to have been common at most levels of educated society in the post-rebellion period. Even the emperor Tai tsung was said to have been concerned to ‘effect the complete development of his nature’.187 Such high ranking post-rebellion scholars as the official historian, reformer of the examination system and grand secretary Yang Wan and the vice-president of the state academy directorate Chang Shen showed a suggestive commitment to contemplative goals, or ‘the complete development of their natures’. But again the private, unofficial character of their interest is indicated by the fact that it is recorded only incidentally, in a commemorative biography for a third official, Tu Ya (725–98, canonized Su).188 Another senior official who may have had an interest in interior questions was Ts’ui Yu-fu, briefly a chief minister under Te tsung; the preface to his collected

187   t t 105/549.2. 188   Ch’üan Tsai-chih wen chi 13/10b, spirit path stele for Tu Ya, who had been appreciated by Fang Kuan before the rebellion, and for whom cf. also cts 146/3962–64, biog. Another example of a high-ranking official with this interest is Ch’i K’ang; see Ch’üan Tsai-chih wen chi 14/1a.

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

Canonical Scholarship

155

works, by Ch’üan Te-yü, describes him as having realized ‘the sincerity and enlightenment of the Chung yung’.189 The south-eastern community of serving officials and refugee scholars, who provided the background for the first phase of the Ch’un-ch’iu school was also important in the development of the hsing ming emphasis. Chang I (d. 783),190 Li Hua (c. 710–c. 767)191 and Tu-ku Chi,192 all of whom held academic office at some point in their careers, for example, all served in the far south-east, and each demonstrated an interest in interior questions. Some of these scholars also showed interest in the Mencius, much as Yang Wan had done. There is no suggestion, however, that their approach was exclusively Confucian; rather they were committed to Buddhism or Taoism, and analysed their experience mainly in Buddhist or Taoist terms. In Te tsung’s reign, interest in the analysis of religious contemplation continued to be fashionable. Religious eclectism and a sustained debt to the south-east is typified by two later figures who came north to serve in the ritual and scholarly agencies at Ch’ang-an. Both were important because they demonstrate the combination of personal political influence and private commitment to religious ideals that marked the first phase of the hsing ming movement. Liang Su (753–93), the first of these two, was from 770 until Tu-ku’s death in 777, a follower of Tu-Ku Chi in the far south-east. Later, he was appointed a historian and Han-lin rescript writer, and was politically influential when in 792 Lu Hsüan kung directed the examinations. His interest in selfcultivation led him to write on meditation, from a lay Buddhist perspective.193 189   Ch’üan Tsai-chih wen chi 33/2a, preface to Ts’ui Yu-fu’s collected works. Ts’ui Yu-fu had a south-eastern background, in common with other scholars who developed an interest in self-cultivation; see Ebrey (1978), pp. 96–8. 190  For Chang I and his commentary on the Mencius, see cts 125/3545–6, biog.; thy 36/659; cf. also Ch’üan Tsai-chih wen chi 14/1b. 191   w yyh 820/4a–5a; wyyh 860/1a–10b, Li Hua; Fo tsu t’ung chi 7/189.1; P’i-ling chi 9/2b. Liang Su, in a sacrificial graveside prayer written on behalf of Tu-ku Chi for Li Hua, wyyh 982/2a, described him as ‘taking a leisured journey (yu yu, Mao shih chu shu 17D/1b) in the realm of hsing ming’. For Li Hua’s interest in Mencius, among other pre-Ch’in texts, see wyyh 742/4b. 192   P’i-ling chi 3/3a; 17/7b–8a; 9/1a–4a; also P’i-ling chi 20/6a for Tu-ku discussing ‘life and death’. Tu-ku was, however, aware of the Analects statement that Confucius was reluctant to discuss ming; see P’i-ling chi 20/5a and Analects 9/1. Cf. also Liu Ho-tung chi 8/115, report of conduct for Liu Hun, for the example of another scholar who declined to discuss such questions. 193  For Liang Su’s debt to Tu-ku Chi, see wyyh 737/8a–b; wyyh 703/4b–6b; wyyh 972/9b– 13a; wyyh 982/3a–4b. For Liang’s interest in ‘effecting the complete development of his

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

156

McMullen

He resisted appointment to Tu Yu’s staff in the south-east, and condemned interventionist political figures in history like Kuan Chung (6th century bc) and Chu-ko Liang (ad 181–234), ‘who concentrated on saving the ages [in which they lived] and called themselves beams and ridge-poles of the state’.194 Liang Su played an important role in the examination career of Han Yü, and Li Ao also acknowledged him.195 Ch’üan Te-yü, the second of these figures, is especially important because of the great range of his connections in the scholar community. He maintained a friendship with Tu Yu until well into Hsien tsung’s reign, when they were both elderly. He also knew the young reformer Liu Yü-hsi and had contact with Liu Tsung-yüan. His wide social circle suggests that neither interest in the hsing ming emphasis, nor indeed commitment to the sort of reform programme some of the Ch’un-ch’iu scholars supported, was an exclusively held position or a barrier to the kind of social activities that the T’ang scholar community valued. Ch’üan Te-yü knew Liang Su and perhaps also the young Han Yü in the south-east,196 and like Liang preferred capital tenures to service there. He held academic office, as an erudit of the court of sacrifices before 780, rescript writer for an exceptionally long period of nine years (Ch. 6, pp. 239–40), and a chief examiner in the period 802–5.197 In his leisure, Ch’üan engaged in meditation in a Taoist monastic setting. In his account of this experience, he suggested that the Confucian ideal of ‘sincerity and enlightenment’ (ch’eng ming), which derived from the Chung yung, was the analogue to both Buddhist and Taoist meditation ideals.198 He also used the term ‘restoring the nature’ ( fu hsing), of nature’ and his identification of this as a Buddhist process, see twt 61/5b–8a. Cf. Fo tsu t’ung chi 7/189.2. His disciple Lü Wen, an adherent of the Ch’un-ch’iu school, also spoke of a process of ‘effecting the complete development of the nature’, as, by implication, a nonConfucian process; see Lü Ho-shu wen chi 1/7a–b. Liu Yü-hsi, in commenting on the early intellectual life of Wei Ch’u-hou, an official scholar, said that he ‘exhausted the source of hsing and ming’, and linked this with his interest in Buddhism; see Liu Yü-hsi chi 19/164. 194   t wt 92/11a–b, pref. by Ts’ui Kung. 195  For Liang Su’s role in Han Yü’s examination career, see Han Ch’ang-li chi 4/17/76–77. 196  For Ch’üan Te-yü’s connection with Liang Su, see wyyh 966/2b–3b, epitaph for Ch’üan’s mother by Liang; also wyyh 983/3a–b, sacrificial graveside prayer by Ch’üan. For Ch’üan’s connection with the young Han Yü, see Ch’üan Tsai-chih wen chi, pu k’o 5a–b. The Chungling referred to is in Kiangsi, where Ch’üan held office in ‘early Chen-yüan’, that is in about 785, when Han Yü was 18 sui. For Ch’üan’s south-eastern background, see also e.g. Ch’üan Tsai-chih wen chi 39/3b–4b. 197   c ts 148/4002 biog. 198   Ch’üan Tsai-chih wen chi 1/11b–12a; cf. 45/8b, memorial of 797, for comparable vocabulary.

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

Canonical Scholarship

157

the contemplative process as early as 787, over a decade before Li Ao wrote his treatise on this topic and gave the term detailed analysis.199 A final figure in this first phase of the movement, again important as a political sponsor, was Lu San, also a south-easterner. Liang Su knew him,200 and Ch’üan Te-yü, who associated with him in the south-east,201 recommended in 780 that he take his own metropolitan post as erudit of the court of sacrifices.202 Lu San played a role in the early careers of the principal figures of the second stage of the hsing ming movement, Han Yü and Li Ao. Li Ao visited him in the south-east, and expounded his theory of the nature (hsing) to him. Lu San, Li Ao recorded, praised his insight in hyperbolic terms. In 802, Lu San assisted Ch’üan Te-yü, who was then chief examiner, and passed on to him the recommendations of Han Yü.203 When in 802 Lu San died, Ch’üan Te-yü wrote his epitaph,204 and Li Ao composed a tribute in which he likened Lu to Yen Hui, Tzu-hsia and Mencius, all as much exemplars for the hsing ming scholars as Kuan Chung and Chu-ko Liang were to the political activists.205 But although Li Ao described Lu San as ‘effecting the complete development of his nature’,206 there is little detailed indication of his position over the Confucian doctrinal problems that were now to become so important. The next phase of the hsing ming movement is marked by the survival of much more documentation, and by a sharp change in its orientation. Its outlook had been eclectic, accommodative and quietist. It had been an 199   Ch’üan Tsai-chih wen chi 32/8b, dated 781; and 32/6b, dated 787. 200   Ch’üan Tsai-chih wen chi 24/11a. When Han approached Lu San with recommendations for the chin shih, he stressed the precedent of Liang Su’s recommendations to Lu Hsüan kung, a point that was the more telling to Lu San because he had known Liang Su; see note 195. 201   Ch’üan Tsai-chih wen chi 36/8a–9a. 202   Ch’üan Tsai-chih wen chi 46/5b; cf. also Ch’üan Tsai-chih wen chi 36/8a–9a, written probably in 802, on the same or a similar occasion as that described in Han Ch’ang-li chi 5/19/6–7. 203   Li Wen kung chi 2/7b; also Li Wen kung chi 17/140b, and cf. Li Wen kung chi 13/110b–111b. For Han Yü’s connection with Lu San, see Han Ch’ang-li chi 4/17/75; cf. 3/11/68–69; 5/19/6–7. 204   Ch’üan Tsai-chih wen chi 24/10a–11b. 205   Li Wen kung chi 13/110b–111b. For Li Ao’s high estimate of Yen Hui, as the disciple of Confucius who attained most and who best understood his doctrines, see below note 210. For Han Yü’s emphasis on Yen, Han Ch’ang-li chi 5/19/21–22; also Chang Chi, letter to Han, in Han Ch’ang-li chi 4/14/38. References to Yen Hui and his inner life are frequent over this period; e.g. Ch’üan Tsai-chih wen chi 42/9b; 1/12a; also Ch. 2, note 255. 206   Li Wen kung chi 17/140b. Li used the same phrase in Li Wen kung chi 2/8a.

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

158

McMullen

uncontroversial exploration of an area of human experience that did not conflict with, but rather complemented, the official career and the political arena, an indication of a scholar’s seriousness as a human being but not much more. Now, however, it became more actively reformist, exclusively Confucian and strongly anti-Buddhist and anti-Taoist. This change not only caused dissention among its adherents; in a period when the emperor and the official community generally accepted eclectic attitudes to the three teachings,207 it also cost the movement the respectability it had previously enjoyed. Its more political demands were now promoted by a small group of scholars. These men tended to be either in low-rank academic or general service posts, or else on the periphery of official life. Their strong commitment to anti-Buddhism marked them as a minority, distinct from their contemporaries in the scholar community. The most important figures in this second phase of the movement, Han Yü and Li Ao, were, however, temperamentally very different. Though it can by no means have seemed to be the case at the time, the differences of position that developed between them on Confucian doctrinal questions were to have major historical importance. The issues were analyzed at an unofficial level; highly polished though their writings were, there is no indication that either intended them for high authority. Where for Li Ao the political aspect of their anti-Buddhism did not adversely affect his career, Han Yü at certain points in his career took major political risks in promoting his anti-Buddhist beliefs, even endangering his life for their sake. Han Yü and Li Ao had probably formulated their different notions of an exclusively Confucian doctrine of the nature before 801.208 Li Ao’s long three-part essay ‘On restoring the nature’ (Fu hsing shu) shows a familiarity with psychological analysis and quietist ideas derived from the supra-mundane faiths. Three features of the position he described in it proved unacceptable to Han Yü: first his universalism, his belief that man’s moral endowment did not vary with the individual;209 secondly his apparently negative view of a central psychological component, the feelings, which he believed clouded or obscured the nature. Finally, Li posited a process of introspection or of ‘effecting the complete development of the nature’ that involved refining away the feelings 207  Eclecticism was the imperial attitude in the final years of Te tsung’s reign; see Ch’üan Tsaichih wen chi 23/2b–3a and cts 135/3728–29, biog. of Wei Chü-mou. Ch’üan Tsai-chih wen chi 45/8b, memorial of 797, is likely to refer to the same event. 208  I am grateful to Professor T. H. Barrett for letting me see his detailed argument dating the composition of the Fu hsing shu to before 800 or 801. Neither Hanabusa (1964), p. 69, nor Hartman (1986), pp. 204–6, dates Han Yü’s Yüan hsing; but cf. Rideout (1948), pp. 406–7. 209   Li Wen kung chi 2/5a; 4/27a–b.

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

Canonical Scholarship

159

through contemplation, and that led to the state of ‘sincerity and enlightenment’ (ch’eng ming) of which the Chung yung had spoken. This was in some respects a goal comparable to that of enlightenment in Buddhism. But Li formulated it in terms drawn from the Confucian canons, and saw the favourite disciple of Confucius, Yen Hui, as its main exemplar.210 Han Yü, who had been Li Ao’s mentor in literary composition,211 wrote a series of briefer and less profound statements on the same general subject. Showing much less concern for the existential values that Li Ao discussed, he adopted a position sanctioned by the Analects and by the pre-Buddhist Confucian tradition. Man was endowed at birth with nature and with feelings in one of three grades. Though those in the middle category might move in either direction, the upper and lower categories were not capable of change.212 In contrast to Li Ao, therefore, Han Yü prescribed no process of self-cultivation by contemplation or of the refining away of the feelings. His perspective on the doctrinal questions Li Ao’s analysis raised was rather that of a social reformer, albeit one who took his stand outside the official institutions in which he served. He believed in the interior value of sincerity (ch’eng); but the concept of quietness (ching), which had strong Buddhist and Taoist overtones, and which was important to Li Ao, did not apparently figure in his analysis. The contrast in attitudes of Han Yü and Li Ao is highlighted by the different attitudes to the texts that, in the long term, they both did so much to promote. Both indicated their ultimate commitment to Confucian political and social ideals by citing the famous eightfold progression from the Ta hsüeh, from the well-ordered state to the cultivation of the individual self. For Li Ao, the starting point, the first of the eight steps, was the sentence, so celebrated in later, Neo-Confucian debate, ‘the extension of knowledge lies in letting things come’. This he saw as applying to the state of liberated quiescence, of ‘sincerity and enlightenment’ in the Chung yung phase that he used, in which the feelings were refined away. Han Yü, using the same passage, stipulated, not a state of quiescence, but a simple process of self-correcting: ‘What the ancients meant 210  For Li’s focus on Yen Hui, see Li Wen kung chi 1/2b–3a; 2/7a, quoting Analects 6/7 and 11/18 and Chou i chu shu 8/8a–b. 211   Liu Yü-hsi chi 19/166. 212   Han Ch’ang-li chi 3/11/59–67; Analects 17/2. This statement had general currency in T’ang: e.g. Han shu 63/2869 and commentary by Yen Shih-ku; thy 40/729, Liu Chih-chi; tt 13/73.1, Tu Yu; Liu Ho-tung chi 33/527–28; Han Ch’ang-li chi 2/4/23. It was even used with reference to the barbarians, who might, ‘with practice’, draw closer to the Chinese; see ttclc 128/689, edict of late 738, ordering them to ‘observe the ritual in the state academy directorate’. See also Po Hsiang-shan chi 7/45/30, Tse-lin 2; 7/46/57, Ts’e-lin 34.

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

160

McMullen

by rectifying the mind and making sincere the purpose was done by taking action’.213 Towards Mencius, however, their attitudes, while coinciding in some respects, contrasted even more in others. Li Ao’s endorsement of Mencius was particularly significant at the doctrinal level. He adopted both the Mencian idea that the nature was good, and also Mencian universalism, that is the idea, voiced by Yen Hui in the Mencius, that any man might become a sage like the emperor Shun.214 Han Yü’s perspective on Mencius, on the other hand, was ambivalent and even inconsistent. He agreed with Li Ao in seeing Mencius as a transmitter of Confucius’s own teaching, through Tseng Shen and Tzu-ssu, and he commended Mencius as ‘the purest of the pure’.215 He also made use of the Mencian idea that any man might, like Shun, become a sage. But he did so in the context of a tilt at what he saw as the prevalent custom of denigrating goodness in others, rather than in the context of a reformulation of contemplative values. He asked simply that men should improve themselves, and ‘by thinking day and night, do away with those points at which they did not resemble Shun’. At the doctrinal level, however, he rejected the Mencius as an authority, saying that Mencius, in characterizing the nature as wholly good ‘had found one [of the three categories] but missed [the other] two’. In his emphasis on ‘wide love’ (po ai), Han, ironically, showed more affinity for the ideas of Mo tzu, whom Mencius was generally praised for having attacked and refuted.216 It was a further irony therefore that Han Yü, Li Ao and others should have united in making Mencius a symbol for the anti-Buddhism of their movement. The view of Mencius here was, however, straightforward, unanalytical and commonplace. Mencius had promoted the teachings of Confucius, and resisted the heterodoxies of his own time, the philosophies of Yang Chu and Mo Ti.217 His achievement ‘was not less than that of [the sage emperor] Yü’.218 But Han Yü’s circle of followers took this conventional and widespread view of Mencius much further. Already in 795, Han Yü’s follower Chang Chi (c. 765– c. 830) associated Han with the figure of Mencius. In the late Chen-yüan period, 213   Li Wen kung chi 2/9a–b; Han Ch’ang-li chi 3/11/62. 214   Li Wen kung chi 2/11a; 2/7b. 215   Han Chang-li chi 3/11/63; 3/11/72–73. 216   Han Ch’ang-li chi 3/11/63–64. For this inconsistency, see the remark by Shao Po (fl. c. 1120) quoted in Ching i k’ao 232/1a–b. 217  E.g. wyyh 980/12b, Li Hua; Huang-fu Ch’ih-cheng wen chi 2/8a, of Ku K’uang. Even the Buddhists laid claim to this idea, see Fo tsu t’ung-chi 7/189.1, quoted by Hou Wai-lu (1959), p. 334. See also twt 86/1b, Wei Ch’u-hou. For further evidence of an interest in Mencius in this period, see Huang-fu Ch’ih-cheng wen chi 2/1a–2a. 218   Han Ch’ang-li chi 4/18/85, dated 820, Hanabusa (1964), p. 69.

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

Canonical Scholarship

161

Li Ao and Lu San made the identification more directly. As much as a quarter of a century later, in a sacrificial graveside prayer for Han, Li Ao repeated the idea that Han was the Mencius of their time.219 Han Yü, Li Ao and others did not, however, continue to debate Confucian doctrinal ideas on interior questions with their initial intensity for so long a period. Rather it was anti-Buddhism at a more practical and political level that marked the movement in the Yüan-ho period and that continued to give life to the Mencian parallel. In this phase of the movement, it was the combative Han Yü,220 rather than Li Ao, who took the bolder political action. In 819, particularly, he risked his life by submitting a memorial expressing sardonic contempt for the Buddhist religion and its attendant extravagances. Though this memorial did not present a profound or original case, it angered the emperor. While the Buddhist community ‘set up a delighted clamour and tapping’, Han Yü, saved from capital punishment only by the intercession of friends, was banished eight thousand li, to Ch’ao-chou in modern Fukien.221 After this, the Confucian inspired intellectual anti-Buddhism of the second stage of the hsing ming movement apparently very rarely resulted in political action. Only a small number of Han Yü’s followers, themselves obscurely placed enough for their attitudes to be of little political consequence, remained actively anti-Buddhist. When, three decades later, harsh measures were taken against the Buddhist church, the pressure came from a much higher level in the official hierarchy, and was justified mainly in Taoist rather than in Confucian terms.222 For the rest of the dynasty, Han Yü’s influence proved greater in literary style than in doctrine,223 and there was only fitful analysis by academic officials of the problem of the moral identity of the nature that he and Li Ao had raised.224 219  For Li Ao’s identification of Han Yü as a Mencius figure, see Li Wen kung chi 7/50a–b; and after Han’s death, 16/120a–b. For Lu San, Han Ch’ang-li chi 3/11/68–69, dated 802, Hanabusa (1964), p. 67. For Chang Chi’s promotion of Han Yü as a Mencius figure in 795, see Han Ch’ang-li chi 4/14/36. See also Ch’en Yin-k’o (1936), pp. 39–43. 220   c ts 171/4454, biog. of Li Han, describes Han Yü as ‘unyielding and censorious; cts 160/4207 describes Li Ao in much the same terms. For a provincial anti-Buddhist initiative by Li Ao, see Li Wen kung chi 10/80b–82a. 221   Huang-fu Ch’ih-cheng wen chi 2/6a–b and 6/5a–b. 222  Kenneth Ch’en (1956), pp. 67–105. For Li Te-yü’s role, see T’ang Ch’eng-yeh (1973), pp. 535–62. 223  For brief accounts of Han Yü’s disciples and followers, see Ch’ien Chi-po (1958), 5/77–113; Lo Lien-t’ien (1977), pp. 212–20. 224   Huang-fu Ch’ih-cheng wen chi 2/46–56; Hartman (1986), p. 206; Fan-ch’uan wen chi 6/106– 7, favouring Hsün tzu’s view. Huang-fu Yung, a chin shih who held senior academic office,

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

162

McMullen

The increased level of interest in the Mencius was, however, sustained. Han Yü continued to be linked with Mencius, for example by Tu Mu, Sun Ch’iao and another admirer, Lin Chien-yen.225 The late ninth-century scholar P’i Jih-hsiu combined respect for Mencius and Yen Hui, whom he saw as exemplars of selfcultivation, with fervent admiration for Han Yü.226 P’i’s friend Lu Kuei-meng also promoted Mencius, having studied the text, with the writings of Yang Hsiung and the canons, since boyhood.227 At the very close of the dynasty, too, a certain Sun Ho, a chin shih of 897, ‘loved the writings of Hsün tzu, Yang [Chu] and Mencius, and imitated Han [Yü] in his compositions’. He even styled himself Hsi-Han, to emphasize his dedication.228 These late T’ang acts of homage to Han Yü, however, were again socially and politically inconsequential. Han’s attempt to restore the dynasty and to regenerate society by promoting a renewed Confucianism, purged of Buddhist and Taoist intrusion, had failed, like the comparable efforts of the Ch’un-ch’iu scholars to cure the illness of the state. Han’s ideas had to wait for the re-assertion of strong dynastic power, under the Sung, to achieve official recognition.

wrote a work entitled Hsing yen (Statements on the nature) in 14 sections. Po Chü-i, his friend for 24 years, wrote of him, in Mencian language, that ‘his mind was not disturbed’; see Po Hsiang-shan chi 9/61/68. 225   Fan-ch’uan wen chi 6/9a–10b; Sun Ch’iao chi 2/7b; twt 86/6b–7a, Lin Chien-yen. 226   P’i tzu wen sou 9/95–6; 9/99–100; 3/23–4. 227   w yyh 690/10b, Lu Kuei-meng. 228  For Sun Ho, see T’ang shih chi shih 61/928; tkck 24/19a.

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

Neo-Confucianism and Neo-Legalism in T’ang Intellectual Life, 755–805 Edwin G. Pulleyblank The Yüan-ho period (ad 806–20) is looked upon as the second flowering of the T’ang dynasty, a time of partial recovery after the K’ai-yüan and T’ien-pao (713–56) glories had been cut short by the disastrous events of the rebellion of An Lu-shan (755–63). In particular it is the time associated with the ku-wen movement for the reform of the content and style of prose writing, and with the beginnings of the revival of Confucian philosophy which culminated in Sung Neo-Confucianism. If one were looking for the era of greatest and most significant intellectual activity, it would, however, be more appropriate to turn one’s attention to the preceding Chen-yüan period (785–805), for it was under the autocratic and miserly Te-tsung (779–805), one of the most individual and forceful of T’ang emperors, who was also one of the greatest patrons of literature of his line, that activity was most vigorous. A galaxy of young men, whose maturity came in the reign of his grandson Hsien-tsung (805–20), congregated in Ch’ang-an seeking fortunes by their talents and spending leisure hours in discussions on political and economic as well as ethical and metaphysical problems. In the fields of prose writing and philosophy, the best known names are those of Han Yü and Liu Tsung-yüan, but they are only two among a numerous crowd. Some, such as Liu Yü-hsi, Lü Wen, and Li Ao, left writings which have survived; others we know only through references in the histories or in the writings of friends. This is a focal period in Chinese intellectual history, for it represents the high point of enthusiasm and hope among the literati for a rejuvenation of the T’ang dynasty, before the successive setbacks and disappointments of the early ninth century brought disillusionment. During Te-tsung’s reign, in spite of the many unsolved problems that still beset the state, some degree of stability and order seemed to have been established, and there was hope that a genuine restoration might be possible. It was this that gave point to discussions on fundamental issues. New institutions were taking shape—a new basic system of taxation and other fiscal devices, Source: “Neo-Confucianism and Neo-Legalism in T’ang Intellectual Life, 755–805,” in Arthur F. Wright (ed.), The Confucian Persuasion, Stanford: Stanford University Press, 1960, 77–114. Copyright 1960, Board of Trustees of the Leland Stanford Jr. University, renewed 1988. Reprinted by permission of the publisher, Stanford University Press, sup. org.

© koninklijke brill nv, leiden, ���9 | doi:��.��63/9789004380158_007

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

164

Pulleyblank

modifications in the structure of the bureaucracy—which were to have a long history into and through the Sung period. There was plenty to stimulate and occupy the minds of would-be statesmen. A few of these men are well known. Han Yü and Li Ao have received attention mainly as precursors of the Sung Neo-Confucians, but judgments of them are mostly based on views of Ou-yang Hsiu or Chu Hsi. To understand them properly, it is necessary to put them in the context of their own time. Moreover, Confucianism was by no means the only philosophy in vogue. Throughout this period men were not overly concerned by the demands of orthodoxy; indeed Han Yü’s authoritarian exclusiveness, his one real claim to originality as a thinker, was quite exceptional. As always, men turned to the Classics for inspiration, but they turned also to Buddhism, Taoism, and the ancient Legalist writers. A new critical spirit was abroad which made men seek in the Classics for interpretations consonant with reason rather than merely consistent with the orthodox commentaries. It may be true to say that the culmination of all this came in Sung, but the immediate significance was for its own time and it is surely a mistake to look only at what the Sung thinkers three hundred years later found to be significant. I have attempted in what follows to outline in a preliminary way the important intellectual developments of the half-century following the An Lu-shan rebellion.

Consequences of the An Lu-Shan Rebellion

A brief account of the half-century from 755 to 805 will set the stage and bring out some of the main political and economic problems which, even when not overtly occupying the attention of writers, lay in the background of their thinking. The first impulse of the rebellious armies which moved rapidly down in the winter of 755–56 from the Peking region to take the Eastern Capital, Lo-yang, was checked for a while at the T’ung-kuan. Loyal uprisings along the rebels’ line of communications and the invasion of Hopei from Shansi through the T’u-men Pass by Kuo Tzu-i’s imperial army seemed on the way to defeating the rebels when, in the summer of 756, the collapse of the T’ung-kuan left the way to Ch’ang-an unguarded. Hsüan-tsung abandoned his court and fled westward. At Ma-wei post-station, west of Ch’ang-an, his mutinous bodyguard killed the Chief Minister Yang Kuo-chung and forced the emperor to put to death his favorite, Yang Kuo-chung’s cousin, Yang Kuei-fei. Hsüan-tsung then went on to Ch’eng-tu in Szechwan, but the Crown Prince proceeded northwest to Ling-wu

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

Neo-Confucianism in T ’ ang Intellectual Life

165

where he assumed the imperial dignity and began gathering forces to restore the fortunes of his house. Helped by an Uyghur contingent, and by the murder of An Lu-shan at the beginning of 757 which destroyed the coherence of the rebels, before the year’s end the new emperor Su-tsung recaptured first Ch’angan and then Lo-yang. The T’ang forces did not, however, have the strength to push their victory to completion, and the rebels were able to re-form, first under An Lu-shan’s son, An Ch’ing-hsü, then under his general, Shih Ssu-ming, and finally under Shih Ssu-ming’s son, Shih Ch’ao-i. For several years the war moved inconclusively back and forth in the Yellow River plain. Finally, at the end of 762, with the aid of a second Uyghur intervention, the T’ang armies were able to defeat Shih Ch’ao-i decisively. Abandoned by his supporters, he tried to flee northward to the land of the Khitans, but was treacherously slain by one of his own men. This marked the official end of the rebellion, but the country was far from pacified. On the west, the aggressive Tibetans had occupied large areas left undefended by the recall of frontier armies to fight the rebels. In 763 they even advanced on Ch’ang-an and forced the emperor to flee east to Shan-chou. They entered the city and attempted to set up a puppet emperor. Though they soon withdrew and did not repeat their past success, they remained a constant menace and large armies had to be stationed on the western frontier to guard against their raids. On the north, the Uyghurs were nominally the allies and restorers of the dynasty, but their friendship was almost as oppressive as the enmity of the Tibetans. Their soldiers disturbed the peace with impunity in Ch’ang-an, and foreign merchants under their protection evaded Chinese control.1 In the northeast, peace was only restored by allowing rebel generals to retain control of their territories as T’ang military governors. Although they were too undisciplined and divided among themselves again to menace the dynasty, they were effectively independent of T’ang authority and combined against any major attack by the central government. They ran their own internal affairs and paid no taxes to the center, except for occasional “tribute” to solicit favors. They were conscious of an analogy with the “feudalism” of the Ch’un-ch’iu (722–481 bc) period, and the language of feudalism was commonly used of them.2 1  h ts, ch. 217A, p. 7b; tctc, ch. 224 (Ta-li 7/1/chia-ch’en, 7/7/kuei-ssu, 8/7/hsin-ch’ou); ch. 225 (Ta-li 9/9/jen-yin, 10/9/wu-shen, 14/7 keng-ch’en). [Dates are here given as: Era name, year of the era, month of the year, followed by the cyclical term indicating the day. Ed.] 2  For example, see tctc, ch. 227 (Ch’ien-yüan 3/11/—), proposal for an alliance between certain of the Chieh-tu shih.

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

166

Pulleyblank

In large parts of the rest of China, conditions were not much more favorable for centralized control. Imperial armies raised to fight the rebels and now stationed throughout the country were almost as untrustworthy as the former rebels in Ho-pei. Generals attempted to emulate their fellows in Ho-pei and establish independent satrapies. Soldiers were unruly and might assassinate or drive out a commander who was too strict a disciplinarian or too niggardly with his largesse. The suppression of militarism and the restoration of civil government were vital tasks, but, although a certain measure of success was gradually achieved in part of the empire, the problem was not successfully solved until the Sung dynasty two hundred years later. The central government itself was in an unhealthy state. Great numbers of bureaucrats had either gone over to the rebels or been scattered far and wide, and Su-tsung, when he had retaken the capitals, had at first to make do with such officials as had followed him to Ling-wu. Later, others came from Hsüantsung’s court in Ch’eng-tu and elsewhere, but it is not surprising that Su-tsung relied heavily on those he knew best, his eunuchs. He had lived in seclusion at his father’s court, constantly afraid of being suspected of involvement in plots, real or imagined, against his father’s throne.3 He can have had little contact with the literati or indeed with any but men like Li Fu-kuo, the former eunuch stable boy4 who was at his side on the flight northwest from Ma-wei to Ling-wu and came to wield immense power over the court after the return to Ch’angan. Although Li Fu-kuo himself fell from power in 762, his place was taken by others. Eunuchs had previously been used as messengers between the inner palace and the ministers. Now this practice was institutionalized by the setting up of the Shu-mi Yüan, a eunuch secretariat which gradually assumed advisory and ultimately executive functions.5 If civil officials were mistrusted, the military were under even graver suspicion, and the emperors turned naturally to eunuchs to act as their eyes and ears in keeping watch. The system of eunuch supervisors (chien-chün)6 was extended to all armies under imperial authority, both in the provinces and at the capital. In the course of time, eunuch supervisors supplanted altogether 3  Cf. E. G. Pulleyblank, The Background of the Rebellion of An Lu-shan (London Oriental Series, Vol. 4, 1955), p. 100. 4  Ibid., p. 144, n. 24. 5  On the origins of the Shu-mi Yüan see Sun Kuo-tung, “T’ang tai san sheng chih chih fa-chan yen-chiu” (“Development of San Sheng System of the T’ang Dynasty”), Hsin-ya hsüeh-pao, iii (1957), 112–16. 6  On the beginnings of the chien-chün system see The Background of the Rebellion of An Lushan, p. 74.

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

Neo-Confucianism in T ’ ang Intellectual Life

167

the regular commanders in the Palace Armies, and this was the basis for their dominance over the court, which was to last until the end of the dynasty. In the provinces the friction and suspicion between eunuch supervisors and generals were frequently a source of trouble and aggravated the problem of militarism. The rebellion resulted in great financial difficulties for the central government. Cut off from most of its tax-producing area, Su-tsung’s court struggled painfully to gather resources for war. Financial exhaustion prevented the mustering of effort necessary to bring the war to a close and this in turn increased the financial exhaustion. A series of desperate fiscal expedients including forced loans from merchants,7 sale of offices and Buddhist and Taoist ordination certificates,8 and the minting of token coinage9 was resorted to. When peace nominally came in 762, most of North China was in ruins and still overrun with armies. Furthermore, the loss of life and the migrations caused by the war had completely upset the local registers of population, which were the basis of taxation. The south was relatively unscathed, but even the prosperous Chiang-nan area had felt the ravages of war in 760–61 when a dissident general from Huaihsi (northern Anhwei) invaded the southern Kiangsu area.10 He was subdued only by armies from the north, and these caused at least as much damage by pillaging as the rebels. In 762, when communications with the capital were somewhat improved, Yüan Tsai, as tax commissioner, tried to collect eight years of back taxes in Chiang-nan and Huai-nan, and to register the many fugitives who had invaded those regions. These harsh measures led to peasant uprisings, which once again required the importing of armies from the north to suppress them.11 In 762, Liu Yen was appointed Commissioner for Revenue, Salt and Iron, Transport and Taxation for the eastern half of the empire, a post which he held with short intervals until 780. He was undoubtedly a genius. By a tactful and farsighted policy he was able to encourage the economic recovery of the areas under his authority and, at the same time, to make their tax yield the financial 7  tt, ch. 11, p. 63.1. 8  See J. Gernet, Les aspects économiques du bouddhisme dans la société chinoise du Ve au Xe siècle (Publications de l’Ecole Française d’Extrême-Orient, Saigon, 1956), pp. 50–51, and E. G. Pulleyblank in Journal of the Economic and Social History of the Orient, I (1957), 155. 9  S. Balazs, in “Beiträge zur Wirtschaftsgeschichte der T’angzeit,” Part 2, msos, xxxv (1932), 30. Note that the rebels also were forced to the expedient of issuing token coinage at the ratio of 1:100 ordinary cash. hts, ch. 54, p. 7a. 10   t ctc, ch. 222 (Shang-yüan 2/1/—). 11   t ctc, ch. 222 (Pao-ying 1/chien-yin month/wu-shen, 1/8/chi-ssu); cts, ch. 11, p. 2b; ch. 110, p. 5a.

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

168

Pulleyblank

foundation of the T’ang government. He did this, first, by restoring the transport canal system from the Yangtze to Lo-yang and Ch’ang-an and, second, by applying and perfecting the salt monopoly, which in 758 had been introduced as a war measure by Ti-wu Ch’i. He paid less attention to the registration of the people and the restoration of direct taxes. Much of his success was due to his skill as an administrator. He was meticulous in his own work and managed to get the same standard from his subordinates. Many of the financial experts of the latter part of the century were men whom he had recruited into his service and trained.12 The reform of direct taxation had to wait till the beginning of Te-tsung’s reign in 780, when Yang Yen, Liu Yen’s great rival, introduced his Two Tax (liang-shui) reform. This abolished the old system of taxation in kind, which had been levied on a head tax basis in connection with the Equal Field (chünt’ien) system of land-holding and in accordance with registration of households in the districts of their origin. Instead, all direct taxes were calculated in money and grain, and were based on household classification by wealth and on the land actually under cultivation. Moreover, tax quotas were established for each locality on the basis of past receipts, and no effort was made to levy taxes, as previously, at standard rates throughout the empire. This reform finally recognized the decrepitude of the old system and was a rationalization of practices which had grown up in chaotic fashion during its decay. In turn, it soon gave rise to evils and abuses that called forth much criticism.13 Liu Yen and Yang Yen were personal enemies. Both had originally been friends of Yüan Tsai, the brilliant but corrupt Chief Minister of Tai-tsung, but Liu was instrumental in Yüan’s downfall in 777. Yang Yen in turn brought about Liu Yen’s banishment and death in 780. Lu Ch’i, who caused Yang Yen’s defeat and death soon after, was reputedly taking revenge for Liu Yen.14 It has been suggested that the feud arose from differences in economic theory reflected 12  Biographies of Liu Yen are in cts, ch. 123, hts, ch. 149. Cf. also the eulogy by Ch’en Chien, “Liu Yen lun,” in ctw, ch. 684. Chü Ch’ing-yüan, Liu Yen p’ing-chuan (“A Critical biography of Liu Yen”) (Shanghai, 1937), is an important study of his economic policies. See also D. C. Twitchett, “The Salt Commissioners after An Lu-shan’s Rebellion,” Asia Major iv (1954), 64ff. 13  Cf. Balazs, op. cit., Part 1; msos, xxxiv (1931), 82–92. Chü Ch’ing-yüan, T’ang tai ts’ai-cheng shih (“History of Government Finance in the T’ang Dynasty”), 1940 (Japanese translation by S. Nakajima, Tokyo, 1944) pp. 49–94; T’ao Hsi-sheng and Chü Ch’ing-yüan, T’ang tai ching-chi shih (“Economic History of the T’ang Dynasty”) (Shanghai, 1936), pp. 152ff. A number of studies of this reform have recently been published by Hino Kaisaburo. 14  Cf. Wei Ch’u-hou, “Ch’ing ming ch’a Li Feng-chi p’eng-tang shu” (“Memorial Requesting the Investigation of the Faction of Li Feng-chi”), ctw, ch. 716.

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

Neo-Confucianism in T ’ ang Intellectual Life

169

in Liu Yen’s use of indirect taxes and development of regional economy as opposed to Yang Yen’s emphasis on direct taxation and centralized control.15 Although personalities and factional rivalries must be given a large share of the blame, especially for the later repercussions of the feud, there does seem to have been a divergence of viewpoint, as we shall see when considering the economic ideas of Lu Hsüan-kung and Tu Yu. There were of course other problems of government in the reigns of Taitsung and Te-tsung besides fissiparous militarism, eunuch influence, and economic policy. The minds of the literati were exercised particularly by the examination system and by the growing complexities of the bureaucratic system in which they found themselves enmeshed. In this period we also perceive the beginnings of a number of institutions which did not form part of the regular T’ang system; for example, a formalized system of recommendation and sponsorship as a supplement to other methods of appointing officials.16 The Han-lin Academy is no longer—as it was under Hsüan-tsung—simply an office within the palace where the emperor kept such experts in chess, calligraphy, juggling, or literary composition as he might wish to have close at hand. Rather, it becomes a small informal body of confidential advisers who sometimes wield greater influence than the titular Chief Ministers.17 Its development, like that of the eunuch secretariat, follows a pattern occurring frequently in Chinese institutional history, whereby, when there is antagonism or lack of confidence between the emperor and the bureaucracy, new informal bodies of advisers appear close to the throne, only in turn to be institutionalized and lose their intimate character. This, without going into details, is a sketch of some of the main developments in the second half of the eighth century. Now let us turn to some of the thinkers and writers of this period.

Refugee Intellectuals

From the time An Lu-shan unleashed his armies on North China and took Ch’ang-an, the great clans and lesser houses of the northeast (Kuan-tung) were thrown into confusion. Some individuals made their way to the courts of 15  Kanai Yukitada, Tōdai no shigaku shisō (“T’ang Dynasty Historiographical Thought”) (Tokyo, 1940), pp. 105ff. Cf. Chü Ch’ing-yüan, Liu Yen p’ing-chuan. 16  See below. 17  Cf. Li Chao, Han-lin chih, in Han-yüan ch’ün shu, 2 ch. (Chih-pu-tsu chai ts’ung-shu); Sun Kuo-tung, “T’ang tai san sheng chih chih fa-chan yen-chiu,” as cited in note 5, pp. 108–12.

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

170

Pulleyblank

Hsüan-tsung at Ch’eng-tu, or to that of Su-tsung at Ling-wu. Some organized resistance in the rear of the rebels. Others fell into rebel hands and collaborated willingly or unwillingly with their captors. Still others took refuge in the hills. But as time went on and prospects of peace and order returning to the Yellow River plain became more remote, more and more made their way south to the peaceful lands across the Yangtze, as in the Yung-chia period (307–12) when North China fell to the barbarians.18 There is no way of estimating the size of this movement, but it is clear from the frequency with which it is referred to in biographies that it must have involved very considerable numbers. The common people were on the move too, but how many of them, as distinct from the great clans and their adherents, were able to escape the ravages of war, famine, and pestilence, and arrive safely in the south can only be conjectured. The immigrants settled in various places in central China. One of the principal escape routes was through Nan-yang and the Han River, and some stopped at places in Hu-pei such as the present Hankow. Others went farther east and south into Kiangsi, and many more came, whether by that route or more directly, to the ancient cities of the Wu-Yüeh region—Jun-chou (Chen-chiang) Su-chou, Ch’ang-chou, Hang-chou, etc. We can, today, only speculate on the economic aspects of this movement. It is probable that many of the clans were able to bring some part of their wealth with them and to use it to establish themselves in the south, which was at the beginning of its great expansion. Others may have been able to return later and recover their property in the north. For very many, however, it must have meant impoverishment. What is evident is that for some time there was a large aggregation of unemployed educated men in the south unable because of the war to find places in the regular bureaucracy. They were frequently recruited to the staffs of local military governors or to the newly established Salt Commission. Among these displaced literati we find the beginnings of the intellectual movements that reach their high point at the end of the century. The rebellion itself and the challenges it presented were no doubt directly responsible for one of the most striking characteristics of post-rebellion thought in contrast to the immediately preceding period—its seriousness, its involvement in the world, its concern, even when dealing with theoretical problems, for application to the present day—its activist (yu-wei) tone. In the fetid political atmosphere prevailing before the rebellion, characterized by the oppressive dominance of Li Lin-fu and by the dangerous power struggle 18  Besides the examples referred to in this article, there are very many more in biographies, epitaphs, etc., of the period and after. There are also references of a more general kind to the migration. Cf. hts, ch. 194, biography of Ch’üan Kao.

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

Neo-Confucianism in T ’ ang Intellectual Life

171

between Yang Kuo-chung and An Lu-shan, men of character had tended to seek to maintain their personal integrity through withdrawal from the world or, at least, in the individualistic cultivation of such things as poetry, fine writing, or philosophy. After 755, we seem to find a new spirit of social concern— even among those of a Taoist turn of mind. There is, for example, Yüan Chieh (719–72), an aristocrat who before the rebellion had held aloof from political life and written satirical attacks from a Taoist point of view on the world’s degenerate state. When the rebels took Loyang, he was first forced to flee with his family to the south and was then drawn willy-nilly into official employment.19 Or we may think of the Taoist writer, Liu Shih, who managed to spend the war years in retirement in the region of Nanch’ang (Kiangsi). Even he was acutely aware of the troubles of the time, and he wrote not only about the pleasures of his country retreat, but also about the sufferings of the peasants sold into slavery because of want, an evil which seems to have touched the consciences of scholar-officials from this time onward in a way that it had not done previously.20 Similar patterns can be found in other biographies of the period. More common perhaps were the men who had held office or would have looked naturally to an official career, but were cast adrift and given both the leisure and the stimulus to think about the problems of their time. For some the reaction was shock and pain—as in Tu Fu’s poems in which private griefs are universalized to express the torment of society. But in the comparative quiet and safety of the lower Yangtze region, groups of young men soon began to congregate, interspersing their drinking parties with discussions on how to reform the world, and scholars began to seek in the Classics and ancient philosophers, or in a revitalized literature, ways to bring order, peace, and moral regeneration.

19  Biography in hts, ch. 143. See also the excellent Yüan Tz’u-shan nien-p’u, by Sun Wang, first published in Chin-ling ta-hsüeh wen-hsüeh yüan chi-k’an in 1935, revised 1948 and reprinted in Shanghai, 1957. 20   h ts, ch. 143. Eight pieces by him are included in ctw, ch. 377. Note especially his “Ts’ao t’ang chi” (“Record of the Thatched Hall”) written in 767, in which he says, “I have made my home above the Hsiu River (near Hsiu-shui hsien in Kiangsi) for ten years. This land is distant and remote; war has not reached it. Yet I still daily see the wrong of poor people being sold …” The freeing of people sold into bondage is a meritorious act recorded of several persons sent to govern southern prefectures at the end of the eighth century and the beginning of the ninth, including both Han Yü and Liu Tsung-yüan. There were, moreover, imperial decrees trying to stop the practice. Previously traffic in “southerners” (nan-k’ou) seems to have been regarded as quite legitimate. It is striking that this antislavery movement should have coincided with the Confucian revival.

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

172

Pulleyblank

For instance, there were the so-called “Four K’uei” at Shang-yüan (present Nanking), one of whom was Han Hui, the elder brother of Han Yü.21 Han Yü’s father, Han Chung-ch’ing, happened to be a magistrate of Wu-ch’ang when the rebellion broke out, so he had no need to flee from the north. When, in 757, he was about to be transferred to the magistracy of P’o-yang, the poet, Li Po, who had been in that region for several years, composed an inscription praising his government of Wu-ch’ang.22 What happened to him thereafter until his death in 77023 is not clear. The Han family later possessed an estate (pieh-yeh) at Hsüan-ch’eng.24 Judging by the analogies in other family histories of the time, I suspect that it may have been acquired by Han Chung-ch’ing during the rebellion period. At any rate, the eldest son, Han Hui, and his companions, Ts’ui Tsao, Lu Tung-mei, and Chang Cheng-tse, are said to have been sojourning at Shang-yüan in the Yung-t’ai period (765) where they “loved to discuss plans for settling the affairs of the world (ching-chi) and considered themselves capable of serving as ministers to princes.” Unfortunately no writings have survived to tell what sort of plans they discussed. But Han Hui was later associated, along 21   c ts, ch. 130, biography of Ts’ui Tsao. Cf. also Han Yü’s epitaph of Lu Tung-mei in Han Ch’ang-li chi, ch. 24. The k’uei is best known as a mythological one-legged creature mentioned in Chuang-tzu, the Shan-hai ching, etc., but it is also the name of one of the ministers of Yao and it is in this sense that it was applied to Han Hui and his companions. Cf. “Yao tien,” ch. 35, “The emperor said: K’uei, I charge you to be Director of Music, to teach the descendant sons, [to be] [apprehensive:] careful, [hard:] firm and yet not tyrannical, great and yet not arrogant. Poetry expresses the mind …” Karlgren, Book of Documents (Stockholm, 1950), p. 7: The appellation thus aptly fits the ambition to serve in ministerial capacity and the ideal of using literature as a means of reforming the world. T’ang Kuoshih pu, ch. 3, p. 58, refers to Han Hui’s singing ability as a reason for the sobriquet, but this may well be anecdotal invention. T’ang chih-yen, ch. 4, ascribes the name “Four K’uei” to four other persons, one of whom was Li Hua. Chu Hsi, no doubt rightly, rejects this (K’ao-i to Han Ch’ang-li chi, loc. cit.), but it shows that the name was current. 22   Li T’ai-po ch’üan-chi (Chung-hua Press, Peking, 1957), p. 29. 23  Han Yü, who was born in 768, lost his father at the age of three (sui), therefore in 770. See his biography with commentary by Chu Hsi appended to Han Ch’ang-li chi. 24  The earliest express statement of this seems to be by Hung Hsing-tsu in his Han-tzu nienp’u (Han Wen lei p’u, ch. 3, p. 9a, in Han Liu nien-p’u, 8 ch.) reprint of Hsiao Ling-lung shan kuan (1875). It is not clear whether this is based on an unnamed separate source of information or whether it is inferred from references to his living in Chiang-nan with his sister-in-law around 781 and to members of his family living there around 820. Even if only the latter, it seems very probable. See the commentary on “Shih Shuang shih” (“Poem Shown to [Han] Shuang”) in Han Ch’ang-li shih hsi-nien chi-shih, xii, 565–66 (Shanghai, 1957), edited by Ch’ien Chung-lien.

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

Neo-Confucianism in T ’ ang Intellectual Life

173

with Yang Yen, in the regime of the Chief Minister, Yüan Tsai, in whose downfall they were implicated in 777;25 Ts’ui Tsao was a close friend of Liu Yen and was himself later a financier.26 It is therefore not difficult to suppose that the ching-chi discussed by the “Four K’uei” may have had its modern meaning of “economics.” In addition, Han Hui was deeply interested in another topic that was in the air, the reform of literature, an interest he was to pass on to his younger brother.27

25   t ctc, ch. 225 (Ta-li 12/4/kuei-wei). Cf. Hung Hsing-tsu, Hantzu nien-p’u in Han Wen lei-p’u, ch. 3, p. 7a. 26   c ts, ch. 130, biography of Ts’ui Tsao. Cf. Kanai Yukitada, Tōdai no shigaku shisō, p. 105. 27  Han Hui wrote a piece called the “Wen heng” (“Balance of Literature”), which is quoted in extenso in a short biography of him written by the Sung writer Wang Chih. It shows him to have been an advocate of ku-wen in the same extreme way that we find in Li Hua. That is, he considers that the decline in letters set in with Chuang-tzu and Lao-tzu in philosophy, with Ch’ü Yüan and Sung Yü in poetry, and with Ssu-ma Ch’ien and Pan Ku in narrative. Wang Chih remarks that it showed how much Han Yü resembled his elder brother. Recently Ch’ien Mu has quoted this to show that Han Yü derived his interest in ku-wen from his own family, not from Tu-ku Chi and Liang Su as stated in his biography, a fact which would explain the absence of references to Tu-ku and Liang in his works. It seems clear that Han Yü did look back to his brother and uncle for inspiration, but it is also true that, as one would expect, both in philosophy and in ideas on literature he shows the influence of the evolution in thinking that had taken place through Tu-ku Chi, Liang Su, et al. For Han Yü accepted all Chou and early Han literature and held the western Han writers Yang Hsiung and Ssu-ma Hsiang-ju in particularly high regard; and though his rejection of Buddhism represents a sharp contrast to Liang Su—and is probably sufficient to account for his failure to say much about him—his emphasis on the Chung-yung, Tahsüeh, and I-ching is in the Liang Su tradition. Cf. Ch’ien Mu, “Tsa-lun T’ang tai ku-wen yün-tung” (“The Classical Prose Reform Movement in the T’ang Dynasty”), Hsin-ya hsüehpao, iii (1957), 123–68.  Wang Chih’s Han Hui chuan originally formed part of the Han Wen lei-p’u, in 10 chüan, edited by Wei Chung-chü and appended to the Hsin-k’an wu-pai-chia chu yin-pien Ch’angli Hsien-sheng wen-chi, published in 1200. Later reprints of this edition omit the supplementary material, and the separate edition of the Han Wen lei-p’u in 7 chüan, forming the major part of the Han Liu nien-pu, 8 chüan, omits the Han Hui chuan. (The Han Liu nien-p’u was separately constituted in 1730 by Ch’en Ching-yün. I have consulted it in the reprint of the Hsiao Ling-lung shan kuan, 1875.) A facsimile reproduction of the edition of Han Yü’s works of the year 1200 was published by the Commercial Press in 1912, but I have been unable to consult it. The Han Hui chuan is, however, also included in Ch’üan T’ang wen chi-shih, ch. 39, pp. 10a ff. Cf. also Kung Shu-chih, Han Yü yü ch’i ku-wen yün-tung (“Han Yü and His ‘Ancient Style’ Movement”) (Chung-king, 1945), p. 14.

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

174

Pulleyblank

The Ku-Wen Movement

The best known intellectual movement of the second half of the eighth century is undoubtedly the ku-wen movement for the reform of prose style. There had been criticisms of the empty euphuism of the parallel-prose style even before the beginning of the T’ang dynasty and advocacy of a return to a simpler style modeled on Chou and Han writers;28 but it is only in the post-rebellion period that we find these ideas taking on the character of a movement. Some northern scholars exiled in Chiang-nan formulated the principles of literary composition and stressed the importance of a reformed prose for moral regeneration and the revival of the Confucian tradition which Han Yü developed and propagandized with a brillance that left his forerunners in the shade. The central figure in the initial period was Li Hua, already a celebrated writer in the reign of Hsüan-tsung. He had held a number of not very high offices, and when the rebellion broke out he was Omissioner of the Left. In the summer of 756, the T’ung-kuan fell and the court fled westward in panic. Li Hua, instead of going to join the exiled emperor, went to seek his mother who was living in rebel-controlled territory in southern Ho-pei. He was captured by the rebels and forced to take office under them. When the capitals were retaken the following year, he was tried for collaboration and was only saved from a severe penalty by the intervention of Liu Chih (of whom more will be said presently). He was given a minor post at Hang-chou. Soon after, his mother died and he retired. For the rest of his life he took no office but settled on a farm at Ch’u-chou (Huai-an Hsien in Kiangsu). He suffered from ill health and, according to Tu-ku Chi who wrote a preface to his works, was so poor that the younger members of his family had to work in the fields. He himself eked out his income by composing tomb inscriptions for the gentry of the region.29 Li Hua and his friends Hsiao Ying-shih and Chia Chih had already, before the rebellion, in the words of Tu-ku Chi, “advocated the style of middle  It should also be noted that Han Hui’s name appears among Liu Tsung-yüan’s list of the friends of his father. Hui is said to have been good at philosophical conversation (ch’ing yen) and famed for his literary ability. See Liu Ho-tung chi, ch. 12 (ts’e 2, p. 89, edition of the Kuo-hsüeh chi-pen ts’ung-shu). 28  On the beginnings of this movement, see inter alia, Kung Shu-chih, op. cit.; Kuo Shaoyü, Chung-kuo wen-hsüeh p’i-p’ing shih (“History of Chinese Literary Criticism”) (rev. ed., Shanghai, 1955); Lo Ken-tse, Chung-kuo wen-hsüeh p’i-p’ing shih (new ed., Vol. ii, Shanghai, 1957); Ch’ien Chi-po, Han Yü chih (rev. ed., Shanghai, 1957); Ch’ien Mu, loc. cit. 29  Biographies in cts, ch. 190C; hts, ch. 203. Preface to his collected works by Tu-ku Chi, ctw, ch. 388, p. 11b. See also Ch’ien Chi-po, op. cit., pp. 12–16.

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

Neo-Confucianism in T ’ ang Intellectual Life

175

antiquity (i.e., Chou and Han) in order to enlarge the influence of literature.”30 Unfortunately Li’s early writings were already lost when Tu-ku Chi wrote, and it would be difficult to judge whether there was development in his later years. His mood and that of his friends before the rebellion seem to have resembled Yüan Chieh’s. They sought relief for a corrupt world through a reform of literature which would present the true way of the ancients. Apparently it was while he was in retirement that Li Hua influenced others to try to revive the ancient style and started the movement which dominated literary discussions during the next half-century. Li’s friend Hsiao Ying-Shih, a descendant of the Southern Liang ruling house, was a proud-spirited man who held office for only a few brief periods and who, it was once said, even dared on one occasion to show disrespect to Li Lin-fu. He anticipated post-rebellion interest in the Spring and Autumn Annals and in the later problem of the legitimate line of successsion (cheng-t’ung), for he wrote, or at least projected, a continuation of that work from Han to Sui in which he expressed judgments on men and events and, in particular, made the line of legitimacy pass directly from Liang to Sui, bypassing Ch’en. Allegedly, he gathered disciples around him who treated him as their master.31 During the rebellion, he fled south like so many others, was for a time in Shan-nan (Hupei), and then farther east. Probably at this time he acquired an estate near Lake P’o-yang. He died at the age of 52 in 768. His son, Hsiao Ts’un, was a friend of Han Hui, and Han Yü, when a boy, knew him.32 Ts’un was intimate with the ku-wen writer Liang Su and the historian Shen Chi-chi, and had a career as a financial official. Another close friend of Li Hua was Liu Hsün, a son of the great historian Liu Chih-chi (661–721), author of the Shih-t’ung, and a brother of Liu Chih who was instrumental in saving Li Hua’s life. A student of the Classics, Hsün wrote a work called the Liu shuo, which, according to Li Hua, was aimed at “harmonizing men’s minds by means of the Six Classics.” Since the work is now lost, we have no way of telling how this was to be brought about, but it is interesting evidence of the beginnings of the revival of Confucian studies even before the rebellion.33 30   c tw, ch. 388. 31  Biographies in cts, ch. 190C; hts, ch. 202. See also Ch’ien Chi-po, op. cit., pp. 5–12; Hiraoka Takeo, Keisho no dentō (“The Tradition of the Confucian Canon”) (Tokyo, 1951), pp. 92–139. 32   h ts, ch. 202; Yin hua lu, iii, 89–90; Han Ch’ang-li shih hsi-nien chi-shih, p. 525. Cf. Ch’ien Chi-po, op. cit., p. 12. 33   c ts, ch. 102; hts, ch. 135; Li Hua, “San hsien lun,” ctw, ch. 317, p. 3b; T’ang kuo-shih pu 1, p. 15.

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

176

Pulleyblank

Although such men as Yüan Chieh, Li Hua, and Hsiao Ying-shih were all affected by the rebellion and reacted to it, they had matured during the prerebellion period. Their ideas for reform of scholarship, literature, and morals were originally prompted by revulsion from the world and, as we have seen, tended to be idealistic and impractical. They were aristocrats, and several of them had been content to live in retirement on their estates rather than to seek office. Too, with the exception of Hsiao Ying-shih, they were northerners, and, as Lo Ken-tse suggests,34 part of their protest may have been the traditional adherence of the north to classical studies and its objection to the florid literature of the south, which had been invading the T’ang court since the beginning of the dynasty. After the rebellion, the young men who followed in their footsteps and took inspiration from their ideas were in a quite different position. It was no longer sufficient to hold oneself aloof and perfect oneself in retirement. The disordered state of the world cried out for practical remedies; many of the scions of aristocratic houses were forced as never before to seek employment and to use scholarship and literary craftsmanship either to attract the notice of provincial governors or to seek office at court through the examination system. Dwelling in the south and coming under direct influence of the southern cultural tradition, some of the displaced northerners modified their protest from a simple, negative, purist reaction to a creative synthesis which, in the field of belleslettres, eventually produced a rich and supple free prose style and, in the field of philosophy, drew on southern forms of Buddhism to supply a metaphysical grounding for a revived Confucianism. Possibly this modification can already be seen in Li Hua’s later works, and it is clear in the writings of Tu-ku Chi. Tu-ku Chi came from an illustrious northern family and shared his family’s strong interest in Taoism. As a young man, in the years before the rebellion, he attracted attention as a writer and met Li Hua and Fang Kuan. He passed the examination in the Taoist Classics in 754. During the rebellion he fled to Chiang-nan and established his home at Suchou, where he served for a short time under the Military Governor of Chianghuai and was later summoned to court and appointed Omissioner of the Right. In this position he is said to have remonstrated sharply on various abuses. After holding other offices at the capital, he went in 768 to be Prefect of Hao-chou (near Feng-yang Hsien in Anhwei). From then until his death in 777, he served in the lower Yangtze region. In 771 he became Prefect of Shu-chou (Ch’ien-shan 34  Op. cit., vol. 2, pp. 113–14.

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

177

Neo-Confucianism in T ’ ang Intellectual Life

Hsien in Anhwei), and finally, in 774, he became Prefect of Ch’ang-chou, the title by which he was posthumously known.35 Like his friend and master, Li Hua, he practiced and advocated ku-wen and gathered disciples. He associated with other leading figures and was regarded as the leader of the movement. According to the modern critic Lo Ken-tse, he was less uncompromising than Li Hua in his demands for simplicity of diction and for rejection of all but the canonical works as models.36 His greatest pupil was Liang Su, who occupied a pivotal position in the history of ku-wen before Han Yü. But before considering Liang Su, let us turn to certain other intellectual trends which were developing in the south in the post-rebellion period.

The New Criticism of the Spring and Autumn Annals

Parallel to the main literary and philosophical movement leading to ku-wen, we find a new trend in criticism of the Confucian Classics which arose in reaction to the orthodoxy established by the standard commentaries (Wu ching cheng-i) in the reign of T’ai-tsung.37 As on other occasions in the history of Chinese thought, it began with a new school of interpretation of the Spring and Autumn Annals. Tan Chu (725–70), the founder of this school, was of northern origin. He had not yet held office when the rebellion broke out in 755, but was then sojourning in Chiang-nan. Because of the troubles in the north, he did not return. He was appointed to minor posts (perhaps by the local military governors?) first at Lin-hai Hsien (Chekiang) and then at Tan-yang Hsien in Jun-chou (Kiangsu). At the end of his second period of appointment, in 761, he retired and settled permanently in Jun-chou. During the next ten years, until 770, he worked on his commentary to the Spring and Autumn Annals. In that year Chao K’uang, who was the person principally concerned with the publishing and transmitting 35   h ts, ch. 162. Cf. the hsing-chuang by Liang Su, ctw, ch. 522, p. 3b; preface to his works by Li Chou, ctw, ch. 443, p. 16a; shen-tao pet by Ts’ui Yu-fu, ctw, ch. 443, p. 16a. The dates of his taking office at Hao-chou, Shu-chou, and Ch’ang-chou can be inferred from his memorials expressing thanks for his appointments at Hao-chou and Ch’ang-chou in ctw, ch. 385. 36  Op. cit., pp. 125–30. 37  Edited by K’ung Ying-ta in 180 chüan and completed in 653. See Nagasawa Kikuya, Shina gakujutsu bungeishi (Tokyo, 1938), pp. 145 ff. (trans. P. Eugen Feifel, Geschichte der chinesischen Literatur, Peking, 1945, pp. 176ff.).

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

178

Pulleyblank

of Tan’s work, was serving on the staff of Ch’en Shao-yu, Military Governor of Hsüan-she (southern Anhwei), and passed through Jun-chou on official business. There he met Tan Chu and became interested in his work. In the same year Tan Chu died, at the age of forty-seven. Later that year Chao was transferred, along with the Military Governor, his superior, to Che-tung (northern Che-kiang). Tan Chu’s son Tan i and Lu Ch’un, who was later sometimes referred to as Chao K’uang’s disciple but who seems to have been a direct disciple of Tan Chu, copied out Tan Chu’s writings and took them to Chao K’uang, who edited them. Lu Ch’un then prepared the work, known as the Ch’un-ch’iu t’ungli, for publication and wrote a preface dated 775.38 There are only meager details available about Chao K’uang. He is known to have been Prefect of Yang-chou (Yang-hsien in Shensi) (presumably later than his post under Ch’en Shao-yu) and he wrote a long critique of the examination system, in which, among other interesting remarks, he expresses his dislike of the standard commentaries on the Classics and his desire that classical studies concern themselves with the “general meaning” (ta-i) of the texts rather than with mere questions of punctuation and textual criticism.39 Lu Ch’un, like most of his surname, came from Su-chou. Liu Tsung-yüan states that he was a friend and disciple of Tan Chu, so he must have begun bis studies before the latter’s death. According to his biography Lu first received employment on Ch’en Shao-yu’s staff when the latter was Military Governor at Yang-chou (773–84). It is possible that Lu received the appointment through Chao K’uang’s influence in order that the two might work together on Tan Chu’s posthumous papers.40 Later he was at Ch’ang-an where he held a number of offices, finally being made Doctor in the University of the Sons of State (Kuo-tzu po-shih). It was no doubt in this capacity that he was able to propagate his ideas among the young 38  There is only a brief mention of Tan Chu in cts, ch. 189B. The biography in hts, ch. 200, is based, rather inaccurately, on Lu Ch’un’s “Ch’un-ch’iu li t’ung hsü” (“Preface to the Compendium of Rules of the Spring and Autumn Annals”), ctw, ch. 618, p. 3b. This piece also appears as p’ien 8 in Lu Ch’un, Ch’un-ch’iu chi chuan tsuan li in the Ku ching chieh hui han. 39  Besides brief mentions of him in cts, ch. 189B, and hts, ch. 200, and in Lu Ch’un’s preface cited in note 38, we know that he (later?) became Prefect of Yang-chou (Shensi). Cf. tt, ch. 17, p. 97. 40   c ts, ch. 189B; hts, ch. 200; mu-piao in Liu Ho-tung chi, p. 9. His name was changed from Ch’un to Chih (Mathews’ number 1009) to avoid a taboo in the name of the Emperor Hsien-tsung. Although this was not homophonous with the name of Lu Chih (Hsüankung) (Mathews’ number 980), it was sufficiently similar that there is occasionally confusion between them. See note 64.

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

Neo-Confucianism in T ’ ang Intellectual Life

179

intellectuals. Besides advocating the works of Chao K’uang and Tan Chu, he also wrote his own commentary on the Annals. In the latter part of Te-tsung’s reign he was serving in the provinces, but his works circulated in the capital and had great influence. He returned to Ch’ang-an in 804 and found himself the revered teacher of an important coterie. Before the time of Tan Chu, students of the Spring and Autumn Annals had generally followed the tradition of one or another of the three commentaries established in the Han period—Kung-yang, Ku-liang, or Tso—and though they might supply deficiencies in their favored commentary by reference to the others, there was little or no attempt to make any fundamental criticism of the accepted opinions about them. Tan Chu, however, proposed to reject the authority of all three commentaries and return directly to the Classic itself to establish on a rational basis the rules of “praise and blame” which Confucius had used in editing it. He thought that all three commentaries had been orally transmitted at first and only after several generations written down by later disciples. Moreover, in the case of the Tso commentary, the tradition of Confucius’ teaching had been greatly amplified, partly at the beginning by Tso Ch’iu-ming’s drawing on the chronicles of various other states, and partly at the time of writing down by the addition of material from heterogeneous sources such as family biographies of leading figures, books of divination, collections of anecdotes, and proposals of the School of Politicians, etc., so that it was difficult to sort out true from false. Chao K’uang went further and rejected the tradition that identified the author of the Tso commentary with Tso Ch’iu-ming. He declared that the Kuo-yü and the Tso-chuan were different in style (wen-t’i) and had such discrepancies that they could not be by the same author.41 41  No separate works by Tan Chu were listed in the bibliographical chapter of the Hsin T’ang shu or of the Sung shih. He was extensively quoted by Lu Ch’un, and these quotations together with one or two others from Sung and Yüan commentators of the Ch’un-ch’iu have been collected by Ma Kuo-han and published under the title Ch’un-ch’iu chi chuan in the Yü-han shan fang chi-i shu. Chao K’uang was the author of the Ch’un-ch’iu ch’an-wei tsuanlei i-shu in 10 chüan, which is not listed in the hts “I-wen chih” but is found in the Sung shih “I-wen chih,” as well as bibliographies from the Sung period such as the Chung-hsing kuan-ko shu-mu. It is no longer extant, but quotations from Chao K’uang in the works of Lu Ch’un, together with one or two more from Sung and Yüan works, have been collected under this title by Ma Kuo-han and published in the above-mentioned collection. Three works by Lu Ch’un on the Ch’un-ch’iu still exist, namely, (1) Ch’un-ch’iu chi-chuan tsuan-li, 10 ch., (2) Ch’un-ch’iu wei chih, 3 ch., (3) Ch’un-ch’iu Tan Chao erh hsien-sheng chi-chuan pien-i, 10 ch. They are all reprinted in the Ku-ching-chieh hui-han.  On their ideas and influence see, inter alia: P’i Hsi-jui, Ching-hsüeh t’ung-lun (Shanghai, 1938), pp. 56ff.; Nagasawa Kikuya, Shina gakujutsu bungeishi (Tokyo, 1938), pp. 152ff. (trans.

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

180

Pulleyblank

It is difficult for us to understand the interest which the dry, crabbed chronicle of Lu aroused among Chinese literati, but we can, I think, appreciate the excitement with which young men learned of such a radical and “modern” critical approach to a canonical work. One can hardly find a parallel in Europe before the Homeric criticism of Bentley and his successors at the end of the seventeenth century. It is not merely that scholars dared to call tradition into question; others such as Wang Ch’ung and Liu Chih-chi had done this before. But we have only to compare Tan Chu, Chao K’uang, and Lu Ch’un with Liu Chih-chi half a century earlier, to see the difference. Liu Chih-chi was as bold or bolder in his attack, for he would even call Confucius to account; but he was frequently partial and prejudiced, and he dealt haphazardly with individual points.42 Most remarkable in the School of Tan Chu is the cool, detached, and methodical rationality with which they attempted to get at the truth. No doubt by our standards they did not go far enough in questioning tradition; we must remember, however, that they stood at the very beginning of the critical scholarship which eventually flourished under Sung. Several leading Sung scholars acknowledged their debt to these three men, whom, in the opinion of the late nineteenth-century critic P’i Hsi-jui, they did not surpass in the justness of their judgment.43 And such a modern scholar as Professor William Hung has found Chao K’uang’s arguments relevant to his discussion of the Spring and Autumn Annals and its commentaries.44 There was, it is true, an ambivalence in their critical approach o£ which they were unaware. Their purpose was to get behind the distortions of tradition to the true meaning of the Sage, and they assumed this to be perfectly consonant with their own highest principles of rationality. Thus we find in them, and in the criticism of such a follower of theirs as Liu Tsung-yüan, a mixture of judgments based on critical historical arguments and of unhistorical, a priori E. Feifel, Geschichte der chinesischen Literatur, pp. 185–86); Takeuchi Yoshio, Chūgoku shisōshi (Tokyo 1954; 1st ed., 1936), p. 238; Honda Nariyuki, Chung-kuo ching-hsüeh shih (Chinese translation by Sun Liang-kung) (Shanghai, 1935), p. 235; Kanō Naoki, Chūgoku tetsugaku shi (Tokyo, 1953), pp. 340–43. 42  See Shih-t’ung, especially ch. 13, “Suspicions about Antiquity”; ch. 14, “Doubts about the Classics”; ch. 15, “Preferring Tso.” Cf. E. G. Pulleyblank, “Chinese Historical Criticism—Liu Chih-chi and Ssu-ma Kuang,” to be published shortly in a collection of papers on the historiography of East Asia by the School of Oriental and African Studies, London. 43  P’i Hsi-jui, op. cit., p. 59, quotes praise of Tan Chu and his followers from Shao Yung, Ch’eng Tzu (one of the Ch’eng brothers), Chu Hsi, and Wu Ch’eng (of Yüan). Honda, loc. cit., also refers to praise by Lu Chiu-yüan. 44  Hung Yeh (William Hung), “Ch’un-ch’iu ching-chuan yin-te hsü,” pp. i–cvi in Ch’un-ch’iu ching-chuan yin-te (4 vols., Peking, 1937).

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

Neo-Confucianism in T ’ ang Intellectual Life

181

judgments about what must be true. The unquestioned assumption about the agreement between the real ideas of the Classics and the metaphysical ideas being developed during the T’ang and Sung was unjustified, so that there were bound to be conflicts in the thought of later men, if not in the initiators of the critical movement. During the Sung we find the two tendencies differently emphasized in different thinkers. The unhistorical side finally won out, for NeoConfucianism in its Chu Hsi synthesis used the interpretation of the Classics as the vehicle for expounding its new metaphysical system. Before this happened, however, criticism had made many solid gains and established a basis from which the Ch’ing philologists could once again react against subjective interpretation of the Classics and try to get a more objective idea of what they really were and really said. One further remark about Tan Chu, Chao K’uang, and Lu Ch’un is worth making. In contrast to the ku-wen writers, who were mostly aristocrats, none of these initiators of the new critical approach to the Classics belonged to the great ruling families. Their personal backgrounds are obscure, but we may safely surmise that they belonged to the stratum of the literati which depended entirely on the examination system for entrance into official position. Chao K’uang’s views on the examinations are set out at length in the T’ung-tien and are referred to below. As we shall see, he criticized the existing system not with a view to abolishing it and reverting to a dependence on recommendation, which was the nostalgic view of some aristocrats, but in order to put substance and meaning into the Confucian curriculum and to ease the candidates’ financial burdens. What can be inferred from this difference in social background is not clear. Both classical studies and training in literary composition were largely directed toward passing the civil service examinations. It may be that new men were comparatively free from family traditions of classical scholarship and so able to look at problems with a fresh eye; on the other hand, the tradition of letters as the avocation of aristocratic courtiers may have played a part. It seems to me no mere coincidence that both movements developed at a time when enforced leisure and remoteness from the capital diverted energies from the scramble for office.

T’ien-t’ai Buddhism

In Buddhism new movements were also developing. Though we are concerned here mainly with the Chinese classical tradition, it would be misleading to give the impression that at this period Buddhism had lost its hold on the minds of

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

182

Pulleyblank

the secular learned, or was kept apart in the monasteries without influence on the world outside. Indeed, a substratum of Buddhist ideas is often to be presumed even when not explicitly present, and it is well known that as NeoConfucian metaphysics developed it drew directly on Buddhism. At Ch’ang-an, official Buddhism with its elaborate and costly ceremonies was now dominated by the newly introduced Tantric School; but in the provinces, and mainly in the south, there were other more significant developments. This was the period of the great movement started by Fa Chao (active 766–805) for the spread of the Pure Land teachings among the people.45 More limited and sophisticated in appeal, the Ch’an and T’ien-t’ai Sects were philosophically opposites but alike in their tendency to assimilate Buddhism to Chinese intellectual tradition.46 Of the two, Ch’an has the more perennial interest, but T’ien-t’ai is the more immediately relevant to the present discussion. The syncretism of this sect, which sought to harmonize sectarian divisions within the faith by treating them not as mutually exclusive but as forming a hierarchy of “levels,” prepared the way for finding the truths of Buddhist metaphysics in Confucian texts as well. Several of the early T’ang Neo-Confucians were lay followers of this sect, and the influence of its metaphysics can be seen clearly in their writings. The T’ien-t’ai Sect was indeed undergoing a rejuvenation under its ninth patriarch Chan-jan (711–82) which amounted to a second founding. His lay surname was Ch’i and he came from a family with a tradition of secular learning. From the age of about twenty, he studied Buddhism under the preceding patriarch Tso-ch’i, and began to gather disciples as a “scholar in retirement” (ch’u-shih), but it was not until after 742 that he became a monk. He wrote prolifically, expounding T’ien-t’ai doctrines and confuting opponents, and gained great fame.47 It is perhaps no more than coincidence that his period of influence coincided with the flight of northern scholars to the south and 45  See Tsukamoto Zenryū, Tō chūki no Jōdokyō (“Chinese Buddhism in the Middle Period of the T’ang Dynasty, with Special Reference to Fa-chao and the Doctrine of the Pure Land”), Kyōto, 1933 (with English summary). 46  On the general development of Buddhism in this period, see Arthur Wright, “Buddhism and Chinese Culture,” jas, 17 (1957), 17–12, especially pp. 31–38. 47  Chan-jan has a biography in Sung Kao-seng chuan, ch. 6 (Taishō Tripitaka, 2061, L, 739; see also the biography of his disciple Yüan-tsao, ibid., p. 740). According to his biography, Chan-jan resolutely declined all invitations to go to Ch’ang-an, but Wei Ch’u-hou’s epitaph for the Ch’an monk Ta-i (important for its account of the four schools of Ch’an after the death of Hui-neng) relates that Chan-jan was in Ch’ang-an shortly before his death, engaged in disputes over doctrine with Ta-i in the Nei Shen-lung-ssu within the palace in the

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

Neo-Confucianism in T ’ ang Intellectual Life

183

with the ku-wen movement, but one wonders whether there may not have been some reciprocal stimulus. It seems legitimate to suppose that his great success resulted at least partly from the attention he attracted among these refugee scholars and the encouragement they gave him. Li Hua and Liang Su were followers of both him and his disciple Yüan-tsao.48 Later, Liu Tsung-yüan was called a disciple in the fourth generation.49

The Return to the North Under Te-tsung

With the accession of Te-tsung (779–805) we find the center of intellectual life moving back from the provinces to Ch’ang-an. Already, under Tai-tsung (762– 79), members of coteries in the southeast had been drawn into political life at the capital. We have already mentioned Han Yü’s elder brother Han Hui, who achieved some fame in the south in 765, came to Ch’ang-an and became involved in the downfall of Yüan Tsai in 777. Clearly the patronage of the leaders of the ku-wen movement was used by men of talent in the Yangtze region as a means, first, to get employment under one of the military governors or financial commissioners of the region and, then, to get valuable recommendations for passing the examinations and obtaining posts in the central bureaucracy, which still remained the ultimate goal of all men of ambition. Thus we find Tu-ku Chi saying farewell to a Mr. Liu and commending him to the financier Han Hui (not Han Yü’s brother) who, he says, has already been instrumental in the advancement of several other friends.50 When Ts’ui Yu-fu, a close friend of Tu-ku Chi and Li Hua, and a strong supporter of the ku-wen movement, became Chief Minister in 779–80, he immediately filled a great many offices with personal acquaintances.51 He defended this on the ground that it was only by personal knowledge that he could vouch for their ability and character. We note with interest that he is in no way condemned for this by the T’ang historians. presence of the future Shun-tsung. There is clearly much more to be investigated about the influence of Buddhism on Shun-tsung and the people around him. (ctw, ch. 715.) 48   Sung Kao-seng chuan, loc. cit.; Fo-tsu t’ung-chi, ch. 10 (Taishō Tripitaka, 2035, xlix, 203C). 49   Fo-tsu t’ung-chi, loc. cit. 50   c tw, ch. 388, p. 16b. The Han Yu-shen mentioned in this hsü is the financial expert Han Hui who was involved, along with the other Han Hui, Han Yü’s brother, in the downfall of Yüan Tsai. (See n. 25.) This style is not given in his biographies in cts, ch. 129, and hts, ch. 126, but is mentioned in his hsing-chuan by Ch’üan Te-yü (ctw, ch. 507, p. 4a). 51   c ts, ch. 119. Cf. hts, ch. 142; tctc, ch. 225 (Ta-li 14/4/kueimao).

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

184

Pulleyblank

Rather, Ts’ui’s rival, whom he supplanted, is criticized for having rigidly kept to the examination procedures in making appointments. This throws an interesting light on contemporary criticisms of the examination system (discussed by Mr. Nivison elsewhere in this volume) and also on the development of a formalized system of sponsorship, the beginnings of which we see at the same period.52 Candidates protested against the narrow formalized curriculum they were forced to submit to, and against the exhausting, costly, long-drawn-out, and, for the majority, ultimately frustrating ordeal of the examination system. They looked back to an idealized past in which true merit had allegedly been recognized and brought forward by simpler procedures. Their attitude to patronage, however, differed according to whether they were successful in obtaining it or not. Men of good family background, who found it relatively easy to make good connections, were naturally more inclined to see the virtues of direct recommendations. Those who found in the examination system, however arduous, their only hope of advancement were more inclined to lay stress on the impartiality of the examiners, but even they saw nothing reprehensible in trying to attract attention of influential persons by sending examples of their writing. What would be currying favor when done by a “small man” would only be obtaining just recognition of merit when done by a true gentleman. Examiners in their turn were anxious to make their names illustrious by selecting young men of authentic talent, and evidently relied on such extracurricular aids to their judgment as much as on the formal examination essays. They regarded this as in no way impairing their impartiality. To the outsider and the disappointed candidates, the line between such practices and favoritism or even corruption must often have seemed finely drawn. Yet, as contemporary critics of the examinations maintained, an all-round estimate of a candidate’s talents and accomplishments by a man of discernment was more reliable than mechanical results obtained in formalized and impersonal examinations. The most illustrious examiner during the twenty-five years of Te-tsung’s reign probably was Lu Hsüan-kung (Lu Chih),53 administrator of the doctoral examinations in Chen-yüan 8 (792). This was an annus mirabilis in which not only Han Yü but also several other noted writers who were his friends, such as

52  See below. 53  I refer to him by his title, by which he is well known in China, instead of his name because of the possibility of confusion with Liu Chih (see below). Biographies in cts, ch. 139; hts, ch. 157. See also preface to his works by Ch’üan Te-yü in ctw.

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

Neo-Confucianism in T ’ ang Intellectual Life

185

Ou-yang Chan,54 and Li Hua’s nephew Li Kuan,55 and future statesmen such as Li Chiang56 and Ts’ui Ch’un,57 received their degrees. In making his choice, Lu Hsüan-kung is said to have relied especially on the recommendations of Tu-ku Chi’s disciple Liang Su, who had since 789 been in Ch’ang-an serving as Han-lin Scholar and tutor to the imperial princes. Lu Hsüan-kung has an important place in intellectual history. A southerner from the prominent Lu clan of Su-chou, he was a statesman who, after passing the chin-shih degree in 773, had risen to prominence as a Han-lin Scholar during Te-tsung’s exile in Feng-hsiang in 783, where he had been in charge of composing edicts. Since that time he had been an important adviser of the emperor. Lu Hsüan-kung was not himself a ku-wen writer; indeed, he was a most elegant practitioner of the balanced prose style. But in his attitude to politics he was a Confucian moralist and his writings—though somewhat long-winded—are full of substance. Unfortunately, his private writings have not been preserved; although we know a good deal about his views on public matters, we know comparatively little about his personal relationships. As a statesman he was a conservative who advocated ancient moral precepts but nevertheless had a realistic view of practical matters. He is noted for the way in which he opposed Te-tsung’s desire for personal rule. For example, though a Han-lin Scholar, he memorialized against the encroachments this palace organ was making on the functions of the regular bureaucracy.58 He is represented as the paragon of Confucian constitutionalists, and was obviously considered the leader and patron of that element in the literati. Liang Su,59 Lu Hsüan-kung’s adviser on examination candidates, came from a northern family which had been forced to flee south at the time of the An Lushan rebellion and had in consequence “somewhat come down in the world.”60 He also studied extensively in all fields of learning—Classics, philosophers, history, Taoism, and Buddhism—and he attracted the notice of Li Hua and Tu-ku Chi by his brilliant writings. He was a lay follower of the T’ien-tai Sect and wrote the Chih-kuan t’ung-li,61 an exposition of chih-kuan meditation, one 54  Biography in hts, ch. 203. See also Ch’ien Chi-po, op. cit., pp. 60–64. 55  Biography in hts, ch. 203. See also Ch’ien Chi-po, op. cit., pp. 58–60. 56  Biographies in cts, ch. 164; hts, ch. 152. 57  Biographies in cts, ch. 159; hts, ch. 165. 58   t hy, ch. 57, p. 979. 59  Biography in hts, ch. 202 (appended to Su Yüan-ming). Epitaph by Ts’ui Yüan-han, ctw, ch. 523, p. 26a. Preface to his works by Ts’ui Kung, ctw, ch. 480. Cf. Ch’ien Chi-po, op. cit., p. 26. 60   c tw, ch. 523, p. 26b. 61   c tw, ch. 517, pp. 16a ff.

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

186

Pulleyblank

of the principal tenets of the sect. Of this work the Sung Kao-seng chuan says, “The Han-lin Scholar Liang Su therefore alone wielded a mighty brush and achieved words of extraordinary virtue…. When we examine Scholar Liang’s discourse we find conclusions and arguments equally cogent. Who but this man could move great Confucians?”62 Ts’ui Yüan-han,63 the Confucian who composed Liang’s epitaph, compared this work to the Great Appendix to the Book of Changes. Nothing illustrates more clearly how little philosophical conflict between Confucianism and Buddhism was felt at this period—until Han Yü came along with his intransigent attitude. After his arrival in Ch’ang-an in 789, Liang Su was the acknowledged leader of the new stylistic movement, and hopeful young men hastened to pay him court and study under his direction. Lü Wen, who studied the Spring and Autumn Annals under Lu Ch’un, was Liang Su’s pupil in literary composition.64 Han Yü, then at the capital trying to take his chin-shih degree, does not appear to have been personally a student of Liang Su, but, according to his biography, “he went around with the followers of Tu-ku Chi and Liang Su.”65 Han Yü himself, in a letter written ten years later when he held the post of Doctor in the Ssu-men Academy and was seeking similar favor for some of his own protégés, referred to Liang Su’s part in presenting his name and those of other examination candidates.66 Liang Su’s death and the fall of Lu Hsüan-kung in 794 did not bring to an end the interest in ku-wen and in revitalized Confucian studies among the examination candidates during the last years of Te-tsung’s reign, nor is there any 62   Taishō Tripitaka, 2061, L, 740 (appendix to biography of Chan-jan). 63  See note 59. 64  Biographies in cts, ch. 137; hts, ch. 160. Eulogy (lei) by Liu Tsung-yüan, Liu Ho-tung chi, ch. 9 (ts’e 2, p. 49). Preface to his works by Liu Yü-hsi, ctw, ch. 605, p. 10b states that he studied the Ch’un-ch’iu under Lu Chih (Lu Hsüan-kung), but this must be a mistake for Lu Ch’un (see note 40), as stated in the hts. Lü Wen refers to the importance of Ch’un-ch’iu studies in a letter to a cousin (ctw, ch. 627, pp. 15a ff.). Ch’ien Mu ranks him above any of his contemporaries as a thinker. (Hsin-ya hsüeh-pao 3/1, pp. 163ff.). 65   c ts, ch. 160. T’ang chih-yen, ch. 7, p. 80, has an anecdote about Liang Su and four of the candidates he recommended to Lu Hsüan-kung in 792, in which he is said to have made predictions about their future careers. Ch’ien Mu (loc. cit., p. 131) shows that it could not have happened. 66  “Yü Tz’u-pu Lu Yüan-wai shu” (“Letter to Lu San”), Han Ch’ang-li chi, ch. 17 (ts’e 4, p. 75). See also “Ou-yang sheng ai-tz’u” (“Lament for Ou-yang Chan”), ibid., ch. 22 (ts’e 5, p. 45). A letter of Li Ao also mentions Li Kuan as having recommended the poet Meng Chiao, who was also a close friend of Han Yü, to Liang Su (ctw, ch. 635, p. 19b).

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

Neo-Confucianism in T ’ ang Intellectual Life

187

reason to suppose that examiners were unsympathetic. Candidates who were unsuccessful, or had to wait for some time to pass, naturally complained of unfairness, but one must take such murmurings with a grain of salt.67 After all, only a few could be chosen each year from the many applicants. Of those who were chosen, many made their mark later. The examiners who followed Lu Hsüan-kung are for the most part praised in the histories. Ku Shao-lien68 (793– 94 and 798), who followed Lu Hsüan-kung and under whom Liu Tsung-yüan took his chin-shih, is not an outstanding figure but his apparent hostility to Pe’i Yen-ling suggests he probably supported Lu Hsüan-kung. Kao Ying69 (799–801) was noted for his uprightness as an examiner, and Ch’üan Te-yü70 (802–3 and 805), a writer of considerable fame, later received Han Yü’s praise for his conduct as an examiner,71 a significant tribute, for Han Yü had been concerned during Ch’üan’s period of examining with getting his own candidates through. Only Lü Wei (795–97),72 who had connections with P’ei Yen-ling, is hostilely treated in his biography, but since his son Lü Wen was a student of Liang Su and Lu Ch’un, and a close friend of Liu Tsung-yüan, it is unlikely that Lü Wei was prejudiced against the “new thought.” Among the young men themselves two groups stand out. There were those who from about 803 onward constituted a “brains trust” for the faction headed by Wang Shu-wen, Wang P’ei, and Wei Chih-i, which took power and came to grief in 805 under Shun-tsung. Liu Tsung-yüan, Liu Yü-hsi,73 Lü Wen, Li Ching-chien,74

67  Cf. A. Waley, The Life and Times of Po Chü-i (London, 1949), p. 23. 68   h ts, ch. 162. 69   c ts, ch. 147; hts, ch. 165. Cf. Waley, op. cit., pp. 18, 23. 70   c ts, ch. 148; hts, ch. 165. 71  “If the words of those who recommended scholars to him were worthy of credence, [these scholars] did not fail to find employment because they were commoners; if the words were not worthy of credence, then even though high officials and men of influence interceded, he did not in the slightest degree alter his opinions.” Oman’s epitaph in Han Ch’ang-li chi, ch. 30 (ts’e 6, pp. 47–49). 72   c ts, ch. 137; hts, ch. 160. 73   c ts, ch. 160; hts, ch. 168. See also his autobiographical sketch, “Tzu Liu-tzu chuan” (ctw, ch. 610, pp. 13b ff.). 74   c ts, ch. 171; hts, ch. 81. He was the author of a criticism of Mencius (Meng-tzu p’ing). See note 83. He also seems to have had connections with Han Yü’s circle, for Li Ao couples his name with that of Chang Chi in a recommendation to the Military Governor of Hsü-chou, Chang Chien-feng (letter cited in note 66 above). A later incident in his rather colorful career is referred to by Waley, op. cit., p. 138.

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

188

Pulleyblank

Han Yeh,75 Han T’ai,76 and Ch’en Chien77 are among the most prominent of them. The rival group included Han Yü and his followers. Liu Tsung-yüan’s group was probably more influenced by Tan Chu’s school of classical studies than by the ku-wen movement proper. Liu says of himself, “I practiced literature for a long time, but I despised it in my heart and did not work at it. I looked upon it as no more than a special skill such as in chess. So when I was at Ch’ang-an I did not try to make a name by this means. My intention was to apply myself to practical affairs and make my way that of reforming the age and affecting people.”78 He describes how his interest in the Spring and Autumn Annals was aroused by hearing a discussion at the house of his brother-in-law P’ei Chin in Ch’ang-an. He heard P’ei and a certain Mr. Yüan arguing with Lü Wen and Han Yeh on the meaning of a passage in the Annals. It came to him as a revelation that the ancient significance of the work which had long been hidden was being laid bare. Later, at Han T’ai’s house, he saw a copy of

75   h ts, ch. 168 (appended to the biography of Wang P’ei). He belonged to the true Ch’ang-li Han clan. (Although Han Yü is referred to, and referred to himself, as from Ch’ang-li, from which the most ancient and illustrious Hans came, his family did not in fact come from there. See Chu Hsi’s commentary to the biography in Han Ch’ang-li chi.) He was related, in the next generation to the two brothers Han Huang and Han Hui, who were both financial experts (cts, ch. 129; hts, ch. 126; see note 50 above). The genealogical table of Chief Ministers in the hts calls him the son of Han Hui (hts, ch. 73A, p. 15a), but cts, ch. 129, refers to him only as the tsu-tzu of Han Huang. 76   h ts, ch. 168 (brief reference appended to Wang P’ei’s biography). According to hts, ch. 73A, p. 9a he belonged to the same clan as Han Yüan, a chief minister under Kao-tsung (cts, ch. 80; hts, ch. 105). Han Yü later recommended him to replace himself as prefect of Ch’ao-chou in 820. After the collapse of the Wang Shu-wen party, Han Yeh had been banished to Ssu-ma of Ch’ien-chou (Kan-hsien in Kiangsi). In 815, when the exiled members of the party were given a partial reinstatement, he was made Prefect of Chang-chou (Lung-ch’i hsien in Fukien). Han Yü was banished to be Prefect of Ch’ao-chou (Hai-yang hsien in Kuang-tung) in 819 on account of his famous memorial on the bone of Buddha. Since both Chang-chou and Ch’ao-chou were prefectures of the lowest grade and Ch’aochou was actually farther south, it is not obvious what superiority the one had over the other. Possibly in recommending Han Yeh, Han Yü was trying to enable him to enter the stream of regular promotion, since Yeh had already been at Chang-chou for five years, much longer than the normal length of time. In any case it is interesting to find Han Yü endeavoring to help victims of the downfall of Wang Shu-wen. (See Han Ch’ang-li chi, ch. 39, ts’e 7, p. 43.) 77   h ts, ch. 168, after the biography of Wang P’ei. There are six pieces by him in ctw, ch. 684, including an essay in praise of Liu Yen, on whose staff he had once served. 78  Letter to Wu Wu-ling in Liu Ho-tung chi, ch. 31, ts’e 4, pp. 101–2.

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

Neo-Confucianism in T ’ ang Intellectual Life

189

Lu Ch’un’s works.79 Finally, when Lu Ch’un returned to Ch’ang-an in 804, Liu Tsung-yüan and his friends were able to talk directly with the man they looked on as their master. The influence of this school of interpretation can be seen in Liu Tsungyüan’s own critical works. Sometimes his argument recalls Tan Chu’s views on the transmission of the commentaries to the Spring and Autumn Annals. This resemblance is to be seen in Liu’s essays on the authenticity of Wen-tzu, Yen-shih ch’un-ch’iu, Kang-ts’ang-tzu, and Ho-kuan-tzu, and in his discussion of the Analects (which, he argues, cannot have been composed by Confucius’ own disciples but was probably put together by the disciples of Tseng-tzu).80 More often, however, his criticism of ancient works is ethical and rationalistic. Old stories are criticized for the conduct that is portrayed in them, judged on a timeless basis, without taking into account any relativity of standards. This is the spirit pervading his critical attack on the Kuo-yü (Fei Kuo-yü) which he composed during his exile.81 The spirit is one of independence, of unwillingness to accept tradition blindly; but the impulse is toward a rationalistic ethic for his own time rather than toward what we would think of as objective historical scholarship, aimed at understanding the past in its own terms. Liu Tsung-yüan and his friends must have shocked people by their boldness. In a letter to Lü Wen about the Fei Kuo-yü, Liu relates the comment that a man named Lu had made about his friend Li Ching-chien’s Critique of Mencius (Meng-tzu p’ing). “It is certainly well done. But did the ancients in writing books attack their predecessors like that?” Liu observed, “His intention was to make clear the Way. It was not to attack Mencius.”82 Liu Tsung-yüan’s group seems to have remained eclectic in its thought. However fervently the value of the Classics was discussed, there did not seem to be a need felt to reject non-Confucian schools of thought. For example, we find Liu Tsung-yüan, Li Ching-chien, and Lü Wen showing great interest in Tuan Hung-ku, an advocate of Legalist doctrines.83 This again shows that their main concern was to construct a valid philosophy, not to preserve an orthodoxy. 79  Letter in answer to Yüan of Jao-chou in Liu Ho-tung chi, ch. 31, p. 98. See also his epitaph on Lu Ch’un, ibid., ch. 9, ts’e 2, p. 45. 80  See the essays on these books in Liu Ho-tung chi, ch. 4. 81  Ibid., ch. 44–45. 82  Ibid., ch. 31, ts’e 4, p. 100. The character chih (Karlgren, Grammata Serica, no. 804e) is not given the meanings “attack” or “destroy” in dictionaries but it is occasionally found in texts where it evidently has such a sense. See P’ei-wen yün-fu under chih-hui (G.S., no. 356a) and ch’iung (G.S., no. 1006g) chih. 83  Ibid., pu i, ts’e 6, p. 86.

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

190

Pulleyblank

By contrast Han Yü, who like other advocates of ku-wen emphasized wen as the means of expressing the Way, adopted an attitude of militant Confucian orthodoxy. There was little in Han Yü’s specific ideas either on literature or on philosophy that had not already appeared in the writings of his predecessors. Interest in such works as the Doctrine of the Mean, the Great Learning, and the Book of Changes as sources of a Confucian metaphysics was by now a commonplace. The idea that the line of succession of Confucian teaching had been broken was expressed also by others, notably Lü Wen.84 Han Yü’s innovation lay in actually assuming the role of teacher in the line of Mencius. According to both his own account85 and that of Liu Tsung-yüan,86 this pretension brought him ridicule and, although he did get a devoted group of followers, was responsible for his ill success in getting into office. After futile attempts to get a post through the Placement Examinations in 793 and 795, he gave up and took service under military governors at Pien-chou and Hsü-chou. Only in 802 did he get a post at Ch’ang-an as a Doctor in the Ssu-men Academy. In 803, he at last won entry into political office; he became an Examining Censor as a colleague of Liu Tsung-yüan, who was his junior in years and had had a much smoother start to his career. Liu Yü-hsi, who had taken his chin-shih degree in 793, was in the same office along with Liu Tsung-yüan, but had since then been serving at Yang-chou on the staff of the Military Governor. This Military Governor, who was none other than Tu Yu, one of the most learned scholars of the age and a distinguished statesman, also came to Ch’ang-an in 803, and it is now time to say something about him.

Political and Economic Thought—Tu Yu

Besides the new ideas in belles-lettres, scholarship, and metaphysics, there was in the post-rebellion period much thought devoted to practical matters of government and to questions of economic policy. Indeed, two of the most notable productions of the second half of the eighth century were Su Mien’s Hui-yao and Tu Yu’s T’ung-tien, both encyclopedic works of institutional history which provided models for later continuations and imitations. The Hui-yao,87 out of 84   c tw, ch. 627, pp. 15a ff. 85  “Shih shuo,” Han Ch’ang-li chi, ch. 11, ts’e 3, pp. 75–77. 86  Letter in reply to Wei Chung-li, Liu Ho-tung chi, ch. 34, ts’e 5, pp. 3–5. 87  Su Mien’s Hui-yao was in 40 chüan. thy, ch. 39, p. 660, appears to say that it was presented to the throne in Chen-yüan 19 (803), but as this merely forms an addition to an item recording the presentation of the T’ung-tien (which was moreover probably presented in

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

Neo-Confucianism in T ’ ang Intellectual Life

191

which grew the existing T’ang Hui-yao, was limited in scope for it dealt only with the T’ang dynasty. But the idea of providing a classified digest of important official documents was a most useful one, to which many generations of students have been immensely indebted. Unfortunately we are not well informed about Su Mien, although his brother Su Pien was an important financial official in the latter years of Te-tsung’s reign.88 The Su family were scholars, and the family library was reputed to be the most extensive private collection in the empire. There are scattered personal comments by Su Mien preserved in the T’ang Hui-yao from which we can deduce that, like his contemporary Tu Yu, he was concerned to explain the troubles that had befallen the dynasty in terms of economic and institutional history.89 His views, however, were traditional and much less interesting than those of Tu Yu, who is seldom mentioned in the history of Chinese philosophy but whose independence and originality deserve close attention. Tu Yu’s T’ung-tien was inspired by, and based on, a previous institutional encyclopedia known as the Cheng-tien, written by the son of the historian Liu Chih-chi, Liu Chih,90 who was, as already mentioned, instrumental in saving Li Hua’s life. Liu Chih’s work has not survived, but there are a number of lengthy quotations from it91 and we have other indications of his way of thinking. The Japanese scholar Kanai Yukitada92 credits him with considerable understanding of the problems of his time, especially of the examination system and of the development of centralized bureaucracy. I find it hard to agree. His ideas are not without interest, but, like others of his contemporaries, he was a wouldbe statesman prevented by the political situation from taking an active part in

801, not 803—see note 103), this is not to be relied upon. In any case, the book was probably completed toward the end of the Chen-yüan period. 88  Su Pien has biographies in cts, ch. 189B, and hts, ch. 103, and Su Mien is mentioned in both, but little is said of him besides the fact that he wrote the Hui-yao. He also edited the literary works of Chia Chih, who had been a friend of Li Hua and Hsiao Ying-shih. This may indicate that Su Mien had an interest in the ku-wen movement—like the majority of his contemporaries. (hts, ch. 60, p. 8a.) 89  Cf., for example, his comment on the development of the practice of appointing special commissioners (shih) (thy, ch. 78, p. 1438, translated in The Background of the Rebellion of An Lu-shan, p. 133). 90   c ts, ch. 102; hts, ch. 135. Funeral address (chi-wen) by Li Hua in ctw, ch. 321, p. 17a. 91   t hy, ch. 47, p. 830, includes his discussion on “feudalism” (feng-chien) from the Chengtien. It is also reasonable to assume that the discussion (lun) on examinations ascribed to him and included in tt, ch. 17, is from the Cheng-tien. 92   Tōdai shigaku shisō, pp. 90 ff.

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

192

Pulleyblank

affairs, expressing his frustration in theorizing insufficiently controlled by reality. His chance eventually came, only to end most ingloriously. Liu Chih had already made a name for himself by his discussions on government policies in the years 734–35 when he held the office of Administrator in the Left Chien-men Guard. His memorial against Chang Chiu-ling’s proposal to allow private coinage of money is noteworthy; Tu Yu considered it to be one of the best expositions of the problems of currency that had ever been made. During Li Lin-fu’s dictatorship, Liu Chih dropped out of sight as a politician, but it was at that time that he wrote the Cheng-tien. He was greatly admired by Fang Kuan, another scholar official, who had achieved limited fame as a provincial administrator before the An Lu-shan rebellion but had failed to advance to high office at the capital. Opportunity came for both of them when, after having protested vainly against the disastrous policy of seeking battle with the An Lu-shan forces at the T’ung-kuan,93 they joined the emperor on his flight to Szechwan in 756. The old emperor rewarded Fang Kuan’s loyalty by making him Chief Minister, and Liu Chih played the role of confidential adviser. One of Fang Kuan’s first acts was to persuade the emperor to divide the empire into regions and put a son in charge of each.94 This must have been Liu Chih’s suggestion, for we know that he was strongly in favor of restoring the ancient “feudal” system of the Chou period and regarded it as a great mistake that T’ai-tsung had not persisted in his plan to establish the T’ang empire on this basis. The policy proved to be a costly fiasco; the only prince to go to his post (“fief”), Prince Lin, who was sent to the south, promptly made plans to set

93  Yang Kuo-chung, who was widely blamed for having been responsible for the rebellion, feared the proximity to the capital of the large, immobile army of Ko-shu Han defending the western end of the T’ung-kuan. He therefore used every effort to have Ko-shu Han ordered to advance, against his better judgment, on the rebels camped at the eastern end. When, as a result, Ko-shu Han did advance, his army was completely routed, leaving the road to Ch’ang-an undefended. tctc, ch. 218 (Chih-te 1/ before 6/kuei-wei). The part played by Li Hua, Fang Kuan, and Liu Chih in trying to prevent Ko-shu Han from being forced to advance is not mentioned in the tctc and has been largely overlooked by historians. Li Hua describes it in his funeral address on Liu Chih (see note 90), and there is an obscure reference to unaccepted memorials in Li’s biography in hts, ch. 203, p. 1a, which seems to refer to it. There seems no reason to doubt the truth of Li Hua’s statement, which would give additional reason for the extraordinary trust placed in Fang Kuan by both Hsüan-tsung and Su-tsung thereafter. 94   t ctc, ch. 218 (Chih-te 1/7/ting-mao).

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

Neo-Confucianism in T ’ ang Intellectual Life

193

up an independent empire based on the Yangtze, and had to be put down by an expeditionary force.95 Later Fang Kuan was sent to join the new emperor, Su-tsung, at Feng-hsiang; he enjoyed such high favor that in spite of being a civil official, he was put in charge of the army and allowed to plan the campaign against the rebels. Once again he adopted a plan based on literal imitation of “the way of the former kings”; he went into battle with chariots drawn by oxen. The result was a disastrous defeat in which the imperial army suffered 40,000 casualties. Liu Chih was Fang Kuan’s staff officer (ts’an-mou) for this operation and, as Kanai surmises, it is extremely likely that he was responsible for the fanciful tactics.96 Although we do not possess any of Liu’s writings on war, we know that, apart from the probable inclusion of a section on war in the Cheng-tien, he wrote two works on military theory.97 Fang Kuan did not immediately lose office because of this disaster, but his influence not unnaturally declined. He ceased to attend to affairs of state and gave himself over to pleasure in the company of his scholar associates. They remained in power long enough to protect Li Hua in the collaboration trials after the recapture of the two capitals, but in the middle of 758 they were all dismissed to provincial posts,98 and Liu Chih died soon after.99 At this time Tu Yu100 was a young man living amid the intellectual ferment of the lower Yangtze, but it was somewhat later that Liu Chih’s book came to his notice, or, at any rate, that he conceived the idea of enlarging it. As a thinker, Tu Yu stands apart from most of his contemporaries. He has been called a 95  A brief account of Prince Lin’s adventure is given by Waley in The Poetry and Career of Li Po (London, 1950), pp. 78–79. 96   t ctc, ch. 219 (Chih-te 1/10/kuei-wei and hsin-ch’ou). See Kanai, op. cit., p. 91. In spite of these disasters, Fang Kuan, who represented the literatus element at a court dominated otherwise largely by eunuchs and military men and was a friend of the poet Tu Fu as well as such men as Li Hua and Liu Chih, has retained his reputation as a paragon of Confucian virtue. This is reflected in the accounts of him by Florence Ayscough and William Hung in their biographies of Tu Fu. 97  The Chih ko chi, 7 ch., and the Chih-te hsin i, 12 ch., both now no longer extant. See cts, ch. 102, p. 7a, and hts, ch. 59, p. 15a. Cf. Kanai, op. cit., p. 91. 98   t ctc, ch. 220 (Ch’ien-yüan 1/6 wu-wu). 99  See Li Hua’s funeral address (note 90). This document is not dated, but it says that Liu Chih died in his place of banishment, just when he was about to be promoted to a more important place. 100  Biographies in cts, ch. 147; hts, ch. 166; inscriptions, etc., about him by Ch’üan Te-yü in ctw, ch. 496, pp. 4a ff., ch. 505, pp. 4a ff., ch. 509, pp. 14a ff. On Tu Yu as a thinker, see Naitō Torajiro, “Ni ts’e i tao,” in Kanō kyōju kanreki kinen Shinagaku ronsō (“Chinese Studies in Honor of Professor Kano”), pp. 5–8, and Kanai Yukitada, op. cit., pp. 96 ff. Cheng Hosheng, Tu Yu nien-p’u (Shanghai, 1934), is useful but not very thorough.

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

194

Pulleyblank

Legalist. Whether or not this is an altogether fitting description, he has undeniable Legalist affinities. His appreciation of political planners and his belief that one should not attempt to imitate antiquity but should adapt the essential truths of older writings to the needs of the present are qualities which are usually associated, if not necessarily with pure Legalism, at least with the Legalist pole within Confucianism. His views as expressed in his encyclopedic history of governmental institutions, the T’ung-tien, have a refreshing down-toearthness and practicality combined often with a clear insight into the nature of contemporary problems and their historical background. Tu Yu came from an aristocratic northwestern family. Although a man of great learning, he did not practice belles-lettres and was almost ostentatious in referring to his lack of literary ability. He entered service through yin privilege and was made an Administrator (ts’an chün) in Chi-nan Commandery (Li-ch’eng in Shantung). The time must have been just before the rebellion of An Lu-shan. His next post was in the south at Shan-hsien (Sheng-hsien in Chekiang). By his knowledge of the law he attracted the notice of his father’s old friend Wei Yüan-fu,101 the Prefect of Jun-chou, who made him his Legal Administrator (Ssu-fa ts’an chün). We have no precise dates for Tu’s appointments in the southeast nor any information about how he came there, but it is likely that he was originally a refugee. It is, of course, possible that he received his appointment at Shan-hsien before the rebellion broke out, but the ten-year interval before the first recorded date, 765, is rather long, and it is probable that there was a period of unemployment in the first part of it. Significantly his career thereafter was in the service of a governor in the region where refugees were congregating. When Wei Yüan-fu was made Inspector of Che-hsi (with headquarters at Jun-chou) in 765,102 Tu Yu remained on his staff. At this time he began the T’ung-tien, working on it from 766 to 801 before presenting it to the emperor.103 Although his completed work belongs to the end of the eighth century and 101  Biography in cts, ch. 115. 102   See Wu T’ing-hsieh, T’ang fang-chen nien-piao, p. 134.2 (Erh-shih-wu shih pu-pien, Shanghai, 1936, p. 7416.2). 103  The date of presentation of the T’ung-tien is given as Chen-yüan 17 (801), in cts, ch. 13, p. 18b, and as both Naitō (op. cit.) and Tamai Zehaku (“Daitō rikuten oyobi Tsuten no Sō kampon ni tsuite,” Shina shakai keizai shi kenkyū (Tokyo, 1942, pp. 429–63) have shown, there are good reasons for accepting this date in preference to the date 803 given in thy, ch. 39, p. 660 (see note 87) or the date Chen-yüan 10 (794) given in the text of the memorial of presentation as recorded in some editions of the T’ung-tien. In this memorial he states that he has worked on the book for thirty-six years, which would mean that he had begun it around 765 or 766. This in turn is consistent with an earlier preface to a preliminary

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

Neo-Confucianism in T ’ ang Intellectual Life

195

made its impact then, we are no doubt right to think of it primarily as a product of the post-rebellion period. Tu Yu had a distinguished career as an official mainly in the provinces. His one important service in Ch’ang-an was in 780 when Yang Yen became Chief Minister and initiated his financial reforms. Yang summoned Tu Yu to the capital and entrusted him with various financial offices, in particular the management of supplies for military operations then under way. Finally he was put in charge of the Public Revenue Department. While in this office he attempted a rationalization of some of the miscellaneous functions that this department had accumulated since the rebellion. He also seems to have been partly responsible for an effort to raise emergency revenues by forced loans on merchants. When Lu Ch’i drove Yang Yen from power, in 782, Tu Yu was also dismissed and for the next twenty years served mostly in provincial posts. In 789 he became Military Governor of Huai-nan, an important province at the lower end of the Transport Canal—a post which his patron Wei Yüan-fu had once held. Except for a break after his mother’s death, he occupied this post until 803, when he was summoned to court and made a titular Chief Minister with the high honorific rank of Ssu-k’ung. Now sixty-eight years old, an elderly and distinguished statesman, he published his Li-tao yao-chüeh in ten chapters, a work consisting of thirty-three essays on political science which seem to have corresponded in the main to the discussions on various topics scattered through the T’ung-tien.104 draft of the work, written by Li Han, printed at the front of modern editions of the T’ung-tien.  The Ch’ing historian Wang Ming-sheng, who was hostile to Tu Yu, minimized his achievement by pointing out that he had had Liu Chih’s text as a basis and that the bulk of the additional material consisted of the Ta T’ang K’ai-yüan li incorporated bodily into the text. It is impossible to tell how much of the T’ung-tien is in fact taken over from the Cheng-tien, but the prefaces and author’s discussions, as well as much of the double-line commentary, are unquestionably his, and it is there that we have the basis for assessing his stature as a historian and thinker. (Shih-ch’i shih shang-ch’üeh, ch. 90, pp. 10a ff. Kuangya shu-chü ts’ung-shu.) 104  This is shown by quotations from it in Wang Ying-lin, K’un-hsüeh chi-wen, ch. 5, p. 20b; ch. 6B, p. 20b; ch. 14A, p. 1b; ch. 14B, p. 7a (sptk). Tu Yu’s preface and memorial of presentation are quoted in Yü hai, ch. 51, p. 28b. There we learn that it had 33 p’ien distributed among the 10 chüan under the following headings: ch. 1–3, Shih-huo (“Economics”); ch. 4, Chü, ming-kuan (“Examinations, Appointment of Officials”); ch. 5, Li chiao (“Rites and Instruction”); ch. 6, Feng-chien, chou-chün (“Feudalism,” “Administrative Geography”); ch. 7, Ping hsing (“Military Affairs, Criminal Law”), ch. 8, Pien-fang (“Frontier Defense”), ch. 9–10, Ku-chin i-chih i (“Discussions on the Differences in Institutions Between

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

196

Pulleyblank

Tu Yu was certainly a law specialist, whether or not he can be called a Legalist in a strict sense. His admiration for the Legalist planners of antiquity, including, for instance, Lord Shang who laid the foundations for the rise of Ch’in, could hardly have been countenanced by a strict Confucianist.105 He admired Legalist statesmen for the way they developed institutions suited to the conditions of their own days, untrammeled by the imitation of past models; ancient books should be consulted for the basic principles they propounded rather than for ideal patterns. He himself valued the economic ideas of Kuan-tzu, but speaking of that work, he remarked: “Whenever one consults the books of the ancients, it is because one wishes to reveal new meanings and form institutions in accordance with present circumstances. Their Way is inexhaustible. How much more are plans for contrivances and expedients subject to a thousand changes and ten thousand alterations. If one imitates in detail, it is like notching a boat to mark a spot …”106 This may seem a self-evident principle, but it was not understood by a great many of the restorers of antiquity. The historian Liu Chih-chi had been opposed to mechanical imitation of past models in historiography.107 Nevertheless, as we have seen, his son Liu Chih was sometimes a literalist concerning the ways of former kings; nor was he alone in this. A good deal of the impetus toward the revival of Confucianism came from ritualists who believed that regeneration could only be achieved by literal adherence to the precepts of the Rites. Liu Tsung-yüan tells the story of a Mr. Sun who brought ridicule on himself by trying to restore the practice of announcing at court the coming of age, “capping,” of sons. When he presented his bamboo tablet and said to the assembled courtiers, “The capping of my son so-and-so is completed,” the Governor of the Capital stood in his path and, trailing the tablet behind him, angrily said, “What has it to do with me?”—to the great amusement of the assembly.108 Tu Yu was not bemused by any such phantoms of the past. Chu Hsi called his Li-tao yao-chüeh “a book that makes the past wrong and the present right.”109 Tu’s introduction to the section on barbarians in the T’ung-tien gives an idea of his views on the progressive development of world history. He attributes Ancient and Modern Times”). This is similar, though not identical, to the arrangement of the T’ung-tien. Cf. Naitō, op. cit. 105   t t, ch. 12, p. 71.3. Cf. Arthur Wright, “The Formation of Sui Ideology,” in Chinese Thought and Institutions (Chicago, 1957), p. 81. 106   t t, ch. 12, p. 68.1. 107  See especially the section entitled “Mo-ni” (“Imitation”) in Shih-t’ung, ch. 8. 108  Letter to Wei Chung-li, Liu Ho-tung chi, ch. 34, ts’e 5, p. 4. 109  Quoted in Yü hai, ch. 51, p. 29a.

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

Neo-Confucianism in T ’ ang Intellectual Life

197

China’s advancement compared with the surrounding peoples to its central geographical position and favorable conditions of life: “The people’s nature is mild, and their wits are intelligent. The products of the earth are abundant, and its creatures multiply after their kind. Therefore it gave birth to sages and worthies who, one after another, gave forth laws and teaching. According to the needs of their times, they remedied evils and adapted things to profit and use.” He goes on to explain that in the Three Dynasties (Hsia, Shang, and Chou) the country was divided into warring states until Ch’in suppressed the feudal lords and established bureaucratic government. Nor was this the end of the story, for population had increased through the Han, Sui, and T’ang, and the achievement of peace and increase of population were the proper objectives of government. He criticizes the Taoist picture of a golden age of simplicity before the creation of social institutions. While he sympathizes with the Taoists’ satire on the degenerate sophistication of civilized life, he feels that their view of antiquity is unrealistic. He maintains that the ancient Chinese lived very much like the barbarians—and he proceeds to list customs such as living in caves or in the open and failure to practice clan exogamy, which were recorded of the ancient Chinese and were also known among the barbarians of his own day.110 In thus giving a positive value to the social institutions invented by the sages, Tu Yu’s position agrees with that of a Confucian like Han Yü in the “Yüan tao.” It differs in (1) substituting an environmental explanation for the element of magic or mystique in the sages, (2) making progress continuous and cumulative, and including such men as Shang Yang and even Ch’in Shih-huang-ti among its agents, (3) adopting a purely materialistic criterion for judging human welfare. Confucius had given due importance to the people’s livelihood, but his followers tended to talk about more spiritual things. Whereas the ku-wen reformers wanted to use literature as a means of promoting “civilizing transformation” or “culture” (chiao-hua), a favorite expression of theirs, Tu Yu says bluntly, “The beginning of the Way of Good Government lies in putting into practice civilizing transformation; the basis of civilizing transformation lies in bringing about a sufficiency of food and clothing.”111 One subject on which Tu Yu differs most noticeably from his model Liu Chih is “feudalism” (feng-chien). As we have seen, Liu Chih, before the An Lu-shan rebellion, strongly advocated this institution hallowed by association with the sage kings. Afterward, with the military governors often likening themselves 110   t t, ch. 185, p. 985. Compare also tt, ch. 48, p. 279; ch. 58, p. 337. Cf. Naitō, op. cit., Kanai, op. cit., pp. 127–31. 111   t t, ch. 1, p. 9.1.

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

198

Pulleyblank

to the “feudal lords” of the Chou dynasty, the faults of the feng-chien system became obvious. Yet feng-chien terminology was constantly used in a literary way of the military governors, and some optimistic literati may have felt comforted by the historical analogy. They may even have looked to the perfecting of a feng-chien system like that of Chou as a solution to the dynasty’s difficulties. For Tu Yu at least the issue was clear. The feng-chien system did not, he says, originate by a deliberate act on the part of the sages. It was the result of custom. The original feudal lords were not chosen by the Son of Heaven but were simply the chiefs of small communities. Originally these were very numerous, but the numbers were reduced in successive dynasties until Ch’in abolished them altogether. Liu Chih and others before him had advocated a feng-chien system as a defense for the imperial house, since if the ruler enfeoffed his sons there would be support in the country when any other family attempted to usurp. The long duration of Chou in comparison to Ch’in was offered as proof. Tu Yu does not directly refute this argument. He says that if the criterion is to be the duration of the royal house, then feudalism may be the answer, but there will be war and the population will be sparse; if the criterion is the populousness of the people, then there must be a bureaucratic system of provincial government. “To set up feudal kingdoms benefits one house. To arrange commanderies benefits all the people.” And he adds, “All laws and ordinances that are established must decay. It is only a question of estimating how long they will go on doing harm [before they can be reformed].”112 To find a Chinese scholar in imperial times stating so frankly that there can be a contradiction between the interests of the ruling house and those of the people is remarkable. Those who know Liu Tsung-yüan’s works will have no difficulty seeing in this discussion the inspiration of the “Feng-chien lun” (“Discourse on Feudalism”). Liu, like Tu, treats “feudalism” not as a perfect institution deliberately conceived by the ancient sages, which the ancient kings had been unable to do away with all at once, but as an unfortunate result of society’s natural evolution. Liu’s essay is even more remarkable than Tu’s, for he gives a more fully worked out theory of society’s origins. Rulers first arose, neither as divinely inspired sages nor as men of superior endowment produced by the favorable environment of the Central Land, but as the result of conflict and struggle which led people to submit to those who could settle their disputes. This happened first on a small local scale; then the conflicts of localities and groups led to the formation of larger and larger political units until the whole empire had been brought under one man. The influence of Hsün-tzu’s conception of men’s 112   t t, ch. 31, p. 177.

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

Neo-Confucianism in T ’ ang Intellectual Life

199

desires giving rise to conflict as the prime factor necessitating social organization is obvious and explicitly acknowledged, but Liu Tsung-yüan goes much further than Hsün-tzu in giving human society a naturalistic origin instead of looking on ancient institutions as the deliberate creations of sages, of men with more than normal intelligence and vision. Liu’s account of feudalism’s evils and its breakdown follows and elaborates Tu Yu’s argument.113 Tu Yu’s ideas on the selection of officials and on economics are of great interest. The problem of the examinations intimately affected all bureaucrats and would-be bureaucrats. The examination system was sufficiently new—in its Sui-T’ang form—so that both its detailed workings and its basic principles were matters of debate. Scarcely anyone writing on it fails to adopt a critical tone, but there are differences in the criticisms, often only subtle nuances, that stem from different interests and outlooks. The debate is, of course, reflected in the “Treatise on Examinations” in the New T’ang History.114 There is much more material in the relevant chapters of the T’ung-tien, and a full translation of these would be of great interest. Tu Yu, who had entered office through hereditary privilege and had no literary pretensions, was among those who thought the examination system had opened the door much too wide and accepted men for accomplishments irrelevant to the business of government. He would have tried rigorously to reduce the number of official posts at the capital and leave it to senior officials to appoint their own staffs or make recommendations.115 Critics who were more closely identified with literature and classical scholarship were more inclined to emphasize the hardships of the candidate under the existing system. For instance, Chao K’uang, already mentioned as a follower of Tan Chu’s school of classical studies, appreciated the relative equality of opportunity provided by a centralized examination system and proposed a scheme whereby candidates would write their essays in the provinces to be forwarded to the capital for grading. He also wished to limit the opportunities for non-scholars to get official posts without examination, and to ensure 113   Liu Ho-tung chi, ch. 3. In the Sung period Liu’s essay was generally considered to be the best discussion on the question. Among those who expressed approval were Sung Ch’i, in the historian’s tsan at the end of hts, ch. 168; Su Shih in “Shih-huang lun, Chung,” Chingchin Tung-po wen-chi shih-lüeh (Peking, 1957), ch. 24, p. 201; and Fan Tsu-yü, in T’ang chien, ch. 4, p. 27 (Ts’ung-shu chi-ch’eng). A certain Mr. K’ung also praised the “Feng-chien lun” as being something to which there was nothing comparable in Han Yü’s works. See the commentary to the title in Liu Ho-tung chi, loc. cit. 114  R. des Rotours, Le traité des examens traduit de la Nouvelle Histoire des T’ang (Paris, 1932). 115   t t, ch. 18, p. 104.

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

200

Pulleyblank

that those who did get in were grounded in classical studies.116 Lu Hsüan-kung did not think the examination system had opened the door too wide. He told Te-tsung on one occasion that, although the Empress Wu (under whom the examination system had for the first time brought a considerable number of new men into the bureaucracy) had no doubt gone too far in making advancement easy, her arrangements for the judgment of character and achievement had in fact brought many good officials to the top. Unlike Tu Yu, he felt that under Te-tsung it had become too difficult to enter officialdom.117 Lu proposed to introduce into the system of promoting officials a formal procedure of recommendations by the heads of bureaus who would then be held responsible for their protégés and receive merit or demerit on the basis of the grades the protégés received in the annual examination of merit. This is not the place to discuss fully the T’ang beginnings of the sponsorship system, which was to play an important role in the Sung.118 To Lu Hsüan-kung it seems to have commended itself as a public procedure for avoiding the evils of secret favoritism. When complaints were made that sponsorship meant favoritism, the emperor wished to entrust all promotions to Lu Hsüan-kung as Chief Minister, but Lu would not accept this responsibility. He defended his plan, which was, however, rescinded.119 The conflict between literatus and aristocrat is again to be seen in the contrast between the economic ideas of Lu Hsüan-kung and Tu Yu. Lu Hsüan-kung is famous for his memorials criticizing Yang Yen’s tax reforms of 780;120 whereas Tu Yu, who had been associated with Yang Yen, continued to take a favorable view of the measures. A full understanding of the differences in the theoretical outlook of Tu Yu and Lu Hsüan-kung would require an intensive study of their backgrounds in earlier Chinese economic thought, particularly as found in the Kuan-tzu, for clearly the two men shared many basic conceptions. Both had 116   t t, ch. 17, p. 97. 117   t ctc, 234/5/jen-tzu. 118  E. A. Kracke, Jr., Civil Service in Early Sung China (960–1067) with Particular Emphasis on the Development of Controlled Sponsorship to Foster Administrative Responsibility (Cambridge, Mass., 1953). 119   t ctc, loc. cit. There was of course nothing new in the idea that high officials should recommend people for office and that they should be held responsible for misdemeanors of those whom they had so recommended. It is the institutionalizing of this as a part of the regular procedures for appointing officials which seems to be an innovation of the postrebellion period. Lu Hsüan-kung’s proposal is not the first time we hear of it; it appears to be only a more extensive use of a procedure that was already being introduced. We find decrees on the subject from 766 onward. See tfyk, ch. 630. pp. 13a, 14b, 15a, etc. 120   c tw, ch. 465. Translation by Balazs, op. cit., Part 3; msos, ch. 36 (1933), pp. 1–41.

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

Neo-Confucianism in T ’ ang Intellectual Life

201

the same physiocratic ideal of a basically agrarian and equalitarian economy, and they drew on the same or similar maxims to illustrate their points. Yet there were profound differences of emphasis. Lu’s economic theory was conservative and traditional. He admitted that after the rebellion of An Lu-shan the fiscal situation of the country had become thoroughly unsound and in need of reform, but he maintained that to remedy present evils Yang Yen should have restored rather than eliminated the fundamentally good system instituted by the founders of the dynasty and based on principles hallowed in antiquity. He attacked the principle of taxing property and production rather than levying equal taxes on all individuals and households. It would take too long to discuss all his ideas in detail, but in general they favor relaxation of state interference to let the economy find its own harmony, and intervention only to suppress abuses such as the encompassment of small holdings in large estates (chien-ping) or to ensure stability in times of need by use of granaries. He was particularly opposed to Te-tsung’s tendency to rely on unscrupulous financial officials who would raise money by all kinds of extralegal means. This stand finally brought about his downfall, for Te-tsung was determined to enrich himself, and when it came to a showdown in 794, he preferred P’ei Yen-ling to Lu Hsüan-kung.121 Tu Yu, on the other hand, argued that it was of the highest importance for the state to see that the people were all settled on the land and registered in localities, and he expressed his admiration for statesmen of the past such as Kao Chiung of the Sui122 and Yü-wen Jung of the K’ai-yüan period (713–42)123 who had sought to effect this. He distinguished them from other officials who had devised unscrupulous schemes for enriching the treasury. Tu Yu praised Yang Yen’s reform primarily from this point of view— because it reduced somewhat the chaos into which the registers had fallen and increased the number of people accounted for on the rolls.124 Lu Hsüankung, however, although he wanted the people settled on the land, was with those who felt that it was best to foster conditions that would make them content and felt that coercive methods would be self-defeating. These different 121   t ctc, ch. 235 (Chen-yüan 10/11/jen-shen, 12/jen-hsü). 122   t t, ch. 7, p. 42.2. 123   t t, ch. 7, p. 41.1. Tu Yu’s favorable treatment of Yü-wen Jung is in marked contrast to that of Su Mien in the Hui-yao and of the official histories. See The Background of the Rebellion of An Lu-shan, pp. 30 ff.; Suzuki Shun, “Ubun Yū no kakko ni tsuite,” Wada Hakushi kanreki kinen Tōyōshi ronsō (“Oriental Studies Presented to Professor Wada”) (Tokyo, 1951), pp. 329–44. 124   t t, ch. 7, p. 42.3.

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

202

Pulleyblank

viewpoints reflect the contrast between Hsün-tzu’s view that human nature is naturally bad and must be coerced, and Mencius’ view that man is naturally good and will respond to good treatment. On a more theoretical level there appears to be a difference in Tu Yu’s and Lu Hsüan-kung’s views on the proper basis of taxation which reflects their different attitudes to Yang Yen’s reform. As already mentioned, Lu advocated taxation on the basis of the individual person and household and decried the principle of taxes based on production or property. But Tu Yu is at pains to argue that in ancient times good systems of taxation were always based on land and not on men.125 Tu Yu shows his Legalist bias in his attitude to merchants; he stresses the primacy of agriculture but, like the Legalist-minded great officials of the “Discourses on Salt and Iron” in the Han Dynasty,126 shows little sympathy for commerce. Mainly his references to it emphasize the necessity of curbing the merchants’ tendencies to exploit the economic weakness of the peasants, to accumulate fortunes and upset the equilibrium of society. He does not condemn the extraordinary levies on merchants made in the An Lu-shan rebellion, commenting, “It was the proper thing in a time of emergency.”127 He was himself partly responsible for a similar measure in 782.128 With this attitude we may perhaps contrast the view expressed in 822 by Han Yü on a proposal to change the organization of the salt monopoly by having officials instead of merchants handle distribution and sales. Han Yü did not exactly praise the salt merchants but he recognized their position as a vested interest cooperating with the state.129 Any difference of attitude toward merchants among the literati was, to be sure, a matter of nuance, but I think one may see here a difference between the lofty and distant contempt expressed by aristocrats like Tu Yu and the more realistic evaluation of merchants by Confucians of less exalted ancestry.

The Reform Party of Wang Shu-Wen

The year 803, in which Tu Yu came to the capital and became a titular Chief Minister, is an important one, for just at this time a reform party with the 125   t t, ch. 4, p. 25. 126  See Esson M. Gale, Discourses on Salt and Iron (Leiden, 1931). 127   t t, ch. 11, p. 63.1. 128   t ctc, ch. 227 (Chien-chung 3/4/chia-tzu). 129   Han Ch’ang-li chi, ch. 40, ts’e 7, p. 55. Cf. tctc, ch. 242 (Ch’ang-ch’ing 2/4/chia-hsü).

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

Neo-Confucianism in T ’ ang Intellectual Life

203

Crown Prince as its center begins to be active in political affairs, evidently in the expectation that he would soon succeed his father. There was widespread resentment of various abuses that had grown up in the latter years of Te-tsung’s reign as a result of his increasingly autocratic and personal rule. His autocratic temper, combined with his avarice, led him to treat the bureaucracy with suspicion and to rely on favorites who used unscrupulous methods to fill the treasury (and their own pockets). These dissatisfactions were combined with a wish to strike radically at the cankers of eunuch influence and militarism which had been undermining the state for the past half-century. The leaders of the reform group were two men of the Crown Prince’s personal entourage, Wang Shu-wen and Wang P’ei, and a Secretary in the Ministry of Civil Office, Wei Chih-i, who had once been a Han-lin Scholar and was now the leader of the liberal element in the bureaucracy. These three gathered around them a number of the most brilliant young men at Ch’ang-an, including, as we have seen, Liu Tsung-yüan and other followers of the ku-wen movement in literature and of the new critical movement in scholarship. In 805 Te-tsung died, Shun-tsung succeeded to the throne, and for a short time the group controlled the government. Unfortunately for their plans, Shun-tsung had shortly before suffered a stroke, making him an invalid unable to speak. This misfortune and the opposition from eunuchs, military governors, and more conservative members of the bureaucracy ensured their downfall after only a few months. The abdication of the emperor in favor of his son, posthumously known as Hsien-tsung, was engineered, and the members of the reform group were convicted of seditious conspiracy and banished to lowly offices in the remote south.130 The most severely dealt with were Wang Shu-wen and Wang P’ei, men of undistinguished family background and no connections; they have both been vilified in the official histories. Wang P’ei131 was a calligrapher from Hang-chou—a man of ugly appearance with a southern accent which made him ridiculous in the eyes of the northern literati. Wang Shu-wen,132 who came from nearby 130  The Shun-tsung shih-lu, which deals with these events, has been translated by Bernard S. Solomon as The Veritable Record of the T’ang Emperor Shun-tsung (Cambridge, Mass., 1955). It purports to be by Han Yü but is certainly not now in the form which finally left his brush. As I have shown elsewhere, there is some reason to think that it may be only the preliminary draft by Wei Ch’u-hou which Han Yü revised and enlarged (bsoas, xix, 1957, 336–44). The question is not unimportant, for inferences about Han Yü’s attitudes have frequently been made from it. 131  Biographies in cts, ch. 135; hts, ch. 168. 132  Biographies in cts, ch. 135; hts, ch. 168.

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

204

Pulleyblank

Yüeh-chou (Shao-hsing), had initially won the favor of the Crown Prince as an expert in the game of wei-ch’i but had more claims to scholarship. Wang Shuwen was put to death after being exiled; this may also have been the fate of Wang P’ei, who is not heard of again. Wei Chih-i133 came from an aristocratic family and his connections protected him for a time, but in the end he was banished to the most pestilential place in the empire, Hainan island, where he eventually died.134 Liu Tsung-yüan and the other lesser criminals were also banished, but not quite so far. In 815 they were pardoned to the extent that they were made prefects of chou, but they were sent to remote spots.135 The Crown Prince seems to have played an active part in gathering these men around him. If we may trust the biography of Wei Chih-i in the Old T’ang History, Wei was first introduced in 785 to Wang Shu-wen by the Crown Prince.136 There is evidence that in the early years of Te-tsung’s reign the Crown Prince, already a grown man, had at his court leading Buddhist monks debating on conflicts of doctrine,137 and we note that Liang Su for a time held the post of Reader to him.138 If we could discover more about the Crown Prince’s personality and his entourage during his father’s reign, we might be in a better position to understand the events of his own reign. The reformist aims of the group seem clear. In general, their official condemnation as a seditious faction was accepted thereafter, but already during the Sung, Hung Mai expressed doubts. He pointed out that during their brief period of power they had executed a number of highly commendable measures, including: (1) steps to curb the oppressive behavior of the eunuchs, procuring goods for the palace in the markets of the capital, (2) steps to prevent the youths of the eunuch-controlled palace stables from terrorizing the streets, 133  Biographies in cts, ch. 135; hts, ch. 168. In the writings of Li Te-yü there is a sacrificial piece (chi-wen) to Wei Chih-i. There is a difficulty about the date, since it begins, “In the fourth year of Ta-chung (850)” and according to the histories Li Te-yü died in the twelfth month of Ta-chung 3 (cts, ch. 18B, p. 8a; ch. 174, p. 11b; tctc, ch. 248 [Ta-chung 3/12/ chi-wei]). Assuming it to be genuine, however, it is a very interesting document. Since Li Te-yü was in exile on Hainan island, the victim of factional attack, it may be no more than an expression of sympathy for someone who had suffered a similar fate. Possible implications for the nature of the feud between Li and the Niu faction cannot be gone into here. 134  Solomon, op. cit., p. 56. 135   t ctc, ch. 239 (Yüan-ho 10/3/i-yu). 136   c ts, ch. 135, p. 12a. It is noteworthy that Liu Tsung-yüan said that his own association with the “guilty men” had gone on for ten years before they came to power. Liu Ho-tung chi, ch. 30, ts’e 4, p. 69. 137  See note 47 above. 138   h ts, ch. 202, p. 16a.

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

Neo-Confucianism in T ’ ang Intellectual Life

205

(3) abolition of the monthly “tribute” from the Salt and Iron Commission, (4) drastic reduction of the number of singing girls in the palace, (5) the recall of good officials such as Lu Hsüan-kung who had been exiled during the latter years of Te-tsung’s reign and kept in distant regions by the ten-year lack of amnesties. Furthermore, if their attempt to deprive the eunuchs of control of the Palace Armies (which was turned into the most serious evidence against them) had succeeded, later tragedies would have been prevented.139 The Ch’ing historians Wang Fu-chih140 and Wang Ming-sheng141 later expressed similar views, and, in recent times, several scholars studying the period have adopted this point of view.142 In retrospect we can see that their aim was to reform and strengthen the central government, and, not surprisingly, they incurred the inveterate hostility of the main body of military governors whom they failed to placate. What might seem less understandable at first sight is the opposition from within the ranks of the bureaucracy. Liu Tsung-yüan attributed it to consternation at seeing young men suddenly promoted to high office, and to the envy and disappointment of those who could not share in their good fortune. I think that this was undoubtedly true. Although there was antagonism between eunuchs and bureaucrats, eunuchs had by now become part of the established order, and relations of mutual dependence and benefit existed between individual bureaucrats and eunuchs. To the conservative-minded, the existing system was no doubt more tolerable than government by upstarts who might threaten their hard-won positions. The Wang Shu-wen group is indeed a classic example of what is often called a “faction” (p’eng-tang). In another volume in this series Mr. Nivison discusses “faction” as a recurrent problem in Chinese political theory.143 It was generally held that any combination was ipso facto wrong. Officials ought to “stand alone” (tu-li) and express their honest views for the benefit of the emperor. In the Sung period there were some who held that it was proper for superior 139   Jung-chai hsü-pi, p. 7.4b (sptk). 140   Tu T’ung-chien lun, p. 25. 1a ff. (sppy). 141   Shih-ch’i shih shang-ch’üeh. 142  See, for instance, Shigezawa Junrō, “Ryū Sōgen ni mieru Tōdai no gōri shugi” (“Rationalism in the T’ang Dynasty as Seen in Liu Tsung-yüan”), Nihon Chūgoku gakkai hō, iii (1951), 75–84; Huang Yün-mei, Han Yü Liu Tsung-yüan wen-hsüeh p’ing-chieh (“A Critical Introduction to the Literary Works of Han Yü and Liu Tsung-yüan”) (Chinan, 1957), first published in Wen-shih-che, 1954; Shimizu Shigeru, “Ryū Sōgen no seikatsu taiken to sono sanzuiki,” Chūgoku bungaku hō, ii (1955), 45–74. 143  David S. Nivison, “Ho-shen and His Accusers” in David S. Nivison and Arthur F. Wright, eds., Confucianism in Action (Stanford, 1959), pp. 209–43.

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

206

Pulleyblank

men to join forces, but even if this were admitted, it could only be on the basis of ethical principles, not of detailed policies. For men to combine and plan privately to try to effect policies would be to subvert the hierarchy of authority flowing down from the emperor. In practice of course this did not mean that men “stood alone.” Officials who in fact did so were an eccentric few who might win reputations for their Confucian virtue but were unlikely to rise rapidly on the bureaucratic ladder. To get ahead one had, not to combine with one’s colleagues, but to curry favor with persons of influence. An infinitely complicated web of one-to-one personal relationships took the place of definable interest groups and parties with programs. No doubt it is often possible to distinguish different interests expressing themselves now in one way, now in another, but it is extremely difficult in practice to correlate interests and personal groupings continuing over any length of time. Such a situation in political life, which is of course not at all restricted to China, naturally tends to a conservative maintenance of the status quo. However strongly a person might feel that certain reforms ought to be made, as long as he remained isolated and dependent for advancement on his superiors, he would be likely to find it advisable to curb his ardor and avoid antagonizing vested interests. By the time he had raised himself to a position of influence, he had become part of the system and was unlikely to try to alter it. It is no accident, therefore, that when we do find reformist movements, they are often initiated close to the throne outside the regular bureaucratic procedures, and often incur the stigma of “faction,” not distinguished in the eyes of Chinese historians from other factional groupings of a purely personal character. It is in the light of such considerations that we must, I think, judge the role of Tu Yu in the affair. It has, so far as I know, never been suggested that he had anything to do with the reform group. On the contrary, he was apparently involved in arranging for the abdication of Shun-tsung, which sealed its ruin; we know that in his role as elder statesman Tu Yu conveyed the abdication document from Shun-tsung to his son.144 The fact that during the reform group’s period of power Tu Yu was made Commissioner for Public Revenue and Salt and Iron, and given general authority over state finances, is explained away by the statement that he was merely a figurehead for Wang Shu-wen, who, as his deputy, kept the real power in his own hands.145 Yet it seems to me there is good reason to think that even if Tu Yu was never personally in the intimate counsels of the group, his ideas were influential among them and he was 144  Solomon, op. cit., p. 53. 145  Ibid., p. 19.

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

Neo-Confucianism in T ’ ang Intellectual Life

207

looked upon as a source of leadership and inspiration. Though he never seems to have been a particularly forceful man in practical affairs, in contrast to his daring as a thinker, Tu Yu’s eminence and prestige were such that he could hardly have been used against his will by Wang Shu-wen. There is every reason for thinking that he would have approved of the aims of the group, though he, like other senior statesmen, may have been shocked, or found it politic to be shocked, by the “excesses” and rash and hasty actions of the group when it became clear that they had taken on more than they could handle. As to the ideological influence of Tu Yu on the group, it should be noted that one of the young men, Liu Yü-hsi, had been for a long time on his staff in Huai-nan and came with him to Ch’ang-an. The publication of the T’ung-tien in 801 and the collection of essays from it, the Li-tao yao-chüeh in 803, seems to have caused a considerable stir in intellectual circles: Ch’üan Te-yü praised it fulsomely;146 and we have seen that Liu Tsung-yüan’s “Feng-chien lun” was certainly derived from Tu Yu. Later, while in exile, Liu expressed views about government economic policy closely resembling those of Tu Yu.147 Furthermore it seems to me not impossible that Han Yü’s “Yüan tao” (“Inquiry into the Way”) may also owe something to Tu Yu’s discussion about the development of civilization. The dates of this and other related philosophical works by Han Yü are not known precisely, but the first probable reference to them is in the year 805, and it seems to me likely that they were the product of his period of exile from the winter of 803 to the winter of 805, coming immediately after his short period in the censorate along with Liu Tsung-yüan and Liu Yü-hsi in the latter part of 803.148 146  See the commemorative inscription set up on Tu Yu’s departure from Yang-chou in 803 (ctw, ch. 496, p. 7a) and also his epitaph (ctw, ch. 505, p. 5a). The latter passage is quoted as by Ch’üan Te-yü, but without mention of the particular piece it comes from, in Yü hai, ch. 33. 147  Letter to Yüan of Jao-chou discussing government, Liu Ho-tung chi, ch. 32, ts’e 4, pp. 104 ff. 148  The only means of dating the “Yüan tao” and its companion pieces, the “Yüan hsing,” etc., appears to be what is very probably a reference to them in a letter to Li Sun in 805 (Han Ch’ang-li chi, ch. 15, ts’e 4, p. 44). See Chu Hsi’s commentary to the title of the “Yüan hsing” (ibid., ch. 11, ts’e 3, p. 63); also J. K. Rideout, “The Context of the Yüan Tao and the Yüan Hsing,” bsos, xix (1948), 408. This gives only a terminus ante quem, and it is of course possible that he had written these pieces somewhat earlier, but I am inclined to think they were produced during his period of exile. Rideout believed that the “Yüan hsing” in particular was a reaction to Li Ao’s “Fu hsing lun,” which he convincingly dates in 799–800 (op. cit., p. 406 n. 4). It seems to me to be possible to accept this and still leave room for the possibility that an additional stimulus may have come from contacts with Liu Tsung-yüan and Liu Yü-hsi in 803, and possibly also from Tu Yu’s writings.

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

208

Pulleyblank

That Han Yü and his friends were actively discussing philosophical problems around this time seems to be indicated by Liu Tsung-yüan’s “T’ien shuo,” in which he expresses his view of the indifference of heaven to human affairs.149 The essay purports to be a record of a conversation between himself and Han Yü, and is referred to by Liu Yü-hsi in terms which also imply this. If we accept this, the occasion cited must have been either before 794, when Liu Tsung-yüan, only in his twenty-second year, left Ch’ang-an on a two-year visit to his uncle in Pin-chou to the west,150 or after 801, when Han Yü returned to Ch’ang-an after serving on the staffs of military governors in the east since 795.151 I think the latter period is much more probable. Liu Yü-hsi at some time after 805 wrote his “T’ien lun” as a development of Liu Tsung-yüan’s essay.152 It is not necessary to suppose that he was present at the conversation between Han Yü and Liu Tsung-yüan, but his interest in the matter suggests that he may have been. At any rate, he came to Ch’ang-an along with Tu Yu in 803, and Han Yü later referred to his intimacy with the two Lius at this time.153 The association with Tu Yu need not mean that the Wang Shu-wen group is to be regarded as pro-Yang Yen and anti-Liu Yen, for they included in their number Ch’en Chien, who had once been on Liu Yen’s staff and had written a eulogy of him and his policies.154 Moreover, one of the first acts of Shun-tsung’s reign was to rehabilitate Lu Hsüan-kung and other victims of the tyranny of Te-tsung’s later years. Probably that old factional dispute had died out. Han Yü’s personal involvement with the reform group is both interesting and important. Though he was a personal friend of Liu Tsung-yüan and Liu Yü-hsi, he does not seem to have been brought into the censorate as their colleague in 803 by the Wang Shu-wen party. Indeed, shortly before bis appointment he wrote an obsequious letter of flattery to Te-tsung’s corrupt and brutal

149   Liu Ho-tung-chi, ch. 16, ts’e 3, p. 63. 150  Shih Tzu-yü, Liu Tsung-yüan nien-p’u (Wu-han, 1958), p. 15. 151  Biography with commentary by Chu Hsi, Han Ch’ang-li chi, ts’e 8, pp. 30–31. 152   c tw, ch. 607 (see also Liu Ho-tung chi, ch. 16). That this was written when the two Lius were no longer together is indicated by the fact that Liu Yü-hsi sent his “T’ien lun” to Liu Tsung-yüan by letter and Liu Tsung-yüan wrote a letter back to him about it (Liu Ho-tung chi, ch. 31, ts’e 4, p. 97). Shih Tzu-yü is no doubt right in assigning this letter to the period of exile and he may also be right in putting the actual writing of the “T’ien shuo” after Liu Tsung-yüan had gone to Yung-chou, though it is merely a subjective judgment. (Op. cit., pp. 83–84.) 153  See note 158 below. 154  See note 12.

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

Neo-Confucianism in T ’ ang Intellectual Life

209

favorite, Li Shih,155 and, embarrassing as the suggestion is to those who see him as a model Confucian, it is not impossible that this was what brought about his appointment. He did not last long in the censorate, for a few months later he and two others were suddenly expelled to govern counties in the south.156 The occasion for this expulsion seems to have been a memorial they sent drawing attention to the people’s distress that resulted from the current drought,157 and it has been assumed by traditional commentators that Li Shih, who had been collecting taxes as usual, was offended and brought about their dismissal. Others have pointed out, however, that Han Yü himself doubted that the memorial was the actual reason for their dismissal, for he seemed to blame it on Wang Shu-wen. He even suggested that his friends Liu Tsung-yüan and Liu Yü-hsi might have revealed private conversations. Although he did not long hold a grudge against them, he continued to feel hatred for their leaders, in whose downfall he exulted.158 155   Han Ch’ang-li chi, ch. 15, ts’e 4, p. 42. Cf. Chu Hsi’s note on Han Yü’s biography, ibid., ts’e 8, p. 31. Ch’ien Chi-po, in his Han Yü chih, p. 33, besides referring to this letter, quotes another letter to a certain Li Shang-shu (Han Ch’ang-li chi, ch. 19, ts’e 5) to show that Han Yü retained his friendship and regard for Li Shih. Chu Hsi’s commentary, however, argues that the Li Shang-shu in question in the latter case is Li Chiang, not Li Shih, and this must be correct. Li Chiang was made prefect of Hua-chou in 815, but there is no record of Li Shih’s having held this post. On the other hand, the severe criticism of Li Shih in the Shun-tsung shih-lu need not mean that Han Yü was deliberately showing ingratitude, or alternatively being vindictive. In the first place, it is doubtful whether he was in any way responsible for the existing version of the Shih-lu, and in any case he was working on the basis of an existing draft which he would hardly have dared to alter for personal reasons. 156   t ctc, ch. 236, Chen-yüan 19/12/-; Han Ch’ang-li chi, ts’e 8, pp. 31–32. 157   Han Ch’ang-li chi, ch. 37, ts’e 7, pp. 18–19. 158  Han Yü’s biographies in the histories attribute his dismissal to a memorial criticizing the abuses in purchasing for the palace in the markets of the capital, but there is no such memorial in his works. Li Ao’s hsing-chuang (ctw, ch. 639, pp. 22a ff.) says that he “was hated by a favorite of the emperor.” A memorial inscription by Huang-fu Shih mentions the memorial about famine conditions and says that he was in consequence hated by those in charge of the government (ctw, ch. 687, pp. 14a ff.). Han Yü’s epitaph for Chang Chien, one of the men who was banished with him, also says they were slandered by a “favorite” (Han Ch’ang-li chi, ch. 30, ts’e 6, pp. 41–43). It is a natural inference that the person offended was Li Shih. (Cf. Hung Hsing-tsu in Han Wen lei-p’u, ch. 2, pp. 4a ff.). On the other hand, in a long poem written at the end of 805, Han Yü refers to his memorial but says that it was appreciated both by the emperor and Tu Yu and wonders if his words have been reported to his enemies by his friends the two Lius (Han Ch’ang-li shih hsi-nien chi-shih, ch. 3, pp. 132 ff.). His hostility to Wang Shu-wen and Wang P’ei is made abundantly clear in other poems written shortly afterward (ibid., pp. 151 ff., pp. 171 ff.). See Ch’ien Chi-po, op. cit., p. 32; cf. Solomon, op. cit., pp. xiii, xiv.

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

210

Pulleyblank

Wang Shu-wen and Wei Chih-i probably were responsible for the expulsion in 803 from the capital of others whom they suspected of opposition to them,159 and if they regarded Han Yü as unreliable—as they may well have done, especially if he was friendly with Li Shih—they may have seized a pretext to get rid of him. Control of the censorate was particularly important for a group seeking to take over the government. There was certainly a sharp difference in ideas between Han Yü’s “conservative” ku-wen movement and the “radical” school which was dominant at the end of Te-tsung’s reign. In contrast to the interest of Liu Tsung-yüan and his associates in the new classical scholarship, Han Yü seems to have concerned himself very little with it. When he meets Lu T’ung, a follower of the Tan Chu school of criticism of the Spring and Autumn Annals, he is mainly interested in him as a recluse and a poet, and his comments on his scholarship, although appreciative, are made from the outside and show nothing of Liu Tsung-yüan’s fervor.160 Later he confessed to an adherent of the Kung-yang school of interpretation that he had paid little attention to such matters since preparing for his chin-shih examination.161 Chinese critics recognized that Han did not show anything like the critical acumen displayed by Liu Tsung-yüan;162 nor are his contributions to philosophy impressive. Most of the new ideas in this field attributed to him were already current among the ku-wen writers, and it is only because of his literary eminence and his efforts as a propagandist that they are known through him. It seems reasonable to suppose that it was this mediocre attainment in scholarship, though not in literary composition, that made people look askance at his setting himself up as a teacher and at his claim to renew the broken line of Confucian tradition. Although Han Yü’s advocacy of the teacher-pupil tradition and adoption of the role of teacher have been regarded as important points in his favor, in

159   t ctc, ch. 236 (Chen-yüan 19/9/chia-yin); Solomon, op. cit., p. 54. 160  Lu T’ung’s collected poems are extant (Yü-ch’uan tzu shih-chi, 2 ch., wai-chi, 1 ch., sptk). He lived as a recluse in Lo-yang, and Han Yü visited him when he was Magistrate of Honan in 810–11. Han Yü liked his long satirical poem “On an Eclipse of the Moon” and wrote an imitation, which, however, has been less admired than the original by the critics (Han Ch’ang-li shih hsi-nien chi-shih, pp. 324 ff.). He also wrote another poem to him, which is a principal source of information about him (ibid., pp. 340 ff.). The brief notice about Lu T’ung in hts, ch. 176, is based on this and appended to Han Yü’s biography. Lu T’ung wrote a commentary on the Ch’un ch’iu in 4 chüan, which survived into Sung. Fragments have been collected in the Nan-ching shu-yüan ts’ung shu. 161  Letter in reply to the Censor Yin, Han Ch’ang-li chi, ch. 18, ts’e 4, pp. 81–82. 162  Cf. Chu Hsi’s comment on “Tu Ho-kuan-tzu,” Han Ch’ang-li chi, ch. 11, ts’e 3, p. 73.

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

Neo-Confucianism in T ’ ang Intellectual Life

211

contrast to Liu Tsung-yüan’s rejection of this role,163 from a modern point of view this judgment should probably be reversed. The authoritarian character of the Confucian teacher-pupil relationship has had a baneful influence on the development of scholarship in the Far East. Liu Tsung-yüan’s attitude is both exceptional and enlightened. He refused to claim the superiority that qualified him to be a teacher of others; instead he invited prospective pupils to come and engage in free discussion with him for their mutual benefit. Han Yü compensated for his weakness in metaphysics and Confucian scholarship by a nationalistic rejection of Buddhism that contrasted sharply with the eclectic attitude of the earlier ku-wen movement. Their political attack on this “foreign intrusion” may have fostered in Han Yü and his followers an attitude of acceptance toward the existing political order, which Liu Tsungyüan and his “radical” friends regarded as urgently in need of reform. The disciples who acknowledged Han Yü’s leadership are not particularly distinguished but they seem to have maintained this opposition to Buddhism. We find Sun Ch’iao protesting at the restoration of Buddhism after its suppression in 845.164 There have been efforts to give a sociological interpretation of the contrasting positions of Han Yü and Liu Tsung-yüan. Marxist writers have tried to show that Han Yü represented the large landowners while Liu represented the small.165 This is far too crude. It is not at all easy to see in the affair of 805 any correlation between social background and political alignment. Professor Ch’en Yin-k’o’s view that the factional struggles of the following reigns reflect the conflict of interest between hereditary privilege of the aristocracy and the aspirations of new men has recently been subjected to severe criticism.166 This latter question must be left aside. Yet one may note the apparently paradoxical fact that in the general development of thought in the post-rebellion period, persons of aristocratic (mainly north-western) background were frequently rationalistic and unorthodox in their political thinking, while the new literati coming up through the examinations system, with whom the future lay, were more inclined to a moralistic traditionalism. 163  Han Yü’s point of view is expressed in “Shih shuo,” Han Ch’ang-li chi, ch. 11, ts’e 3, pp. 75–77. Liu Tsung-yüan’s is most fully given in his letter in reply to Wei Chung-li, Liu Ho-tung chi, ch. 34, ts’e 5, pp. 3–5. 164   c tw, ch. 794, pp. 6b ff. Sun Ch’iao, in a letter to a friend (ctw, ch. 794, pp. 14a ff.) describes himself as a disciple of Han Yü through Huang-fu Shih and a certain Lai Tse (style Wutse)—cf. Teng-k’o chi-k’ao, ch. 20, p. 4b (Nan-ching shu-yüan ts’ung-shu), and hts, ch. 50, p. 9a. 165  See Lü Chen-yü, Chung-kuo cheng-chih ssu-hsiang shih (Peking, 1955), pp. 397 ff., 413 ff. 166  Ts’en Chung-mien, Sui T’ang shih (Peking, 1957), chap. 45. Cf. Ch’en Yin-k’o, T’ang tai cheng-chih shih shu-lun kao (Shanghai, 1947), pp. 53–93.

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

212

Pulleyblank

Conclusion This essay has tried to place in perspective the thought, mainly the political thought, of the half-century following the An Lu-shan rebellion. I have been especially struck by the way the most significant intellectual movements can be traced back to the refugee scholars who congregated in the lower Yangtze region during the rebellion period. The next generation developed further the ideas of their predecessors and, carrying the movement back to Ch’angan, tried, if my picture is correct, to apply themselves to a renaissance of the dynasty. The defeat of Wang Shu-wen’s party in 805, which strengthened the power of the eunuchs and removed from political life some of the most ardent spirits, must have discouraged would-be reformers. The way eunuchs were able, thereafter, to murder one emperor after another added to the gloom, no doubt. In 835 came the disaster of the Sweet Dew incident. A plot against the eunuchs which had the connivance of Emperor Wen-tsung was discovered by the eunuchs; many officials were massacred and the three Chief Ministers were put to death as traitors. It is small wonder that the literati retreated more and more into the consolations of poetry and belles-lettres, and into Taoism. If the philosophy of the period I have discussed was attempting to deal with a particular historical situation, it was also, in a wider sense, attempting to deal with the problems of the evolving bureaucratic imperial system. It is therefore not surprising that when the Confucian literati again came into their own in the Sung dynasty they should have to some degree started where the T’ang thinkers stopped. Han Yü became the patron saint of a new ku-wen movement and is in the direct line of Neo-Confucian philosophy. Liu Tsung-yüan, on the other hand, was appreciated for his descriptions of scenery but was not valued as a thinker. Yet, more radical, rationalistic types of thinking did not disappear, and it is perhaps legitimate to think of the reform movements of Northern Sung, culminating in Wang An-shih, as in some sense corresponding to the abortive movement of Wang Shu-wen’s party in 805. In the Sung, too, we find the “Legalist” tradition drawn upon and introduced in Confucian dress, perhaps more covertly than in T’ang and with more sophistication. No doubt the specific problems and aims of the eleventh century were quite different from those of the ninth, but it may be possible to see an underlying similarity in the mood which inspired both, and in the type of opposition both aroused. One might also broach the question whether it was inevitable in the Chinese situation that rationalism should always come off second best in the contest with traditional moralism. It is possible that the kind of clearsightedness toward the traditional mystique of Chinese society we find in Tu Yu or Liu­

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

Neo-Confucianism in T ’ ang Intellectual Life

213

Tsung-yüan came too close to undermining the foundations of that society. If the emperor system was to be maintained—and there was no conceivable alternative in anyone’s mind—it was perhaps necessary to buttress and maintain its mystique with the kind of authoritarianism that Chu Hsi and his followers preached. It is, however, a fallacy in history to suppose that what in fact happened was necessarily the only thing possible. In more favorable circumstances it is conceivable that the stultification of the Ming might have been avoided. But we shall understand what did happen better by trying to understand those who attempted to stand in the way of history and deflect it into another course. Abbreviations cts Chiu T’ang shu (Ssu-pu pei-yao edition). ctw Ch’üan Tang wen (1814 edition). hts Hsin T’ang shu (Po-na edition). sptk Ssu-pu ts’ung-k’an. tctc  Tzu-chih t’ung-chien. References are given to chüan and date. The punctuated edition of The Ku-chi ch’u pan she, 1957, has been used. tfyk Ts’e-fu yüan-kuei (1642 edition). thy T’ang Hui-yao (Kuo-hsüeh chi-pen ts’ung-shu edition). tt T’ung-tien (Shih-t’ung edition). The Yin hua lu, T’ang kuo-shih pu, T’ang chih-yen are quoted from the Chung-kuo wenhsüeh ts’an-k’ao tzu-liao hsiao ts’ung shu (Shanghai, 1957).

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

The Structure of Tang Selection P. A. Herbert

Patterns of Recruitment

Broadly speaking, there were three patterns of recruitment in the Tang civil service. Officials of the fifth rank and above were nominated to office by the chief ministers. These were a relatively small number of individuals whose careers and personalities would be well known and who would be hand picked for their individual merits or political affiliations. Selection to such posts is outside the scope of this study and belongs more properly to studies of political history or individual careers. The main concern of this study of civil service selection (xuan) is with regular officials of the sixth to ninth ranks, the lower executive ranks. They were chosen through the system of regular, centralised examinations about to be described. Petty officials crossing the bar between the lower and regular bureaucracies were also subject to the selection examinations. A separate “minor selection examination” was used to select men for posts as clerks in offices in the capital. There is no clear information on how provincial clerks were selected, but presumably they were recruited locally by their superiors.

The Selection Examinations

Selection candidates were drawn from the whole range of official status holders of ranks six through nine and petty officials eligible for entry into the regular bureaucracy. They included men who had recently acquired official status and who had served the required waiting period for eligibility for selection to their first posting. Besides these newcomers, all serving officials who had completed their terms of office had to sit for the selection examinations again, after a waiting period, to obtain a further posting. Information on the waiting periods required before men became eligible for their first offices and between offices is scanty, for the exact regulations are not extant, but it is clear that Source: “The Structure of T’ang Selection,” in P. A. Herbert, Examine the Honest, Praise the Able: Contemporary Assessments of Civil Service Selection in Early T’ang China, Canberra: Faculty of Asian Studies, Australian National University, 1988, 27–56.

© koninklijke brill nv, leiden, ���9 | doi:��.��63/9789004380158_008

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

The Structure of Tang Selection

215

would-be office holders frequently had to wait quite long periods before they became eligible for office. It was axiomatic that the lower down the official scale a man was, the more selection waiting periods and examinations he had to go through, while the higher up the scale he rose, the more in-service assessments he received.1 In other words, at the lower level, there was a classic under-employment situation, with men spending long periods without duties, draining their own and the state’s resources and experiencing frustration, while the higher levels aptly illustrated the Chinese proverb “the able receive the most work” (nengzhe duo lao). Selection candidates were not allowed to submit themselves for examination.2 They were submitted by the organ of the administration in which they had previously served or by the local authority of their native district. The exact regulations for selection, which took place in winter, were issued annually to the various organs of the metropolitan and provincial bureaucracy in the fifth month prior to the selection examinations.3 To be eligible, all candidates had to reach the capital and register with the Department of Affairs of State’s Board of Personnel by the end of the tenth month. Exact time of arrival was determined by the distance to be travelled. The Board was busy registering candidates throughout the tenth month.4 Along with each candidate, the organ of the administration submitting him sent a detailed set of credentials known as a jiezhuang or candidate’s dossier. The dossier included the following information: place of registration registered name names and offices (if any) of father, paternal grandfather and great grandfather (including eligibility to hereditary privilege, if any) details of relatives by marriage age appearance 1  T D 15, 40b–41a (A I, #39). 2  Scattered references to “self-recommendation” (zi ju 自舉) appear to date from times when there was a critical need for officials. Fukushima Shigejirō 福島繫次郎 in his Chūgoku Nanbokuchōshi kenkyū 中国南北朝史研究, Tokyo, 1962, 92–99, suggests self-recommendation was the norm in decree examinations. 3  T D 15, 40a (A I, #32). 4  Candidates within 500 li (approximately 265 kilometres) of the capital had to arrive in the first ten day period of the tenth month, candidates within 1000 li in the second ten day period and those more than 1000 li from the capital in the third ten day period. See DTLD 2, 4b–5a and Niida, Tōryō shūi, 283.

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

216

Herbert

details of assessment record and reason for termination of previous office (in the case of men with a prior service record) and/or doctoral degree (if any) All this information was carefully checked by the Board of Personnel and as a further precaution against violation of selection regulations, candidates were placed in mutually responsible groups of five, required to report to the authorities any misrepresentation contained in their fellows’ credentials or any attempt to cheat in the examination.5 Candidates had, besides, to be vouched for by other, more illustrious persons. Each candidate had to name in his dossier five metropolitan officials as guarantors and one personal referee. From 716, the guarantors themselves were required to submit to the Board of Personnel a form containing personal details of themselves and their protégés.6 The provision of guarantors satisfied the desire for personal recommendation and character reference deeply engrained in the Confucian system and represented the ideal handed down from the Golden Age. Anecdotal material shows that the system of guarantors frequently violated the spirit of its classical precedent and became a free-forall in which candidates touted for references and would-be sponsors angled for promising protégés. Once the candidates’ credentials had been checked, the selection examinations began. They were conducted by three examiners, the president and two vice-presidents of the Board of Personnel. It appears that at first the president was responsible for selection of officials in ranks six and seven and the vicepresidents for selection of ranks eight and nine. From 710, they shared overall responsibility for selection, with the vice-presidents usually carrying out the actual administration and the president authenticating the necessary documentation.7 By 650, a separate selection examination was held in Luoyang, the eastern capital, and was administered by one of the vice-presidents.8 Sometimes other officials were seconded to serve as examiners in the Board of Personnel in addition to their primary duties. For instance in 646, Du Yan was praised for his work as acting selection examiner after he quizzed the

5  T D 15, 40a (A I, #32); Cefu yuangui 冊府元龜 (CFYG), 1643 edition, reprinted Beijing, Zhonghua shuju, 1960, 629, 15b–16a. 6  C FYG 630, 2b. 7  C FYG 629, 4a–b; THY 58, 1004. 8  D TLD 2, 5a; CFYG 629, 3b.

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

The Structure of Tang Selection

217

candidates with unusual thoroughness and selected over seventy men who subsequently proved most suitable in their offices.9 The selection examination ground was ringed with a bramble fence and candidates were thoroughly checked before the examinations began, to prevent the smuggling in of substitute candidates and cribs.10 The selection examinations evidently continued over more than one day, for in the early eighth century, Wei Sili requested that beds and quilts be provided for the candidates.11 Selection tests were of four kinds.12 Two of the tests must obviously have been administered as part of an interview with the examiners. These were “deportment” (shen) and speech (yan). The former might more properly be termed “presence”, for it required that candidates be impressive looking. Such men would presumably command more respect from underlings and commoners when in office as well as meeting the superstitious requirements of an age which believed in physiognomy as an indication of fate. “Speech” is more readily understandable as a criterion for selection. Candidates had to be able to express themselves lucidly and, presumably, with an acceptable accent. A written test was administered to judge candidates’ calligraphy. As in the case of speech, clarity in handwriting was valued in the Tang dynasty for practical as well as aesthetic reasons. In an assessment of the influence of this test on Tang calligraphic art, Ishida Hajime points out that Tang scholars studied calligraphy from their youth, so that mastery of the formal style of kaishu was common among great officials. Taizong promoted the study of calligraphy in the style of China’s most famous calligrapher, Wang Xizhi. This situation contrasts greatly with the Song period, when serious study of calligraphy was taken up as a hobby by older scholars and more informal and individual styles were valued. The great Song calligrapher, Su Dongpo, noted with regret that in consequence the great men of his day could not write as well as Tang minor provincial officials.13 The most important test in the selection examinations was pan (“judgements”).14 Pan was an old literary genre with a practical origin, closely 9  THY 74, 1334. 10   T D 15, 40a (A I, #32); XTS 45, 1171. 11   Sui Tang jiahua 隋唐嘉話, contained in a 1496 edition of Shuofu 說郛 held in the Seikadō bunko, Tokyo, 26b. 12  See TD 15, 40a (A I, #33); XTS 45, 1171; CFYG 629, 16a; DTLD 2, 5a. 13  Ishida Hajime 石田肇, “Shin gen sho han no sho to Tō Sō jidai no sho 身言書判の書と 唐宋時代の書, Shoron 書論 8 (1976), 100–106. 14   Pan are described in Ichihara Kōkichi 市原亨吉, “Tōdai no han ni tsuite 唐代の判につ いて,” Tōhō gakuhō 唐方学報 (Kyoto) 33 (1963), 119–198.

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

218

Herbert

related to the role an executive official would have to play in his actual career, particularly in the provinces. It was especially prized in the Tang dynasty with its emphasis on the legal Code and the norms of ritual. From encompassing purely legal decisions, pan went on to include judgements decided or influenced by moral criteria. The Judgement of Soloman might be considered a pan within the Judeo-Christian tradition for, like the best Chinese judgements, it combines elements of law, ethics, psychological insight and simple common sense. A similar decision over a family dispute was made by Yan Zhenqing in a judgement he gave while a prefect. The snobbish wife of a scholar named Yang Zhijian took the bold step of petitioning for a divorce, on the ground that Yang had achieved no fame. Yan granted Mrs Yang her divorce, but to show how shameless she was according to Confucian morality, he heaped honours on Yang and raised him out of his obscurity. Yan’s judgement, following the prescribed pattern, was written in highly formal parallel prose, in four and six character phrases, including a number of classical allusions used in a similar way to legal precedents in common law. Its highly formal structure and allusive character render it virtually untranslatable. From a practical legal expression and a dry examination subject, pan developed into a flourishing literary genre and even an international money-spinner. Judgements were written in a set form which could best be learned by reading and practising many examples. The study of acceptable judgements would also help in deciding the solutions to new problems, since they were based on legal and ethical precedents and required the citation of historical examples. Candidates practised writing pan and, if they were successful in the selection examinations, there was a ready market for published editions of their compilations of judgements as examination crammers for less able or less studious candidates. Bo Juyi’s Bodao pan (One hundred judgements) became influential after the period under study,15 but an earlier compilation, Zhang Zhuo’s Longjin fengsui pan would have been available to candidates in the eighth century. Zhang’s model judgements were even studied in Korea and Japan, which had copied the Tang examination system. It is evident that he derived considerable financial benefit from his publication, for he was nicknamed Qingqian xueshi “The Copper Coin Scholar” or, more loosely, the scholar-millionaire. Ichihara Kōkichi even traces a connection between judgements and the newly emergent Tang literary genre of short stories. Zhang Zhuo was a writer of anecdotes and Bo Juyi’s new style ballads attest to his powers as a story teller. A satirical form of humorous pan emerged among the educated public, presumably as a 15  A judgement composed by Bo Juyi is translated by Robert des Rotours in Le traité des examens traduit de la Nouvelle histoire des T’ang, second edition, San Francisco, 1976, 345–347.

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

The Structure of Tang Selection

219

form of light relief for initiates into the finer points of the genre. These judgements were couched in the same pompous legalistic language, but they dealt in a tongue-in-cheek way with trivial everyday subjects. With such a rich genre to draw on, candidates became very skilled at judgements and examiners asked increasingly difficult questions drawn first from the classics and then from less orthodox works. The close connection between pan as a literary genre and the examination system may be seen by the fact that the genre virtually ceased to exist in the Song dynasty, when it was dropped from the examinations. There is every indication from the sources that pan, like shu (calligraphy), was a written test. Judgements and calligraphy may even have been tested at one and the same time.16 Most sources state that calligraphy and judgements, the written subjects, were tested first, followed by presence and speech.17 It may be assumed that, to save effort, the examiners weeded out a large number of candidates in the written test and only interviewed those they considered likely to pass.18 Da Tang liudian, however, states that pan was an oral test, held in the early morning when both candidates and examiners were fresh and alert, and that while the vice-presidents of the Board of Personnel asked the questions, an examining board of famous scholars decided the result. This source does, however, refer to a written alternative test.19 The oral test was presumably the ideal norm, while the written test was a practical compromise, given the large number of candidates. In a highly competitive examination sat by serving officials, where candidates and their sponsors had so much at stake, objective assessment would have been difficult and preservation of candidates’ anonymity desirable. The practice of candidates being required to paste a slip of paper over their names (zi hu qi ming) was therefore introduced by Empress Wu to prevent corruption.20 In 695, however, she reversed her policy on anonymity in the selection examinations. The reason for the change was that: Candidates are sometimes chosen for their service record and sometimes for their scholarship and conduct. The pasting over of names in the test

16   C FYG 629, 16a. 17   T D 15, 40a–b (A I, #34); XTS 45, 1171–72; CFYG 629, 16a–b. 18   T D 15, 41a (A I, #40) suggests that pan was the deciding factor in whether a candidate passed or failed selection. 19   D TLD 2, 5b. 20   T D 15, 42a (A I, #51).

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

220

Herbert

on judgements as a way of nominating men to office is a travesty of the appointment system and is very different from the orthodox system.21 These strong words imply that pasting over names placed undue emphasis on the testing of literary skill in the form of judgements. Talent (cai) which could be assessed in this way was only the second criterion supposed to be used in determining success in the selection examinations. The first was “virtue and good conduct” (de xing), subjective qualities which could only be assessed from the candidates’ dossiers and possibly from the interview test. The third criterion was “merit acquired in service” (lao), which again required inspection of the candidates’ dossiers.22 Neither conduct nor service could be assessed if the identity of the candidate was suppressed. Empress Wu’s switch in policy was probably the result of her changing circumstances. When she advocated anonymous testing of judgements she was trying to win wide support, but once she was firmly in control and wanted to keep careful check on her officials, she attacked suppression of identity. A decree of 727 suggests that the suppression of candidates’ names in the judgement test was considered normal practice, but enforcement was lax, for it calls for the pasting over of names in that year’s test “according to the regulation” (yi li).23 While regular officials of ranks six through nine were being selected in the way described above, would-be clerks in the metropolitan organs of administration were being chosen in a separate “selection outside the current” (liuwai xuan) or “minor selection” (xiao xuan). Up until 736, the minor selection was supervised by one of the secretaries to the Board of Personnel. Thereafter minor selection was incorporated into the selection system administered by the vice-presidents of the Board of Personnel. Those eligible for the minor selection were the sons of low ranking officials, who were not entitled by hereditary privilege to regular nominal rank, and commoners. Candidates were submitted by their local authorities in the same way as candidates for regular office and were tested in a similar way. They received ranks in the nine grade petty official ranking system and after a certain period of service, could compete in the regular selection examinations for less prestigious regular official posts.24 21   C FYG 629, 20b. 22  Criteria for success in selection are given in TD 15, 40a (A I, #33); XTS 45, 1171; CFYG 629, 16a. 23   T HY 75, 1361; CFYG 630, 6a. 24   D TLD 2, 31a–32a.

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

The Structure of Tang Selection

221

The third month of spring saw the end of the selection examinations and the announcement of nominations to office (zhuni). At the time of the selection interview, candidates were asked their preferences for nomination. These were not always observed. One examiner, for instance, sent men to serve in diametrically opposite parts of the empire from those in which they showed an interest, until candidates got wise to his strategy and asked for the opposite of what they actually wanted.25 Candidates’ dossiers provided a great deal of information pertinent to offices which they might or might not hold. The dossiers were therefore carefully checked by an undersecretary to the Board of Personnel in charge of the Southern Office (nan cao), the office where all personnel files were stored. The Southern Office passed the dossiers they had checked to the examiners for their reference.26 Candidates could not be given posts in their own or their wives’ native places. Nor could they serve under close relatives or in posts in which they would work in liaison with relatives.27 These prohibitions were obviously designed to prevent the build up of power bases in the provinces or factions at court. Less obvious to the modern reader, but nonetheless important in view of Confucian veneration for ancestors, was the requirement that men could not serve in a department or office which violated an ancestral taboo. In other words, if the name of the department or the designation of the office contained a character identical to one in his father’s, grandfather’s or great grandfather’s given name, a candidate for office could not accept the post offered.28 Once decided, nominations to office were publically announced. Candidates were not forced to accept the first posting offered to them. They had two chances to refuse offices, but if they refused the third, they had to wait to take the selection examinations again the following winter.29 Successful candidates who had accepted the posts offered to them were placed by the examiners on pass lists ( jia) classified according to the way in which the candidates had acquired official status.30 The pass lists were sent for verification to the Chancellory and, if satisfactory, were submitted to the emperor. Cumulative sets of the information contained in candidates’ dossiers and annual pass lists ( jia li) were kept 25   Fengshi wenjian ji 封氏聞見記 in a combined edition with Sheshi suibi, Taipei, Guangwen shuju, 1968, 3, 5b–6a. 26   D TLD 2, 32a. 27  Niida, Tōryō shūi, 286–87. 28   Tang lü shuyi 唐律疏議, Kyoto, 1968, 3, 226, translated in Wallace Johnson, The T’ang Code, Vol. I, General principles, Princeton, 1979, 129–30. 29   T D. 15, 40a–b (A I, #34); XTS 45, 1171–72; CFYG 629, 6a–b. 30   C FYG 630, 1a.

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

222

Herbert

in file stores ( jia ku) in the Board of Personnel, Chancellory and Secretariat.31 These personnel files were colloquially known as “registers” (bushu).32 Such personnel registers were kept from the very beginning of the dynasty, even before the selection system was formalised.33 During the An Shi Rebellion, in the widespread civil disruption and scattering of officials and of the common population, the official registers fell into confusion. Li Xiyun patched them together during the rebellion while serving at Suzong’s court-in-exile, but by 788, the Board of Personnel was again unsatisfied with the state of the records which were so chaotic that people were even able to “borrow” the identity of dead officials. The Board therefore called for a standardised form of candidate dossier, from which data would be easy to transfer into the files, and for submission of dossiers not later than the end of the first month, long after the statutory limit of the end of the tenth month and well into the selection period.34 Formal accreditation to office was granted in different ways, according to the rank of the official concerned. Offices of the third rank and above were awarded by imperial letter (ce shou). Officials of the fourth and fifth ranks were given their offices by imperial decree (zhi shou).35 They were nominated for these offices by the chief ministers and had to have served for a certain number of years without being convicted of corruption.36 Lower officials passed through the regular selection examination system described above and were appointed by imperial order (zhi shou) except for certain holders of low ranking “pure” offices who from 716 were treated the same as officials of the fourth and fifth ranks.37 Each official was given a tally authenticated by the Board of

31  For detailed information on pass lists and personnel files, see Zhang Qun 張群, “Tangdai jia zi shiyi” 唐代‘甲’字釋義, Dalu zazhi 大陸雜誌 9, 1 (1954), 8–10. 32   D TLD 2, 27a. 33  In 622, Zhang Daoyuan 張道源 complained that the registers (bushu 簿書) of the Board of Personnel were so complex that it was easy for corrupt practices to arise. Emperor Gaozu called for comments, but since the majority of officials did not share Zhang’s concern, the matter was dropped. See THY 74, 1333. 34   T HY 74, 1340–41. 35   T D 15, 39b–40a (A I, #31). 36   C FYG 629, 20b–21a, quoting a decree of 696: Officials due to enter the fifth rank had to hold offices of the sixth or seventh rank and to have gone through twelve or more assessments since gaining official status. Officials due to enter the third rank had to hold fourth rank office and have gone through twenty five or more assessments. Both must have no record of corruption. Provisions in DTLD 2, 21a–b, referring to the late 730s, are similar, except that the assessment requirements are increased to sixteen and thirty respectively. 37   T D 15, 39b–40a (A I, #31).

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

The Structure of Tang Selection

223

Personnel as his certificate of appointment (gaoshen).38 He was then required to take up his appointment within a given time limit.39 In the century that followed the adoption of the regular annual selection system in 645, certain modifications were made to render the system more workable.40 In 669, the then vice-president of the Board of Personnel, Pei Xingjian, set up something called “the long list of names” (changming xingli bang or changming bang).41 The nature of this list has provoked considerable speculation. Des Rotours comments that its exact nature is unclear, but concludes that it was a register of successful candidates, used as a guideline in nominations to office.42 Guisso believes that the list was a kind of social register, used to keep undesirable social inferiors out of office.43 I suspect that it was simply a listing drawn up from examination results and other criteria in candidates’ dossiers to indicate who had passed and who had failed the initial stages of selection and who was therefore eligible for nomination to office. It is significant that at the same time, Pei Xingjian fixed the qualifications required for all provincial posts and their relative placing on the promotion scale. Later, the list was made available for public scrutiny to avoid secret manipulation of selection. The “long list of names” allowed scope for a variety of criteria in selection. I do not accept that it was slanted towards selection by class, although a candidate’s social origin must have affected his chances in the selection interview tests on presence and speech. By the early eighth century, the numbers involved in the selection process were so overwhelming that thorough testing was extremely difficult. In 730, therefore, Pei Xingjian’s son, Pei Guangting, introduced the “regulation for advancement by seniority” (xunzi ge).44 As its name suggests, this regulation took into consideration one criterion alone, a person’s position on the official ladder. It was much easier to administer than Pei Senior’s complex system and was hailed by untalented pen-pushers in the bureaucracy as “holy writ” (sheng shu), but it was violently attacked by Pei Guangting’s chief rival, Xiao Song, who had it abolished immediately after Pei’s 38   T D 15, 40a–b (A I, #34); XTS 45, 1171–72; CFYG 629, 16a–b. 39   Tang lü shuyi 9, 491. 40  These modifications are examined in detail in Toriya Hiroaki 鳥谷弘昭, “Han Kōtei no junshikaku ni tsuite” 裴光庭の循資格について, Rishō shigaku 立正史学 47 (1980), 47–62. 41   T D 15, 40b–41a (A I, #39); XTS 45, 117 42  Des Rotours, Le Traité des examens, 243–44. 43  R. W. L. Guisso, Wu Tse-t’ien and the politics of legitimation in T’ang China, Bellingham, Washington, 1978, 94 and 259, notes 55 and 56. 44   T D 15, 40b–41a (A I, #39); CFYG 630, 6b; XTS 45, 1177.

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

224

Herbert

death in 733.45 Nevertheless, because of its administrative convenience, the regulation for advancement by seniority continued to be followed by the Board of Personnel in making appointments. Special provisions existed within the selection system for drawing off the exceptional talents needed for high powered postings, but nevertheless, court advisers continued to echo Xiao Song’s argument that talent should be properly examined.

Accelerated Appointment

Normally, would-be bureaucrats had to wait for a number of years after acquiring official status before they were granted their first post and in service officials had to wait for some time after completing one term of office before they received another. The waiting period (nianshu) was computed in numbers of selections (X number of xuan). The waiting period, together with the actual period in-service and position on the official ladder, was taken into account in assessing each candidate’s seniority qualification (zi) for selection.46 In 669, Pei Xingjian regulated the rules regarding qualification for office and incorporated them into the system governing the compilation of the “long list of names.” His son, Pei Guangting, made the system even more rigid in 730, when he instituted selection solely by seniority. The selection system did, however, contain provisions for accelerated appointment of the particularly talented. These provisions allowed certain candidates to pass through the selection system more quickly and were referred to as “exemption from waiting a certain number of selections” (jian X number of xuan). Even before the Peis’ regulations, special provisions were made for mingjing candidates who had studied the Zhou li (Rites of Zhou) and Yi li (Ceremonies and rites), two classical compilations of the rites, to be exempted from waiting one selection (jian yi xuan).47 Ritual texts were considered particularly important in the early Tang, when ritual, which provided legitimacy to the dynasty, was being reformulated after the confusion of the Period of Division, but they never seem to have been popular with doctoral candidates.48 45   T HY 74, 1348; CFYG 630, 7b–8a. 46  See Zhao Kuang’s 趙匡 tenth fault of the examination system, TD 17, 17a–18a (A II, #15) and Quan Tang wen 全唐文 (QTW), palace edition with preface dated 1814, reprinted Taipei, Wenyou shudian, 1972, 355, 14b–17a. 47   T HY 75, 1375. 48   T D 15, 38a (A I, #12); THY 75, 1376; XTS 44, 1160–61.

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

The Structure of Tang Selection

225

In 675, a further provision was made to assist particularly promising doctoral candidates in their careers. This was the introduction of the “great success” (da cheng) test. The test consisted of recitation of a thousand words from a classical text studied for examination purposes and a number of essays. The maximum quota of passes in this examination was twenty. A pass entitled one to immediate nominal office with salary and treatment equivalent to that of a holder of provisional office.49 This arrangement would suggest that doctoral candidates who acquired official status normally had to wait for their nominal rank, or at least for a salaried nominal office, even before they began the waiting period for a functional office. The most frequently mentioned provisions for accelerated appointment were two special examinations held alongside the regular selection examinations which could be taken by men of any category who had not yet completed their waiting period. These highly competitive examinations were known collectively as the tests for those who were “exceptional and stood apart” (zhaojue). There were two such tests, the “test in calligraphy and judgements: plucking up and collecting” (shu pan bacui ke or more simply bacui ke) and the “test of vast learning and great literary skill” (boxue hongce ke, abbreviated to hongce ke). The bacui ke was first held in 673, soon after Pei Xingjian’s measure to regulate selection qualifications. As its full name suggests, it tested three judgements and also assessed calligraphy. At the beginning of Xuanzong’s reign, the hongce ke was inaugurated as a test of literary skill. Like the decree examinations, it consisted of three essays.50 Indeed, one of the reasons why information on these two tests of extraordinary merit is difficult to locate is that they are often confused with the decree examinations because of their similar titles. Passes in the bacui and hongce tests led to immediate nomination to office without serving the full waiting period. They were extremely difficult to obtain, as is attested in the full list of passes given in Matsumoto Akira’s study, from which Chart V is adapted.51

49  See XTS 44, 1163 and the translation and comments by des Rotours in Le Traité des examens, 165 and 165–66, note 4; WXTK 41, 392a–b. 50   T D 15, 41a (A I, #41); XTS 45, 1172; ZZTJ 201, 6362. 51  Matsumoto Akira 松本明, “Tō no senkyosei ni kan suru sho mondai—toku ni Ribu kamokusen ni tsuite” 唐の選挙制に関する諸問題—特に吏部科目選について in Suzuki Shun sensei koki kinen Tōyōshi ronso 鈴木俊先生古稀記念東洋史論叢 (1975), 391–414.

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

226 Chart V

Year

673 690 697–700 701 707 713 721 722 725 728 730 731 734 736 713–742 743 749 755 742–756 712–756 756–762 763 769 770 779 766–779 762–779 783 790 792 793 794 795

Herbert Passes in the special selection examinations for accelerated appointment (673–800) (after Matsumoto Akira) No. of passes in judgem’t exam (shu pan bacui ke)

1 3 1 6 1 1 3 1 1 1 1 1 3 65 1 1 1 33 26 10 2

2 9 3

No. of passes in essay exam (boxue hongce ke)

1

2 4 2 1 1

2 1 2 3 2 1 3 4 3 1

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

227

The Structure of Tang Selection

Year

796 799 Total*

No. of passes in judgem’t exam (shu pan bacui ke)

No. of passes in essay exam (boxue hongce ke)

177

6 2 41

* For the 127 years between 673 and 800.

In 781, Zhao Zan, a doctoral examiner, requested the revival of a provision of 763 that mingjing candidates who had offered five classics and answered well in the tests on general meaning of the classics and essay writing be exempted from waiting two selections ( jian er xuan). He did however add the rider that the number of men thus exempted should be kept to a minimum. Zhao’s suggestion was merely shelved, for an edict following his memorial requested that such candidates be reported to the throne and their treatment be discussed.52 The year 785 saw adoption of a suggestion originally made in 728 that ming­ jing candidates who had studied the classics on the rites and all three commentaries to the Chun qiu (Spring and autumn annals) be exempted from a period in nominal office (mian ren sanguan), in other words, that they proceed immediately to selection for functional office.53 Again, as in 635, emphasis was placed on special provision for students of ritual. In 786, it was the turn of students of the Code to be exempted from waiting two selections.54 Students of the Tang ritual text, the Kaiyuan li, which contained the definitive rules of the Tang court, were also allowed exemption from normal selection limitations and immediate office upon success in the regular selection examinations.55 Three years later, in 789, further provisions allowed for accelerated appointment of students of classical works on the rites.56 Clearly if a Tang selection candidate wanted to gain office quickly, he did well to study ritual or the legal Code.

52   C FYG 640, 11b–12a. 53   T HY 75, 1373. 54   T HY 75, 1375. 55   T HY 76, 1396–97; TD 15, 39b (A I, #29). 56   T HY 76, 1397–98; TD 15, 39b (A I, #29).

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

228

Herbert

Passing the jinshi examination did not in itself lead to accelerated appointment, but it did place a man on the first rung of a ladder leading rapidly to high court office and even to the premiership. Together with the graduates of decree examinations, jinshi followed a career pattern which took them through the “pure” stream of office (ging guan). These were offices which, in line with the elitism carried over from the Period of Division, carried the greatest social prestige. They also frequently brought their holders close to the emperor and, even if they were of relatively low rank, carried great responsibility and taxed ability. The conferring of these offices on jinshi graduates rather than on the sons of the establishment marks the new face of Tang elitism, based on office conferred by the imperial house rather than on birth. A list of “pure and renowned offices” (qingwang guan) and “pure offices” (qing guan) throughout the regular bureaucracy is given in Da Tang liudian57 and is reproduced in Chart VI. A further list of “eight eminent offices” (ba jun) is given in Fengshi wenjian ji and Tang yulin, referring to the offices, open to jinshi and decree examination graduates, leading to the premiership.58 These eight eminent offices are given in Chart VII. Once men had entered the “pure” stream of office, they could skip grades and rise rapidly in rank.59 The two lists, given in Charts VI and VII, overlap to a large extent. Between them, they show the “pure” stream of office and those key offices within the pure stream which were considered most desirable. All other offices were classified as “impure” (zhuo) and may be regarded as routine, pen-pushing jobs. Former petty officials were allowed to serve in impure posts, but they were debarred from pure posts, for which they were considered unsuited in training, temperament and social background.60 In 779, responsible provincial posts were added to those offices which might not be held by ex-petty officials,61 proving that concern over lack of Confucian training and lack of Confucian wisdom gained through experience went hand in hand with social snobbery in creating a prejudice against this class of regular office holders.

57   D TLD 2, 24a–25a. 58   Fengshi wenjian ji 3, 4a–5a; Tang yulin 唐語林, in a combined edition with Tang zhiyan, Taipei, Shijie shuju, 1967, 8, 277–78. 59   D TLD 2, 6a. 60   D TLD 2, 6a–b. 61   T HY 58, 1005.

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

229

The Structure of Tang Selection



Recruitment and Promotion of Military and Craft Specialists

In face of an increasingly civilian minded establishment, Empress Wu in 702 set up military examinations administered by the Board of War (wu bu), parallel to the doctoral examinations. A form of military selection by the Board of War also existed. Aptitude tests for military personnel were physical rather than mental. They included target practice and weight lifting. Naturally, candidates had to be physically fit and there was a minimum height requirement of six feet for military selection candidates,62 who had to be under forty years of age. Military selection did, however, involve some degree of mental facility, for candidates went through similar tests to their civil counterparts and were required to give clear and detailed answers. Textbooks are not specified.63 In 678, a student of the state University, who was later to achieve high office, Wei Yuanzhong, discussed the problem of the current Tibetan incursions. He objected that military officers were chosen only for physical skills and not for knowledge of strategy.64 It is not clear what offices were awarded in the military selection, but presumably they were in the officer ranks of the imperial guards and other prestigious army posts. Generalships were awarded on individual merit and according to political considerations. Like high ranking civil offices, they are beyond the scope of this study. There were provisions for military men to transfer to the civil service. Men who were serving in the army could take the selection examinations at their posts and their papers would be sealed and forwarded to the Board of Personnel.65 Specialists in arts and crafts, such as cuisine and music, and in sciences, such as astronomy, divination and medicine, were by a decree of 697 restricted to promotion within their own craft.66 There was provision for certain specialist agencies of the administration, namely the imperial household agency, the imperial library and the stables, to regulate the promotion of their own staff,67 presumably because the criteria for promoting specialists were so different from those normally employed in selection. 62  A surprisingly tall stature, since the Tang foot was approximately the same as the imperial measure. See Edwin G. Pulleyblank, The background of the rebellion of An Lu-shan, 227. 63  For military examinations, see TD 15, 37b (A I, #10); XTS 44, 1170. For military selection, see TD 15, 40b (A I, #35); CFYG 629, 16b. 64   Z ZTJ 202, 6386–89. 65   D TLD 2, 8a. 66   X TS 45, 1174; TD 15, 42a (A I, #52); THY 69, 1183. 67   D TLD 2, 7a.

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

230 Chart VI Rank

Herbert Pure and renowned offices and pure offices throughout the bureaucracy Office

Pure offices

8 8 8 7 7 7 7

7 7 6 6 6 6 6 6 6 6 6

Remembrancer of the Left and Right Examining Censor Assistant Instructor in the College of Four Doors Omissioner of the Left and Right Censor in Attendance Concerned with Palace Affairs Doctor of the Court of Imperial Sacrifices Judicial Inspector in the Directorate of General Affairs in the Household of the Crown Prince Doctor of the College of Four Doors Assistant Instructor in the University Secretary Responsible for Noting the Acts of the Emperor Official Responsible for Noting the Acts of the Emperor Secretary of the Right of the Secretariat to the Crown Prince Secretary of the Left of the Secretariat to the Crown Prince Undersecretary (all offices) Attendant Censor Secretary to the Imperial Library Deputy Secretary to the Bureau of Composition Doctor of the University

zuo you shiyi 左右拾遺 jiancha yushi 監察御史 simen zhujiao 四門助教 zuo you buque 左右補闕 dianzhong shiyushi 殿中侍御史 taichang boshi 太常博士 danshi sizhi 詹事司直

simen boshi 四門博士 taixue zhujiao 太學助教 qiju lang 起居郎 qiju sheren 起居舍人 taizi siyilang 太子司議郎 taizi sheren 太子舍人 yuanwai lang 員外郎 shiyushi 侍御史 bishu lang 秘書郎 zhuzuo zuolang 着作佐郎 taixue boshi* 太學博士

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

231

The Structure of Tang Selection

Rank

Office

Pure offices

6

6 6 5 5 5 5 5 5 5 5 5 5 5 5

Assistant in the Directorate of General Affairs of the Household of the Crown Prince Master of Literature in the Crown Prince’s Library Assistant Instructor in the College for Sons of the State Vice-president of the Censorate Counsellor to the Emperor Secretary to the Chancellory Secretary to the Secretariat Counsellor to the Crown Prince Librarian to the Crown Prince Doctor of the College for Sons of the State Secretary (all offices) Assistant Librarian in the Imperial Library Secretary to the Bureau of Composition Assistant in the Court of Imperial Sacrifice Colonel of the Left/Right of the Imperial Guards

danshi cheng 詹事丞

taizi wenxue 太子文學 guozi zhujiao 國子助教 yushi zhongcheng 御史中丞 jianyi dafu 諫議大夫 jishizhong 給事中 zhongshu sheren 中書舍人 canshan dafu 賛善大夫 taizi xima 太子洗馬 guozi boshi 國子博士 langzhong 郎中 bishu cheng 秘書丞 zhuzuolang 着作郎 taichang cheng 太常丞 zuo/you wei langjiang 左右衛郎將 zuo/you shuaifu langjiang 左右率府郎 將

4 4

Grand Counsellor of the Left/ Right to the Crown Prince Colonel of the qian niu Guards

4

左右千牛衛中郎將 Lieutenant General of the Guards zuo/you fushuai 左右副率

4

zuo/you shuaifu 左右率府

of the Left/Right of the Crown Prince Colonel of the Guards of the Left/Right of the Crown Prince

taizi zuo/you yude 太子左右諭德 zuo/you qian niu wei zhonglangjiang

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

232 Chart VI Rank

Herbert Pure and renowned offices and pure offices throughout the bureaucracy (cont.) Office

Pure and renowned offices

The following offices of rank 4: Vice-president of the Secretariat Vice-president of the Chancel lory Chief Assistant of the Left/Right of the Department of Affairs of State Vice-presidents (all offices) Vice-president of the Court of Imperial Sacrifices Deputy Director of the Imperial Library Deputy Director of General Affairs in the Household of the Crown Prince President of the Left/Right of the Secretariat to the Crown Prince General of the Guard of the Left/ Right to the Crown Prince Deputy Dean of the Directorate of Education

zhongshu shilang 中書侍郎 huangmen shilang 黃門侍郎 shangshu zuo/you cheng 尚書左右丞

shilang 侍郎 taichang shaoging 太常少卿 bishu shaojian 秘書少監 taizi shao danshi 太子少詹事

taizi zuo/you shuzi 太子左右庶子 taizi zuo/you shuai 太子左右率 guozi siye 國子司業

* Given incorrectly as chang 常 in Da Tang liudian. Source: Da Tang liudian 2, 24a–25a

All metropolitan and provincial offices of ranks 1 through 3.

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

233

The Structure of Tang Selection Chart VII The “eight eminent offices” No.

Office

Grade

I II

graduation with jinshi degree Collating Secretary in the Imperial Library Corrector of Characters in the Imperial Library Junior Officer in a county adjoining the capital Examining Censor Censor in Attendance Concerned with Palace Affairs Remembrancer Omissioner Undersecretary (to a Board) Secretary (to a Board) Secretary to the Secretariat Secretary to the Chancellory Vice-president of the Secretariat President of the Secretariat

jiaoshu (lang) 校書[郎]

(B9L) A9U

zhengzi 正字

A9L

jiwei 畿尉

A9L

jiancha yushi 監察御史 dianzhong 殿中[侍御史] (shiyushi)* shiyi 拾遺 buque 補闕 yuanwai lang 員外郎 langzhong 郎中 zhongshu sheren 中書舍人 jishizhong 給事中 zhongshu shilang 中書侍郎 zhongshu ling 中書令

A8U B7U

III IV

V VI VII VIII

B8U B7U B6U B5U A5U A5U A4U A3

* Tentative interpretation. I take cheng (丞) in Tang yulin to be an interpolation. Sources: Fengshi wenjian ji 3, 4a–5a and Tang yulin 8, 277–78

Assessment Besides sitting the selection examinations to obtain their first and subsequent offices, Tang officials were subject to a form of continuous assessment, or kao. While the length of waiting periods between offices was calculated in numbers of selections, the length of a term of office was calculated by number of assessments. Completion of a term of office was called “fulfilling the required number of assessments” (kao man). Assessments were supposedly annual, although the period between them varied considerably in practice. In 716, the minimum period to be served before assessment was set at two hundred days.68 68   C FYG 630, 2a. See also DTLD 2, 45b.

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

234

Herbert

The normal tenure of office required the fulfilment of four assessments (man si kao),69 but again, Tang sources show that actual terms in office varied from under a year to ten years or more. Informal assessments of subordinates discussing their merits and demerits were made by superiors, who read them out publically. The highest officials in the bureaucracy were assessed by the emperor. Provincial assessment reports were regarded as a kind of tribute70 and were brought to the capital by commissioners to the New Year court assembly (chaoji shi), high ranking provincial officials deputed to bring tribute offerings to court. These commissioners were also known, for obvious reasons, as “assessment commissioners” (kao shi). A complex list of deadlines for the preparation of provincial assessment reports existed, based on distances from the capital.71 Assessment reports were handed in to the Department of Affairs of State. Each year, two high ranking metropolitan officials checked them, while secretaries of the Chancellory and Secretariat oversaw the whole assessment process. Formal assessments were drawn up by the secretary to the Bureau for the Examination of Merit (kaogong langzhong), a subordinate office of the Board of Personnel, and were publically announced to the metropolitan officials and to the commissioners, on behalf of their provincial colleagues. Only after the public announcement were assessments formally reported to the emperor.72 Assessments were expressed in highly formalised language, in terms of four “credits” (shan) and twenty seven forms of “commendation” (zui). Since the full list of credits and commendations has been translated by des Rotours,73 they will not be repeated here. Besides these plus factors, there were “demerit points” ( fu) and “censures” (dian) calculated in monetary terms as fines, ten demerit points equalling one censure.74 When all the credits and commendations had been totted up, officials were placed in one of nine categories, from upper first class (shangshang) to lower third class (xiaxia). Since not all official conduct could be classified within the strictures of formal credits and 69   T D 15, 40b (A I, #37); XTS 45, 1173; CFYG 629, 17a; ZZTJ 201, 6362. CFYG 630, 21a cites a provision of 793 for magistrates to serve a period of five assessments if replacements are not available. 70  The symbolic connection between tribute, doctoral candidates and assessments is pointed out by Zhang Qun 張群, Tangdai kaoxuan zhidu kao 唐代考選制度攷, Taipei, Zhongyang wenwu gongyingshe, 1954, 39–41. 71   D TLD 2, 45a–b; THY 58, 1008–09. 72   D TLD 2, 45a–46a. 73  Robert des Rotours, Traité des fonctionnaires et traité de l’armée, traduits de la Nouvelle histoire des T’ang, second edition, San Francisco, 1974, 60–64. 74   D TLD 2, 48a–b.

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

The Structure of Tang Selection

235

c­ ommendations, a certain amount of leeway was allowed to superior officials in recommending assessment grades.75 Petty officials were similarly assessed, but the assessment grades (kaodi) to which they might be assigned were only four in number, namely: 1. Pure, attentive, conscientious and just; proper in making investigations and enlightened in conducting examinations 2. Neither lazy in office nor partial in handling affairs 3.  Not conscientious in carrying out duties; guilty of several transgressions 4. Oblivious of the public good and self-seeking; accused of corruption Petty official assessments were processed in a similar way to those of regular officials. Those who received the fourth grade of assessment were relieved of office.76 The actual relationship between assessment and selection is not altogether clear. Ideally, assessments were supposed to contribute the element of “merit acquired in service” (lao) in the selection examinations. The Tang statutes laid down that after a period of four assessments, the normative period of tenure for a metropolitan or provincial official of the sixth rank or below, a full record of middle second class (zhongzhong) assessments was to bring one grade of promotion and that each upper second (zhongshang) or lower first (shangxia) grade was to bring two grades of promotion. Furthermore, a third class grade could be cancelled out by a first class grade.77 Indications are that with the introduction of Pei Xingjian’s and Pei Guangting’s modifications to selection, the criterion of merit acquired in service, like talent and scholarship, was eclipsed in importance by seniority, unrelated to performance. In 701, for instance, a decree stated that assessment records were not to be considered in passing selection candidates except that, when all other considerations were equal, those with first grade assessments were to receive priority. A first grade assessment was not, however, to be weighed against poor examination performance.78 Even if they weighed lightly or not at all in selection, assessments still served a useful purpose as a psychological means of controlling the behaviour of subordinate officials. In 651, for instance, an imperial relative serving as a prefect 75   T D 15, 45b (A I, #65). 76   D TLD 2, 49a–b. 77  Niida, Tōryō shūi, 288–90. 78   T HY 75, 1359.

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

236

Herbert

practised cruelties on the people under his jurisdiction. Emperor Gaozong, who was assessing him, gave him a low rating to shame him into better behaviour.79 A few years later, Lu Chengqing used a similar ploy on an official in charge of transporting tax grain to the capital by canal barge: Lu Chengqing….. was assessing metropolitan and provincial officials. One official was in charge of transport and because of storm winds had lost some grain. In assessing him, Lu said, “When you were in charge of transport, you lost part of the supplies, so I am assessing you lower second grade (zhongxia).” The man showed no change in expression and withdrew without uttering a word. Lu appreciated his gravity, so he changed his comment to “He could not help what happened, so I assess him middle second grade (zhongzhong).” When the man neither looked pleased nor said he was sorry, Lu again changed his comment to, “He is unmoved by favour or disgrace. I assess him upper second grade (zhongshang).”80 This anecdote is worth quoting in full because it illustrates several important points. Firstly, it provides a real life example of the public proclamation of assessment, in front of the man being assessed, as laid down in the regulations. Secondly, it shows that an assessment could be changed at will by the official making it and that superiors did so in order to influence and encourage subordinates to follow acceptable behaviour.81 Thirdly it shows that gravity and inability to be moved by flattery or the threat of punishment were considered Confucian virtues.

79   Z ZTJ 199, 6274. 80   Sui Tang jiahua, 18b. See also ZZTJ 201, 6358. 81  Compare TD 15, 45b (A I, #65).

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

Decree Examinations in T’ang China P. A. Herbert I Introduction The T’ang examination system comprised two major types of examinations, namely those held for civil-service selection (hsüan 選) and the doctoral examinations (chü 舉).1 The decree examinations are often regarded as a mere adjunct to the regular doctoral examinations, since the numbers involved in them were relatively small and they were not held on a regular basis. Nevertheless, their historical and contemporary importance to T’ang intellectuals, their timing and the titles given them, as well as the subsequent prominence of many of their graduates, indicate a significance far outweighing that suggested by mere numbers. This paper is devoted to the rise and fall of the decree examinations as an institution and as a barometer of T’ang politics. The segment of the T’ang examination system rendered, somewhat loosely, in English as either “decree examinations” or “palace examinations” covers the T’ang institutions known in Chinese as chih-chü 制舉 and chih-k’o 制科. The precise meanings of both Chinese terms will be discussed below but, for the sake of convenience, the translation “decree examinations” will be used in this paper to cover the chih-chü/chih-k’o system as a whole. In an article on various types of irregular T’ang examinations published in 1975, the Japanese scholar Matsumoto Akira 松本明 gives a succinct definition of the twin aims of the decree examinations. He notes that they were designed both as an extraordinary means of entry into the bureaucracy for new and unrecognized commoner talent and as an effective means to rapid promotion for men already serving as officials.2 With one or another of these purposes in mind, emperors would from time to time issue decrees calling for Source: “Decree Examinations in T’ang China,” T’ang Studies 10–11 (1992–93): 1–40. 1  A general account of the T’ang examination system is given in Huang Ch’ing-lien, “The Recruitment and Assessment of Civil Officials Under the T’ang Dynasty,” Ph.D., diss., Princeton Univ., 1986; see also P. A. Herbert, Examine the Honest, Appraise the Able: Contemporary Assessments of Civil Service Selection in Early T’ang China (Canberra: Australian National Univ. Press, 1988), and P. A. Herbert, “Curriculum and Content of Doctoral Examinations in Tang China,” Studies in Language and Culture (Osaka University) 12 (1987), 149–63. 2  Matsumoto Akira 松本明, “Tō no senkyosei ni kan suru sho mondai: toku ni Ribu kamokusen ni tsuite” 唐の選挙制に関する諸問題—特に吏部科目選について, in Suzuki Shun

© koninklijke brill nv, leiden, ���9 | doi:��.��63/9789004380158_009

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

238

Herbert

submission of candidates for examination within the palace. The titles of the decree examinations clearly indicated their purposes and/or the type of candidates required. II

Sources and Statistics

The richest Chinese source of material on the T’ang decree examinations is the mid-nineteenth century compilation Teng-k’o chi k’ao 登科記考 (A Study of the Records of Examination Passes), which covers the T’ang and the Five Dynasties periods (early seventh to mid-tenth centuries).3 Teng-k’o chi k’ao is the major work of the Ch’ing scholar Hsü Sung 徐松 (1781–1848), who dedicated much of his life to the study of history and geography. The T’ang administration maintained detailed personnel files, including annual lists of the candidates who passed the various examinations. These lists were unfortunately lost in the Sung dynasty, but fragments have survived, notably a table of numbers of successful candidates in the regular doctoral examinations in Wen-hsien t’ung k’ao 文獻通考 and a partial list of decree examinations and graduates in T’ang hui yao 唐會要.4 Hsü Sung painstakingly searched all available historical and literary sources on the T’ang dynasty for references to examinations and in particular for the names of successful candidates and recorded them in his study chronologically, along with the material included in the table and list mentioned above, and all extant examination questions and answers. Robert des Rotours, in the most widely consulted Western-language study on the T’ang examinations, appraises Hsü’s work extremely highly. He characterizes it as “un travail définitif auquel on peut se fier.”5 In this paper I have treated it as the most reliable, definitive work on the T’ang examinations in general and on the decree examinations in particular. While not wishing to disparage Hsü Sung’s excellent work, it should be stressed that since so much material was lost in the centuries between the sensei koki kinen: Tōyōshi ronso 鈴木俊先生古稀記念:東洋史論叢 (Tokyo: Yamakawa Shuppansha, 1975), 391–414. 3  The extremely convenient three-volume edition of tkck, punctuated and annotated by Chao Shou-yen 趙守儼, is cited in this paper. The photolithographic reproduction of the Nan-ching shu-yüan ts’ung-shu 南京書院叢書 edition of 1888, with supplementary entries compiled by Lo Chi-tsu 羅繼祖 and an index by Nasu Kazuko 那須和子, was also consulted. 4  Respectively, whtk, 29.15a–22a and thy, 76.1386–90. 5  Robert des Rotours, Le Traité des examens, traduit de la Nouvelle Histoire des T’ang, rev. edn. (San Francisco: Chinese Materials Center, 1976), 118.

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

Decree Examinations in T ’ ang China

239

T’ang and the Ch’ing, Teng-k’o chi k’ao provides the modern scholar with only a partial reconstruction of the T’ang examination lists, along with much valuable documentation on the institutional development of the T’ang examination system. For the chin-shih 進士, the regular literary doctoral degree, Teng-k’o chi k’ao gives what appear to be reasonably accurate figures for most years as preserved in the Wen-hsien t’ung-k’ao table, along with some names.6 For the decree examinations, however, the figures given are less reliable, since the material available to Hsü Sung himself was more scattered and sketchy. For most of those decree examinations that he has succeeded in tracking down, Hsü is able to name one or two, or at most five or six, successful candidates. A text reproduced in the encyclopedia T’ung tien 通典, another major source on T’ang examinations, gives an intriguing glimpse of the numbers involved in the decree examinations during the early eighth century, a period when they were held relatively frequently. The text reads: Those submitted as candidates in response to the imperial summons were at the most two thousand men, at the least one thousand. Those accepted were only one in a hundred.7 Statistics quoted in old Chinese texts are notoriously unreliable. One to two thousand candidates with a pass rate of one per cent could be taken for conventional figures, but there are some indications in Teng-k’o chi k’ao that these figures may in fact be close to the truth. For two examinations, held in 738 and 742 respectively, Hsü gives the following figures: twenty-four successful candidates, of whom three can be named, and twenty successful candidates, of whom two can be named.8 Later, from the 780s onwards, when the extant records on decree examinations become more detailed, he mentions several instances in which fifteen or so candidates passed. Probably there was considerable variation in the number of candidates and passes in decree examinations according to their type, in particular according to whether they were designed to draw in new talent or to promote high-fliers 6  See the graph appended to this article and P. A. Herbert, “Curriculum and Content of Doctoral Examinations in Tang China,” 163, chart iv. 7  t t, 15.39a, translated in P. A. Herbert, Examine the Honest, 178. My earlier translation fails to bring out the fact that this passage refers to decree examinations. 8  t kck, 8.290–91 and 9.301–2 respectively. Compare the reference in tkck, 3.92 to a certain Chang Chien-chih 張柬之, who is said to have gained the only top-class pass in a field of one thousand candidates for the hsien-liang fang-cheng examination in 689. The names of two or possibly three other successful candidates, presumably lower passes, are known.

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

240

Herbert

within the bureaucracy. It is by no means certain that Hsü Sung has traced all the decree examinations that were held in the T’ang dynasty, so that a guess at the total number of successful candidates is difficult and an accurate estimate impossible. It is, however, possible to indicate periods in T’ang history when emperors made particularly frequent or regular use of decree examinations. Teng-k’o chi k’ao also provides a valuable database of approximately four hundred and fifty names of men who passed civil decree examinations during the two centuries preceding 828.9 III

The Rise of the T’ang Decree Examination System

The T’ang examination system as a whole was innovative and can be traced back in part only to the preceding Sui dynasty. But the decree examinations could claim a much more venerable history. Chinese sources make clear that they are directly descended from civil-service tests held in the Han dynasty.10 During the Han, the majority of officials were chosen directly by their bureaucratic superiors, but the most prestigious posts in the capital were filled by imperial summons or on the recommendation of leading metropolitan or provincial officials. At first, Han emperors issued pronouncements calling for recommendations at irregular intervals, as and when more officials were needed, as their T’ang successors were to do through the medium of the decree examinations. Hans Bielenstein points out that: The [Han] government normally sought for men with special moral traits, and the edicts soon began to express these in stereotype formulas, such as men who were Capable and Good (hsien-liang 賢良), Sincere and Upright ( fang-cheng 方正), who possessed Flourishing Talent (hsiu-ts’ai 秀才), and who Spoke Frankly and Admonished Unflinchingly (chih-yen chi-chien 直言極諫). More rarely, the edicts made military ability or legal expertise the chief requirement.11

9  This figure does not include passes in military decree examinations, which are also included by Hsü Sung. See D. L. McMullen, “The Cult of Ch’i T’ai-kung and T’ang Attitudes to the Military,” T’ang Studies 7 (1989), 59–103. 10  For instance in tfyk, 645.1a–6b, especially the introductory passage. 11  Hans Bielenstein, The Bureaucracy of Han Times (Cambridge: Cambridge Univ. Press, 1980), 133. For a full description of Han civil-service selection and examinations, see pp. 132–42.

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

Decree Examinations in T ’ ang China

241

From 130 bc onwards, provincial authorities were required to make regular annual recommendations of worthy candidates. The men thus recommended were known as hsiao-lien 孝廉, which Bielenstein translates “Filially Pious and Incorrupt.”12 In the mid-second century ad, they were joined by the hsiao-t’i 孝悌 (“Filially Pious and Fraternally Respectful”), while the title hsiu-ts’ai 秀才 was changed, because of an imperial taboo name, to mao-ts’ai (“Abundant Talent”).13 As their numbers increased, candidates belonging to all these categories, upon reaching the capital, were required to undergo examinations. The term hsiu-ts’ai, which was to become so well-known in late imperial China, was used for the name of a regular doctoral degree in the Sui and early T’ang dynasties. Almost all the other Han terms for candidates for office (hsienliang, fang-cheng, chih-yen chi-chien, hsiao-t’i and mao-ts’ai) occur independently or in combination in the titles of T’ang decree examinations, especially after the mid-eighth century, reminding T’ang intellectuals and modern readers of the close link between Han selection and the T’ang decree examinations. The term hsiao-lien (filial and incorrupt) is associated with Yang Wan’s 楊綰 abortive attempt to reform the examination system, discussed below. Calls for recommendations of men possessing these and similar traditional Chinese virtues were made from time to time by emperors in the centuries between the fall of the Han and the foundation of the T’ang. In the brief Sui dynasty, six such instances are recorded in the section on chih-chü in Ts’e-fu yüan-kuei 册府元龜, for the years 583, 598, 603, 607, 609, and 614.14 The entries for 598 and 609 are particularly interesting, since they include the character k’o 枓, which means “category” and, in an academic context, comes close to the English word “degree.” It is used of examinations leading to degrees conferring official status and is thus associated with the subsequent T’ang topical decree examination (chih-chü k’o-mu 制舉科目 or chih-k’o 制科). In the Sui references, no specific mention is made of a formal examination (shih 試 “test” or ts’e-shih 策試 “essay test”), so “degree” may be the best translation for k’o in this context. Evidence of the date when decree examinations began in the T’ang dynasty is confusing in the older Chinese sources, causing much disagreement among modern scholars. The list of named decree examinations in T’ang hui yao begins with an examination dated 658, so des Rotours and a number of modern Chinese scholars accept that date as the earliest instance of a decree examination in the T’ang. Some Japanese scholars have argued, on the basis of less clear references to “submission in response to imperial decree” (ying chih chü 12  Ibid., 134. 13  Ibid., 136. 14   t fyk, 645.9b–10b.

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

242

Herbert

應制舉) and similar expressions, that decree examinations can be traced right

back to the beginning of the T’ang. Hsü Sung believed that “submission following an imperial decree” (chih-chü 制舉) began in 622, the fifth year of the Wu-te 武德 period of the reign of Kao-tsu 高祖 (r. 618–626), the founder of the T’ang dynasty, the fifth year of the dynasty itself.15 Understanding the genesis of the T’ang institution of the decree examination requires close attention to the terminology used to describe early T’ang instances of recommendations and examinations that followed upon imperial pronouncements. One must also appreciate the importance of early Han precedents as perceived by the first T’ang emperors.16 Let us consider, for example, the term ying chih chü 應制舉. Matsumoto Akira pointed out that the career of Ts’ui Jen-shih 崔仁師, a historian who was to distinguish himself as a high official in the service of emperor T’ai-tsung 太宗, began, according to his Chiu T’ang shu 舊唐書 biography, at the very start of Kao-tsu’s reign with “submission in response to an imperial decree” (Wu-te ch’u ying chih chü 武德初應制舉) and appointment to a post as administrative supervisor (lu-shih ts’an-chün 錄事參軍) in Kuan-chou 管州.17 Fukushima Shigejirō 福島繁次郎, writing in 1962, drew attention to an entry in T’ang hui yao quoting a decree of Wu-te 5 (622) in which Kao-tsu called upon every high official to recommend one man. Fukushima therefore concluded, as had Hsü Sung, that chih-chü began in 622, the same year as the regular doctoral examinations. Fukushima also noted that this decree made unprecedented provision for those who deemed themselves suited for office because of talent and moral behavior but had not been recommended by a third party, to apply on their own behalf by setting out all their skills and abilities (i hsü-t’ing tzu-chi chü ch’en i-neng 亦許聼自己具陳藝能).18 It thus appears that ying chih chü in the reign of Kao-tsu did not mean “to sit for a formal decree examination.” Rather it indicates a conscious attempt to copy the early Han practice of recruiting 15  For a discussion of various scholarly interpretations of the initiation of decree examinations in the T’ang, see Matsumoto, “Tō no senkyosei ni kan suru sho mondai,” 392–96. For Hsü Sung’s view, see tkck, 1.3–4. For des Rotours’ view, see Traité des examens, 41. 16  The importance of Han precedent to the early T’ang emperors is stressed by Howard Wechsler in Offerings of jade and Silk: Ritual and Symbol in the Legitimation of the T’ang Dynasty (New Haven: Yale Univ. Press, 1985), passim. 17   ChTS, 74.2620–22. 18   t hy, 26.490, quoted in Fukushima Shigejirō, Chūgoku Nambokuchōshi kenkyū 中国南 北朝史研究 (Tokyo: Meicho Shuppan, 1962), 93. A fuller version of this imperial pronouncement is quoted in TKCK, 1.3–4, for which see below. Fukushima discusses the decree examinations, related examinations, and the reconstruction of T’ang results-lists on pp. 77–129 of his work. “Self-submission” (tzu-chü) is discussed on pp. 92–99.

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

Decree Examinations in T ’ ang China

243

men upon recommendation. It is unlikely that any test, let alone a formal examination, was administered in the capital at the very beginning of the dynasty. On the contrary, T’ung-tien points out that until around 622 it was difficult to persuade scholars to come out of retirement; the Board of Personnel (li-pu 吏部) had to call upon provincial authorities to send anyone they considered worthy, and all were immediately given office.19 It is therefore misleading to translate in the first years of T’ang as “decree examination.” Chih-chü simply means “to be submitted following a decree.” Testing began in earnest in the Chen-kuan 貞觀 period, during the reign of emperor T’ai-tsung (r. 626–649). The biography of Hsieh Yen 謝偃, an undistinguished official who in his own day was a noted writer of rhyme-prose ( fu 賦), contains the sentence, Chen-kuan ch’u ying-chao tui-ts’e chi-ti 貞觀初 應詔對策及第, which may be translated: “At the beginning of the Chen-kuan period, in response to an imperial edict, [Hsieh] answered with an essay (or plan) and gained a degree.” Hsü Sung tentatively dates Hsieh Yen’s success to 627.20 He gives another instance of a candidate who in 633 “[in response to] a decree offered an essay (or plan) and gained a high degree” (chih ts’e kao-ti 制策髙第).21 Clearly something like a “decree examination” was emerging in T’ai-tsung’s reign. On the other hand, for the year 632, Hsü Sung quotes the Chiu T’ang shu biography of Ts’ui Hsin-ming 崔信明, who was “submitted in response to an edict” (ying chao chü 應詔舉) and appointed to a post as vicemagistrate (ch’eng 丞).22 It may therefore be concluded that T’ai-tsung took the logical step of following the later Han precedent in testing those persons brought to his attention after imperial pronouncements calling for recommendations. At least some of the tests he administered appear to have been in the form of essay questions. Hsü Sung traces the nomenclature of T’ang decree examinations to an edict of the fourth month of 629. This interesting edict begins with T’ai-tsung’s affirmation of his conscientious attitude to government, augmented by filial respect for his retired father. He insists on the importance of employing honest officials as a channel for learning the feelings of the common people. He goes on to speak of the improvement in the social order, now that stable government has been achieved, and applauds those who live in accordance with traditional Chinese family morality. Having set out rewards for such people and 19   t t, 15.41b. 20   ChTS, 109A.4989. For Hsü Sung’s dating, see tkck, 1.10. 21   t kck, 1.16. 22  Ibid., 1.15, quoting ChTS, 140A.4991.

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

244

Herbert

for the elderly and other worthy souls, he calls upon high provincial officials to send in confidential reports about worthy and unworthy officials. The edict concludes as follows: As for men from the villages and hamlets who come from within commoner households, if they have “civil or military talent or ability clearly qualifying them to be chosen” (wen-wu ts’ai-neng cho-jan k’o ch’u 文武材 能 灼然可取), or are “loyal and respectful in word and deed, able to control current affairs” (yen-hsing chung-chin k’an li shih-wu 言行忠謹堪理 時務), or who, “in time of disorder, show an attitude of striving, and when encountering a time of peace, practice self-denial” (tsai hun-luan erh ssuch’ing ou t’ai-p’ing erh k’o-chi 在昏亂而肆情禺太平而克己), they should let a record be made of their names and credentials (ming-chuang 名狀) and be reported in the same way as the officials.23 The term chuang 狀 or chü-chuang 舉狀 refers to a dossier setting out the personal details of the doctoral-examination candidate. It appears also to have included comments on his moral qualifications.24 This is not, in fact, the earliest reference quoted by Hsü Sung to nomenclature from T’ai-tsung’s reign that puts one in mind of the decree examination. Immediately after he ascended the throne in the eighth month of 626, T’aitsung proclaimed an Act of Grace containing the words: As to elderly men of learning, frank in speech and upright in giving admonition (chih-yen cheng-chien 直言正諫), let their local chief official recommend them for submission in accordance with their credentials (sui chuang chien-chü 隨狀薦舉).25 The phrase chih-yen cheng-chien is almost identical to the Han formula chihyen chi-chien mentioned above. A similar call for the recommendation of scholars and commoners (shihshu 士庶) in six categories of moral virtue, Confucian scholarship, and literary talent was issued in 641, and an edict of T’ai-tsung declaring his intention to carry out a major sacrifice at T’ai-shan 泰山.26 None of these pronouncements 23   t kck, 1.13–14. 24  See P. A. Herbert, “Recording and Processing Data on Candidates in the Tang Examination System,” Studies in Language and Culture 14 (1988), 143–56, esp. 148–49. 25   t kck, 1.8. 26  Ibid., 1.21–22.

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

Decree Examinations in T ’ ang China

245

mentions the formal examinations that are normally specified in later edicts and decrees calling for recommendations in similarly named categories. It may therefore be concluded that during T’ai-tsung’s reign the emphasis was still on recommendation and not upon examination, although on some occasions, at least, oral and written tests were administered to those who had been recommended.27 The strategy of publishing edicts and decrees for recommendations, followed by relatively informal testing, was one of T’ai-tsung’s techniques for promoting “good government” by means of recruiting worthy officials, a major aspect of his style of government. Consciously copying Han practice, Kao-tsu and T’ai-tsung saw chih-chü as lending legitimacy to their regime. As mentioned above, Fukushima has drawn attention to another aspect of T’ang chih-chü that was a T’ang innovation. This innovation in the system of submission was tzu-chü 自舉, literally “self-submission” or, in modern parlance, “personal application.” Like T’ai-tsung’s abortive attempt at reviving feudalism, it represents an early T’ang experiment with varying modes of recruitment and administration. “Personal application” appears in the imperial pronouncement of Wu-te 5 (622), which Hsü Sung first suggested and Fukushima accepted as being the beginning of chih-chü. This is a representative decree from the earliest stage in the development of the T’ang decree submission/examination system. In it, Kao-tsu states that as part of the process of setting up the new regime and relieving the sufferings of the people who are recovering from civil war, he wishes to obtain men with talents suited to immediate needs. In particular, he is seeking men who dwell hidden away as recluses in caves, in rustic obscurity, wearing [the] humble serge [of commoners] but with [the] pearls [of wisdom] hidden in their bosoms, and lack the means to realize their own potential (wu-yin tzu-ta 無因自達).28

27  The only possible hint that formal decree examinations may have been held under T’aitsung is in an entry under the year 640, in a note following the name Li Ch’u-ts’ai 李楚 才 (tkck, 1.21). His funerary inscription by Yang Chiung 楊炯 reads: “In response to an edict calling for submission of candidates for four examinations, he submitted an essay and gained a first-class pass” (應詔四科舉射策登甲策), where k’o represents a named examination and is anachronistic when referring to the first half of the seventh century. 28   t kck, 1.3–4.

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

246

Herbert

In the course of the text as given in Teng-k’o chi-t’ao, such worthy recluses are urged to “advance themselves” (tzu-chin 自進) and to “make personal applications” (tzu-chü). This imperial pronouncement, with its emphasis on “that which is of value to current needs” (so-kuei shih-shih 所貴適時), sets the tone for later appeals to candidates for office to “make personal applications” (tzu-chü). In 659, at a critical juncture in T’ang court politics, which marked the passing of T’aitsung’s old guard from power and the establishment of empress Wu’s 武后 dominance, nine hundred commoner candidates who had submitted their own applications sat an essay examination in the palace.29 Later, when she was preparing to displace the Li-T’ang 李唐 dynasty, empress Wu needed officials beholden to herself on whom she could rely. In 685, she therefore issued an edict calling upon civil and military officials and commoners to make personal applications, most likely for two decree examinations designed to recruit high officials and border generals.30 During empress Wu’s reign, north China was beset by attacks mounted by the Northern Turks under their leader Qapaghan (Chinese name: Mo-ch’o 黙啜). Shortly after empress Wu’s fall, Qapaghan inflicted a stunning defeat on the Chinese at Ming-sha 鳴沙 in the twelfth month of 706. Early in 707, emperor Chung-tsung 中宗 asked his officials to submit plans on how to cope with the Turkish menace.31 At the same time, he called upon “bold men with surpassing martial skills” to submit personal applications.32 Further calls for personal applications came in 757 and 758 from Su-tsung 肅宗, after recapturing the two capitals, Ch’ang-an and Lo-yang, during the An-Shih Rebellion. In 757, four categories of men—men with executive abilities; scholars of the classics, history, and literature; administrative and legal experts; and men with military skills—were to make personal application to their local authorities in order that they might be tested and, if successful, be recommended to the emperor. This exercise represents an attempt by Su-tsung at a rapid restoration to normality of the bureaucratic and military structure thrown into chaos by the rebellion. In 758, the emperor complained that many

29   t hy, 26.507. 30   t kck, 3.77–78. 31   t ctc, 208.6607–9. For an English account of Qapaghan’s career, see Richard W. L. Guisso, “The Reigns of the Empress Wu, Chung-tsung and Jui-tsung (684–712),” in The Cambridge History of China, vol. 3, Sui and T’ang China, Part I, ed. Twitchett (Cambridge: Cambridge Univ. Press, 1979), 313–18. 32   t kck, 4.142, quoting from ChTS, 7.143.

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

247

Decree Examinations in T ’ ang China

of the men who had come in response to his call for talent had been unsuitable. He therefore made a second appeal for men to apply in person.33 An interesting reference to tzu-chü outside the context of national emergency exists in the title of a decree examination held in 727 to recruit men “of elevated talents sunk in the obscurity of the rural wilds who are to make personal application” (高才沈淪 草澤自舉).34 At this time, Chang Yüeh 張說, who is known to have cultivated obscure scholars, was the power at court. It seems likely that he realized that the chances of such men of “elevated talent” being recognized by idealistic officials were slim in the real world in comparison to the securing of recommendation through personal connections, so he suggested this decree examination to emperor Hsüan-tsung 玄宗. Other T’ang intellectuals had a low estimation of the submission of personal applications. In T’ai-tsung’s reign, Wei Cheng 魏徵 argued forcibly against “selfsubmission” (tzu-chü), since very few people could take a true measure of their own worth and most would regard “self-submission” as a license to inflate their pretensions and compete in an indecorous manner.35 Yang Wan took a very similar view of personal applications in 763. Permission to submit personal applications appears to have been an exceptional provision in the decree submission/examination system, normally found only in times of national emergency or political crisis. The history of T’ang decree submission/examination as a whole, particularly in its early phase, reflects political trends and the desire to establish legitimacy. Under Kao-tsu and T’ai-tsung, chih-chü represented an appeal to Han precedent and meant recommendation and relatively informal testing. In its fully developed form, the T’ang system of irregular formal examinations following imperial decrees was created under Kao-tsung 高宗 or, more precisely, under the influence of Wu Chao 武曌, his empress. IV

Decree Examination System under Empress Wu and Hsüan-tsung

In the ninth month of 649, following the death of T’ai-tsung and his own accession, Kao-tsung issued an edict calling for the recommendation by leading metropolitan and provincial officials of specific numbers of worthy subordinates in eight named categories. These categories, in all cases but one, emphasize exemplary moral behavior and the potential for statesmanship. Particular 33   t kck, 10.344 and 345–46. 34  Ibid., 7.250, and the pass list, 7.251. 35   Chen-kuan cheng-yao 貞觀政要 (Shanghai: Ku-chi, 1978), 3.93, and tt, 15.41b.

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

248

Herbert

priority was to be given to officials who were descendants of loyal followers of the T’ang founders and of those who had died for the T’ang cause. Those recommended would be given extraordinary promotions.36 This edict is noteworthy for two reasons. Firstly, it shows continuity with the policy of Kao-tsu and T’ai-tsung of using the Han style of recommendation following imperial pronouncements. Examination is not mentioned. If a test was applied, it was clearly subordinate to the act of recommendation. The emphasis on promoting descendants of loyal servants of his grandfather and father and martyrs to the T’ang cause is, of course, a mark of the new emperor’s conventional filial piety, but like the continuity of policy, it shows the influence of the “old guard” of statesmen who had advised T’ai-tsung and who still dominated the court. On the other hand, the timing of the edict, its neat categorization of candidates, and specification of fixed numbers suggest a certain institutionalization of recommendation which contrasts sharply with the sincere search for hidden talent in the edict quoted above that was issued early in T’ai-tsung’s reign. Rather than seek new talent among recluses, Kao-tsung and his old guard advisers rewarded those who already had a place in the bureaucratic and social establishment. Similarly, a few months later, at the beginning of 650, Kao-tsung called for the metropolitan colleges to recommend to him their best students.37 The colleges were a bastion of the establishment, with an intake largely restricted to the sons and grandsons of relatively high-placed officials and holders of honorary ranks and noble titles. A marked change is noticeable within a few years of Kao-tsung’s accession and coincides with the period when Wu Chao became a force to be reckoned with at court and the influence of the old guard waned. In the winter of 655, Wu Chao was made empress.38 In the tenth month of 656, an edict called upon provincial authorities in the Ho-nan 河南, Ho-pei 河北, and Chiang-Huai 江准 regions—in other words in northeastern and eastern central China—to seek out men worthy of executive office, treat them with the greatest courtesy and send them to the capital.39 By the year 659, empress Wu’s opposition, the old guard of T’ai-tsung’s statesmen, had been totally crushed40 and in 660, empress Wu took over her husband’s duties because of his ill health.41 The power base of the old guard 36   t kck, 1.34–35, quoting from tfyk, 645.11a. 37   t kck, 2.37. 38   t ctc, 200.6293–94. 39   t kck, 2.44, quoting from tfyk, 645.11a. 40   t ctc, 200.6317. 41  Ibid., 200.6322.

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

Decree Examinations in T ’ ang China

249

was Kuan-chung 關中, the area around the principal capital, Ch’ang-an, in northwest China. The edict of 656 calling for recruitment of officials in eastern China may therefore be interpreted as an attempt to counterbalance the northwestern influence in the bureaucracy. Similarly, in 676, an edict called for the recommendation of men from the regions of Shan-tung 山東 and Chiang-tso 江左 (eastern and central China), which were characterized as populous but poorly represented in the higher levels of the bureaucracy.42 In his study of empress Wu, R. W. L. Guisso points out that the empress deliberately set out to create an elite dependent for its political and social status upon service in the bureaucracy. She was determined that the great clans of the northwest would no longer dominate the court to such an extent that their authority was, in practice, equal to that of the monarch. Nor could she tolerate the pretensions of the great clans of the northeast, the so-called Shantung aristocracy, who regarded themselves as a social elite and refused to marry outside their own group. Rather than simply bringing unconnected, obscure “new men” into the bureaucracy, Guisso argues that empress Wu drew on a wide spectrum of “new blood,” emphasizing demonstrated ability rather than birth or high connections as the criterion for official appointment.43 The edicts of 656 and 676 may be viewed as consistent with this policy. It is not without significance that the first clearly authenticated instance of a formal decree examination (chih-k’o 制科), as opposed to recommendation followed by a test, is dated 658, near the beginning of empress Wu’s period of power.44 The year 657 had seen a crisis in the regular system of bureaucratic selection, with numbers of qualified candidates reaching such huge proportions that the as yet new and relatively untried system had broken down. The emperor called for discussions of selection policy, but officials at court feared

42   t kck, 2.63. 43  R. W. L. Guisso, Wu Tse-t’ien and the Politics of Legitimation in T’ang China (Bellingham, Wash.: Western Washington Univ., 1978), 87–106. The fact that empress Wu’s examination policy did not discriminate against the establishment, including the northwestern aristocracy, in favor of “new men” is shown by the results of two decree examinations held in 696 in connection with two imperial ceremonies. The five men who passed in the nan-chiao chü 南郊舉 and feng-shen hsien-liang fang-cheng k’o all came from eminent families: Su 蘇 (a metropolitan family) Pei 裴 (from Chiang-chou 絳州, in modern Shansi province), and Ts’ui 崔 (originally a branch of the famous Po-ling Ts’uis 博陵崔, settled in the metropolitan region). See tkck, 4.118–19. Such families were known to be the custodians of ritual traditions, so their success in these examinations is hardly surprising. 44   t kck, 2.46–47; tfyk, 645.11b; thy, 76.1386.

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

250

Herbert

for their vested interests, so the matter was dropped.45 The graph of chin-shth passes included in Wen-hsien t’ung-kao and Teng-k’o chi k’ao appended to this paper shows violent fluctuations in numbers during Kao-tsung’s reign, with frequent suspension of examinations. Thus the regular doctoral examination system was also undergoing teething troubles. It is therefore not surprising that empress Wu turned to another examination strategy to further her policy of bureaucratic and social control. I conclude that she deliberately chose to promote decree examinations in the late 650s because of their connection with the Han institution of recommendation and testing. Their Han origin conferred on the decree examinations (chih-k’o), however different they might be in spirit from Kao-tsu’s and T’ai-tsung’s chih-chü, a legitimacy and ideological acceptability which court conservatives could not dispute, while the flexibility of their irregular nature made them politically expedient for the progressive empress. Thus under empress Wu, recommendation followed by tests evolved into a type of formal examination, just as in the Han system of recruitment through recommendation, examinations had gradually become more important. During her period of supreme power, after the death of her husband and the ousting of emperors Chung-tsung and Jui-tsung 睿宗, empress Wu appears to have made relatively regular use of decree examinations (see the graph). Particularly noteworthy is her call in the sixth month of 689, preparatory to overthrowing the T’ang dynasty and setting up her own Chou 周 dynasty, for recommendations of men with the ability to be “joists and beams” (liang-tung 梁棟), in other words, able ministers of state. The edict in which this request appears mentions eight categories of persons and specifies that each category of examination (k’o 科) is to consist of three essay questions,46 the pattern frequently found in extant T’ang decree-examination questions and answers. Two examinations with titles almost identical to two of the eight mentioned in this edict are recorded, with one successful candidate named as passing in each. The date of these examinations is given as the first month of Yung-chang 1 (689) in Ts’e-fu yüan-kuei and T’ang hui yao, but Hsü Sung believes they are an outcome of the edict of the sixth month of 689 and should therefore be dated the first year of Tsai-ch’u (689–90).47 Hsü’s dating is confirmed by a reference, which he quotes, to empress Wu’s initiating the testing of provincial candidates over a period of several days within the palace in the second month of 45   t ctc, 200.6308. See also the memorial by Liu Hsiang-tao 劉祥道, tt, 17.9a–10b, and thy, 74.1334–35. 46   t kck, 3.90–91. 47  Ibid., 3.97. See also tfyk, 645.13a and thy, 76.1386.

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

Decree Examinations in T ’ ang China

251

690.48 Thus empress Wu took two decisive steps in establishing the decree examination as part of the T’ang examination system.49 In 658 she initiated formal examinations held at irregular intervals, and in 689 to 690 she gave them the form of palace examinations in which three essays were set, plenty of time was allowed for thought, and the monarch acted as an honorary (or even actual) chief examiner. More will be said below about the pattern for the conduct of decree examinations established by empress Wu. The decree examinations were one means by which empress Wu manipulated political control of bureaucratic personnel and established the legitimacy of her position first as empress and later as monarch of the Chou dynasty. When Li Lung-chi 李隆基 (i.e., Hsüan-tsung) strove to re-establish T’ang authority and purify and strengthen T’ang administration, he used the precedent so ably created by empress Wu to good effect. In Teng-k’o chi k’ao, nine decree examinations are dated to the year 712 in which Jui-tsung abdicated and his son Li Lung-chi, the real power behind the T’ang restoration, took the throne; a further five are recorded for 714.50 Thereafter, decree examinations continued with a fair degree of regularity, if not such great frequency, throughout Hsüantsung’s reign (see graph). The dictator Li Lin-fu 李林甫, who controlled the administration from 736 until his death in 752, is said to have interfered in a decree examination in 747 in order to prevent the candidates from revealing to the emperor the nature of his reign of terror.51 Tzu-chih t’ung-chien states that, as a result, not a single candidate passed the examination5;52 but in fact one pass is recorded for a decree examination held in that year.53 It was in the year 751 that all candidates were deemed incompetent and were “sent down” to apply themselves to further study ( fang keng hsi-hsüeh 方更習學) in one of the three examinations held in that year, after an incident in which some candidates were caught cheating.54 48   t ctc, 204.6463; tfyk, 639.20a; thy, 76.1390. Quoted in tkck, 3.96. See also des Rotours’ comment on the phraseology of this entry in Traité des examens, 209, n. 1. 49  Guisso, Wu Tse-t’ien, 92–93, points out that the selection examinations and the regular doctoral examinations also took their definitive T’ang forms under empress Wu, in 669 and 681 respectively. 50   t kck, 5.158–60, and 5.172–74. 51  See the detailed discussion of this incident in P. A. Herbert, “A Hawk Among Rabbits: An Appraisal of the T’ang Chief Minister Li Lin-fu,” in Higashi Ajia no hō to shakai: Nunome Chōfu hakase koki kinen ronshū 東アジアの法と社会：布目潮渢博士古稀記念 論集 (Tokyo: Kyūko Shoin, 1990), 37–38. 52   t ctc, 215.6876. 53   t fyk, 645.16a; thy, 76.1388. Quoted in tkck, 9.313. 54   t kck, 9.321.

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

252

Herbert

Li Lin-fu had reformed the regular doctoral examinations in 737 as a deliberate ploy to regulate their functioning and to strengthen the controls placed upon candidates. It is possible that the regularity with which the decree examinations were held during Li Lin-fu’s period in power is another indication of his tidy-minded efficiency. Decree examinations were called once every two or three years during the period 736 to 752, with three examinations marking the change of reign-period name in 742 and three examinations again in 751, after Li Lin-fu had lost his total control over court politics. The titles of decree examinations frequently give a clear indication of the types of moral qualifications or skills required of the candidates who were to be submitted or to apply. Under Kao-tsung, one-fourth of the decree examinations call for men who have been living in obscurity or in retirement in the provinces. All these instances date between the years 659 and 666, early in empress Wu’s period of dominance. Another one-fourth call for men qualified for office as statesmen, metropolitan or provincial officials, administrative law experts, or confidential secretaries. Of the rest of the decree examination titles, the highest number request men of literary talent. During empress Wu’s reign, the titles of decree examinations place slightly less emphasis on statecraft and slightly more on literature, perhaps indicating a shift away from urgent need for counsellors close to the throne to more routine head-hunting of potential high-flyers. A notable change is in the introduction of titles connected with specific events. In 689, the Ming-t’ang ta-li k’o 明堂大禮科 (“Examination marking the great ceremony at the bright hall”) was held,55 while 691 saw an examination celebrating the “mysterious great sutra” 奧大經, that is, the Great Cloud Sutra (Ta yün ching 大雲經) “prophesy” of empress Wu’s rise to power as the incarnation of Maitreya Buddha.56 Hsüan-tsung’s early decree examinations place most emphasis on statesmanship, an urgent requirement for restoring T’ang greatness. Five out of the fourteen examinations held from 712 to 714 call for qualifications leading to high office. Another is for potential provincial officials, reflecting Hsüantsung’s concern for the poor caliber of provincial officialdom, neglected since T’ai-tsung’s reign. Four examinations are for literary men and one is for talent hidden away in the wilds. In the first half of Hsüan-tsung’s reign, when a number of eminent literary figures served at court and won the emperor’s 55  Ibid., 3.92. 56  Ibid., 3.97. For a detailed discussion of the Great Cloud Sutra and its use in the legitimation of empress Wu’s reign, see Antonino Forte, Political Propaganda and Ideology in China at the End of the Seventh Century (Naples: Istituto Universitario Orientale Seminario di Studi Asiatici, 1976), passim.

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

253

Decree Examinations in T ’ ang China

confidence and esteem, examinations searching for literary talent were most frequent. During Li Lin-fu’s dictatorship, however, the literary category diminishes markedly in importance, reflecting his impatience with literary niceties. No category of examination title is dominant during these years, suggesting that decree examinations were regarded as something of a routine exercise, but a new category is introduced. Two of Hsüan-tsung’s decree examinations held during Li Lin-fu’s dictatorship and one after his fall were Taoist in nature.57 They reflect Hsüan-tsung’s growing preoccupation with Taoist philosophy and mysticism and perhaps indicate a means by which Li Lin-fu drew his attention away from political activity. V

Challenges and Stultification: After the An-Shih Rebellion

All examination and personnel selection activity was severely disrupted, along with other administrative business, during the An-Shih Rebellion of 755 to 763. Appeals dating from 757 and 758 for men to come forward on their own initiative to serve as officials have already been mentioned. In 757, over half the passes awarded in the chin-shih examination, the regular doctoral examination of literary talent, were given in eastern China or Szechwan, areas remote from the theater of civil war.58 The fifth month of the year 759 saw Su-tsung hold four decree examinations, the first such examinations to be held in a regular, formal manner since the outbreak of the rebellion.59 After the year 757, when chin-shih candidates were selected decentrally and a large number of passes were awarded, presumably because the various examiners based in widely separated parts of China acted unilaterally and had no way of telling what was being done elsewhere, the annual number of chin-shih passes dropped to its normal level of between twenty and thirty (see graph) until 762, when the regular doctoral examinations were suspended. Later in the same year, Tai-tsung 代宗 came to the throne against a background of confusion, doubt and questioning of the whole personnel selection process. Scholars had been profoundly shaken intellectually by the rebellion, in addition to having their lives and careers physically disrupted. Some seriously questioned basic T’ang assumptions about bureaucratic selection. One of these scholars was Yang Wan, a brilliant but profoundly introverted poet, 57  The ming ssu-tzu k’o 明四子科 of 741, the yu-tao k’o 有道科 of 749, and the tung-hsiao hsüan-ching k’o 洞曉玄經科 of 754. 58   t kck, 10.344. 59  Ibid., 10.347.

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

254

Herbert

whose sensibilities had been wounded by his experiences while competing for the chin-shih degree. Later, he placed first in a decree examination in 754, which stressed literary skill. In 763, as vice-president of the Board of Rites, with responsibility for the regular doctoral examinations, he sparked off a vigorous debate at court by calling the whole T’ang examination system into question.60 Yang Wan thoroughly condemned the doctoral examinations, with their emphasis on exhibitionist behavior and vulgar competition. He wanted the chin-shih and the doctoral examination on the classics, the ming-ching 明經, abolished. In their place, he appealed to the new emperor: I request that Your Majesty order magistrates to scrutinize who is filial and incorrupt (hsiao-lien), to select those noted for good behavior in their villages and those who have studied and gained knowledge of the classics and to recommend them to the prefectures for the prefects to assess and examine (k’ao-shih 考試) them and send them up to the Department [Affairs of State]. Let each have a thorough knowledge of one classic. Confucian scholars at court are to be chosen to ask them twenty questions on the meaning of the classic and they are to answer three essay questions (tui-ts’e san-tao 對策三道). The first-class passes are to be given office immediately, the second-class passes to acquire official status (ch’u-shen 出身) and those placed in the third class are to fail and be sent home.61 Not only does Yang Wan’s suggested reform appeal to Han precedent and closely resemble the decree examinations in requiring candidates to compose three essays, but it even uses throughout terminology that closely resembles that used in the titles of decree examinations and the imperial announcements of the exams. It is scarcely, therefore, an exaggeration to say that Yang Wan advocated abolishing the regular doctoral examinations and transforming the irregular decree examinations into a regular institution to take their place. Yang Wan did not succeed in his bid to have the regular doctoral examinations abolished, but it is noteworthy that the titles of post-An-Shih Rebellion 60  For details on Yang Wan, see ChTS, 119.3429–37 and hts, 142.4664–66. For his decreeexamination pass, see also tfyk, 645.16a/b; thy, 76.1388; tkck, 9.333. For the debate, see also tctc, 222.7143–44. 61  The summary of Yang’s argument in tctc, 222.7143–44. A more detailed version is given in ChTS, 119.3430–32. Hsü Sung comments that Yang Wan was deputy director of the Court of Imperial Sacrifices (太常少卿), not vice-president of the Board of Rites (吏部 侍郎), but I have been unable to corroborate this from other sources.

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

255

Decree Examinations in T ’ ang China

decree examinations, especially under Tai-tsung and his immediate successor Te-tsung 德宗, make frequent allusion to the stereotyped phraseology of the Han recommendation system. Tai-tsung and his successors called for decree examinations at reasonably regular intervals, particularly in the early years of their reigns, suggesting that the precedent set by empress Wu and Hsüantsung was becoming institutionalized. Apart from greater use of “Han”-style terminology and more frequent repetition of the same titles, no particular category of candidate or qualification is preponderant in post-rebellion decree examinations. Neither does their dating reveal any particular relationship with events of great political significance, except for the fact that they are more frequent early in emperors’ reigns. All these characteristics suggest that the holding of decree examinations was becoming a convention. Institutionalization of irregular recommendations and examinations designed to fill specific needs within the bureaucracy at specific times is a contradiction in terms. Frequently in Chinese institutional history the irregular mode becomes regular and supplants its precursor, but despite Yang Wan’s efforts, this did not happen to the T’ang decree examinations. Instead, they lost their raison-d’être and died a sudden death in the early ninth century. VI

The Sudden End to the T’ang Decree Examination System

The last recorded decree examintion of which there is clear evidence took place in 828. In 827, following the accession of Wen-tsung 文宗, an Act of Grace was promulgated. The decree announcing this Act of Grace included a call for leading provincial officials to recommend men in the four categories of “capable and good, sincere and upright, able to be frank in speech and unflinching in admonition” (賢良方正能直言極諫); “in classical studies excellent and profound, capable of acting as an exemplar” (經學優深 可為師法); “with detailed experience in governing as an official to the extent of being able to exert civilizing influence” (詳閑吏理 達於教化); “in military strategy having broad skills, able to be employed as a general” (軍謀宏遠 堪任將帥). Men who had no one to recommend them might make personal applications. All candidates were to reach the capital by the first month of the following year.62 In the third month of 828, Wen-tsung personally received the candidates in the palace and impressed upon them the venerability of decree examinations, 62  In the title of one of the subsequent examinations, the phrase hsiang-hsien li-li is rendered hsiang-ming li-li 詳明吏部 “having a detailed understanding of governing as an official.”

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

256

Herbert

with their Han and Wei precedents and their seriousness as a means of finding talent for employment in the bureaucracy. The examinations continued according to the usual pattern, with the emperor as nominal chief examiner and three eminent court officials, Feng Su 馮宿, Chia Su 賈餗 and P’ang Yen 龐嚴, presiding on his behalf. Nothing in the description of the examinations suggests that there was any intention at this time to downgrade or phase out decree examinations. Even the premiers put in an appearance at this important state occasion.63 Results lists are given in Teng-k’o chi-k’ao for two of the civil examinations and the military examination. Only the examination on the classics is not mentioned.64 Perhaps decree examinations had become mere state occasions and candidates were not expected to take seriously the august decree titles for which they were competing or even the pious exhortations of the emperor. One of the candidates who competed in one of these examinations did take the title of his examination seriously, as is shown in a comment following the edict announcing the results: At the time, there was a certain Liu Fen 劉賁 who took the “Frank in speech and unflinching in admonition” examination. He answered point by point, taking great pains and writing a total of several thousand characters. Although he did not pass, the gist of his essay became known at the time.65 We are told that the hundred or so other candidates who took the same examination wrote commonplace, conventional answers. Although the examiners recognized Liu’s extraordinary talent and could not find fault with his argument, they dared not pass him. His essay caused a sensation among the intelligentsia. It was passed from hand to hand and men of good will who read it wept over its contents, but officialdom tried to hush up the embarrassing incident.66 Liu Fen’s notorious answer is preserved in his standard biographies. It is a scathing critique of contemporary politics and an attack on corruption and in-fighting among those in power, which court officials would not wish the

63   t kck, 20.740–41. 64  Ibid., 20.742–43 and 20.746–48. 65  Ibid., 20.743. 66   t hy, 76.1394, quoted in tkck, 20.743.

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

Decree Examinations in T ’ ang China

257

emperor to read.67 One brave scholar did dare to draw Wen-tsung’s attention to Liu Fen’s plight. He was Li Ko 李郃, a fellow candidate, one of the twentytwo who did pass the examination. He wrote a memorial requesting that he be allowed to relinquish the office awarded to him in order that it might be given to Liu Fen, who was more deserving of it. Li wrote: What is more, [the essays] which I wrote in answer fall very far short of [Liu] Fen’s. In my own heart, I feel ashamed to call myself “capable and good” (賢良, the opening phrase of the examination title). So what will others say? I beg leave to give back [the appointment] which I have received, that it may be used to reward [Liu] Fen’s honesty.68 The early ninth century was a period of frequent rebellion in the provinces and of factional strife at court. It was also a period of rigorous scrutiny of the regular examinations. It has been suggested that the “Niu-Li factional strife” between Niu Seng-ju 牛僧孺 and Li Te-yü 李德裕 and their respective followers was a fight between aristocrats who disliked the examination system and “new men” who favored it, but that simplistic argument has been discredited.69 The demise of the decree examination after 828 has nothing to do with antagonism between Niu Seng-ju and Li Te-yü. It seems most likely that the sensation caused by Liu Fen’s answer convinced court officials that the decree examinations had outlived their usefulness in an era when regular examinations could be used to bring in all the candidates needed for office. Decree examinations were less easy for officials to administer and their candidates less easy to control, since the emperor was directly involved. After 828, therefore, it appears that whenever emperors felt inclined to call for examinations by decree, excuses were found for not holding them. In 830, Wen-tsung called for recommendation and personal application of men in the same three civil categories covered in the 828 decree examinations. They were to come to court by the first month of 831.70 Early in 831 an order went out to suspend submission of men because of “military activities”—in 67   ChTS, 190C.5064–77; hts, 178.5293–306. According to his hts biography, Liu was later rehabilitated by Niu Seng-ju’s faction, but his downfall was engineered by court eunuchs and he died in exile. 68   t kck, 20.743–44. 69  See Michael T. Dalby, “Court Politics in Late T’ang Times,” in Cambridge History of China, vol. 3, 561–681, citing Tonami Mamoru, “Chūsei kizokusei no hōkai to hekishōsei: Gyū-Ri no tōsō o tegakari ni,” Tōyōshi kenkyū 21.3 (1962), 1–26. 70   t kck, 21.751.

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

258

Herbert

other words, because of the Nan-chao 南詔 invasion of Szechwan and possibly because of civil strife in the provinces.71 Previously, national emergencies had been reasons for holding decree examinations, not excuses for cancelling them. In 885 there appears to have been an attempt to revive the institution of decree examinations following Huang Chao’s 黃巢 rebellion, when Hsi-tsung 僖宗 was desperate to recruit scholars who had chosen or been forced to live in retirement because of the political situation. Although an edict of the fifth month of 885 states that provisions have been made for Confucian scholars or military men to sit for decree examinations,72 there is no record of such examinations taking place. Similarly, in the fourth month of 901, near the end of the T’ang dynasty, there is clear evidence of an attempt to revive decree examinations. Chao-tsung 昭宗, at the time of a change in reign-period, asked the regular officials and provincial governors of the empire to recommend and send in to the capital by the eleventh month commoners under three stereotyped categories. The emperor intended to set them an essay examination.73 In fact, in the eleventh month, Chao-tsung was not resident in the capital himself,74 so presumably the projected examinations, of which there is no further mention, were cancelled. Decree examinations were revived in the shortlived dynasties which followed the fall of the T’ang. They are recorded in 909, under the Later Liang dynasty and in 957 under the Later Chou.75 VII

Administration of the Decree Examinations

Turning from the history of the T’ang decree examinations, the question arises as to how such examinations were conducted. As has been mentioned above, the regular form for decree examinations, as set forth by empress Wu, was three essays (ts’e 策). Before the initiation of the formal decree examination system, however, various tests were sometimes administered to men recommended to the emperor in response to a decree.

71  For the order, see tkck, 21.753. For the Nan-chao invasion and civil strife, see tctc, 244.7872–74. 72   t kck, 23.883–84. 73  Ibid., 24.924. 74  Ibid., 24.924, quoting ChTS, 20A.773. 75   t fyk, 645.20b–21a and 22b respectively.

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

Decree Examinations in T ’ ang China

259

Allusion is made in T’ang hui yao and Teng-k’o chi k’ao to the tests undergone by candidates recommended in 644. An edict of the fifth month of the previous year called on prefectures and counties to submit outstanding scholars in the category “filial and incorrupt, possessed of abundant talent, devoted to scholarship, and of unusual ability” (hsiao-lien mao-ts’ai hao-hsüeh i-neng 孝廉 茂才 好學異能).76 In the third month of 644, candidates duly arrived from Fu-chou 酆州 (in modern Shensi province) and Pien-chou 汁州 (modern K’ai-feng 開封). T’aitsung treated them with great courtesy, allowing them to sit in his presence. The emperor then proceeded to conduct a test or interview. The first question he asked was about the art of government as practiced by the ancient emperors. The Teng-k’o chi k’ao text uses the expression ch’en liang-ts’e 陳良策: the emperor asked the candidates to “set out a good plan.” The context in which the character ts’e is used here suggests the earlier meaning “plan” or “policy statement” rather than the more specific meaning of “written examination answer.” The candidates did not come up with answers to this question (wu-tui 無對).77 The crown prince asked more questions about the classics and historical subjects, but still the candidates were incapable of answering (pu-neng ta 不能答). T’ai-tsung was saddened by the fact that, after he had sent for the most outstanding scholars in the empire, the men submitted for his scrutiny proved of such poor quality. The candidates then proceeded to the Secretariat, where they submitted essays (she ts’e 射策)—in this phrase, ts’e always means an examination essay—but their answers were still unsatisfactory. It is not altogether clear whether the tests in 644 were all conducted in writing or whether T’ai-tsung gave the preliminary test to candidates orally. From the usage of ts’e, I suspect that the preliminary test was oral. An incident involving the yüeh-mu chü 岳牧舉 (“submission of candidates for metropolitan and provincial office”) in 680 suggests that at least on this occasion the emperor did interview the candidates before they underwent a written test. The successful candidate Yün Pan-ch’ien 員半千 had shown remarkable promise in his youth and had been singled out by his teacher as “a worthy man of the kind who appears only once every five hundred years,” hence Yün’s choice of his new name Pan-ch’ien (“Half-millennium”). Yün had already passed a decree examination in 676 and had been made a district junior officer (hsienwei 縣尉). In this post, he was outspoken, asking his immediate superior, the magistrate, to open the grainstore to the starving people during a drought. 76  Ibid., 645.10b–11a, quoted in tkck, 1.23–24. 77   t kck, 1.24–26. Hsü Sung states that the characters wen chih 文之. are an interpolation. A slightly different version of this incident is given in thy, 76.1395.

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

260

Herbert

The prefect ordered that the granary remain closed, but was reprimanded by the commissioner for relief sent to inspect the area, who shamed the prefect by praising Yün’s initiative. Perhaps the commissioner recommended Yün for the examination, as he seemed so obviously suited for higher office. In its entry on the 680 submission of candidates, Teng-k’o chi k’ao carries a quotation from Yün’s Chiu T’ang shu biography to the effect that Kao-tsung received all the candidates in the palace and personally asked them questions. Yün, who had previously shown his initiative, articulateness, and lack of shyness, replied out of turn (yüeh-tz’u erh tui 越次而對); Kao-tsung was impressed and rewarded him, and when he came to sit for the essay test (chi tui-ts’e 及對策), he was given a high pass.78 The expressions “replied out of turn,” as well as “and when he came to sit for the essay test,” imply that a preliminary test was involved and that that test was conducted orally. A detailed picture of the conduct of a typical decree examination can be built up with more certainty from a number of references to decree examinations in the reigns of Jui-tsung and Hsüan-tsung and in the post-An-Shih Rebellion period. A decree of the twelfth month of 710, during the short reign of Juitsung, specifies how candidates for recommendation were to be pre-selected. A wide range of qualifications is demanded from seven categories of candidate: experts on the classics; history; the three ideologies of Confucianism, Taoism and Buddhism; calligraphy and phonology; musical composition and performance; strategy; and literary skills. The relevant authorities, including presumably those in charge of professional work in some of the fields mentioned, were to “select them widely and test them to make clear [their qualifications] and report” to the emperor, after which the emperor would himself make a selection from among them.79 A more orthodox decree examination was held in 738 to select men whose “literary expression was refined and beautiful” (wen-ts’u ya-li 文詞雅麗). Hsüantsung called upon the authorities to set out a feast for the candidates. He then addressed a homily to them, reminding them that written decree examinations dated back to ancient times and were designed to select talent suited to government service. Candidates were supposed to possess real talent but, of recent years, many candidates had sat for decree examinations, but few had merited passes because answers had mostly been hackneyed and irrelevant to the questions set. Hsüan-tsung went on to say:

78   t kck, 2, 69. For biographical information on Yün Pan-ch’ien, see ChTS, 190B.5014–15. 79   t fyk, 645.14a; thy 76.1392–93. Quoted in tkck, 4.151.

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

Decree Examinations in T ’ ang China

261

Our questions are divided into [topical] sections. You should answer as indicated. Do not employ frivolous language. He then invited all the candidates to sit down to a feast and, when the feast was over, they were to proceed to the examination. On this occasion, twenty-four men passed and “all received offices in accordance with their qualifications” (chieh liang-tzu shou-kuan 皆量資授官),80 The holding of a feast before the essay test was a regular feature of decree examinations. The emperor, or exceptionally the crown prince, was always the nominal examiner for decree examinations and always, except in the most exceptional circumstances, as noted below, put in an appearance. Few emperors, however, took their duty toward decree examination candidates as seriously as did Taitsung in the fourth month of 771. On this occasion, there were four categories of candidates assembled in one of the halls of the palace. The emperor personally encouraged the candidates, treating them with great courtesy. Besides calling for the regular feast, he presented them with delicacies, tea, and wine from his own table. He noticed with regret that some of the candidates were wearing worn-out clothing and had an unhealthy, undernourished look about them. Tai-tsung was moved to tears when he realized that his people, including these scholars, were suffering the grave after-effects of civil war. As well as feeding them, Tai-tsung provided new clothes for the poor scholars. All day during the examination, the emperor sat in a dignified attitude reading the account of his ancestor T’ai-tsung’s pursuit of good government, the Chen-kuan cheng-yao 貞觀政要, reading matter most appropriate to the occasion, since it contains many references to T’ai-tsung’s concern over finding and consulting wise ministers and worthy officials. Tai-tsung personally looked over a total of more than a hundred completed essays submitted for the examinations. When dusk came and some of the candidates had not finished writing their essays, Tai-tsung ordered candles for them so that they could continue writing until midnight without spoiling their train of thought.81 This is an important passage, since it reveals a number of significant points concerning the conduct of decree examinations. It shows, in the first place, that the examinations were the emperor’s personal concern and that on this occasion at least, he took them very seriously, remaining in the examination hall throughout the examination, and that he took a hand in the grading. The feast and the polite treatment of the candidates as honored guests are 80   t kck, 8.290. Tzu (“qualifications”) usually refers to seniority when used in the context of the selection examination system, but may have a wider meaning here. 81  Ibid., 10.376.

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

262

Herbert

emphasized. Candidates for decree examinations were obviously not always from wealthy or well-connected families and brought the emperor into direct contact with aspects of ordinary life outside the palace otherwise unfamiliar to him. Such revelations could come as a shock. Finally, decree examinations, unlike modern examinations, had no fixed time-limit and could last well into the night. Sometimes, particularly in later decree examinations, the de facto examiners are named, as in the case of the last examination held in 828 cited above and an examination of 785, nominally supervised by Te-tsung, when the actual examiners (shih-kuan 試官) were Pao Fang 鲍防 and Tu-ku Mien 獨孤愐.82 There is an interesting instance in 821 of an emperor holding a decree examination that had been called by his predecessor. In the seventh month of 819, Hsien-tsung 憙宗 announced that he would hold four decree examinations during the first month of 820. He died shortly after, before he could carry out his intention.83 In the second month of 820, his successor, Mu-tsung 穆宗, declared that he did not wish to conduct the four examinations himself, so he ordered some of his highest officials to take his place and hold the examinations in the Department of State Affairs late in the third month.84 In the third month, the president of the Board of Personnel, Chao Tsung-ju 趙宗儒, commented that many of the candidates had already given up hope and returned home, so the examinations should be abandoned. The emperor agreed.85 In 821, however, Mu-tsung held examinations for candidates in the same four categories.86 In a decree examination held in the fifth month of 745 for “men who have risen above the ordinary to live as recluses and have not received office” (kaotao pu-jen 高蹈不任), special courtesy was extended to five candidates because of their extreme old age. They were given official robes and other presents, but they were allowed to resume the hermit life to which they had chosen to dedicate themselves. On this occasion, official transport was provided to and from the capital for the candidates.87 Normally, decree examination questions were restricted to essay questions, usually three in number. In the decree examination of 754 for men whose “literary style was expansive and beautiful” (tz’u-ts’ao hung-li 詞藻宏麗) for which 82  Ibid., 12.423. 83  Ibid., 18.675. 84  Ibid., 18.680. 85  Ibid., 18.680. 86  Ibid., 19.685 and 19.693–95. 87  Ibid., 9.307.

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

Decree Examinations in T ’ ang China

263

Yang Wan sat, tests included essays, one shih-poem, and one rhyme-prose ( fu). This is said to have been the first occasion on which poetic compositions were tested in a decree examination.88 Literary compositions of some kind were however included in the “multi-talent examination” (to-ts’ai k’o 多才科) of 733, which specified that candidates “were to be tested by means of three major essays deliberating upon governing the state and also tested by means of three miscellaneous compositions; those who show elevated literary style are to be chosen” (試經國高略大策三道并 試雜文三道取其詞氣髙者).89 Corrupt practices and even cheating were sometimes associated with the decree examinations, as with all forms of examination in T’ang China and elsewhere. Hsüeh Ch’ien-kuang 薛謙光, an omissioner (pu-ch’üeh 補闕), complained in a memorial of 692 that candidates sought favors from officials and powerful persons. Such candidates were known as “seekers after recommendation” (mi-chü 覓舉), in other words, “self-seekers” (tzu-ch’iu 自求).90 The phrase tzu-ch’iu is very different in meaning and implication from tzu-chü (“personal application”). As suggested in the English translation, by tzu-ch’iu, Hsüeh implied pushing oneself forward with the aim of being selected, rather than the lofty classical ideal of waiting to be sought out for one’s superior qualities as was implied in the titles of decree examinations aimed at recruiting recluses and unknown men of talent. Hsüeh goes on to claim that some candidates even falsified their credentials and hid criminal records in order to be recommended by their local authorities. In 751, as mentioned above, some candidates cheated in a decree examination. They carried cribs in the form of model essays into the examination hall. They were punished and their guarantors (so-pao chih kuan 所保之官) were exiled. Candidates for selection were always required to name metropolitan officials as guarantors.91 Evidently decree examination candidates also had guarantors. The guarantors were presumably the officials who recommended them. A decree of 808 details the precautions taken on at least this one occasion to make sure that essays were genuine, that candidates and examiners were kept apart, and that nothing regarding the examination was revealed to outsiders:

88   t fyk, 643.14b–15a, quoted in tkck, 9.330. See also thy, 76.1393. 89   t kck, 8.264. From 781, if not earlier, “miscellaneous compositions” (tsa-wen) in the context of the chin-shih examination meant shih and fu. 90   t t, 17.12a–13b. 91  For the 751 examination, see note 54. Provisions for guarantors in the selection examinations are specified in tt, 15.40a.

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

264

Herbert

When the candidates have completed their test, as night comes on, they are to hand over their essays. A count is to be taken of those who cannot return home and they are all to spend the night in the Kuang-chai Temple 光宅寺. There must be policing of officials and petty officials responsible for the examination and their attendants. After the candidates have given in their papers, they (the officials, etc.) are to proceed to the Pao-shou Temple 保壽寺 to spend the night. Furthermore, the Chin-wu Guard 金吾衛 are to depute a runner to take charge of leading each group, to accompany them to the place where they are to spend the night. [Officials] responsible are not to be garrulous about the examination.92 Candidates who passed decree examinations, particularly those who achieved high passes, were often given office immediately or shortly afterwards. Such awards of office are consistent with the two aims of the decree examinations stated in the introduction to this paper: drawing unrecognized talent into the bureaucracy and promoting promising officials. The T’ang work Feng-shih wen-chien chi 封氏聞見記 discusses the decree examinations and their relationship to career patterns. Passing a decree examination, like passing the chih-shih, the regular literary degree, placed a man on the first rung of the ladder of “eight eminent offices” (pa-chün 八雋), in other words, pure offices (ch’ing-kuan 清官), which brought him into close contact with the emperor and might eventually lead through accelerated promotion to a premiership.93 Feng Yen 封演, author of Feng-shih wen-chien chi, also draws attention to the anomaly that whereas a chin-shih pass and a pass in a decree examination were both accounted the first of the eight steps to high office, a chin-shih pass carried more prestige. Immediately after detailing the “eight eminent offices,” Feng gives an account of eight brothers, seven of whom gained chin-shih degrees while the eighth passed in a decree examination. The seven chin-shih refused to have their brother sit with them and made humiliating jokes about him.94 This incident recalls a comment made by Shen Chi-chi 沈旣濟, writing shortly after the An-Shih Rebellion. In a passage on the T’ang examination system, he says: 92   t hy, 76.1393, quoted in tkck, 17.628. 93   Feng-shih wen-chien chi chiao-cheng 封氏聞見記校證 (Peking: Yenching Univ. Library edition, 1933), 3.11a. Also quoted in T’ang yü-lin 唐語林 (Taipei: Shih-chieh, 1967), 8.277–78. 94   Feng-shih wen-chien chi chiao-cheng, 3.11a.

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

Decree Examinations in T ’ ang China

265

Some dragged men out of seclusion, mocking them in verses, and as they were lined up by the roadside [ready to go to the capital to take the examination], continued to make them the butts of ribaldry.95 In the passage as a whole, Shen refers mainly to the chin-shih candidates of Hsüan-tsung’s reign, but presumably here he refers also to decree examination candidates, particularly those of the reclusive type who were persuaded, perhaps against their better judgment, to take the examinations. The lower estimation accorded decree-examination graduates in comparison to chin-shih graduates applied only to those officials whose route of entry into the bureaucracy was via decree examinations. They were figuratively, if not literally, the hungry men in tatters whom Tai-tsung pitied and at whom less sensitive people jeered. The high-flyers who achieved the greatest success in the decree examinations were those serving officials who often already possessed chin-shih degrees, came to the personal attention of emperors through passing the decree examinations, and received promotions to posts in close contact with the emperor and his highest officials. For them, the decree examinations represented the narrow gateway to preeminence. Abbreviations ChTS Chiu T’ang shu 舊唐書. Peking: Chung-hua, 1975. hts Hsin T’ang shu 新唐書. Peking: Chung-hua, 1975. tctc Tzu-chih t’ung-chien 資治通鑑, Peking: Ku-chi, 1956. tfyk Ts’e-fu yüan-kuei 册府元龜. 1643 edition; rpt. Peking: Chung-hua, 1960. thy  T’ang hui yao 唐會要. Basic Sinological Series; rpt. Peking: Chung-hua, 1957. tkck Teng-k’o chi k’ao 登科記考. Peking: Chung-hua, 1984. tt  T’ung tien 通典. Sung edition held in the Kunaichō Sho-ryōbu 宮內廰書 陵部, Tokyo, compiled with introduction by Nagasawa Kikuya 長澤規矩也 and Ozaki Yasu 尾崎康; rpt. Tokyo: Kyūko Shoin, 1980–81. whtk  Wen-hsien t’ung-k’ao 文獻通考 Yüan edition held in the Sei-kadō Bunko 靜嘉堂文庫, Tokyo.

95   t t, 15.39a.

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

266

Herbert

KEY: A = number of passes in original chin-shih examination. B = number of passes in retake examination. C = figure for the year 757 including 29 passes in the provinces. D = figure for the year 757 omitting provincial passes. NOTES: 1. Up to the year 650, hsiu-ts’ai pass figures are added to chin-shih pass figures. 2. Regular doctoral examinations were suspended in the years 628, 642, 652–53, 663, 669, 671–72, 676–79, 694, 709, 762, and 804. All chin-shih candidates failed in 665. 3. For 666, the number of chin-shih passes is not recorded. Hsü Sung suggests the names of two possible successful candidates. There are no extant records for the years 692, 711, and 725.

Graph showing incidence of named decree examinations compared with numbers of chin-shih passes: Years 620 through 840 Sources: Wen-hsien t’ung-k’ao and Teng-k’o chi k’ao

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

The Bureaucratic Apparatus [of T’ang Historians] Denis Twitchett Before the T’ang period, the composition of official history had been the concern of two separate groups of officials: the court diarists, responsible for keeping the day-by-day record of the activities of the emperor and the court, and the Office of Literary Composition, responsible for the actual writing of the national record. This division of functions continued under the T’ang but became far more complex.1

The Court Diarists

The existence of court diarists can be traced back to the dawn of Chinese recorded history. The diarists of T’ang times2 saw themselves as the successors to the various recorders mentioned in the Chou li3 and to the recorders of the left and of the right (Tso-shih, Yu-shih) mentioned in the Li chi4 and in the Tso chuan.5 Source: “The Bureaucratic Apparatus,” in Denis Twitchett, The Writing of Official History under the T’ang, Cambridge: Cambridge University Press, 1992, 5–30. © Cambridge University Press, reproduced with permission. 1  See tt 21, p. 124a; tfyk 554, p. 3b; Wang Ying-lin, Yü hai (Taipei, 1964) 48, p. 3b. The division of functions was first formalized under the Northern Chou (559–79). See Wang Chung-lo, Pei-Chou Liu-tien (Peking, 1979), vol. 1, p. 191. 2  See, for example, Liu Chih-chi, Shih t’ung, “wai-pien” 1, Shih t’ung t’ung-shih (Shanghai, 1978) 11, pp. 304ff. 3  The Chou li mentions four different categories of recorders, the T’ai-shih, Hsiao-shih, Nei-shih, and Wai-shih. See Edouard Constant Biot, Le Tcheou-li; ou, rites des Tcheou (Paris, 1851), vol. 2, pp. 104–20. 4  See Li chi, “Yü-ts’ao”; Seraphim Couvreur, Li Ki ou mémoires sur les bienséances et les cérémonies (Ho Kien Fou, 1913; reprint, Paris, 1951), vol. 1, pp. 678–9; James Legge, The Li Ki, Book of Rites (Oxford, 1885), vol. 2, p. 2. The T’ang commentary to Li chi identifies the recorder of the left with the T’ai-shih of the Chou li, and the recorder of the right with its Nei-shih. 5  Tso-chuan mentions the recorder of the left under Duke Hsiang, fourteenth year, and Duke Chao, twelfth year. See Seraphim Couvreur, Tch’ouen-ts’iou et Tso tchouan; La chronique de la principauté de Lou (Ho Kien Fou, 1914; reprint, Paris, 1951), vol. 2, p. 297; vol. 3, p. 207; James Legge, The Chinese Classics, vol. 5, The Ch’un-ts’ew with the Tso Chuen (London, 1872; reprint, Hong Kong, 1960), pp. 464, 641. There seems to be no early mention of the recorders of the right. William Hung, “A T’ang Historiographer’s Letter of Resignation,” hjas 29 (1969): 5–52,

© koninklijke brill nv, leiden, ���9 | doi:��.��63/9789004380158_010

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

268

Twitchett

According to a tradition generally believed in T’ang times, the functions of these officials had been strictly distinguished, the recorder of the left having been responsible for noting down the emperor’s actions, the recorder of the right for taking down his words. T’ang scholars believed that these two officials had produced distinct types of records from which had originated two separate genres of ancient historical writing. The Spring and Autumn Annals (Ch’un ch’iu) and the Tso Tradition (Tso chuan) were essentially concerned with the actions6 of rulers, the traditional business of the recorders of the left, whereas the Book of Documents (Shang shu) and the Discourses of the States (Kuo yü) merely recorded their utterances, as had the diary kept by the recorder of the right.7 This somewhat naive theory of a division of roles between the diarists and its connection with the origins of the earliest genres of historical writing does not bear very close examination, but the scholar-historians of T’ang times believed it and took it seriously as an ideal institutional arrangement to which they themselves tried to conform.8 In the period of division following the collapse of the Han the titles and organization of the court diarists underwent many changes.9 The deliberate separation of the routine keeping of the Court Diary from the process of statesponsored historical composition was finally formalized under the Northern Chou and under the Sui.10 Under the Sui the composition of the national esp. 17–18, n. 18, suggests that the original meaning of Tso-shih was not “recorder of the left” but “assistant scribe” and that the term “recorder of the right” (Yu-shih) was a Han invention based on a misunderstanding of this original meaning. For a detailed discussion of these early recorders or annalists, and an attempt to make coherent sense of the relations between the recorders of the left and the right and the other titles mentioned in early literature, especially the Chou li, see Chin Yü-fu, Chung-kuo skih-hsüeh shih (Peking, 1962), pp. 3–19. 6  Although the Tso chuan is full of speech and dialogue, the Spring and Autumn Annals itself contains no direct speech whatever. 7  This division of functions between the different recorders, and the identification of their work with the Spring and Autumn Annals, on the one hand, and the Book of Documents, on the other, had already been made in the bibliographical monograph of Pan Ku’s Han shu. See hs 30, p. 1715; Han shu i-wen chih (Shanghai, 1955), p. 13. However, Pan Ku reverses the respective functions ascribed to the recorders of the left and the right in the Li chi, the account usually followed by T’ang scholars. 8  See, for example, the response of Li Chi-fu (758–814) to Hsien-tsung’s enquiries about the Record of Administrative Affairs (Shih-cheng chi) in the tenth month of 813, recorded in thy 64, p. 1109. See also the edict dated the ninth month of 817, and the memorial from the diarist Yü Ching-hsiu that followed its promulgation, in thy 56, pp. 962–3. 9  See tlt 8, pp. 23b–24a; tt 21, p. 124a; tfyk 554, p. 3b. These changes are set out in tabular form in Chin Yü-fu, Chung-kuo shih-hsüeh shih, pp. 88–9. 10  However, in early T’ang times this rule was not very strictly enforced. Under Kao-tsung, for example, Ku Yin, a professional historian who was one of the compilers of the Veritable Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

The Bureaucratic Apparatus [ of T ’ ang Historians ]

269

record was undertaken by a subdepartment of the Imperial Library (discussed later), as was responsibility for the calendar, while the keeping of the Court Diary became the responsibility of officials entitled Ch’i-chü she-jen on the staff of the Imperial Secretariat (Chung-shu sheng). At first the T’ang continued this practice,11 but in 628 the diarists of the Secretariat were replaced by similar officials with the title Ch’i-chü lang serving in the Imperial Chancellery (Menhsia sheng). This change presumably reflects the increasing prestige of the Chancellery at the expense of the Secretariat. In the period 656–61 the diarists attached to the Secretariat were also revived, and from this time onward the T’ang court maintained two parallel groups of official diarists, who theoretically divided their functions on the lines described in the Li chi and other early texts: The Ch’i-chü lang attached to the Secretariat were responsible for recording the emperor’s actions, whereas the Ch’i-chü she-jen under the Chancellery recorded his words. The identification of these posts with the models provided by antiquity was made specific when first during the period from 662–70 under Kao-tsung, and again during the “Chou dynasty” under Empress Wu Tse-t’ien (690–705) the titles of many posts in the bureaucracy were changed to follow the ancient models of the Chou li. Under these reforms the diarists again became recorders of the left and of the right.12

Record for T’ai-tsung’s reign and of the National History of the first two reigns, the Wu-te Chen-kuan liang-ch’ao kuo-shih, was appointed concurrently as a court diarist and as a compiler of the National History at some time between 650 and 656. See cts 73, p. 2600. Similarly, in the 670s, Li Jen-shih, while holding the post of diarist of the left (Tso-shih), also worked on additions to the unsatisfactory National History that had been completed in 659 under Hsü Ching-tsung. See thy 63, p. 1094; cts 73, p. 2601; Shih t’ung 12, p. 373. During the reign of Chung-tsung (705–10) after the completion of the Veritable Record of Empress Wu’s reign (Tse-t’ien Shih-lu) in 706, the assistant in the Historiographical Office, Wu Ching, who had been one of the compilers, was appointed court diarist (Ch’i-chü lang) and apparently continued to work as an official historiographer as before. See cts 102, p. 3182. 11  On the court diarists under the T’ang, see tlt 8, pp. 23a–25a; cts 43, p. 1845; hts 47, p. 1208 (Robert des Rotours, Traité des fonctionnaires et traité de l’armée [Leiden, 1947], vol. 1, pp. 152–63) for the Ch’i-chü lang, and tlt 9, pp. 19a–20b; cts 43, pp. 1850–1; hts 47, p. 1212 (Rotours, Fonctionnaires, vol. 1, p. 187), for the Ch’i-chü she-jen. See also thy 56, pp. 961–5, and tt 21, pp. 123c–124a, for general discussion. The T’ang court was not alone in continuing the Sui practice. Their rival successor to the Sui in Hopei, Tou Chien-te, selfstyled emperor of the Hsia Dynasty (617–21) employed as Ch’i-chü she-jen Wei Cheng, later to become famous as a minister and historian during T’ai-tsung’s reign, who had fallen into his hands in 619. See cts 71, p. 2546; hts 97, p. 3868. 12  See tlt 8, p. 24b; tlt 9, p. 19b; tt 21, p. 124a; tfyk 554, p. 4a, and the other sources listed in n. 11. From 662 to 670 the Ch’i-chü lang in the Chancellery were given the title Tso-shih recorders of the left. They were again given this title under Empress Wu from 690 to 705. Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

270

Twitchett

Although the conception of the diarists’ functions was an ancient one, the titles Ch’i-chü she-jen and Ch’i-chü lang were recent innovations.13 These titles derived from the name of the Court Diary, the Ch’i-chü chu, for which they were responsible.14 This title for the Court Diary, the “Diary of Activity and Repose,” was a long-established one that can be traced back at least to Han times.15 In the pre-Sui period of division the Ch’i-chü chu had been the standard chronological record compiled for each reign.16 The Ch’i-chü she-jen under the Secretariat were entitled Yu-shih, recorders of the right, during these same periods. 13  The title Ch’i-chü she-jen had first been established under the Sui and was subsequently taken over by the T’ang. The parallel Ch’i-chü lang were first established in 628. See tlt 8, p. 24a–b; cts 43, p. 1845. 14  Until 628 it seems that the Office of Literary Composition (Chu-tso chü) actually composed the Court Diary, reducing the record kept by the diarists to a final polished form. See cts 43, p. 1845. 15  The ss Monograph on Literature traces the Diaries of Activity and Repose back to the Ch’i-chü chu said to have been kept in the Inner Palace during the reign of Han Wu-ti (141–86 bc), and to that kept for Ming-ti of the Later Han (ad 57–75) by his empress née Ma. See hhs 10A, p. 410. The monograph also speculates that they may have originated as Inner Palace records kept by women scribes. See ss 33, p. 966; Chang-sun Wu-chi, Sui shu ching-chi chih (Shanghai, 1955), p. 49.  Modern scholarship rejects this account of the origins of the Diaries of Activity and Repose. Charles S. Gardner, Chinese Traditional Historiography (Cambridge, Mass., 1938), p. 88, disposes of the record connecting the diaries with Wu-ti’s time, pointing out that the only evidence for this comes from a spurious text, the Hsi-ching tsa-chi. Hans Bielenstein, “The Restoration of the Han Dynasty: With Prolegomena on the Historiography of the Hou Han shu,” bmfea 26 (1954): 1–210, esp. 21–2, also disputes the idea that the diaries were kept by female clerks and makes the important observation that the context in which the diaries kept by Empress Ma in Ming-ti’s reign are mentioned suggests that such records were already commonplace. A. F. P. Hulsewé, “Notes on the Historiography of the Han Period,” in Beasley and Pulleyblank, Historians of China and Japan, p. 41, following Chu Hsi-tsu’s “Han Shih-erh shih chu-chi k’ao,” Kuo-hsüeh chi-k’an 2, no. 3 (1930): 397–409, points out that a more important origin of the later Court Diaries was the Recorded Notes (chu-chi) kept by the imperial astronomers of Han times. This same idea was already expressed in early T’ang times by Yen Shih-ku (581–645) in his commentary to the hs Monograph on Literature: It was repeated by Wang Ying-lin in Yü hai 48, p. 20a–b.  The ss account of their origin, however, was generally believed in T’ang times. See, for example, Liu Chih-chi, Shih t’ung, 11, p. 324, a passage that closely follows the account in the ss monograph, and Tu Yu, T’ung tien (tt 21, p. 124a). For a general discussion of the Court Diaries, see also Chu Hsi-tsu, “Han, T’ang, Sung Ch’i-chü chu k’ao,” Kuo-hsüeh chik’an 2, no. 4 (1930): 629–40. 16  The ss bibliographical monograph (ss 33, pp. 964–66; Sui shu ching-chi chih, pp. 48–49) lists forty-four works in this category, the earliest of them dealing with the reign of the last Hou Han emperor, Hsien-ti (reigned 189–220). Some of them, like the Chin Ch’i-chü chu in Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

The Bureaucratic Apparatus [ of T ’ ang Historians ]

271

After the middle of the seventh century the two pairs of diarists were of equal rank, holding the lower six rank upper division.17 They were thus quite important officials, of equal standing with the under secretaries (yüan-wai lang) of the six executive boards. From Sui times onward, they had been classed as “officers in constant attendance” (shih kuan), that is, part of the emperor’s personal retinue, as his ceremonial guard of honor (i-chang) was. The posts were also classified as “pure offices” (ch’ing-kuan), that is, offices reserved for men of notable family, proved rectitude, and high moral repute.18 They were thus members of an elite group of officials, which included the higherranking grand counselors (San-ch’i ch’ang-shih) and counselors (Chien-i ta-fu) and the lower-ranking omissioners (Pu-chüeh) and remembrancers (Shih-i), all of whom were also on the staff of the Chancellery and Secretariat. All these officials, whose court functions were closely linked to the emperor himself, were expected to exercise criticism and a sort of moral censorship over the emperor’s pronouncements and actions. They were allowed considerable freedom of speech to exercise this function and had influence far greater than their relatively low ranks would suggest. The diarists were not expected to be mere passive observers, or high-class stenographers taking minutes of the court meetings. Like the remembrancers and omissioners, they were expected to remonstrate against policies they felt to be unwise or ill-considered, particularly when matters of historical precedent or analogy were involved. For example, in 727 the famous historian Wei Shu, then employed as a diarist, Ch’i-chü she-jen, remonstrated successfully with the emperor for neglecting a long-established practice by not observing mourning and suspending the court as a mark of respect for the deceased former chief minister, Su T’ing.19 317 chapters by Liu Tao-hui and the Hou Wei Ch’i-chü chu in 336 chapters, were very large works indeed. No fewer than twenty-one of them dealt with various reigns of the Chin dynasty. The bibliographical monograph of cts also lists twenty-six works in this genre, twenty-one of them relating to the Chin, as being included in the Imperial Library in 721. See cts 46, pp. 1997–8; T’ang shu ching-chi i-wen ho chih (Shanghai, 1956), pp. 91–3. 17  See tlt 8, p. 23a; tlt 9, p. 19a, etc. See also Chang Jung-fang, T’ang-tai ti Shih-kuan yü shihkuan (Taipei, 1984). 18  See the manuscript table of official posts dating from the T’ien-pao reign period (742–56), discovered at Tun-huang and now in the Bibliothèque Nationale, Fonds Pelliot Chinois Touen-houang P. 2504, published in Yamamoto Tatsuro, Ikeda On, and Okano Makoto, eds., Tun-huang and Turfan Documents concerning Social and Economic History, vol. 1 (B), Legal Texts (Tokyo, 1978), p. 90 (Document xxii (4), section 23). 19  See thy 56, p. 962. Su T’ing died on either the fourth (according to cts 8, p. 91), or the ninth (according to tctc 213, p. 6778) day of the seventh month of 727. The incident is curiously omitted from Wei Shu’s biographies. Because Wei was a protégé of Chang Yüeh, Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

272

Twitchett

In 819, another diarist, P’ei Lin,20 was exiled to a provincial post in disgrace for presenting a lengthy and scathing memorial attacking the emperor for his increasing reliance on alchemists and their elixirs, which ended with the following justification of his doing so: I submit that since the Chen-kuan era there have been numbered among the Diarists of the Left and Right men such as Ch’u Sui-liang, Tu Chenglun, Lü Hsiang, Wei Shu and so forth, all of whom have given the utmost of their loyalty and sincerity, who have with all their hearts remonstrated with their emperors. Your petty servant has by error been placed among your majesty’s retinue, and has received this post among the “officers in attendance,” as one of your close attendants [tsui-chin tso-yu]. Since the Tso chuan says “The intimate officers of the King should devote themselves entirely to remonstrating with him,” for an officer in close attendance on the emperor it is truly his fundamental duty to express himself to the throne with the utmost of loyalty and sincerity.21 Not every diarist measured up to this ideal. But it is important to remember, in the context of the didactic preoccupations of traditional Chinese historiography, that the basic material for the historical record, the Court Diary, was written not by mechanical reporters of what occurred but by officials holding posts with serious political and moral responsibilities, who saw themselves and were perceived by others as active participants in, and commentators on, state affairs. The introduction of moral criteria into the historical record who had been removed from office as chief minister in the fourth month of 726, and then forced to retire in the second month of 727, Wei Shu may have been protesting on his patron’s behalf about this slight offered to the memory of a statesman of unblemished reputation. Wei Shu’s personal connection with Chang Yüeh continued: In the second month of 728 Chang Yüeh, though retired, was appointed to the Chi-hsien Academy and given overall responsibility for the National History on which Wei Shu was also engaged. 20  P’ei Lin (d. 838) has biographies in cts 171, pp. 4446–50; hts 118, p. 4287; and a brief note in ctw 713, p. 19a. 21  See thy 56, pp. 963–5; tfyk 546, pp. 17a–19a; cts 171, pp. 4447–8. The incident is dated in the tenth month by thy 56, p. 963, and in the eleventh month by cts 15, p. 471. Among other suggestions, P’ei Lin proposed that before the emperor took any elixir the alchemist who had compounded it should be made to ingest it for twelve months to prove whether it was efficacious. His concerns proved only too real. Two months later the emperor was dead, almost certainly poisoned by elixirs containing massive doses of mercuric salts compounded by an alchemist called Liu Pi. On this incident, see Joseph Needham, Clerks and Craftsmen in China and the West (Cambridge, 1970), pp. 317–19.

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

The Bureaucratic Apparatus [ of T ’ ang Historians ]

273

was therefore not simply an element introduced in the later stages of the historiographical process when the historians were writing up a considered verdict on the events of a given reign. The application of Confucian22 moral judgments to the recording of events began with the men responsible for the very first stage of the record.23 The identification of the diarists with the other “pure officials” in attendance is underlined, particularly in the eighth and ninth centuries, by the frequent promotion to diarist of men who had served as omissioner or remembrancer,24 and by the frequent appointment of remembrancers, omissioners, grand counselors, and counselors to concurrent duties in connection with the compilation of the National History.25 There was a high degree of specialization and professionalism in the careers of T’ang historians, and one aspect of such specialization was the regular transfer of officials between posts connected with historiography and the “pure” remonstrative posts that required a similar knowledge of precedents and exercise of moral judgment on events. Such 22  Although the histories made their judgments in the light of conventional Confucian ethical standards, not all historians were necessarily Confucians. For example, Li Ch’un-feng, the author of the Monographs on Astronomy, on the Calendar, and on Portents in both the Chin history and the Monographs on the Five Dynasties (later incorporated into the ss) who also served as an official historian in the early years of Kao-tsung’s reign, was strongly influenced by Taoism and was the son of a noted Taoist scholar, Li Po. He, his son, Li Yen, and his grandson, Li Hsien-tsung, all served as astronomers royal in the seventh century. See cts 79, pp. 2717–19; hts 204, p. 5798. Under Hsüan-tsung a Taoist priest, Yin Yin (also read Yin An), was appointed in 737 to the control of the Historiographical Office (as chih Shih-kuan shih) and even received a special dispensation to perform his duties wearing his Taoist religious robes. See thy 63, p. 1101; hts 200, p. 5702. 23  The didactic purpose of the Court Diary is spelled out in the specification of the duties of the diarists given in cts 43, p. 1845. These duties were of course defined in the section of the Statutes relating to the office, from which the passage in cts derives. 24  For examples, see Sun Kuo-t’ung, T’ang-tai chung-yang chung-yao wen-kuan ch’ien-chuan tu-ching yen-chiu (Hong Kong, 1978), pp. 314–25, tables 2, 4, 5. These show that more than half (twenty-seven of fifty-one known cases) of the diarists whose careers are analyzed had previously served either as omissioner or as remembrancer. Table 6 (pp. 326–27) also shows that almost all the diarists, particularly after the end of the seventh century, were next promoted as under secretaries of one or other of the six boards. Of the sixty-one individuals whose careers are analyzed, all but eight were promoted directly into higherranking “pure offices” in the central government. 25  Such a combination of roles was, however, sometimes the object of protests. In 795, for example, P’ei Yen-ling objected to the appointment of one of the grand counselors as a compiler of the National History on the grounds of an implicit conflict of interest between the two offices. See thy 63, p. 1101.

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

274

Twitchett

a career track tended to keep the individual within the inner circle of court advisers. The record kept by the diarists was thus intimately connected with those in power and was almost exclusively concerned with events that affected the central administration.

The Office of Literary Composition

Until the early T’ang the composition of the dynastic record as distinct from the actual day-to-day recording of court events as they occurred had been the responsibility of the Office of Literary Composition (Chu-tso chü), a subdepartment of the Imperial Library (Pi-shu sheng).26 The Imperial Library also controlled the Imperial Observatory (T’ai-shih chü), headed by the astronomer royal (T’ai-shih ling), the officer responsible for the calendar.27 This officer had an ancient and intimate connection with the work of the historians28 and continued to provide a variety of information for the historical record throughout the T’ang period.29 The Office of Literary 26  For details on the Office of Literary Composition, see tlt 10, pp. 19b–23a; cts 43, p. 1855; hts 47, p. 1215 (Rotours, Fonctionnaires, vol. 1, pp. 207–8). See also tt 26, pp. 155c–156a; thy 65, p. 1123, and Li Hua, “Chu-tso lang t’ing pi chi,” in ctw 316, pp. 5a–7b; wyyh 799, pp. 2b–4b. See also David McMullen, State and Scholars in T’ang China (Cambridge, 1988), p. 22. 27  For details on the Imperial Observatory, see tlt 10, pp. 23a–34b. The descriptions given in cts 43, pp. 1855–6, and hts 47, pp. 1215–17 (Rotours, Fonctionnaires, vol. 1, pp. 208–17), describe the office after its complete reorganization in 758, when the Imperial Observatory (T’ai-shih chü) became the Ssu-t’ien t’ai, a large astronomical service totally independent of the Imperial Library. After 758 its director held equal rank with the director of the Imperial Library and controlled a staff of 138 ranking officials, with many assistants. 28  In the Chou li the T’ai-shih was responsible both for the historical record, astronomical observations, the calendar, and the various rituals marking the seasons. Under the Han the T’ai-shih ling, the “prefect grand astrologer,” was likewise responsible for the calendar, arrangements for the appropriate dates for ritual observances, and recording omens and portents. See H. Bielenstein, The Bureaucracy of Han Times (Cambridge, 1980), pp. 19 and 163 (n. 61–62). Wang Yü-ch’üan, “An Outline of Central Government of the Former Han Dynasty,” hjas 12 (1949): 151, says that he kept the Court Diary. Bielenstein finds no evidence to support this, however. 29  See thy 63, p. 1089; wthy 18, pp. 293–4. See also tlt 10, p. 26a, according to which at the end of every quarter the Imperial Observatory should record all the omens and portents it had observed and send these to the Chancellery-Secretariat so that they might be entered in the Court Diary. At the end of every year they should also make a separate comprehensive

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

The Bureaucratic Apparatus [ of T ’ ang Historians ]

275

Composition was called the Chu-tso ts’ao, as under the Sui, until 621. Under the Sui it had employed two chief secretaries and eight assistant secretaries: the latter increased to twenty under Emperor Yang-ti. When the T’ang renamed the office Chu-tso chü in 621, its size was drastically reduced to two chief secretaries and four assistant secretaries. This reduction is inexplicable, because the T’ang government shortly after began an ambitious program of compiling standard dynastic histories for the Sui and the preceding dynasties in which the Office of Literary Composition played a leading role.30 In 629, however, with the establishment of the Historiographical Office (Shih-kuan), the Office of Literary Composition ceased to have any direct formal responsibility for the compilation of the historical record,31 although members of its staff, and also members of the staff of the Imperial Library, frequently appear among the lists of editors and compilers of historical works.32 From 662 to 670 the office was renamed the Ssu-wen chü. In 738 the office was still further reduced in size, and by this time its functions were restricted to the composition of funerary memorial inscriptions, the texts of prayers addressed to the spirits, and the liturgical texts employed in the state sacrifices.33 Nevertheless, posts in the office, like all positions in the Imperial Library,34 remained prestigious appointments for a man of literary talent and were a desirable step in the elite career of

record to be sent under seal to the Historiographical Office, thus providing a double check on such information. 30  See William Hung, “The T’ang Bureau of Historiography before 708,” hjas 23 (1960–1): esp. 94–98. 31  See thy 65, p. 1123; thy 63, p. 1089; tt 21, p. 126c; tlt 9, p. 29b, commentary. 32  See Li Hua, “Chu-tso lang t’ing pi-chi,” wyyh 799, p. 4a. 33  See Li Hua, “Chu-tso lang t’ing pi-chi,” wyyh 799, p. 4a. The reduction in the size of the office was probably the result of the growing influence under Hsüan-tsung of the Chihsien and the Han-lin academies, the scholars of which were increasingly used to compose state documents of all kinds. This further eroded the responsibilities of the Office for Literary Composition. 34  See Feng Yen, Feng-shih wen-chien chi (edition of Chao Chen-hsin, Feng-shih wen-chien chi chiao-chu [Peking, 1958]) 3, p. 16, which gives a first appointment as a collator (chiao-shu) or corrector (cheng-tzu) as an important stage in an elite career track for a man of literary talent.

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

276

Twitchett

a “pure” court official.35 Both the secretaries and the assistant secretaries were classified as “pure” officials (ch’ing-kuan).36

The Historiographical Office

The Historiographical Office (Shih-kuan) is said to have been established as a separate bureau in the palace in 629.37 According to the T’ang liu-tien its duties were as follows: The historiographers38 [shih-kuan] are responsible for the compilation of the National History. They may not give false praise, or conceal evil, but must write a straight account of events. The portents of heaven, earth, sun and moon, the distribution of mountains and rivers, fiefs and cities, the precedence between junior and senior lines of descent, ritual and military affairs, changes of reward and punishment, between prosperity and decline, all should be first recorded. The historians should base themselves on the Court Diary [Ch’i-chü chu] and the Record of Administrative Affairs [Shih-cheng chi] to make a Veritable Record [Shih-lu], setting this

35  Po Chü-i, supplemented by K’ung Ch’uan, Po-K’ung liu tieh (Taipei, 1969) 74, pp. 176–89, quotes a story about the historian-scholar Hsü Ching-tsung, who, upon being appointed secretary of the Office of Literary Composition, said to his kinsmen, “If a scholar-official does not serve in the Office of Literary Composition, there is no way to perfect his family’s standing.” Where this anecdote came from is unclear; it was not in Po Chü-i’s original Poshih liu tieh shih lei chi (Tapei, 1969) 21, p. 506, and was thus inserted by K’ung Ch’uan in the mid twelfth century. 36  See Fonds Pelliol Chinois Touen-houang P. 2504 (as cited in n. 18), p. 89, section 19; p. 90, section 23. The secretaries held the lower fifth rank, upper division; the assistant secretaries the lower sixth rank, upper division. 37  For details of the Historiographical Office, see tlt 9, pp. 28a–30b; cts 43, pp. 1852–3; hts 47, p. 1214 (Rotours, Fonctionnaires, vol.1, pp. 199–204). See also thy 63, pp. 1089–1104; thy 64, pp. 1105–14; tt 21, pp. 126c–127a. 38  The term shih-kuan is usually used as a general term for all officials engaged in the writing of history, irrespective of their rank and of the official bureau to which they were affiliated. However, it is also sometimes used in a more specialized sense meaning The Historiographer, that is to say the director of the National History, elsewhere known as the Chien-hsiu Kuo shih or Chien Kuo shih. See, for example, cts 43, p. 1852; and thy 63, p. 1092, where Chang-sun Wu-chi, the senior chief minister, is entitled Shih-kuan T’ai-wei in connection with the compilation of the National History and the Veritable Records for T’ai-tsung.

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

The Bureaucratic Apparatus [ of T ’ ang Historians ]

277

out in chronological form and incorporating the principles of praise and blame. When this is completed it is to be stored in the official storehouse.39 Neither the T’ang liu-tien nor any other source gives a complement of regularly established posts in the office; they list only its staff of clerical assistants.40 The official historiographers, the directors of the National History (Chien-hsiu Kuo shih) and the compilers (Hsiu-chuan), were in fact always appointed while concurrently holding other offices; exceptionally talented younger men could also be seconded from their substantive posts to work as assistants in the department (chih-kuan). During the early years of the reign of Hsüan-tsung (713– 36) those compilers who were not already “officers in attendance” (kung-feng) in their substantive office were required to attend court daily, together with the court diarists. This practice ended late in 736 when Li Lin-fu became chief minister.41 The senior historian sometimes held the title “in charge of the affairs of the historian officials” (chih shih-kuan shih).42 In 809, following a memorial from Chief Minister P’ei Chi, there was a systematization of the status of the members of the Historiographical Office. All officials who in their substantive office were entitled to attend court were to be appointed as compilers, while those who did not attend court were to be made assistants. The compiler with the highest-ranking substantive office was put “in charge of the affairs of the office” (p’an kuan-shih) as a sort of project manager.43 39   t lt 9, p. 296. 40  The exception is hts 47, p. 1214, which lists four compilers (hsiu-chuan). This figure, however, almost certainly refers, like much else in the monograph on officials in hts, to the situation at the end of the T’ang period. cts 43, pp. 1852–3, merely lists the titles of the Shih-kuan (which here, from the commentary, refers to the director of the National History), the compilers (hsiu-chuan), and the assistants (chih-kuan). It says that compilers were first appointed during the T’ien-pao period (742–56) and were concurrent holders of offices in other bureaus who worked in the Historiographical Office, whereas those first entering the office were given the title of assistant. 41  See hts 47, p. 1208. 42  For example, see cts 102, p. 3184, which tells us that Wei Shu held this title from 730–9. 43  See thy 63, p. 1101, which dates the reform in the sixth month of 811; cts 43, p. 1853, which confirms the year as 811; hts 169, p. 5150, which is undated; and cts 148, p. 3990, which dates it in the autumn of 809. The date given in thy 63 cannot be correct since P’ei Chi died on the thirteenth day of the fifth month of 811 (see cts 14, p. 435), having been seriously ill since the end of 810 (see cts 148, p. 3990). P’ei Chi’s reform also attempted to regularize the titles of scholars in the Chi-hsien Academy. According to hts 169, this reform was incorporated into the Statutes (ling).

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

278

Twitchett

Later in the eighth century there came to be a regular establishment of four compilers of the National History, but as late as 832 court opinion objected to the appointment of as many as four compilers, favoring only two or at most three compilers employed at any one time.44 In 852, following a memorial from Cheng Lang, the number of compilers was permanently increased to four, each in theory responsible for one of the quarters of each year, while the position of assistant was abolished.45 The compilers’ and assistants’ concurrent offices give a clear indication of their place in the bureaucratic structure.46 All these offices were prestigious “pure offices” reserved for the scholarly elite within the civil service, men of proven intellectual talent and impeccable family. The established offices and ranks of the compilers before Hsüan-tsung’s time tended to be higher than in the second half of the dynasty, usually in the range from the seventh to the fourth rank, though a few, mostly under Hsüan-tsung, held basic posts of the third rank. After the accession of Su-tsung their basic offices were slightly lower, from the eighth to the fourth rank, and in the ninth century none held higher than the fifth basic rank. However, the type of basic appointment, “pure offices” in the elite stream of central offices, remained unchanged. The assistants, too, formed part of the same charmed circle. Their rank was very low, mostly in the ninth rank. But their basic posts, normally marshals (wei) of counties within the metropolitan administration, or posts of collator (chiaoshu lang) in the Imperial Library, were the most desirable of first appointments for a promising young man. The compilers worked under the general supervision of one or more of the chief ministers who were designated director of the National History. There was normally a single director in charge of the National History. But there were exceptions to this rule. The greatest number of concurrent directors was in the early 650s, when there were six or seven acting simultaneously,47 and in the first decade of the eighth century, when there were again several directors. Liu Chih-chi (661–721) sarcastically describes them as “nine shepherds for ten

44   See thy 63, p. 1102. 45  See thy 64, p. 1114; hts 47, p. 1214. 46  See the materials conveniently tabulated in Chang-Jung-fang, T’ang-tai ti Shih-kuan yü shih-kuan, pp. 253–69. 47  See the edict of 651, mentioning three names; ctw 11, pp. 22b–23b; wyyh 464, pp. 5b–7a. The edict of 653 presenting the subcommentaries to five Confucian canons mentions six, ctw 136, pp. 7b–9b, and another edict of 653 presenting the completed law code mentions seven, ctw 136, pp. 9b–12b.

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

The Bureaucratic Apparatus [ of T ’ ang Historians ]

279

sheep.”48 After Hsüan-tsung’s reign there was always a single director, except for the period from 788 to 800, when Te-tsung allowed the post to remain empty. In the confused days of the end of the dynasty, from 889 on, it became customary to appoint more than one director,49 even though by then the functioning of the Historiographical Office was largely suspended, not least because the imperial archives and libraries had been completely destroyed or dispersed.50 According to Sung Min-ch’iu (1019–79), the post of director was one of a number of concurrent responsibilities distributed among the chief ministers. He claimed that the T’ang tradition was for the senior of the four chief ministers to be appointed concurrently as commissioner for the T’ai-ch’ing Palace, the Temple of Lao-tzu in Ch’ang-an: his colleagues were appointed, in order of precedence, as chief scholar (ta hsüeh-shih) of the Hung-wen Academy, director of the National History, and chief scholar of the Chi-hsien tien Academy.51 The reality was less orderly than this ideal pattern.52 It is unclear what period Sung Min-ch’iu refers to; presumably it is to very late T’ang, but I cannot identify any group of chief ministers to whom it applies, and there was never a set number of four chief ministers. The director was, nonetheless, usually one of the junior chief ministers. He was also, like the historians under his charge, normally a member of the scholarly elite within the bureaucracy. After the An Lu-shan rebellion the great majority were examination graduates: by the ninth century almost every director had passed the chin-shih examination. Moreover almost half of them had, 48  See Shih t’ung 20, p. 591. 49  For details, see Chang Jung-fang, T’ang-tai ti Shih-kuan yü shih-kuan, pp. 47–58, and table, pp. 270–80. 50  See a memorial by Lo Kun, probably written either in 891 or 898, which suggests the establishment at the capital of a special agency, funded by the Palace Treasury, to buy books for the government and to make up for the destruction of the Imperial Library and the collections of the Three Academies. ctw 828, pp. 6a–7a; wyyh 694, pp. 15a–16a. 51  See Sung Min-ch’iu, Ch’un-ming t’ui-ch’ao lu (Shanghai, Ts’ung-shu chi-cheng edition, 1936) A, pp. 10–11. 52  See hts 46, p. 1183. Wang Ying-lin (1223–96) also says that the Hung-wen Academy (which he mentions under the alternative name Hsiu-wen Academy), the Historiographical Office, and the Chi-hsien Academy were grouped together as the Three Academies (sankuan), and that the system under which each was headed by a chief minister was established in 726. See Yü hai 165, p. 11a–b. The use of the term san-kuan in this unusual sense (it normally referred to the three principal schools in the Imperial University system) is confirmed by an address from Lo Kun to the “Chief Minister Director of the National History” dating from the 890s in which it is implied that these were the three major offices involved in the recording of history. See wyyh 653, p. 7a–b; ctw 828, pp. 7b–8b.

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

280

Twitchett

earlier in their careers, either served in the Historiographical Office or held office as diarists. They thus had professional experience of historical compilation.53 The post of director was usually, but not invariably, a purely nominal “political” appointment. All the actual writing was done by the compilers. However, the directors frequently interfered with their work, and the appointment of chief ministers to direct the office underlines the fact that the compilation of the historical record was thought of as one of the ongoing functions of government. Official historiography was as much a political as a scholarly activity. The historians were not simply recorders of events. Like the court diarists, they were held to be the custodians of precedents and traditions. For example, in 717 Hsüan-tsung ordered the historians, in the months preceding the great seasonal sacrifices, to memorialize in detail about the ceremonies that ought to be carried out.54 Although the Historiographical Office had a somewhat anomalous and quasi-independent status, standing outside the regular bureaucratic hierarchy, it was attached for administrative purposes to one of the central ministries.55 53  See the information tabulated in Chang Jung-fang, T’ang-iai ti Shih-kuan yü shih-kuan pp. 270–80. The increasing proportion of chin-shih among the chief ministers acting as directors during the late T’ang may not be the result of deliberate policy but may simply reflect the steadily growing preponderance of men with a strong scholarly background among the highest court officials. The large number with historical experience does suggest conscious choice, however. 54   t hy 64, pp. 1107–8. 55  See tlt 9, p. 28a; cts 43, p. 1852; hts 47, p. 1214, all of which show it as a subordinate bureau under the Chancellery. Hung (“The T’ang Bureau of Historiography before 708,” p. 100) has questioned whether, at least in the earlier part of the dynasty, this was the case. However, his arguments are not conclusive, being largely based on the absence of confirmatory evidence, and in any case apply only to the seventh century. There is no doubt that the origins and early activities of the Historiographical Office remain obscure. However, by 663 the office is definitely mentioned by name. See thy 63, p. 1089. Hung’s objection that in the tenth month of 670 (8 January 671 in the western calendar) an edict uses the term Shih-ssu rather than Shih-kuan (see thy 63, p. 1100, some texts of which write not Shih-ssu but suo-ssu, the “office responsible”; and ttclc 81, p. 467) has little force, because the edict dates from a period of Kao-tsung’s reign during which the names of all offices were changed to archaic forms, many of them ending in -ssu. The term Shihssu was thus probably (although this is unconfirmed by other evidence) the name used for the Historiographical Office from 662–70. The old names for offices were restored only on the twentieth day of the twelfth month of 670 (February 671 in the western calendar), some weeks after the edict in question had been promulgated. By the reign of Hsüantsung the office was certainly subordinated to the Chancellery, just as were the Chi-hsien

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

The Bureaucratic Apparatus [ of T ’ ang Historians ]



281

Public business area of the Ta-ming Palace

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

282

Twitchett

At first it was placed under the Chancellery, and its premises were next door to or to the north of the main56 Chancellery building in the palace. When, in 663, the Ta-ming Palace was rebuilt and became the permanent residence of the emperor and the site of the court assemblies, the premises of the Historiographical Office were placed south of the new Chancellery building.57 There was also a branch Historiographical Office in the eastern capital, Loyang, where the court spent much of the late seventh and early eighth centuries. In the Loyang palace the historiographical office was sited next to the Secretariat building.58 Later, as the Secretariat gradually replaced the Chancellery as the most important of the central ministries, the location of the Historiographical Office next to the Chancellery building in the Ta-ming Palace in Ch’ang-an became both inconvenient and inappropriate. In 73759 Li Lin-fu, who was then chief minister and concurrently director of the National History, suggested that because the Secretariat was now the most crucial and important organ of government it would be appropriate if the Historiographical Office was moved. In response to a memorial from one of the historiographers, Yin Yin,60 the office was moved into a building near the Secretariat that had formerly

Academy and the office responsible for the Urns for Communication with the Throne (Kuei-shih yüan). None of these bodies had a regular establishment of ranked posts, and their personnel, like the historians, held their substantive offices elsewhere. 56  The Chancellery and Secretariat had their main buildings inside the Palace City (Kungch’eng) close to the main audience halls. See Hiraoka Takeo, Chōan to Rakuyō (Kyoto, 1956), maps 19, 20. They also maintained Outer Offices (wai-sheng) in the outer Administrative City (Huang-ch’eng) on either side of the main street leading to the great gateway to the palace. See ibid., maps 17, 18. 57  See thy 63, p. 1089. 58  See Shih t’ung 11, p. 318. Hung suggests that this is evidence that at this time (circa 703) the Historiographical Office was not subordinate to either ministry but placed close to one or other of them for convenience. The two ministries were responsible for receiving all memorials and other documents addressed to the throne and for drafting and promulgating edicts. They were thus essential sources of documentation for the historiographers. Liu Chih-chi says that the premises of the office in Lo-yang were much superior to those in Ch’ang-an, and that the working conditions too were better there. 59  See tt 21, p. 127a; cts 43, p. 1852. This date is wrongly given as 727 in thy 83, p. 1089, and as 732 in hts 47, p. 1214. Both dates are impossible, because Li Lin-fu became chief minister only in late 736. 60  See n. 22 in this chapter.

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

The Bureaucratic Apparatus [ of T ’ ang Historians ]

283

housed the Imperial Pharmacy.61 Responsibility for the supervision of the Historiographical Office was also formally transferred to the Secretariat.62 Either the office was subsequently moved yet again or, in addition to its business office in the Ta-ming Palace and its branch office in Lo-yang, it also kept at least some of its records in another building, for when the Historiographical Office and its records were burned during the An Lu-shan rebellion, it is said to have been in the little-used secondary Hsing-ch’ing Palace in the east of Ch’ang-an.63

The Palace Department of the Imperial Library

When at the beginning of T’ai-tsung’s reign the new Historiographical Office was set up under the control of the Chancellery, it did not immediately take over all the responsibilities for historical writing formerly exercised by the Office of Literary Composition. A separate organization, the Palace Department of the Imperial Library (Pi-shu nei-sheng) is also said to have been set up in the palace, at the Secretariat.64 This office, about which we have little information, 61  See tt 21, p. 127a; thy 63, p. 1089; cts 43, p. 1852. On the office’s new location, see Hsü Sung (1781–1848), T’ang liang ching ch’eng-fang k’ao (edition of Lien-yün-i ts’ung-shu, 1848, reprinted in Hiraoka Takeo, Chōan to Rakuyō) 5, p. 4a, which says that their building was to the west of the Secretariat building and north of the Imperial Pharmacy. For the position of the Historiographical Office in Lo-yang, where it was also to the west of the Secretariat, see the anonymous Ho-nan chih (Ou-hsiang ling-shih, 1908 ed., reprinted in Hiraoka Takeo, Chōan to Rakuyō) 4, p. 2b. Hsü Sung’s plan of the Ta-ming Palace (Chōan to Rakuyō, map 27, 1) shows the Historiographical Office to the west of the Chancellery, as does the modern reconstruction of the plan of the palace by Sekino Tadashi (1867–1935) (ibid., plan 29). 62  See thy 63, p. 1089; hts 47, p. 1214. At this same time the compilers ceased to attend court together with the diarists, although they might do so by reason of their substantive office. See hts 47, p. 1208. 63  See thy 63, p. 1094. 64  See thy 63, p. 1091; tfyk 554, p. 4b; tfyk 556, p. 11a–b; Yü hai 121, pp. 31b–32a. The text says that “The Palace Department of the Imperial Library was set up at the Secretariat [yü Chung-shu].” Whether this means it was located in the premises of the Secretariat, or that it was placed under its administrative control is unclear. Because the Imperial Library was an important organization quite independent of the Secretariat, the latter seems unlikely. According to Sung Min-ch’iu, Ch’ang-an chih (Ching-hsün t’ang ts’ung-shu, 1784 ed., reprinted [chapters 6–10 only] in Hiraoka Takeo, Chōan to Rakuyō) 6, p. 2a, it was set up in 629 and abolished “shortly afterward.” Hung, “The T’ang Bureau of Historiography

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

284

Twitchett

seems to have been specially assigned responsibility for the compilation of the standard histories of the earlier dynasties, work on which had begun some years earlier65 and which were a major preoccupation of T’ai-tsung. They before 708,” pp. 96–8, has already pointed out that there is very considerable confusion here. First, the Imperial Library had already had a “Palace Department” (nei-sheng) since Sui times. See ss 58, p. 1413; hts 98, p. 3887. Like the Secretariat and Chancellery and other ministries, the Imperial Library had offices both inside the palace compound, and in the Administrative City. The latter was called its “Outer Department” (wai-sheng), and its premises are shown on the fragmentary map of Ch’ang-an by Lü Ta-fang engraved on stone in Sung times and dated 1080 that survives in the Pei-lin in Sian (see Hiraoka Takeo, Chōan to Rakuyō, map 2). The “Outer Department” had a book depository shu-ko, which was to the east of the business office of the director (chien-yüan). See Ch’ang-an chih 7, p. 4a. It also housed an elementary school for the children of the imperial family and very high ranking officials. See thy 35, p. 633. The “Outer Department” is mentioned as early as 618, and this would imply that the Palace Department also existed at the beginning of the T’ang. Certainly the Palace Department existed in 628 or 629, when Ching Po was sent there to work on the Sui history with Yen Shih-ku and K’ung Ying-ta. See cts 189A, p. 4954; hts 198, p. 5056. Second, there is some doubt whether in fact the work of the Palace Department on the histories of the preceding dynasties was in fact clearly separated from that of the newly founded Historiographical Office. It was not until some years later that the work on the T’ang Veritable Records was begun. Before that, what did the Historiographical Office do, if not work on the earlier periods? Reworking the Court Diaries into polished form would in itself hardly have justified a whole new organization and the Office for Literary Composition was already doing this. Third, there is the question why the work was to be done at the Secretariat building. The Imperial Library certainly had its own premises within the palace, although the site has never been identified. These buildings must have been rather large, simply to accommodate in scroll form the very large collection we know it to have held. Does the term “Imperial Library” here perhaps refer to the Office for Literary Composition, which was its main subsection responsible for historical writing? Officials from the Imperial Library, and also from the Heir Apparent’s Library, had been engaged on the writing of the histories of earlier dynasties since 622. Some historians were already working at the Office for Literary Composition before 629. It seems that the innovation initiated in 629 was that a new coordinated program for the compilation of the earlier histories was put into motion, with the Imperial Library in charge, and that the work was perhaps accommodated in the building of the Secretariat, where some of the principal compilers held their substantive offices. 65  See cts 73, p. 2597; thy 63, p. 1090. In the eleventh month of 621 Ling-hu Te-fen, then a court diarist, made an informal address to Kao-tsu reminding him that Standard Histories of the recent dynasties were lacking, and pointing out that although documents were available from the Liang and Ch’en dynasties in the South, and from the Northern Ch’i, there were many gaps in the records of the Northern Chou and the Sui. He urged that work be begun on these histories while people with direct experience of the period were still alive. After another decade the traces of these events would be irretrievably lost. A

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

The Bureaucratic Apparatus [ of T ’ ang Historians ]

285

first undertook the compilation of the histories of the “Five Dynasties” (the Northern Chou, Sui, Liang, Ch’en, and Ch’i), which were completed in 636.66 Later, in 646, they were ordered to produce a new history of the Chin.67 By the beginning of Kao-tsung’s reign, however, responsibility for the compilation of the histories of earlier dynasties seems to have been transferred to the Historiographical Office.68 In 656 the historiographers presented the newly year later, on the twenty-sixth day of the twelfth month of 622, an edict appointed a team of historians to compile histories of the Wei, Northern Chou, Sui, Liang, Ch’i, and Ch’en dynasties. See ttclc 81, pp. 466–7; cts 73, pp. 2597–8. But although the work continued for several years, these histories remained uncompleted and work on them was suspended. See thy 63, pp. 1090–1. 66  See cts 3, p. 45; cts 73, p. 2598; thy 63, p. 1091. See also Wu Ching, Chen-kuan cheng-yao (Harada Taneshige’s variorum edition, Jōgan seiyō teihon [Tokyo, 1962]) 7, p. 218. This section is missing in most current editions of Chen-kuan cheng-yao, for example in the edition published by the Shanghai Ku-chi ch’u-pan she in 1978.  The histories commissioned under T’ai-tsung and now completed did not include a new history for the Northern Wei, although one had been begun in 622. This was excluded because there were already two histories of the period that were considered adequate, one the surviving Wei shu by Wei Shou (listed under the title Hou Wei shu in 130 chapters by the bibliographical chapter of cts 46, p. 1990), and the other by Wei T’an (this probably refers to his Wei Shu in 107 chapters, listed in cts 46, p. 1990, but may refer to his Wei chi in twelve chapters, listed in cts 46, p. 1991; both works are lost). It was therefore thought unnecessary to compile the Wei history afresh. Not everybody agreed; Ling-hu Te-fen subsequently produced a Hou Wei shu of his own in 50 chapters. Yet another Wei shu in 100 chapters was written by Chang Ta-su, who was a compiler in the Historiographical Office during the 650s, and presumably wrote his book as a result of his dissatisfaction with the existing histories. See cts 68, p. 2507. Both works were in the Imperial Library in 721. See cts 46, p. 1990. Small fragments of each survived into Sung times: Wang Yaoch’en, Ch’ung-wen tsung-mu (Ssu-k’u ch’üan-shu chen-pien, pieh-chi [Taipei, 1975]) 3 lists a single chapter of the annals of Wei T’an’s history, and two chapters of the Monograph on Astronomy from that of Chang Ta-su as having been in the imperial collections in 1042, but lost by 1144. 67  See thy 63, p. 1091; tfyk 554, pp. 16a–b; tfyk 556, pp. 12a–13b; ttclc 81, p. 467; cts 73, p. 2598. This work was entrusted by Fang Hsüan-ling to a team of eighteen compilers led by Ling-hu Te-fen, who had been disgraced and removed from office shortly before, owing to his close involvement with the former heir apparent Li Ch’eng-ch’ien who had been degraded in 643. Ling-hu Te-fen had been the president of the heir apparent’s Secretariat of the Right, a very important office in his administration. The work on the Chin shu seems to have been begun in 644 and completed in 646. The edict ordering its compilation is dated 646 in ttclc 81, however, so the dates given elsewhere may possibly be incorrect. 68  It is also quite possible that the Palace Department of the Imperial Library and the Historiographical Office were not in fact completely separate entities at this time. Li Hua (see wyyh 799, p. 4a) says that the Historiographical Office was set up early in the reign

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

286

Twitchett

completed monographs on the Five Dynasties (Liang Sui Ch’i Chou Ch’en Wutai chih) to the throne.69 There is no further mention of the Palace Department of the Imperial Library as a separate office. In T’ai-tsung’s reign, moreover, members of the staff of the Imperial Library and of its subordinate bureaus, the Office for Literary Composition and the Imperial Observatory, the offices that traditionally had been responsible for the compilation of historical records, are still to be found not only on the editorial teams appointed for each of the histories of the earlier dynasties, but also among the compilers of the first official histories of the T’ang, such as the Veritable Records for the reign of Kao-tsu and for the first years of T’aitsung, completed in 643;70 the Veritable Records for the last years of T’ai-tsung, completed in 650;71 those for the first years of Kao-tsung, completed in 659;72 and the first National History, completed in 656.73 Although this arrangement may have been due in part to administrative convenience, in that each of these departments held some of the required documentation, the old involvement in historical compilation of various subordinate sections of the Imperial Library apparently continued for many years after the establishment of the Historiographical Office.

The Diarists of the Heir Apparent

Under the T’ang, the heir apparent presided over his own administration in his official residence, the Eastern Palace, and held his own court assemblies at which he received petitions and issued orders. This administration, which was a sort of microcosm of the central government, was designed primarily to train and accustom the future emperor to the routines and procedures of his future role. His orders and acts, however, had real consequences. Among the officials of his Secretariat of the Left (Tso Ch’un-fang) were four grand secretaries (Ssu-i of T’ai-tsung to undertake the histories of the five former dynasties, and that many of the historians appointed to it were secretaries from the Office of Literary Composition. 69  See cts 4, p. 75; tfyk 554, p. 4b. thy 63, p. 1092, mistakenly writes the title as the Liang Ch’en Ch’i Chou Sui Wu-tai shih, not chih, but undoubtedly refers to the same work in thirty chapters. This book was originally an independent work, but was subsequently incorporated into the Sui shu to form its monographs. 70  See thy 63, p. 1092; tfyk 556, p. 12a. For further details, see Chen-kuan cheng-yao 7, p. 219– 20 (in Harada Taneshige’s edition). 71  See tfyk 556, p. 14a. 72  See thy 63, p. 1093; tfyk 554, p. 16b; tfyk 556, p. 14a–b. 73  See thy 63, p. 1093; tfyk 556, pp. 14b–15a.

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

The Bureaucratic Apparatus [ of T ’ ang Historians ]

287

lang) who were expected both to act as the prince’s moral censors and also to keep a record of the prince’s performance of his ritual and administrative duties. This diary, the Tung-kung chi-chu, was forwarded to the Historiographical Office at the end of every year.74

The Involvement of Other Offices

The offices discussed thus far were primarily responsible for the compilation of historical records and the writing of history. As we have seen, however, the Historiographical Office was a somewhat nebulous organization. Its personnel held no established ranks and all were concurrently employed in other posts. Various other offices thus became involved in the historiographical process from time to time as a result of the overlapping responsibilities of those of their staff members who held concurrent posts as historians. Often this seems to have been done by design, and official historians were appointed to substantive posts that were either connected in some way with their historical duties or were relative sinecures, so as to free most of their time for their work on the histories. Some career historians retained their concurrent posts in the Historiographical Office for many years while they were promoted through a series of substantive offices and, even when transferred to posts with no specific responsibilities for historical work, continued to work on official historical projects, sometimes with official authorization, sometimes without. One example was Wu Ching (ca. 665–749), a long-term historiographer who continued to work on the National History even after he was disgraced and transferred to a series of provincial posts during Hsüan-tsung’s reign.75 Another was Ling-hu Huan. When in 789 he was exiled to Chi-chou in Kiangsi following the death of his patron Li Mi (722–89), he was allowed to take with him the Veritable Record of Tai-tsung, on which he had been working, so that he could complete it in his provincial post.76 Another later case was that of Shen Ch’uan-shih (769–827), 74  See tlt 26, p. 21a–b; tt 30, p. 173a; cts 44, p. 1907; hts 49A, p. 1293 (Robert des Rotours, Traité des fonctionnaires et traité de l’armée (Leiden, 1948), vol. 2, pp. 579–80); Yü hai 128, pp. 26b–27a. According to hts there were only two of these grand secretaries. The posts were first established in 631. In the early years of the dynasty, these secretary-diarists were sometimes employed concurrently in compiling official histories. The posts were considered very prestigious and exceptionally “pure” offices. When Ching Po was appointed to one of these posts in 643, one of the chief ministers, Ma Chou, expressed his regret that he himself was of too high a rank to serve in the same office. See cts 189A, p. 4954. 75  See cts 102, p. 3182; hts 132, p. 4529. 76  See cts 149, p. 4014.

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

288

Twitchett

who was permitted to take with him the unfinished Veritable Record of Hsientsung’s reign when he was posted as civil governor of Hu-nan in 823.77 Many of the official historiographers were also notable scholars who wrote extensively in other genres. Throughout the T’ang a great many of them not only held concurrent appointments in the Historiographical Office, but were at the same time also members of one or other of the academies in which the government maintained talented scholars whose services were at the disposal of the emperor. The academies were thus drawn into the process of historical compilation. In the early years of the dynasty the most important of the academies were the short-lived Wen-hsüeh kuan attached to the household of the future T’ai-tsung when he was Prince of Ch’in78 and the Hung-wen kuan. The Hungwen kuan, founded in 621 under the name Hsiu-wen kuan, and renamed in 626, was attached to the Chancellery.79 Like the Historiographical Office, and like the later academies, it had no fixed complement of scholars; its scholaracademicians (hsüeh-shih), auxiliary scholars (chih hsüeh-shih), and assistants (wen-hsüeh chih-kuan) had no fixed ranks and held their substantive offices elsewhere.80 After T’ai-tsung ascended the throne in 626 he built up a substantial library at the academy, where he maintained a large group of gifted scholars with whom he regularly discussed state affairs.81 A number of them were also engaged in the various historical projects of his reign. The Hung-wen kuan was a permanent institution. Other less closely organized groups of scholars were also recruited from time to time to act as imperial secretaries and to assist in drafting state papers. During the 660s Empress Wu recruited such a group, known as the “scholars of the Northern Gate” 77  See cts 149, p. 4037; hts 132, p. 4541. 78  See thy 64, p. 1117. 79  On the Hung-wen kuan, see tlt 8, pp. 41a–44b; cts 43, pp. 1847–8; hts 47, pp. 1209–10 (Rotours, Fonctionnaires, vol. 1 pp. 169–73); tt 21, p. 124c; thy 64, pp. 1114–5; also the wall record for the Chao-wen kuan by Ch’üan Te-yü, dated 807, Ch’üan Tsai-chih wen-chi (sptk edition, Shanghai, 1936) 31, pp. 4b–6a; wyyh 797, pp. 5a–6a. See also McMullen, State and Scholars in T’ang China, p. 15. Unlike the other later academies, the Hung-wen kuan was not only a center for government-sponsored scholarly and literary work. It was also a school, with lecturers who taught a small group of students selected from among the sons of high-ranking metropolitan officials, These students were presented for a modified version of the ming-ching examination. 80  The titles given to these scholars grew more and more confused as time went on. A final attempt to regularize them was made in 823. See thy 64, p. 1116. 81  See thy 64, p. 1114; thy 57, p. 977. See also Liu Po-chi, T’ang-tai cheng-chiao shih (Taipei, 1954), pp. 92–6.

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

The Bureaucratic Apparatus [ of T ’ ang Historians ]

289

(Pei-men hsüeh-shih), who acted as her confidential secretariat, and also compiled a long series of literary works under her name.82 During Chung-tsung’s reign the Hung-wen kuan fell into disrepute. On Juitsung’s accession in 710 many of its scholars who had been associated with the discredited former regime were dismissed or degraded, and it was reorganized.83 Its place as the chief center for important government-sponsored scholarship was taken by a new academy established by Emperor Hsüan-tsung in 718. This was at first called the Ch’ien-yüan yüan and was placed under the loose supervision of the Secretariat. It was renamed the Li-cheng yüan or Li chenghsiu shu yüan in 719, and in 725 became the Chi-hsien yüan, the “Academy of Assembled Worthies.”84 This new academy rapidly became the major center of government-sponsored literary and scholarly activities. It built up a very large library and a great collection of paintings. Its scholars included most of the prominent writers and scholars of the time, and under its auspices they 82  This group is first mentioned under this name in 675 (see tctc 202, p. 6376) but were almost certainly already in existence as early as 666 (see tfyk 550, p. 3a; thy 57, p. 977). According to tfyk they ceased to be employed as drafting officials for state papers from 682 onward. They are last mentioned under the name of the “Scholars of the Northern Gate” in 688 (see tctc 204, p. 6447). 83  See thy 64, p. 1115. 84  The best account of the Chi-hsien yüan, although it stops short at the An Lu-shan rebellion, is Ikeda On, “Sei Tō no Shūken’in,” Hokkaidō Daigaku Bungakubu kiyō 19, no. 2 (1971): 47–98. This assembles a wide range of source material, including many quotations from Wei Shu’s lost account of the academy, the Chi-hsien chu-chi, which are preserved in the late twelfth-century Chih-kuan fen-chi of Sun Feng-chi (1135–99), (reprint, Taipei, 1983) and in Yü hai. On Wei Shu’s Chi-hsien chu-chi, see hts 58, p. 1477; Ch’ung-wen tsungmu 3, p. 25a; Ch’ao Kung-wu, Chün-chai tu-shu chih (Yüan-pen) (spyk edition, Shanghai, 1936) 2B, p. 5a, (Ch’ü-pen) (Wan-wei pieh-tsang edition, Taipei, 1981) 7, p. 14a; Ch’en Chensun, Chih-chai shu-lu chieh-t’i (Ssu-k’u ch’üan-shu chen-pen edition, Pieh chi, vols. 150–1. Taipei, 1975) 6, pp. 4b–5a; Yü hai 48, p. 34a–b; whtk 202, p. 1687c. See also Chao Shihwei, ed., Chung-hsing Kuan-ko shu-mu chi-k’ao (Peking, 1932) 2, p. 20b. It is said to have been in three chapters by hts, by Ch’ung-wen tsung-mu, and by Chih-chai shu-lu chieht’i; in two chapters by Yü hai (citing Ch’ung-wen tsung-mu) and whtk, and in a single chapter by both recensions (the Yüan-chou and Ch’ü-chou versions) of Chün-chai tu-shu chih. The latter informs us that it was written in 756. Formal descriptions of the academy are to be found in tlt 9, pp. 22a–28a; cts 43, pp. 1851–2; hts 47, pp. 1212–13 (Rotours, Fonctionnaires, vol. 1, pp. 189–98); tt 21, p. 126c; thy 64, pp. 1118–21. The Chi-hsien yüan had its main premises in the Ta-ming Palace, immediately west of the Historiographical Office. It also had a building or buildings in the Hsing-ch’ing Palace in Ch’ang-an, and premises at the detached palace at the hot spring east of the capital, the Hua-ch’ing kung. There was also a branch of the academy in Lo-yang.

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

290

Twitchett

produced a flood of major works: bibliographies and catalogs, ritual compendiums, canonical commentaries, calendrical and legal works, literary collections, and encyclopedias. There was also much historical activity at the academy, a large part of whose library consisted of historical works.85 In 725, when it received the name Academy of Assembled Worthies, it was placed under the control of Chang Yüeh, who was then a chief minister and a compiler of the National History. He worked there on his “Veritable Record of the Reigning Monarch” (Chin-shang Shih-lu), in which work he was assisted by T’ang Ying, a young scholar of the academy.86 Wu Ching, who had been an active historiographer since Empress Wu’s reign, was also engaged as a scholar of the academy, where he worked both on the compilation of the official National History (Kuo-shih) and on his private chronological history of the T’ang entitled T’ang Ch’un-ch’iu.87 Wei Shu, a young man destined to become a major historian, was also appointed to the academy by his patron Chang Yüeh, and worked there on the National History and on various private historical projects.88 So much historical work was going on at the Academy of Assembled Worthies that for a while in the late 720s it seemed that the status of the Historiographical Office was threatened. In the sixth month of 727 Li Yüan-hung, who had replaced Chang Yüeh as chief minister, memorialized the throne complaining that the work of Chang Yüeh and Wu Ching on the National History was causing important documents to be transferred from the Historiographical Office to the academy, with a serious risk of loss. Orders were issued that the historians working at the academy were to go to the Historiographical Office to consult their sources.89 The academy remained important well into the ninth century, although its days of literary glory ended with the An Lu-shan rebellion, and members of the Historiographical Office continued to be concurrently engaged as scholars there.90 85  See thy 64, p. 1119. In 731 the collection comprised 89,000 chüan of which 26,820 chüan were historical works. The library contained many old books, but the greater part were newly copied specially for inclusion in its collection. The academy suffered severe losses during the An Lu-shan rebellion. In 763 its chief scholar memorialized the throne asking that a reward of one thousand cash per volume be offered to replace works that had been stolen. See tfyk 50, p. 12a. 86  See cts 97, pp. 3054–5; Wei Shu’s Chi-hsien chu-chi as cited by Yü hai 48, pp. 4b–5b. 87  See cts 102, p. 3182; hts 132, pp. 4529–31. 88  See cts 102, pp. 3183–5; hts 132, pp. 4529–31. 89  See thy 63, p. 1099. 90  The historian Chiang I, fur example, began his career in 781 or 782 as a clerical assistant (hsiao-li) in the academy at the recommendation of then Chief Minister Chang I. See tfyk 608, pp. 28a–b; hts 132, p. 4531. Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

The Bureaucratic Apparatus [ of T ’ ang Historians ]

291

Occasionally, too, special personal arrangments were made to supplement the offical record. During the first half of Hsüan-tsung’s reign, for example, his elder brother, Li Ch’eng-ch’i (679–741), the prince of Ning, who although banned from participation in court affairs had a profound understanding of politics, was given a special dispensation to prepare his own annual summary of events for presentation to the Historiographical Office. This was no formality but a serious work running to several hundred pages each year.91

Information Reported from Various Administrative Offices

It is often stated that the sources for the Veritable Records were almost exclusively the formal records of events and decisions at court incorporated in the Court Diaries, Ch’i-chü chu, and the Records of Administrative Affairs, Shihcheng chi. However, it is important to remember that the Historiographical Office had many other sources of information. About those of a personal, informal nature we can only surmise. But a wide variety of governmental agencies and ministries were legally required to collect and to make regular returns of specific types of information directly to the Historiographical Office for incorporation in the dynastic record. The rules on this subject are as follows:92 Fortunate omens: The Board of Rites [Li-pu] should draw up a list of these and report them every quarter. Heavenly portents: The Astronomer Royal [T’ai-shih]93 should report these together with his predictions and evidences of coming good fortune. Appearance at court of tribute-bearing missions from foreign countries: Whenever such a foreign mission arrives, the Court for Diplomatic Reception [Hung-lu ssu] should examine them on the natural conditions and customs of their country, on their dress, and the products brought as tribute, and on the distance and route by which they have come. These facts are to be reported together with the names of their leaders. Changes in the musical scales and newly composed airs: The Court of Imperial Sacrifices [T’ai-ch’ang ssu] should draw up an account of those responsible, and of the words and music, and report it. 91  See cts 95, p. 3012. 92   t hy 63, pp. 1089–90. 93   w thy 18, p. 293, gives the responsible office as the Imperial Observatory (Ssu-t’ien t’ai), which replaced the T’ai-shih chü in 757. Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

292

Twitchett

Establishment and suppression of Prefectures and Counties; the granting of marks of distinction to “filial and righteous families”:94 The Board of Finance [Hu-pu] should report such matters whenever they occur. Changes in the laws, and decisions of cases involving new principles: The Board of Justice [Hsing-pu] should report these as they occur. Good harvests, famines, flood, drought, locusts, hail, wind, and frost, together with earthquakes and the flooding of rivers: Whenever such things occur the Board of Finance [Hu-pu] and the local Prefecture and County should investigate the exact date, and report it together with an account of the relief measures taken to relieve the people. Enfiefment of various types:95 The Ssu-fu96 should investigate and report on such events. Succession to a fief by inheritance need not be reported. Appointment to office of senior officials in the various offices at the Capital together with Prefects, Governors-general, Protectors-general, Grand Commanders of Armies in the Field, and Deputy Commanders: All such appointments should be reported together with the text of the edict of appointment. In the case of civil officials the Board of Civil Office [Li-pu] should send in the report. In the case of military officers the Board of War [Ping-pu] should report. Details of unusual evidences of good administration by Prefects and County Magistrates: If there is anything particularly outstanding of this sort the Prefecture responsible should report it, entrusting their report to the commissioner responsible for local merit assessments [K’ao-shih]. Eminent scholars, men of unusual ability, great men, retired scholars, righteous husbands and chaste widows:97 Whenever there are exemplary persons of this sort in a Prefecture or County, without ques‑ tion whether or not they hold official rank, the authorities should 94   w thy 18, p. 294, in place of “filial and righteous families” has “filial sons, obedient grandsons, righteous husbands, and chaste wives.” 95   w thy 18, p. 294, specifies the “enfiefment and establishment of ancestral temples in the empire.” 96  The vague term Ssu-fu in the text of thy would normally refer to the various household administrations that were attached to each prince’s fief. In the context, however, it seems rather more likely that it derives from a graphic error for Ssu-feng, the Department of Noble Fiefs, a subdepartment of the Board of Civil Office that was directly responsible for the nobility, as in wthy 18, p. 294. 97   w thy 18, p. 294, has a special section for “men of great virtue, unusual ability, great men, retired scholars, hermits who have lived for many years in the mountains, and writers.”

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

The Bureaucratic Apparatus [ of T ’ ang Historians ]

293

investigate the truth of what they have heard, and each year record it and send this in, entrusting it to the commissioner responsible for the local merit assessments [K’ao-shih]. The deaths of senior officials employed in the various offices of the Capital: The office concerned shall ascertain the facts according to the individual’s curriculum vitae and send it in. The deaths of Prefects, Governors-general, Protectors-general, Deputy Commanders of Armies in the field and below: The Prefecture or Army concerned should investigate the deceased’s curriculum vitae [lichuang] and send it in, entrusting it to a convenient envoy. The fixing of Posthumous canonizations [Shih] of Princesses and Officials: The office of Merit Assessments [K’ao-kung] should record it, and send in the recommendation on the granting of a posthumous canonization [Shih-i] together with the Account of conduct [Hsingchuang] of the individual.98 Attendance of Princes at court: The Court of the Imperial Clan [Tsungcheng ssu] should investigate and report on this.99 All the above matters should be investigated and reported to the Historiographical Office as they occur by the responsible authority specified in the appropriate section. If the Historiographical Office discovers any matter which is appropriate for inclusion in the record, but not analogous with the matters specified above, it may send an official communication directly requesting information about it. Any place receiving such a communication should make enquiries in accordance with the schedule of requests, and make its report within a month.100 It is far from clear how seriously this system was enforced. There is ample evidence, for example, that the Historiographical Office actually did receive the Accounts of Conduct and the recommendations on posthumous canonization for deceased high officials. Population and tax figures reported by the Board of 98   w thy 18, p. 294, does not mention princesses, and also adds a note that the deceased’s family is permitted to submit an Account of Conduct. 99   w thy 18, p. 294, also makes the Court of the Imperial Clan liable for reporting the activities of members of the imperial clan employed in office, and for the ceremonies attending the weddings of imperial princesses. 100  See thy 63, p. 1090. Compare the slightly more detailed version of these arrangements given in a memorial from the Historiographical Office under the Later T’ang in 924 describing the “old precedent of our dynasty”; see wthy 18, pp. 293–4. The Later T’ang considered themselves a continuation of the T’ang, and the “old precedents” were those in force during the T’ang.

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

294

Twitchett

Finance were certainly sent in and included in the record. The responsibility of the astronomer royal for calendrical information was also taken seriously and was specifically mentioned among the duties of his post in the T’ang liu-tien.101 After the An Lu-shan rebellion, this system for reporting information fell into disorder and disuse. In 779–80 the historians requested that it be revived.102 There is, however, no means of checking whether this was actually done. The system certainly fell into disuse during the latter decades of the T’ang and during the political confusion of the Liang dynasty (907–23). In 924, under the new Later T’ang dynasty (923–36), the revived Historiographical Office memorialized the throne, asking that all official documents and reports be sent in to the office as had been done under the T’ang.103 Thus there existed throughout the T’ang dynasty a group of court diarists and other officials responsible for the collection of information and the keeping of a court record. There were also official historiographers with the ongoing responsibility for transforming this material into a polished history of the dynasty, and for writing the histories of earlier periods. But these arrangements remained fluid and changed repeatedly during the course of the dynasty. The organization for producing the historical record had by no means yet reached a final form such as was to be achieved under later dynasties.

101  See tlt 10, p. 26a. 102  See thy 63, p. 1090. 103  See wthy 18, pp. 293–4. On this attempt to revive the Historiographical Office, see Wang Gungwu, “The Chiu Wu-tai Shih and History writing during the Five Dynasties,” Asia Major, n.s., 6, no.1 (1957): 1–22; Chin Yü-fu, “T’ang-Sung shih-tai she kuan hsiu-shih chih-tu k’ao,” Kuo-shih-kuan kuan-k’an 1, no. 2 (1948): 6–18.

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

Bureaucracy and Cosmology: The Ritual Code of T’ang China David McMullen Introduction* The historical and literary sources of the T’ang period (ad 618–907) in China convey the impression that Confucian rituals were central to the life of the state. The official histories describe the emperor and members of the imperial clan performing a wide range of rites. Elaborate, specially drafted and costly rituals, some involving long journeys, marked important political developments. Rituals followed crises in dynastic rule as well as periods of stability and achievement. There was also an extensive programme of regular, annually conducted rites. Most of these involved not only the emperor and imperial clan members, but the chief ministers, general service officials and official scholars. They were often attended by onlookers from the general population, so that large numbers of people were involved. The grandest of the imperial rites had a climactic quality. One or two of the most successful performances were considered the greatest achievements of the reigns in which they took place. They strained the wealth, logistical capacities and even security of the T’ang state to near its limit.

Source: “Bureaucracy and Cosmology: The Ritual Code of T’ang China,” in David Cannadine and Simon Price (eds.), Rituals of Royalty: Power and Ceremonial in Traditional Societies, Cambridge: Cambridge University Press, 1987, 181–236. © Cambridge University Press, reproduced with permission. * I am grateful to the editors for very helpful remarks on two preliminary drafts of this article, and to my colleague Dr Anne Birrell for a series of useful observations on the first draft. In the following annotation, selected references only are supplied.  Since the completion of this article a major study of T’ang ritual has been published in the late Howard J. Wechsler’s Offerings of Jade and Silk: ritual and symbol in the legitimation of the T’ang dynasty (New Haven, 1985). Professor Wechsler’s main focus is on the role of imperial ritual in the first two reigns of the dynasty, but he provides much invaluable detail in accounts of most of the principal rituals mentioned in my article. He also analyses pre-T’ang theories of ritual and describes some of the main ritual controversies in which T’ang scholars were engaged.

© koninklijke brill nv, leiden, ���9 | doi:��.��63/9789004380158_011

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

296

McMullen

Official scholars, on whose expertise the emperor was entirely dependent for the drafting, performance and documentation of the rites, emphasised the importance of ritual by making it a central element in their world view. They believed that the ritual tradition extended back to the golden age of remote antiquity and that the paragon emperor Shun (reigned traditionally 2255–2205 bc) had commissioned an official to organise and direct the kingly rites. They held that later rulers had established an imperial ritual programme and set up institutions to maintain it.1 For them the early Chou dynasty (1121– 220 bc) was the source of most ideals about ritual. The late Chou and Former Han (206 bc–ad 8) were also important, for this was the age in which the Confucian canons, the authority for much medieval thinking on the state, were identified and promoted. T’ang scholars located the sanction for the imperial ritual programme in the Confucian canonical corpus as a whole, but especially in the three rituals canons. The Record of ritual (Li chi), Rites of Chou (Chou li) and Observances and rites (I li) were a heterogeneous collection, both in origin and in content. The Record of ritual was a group of treatises dealing with the theory and function of ritual and containing detailed directives; the Rites of Chou consisted of idealised institutional blueprints; and the Observances and rites comprised mainly idealised ritual directives.2 Much of the material these three canons contained had little to do with the historical Confucius. But in the seventh and eighth centuries scholars accorded them enormous prestige, calling them ‘inerasable authorities’,3 and they kept them and later exegesis of them in the forefront of discussion of state rituals. Scholars also frequently cited later ritual usage, especially since the Later Han (ad 23–220), when the imperial ritual programme attained approximately its medieval form, as a sanction for contemporary practice. T’ang scholars also indicated how central they considered the Confucian state ritual programme to be by the elaborate theoretical justification that they advanced for it. Again, the learning involved here derived its authority from its 1  For early T’ang surveys of the pre-T’ang state ritual tradition, see Sui shu (History of the Sui) (Peking, 1973), ch. 6 pp. 105–7; Chin shu (History of the Chin) (Peking, 1974), ch. 19 pp. 579–80. cf. T’ung tien (Comprehensive compendium, hereafter tt, Taipei, 1959), ch. 41 pp. 233A–C; Chiu T’ang shu (Old T’ang history, hereafter cts, Peking, 1975), ch. 21 pp. 815–19. 2  James Legge, The texts of Confucianism, Parts iii–iv, The Li Ki (Oxford, 1885); Edouard Biot, Le Tcheou-li, 2 vols. (Paris, 1851); S. Couvreur, S. J. Cérémonial, texte, chinois et traduction, 2nd edition (Sienhsien, 1928). 3  c ts ch. 21 p. 818, cts ch. 8 p. 180, cts ch. 27 p. 1031; Morohashi Tetsuji, Dai Kan-Wa jiten (Great Sino-Japanese dictionary, Tokyo, 1955), entry 19.124.

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

The Ritual Code of T ’ ang China

297

antiquity. For ritual had very early become conceptualised, and linked with the condition of mind that correct observance of rituals brought about. Already in the late Chou, propriety, as the term li is often translated in this context, had become not just one of the five cardinal virtues but a major element in theories of society. Under the influence of the New Text school of cosmology and ethics in the second century bc, it was held to be a principle of the cosmos itself.4 Even in the vigorous and outgoing early T’ang period, in which attitudes tended to be practical rather than theoretical, official scholars endorsed, with significant variations of emphasis, this legacy of theoretical thinking on ritual, and used it to emphasise the importance of the imperial ritual programme. Finally the official view emphasised the importance of state ritual by the stress it gave to its institutional history. Scholars traced the posts and official titles that concerned the rites back to remote antiquity and linked the ritual offices inseparably to the rites themselves. Institutional continuity reaching, with few interruptions, back into the canonical age was to the T’ang world further proof of the authority and importance of the tradition. The great importance of Confucian sanctioned state ritual to the dynasty and the central place it occupied in the world view of scholars meant that official scholarship was particularly conscientious in documenting it. Thus the Old T’ang history devotes the first and longest of its eleven treatises, monographs reserved for different aspects of government, to imperial rituals. This treatise records the major changes that took place in the imperial ritual programme in the course of the dynasty.5 State rituals occupy a significant proportion of the Comprehensive compendium and the Gathering of essentials for the T’ang, two unofficial compendia that survive from the early ninth century and that closely follow official compilations in their perspective.6 Imperial rituals were also a major theme in the surviving works of individual scholars of all periods in the dynasty, whether prose or verse. More important for the present article, official scholars of the medieval period collected together their prescriptions for imperial rituals in the form of dynastic ritual codes, each of which was intended to be normative for the dynasty in which it was compiled. These successive ritual codes defined the 4  Fung Yu-lan, tr. Derk Bodde, A History of Chinese Philosophy, 2 vols. (Princeton, 1952–3) i, pp. 297–9; 337–57. ii, pp. 40–2; Ch’ü T’ung-tsu, Law and society in traditional China (Paris, 1961), pp. 230–41. 5  c ts ch. 21–2 p. 815–1038. cf. Hsin T’ang shu (New T’ang history, hereafter hts, Peking, 1975), ch. 11–20 pp. 307–457. 6  t t ch. 41–140 pp. 233–732; T’ang hui yao (Gathering of essentials for the T’ang, hereafter thy) (Peking, 1955), ch. 11–23, pp. 271–448; ch. 37–8, pp. 669–99 etc.

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

298

McMullen

imperial ritual programme systematically and to a remarkable degree of detail. In this respect, the sources for the Confucian state ritual of the eighth century are superior to those of earlier periods. For from this middle century of the T’ang there comes the first full state ritual code to have survived in the Chinese tradition. This code, which was completed in 732, gives a set of very exact directives for the entire Confucian sanctioned imperial ritual programme, including the rarely conducted rites as well as the annually recurrent ones. Explored fully and related to the other sources that contain information about state ritual, the code will ultimately provide an extremely detailed picture of how the imperial rites were performed and of their place in the T’ang state and in T’ang society.7 Enormous energy, then, on the part of emperors, chief ministers and general service officials, was expended on the state ritual programme, and enormous erudition was devoted to it by scholars who surveyed the pre-T’ang state itself. The centrality of the rites to the T’ang state is beyond contention. To those interested in assessing the varying role of royal rituals in different cultural contexts, their obvious importance to the T’ang raises a number of questions. These concern T’ang official understanding of the purpose of state ritual in relation to other purposes of government, or the official justification for the code. Behind official pronouncements lie the questions of the level of political and sectional interest in the state ritual programme, and the effect this had on the shape of the programme itself. How the official view of ritual related to unofficially or privately expressed assessments of it, and how these unofficial views changed in response to changing political realities are also aspects of the T’ang ritual tradition as a whole. In attempting to answer these questions, comparison with other states will be helpful, for it provides a spectrum on which to place the medieval Chinese tradition. At one extreme there is the case of nineteenth-century Bali, described recently in an anthropological context. Here the king’s role was seen above all in terms of dramaturgical or ritual performance. ‘Royal rituals enacted, in the form of pageant, the main themes of Balinese political thought’, and ‘statecraft [was] a thespian art’. Administration, on the other hand, was relatively unimportant and was relegated to the village level.8 The other extreme is provided by the nineteenth-century British constitution. Here the monarchy

7  Ta T’ang K’ai- yüan li (Ritual code of the K’ai-yüan period in the great T’ang dynasty, hereafter ttkyl), photolithographic reprint of edition of 1886 published with introduction by Ikeda On as Dai Tō Kaigen rei (Tokyo, 1972). 8  Clifford Geertz, Negara: the theater state in nineteenth century Bali (Princeton, 1980), pp. 120–2.

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

The Ritual Code of T ’ ang China

299

in its ceremonial role was strictly demarcated from political activity, and was held to be secondary, a dignified rather than efficient aspect of the state.9 At first sight, T’ang China seems indeed to be close to the Balinese case. The time, energy and expense and the intellectual priority accorded to state ritual in the period in which the code was produced was such as to suggest that T’ang imperial ceremonials, far from being secondary, might indeed constitute the most important function of the state. Yet a more detailed account of T’ang state ritual in the early and mid-eighth century leads inevitably to the conclusion that, though any parallel with nineteenth-century Britain seems inappropriate, comparison with Bali needs qualification on several counts. In the first place, the T’ang bureaucracy had many functions in addition to imperial rites. The very organisation of the central bureaucracy, in which ritual offices formed a proportion, perhaps one sixth only, of the total bureaucratic structure, indicates its commitment to a range of administrative operations. Official scholars accorded activities of government other than ritual a comparable importance and allowed them a comparable history and antiquity. The central government included administrative bodies concerned with education; the enforcement of criminal law; defence; taxation; and the selection and promotion of officials. These activities were the responsibility of the same officials who administered the rituals and ultimately, like the rituals, of the emperor himself. Far from being relegated to a subordinate level, these operations also commanded prestige and were not, in practical political terms, so different in status from the regular ritual programme.10 Secondly, the Confucian imperial state ritual programme as it was formulated in the code did not represent the totality of the emperor’s religious or ritual commitments. The wealth of rituals the code embodied were almost without exception sanctioned by the traditional Confucian outlook on the cosmos and on society. Besides this Confucian outlook however, the transcendental faiths of Buddhism and Taoism also exerted claims on the emperor’s commitment, time and resources. The mid-eighth-century world did not consider these three belief systems mutually exclusive. For Buddhism and Taoism differed from the 9  S. R. F. Price, Rituals and power: the Roman imperial cult in Asia Minor (Cambridge, 1984), pp. 239–40. Walter Bagehot, The English Constitution, ed. with introduction by R. H. S. Crossman M. P. (London, 1963), pp. 61–5. 10  Robert des Rotours, Le traité des examens, traduit de la Nouvelle histoire des T’ang (Paris, 1932), pp. 3–25, summarises the structure of the T’ang government. The same scholar’s Traité des fonctionnaires et traité de l’armée, traduits de la nouvelle histoire des T’ang, 2 vols. (Leiden, 1947), provides a detailed and well-indexed translation and annotation of the New T’ang history monograph on the ‘offices and posts’ of the dynasty.

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

300

McMullen

Confucian outlook to such an extent that in practice they affected an entirely different area of experience. The Confucian tradition was above all concerned with the political and social conduct of the human hierarchy, supplying for it a comprehensive system of ethical norms. As a religious system, however, it retained some features that an important scheme for religious evolution has termed ‘archaic’.11 It had no developed sense of the transcendent; it focused on this life rather than life after death. It involved a plurality of cosmic agents and divinities whose chief function was to affect conditions in the natural and human worlds. It posited the emperor as the main link between the people and the cosmic order, and accorded the nobility and the official classes, who also monopolised political power, a superior standing in rituals. There was no Confucian process of ordination and consequently no Confucian priesthood. Rather, rituals were performed by the emperor himself and the imperial clan, or by officials representing them, and were drafted and administered by a highly structured bureaucracy that had its range of other duties to discharge. Buddhism and Taoism on the other hand belonged to a later stage in religious evolution, one that has been called ‘historic’. They promoted belief in an entirely different realm of reality. They both had priestly hierarchies that existed almost completely outside the official Confucian sanctioned bureaucratic establishment. They advocated monasticism, and by the mid-eighth century, Buddhist monasteries in particular had come to be substantial holders of land and wealth. Their social outlook was, especially by the mid-eighth century, more universal than that of Confucianism and they addressed the individual and his own search for answers to questions of ultimate value. Though their clergy might be politically active, they did not, to the extent that Confucians did, provide values and traditions for the state and its administration. The Confucian oriented official scholars who compiled the code of 732 may have been adherents of the transcendental faiths in private. But in their official capacities they regarded both Buddhism and Taoism as outside their concern. In documenting and codifying Confucian sanctioned imperial ritual, they demarcated it strictly from the Taoist and Buddhist activities of successive T’ang sovereigns. But these faiths, their clergy and their wealthy monastic institutions remained powerfully present in the mid-eighth-century T’ang state. Despite the lofty claims that official Confucian scholars made for the Confucian world view and for Confucian imperial ritual, the Confucian ­programme

11  Robert N. Bellah, ‘Religious evolution’, American Sociological Review xxix, 1964, pp. 358–74, and Beyond Belief: essays on religion in a post-traditional world (New York, 1970), pp. 20–50.

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

The Ritual Code of T ’ ang China

301

was forced to compete with the very different demands and outlays of the transcendental faiths. In the following pages, it will be argued that Confucian sanctioned imperial ritual constituted one of the principal, but by no means the only function of the T’ang dynastic state. The ritual programme represents a unique combination of an ‘archaic’ belief system administered by a learned and highly structured bureaucracy. A more detailed account of the imperial ritual programme as it was embodied in the K’ai-yüan ritual code (pp. 304–316) will be followed by a brief description of the institutions within the bureaucracy that administered the code and drafted the rituals (pp. 317–25). It will be emphasised that these institutions were integrated into the bureaucracy as a whole, and that the ritual tradition was accessible to the entire official body. Then an account of the official justification for the ritual programme will show how the officially accepted theory of ritual contained two distinct, though not mutually exclusive, emphases, the cosmic on the one hand and the social and moral on the other (pp. 325–32). In medieval China the activities and even the affirmations of support that surround a traditional operation of the state cannot always be taken simply at face value, for they very often disguise other interests. The political picture behind the state ritual programme shows scholars involved in harsh competition, in sectional conflict, factionalism and individual ambition (pp. 332–40). The very vigour and variety of this ulterior interest in the rites is essential to a full understanding of the code and its importance in the T’ang world. Finally, it will be suggested that the unique character of Confucian state ritual, its retention of an ‘archaic’ belief system combined with its sophisticated bureaucratic framework, gave rise, as towards the end of the eighth century the political power of the T’ang declined, to a shift in the attitudes of some intellectuals towards the function of the rites (pp. 340–45). This change took the form of a stronger unofficial emphasis on the social and moral function of rituals, and an occasionally striking scepticism towards their traditional and publicly accepted cosmological underpinnings. An apology for royal rituals that sees their function mainly in social terms anticipates a common modern defence of royal ceremony. That scholars should have arrived at a position of this kind nearly twelve hundred years ago in China points to the vitality and sophistication of the community in which they lived. The reign of the emperor Hsüan tsung of the T’ang (712–756), in the middle decades of which the code was produced, was traditionally a high point in the history of medieval China. The T’ang dynasty had been founded by Li Yüan (r. 618–626) in 618 and was heir to the achievement of the brief Sui dynasty (589–618) in re-unifying China after four centuries of division. The T’ang

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

302

McMullen

rapidly developed into politically the most stable and economically the most prosperous state that China had seen. The T’ang sovereigns governed through the mandarinate, a hierarchy of officials in nine degrees and twenty-nine classes, who numbered about 17,000, and through a substantial clerical bureaucracy below them. The T’ang empire, extending from beyond the Great Wall in the north to Hai-nan island in the south, had a registered population of nearly 50,000,000. The court resided either at Ch’ang-an, a great walled, grid-plan metropolis of some two million inhabitants and the cultural and intellectual capital of East Asia, or at Lo-yang, the smaller second capital. Up to the rebellion of An Lu-shan in late 755, there were seldom serious threats to the state’s security. The empire-wide registration and taxation systems, though in decline towards the end of this period, were effective enough to supply the court and metropolis with increasing wealth. The scholarly and educational institutions established by successive T’ang rulers, benefiting from stability, were more successful than any before them. The community of scholars that supplied these academic institutions, though numerically small and concentrated on the metropolis when compared to the larger, more devolved intellectual community of the pre-modern era, was nonetheless larger than in pre-T’ang times.12 This long period of prosperity and institutional stability permitted the royal ritual programme inherited from the period of disunion and embodied in successive pre-T’ang ritual codes to expand and to attain new levels of sophistication and grandeur. Indeed this very expansion was an underlying reason for the production of the code of 732. For the K’ai-yüan ritual code, known even in the eighth century by the reign period in which it was compiled,13 was the third that had been produced under T’ang rule. The first, completed in 637, and the second, finished in 658, had each in turn been considered unsatisfactory. The reasons for their rejection were complex in a way characteristic of the T’ang ritual tradition; a compound of political as well as technical ritual motives brought about their replacement. Underlying these motives, however, was a confidence resulting from unprecedented prosperity.14 12  D. C. Twitchett (ed.), The Cambridge history of China, vol. iii, Sui and T’ang China 589–906, Part I (Cambridge, 1979), pp. 150–463. Arthur F. Wright, ‘Symbolism and function: reflections on Ch’ang-an and other great cities’, Journal of Asian Studies xxiv.4 (1965), p. 668. 13   t hy ch. 37 p. 671. But cf. ttkyl, preface p.4a, quoting thy: and Ta T’ang liu tien (Administrative regulations of the great T’ang dynasty, hereafter ttlt), edn of 1724 edited by Hiroike Senkurō and Uchida Tomoo, published as Dai Tō rikuten (Tokyo, 1973) ch. 4 pp. 14a–b; and thy ch. 9A p. 164. 14  For the two earlier codes, thy ch. 37 pp. 669–70; tt ch. 41 pp. 233A–C; cts ch. 21 pp. 816– 18; hts ch. 11 pp. 305–9. Robert des Rotours, Le traité des examens (see n. 10), p. 149, n. 1, dates the first code to 633; this date, given in thy ch. 37 p. 669, probably refers to an

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

The Ritual Code of T ’ ang China

303

The immediate decision to compile a new code was the indirect result of traditional veneration for, combined with practical dissatisfaction with, one of the three ritual canons. It had been suggested in 726 that the prescriptive parts of the Record of ritual, the least homogeneous of the three ritual canons, might be systematically re-arranged to form a prescriptive authority for the imperial ritual calendar. However, one of the chief ministers of the time, Chang Yüeh (667–731), called the Record of ritual ‘an inerasable authority over the ages’, and blocked the proposal. Chang Yüeh held concurrent office as the director of the Chi-hsien yüan, an advisory scholarly college founded by the emperor under another name in 717 and the most successful institution of its kind in the dynasty. His opposition prevailed, and instead it was decided to commission a new ritual code, to ‘find the mean’ between those of 637 and 658.15 The commission appointed to draft the new code consisted of at least four scholars and was under the direction of the same Chang Yüeh. The operation took place, however, not in one of the agencies that had responsibility for running the royal ritual programme but in the Chi-hsien yüan itself. Indeed there is a hint that the emperor had originally intended that one of the main functions of this college should be the re-drafting of state ritual directives.16 Although the commission included experienced scholars, there were difficulties and delays, frequently attendant on state initiated scholarly projects in the seventh and eighth centuries. By 731, Chang Yüeh and another senior member of the commission were dead. The emperor then brought in an outside official, Hsiao Sung, a successful general service administrator rather than a scholar, to take charge of the entire programme of the Chi-hsien yüan.17 Hsiao Sung drafted in probably at least three other academic personnel to work on the code. In 732 it was completed, submitted, approved by the emperor and promulgated.

interim set of directives only. Tzu-chih t’ung chien (Comprehensive mirror for aid in good government, hereafter tctc (Peking, 1956) ch. 194 p. 6217, 637.10, and cts ch. 73 p. 2595 both give a date of 637. 15   t tkyl preface 4a–6a; thy ch. 37 pp. 670–71; tt ch. 41 p. 233C; cts ch. 97 pp. 3049–57; and hts ch. 125 pp. 4404–10, biographies of Chang Yüeh. Chang blocked another attempt to revise the Li chi in 726; see K. P. Kramers, ‘Conservation and the transmission of the Confucian canon’. Journal of the Oriental Society ii.1 (1955), pp. 118–32. For the Chi-hsien yüan, see Ikeda On, ‘Sei Tō no Shūkenin’ (The Chi-hsien yüan in the high T’ang), Hokkaidō Daigaku Bungakubu kiyō xix.2 (1971), 47–98. 16   Ch’üan T’ang shih (Complete T’ang poems) (Peking, 1960), ch. 3 p. 35. 17   c ts ch. 99 pp. 3093–5 and hts ch. 101 pp. 3953–4.

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

304

McMullen

The Ritual Code

The K’ai-yüan ritual code contains detailed directions for just over 150 rites. Many of these had well documented histories going back centuries. Some had been controversial in the T’ang itself. One or two were innovations, established just in time, as it were, for inclusion in the code. The rites are organised into five divisions, following, though re-arranging the order of, a scheme mentioned in the Confucian canons and believed by T’ang scholars to have existed in remote antiquity.18 Each division contained its own category of rituals, which were distinct in terms of their subject matter or structure. The first division was for ‘propitious rituals (chi li)’. These concerned sacrifices (ssu or chi) to the supernatural powers or cosmological forces, which were made on the main altar sites at the capital and in the suburbs or at altars elsewhere in the empire. The recipients of these offerings were the gods of high heaven and earth, the gods of the five directions and of the harvest, the sun, moon, stars, sacred peaks, seas and great rivers. The offerings (hsiang) to ancestors of the emperor were also ‘propitious rituals’. The imperial ancestral cult centred on the imperial mausolea, five in number at this point in the dynasty, with their tombs cut in natural rock in hills to the west of Ch’ang-an,19 but more crucially on the dynastic ancestral temples (tsung miao), of which there were two, one at Ch’ang-an and a second at Lo-yang. The ancestral precincts and the rituals that took place at them stood, in a way no other rites could, for the dynasty itself, and in both ritual and political terms were extremely sensitive.20 A less prestigious ritual in this division was the twice yearly offering to Confucius, his immediate disciples and figures in the Confucian exegetical tradition. In the capital this ritual, the Shih-tien, was held at the state academy directorate, the institution responsible for the school system empire-wide.21 A parallel offering was prescribed for the ancient teacher of military skills Ch’i T’ai kung, a figure held to have lived at the start of the Chou and promoted as the military counterpart to Confucius, the exemplar of civil virtues. Though 18   Shang shu chu shu (Book of documents with commentary and sub-commentary) (Ssu-pu pei-yao edn) ch. 3 pp. 5b, 6a and 7a; Li chi chu shu (Record of ritual with commentary and sub-commentary) (Ssu-pu pei-yao edn.) ch. 1 p. lb; tt ch. 41 p. 233A. 19   c ts ch. 25 p. 973; cf. ttlt ch. 14 p. 27a which lists six. Hsüan tsung chose the site for his own burial late in 729; see thy ch. 20 p. 397. 20   t t ch. 47 p. 271B; cts ch. 26 p. 979; thy ch. 17 p. 352–5; cts ch. 8 p. 177, ch. 25 p. 952. 21   t tkyl ch. 53–4; 7 tt ch. 53 pp. 303C–304C, 305C–306C; thy ch. 35 pp. 635–43; Ta T’ang chiao-ssu lu (Record of suburban and temple offerings under the great T’ang), in Dai Tō Kaigen rei, pp. 725–820, ch. 10 pp. 5b–9b. The terms chi, ssu, hsiang and Shih-tien are officially defined in ttlt ch. 4 p. 33a.

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

305

The Ritual Code of T ’ ang China

Figure 23

A traditional Chinese wood-block picture map of Chao-ling, the mausoleum of T’ai tsung, the second T’ang emperor (reigned 626–649). From a gazetteer for Li-ch’üan county, Shensi province, prefaced in 1535. The mound in the centre represents T’ai tsung’s mausoleum, and to its south are the tombs of the senior civil and military officials who were given special permission to be buried near him. Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

306

Figure 24

McMullen

Representations of the positions required for dancers in a ceremony to Confucius. The final four characters read, ‘Ah! We complete the ritual’. From a historical compendium on Ch’üeh-li, Confucius’ birthplace, originally prefaced in 1505 and published in an amplified edition in 1870.

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

The Ritual Code of T ’ ang China

307

T’ai kung had had temples before, a cult at this level was a T’ang innovation, established by edict only one year before the completion of the code.22 This first division of the code also contained the directives for a number of rituals which were non-recurrent in the sense of not being conducted annually or at fixed intervals but rather when the emperor, ministers and ritual scholars agreed conditions to be appropriate.23 The greatest of these, and the least frequently performed, were the Feng and Shan rites, which in T’ang tradition might be held on Mount T’ai, the sacred mountain in Shantung that was believed to link the human and spirit worlds, or on the other four sacred mountains of China. In T’ang times these rites had been performed only twice before 732, in 666 and 725 on Mount T’ai and once, early in 691, by the usurping Empress Wu, on Mount Sung, the central peak.24 Other non-recurrent rituals concerned the reports the emperor made at the ancestral temple and at other capital altar sites on returning from journeys. In this division of the code there were also prescriptions for certain of the important rites to be conducted at local level throughout the empire, in humbler versions of their grand imperial counterparts.25 The programme of recurrent and non-recurrent rituals this division of the code contains was conceded to be too onerous for the emperor to discharge every year. For twenty-five of the rituals in this division the code therefore provided prescriptions for proxy celebrations in the absence of the emperor. The role of taking the place of the sovereign devolved first on the grand pacifier [t’ai wei], one of the ‘three dukes [san kung]’, the highest officials in the hierarchy, then, if this post was vacant, on officials of the third grade and above. The senior officials of the agencies that ran the state ritual programme also participated.26 The second division of the code was the very much shorter ‘rituals for guests [pin li]’. These were concerned with the reception of envoys from outlying 22   t tkyl ch. 55; tt ch. 53 pp. 306C–307B; thy ch. 23 pp. 435–9; Ta T’ang chiao-ssu lu ch. 10 pp. 9b–17b. 23   h ts ch. 11 p. 3120 makes the distinction between recurrent (ch’ang) and non-recurrent ( fei ch’ang) rites. 24   t tkyl ch. 63–4; tt ch. 54 pp. 310C–314A; cts ch. 23 pp. 881–907; hts ch. 14 pp. 349–53. Edouard Chavannes, Le T’ai Chan (Paris, 1910), pp. 169–235 translates cts ch. 23. For the interest of T’ang sovereigns in celebrating the Feng and Shan rites on sacred peaks other than Mount T’ai, see my note in T’ang studies, ii (Winter, 1984), pp. 37–40. For the traditional position of Mount T’ai, see Shang shu chu shu ch. 2 p. 6b. 25   t tkyl ch. 56–61; 68–78. 26  For the regulations for proxy celebrations, see ttlt ch. 4 p. 40b; cts ch. 21 p. 819, ch. 43 p. 1815 and 1831, cf. also ttlt ch. 14 p. 12b, 156a; ttkyl preface 1a–4b; hts ch. 11 p. 310.

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

308

McMullen

territories, their procedure for their audiences, feasts, presentation with gifts and departure. It comprised only six observances. The code does not name the states involved, but refers to princes of foreign states, or ‘envoys from foreign states’, and these states were graded into great, medium and ‘small below’. The Administrative regulations of 738 or 739 lists over seventy states as existing at this time, but does not grade them.27 The third division, only slightly longer than the second, was for ‘army rituals [chün li]’. These included the rites the emperor was to perform at precincts at the capital, among them the dynastic ancestral temple, before he went on campaign; regulations for the proclamation of victories and the rewarding of officers. There were also prescriptions for ‘rehearsals at war’, to be held in the middle month of winter, for hunts and for imperial attendance at, or even participation in, archery competitions.28 The fourth division of the code was much longer and was more concerned with rites of passage than with cosmic or supernatural forces. The ‘felicitation rituals [chia li]’ contained prescriptions for marriage ceremonies, from those for the emperor himself, the crown prince and princesses to those for officials of the sixth grade and below. There were directives for the procedures by which the emperor and empress were to receive congratulations from the hierarchy of imperial family members and from officials. Ceremonies for the investiture of the crown prince, other princes and senior ministers and for the appointment of envoys to the provinces formed another group in this division. Yet another included coming of age ceremonies for all levels of the hierarchy to the sons of sixth grade officials and below. Finally, a number of rituals were for local performance, among them the declaration of amnesties, the announcement of the imperial will and the despatch by provincial officials of memorial submissions to the court.29 The fifth and last division of the code was for ‘rituals of ill omen [hsiung li]’. This contained procedures to be conducted after bad harvests, illness and mourning. The code prescribed details for those involved down to officials of the sixth grade and below. The first T’ang codification, that of 637, had included a sixth section, containing the prescriptions for imperial funerals, but this had been removed from the second code, for reasons of tact, and it was absent also from the K’ai-yüan code. This omission meant that whenever an 27   t tkyl ch. 79–80; ttlt ch. 4 pp. 56a–b. 28   t tkyl ch. 81–90. 29   t tkyl ch. 91–130. For a vivid eyewitness description of an announcement of the imperial will, at Teng-chou on the northern coast of Shantung province in 840, see Edwin O. Reischauer tr., Ennin’s diary (New York, 1955), pp. 180–2.

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

The Ritual Code of T ’ ang China

309

emperor died the funeral procedures had to be drafted anew. It was not to be until 806, at a time when individual officials wrote sardonically about the sycophancy of their seventh-century predecessors over this matter, that the dynasty officially permitted directives for imperial funerals to be preserved for future use. Perhaps again because to do so would have involved ‘anticipating untoward events’, in the phrase used when imperial funerals were withdrawn from the code, no prescriptions for accession, or for ‘proceding to the throne’ were included.30 In a system that pre-dated the T’ang, each of the observances in the first, the ‘propitious rituals’ division, was ranked in one of three grades, major, medium or minor.31 In the code as a whole, over sixty rituals specified the participation of the emperor in their titles; but he acted as principal in a considerably larger proportion than this figure implies. The titles of a further twenty-two mentioned the empress or the crown prince. Some of these royal rituals were specific to the emperor, in the sense that no ritual resembling them was performed at lower levels in the hierarchy. Others were imperial versions of rituals that were repeated, in carefully stipulated and decreasing degrees of grandeur down the mandarin scale or in the provincial administration. The main rites of the ‘propitious rituals’ division of the code were specific to the emperor: for example the offerings at the suburban altars or the Feng and Shan rites on the sacred mountains. The highly sensitive dynastic ancestral cult on the other hand, although its rites were held to be distinct in name and in their organisation, was in essence a grand imperial version of the ancestral cults that officials of the fifth grade and above were permitted to maintain. The T’ang ritual tradition as a whole offered varying explanations of the emperor’s position before heaven, earth and man. But this provision, which was sanctioned by the canons, at least suggests that he differed in degree only from other members of the noble or official hierarchy. On the other hand to some T’ang scholars the imperial ancestral cult was different from other ancestral cults in being a ‘public [kung]’ concern, while the ancestral cults of high ranking families were ‘private [ssu]’. This distinction was largely ethical rather than religious. Some scholars thus used the 30   t tkyl ch. 131–50; Liu Ho-tung chi (Collected works of Liu Tsüng-yuan (773–819)) (Peking, 1958) ch. 21 p. 367–9; thy ch. 37 pp. 669–70. tt, despite this prohibition, contains directives for Tai tsung’s funeral in 779. 31   t t ch. 45 p. 260B; ttkyl preface 1a–b; ttlt ch. 4 p. 33a; tt ch. 106 p. 561B; cts ch. 21 p. 819 hts ch. 11 p. 310. See also the article by Kaneko Shūichi, ‘Tōdai no daishi, chūshi shōshi ni tsuite (Concerning the major, medium, and minor offerings in the T’ang)’, Kōchi Daigaku Gakujutsu kenkyū hōkoku, 25, Jimbun kagaku, 1, pp. 13–19.

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

310

McMullen

term ‘public’ when commenting critically on the emperor’s reluctance to allow them to influence aspects of the imperial cult.32 One observance repeated in humbler versions through the T’ang empirewide administrative structure was the twice yearly offering to Confucius, his disciples and the main figures of the Confucian exegetical tradition. In its grandest, metropolitan version, this ceremony, the Shih-tien, was conducted by the crown prince as chief celebrant, with the entire official body and the students of the state academy in attendance. In the provincial versions, the chief celebrants were the local prefects, or in the administrative subdivisions of the prefecture, the county magistrates. Their brief addresses to Confucius, stipulated in the code, were however identical, but for the declaration of their names, to that given by the crown prince. This too was a ‘public’ ceremony, in that private celebrations of the Shih-tien seem likely to have been forbidden by law. Epigraphical evidence, and the evidence of texts that were both inscribed and transmitted in literary collections, indicates that this ritual was widely conducted.33 There were local variations however, and accounts of how the Shih-tien was conducted on the remote southern coast of China caused only contemptuous amusement in the north.34 For the code’s purposes, the lowest level in the administrative hierarchy was the village [li]. The code stipulates one ceremony only for village level, an offering to gods of soil and grain. This, like the Shih-tien, was a public ritual to be performed empire-wide. The grand, imperial version of this ritual was a medium grade observance, for which proxy directives were also supplied. Called in the code the ‘sacrifice to the grand altar of the god of the soil’, it involved standard offerings of cooked meat and of jade and silk to both the god of the soil and the god of the harvest in the middle month of spring and autumn. At the imperial level, elaborate equipment and the attendance of all grades of officialdom, besides the principal celebrants, were required. For the humble village version, however, the code stipulated only the participation of the head of the community and the community members, while the site of the o­ bservance 32  For the laws for private ancestral temples see ttkyl ch. 3 p. 12a; ttlt ch. 4 p. 40a; also thy ch. 19 pp. 387–92 and T’ang ta chao ling chi (Great collection of edicts and commands of the T’ang) (Shanghai, 1959), ch. 68 p. 382. For the idea that the ancestral temple was governed by ‘public’ considerations, see cts ch. 25 pp. 954–5; thy ch. 14 pp. 315–16; tctc ch. 244 pp. 7876–7, 831.6. 33   Wen yüan ying hua (Flowers in the garden of letters) (Peking, 1966) ch. 509 pp. 2b–3b; cts ch. 187A p. 4874; Liu Ho-tung chi ch. 5 p. 74–7; Han Ch’ang-li chi (Collected works of Han Yü (768–824)) (Kuo-hsüeh chi-pen t’sung-shu, edn.), vol. 6, ch. 31 p. 58. 34   Ch’ün-chü chieh-i (Humorous stories for living in a crowd), in Li-tai hsiao-hua chi (Chronological anthology of humour) (Shanghai, 1956), p. 59.

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

The Ritual Code of T ’ ang China

311

was to be ‘beneath a sacred tree’. As for the equipment, ‘if there are no ritual vessels, other vessels are to be used as best the situation allows’. The community headmen were, however, required to celebrate in the middle month of spring and autumn, just as the emperor or his proxy were at the capital.35 Among other rituals that were repeated down the hierarchy were rites of passage in the fourth division of the code, like coming of age and marriage rites.36 Such ceremonies were therefore not exclusively royal rituals, but were rather general ones, of which the imperial performance was the grandest. In their non-imperial versions at least they were considered private. In the case of rites of passage the code prescribed only to the rank of ‘sixth degree (officials) and below’. In stopping its coverage at the ninth, the lowest degree of officialdom in this way, the code again underlined its concern with the apex of the T’ang social hierarchy only. Here its compilers tacitly endorsed the principle, first formulated in the Record of ritual, that ‘the rules of ceremony do not go down to the common people. The penal statutes do not go up to the great officers’. The code thus bore out the hierarchical view of society that was essential to the Confucian perspective and that contrasts with the more universal outlook of the transcendental faiths. Concerning the many popular or unofficial rites and festivals that existed alongside state rituals, the code therefore supplies no information at all, just as it remains silent about the flux of religious practices among the population at large.37 Inevitably, within each of its five divisions rituals tended to echo one another or to have shared features. The code, rather than repeat prescriptions, acknowledged this by supplying cross-references when particular rites used procedures common to others. The internal consistency to which this feature of the code bears witness is an indication of the systematic approach of the compilers to the tradition. In the ‘propitious rituals’ division, rituals were treated typically under seven headings.38 First there were directives for preliminary abstinence, for which there were two phases, ‘relaxed’ and ‘intensive’. General regulations at the front of the code stipulate four days of ‘relaxed’ and three 35   t tkyl ch. 71 pp. 7a–9b; ttlt ch. 4 p. 39b and 51b; Han Ch’ang-li chi vol. 6 ch. 31 p. 58 and hts ch. 164 p. 5058. For the imperial version see ttkyl ch. 33–4, and for confirmation of its middle grade status, cts ch. 24 p. 933. Rituals involving prayers for rain were also prescribed for local level; these were by their very nature non-regular; ttkyl ch. 67, 70 and 73; ttlt ch. 4 pp. 41b–42a. 36   t tkyl ch. 114–22 pp. 123–5. 37  The orthodox view towards these was one of disapproving condescension; see T’ang kuo shih pu (Supplement to the T’ang dynastic history) (Shanghai, 1957), p. 65; or sometimes of harsh intervention, cts ch. 89 p. 2887, ch. 156 p. 4129. 38  Cf. hts ch. 11 p. 310.

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

312

McMullen

of ‘intensive’ abstinence for a major ritual; three days of ‘relaxed’ and two of ‘intensive’ for a medium ritual; and two days of ‘relaxed’ and one of ‘intensive’ for a minor one.39 If these ‘periods of separation’, in modern anthropological terms, proved impracticably long, the code prescribed a single night of ‘pure abstinence’. During ‘relaxed abstinence’, business was to be done as usual, except that participation in mourning, enquiring after sickness, or deciding on or signing documents concerned with punishment or killing was not permitted, nor was the imposition of punishment on criminals or participation in ‘unclean or evil activities’. ‘In intensive abstinence (the participants) are permitted to attend to the business of the ritual only; the other activities are all barred’. Abstinence was not normally prescribed for the rituals whose primary purpose was social, those in the ‘rituals for guests’ and the ‘propitious rituals’ divisions. A night of ‘pure abstinence’ was however required of all participants in those army rituals that took place at the main capital altar precincts or in the imperial ancestral temple.40 The next heading concerned the deployment of the equipment to be used in the observance. Usually three days beforehand, tents (tz’u) would be set out where participants were to await the ritual. Two days before, appropriate musical instruments, themselves the subject of much attention and historical erudition and the responsibility of specialist offices, would be deployed. Placements (wei) for the ritual itself, with the position of participants and the direction they faced stipulated, would be set out. A third heading usually concerned ritual inspection of the food offerings and vessels.41 Then there was a heading for the directives for the imperial procession as it left the palace or relevant premises and went, probably at a brisk pace, to the precinct concerned.42 The fifth and sixth sections provided directives for the offering of the jade and silk, and the presentation of sacrificial meat and wine, by the emperor or his representative. In these, the central episodes for rituals in the first division of the code, there were again recurrent features. The code invariably prescribed three presentations or offerings to the divinity, cosmological agent, ancestor 39   t tkyl ch. 3 pp. 7b–8b; ttlt ch. 4 pp. 41a–b; cf. Lung chin feng sui p’an ( Judgements of dragon sinew and phoenix marrow) (Hu-hai Lou ts’ung-shu edn.) ch. 4 pp. 4a–5b. 40   t tkyl ch. 81 p. 2a, ch. 82 p. 1a, ch. 83 p. 1a, ch. 84 p. 1a. 41   h ts ch. 11 pp. 312–16. 42  The pace may be inferred from a memorial dated 711 by the great historian–critic Liu Chih-chi (661–721); see tt ch. 53 p. 304C and cts ch. 102 p. 3172. Han Yü implied that acolytes rushed about during ceremonies; Han Ch’ang-li chi vol. 4, ch. 14 p. 28 and Shang shu chu shu ch. 11 pp. 11b–12a.

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

313

The Ritual Code of T ’ ang China

Figure 25

Drawings of traditional Chinese ritual vessels, of the sort stipulated in the code. From a historical compendium on Ch’üeh-li, Confucius’ birthplace, originally prefaced in 1505 and published in an amplified edition in 1870.

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

314

McMullen

or deceased exemplar involved. For it was an accepted principle that ‘ritual is perfected in threes’.43 If the emperor participated, he made the first offering, to be followed by one of the very high ranking officials, with the president of one of the ritual agencies typically making the third presentation. In many cases, each presentation was accompanied by the reading of a ‘prayer text [chu wen]’. The code supplies texts in the case of some rituals; for others it gives a note stating that the text was to be composed ‘near the time’. The texts the code supplies are notable for the laconic quality: they usually involved the participant stating the year, month and day, identifying himself, and then giving a very brief address in archaic and honorific language. In the case of the rite to Shen Nung, a legendary paragon emperor and originator of husbandry, for example, the emperor stated, ‘I make bold to announce to Shen Nung that, at the start of the first spring month, the ploughing work begins. In accordance with the laws and precepts, reverently we attend our thousand acres.’ He then used a standard formula to present the offerings: ‘Diligently with a length of silk and with sacrificial grain, I reverently make replete the unvarying offering and set forth the holy oblations.’44 Normally even for the grandest rituals the prayer text did not involve more than a very brief report of this kind. Any ‘transformative’ element in modern anthropological terms, involving explicit change in the status of the performer in relation to the divinity or cosmological agent addressed was therefore missing. However, in one ritual the emperor or his proxy prayed for the grain harvest, and in times of drought or flood, a number of supplicative rituals were prescribed, down to prefectural and county levels. In these, the emperor or his local representative requested a response from the cosmic forces or from local divinities.45 By and large, therefore, the organisation of this the central episode of the rituals in the propitious division amply bears out the this-worldly, nontranscendent or ‘archaic’ focus of the official Confucian religious outlook. Another feature of the central episode of the rites in the ‘propitious rituals’ division of the code, again sanctioned in the Record of ritual, was that the recipient of the offering named in the title of the ritual was often not the only one involved. Another, ‘correlative’ recipient was frequently included. Sometimes this was a divinity or deceased exemplar who was loosely associated with the main recipient. Thus when the emperor made offering to Shen Nung, the code prescribed that Hou Chi, the legendary ancestor of the Chou house and

43   t hy ch. 8 p. 114. 44   t tkyl ch. 46 p. 11a. 45   t tkyl ch. 6–7, 65–7, 70, 73; ttlt ch. 4 pp. 41b–42a.

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

The Ritual Code of T ’ ang China

315

s­ uperintendant of grain under the paragon emperor Yao,46 should also receive offerings. In a number of the rituals, the correlatives were deceased members of the T’ang imperial line. In this way, the dynasty gave to its ritual programme, almost all of which was of far greater antiquity than the T’ang itself, a specifically T’ang identity. Thus in the Feng rites on Mount T’ai to high heaven, the code prescribed Kao tsu, that is Li Yüan, the founder of the dynasty, as the correlative, while in the Shan rites that followed it was Jui tsung, at the time the code was compiled the most recently deceased sovereign and Hsüan tsung’s own father, who was stipulated. The detailed narratives of the Feng and Shan rites on Mount T’ai in late 725 show that here the code was endorsing the precedent of the most recent celebration of this highly prestigious ritual.47 In some cases more than one correlative recipient was prescribed. In its version of the Shih tien ritual to Confucius, the code stipulated as correlatives the seventy-two Confucian disciples, and twenty-two exegetical scholars, whose lives had spanned a period of nine hundred years till the fifth century ad.48 The counterbalancing military cult was even more accommodating, for it included, at least in middle and late eighth century times, offering to a small number of very recently deceased T’ang generals.49 But the most immediate provision of this kind concerned the imperial ancestral temple. For a very few of the greatest officials of a reign were posthumously represented in the shrine of the deceased emperor whom they had served. To be installed as a ‘correlative meritorious official’ in the ancestral temple was one of the rarest of the posthumous honours the state offered.50 The principle of the correlative recipient was thus clearly very useful, for it permitted expansion and a pluralistic approach to both divinities and past exemplars. In this way, the scope of the symbols to which the state gave approval was kept flexible, and the ability of the rites to integrate a wide range of beliefs and experiences in the interests of the dynasty was reinforced. The sixth and final heading in the codes’ directives for rituals in ‘propitious rituals’ division typically concerned the procedure for the return of the imperial party or the principal celebrant. This tended to be standard for a high 46   t tkyl ch. 46 p. 11a; ttlt ch. 4 p. 36b. 47   t tkyl ch. 63 p. 13a and 15a; ch. 64 p. 7b and 9b; cts ch. 23 p. 889, 900 and 902; thy ch. 7 pp. 114, 115 and 117. 48   t tkyl ch. 53 pp. 9a–10a; ttlt ch. 4 pp. 36b–37b; Ta T’ang chiao-ssu lu ch. 10 p. 7a. 49   t tkyl ch. 55 p. 5a; thy ch. 23 pp. 435–6; Ta T’ang chiao ssu lu ch. 10 pp. 9b–12a. 50  The ‘correlative meritorious officials’ up to Jui-tsung’s reign (710–712) are listed in ttlt ch. 4 pp. 35b–36b and thy ch. 18 pp. 370–1. Neither text gives lists for the reign of the Empress Wu (690–705).

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

316

Figure 26

McMullen

Eighteenth-century picture diagram of the round altar. The places for participating officials, e.g. members of the board of rites, the court of ceremonial for foreigners, and the censors are marked, as are the positions of offering vessels, censers, etc. From the Chinese Encyclopaedia of 1726.

proportion of the observances, and the code therefore simply referred to the directions given for the first ritual, that for the offering at the round altar held at the winter solstice. In one or two cases the code required an assembly or feast for the officials who had participated, to be held a day after the ceremony concerned. The ploughing rite which followed the emperor’s offering to Shen Nung and Hou Chi was followed by a feast in the palace of the Supreme Ultimate for all participants. The authority for this apparently highly formal celebration came from the Monthly commands, a section of the Record of ritual, and one of the most important sanctions for the cosmological element in the imperial ritual calendar.51 51   t tkyl ch. 46 pp. 15b–16a; James Legge, The Li Ki (see n. 2), pp. 254–6. Cf. tctc ch. 214 p. 6810, 735.1. Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

The Ritual Code of T ’ ang China



317

The Ritual Institutions

The amount of detail in the K’ai-yüan ritual code is prodigious. The Chinese text as it was copied for the imperial library in or just before 1782 runs to approximately 1000 traditional pages,52 while the commonly available edition of 1886 approaches 1500. Rendered into English, the code’s length would exceed this several fold. But T’ang official scholarly compilations tended to be highly compartmentalised. The purpose of the code was prescriptive, to provide sound directives for a programme of rituals that the compiling officials themselves had delimited. It restricted itself almost exclusively to directives for performances, not to the usages and sumptuary regulations that have also been seen as an aspect of the state ritual tradition.53 It therefore includes little background or explanatory information. Only in the opening three chapters does it depart from its concern for specific detail and supply more general directives. This prefatory section covers matters such as divining auspicious days, imperial and official processions, the rules for dress, abstinences, definition of droughts, disposal of deficient ritual equipment, care of animals intended for sacrifice and the regulations for furnishing the tombs of officials. Modern Japanese scholarship has established that this material reflects closely contemporary statutes and ordinances governing state rituals. Its different purpose explains why it differs in tenor from the bulk of the code.54 This admirably tidy and disciplined approach means that other T’ang sources, mainly official histories and institutional compendia, must supply what the code refrains from giving: a more detailed description of the institutional framework that supported the ritual programme, and of its place in the bureaucracy as a whole. The implementation of the T’ang ritual programme was the responsibility of a number of special agencies in the central government, while in the provinces it normally devolved either on officials despatched from the central government or on the general prefectural or county administrations. The ritual agencies in the central government were, however, integral parts of the general bureaucracy. Personnel appointed to them were governed by the same laws as officials elsewhere in the civil service. Length of tenure, regulations governing promotion and so on were therefore in no way different. Though those in state ritual posts were often referred to as ‘ritual officials 52  Photolithographically reprinted in Ssu-k’u ch’üan-shu chen-pen (Rare volumes from the complete collection in four treasuries), eighth series (Taipei, 1978). 53  Etienne Balazs, ‘History as a guide to bureaucratic practice’, in Chinese civilization and bureaucracy, tr. H. M. Wright, ed. Arthur F. Wright (New Haven, 1964), pp. 136–7. 54  Ikeda On, ‘Dai Tō Kaigen rei no kaisetsu’ (Explanation of the Ta T’ang K’ai-yüan li), in ttkyl p. 823. Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

318

McMullen

[li kuan]’. there was no question of their being a privileged group, still less of their having any form of special initiation. The institutional structure to which they were appointed was remarkably stable, more so, it could be argued, than the ritual programme itself, for it underwent relatively little evolution in the course of the dynasty. The main central agencies concerned were four ‘courts [ssu]’, all having premises in the imperial city, the great enclosed precinct in which the principal institutions of the government were sited. The most important of the ritual courts was the court of sacrifices. As a title at least, this went back to the Former Han dynasty. It was headed by a president, whose grade was third degree first class on the mandarin scale. At the time the code was compiled, besides his role as a participant in imperial, proxy and other rituals, he bore responsibility for administering and maintaining the great majority of the precincts in which the code’s rituals were conducted. Beneath him there were eight ancillary bodies, whose names indicated their function. They were for the upkeep of altar sites, for music, for ‘drums and wind (instruments)’, for medicine, divination, the supply of sacrificial animals, for the numerous lesser cult shrines that the dynasty maintained, and finally for the maintenance of the imperial mausolea. There were also four services, concerned with keeping all equipment used in the rituals. The presidency of the court of sacrifices was considered a high status position. Though it did not involve activity in the mainstream of political life, it was sometimes used as a holding post for senior officials. There are thus several examples from late in the eighth or early in the ninth century of appointment to chief ministerships following tenure of the presidency.55 Another role that the president of this or of the other ritual courts was occasionally given was that of diplomatic commissioner to foreign states. The relations between China and its neighbouring states, this arrangement seems to imply, were conceived of as an extension of the ritual programme, of the highly regulated ceremonial framework prescribed in the ‘rituals for guests’ division of the code.56 Beneath the president there were two deputy presidents, whose grade was fifth degree third class and whose posts were also considered prestigious.57 Below these there were two registrars and four doctors. The latter had special responsibility for drafting and revising the directives for the rituals in the code. 55   c ts ch. 119 p. 3435, Yang Wan in 777; cts ch. 136 p. 3756, Ch’i K’ang in 800; cts ch. 147 p. 3976, Kao Ying in 803; cts ch. 147 p. 3973, Tu Huang-shang in 805. Cf. cts ch. 67 p. 248.7, Li Chi in 648. 56   c ts ch. 196B p. 5245. cf. cts ch. 196B pp. 5263–4. 57   P’i-ling chi (Collected works of Tu-ku Chi (725–777)) (Ssu-pu ts’ung-k’an edn), ch. 17 p. 6b.

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

The Ritual Code of T ’ ang China

319

They also had an entirely separate, and politically sensitive, function, that of proposing canonisation titles for deceased mandarins of the third degree and above. Below them again were a number of low ranking, low prestige offices bearing responsibility for minor aspects of the ritual programme. Several appear frequently in the code, having roles as heralds or ushers. In general sources they are sometimes mentioned as first appointments to junior officials. In these cases they typically led to general service appointments rather than careers specialising in state ritual. Another, very junior role was assigned to acolytes, who might be as young as twelve and who were the sons of officials of the sixth degree or higher in the general bureaucracy. These seem to have numbered as many as 862, and were attached to the offices responsible for altar sites and to the dynastic ancestral temple. Under the influence of the usurping Empress Wu, as an aspect of her promotion of the female interest in all aspects of government, in the first decade of the eighth century there were girl acolytes. Acolytes formed one of the groups that were set facilitated examinations leading to official status. In the late eighth century there was discussion of the idea that students from the state academy might take the role of regular acolytes when the number was later insufficient, and the problem was set as an examination question. Service as an acolyte may have had its irksome aspects, for it was sometimes excused in imperial acts of grace. But the use of the ritual institutions and programme to provide sons of ranking officials with a familiarisation with the state and its procedures, though certainly a pre-T’ang practice, indicates how central the Confucian sanctioned ritual programme was to the official hierarchy.58 Of the other courts with state ritual functions, the court of imperial banquets had particular responsibility for the wine and meat used in sacrifices. Its president, with the rank of third degree second class, offered the third and final presentation in proxy imperial celebrations. Like the court of sacrifices, the court of imperial banquets ran ancillary institutions, in this case six, concerned with the maintenance of ritual kitchens and the preparation of ritual foods, drink and equipment. A third court with a role in the state ritual programme, one that grew in importance in the course of Hsüan tsung’s reign, was the court of the imperial family. With a structure similar to that of the court of sacrifices, its main 58   t t ch. 15 p. 85A and hts ch. 45 p. 1180 give a total of 862; ttlt ch. 14 pp. 2a, 3a, suggest a total of only 370, ttlt ch. 4 p. 7a. For girl acolytes, see cts ch. 51 p. 2173; ch. 92 p. 2965; hts ch. 13 p. 337; tctc ch. 209 pp. 6636–7, 709.14. For the irksome nature of acolytes’ duties, see Han Ch’ang-li chi vol. 4 ch. 14 pp. 28–9; and remission of service, T’ang ta chao-ling chi ch. 74 p. 416.

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

320

McMullen

function at the time the code was compiled was the upkeep of the records of the imperial Li family. From 737, however, five years after the completion of the code, it assumed responsibility for both the imperial mausolea and the imperial dynastic temple. In the second half of Hsüan tsung’s reign its functions were substantially increased by another development. The emperor, honouring the claim that the T’ang royal house was descended from Lao tzu, one of the principal deities of Taoism, greatly promoted Taoist texts, education and examinations. He assigned to the court of the imperial family the management of the institutional aspect of this campaign to glorify the teachings of his supposed ancestor. This development represents one of the few instances when the administration of one of the rival belief systems was imported into the Confucian sanctioned institutional framework. Finally, yet another court, that of ceremonial for foreigners, had ritual responsibilities. These concerned the reception of representatives from foreign states and the hospitality, strictly regulated, to be accorded them. This agency would seem to correlate neatly with the ceremonials prescribed in the brief ‘rituals for guests’ section of the code. In fact, the president of this court is not mentioned in the directives for the rituals concerned, for which the main role was taken by the president of the court of sacrifices. But the presidency was sometimes used as a title for diplomatic commissioners visiting foreign states.59 These specialist ritual courts were not, however, the only institutions with responsibility for the imperial ritual programme. In the executive ministry of the bureaucracy, the department of affairs of state, one of the six boards, the board of rites at the time the code was compiled was exclusively concerned with ritual. This body defined and administered rules of propriety and ceremonial usage covering dress, styles of address and other aspects of official life. It also received and processed reports of auspicious omens sent in from the provinces. From 735, however, it had sole responsibility for one of the most prestigious of all the scholarly operations of the dynasty, the regular examination system. But the secretary and supernumerary secretary of the board of rites, nominally at least, oversaw the whole of the ritual code. One of the dependent bodies of the board, the bureau of sacrifices (tz’u pu), had as part of its charge the administering of the ‘propitious rituals’ contained in the code. A second dependent body, the bureau for provisions for sacrificial offerings, was in charge of meat and wine used in sacrifices and offerings. Just what functioning relationship the officials in these bureaucratic bodies had with those in the specialist ritual courts is not clear. However there was 59   c ts ch. 67 p. 2479, ch. 194A p. 5175, ch. 196B p. 5246 and 5251. The presidency of the court of the imperial family was also used in this way. cts ch. 194A p. 5177.

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

The Ritual Code of T ’ ang China

321

evidently some coincidence of role, for the court of sacrifices is sometimes mentioned in conjunction with the board of rites. Moreover the Administrative regulations, in describing the duties of the board of rites, states that regulations for certain of its operations are recorded under the entries for the court of sacrifices, the court of imperial banquets, or the court of ceremonial for foreigners.60 But there was one difference between the senior posts in the ritual courts and those in the board of rites. For the emperor sometimes emphasised the dynastic element in the state ritual programme by appointing members of the imperial family, often in a supernumerary capacity, to the presidencies of the ritual courts.61 The board of rites on the other hand was staffed exclusively from the regular bureaucracy. There is, however, no clear evidence, as far as appointment to high level ritual office is concerned, for a conflict of interest between the imperial Li clan and the regular bureaucracy. The senior posts in the ritual courts, especially those involving scholarly knowledge of the ritual tradition, seem to have been quite as often given to scholar officials from the general service as to imperial princes. Nor was the drafting of rituals seen as an inaccessible or arcane activity. One of the middle ranking offices in the court of sacrifices, that of doctor, demanded scholarly expertise. But, like the senior posts, it was typically given to scholarly general service officials. Moreover its duties involved exposure to general publicity within the official community, and as a result it was much more respected than posts carrying the title of doctor elsewhere in the bureaucracy, in the state academy for example.62 That specialisation in state ritual did occur is undeniable. Some of the great aristocratic clans of the seventh and eighth centuries preserved a family tradition of expertise in state ritual; but they did so partly because such expertise gave, and indeed was openly acknowledged as giving, general career advantages.63 There are also a very few examples of individuals remaining in middle echelon ritual office for long periods. But these were the obsessive scholars of the subject, or the otherwise politically ineffective or unambitious, and they are reported as idiosyncratic cases, or, more 60   t tlt ch. 4 p. 26b and 52b, for the court of sacrifices; ch. 4 p. 52a for the court of imperial banquets; ch. 4 p. 31b and 57a for the court of ceremonial for foreigners. The board of rites and the court of sacrifices are mentioned together in cts ch. 27 p. 1032. But their officials could disagree with each other; cts ch. 26 p. 999. 61   c ts ch. 86 and 95, biographies of the sons of Kao tsung, Chung tsung and Jui tsung provide over twenty examples of this. For the education of Kao tsung’s son Li Hung, crown prince in the 650s, in the Record of ritual, see cts ch. 86 p. 2828. 62   t hy ch. 65 p. 1136; tt ch. 25 p. 148A. 63   Liu Ho-tung chi ch. 21 p. 369; cts ch. 189B p. 4964.

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

322

McMullen

likely, drop out of official biographical sources altogether.64 Tenure of senior and middle ranking ritual office is thus to be seen as a normal feature in the careers of many scholarly general service officials, while the official scholars who staffed the metropolitan academic institutions, the state academy, the imperial library and the history office were even more often appointed to ritual office. The specialised role of the ritual posts is thus to be balanced against the fact that those appointed to them were generalists, if usually scholarly ones. That the administration of the ritual programme was considered one of the operations of the general bureaucracy is also indicated by other bureaucratic provisions in this period. One of the conspicuous developments in the reign of Hsüan tsung was the consolidation of a system of commissionerships (shih) to over-ride the established bureaucracy and address specific problems. In the early eighth century there were commissioners for such widely different areas as transport, revenue, minting and the collection of books and the re-organisation of the imperial library. As early as 666 there had been a grand commissioner for the Feng and Shan rites on Mount T’ai. Probably from 710 there were commissioners for rites (li i shih), to manage the ever difficult drafting problems of imperial obsequies or the funerals of members of the imperial family. In 713 a ritual commissioner was appointed for a ‘rehearsal at war’ ceremony near the capital. Chang Yüeh was a ritual commissioner in late 723, in connection with the sacrifice to the god of the earth of that year. In 725, he was commissioner for the Feng and Shan rites. Another example was Wei Shu, a lifelong official scholar and member of several scholarly commissions, appointed in 750. Ritual commissioners are much more conspicuous in the records for the post-rebellion period than in those of Hsüan tsung’s reign. But their appearance in the documentation for the first half of the eighth century indicates that the state ritual programme was subject to exactly the same evolutionary trends as other, more obviously practical areas of administration.65

64   E.g. in the first half of the century, Wang T’ao, omitted from cts, but included in hts ch. 122 pp. 4354–60, for whom see also thy ch. 13 pp. 304 and 306 and ch. 36 p. 491. Also P’eng Ching-chih, hts ch. 199 pp. 5669–70, and thy ch. 21 pp. 405–6. Later in the dynasty there are the examples of Wang Ching, compiler of the invaluable Ta T’ang chiao ssu lu, for whom see hts ch. 58 p. 1493, cts ch. 25 p. 956 etc. and P’ei Ch’ai thy ch. 65 p. 1136; cf. hts ch. 58 p. 1493. 65  For ritual commissioners, see cts ch. 67 p. 2487; tctc ch. 206 p. 6656, 710.7; thy ch. 26 p. 503; tt ch. 43 p. 248B; cts ch. 102 p. 3148; cf. thy ch. 37 pp. 671–2. It is possible that the use of the term ritual commissioner prior to 750 is anachronistic; but it remains certain that specific ritual tasks were often given to selected officials whose posts were not in the ritual agencies. See cts ch. 189B p. 4964 for a clear example of this.

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

The Ritual Code of T ’ ang China

323

Finally, the integration of the ritual programme in the bureaucracy as a whole is indicated by the fact that the censorate, the body that monitored the operation of the whole of the metropolitan and provincial government, also had explicit responsibility for the imperial rituals. One of its ranks, that of investigating censor, was charged with observing the rituals and maintaining their decorum.66 This was, incidentally, a necessary function, for just as the T’ang court itself was sometimes boisterous, so ritual performances might be attended by clamour and disorder, among participants or onlookers. No brief summary of the institutional framework supporting the imperial ritual programme can do justice to the complexity of the organisation the histories and institutional compendia describe. Other agencies in the central government, like the remaining five courts, the history office or the imperial palace services, might be involved in aspects of the ritual programme. Court diarists were required to note performances of rituals; the bureau of compositions, a department of the imperial library, produced prayer texts. Some posts in the crown prince’s extensive household also had ritual functions. A list of all officials mentioned by the code would of course extend far beyond the main ritual agencies. The attendance of such symbolically important figures as the ‘three dukes’, the highest ranking officials in the bureaucracy, is prescribed with great frequency. So too is that of the direct descendants of the two dynasties that had preceded the T’ang, the duke of Hsi, descendant of the Sui royal house, and the duke of Chieh, descendant of that of the Northern Chou (557–581). Their presence stood for the ideal, which had never been remotely realised, of a voluntary and non-violent transmission of power from dynasty to dynasty, in this case from the two pre-T’ang northern based dynasties to the T’ang itself.67 Likewise the attendance at certain rituals of the lineal descendant of Confucius, at this point in the dynasty in the thirty-fifth generation, had clear symbolic value.68 For important rites, the code required the attendance, in rows according to seniority, of all civil and military officials in the capital. For the twice yearly Shih tien ritual to Confucius, the entire staff and 66   t tlt ch. 13 p. 14b; tt ch. 24 p. 144C; cts ch. 44 pp. 1861–3, esp. p. 1863; Liu Ho-tung chi ch. 26 p. 432, Lung chin feng sui p’an ch. 4 pp. 4a–5b. 67   E.g. ttkyl ch. 4 p. 2b–ch. 46 p. 3b. For the cults to the imperial ancestors of the Northern Chou and Sui houses, see ttlt ch. 4 pp. 39b and 55b, Ta T’ang chiao-ssu lu ch. 10 p. 20a; also cts ch. 43 p. 1832 for their descendants’ ritual court appearances; des Rotours, Fonctionnaires, p. 93. From 750 to 753 however, this system was temporarily abolished; see tctc ch. 216 p. 6899, 750.11, tctc ch. 216 p. 6918, 753.9, thy ch. 24 pp. 462–3. 68   t tkyl ch. 16 p. 2b; tt ch. 53 p. 305; Ts’e-fu yüan kuei (Compilation of most precious documents) (Peking, 1960), ch. 50 p. 6a.

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

324

McMullen

student body of the state academy were paraded, and a guard was provided to keep onlookers in order.69 No figures survive from T’ang times to indicate precisely the cost to the treasuries of the great imperial rituals or the numbers of those present. But the extensive distribution of imperial largesse and of tax concessions, announced in acts of grace that followed great ritual performances, involved enormous expense.70 From the evidence of the Sung period (960–1279), it would seem that the expenditure on the principal rites could be colossal and that the total numbers of those involved might be in the order of thousands.71 A T’ang scholar official’s career was likely to involve, in turn with general administrative tenures at the capital or in the provinces, appointment to a ritual post or an office that carried ceremonial functions. The state’s ritual programme impinged on his professional life and took up his time at many points. As their surviving works demonstrate, numbers of eighth-century scholar officials submitted memorials on ritual questions in turn with proposals on other, quite practical administrative issues, for example on the reform of the taxation or registration systems or on aspects of the political or even military structure. Chang Yüeh himself, as an outstanding general service official with provincial and military experience, was also an active ritual scholar in this way; but there are many other examples from eighth-century times.72 This integration of the rituals into the general bureaucratic service had an important intellectual concomitant. For it meant that, officially at least, all the activities in which scholars were engaged, not merely state ritual, were to be 69   t tkyl ch. 54 p. 2b, ch. 55 p. 2b. Cf. tctc ch. 214 p. 6810, 735.1, for the inability of the Chinwu guard to control the clamour of onlookers after the ploughing rite of 735. 70   E.g. following the ploughing rite of 735, see T’ang ta chao-ling chi ch. 74 pp. 415–16. Cf. thy ch. 21 p. 406 and hts ch. 199 p. 5470 for a rare example of mention of the cost of an individual rite. 71  J. T. C. Liu, ‘The Sung emperors and the Ming-t’ang or Hall of Enlightenment’, Etudes Song, in memoriam Etienne Balazs série 2, Civilization I (Paris, 1973), pp. 35–56. 72  For Chang Yüeh, see E. G. Pulleybank, The background to the rebellion of An Lu-shan (London, 1955), esp. pp. 50–2. Other examples are Ts’ui Jung (643–708), cts ch. 94 pp. 2996–3000, hts ch. 114 pp. 4194–6; thy ch. 26 p. 491 (cf. tt ch. 40 pp. 3285A–C and cts ch. 22 p. 873), and D. C. Twitchett, ‘A Confucian’s view of the taxation of commerce: Ts’ui Jung’s memorial of 703’. Bulletin of the School of Oriental and African Studies xxxvi (1973), pp. 429–45; Liu Chih (d.c. 759), cts ch. 102 p. 317; hts ch. 132 p. 4524; cts ch. 25 p. 971, and ch. 27 p. 1035; and Edwin G. Pulleybank, ‘Neo-Confucianism and Neo-Legalism in T’ang intellectual life, 755–805’, in Arthur F. Wright ed., The Confucian persuasion (Stanford, California, 1960), pp. 98–9.

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

The Ritual Code of T ’ ang China

325

related to the same comprehensive Confucian view of the cosmos and the state. On the other hand, the fact that the imperial ritual programme was so firmly integrated into the general bureaucratic structure may also be seen to have reinforced two developments that in effect detracted from the religious aspect of the state ritual tradition over the T’ang: the politicisation of the tradition and the growing unofficial emphasis on the social rather than cosmological function of the rites.

T’ang Concepts of State Ritual

The T’ang emperors, despite their lofty claims of descent from the deified Lao tzu, had immediate origins in the semi-sinicised milieu of the northern dynasties in the period of disunion. The early T’ang court had been less cultured, less sophisticated in its ritual tradition, and more dominated by soldierly values than the courts of the vanquished southern dynasties. Practical and out-going attitudes also coloured the court’s outlook on the imperial ritual programme. There was thus almost no independent speculation on the theoretical aspects of the royal rituals that pre-T’ang tradition required the dynasty to perform. Rather, the emperor, high ministers and ritual officials had an intense and practical interest in their correct, and, from the mid-seventh century on, more grandiose performance. An understanding of the theoretical importance of royal rituals to the T’ang state must therefore be gleaned from the documentation produced by the controversies and discussions surrounding individual rituals. The dynasty’s need, again determined by practical as much as theoretical considerations, to impose its own identity on the scholarly tradition as a whole also resulted in the mid-seventh century in the re-affirmation of traditional theoretical positions with regard to ritual. Likewise the questions that examiners set, in T’ang times reflecting the general intellectual climate much more faithfully than did those of the enormous and inflexible operation of late imperial times, contained some indications of what current attitudes to state ritual were and how they changed. Much of early T’ang theorising on ritual involved repeating traditional notions. Despite their lack of fresh perceptions, however, seventh-century official scholars offered striking testimony to the sophistication of the Chinese ritual tradition. Their first justification for ritual was cosmological. They asserted that ritual was an integral component of the universe itself. ‘It forms the warp of heaven and earth and provides the structure of human conduct.’ ‘Its function in the

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

326

McMullen

good ordering of the universe is co-eval with heaven and earth.’ Ritual drew its elements from the physical universe and the succession of the seasons.73 The same early T’ang scholars who were skilled in the ritual tradition were also versed in the calendar and in cosmology. They therefore integrated the ritual programme with their view of the universe and with their numerological and calendrical systems. The official T’ang world-view was largely inherited, drawn up in the mid-seventh century by official scholars who culled as nearly a self consistent view as they could from the mass of conflicting sources they reviewed. This was a comprehensive world-view, and it bound the human and cosmological processes together in a hierarchy of complex and highly detailed dependent relationships. As a component of the cosmos, man was governed by the same series of elements, numbers, colours, directions, tastes, musical notes and cycle of seasons. Man’s own conduct, particularly that of the emperor, therefore affected this cosmic system.74 Cosmological events, climatic irregularities, epidemics and other catastrophes, as well as good omens in this view were relevant to the ritual programme.75 Interpretation of bad omens and of disasters might vary, and there were frequent controversies. The question, in particular, of when to schedule the major non-recurrent rituals for which the code supplied directives was seen in these cosmological terms. The frequency of requested, planned and aborted Feng and Shan rites not only on Mount T’ai but also on Mount Sung and Mount Hua indicates the sense of uncertainty and dependence on cosmologically interpreted factors that might surround the principal rituals in the first division of the code.76 The whole programme 73   Li chi chu shu, preface, 1a. 74  The leading official scholar of the early T’ang who combined calendrical, numerological and ritual expertise was K’ung Ying-ta (547–648), held to be the 32nd generation descendant of Confucius. He was director of the first commission that worked on the Wu ching cheng i (True meaning of the Five canons), completed in 653; he also worked on the ritual code of 637. For key statements on the shape of the universe, see Sui shu ch. 19 p. 505 and Shang shu chu shu ch. 3 p. 3b, in both of which K’ung Ying-ta was involved. See also C. Cullen, ‘A Chinese Eratosthenes of the flat earth’, Bulletin of the School of Oriental and African Studies xxxix (1976), pp. 107–9. 75  For the classification of good omens and their processing by the board of rites, see ttlt ch. 4 pp. 15a–21a. Also, Edward H. Schafer, ‘The auspices of T’ang’, Journal of the American Oriental Society lxxxiii (1963), pp. 197–225, esp. pp. 199–202; also Lung chin feng sui p’an ch. 2 pp. 1a–2a, suggesting a problem of falsified omen reports. 76  In 641 the rites planned for 642 were abandoned because of a comet; thy ch. 7 p. 87. In 647 those planned for 648 were stopped because of irregular tides at Ch’üan-chou on the coast of modern Fukien; thy ch. 7 p. 93. In 682 rites on Mount Sung were aborted because

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

The Ritual Code of T ’ ang China

327

may also be assumed to have been subject to the highly developed tradition of divination as far as auspicious timing was concerned. This, the cosmological view of ritual, was part of the orthodox ideology in which serving officials were schooled. The great majority of them publicly adhered to it, out of deference to imperial power and from individual ambition. Traditional theory, however, also permitted a more social and less cosmological emphasis in approaching the rites. This did not necessarily conflict with the cosmological view although it might sometimes coexist uneasily with it. For there was throughout the T’ang a tradition, in admonition to the emperor and in unofficial writing, of marking off the human from the supernatural and cosmological processes. But whether sceptical about the role of heaven in this way or not, in discussing the role of ritual in the state or in society T’ang scholars had certain unquestioned assumptions. An essential feature of their view of the human world was that it formed a hierarchy, a social pyramid, the strata of which were to be regulated by rules of extraordinary precision. These rules affected far more than the imperial ritual programme; but the ritual tradition, like that of criminal law, provided a context for generalising about them. T’ang scholars, reiterating a view that originated in the canons emphasised that distinctions of status were an essential feature of ritual.77 In the state ritual programme itself and in the code, their concern was above all with the emperor, then with the imperial family and with the court and high ranking officials. The lower levels of officialdom and provincial officials were mentioned only to stipulate that ritual provisions were reduced in grandeur with descending rank on the mandarin scale and in the empire-wide administrative hierarchy. In speaking more generally of the social function of ritual, seventh- and early eighth-century official scholars repeated the traditional view that ritual exercised a restraint on the social hierarchy and on the unruly appetites of man. It provided ‘defensive dykes for the control of great floods’, or ‘the bit and spurs to ride a mettlesome horse’.78 Ritual thus provided a means to control man’s potential for socially disruptive behaviour. It functioned through the

of famine, barbarian threat and the emperor’s own illness; thy ch. 7 pp. 101–2. In 751 a fire at a temple precinct caused the rites on mount Hua to be abandoned at a late state; thy ch. 8 p. 138. In 725 the secretary of the board of war questioned whether it was appropriate for military manoeuvres to take place before the great celebration of that year; cts ch. 194A p. 5175. 77  E.g. Li chi chu shu ch. 50 p. 3b; cts ch. 21 p. 815–00, ch. 86 p. 2828. Also, Ch’ü T’ung-tsu, Law and society in traditional China, pp. 226–41. 78   Li chi chu shu, preface, 1a.

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

328

McMullen

prescribed ceremony, which appealed to the affective side of man’s nature, but channelled his emotions to make them socially and morally constructive. This notion of ritual as a positive social value also involved understanding the concept in the broadest possible sense. For medieval scholars extended this ideal to all administrative operations. They used a traditional dichotomy, mentioned by Confucius himself, that opposed administration by ‘ritual’ to government by ‘punishments’. Ritual government was thought of as voluntary and positive. Punishments involved coercion; and the use of the criminal law, and still more of military force, was an indication of failure. This distinction was brought up in many different contexts, including those of the major cosmically interpreted rituals themselves. For punishments, though primarily a social and political category, were also held to have cosmological implications. The greatest of the non-recurrent rites, the Feng and Shan, should therefore only be performed when ‘punishments were not being applied’. Ritual occasions also enabled the emperor to affirm the benevolence of the cosmic order and of his own role in mediating between it and the human hierarchy. The great ritual celebrations were customarily followed by acts of grace (she). In these, the emperor might extend pardon to criminals in all categories of wrong-doing, although those guilty of the most serious crimes tended merely to have their sentences reduced.79 There are grounds for seeing in this frequently cited opposition between ritual and punishments one of the main themes in medieval Confucian ideology. Early T’ang Confucianism did not have, as neo-Confucianism was to develop, a generally accepted, central and highly analytical belief in the innate goodness of human nature. But its often repeated affirmation of the priority of ritual over punishments was a comparable expression of optimism over government and society.80 This assertion in official and unofficial writing of the primacy of ritual was moreover characteristically medieval in another sense. For it expressed medieval concern for the state and for government, rather 79  D. C. Lau tr., Confucius; the Analects (Harmondsworth, 1979), p. 63. thy ch. 8 p. 108; Brian E. McKnight, The quality of mercy, (Honolulu, 1981), pp. 64–5. 80  The priority of ritual over punishments is asserted in e.g. Sui shu ch. 25 p. 695, T’ang lü shu-i (Commentarial discussions on the criminal laws of the T’ang) (T’sung-shu chi-ch’eng edn) preface p. 2, in the context of criminal law; Sui shu ch. 74 p. 1691; cts ch. 90 pp. 2913– 14, in that of general administration; Teng-k’o chi k’ao (An enquiry into the record of successful examination candidates (Nan-ching shu-yüan ts’ung-shu, edn) ch. 1 p. 9a; T’ang ta chao-ling chi ch. 106 p. 542; Po Hsiang-shan chi (Collected works of Po Chü-i (772–846)) (Kuo-hsüeh chi-pen ts’ung shu edn) vol. 7 ch. 47 pp. 72–3, in examination questions; Han Ch’ang-li chi vol. 7 ch. 5 p. 84 in the context of education. See also Ch’ü T’ung-tsu, Law and society in traditional China, pp. 267–79.

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

The Ritual Code of T ’ ang China

329

than for interior or psychological problems relating to the individual. In this age, when the Confucian intellectual outlook was much more court-centred than was later to be the case, it was to be expected that scholars would locate the highest good in a largely court managed activity rather than in the individual considered more universally. The view of the rites as having a social function was also given a more downto-earth emphasis. Emperors and ritual scholars alike saw in certain rituals a valuable and practical exemplary function. This is clearly true of the ploughing rite that might follow the offering to the paragon emperor Shen Nung, for emperors considered it a means of encouraging good husbandry throughout the empire. In performing it, moreover, they sometimes emphasised its exemplary value by setting aside the status specific aspect of the rite. Thus in 667 Kao tsung, the third T’ang emperor, rejected a decorated ploughshare, stating, ‘agricultural equipment is for the use of the farmer and it depends on its functionality. How can it set value on decoration?’ In the same performance, Kao tsung also deliberately exceeded the number of furrows required by canonical authority of him as emperor, in order to emphasise the importance of husbandry. In the ploughing rite of 735, Hsüan tsung also went beyond the traditional stipulations for the same reason.81 Another, though rather seldom performed, ritual that was seen as having practical aims was the ‘rehearsal at war’ in the third division of the code.82 Again, emphasis on social or even military function did not necessarily conflict with the cosmological perspective on rituals. Both the ploughing and the military rehearsal rites were also discussed in cosmological terms, at least as far as their place in the annual cycle was concerned.83 But even when they reiterated high ideals about the superiority of government by ritual or endorsed the suasive function of the rites, Confucian scholars also conceded that the constraint of criminal law and the coercion of military force were necessary. The appropriate balance that the dynasty should strike between government by ritual and government by punishments was a

81   t hy ch. 10B p. 244; cts ch. 24 p. 913. Cf. also cts ch. 24 pp. 913–14, for a similar move by Su tsung in 759. 82   t hy ch. 26 pp. 501–4; cts ch. 8 p. 171; Chen-kuan cheng yao (Essentials of the good government of the Chen-kuan period) edn prepared by Harada Taneshige and published as Jōkan seiyō teihon (Tokyo, 1962), ch. 9, p. 266. Neither cts ch. 21–2 nor tt ch. 76–8 documents this ritual; cf. below at n. 103. 83   t t ch. 46 p. 236C and thy ch. 10B p. 243, ch. 10B p. 347, ch. 26 pp. 502–3 and Li chi chu shu ch. 14 p. 14b; thy ch. 26 pp. 403–4.

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

330

McMullen

recurrent examination question in the seventh and eighth centuries.84 More­ over the ritual programme and the main ritual precincts were protected by some of the severest articles in the codified criminal law of the dynasty.85 In the ritual code itself also, the preliminary announcement that a rite was to be conducted included the statement, delivered in the department of affairs of state to the participating officials that ‘each is to uphold his office; if they do not provide their services, the state has unvarying punishments …’ Belief in the social value of state ritual nonetheless gave intellectual dignity and conviction to the approach of scholars to the great rituals in the code. Here the Confucian tradition demonstrated an ability to contain different intellectual emphases, religious and social or ethical, that was one of the reasons for its durability. Emphasis on the social function of rites is apparent in one or two of the controversies of the reign in which the code was promulgated.86 Later, however, in the post-rebellion period, it combined with a new scepticism about the role of ‘heaven’, and became the predominant emphasis among those scholars who reassessed the function of state rituals. A final and central feature of the official view of state rituals in the eighth century was that the programme it involved, though ultimately sanctioned by the ‘inerasable authority’ of the Confucian canons, was far from being unchanging. Again, this view of ritual as an evolving tradition was amply outlined in canonical literature, by Confucius himself,87 and in the exegetical tradition. The idea of historical change in ritual was basic to the activity of revising and re-drafting rituals, in which so many eighth-century scholars were engaged. Early T’ang official scholars had documented the evolution that had taken place in individual rituals over the Chin dynasty (ad 265–420) and the period of disunion. The changes that occurred then and in the Sui and early T’ang may prove to be small when compared to the changes between the Former and Later Han. Yet the admissibility of ongoing change was a vital aspect of the medieval tradition. The T’ang emperors T’ai tsung, Chung tsung and Hsüan tsung all committed themselves explicitly to a view of state ritual as evolving. When T’ai tsung rejected the suggestion that the ploughing rite should take place in the southern rather than the eastern suburb and endorsed the sanction of the Book of 84  See above, n. 80 and n. 87 below. 85  In the dynastic criminal code, of the most serious crimes, the ‘ten abominations’, the second, ‘plotting great sedition’ and the sixth ‘great irreverence’, had specific reference to ritual precincts and equipment; see The T’ang code. tr. with an introduction by Wallace Johnson (Princeton, 1979), i, pp. 17–23, 63–5, 69–70. 86   c ts ch. 25 pp. 965–73: thy ch. 37 pp. 680–4; cts ch. 27 pp. 1032–6. 87  D. C. Lau, Confucius: the Analects (see n. 79), p. 66.

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

The Ritual Code of T ’ ang China

331

documents (Shang shu) rather than that of the Record of ritual which conflicted with it, he stated, ‘Ritual follows the feelings of man; what question is there of it being unvarying?’88 Likewise the great majority in the scholarly community, though always apt to adduce canonical sanction in re-drafting rituals, supported the idea of evolution.89 They advocated change not least because as individuals they often stood to gain by identifying themselves with the innovation that they knew the sovereign wanted. Scholars had therefore to reconcile the bookish element in the tradition, their instinct to condemn what was ‘not ancient’ in ritual and their ideal of a ‘clear text’ in the canons or a ‘canonical basis’ as a sanction for particular rites on the one hand with the expanding resources of the dynasty and the ambition of its sovereigns on the other. The result was the casuistry that is so apparent in records of ritual discussions.90 The pragmatic attitude of the T’ang scholarly community to the ritual programme combined with the recognition of evolution to make T’ang attitudes to individual rites strikingly flexible. A ceremony that might have lapsed for many centuries might be revived for reasons that seem little more than fortuitous. A case in point is a local ritual to the god of the earth associated with the great emperor Wu ti of the Former Han (reigned 140–87 bc) at Fen-yin in modern Shansi. In 722 Hsüan tsung passed through Fen-yin and, at Chang Yüeh’s suggestion, revived it. From a note in the Comprehensive compendium of 801, it seems likely that he commandeered and adapted an existing temple for the purpose.91 Another example is a cult to the ‘precious spirits of the nine palaces (chiu kung kuei shen)’, the justification for which was cosmological. This cult had originally been established in ad 134, but had long since lapsed. In 744

88   c ts ch. 24 p. 912; hts ch. 199 p. 5670; T’ang ta chao-ling chi ch. 74 p. 416; cts ch. 25 p. 9531; Ch’üan T’ang shih ch. 3 p. 35. 89   c ts ch. 26 pp. 1024–6 and 1030; tt ch. 53 p. 304C; cts ch. 102 p. 3172; thy ch. 12 p. 298; cf. Teng k’o chi k’ao ch. 5 pp. 10a–b and Po Hsiang-shan chi vol. 7 ch. 48 pp. 79–80. 90  For condemnation of what was ‘not ancient’, see Ta T’ang chiao-ssu lu ch. 5 pp. 3b–4a, tt ch. 70 pp. a385A–386A, thy ch. 12 p. 249. For the need for a ‘clear text (ming wen)’ thy ch. 11 p. 273 (cts ch. 22 p. 853 reads only ‘without a text’); cts ch. 21 p. 826, ch. 20 p. 980, 27 p. 1036. For the desirability of a ‘canonical basis (ching chu)’, cts ch. 25 pp. 953 and 968; thy ch. 6 p. 69. 91   Shih chi (Historical records) (Peking, 1959), ch. 12 pp. 464–5 and ch. 28 p. 1392; tt ch. 45 pp. 260B–C; ttlt ch. 4 p. 34b. cts ch. 8 p. 185, ch. 24 p. 928, ch. 97 p. 3054 and tctc ch. 212 p. 6755, 723.3. This cult was mentioned in a decree examination answer of 726; see Teng k’o chi k’ao ch. 8 p. 20a.

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

332

McMullen

Hsüan tsung, responding to a memorial from a magician, re-established it and it lasted well into the following century.92 The divinities or correlatives that received sacrifices in well-established rituals might also be changed as the result of argument. Thus in the case of the ritual for husbandry that preceded the ploughing rite, there had been from 685 to 705 different positions with regard to which gods were to be the recipients of sacrifices. From 705, the gods of the soil and of grain had been prescribed, with Kou-lung shih, a divinity associated with the earth as correlative. But in 731, a year before the completion of the code, the paragon emperor and first teacher of husbandry, Shen Nung, was substituted, with the Hou Chi as correlative. Accordingly it was to these two exemplars that the code of 732 prescribed offering. The arguments involved in this change of practice imply that for T’ang ritual scholars it was the context provided by existing altars and a ploughing site and by a place in the ritual calendar that provided premises for the debates. The problem of which divinities might be involved resulted from these practical administrative considerations. This flexibility over the selection of cult figures was again demonstrated in 735, three years after the promulgation of the code. The ploughing rite was redrafted and yet another divinity, Chumang, a god associated with spring and the element wood, was substituted for Hou Chi. In 759, however, Hou Chi was restored as correlative.93 Belief in change and adaptation was, therefore, an inbuilt feature of the state ritual tradition. In the course of the eighth century this feature was to combine with unofficially expressed scepticism about the cosmological aspects of state ritual, and was indirectly to reinforce the marked shift that took place towards understanding the rites principally in terms of their social and moral function.

Political Dimensions

A certain blandness in the traditional justification for the state ritual programme that T’ang scholars reiterated, and the fact that their ritual institutions were not original to the dynasty, should not give the impression that eighthcentury ritual practice was uneventful or placid. On the contrary, accessible as ritual issues were to the scholarly bureaucracy at large, they elicited fierce competition. They were also fought over by other competing interests, religious, political and even military. The political vitality of the early and ­mid-eighth 92   t hy ch. 10B pp. 256–60; Ta T’ang chiao-ssu lu ch. 6 pp. 1a–6b; cts ch. 24 pp. 929–34. 93   Ta T’ang chiao-ssu lu ch. 10 pp. 1a–b, summary only; tt ch. 45 pp. 262B–C: thy ch. 10B p. 245, ch. 42 pp. 421–4; cts ch. 24 pp. 913–14.

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

The Ritual Code of T ’ ang China

333

century was thus expressed in the ritual tradition, and this political dimension added to the importance of imperial rituals in the T’ang state. A major source of tension for the Confucian ritual tradition concerned the relation of the rites contained in the code, for Confucians the orthodox programme, with the other principal belief systems and the wealth of lesser cults that existed. The most obvious threat to the regular implementation of the rituals in the code, particularly those in its first division, came from the transcendental teachings of Buddhism and Taoism. The Confucian rituals as they are prescribed in the code, with their laconic prayer texts, can have done little to satisfy deeper religious instincts. At best they might move the individual to reverence for the unseen cosmological powers or for past exemplars, or to pride in the dynasty. Or they might stir the emperor to compassion for deceased members of his family, as Hsüan tsung was said to have been stirred when making offering in the ancestral temple to Jui tsung, his father in 717.94 The Buddhist and Taoist traditions on the other hand provided copiously for the religious imagination, and in particular promised abundantly for the afterlife. Successive T’ang sovereigns encouraged the idea of the fundamental compatability of Confucianism, Buddhism and Taoism. But, in fact, representatives of the three teachings at court competed against each other for the patronage of the emperor, the imperial family and the court. Individual clerics urged the requirements of their own faiths, sometimes openly deriding Confucian religious traditions.95 They were often successful, for despite the protest of Confucian orientated officials, under Hsüan tsung’s immediate predecessors, resources that might have gone into the upkeep of ritual precincts or buildings were used on a lavish scale for Buddhist and Taoist temples.96 In the second half of Hsüan tsung’s reign, the emperor’s promotion of Taoism and establishment of Taoist schools and temples, to be managed by the court of the imperial family, again must have been to the detriment of the programme prescribed by the Confucian ritual code. Emperors were also liable to take up and develop their own religious interests, perhaps new or dubiously canonical cults, on the margins of orthodoxy. They were often supported in this by the ambitious or opportunistic, from within the bureaucracy or outside it. Such cults might also result in ritual obligations that competed with the orthodox programme.

94   c ts ch. 25 p. 952; cf. cts ch. 25 p. 973. 95  Lo Hsiang-lin, ‘T’ang tai san chiao chiang-hua k’ao (An examination of the debates between the three teachings in the T’ang age)’, Journal of Oriental Studies i (1954), pp. 85–97; cts ch. 192 p. 4128. 96   c ts ch. 89 pp. 2893–4, ch. 101 pp. 3155–61, ch. 88 pp. 2870–1.

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

334

McMullen

In this situation, the almost complete precision with which the compilers of the K’ai-yüan code resisted the intrusion of rituals from the other teachings and from other cults is both striking and historically important. Only one of the cults introduced in the decades prior to the code and included in it was clearly unsanctioned by the canons or the exegetical tradition. This, the ‘cult of the five dragons’, was promoted by Hsüan tsung himself, and celebrated the favour he believed the supernatural powers had shown him as crown prince.97 The code’s compilers were largely successful also in excluding the syncretic approach towards Taoism and Confucianism that Hsüan tsung developed in the second, fervently Taoist half of his reign. The enforcement of rigid classificatory lines like this was, of course, a feature of the medieval Chinese official approach to any scholarly activity; but the compilers of the code may also have been helped in maintaining its discrete Confucian character by a bureaucratic factor, by the institutional framework in which they operated. For the main institution concerned with implementing the code, the court of sacrifices, had a long tradition of providing administrative support for Confucian rituals. There were, at least at the time when the code was compiled, separate institutional provisions for the other teachings. Only the board of sacrifices (tz’u pu), a department of the board of rites, was administratively concerned with all three teachings. And it is not clear how much, in practice, the officials of the board of sacrifices were active in the imperial ritual programme the code prescribed. Even though the compilers of the code formed an ad hoc commission that worked not in the court of sacrifices but in Hsüan tsung’s own advisory college, the Chi-hsien yüan, they observed the lines of demarcation that these separate institutions provided. The code’s compilers also set aside their own religious beliefs, for it is very probable that as individuals and in private they were adherents of one or other of the transcendent faiths. For them to have promoted the rival teaching of Buddhism, or indeed to have expressed their own personal belief in Buddhism, in the context of the code, however, would have been both to have disregarded traditional lines of demarcation and to have offered help to forces within the bureaucracy or outside it with whom they were in competition for the emperor’s interest and patronage. The lines of demarcation between the Confucian outlook and Taoism were never so clear, for Confucian orientated scholar officials were also scholars of classical Taoist texts. But in their official capacities they generally stopped short of promoting religious Taoism, and they certainly 97   t tkyl ch. 51 pp. 5b–8a; thy ch. 22 pp. 433–4; Ta T’ang chiao-ssu lu ch. 7 pp. 6b–9b; Robert des Rotours, ‘Le culte des cinq dragons sous la dynastie des T’ang (618–907)’, Mélanges de sinologie offerts à Monsieur Paul Demiéville (Paris, 1966), pp. 261–80.

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

The Ritual Code of T ’ ang China

335

did not permit it to intrude into the imperial ritual code. Again this was both because they were observing traditional distinctions and institutional lines of demarcation and because they did not wish to lend support to rivals of whose attitudes and motives they were suspicious. Hsüan tsung’s promotion of Taoism through the court of the imperial family came after the promulgation of the code. But there is relatively little evidence that it intruded into the imperial ritual programme embodied in the code.98 A scarcely less important expression of the broad sectional interest of the compilers of the code concerned their attitude to the military element in it, to the ‘army rituals’ and related ceremonies. The scholars who staffed the academic agencies stood above all for civil virtues. The military element in the T’ang state, on the other hand, was enormous. In the mid-K’ai-yüan period, the Old T’ang history records, the dynasty could call on at least 120,000 men at the capital, who served 20,000 at a time for two months of the year. The armies at the frontiers numbered in excess of 400,000. The generals or commissioners who controlled these armies were often drawn from the general service of the bureaucracy.99 Chang Yüeh himself had military experience in three high-ranking frontier posts. But the scholars who staffed the Chi-hsien yüan and those who compiled the ritual code were typically metropolitan academic officials whose careers had not involved military appointments. They were ‘book men [shu sheng]’ and ‘Confucians [ju sheng]’. Highly selected and identified as the intellectual and scholarly élite, they had great confidence in their own value system. They and many Confucian orientated general service officials, like Chang Yüeh himself, had a political interest in controlling the power of the armies. They were also strongly opposed to imperial military adventurism beyond the frontiers, for they saw in this historically one of the most disruptive forces of internal political and economic stability that unwise imperial conduct might unleash. Though they might occasionally come forward with ideas on strategy, they did not generally have good relations with

98  One indication that most Confucian orientated scholars may have been willing to accept the classical Taoist element in the Taoist programme Hsüan tsung promoted may be seen in the fact that aspects of the Taoist cult to Lao tzu were documented in some official and unofficial surveys of the Confucian ritual programme, eg. cts ch. 24 pp. 925–8; but especially tt ch. 53 p. 305B. 99   c ts ch. 97 p. 3049. cf. E. G. Pulleybank, The background to the rebellion of An Lu-shan, pp. 65–6, and ‘The An Lu-shan rebellion and the origins of chronic militarism in late T’ang China’, in John Curtis Perry and Bardwell L. Smith (eds.), Essays on T’ang society (Leiden, 1976), pp. 33–60.

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

336

McMullen

­ rofessional soldiers or even with some of the officials in the general service p who had practical military experience.100 The fact that the drafting of the state ritual programme was dominated by these Confucian orientated academic officials, with their civil ideals, may explain why the military element in the state was as little represented in the code as it was. The Confucian canons themselves supported two notions as far as the proportion of military and civil in the state was concerned. One was that of a symmetry or balance between the military (wu) and the civil (wen). The second was that of three seasons given to agriculture against one for war. Neither ideal was fully realised in the code, though scholars brought up both in the course of ritual debates. Thus the ‘army rituals’ divison of the code was the second shortest, exceeded in brevity only by the section prescribing ceremonies for foreigners. Moreover although most of the army rituals involved offerings at altar sites and in the ancestral temple, the preliminary abstinences the code prescribed for them were much shorter than for rituals involving the same precincts in the first, the ‘propitious rituals’ division of the code. The hostility or at best indifference of the Confucian orientated scholar community generally towards the military is also seen in the relative failure of the military cult to Ch’i T’ai kung. This had been reconstituted in 731 on a scale to balance the expanding civil cult to Confucius. Some scholars at least believed that a cult for T’ai kung had the sanction of the Monthly commands section of the Record or ritual.101 But in the aftermath of the An Lu-shan rebellion it provoked opposition. It was re-established and expanded in 760, lapsed, to be revived and expanded yet again in the early and mid-780s. Surviving memorials concerning the cult at this later stage show that opposition, expressed in terms of the priority of ‘ritual’ over ‘expedient action [ch’üan]’, was by now highly articulate.102 Another ritual, the ‘rehearsal at war’, scheduled for the winter and designed to express the canonical ideal of three seasons for agriculture and one for war, although it had its advocates, the Emperors T’ai tsung and Hsüan tsung and the Empress Wu among them, is poorly documented. It too lapsed after the An Lu-shan rebellion.103 The conclusion must be that the 100  For this distinction between official scholar and general service official see e.g. cts ch. 102 pp. 3164 and 3177. Cf. also cts ch. 157 p. 4158. For the scholars’ contempt of professional soldiers, tctc ch. 214 p. 6822, 736.16. The T’ang tradition of admonition against military adventurism is represented in tt ch. 200 pp. 1085B–1087B and Chen-kuan cheng-yao ch. 9 pp. 271–81. 101   Ta T’ang chiao-ssu lu ch. 10 p. 14b. 102   Ta T’ang chiao-ssu lu ch. 10 pp. 12b–15b. 103   Chen-kuan cheng-yao ch. 9 p. 266, thy ch. 26 pp. 501–4, cts ch. 89 p. 2900.

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

The Ritual Code of T ’ ang China

337

scholars who drafted and managed the ritual programme were opposed for reasons of sectional interest to giving more than minimal recognition to the military aspect of the state. Official scholars were, however, less successful, and certainly less united in resisting another of the forces that helped shape the ritual programme the code embodied. The growing wealth of the dynasty in the seventh and early eighth centuries fostered in at least the longer reigning sovereigns the ambition to celebrate the great traditional rituals in ever more grandiose ways. And again their motives were less than straightforward. By performing her own versions of the greatest non-recurrent rituals, the usurping Empress Wu claimed the support of the cosmological and supernatural powers and asserted the legitimacy of her own regime.104 Hsüan tsung’s celebration of the southern suburban rite in 723, although the emperor described it in self-deprecating language, was intended to mark twelve years of stable and successful rule.105 The increase in the scale of the major imperial rituals from the seventh century to the mid-eighth, however, ran through the reigns of successive sovereigns and is sometimes clearly recorded in the official histories as doing so. Describing a grandiose offering to the river Lo made by the Empress Wu in 688, the Old T’ang history relates, ‘The imperial sons all followed; all officials, civil and military, the headmen of the barbarian (states), all stood according to their positions. Rare birds and beasts were all deployed before the altar. Since the start of the T’ang, decorated items and processional equipment has never been so resplendent’.106 The celebrated Hall of Light (Ming t’ang) that the Empress Wu constructed at Lo-yang in 688, and then, after a disastrous fire, rebuilt on the same site was envisaged as a central ritual precinct. The edifice was on a scale and used materials and furnishings far more lavish than those envisaged by the more cautious ritual scholars who had discussed it in the reign of T’ai tsung. Descriptions of the Hall of Light imply that it surpassed all previous buildings in the use of architectural techniques and resources.107 Under Hsüan tsung, expansion was also the theme in the dynastic ancestral temple. Canonical authority required that the temple contain not more than seven shrines, representing seven generations of deceased ancestors of the imperial

104   c ts ch. 6 p. 119– ch. 24 p. 922. cf. tctc ch. 204 p. 6454, 688.14; R. W. L. Guisso, Wu Tse-t’ien and the politics of legitimation in T’ang China (Western Washington University, 1978), p. 65. 105   c ts ch. 8 p. 186, T’ang la chao-ling chi ch. 68 pp. 380–1. 106   c ts ch. 24 p. 922; cts ch. 6 p. 119. 107   t t ch. 44 pp. 251A–45B; thy ch. 11–12 pp. 271–92; cts ch. 22 pp. 849–79; R. W. L. Guisso, Wu Tsu-t’ien, pp. 35, 46, 65, 129.

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

338

McMullen

line. In 722, however, it was expanded to nine.108 In 734, after a debate in which at least five scholars from outside the ritual agencies, and one from the board of rites, disagreed with the president of the court of sacrifices, the provisioning of offerings in the ancestral temple was also expanded.109 The amounts involved in imperial largesse to participants after major ritual celebrations and the range of the pardons included in acts of grace also increased in the course of the first half of the dynasty. The celebration of the Feng and Shang rites on Mount T’ai by Hsüan tsung in late 725 was particularly lavish. It was called at the time ‘an event in a thousand years’. The ‘whole state’ made the journey east to the mountain, and the procession included representatives of foreign states, recorded in more detail than on any previous occasion.110 The performance thus invited favourable comparison not only with Kao tsung’s celebration of 666, but also with those by past emperors extending back to the great Wu ti of the Han in 110 bc. Under Hsüan tsung expansion of the Confucian shrine, too reached a high point: Confucius and his immediate disciples were represented by statues, while pictures of the larger circle of his followers, numbering over seventy, were drawn on the temple walls. The dynasty also honoured all these figures with grandiose posthumous titles.111 This expansion of the state ritual programme and of its precincts was sustained till towards the end of Hsüan tsung’s reign. Some scholars asked for restraint, and their submissions were carefully preserved by official scholars of the post-rebellion period. But by and large only after the rebellion of An Lushan had drastically reduced the resources and political power of the dynasty did scholars more generally condemn the period as one in which the classical virtue of austerity had been set aside. By late eighth-century times, however, this post-rebellion condemnation of the excesses of the second half of Hsüan tsung’s reign became an aspect of a radical reassessment by certain intellectuals of the function of imperial rituals in the state. The flourishing ritual tradition of the first half of the eighth century provided a medium for competition among officials and a means to gain access to officialdom for those as yet commoners. For the dynasty exploited in several distinct ways the enormous prestige of the official career in its service and that 108   t t ch. 47 p. 270C; thy ch. 12 p. 298; cts ch. 8 pp. 183 and 185; tctc ch. 212 pp. 6750, 722.10 and 6756, 723.11. 109   c ts ch. 25 pp. 969–72; thy ch. 17 pp. 349–52. 110   c ts ch. 99 p. 3098; T’ang shih chi shih (Record of events relating to T’ang poems) (Peking, 1965), ch. 3 p. 35; cts ch. 194A p. 5175. 111   Ta T’ang chiao-ssu lu ch. 10 pp. 6b–8b; tt ch. 53 pp. 306A–C; thy ch. 35 pp. 637–8; cts ch. 24 pp. 919–21.

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

The Ritual Code of T ’ ang China

339

of the great state rituals. The major non-recurrent ritual events, held in celebration of propitious conditions, good harvests and general peace, became focal points for support for the dynasty. They were used to hold special examinations for recruiting to the public service, known as ‘great ritual candidatures (ta li chü)’.112 The Feng and Shan rites were thus accompanied by special decree examinations or intakes into the bureaucracy. These were sometimes announced in anticipation of performances that in the event did not take place. This happened in 647, while examinations were held in connection with the successful performances of 666 and 725.113 The major rituals, or aspects of them, were also made the topics for composition in the annually held chin-shih examination, the most prestigious of all examinations. The requirement here was for a highly wrought and high-flown panegyric style in praise of the dynasty. In 713, for example, candidates were required to compose a prose-poem on the ploughing rite. They would have been aware that they were treating a topic on which Kao tsung, the third T’ang emperor, had composed a prose-poem, and for which the great scholar-emperor Wu ti of the Liang (r. ad 502–550) had written a poem.114 The dynasty also encouraged open and competitive discussion of specific ritual issues. If a scholar risked imperial irritation by offering criticism in terms of unwelcome Confucian austerity, he was normally protected from drastic punishment by the prestige attached to the tradition of frank admonition. But in fact the emperor could assume that all contributions to the resulting debates would fall within acceptable limits. No radical dissent would be voiced by a scholar community the lower ranks of which were dominated by the hope of promotion and the higher levels of which might be, as Chang Yüeh was, reluctant to allow change. Irregularities or sharp departures from conventional practice were, therefore, more likely to come at the wish of emperors themselves or their sycophants than from the regular academic establishment. Typically a ritual problem was discussed initially on imperial order by the ‘ritual officials’. But, as in the case of the debate on the provisioning of the imperial ancestral rites in 734, others whose posts had nothing to do with the ritual agencies might participate. Even a commoner might make a 112   c ts ch. 100 p. 3116. 113   T’ang ta chao-ling chi ch. 66 pp. 368–9; Teng k’o chi k’ao ch. 2 pp. 12b–14a; ch. 7 pp. 16a–17a. A special intake was also announced after the ploughing rite of 735; see T’ang ta chao-ling chi ch. 74 p. 415 and Teng k’o chi k’ao ch. 8 pp. 11a–b. 114   Teng-k’o chi k’ao ch. 5 pp. 12b–14b; thy ch. 10B p. 245; I wen lei-chü (Literary encyclopaedia) (Peking, 1965), ch. 39 pp. 702–5; Ch’u hsüeh chi (Record for early learning) (Peking, 1962), ch. 14 pp. 339–41.

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

340

McMullen

submission.115 Difficult decisions, however, were sometimes taken, after general discussion, by the chief ministers.116 Here again was an indication that the ritual tradition, though represented by specialist bodies, was the concern of the bureaucracy as a whole. The state ritual tradition, therefore, was a vehicle not only for the display of scholarly erudition and judgement by ritual specialists, general service officials and aspirants to office; it was also at the same time a camouflage for general political ambition. Not surprisingly it also became a medium for factional activity. It is no coincidence that in the course of the ‘event in a thousand years’ on Mount T’ai in 725, the scholarly chief minister Chang Yüeh, by allowing very few officials to accompany the emperor up the mountain, and by showing anti-military bias in the distribution of largesse that followed, first incurred the enmity of his aristocratic rival for power Yü-wen Jung. In doing so he helped draw for the first time the deep factional lines that were to dominate political life for the remainder of Hsüan tsung’s reign.117 But this celebrated rite on the eastern peak provides one instance only of a ritual event acquiring a factional dimension or calling forth a response from the scholarly bureaucracy at large. Indeed, the K’ai-yüan code itself, like other medieval official scholarly compilations, may be understood on several levels. A set of highly detailed, practical directives, it was also the result of the vigorous sectional and political forces that gave it the shape it has.

The Later History of the Code

From its promulgation in 732 until the end of the dynasty, the code held an unchallenged official position. Its high standing is to be seen in the frequency with which, from about 780, extant sources show scholars appealing to it or citing it in debates on ritual issues.118 On both official and unofficial initiative, a number of supplements to the code were produced; these too paid homage to it either in their titles or in the memorials with which their compilers presented them.119 Scholars thus never compiled another complete code under T’ang 115   t t ch. 47 p. 271A; thy ch. 17 pp. 353–5; cts ch. 24 pp. 952–3. 116   c ts ch. 25 pp. 948–9; thy ch. 17 pp. 349–52; cts ch. 25 pp. 969–72, ch. 27 p. 1030. 117   c ts ch. 97 pp. 3054–5; E. G. Pulleybank, The background to the rebellion of An Lu-shan, p. 50. 118  E.g. cts ch. 21 pp. 843–4; tt ch. 41 p. 233C; Teng-k’o chi k’ao ch. 16 p. 2b. 119   t hy ch. 83 pp. 1529–30 and cts ch. 150 p. 4046; Liu Ho-tung chi ch. 4 p. 44; cts ch. 171 p. 4454, ch. 157 p. 4155.

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

The Ritual Code of T ’ ang China

341

rule. In 789, moreover, the dynasty established a regular annual examination in the code. This made it, remarkably, the only T’ang official compilation to be accepted as an official syllabus in this way. The examination lacked the standing of the chin-shih, the most prestigious of the regular examinations, and is poorly documented in consequence. But it continued into the Five Dynasties period (906–960). No questions or answers survive, though the names of a handful of successful candidates are known.120 The continued high standing of the code, however, did not mean that the full range of its directives was in any sense mandatory in the imperial ritual programme. From almost immediately after the code’s promulgation, ritual officials were again involved, sometimes at the emperor’s behest, in re-drafting individual rites.121 It is thus abundantly clear that the K’ai-yüan code, like its predecessor codes drawn up under the T’ang, was never conceived as providing permanent or definitive answers for all the problems that arose, or as preventing continued evolution. Even in the period after the An Lu-shan rebellion of 755, when the political and economic power of the T’ang was drastically reduced, there were long-running ritual controversies, and continual readjustment in the ritual programme took place. The public deference of scholars towards the K’ai-yüan code in the postrebellion period is thus to be understood largely as a public expression of reverence and nostalgia for the politically prosperous, stable and expansionist era in which it had been produced. For by the closing decades of the eighth century the intellectual world recognised that the remarkable security and wealth the court had enjoyed in the reign of Hsüan tsung were beyond recovery. To scholars writing in an official context and to examination candidates, the code, like the great celebration on Mount T’ai in the winter of 725, had come to stand for the highest point in the history of the dynasty. These public affirmations of dynastic loyalty and of nostalgia were, however, accompanied by a change in the unofficial attitudes of at least an element in the scholar community towards the ritual programme. This shift may be understood in part as a natural result of what this article has attempted to establish as a central and unique combination of features in the T’ang imperial ritual tradition: the Confucian sanctioned belief system the code embodied retained ‘archaic’ characteristics. Its focus was this worldly; despite its maintenance of the plurality of divinities and cosmic agents, it had no developed sense of the transcendent and no priesthood. At the same time those who 120   t t ch. 15 p. 84B; thy ch. 76 pp. 1396–7; des Rotours, Le traité des examens, (see n. 10), pp. 148–9. 121  E.g. thy ch. 36 p. 658, ch. 20 p. 403.

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

342

McMullen

drafted and administered the rites were members of a highly structured and increasingly sophisticated bureaucracy. The interaction of these disparate features was bound in the course of time to produce a change of attitude towards the rituals. In the post-rebellion period moreover, precisely because of the deterioration in the political and economic situation, scholars were less bound by conventional or official attitudes than at any earlier period in the dynasty, and more aware of the relative and perishable nature of many of the traditions surrounding the state. In this intellectually vigorous climate, a few radical reformist thinkers, writing unofficially, came to apply to the imperial ritual programme and to the belief system it represented the same demands for practicality, austerity and effectiveness that they had applied to the range of problems they encountered in their general administrative roles. Their urgent priority was now, not the ritual celebration of unprecedented success, but the restoration of effective government, or, in the metaphor used in this period, the curing of the state’s illness. The resulting critique of imperial ritual developed the idea that the activities of heaven or of the cosmos and of man were separate. It led one or two intellectuals at the end of the eighth century to a strikingly agnostic view of the Confucian sanctioned ritual programme. Liu Tsung-yüan (773–819), though at least in his later life a committed lay Buddhist, was the most determined of these sceptics. In effect he queried the ‘inerasable authority’ of the ritual canons. For he even went so far as to criticise parts of the Monthly commands section of the Record of ritual. This was the canonical authority that most clearly integrated the cosmological view of ritual with the annual ritual programme, and it had been of particular concern to T’ang emperors and to official scholars.122 Liu Tsung-yüan charged that the functioning correlations between man and the cosmos that the Monthly commands posited were baseless, and insisted that ‘it is not necessary to regard as supernatural’ the activity of making offering to divinities or to the spirits of deceased exemplars. Rather, he maintained, rituals had a moral function, that of increasing the respect and admiration of society for good conduct and elevating the position of the emperor.123 A modified form of this agnostic or sceptical view of Confucian sanctioned ritual was probably widely accepted in the late eighth and early ninth 122   Chen-kuan cheng-yao, ch. 1 pp. 29–31; thy ch. 26 pp. 291–2, ch. 75 p. 1374, ch. 77 pp. 1410–11; cts ch. 22 pp. 867–73, ch. 24 p. 914; tt ch. 70 p. 385C; thy ch. 26 p. 492 ch. 35 p. 645, ch. 77 p. 1411; T’ang yü lin (Anthology of T’ang remarks) (Shanghai, 1957), ch. 2 pp. 48–9. 123   Liu Ho-tung chi ch. 3 pp. 52–6; ch. 16 pp. 296–7, ch. 26 pp. 432–4, ch. 31 pp. 503–4, ch. 45 p. 787.

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

The Ritual Code of T ’ ang China

343

centuries. The celebrated verse writer Po Chü-i rehearsed it in a specimen examination answer he drafted in 806. Tu Yu, the compiler of the Comprehensive compendium of 801, referred indirectly to it.124 Such an outlook might involve its advocates in re-assessing and down-grading particular rites in the code on the grounds of their extravagance or pointlessness. Liu Tsung-yüan, again the most extreme proponent of this outlook, even went so as to suggest that the Feng and Shan rites were meaningless.125 But these criticisms never resulted in rejection of the idea of a dynastic ritual code. For even in this period of radical reassessment, the most sceptical thinkers did not attack the code as a whole. Liu Tsung-yüan indicated that he did not want to see the imperial ritual programme lapse. He also expressed gratification when in 811, after a long interval, the ploughing rite was scheduled.126 Po Chü-i and Tu Yu were among those who, in an official or semi-official context, praised the K’ai-yüan code most fulsomely, and their knowledge of the state ritual tradition was not in doubt. In terms of the cultural spectrum suggested at the start of this article, these late eighth-and early ninth-century scholars had moved significantly further than their predecessors in the K’ai-yüan period from the Balinese position, in which royal rituals were accorded a place as the dominant activity of the state. Their outlook might even be thought to have approached that of nineteenth-century Britain, in which royal ceremonial was allocated a dignified and secondary rather than an efficient role in the constitution. But such an interpretation would be misleading; these figures remained firmly Confucian. They never surrendered the idea that an imperial ritual programme expressed canonically sanctioned and essential social and moral priorities and therefore performed a vital function for the state. In Liu Tsung-yüan’s words, those parts of the programme that related to ‘the surburban and ancestral temples and the hundred divinities are the transmitted records of antiquity and should not be set aside’. Despite the development of these critical attitudes towards state ritual in the second half of the T’ang, the influence of the K’ai-yüan ritual code extended far beyond the dynasty. For the re-establishment of strong dynastic power tended to be accompanied by the re-assertion of the traditional view of the emperor, the state and the cosmos. All later dynasties of any duration, therefore, followed the T’ang in producing one or more ritual codes. The Sung did so not 124   Po Hsiang-shan chi vol. 7 ch. 47 pp. 72–3, ch. 48 pp. 81–2; tt ch. 41 p. 233C. 125   Liu Ho-tung chi ch. 1 p. 18; cf. T’ang chien (Mirror for the T’ang) (Ts’ung shu chi-ch’eng edn) ch. 4 p. 29. 126   Liu Ho-tung chi ch. 43 p. 736; thy ch. 10B p. 255: hts vol. 14 ch. 359 p. 60. The performance was eventually cancelled and Liu expressed regret, Liu Ho-tung chi ch. 43 p. 533.

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

344

McMullen

only in the period 968–75,127 but again in 1111.128 Even the non-Chinese Jurchen dynasty, the Chin (1115–1234) approved a code in 1195.129 The Ming code, the Collected rituals of the great Ming (Ta Ming chi li), was completed in 1370 and re-published in 1530.130 The Manchu Ch’ing dynasty commissioned a code in 1736, which was completed in 1756.131 These Sung, Jurchen, Ming and Ch’ing codifications are merely the extant examples of a tradition that lived through until the end of the imperial era. The extant codes may be assumed to bear the same sort of relation to the many other prescriptive treatises, compendia and the background documentation produced in the dynasties concerned that the K’ai-yüan code had to the abundant documentation of the T’ang. In the same way, the imperial ritual programme of later dynasties was administered by ritual institutions that perpetuated the T’ang nomenclature. There seems, on superficial inspection, to be considerable continuity in the content of these extant later imperial ritual codes. With the exception of the Jurchen code of 1195, they are all organised in the same five divisions that the T’ang scholars had used, though later compilations placed the five divisions in a different order. Many of the rituals contained in the K’ai-yüan code recur down to Ch’ing times. The compilers of the Ch’ing code acknowledged this continuity, for in their preface they cited the K’ai-yüan code as the first extant authority of its kind. These continuities should not however preclude an understanding of the K’ai-yüan code and of the milieu in which it was produced as essentially medieval. For the great change in the Confucian tradition that has enabled intellectual historians to characterise a neo-Confucian era in Chinese thought from Sung times resulted in changed attitudes to the imperial ritual programme. The principal social trend here was towards a much larger, less court-centred and more devolved intellectual community. As a result, perspectives on man in society became more universal. Where analysis of ethical and social questions was concerned the level of generalisation became higher. Neo-Confucianism also developed a theme that had been present in early, pre-medieval 127   Sung shih (History of the Sung) (K’ai-ming shu-tien ed., Taipei, 1961) ch. 95 pp. 4731C–D; Ikeda On, ‘Dai Tō Kaigen rei no kaisetsu’, p. 827. 128   Cheng-ho wu li hsin i (New ritual directives for the five divisions of ritual of the Cheng-ho period), reprinted in Ssu-k’u ch’üan shu chen-pen, first series (Shanghai, 1934). 129   Ta Chin chi li (Collected rituals of the great Chin dynasty), reprinted in Ssu-k’u ch’üan shu chen pen, eighth series, (Taipei, 1978). 130   Ming chi li (Collected rituals of the Ming), reprinted in Ssu-k’u ch’üan shu chen-pen, eighth series, (Taipei, 1978). 131   Ch’in-ting ta Ch’ing t’ung li (Imperially commissioned comprehensive rituals of the great Ch’ing dynasty), reprinted in Ssu-k’u ch’üan shu chen pen, eighth series (Taipei, 1978).

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

345

The Ritual Code of T ’ ang China

Confucianism. Self-cultivation and contemplation, the individual’s quest for sage-hood and enlightenment now assumed major importance in the teaching of neo-Confucian masters. There was also a more agnostic spirit, and less interest in the supernatural or cosmological justification for the state. The effect of these changes on the creative mainstream of the Confucian tradition was that the great majority of individual philosophers had much more distanced attitudes as far as the detailed ceremonial and ritual traditions surrounding the apex of the state were concerned. The shift was, moreover, a general one, for it affected the official scholarly milieu as well as the now much larger unofficial intellectual world. Expressed in terms of proportion, it meant that royal rituals took up a smaller part of the total official documentation that the history office and other official academic agencies generated.132 The sense of immediacy and of involvement in state ritual by a large proportion of the intellectual élite that the T’ang sources so strikingly convey had receded. The balance between belief in cosmological and supernatural elements in government on the one hand and more sceptical and society orientated attitudes on the other, had already, in the second half of the T’ang, the age of Liu Tsung-yüan and Tu Yu, begun to change. In the neo-Confucian era it shifted still further towards the ethical and social. The K’ai-yüan ritual code, longer and more detailed than almost all its extant successors, strikes a characteristically medieval balance between traditional religious beliefs and bureaucratic attitudes. As late eighth-century scholars themselves had already suggested, it stands for an epoch in the long history of imperial China. 132  Etienne Balazs, ‘History as a guide to bureaucratic practice’ (see n. 53), pp. 135–42.

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

Wei Cheng’s Thought [esp. Regarding Government] Howard J. Wechsler As the preceding chapters may have suggested, the corpus of Wei Cheng’s extant writings is rather extensive: a considerable number of remonstrances and memorials, replies (tui) to simple queries or to questions of greater political or philosophical consequence posed by the emperor, minor literary works, poetry, biographical and topical treatises in the Liang, Ch’en, and Sui Standard Histories, and recorded dialogues at court with T’ai-tsung and fellow officials. The subject matter of these materials is quite heterogeneous. Nevertheless, they are all held together by certain unifying strands that comprise the core of Wei’s political beliefs, and we might pause here to treat them at some length. First, however, it may be useful to summarize some of the developments that had taken place in the realm of Chinese thought during the centuries immediately preceding Wei’s time.1 During the Han dynasty, classical studies had split into two broad “schools,” the so-called New Text and Old Text, although members of each of these schools by no means disagreed on all questions. The New Text School (chin-wen chia) was based on versions of the classics, some of which were allegedly set down from memory following the Ch’in burning of the books in 213 bc. The Old Text School (ku-wen chia) was based on supposedly newly discovered classical texts written in an archaic script dating from the pre-Han period. Sharp distinctions between the two schools cannot always easily be made but, generally speaking, their philosophical beliefs divided as follows. The New Text School, of which Tung Chung-shu (ca. 179–ca. 104) is the most renowned exponent, placed a special emphasis on the Spring and Autumn Annals as interpreted by the Kung Yang commentary and on the assumption that the Annals contained hidden meanings by which Confucius sought to convey his moralistic teachings. The thought of the Kung Yang Source: “Wei Cheng’s Thought,” in Howard J. Wechsler, Mirror to the Son of Heaven: Wei Cheng at the Court of T’ang T’ai-tsung, New Haven: Yale University Press, 1974, pp. 166–87. 1  The following discussion is based for the most part on P’i Hsi-jui, Ching-hsüeh li-shih [A History of Classical Learning] (1924; reprint, Hong Kong, 1961), pp. 101–219; Yang Hsiangk’uei, “T’ang-Sung shih-tai ti ching-hsüeh ssu-hsiang” [Concepts of Classical Learning in the T’ang and Sung Dynasties], Wen-shih-che [Literature, History, and Philosophy] (1958, no. 5), pp. 7–16; Tjan Tjoe Som, Po Hu T’ung, 2 vols. (Leiden, 1949), 1:95–165; and Fung Yu-lan, A History of Chinese Philosophy, 2 vols. (Princeton, 1953), 2:7–167.

© koninklijke brill nv, leiden, ���9 | doi:��.��63/9789004380158_012

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

Wei Cheng ’ s Thought [ esp. Regarding Government ]

347

commentary was heavily influenced by five elements and yin-yang cosmology and by prognostication and apocryphal texts (ch’en-wei shu) which, among other things, presumed the divinity of Confucius. The Old Text School arose essentially as a reaction to the New Text School. It favored the Tso commentary to the Annals, which was written with far less assumption of the principles of praise and blame than the New Texters saw in the Annals; purged Confucianism of much of its cosmological and supernatural beliefs; and leaned toward a more naturalistic view of the universe somewhat akin to that of Taoism. Its two best known exponents are Yang Hsiung (53 bc–ad 18) and Wang Ch’ung (27–ca. 100). Toward the end of the Later Han there occurred a short-lived syncretism of Old Text and New Text ideas in the work of Cheng Hsüan (127–200), a syncretism that nevertheless minimized the five elements and yin-yang approach of the New Text School. In the post-Han period, the New Text School sank into rapid and almost complete oblivion. The long-range trend, however, was for a continued division in Chinese thought, reinforced even further by the Period of Disunion, which created a gulf between North and South China not only in the political realm but in the philosophical as well. The school of Northern Learning (pei-hsüeh) followed for the most part the commentaries to the classics written by Cheng Hsüan and placed a heavy emphasis on such ritual texts as the Record of Rites. The school of Southern Learning (nan-hsüeh) represented a continuation of Old Text Confucianism heavily overlaid with hsüan-hsüeh, “dark learning” or Neo-Taoism, as interpreted most notably by Wang Pi (226–49). Both schools utilized the Tso commentary to the Annals but in differing interpretations. Over the years, however, whatever geographical distinction Northern and Southern Learning originally possessed gradually disappeared. What determined whether an individual subscribed to the ideas of one school over the other was personality, temperament, and philosophical and political outlook rather than, simply, geographical origin. The Sui and T’ang reunifications of China brought with them an earnest attempt to reunify Chinese thought as well. The effort to create a new philosophical synthesis was most clearly revealed in the Five Classics with Orthodox Commentaries (Wu-ching cheng-i) in 170 (variant 180) chüan, compiled by imperial order between 638 and 653 under various supervisors.2 The Orthodox Commentaries, which served as the basis for the popular ming-ching (“classics”) examinations during the early T’ang, contains commentaries to the classics from both the schools of Northern and Southern Learning and thus 2  See thy ch. 77, p. 1405, and cts 73.16b.

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

348

Wechsler

effected, at least superficially, an amalgamation of Chinese thought under imperial auspices. But this amalgam may not have been an accurate reflection of the philosophical attitudes of a majority of the seminal thinkers of the period. K’ung Ying-ta, the chief compiler of the Orthodox Commentaries and a northerner, appears to have favored scholars of the school of Southern Learning over those of their northern counterparts, and the definitive edition (ting-pen) of the Five Classics compiled early in T’ai-tsung’s reign by Yen Shih-ku (a northerner, but a descendant of the southerner Yen Chih-t’ui) also more frequently employed southern versions of the texts,3 all of which has prompted P’i Hsi-jui’s somewhat hyperbolic remark that “Following the unification of classical studies there was Southern Learning but no Northern Learning.”4 Where did Wei Cheng stand on the matter of Northern vs. Southern Learning? This is a difficult question to answer, given our still deficient understanding of the philosophical and, especially, the political content of the two schools, and also Wei Cheng’s own distaste for pursuing questions of ideology divorced from their practical contexts. One of the rare instances in which Wei does comment on the problem is in the “Confucian scholars” section of the Sui History, where he very briefly traces differing interpretations of the classics employed by the two schools of thought during the Period of Disunion. Here he adopts a neutral position, merely noting that Northern and Southern Learning each had its own strong points and that thus each in its own way contributed to an understanding of classical doctrine.5 At the same time there appears to have been a strong early T’ang trend toward philosophical eclecticism that Wei also shared. The contents of the Essentials of Government from Divers Books—which Wei helped compile and which contains materials culled from schools of thought as varied as Confucianism, Taoism, Mohism, and Legalism—provide a good example of this eclecticism. Thus, while the doctrines of Confucius and Mencius constitute the core of Wei’s thought, we also discover a rich overlay of elements from schools of Chinese philosophy that had acted upon and modified classical Confucianism all during the previous millenium.

3  P’i Hsi-jui, Ching-hsüeh li-shih, pp. 176, 207. 4  Ibid., p. 196. 5  SuiS 75.1b.

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

Wei Cheng ’ s Thought [ esp. Regarding Government ]



349

Basic Principles of Good Government

At the very basis of Wei’s thought, around which all other elements in his political philosophy revolved, was the Confucian article of faith that good government would be achieved when a ruler’s conduct had been set right and he became a moral exemplar for his people. The idea is an ancient one. In the Analects we read, “Master K’ung said, Ruling (cheng) is straightening (cheng). If you lead along a straight way, who will dare go by a crooked one?”6 The Mencius (Meng-tzu) says, “Let the prince be benevolent, and all his acts will be benevolent. Let the prince be righteous, and all his acts will be righteous. Let the prince be correct, and everything will be correct. Once rectify the ruler, and the kingdom will be firmly settled.”7 By the time of the appearance of the Great Learning (Ta-hsüeh) and Doctrine of the Mean (Chung-yung) in the last centuries before our era, cultivation of the person (hsiu-shen) and rectification of the person (cheng-shen) were becoming fundamental themes in Confucianism.8 A corollary to this principle was that once the ruler had attained moral perfection the people would naturally emulate him, and in so doing be morally transformed themselves. This transformation, known as chiao-hua, was a basic Confucian goal. Happily, Wei Cheng argued early in Chen-kuan, T’ai-tsung had an excellent opportunity to transform all his subjects since they were now most susceptible to his moral influence. The emperor and one of his officials, Feng Te-i, had argued just the opposite: that after a long period of disorder, such as China had recently suffered, it would be a difficult task to change the people for the better. But Wei countered with the observation that after a long period of peace people tended to become arrogant and difficult to instruct; after a long period of chaos, what the people craved most was order, so that teaching them would be as easy as feeding the starving or giving water to the thirsty.9 Chiao-hua was not a concept that implied positive action; rather, according to the Confucians, the ruler had but to manifest his superior virtue and 6  Analects 12.17; Arthur Waley, Analects, p. 167. 7  Mencius 4.1; James Legge, The Chinese Classics, vol. 2, The Works of Mencius, pp. 310–11. On the theme of rectification of the ruler in Wei’s writings, see Lung Shou-t’ang, “Wei Cheng chih cheng-chih ssu-hsiang” [Wei Cheng’s Political Thought], Hsiang-kang ta-hsüeh chung-wen hsueh-hui hui-k’an [Journal of the Chinese Literature Society, the Chinese University of Hong Kong] (1960), 46–47. 8  See, for example, Legge, The Chinese Classics, vol. 1, The Great Learning, p. 359, “From the Son of Heaven down to the mass of the people, all must consider the cultivation of the person the root of everything besides.” 9  t ctc ch. 193, p. 6084; wckcl 3.2b–5; ckcy 1.16–17b. Feng Te-i’s death shortly after T’ai-tsung came to power (cts 63.3b; hts 100.6) dates this episode early in Chen-kuan.

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

350

Wechsler

the people would automatically and voluntarily respond to his beneficent influence. For Wei, quiescence (ching) was the foundation of transformation.10 The themes of folding one’s hands yet accomplishing everything, quiescence, and good government by means of non-activity (wu-wei erh chih) that are so frequently encountered in Wei’s writings are best known in their relationship to Taoism. Yet in Analects 13.6 Confucius says, “When a prince’s personal conduct is correct, his government is effective without the issuing of orders,” and Analects 15.4 actually contains the phrase wu-wei erh chih; thus the concept is by no means limited exclusively to classical Taoism.11 As early as the Han dynasty, wu-wei erh chih had become a common theme of Confucianism and, as we have seen, occupied an important place as well in the thought of such essentially Confucian philosophers of the post-disunion period as Wang T’ung. Its employment by Wei thus appears to have been more a reflection of the existence of a Taoist pole in early T’ang Confucianism than an expression of his own deep commitment to Taoist ideology per se. Moreover, it was a powerful tool traditionally employed by bureaucrats who were attempting to restrict their ruler’s scope of activity while at the same time seeking to enhance their own.12 Wei’s use of the wu-wei theme was further intended to prod T’ai-tsung into making wen (“civil virtue”) predominate over wu (“military virtue”) in his administration. The minister’s repeated opposition to his emperor’s military plans was based on the belief that nothing could be gained by warfare that could not be gained by moral suasion. It was for this reason, for example, that he called on the emperor not to send troops against the “barbarian” rebel Feng Ang, arguing that if Feng were treated with sincerity and trust, he would “by himself come to court.”13 Wei’s implacable hostility toward things military was also shown in his reaction to a martial dance performed at court during times of feasting and celebration. Entitled “The Prince of Ch’in Smashes the Ranks” (Ch’in-wang p’ochen) and dating from the time of Li Shih-min’s campaign against the rebel Liu Wu-chou early in Wu-te, it was performed to music in the Kuchean style by a 10   wckcl 3.2b. 11  James Legge, The Confucian Classics, vol. 1, Confucian Analects, p. 266; H. G. Creel, “The Fa-chia: ‘Legalists’ or ‘Administrators’?”, in Studies Presented to Tung Tso-pin on his Sixtyfifth Birthday, Bulletin of the Institute of History and Philology, Academia Sinica, extra vol. 4 (Nankang, Taiwan, 1961), p. 614. 12  See, for example, James T. C. Liu, “An Administrative Cycle in Chinese History: The Case of the Northern Sung Emperors,” Journal of Asian Studies, 21 (1962), 143. 13   wckcl 1.8b.

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

Wei Cheng ’ s Thought [ esp. Regarding Government ]

351

company of 120 (alternately 128) men in full armor wielding halberds. At such times it is said that Wei would hang his head to indicate his displeasure at its military theme.14 Naturally, Wei’s abhorrence of aggressive warfare had a very practical as well as an ideological basis. He was keenly aware that Yang-ti’s abortive Koguryŏ campaigns had been a major factor in his demise and that warfare in general could weaken a state as easily as it could strengthen it, a view he neatly summed up in his commentary on the “Eastern Barbarians” (tung-i) section of the Sui History with an old saying, a rhymed couplet, which went: “Striving to broaden one’s virtue leads to prosperity; striving to broaden one’s territory leads to destruction” (wu kuang te che ch’ang; wu kuang ti che wang).15 Wei must also have viewed any increased role of military officials in government with great alarm, since it would have served to undermine civil official preeminence. Indeed, when the word “official” (kuan) appears in Wei’s writings, we are always to understand it in the context of “civil official.” The civil official, not his military counterpart, aided the ruler in achieving moral perfection. Good government came about only as a result of the cooperation between the ruler and his civil officials. As Wei once wrote to T’ai-tsung: I have heard that the ruler is the head and his officials are the arms and legs. If they are well coordinated and of the same mind they will combine to form a [whole] body. If [the body] is formed but is incomplete, it will not yet constitute a whole man. Although the head is exalted, it still must depend on the arms and legs to form a [whole] body. [In the same way], even if a ruler is enlightened and wise, he must depend on his arms and legs [i.e., his officials] to bring about good government.16 How could the ruler’s behavior best be rectified?—by his cheerful acceptance of remonstrance and advice from his officials, who would correct any errors of 14   h ts 97.15; tctc ch. 192, p. 6030, note of Hu San-hsing, ch. 194, p. 6101; thy ch. 33, p. 612; Hsiang Ta, T’ang-tai Ch’ang-an yü Hsi-yü wen-ming [T’ang Ch’ang-an and Western Regions Culture] (1933; reprint, Peking, 1957), p. 62. Cf. C. P. Fitzgerald’s rather unconvincing thesis that Wei bowed his head because he was ashamed of his previous role under the crown prince Chien-ch’eng; Son of Heaven, p. 154. T’ai-tsung appears to have had the last laugh, though, by commanding that Wei Cheng and others furnish the verses to be sung in accompaniment to the Ch’in-wang p’o-chen dance music; wckcl 4.15b–17. 15   SuiS 81.17. This was an echo of themes in the Mo-tzu; see Burton Watson, Mo Tzu: Basic Writings (New York, 1963), pp. 54–55. More condemnations of warfare by Wei are found in SuiS 82.8b–9 and 83.18b–19. 16   w ckwc ch. 1, pp. 10–11; also ckcy 3.4; hl 2.21b–22.

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

352

Wechsler

judgment he had made and ensure that he would never act in such a way as to jeopardize his rule. Early in Chen-kuan Wei observed: Although the ruler is wise, he should still humbly accept [the opinions of other] men. Thereupon wise men will offer him their plans and brave men will exhaust their strength [on his behalf]. Yang-ti relied on his own talents and arrogantly used his own [ideas], so although he spoke of Yao and Shun, he acted like Chieh and Chou. Because he did not know this he was destroyed.17 Wei was fond of pointing out ways in which the sage-rulers of ancient times had made provision for remonstrance, such as the drum used by Yao and the wooden boards set up at the roadside by Shun.18 Because of the important place occupied by civil officials as administrators and as advisers to the throne, they naturally had to be chosen with the utmost care. The worth of an official, Wei observed in good Confucian fashion, was not merely related to his administrative capabilities. Equally or more important, he had to be of good moral fiber. As he once counseled T’ai-tsung: Since ancient times the selection of officials has been a difficult task. Therefore, we examine their accomplishments and investigate their characters. Today, if we seek a man [for office], we must first investigate his conduct, and only when we are certain that he is good will we employ him. If the man is unable to do his job well, it will only be because his talents and energy are not up to par and will cause no great harm. But if we mistakenly employ an evil man, if he is capable he will cause much harm. Only in times of disorder do we seek talent but pay no attention to conduct. In this time of great peace and plenty (t’ai-p’ing), before we employ anyone we must see to it that he possesses a combination of talent and good conduct.19 In the process of selection, the ruler had to advance the morally superior man (chün-tzu) and reject the petty man (hsiao-jen). At the same time, differentiating one from another was not an easy task, Wei warned, and the ruler had to learn not to allow the defects of worthy men and the merits of inferior men to delude him about their ultimate usefulness: 17   t ctc ch. 192, p. 6053. 18   w ckwc ch. 1, p. 16. 19   wckcl 3.15–15b.

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

Wei Cheng ’ s Thought [ esp. Regarding Government ]

353

Petty men are not without their minor virtues and superior men are not without their minor faults. The minor faults of superior men are like flaws in white jade; the minor virtues of petty men are like the single cuts lead knives can make [before they must be rehoned]. The good workman does not attach any importance to the single cut a lead knife can make because its minor virtue is incapable of obliterating a multitude of defects. The good businessman does not reject white jade with flaws because its minor defects are insufficient to spoil its great beauty.20 A further cardinal rule of good government was that penal law and punishment should not be oppressive. According to the theory behind chiao-hua, if the ruler was able to provide a perfect moral model, laws and punishments designed to regulate the people would be entirely unnecessary. As Wei once put it: One cannot serve the three-foot laws [in ancient times, the bamboo slips on which laws were written were three Chinese feet long] to constrain the people of the Four Seas and still seek to fold one’s hands and rule by non-activity (wu-wei). Therefore, when the sage and wise rulers altered the customs and changed the habits [of the people], they did not depend on strict punishments and laws, only on benevolence and righteousness…. Benevolence and righteousness comprise the trunk of rule, punishments comprise the branches. Ruling with punishment is like driving horses with a whip. When the people have all been transformed no strict punishments are meted out. When horses have completely exhausted their energy no whips are used. Speaking like this, punishment cannot result in good rule. This is already clear.21 At the same time, Wei was also realistic (or Legalistic) enough to conclude that law and punishment were necessary, if evil, tools of a ruler whose morality was frequently imperfect. His preface to the “Biographies of Harsh Officials” (k’u-li chuan) in the Sui History notes that there are four ways through which a nation could be administered: the first, benevolence and righteousness; the second, restraint of the people by a ritual code; the third, law; the fourth, punishment. Benevolence and righteousness and restraint by ritual were the essentials, or “trunk”; law and punishment were the non-essentials, or “branches.” But if law and punishment alone could not bring about chiao-hua, they 20   w ckwc ch. 1, p. 7. 21  Ibid., pp. 13–14.

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

354

Wechsler

might nevertheless ultimately aid its attainment: “Without the trunk [the tree] does not stand; without the branches it is incomplete.”22 Once having acknowledged the utility of these Legalist tools, Wei made certain to point out that they worked best during times of great disorder, such as in the Warring States period, but that during times of peace, such as China was experiencing during Chen-kuan, laws and punishments were best when they were moderate and compassionate.23 In much of Wei’s thought there are echoes of Mencius’ economic doctrines. Mencius had said, “If the seasons of husbandry be not interfered with, the grain will be more than can be eaten.”24 Like all Confucians, Wei believed in the primacy of agriculture over trade and industry. Once, upon hearing that a prefecture in Chien-nan (Szechuan) and an imperial atelier at the capital were manufacturing silk gauze, embroidered silk, and metallic baubles for the pleasure of the emperor, Wei wrote a remonstrance that well illustrates his views on economic priorities: [Working with] gold, silver, pearls, and jade interferes with agriculture. Embroidering with metallic and colored threads harms a woman’s work. “If one man does not till the soil, [someone in the empire] will suffer hunger. If one woman does not weave, [someone in the empire] will suffer cold.”25 Men of antiquity either cast these things into deep springs or burned them on the thoroughfares, but Your Majesty is fond of them. I cannot bear the shame of it.26 Thus, the state would flourish only when the ruler did not interfere with the agricultural labor of the peasantry by making excessive or unseasonal demands on them in order to wage war, construct palaces, manufacture useless articles of luxury, and the like. As we have seen, Wei’s anxiety about the economic consequences of T’aitsung’s policies is evident in many of his memorials and remonstrances. He opposed the plan to conscript the chung-nan into the army because manpower for agricultural labor and corvée would thereby be reduced. He opposed 22   SuiS 74.1. 23  See wckcl 3.8, 3.21–21b, 4.3–3b; wckwc ch. 1, pp. 13–17, ch. 2, pp. 24–25. Lung Shout’ang, “Wei Cheng,” 56–57, discusses certain parallels between Wei Cheng’s thought and Legalism. 24   Mencius 1A.3; Legge, The Chinese Classics, vol. 2, The Works of Mencius, p. 130. 25  From hs 24A.8b. 26   wckcl 1.27.

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

Wei Cheng ’ s Thought [ esp. Regarding Government ]

355

the Feng and Shan sacrifices because China had not yet recovered from the economic dislocations of the late Sui period and the people, he felt, should not be obliged to support the pomp and ceremony the sacrifices required. It would be wrong, though, to assume that Wei’s economic concerns stemmed solely from his deep compassion for the masses. Nor were they caused, as some Chinese Marxist historians would have us believe, merely by his fear of the consequences of peasant discontent, which, they say, he developed during his service under the late Sui rebels Yüan Pao-tsang and Li Mi. On the contrary, Wei ultimately viewed the Chinese economic order from the top downward and was always concerned foremost with the economic well-being of the central government and the dynasty in general. Although he held that the people should not be taxed oppressively, he argued even more vigorously that policies which tended to reduce total productivity and tax receipts (like the plan to conscript the chung-nan) and thereby weaken the central government were to be avoided. Illustrating this point was Wei’s response early in Chen-kuan to the emperor’s plan to create a modified “feudal” (feng-chien) system by enfeoffing princes of the blood and selected court officials with territory in various parts of China.27 Wei gave several reasons why he opposed the plan, but the most important of them were related to the effect it would have had on the central government and the bureaucracy: the people had not yet recovered from the Sui disorders and if they were placed under the administration of princes and officials on the fiefs, they would fear increased exactions on their wealth and labor and would abscond; if a large amount of territory were distributed as fiefs, the central government would be able to levy direct taxes only on a shrunken royal domain and would soon become impoverished; if government revenues fell, the great ministers and other bureaucrats, who were dependent on their cash and food stipends, would have no way to subsist.28

Consolidation Themes

Even if a ruler followed all of Wei’s ideas concerning the conduct of government discussed above, there was still the possibility that initial success might cause him to become less diligent and to slacken his efforts. In Wei’s view, by far the greatest obstacle to a ruler’s success was complacency. He thus liberally sprinkled his writings and speeches with three mottoes or principles drawn 27  For court discussions on the feng-chien system, see thy ch. 46, pp. 824–30. 28   w ckwc ch. 1, pp. 26–27; thy ch. 46, pp. 826–27.

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

356

Wechsler

from the Chinese classics, by means of which he exhorted T’ai-tsung to maintain a constant vigilance over his administration. The first principle comes from the Book of Odes (it is also quoted twice in the Tso-chuan): “All are [successful] at first, But few prove themselves to be so at the last” (mo pu yu ch’u, hsien k’o yu chung).29 The second is based on three passages in the Book of History and the Tso-chuan that are similar in meaning: (1) “Be careful for the end at the beginning” (shen chung yu shih), which Wei alters slightly to “Be careful at the end as you were in the beginning” (shen chung ju shih);30 (2) “Be careful of the beginning and fearful of the end (shen shih erh ching chung); then in the end you will have no distress”;31 and (3)’. “He who at last, as at first, is careful as to whom and what he follows is a truly intelligent sovereign” (chung shih shen chüeh yu wei ming-ming hou).32 The third principle, which Wei employed most often of the three, is also from the Tso-chuan: “In a position of security think of peril (chü an ssu wei). If you think thus, you will make preparation against the danger, and with the preparation there will be no calamity.”33 Wei appears to have been somewhat pessimistic concerning the ability of monarchs to avoid the dire consequences of complacency, as can be seen from the comments he made during a court debate that took place in 638. T’ai-tsung had inquired of his officials whether founding a dynasty or preserving it was more difficult. Fang Hsüan-ling had replied that when a dynasty was established all the contenders for the throne struggled violently for power; therefore the founding was more difficult. Wei, however, argued just the opposite. “From 29  Legge, The Chinese Classics, vol. 4, The She King (3.1), p. 505; vol. 5, The Ch’un Ts’ew with the Tso Chuen, Duke Hsüan, second year, p. 288, and Duke Hsiang, thirty-third year, p. 562, all slightly emended. For Wei Cheng’s use of this phrase, see, for example, ckcy 6.6 and hl 2.32b. 30  Legge, The Chinese Classics, vol. 3, The Shoo King, p. 211. For Wei Cheng’s use of the phrase, see tctc ch. 192, p. 6048. It is occasionally quoted by other of T’ai-tsung’s officials, such as Yü Shih-nan (cts 72.3b) and Chang Hsüan-su (ctw 148.8b), suggesting that it may have been part of the Confucian vocabulary of the time. 31  Legge, The Chinese Classics, vol. 5, The Ch’un Ts’ew with the Tso Chuen, p. 517, slightly emended. It, in turn, was based on a passage in the Book of History: “To give heed to the beginning, think of the end;—the end will then be without distress” (shen chüeh ch’u, wei chüeh chung; chung i pu k’un); Legge, The Shoo King, p. 490. For Wei’s use of the shen shih erh ching chung phrase, see ckcy 1.7. 32  Legge, The Chinese Classics, vol. 3, The Shoo King, p. 210. 33  Ibid., vol. 5, The Ch’un Ts’ew with the Tso Chuen, p. 453. For Wei’s use of this phrase and its variations, see wckcl 3.17, 3.19, 4.5b, 4.21b; ckcy 10.11, 10.18b. See also its use by Wei’s contemporaries Ts’en Wen-pen, ckcy 10.9, Chang Hsüan-su, ctw 148.12b, and even the Taoist magus Sun Ssu-mo, as quoted in Joseph Needham, Clerks and Craftsmen in China and the West (Cambridge, 1970), p. 345.

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

Wei Cheng ’ s Thought [ esp. Regarding Government ]

357

the time of the rulers of old, none failed to gain [the empire] by hardship or lose it by indolence. Preserving what has been gained is the more difficult.”34 Some eight centuries before Wei Cheng’s time, Lu Chia had cautioned his monarch, Han Kao-tsu, that although he had won the empire on horseback he could not rule it from horseback,35 thereby inaugurating a great Confucian cri de coeur. Like Lu Chia, Wei appears to represent a recurrent type of official in Chinese history whose tenure of office comes soon after the founding of a dynasty. Such an official exhorts his prince to dismount from his horse of conquest and make civil rather than military virtue his central concern. He stresses that the overriding need of the regime is the preservation in the civil realm of the gains it had already won on the battlefield. He therefore advises extreme caution in the making of domestic and foreign policy with a view toward reinforcing the foundations of dynastic power. An analysis of Wei’s role under T’ai-tsung suggests that he embodied many of the characteristics associated with what we might call the “consolidation minister.”36 Too many rulers had failed to consolidate their regimes before embarking upon various ill-conceived domestic and foreign ventures, and as a consequence had been toppled from their thrones. Yet the failures of past history were nevertheless useful, Wei believed, in that they could be employed as mirrors which might be held up to guide the rulers of the present. By early T’ang times the tradition of the mirror image (which Sung Shen-tsung later employed in selecting a title for Ssu-ma Kuang’s magnum opus) was already more than a millenium old. The Book of Odes contains what is perhaps the prototype in Chinese literature of all mirror imagery: “The mirror of Yin is not far-distant;—It is in the age of the [last] sovereign of Hsia.”37 Here, the evil conduct of Chieh, “bad last” ruler of Hsia, is held up as a warning example to the rulers of the Yin (Shang) dynasty. The passage from the Book of Odes was one of Wei’s favorities, and in his writings he made copious use of both it and mirror imagery in general.38 34   t ctc ch. 195, p. 6140; also wckcl 4.11b–12b; hl 1.22–22b; ckcy 1.3b–4. Cf. a similar episode in tctc ch. 195, p. 6161, and ckcy 10.18–18b, 1.8b–9. 35   s c 97.7–7b. 36  Perhaps one might even extend this typology to include those officials appearing at the beginning of any reign who tend to stress the consolidation of gains achieved during the previous administration over any expansionary ideas their own rulers might entertain. 37  Legge, The Chinese Classics, vol. 4, The She King, p. 510, translates the character chien as “beacon.” On mirror imagery see also Achilles Fang, The Chronicle of the Three Kingdoms, 2 vols. (Cambridge, Mass., 1952, 1965), 1: xviii–xix. 38  See, for example, wckcl 3.20; wckwc ch. 3, pp. 29, 30; SuiS 45.19b. Hung Mai (1123–1202), Jung-chai sui-pi (Wan-yu wen-k’u ed., Shanghai, 1939), ch. 16, pp. 154–55, demonstrates

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

358

Wechsler

The Civil Official as a Restraint on Imperial Power

In Wei’s eyes the civil official served his prince as a human mirror whose counsel reflected the collective wisdom of ages past. Knowing well that civil officials had the potential to exert a strong restraining force on untrammeled monarchical authority—the source of ruin for many a dynasty—he was thus fiercely protective of their powers and prerogatives. A considerable number of episodes in the sources reveal how strongly Wei resisted every move by T’ai-tsung to impinge upon the prudential and administrative roles of his civil officials. In one such case, the emperor sharply reprimanded Fang Hsüan-ling and Kao Shih-lien for daring to raise inquiries concerning articles being manufactured for imperial use at ateliers located in the vicinity of the Hsüan-wu Gate. When the two officials pleaded with the emperor for forgiveness, Wei became outraged at their obsequious attitude and bravely stepped forward to speak out: I do not understand why Your Majesty reprimands them, nor do I understand why Hsüan-ling and Shih-lien beg forgiveness. Since Hsüan-ling [and Shih-lien] are employed as high officials, they are Your Majesty’s legs and arms and eyes and ears. If there is something being manufactured, how can they not be allowed to know about it? If you reprimand their inquiry, none of your officials will comprehend it.39 Another time several memorialists suggested that T’ai-tsung ought to receive all memorials directly from his officials rather than allow them to pass through regular bureaucratic channels. In this way, it was said, he would avoid concealment of facts. Wei saw in this yet another attempt to hack away at civil official participation in government. When asked his opinion of the suggestion, he sarcastically replied that if the memorialists were requesting that the emperor do away with his officials and personally deal with all the administrative trivia at court, then the emperor ought personally to deal with all provincial affairs as well!40 In a similar vein, Wei was ever ready to protest any sign that the emperor was not treating his officials with the respect due their privileged status in that Wei belonged to a whole group of Han and T’ang dynasty officials who advised their rulers to regard former dynasties as mirrors. But whereas Han officials like Lu Chia pointed to Ch’in Shih-huang-ti as their bête noire, Wei and his fellow T’ang officials could make use of a minatory example much closer to them in time—Sui Yang-ti. 39   c kcy 2.43b; tctc ch. 196, p. 6173; Wang Tang, T’ang yü-lin, ch. 1, p. 16. 40   wckcl 3.5–5b; tctc ch. 195, p. 6163.

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

Wei Cheng ’ s Thought [ esp. Regarding Government ]

359

government. “I have heard,” he would often say, “that ‘A ruler in employing his ministers should be guided solely by the prescriptions of ritual.’ ”41 A case in point was his objection to the sentencing of the prefect (ling) of Ch’ang-an county, Wang Wen-k’ai, to thirty strokes of the rod because he did not order the official P’ei Chi to quit the capital after being dismissed and sent home by the emperor. “P’ei Chi’s affair deserved ten thousand deaths. Now Your Majesty, remembering his old merit, did not follow the letter of the law; you merely dismissed him from office and went no further than reducing his enfeoffment by half. But even people sentenced to banishment are given some time [in which to prepare to leave]. Why was this not even more true in the case of Chi, who was [only] dismissed and sent back home? The Ancients said: ‘Advance men with propriety (li), and send them away with propriety.’ I think that Wen-k’ai did not incessantly force [Chi to leave] because he was aware of Your Majesty’s gracious pardon and knew that Chi was a great official. Then, if we discuss the facts of this case, there has been no crime.” T’ai-tsung replied, “When I ordered Chi to pay his respects to his ancestors [i.e., sent him home without additional punishment], was this not propriety?” He then pardoned Wen-k’ai and did not [further] inquire [into the matter].42 In several other of his remonstrances discussed in this study, Wei Cheng similarly argued that the emperor had to follow to the letter the rules of propriety governing his relations with his ministers. By such means he attempted to ensure that the emperor would not capriciously trample on official class prerogatives and that officials would not be left to suffer unaided the vagaries of imperial authority. It would be enlightening to know just how far Wei’s own contributions to T’ang ritual were similarly aimed at tightening the screws on imperial freedom of action relative to the official class.

Foxes on the City Walls, Rats on the Altars of State

Although the emperor represented a formidable threat to official class power and prestige, he was merely one among several such threats at court; following close upon his heels came members of the imperial house and their relatives by marriage. We have already seen how Prince Li T’ai began demanding greater deference from court officials once he felt he had received T’ai-tsung’s support and how Wei Cheng had opposed him. Sometime later, in 638, Wei 41   wckcl 2.22; wckwc ch. 1, pp. 5, 20; tctc ch. 199, p. 6160. The quote derives from Analects 3.19; Waley, Analects, p. 99. 42   wckcl 1.11b–12.

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

360

Wechsler

once again had to defend members of the official class against the imperial princes. This year the president of the Board of Rites, Wang Kuei, memorialized that according to T’ang law, officials of the third rank and above, upon encountering princes of the first rank (sons or brothers of the emperor), were not required to dismount from their horses as a sign of respect, but that lately everyone was disregarding this rule. T’ai-tsung angrily exclaimed, “Will you all honor yourselves and denigrate my sons? This is improper and cannot be done!” Wei thereupon remonstrated that since antiquity princes of the first rank had always been inferior to officials of the third and above. As for the latter dismounting for the princes, “If we inquire of ancient precedent, then there is nothing that can serve as a model. To practice [dismounting] at the present time would naturally violate the laws of the state.”43 The emperor then accepted Wang Kuei’s memorial. Wei also did not have to be reminded that since Han times relatives by marriage of the imperial family (wai-ch’i) had energetically vied with court officials for imperial favor, on occasion even usurping power for their own houses. He was thus strongly anti-wai-ch’i in sentiment. In chapter 5 we saw that Wei protested against the sentencing of Hsüeh Jen-fang to one hundred strokes of the bamboo and dismissal from office for detaining the father of a concubine of an imperial prince for questioning. The remonstrance he wrote at the time reads in part as follows: Foxes on the city walls and rats on the altars of state are all petty creatures, [but] because they rely on [these essential structures] getting rid of them is not easy. How much more is this the case with relatives by marriage and imperial princesses? In former ages they were all difficult to control, and since the Han and Chin dynasties they have been out of hand. By the middle of the Wu-te period most of them were already arrogant and indolent. Only when Your Majesty ascended the throne did they become respectful. In doing his duty Jen-fang was able to preserve the law for the sake of the state. How can you unreasonably mete him out a severe punishment in order to achieve the selfish ends of your relatives by marriage?44 Wei’s preface to the wai-ch’i section in the Sui History similarly reveals his strong bias against that group.45 43  Ibid. 2.6; see also ckcy 7.14–14b; tctc ch. 195, p. 6135. 44   wckcl 2.4b; see also ckcy 2.42b–43; thy ch. 51, p. 886. 45   SuiS 79.1–1b.

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

Wei Cheng ’ s Thought [ esp. Regarding Government ]

361

Eunuchs, because of their proximity to the throne, their historical propensity to appropriate power at the expense of other court groups, and the tendency of Chinese monarchs to use them as a counterweight to the regular bureaucracy, were yet another source of anxiety for Wei. Early in the T’ang, eunuchs were for the most part confined to the palace and the environs of the capital; there was as yet little sign that they would attain a dominant position in T’ang government or that they would manipulate the throne itself, as they did during much of the eighth and ninth centuries. But by the middle of Chen-kuan, eunuchs were already being assigned as messengers to the provinces and even beyond the frontier, thus provoking the resentment of regular bureaucrats. Wei’s hostile attitude toward the eunuchs is revealed by an event that took place in 637. At this time a eunuch messenger on assignment had gone to the Bureau for the Surveillance of the Frontier (ssu-men) to obtain a passport required of all those venturing through one of the twenty-six checkpoints along the border. A secretary of the Bureau, probably with some malice, had delayed the request. The eunuch had reported the matter to the throne, causing the immediate demotion of the secretary to a provincial post. Angered that a mere eunuch had precipitated the dismissal of a regular official, Wei quickly remonstrated: It has always been difficult to treat the likes of eunuchs with any intimacy, since they lightly make up stories and find it easy to provoke trouble. The practice of employing them to travel alone as envoys to far-off places is unwise and should not be broadened. [The matter] should receive your most careful consideration.46 The emperor was persuaded to rescind his dismissal of the secretary, and promised that from that time forward eunuchs would no longer be employed as provincial envoys.47 Lastly, Wei blocked an attempt by still another rival power group, the Buddhist clergy, to enhance its influence over the throne. Memorialists, undoubtedly encouraged by the clergy itself, had suggested to the emperor that he daily receive a delegation of Buddhist monks in private audience to aid him in his religious worship. It was obvious to Wei that this was merely a ploy to

46   wckcl 2.13; also ckcy 5.18b; tctc ch. 195, p. 6158; thy ch. 65, p. 1132. wckcl 2.23b contains a somewhat different version of the same episode. The incident is discussed in J. K. Rideout, “The Rise of the Eunuchs during the T’ang Dynasty,” Asia Major, n.s., 1 (1949), 58–59. 47   c kcy 5.18b.

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

362

Wechsler

allow the Buddhists a hand in determining state policies, and he would have none of it: Buddhism basically esteems purity in order to avoid worldly frivolities and strife. Furthermore, religious and secular matters are unalike. Long ago Shih Tao-an was the most eminent monk of the age. When Fu Yung-ku [Fu Chien of the Former Ch’in, rg. 351–85] rode with him in the same carriage, Ch’üan I regarded it as improper. Shih Hui-lin [another monk] was not without talent and refinement. When Sung Wen[-ti] [rg. 424–53] led him up to the palace hall, Yen Yen-chih said: “As to this exalted position, is it proper to cause a man maimed by punishment [i.e., the tonsure] to occupy it?” Now if Your Majesty wishes to honor a belief in Buddhism, why would it be necessary to receive the monks daily in separate audience?48 The proposal was quietly shelved. Yet Wei was no Han Yü, the great literatusofficial of the late eighth/early ninth century renowned for his eloquent attack on Buddhist influence in Chinese society. Rather, he appears to have had no strong antipathy toward Buddhism per se and made no criticism of the routine Buddhist observances that filled court life. Indeed, when T’ai-tsung learned that kingdoms on the Korean peninsula had landed some Buddhist monks at a place in modern Shantung province and surmised that they were spying on China, Wei was quick to allay his fears.49

On Correct and Evil Officials

Officials were susceptible to erosion of their authority not only by the emperor and those court groups just discussed, but also by the more unscrupulous of their colleagues who were ever ready to further their own careers at the expense of others. In Wei’s eyes the censor Ch’üan Wan-chi had been such a person.50 So too, probably, was Wei Hung-chih, an obscure, low-ranking official who for some reason unspecified in the histories dared to criticize the chief ministers, thereby overstepping his position and incurring Wei’s wrath. In retaliation, Wei wrote a memorial in which he implied that Hung-chih had 48   wckcl 3.24–24b. See also hl 1.13b–14; thy ch. 47, p. 836. I am indebted to Professor Arthur F. Wright for his aid in this translation. 49   wckcl 4.23. 50  See above, p. 152.

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

Wei Cheng ’ s Thought [ esp. Regarding Government ]

363

formed a faction (p’eng-tang),51 a term that in Chinese politics generally bore a strong pejorative connotation. As we see in another of his memorials, according to Wei there were two types of associations of men at court, one good, the other bad. If men came together for a good purpose, they were called t’ung-te, or “united in virtue”; if they came together for an evil purpose, they were called p’eng-tang.52 Interestingly, the distinction Wei makes here between the two types of associations precedes by four centuries Ou-yang Hsiu’s famous essay “On Factions” (P’eng-tang lun), in which Ou-yang made the unorthodox proposal that his sovereign not indiscriminately label all factions bad, but rather distinguish between factions composed of superior and petty men.53 What Wei was implying in his own memorial, naturally, was that T’ai-tsung should discern who among his officials had merely “united in virtue” and who had formed actual factions, and that only the latter deserved punishment for deluding the throne and interfering with the policies of upright officials. Of what modes of official behavior did Wei Cheng approve? As there was an ideal ruler who could serve as a moral exemplar for his people, so too was there an ideal official who could serve as a model for all men of his class. In a memorial written in 640, Wei described at length the qualities possessed by such an official, quoting from the description of the “six correct officials” (liu-cheng) by the Former Han scholar Liu Hsiang (ca. 79–ca. 8 bc): Who are known as the liu-cheng? The first: When sprouts have not yet stirred and signs have not yet been perceived, he alone clearly sees the possibilities of preservation and destruction and the essentials of success and failure, makes preparations against [destruction and failure] before they have appeared, and causes his ruler to excel and to occupy a glorious position. Such a one is a divinely inspired official (sheng-ch’en). The second: With a humble mind he exhausts his ideas. Daily he offers good advice. He exhorts his ruler with propriety and righteousness; he advises his ruler with far-sighted plans. He accords with his ruler’s excellences and corrects his evils. Such a one is an excellent official (liang-ch’en). The third: He rises early and retires late. He is not remiss in recommending the worthy and repeatedly speaks about events of old to encourage his ruler. Such a one is a loyal official (chung-ch’en). 51   w ckwc ch. 1, p. 7. 52  Ibid., p. 8. 53  See James T. C. Liu, Ou-yang Hsiu (Stanford, 1967), pp. 52–55.

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

364

Wechsler

The fourth: He clearly ascertains [the roots of] success and failure. At an early time he prevents [failure] and corrects [faults]. He stops up the leaks [through which failure may flow] and cuts off their source. He turns misfortune into prosperity and in the end causes his ruler to be without anxiety. Such a one is a wise official (chih-ch’en). The fifth: He protects civil virtue and serves the law. When he is employed as an official and exercises power, he does not accept presents, refuses emoluments, eschews gifts, and is frugal with food and drink. Such a one is a pure official (chen-ch’en). The sixth: When the nation is in confusion and in chaos he does not flatter, but dares to oppose his ruler’s stern countenance and to speak of his ruler’s faults to his face. Such a one is an upright official (chih-ch’en).54 Obviously, Wei Cheng had taken the liu-cheng as a guide for his own official conduct. At the same time Wei also quoted from Liu Hsiang’s description of six types of “evil officials,” the liu-hsieh. The ordinary official (chü-ch’en) thinks only of his office and is covetous of his salary. He goes with the times and has no fixed principles. The flattering official (yü-ch’en) says “good” to whatever his ruler says and “you may” to whatever his ruler proposes to do. He ascertains what his ruler likes and advances it in order to please him; he agrees with his improper conduct in order to make him happy, without taking account of future harm. The treacherous official (chien-ch’en) is inwardly designing but respectful on the surface. He watches his ruler’s countenance and employs artful speech accordingly. He is jealous of worthy and capable men; he reveals the good points and hides the defects of those he wishes to recommend and reveals the faults and hides the good points of those he wishes to degrade. He causes his ruler to reward and punish unfairly and the ruler’s orders to be disobeyed. The slandering official (ch’an-ch’en) has knowledge sufficient to hide his faults and eloquence sufficient for him to advise others. In the palace he separates flesh and blood and at court incites disturbances. The base official (chien-ch’en) monopolizes and usurps power and makes mountains out of molehills. He forms a faction with members of his own family to enrich his house. He alters the meaning of his ruler’s orders to enhance his own position. The official who destroys the state (wang-kuo chih ch’en) fawns upon his ruler with artful talk. He beguiles his ruler into unrighteousness. He surrounds his ruler with a faction so as to hinder his right perception of things. He causes black and white, 54   w ckwc ch. 1, p. 9. The original text is found in Liu Hsiang, Shuo-yüan (sptk ed., Shanghai, 1929), 2.1–2.

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

Wei Cheng ’ s Thought [ esp. Regarding Government ]

365

right and wrong to be undifferentiated. He causes the bad name of his ruler to be broadcast at home and abroad.55 Thus, the wise ruler is one who can distinguish those among his officials who follow the path of the liu-cheng from those who follow the path of the liu-hsieh. Ultimately, good government depends on the ruler’s understanding of human nature—one of the most fundamental of all Confucian themes. Indeed, herein lies a basic characteristic (basic flaw, some might argue) of Wei’s thought: its almost complete lack of originality. Wei was not a seminal thinker, the creator of a new philosophical system. His extant writings and recorded speeches, mostly responses to practical problems besetting the conduct of his prince and the administration of government, do not concern themselves with purely ideological or doctrinal matters. Where Wei’s ideas on politics, administration, interpersonal relations in government, and so on, do emerge, they represent for the most part a reworking of or a new emphasis on some rather venerable Confucian themes, both classical and post-classical. Wei’s writings are predictable, often tedious, relieved from time to time only by the liveliness of his invective and the ingenuity of the devices that he used to provoke the emperor. If studies on Confucianism typically skip in time all the way from the Later Han dynasty up to the appearance of the celebrated Han Yü (768–824), precursor of the Neo-Confucian revival, it may well be with good reason. From the purely philosophical point of view at least, the intervening period appears to have been a wasteland, when no one save the much discredited Wang T’ung came forward to take up the great Sage’s mantle. Yet even Wang T’ung was more imitator than innovator, content to expound upon the Way without expanding upon it. Nevertheless, following the long Period of Disunion, which saw the eclipse of Confucianism by other, more vigorous philosophies, perhaps it was enough, as Wei did, to passionately proclaim the great principles of Confucianism everywhere throughout the halls of government. From the moment of his arrival at the Chen-kuan court, Wei lived and breathed Confucianism, stale as this Confucianism may seem to us in retrospect.56 In the doctrines he so forcefully espoused, Wei saw the means of arresting the premature dynastic decay that had plagued China for the past several centuries. But Wei’s intent was not 55   w ckwc ch. 1, pp. 9–10; Shuo-yüan 2.2–3b. 56  Given the examples of both Wang T’ung and Wei Cheng, I cannot agree with the assertion that before the Sung “while there were Confucian scholars, there were virtually no Confucianists; that is, persons who adhered to the teachings of Confucius as a distinct creed which set them apart from others”; Wm. Theodore de Bary et al., Sources of Chinese Tradition (New York, 1960), p. 411.

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

366

Wechsler

merely to strengthen the power of the throne; rather it was to provide restraints on this power at the very moment he was seeking to increase the advisory and policy-making roles of the civil official. Wei’s writings and speeches, and his political behavior in general, were designed to demonstrate to his colleagues (and their successors) how they might defend themselves against the throne and other court groups, how they might curb imperial excesses, and how they might wage the struggle to gain a paramount voice in decision making. More than anyone else during his time, Wei pointed to the power the civil official potentially could wield in government and more than anyone else he attempted to translate that potential into reality.

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

Imperial Power and the Ruling Class [under Empress Wu] Richard W. L. Guisso The single most intriguing theme discernible in the Empress Wu’s regime, at least by modern historians, has been that of social change. Though the debate continues, Maoist historians seem to have reached a broad consensus that her opposition to the aristocracy and to the large-landlord class intensified class struggle both inside and outside the ruling group, and so constituted an important stage in the dialectical process.1 Western scholars who are interested in the problem have usually taken as their starting point the work of two major scholars. From Naitō Torajiro has come the idea that the late T’ang saw social changes which made it the transitional period between medieval-aristocratic and modern-bureaucratic China,2 and from Ch’en Yin-k’o the perception of T’ang history prior to the An Lu-shan rebellion as a struggle between two blocs—an entrenched aristocracy and a newly-rising group of examination bureaucrats.3 Professor Ch’en elaborated his thesis by the division of the aristocracy into contending northeastern and northwestern groups, and by suggesting that the Empress Wu was responsible for the most significant change: From the establishment of the T’ang by Kao-tsu and T’ai-tsung until just prior to the accession of Kao-tsung, those who held the most important civil and military posts in the state were mostly the successors of the Sui, Northern Chou, and Western Wei. This is to say that they were the descendants of the bloc assembled under the banner of Yü-wen T’ai’s “Kuan-chung first” policy. From the time the Empress Wu seized power, she destroyed this traditional policy step by step and, because of her ambition to found her own dynasty, she began the destruction of the fu-ping Source: “Imperial Power and the Ruling Class,” in Richard W. L. Guisso, Wu Tse-t’ien and the Politics of Legitimation in T’ang China, Bellingham: Program in East Asian Studies, Western Washington University, 1978, 70–86. 1  See Chapter 1, notes 46–49. 2  See Miyakawa Hisayuki, “Outline of the Naitō hypothesis,” and Miyakawa Ichisada, “Four ages.” 3  The thesis is presented most systematically in Parts i and ii of Cheng-chih shih, pp. 1–94.

© koninklijke brill nv, leiden, ���9 | doi:��.��63/9789004380158_013

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

368

Guisso

system which was the traditional prop of the Kuan-chung bloc. [So] in this period social classes were in a state of flux, rising and falling. As for the chin-shih and other examinations, although established in the Sui, they were unnecessary for the attainment of civil office. Under the Empress Wu, great emphasis was placed upon selecting officials for their literary abilities, and [because] she broke the old rules of selection, the chin-shih degree became the object of great competition everywhere. At that time there existed many persons in Shan-tung and Chiang-tso who had devoted themselves to literary pursuits, but lack of membership in the Kuan-chung bloc had disqualified them from rising to official positions. The political revolution of the Empress Wu brought them to court. After this the high positions formerly monopolized by the old [elite] of Wei, Chou, and Sui were necessarily taken and occupied by this newly rising class. The replacement of Li-T’ang by Wu-Chou was thus not merely a political change. It was a social revolution and, seen in this light, was a more significant change than the dynastic succession from Sui to T’ang.4 … Both these hypotheses are persuasive ones, though the former has proven to be somewhat more durable than the latter. In Ch’en’s work certain weaknesses in premise and definition, as well as internal contradictions, have led to the successful challenge of some of his conclusions,5 but several of his insights 4  Ch’en, Cheng-chih shih, p. 14. 5  See, for instance, Wechsler, “Factionalism,” in Perspectives, pp. 87–120, and the several remarks in Johnson, Oligarchy, pp. 128–130. In an earlier version of his work, Johnson listed three broad categories of objection to Ch’en’s thesis, and these seem worthy of mention here. In the first place, Ch’en’s belief that the ruling class polarized along historico-geographical lines because of the rivalry between Chou and Ch’i does not give sufficient weight to the fact that the imperial clans of both originated in the so-called Four Garrisons, and that the ruling group in both was composed of precisely the same type of clan; that is to say, of Hsien-pei and collaborating Chinese who were not numbered among the preeminent surnames of the time. This latter fact is important because while the refusal of the northeastern elite to collaborate with the Ch’i and the Sui is well-known, the corresponding attitude of the northwestern clans usually goes unremarked. Liu Fang, in his Hsing-hsi lun (see note 8), identified as the preeminent of the Northwest the six clans of Wei, P’ei, Liu, Hsüeh, Yang and Tu. While they were prominent in the government of Sui and T’ang, they seemed to play hardly any role at all in the Chou central administration and so can hardly be identified with Yü-wen T’ai’s “northwestern bloc.” The imperial clans of Sui and T’ang both claimed kinship with the purely Chinese elite of this area, but it is well established that these were spurious claims and that both were of Six Garrison (Wu-ch’uan) origins. For this reason there seems no cause

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

Imperial Power And The Ruling Class [under Empress Wu]

369

remain valuable, and the emphasis he places upon regionalism, examinations, and imperial policy as a determinant of change seem still to be warranted. Moreover his work raises some questions which might profitably be examined in the present context. Did the Empress Wu, for instance, consciously foment a “social revolution” even in the narrow sense in which Professor Ch’en uses the term? Did her reign see the rise of any new “class”? What role was played by the examination system and the decline of the fu-ping in societal change? And finally, what was the nature of the T’ang ruling class and, within it, the relationship between status and office? While the nature of our sources for the early T’ang is such that all statements regarding social mobility contain a certain degree of speculation,6 partial or interim answers to all these questions are possible. Of the several contemporary observations on the subject ot the T’ang ruling class and its development, that of the genealogist-historian Liu Fang is perhaps

to assume a political polarization of the T’ang ruling class on purely regional grounds. I shall later show that friction arose more from perceptions of cultural and racial purity than from geography. The second criticism of Ch’en’s formulations is one applied to his definitions. To him the T’ang “ruling class” consisted of “the [imperial] Li clan and the high civil and military officials,” so that the local and semi-official power structure receives too little consideration. His use of such terms as “class” and “social revolution” are rather vague and, more seriously, his attempt to divide the so-called Shan-tung and Kuan-chung blocs by the existence of the T’ai-hang Mountains is hardly adequate. To cite only one example, the native place of the Empress Wu lies to the northwest of the range yet, as I shall show, she was considered by contemporaries to be an “easterner.” Finally, Professor Ch’en seems never to resolve a basic contradiction. After depicting so clearly the regional rivalry in Part I of his book, he moves on in Part ii to the struggle between the new [post-Empress Wu] and the old elites, writing on p. 53, “Before the Empress Wu destroyed the old ‘Kuan-chung First’ policy, the T’ang ruling class had been made up not only of the Chinese, non-Chinese, and semi-Chinese clans of Yü-wen T’ai’s northwestern bloc, but also of the great Shan-tung clans who in tradition followed that of the northern dynasties. Men of this type composed almost the whole of the upper bureaucracy….” [My emphasis]. A comparison of this passage with that quoted in the text shows a direct contradiction, caused by an initial failure to define closely the ruling class. These criticisms, of course, do not invalidate Ch’en’s work, but they do suggest that some modification is required. 6  Of several reasons which could be cited, the most important is that extant sources provide information on only a tiny proportion of the T’ang population, and that with the exception of the Tun-huang manuscripts, no local history has survived. Twitchett, “The composition of the T’ang ruling class,” in Perspectives, pp. 47–85.

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

370

Guisso

the most comprehensive.7 In his Discussion of Surnames and Lineages (Hsinghsi lun),8 dating from about 719, he wrote that clans and lineages had been a state concern from the most ancient times: There existed a work called the Shih-pen which recorded the names, titles, and lineages of the feudal lords and great officers from the [time of the] Yellow Emperor down to the Spring and Autumn period…. After the Han arose Ssu-ma Ch’ien and his father edited the Shih-pen [for inclusion in] their work, the Shih-chi. They clarified the [origins of the] hereditary houses, following the Chou genealogies, and so came to know the origins of their surnames and clan names…. [Now] when Han Kao-ti had risen and, on foot, had gained control of the empire, he appointed his officials because of their wisdom, and conferred noble rank for merit…. He employed [only] those descendants of former princes, dukes and great ministers who were talented, and he rejected those without ability. He made no distinction between shih and shu,9 and so for the first time office [rather than lineage] was much 7  See CTS 149:3479:3 and HTS 132:3975:3. 8  The full text is found in CTW 372:7a–11b and in an abbreviated form in HTS 199:4093:4. It is the only lengthy discussion of the subject remaining from T’ang times. 9  This distinction was the most crucial in medieval status terminology and was sanctioned both by law and by custom. While shu is rendered conventionally by “commoner,” the term shih has yet to find a suitable English equivalent, and rather than use “scholar,” “gentry” or “literati,” I prefer to leave the word in its Chinese form and define the characteristics of the group. The best studies on the term are Niida’s Shina and his supplementary remarks in “Rikucho.” Miyakawa, Rikuchōshi kenkyū (Tokyo, 1956) attempts to make the distinction a precise one, and Mao, Liang-Chin demonstrates the wide range of status terminology used even within the shih class. Miyazaki’s Kuhon shows that the meaning of shih varied over time and differed in North and South. Johnson, Oligarchy, Chapter 1, pp. 5–17 elaborates upon some of this research. Working from these studies, the following salient characteristics of the shih in the early T’ang might be isolated: (a) They were generally and by social convention recognized to constitute a class distinct from the shu. (b) The criminal code granted them no concessions on the basis of social as opposed to occupational status (Johnson, Oligarchy, p. 8). (c) They were exempted from corvée, and this privilege was a mark of their social status (Miyazaki, Kuhon, p. 249, and Mao, Liang-Chin, p. 283). (d) State schools, both central and provincial, tended to admit only the shih, but Miyazaki’s view (p. 247) that education was reserved to the shih is an overstatement.

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

Imperial Power and the Ruling Class [ under Empress Wu ]

371

esteemed. Still, however, the northeastern hao-chieh10 were transported to fill the capital city. At this time the Chu and T’ien of Ch’i and the Ch’ü and Ching of Ch’u were all [considered to be] the preeminent surnames (yu-hsing)….11 The Wei [dynasty] set up the ‘nine rank’ system and established [officials called] the ‘impartial and upright’ (chung-cheng) to recommend [potential officials], and they esteemed men of noble descent (shih-chou) and looked down upon those scholars of low birth (han-shih). Power reverted again to the preeminent surnames…. The Chin and Sung followed this method and so for the first time, surnames (hsing) reached the pinnacle of importance. Therefore, the shih

(e)

Shih status depended principally upon the sanction of other shih and was beyond the power of the emperor to confer (Chou, Wei Chin, pp. 99–100). (f) The hallmark of the shih was culture: morality, high standards of conduct, loyalty to discipline and tradition, a certain commitment to noblesse oblige and to the self-consciousness of class. Within these limits, it was an hereditary status. (g) The shih were distinguished by an exclusive marriage circle and by a determination not to marry below their station (Mao, Liang-Chin, pp. 230–37). There was no statutory ban on marriage with commoners (Niida, “Rikuchō,” p. 22). (h) Status and officeholding were closely connected. Commoners found it exceedingly difficult to enter the bureaucracy and, although never barred de jure from officialdom, tended to rise only in unusual circumstances, chiefly through military prowess in periods of dynastic change. For the shih, state service was an important means of preserving their status, but it was not all-important, and culture and lifestyle combined with social recognition permitted some clans to eschew office while retaining prominence. See, for example, Ochi Shigeaki, “Nancho no koseki mondai,” Shigaku zasshi, 69:8 (1960), pp. 940–64. Shih status could be gained by office, but office was granted generally only to those who already possessed shih status. It is this final point which has led Japanese scholarship in this century to characterize the entire post-Han era as that of the “aristocratic society” to which Toyama’s book of readings Kizoku shakai provides an excellent short introduction. Johnson, Oligarchy, pp. 5–45, is a most useful new addition to the literature on the subject. 10  A term discussed later in the chapter. I prefer the translation “magnate clans” which perhaps conveys something of the close regional ties, military prowess and “feudal” organization which seem to have been common to their tradition. Liu Fang provides a useful insight here, for by T’ang times the term seems to have included a connotation of opposition to centralization and was applied most frequently to the northeastern clans. 11  The term yu-hsing is used consistently throughout the essay as the generalized description of the most important clans on the national level.

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

372

Guisso

and the shu, the noble (kuei) and base (chien 賤)12 were clearly separated once and for all. At that time when officials came to select men for promotion, they were certain to examine their genealogy (p’u) and registration (chi) and check to see that there was no fraud. Thus every official possessed his genealogical table (shih-chou p’u) and it listed the offices [of family members] for generations…. South of the Yangtse [the greatest clans] were [called] the ch’iao-hsing and the greatest were the Wang, Hsieh, Yüan and Hsiao. In the Southeast were the wu-hsing, and the greatest were the Chu, Chang, Ku and Lu. East of the mountains (shan-tung) were the chün-hsing, and the greatest were the Wang, Ts’ui, Lu, Li and Cheng. In Kuan-chung, they were also [called] the chün-hsing, and the greatest were Wei, P’ei, Liu, Hsüeh, Yang and Tu. In northern Shansi (Tai-pei) were the lo-hsing, and the greatest were the Yüan, Chang-sun, Yü-wen, Yu, Lu, Yüan and Tou. These lo-hsing were the ‘eight clans and ten surnames’ and the ‘eighty-six lineages and ninety-two surnames’ of the time when [the Northern] Wei [emperor] Hsiao Wen-ti moved his capital to Lo-yang [in 494]…. and they called themselves men of Lo-yang in Ho-nan….13 [Liu goes on here at some length to explain how the chün-hsing were ranked according to the offices held by family members into four groups, and how this was the true origin of the term ‘four surnames’ (ssu-hsing) later appropriated by the northeastern clans alone. He points out that the Northern Ch’i restricted office to those numbered in the ssu-hsing.] Formerly [during] the Southern Dynasties (Chiang-tso), their system of status for clans and lineages was to consider the first of the ranking surnames in each commandery the ‘preeminent surnames.’ During the t’ai-ho period [477–500] of the Wei, all the ‘four surnames’ of each 12  These terms have received less attention than shih and shu, probably because the latter was the more common distinction until the middle T’ang. Several translations, such as “noble and base” or sometimes “free and unfree” have been used, but these are not altogether satisfactory since chien is often used for servants, slaves, artisans, merchants, prostitutes, etc., and kuei seems often to have been highly relative and to have meant simply “non-chien.” It was replaced by the term liang “free” in T’ang law. There now exists a full study of the chien class in the T’ang. The opening chapter in Hamaguchi Tō ōchō discusses the distinction. 13  It was members of this group who were most prominent in the high officialdom of the early T’ang and who constituted Professor Ch’en’s “Kuan-chung bloc.” On the ssuhsing question which follows in Liu’s essay, see Johnson, Oligarchy, pp. 28–30, and note TCTC 140, p. 4394, where Hu San-hsing anachronistically applies the term to the first four clans of the realm. This seems to have been T’ang usage.

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

Imperial Power and the Ruling Class [ under Empress Wu ]

373

commandery were considered ‘preeminent surnames.’ The same term [‘preeminent surnames’] was used for those classified as chia-men in the Lei-li of the monk T’an-kang14 of the Northern Ch’i; for the ‘clans respected by all’ (ssu-hai t’ung-wang) in the chien-te [572–78] compilation of the Northern Chou;15 for those called the mao-hsing in the Sui compilation of k’ai-huang [581–601];16 and [finally] when the Shih-tsu chih was compiled in the chen-kuan period of the T’ang, the same term was used for those there classified as the first-rank families. Those whom Lu [Ching-ch’un] calls the sheng-men in his Chu-hsing lüeh17 were also considered as ‘preeminent surnames.’ Any [classification] which fails to accord with these practices of successive dynasties should not be called ‘genealogy’ with them. The presentday custom of referring to the Ts’ui, Lu, Li, and Cheng as the ‘four surnames’… certainly does not follow this standard rule. When culture (wen) is in decay, honor is given to false [genealogies]. The Sui was heir to this sort of decay but did not understand why there was corruption and so went against the way of antiquity. [The Sui] abolished the way of local recommendation (hsiang-chü) and did not consider persons [for selection] in reference to their place of origin (ti-chu). It valued officials only for their administration. The shih thus had no 14  The only mention I find is in T’ai-p’ing kuang-chi 184:41a, and the bibliographical treatises of the T’ang histories make no reference to the monk’s work. The T’ung-chih, chüan 66 which is probably our most complete listing of genealogical works, mentions neither this nor the next two works listed by Liu Fang. 15  I find no other reference, though the title seems to indicate that this was a statewide compilation. 16  Again I find no details of this work, and it seems odd that it goes unmentioned at the time the Shih-tsu chih was compiled. As the first work since the Han to apply to the entire country, it should have been a primary source of reference. In this connection, we might note the growing tradition in northern dynasties of compiling national status lists, and the parallel development whereby the definition of preeminent surnames came increasingly to be divorced from local prominence. It seems clear that whereas in the Wei each prefecture had its preeminent surnames, these later compilations listed only clans of wider prominence and, therefore, some prefectures would go wholly unmentioned.  It is important to note here that the term mao-hsing seems to refer to Chinese clans of long and proud lineage and suggests that the Sui was more ambiguous than is generally thought in its “anti-aristocratic” policies. 17  On Lu, see CTS 189 hsia:3577:1 and HTS 199:4092:3. His work is attested in the CTS, HTS, TFYK, T’ung-chih and Yü-hai which also mention an earlier compilation called the I-kuan p’u. He died in prison in 696 and is considered by Liu Fang as the father of T’ang genealogical studies.

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

374

Guisso

locality or village of their own; villages had no officials, and the people lost integrity and their sense of shame. The shih lineages were thrown into disarray and the shu-jen went beyond their appointed station…. When the people have nothing left to preserve, the shih lineages will be destroyed. And when the shih lineages are destroyed, the state perishes as a result.18 Liu Fang’s analysis, impressive in its detail and in its familiarity with a full range of status terminology, is important in several respects. In the first place, it identifies a stratum of local shih, a minor aristocracy which was recognized by the state to exist “in every commandery” during post-Han times and which was, in Liu’s view, a crucial factor in the maintenance of social order at a local level. Without them the state would perish and, as a result, they had been an object of concern to the throne since Han times. He points out that although rulers like Han Kao-tsu had successfully asserted their right to define the criteria of status, the practice of respect for lineages distinguished either by nobility or by a tradition of state service was a very ancient one and that as time went on, the growing contiguity between title and office produced a class with a virtually hereditary charisma which was highly impervious to imperial control. The gradual establishment of the ethic of Han Confucianism made such qualities as superior humanity and refinement, scholarship and service to inferiors the hallmark of prestige, and the means of distinction between shih and shu. There is ample evidence in the period of division that conferral of shih status stood beyond the imperial prerogative, and that heredity was an essential element in its attainment.19 When a high minister of the Sung (420–477) remarked that 18  CTW 372:7a–11b. 19  Numerous anecdotes from the dynastic histories could be cited to demonstrate this, and some are to be found in Ikeda On’s draft chapter for Volume 3 of the forthcoming Cambridge History of China, “The decline of the T’ang aristocracy.” We might also note, for instance, Nan-shih 56:1754:4 which tells of a man who had the “bearing, [official] qualification and style” of the shih and requested the Southern Ch’i emperor to confer the status. After consultation with his social arbiters, the emperor had to confess that he could do nothing since “the shih-ta-fu really stand beyond my authority [in this matter].”  Chou, Wei Chin, pp. 98–99, cites this anecdote to demonstrate the snobbery and the necessity of peer recognition which were part of the shih group, but it seems equally valid to suggest that it was lack of lineage which led to the rejection of the aspiring shih.  Wei-shu 60:2035:2 contains a memorial from Han Hsien-tsung asking that shih 士 and shu 庶, have separate living quarters in Lo-yang rather than being assigned their places on the basis of official rank. In it, Han contends that official rank is transient, its holders rising and falling. On the other hand, if artisans were to cultivate the conduct and standards

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

Imperial Power and the Ruling Class [ under Empress Wu ]

375

“between shih and shu there is a gulf ordained by Heaven,”20 he was affirming not only the validity and necessity of distinction, but also an hereditary source of charisma beyond definition by man. Liu Fang refers secondly to state recognition of shih status, most notably through the bureaucratic and legal privileges it granted in the nine-rank chung-cheng system. Although much could be said about this system and the aristocratic monopoly of power it produced, the important result for present purposes is that it helped to make distinctions within the aristocracy: Liu’s elite of “preeminent surnames” and the “blocs” identified by Ch’en Yin-k’o. Genealogical compilation became common practice both for individual clans and for the state,21 but the motives were different and, as Liu suggests, there arose a parallel divergence between the public and private criteria of status. To him this was a serious matter and, as we shall see, the first T’ang emperors shared his concern. In specific terms, Liu saw the ideal state as one characterized by wen, a word of wide connotation referring both to civil virtue and high culture. Honor, both from above and below, should be bestowed on the basis of wen, and its twin foundations were officeholding (kuan) and lineage (hsing). Moreover, as he suggests in the same essay, the clans of certain areas preserved through the generations different types of virtue: of the shih, they would not reach the status in a hundred years. He went on to say that if the sons of the shih followed the behavior of the artisans, they would become like them in a single morning.  Some scholars, for instance Johnson, Oligarchy, p. 8, see this as a denial of the hereditary principle. I believe that Han is suggesting that shih status can be lost but that “conduct and standards” are in themselves insufficient to gain the status. The emperor might bestow high office or expose non-shih to their superiors in the hope of raising their status but, according to Han, both tactics are futile.  These examples suggest, therefore, that while office and culture are attributes of the shih, they are not the determinants of the status. I shall later attempt to show the tenacity of the hereditary principle in relation to the early-T’ang attempts at status fixing and even in the form taken by its regulations concerning the examination system. 20  Said in relation to Wang Hung, Sung-shih 42:1550:4, an imperial favorite who sought shih status. Similar statements are found in Miyakawa, Rikucho, pp. 376 ff. and Chou, Wei Chin, pp. 99 ff. 21  The secondary literature on this subject is large, and T’ung-chih 66, pp. 783c ff. gives some idea of the volume of these compilations, listing 131 works in the four categories of arrangement, those by rank, rhyme, single area and single clan. Cheng Ch’iao, T’ung-chih erh-shih-lüeh 1:1b, tells us that under the T’ang, “the study of lineage and clan reached its zenith.” A good overview of the subject is the important article by Utsunomiya, “Tōdai kijin.” A number of studies cited in later notes shows the prevalance of falsification.

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

376

Guisso

The people east of the mountains are unsophisticated and honest (chih 質), and so they esteem marriage connections. Their sincerity (hsin) is worthy of praise. Those of Chiang-tso [the Yangtze valley] are highly cultured (wen) and so esteem individual worth. Their wisdom (chih 智) is admirable. In Kuan-chung, the people are brave and manly (hsiung) and so esteem officeholding. Their perception (ta) is admirable. The people of Tai-pei [northern Shensi] are martial (wu) and so esteem noble relationships (kuei-ch’i). Their breadth [of mind] (t’ai) is admirable.22 In this statement lies the problem. While great clans could base their claim to status on regional values, genuine centralization was illusory, In other words, a dynasty seeking to consolidate its rule had first to unify the definition of wen since this was the acknowledged basis of respect. Ideally, it should itself personify the highest wen and enforce conformity upon rival claimants. It seems clear from Liu’s statement, however, that intangible and long-standing clan tradition was relatively immune from imperial control, and that the successive attempts by ambitious dynasties to compile state genealogical lists had failed to achieve their aims. Only by revising this triadic relationship, by emphasizing kuan over hsing as a component of wen and as a claim to the status and perquisites of the shih class could a dynasty triumph over regionalism and incipient rivalry. And for the T’ang, the political history of post-Han China and the disintegration which accompanied the fall of the Sui made priorities clear. The following discussion is largely the story of how the dynasty sought to deal with the centrifugal tendencies of the aristocratic elite, and how under Kaotsung and the Empress Wu a solution was found which ultimately effected a change in the nature of the ruling class in China and so hastened the end of the aristocratic government which had been perhaps the outstanding feature of “medieval” China. The climate of the court in the early days of the T’ang was highly aristocratic and was seen in that light by others than Liu Fang. Su Mien, in his important encyclopedia (Hui-yao) of 803 wrote: Of the great officials who were founders of the present dynasty, all were of aristocratic clan (kuei-tsu). Since the Three Dynasties of antiquity, has there ever been a dynasty [so aristocratic] as ours?23 22  CTW 372:10a. This passage offers an interesting insight into Ch’en’s Northwest-Northeast hypothesis. The greater role of Kuan-chung in the government of the early T’ang was perhaps as much a matter of choice and regional proclivity as of dynastic policy. 23  THY 37, p. 663.

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

Imperial Power and the Ruling Class [ under Empress Wu ]

377

The imperial house, of course, was aware of this phenomenon, but between the first two emperors, Kao-tsu and T’ai-tsung, there was a great difference in attitude. The former, judging by remarks attributed to him, was both confident and proud in his lineage. Speaking once to the scion of a great northwestern clan, he said; Our clan, the Li, was in antiquity wealthy and influential in Lung-hsi and in my grandfather’s time [even] married into the imperial family. When  I raised my righteous troops, the whole empire joined me and in a few days, I rose to become Son of Heaven. Looking at the rulers of former times, many rose from humble origins, toiling in the ranks and scarcely able to sustain their lives. You, sir, are also from an eminent clan (shihchou ming-chia) which has successively held pure and illustrious (ch’inghsien) [posts]…. After a thousand years, only the descendants of you and me will be unashamed of their antecedents!24 Kao-tsu in this speech makes clear the degree to which a proud genealogy had become the source of pride, satisfaction and even self-respect by the seventh century. In his claim of relationship to the Lung-hsi Li clan—a claim almost certainly false25—he also betrays a certain anxiety. For although the wealth, the regional influence, and the status of a former consort clan with its implied concomitant of title and office which he claimed would seem an impressive enough background even for an imperial house, in the early T’ang it was not. The reason for this is the existence of a regional aristocracy whose prestige and pride was much greater. In 632, only a dozen years after his father’s boast, T’ai-tsung defined the rivalry as follows:

24  CTS 57:3292:4. The minister is P’ei Chi (Appendix B, no. 2) who, though poor in his youth, considered himself a member of the eminent Ho-tung P’ei clan which is systematically studied by Yano, “Kizoku seiji.” Note also Kao-tsu’s conversation with Tou Wei (no. 5), CTS 61:3301:1, in which he speaks in a similar vein. 25  The lengthy debate on the origins of the T’ang clan has centered chiefly on whether their lineage was purely Chinese or Hsien-pei, and Ch’en, Kuo-shih chiu-wen, v. 2, pp. 2 ff. collects most of the important data. Ch’en Yin-k’o has three articles on the subject in CYYY but his suggestion that they were a decayed branch of the Chao-chün Li clan seems to have been effectively rebutted by Chu Hsi-tsu, “Po Li-T’ang,” and Liu P’an-sui, “Li-T’ang.” Though conflicting evidence makes the problem very vexing, I am of the view that the Lung-hsi connection was false and that it should be seen in the same light as the Sui attempt to claim connection with the eminent Hung-nung Yang clan.

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

378

Guisso

At the time, the court discussed how the shih of the Northeast liked to boast of themselves. Although in successive generations they had declined, they still relied upon their former status (ti) and when their daughters married into other clans, they had [to be paid] a bride gift (p’in-ts’ai).26 T’ai-tsung disliked this, considering it very detrimental to the teachings of rectitude (chiao-i). As a result, he decreed that [Kao] Shih-lien,… Wei T’ing,… Ts’en Wenpen… and Ling-hu Te-fen27 should revise and correct the [state list of] clans. Thereupon the genealogies (p’u-tieh) of the whole empire were investigated and checked against historical records to verify their truth. The loyal and virtuous were commended and advanced, while the disloyal and recalcitrant were censured and demoted. The compilation was called the Shih-tsu chih and Shih-lien then grouped the clans in ranks and categories (teng-ti) and submitted [the work]. 26  On the bride gift, see Okamoto, “Tōdai heizaikō.” Okamoto shows that the gifts were often of very great value and as important as the letters of engagement in formalizing the marriage. Since the northern dynasties the state had attempted to define the gifts in terms of official rank, but social status seems to have been of greater importance. See also TRSI 9:27 and 30, pp. 245 ff. and Nien-erh shih cha-chi, chüan 15, v. 1, p. 197. 27  On the compilers, see respectively Appendix B, no. 15; CTS 77:3333:2 and HTS 98:3910:2; Appendix B, no. 26; and CTS 73:3325:4 and HTS 102:3918:3. The large literature on the subject is extremely well reviewed by Ikeda’s “Tōchō.” In this literature the background of the compilers is generally neglected.  The reason for their choice is difficult to know. Te-fen and Wen-pen had been the compilers of the Chou-shu in the large project of 629 and so could claim historical expertise, but neither of the other two seem to have been distinguished scholars. Both of them, however, were from relatively eminent clans, both held the highest positions in the Board of Civil Office, and both had present marriage connections with the imperial clan. The Ling-hu and Wei clans shared the background and regional origins of the T’ang, and the grandfather of Ts’en Wen-pen had been closely identified with the Later Liang and so may have been included to represent the South in an attempt to achieve regional balance.  Kao Shih-lien was probably the key figure, not only as head of the project but as the only northeasterner in it. He claimed descent from the imperial clan of Ch’i, and thus his Six Garrison origins and the fact that he was the uncle of T’ai-tsung’s wife would effectively set him apart from the four great northeastern clans against whom the list was directed.  Therefore, although the chief compilers reflected the regional balance typical of the upper bureaucracy in the early T’ang, it is difficult to escape the conclusion that the group constituted a “packed jury.” Their first verdict must have been a great surprise and attests to the overwhelming prestige of the northeastern clans and the dichotomy between office and social status which was still so strong.

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

Imperial Power and the Ruling Class [ under Empress Wu ]

379

T’ai-tsung said, ‘For a long time, I have borne no resentment against the Ts’ui, Lu, Li and Cheng [clans] of the Northeast. I consider that over the generations they have declined to insignificance and have wholly failed to produce officials. Yet still they call themselves shih-ta-fu, and in their marriages they demand huge sums in money and silk. In ability and knowledge they are deficient, but [still] they esteem themselves. They [are forced to] sell the trees from ancestral graves and rely upon [marriage to] the rich and noble. I cannot understand why the people respect them [so greatly]. Previously, the Ch’i dynasty was restricted only to Ho-pei while the Liang and Ch’en possessed no more than Chiang-nan. At that time, though they had talented men, they were minor insignificant states and not worthy of respect. Yet today the Ts’ui and the Lu, the Wang and the Hsieh are admired.28 I have pacified the [entire] empire and made one family of all the world. The shih of my court are all renowned for their achievements and the loyalty and piety of some is praiseworthy while the learning and talent of others is wide. [For this] they were chosen and employed. [Yet even] those who occupy positions above the third rank seek to become the relatives [by marriage] of those decayed, ancient families, sending them lavish [gifts of] money and silk, paying them homage.29 Now when I specially decreed the settlement of clan [rankings], I desired to honor the officials of this court. Why then has Ts’ui [Min-] Kan still been placed in the first rank?…30 Do you gentlemen not respect the 28  Ikeda, “Todai,” in perhaps the most important study to date correlates Tun-huang evidence with later sources to reconstruct the fullest list of eminent T’ang clans. Interestingly, the Wang and the Hsieh do not seem to appear in the lists. Moriya’s classic study of the T’aiyüan Wang clan, Rikuchō, identifies two branches early established in the South and remaining prominent for three to five centuries, so it may be to one of them that T’ai-tsung refers. Both of the northern Hsieh clans mentioned by Ikeda had southern branches. 29  T’ai-tsung is reported to have chosen marriage partners for the imperial clan only from the households of his meritorious ministers and to have refused even to discuss the possibility of a northeastern marriage (TCTC 200, p. 6318). His great ministers, however, men like Wei Cheng, Fang Hsüan-ling and Li Chi, did not hesitate to seek brides from these clans once their fortunes were made. This was an important phenomenon since it was not only the bride gift but the influence of and aid from in-laws at court which helped keep these clans solvent.  See also Nunome, Zui-Tōshi, pp. 358–9 and 362–3, which offers a complete list of the marriages of T’ang princesses, confirming the boycott of the northeastern elite. 30  This was Ts’ui Min-kan whom the dynastic histories call Ts’ui Kan since “min” was a taboo character under T’ai-tsung. The Ts’ui clan was composed of ten branches and

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

380

Guisso

offices and titles of my court? You were not instructed to discuss former generations but simply to choose present offices and titles for your ranking. Subsequently Ts’ui [Min-] Kan was placed in the third rank. The [revised] work [completed in 638] was made up of one hundred chüan and was promulgated by decree to the empire.31 This important incident, which permits great insight into T’ai-tsung’s vision of a united China, has wide implications. He and his father had indeed united the empire into a single family, and he was determined that status should derive from himself as head of that family and should be earned in service to the T’ang. As emperor he could hardly have been unaware of the existence of the four regional super-elites identified by Liu Fang, but he directed his attack only against the clans of the Northeast. His ostensible reasons were their pretensions, their exclusivistic posture which failed to produce central officials, and marriage customs he regarded as detrimental to established morality. It seems unlikely, however, that these exhaust his reasons and, his disclaimer notwithstanding, he was probably aware of the northeastern clans’ source of prestige and felt both inferior and insecure before it. Min-kan belonged to that of Ch’ing-ho. At the time of the list, he was vice-president of the Chancellery, a post of the fourth rank, and Wei T’ing was his opposite number.  The diction of CTS 65:3309:3 creates something of a problem here, since the passage in question is usually interpreted to mean that Ts’ui was placed above the imperial clan. I am not sure this was the case. The rather sketchy descriptions of the Shih-tsu chih which remain tell us that the work contained 100 chüan and placed all the preeminent clans of the empire in nine ranks. It seems unlikely, therefore, that there would be only one clan in the first rank and T’ai-tsung was probably objecting to equality rather than superiority.  It would be well to take note here of two possible interpolations in the parallel HTS account found in 95:3903:4. In the first T’ai-tsung is said to have instructed his compilers to “rank first the imperial house and follow it with the consort families. Demote new clans and promote the old, placing the great (kao-liang) on the right and the lesser (hanchün) on the left.” Second, just before pointing out that he wished to honor the officials of his own court, he is reported to have said, “For many unbroken generations to be kung, ch’ing, or ta-fu—that is what should be defined as a great family (men-hu).”  Both these statements are somewhat suspect since they contradict the main criterion for inclusion, that is, present official position. Since this is confirmed in every other source, the HTS statements must be rejected. They do, however, illustrate the quandary in which the compilers found themselves and suggest the possibility that the finished work was organized in such a way that the first two ranks were monopolized by the imperial and consort families with the other clans spread through the remaining seven. 31  CTS 65:3309:2. See also TCTC 195, pp. 6135–36 and THY 36, p. 664.

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

Imperial Power and the Ruling Class [ under Empress Wu ]

381

In the seventh century the northeastern ssu-hsing quite simply represented the longest, purest, and most uncompromising strain of the Confucian tradition. Centered around the historical birthplace of the sage, they had made li 禮, a respect for the Way, the core of their pride, and when North China fell into foreign hands they had refused both to flee south and leave their ancestral graves untended and, in most cases, to sully their bloodlines by marrying with the conquerors.32 When Hsiao Wen-ti (Northern Wei), for instance, accepted a consort from the Lung-hsi Li clan in 496, he had to compel the same four clans mentioned by T’ai-tsung to provide inmates for his harem and wives for his sons.33 For reasons like these, the elite of the Northeast had distinguished themselves from their counterparts in other regions of North China and had added to their prestige by the traditional severity of their clan rules and the value they placed on the maintenance of morality and ceremonial within their families.34 Their prestige arose, therefore, more from social than bureaucratic causes, and their general unwillingness to enter state service under the 32  Moriya, Rikuchō, especially pp. 135–36, shows how narrow the marriage circule of the T’aiyüan Wang, who were often grouped with the northeastern ssu-hsing, remained throughout the T’ang; and an unpublished m. a. thesis by Saitō Aiko for Tokyo University has shown in meticulous detail that this was true for the other super elites of the Northeast. The contrast with the Li-T’ang clan, whose first rulers had married into the Tu-ku, Tou, and Chang-sun clans, should be noted.  The refusal to marry with foreigners was principally a matter of pride, but a suggestion by Ts’en, Sui-T’ang shih, v. 1, p. 117, has merit. He points out that dissociation from the central power of the foreign regimes was also dissociation from exploitation of the peasantry and so became a source of popularity among the people. 33  TCTC 143, pp. 4393–6. The passage also offers some information on the Wei perception of the ssu-hsing and their relation to the Lung-hsi Li clan. The northeastern elite was not wholly successful in preserving its daughters from the foreigners. 34  For a general exposition of the clan rules, see Liu, Clan Rules. In the T’ang the two most celebrated sets of rules were those of Mu Ning and Han Shou, both from prominent northeastern clans. See CTS 155:3487:3 and HTS 163:4026:1. Among the ssu-hsing, the Ts’ui were most famous for their life style, cohesiveness and piety. See, for instance, HTS 182:4062:2, the biography of Ts’ui Tzu-yüan, and note how the qualities they instilled in members suited them for official careers and maintained them as the predominant T’ang clan. The biography of Ts’ui Hsüan-wei (Appendix B, no. 132) contains an admonition from his mother, née Lu, that as he embarked upon his official career he should not seek profit for himself or his clan but seek only honesty and integrity. Another fine example of clan rules are those of Liu Tz’u (CTS 165:3508:2) whose seat was near the capital but who had close marriage ties with the Ts’ui. See also Utsunomiya, “Todai kijin,” pp. 493 ff. and Takeda, “Tōdai shizoku.”

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

382

Guisso

foreign regimes which preceded the Sui provided them with the appearance and sometimes the actuality of martyrdom.35 Reluctance to hold offices in the capital accentuated their regional characteristics, intensified their cohesiveness and exclusivity, and dissociated them both from the exploitative nature of the foreign regimes and from the parasitical character of the collaborating Chinese elite.36 We might note, finally, that it was in the Northeast that the T’ang had found the greatest resistance to its dynastic foundation and, as we will see, had reason to suspect the area of separatist tendencies for many years to come.37 Although Liu Fang had associated the military arts with the clans of Tai-pei, there is ample contemporary evidence to suggest that, as a whole, the Northeast possessed the finest fighting men and that the term hao-chieh or hao-tsu, which seems to have been a wide generic classification for the great clans in Han times, came to connote such characteristics as self-sufficiency and martial skill, and was reserved increasingly for the people of the Northeast.38 Both in economic and 35  The best-known case is that of Ts’ui Hao whose execution in 450 led to the decimation of the entire northeastern aristocracy. See Wang I-t’ung, “Ts’ui Hao.” 36  Evidence could be adduced to support each of these points, and I suspect that these are what T’ai-tsung had in mind when he suggested that the people of Kuan-chung and Shantung were different in their way of thinking (CTS 78:3336:4). An incomplete study I have made of the data in the HTS tsai-hsiang table suggests that in the two generations from the inauguration of the Sui to the death of T’ai-tsung, the number of officials produced by the northeatern ssu-hsing was proportionately almost the same as that of six super elite clans of Kuan-chung identified by Liu Fang but, as several scholars have suspected, the proportion of central officials was appreciably lower. Interestingly, a fairly high proportion of these provincial appointments was made in the Northeast in spite of the fact that since the Sui attempts had been made to restore the Han practice of avoidance. The reasons, I think, are the same as those in the northern and southern dynasties recently explored by Kubozoe’s “Gi-Shin.” The long resistance of the Northeast to the T’ang and its continuing hostility meant that Kao-tsu and T’ai-tsung not only used intermediaries like Wei Cheng and Ts’ui Kan there but seem also to have permitted the region a greater degree of autonomy. See, for instance, TCTC 186, p. 5823.  A current debate in Japanese scholarship concerns the precise relationship between the aristocracy and the rural community (hsiang-t’ang). What seems clear, however, is that the peasants, even in the T’ang, were still in the habit of looking to the great clans in their area for relief in times of calamity. The early T’ang seems to have been slow to usurp these relief functions in the Northeast, and I shall show that Kao-tsung and Empress Wu were more active in this respect. 37  See Pulleyblank, Background, pp. 75 ff. and the literature reviewed there. 38  See, for instance, Ch’en Yin-k’o, “Hao-chieh,” which stresses the cohesion and military prowess of the group, demonstrating that T’ai-tsung’s coup depended on them. Matsui, “Tōdai,” shows that the term was also used commonly for the aristocracy of the Southwest which shared many characteristics with those in the Northeast. Kikuchi, “Setsudoshi” is Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

Imperial Power and the Ruling Class [ under Empress Wu ]

383

strategic terms, the area was a key one in the survival of the dynasty and so could not be permitted to remain aloof from the centralizing policies which T’ai-tsung considered so important. His special treatment of the area sprang therefore from two causes. The greatest clans of the Northeast enjoyed a status in the popular view which was beyond imperial control, and with their superior prestige, marriage revenues and court connections as well as their traditional military power, could threaten not only national unity but the very life of the dynasty. Status fixing by the state was an old solution but, in the newly united T’ang Empire, one with a new importance. The revised versions of the Shih-tsu chih, according to the old T’ang History, was greeted with general approval,39 but there are indications that in its stated aim of equating status with present success in the service of the central government, its effects were minimal. There was no rush of northeastern aristocrats to the capital, and in 642 the emperor was forced once more to attack the same clans, charging that they had abandoned the profession of official. He struck at a key source of their economic viability with a decree forbidding the sale of marriages (mai-hun).40 In the same edict he again commented on the pretensions of the group, and complained that new officials and wealthy families still contended to marry with them, a custom which was now well-established. Perhaps in order to justify what must have seemed a harsh order, he claimed that marriages of this sort degraded the great clans, led to friction among inlaws because of status differences and even caused the “buyers” of marriage to adopt the pretensions of the northeasterners. His deeper motives were perhaps punitive, reflecting both his frustration that the northeastern clans still remained aloof, and his determination to force them into acquiescence. If this is the case, his tactics were wrong, for as Ssu-ma Kuang remarks of a later ban, the best short treatment of the hao-chieh as a distinct group and their influence on the societal structure. 39  CTS 82:3343:4 and TCTC 200, p. 6315. I am sceptical about this in view of the forced revision, the treatment of the northeastern ssu-hsing, and the fact that the quoted passage occurs only in the biography of Li I-fu. It met approval, I think, only in relation to the Hsing-shih-lu. 40  TTCLC 110:4a, TFYK 159:7a, and THY 83, p. 1528. The ban was issued in the sixth month of 642 and is discussed with its secondary literature in Ikeda, “Tōchō,” pp. 51 ff. In it T’aitsung again attacks the northeastern clans saying that traditions and standards of conduct there have declined, and while the clans still enjoyed local prominence and renown, they had fallen into poverty and rejected the propriety of marriage by interesting themselves primarily in the wealth of those clans who wished to marry with them. They turned weddings into a business proposition. Although T’ai-tsung accused the clans of Shansi (Yen-Chao) of the same fault, his regional bias is clear. Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

384

Guisso

Since [these] great clans (tsu-wang) were the object of contemporary respect, in the end the prohibition could not succeed. Some [of the clans] secretly sent their daughters to the [intended] groom’s house, and others [permitted their] daughters to grow old as spinsters.41 In short, these two measures of T’ai-tsung must have struck the northeastern elite as expressions of regional discrimination on the part of a dynasty whose own origins were northwestern and, as a result, can hardly have persuaded them to contribute their services to it for reasons other than economic necessity. Secure in their social prestige and realizing that if status were dependent on office, the northwestern clans already had a thirty-year head start over them, they seem for the most part to have continued their policy of non-cooperation. Of the northeastern representatives among T’ai-tsung’s chief ministers not a single one was of the group of families he identified as objects of his centralizing concern.42 I mentioned earlier that Kao-tsung inherited his father’s bureaucracy and that the preponderance of northwestern influence in its highest levels has led some scholars to interpret the political struggle over Empress Wu’s evaluation as a conflict between regional aristocracies. While I have attempted to show the deficiencies of this view, I must also concede that it contains some measure of validity. The fact that it has been so widely discussed indicates at least that the problem of regional jealousy within the ruling class had not wholly been solved. The events following Wu’s elevation, most particularly her destruction of the principal northwestern tsai-hsiang and the declaration of Lo-yang as the eastern capital, created a favorable climate for another attempt to conciliate the northeastern elite, and in 659, with a new clan list and marriage ban, the attempt was made. The immediate inspiration for the revision of the national clan list in 659 was, we are told, the fact that Empress Wu’s clan was unranked or “not clearly” ranked in the Shih-tsu chih.43 While this may be true, it seems from such facts 41  TCTC 200, p. 6318. 42  See Appendix B. Ts’ui Jen-shih was from Ting-chou in Ho-tung, and I find no relationship with the eminent Ts’ui of Ho-pei. 43  The sources disagree. CTS 82:3343:4 claims that Li I-fu initiated the recompilation because his clan was unranked, and TCTC 200, p. 6315 says Hsu Ching-tsung asked for a revision because the empress’ clan was unranked. HTS 95:3903:4 and THY 36, p. 665 cite both reasons. I find it hard to believe, in view of Shih-huo’s career, that the Wu clan was unranked and prefer the THY wording: “was not clearly ranked.” This would suggest that the empress was unsatisfied.

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

Imperial Power and the Ruling Class [ under Empress Wu ]

385

as the comparatively low rank of the compilers of the new list44 and from the extant descriptions of the completed work that the motives for the revision paralleled those of T’ai-tsung two decades earlier. The chief differences between the two works is to be found in their size and scope, in the exact correlation of official and treatise rank in 659 and in the omission in the later work of all but the immediate family of the individual ranked. These differences are presented in tabular form as follows:45  Li I-fu was very conscious of his inferior clan status and falsely claimed membership in that of the Chao-chün Li clan. He had to resort to intimidation to enforce the claim but when he attempted to use his false status to marry his son into the northeastern elite, he was rebuffed. His motives for desiring a new list were therefore strong, and he may also have inspired the marriage ban issued the same year that he was attempting to marry his son. See TCTC 200, p. 6318. 44  The compilation of 659 was delegated to twelve men of whom we have the names of only five. Of these, only Li I-fu and Lü Ts’ai (CTS 79:3338:4 and HTS 107:3926:4) are known in any other context. With the exception of Li I-fu, none occupied a position above the third rank, and one even held a liu-wai post! Since the compilers represented neither great clans nor the upper bureaucracy, a new concept of status was reflected in their choice. 45  The table is based on information found in all the standard sources and follows that of Ikeda, “Tōchō,” p. 48.  The chief differences lie in the size and scope of the works. In addition, the Hsingshih-lu makes an exact correlation of official rank and rank within the treatise and deletes all but the immediate family of the man ranked. The list of 659 was much larger than its predecessor and reflects the growth in officialdom and in the number of persons rising through the military in a time of foreign expansion. The approximate number of clans and persons ranked is difficult to know because we are by no means sure that the term chia ‘家 is similarly used in both lists, nor indeed are we certain of the exact meaning of the term. Johnson, Oligarchy, pp. 91–92, sees chia as a subdivision of hsing 姓 which might be translated as “surname” or “descent group,” and this seems justified in view of the form of the list. We must be cautious, however, about equating chia here with “family” or “household,” a unit in T’ang times of five or six members, since if we did so we would find ourselves with an unbelievably small ruling class—less than .02% of the population. In the table I have chosen to translate the term chia as lineage in full realization that the rendering is unsatisfactory, but in the hope that it will underline the fact that we are dealing with a special sort of “family.” Imabori, “Todai shizoku,” convincingly establishes a difference in the size of shih and shu households, and Niida, Shina, p. 337, suggests an average figure of ten persons in what would here be termed a chia. This, of course, is about double the size of an ordinary hu. Moriya, Rikuchō, pp. 143 ff. assembles evidence of a much larger figure but, after finding that a hundred was not uncommon, is reluctant to use the figure. Judging from specific cases of clans punished by Empress Wu, I would tend to put the number of individuals in a chia at between ten and thirty in this case. See, for instance, TCTC 200, p. 6317, and 206, p. 6513.

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

386

Guisso

Title of work

Shih-tsu chih (final form 638)

No. of chüan Contents

100 293 surnames 姓 1,651 lineages 家 (a) Arranged by rank 等 (b) Imperial clan placed first, followed by the consort families. The Ts’ui clan, which had originally ranked first, was placed third.

Features

Hsing-shih lu (completed 659)

200 245 surnames 姓 2,287 lineages 家 (a) Same. (b) Ranking reputedly decided by the emperor, who also wrote a preface. [The imperial clan probably occupied a special rank at the top.] In the first rank were placed the four consort families (Tu-ku, Tou, Chang-sun, and Wu),1 the imperial clans of Sui and Chou (Yang, Yü-wen),2 the clans of those who had [in the T’ang] held the posts of san-kung,3 of t’ai-ssu,4 of t’aitzu san-shih,5 of k’ai-fu-i-t’ung san-ssu,6 of p’u-she.7 Civil and military officials of the second grade (二品) and [active] civil officials of the third grade (chih-chengshih che san-p’in) were placed in the second rank. Every [other] official was ranked according to his grade. (c) A total of 9 ranks. (c) Same. (d) Aim to distinguish shih 士 (d) Inclusion of all who had reached from shu 庶. at least the fifth grade in the T’ang period. No relatives of the man ranked, except his brothers, his sons and grandsons, were included. (e) Greeted generally with (e) Since all who had achieved the approval. fifth grade were elevated to shih 士 status, even soldiers, the gentlefolk (chin-shen shih ta-fu) were all ashamed to be ranked. They called the work the Rule of Meritorious Service to the State (Hsün-ko).

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

Imperial Power and the Ruling Class [ under Empress Wu ]

387

Title of work

Shih-tsu chih (final form 638)

Hsing-shih lu (completed 659)

Notes

(1) An observation found in the Comprehensive Mirror, and in the old T’ang History, 189, 3578:1, biography of Liu Ch’ung.

(1) The respective empresses of Li Ping, Kao-tsu, T’ai-tsung, and Kao-tsung. (2) Referred to as hsi-kung and chieh-kung. (3) The posts of t’ai-wei, ssu-t’u and ssu-k’ung. In addition to the Imperial clan, members of the P’ei, Chang-sun, Fang and Li clans had held one of the posts. (4) The heads of the Chancellery, Secretariat and Department of State which were second and third grade posts. (5) The Three Preceptors of the Crown Prince, who were all first-grade officials. Four men had held the post. (6) An honorary official title of the first class. (7) The two vice-presidents of the Department of State Affairs. Six men not included elsewhere in the first rank would be found here.

The substantive differences between the Shih-tsu chih and the Hsing-shih lu reflect as much as anything else a different tactical approach to an identical problem. As we have seen, the first work is said to have been greeted with aristocratic approval and, moreover, to have had as one aim the separation of shih and shu. Both of these attest to its traditional character, and the necessity for a marriage ban four years later suggests that perceptions of status were little changed by it. T’ai-tsung clearly was not anti-aristocratic, nor did he seek by the treatise to weaken or destroy the northeastern clans. He sought rather to woo them into direct service to the T’ang, but his openly indiscreet and

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

388

Guisso

hostile attitude and his arbitrary reduction of the Ts’ui clan’s ranking could only have had the effect of polarizing the two sides. Moreover, since he decreed that the list was to be a permanent ordinance (yung-shih),46 the proud clans of the Northeast would be forever inferior in the eyes of the state to those whom they regarded as the upstart clans of the Northwest. The Hsing-shih lu was at once more conciliatory and more firm. Like its predecessor, it emphasized state service over family as a determinant of status but even more greatly stressed the individual over the clan. The truly important change lay in the fact, therefore, that heredity was no longer a factor in obtaining shih status and that any individual who reached fifth-rank officialdom needed no other qualification to be accounted a shih. At a stroke officeholding became more desirable and more necessary, and marriage into one of the great northeastern clans no longer held the same attraction for those who sought rapidly to raise their status. At the same time, however, the olive branch of conciliation was held out to the Northeast with the collection and burning of all available copies of the Shih-tsu chih47 and the elimination thereby of the clan superiority of the Northwest. Probably for the same reasons, the new work recognized the validity of marriage as status in the high ranking of consort families and removed the imperial clan from competition with the Northeast by creating a special rank for it. The Northeast, moreover, had produced its share of regional officials above the fifth rank48 and, by providing most of the manpower for T’aitsung’s Korean campaigns, had probably come to hold most honorific ranks. On these two grounds, its representation in the second work was, in relative terms, likely to have been better than in the first. Although the new list took even less account of genealogy than the old, on balance it was probably more palatable to the northeastern clans. The same was true of the marriage ban issued four months later. This document forbade the intermarriage of seven surnames (hsing) and specified 46  CTS 82:3343:4. HTS 223 shang:4163:4 uses the term ch’ang-shih. This is rather puzzling. Since a stated aim of the compilation was to form a nexus between status and official rank, it could hardly be left unrevised and was probably intended to serve as the basis of future lists. 47  CTS 82:3343:4, HTS 223 shang:4163:3, and THY 35, p. 665. Ikeda and Johnson are both of the view that the burning did not serve its purpose since the list of 707 was to be based on the Shih-tsu chih. This shows, however, only that some copies survived and underestimates the symbolic importance of the act. In any case, the Hsing-shih lu was the standard of status measurement for half a century of rapid social change and so of greater importance than T’ai-tsung’s compilation. 48  Note 36 above.

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

Imperial Power and the Ruling Class [ under Empress Wu ]

389

limitations on the bride gift.49 Pragmatic and free of moralizing, it listed first the Lung-hsi Li clan, thereby implicitly bonding the imperial family to the six greatest clans of the Northeast who made up the remainder of the list. It tied the bride gift to present official rank, thus providing additional incentive for the aloof clans to seek high capital office. In this way the ban was directed not so much against their economic base as against their exclusive marriage circle and the cohesion which fostered their sense of regional independence. If this is the correct interpretation of the ban’s purpose, and if it succeeded in drawing more of the northeastern elite to the capital, it would seem logical to conclude that there was no real need for enforcement of its specific prohibitions. As we have seen, traditional historians record that the response of the great clans was to marry secretly or to keep their daughters unwed, and recent scholarship has empirically demonstrated that this was indeed the case.50 In spite of this, there is no evidence until after the Chou dynasty that the throne insisted on enforcement,51 and the principal reason seems to have been that there was no need to do so. From all we can gather, the two status measures of 659 seem to have succeeded in wiping clean the slate. By permitting a limited bride gift rather than outlawing the practice as T’ai-tsung had done, Kao-tsung and the Empress Wu were recognizing that the northeastern clans had at least a social claim to superior status. The Hsing-shih lu at the same time served notice that status would henceforth be tied principally and definitively to office in the T’ang. In comparison to the Shih-tsu chih, it was more in the nature of a Who’s Who than an Almanach de Gotha and so acted as an invitation to the northeastern clans to diversify their claims and compete for status on the same terms as the other regional elites. The fact that the empress, considered by contemporaries to be

49  CTS 82:3343:4, HTS 95:3904:1, THY 83, pp. 1528–29, TCTC 200, p. 6318, and Yü-hai 50:25a. In the ban seven surnames and eleven chia were forbidden to intermarry, and the bride gift was limited so that chia of the third rank or above could receive 300 p’i of silk [a p’i was 1.8 by 40 feet] and one of the eighth rank received fifty p’i. It had also been the practice, if the groom’s house were higher than the bride’s, for his chia to accept double these gifts. This was now forbidden. 50  See the studies of Moriya and Saitō cited above, and Tsukiyama, Tōdai, pp. 163 ff. 51  Chung-tsung reissued the ban, probably in 707 when the new clan list was compiled. The standard sources do not mention the event but, judging from WYYH 900:7a, this final prohibition specified five surnames, mentioning the Northern Wei head of the clan by name and numbering the sons of each, so that the descendants of a total of forty-four men were forbidden to intermarry. In the case of the Ts’ui, eight generations were represented so that the scope of the ban was wide, including perhaps 300 chia.

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

390

Guisso

an easterner and known to favor Lo-yang over Ch’ang-an,52 was by 659 being seen almost as an equal partner in Kao-tsung’s government, must have served as an additional attraction to the Northeast. It seems more than coincidence, therefore, that before Kao-tsung’s death, all but one of the lineages specified in the marriage ban of 659 had provided the T’ang with a chief minister; and in the Chou dynasty the high proportion of these same surnames in tsai-hsiang ranks further attests to the success of her reconciliation.53 Finally, we might note in this regard the form of what seems to have been the last major attempt by the T’ang at status fixing on a nationwide scale, the

52  On the fondness of the empress for Lo-yang, see Ch’üan, T’ang-Sung and Ts’en, Sui-T’ang shih, v. 1, pp. 142–47.  Both political and economic factors were present in the lengthy sojourn of the court at Lo-yang from the time of the empress’ rise, and while I accept the belief that there existed antipathy between her and the northwestern clans concentrated near Ch’ang-an, I also believe that she and Kao-tsung were attempting consciously to reconcile the Northeast. It is significant that immediately after the marriage ban in the intercalary tenth month of 659, they went to Lo-yang where they were to stay for two years. This was, incidentally, a prosperous time in the Northwest and a period of drought in Ho-pei. In 660 there was a good deal of construction in Lo-yang—palaces, bridges, marketplaces, etc., and it was also from here that Su Ting-fang’s great Korean expedition was launched. See THY 30, p. 560 and 86, p. 1577, and TCTC 200, p. 6320.  The growing importance of Lo-yang had been evident from about 656 when the first new palace in many years was built there (THY 30, pp. 551–2 and note the reasons why the first T’ang emperors did not build there), and when in early 657 the officials at Lo-yang were declared in all respects equal to their Ch’ang-an counterparts. THY 68, p. 1190. As noted earlier, Lo-yang became the Eastern Capital at the end of that year. It seems clear that the emphasis on Lo-yang reflected not only an appreciation of political and economic realities but, more specifically, was an attempt to reconcile the Northeast. In 660 when Kaotsung and the empress made a well-publicized trip to Ping-chou, the geographical origins of the empress were further emphasized. 53  The exception was the Cheng clan of Hsing-yang, and T’ang-kuo shih-pu, shang, p. 21, tells us that even in the T’ang they were considered reluctant to move from their native place. My calculations show the following distribution of tsai-hsiang from the ssu-hsing:

Ts’ui Lu Li Cheng

Kao-tsu

T’ai-tsung

Kao-tsung ( from 659)

Empress Wu

0 0 0 0

0 0 0 0

2 1 2 0

4 0 6 0

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

Imperial Power and the Ruling Class [ under Empress Wu ]

391

Hsing-tsu-hsi lu of 714.54 Though details of the work are scarce, it was altogether a more “professional” compilation than its predecessors55 and, as much as anything else, reflects a general acknowledgment that the state had won the right to set the criteria of status. The proposal for its composition in 707 remarked, first of all, that “official families (kuan-mien chih chia) have changed, rising and falling…. Judgments of former times can hardly be used as a rule for the present.”56 Chung-tsung concurred, decreeing that a treatise of eminent clans (yu-hsing) be compiled to include “men of high repute and great virtue, of pure occupation and good breeding; [men] whose titles are inherited, of the standard of well-known scholars, whose bravery has been often demonstrated, who have received honorific rank, who are of supreme distinction at this court, and who are honored in court and countryside, their fame united with general admiration…. Barbarian chieftains of notable official service should be placed in a separate category.”57 The work was to be based on the Shih-tsu chih which seems, therefore, to have escaped total destruction and was completed in 200 chüan.58 Although it is sometimes suggested that this compilation reflects the “aristocratic revival” said to follow the deposition of Empress Wu, these surviving details seem to indicate that it resembled the Hsing-shih lu rather more than the Shih-tsu chih. Once again, for instance, the emphasis was placed on the 54  CTS 189 hsia:3578:1, and see Utsunomiya, “Tōdai kijin,” p. 59 for other sources. The work was suggested in 707 and completed seven years later. See also Ikeda, “Tōchō,” pp. 48 and 56. Lü, Sui T’ang, p. 788, contends that this was the last of the state-sponsored treatises, and although there is evidence of other, much smaller works (Utsunomiya, “Todai kijin,” pp. 59 ff.), this was certainly the last large-scale manifestation of state concern with status fixing. 55  The compilers were all high officials. Headed by Wei Yüan-chung (Appendix B, no. 121) and Chang Hsi (no. 124), they included five Board vice-presidents, most of whom were concurrently appointed to the History Office. Wei’s death put a temporary stop to the project but Liu Ch’ung, who made the original proposal, continued to work with the historians Wu Ching and Liu Chih-chi and five others to complete the project. Hsüan-tsung had it corrected and brought up-to-date by Liu Ch’ung, Liu Chih-chi and Hsüeh Nan-chin. See THY 36, p. 665. The same source notes that Liu Chih-chi had recently created a stir by publishing the first honest and critical genealogy of his clan. 56  TFYK 560:20b. See also CTW 235:2a. 57  TFYK 560:21a. 58  CTS 189 hsia:3578:1. The use of the Shih-tsu chih as a model is usually explained in terms of an aristocratic revival after the Chou. I doubt, however, that this was the case as early as 707, and I suspect rather that it reflects a nostalgic wish to associate the era with “the good government of chen-kuan” and to eradicate the influence of Empress Wu in as many respects as possible.

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

392

Guisso

individual rather than the clan, and once again it listed recipients of honorific rank and foreign chieftains—the groups whose inclusion in the compilation of 659 had led to widespread criticism. On this occasion we hear of no such protest and, as a result, might be justified in assuming that in the intervening seventy years, a principle had been successfully established. The Hsing-tsu-hsi lu was designed not so much to create as to identify present membership in the shih class. In 714 success and reputation in the service of the T’ang were accepted as the sole criteria of eminence, and scholars, soldiers and “barbarians” were all eligible seemingly without reference to their antecedents. The fact that during the two generations which had passed since the Shih-tsu chih, status had come to depend more on office than on clan, on achievement rather than birth, represents a social change of truly grand significance. It is perhaps the most important historical change for which Kao-tsung and the Empress Wu were responsible. They had done no less than redefine the ruling group, the shih class, changing its principal characteristic from “aristocratic” to “bureaucratic,” and winning popular acceptance for the new view of societal stratification. One of Kao-tsung’s chief ministers summed up the change when he retired from office in 683, reportedly remarking on the three disappointments he had suffered in his life: his failure to win the chin-shih degree, to marry a daughter of the northeastern wu-hsing, and to be a compiler of the National History.59 The fact that he was the scion of one of the preeminent clans makes his preferences the more meaningful. The means by which the change was wrought, most particularly the growth of the examination system, belong properly to the next chapter, but two other aspects of it should be noted here. In the first place, the changes which occurred in the ruling class were as much qualitative as quantitative. Second, and in a directly related process, the new, more highly bureaucratized shih began to show signs of melding with its former antithesis, the shu, to form a single class of subject in a more “modern” societal configuration. Under Hsüan-tsung the change was embodied formally in law. To understand these two developments, it is necessary to return to the state registers of status. From the descriptions provided of the two works of the 59  The story is told of Hsüeh Yüan-ch’ao (Appendix B, no. 61) and is found in T’ang yü-lin 4, pp. 140–41. Interestingly, his biography tells us that he was a compiler of the kuo-shih while attached to the Hung-wen kuan and also that he was married to a T’ang princess. The anecdote may therefore be apocryphal, but if not, his preference in wives is doubly interesting. The T’ang kuo-shih pu contains several anecdotes of the sort.

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

Imperial Power and the Ruling Class [ under Empress Wu ]

393

seventh century, it seems abundantly clear that the clans counted in the nationwide elite comprised only a miniscule proportion of the population and indeed as early as the Sung, traditional scholars had pointed out that they were really the elite of an elite, existing above a generalized, locally prominent shih class who were also exclusive in their marriages and able to claim both social and bureaucratic preferment.60 As I already suggested, this situation is reflected only slightly in the state registers, with the clans listed in the Shih-tsu chih, for instance, comprising no more than .05 percent of the entire population.61 Although the Hsing-shih lu increased its coverage to 245 clans and 2,287 lineages, the proportion was probably not much greater, and since status ranking was to correspond to official rank and there were over 16,000 liu-nei officials at the time,62 this seems a situation difficult to explain. Even considering the nepotistic nature of the early T’ang bureaucracy, the majority of officials must have gone unlisted, and from this indirect evidence, it seems reasonable to assume the existence of a much larger shih stratum whose lineage made them desirable as prospective officials, but whose social standing was still too low for the national standard of registration. Until the discovery of the Tun-huang manuscripts, further development of this hypothesis was difficult. Several documents in various collections, however, have recently thrown light on the problem, though there remains a certain controversy over their exact interpretation.63 These conflicts aside, they constitute an extremely valuable source. The documents in question, four or possibly five in number,64 date from the middle of the eighth century and offer detailed information on those clans 60  See, for instance, the essay from Shen Kua’s Meng-hsi pi-t’an translated by Twitchett, “T’ang ruling class,” in Perspectives, pp. 54–56. 61  My own estimate. See note 45 above. 62  TCTC 200, p. 6308, from a memorial by Liu Hsiang-tao, discussed in the following chapter. Liu counts 13,465 liu-nei officials in 657 and says that 1,400 more entered the bureaucracy annually. The list of 659 was to include those who had reached the fifth rank, perhaps one-tenth of this figure. Of these higher posts there were a total of about 3,000 in capital and provinces so that, considering the figures in the Hsing-shih lu, it seems clear that we are dealing with a highly inbred bureaucracy. 63  There is a fairly full literature on the subject, including the studies of Ikeda, Takeda and Twitchett mentioned earlier. In addition, see Niida Noboru, “Stein,” and Mou Jun-sun, “Tun-huang.” 64  The most important are Peking wei 79 and Stein 2052. Four fragments collectively numbered Stein 5861 and Pelliot 3191 may, according to Ikeda, be part of a whole. Another Pelliot fragment, no. 3421, may be related. I had intended to discuss these documents in detail because I am reluctant to accept all of Ikeda’s conclusions, but Johnson, Oligarchy,

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

394

Guisso

considered to possess social prominence in the various chou and chün of the empire. Collation of them with later geographies and rhyme dictionaries65 reveals the existence of a second level of shih lineages which was perhaps five times the size of the super elite of the Shih-tsu chih.66 This configuration is the chief explanation for the form taken by social development in the second half of the seventh century. Let me explain further. In the early T’ang, prestige, influence, and the perquisites of rule belonged to the shih class, but an inordinate share of social status was monopolized by a tiny, nationally renowned super elite concentrated east of the T’ai-hang Mountains. The following map, based on manuscript wei 79 in the Peking National Library, illustrates the regional disparity and shows the distribution of the aristocratic groupings identified by Liu Fang. As I suggested earlier, the northeastern clans were exclusivistic, proud and, largely through marriage practices, self-sufficient. Their concern with genealogy, clan solidarity and Confucian ritual was at least partly an attempt to remain distinct not only from the other regional super elites, but also from a wider stratum of local shih which aspired to the same heights. Provided the criteria of status remained unchanged, their security was assured. The attempt begun by T’ai-tsung to equate status with present office, essentially a measure designed to balance rather than weaken the great clans, changed the situation. Whatever the failings of his policy, the emperor had at least recognized the use to which the local shih could be put, recognized their aspirations and their adaptability and thus, it will be recalled, had canvassed the genealogies of the entire empire rather than simply concentrating on the areas where the super elites were concentrated. Kao-tsung and Empress Wu had extended this principle in 659, attempting further to blur distinctions within the aristocracy while at the same time conciliating the alienated Northeast. For them the second part of the policy was of greater immediate importance, and until the 670s it reflected the need for unity, for acceptance on the part of an upstart dynasty, and for a cooperative Northeast to support the Korean pp. 64–87, has recently outlined the relationship and filiation of the documents with great care. I am not sure, however, that his view of the purpose of the lists is convincingly established. 65  Principally the T’ai-p’ing huan-yü chi, a Sung geography by Yüeh Shih, the Kuang-yün rhyming dictionary, and the genealogical work Ku-chin hsing-shih shu pien-cheng. See Ikeda, “Tōdai,” pp. 314–18. 66  Because the Tunhuang documents give only hsing and not chia and are so very fragmentary, comparison is difficult. Still, they list about five hsing per prefecture, while the state lists of 638 and 659 average less than one, so that even allowing for uneven distribution, the multiple of five seems reasonable.

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

Imperial Power and the Ruling Class [ under Empress Wu ]

395

campaigns. Imperial success in these aims was not total, but a good deal of evidence can be adduced to show that by Kao-tsung’s death representation of the national elites was more balanced in the chief ministerial ranks than had been the case in the 650s and, as will be shown later, that they came to dominate such governmental organs as the Chancellery and the Board of Civil Office.67 67  To avoid a major inconsistency of Professor Ch’en Yin-k’o, I must specify here that the term “national elite” henceforth includes the preeminent clans of the Northeast. These were, I have suggested, well represented in the Hsing-shih lu and increasingly accepted high posts in the central government from 659.  Analysis of the T’ang tsai-hsiang, or chief ministers, has recently become conventional among scholars despite the shortcomings of the group as a representative sample. Johnson, Oligarchy, Appendix I, pp. 204–206, outlines some of the problems and advantages of using the sample and on p. 138 offers the following breakdown in “chief ministerial man-years”: Period

Great clan (claimed)

Non-great clan

618–626 (Kao-tsu) 627–649 (T’ai-tsung) 650–659 (Kao-tsung) 660–683 (Kao-tsung and Wu-hou) 684–704 (Wu-hou) 705–711 (Chung-tsung and Jui-tsung) 712–735 (Hsüan-tsung) 736–755 (regime of Li Lin-fu)

91.8 (38.8)% 63.2 (15.9) 40.0 (8.0) 53.3 (1.8) 53.1 (0.4) 56.3 (1.0) 52.2 (4.4) 50.0 (38.6)

8.2% 36.8 60.0 46.7 41.7 40.6 47.8 50.0

Although Professor Johnson seems to recognize the existence of unusual situations in 618–626 and 736–755 which might distort his conclusions, he uses these figures as his principal evidence in refuting Ch’en Yin-k’o’s notion of the rise of a new class during and after the regime of Empress Wu. I do not accept his view.  In the first place, the use of “man-years,” even if combined with the very different matter of individual appointment, can be extremely misleading and is valid only when placed in the context of the evolution of tsai-hsiang policy and the differing views of the institution held, for instance, by T’ai-tsung and the Empress Wu. T’ai-tsung’s “usurpation” and Empress Wu’s destruction of the entire tsai-hsiang group who had opposed her rise were events that were perhaps atypical but nonetheless distort the percentages of such a small sample to an unacceptable degree.  Finally, and this is a more serious methodological objection, the possession of great-clan status did not prevent a man from being at the same time an “examination bureaucrat.” Appendix B of the present work shows how rapidly even members of the greatest clans turned to examination as a qualification for office in the later part of the seventh century.

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

396

Guisso

During this same period, however, the bureaucracy was expanding rapidly,68 and if the regional distinctions within the super elite were being diluted by their common bureaucratization, the group as a whole found itself pressured from below by the inferior layer of shih lineages who were rising in the bureaucracy through examination and the decreasing imperial emphasis on genealogy. Having committed themselves to court politics and become vulnerable to their vicissitudes, the great clans found themselves on the defensive in the maintenance of the prestige they had so long monopolized. To compensate for their surrender to one form of definition of status in terms of state service, they seem to have responded by intensifying what they still monopolized, the purity of their Confucian tradition.69 When Kao-tsung died and the Empress Wu seized control of the government, they found themselves, therefore, in an impossible position. As representatives of the Confucian opposition to a woman ruler and as possessor of great political influence, they became the primary objects of the terror, and over the next decade hundreds of families suffered extinction or eclipse, in many cases taking generations to recover.70 Though monographic studies on  Professor Johnson concludes on p. 139 that “the examination system had very little effect on the social composition of the political elite in the T’ang dynasty.” This is true only in the narrowest sense. Equation of the tsai-hsiang group with “political elite” is tendentious, to say the least and, moreover, ignores the changes which the examination system began to occasion even in this narrowly-defined political elite. The typical great-clan chief minister during Kao-tsu’s reign was very different from his counterpart during Hsüantsung’s reign. To obscure this fact does no service to our understanding. It seems to me that in dealing with the tsai-hsiang sample, questions of a different sort might be asked and I attempt to do so in subsequent sections. I deal with aristocratic domination of the Chancellery and the Board of Civil Office in Chapter 7. 68  See Chapters 5 and 9. 69  See Utsunomiya, “Todai kijin,” pp. 81–95. The family codes, private genealogies, and continued exclusivity of marriage were symptoms of the phenomenon, and anecdotal collections offer many examples of the pride of clan tradition continuing to flourish. See, for example, T’ang kuo-shih pu, shang, p. 20, for Li Chen 李稹 who refused to use his official titles, always signing himself “Lung-hsi Li Chen.” Several other interesting points are raised in the excellent article by Takeda, “Tōdai shizoku.” 70  In order to clarify somewhat the full extent of the tragedy of so many great clans during the period of the Empress Wu, I provide below a list of those exiled and executed for real or alleged crimes. In most cases their immediate families were exiled or enslaved and their property confiscated. Sources are given only for those who do not appear in Appendix B.  (a) In 659, at the instigation of Hsü Ching-tsung, a final revenge was carried out on the former tsai-hsiang Chang-sun Wu-chi, Ch’u Sui-liang, Han Yüan, and Liu Shih. In every

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

Imperial Power and the Ruling Class [ under Empress Wu ]



397

Distribution of Great Clans During T’ang. Locations of chou with one or more great clans from Peking wei 79; locations in parentheses from Ikeda On’s reconstruction from T’ai-p’ing huan-yü chi.

case, their immediate families were enslaved, those of sons and grandsons exiled to the far South, and several collateral relatives, nephews and cousins, also executed (see CTS 183:3553:1 Chang-sun Shang). When the empress pardoned Chang-sun in 674, there remained only a single grandson to carry on his ancestral sacrifices, and the others were pardoned only in 706, in the last will of the empress (CTS 7:3078:1). It may be noted that the Liu family, for example, was in desperate straits economically when the pardon came, and Liu Shih’s grandnephew had to beg from Hsüan-tsung the money to bury properly his ancestor. In the same memorial, he pointed out that all five families punished at the time [including, I think, that of Lai Chi, who suffered only exile] were in a similar position. See CTS 77:3334:4.

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

398

Guisso

the subject are lacking, it seems safe to assume that the great aristocracy suffered a severe blow under the empress’ regime and that had they clung more tenaciously to their regional bases, they might have attained a much higher rate of survival. This, of course, lies in the realm of speculation and, in any case, it is difficult to see continued aloofness as a viable alternative for them. The message  In 665, the adopted son of the deposed Empress Wang, who was degraded to commoner when tied to the “treason” of Chang-sun Wu-chi and the others, was executed for complicity with Shang-kuan I (TTCLC 39, and CTS 4:3072:4).  (b) In 665 Shang-kuan I, who was a poet of renown as well as scion of an eminent Shan-chou clan, was executed for his part in persuading Kao-tsung to depose the Empress Wu. His only son died with him and the rest of the family was enslaved. Shang-kuan’s son had married into the Cheng clan, and it was their daughter Shang-kuan Wan-erh who, after rising from palace slave to private secretary to the Empress Wu, came to exercise such power in the reign of Chung-tsung and was able then to rehabilitate her grandfather. She, of course, perished in the coup against the Wei faction.  (c) In 681 a small rebellion in Ch’ang-chou led by Li Lung-tzu was snuffed out and the leader executed (HTS 3:3640:2). I am unable to establish his lineage.  From 659 until the rebellion of Li Ching-yeh, with the sole exception of the Shangkuan affair, the regime of the empress was singularly bloodless and compares favorably with that of almost any other T’ang ruler. It is difficult not to notice that this was the period in which she sought, above all, consolidation and centralization, when she was wooing the elite of the Northeast to the center.  (d) In the aftermath of Li Ching-yeh’s rebellion of 684, the carnage was great. The most prominent of the clans to fall were those of Li Chi whose own tomb was desecrated, T’ang Chih-ch’i (CTS 85:3348:4), Lo Pin-wang (CTS 190 shang: 3581:3), and Tu Ch’iu-jen (HTS 106:3924:3). Only Lo Pin-wang was of undistinguished lineage. Of capital officials the list is much greater; P’ei Yen and the general Ch’eng Wu-t’ing were largely self-made men but at the height of their influence, while Yang Shen-jang (CTS 190 shang:3581:1) and Yüan Wan-ch’ing (CTS 190 shang:3581:1) and Yüan Wan-ch’ing (CTS 190 chung:3581:4) were of old and eminent families, the former a maternal relative of the empress. Several of the other victims (Hsüeh Chung-chang, Wei Chao, Liu Ch’i-hsien) are without biographies but bear surnames appearing in the clan lists and appear in the HTS treatise as members of great clans.  The most interesting aspect of this incident is that, with the exception of the Yangchou ringleaders, officials above the fifth rank suffered the ultimate penalty, while those below that rank were either demoted or exiled with their immediate family. In Yangchou itself, those below the sixth rank were simply disenrolled (ch’u-ming). See CTS 190 shang:3580:2, biography of Liu Yen-yu.  There is no complete list of the victims of the long series of reprisals which followed the suppression, since the evil officials in charge of the investigations “forced every man indicted to implicate tens and hundreds of others. The eminent families (i-kuan) trembled in fear …” (CTS 186 shang:3565:1).

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

Imperial Power and the Ruling Class [ under Empress Wu ]

399

of the state registers and marriage bans was difficult to ignore and, as we will shortly see, from about the time of the examination reforms of 669, imperial recruitment policies were directed largely toward the provincial elite whose  The persecutions concerned with the uprising ended with the Act of Grace of 690 when all but the ringleaders received a complete pardon (yüan-mien), and it is interesting to note that by 801 even Ching-yeh’s descendants had flourished and became prominent in the semi-foreign area to which they were exiled (CTS 87:3314:3). The major reason for the higher recovery rate of the families involved was the relatively short period before amnesty, and a legal precedent supported by the Empress Wu on a personal plea from the celebrated censor Hsü Yu-kung (CTS 85:3349:1; and HTS 113:3939:3). By contending that the penalty of chi-mo which entailed confiscation and registration of household members as chien 賤, “unfree,” in the census registers could not be applied to the family of a man already dead when the charge was brought, he saved over 300 families, reducing their punishment to exile. See TT 169, pp. 8940 ff. and TFYK, pp. 616–17, which both contain several other examples of how Hsü Yu-kung successfully pleaded for clemency before the empress and so set several legal precedents. Paradoxically, the law seems to have become more lenient during the terror!  (e) The week-long Princes’ Rebellion of 688 ended in the wholesale destruction of the T’ang family. At least forty-four princes were executed and over one hundred chia destroyed, with many others barred from holding office (mei-kuan). The sources, with some exaggeration, claim that only a few children, exiled in the remote South, survived. At the same time, the evil official Ch’iu Shen-chi (CTS 186 shang:3564:1) seized the opportunity to implicate in the rising former exiles in Ling-nan and, as commissioner to the area, carried out a brutal massacre, drowning over 1,000 families.  A complete list of casualties is impossible to draw, but we are told that particularly in the two centers of the rebellion, Po-chou and Yü-chou, they were great. In the former over 1,000 families were killed, in the latter 6,700 persons were convicted of complicity which usually carried a sentence of exile, and 5,000 persons suffered the penalty of chi-mo (CTS 89:3356:2, biography of Ti Jen-chieh). Several capital officials, including the empress’ own son-in-law, were also executed. TT and TFYK, loc. cit., show that the empress presided personally over many of the trials at the time, and she appeared both competent and just.  (f) In 689 several of the highest capital officials, including the tsai-hsiang Chang Kuang-fu, were executed for conspiracy with the brother of Li Ching-yeh who had escaped his place of exile. Several more, including Wei Yüan-chung, were sentenced to death and amnestied only on the day of execution. There are a number of historiographical problems surrounding the event; see the k’ao-i notice for TCTC 204, p. 6460.  After the declaration of the Chou dynasty and the elimination of the T’ang family, there occurred a respite in the terror, and only in 697 was there another large incident.  (g) In the first month of that year, a provincial prefect, Liu Ssu-li, conspired to rebel and, when word leaked out, was executed with his closest associates. At the time, Lai Chün-ch’en, seeking to regain his lost favor, managed to convince the empress that numerous high officials in the capital were involved, with the result that over thirty chia were destroyed. CTS 57:3293:2 has a good account of the incident.

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

400

Guisso

self-interest, traditions, and competence made them natural alternatives to the great aristocrats in the highest ranks of officialdom. Lest this group displace them by default, the latter were forced to respond to bureaucratization and, judging from extant lists of high officials, this was as true of the other regional super elites as of the northeastern wu-hsing. As we shall see, even the scions of great clans applied with increasing frequency for the state examinations. These developments served, therefore, not only to initiate change within the aristocracy but also within the bureaucracy. By the Chou and thereafter, officialdom was not only growing but, in the lower ranks at least, its highly aristocratic character was becoming progressively diluted. The super elites, still well entrenched in high office, found themselves pressured from below by the lesser shih and at the same time subject to the systematic policies by which the empress attacked the upper ranks of the bureaucracy. There was no need for status registers and marriage bans between 659 and the end of the Chou principally because the great clans had been enticed into a larger and more “egalitarian” bureaucracy and so had ceased to threaten the state from a position of independence. Representative of the emerging pattern is Li Ching-hsüan who was a member of the eminent Chao-chün Li clan, entered the bureaucracy through recommendation, recruited “myriads” each year, mostly from the Northeast, as president of the Board of Civil Office married three times into the northeastern aristocracy, and was disgraced in 680. His whole family was later exterminated.71  I count during this incident a total of thirteen high officials executed, among them members of such eminent clans as the T’ai-yüan Wang, the Lung-hsi Li, and the Yü-wen. The celebrated genealogist Lu Ching-ch’un also died. In many ways this event was a turning point in the dynasty. Ssu-ma Kuang remarks that so eminent were the victims that “the whole empire was angered,” and the empress, of course, destroyed the evil officials later in the same year and almost immediately summoned Chung-tsung back to the capital.  The six incidents cited above account for most of the empress’ victims, though the list is by no means complete. The years 690 and 691 saw the execution of eighteen other officials above the fifth rank, several of them tsai-hsiang and many, members of great families. Most perished for opposing Wu Ch’eng-ssu’s ambition to be named crown prince— a most natural cause and a fine example of the intensification of the tradition which initially had given the elite members their status. In almost every case the penalty of chi-mo was applied to the victim’s family, but in several Acts of Grace before her deposition the empress specifically amnestied those who for complicity had been exiled or enslaved (ch’ang-liu chi yü-nu). I hope to complete soon a separate study on the recovery of these families. 71  Appendix B, no. 59. Interestingly, Li first became vice-president of the Board of Civil Office in 669 when the first major reform of the period was undertaken. Just as Ch’en Shu-ta (Appendix B, no. 7) was earlier credited with bringing many southern shih to Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

Imperial Power and the Ruling Class [ under Empress Wu ]

401

The second major social change which occurred in the last half of the seventh century, the blurring of the distinction between shih and shu, was closely related to the policies discussed above and was equally the result of conscious imperial aim. Although the change was not completed until the k’ai-yüan period (713–42) and, although it is difficult to say whether T’ang policy generally or the special situation of the Empress Wu was a more important factor in the change, the end of the distinction may be viewed as an event which truly marked the end to the “aristocratic” epoch. Until the Sui dynasty it seems clear that the shih-shu division had had at least a quasi-legal basis72 and, as we noticed earlier, T’ai-tsung’s Shih-tsu chih had as one of its aims the preservation of the distinction. The Hsing-shih lu, however, embodied a major departure from this principle and had included in its rankings such groups as the recipients of honorific rank and foreign chieftains who were, in most cases, not of the shih class. And from 659 the policy of the throne seems to have sought further to erode the distinction. The principal tactic, as we shall shortly see, was the extension of the examination principle and the wooing of liu-wai officials, but after 684 we find a much wider range of specific measures designed to achieve the same end. Several edicts, for instance, called specifically for the recommendation of virtuous men, “whether shih or shu” to fill official positions, and in 685 talented commoners (pai-hsing) were given the right to recommend themselves for posts.73 During the terror servants and even slaves were given the right to denounce their masters and to travel at state expense to the capital to make their reports, while in the great Khitan rebellion, slaves and convicts were permitted for the first time to enlist in the army and to be eligible, therefore, for the various military rewards.74 These measures were of the sort responsible for what Professor Ch’en has termed the “social flux” of the empress’ period but, generally, were measures of expediency and need not be seen as any more than the most striking examples of an evolving new perception of status dating at least from 659. A systematic analysis of the edicts of the Chou period shows that the term shih is almost invariably used in its bureaucratic sense, the connotation shifting from social class to official

accept capital office, Li is credited with recruiting numerous men from the Northeast. In all, he served in the Board for eight years. 72  As, for instance, in the form of census registration. See the standard study by Masumura, “Kōhakuseki,” and the studies cited in note 9 above. 73  WYYH 462:9a and CTW 95:14b. See also TCTC 203, p. 6435. 74  On the informers, see TCTC 203, pp. 6438–39, and on slaves and convicts, TCTC 205, p. 6507 and THY 86, p. 1569. The last measure inspired a strong protest from Ch’en Tzu-ang. See Ch’en Tzu-ang chi, pp. 178 ff. Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

402

Guisso

function.75 Conversely, the term shu, which appears in edicts much less frequently than such terms as min or pai-hsing, is used for any person who held no rank in the official hierarchy. The effect of this practice, which included the substitution of a distinction between “free” (kuei or liang) and “unfree” (chien 賤), was soon to be recognized in law and embodied in the Code of the T’ang (T’ang-lü shu-i) by 737.76 It seems fair to say, therefore, that the regime of Wu Tse-t’ien marks a definite stage in the evolution of status perceptions in medieval China. In specific terms, it was the state status register she inspired, and her own edicts (ling) which were largely responsible for the creation of the simplified, two-level model of stratification upon which the lü-ling system of the T’ang was thenceforth to be based.77 In summarizing the role of the Empress Wu in the social change which occurred in her era, we must take care not to give her too much credit for acting consciously to achieve this end. While her unorthodox governmental role was clearly a catalyst, she remained in many respects typical of her times. In her regime we find the customary sumptuary regulations directed against merchants and artisans and at least one measure designed to limit the official advancement of such groups as astronomers, musicians, physicians and diviners.78 We have seen how in 683 she took great pains to embellish her own genealogy, and on other occasions she was not above crude snobbery. Objecting once to the family into which her daughter planned to marry, she asked, “How can my daughter become the sister-in-law to farm girls (t’ien-she nü)?”79 On several occasions she affirmed her belief in the hereditary nature of good and evil character and the effect of geography upon it.80 Nonetheless she was a pragmatist 75  See, for instance, CTW 96:4b, 96:6a, and 96:11a. 76  TLSI 22:3, pp. 93 ff. The material is arranged in a very convenient chart in T’ao Hsi-sheng, Chung-kuo cheng-chih ssu-hsiang shih, v. 3, pp. 249–51. 77  The most important recent work on the system has been done by Ikeda On. See his fine summary, “Ritsuryō” where on p. 289 he offers some cogent remarks on the interaction between status and lü-ling systems. See also, in the same volume, Ogata, “Ryōsensei,” which offers a tabular presentation (p. 354) of the shift in usage from shu to min. 78  TTCLC 100:3a, CTW 95:9b, and THY 67, p. 1183. For instances of sumptuary regulations directed at merchants and artisans, see THY 31, p. 572 and CTW 96: 19b. 79  TCTC 202, p. 6402. 80  For the former, see CTW 95:12a and TTCLC 114:7b, an edict of 701, forbidding the descendants of Yang Su to hold capital office, and CTShih 5, p. 57 where she speaks of the virtues of her own ancestors. The influence of geography on character was a prevalent idea of the age as we have seen in Liu Fang’s essay and Lo Pin-wang’s polemic. Pulleyblank, in Historians, pp. 135–66, shows that Liu Chih-chi expressed the view in his historical

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

Imperial Power and the Ruling Class [ under Empress Wu ]

403

before all else and because of her background, her Buddhism and her evolving dynastic aims, she was less inflexible than her predecessors in promoting the rise of worthy and talented men. Finally, the direction of social change and status perception must be related to the most basic of the empress’ aims, her attempt to establish first the T’ang, then the Chou in an age when regional and clan-centered traditions were paramount and when dynasties originated in the great clans. This meant that the status claims and the prestige of the super elites not only in the Northeast but everywhere had to be subordinated to those of the state. She and Kao-tsung together took the first step, following T’ai-tsung’s lead, but the development of the policy, particularly the patronage of the examination system and the lower stratum of shih whose rise it fostered, seems to have been more her own policy than her husband’s. This theme we shall discuss in the following chapter. Once officeholding had become the sine qua non for achieved status and the road to high office had been opened more widely, the greatest clans found themselves on the defensive. They possessed, after all, no monopoly on talent. When they finally acknowledged state control of status by their rush to gain office, their invulnerability was ended. As already mentioned, from 684 onward so many suffered execution or the penalty of “eradication of name” (ch’u-ming) that entire clans were suddenly plunged into economic hardship and social obscurity.81 In spite of this, the empress’ policies were not intrinsically anti-aristocratic. She was opposed only to those characteristics of the great clans which she saw as detrimental to state unity and to the full achievement of imperial power and administrative efficiency. In this opposition, however, lay the seeds of the conflict within the ruling class which would be so pronounced a feature of her own dynasty, and which would help to cement the changing perceptions of status which had evolved in the preceding decades. There is little evidence that her regime fostered the rise of any new class, but it did facilitate the advancement of an already defined secondary shih class and, perhaps of greater importance, it hastened the decline of the proud oligarchy whose monopoly of political power was the outstanding characteristic of post-Han China. works. Empress Wu exhibited the same feeling on several occasions as, for instance, in her decree executing Lai Chün-ch’en, and sometimes when she honored other officials. See CTW 95:11a and HTS 109:3930:2.  This idea may help to account for the popularity of northeastern marriages and the propensity of so many officials to claim native places with which they had only the most remote connection. 81  Explored in earlier notes, especially note 70 above.

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

The Chou Dynasty [of Empress Wu] Richard W. L. Guisso The Chou dynasty, which lasted from October 690 to February 705, was one of the shortest in Chinese history. The restoration of the T’ang, moreover, occurred not with the death of the Empress Wu but with her deposition. Why, after capping so long and successful a political career with China’s only female usurpation, did Wu Tse-tien end her career in failure? But, one may ask, was this denouement actually a failure? Was the empress, first of all, attempting to establish a dynasty which would endure for the conventional “myriad generations”? In all likelihood, it seems not. I stated earlier that her loyalty to the T’ang, or at least to its aims and achievements, was genuine, and that her regency and possibly her usurpation were undertaken in response to perceived crises. Beyond this she seems to have given signs at the dynastic changeover that the T’ang was not to be deprived permanently of the empire, and that her aim was to legitimate her own reign rather than to found a dynasty which the Wu clan would continue. These signs were both symbolic and substantive, and the first of them appeared at the very moment of the founding since there is no evidence that Jui-tsung formally abdicated. Since the third century AD dynastic founders had invariably used ritual abdication as a formal preface to the transfer of power,1 and thus from the beginning the source of the Chou mandate was questionable. In substantive terms, it is important to note further that Jui-tsung was immediately declared his mother’s successor, and although his title was huang-ssu rather than Source: “The Chou Dynasty,” in Richard W. L. Guisso, Wu Tse-t’ien and the Politics of Legitimation in T’ang China, Bellingham: Program in East Asian Studies, Western Washington University, 1978, 126–54. 1  The question of ritual abdication is discussed by Carl Leban in a paper titled “The accession of Ssu-man Yen, ad 265: Legitimation by ritual replication” (Asilomar, 1975). In the case of the Chou, it is important to note that the annals of the empress in both T’ang histories state simply that she “raised her title” to emperor, and Ssu-ma Kuang records that “the empressdowager accepted (k’o) the request of the emperor and the ministerial body … and elevated her title to sheng-shen huang-ti.” TCTC 204, p. 6467. Jui-tsung does not seem to have signed the petition which requested the foundation of the Chou, but rather to have asked simply that his surname be changed from Li to Wu. TCTC, loc cit. In 705 the CTS, HTS, and TCTC use the term ch’uan-wei, “transmit the throne” for Chung-tsung’s reassumption of the position of emperor.

© koninklijke brill nv, leiden, ���9 | doi:��.��63/9789004380158_014

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

The Chou Dynasty [ of Empress Wu ]

405

t’ai-tzu,2 and although he had adopted the Wu surname, he held his position until 698, and then it was his elder brother Chung-tsung who replaced him.3 Although the latter also accepted the Wu surname, contemporaries must have seen this as simple expediency, and the important fact is that throughout the Chou the succession never left the hands of the Li. The ambitions of certain members of the Wu clan were to make the succession question the most acrimonious and bloody of the period but, as I shall show, the empress only briefly considered the possibility of a Wu successor. She may in fact have anticipated both the struggle and its form, since even prior to the restoration she began to construct a close web of familial relationships between the Wu and Li clans,4 and was to continue to do so in subsequent years. In 699 she formalized this policy by having the leading members of both clans swear an oath of concord at the ming-t’ang.5 A second important sign which pointed throughout the dynasty to the likelihood of a T’ang restoration was the lack of genuine instructional-symbolic innovation by the empress. In this she broke the well-defined pattern both of conquering dynastic founders like Han Kao-tsu and usurpers who planned an enduring dynasty like Sui Wen-ti. We find, for instance, no new law code, the foundation of no institutions of lasting importance, and no attempt at widespread social reform. With few exceptions, even the ritual coloration of the Chou varied little from the T’ang, and though red banners replaced the yellow of T’ang,6 and “Wu” replaced “T’ang” in place names throughout the empire, court dress and ceremonial, official titles and office names remained largely unchanged after 690. In view of the empress’ demonstrated concern with ritual and nomenclature, it is particularly striking that she seems never to have chosen a “virtue and element” (te-yün) for the Chou.7 Chung-tsung, as we have noted, could remark at the restoration that the empress “did not disregard 2  C TS 6:3076:1, HTS 4:3641:3, and TCTC 204, p. 6467. Jui-tsung’s heir was at the same time made huang-sun. 3  Notes 124 and 139 below. 4  Some of these relationships are outlined in Ch’en, “Hun-yin chi-t’uan,” and Fitzgerald, Empress Wu, p. 147. See also TCTC 204, pp. 6466–67 for T’ai-p’ing’s marriage to the empress’ nephew and the conferral of the Wu surname on the Ch’ien-chin princess in 690 just prior to the usurpation. 5  T CTC 206, p. 6540. 6  H TS 4:3641:1 and TCTC 204, p. 6470. 7  Professor Hok-lam Chan has recently pointed out to me that the T’ang did not formally adopt the earth power until 750. Since this is the case, we might assume that the empress had no need to formally choose an element for her dynasty. On the other hand, the tradition of doing so had become well-established in the period of division, and in the early T’ang there were

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

406

Guisso

the old things,” and on grounds like these it seems difficult to avoid the conclusion that the Chou is best seen as a caretaker regime and, indeed, was regarded in this way by the Empress Wu herself. I am not saying, of course, that this constitutes a full explanation for what seems the relative inactivity of the empress during the Chou in comparison with the decades prior to 690. Other factors were certainly present, the most important being the growing inflexibility of the central government as the T’ang gained greater security, and also the relative success of its gradualist approach to the paramount state concerns of consolidation of imperial authority at the expense of the aristocracy. I have earlier attempted to show how closely the empress identified herself with these policies. In addition, we must recognize that throughout the Chou the empress’ position remained rather tenuous. She must have known that the acceptance of her claim to be emperor was based chiefly on personal or charismatic factors, but that at the same time a female Son of Heaven was a semantic and cultural violation of the Confucian tradition to which the elite among her subjects, the group through whom she had to work, gave perhaps their deepest loyalty. For this reason, her freedom of action was restricted, and she found herself more subject to the wishes of the bureaucracy than might have been the case with a male ruler of similarly absolutist temperament. Her political defeats, which ranged from matters like a failure to effect a planned reorganization of the provinces in the mid-690s to her inability to protect her favorites, were more numerous than is generally recognized, and although she was able to continue the trend toward dilution of ministerial power, it is of great importance to recognize that without the presence of her husband or son on the throne, she was less effective in government than she had been prior to 690. These considerations, although suggestive, still do not throw much light on our earlier question of how the empress was able to endure even for fifteen years in her anomalous position, and how eventually she invalidated her claim. As will be pointed out, her sensitivity to bureaucratic loyalties and her skillful manipulation of the succession were factors in the maintenance of her position, but for a wider explanation we might turn to a traditional source. The New T’ang History, comparing her success to the failure of Empress Wei, says “reward and punishment came from her alone and she did not borrow [the authority of] ministers. She usurped above but governed well below (chien yü shang erh chih yü hsia) and so could rule the empire.”8 If this is the case, extensive debates on the matter in which Wang Po (648–75) proposed the scheme which was eventually adopted. See Lü Ssu-mien, Liang-Chin, pp. 1469–72. 8  H TS 76:3869:2.

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

The Chou Dynasty [ of Empress Wu ]

407

it might be argued that the empress used the same combination of political acumen and popular support to which she owed her rise as the basis of rule during most of the Chou. It was only in her last years, after she had delegated authority to unworthy favorites, permitted corruption, and neglected the state, that she was forced to retire. It is difficult to tell if her deposition occurred because she had lost popular support, though what evidence there is would suggest this was not the case.9 It is equally difficult to judge with certainty whether any but the very small number of ministers actually involved in the coup of 705 would have participated in it had they been offered the chance. The possibility exists, therefore, of a very simple explanation for the deposition—superior planning on the part of a small number of alienated bureaucrats. Still the absence of identifiable support for the empress at the end would seem to indicate that one of my earlier suggestions remains valid. The source of her legitimacy was personal, not normative, and her occupancy of the throne was an aberration accepted partially because it was seen to be temporary. The level of her performance was, therefore, subject to a scrutiny more rigorous than might be the case early in more conventional dynasties, and only while she continued to fulfill successfully the well-defined role of emperor10 were her claims actually to be emperor accepted. In sum, Wu Tse-t’ien was emperor by sufferance, her authority deriving not from ritual abdication and, therefore, not in the fullest sense from Heaven. Undoubtedly certain of her ministers continued to regard her as a usurper or, at best, as a self-proclaimed regent. In these circumstances she needed to constantly demonstrate that the “young prince” was not ready for rule, or that she was more capable than he and was preserving rather than changing his heritage, or both. In these circumstances there also existed ample precedent for ministers, or for the young prince himself, to decide that the regency should be brought to an end. What the de facto ruler in this situation might consider treason, could be regarded from the opposite perspective as an act of Confucian loyalty and, indeed, of ministerial duty. 9  See, for instance, Kuo Mo-jo, Wu Tse-t’ien, pp. 135–37 and 154–55 which outline the findings of an academic expedition to the old Li-chou area which discovered epigraphical evidence of her popularity and also the existence of an agricultural festival said to be held in commemoration of her birthday. See also TCTC 208, p. 6587, for Ts’ui Chiao’s secret memorial to Chung-tsung pointing out that the hearts of the people still adhered to the empress. 10  For a recent and succinct discussion of the importance of role in legitimation, see J. D. Langlois Jr., “Ritual and law in the legitimation of the Ming dynasty” (Asilomar, 1975).

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

408

Guisso

The history of the Chou may perhaps be more readily understood against this background. In spite of its short duration, the Chou might be divided for analytical purposes into four stages which seem to conform, admittedly with some artificiality, to the traditional rhythm of the dynastic cycle. The division between these periods is not always clearcut, but certain turning points may be identified and, using this form of analysis, we might posit that the first five years made up the great days of the dynasty. In this period the state was prosperous, expansive and confident, and the court was characterized by administrative efficiency. In 696, however, a downturn began, and from China’s massive defeat by the Tibetans early that year, serious problems of a financial and military nature arose and were not arrested until late 697 with the gradual suppression of the foreign threat and the rise to ministerial hegemony of Ti Jen-chieh. This “dynastic revival” gave the Chou a second period of at least superficial prosperity, but after Ti’s death in 700 and the rise to unbridled supremacy of the Chang faction, corruption began to run rife and a rapid decline set in which even the removal of the empress in 705 was unable to stem. Throughout the period, and as counterpoints to this rhythm, there were also to be seen the general continuation of T’ang policies upon which I have already commented, and the gradual decline of the empress’ health and capabilities, especially in the final period as she approached her eightieth birthday. The Chou began, as might be expected, with several new ceremonial arrangements designed to emphasize the transfer of the Mandate. These consisted chiefly of such symbolic measures as the rearrangement of ancestral temples and sacrifices,11 the changing of the banners to red, the substitution of “Wu” for “T’ang” in place names and, most importantly, the transfer of the seat of government to Lo-yang which had become the “Sanctified Capital” in 684. The declaration of a new capital was, of course, an act redolent with the traditional symbolism of imperial prerogative, and we have seen earlier how in economic terms Lo-yang was a logical capital site.12 At this time practical motives of a different sort were also present, for it is likely that the empress was attempting to dissociate herself from the T’ang capital and from the loyalist clans who had been drawn by official posts to settle there. As an extra measure 11   T CTC 204, pp. 6467–68. Arthur Wright, in his paper on the formal procedures by which Sui Wen-ti sought legitimation, emphasized the importance of the foundation of ancestral temples. For examples of the substitution of “Wu” for “T’ang” in place names, see THY 70, p. 1257 and 71, p. 1261. 12  Chapter 4, n. 92, and Chapter 6; see also Twitchett’s remarks in Financial Administration, pp. 84 ff. A northern capital was founded in 692, TCTC 205, p. 6487.

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

The Chou Dynasty [ of Empress Wu ]

409

of insurance, at least 100,000 households from seven chou surrounding Ch’angan were resettled near Lo-yang to strengthen the new center, and it is probably more than coincidence that they were chosen from areas where the concentration of fu-ping units was highest.13 Ch’ang-an remained the western capital and was placed in the immediate charge of one of the Wu princes.14 It is tempting also to see in this eastward shift of the capital a replication of the pattern of the Han restoration or even to suggest that Lo-yang’s Buddhist associations and the prophecy of the commentary on the Great Cloud Sutra played a part. Whatever her motives, however, the empress rapidly set about turning Lo-yang into a worthy rival to Ch’ang-an, beginning the huge task of walling the city in 692,15 and continuing her earlier program of building temples, monuments and public works. The ming-t’ang complex became larger and more elaborate, and in early 695 she completed the “Pivot of Heaven” (t’ienshu), a pillar of iron and bronze over a hundred feet high erected at the South Gate of the Forbidden City to commemorate the “ten thousand virtues of the Great Chou.”16 These constructions were enormously expensive—the cost of the pillar, for instance, was said to be beyond estimation17—but, at the same time, their symbolic value was great. They constituted a demonstration of selfconfidence by the empress, contributed to the “imperial air” (wang-ch’i) of her capital and, in a more concrete sense, were often measures of patronage of the type we have seen to be so important in the creation of state ideology. I showed earlier that this was true particularly in the case of Buddhism, and fine examples of financial patronage are found not only in the expansion of the ming-t’ang complex, but also in the fabrication of the “seven treasures” 13   T CTC 204, p. 6473, TFYK 486:336, THY 84, p. 1553, and CTS 6:3076:1. This occurred in the seventh month of 691 with the households from such chou as Yung, T’ung, and Ch’in where the concentration of fu-ping units was high. There are some source contradictions, with CTS and TCTC saying that those resettled were drawn exclusively from Kuan-nei while the others say Kuan-wai. TCTC limits the number transported to 100,000, and CTS goes on to say that Yung-chou was divided immediately into four parts.  The basic motives, it seems, were to provide a larger base of support for the new capital and the official class concentrated there, and to weaken the old Ch’ang-an area. 14   T CTC 204, p. 6473. Wu Yu-i, in fact, became liu-shou prior to the movement of the households. 15   T CTC 205, pp. 6478–79. Li Chao-te (Appendix B, no. 99) was placed in charge of the project. 16   T CTC 205, pp. 6496 and 6502–3. Jao’s excellent study, “Tsung-chiao hsin-yang” convincingly connects the t’ien-shu with Manicheanism. 17   T CTC 205, p. 6496. Note also Ssu-ma Kuang’s comments on p. 6498 suggesting that the treasuries were emptied to pay for the construction of the ming-t’ang complex.

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

410

Guisso

(ch’i-pao, Sansk. sapta ratna) in 693 after Hsüeh Huai-i presented a second sutra, that of the “Precious Rain” (Pao-yü ching), which reemphasized the empress’ identity as a ćakravartīn.18 The Buddhist church was to benefit in numerous other ways during early Chou—by gaining precedence over Taoism, encouraging the long ban on the butchering of animals, and the stiffening of penalties for crimes against church property—but it is a measure of the empress’ experience and skill that she did not forget Confucian sensitivities during this early period. As we shall soon see, her substantive concern with the linked issues of education, examinations and official appointments continued to grow at least superficially, and at the same time she distributed symbolic favors even-handedly. It is interesting to note, for example, that in the same ten-day period in which she founded the Great Cloud temples, she conferred a new honorific on Confucius, and later in the dynasty was similarly to honor his favorite disciples.19 She modestly refused early requests to magnify her own titles and hesitated when it was suggested in 691 that she perform the fengshan sacrifices at Mount Sung in Honan.20 When she eventually performed the rite in early 695, it was immediately after carrying out the great sacrifices to Heaven and Earth at the southern suburb, and it was followed by the declaration of a year’s freedom from the tsu tax for the entire empire!21 A second major way in which the empress appealed to Confucian sentiment was in her treatment of the few surviving members of the T’ang clan. There were well-established precedents, though too often honored in the breach, whereby a former imperial house was permitted by its successor to exist in dignity, though usually under surveillance. In the early Chou Jui-tsung and his immediate family seem to have lived under virtual house arrest,22 but as early as 691 the empress began to confer princely rank on her grandsons, beginning with the sons of the former Crown Prince Hsien and later extending the privilege to the five sons of Jui-tsung.23 Because various members of the Wu clan also bore princely rank, this created a situation unprecedented at least since

18   T CTC 205, p. 6493, which describes these treasures, and also Chapter 4. 19   T HY 35, p. 637. 20   T CTC 204, p. 6471. The petitions appeared in early 691, and the one suggesting the fengshan sacrifice included 2,800 names. 21   T CTC 205, p. 6503. See also CTS 6:3076:2 and Jao, “Tsung-chiao hsin-yang,” pp. 402–405. On the forms and purposes of the suburban sacrifice, THY 9 shang, pp. 141 ff. 22   T CTC 204, p. 6473. The sons of the former Crown Prince Hsien were given the Wu surname and imprisoned with Jui-tsung’s family. 23   T CTC 204, pp. 6473 and 6489.

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

The Chou Dynasty [ of Empress Wu ]

411

the Sui reunification but, considering the strength of T’ang loyalism, it was clearly a politic move to permit the coexistence of two “imperial” clans. In measures like these it seems clear that the empress was concerned early in the new reign principally to reinforce her credentials to rule, and that she not only made provision for her own support and security, but was careful simultaneously to balance measures which emphasized dynastic change with others which could be construed as indications that the change was not necessarily permanent. Her purpose, of course, was to broaden her appeal to those literati who might be willing to serve the Chou in the belief that they were protecting the T’ang heritage. As a further enticement, she moved very quickly to alleviate the terror which had been the greatest source of bureaucratic alienation since its initiation in 684. The Chou, for instance, was scarcely six months old when Chou Hsing, the leader of the so-called evil officials, was charged with treason, found guilty and murdered on the road to exile.24 According to one source, he had been responsible for several thousand deaths,25 and though we might expect a certain exaggeration in the figure, it does not seem unreasonable when we consider that an upright censor, permitted to investigate in mid-692, found him responsible for no fewer than 850 miscarriages of justice.26 The empress, of course, had been long aware of this danger, but prior to the usurpation had shown little scruple about using it to her advantage. Now, however, while she permitted Lai Chün-ch’en to take over Chou Hsing’s apparatus of denunciation and punishment, she responded to criticism of her harshness by the appointment of good and courageous judicial officers to balance the bad, and by herself taking a personal interest in important cases. Most famous of the good censors was Hsü Yu-kung, a man she appointed in spite of his fearful refusals of office immediately after the dynastic change,27 and she soon added to the censorate men like Yen Shan-ssu, Chu Ching-tse, and Chou Chü28 who spoke with such conviction against the system of delation that after mid-691, the Comprehensive Mirror reports, punishments began

24   T CTC 204, p. 6472 and CTS 186 shang:3564:4. 25   C TS, loc. cit. See also TCTC 205, p. 6485 for Ssu-ma Kuang’s estimate of the toll taken by the so-called evil officials. 26   T CTC 205, p. 6485. 27  On Hsü, see Chapter 5, n. 88. His early refusal of office is found in TCTC 204, pp. 6469–70. 28  For examples of the activities of these, see TCTC 205, pp. 6485–86. It is of interest to note that in mid-692 Chu acknowledged in a memorial that “the minds of the people are already made up” (chung-hsin i-ting), presumably in favor of the Chou.

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

412

Guisso

to diminish.29 In that same year, when Lai accused seven high officials of rebellious conspiracy and managed to extract false confessions from them, the empress listened to the plea of a ten-year-old boy and then released them from prison to conduct their defense before her and prove themselves innocent.30 The question which must be asked here is why the empress permitted Lai Chün-ch’en to remain in office and continue his activities after his dishonesty had been exposed. Several conjectures were earlier offered about the reasons for her initiation of and support for the terror in general, and so long as a woman occupied the throne most of these remained valid. At this time, however, a more concrete reason existed in the appearance in late 691 of a campaign advocating the establishment of Wu Ch’eng-ssu as crown prince.31 Initially the empress was cool to the proposal, but seems at the same time to have flirted briefly with the idea, since she permitted open representations in its favor to be made, and when the expected storm of ministerial opposition arose, Ch’eng-ssu allied himself with Lai to eliminate the most inflexible of his opponents. In subsequent months several “plots” were discovered, and numerous officials were exiled or executed. It is all too likely that the empress was not really deceived and permitted the destruction of so many persons principally to reduce the influence of T’ang loyalism and to see whether the court could be intimidated into the support of Wu claims. She was conscious, however, that she could not go too far, and so balanced the influence of the Lai-Wu faction by conferring virtual invulnerability on their most prominent enemy, Li Chao-te. Chao-te was a ming-ching graduate from a family of officeholders who was vice president of the Secretariat in 691. An uncompromising and outspoken Confucian even when in office and hence unpopular with his colleagues, he nevertheless survived as paramount minister for two years from the time of his elevation to tsai-hsiang status in mid-692.32 He was strongly opposed to a Wu succession, and shortly after the matter arose he told the empress that “from ancient times I have never heard that a nephew became Son of Heaven and 29   T CTC 205, p. 6486. 30  Details are found in TCTC 205, pp. 6479 ff. and Fitzgerald, Empress Wu, pp. 138–40. 31   C TS 183:3553:2 and TCTC 204, pp. 6474–75. See also CTS 87:3352:4. 32  On Li, see Appendix B, no. 99. He is depicted as a courageous official opposing, for instance, the attempt of Lai Chün-ch’en and his cohorts to marry into the T’ai-yüan Wang and Chao-chün Li clans (CTS 87:3352:4), and once having one of the most powerful of the “evil officials” flogged to death, ostensibly for a violation of the sumptuary regulations (TCTC 205, p. 6491). In the matter of the succession, he rapidly showed his position by having Ch’eng-ssu’s spokesman beaten to death. TCTC 204, p. 6475.

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

The Chou Dynasty [ of Empress Wu ]

413

[then] set up ancestral temples for his aunt. Your Majesty, moreover, received the guardianship (ku-t’o) [of the empire] from Kao-tsung and if you give it to Ch’eng-ssu, then he will be without ancestral sacrifices!”33 A few months later the empress accepted his argument that Ch’eng-ssu was a threat to her own power, and she abruptly dismissed her nephew along with his relatives and supporters from their high posts, necessitating the governmental reorganization which left Li Chao-te and his adherents supreme.34 Wu Ch’eng-ssu was never again to hold tsai-hsiang status except for one month in 697. His fall illustrates one of the most remarkable of the differences between Wu Tse-t’ien and other female powerholders in China. Although she frequently appointed Wu princes to positions of command in expeditionary armies and the Yü-lin Guard, she cannot be said to have used male relatives in other ways to underwrite her political power.35 The career of Li Chao-te served the empress’ purposes even with its unhappy end. During early Chou Li served as a vigilant counteracting force to Lai Chün-ch’en, curbing Lai’s excesses and even frustrating his attempts to marry into the great northeastern aristocracy,36 but the empress also clearly saw that Li’s role could last no longer than Lai’s. When Lai was exiled for corruption in late 694, she acted simultaneously on complaints about Li’s excessive power and arbitrary nature to also sent him from the capital,37 and in 697 she had both men executed on the same day even though Li was probably innocent of any capital crime.38 Because Li Chao-te was also the most prominent defender of T’ang interests during the succession dispute, his fate shows how the empress could manipulate the court factions to simultaneously keep both her own and the T’ang restorationist partisans from dominance over the throne. 33   T CTC 204, p. 6476. 34   T CTC 205, pp. 6483–4. Three others, Ts’ui Shen-chi (Appendix B, no. 97), Yao Shou (no. 100) and Li Yüan-su (no. 101), were all raised to tsai-hsiang status with Chao-te on this occasion and so might be identified as his adherents. All were eventually disgraced or executed. 35  On the posts held by the Wu clan, see CTS 183:3553:1. After 695, with the fall of Hsüeh Huai-i, the use of family and favorites in military posts was very rare, the only exceptions being Wu I-tsung and Wu Yu-i. This is discussed passim in the text. I do not suggest, however, that members of the Wu clan became a negligible factor in the political equation. The empress used them on subsequent occasions to further her political aims so that, for instance, Ch’eng-ssu and San-ssu assisted her in the New Year’s ceremony of 693 from which Jui-tsung was barred. (TCTC 205, p. 6488 and HTS 76:3867:4). We may surmise that her aim was to confuse the advocates of a T’ang restoration. 36  See note 32 above. 37   T CTC 205, p. 6496. 38   T CTC 206, p. 6519.

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

414

Guisso

Li Chao-te’s victory in expelling the Wu clan from their powerful positions in 692 was a real one, and there are indications that the empress feared that it would lead the court to regard the succession as settled and so deprive her of a valuable political weapon. Thus she began the year 693 with a magnificent ceremonial which was remarkable not so much because she personally composed the music for the 900 dancers, but because she chose her two nephews to make the secondary offerings, excluding Jui-tsung from the proceedings.39 Shortly thereafter she secretly had Jui-tsung’s two favorite consorts executed on charges of sorcery laid by a slave girl. Jui-tsung was either too cowardly or too diplomatic even to raise the matter, but since one of the consorts was the mother of the future emperor Hsüan-tsung, the incident undoubtedly had its effect on his later attitude toward his grandmother.40 As heir apparent, Juitsung was in the delicate position of being the most likely focus of any restoration plot, and the empress seems to have deliberately emphasized the insecurity of his position two months later when she first changed the princely titles of his five sons, and then publicly executed two officials who had paid him a secret visit.41 On this occasion he himself was denounced as a traitor, and for a brief period the T’ang family’s prospects hung by a very thin thread. He was saved, however, when one of his retainers, while being interrogated by Lai Chün-ch’en, attempted to cut out his own heart to show his master’s loyalty. The empress is said to have sighed at her inability to know her own son when she heard of it and halted the proceedings so that, as the Comprehensive Mirror remarks, “Jui-tsung, from this [time] was safe (te-mien).”42 The empress seems at this time to have decided definitively in favor of a Li succession, and she did not allow any further discussion of the matter until 698 when Chung-tsung’s appointment was being considered. I suggested earlier an element of artificiality in the succession dispute, contending that the empress’ attitude was determined not by any serious intent to have a nephew succeed her, but rather by her desire to attain certain political ends. Among these the most important was undoubtedly the continued expansion of imperial prerogatives at the expense of those of the tsai-hsiang group. This becomes abundantly clear when we place the succession dispute

39  Note 35 above. 40   T CTC 205, p. 6488. Hsüan-tsung’s mother, the Virtuous Concubine, was the greatgranddaughter of Tou K’ang (Appendix B, no. 6). 41   C TS 186 shang:3564:3 and TCTC 205, pp. 6489 and 6490. Ssu-ma Kuang remarks that “from that time none of the nobles or ministers were able to see him.” 42   T CTC 205, p. 6490.

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

The Chou Dynasty [ of Empress Wu ]

415

within the context of other measures more overtly directed toward the same purpose. Scholars have for some time recognized that during Sui and early T’ang a key problem for the imperial house was the reduction of the pretensions and the actual power of the great houses whose members had not only monopolized governmental authority but had also so often usurped the throne during the period of disunion. An earlier chapter discussed at length some of the ramifications of this problem and showed that the T’ang solution lay essentially in the dilution of the power of the great clans through the “bureaucratization” of status. T’ai-tsung, as we have seen, attempted to establish the principle that status was dependent upon service to the T’ang, and as a result he could attract to the highest posts in his bureaucracy men of aristocratic or at least officeholding background. Particularly in the case of his tsai-hsiang, men of proud genealogy were predominant, but because of his unique talents, the collegial-style government which marked most of his reign was highly successful. T’ai-tsung was a big enough man to coopt the best of the well-born of his generation. Inevitably, however, the chief ministers developed a sense of initiative and group consciousness which T’ai-tsung’s successors, Kao-tsung and Empress Wu, seem to have regarded as highly dangerous and, as suggested earlier, the united opposition of the tsai-hsiang to her rise led to the destruction of their leadership and a period of over two decades when they were regarded by the throne as rivals rather than partners in government. Only in the last years of Kao-tsung, and under P’ei Yen’s competent leadership,43 did signs of collegiality begin to reappear. It seems possible that had Chung-tsung supported the aspirations of this group and allied himself with it against his mother, he might have avoided deposition. Instead he alienated P’ei Yen by threatening to continue, in the persons of himself and his wife, the pattern of emperor-consort family government, and thus threw the tsai-hsiang into the arms of the empress during her seizure of power in 684. P’ei’s role in Chung-tsung’s deposition was viewed by the empress as having the unwelcome effect of reinforcing the prestige and ambitions of the chief ministers and, as already suggested, this may have played a role in P’ei’s downfall. His successor as paramount minister, Liu Wei-chih, perished in 687 largely because he fought the empress’ assertion of imperial over ministerial prerogatives.44 Until the rise of Li Chao-te in the early 690s and Ti Jen-chieh in the late 690s, no dominant personality rose from

43  See Chapter 5, nn. 7 and 8. 44  Chapter 5, n. 79.

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

416

Guisso

among the tsai-hsiang. The tenuous nature of the empress’ legitimacy required that this be the case. The manner in which the empress weakened ministerial power was not original nor, indeed, did it create an irreversible trend, for Hsüan-tsung’s government, at least in its early years, bore greater resemblance to that of T’aitsung than to that of the empress.45 This is just one more example of the general lack of innovation during Chou mentioned earlier. The insecurity of the empress’ legitimacy dictated that she work as closely as possible within the traditional framework. Despite occasional appearances of fundamental reform, she changed only the de facto balance of power at the top, and when her two immediate successors proved unable to fill her shoes, the status quo ante of T’ai-tsung’s time reappeared. The New T’ang History “judgment” quoted earlier demonstrates that her methods did not go unnoticed by the dynastic historians. Her chief tactic was the manner in which she used her unquestioned right of appointment and dismissal. To begin with, she expanded the number of ministers holding tsai-hsiang status while at the same time shortening their tenure. Thus we find that in the late 690s as many as sixteen men held the rank, compared to a previous high of ten45 and a much lower average. The highest estimate for T’ai-tsung’s reign is twenty-nine tsai-hsiang in twenty-one years while the corresponding figure for the years 684–705 is sixty-six.46 Under T’ai-tsung, the average tenure was seven years and under the empress, just over two. This contrast is suggestive, of course, of two highly different political styles, and it seems no accident that factionalism was a much more pronounced phenomenon in the Chou. During the Chou only a tiny proportion of ministers reached tsai-hsiang rank before the age of sixty and even then ministers were shifted with such 45  See Pulleyblank, Background, pp. 48 ff., and more particularly, Chou, Han-T’ang, pp. 373 ff. 46  By my calculations, ten men held tsai-hsiang rank in 645. The totals vary according to source and, although my own are different, I am here following the figures given in THY 1, pp. 3–4. Those for the reign of Kao-tsung, divided to correspond with the expansion of Empress Wu’s power, are as follows: Period

appointments

tenure

650–659 (4th month) 659–664 (12th month) 664–684 Total

14  4 25 43

4.5 years 2.2 years 5.5 years 3.1 years

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

The Chou Dynasty [ of Empress Wu ]

417

frequency that consolidation and entrenchment were impossible. Lou Shih-te, whose career was otherwise remarkable for lengthy tenure in high posts without demotion,47 held four different positions and was four times sent away from the capital on military duty during his six years as tsai-hsiang. Even Ti Jen-chieh, who enjoyed the fullest trust of the empress, held no fewer than seven different positions during his three years as tsai-hsiang! Furthermore, a fairly large number of such appointments were preceded by such terms as “provisional” or “acting” (chien-chiao, shih 試)48 so that appointees were openly made conscious of a certain instability in regard to their tenure or qualifications. An astounding eighty percent of the Chou tsai-hsiang suffered dismissal, some to exile or death and others simply to a lower post. For the reign of T’aitsung, the corresponding figure is thirty-three percent.49 It is interesting to note that among the early writings of the perspicacious historian Liu Chih-chi is a fu of 695 which warns of the dangers of aspiring to high office. He entitled it, “Think Carefully” (Ssu-chen fu).50 The overlap of function and responsibility inherent in the role of the tsaihsiang could be exploited by the empress to foster dissension and rivalry. She did this by simultaneous, and sometimes also parallel, appointment of political enemies. Ti Jen-chieh, for instance, was first elevated to tsai-hsiang on the same day as Wu Yu-ning, and when Ti became Chancellery president in mid698, Wu San-ssu was made his opposite number in the Secretariat. Imperial relatives, sycophants, and the occasional mystic expert only in Taoist magical practices51 were seldom to have lengthy tenure, but they were often appointed simply to balance the opinions of their better qualified colleagues. Their very appointment tended to devalue tsai-hsiang status and, by contrast, to elevate the imperial institution. 47  See Appendix B, no. 104, and note TCTC 206, p. 6541), which comments on his remarkable longevity in spite of the terror and the confused politics of the time. 48  While I have made no exact count, I have found that virtually every kind of irregular appointment occurred within the tsai-hsiang group in the Chou. The terms used to designate this type of position are several: chien-chiao “acting,” shih 試 “probationary,” shou “temporary,” and she “provisional.” We find also a large number of concurrent (chien 兼) appointments, supervisory duties (p’an) and those appointed to one post but designated to fill the function (chih 知) of another. This led not only to uncertainty of tenure among the tsai-hsiang but severely restricted their freedom of action and created an atmosphere of distrust in the highest ranks of officialdom. 49  My calculation. See Appendix B. 50   T CTC 205, p. 6501. 51  On these appointments see, for instance, TCTC 206, p. 6533 and 205, pp. 6496–95 and 6499–6500.

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

418

Guisso

Other traditional control methods were directed to the same end. Special rewards, usually in the form of personal commendations, silks or official rank for the recipient’s sons were conferred in ways that would remind ministers that distinction was earned not merely by passing an examination or gaining high office, but by pleasing the ruler while in office. The empress used ministerial remonstrance, something for which she is often praised,52 to encourage criticism directed against her as a person, and discourage attempts to limit the scope of imperial power. She permitted ministers to openly ridicule her fondness for auspicious omens or her penchant for making excessive numbers of appointments,53 but at the same time she discouraged the use of arguments based on precedent in memorials to her, lest such precedents limit her prerogatives. She also punished participants in unseemly quarrels which occurred in the imperial presence.54 Certain questions, like the succession or the role of her favorites, she regarded as “household matters” and throughout the Chou refused to permit ministers to raise them, often on pain of death.55 This, of course, made nonsense of the ancient tsai-hsiang claim of omnicompetence, but skillfully preserved the appearance of free discussion in the court and undoubtedly enhanced the empress’ popularity among those too far removed to perceive the reality. Finally, there were times during Chou when the empress simply excluded the tsai-hsiang from her counsel, preferring to seek advice from outside their circle. Early in the dynasty her chief alternative seems to have been a group of non-official advisers, popularly called the “North Gate Scholars” (pei-men hsüeh-shih). Often considered the predecessors of the Hanlin Academy,56 these men are first mentioned as a group in the year 674, when Ssu-ma Kuang tells us that they began not only to compile the literary works later ascribed to the empress but were “secretly ordered to participate in decisions in order to divide the power of the tsai-hsiang.”57 it is difficult to know how extensively their 52  See Chapter 1, n. 42, and CTS 6:3077:1. 53  For instance, TCTC 205, p. 6484 and p. 6478, respectively. 54  See the biography of Chi Hsiu in Appendix B, no. 120 and TCTC 206, p. 6544. 55  On the question of the succession as a “household matter,” see TCTC 206, p. 6526. Criticism of the Changs brought death in 701 to the son and daughter of Chung-tsung and the son of Wu Ch’eng-ssu. TCTC 207, pp. 6556–57 and CTS 86: 3350:4. Other examples of both are cited in the text. 56  Chou, Han-T’ang, pp. 506 ff. 57   T CTC 202, p. 6376. Judging by the dates of the literary works for which they were responsible, and by a notice in TFYK 550:2b, the group seems to have existed prior to 674, perhaps as early as 667. Most of those whose names are preserved, Liu Wei-chih (Appendix B, no. 72), Fan Li-ping (no. 85), Yüan Wan-ch’ing (CTS 190 chung:3581:4), Miao Shen-k’o (CTS,

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

The Chou Dynasty [ of Empress Wu ]

419

services were used, though the empress did turn to them, for instance, when her regular ministers balked at the grandiose nature of her ming-t’ang plans, and she had the group’s leader, Liu Wei-chih, second P’ei Yen in recommending Chung-tsung’s deposition.58 The standard sources make little mention of them during the Chou, and perhaps because many entered the regular bureaucracy, their coherence as a group was destroyed by then. It is also possible that the empress found less use for them as she turned more to her own family and favorites, some of whom she realized were not without talent. Wu San-ssu, for instance, came almost to monopolize capital architectural projects, and the Princess T’ai-p’ing and Shang-kuan Wan-erh seem to have served as confidantes and secret advisers.59 In the last years of the dynasty the Chang brothers and their many hangers-on played a similar role, enjoying privileged access to the empress and sometimes refusing access to her even to the tsai-hsiang! It is important, as we view this process, to recognize two things. First, the empress enjoyed only a partial success in reducing the prerogatives of her ministers and was forced on occasion to accept defeat in matters as trivial as the curbing of her extravagance and as fundamental as her ultimate admission that the succession concerned not just her household but the entire empire.60 Second, although the bureaucracy was progressively weakened and restricted at the top, it continued to expand and develop in the lower levels as both entry into and promotion within it were facilitated. As a consequence, an ever-larger sphere of bureaucratic activity tended to slip out from under imperial control. Early in 692, for instance, when the empress appointed to lower capital positions a large number of provincials recommended by special commissioners and did so “without inquiring [whether they were] worthy or stupid,” there was a common saying to the effect that even sawyers and pottery menders had become officials.61 In 695 Liu Chih-chi pointed out in a much admired memorial that the empress’ yearly Acts of Grace conferred so many promotions and honorific ranks that official robes and insignia were encountered everywhere. He compared her “excessive” appointments to gravel in need of sifting.62 loc. cit.) and Hu Ch’u-pin (CTS 190 chung:3582:1) received their first capital appointment at about this time. TCTC 204, p. 6447, a notice of 688, is the last mention of the group. 58   T CTC 203, p. 6417. 59   T CTC 204, p. 6466. On Wan-erh, see CTS 51:3279:4. 60   T CTC 206, p. 6526. 61   T CTC 205, pp. 6477–78. See also THY 67, pp. 1180–81, which says that 132 appointments were made at this time. The highest posts, acting grand secretaries of the Secretariat, were of the fifth rank, first grade, and the lowest were clerks of the ninth rank, third grade. 62  See TCTC 205, pp. 6500–01 and THY 40, pp. 728–29.

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

420

Guisso

Without exact figures it is difficult to estimate, except in this impressionistic sense, how greatly the bureaucracy expanded during the Chou, and the task is made even more difficult by the fact that the new men seem to have been absorbed either into nominally menial jobs or through auxiliary appointments, rather than through the creation of new positions.63 But for our present purpose it is perhaps sufficient to demonstrate the existence of the trend. The Chou bureaucracy may not have been characterized by quality below the fifth rank, but it was opening a way to genuine social mobility. Its lower ranks constituted a huge pool from which those who showed ability might be drawn out and elevated to the charmed circle of fifth-rank officialdom. For approximately one-fifth of the Empress Wu’s tsai-hsiang, we have no information whatever on their antecedents or previous careers, a fact which suggests that men of obscure origin were at last reaching the highest posts and that at least some succeeded in doing so even without formal examination. The pressure from below which this phenomenon brought to bear upon those already occupying high posts helps to explain, for instance, why the coup which deposed the empress was the work exclusively of high officials, and it also helps to explain why the Chou saw no further attack on the great aristocracy in the form of clan lists or marriage bans. Such measures were no longer needed. If the tsai-hsiang group in the Chou constituted a representative sample, it seems clear that the scions of the great clans were coming to define their status in bureaucratized terms by writing examinations and accepting capital positions but, at the same time, were able to control only a small proportion of the most pivotal offices.64 In other words, even if they continued to fight for their own interests, they now did so on a field chosen by the throne and which 63  See Chapter 7. The bureaucratic expansion in the early Chou seems to have equaled that in the period just prior to the usurpation. Although we are offered no estimates so specific as Wei Yüan-t’ung’s statement of 685 that over 1,000 chu-se entered the “current” each year (THY 74, p. 1336), we are told, for instance in 692, that vast numbers had their selection rescinded. THY 74, p. 1345 and T’ai-p’ing kuang-chi 185:5a. It is possible that some of the excess was absorbed in the expansion of military administration: eighty-six new fu-ping units were created in 690 (WYYH 464:1b), and later new positions were set up in existing units (Yü-hai 138:4a). As we shall see, the empress had to respond to pressures for bureaucratic reduction toward the end of the Chou. 64  Taking once again our tsai-hsiang sample from the Chou, and excluding those who rose as family or favorites of the empress, we find that of fifty-nine chief ministers, at least thirtythree came from clans which were either aristocratic in background or had a tradition of office holding. Of these, however, only fourteen were from clans specifically regarded as preeminent. This means that less than a quarter of the Chou tsai-hsiang came from the great clans.

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

The Chou Dynasty [ of Empress Wu ]

421

undermined their old regional power bases. Little attempt seems to have been made to balance appointments on a regional basis, and this would suggest that the great aristocracy was coalescing into a more coherent, less divided interest group. Subject to pressures both from above and below in their bureaucratic role, the great clans were clearly on the defensive during the Chou and sought more to maintain than to advance their positions. They were by no means impotent and were revived under Hsüan-tsung,65 but by the eighth century power struggles had become more internalized than before, and the throne now played a more powerful role thanks to the work of the Empress Wu. One more consequence of the widening of the road to office was the extremely high quality of those officials who reached the top. The traditional historians who so frequently criticize the empress are also consistent in their admission that her abilities were great enough to attract into her service men like Ti Jen-chieh, Li Chao-te, Hsü Yu-kung, Wei Yüan-chung, and many others whose talents were hardly inferior to the ministers who had surrounded T’aitsung. Ssu-ma Kuang, summarizing earlier views found in the dynastic histories, says that although the empress made too many appointments, she rapidly dismissed or executed the incompetents, and that because of her decisiveness and her abilities to use reward and punishment and to recognize talent, “the brilliant and worthy of the time were all in rivalry to be employed by her.”66 This is undoubtedly true, but surely it was also the case that these talented men recognized the validity and the value of many of her wider aims and desired to contribute to their achievement. Perhaps the strongest indication that this was so is that the political uncertainty, factionalism, and the continued existence of the terror of early Chou could not undermine its efficient if unambitious administration. The sharpest contemporary criticism of the government, as we have seen, centered upon excessive number of punishments and appointments. There were no signs of unrest outside the capital. Corruption, even when it involved the empress’ closest supporters, was dealt with swiftly.67 For reasons to be mentioned shortly, a demographic shift was beginning to come to the notice of the court, and in this period attempts were made to tighten the census procedure and to 65  Pulleyblank, Background, pp. 47–60, and Tonami, “Chūsei,” pp. 9 ff. 66   T CTC 205, p. 6478. 67  See, for instance, the case of Tsung Ch’u-k’o (CTS 92:3365:4 and HTS 109: 3930:4) who had not only a marriage connection with the empress, but whose brother had been one of her firmest supporters in the usurpation and is said to have invented her new characters. The Tsungs’ first demotion occurred in the tenth month of 690. TCTC 204, p. 6468. A relative of Lai Chün-ch’en lost his position in the Yü-lin Guard in late 692. TCTC 205, p. 6487.

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

422

Guisso

create new administrative units to deal with the growth of population in the Yangtze valley.68 Both the Boards of Civil Office and of War were expanded, presumably to keep pace with their work load, and decree examinations were held, sometimes at the rate of two a year.69 We find little evidence anywhere of maladministration but rather that “the minds of the people were made up [in the empress’ favor].”70 Another sign of her self-confidence was the creation of the Record of Current Politics (Shih-cheng chi) which recorded her every pronouncement and sent it monthly to the History Office.71 The Chou suffered no setback in foreign affairs until 696 and, indeed, China’s prestige among its neighbors seemed to rival its mid-century height. Tribute missions appeared from Central Asia, India and Indochina, a new king was invested in Silla in 693,72 and by 696 over 1,000,000 foreigners had sought and been granted permission to settle within the empire. The largest group to enter were 670,000 Western Turks of the Shih-hsing tribe, and large numbers of Ch’iang and Man tribesmen were also incorporated.73 In December of 692 two of the finest generals of their age, Wang Hsiao-chieh and T’ang Hsiu-ching,74 who had earlier persuaded the empress that the time was ripe for the recovery of the Four Garrisons, inflicted heavy defeats on the Tibetans and reestablished the An-hsi protectorate at Kucha for the first time in twenty-two years. It is interesting to note that the empress seems not to have regarded this as expansionism, for on being informed of the victory, she remarked that the old borders of the chen-kuan era [627–650] had been restored completely and now 68   T HY 85, p. 1555 and TFYK 486:12a. 69  On the decree examinations, see Lin, Sui-T’ang shih, pp. 210–211, and on the Board increases, THY 58, p. 1006, and 59, p. 1030. 70   T CTC 205, p. 6485, from a memorial of 692. 71   T CTC 205, p. 6489, THY 63, p. 1104, and Yü-hai 48:32a. 72  See, for instance, the list of tributary missions in TFYK 970:17a ff., and the investiture of the Sillan king in TFYK 964:10a. 73  My own calculation. When the Shih-hsing sought refuge in China in 690, it was after five years of constant attack by the Eastern Turks. Their leader, the Hu-se-lo Khan, was made a general (TCTC 204, p. 6469). THY 98, p. 1756 remarks on the settlement of the Ch’iang, and TCTC 205, p. 6494, that of the Man. 74  On Wang, see CTS 93:3366:2 and HTS 111:3935:3, and on T’ang, CTS 93:3366: 3 and HTS 111:3935:4. Wang had been captured by the Tibetans in Hsüeh Jen-kuei’s defeat at Ta-feich’uan in 670 and, after living among them for a decade, returned to become China’s greatest expert on the Western Regions. T’ang came from a family whose military tradition went back to the Northern Chou, but which had chosen the wrong side at the time of the T’ang foundation so that to establish himself, Shou-ching had taken the ming-ching degree. His early experience was against the Turks.

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

The Chou Dynasty [ of Empress Wu ]

423

were quiet.75 The frontier remained peaceful until early 694 when the Eastern Turks found themselves a brilliant if erratic leader in the new khan, Mo-ch’o (Qapagan), who soon dominated much of Central Asia and became the most formidable enemy of the Chou.76 He wasted no time in testing China’s border defenses but Chinese arms, even when led by Hsüeh Huai-i or one of the Wu princes, were still formidable. These early raids were sporadic and brought Mo-ch’o so little profit that he soon asked to submit, and in the winter of 695 he was permitted to do so with a generalship and a ducal title.77 Mo-ch’o’s submission was in a sense the high-water mark of the Chou. It was the last foreign policy success the dynasty was to enjoy for quite some time and, moreover, was symbolic of the success of the policy of buying acceptance through growth of rewards which had characterized domestic administration and, with the single exception of the Four Garrisons, foreign affairs also. In the widest sense, the aim of this policy for the empress was to secure her legitimation, permitting prosperity and inertia to do their work of reconciling bureaucrat and commoner to the unprecedented rule of a female emperor. This policy was, however, costly. Bureaucratic expansion, public works and patronage of Buddhism accounted for much of the budgetary increase, but defense costs also mounted rapidly as some of China’s neighbors began to take advantage of internal dissension. The empress had been successful in augmenting the fu-ping system and in the establishment of some new garrisons, especially in the South,78 but it rapidly became evident that the demands on the state treasury were simply too great. Thus as early as mid-694 the empress had tried to assess her officials the equivalent of two months’ salary to support the military, but had been forced to withdraw the assessment in the face of general opposition.79 In the same month, the growing shortage of metal was 75   C TS 93:3366:2. The campaign is outlined in the biographies above and discussed in some detail in Ise, Chūgoku, pp. 197 ff. See also Fitzgerald, Empress Wu, pp. 144–45. 76   T CTC 205, p. 6493. On Mo-ch’o, see CTS 184:3598:1, HTS 140:4132:1, and THY 94, pp. 1691 ff. Ts’en, T’u-chüeh, pp. 324 ff. outlines Mo-ch’o’s raids from 689, even before he became khan. In Western languages, see Giraud, Turcs célestes, pp. 49 ff., and Grousset, L’empire, pp. 155 ff. Mo-ch’o was the brother or half-brother of the Khan Ku-ch’u-lu, usurped the leadership on his death in late 693, and immediately invaded Ling-chou. 77   T CTC 205, p. 6503. The request to submit may have been in response to the fact that Wang Hsiao-chieh had just been sent against him with a large force. 78  Note 63 above and TFYK 532:14a. It is interesting to note also a decree of 694 (THY 78, p. 1438) whereby each of the major provincial officials, governors, prefects and garrison commanders were required to leave one son in the capital guard. It seems unlikely that any new posts were created for this but rather that a hostage system was being established. 79   T CTC 205, pp. 6495–96.

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

424

Guisso

demonstrated when farm implements had to be melted down for the construction of the t’ien-shu.80 At about the same time, officials began to identify still other causes of the financial problem, foremost among them the failure of earlier attempts to tighten census registration. In 695 Li Chiao, a censor of distinguished lineage, a chin-shih graduate and a man of recognized literary talent, drew attention to the growing number of vagrant households, a problem which he asserted had existed for some years.81 He did not directly blame the central government for the phenomenon but rather the indifference of local officials who were lax about identity cards (chien-ch’a), and he pointed out the futility of recent attempts to have local officials reform the situation. He offered four solutions. First, the present law should be clarified and the government should show a tolerant attitude to present offenders. Second, a time limit of 100 days should be set, during which the squatters could give themselves up and opt either to remain where they were or return to their original homes. Third, in either case they were to receive amnesty and financial aid to get a new start. Fourth, the whole reform should be supervised by special censors sent out from the capital. The empress is said to have approved this plan and actually to have appointed the censors before she was forced to shelve it because of ministerial objections whose grounds we are not given.82 There is evidence from Tun-huang that a similar plan was implemented about 702–3, but its effect could not have been great since Hsüan-tsung faced the problem in greatly magnified form early in his reign.83 It seems likely rather that the Chou continued to suffer from the problems Li Chiao connected with the vagrants—tax evasion, local conflicts, and the too-rapid growth of urban population—but that they were not viewed as serious enough to require a concerted effort at reform. These first signs of economic strain, followed by two successive political defeats, coincided for the empress with the severe personal disappointment and weakening of prestige caused by Hsüeh Huai-i’s burning of the ming-t’ang complex in early 695. We have already noted how this event led to a shift in her ideological priorities from Buddhism to Confucianism, particularly in symbolic 80   T CTC 205, p. 6496. It is of interest to note that no protest is recorded. 81  On Li Chiao, see Appendix B, no. 123, and on his memorial, Chapter 5, n. 37 and THY 85, pp. 1560–61. The memorial, along with others relating to the problem, is translated by Balazs in “Beiträge.” 82   C TS 94:3368:2. 83  See Tonami, “Tō no ritsuryō,” and Pulleyblank, Background, pp. 27 ff. On the Tun-huang evidence, see T’ang Ch’ang-ju, “Fo-t’ao-hu.”

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

The Chou Dynasty [ of Empress Wu ]

425

terms.84 Moreover, it seems to have fostered an atmosphere of freer speech at the court. Frank criticism of her plans to rebuild the ming-t’ang85 were voiced as well as complaints, already noted, about excessive appointments and rewards and lax provincial administration. The empress’ response varied: she agreed to build a smaller version of the ming-t’ang, rid herself of Hsüeh Huai-i and other favorites86 and, toward the end of the year, abolished the principle of anonymity (hu-ming) she had herself introduced into the examinations, probably in 690.87 While her reasons for this are not altogether clear, it seems likely that such a measure would have appealed to those already in office who would now find it easier to help their friends and relatives gain entry into the bureaucracy. Although her response to the problem of vagrancy was a positive one, it seems that in these other areas she continued to rely on palliatives rather than take the positive steps which might have jeopardized her popularity. As we have seen, and in spite of the fiscal problems so evident in 695 when she offered the feng-shan sacrifice at Mount Sung, at the end of that year she declared an empire-wide holiday from that year’s tsu tax!88 What impact this might have had we are not told, for domestic problems were soon overshadowed by the renewal of the foreign threat with the massive defeat sustained by the hitherto invincible Wang Hsiao-chieh and his deputy, Lou Shih-te, at the hands of the Tibetans in April 696.89 The underlying causes of this defeat are hard to identify since the sources provide no more than the 84  Most notably, the construction of the nine tripods (TCTC 205, p. 6499); on whose Confucian associations, see Tz’u-y’uan, p. 58a. 85   T CTC 205, p. 6500 and THY 11, pp. 278–79. 86  On the reconstruction, see THY 11, pp. 279–80, TFYK 564:7a, and TCTC 205, p. 6505. The new building was called the T’ung-t’ien t’ang and was smaller in all dimensions than the original. On the death of Huai-i, see TCTC 205, p. 6502, CTS 183:3554:4, and note the contradictions pointed out in the k’ao-i notice. The nun who ran a house of ill repute was enslaved with all her followers. TCTC 205, pp. 6499–6500. 87  The hu-ming principle was abolished in the tenth month of 695. THY 75, p. 1358. The full decree is found in CTW 96:6a but offers little enlightenment on the reason for ending the practice. The empress acknowledges that she initiated the principle of anonymity so that her officials could be chosen from as many sources as possible, but then goes on to characterize the practice as contrary to principle and calls for its end in the name of efficiency and good government. The reason I suggest for the change is conjectural. 88   T CTC 205, p. 6503. 89  The best account of this campaign is found in the biography of Lou (CTS 93:3366:2, HTS 108:3929:4). See also TFYK 443:5a and TCTC 205, pp. 6504–5. The empress was extremely angry, demoting Wang to commoner’s status and Lou to a low provincial post.

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

426

Guisso

briefest outline of events among the Tibetans since the accession of the young btsan-po in 679. From what we can gather, the young ruler had initially lacked the strength to wrest state control from the mGar clan who had been in power since his grandfather’s time. It was a member of this clan, Lun Ch’in-ling, who was responsible for the victory.90 Shocked by her first defeat during the Chou, and with the Tibetans only a hundred miles from Ch’ang-an, the empress had no time for retaliation before a new threat arose on her northern flank. This was the great Khitan rebellion which broke out in early summer. This event seems to have been totally unexpected. The Khitans had been quiescent since 648 when T’ai-tsung had incorporated them into the Ying-chou tu-tu-fu91 under which they had come to enjoy a good deal of self-government. The two rebel leaders Li Chin-chung and Sun Wan-jung had risen to high administrative posts there but, as was sometimes the case, the Khitans had not been fortunate in the Chinese officials placed over them, and the current governor-general seems to have been a hard and self-willed man who not only treated the chieftains like slaves but had refused relief during a recent famine.92 This is usually given as the immediate cause of the rebellion, but the dissatisfaction with Chinese rule may have run deeper, since within a week Li and Sun had thousands of rebels under arms. Since the Khitans were a relatively sophisticated enemy and already lived inside the Wall, official response was both swift and in earnest. Twenty-eight generals at the head of an unspecified force were sent against them, but in the first major battle, near present-day Peking, the imperial troops were almost annihilated. The disaster was compounded when the Khitans used captured Chinese seals to forge orders which led the relief force into an ambush.93 In what was admittedly a desperate situation the empress initially overreacted. She made an unprecedented offer of official rewards to criminals and private slaves who would enroll in the army,94 she quickly began to transfer all the troops from the Northwest as well as from Huai-nan and Shan-nan to Yü-chou, and ordered the latter to make the 4,000-li march in a hundred days! 90  On the history of Sino-Tibetan relations, see Chapter 8, n. 115. 91  See CTS 199 hsia:3618:3 and TCTC 198, 6252–3. The details of early Khitan history are extremely sketchy, and since the Sillan annals in the Samguk sagi make fairly frequent reference to them, the lost annals of Koguryŏ would probably have been our best source. TCTC 205, p. 6505, k’ao-i, gives a brief administrative history of Ying-chou. 92   T CTC 205, p. 6505. The rebel leaders descended from tribal chieftains, and one had won the imperial surname for dynastic service. They were related by marriage and held the positions of governor (tu-tu) and prefect at the time of the rising. 93   T CTC 205, pp. 6506–7, CTS 199 hsia:3618:4, and HTS 219:4145:4. 94   T CTC 205, p. 6507.

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

The Chou Dynasty [ of Empress Wu ]



427

The northwest in the Chou Dynasty.

Even the grain transport ships from the Yangtze valley were to be diverted and continue with their present crews to the front.95 Critics at court objected to all of this, and it is difficult to tell how many of these measures could be carried 95  The only full description of the empress’ actions are contained in a memorial from Ch’en Tzu-ang, abstracted in TCTC, loc. cit., and found in full in Ch’en Tzu-ang chi, pp. 178–82.  Ch’en’s basic purpose in the memorial was to protest the levy, regarding it as unnecessary, dangerous, and an affront to the national dignity. He took the opportunity, however, to raise a number of other matters, both general and specific. Pointing out that the Khitan ambush was unfortunate and had created a very serious problem, he begged the empress to remain calm and to keep in mind that the Tibetans and Mo-ch’o’s Turks posed a more serious threat than the Khitans. To send troops from the northern and western frontiers to meet the latter was folly. Ch’en was critical of the standard of ministerial advice which the empress received, particularly that which permitted the “barbarization” of the armies, and he urged her to seek out the many competent men who had been neglected

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

428



Guisso

The northeast in the Chou Dynasty.

in appointment to high office. He spoke of self-seeking and timid ministers and conflicts between civil and military officers which the empress should no longer tolerate.  Turning to the situation at hand, he asked that the troops levied from Shan-nan and Huai-nan be given a more reasonable length of time to reach the frontier and not be threatened with the death penalty for tardiness. The diversion of grain to the front, he believed, would not only cause enormous hardship and resentment among the carriers for whom no special reward seems to have been arranged, but would cause a great price rise in the Northeast and general confusion. The actual situation at the front, he pointed out, was not even known!  Finally, and perhaps with some regional jealousy, Ch’en pointed out that because the Northeast supplied the armies, its people were not conscripted. He had recently heard that in their arrogance and rudeness, they had claimed that the government could not compel them to serve, and this was being discussed everywhere. Furthermore, the area had recently found itself plagued by a growing number of bands composed of the lawless and unemployed. Central officials should be sent out immediately to conscript them into Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

The Chou Dynasty [ of Empress Wu ]

429

out, for in the same month, June 696, Mo-ch’o’s Turks drove into Liang-chou, and the Tibetans, still in Lung-hsi, threatened to advance unless Chinese forces the forces, thus augmenting the national strength and ensuring that their relatives remain loyal to the state. He ended with a plea that the court be more generous with incentives and rewards for military service lest the task of recruitment become impossible. The common people of the Northeast were exhausted.  From this memorial, submitted less than a month after the defeat, it would seem that the neglect of the military had reached serious proportions. Ho-pei had suffered only two minor invasions since 679, the Turkish raids of 683 and 687 (TCTC 203, p. 6413, and 204, p. 6443), and when Ch’en mentioned the “magnates and roving warriors who have become bandits, the jobless drifters, the wealthy families and strong clans,” he was commenting also on the military potential of the area which we noted earlier. This seems to confirm my finding that the fu-ping declined earliest there, and it might be reasonable to suggest that the dissident elements were deserters or men who might otherwise have found niches within the system.  Two additional facts support this view. In 696 the empress set up in the frontier areas of the Northeast special units called wu-ch’i t’uan-ping, and the next year these were extended throughout Ho-nan and Ho-pei to defend against Mo-ch’o. In these units 150 households provided fifteen soldiers and one horse among them. THY 78, p. 1438 and TCTC 205, p. 6507. See also Hamaguchi, “Fuhei,” pp. 1466 ff. These units were almost certainly the prototypes of the later t’uan-chieh, and their creation seems to reflect a concession to regional feeling and an alternative to fu-ping service. Moreover, in Chang Yüeh’s report on Wu I-tsung’s Ho-pei campaign of 697 (WYYH 647:8a) we find the names of those “mentioned in dispatch.” Twenty-two men are mentioned, and the origin of their units may be tabulated as follows: Ho-pei Ho-tung Kuan-nei Lung-yu Chiang-nan Chien-nan Ho-nan

7 5 4 3 7 1 1

The table suggests that the fu-ping in Ho-pei was weak, especially in view of the fact that of the northeastern units mentioned, only one was among the 86 established in 690 when the empress augmented the forces there. It is difficult to escape the conclusion that these new units were not viable from the start.  We need not, however, read too much into these facts for they seem to say little about the oft-mentioned “separatism” of the area. They suggest more, I think, that because the Northeast had borne the brunt of the T’ang’s early imperialism, the inappropriateness of the fu-ping was revealed there first. The response of the area to alternative service and its refusal to respond to Khitan and Turkish calls to dethrone the empress support this view. Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

430

Guisso

were withdrawn from the Four Garrisons.96 Though now beset on three sides, the empress recovered from her initial panic, and decided to rely upon the calm expertise of officials like Kuo Yüan-chen and Ch’en Tzu-ang who had firsthand experience of the frontiers.97 These men rapidly evolved a series of policies which were ultimately to prove successful. Turning first to the Tibetans, and acting on Kuo’s perception of serious rivalry between the ruler and Lun Ch’in-ling, the empress immediately sent envoys to parley about a marriage alliance and an exchange of territory, but with secret instructions to exploit the factional rivalry.98 Ch’en Tzu-ang had considered the Turks and Tibetans to be a greater threat than the Khitans and, acting on his advice, the empress turned next to Mo-ch’o. Fortunately for her, the Turk had already revealed the nature of his ambitions, requesting not only that the Turks settled inside the Wall be returned to him, but also that the empress adopt him as a son and arrange an imperial marriage for his daughter.99 Knowing, therefore, tht he had designs on the Chinese throne, the empress was able to temporize while still encouraging his hopes, and to this end she promised to consider his requests while meanwhile confirming him as khan and appointing him an imperial general. These marks of favor coincided with the death of the Khitan leader Li Chinchung, and Mo-ch’o seized the opportunity to attack the unprotected Khitan base and capture the families of the absent leaders, as a result of which he was granted still more distinctions from the empress.100 It may be that the Turk had his own interests at heart in this attack on a rival foreign power but, whatever his motives, he substantially weakened the Khitans who now had no option but to advance further south, and they rapidly reached Wei-chou, less than two hundred miles from Lo-yang. Here they halted for the winter, providing the empress with a breathing space, but ominous news soon arrived that the Khitans were seeking local support for her deposition and the restoration of Chung-tsung.101 96   T CTC 205, pp. 6507–9. 97  On Ch’en’s background, see Chapter 5, and on Kuo, CTS 97:3373:2 and HTS 122:3957:1. The latter, a chin-shih graduate of undistinguished lineage, had just returned from personal negotiations with Lun Ch’in-ling and seems to have been familiar with the situation among the Tibetans since it was he who conceived the plan adopted by the empress. 98   T CTC 205, pp. 6508–9. The k’ao-i reproduces a notice from the Yü-shih-t’ai chi outlining subsequent events. 99   T CTC 205, pp. 6509–10. 100   T CTC, loc. cit. 101  The Khitan call for Chung-tsung’s restoration seems to have been made in early 697, and judging from the k’ao-i notices in TCTC 206, pp. 6526–28, the source of the report is the

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

The Chou Dynasty [ of Empress Wu ]

431

The year 697 opened gloomily. Mo-ch’o had become impatient, and he began once more to plunder the border prefectures. At the same time, a wellconnected prefect named Liu Ssu-li was persuaded by a fortuneteller that he possessed signs of greatness, and he rapidly began to form a party at court to help him seize the throne. The conspiracy seems to have been well advanced by the time it came to the ears of Lai Chün-ch’en who saw in it a chance to regain his lost favor and reported it immediately by the special (shang-pien) procedure to the empress. Not fully trusting Lai, she put the investigation in the hands of a grandnephew, Wu I-tsung, but he soon proved to be fully as unscrupulous as Liu, forcing the guilty Liu to implicate many innocents, in all probability those most opposed to the Wu interests. In all, we are told, thirtysix families, “all distinguished shih of the empire,” and over a thousand of their adherents and relatives suffered the penalties of exile or death in this final purge.102 It also brought Lai back to power, and although the empress used her typical ploy and elevated his enemy Chi Hsü at the same time,103 Lai’s power was great enough to again create a climate of fear and dread in the court, and once again to raise questions about the succession. The atmosphere of the court was becoming even more confused just then because of the meteoric rise of two unknowns, the half-brothers Chang, I-chih and Ch’ang-tsung.104 These young men, probably only in their twenties at the time, were the grandsons of a distinguished minister, Chang Hsing-ch’eng, and the elder of them had used the yin privilege to take up an official career. He Ch’ao-yeh ch’ien-tsai. Corroboration is lacking, but in view of the harmony of Khitan relations with T’ai-tsung and Kao-tsung, it seems perfectly reasonable.  There seems to have been a certain amount of collaboration with the enemy in the area, since we are told that Wu I-tsung executed several guilty persons (TCTC 206, p. 6522) before Ti Jen-chieh’s pleas for clemency were heeded. Ti’s statement (CTS 89:3356:4) mentions both those who were coerced and those who joined the Khitans willingly, but points out that the area suffered from economic depression and the unreasonable demands of harsh officials, so that under the circumstances it had been highly loyal. See also TCTC 206, pp. 6535–36. His remarks were made in late 698 after Mo-ch’o had withdrawn and refer both to Turkish and Khitan invasions. 102  On this affair, see TCTC 206, pp. 6512–13, Fitzgerald, Empress Wu, pp. 152–3, and Chapter 6, n. 70(g). 103  Lai had gone so far as to accuse Chi Hsü, the man who had brought the conspiracy to his notice, of being implicated in it. After Chi had cleared himself he, of course, hated Lai, and the empress used their hatred once again to balance her appointments. On Chi Hsü, see Appendix B, no. 120. 104   T CTC 106, p. 6517. These two figures, about whom we would like to know a good deal more, have biographies in CTS 78:3337:1 and HTS 104:3921:4; and for their ancestry, see Appendix B, no. 34. Fitzgerald, Empress Wu, pp. 163 ff., outlines their careers in some detail.

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

432

Guisso

was also “white of complexion, of great beauty and skilled in music and song,” qualities seemingly shared by his younger brother Ch’ang-tsung who attracted the attention of Princess T’ai-p’ing and won her recommendation. Taking advantage of his access to the palace to speak to the empress of his brother, he secured an introduction for him, and before long both were awarded high rank and showered with brocades, horses, slaves and other gifts. In the popular literature of later times and, indeed, even at the time, the rise of the brothers was attributed to a sexual liaison with the empress,105 yet the standard sources not only portray them as somewhat effeminate but specifically state that they were castrated (ch’eng p’i-yang chih ch’ung). Moreover, the empress was seventy at the time and, although she was “skilled with cosmetics” and so able to conceal her age,106 it seems more likely that she favored the youths because they were diverting company and, more particularly, because she believed that Ch’angtsung was the reincarnation of the Taoist immortal Wang Tzu-chin.107 The empress had apparently begun to show the interest in personal immortality which characterized so many rulers late in their lives.108 Whatever the reasons for their rise, the Changs were soon the dominant personalities at court, with even the various members of the Wu clan waiting at their mansion gate and contending for the privilege of holding their bridles!109 The Changs’ relatives were also awarded gifts and honors seemingly without limit, and life at court began to take on an almost frivolous cast. The empress was pondering whether to gild her nine recently cast tripods110 even as the 105  See, for instance, Fitzgerald, Empress Wu, p. 163 for his suggestion that the terminology of HTS and TCTC is such as to indicate that both became the empress’ lovers. For contemporary evidence, see note 159 to this chapter. 106   T CTC 205, p. 6487. A notice of the ninth month of 692 when the empress grew some new teeth and changed the era name to ch’ang-shou, “long life.” 107   T CTC 206, p. 6546 and CTS 78:3337:2. It was for this reason probably that the literary institute, founded in 699 (TCTC 206, p. 6538) to provide a sinecure for the Changs, was called Office of the Crane (K’ung-hao fu). Wang Tzu-ch’ in (6th century BC), also known as Wang Ch’iao, ascended to the immortals on the back of a white crane and so became a symbol of longevity. See also the remark of Yang Tsai-ssu in TCTC 207, p. 6572, suggesting that Chang Ch’ang-tsung had made an elixir of immortality (shen-tan) for the empress. 108  See preceding note; also TCTC 206, p. 6517 for the account of a man who won lavish rewards from the empress for claiming that he dreamt she would live forever. Jao, “Tsungchiao hsin-yang,” pp. 402–5 discusses the question in some detail. 109   T CTC 206, p. 6514 and CTS 78:3337:2. 110   T CTC, loc. cit. See also the biography of Yao Shou (Appendix B, no. 100) who dissuaded the empress from the project. The verses said to be written by the empress for the erection of the tripods are found in CTShih 5:1a. These tripods, incidentally, are said to have been cast from 800 tons of bronze.

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

The Chou Dynasty [ of Empress Wu ]

433

Khitans smashed an army of 170,000, killed Wang Hsiao-chieh, and occupied Yü-chou with great carnage.111 As if this were not enough, Mo-ch’o at the same time presented what was virtually an ultimatum threatening rebellion unless he were awarded the entire Shan-yü protectorate and unless all the Turks who had previously settled in six prefectures inside the Wall were returned to him.112 The ensuing debate seems to have jolted the court back to reality, and the empress inclined first to the “hard line” of Li Chiao and T’ien Kuei-tao,113 but was finally forced to the side of the appeasers by the magnitude of the Khitan threat. Thus she returned to Mo-ch’o several thousand tents of settled Turks, along with huge bribes of textiles, seed grain, farm implements and iron, as well as the promise of an imperial marriage,114 hoping this would temporarily satisfy him. As summer approached, she sent two more huge armies, one 200,000 strong, against the Khitans. The source of these troops is unknown, but considering the enormous levies of the past two years, these armies might be seen as one last great effort at national salvation. The overall commander was Wu I-tsung, a grandnephew of the empress whose military career was not a distinguished one.115 His appointment suggests that the troops might not have been wholly reliable and that the empress was now more concerned with keeping close rein on the military than with a frontier army’s efficiency. Fortunately for her, the tide now turned. The Khitans seemed to have no appreciable success in winning over the Chinese in occupied areas and so sought an alliance with Mo-ch’o, intending later to betray him. The Turk, however, anticipated their bad faith by seizing their newly built storehouse and base in Manchuria, taking all the booty they had hitherto gained. It seems unlikely that he planned any direct attack on the Khitans but news of his raid reached them while they were engaged in battle with the Chinese and, fearing attack 111   T CTC, loc. cit. 112   C TS 185 shang:3560:1 and TCTC 206, pp. 6515–16. In the early 670s, many of the Western Turks and the T’u-yü-hun had submitted to China (TCTC 201, p. 6363, and 202, pp. 6371–72) and had been settled in the six prefectures of Feng, Sheng, Ling, Hsia, Shuo, and Tai. To send these back to Mo-ch’o would weaken the entire North. Ts’en, T’u-chüeh, pp. 333–39 offers a good examination of the events of this year. 113  On Li Chiao, see Appendix B, no. 123 and on T’ien Kuei-tao, CTS 185 shang: 3560:1 and HTS 197:4088:3. Their position was that the Turks were covetous and lacking in good faith and that granting their demands was an invitation to invasion. Their opponents were Yang Tsai-ssu (no. 111) and Yao Shou (no. 100). 114   T CTC 206, p. 6416 and THY 94, p. 1691. The latter remarks “from this Mo-ch’o grew even stronger.” 115  On I-tsung, see CTS 183:3554:1 and HTS 206:4110:1.

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

434

Guisso

from the rear, their Hsi allies mutinied and the rest of the army scattered. Sun Wan-jung was killed by his subordinates while in flight and the remnants of the rebellious tribes joined Mo-ch’o.116 Thus at an enormous cost in men and money and after the devastation of much of the Northeast, the Khitan rebellion came to an end. As was often the case with the empress, success brought a relaxation of the harsher aspects of her rule. Even as this final campaign against the Khitans was in train, she had moved, albeit unwillingly, against Lai Chün-ch’en and had destroyed him before the final victory. His fall was of his own making, for since his return to favor he had, in open disregard of the law, denounced several high officials in order to seize their wives, concubines and possessions, and had become ever more arrogant and unrestrained. Either because he sought supreme power for himself or because he was by now demented, he accused the Wu princes and Princess T’ai-p’ing of engaging in a conspiracy with Chung-tsung and Jui-tsung. The princess, whose shrewdness seems almost to have matched her mother’s, responded by leading the accused in a countersuit. Lai’s rise had been based, of course, on the fact that among his many false denunciations, he had identified the occasional genuine plot, and on the basis of this service to the state, the empress thought to pardon him. Ironically, it was Chi Hsü who had once been accused by Lai, who persuaded her otherwise and she permitted his execution though, as we have seen, she sent his chief enemy Li Chao-te to the same block.117 She also made her peace with Lai’s faction. The decree outlining Lai’s crimes also pardoned the many officials whom he had intimidated into making false appointments.118 Later that year she called together her ministers and, explaining why she had placed such trust in Lai and in Chou Hsing before him, acknowledged that she had been deceived and promised henceforth to eschew such methods.119 It is impossible to gauge the sincerity of the empress’ repentance, but it is a fact that in mid-697 the terror finally came to an end. Moreover, contemporaries seem to have sensed the change of climate, for when Wu Ch’eng-ssu and San-ssu renewed their succession claims early the next year, the ministers united in opposition, showed no fear of remonstration, and in fact were able to speak frankly with impunity. Led by Ti Jen-chieh, who was henceforth to enjoy

116   T CTC 206, pp. 6520–22. 117  On Lai’s activities at this time and his fall, see CTS 186 shang:3564:3 and TCTC 206, pp. 6518–19. 118  See CTW 95:8b for the edict of punishment, and TCTC 206, p. 6520. 119   T CTC 206, p. 6523.

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

The Chou Dynasty [ of Empress Wu ]

435

the complete trust of the empress,120 several went so far as to suggest the regularization of the succession by the elevation of Chung-tsung, and Ti disposed of the old argument that the question was a “household matter” by pointing out that for a ruler, all within the empire constituted his household.121 The empress took time over her decision and, although in Ssu-ma Kuang’s judgment it was Ti who persuaded her definitively against a Wu succession,122 ironically enough, it was the Chang brothers who successfully urged Chung-tsung upon her.123 It is unlikely that the initiative for this lay with them, but although they are seldom credited with any political acumen at all, they could well have at least recognized that their survival would depend on others once their protector died. In the third month of 698, ostensibly for medical reasons Chungtsung was brought back to the capital.124 These events raise interesting questions about the nature of Wu Tse-t’ien’s power, foremost among them perhaps the question of how important coercion was in its maintenance. The answer seems to be that until the bureaucracy was satisfied that the will of Kao-tsung was to be carried out and Chung-tsung was assured succession to the throne, punishments were an integral part of the Wu regime, as important to its survival as its state ideology or the winning of popular support. If this is so, my earlier suggestions on the highly personal nature of the empress’ legitimation are strengthened for, at the risk of repetition, it must 120  On Ti, see Appendix B, no. 93, and TCTC 207, pp. 6550–1. Ti was the most famous and perhaps the greatest of the empress’ ministers and is the subject of a large secondary literature. A man of uncompromising integrity and infinite subtlety, he won the empress’ complete trust and, “although he was fond of arguing with her [in court], she accepted it and followed his advice.” She is said to have called him the “old man of the state” (kuo-lao) and in one incident, recounted in TCTC, loc. cit., is said to have valued his safety more than that of the crown prince. She was disconsolate on his death in 700.  It must be acknowledged, however, that Ti’s power was never total, partly because Wu San-ssu usually occupied a parallel position in the hierarchy and partly because the empress seems to have insisted on the invulnerability of the Changs. Probably because of their relative harmlessness, Ti never sought a confrontation with them. His biography suggests a different reason for this, pointing out that he recommended “several tens” of T’ang loyalists for high posts, so that these were in a position to carry on his work when he died. Most notable among them were three leaders of the restoration coup, Chang Chienchih, Huan Yen-fan and Ching Hui. See also TCTC 207, pp. 6551–52. THY 75, p. 1357, says he recommended all five leaders of the coup. 121   T CTC 206, p. 6526. 122   T CTC, loc cit. “From this [remonstrance] the empress-dowager had no intention of establishing Ch’eng-ssu or San-ssu.” 123   T CTC 206, pp. 6526–27. 124   T CTC 206, p. 6537, and see the lengthy examination of the event in the k’ao-i notice.

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

436

Guisso

be pointed out that her rule had to produce divided loyalties within officialdom. Though most high officials recognized that her ability exceeded that of the rightful heir and therefore deferred to her claims, they also saw themselves as guardians of the well-defined tradition requiring the Son of Heaven to be a male. Thus the empress, whose whole career was in violation of this tradition, could never be certain whether loyalty to her ability would remain uppermost in her officials’ minds. Perhaps more than anything else, it is a measure of the flexibility of T’ang Confucianism that compromise was possible but, since this compromise came about only with the second decade of the empress’ de facto rule and since the compromise was essentially a Confucian victory, it too was a measure of the strength of tradition. With the succession now regularized, the empress would be permitted to remain in power but only so long as she continued successfully to fulfill the obligations of ruler. This is not to say, however, that she became a “lame duck,” a prisoner of the upper bureaucracy, in the last years of her reign. For instance, when Ti Jen-chieh became paramount minister, she took care, as always, to balance his appointment this time with a new position for Wu San-ssu.125 She also stopped short of proclaiming Chung-tsung as crown prince immediately and, as we shall see, she insisted on the invulnerability of the Changs in spite of their increasingly flagrant corruption. To increase her base of support in the lower bureaucracy, she put Li Chiao in charge of the selection system in late 697 where he “began the establishment of several thousand yüan-wai-lang,”126 and permitted the more rapid advancement of lower officials within the bureaucracy.127 In spite of these measures, however, the effect of the terror’s end and the return of Chung-tsung was still enough to produce a greater spirit of cooperation between the empress and her ministers and one which led to happy results, particularly in foreign policy. The defeat of the Khitans, and the necessity of placating Turks and Tibetans to achieve it, seems to have led to the rise of a strong and articulate pacifist party at court, and one which advocated for the first time the abandonment of some previously gained territory. In the winter of 697 Ti Jen-chieh put the case for this into an erudite memorial128 which first outlined China’s natural 125   T CTC 206, p. 6532. 126   T CTC 206, p. 6525. As pointed out earlier, it was Li Chiao who first requested the reduction of yüan-wai officialdom in the next reign. 127  See Chapter 7. See also TFYK 629:20b–21a for a decree of 696 which recognized that advancement “within the current” had been much too rapid. 128  The full text is in CTW 169:2b, and the memorial is abstracted in THY 73, pp. 1326–27 and TCTC 206, pp. 6524–25.

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

The Chou Dynasty [ of Empress Wu ]

437

barriers against the foreigners, and pointed out that the wise rulers of ancient times had recognized that conquest beyond these borders was the use of national wealth to acquire worthless and unproductive land which would yield no revenue. Ambitious rulers like Ch’in Shih-huang-ti and Han Wu-ti had impoverished the state with their expeditions. Recently, he pointed out, successive campaigns had eaten up huge sums of money and still continued to do so, in spite of the fact that “in Kuan-tung there is famine, in Shu and Han [households] flee their registration, and south of the Chiang and Huai the levies never cease, [so that] people cannot return to their livelihood and one after another turn to banditry.” He recommended that the empress abandon her present policy and return to that of T’ai-tsung. For the North and West,129 this would mean utilizing largely native administrations, eschewing far-off garrisons and attacking foreigners only if they rebelled. A-shih-na Hu-se-lo130 should be made khan and placed in charge of the Four Garrisons, while a scion of the old ruling family of Koguryŏ should be given An-tung to hold. Chinese troops should store up supplies and guard the borders against attack while spying on potential enemies. If the Turks and Tibetans attacked in force they would risk defeat, and mere raids would reap little profit. According to the Comprehensive Mirror, those conversant with foreign affairs agreed with Ti.131 Although the empress did not implement Ti’s specific recommendations, she tended during the remainder of the Chou to follow the general policy lines he had set down. Judging from memorials which Ch’en Tzu-ang submitted that year, the empress began to economize as early as 698. She abolished the T’ung-ch’ang army and the huge annual corvée which transported supplies to the northwestern garrisons, and ended the levies and corvée in Chien-nan.132 Even with Mo-ch’o, whose presence had made Ti’s specific 129  For Ti’s reference here, see TCTC 195, p. 6148. 130  I cannot identify this figure. 131   T CTC 206, p. 6525. 132   Ch’en Tzu-ang chi, pp. 173–78. In general Ch’en approved of the demobilization but warned the empress that the self-seeking officials in the prefectures of Sung and Mao, deprived of the profit they reaped from the supply of the military, might falsely stir up trouble with the Ch’iang in order to restore their fortunes. He spoke also of 30,000 vagrant households who had fled to the West from the Szechwan region, attributing their migration to the harshness of officials there and requested administrative reform there. Finally, since he saw the necessity for keeping the T’ung-kuei Army stationed in Sung-chou active, he recommended a much less costly and onerous means of keeping them supplied.  Perhaps because he was a westerner, Ch’en was more conscious than most of the Tibetan threat. The fact that even he advocated a relatively passive stance toward the foreign neighbors suggests that the attitude was becoming widespread in officialdom.

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

438

Guisso

proposals impossible to grant, the empress attempted conciliation. Obviously seeing him as a greater threat than the Tibetans whose invasion route she was content to garrison,133 she abruptly agreed to a marriage alliance in the sixth month of 698, sending Wu Yen-hsiu, the second son of Ch’eng-ssu, to marry Mo-ch’o’s daughter.134 The dispatch of a prince of the blood to marry a foreign woman (rather than a Chinese princess to marry a foreign male) was wholly unprecedented, a fact quickly pointed out by orthodox officials. The empress quickly stilled her critics by demoting one of them, Chang Chien-chih, to a remote southern prefecture.135 He would return in 704 as tsai-hsiang and then take charge of the restoration coup. The sending of a male “bride” to the Turks turned out to be a serious miscalculation, for Mo-ch’o saw immediately that the empress had no intention of granting his major demands and by this time he must also have realized that the succession was to be the Li and not the Wu. Thus he pointed out to the envoy who had accompanied Wu Yen-hsiu that he had wanted his daughter to marry into the Li clan who had always treated his people well, and he threatened to restore the T’ang. In the eighth month of 698 he sent forth troops for that purpose and achieved several rapid successes, with one of the three frontier armies upon which he made a surprise attack submitting voluntarily and without a fight. At the same time, he sent angry missives to the court, listing five grievances against the empress and threatening to seize Ho-pei for himself.136 Her response was to rouse the empire for yet another great effort, recruiting four armies said to total 450,000 men and drawn in all probability from throughout the country.137 Even a force of this size, however, proved unable to pin down the mobile foreigners, and she was forced to call for even more troops. This time the people balked. In over a month, we are told, no more than

133  See the preceding note and TCTC 206, p. 6530. Lou Shih-te was appointed commissioner to assess the garrisons in the area. 134   T CTC 206, pp. 6530–31 and CTS 183:3553:4. Yen-hsiu, the second son of Ch’eng-ssu, reached his destination in the eighth month of 698. He did not return until 704, and in 708 married Princess An-lo. 135  See Appendix B, no. 133 and TCTC 206, p. 6530. 136   T CTC 206, pp. 6530–32, CTS 194 shang:3598:2, TFYK 964:10b and 998:8b. 137   T CTC 206, p. 6533; see also Ts’en, T’u-chueh, pp. 339–50 which offers a full examination of the events of this year, introducing two inscriptions which contradict the standard account. We have no details on the constitution of the force, but since the three commanders were designated as tsung-kuan of the eastern, western, and central provinces, it seems likely that the troops were raised throughout the empire.

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

The Chou Dynasty [ of Empress Wu ]

439

1,000 men responded to the call to arms,138 and she was finally forced to play her trump card. In form acceding to a request from Jui-tsung, she declared Chung-tsung heir apparent and two days later made him supreme commander (yüan-shuai) of the Ho-pei armies, supplanting her grandnephew Wu Chung-kuei.139 The appointment of Chung-tsung was, in fact, a nominal one with Ti Jen-chieh actually filling the post, but its propaganda value seems to have been great since the sources tell us that before long 50,000 men joined the army.140 Ti was able to lead an additional 100,000 men into the field, and this seems to have convinced Mo-ch’o that the combined Chinese forces were too strong to fight and so he withdrew to the steppes, leaving behind a death toll estimated by one source at 80,000 to 90,000 persons.141 His own army had swollen to 400,000 and as the various tribes north of the Wall submitted to him, he briefly became the greatest power in Central Asia, almost reconstituting the great T’u-chüeh empire of the mid-sixth century and exhibiting “a highly contemptuous attitude toward China.”142 The empress’ failure to complete the subduing of Mo-ch’o is the principal reason traditional historiography tends to regard the Chou as a weak dynasty, but in retrospect this seems unfair. The Turks, after all, were driven from China and, although Mo-ch’o was to return once more, his last raid was in no sense so serious a threat as the invasion of 698. It is unlikely, moreover, that the empress had either the ability or the inclination to go onto the offensive. As mentioned earlier, she faced a strong peace party at court, and she also had some doubts about the reliability of her troops when those she did send in pursuit 138   T CTC 206, p. 6534. It is difficult to know whether or not this is true. The dynastic histories reproduce the story only in Ti’s biography and there as his claim rather than as fact. It is found also in Ta-T’ang hsin-yü 1:11a. I am prepared to believe that Chung-tsung’s appointment increased the levies, but the story still seems too contrived to be fully credible. 139   T CTC, loc. cit. It is important to note that the new volunteers came from Ho-pei since it is a strong indication that the area was not “separatist” or disloyal to the T’ang. The other possibility, of course, is that Ho-pei had been alienated by Wu I-tsung’s cruel swoop through the province in 697 to punish those suspected of collaboration with the Khitans. He acted so harshly then that the inhabitants composed a bitter proverb about him. TCTC 206, p. 6522. Undoubtedly the empress’ popularity suffered. 140   T CTC, loc. cit. It is interesting to note that while Ti Jen-chieh led the new army against Mo-ch’o, the empress placed command of the remaining troops (t’un-ping) at the capital in the hands of Wu I-tsung and Wu Yu-kuei, the first sign that she was concerned about her security. TCTC 206, p. 6535. 141   C TS 194 shang:3598:2. Ssu-ma Kuang uses the conventional term “over 10,000” but adds that casualties in the region from Mo-ch’o’s five invasions were countless. 142   T CTC 206, p. 6535. See also Grousset, L’empire, pp. 150 ff.

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

440

Guisso

of Mo-ch’o “did not dare to press [him].”143 She undoubtedly interpreted the reluctance of the people to enroll in the army in 698 as a sign of their desire for rest and, always sensitive to the popular mood, when Mo-ch’o withdrew she seems to have decided that she would henceforth play a waiting game and use diplomacy rather than war to deal with her enemies. She was fortunate, therefore, when her Tibetan policy of fomenting internal dissension at last bore fruit in the summer of 699. While Lun Ch’in-ling was out of his capital, the young btsan-po used subterfuge to escape confinement and struck hard at his enemy, seizing and executing 2,000 of Lun’s supporters in the capital. He then attacked and routed Lun himself. Lun committed suicide and his followers, including over 7,000 tents of T’u-yü-hun, sought refuge in China.144 Perhaps in an attempt to consolidate his position, the btsan-po invaded China the next year but was driven back when the experienced general T’ang Hsiuching bested him in six successive battles.145 Rather than follow up these victories, the empress chose to consolidate, putting Kuo Yüan-chen in charge of strategic Liang-chou. During the next decade his capable administration is said to have secured the borders, brought prosperity to the area, and won the respect of the Tibetans.146 In 703 after yet another defeat, they resumed tribute payments to China and requested a marriage alliance.147 The sources do not record whether or not this was granted, but the accession of a seven-yearold ruler in the following year (704) ensured that the Tibetans would pose no threat for the remainder of the Chou. Of Mo-ch’o little was heard in China until he again invaded early in 702. He swept deep into Shensi and reached the empress’ home prefecture of Ping-chou before halting. Chinese countermeasures were swift. The “Shan-tung Defense Army,” made up entirely of northeastern troops had earlier been established for just such an eventuality148 and, shortly thereafter, a still larger force was levied, and nominally headed by Jui-tsung who received the title of supreme commander. Although the Turks could not be pinned down, the rapidity of the Chinese response seems to have impressed Mo-ch’o, and he retired without engaging in a major battle. In the sixth month of 703 he sent an envoy to request that his daughter be married to a son of Chung-tsung.149 The sources are not 143   T CTC, loc. cit. 144   T CTC 206, pp. 6539–40 and CTS 196 shang:3604:2. 145   C TS 93:3366:3 and TCTC 207, p. 6549. 146  On Kuo, see CTS 97:3373:2, HTS 122:3957:1, and TCTC 207, pp. 6557–58. 147   C TS 196 shang:3604:2 and TCTC 207, p. 6562. 148  See TCTC 207, p. 6558 and 206, p. 6539; also note 95 above. 149   T CTC 207, p. 6562.

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

The Chou Dynasty [ of Empress Wu ]

441

specific on whether or not his request was granted, but a marriage of some sort was certainly arranged since Mo-ch’o sent a message of gratitude at the end of the year.150 That he remained active outside the Wall is evidenced by certain administrative changes on the frontier,151 but after one more great victory at Ming-sha near Tun-huang in 706, he caused no further trouble in China. He was assassinated by a rival in 716.152 In view of these facts, the later stages of the Empress Wu’s foreign policy should be seen neither as unrealistic, overambitious, nor as a failure. The diminution of foreign threats after 699 should perhaps have made possible an attempt to solve some of the longstanding internal problems, but this was not the case. Part of the reason lies, of course, in the empress’ age. Although early in 699 she is said to have grown new eyebrows, shortly afterwards she suffered her first recorded illness, while visiting Mount Sung, in Taoist legend the site of a miraculous ascent to the land of the immortals by Wang Tzu-chin in the sixth century BC.153 Although she recovered slightly, she suffered a relapse the next year. She improved again when she took an elixir of long life fabricated by a Buddhist monk.154 Her health seems gradually to have declined thereafter and with it her extraordinary administrative abilities seemed to diminish. She may also have felt that the time for relaxation had come, that she was at last secure in her legitimation. Indicative of this is the fact that immediately after her second illness she divested herself of the last of her grandiose titles.155 A second reason for the administrative decline which characterized the last years of the Chou is the death of Ti Jen-chieh in the autumn of 700. As the empress put it, his death left the court empty156 and, though she lamented his absence each time an important decision was called for, she was to find no one to replace him. It was partially for this reason that she made perhaps the 150   T CTC 207, p. 6568. CTS 194 shang:3598:2 says that the empress brought two of her grandsons, Ch’ung-chün and Ch’ung-ming, into the audience hall to show them as prospective bridegrooms to the Turkish envoy. Ch’ung-chün was shortly to be active at court, and no other source mentions that Chung-tsung had a son called Ch’ung-ming. Perhaps Mo-ch’o’s new son-in-law was not a prince at all. 151  See TCTC 207, pp. 6562–63. This incident, in fact, involved only the Western Turks but, according to CTS 194 hsia:3600:3, the An-pei protectorate was set up at this time at T’ing-chou. 152   C TS 194 shang:3598:2 and Grousset, L’empire, pp. 157–58. 153   T CTC 206, p. 6539. 154   T CTC 206, p. 6546. The elixir was a costly one and had taken three years to synthesize. 155   T CTC, loc. cit. and CTS 6:3076:4. In celebration of the event she changed the era name to chiu-shih, “everlasting youth,” from the phrase ch’ang-sheng chiu-shih in the Tao-te ching. 156   T CTC 207, p. 6551.

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

442

Guisso

worst decision of her career, turning more and more to the Chang brothers for company and for stimulation. At the beginning of 699, and probably in response to complaints that the Changs did nothing to deserve their favor, she had established for them a new institution, the k’ung-hao fu or Office of the Crane. The crane, of course, was a symbol of longevity and the means by which Taoist immortals were conveyed to Heaven,157 and the purpose of the office was ostensibly that of literary compilation. Not only were a number of competent scholars appointed to it, but in 701 its members produced at least one large and important work.158 In mid-700 its name was changed to the feng-ch’en fu or the Office of Imperial Attendants, but by that time it had already acquired a rather unsavory reputation as the site of evenings of drinking, gambling, and other unseemly and undignified activities in which the empress herself and high officials who wished to curry favor with the Changs often participated. Needless to say, this gave rise to scandalous rumors and, the old T’ang History tells us, the empress went so far as to order that handsome youths be recruited as attendants, an act which caused one official to remonstrate: Your Majesty has already granted intimate favor (nei-ch’ung) to Hsüeh Huai-i and to Chang I-chih and Ch’ang-tsung. Surely this should be enough. Recently I heard that a head of the Servants of the Imperial Apartments, called Liu Mu, himself claimed that his son Liang-pin was white and pure, beautiful in beard and eyebrow. The chief administrator of the Gate Guard of the Left, Hou Hsiang, said that his virility (yang-tao) and robustness surpassed that of Hsüeh Huai-i, and unassisted he wanted to recommend himself as fit for membership in the Office of Imperial Attendants.159 It seems clear, therefore, that in the popular perception the office was viewed as something akin to a male harem, and the empress’ reputation suffered. 157   T CTC 206, p. 6538 and CTS 78:3337:2. Among the best-known scholars appointed to the institute were Li Chiung-hsiu (Appendix B, no. 128), Chi Hsü (no. 120), T’ien Kuei-tao and Yün Pan-ch’ien. The latter, disgusted at the unprecedented nature of the office and the quality of its incumbents, immediately resigned and was demoted for his presumption. See CTS 190 chung:3582:2. 158   T HY 36, p. 657 and TCTC 206, p. 6546. This was the San-chiao chu-ying and probably marks the first time Buddhism was officially classified as a teaching equal to Confucianism and Taoism. 159   C TS 78:3337:2. Fitzgerald, Empress Wu, p. 167 seems to put a different interpretation on the matter. The memorial is abstracted in TCTC 206, p. 6546.

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

The Chou Dynasty [ of Empress Wu ]

443

She is said to have rewarded the author of the blunt memorial quoted above and remarked that had it not been for his plain speaking, she would not have known the situation existed.160 In all likelihood this was true, for the empress’ attraction to the Changs seems to have been based more than anything else on their youth and her belief that Ch’ang-tsung, whom she dressed in feathers and mounted on a wooden crane, was the reincarnation of Wang Tzu-chin161 and hence in possession of the secret of longevity. When we recall that even the great T’ai-tsung, who seems to have been far less superstitious than many of his contemporaries, was not immune from the lure of immortality, her infatuation becomes more comprehensible. These activities of the Changs, aside from the fact that they distracted the empress from more important concerns and somewhat lowered her prestige, were relatively harmless, and there is no record of Ti Jen-chieh, for instance, offering any protest. More harmful, however, was their growing tendency to interfere in politics and to indulge in corruption on a grand scale. This latter activity became apparent almost from the beginning of their careers at court. In fairness it must be acknowledged that the two favorites were less corrupt than other members of their parvenu clan. Chang Ch’ang-i, for instance, sold posts lavishly. On one occasion, when he forgot the given name of a client, he had all sixty persons of the same surname on his lists appointed to office!162 When he and two other brothers were tried for corruption in 704, the three of them were found guilty of accepting over 4,000,000 cash in bribes, and others of the Changs’ proteges were just as bad.163 None of the Changs, it seems, possessed any degree of political talent or literary ability and had been notable failures even in relatively unimportant provincial appointments.164 The old T’ang History explicitly remarks that when Ch’ang-tsung was vice-president of the Board of Rites, he had to have documents ghost-written for him.165 In spite of this, however, the two favorites had no qualms about intervening in state affairs, and the empress, says one source, 160   T CTC 206, p. 6547. 161   T CTC 206, p. 6546 and CTS 78:3337:2. 162  It is difficult to get a complete picture of the Chang clan, but since the principals were referred to as “fifth and sixth master,” they must have had several brothers. Two elder and one younger are mentioned in the histories, all of them involved in corruption. On the case of bribery in the hsüan, see TCTC 206, p. 6547. 163   T CTC 207, p. 6572. Both Li Chiung-hsiu (Appendix B, no. 128) and Chang Hsi (no. 124), who were otherwise reputable officials, seem to have fallen into the same ways. 164  See, for instance, TCTC 207, p. 6563. 165   C TS 78:3337:2.

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

444

Guisso

often entrusted political matters to them.166 As a consequence late in 701 they were responsible for an inner-court tragedy with wide ramifications when they denounced Chung-tsung’s son and daughter along with her husband, a Wu prince, for criticizing the favor they enjoyed. The empress, in a burst of anger, ordered all three to commit suicide, thus leaving Chung-tsung without an heir. The existence of such an heir might have helped prevent the intrigue and bloodshed which marred the next two reigns.167 The Changs allied themselves with the shrewd Wu San-ssu who was himself playing a double game by this time,168 and this formidable alliance so intimidated even the Li clan that in 701 Chung-tsung, Jui-tsung and T’ai-p’ing joined together in the sycophantic request that Ch’ang-tsung be granted princely rank!169 Most of the ministers, needless to say, now feared to raise their voices against the favorites, with the single exception of Wei Yüan-chung, an outspoken and upright man who had four times endured disgrace at the hands of unworthy officials, but whose integrity and military ability had gradually brought him to the somewhat diminished position of paramount minister after the death of Ti Jen-chieh.170 He had twice indirectly attacked the favorites and now openly warned the empress against the “inferior men” (hsiao-jen) with whom she surrounded herself. The Changs accused him of treason and in 703 his trial involved much of the court.171 The details of the trial have been recounted elsewhere,172 but its outcome illustrates how high the Changs had risen and the effect their rise had on the court. Although in the course of the trial it became clear that Wei Yüan-chung was innocent, and although the empress was warned that her behavior was beginning to cause good men to “slap their thighs at home” even though they had to “gag their mouths at court”173 for fear of the Changs, she knowingly overrode 166   T CTC 207, p. 6556. 167   T CTC 207, pp. 6556–57, and CTS 78:3337:2. Chung-tsung’s crown prince was of necessity his son by a concubine, and Empress Wei is said to have been so unhappy with this situation that she sought to have her daughter An-lo replace him. Her actions goaded the Crown Prince Ch’ung-chün into an attempted coup in the summer of 707, initiating a period of unremitting palace intrigue which ceased only with Hsüan-tsung’s accession. 168  On Wu San-ssu, see Appendix B, no. 118. He is said at this time to have formed a liaison with Empress Wei in preparation for her husband’s succession. 169   T CTC 207, p. 6559. The empress refused, awarding instead dukedoms with revenues of 300 households to both brothers. The TCTC notice is somewhat misleading here since the titles were a reward for literary compilation. CTS 78: 3337:2. 170  On Wei, see Appendix B, no. 121. 171   T CTC 207, pp. 6563–66. 172  See Fitzgerald, Empress Wu, pp. 173–177, which follows closely the TCTC account. 173   T CTC 207, p. 6565.

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

The Chou Dynasty [ of Empress Wu ]

445

justice for the first time in her career, exiling Wei Yüan-chung to a low provincial assignment in the Far South. Even in the face of a concerted effort by ministers to have the verdict reversed,174 she insisted it be carried out and by so doing she began to convince some key people that she was no longer competent to rule, and for a woman that was enough to undermine her legitimacy. Deficiencies began to be noticed in other areas and, as problems went unsolved, the support of various constituencies began to dissipate. The upper bureaucracy, already somewhat alienated by her appointment policies and her support of unworthy favorites was the first to defect. Though the empress was highly successful in the first years of the dynasty in minimizing high official opposition by her patronage of Confucianism, by the late 690s it was becoming widely recognized that her support had been more apparent than real. In 699, for instance, one of the grand secretaries, Wei Ssu-li,175 memorialized on the state of education, inspired, according to the Comprehensive Mirror by the fact that the empress often used unscholarly (fei ju-shih) members of her own clan and relatives by marriage for administrative and teaching positions in the imperial schools and supplemented them with their own poorly trained students who gained their positions as rewards for assisting her in ceremonials.176 Largely for this reason, Wei contended, over the last two decades the youth of the empire had come to have contempt for those charged with Confucian learning and, as a result, shunned both the schools and the doctrines they taught. This made little difference to the scions of powerful clans (kuei-men) who could still easily gain appointment through favoritism, but poorer people were unable to rise and thus the bureaucracy was becoming riddled with unworthy and incompetent officials. From the time of Chung-tsung’s deposition the bureaucracy had expanded rapidly, and at the same time evil officials had risen to intimidate or execute those honest officials who remained. Corruption in the selection system and the avarice of local officials had aggravated a bad situation. Wei’s solution was to expand the school system and order the sons of the nobility and upper classes to enroll so as to ensure that they could use no other method to gain office. At school they should be required to obey the rules, to venerate Confucianism, and assist at the lectures of virtuous and erudite men. As a practical inducement to the adoption of this plan, Wei pointed out that it would assure a supply of highly qualified officials at all levels, and instead of the present situation where “more than half the empire’s households have 174   T CTC 207, pp. 6565–66. 175   T FYK 604:4a, TCTC 206, p. 6542, and the full text in CTW 236:3b. On Wei, see Appendix B, no. 131. 176   T CTC, loc. cit.

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

446

Guisso

fled their registration, the tsu and tiao [taxes] shrink, and our budget is insufficient,” the state would be prosperous and happy.177 There is probably an element of exaggeration in Wei’s argument, but evidence from other sources tends to confirm his views on the extent of corruption in the selection system and the declining standards of provincial administration.178 The nature of our sources does not give us access to details of appointments to the capital schools, but Wei is probably correctly identifying a trend since, as we have seen, the literary emphasis in the examinations was regarded by some as detrimental to a more substantive Confucianism, and the several “decree” examinations of the Chou seem also to have been literary in bias.179 For these reasons the last years of the empress may have seen the beginnings of a progressive failure of the Confucian component of her state ideology, and although she made Wei Ssu-li rector of the state university, the imbalance was exacerbated as she turned more frequently to the mystical comforts of Taoism and Buddhism. Recent scholarship has shown that quite aside from her relationship with the Changs and her visits to Mount Sung, the empress seems to have developed a greater interest both in Taoist festivals and literature in her last years, and she bestowed her patronage on both.180 Still more conspicuous and more unpopular with Confucians, however, were the fruits of her long support of Buddhism, a patronage of whose dangers, I have tried to demonstrate, she was consistently aware. Nonetheless, by the year 700 officials could complain of how Buddhist foundations surpassed even the imperial palaces in magnificence and of how monks squeezed the people and disregarded their own Law.181 There were signs that after 700 the empress sought to compensate for her earlier patronage by, for example, ending the ban on butchery and attempting to tax the clergy rather than the people at large for new images she wished to erect.182 These, however, were palliatives and, as we have seen, she remained intensely Buddhistic until the end of her life. The effect of this partial withdrawal of support may, however, have further weakened her position, by raising doubts among the Buddhists about the reliability of her support without 177   C TW 236:3b ff. 178  See, for instance, TCTC 207, p. 6557, on selection. The problem of provincial administration is discussed on p. 151 in the text. 179  See Chapter 7. 180  Jao, “Tsung-chiao hsin-yang,” pp. 403–405. 181  See TCTC 207, pp. 6549–50, and my remarks in Chapter 4. 182  Chapter 4. On the clerical tax, see TCTC 207, p. 6571.

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

The Chou Dynasty [ of Empress Wu ]

447

really appeasing the Confucians. Certainly the report of a miraculous appearance by the Buddha immediately after the ending of the butchery ban suggests an attempt by perhaps opportunistic Buddhists to warn the empress not to abandon her Buddhist loyalties.183 Some dramatic act was necessary to restore this crumbling ideological hybrid. This may be why in November 701 the court returned to Ch’ang-an for the first time in twenty years,184 and remained there for almost exactly two years. Although it is sometimes suggested she made the move for reasons of health, it is far more 1ikely that the empress’ motives were political and ideological. Just prior to the move, she had ordered the suicide of Chung-tsung’s children, and in that same month there had appeared the first open suggestion that she abdicate, in the form of a memorial from an otherwise unknown scholar, Su An-heng. Su pointed out that she had held the empire in trust for over twenty years and that since Chung-tsung had shown himself to be both respectful and mature, it was time the throne be returned to him. She had, moreover, more than twenty grandsons who should occupy the princely ranks now held by the Wu clan, thus ending the anomalous situation of having two imperial clans and avoiding trouble in the future.185 While the empress failed to act on his suggestion, she rewarded rather than punished him, thus showing a clear recognition that his viewpoint was not an isolated one. The return to the T’ang capital, accompanied by the proclamation of the new era name of ch’ang-an, was designed, therefore, both to reduce political pressures and to show the sincerity of her Confucianism, for like the Duke of Chou she was preparing now to “hand over the throne to the intelligent prince.” Although once back in Ch’ang-an the principal concern of the court was how to deal with Mo-ch’o, the empress was also able to make some repairs in her ideological policy and to tighten the administration somewhat. She now extended the examination principle to its logical conclusion by the establishment of the military examination,186 and she declared an absolute moratorium on punishments concerned with the early rebellions against her while ordering a reinvestigation of Lai Chün-ch’en’s cases, an action which resulted in the

183   T CTC 207, p. 6554 and see the k’ao-i notice. 184   C TS 6:3076:4 and TCTC 207, p. 6557. It may be significant that Jui-tsung was given command of the Yü-lin Guard immediately before the empress’ departure. 185   T CTC 207, p. 6556. Su memorialized again on the same subject the next year. See TCTC 207, p. 6559, also TFYK 544:6b. 186  Chapter 7. See Examens, pp. 36 and 209–212.

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

448

Guisso

rehabilitation of many of his victims.187 Several popular appointments including those of Li Chiao and Wei An-shih were made, Wei Ssu-li was put in charge of the state university (kuo-tzu chien), and a new type of censor was established.188 The order for the composition of a national history in 703189 was a Confucian act, and the placing of Wu San-ssu at the head of the compilers may indicate that the empress was about to withdraw from active life and wished to determine her place in history before doing so. Unfortunately, any restoration of her popularity among her ministers which might have resulted from these measures and from the successful resolution of the foreign problem was dissipated by her actions in the aforementioned trial of Wei Yüan-chung, and when she returned to Lo-yang immediately after the trial, she deemed it prudent to leave her grandnephew Wu Yu-i in charge of (liu-shou) Ch’ang-an.190 Back in Lo-yang, and with an energy and competence surprising in a beleaguered and ailing woman nearing her eightieth birthday, the empress at last confronted some of the long-standing problems of her dynasty. Probably at the suggestion of Li Chiao, who had regained tsai-hsiang status in 703, she sent special commissioners throughout the empire to examine the standards of provincial administration on the basis of the six categories (liu-t’iao) used by the Han and perhaps to deal directly with the vagrancy problem.191 Shortly afterwards she held a discussion on the problem of local administration. Li Chiao and T’ang Hsiu-ching argued that she was largely to blame for bad local government because she used demotion to the provinces as a punishment for unworthy capital officials. As a consequence, officials “all value capital posts and have contempt for provincial posts.” Her remedy was to take twenty experienced and able capital officials and concurrently appoint them as prefects.

187   T CTC 207, pp. 6560–61. It was not until 704, however, that complete rehabilitation occurred and the road to office was reopened to those whose relatives had been found guilty of crime. TCTC 207, p. 6574. 188  On the appointments of Li and Wei to tsai-hsiang rank, see TCTC 207, p. 6571. Wei An-shih (Appendix B, no. 125) was already well-known as an opponent of the Changs. Wei Ssu-li (no. 131) held his position in the university until the twelfth month of 704 and, in view of his concern with education, was probably a successful administrator. On the new censors, see THY 60, p. 1053 and TFYK 512:6a. 189   T FYK 554:17a and THY 63, p. 1094. 190   T CTC 207, p. 6567. 191  The concern perhaps arose from a disturbance in the far South in the eleventh month of 703. TCTC 207, pp. 6568–69. I have been unable to find a definition of the “six categories” of the Han.

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

The Chou Dynasty [ of Empress Wu ]

449

According to Ssu-ma Kuang, this experiment was a failure,192 but it demonstrated the empress’ consistent concern for the people and her willingness to attempt unusual tactics to better their lot. It was on these grounds, too, that she was dissuaded from completing a huge image of the Buddha in mid-704193 and, perhaps because of measures of this sort, we find a good deal of evidence to suggest that to the very end she retained her great popularity among the people.194 The same cannot be said, however, of the feelings of her bureaucrats. Although her staff increase in the second university at Lo-yang in early 704195 and the amnesty and rehabilitation of all the earlier victims of Chou Hsing and Lai Chün-ch’en at the beginning of 705196 were undoubtedly popular, her expansion of the number of sub-fifth-rank officials within the bureaucracy still frustrated all her efforts to win the united support of the higher officials. The empress had always seen as one of her tasks the reduction of ministerial power relative to that of the throne. Her “excessive” appointments and rapid promotions within the lower bureaucracy had been a useful weapon in this conflict and one which had evoked numerous protests from high-ranking officials. With the rise of the Changs, however, and the death of Ti Jen-chieh who seems to have served as a kind of buffer, there are signs that the empress finally reversed this policy. In 701, for instance, the promotion process (k’ao) was tightened.197 Acts of Grace, of which there had been eleven in the preceding eleven years, ceased and when new posts were established, as in the case of the provincial censors, the number of officials to hold them was specified.198 At the end of 704 she abolished all the new posts established since 701.199 It is unlikely that economic motives alone were behind these decisions for she was still to suggest the erection of new Buddhist images and build new palaces.200 Rather the policy seems to represent an attempt, belated though it was, to woo the upper bureaucracy. In so doing, of course, she risked the alienation of her 192   T CTC 207, p. 6570 and THY 68, pp. 1198–99. Some of the best-known officials assigned to the task were Wei Ssu-li, Hsüeh Ch’ien-kuang and Yang Tsai-ssu. In some cases, at least, the officials retained their current posts and became only acting prefects so that the experiment was perhaps not a wholehearted one. 193   T CTC 207, p. 6571. 194  Note 9 above. 195   T HY 66, p. 1157 and Yü-hai 59:38b. 196   T CTC 207, p. 6578. 197   T FYK 629:22a and THY 75, p. 1359. 198   T HY 60, p. 1053 and TFYK 512:6a. 199   T CTC 207, p. 6574. 200  See, for instance, on the palaces, THY 30, p. 557 and TCTC 207, p. 6569.

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

450

Guisso

allies below the fifth rank, and it is perhaps because of this policy shift that she found no support among them at the time of her deposition. In retrospect, we may conclude that any attempt at this stage to win over her ministers became impossible for the simple reason that they could not be reconciled to the role of the Changs at court. To the empress, of course, her favorites were a solace and a diversion, not to be taken seriously. The perspective of her ministers, however, was different. To them the brothers were dangerous. They had been responsible already for more than one miscarriage of justice and for the spread of corruption and scandalous rumors which diminished the prestige both of the empress and those whose duty it was to remonstrate with her. They were, moreover, dangerous in a more direct way, for their party of hangers-on was large, their association with the shrewd and ambitious Wu San-ssu was close, and increasingly they alone determined who should have access to the empress.201 In the event of her sudden death, it was not impossible that they could yet overturn the succession. It seemed imperative to certain ministers that action be taken quickly. In the end, it was the continuing attempt of the empress to conciliate her great ministers which brought about her downfall though short of returning to a policy of terror, it is difficult to know what other course she could have taken. In mid-704 she permitted the trial of three lesser members of the Chang clan for corruption and unwisely suggested that I-chih and Ch’ang-tsung be investigated at the same time. She was probably of the opinion that they were not directly involved in their kinsmen’s bribery and that by permitting the conviction of lesser members of the clan, she would demonstrate her impartiality and thereby disarm the critics of the favorites. The tribunal found these men guilty. It also found that Chang Ch’ang-tsung had illegally occupied someone else’s land and imposed upon him a light fine which the empress was happy to confirm. But when the tribunal went further and pointed out that the law demanded he be dismissed from all posts because of his brothers’ proven crimes, she balked. Securing the agreement of one of her tsai-hsiang that Ch’ang-tsung “had meritorious service to the state,” she immediately amnestied him, and before Wei An-shih and T’ang Hsiu-ching could complete their simultaneous investigation of I-chih she transferred them to the provinces.202 T’ang Hsiuching was probably speaking for all the other high officials when he secretly 201   T CTC 207, p. 6575 points out that for several months the empress saw only the Changs, receiving neither chief ministers nor her own sons. The many demotions at the restoration give some idea of the size of the Chang faction. 202   C TS 78:3337:2, TFYK 515:9a, and TCTC 207, pp. 6572–73. Fitzgerald, Empress Wu, p. 182 construes “meritorious service” as a possible double entendre.

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

The Chou Dynasty [ of Empress Wu ]

451

warned Chung-tsung before his departure that the Changs’ abuse of favor was certain to cause a revolution.203 This thought must by now have occurred both to the empress and to the Changs. Her response was an unfortunate mixture of conciliation and overconfidence. She appointed T’ang loyalists like Huan Yen-fan204 and, most notably, Chang Chien-chih to tsai-hsiang rank, and while she attempted to keep the customary balance,205 she thus provided her enemies with a focus of leadership. The Changs, undoubtedly aware that in their recent crisis Wu San-ssu and his faction had offered no support, responded by abandoning their policy of using Wu against Li and decided instead to act on their own. Whether or not the Changs made actual preparations to seize the throne is difficult to know, but in January of 705 wall posters accusing them of conspiracy began to appear. The empress refused to heed reports of these until on the nineteenth of the month, specific charges of treason were laid against Ch’ang-tsung.206 According to his accuser, he had consulted a fortuneteller who told him he had the physiognomy of an emperor and that if he erected a Buddhist temple at Ting-chou, the empire would turn to him in support. This, of course, was a capital charge and, taking no chances, the empress appointed a tribunal of three, two of whom could be expected to vote for acquittal.207 As was also expected, the third judge, Sung Ching,208 dissented and as the court rallied around him, she used her former tactic of appointing him as a provincial commissioner in order to eliminate his opposition. His courageous response was to point out that such a commission was improper for one of his rank and when the empress, returning to her customary respect for the law, permitted him to remain in the capital, Ch’ang-tsung’s case was lost. As Sung Ching continued to press his suit, he was joined by others, most notably Huan Yen-fan, and even by a genuine supporter of the empress, Ts’ui

203   T CTC 207, p. 6576. 204  On Huan, see CTS 91:3360:4 and HTS 120:3951:3. He was the scion of an official clan from Huai-nan obtaining his first post through the yin privilege. Much admired by Ti Jen-chieh, he rose to high positions in the Censorate where he was responsible for the restoration of the political rights of the victims of Lai Chün-ch’en and Chou Hsing. At Ch’ang-tsung’s treason trial, his was the strongest call for punishment, and his biography contains the fullest account of the coup. 205  On Chang, see Appendix B, no. 133. 206   T CTC 207, pp. 6574 ff. Fitzgerald’s account is found in Empress Wu, pp. 184–88. 207  The “tame” judges were Wei Ch’eng-ch’ing (Appendix B, no. 135) and Ts’ui Shen-ch’ing (CTS 77:3335:3 and HTS 109:3930:3). 208  On Sung, see CTS 96:3372:1 and HTS 124:3960:1.

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

452

Guisso

Hsüan-wei.209 Both attempted to convince her that it was in her own interest and, indeed, necessary for her survival to let the law take its course and, reluctantly, she sent Ch’ang-tsung to be tried once more by Sung Ching at the Censorate. While the trial was still in progress, however, she changed her mind and dispatched a messenger with an extraordinary pardon, freeing Ch’angtsung and thus making him immune from further prosecution. This act convinced the court officials that the ascendancy of the favorites was complete and, as the empress returned to seclusion in a remote section of the palace (ch’ang-sheng yüan) where the Changs alone were permitted to see her, they determined to act. According to most sources, Chang Chien-chih, who was now tsai-hsiang and chief imperial librarian, had determined to restore the T’ang during his provincial exile,210 and now he set about recruiting a party to destroy the Changs by force. Easily persuading Ts’ui Hsüan-wei and Huan Yen-fan to join him, he added to his conspiracy the censor and prominent opponent of both the Chang and Wu parties, Ching Hui,211 as well as Yuan Shu-chi who had formerly served in Chung-tsung’s household and was currently a member of Jui-tsung’s staff.212 His presence was probably necessary to associate the princes with the conspiracy, and it was he who gave Chung-tsung what little advance warning of the coup he received. All that remained now was to muster sufficient force, and Chang Chienchih turned naturally to the Yü-lin Guard, suborning the General of the Right, Li To-tso, who some twenty years earlier had aided the empress in effecting Chung-tsung’s deposition. On both occasions he seems to have been motivated 209  See Appendix B, no. 132. Ts’ui was in many ways typical of the northeastern ssu-hsing. After he passed the ming-ching, his mother, who was a member of the eminent Lu clan, instructed him on the integrity necessary to an official calling, and it was his incorruptibility while charged with the examination system which brought him to the personal notice of the empress. His biography shows that he was one of the empress’ most sincere supporters, never suggesting abdication. It was the complete lack of alternatives which finally brought him into the conspiracy. 210  See TCTC 207, p. 6579. 211   C TS 91:3361:3 and HTS 120:3952:2. Ching was one of the “new” officials prominent in the Chou, a member of a family with no tradition of office, a ming-ching graduate and a man whose reputation had been made in the provinces. His opposition to the Wu interests, carried into the next reign, precipitated the destruction of himself and his four co-conspirators by Wu San-ssu. 212   C TS 91:3362:3 and HTS 120:3951:3. Yuan was another of the “new” type of official. A northeasterner, his early career is nowhere documented, and we know only of his capital appointments. He may have had some military experience since during the actual coup, he was in charge of the Southern Guard whose task was to deal with unforeseen resistance.

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

The Chou Dynasty [ of Empress Wu ]

453

by concern for the welfare of the state.213 Chang also insinuated Huan Yen-fan, Ching Hui and Li Chan, the son of Li I-fu, into positions of command in the Yülin Guard, almost “tipping his hand” for the Changs immediately countered by having Wu Yu-i reappointed as general.214 More than anything else this move demonstrates the skill of the Wu interests. Since “several tens” of officials were later to be disgraced as their adherents,215 the Changs might well have chosen another of their supporters for the post instead of a member of the Wu clan which had already demonstrated its unreliability as an ally. The appointment smacks of manipulation by Wu San-ssu, who had already assured his own future through a clandestine relationship with Chung-tsung’s wife and the marriage of his nephew to the heir apparent’s favorite daughter.216 In any case, Wu Yu-i seems to have remained inactive while the coup unfolded, thereby ingratiating the Wus with the conspirators. Their preparations complete, the conspirators moved quickly. On the night of February 20, 705 they gathered a force of just over 500 men at the Hsüan-wu Gate, and sent for Chung-tsung. Although he had earlier approved the plan, Chung-tsung now lost his nerve. While agreeing that the Changs should be destroyed, he did not wish to alarm his sick mother.217 Li Chan, one of the messengers, persuaded him that only his presence could now halt the action, and thus he agreed to accompany the conspirators to the northern gate. As soon as they caught sight of him, however, the others smashed the gate and swept into the palace grounds. Reaching the western pavilion where the empress slept, they ran into the startled Changs in the courtyard and decapitated them on the spot. Awakened by the clamor, the empress confronted the plotters, but must immediately have recognized that the coup was a success. After addressing words of scorn to her son and certain others of the plotters,218 she returned to bed, her half century of power at an end. 213   T CTC 207, pp. 6578–79. Li found himself on the losing side in 707 when he joined the crown prince in an attempted coup against Empress Wei. 214   T CTC 207, p. 6579. 215   C TS 78:3337:2. These included Li Chiao, Ts’ui Shen-ch’ing, Ts’ui Jung, Sung Chih-wen, Shen Ch’üan-ch’i and several others who not only held high posts but were among the bestknown figures of their time. Their connection with the Changs shows how influential the faction had become. 216   C TS 193:3553:4. Wu Yen-hsiu had returned to the capital in 704 and, although he was not married to Princess An-lo until later, their relationship seems to have existed from this time. 217   T CTC 207, p. 6580. He was persuaded by Li Chan, the son of Li I-fu. 218   T CTC 207, pp. 6580–81. The empress had special scorn for Chung-tsung whom she thought perhaps to intimidate, for Li Chan whose father had been her earliest supporter, and for

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

454

Guisso

The next day Chung-tsung took charge of the state and sent commissioners through the provinces to announce the restoration. On the twenty-second of the month he became emperor once more, though seemingly without a formal ceremony of abdication by his mother. He granted her the title “Follower of Heaven, Great Sage, Emperor” (Tse-t’ien ta-sheng huang-ti)219 and transferred her to the Shang-yang Palace west of the city where for the next year she remained an ill but honored political spectator receiving bi-monthly homage from the court. There is ample testimony, as mentioned earlier, that she remained popular among the people and, it seems, even with some ministers who remained at court.220 It must have been with sardonic amusement that she saw Wu San-ssu so rapidly maneuver himself into a commanding position in government, reenact some of her old provisions, and render the five chief conspirators against her impotent.221 Nevertheless she was probably relieved to see the Li and Wu clans avoid strife and instead join in close alliance to impose their will on court and empire. On December 16, 705, as she prepared to die quietly in her bed at about the age of eighty, she could afford to magnanimously renounce the title of emperor which she had held so long and finally to forgive her earliest enemies, those who had stood in her way so long ago when by becoming Kao-tsung’s empress, she began the climb which would end on heights greater than any other woman in China would ever scale.

Ts’ui Hsüan-wei who had been her personal and independent choice for high office. Ts’ui, perhaps with some sadness, could attribute his presence only to “Your Majesty’s great virtue.” 219   C TS 6:3077:1 and TCTC 207, p. 6581. 220  See, for instance, TCTC 208, p. 6587, and 207, p. 6582 for T’ien Kuei-tao’s refusal to cooperate with the conspirators. 221   T CTC 208, p. 6596.

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

The Career of Yang Kuei-fei Howard S. Levy A Introduction Yang Kuei-fei died more than twelve hundred years ago, but her memory has been perpetuated by the composers of Chinese literature and drama. She achieved her fame not because she was a paragon of virtue but rather because she typified the epitome of sensuous beauty, a beauty so devastating in its appeal to an emperor that it shook the T’ang dynasty to its foundations. She was the consort of an imperial bohemian who used to select his sleeping partners in the harem through games of chance, and who took her as his own when she was the legal wife of his eighteenth son. An emperor properly schooled in Confucian mores would have been shocked if an imperial consort had dressed a court official in diapers and promenaded him about the harem in celebration of his birthday, but Yang Kuei-fei’s imperial husband joined in the amusement and rewarded the “lucky baby” involved. The love affair of Emperor Hsüantsung 玄宗 and Yang Kuei-fei lasted for more than a decade. It was not based solely on eroticism, but brought together two aesthetes who shared an enthusiastic devotion to the arts. They were not spectators but accomplished artists. Both played musical instruments in the court orchestra, and the skill of the emperor as a composer was matched by the consort’s mastery of singing and dancing. Yang Kuei-fei was the emperor’s constant companion in the waning years of his life, and it is probably this relationship based on mutual tastes which the emperor missed so keenly when the consort died, and which caused him to mourn for her unceasingly until the day of his death. Hsüan-tsung’s attachment to Yang Kuei-fei gave many of her close relatives a chance to share the imperial patronage which she enjoyed. Several of her cousins were presented with official sinecures and linked in matrimony to the royal family. She had three sisters who also were noted beauties. They married prominently and became influential figures at court. They too were imperial favorites who had free access to the harem and possibly had intimate relations there with the emperor. Yang Kuo-chung 楊國忠 was the most famous cousin of Yang Kuei-fei, because he became Chief Minister and wielded power until the rebellion of An Lu-shan. This branch of the Yang clan rose swiftly to Source: “The Career of Yang Kuei-fei,” T’oung Pao 45 (1957): 451–89.

© koninklijke brill nv, leiden, ���9 | doi:��.��63/9789004380158_015

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

456

Levy

prominence but hovered on the brink of disaster. Its continued survival depended solely on the ability of Yang Kuei-fei to remain the dominant figure in the imperial harem. This is why her relatives at Ch’ang-an 長安 were terrorstricken and in tears everytime she incurred the emperor’s wrath and was expelled. They were saved from disaster when she was recalled by Hsüan-tsung, but they must have realized that these might be only temporary reprieves. The Heir Apparent and presumed successor to the aging Hsüan-tsung hated the Yangs, and Chief Minister Yang Kuo-chung was probably correct in his belief that he would murder them all if and when he replaced his father on the throne. While there is no convincing proof that the Yang clique was planning a revolution, it is possible that they were thinking of seizing power before the death or abdication of the emperor led to their own annihilation. The objectives of this paper are to trace the career of Yang Kuei-fei and to consider her actions and the effect which they had on court developments. B Sources The main early sources for the life of Yang Kuei-fei are her biographies in the new and old standard histories of the T’ang dynasty. I have translated these biographies and appended them to the present study. They vary in several important respects because of the different source materials which provided the basis for their interpretations. The Chiu T’ang-shu 舊唐書 (“Old History of the T’ang”) relied primarily on the shih-lu 實錄 (“true records”) for its record of the first two centuries of T’ang rule.1 The so-called “true records” were composed by court historians after the death of an emperor, and they dealt with various aspects of his rule. As one might expect, there was at times a tendency for these officials to gloss over incidents which placed the ruler in an unfavorable light.2 The biography of Yang Kuei-fei in Chiu T’ang- shu was based on the “true records”, and it failed to reveal the way in which she became an imperial consort. Emperor Hsüan-tsung exposed himself to the criticism and censure 1  This is stated and substantiated in Nien-erh-shih ch’a-chi 廿二史劄記 (World Bookstore, Taipei, 1956), vol. 1, pp. 214–16. The history of the “true records” is discussed by Li Tsung-t’ung 李宗侗 in his work on Chinese historiography entitled, Chung-kuo shih-hsueh shih 中國史 學史 (Taipei, 1955), pp. 74–76. 2  Nien-erh-shih ch’a-chi vol. 1, p. 214, gives many examples in which the Chiu T’ang-shu omitted incidents which occurred during the reigns of Emperor Kao-tsung and Empress Wu 武后 and which placed these monarchs in a bad light. It did this because it relied solely on the ‘‘true records” in these instances and did not compare other sources.

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

The Career Of Yang Kuei-fei

457

of later historians because he took the Lady Yang into his palace when she was still the legal wife of his eighteenth son.3 The Chiu T’ang-shu biography omitted this fact and gave the misleading impression that the consort entered the palace of her own volition and was not married. It probably copied this interpretation from the Hsüan-tsung shih-lu 玄宗實錄 (“True Records of Hsüan-tsung”). The Hsin T’ang-shu biographers consulted more materials than their predecessors, and in their description of Yang Kuei-fei they set the record straight by showing she was first the wife of the emperor’s son and entered the palace at imperial request. The most complete biography of Yang Kuei-fei, entitled Yang T’ai-chen waichuan 楊太真外傳 (“Unofficial Biography of Yang T’ai-chen”), was compiled early in the Sung dynasty. Yang Kuei-fei was given the religious name T’ai-chen 太真 (“Grand Verity”) when she became a Taoist nun, prior to receiving her title as an imperial consort. The compiler of this biography was Yao Shih 樂史 (930–1007), a high official and a prolific writer on many subjects. He was criticized for having written “extensively but not to the point”.4 This type of criticism is applicable to the “Unofficial Record of Yang T’ai-chen”, Its authorship was attributed to him by the compilers of T’ang-tai ts’ung-shu 胃代 叢書 (“Collected Works of the T’ang Dynasty”) and others. It is interesting to note that this work is not listed in Yao Shih’s biography in the Sung-shih 宋史 (“History of the Sung”). However, he is credited in the bibliograpical section of the Sung History with authorship of a monograph in two chapters entitled, Yang Kuei-fei i-shih 楊貴妃遺事 (“Bequeathed Record of Yang Kuei-fei”). I believe that this may have been identical to Yang T’ai-chen wai-chuan, and that later editors may have revised the title and changed the chapter arrangement by reducing the number of chapters from two to one and by subdividing the chapter into “first” (shang 上) and “last” (hsia 下) sections. The modern T’ang historian Ch’en Yin-k’o 陳寅恪 criticized Yao Shih’s biography because it was sometimes chronologically inaccurate and factually unreliable.5 Yao Shih was interested in the occult and the supernatural,6 and at certain places in his account of Yang Kuei-fei he failed to distinguish between fact and fiction. For example, he presented as actual occurrence the materialization of a dragon 3  My monograph entitled, “The Selection of Yang Kuei-fei”, considers in detail the way in which she became imperial consort. It has been accepted for publication in Oriens. 4  Sung-shih 306. 18b. References to the standard histories are to the po-na 百衲 edition. Yao Shih’s biography is briefly described by E. D. Edwards in Chinese Prose Literature of the T’ang Period (London, 1938), vol. 2, pp. 120–2. 5  Sung-shih 203. 23b. 6  Sung-shih 306. 18b. He wrote many works on occult subjects.

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

458

Levy

lady and the coming forth of courtesans from painted screens.7 Despite its occasional lapse into fantasy, Yang T’ai-chen wai-chuan has merit as a historical document. It supplements the unadorned accounts in the standard histories and places emphasis on the aesthetic life at court. Its author describes the players in the imperial orchestra, the musical compositions which they played, and the famous dancers of the age.8 The author succeeds through these descriptions in recapturing some of the atmosphere of the T’ang court. Yao Shih assembled and edited this additional information by consulting material in several minor T’ang works which devoted a few lines to Yang Kuei-fei, Emperor Hsüan-tsung, or their associates.9 Towards the end of his account, he writes of the successful efforts of a Taoist priest to contact Yang Kuei-fei after her death, and records their conversations together.10 He may have adopted this version from the early ninth century descriptions of the spiritual regeneration of Yang Kuei-fei in the famous poem by Po Chü-i 白居易 entitled, Ch’ang-hen ko 長恨歌 (“Song of Prolonged Regret”) and in the complementary prose record by Po Chü-i’s associate Ch’en Hung 陳鴻 called Ch’ang-hen ko chuan 長恨歌傳. Po Chü-i did not limit himself to a factual portrayal of the life of Yang Kuei-fei, but achieved a greater dramatic effect by implying in his poem that she was a properly concealed maiden before she came to the notice of the emperor.11 It was not the poet’s intention to allow the sublimity of his theme of eternal love to be sullied by the sordid truth, since the true facts were supplied in the prose record of Ch’en Hung.12 Poem and prose added little new information to that 7  Yang T’ai-chen wai-chuan (abbreviated YTCWC) vol. 9, A. 4b, 6a. (T’ang-tai ts’ung- shu ed., printed by Mien-chang t’u-shu-chü 錦章圖書局.). All other references to works in this collection are to the same edition. 8  YTCWC vol. 9. A. 4b, 5b, 6a. 9  For example, the following works have material which may have been adopted by Yao Shih: Ming-huang tsa-lu 明皇雜錄 (9th c.) vol. 2. 37a has the story about the precocious child who improvised verse to the music of a flute, repeated in YTCWC A. 5b; K’ai T’ien ch’uan-hsin chi 開天傳信記 (late 9th c.) may have first quoted An Lu-shan as saying that he bowed to Yang Kuei-fei and later to the emperor because barbarians only knew that they had mothers (see YTCWC A. 4b); Chi-shih-chu 記事珠 (early 10th c.) vol. 4. 21b alleged that an old woman found a pair of Yang Kuei-fei’s stockings which she had worn on the day of her death and let people see them for 100 cash, a statement adopted in YTCWC B. 8b. There are indications that other earlier references were also copied in part. 10  YTCWC B. 8a–b. 11   Ch’ang-hen ko (Ssu-pu ts’ung-k’an 四部叢刊 ed.). 12  Ch’en Yin-k’o comments on the relationship between Po Chü-i’s poem and Ch’en Hung’s prose in Yüan Po shih chien-cheng kao, 元白詩箋證稿 (private reprint; Taipei, 1956), pp. 11–12.

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

The Career Of Yang Kuei-fei

459

contained in the original biographies, but the love of emperor and consort was effectively dramatized and projected beyond the physical realm to that of the spirit world beyond. There is a ninth century account of the rebellion of An Lu-shan called An Lu-shan shih-chi 安祿山事迹 (“Record of the Lu-shan Affair”), and it has scattered references to Yang Kuei-fei. These are important because many of them are based on the “true records”. As Professor Pulleyblank has shown, this source on An Lu-shan frequently copied the “true records” more fully than did the compilers of the T’ang standard histories.13 It describes in detail how An Lu-shan was swaddled in diapers on his birthday in 751 and promenaded about the harem, a description of his behavior which is lacking in the offcial biographies of Yang Kuei-fei.14 The Sung historian Ssu-ma Kuang 司馬光 largely follows the thematic structure of the official biographies of Yang Kuei-fei in his Tzu-chih t’ung-chien 資治通鑑 (“Comprehensive Mirror of Governing”) and presents additional material about her from T’ang sources, some of which are no longer extant. Since Ssu-ma Kuang is usually referred to as one who consulted the Chiu T’ang-shu in preference to the later revision,15 it is of interest to note that he accepted statements in Hsin T’ang-shu that Yang Kuei-fei was the wife of the emperor’s eighteenth son before she became imperial consort, and that she was summoned into the palace at the emperor’s command.16 His supplement to Tzu-chih t’ung-chien, distinguished by the addition of the words k’ao-i 考異 “(Examination of Discrepancies”) to the original title, discusses chronological discrepancies and comments on the question of whether An Lu-shan and Yang Kuei-fei were secret lovers.17 Ssu-ma Kuang lists many T’ang sources in the k’ao-i, some of which are no longer extant. The main sources for relevant anecdotes which still remain, apart from works which I have mentioned previously, are Ch’ang- shih yen-chih 常侍言旨 (late 8th c.), Tz’u Liu-shih chiu-wen 次劉氏舊聞 (9th c.) and K’ai-yüan T’ien-pao i-shih 開元天寶遺事 (early 10th c.).

13  See E. G. Pulleyblank, The Background of the Rebellion of An Lu-shan, (Oxford Univ. Press, 1955), pp. 4, 5, 169. 14   An Lu-shan shih-chi A. 14a (Hsüch-hai lei-pien 學海類篇 ed.). 15  See, for instance, Chih-chai shu-lu chieh-t’i 直齋書錄解題 (Kuang-ya ts’ung shu 廣雅叢 書 ed.) 4. 11a; Shih-ch’i-shih shang-ch’üeh 十七史商権 69. 2b. 16   Tzu-chih t’ung-chien (abbreviated TCTC) (Taiwan, 1955, photolithograph of Ming 明 ed.). T’ien-pao 3/12/ chia-wu. 17   Tzu-chih t’ung-chien k’ao-i (abbreviated TCTCKI) (Taiwan, 1955, photolithograph ed.) T’ien-pao 10/1.

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

460

Levy

In the centuries which followed the T’ang. dramatic works were produced which elaborated on the twin themes of the doomed love of Hsüan-tsung and Yang Kuei-fei and the perfidy of An Lu-shan. The most famous of the plays were composed in the Yüan and Ch’ing dynasties, respectively. The Yüan dramatist Po Jen-fu 白仁甫 entitled his production Wu-t’ung yü 梧桐雨 (“Rain by the Wut’ung trees”), taking his title from the line in Po Chü-i’s poem about “… when the leaves of the Wu-t’ung trees fell amidst the autumn rain”.18 Hung Sheng 洪昇 wrote a play called Ch’ang-sheng tien 長生殿 (“Palace of Longevity”) in the Ch’ing dynasty. He selected his title from Po Chü-i’s poetic reference to the lovers having exchanged secret vows at the Palace of Longevity.19 The tragic death of Yang Kuei-fei, which has inspired poets and playwrights for more than a millenium in China, still evokes a sympathetic response from contemporary writers. Two have expressed their belief that Yang Kuei-fei did not die at Mawei, as popularly supposed, basing their arguments almost entirely on a new analysis of the hidden meaning in Po Chü-i’s poem, without recourse to earlier materials. Both authors believe that she escaped but differ as to her eventual fate. One states that she became a. prostitute, the other that she married a barbarian.20 A number of western writers have also explored the dramatic possibilities of the Yang Kuei-fei theme. The French writer Georges Soulié de Morant composed perhaps the most literary of these fictionalized works, appropriately called, “La Passion de Yang Koei-fei”. Several similar versions in English appeared during the twenties.21 In 1946, Hope Danby wrote a book on the subject of Yang Kuei-fei which proved to be more historically accurate than those published by her western predecessors. Bernard Llewellyn has recently published a collection of essays which includes a chapter on Yang Kuei-fei. He makes no claim to scholarship, and limits his research on consort Yang to one of the

18   Ch’ang-hen ko. 19  Ibid., loc. cit. 20  See Yü P’ing-po 前平伯, “Ch’ang-hen ko chi Ch’ang-hen ko to chuan i” 長恨歌傳疑, cited in Fu Meng-chen hsien-sheng chi 傳孟真先生集, vol. 2 (Taipei, 1955), sect. 丁, pp. 42–52, for the prostitute thesis. The “barbarian” hypothesis is set forth by Ko Hsien-ning 葛賢寧 in the journal Wen-shih yüeh-k’an 文史月刊, (Taipei, April, 1957), pp. 62–5. 21  The western novels, by order of appearance, are: Georges Soulié de Morant: “La Passion de Yang Koei-fei”, in Mercure de France, Aug. 15, 1922, pp. 439–62; Sept. 15, 1922, 724–62; Nov. 15, 1922, pp. 116–79; Shu-Chiung (Mrs. Wu Lien-teh): Yang Kuei-fei (Most Famous Beauty of China), Commercial Press, Shanghai, 1923, 102 pp.; Maude Meagher, White Jade, Scholars Press, London, 1930, 160 pp.; Hope Danby, The Illustrious Emperor, Chicago, 1946, 281 pp.

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

The Career Of Yang Kuei-fei

461

English novels published in the twenties.22 The possibility of illicit relations between An Lu-shan and Yang Kuei-fei, dismissed by Professor Pulleyblank as “too grotesque”, becomes one of the main themes in Mr. Llewellyn’s essay.23 The T’ang authority Ch’en Yin-k’o has contributed a critical study of the Ch’ang-hen ko by Po Chü-i. He analyzes the content of the poem from the standpoint of historical accuracy, and reveals the places where poetic license and actual fact do not coincide. The merit of this study is that it traces the development of T’ang fiction, evaluates Ch’ing textual criticism and proposes original solutions.24 Professor Pulleyblank’s scholarly monograph on An Lu-shan has scattered references to Yang Kuei-fei. I have written three monographs dealing with her family background, sisters and selection as consort.25 This is my first paper to deal with the unfolding of her career in the middle T’ang. C

Yang Kuei-fei

1 Early Influences Yang Kuei-fei’s original given name was Yü-huan 玉環. She was born in Szechwan about 719, the daughter of an official census-taker who was on duty there. It is clear that her family was not native to Szechwan, but came from a district located in east-central Shensi called Hua-yin 華陰.26 The emperor was favorably disposed towards the Yang clan of Hua-yin, since she was the second woman née Yang from this district to secure distinction in the harem. The first Yang bore the child who succeeded Hsüan-tsung as emperor, and at her death was posthumously ennobled as empress.27 His high regard for her may have influenced the emperor to select another woman from her clan a decade after

22  Mr. Llewellyn’s essay on Yang Kuei-fei is the seventh of fourteen in his book entitled, China’s Courts and Concubines. (London, 1956). 23  See Pulleyblank, The Background of the Rebellion of An Lu-shan, p. 97; Llewellyn, China’s Courts and Concubines, pp. 97–99. 24  See Yuan Po shih chien-cheng kao. Ch’en Yin-k’o’s study was first issued as a separate monograph called, ‘‘Ch’ang-hen ko chien-cheng” and was published in Canton in 1950. 25  Of these three articles, “The Family Background of Yang Kuei-fei” has been published in Sinologica, V (1957), pp. 101–118.; “The Sisters of Yang Kuei-fei” has been accepted for publication in Phi Theta Annual, University of California, while “The Selection of Yang Kuei-fei” has been accepted for publication in Oriens. 26  YTCWC vol. 9, A. 4b; Yang Yü-huan’s great-great grandfather Yang Wang 楊汪 was a native of Hua-yin (HTS 76. 16a; see HTS 71b. 34b; Sui-shu 隋書 56. 11a.). 27  CTS 52. 1b, 2a.

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

462

Levy

her death. Hua-yin was also associated with supernatural blessings, since it was the discovery of treasure there in 739 which induced Hsüan-tsung to revise his reign style to T’ien pao 天寶 (“Heaven’s Treasure”).28 We discover little about the early years of Yang Yü-huan, since the official historians were interested primarily in her later relationships with Hsüan-tsung and her achievement of a dominant position in the harem. Her father died in Szechwan and she was raised by an uncle who also served in the official hierarchy. We do not know when her mother died, but it was sometime before Yang Yü-huan came to the attention of the emperor.29 The emperor’s eighteenth son went to Szechwan on a military tour of duty in 727,30 and while the references are mute on this subject, he probably came into contact with the family of Yang Yü-huan at that time. This meeting was to be expected, since her clan was prominent and traced its lineage back to the emperors of the Sui Dynasty.31 She married the prince nine years later.32 It is impossible to reconstruct her face and figure at the time she reached maturity, but careful reading of the early texts reveals that she was fully developed and sensuous in appearance An impression still exists among contemporary Chinese that she was stout, but I do not think that this was necessarily the case. We know that the typical Chinese beauty was slender to the point of fragility. Hsüan-tsung once visited the harem of the Heir Apparent and was dismayed to find that it was unkempt and deserted, with its musical instruments gathering dust. He ordered his chief eunuch Kao Li-shih 高力士 to secure five women from the populace and to present them to his son, and requested that he secure those who were slender and light-skinned.33 This was the norm for a harem beauty, and it was one which the lady Yang exceeded. She was an excellent dancer and musician, and must have been a woman endowed with a healthy physique. By modern standards, she might be considered pleasingly plump, but not stout. A story of doubtful authenticity tells of the harem rivalry which 28  TCTC. 29  Her mother was possibly née Li 李. (YTCWC loc. cit.). Both parents were enfeoffed posthumously after she became imperial consort. 30  CTS 107. 6b. 31  Yang Kuo-chung 楊國忠 claimed that he was a direct descendant of the Sui rulers. (Fengshih wen-chien chi) 封氏聞見記 4. 2a (Hsüeh-chin t’ao-yüan 學津討源 ed; this may have been copied into HTS 201. 7a). 32  They were married Feb. 10, 736 (T’ang ta chao-ling chi 唐大詔会集 40. 3a (Shih-yüan ts’ung-shu 適園叢書 ed.). 33   Tz’u Liu-shih chiu-wen vol. 2, 15a.

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

The Career Of Yang Kuei-fei

463

developed between Yang Kuei-fei and an earlier imperial favorite called Meifei 梅妃 (“The Plum-blossom Consort”). “Plum-Blossom” was ousted from the main harem because of her rival’s jealousy, we are told, and sent to an auxiliary harem in Lo-yang. There she pined away in solitude and jealousy, and once exclaimed, apropos of her abandonment, “If he is afraid to love me for fear of arousing that fat woman’s ire, is that not dismissing me?”34 We know that Yang Kuei-fei suffered from the summer’s heat, a discomfiture associated with overweight which may, however, have been aggravated by excessive alcoholic indulgence.35 One might say, then, that her figure has been criticized by the Chinese because it did not conform to the accepted standard of fragile femininity. 2 Harem Life Emperor Hsüan-tsung was the most prolific monarch to live during the T’ang dynasty. His wives bore him fifty-nine children during his lengthy reign, thirty sons and twenty-nine daughters.36 Wu Hui-fei 武惠妃 (“Favored Consort Wu”) was an early favorite and the most important woman in the harem prior to the entry of Yang Kuei-fei. She too came from a powerful clan, and was related to the Empress Wu 武后 and also to the Sui imperial house. She had seven children, three of whom died in infancy. It was her son named Mao 瑁 and titled Prince Shou who became the first mate of Yang Kuei-fei. Hsüan-tsung relieved the monotony of polygamy by introducing parlor games to his harem. His concubines were allowed to flip coins in order to compete for the pleasure of serving at his bedside. Towards the end of the K’ai-yüan era (713–42), the emperor is described as having devoted himself every spring to day and night revels. He had his concubines compete with one another to see who could pick the most beautiful flower. The emperor then let loose a butterfly which he held in his hand and watched to see the flower upon which it first alighted. The woman who owned the flower became the next to enjoy the imperial favors.37 This flitting from flower to flower was an imperial past-time only before the advent 34  Edwards, Chinese Prose Literature, vol. 2, p. 118. Miss Edwards translates the biography called Mei-fei chuan 梅``傳 (“Biography of the Plum-blossom Consort”). My researches disclose that this consort is not mentioned in any other T’ang reference, and I believe that the biography may be entirely fictional and may have been composed in order to discredit Yang Kuei-fei and her clan. 35  See K’ai-yüan T’ien-pao i-shih vol. 2, 26a. 36  HTS 82. 1a, 3b, 5a; 83. 9b, 10b. 37  See K’ai-yüan T’ien-pao i-shih vol. 2, 24b; 18b, 19a.

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

464

Levy

of Yang Kuei-fei. When she entered his life, Hsüan-tsung was then in his fifties and perhaps somewhat physically debilitated by his many years of indulgence and dissipation. When An Lu-shan first came under imperial patronage, we are told, he presented Hsüan-tsung with one hundred grains of an aphrodisiac called the “flower which aids the emotions” (chu-ch’ing hua 助情花). Each grain was about the size of a grain of rice and reddish in color. The emperor swallowed one grain before going to bed, with Yang Kuei-fei, and this helped him to maintain his sexual vigor.38 The women in the harem who were selected to attend the imperial bedside were relatively few in number and more fortunate than the many concubines of Hsüan-tsung who lived out their lives in cloistered solitude. It was then the custom for the women in the harem to make clothing for troops stationed along the frontiers. A soldier once discovered a note in the sleeve of his coat on which the maker described her present life as having already passed, and expressed the hope that she might be united with him in their next life. He showed the note to his commander and it then came to the attention of the emperor. Hsüan-tsung ordered the composer in the harem to show herself, and promised that she would not be punished. When she came forth and admitted her guilt, the emperor displayed compassion and allowed her to marry the soldier.39 The best depiction of harem confinement during Hsüan-tsung’s reign was given by Po Chü-i in a poem entitled, “A Person of Shang-yang” (Shangyang jen 上陽人). Shang-yang was the name of an auxiliary harem in Lo-yang. The poet wrote about a woman who had been selected for the harem when she was sixteen years old, only to remain alone and unseen by the emperor throughout her long life-span. According to the poet, her beauty was noticed by Yang Kuei-fei, who saw to it that she was sent away to an auxiliary harem where she could not possibly come to the notice of Hsüan-tsung.40 The fact that Yang Kuei-fei could achieve a dominant position in the harem and maintain a favored position there for more than a decade is adequate testimony to her abilities as a palace intriguer and strategist. 3 Monopoly of Imperial Favor We do not know when the emperor first noticed Yang Kuei-fei, but it was probably at the time she married his eighteenth son. According to the marriage rites

38  See K’ai-yüan T’ien-pao i-shih vol. 2, 20b. 39  See Ch’en Tung-yüan 陳東源, Chung-kuo fu-nü sheng-huo shih 中國婦女生活史 (Commercial Press, Shanghai, 1937), p. 94. 40   Po-shih ch’ang-ch’ing chi 白氏長慶集 (Ssu-pu ts’ung-k’an, ed.).

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

The Career Of Yang Kuei-fei

465

observed by princes during the K’ai-yüan era, the new bride paid her formal respects to the emperor on the day after the marriage was consummated.41 In the case of Yang Yü-huan, this would have been on Feb. 11, 736.42 Hsüan-tsung’s early favorite Wu Hui-fei died in 738, and he found no satisfactory substitute for her in the harem. He may have already coveted his son’s wife, but felt that he would have to act carefully and with prudence in order to appropriate her for his own without arousing strong opposition from the imperial family and influential court ministers. He waited until 741, and then acted to have Yang Yü-huan registered as a Taoist official and moved to a residence within the palace grounds. Her religious name was T’ai-chen 太真 (“Grand Verity”). The ordination rescript asserted that he had taken this action to satisfy her personal request.43 It was not until the summer of 745 that Hsüan-tsung made Yang T’ai-chen an official member of his harem. He rewarded his son Prince Shou with another wife to make up for the one that had been taken away from him, and a few weeks later entitled Tang T’ai-chen as kuei-fei 貴如 (“Precious Consort”). The consort’s relatives shared in her rise to prominence, and were given official sinecures and honorific titles. The beauty of Yang Kuei-fei was a family inheritance, shared by three of her sisters. These enfeoffed ladies married members of prominent clans and became influential at court. They had free access to the harem, and were so warmly patronized by Hsüan-tsung that they may have been, in effect, occasional paramours of the emperor. 4 Expulsion and Recall The path of love trodden by Hsüan-tsung and Yang Kuei-fei was not completely smooth, but was marred by at least two important lover’s quarrels. The first of these recorded incidents occurred about a year after the Lady Yang had received her elevation to first rank consort status. We are not told what she did to anger the emperor, but do know that Yang Kuei-fei was expelled from the harem and sent to the home of a cousin. The news of her expulsion caused panic among her relatives in the capital, and they gathered together and bewailed their fate.44 They fully realized that the dismissal of the consort meant that their lives were in extreme jeopardy. Hsüan-tsung may have acted in haste, but he soon demonstrated by his actions that he could not so easily rid himself 41  These marriage rites were recorded in K’ai-yüan li 開元禮} (“Rites of K’ai-yüan”). They are preserved in the T’ung-tien 通典 , chapters 106 to 140, and are described in HTS 11. 2a–b. 42  Yang Yü-huan married Prince Shou on Feb. 10, 736. (T’ang ta chao-ling chi 30. 3a.). 43   T’ang ta chao-ling chi 40. 8b. 44  YTCWC vol. 2. A. 5a.

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

466

Levy

of his harem favorite. He pined for her presence, and by noon of the day of her departure had still not eaten. He became increasingly irritable with the passage of time, and flogged his subordinates in a rage of frustration. His chief eunuch Kao Li-shih was the emperor’s confidant in matters of the heart, and he realized that the recall of Yang Kuei-fei was the only action which would restore Hsüan-tsung to reason. He first made a preliminary request, and asked that the consort be provided with sufficient provisions to ensure minimum comfort in her place of exile. Hsüan-tsung acquiesced and went so far as to divide up his own meal and send it along to her. The emperor’s ready and affirmative response convinced Kao that he wanted the “Precious Consort” back immediately, and her return was quickly arranged. While the cause of this lover’s quarrel is nowhere stated, I think that it arose over the issue of the emperor’s having relations with other women in the harem. Yang Kuei-fei probably did not mind if the emperor had an occasional affair with her sisters, but his frequenting the company of women of other clans posed a direct threat to the influence which she and her relatives were able to enjoy. If this hypothesis is correct, she was a woman of unusual courage who gambled with her life in order to strengthen her emotional hold on the emperor. She did this in the correct belief that Hsüan-tsung’s utter dependence on her would ensure her triumphant return to the harem. A Sung reference supports this thesis by stating that after she was recalled no other woman was able to share the imperial favors.45 Yang Kuei-fei was now unrivalled in the harem and more influential than ever. She became the emperor’s inseparable companion. It was their practice to spend a good part of the winter months in palaces erected at hot springs near Ch’ang-an. They went every year and sometimes stayed there for as long as three months at a time. This custom of frequenting the springs was ended only by the rebellion of An Lu-shan.46 Hsüan-tsung and Yang Kuei-fei were both skilled musicians, and played together in the royal orchestra. He beat the drums, while she played a guitar-like instrument of Central Asiatic origin known as a p’i-pa 琵琶. Members of the imperial family and her sisters tried to emulate her by learning to play this instrument. She was also skilled at striking musical stone chimes, and composed original tunes for them. It was said that even the most musically gifted prostitutes in Ch’ang-an could not match her dexterity. Other musicians in the orchestra played lutes, cymbals, clappers and flutes. On one occasion, we are

45  Ibid., loc. cit. 46  E. H. Schafer, “The Development of Bathing Customs in Ancient and Medieval China and the History of the Floriate Clear Palace”, (JAOS, vol. 76, Apr.–June 1956), pp. 74–5.

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

The Career Of Yang Kuei-fei

467

told, the orchestra played from day-break to noon and was in unusually high spirits. A sister of Yang Kuei-fei, called the Lady of Ch’in County 秦國夫人, was in the audience. When the music ended, Hsüan-tsung asked the Lady to reward him for having entertained her. She responded by giving him three million cash for the one performance.47 This figure may not have been exaggerated, since we know that the emperor lavished gifts upon the Yangs and gave them enormous annual stipends for cosmetic expenses alone. Local officials followed his example and deluged them with rare expensive gifts. They did this in the hope that they would then be quickly promoted within the bureaucracy. Before couples of the imperial family could get married, they had to bribe Yang Kuei-fei’s sisters in order to secure the emperor’s consent to the marriage. The consort’s cousin Yang Kuo-chung became wealthy through his financial manipulations, holding more than forty official posts.48 Yang Kuei-fei again incurred the imperial displeasure in the spring of 750 and was sent out of the harem in disgrace. A Sung author has stated that she was expelled for covertly playing a flute belonging to Prince Ning 寧王, an elder brother of the emperor.49 Another alleged that her elder sister the Lady of Kuo County 虢國 stole the flute, and that it was owned by a different brother.50 A purloined flute may have been involved, but it is improbable that it was a possession of either of Hsüan-tsung’s brothers, since both had died many years before.51 At any rate, the dismissal of Yang Kuei-fei again caused consternation among her relatives. They summoned one of their ministerial confidants named Chi Wen 吉温 to their side, and relied on him because he was on good terms with the eunuchs and understood the machinations of the harem. Chi Wen spoke to Hsüan-tsung and advanced the ingenious argument that it would be better for the emperor to kill the Lady Yang in the palace than to lose face by exhibiting her shame to outsiders. Hsüan-tsung, who may have already regretted the impulsiveness of his action, ordered a eunuch to bring Yang Kuei-fei back to the harem. She in turn wept before the eunuch, admitted her guilt and stated that she deserved the death penalty for insubordination. She said that she owed everything to the emperor but her own person, and to show the sincerity of her words cut off a lock of her hair to be transmitted to the throne. The cutting off of her hair may have implied a possible decision

47  HTS 206. 8a. 48  YTCWC A. 5b. The remarks about the orchestra are also repeated in Shin Tōshi Shuhen. 49  YTCWC A. 5a. 50   Ting-wei lei-pien 訂僞類編 2. 12a–b. (Chia-yeh-t’ang ts’ung-shu 嘉業堂叢書 ed.). 51  Ibid., op. cit.

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

468

Levy

to commit suicide, because Hsüan-tsung was alarmed by the sight of it and ordered her immediate return. He then became even more attached to her. His forgiving attitude later aroused the wrath of the Sung historian Ssu-ma Kuang, who asserted that ridding oneself of a woman was as trivial a matter as spitting on the ground. Ssu-ma Kuang cited the exemplary behavior of Emperor Wu 武帝 of the Han dynasty, because he put a consort to death on the grounds that she was overbearing and might dominate the Heir Apparent.52 5 Association with An Lu-shan The influence of Yang Kuei-fei at court can best be seen by the deferential way in which she was treated there by the distinguished northern military leader An Lu-shan. He once entered the palace to see the emperor and his consort, but bowed first to the imperial mistress. The annoyed monarch asked the reason for this bizarre behavior and was informed that, “Your subject is a barbarian. A barbarian puts his mother first and his father afterwards”. According to one anecdotal variant, he said, “A barbarian only knows that he has a mother, but does not know that he has a father”.53 Hsüan-tsung was amused by the reply. An Lu-shan may have acted in this way in order to ingratiate himself with the consort and secure her support. He further asked to be adopted by her as a foster son. This request was granted, and her sisters and cousins were ordered to enter into a family compact with him. There is a possibility, albeit grotesque, that the relationship of An Lu-shan and Yang Kuei-fei was more than that of a solicitous foster mother and a filial adopted son. Authenticated records reveal that they were intimate friends, if not lovers. When An Lu-shan celebrated his forty-eighth birthday at court on February 10, 751, both emperor and consort deluged him with gifts. These included clothing, jade utensils, wine and provisions. Yang Kuei-fei had a huge diaper made out of elegant brocaded material. An Lu-shan appeared three days later, and his “mother” had him washed, dressed in the diaper and promenaded about the Forbidden Palace grounds in a gaily-colored sedan carriage. Hsüan-tsung heard the commotion and uproarious laughter, and made inquiries as to what was taking place. He was told that Yang Kuei-fei was performing the ceremony of washing her “lucky baby” (lu-erh 祿兒) An Lu-shan on 52  TCTCKI TP 9/2. 53  Both T’ang standard histories have the first account, (see the biographies of An Lu-shan), while K’ai T’ien ch’uan-hsin chi 開天傳信記 vol. 2, 28b (T’ang- tai ts’ung-shu ed.) has the second version. This was apparently copied into YTCWC A. 4b and was reproduced verbatim in T’ai-p’ing kuang-chi 太平廣記 238. 22b (Sao-yeh-shan fang ts’ung-shu 掃葉 山房叢書 ed.).

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

The Career Of Yang Kuei-fei

469

the third day after his birth. Hsüan-tsung personally went to see the festivities and bestowed further gifts of gold and silver upon his consort’s “lucky baby”.54 The emperor’s favorable reaction to this rather bizarre behavior seems to indicate that Hsüan-tsung not only tolerated licentiousness but encouraged it. According to one anecdote, sometimes he and Yang Kuei-fei became intoxicated together in the harem. One of their chief amusements at such times was to simulate a large scale mock battle. The emperor would personally command an “army” of over one hundred eunuchs, while his “Precious Consort” opposed him with a similar number of courtesans. The girls then removed their collars, made battle-standards out of them, and fought with the eunuchs. The first side to admit defeat had to drink large quantities of wine from ram’s horn goblets as their penalty. The scene they created was described as a “battle-line of dissipation”.55 After the birthday episode, Hsüan-tsung allowed An Lu-shan free access to the imperial harem. There An Lu-shan shared meals with Yang Kuei-fei, while sometimes he spent the entire night in the harem. The usually discerning Ssuma Kuang states that while there was unspeakable gossip about their conduct, the emperor still did not suspect irregularities. To support his insinuation that they were secret lovers, he cites testimony from a Five Dynasty reference which states that An Lu-shan not only ate with Yang Kuei-fei, but was completely unrestricted in his behavior. He also quotes a no longer extant T’ang source which describes the situation in the harem. It states that all of Yang Kuei-fei’s sisters and the princes made fun of the “lucky baby” and were on the friendliest of terms with him. They banqueted together in mixed company at the harem, something which the author insists had never been tolerated up to that time. The waiting girls in the harem were amply bribed by the nobility so that they could spend the entire night there and have sexual relations with the imperial concubines.56 This description of debauchery is recorded in a near contemporary source, and it therefore gives rise to the possibility that An Lu-shan had opportunity amidst the general conditions of lewdness to have intimate relations with either Yang Küei-fei or other lesser ranking concubines of the emperor.

54   An Lu-shan shih-chi A. 14a. 55   K’ai-yüan T’ien-pao i-shih 開元天寶遺事 vol. 2. 28a. 56  TCTCKI 10/1. The sources cited are Wang Jen-yü’s 王仁裕 K’ai-yüan T’ien-pao i-shih and the non-extant monograph by Wen Yü 温番 called T’ien-pao luan-li hsing-Shu-chi 天寶 亂離幸蜀記.

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

470

Levy

6 Rebellion and Death An Lu-shan was aided considerably by a combination of imperial, consort and ministerial patronage, and came to command three circuits. He controlled the allegiance of about 200,000 troops, or almost 40 percent of the total for the empire. However, his sponsor, the Chief Minister Li Lin-fu 李林甫 died in 752. Li was replaced by Yang Kuei-fei’s cousin Yang Kuo-chung. The new Chief Minister was hostile to the more influential supporters of the deceased Li Linfu. He criticized An Lu-shan at court in his absence, and was joined in this criticism by the Heir Apparent whom An had once snubbed. They accused him of subversion and made his position extremely insecure. The Chief Minister who had befriended An Lu-shan was dead and the emperor who had protected him was now almost seventy years of age. The swift crumbling of his influence at court may have been an important factor in the decision of An Lu-shan to rebel against the throne.57 At the insistence of Yang Kuo-chung, the emperor ordered An to appear at court. Although Yang Kuo-chung warned Hsüan-tsung that this northern renegade would refuse an imperial summons, An came at once. He tearfully reaffirmed his unswerving loyalty, and expressed his fear that Yang Kuo-chung was determined to kill him. This loyalty oath reassured the sovereign, and he allowed An Lu-shan later to return to his northern bastion at Fan-yang 范陽, in the vicinity of modern Peiping.58 In the following year the emperor, who was now being warned on all sides about a coming insurrection, again summoned his favorite to Ch’ang-an. This time An Lu-shan refused the emperor’s invitation. Instead he flourished the banners of rebellion at Fan-yang on December 16, 755, and falsely declared to his troops that he was acting in support of an imperial edict which ordered him to exterminate the disloyal culprit Yang Kuo-chung.59 The news of rebel victories dismayed Hsüan-tsimg and many of his close advisers. The septuagenarian monarch’s first impulse was to abdicate in favor of the Heir Apparent, who might then carry out counter-revolutionary operations with greater effect. Chief Minister Yang Kuo-chung and his relatives were terrified when they learned of the imperial resolve to relinquish the throne. They justifiably feared that the Heir Apparent might murder them 57  Hori Toshikazu 屈敏一 outlines these and other factors behind the rebellion of An Lushan in his article, “Tōmatsu shohanran no seikaku” 唐末諸反乱の性格. Tōyō bunka 7 (1951). 58  CTS 200A. 2b. 59  Ibid., loc. cit.

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

The Career Of Yang Kuei-fei

471

all after he succeeded to the throne. They all wept together and urged Yang Kuei-fei to use her influence with Hsüan-tsung. She abjectly humbled herself before him, holding a clump of dirt in her mouth, and dissuaded him from abdicating. Rebel forces approached Ch’ang-an in the summer of 756, and inflicted a crushing defeat on those defending the passes. A council of the emperor’s top advisers hurriedly assembled and decided that Hsüan-tsung should depart from the capital at once and take refuge to the southwest. On 13 July 756, a small cortege consisting of members of the imperial family, Yang Kuei-fei and her relatives, eunuchs and a troop escort secretly abandoned the city ahead of the victorious rebel entry. The imperial refugees suffered from a shortage of food and arrived at Ma-wei 馬嵬 on the second day of their flight in a state of exhaustion and hunger. A group of about twenty Tibetans stopped Chief Minister Yang Kuo-chung there to complain that they were suffering from starvation. The troops may have feared that Yang Kuo-chung was going to seize power when he reached his strong base of support in Szechwan, and they took advantage of the immediate situation in order to accuse him of plotting with Tibetans to seize the throne. One of the soldiers fired an arrow at Yang which struck the saddle of his horse. The terrified minister dismounted and tried to flee on foot, but was caught and killed by his pursuers. They butchered his corpse, placed his head on a spear, and displayed it above the city gates. The soldiers also killed other relatives of the Yang clan. They surrounded the emperor and refused to withdraw. Hsüan-tsung ordered his eunuch confidant to find out why the troops were still refractory. Kao spoke with their commander, and learned that the soldiers demanded the death of Yang Kuei-fei. They apparently felt that An Lu-shan had been encouraged to rebel because consort Yang had caused Hsüan-tsung to neglect affairs of state. They further feared that she would someday avenge herself upon them for having murdered her cousins and sisters, if allowed to live. The eunuch reported to the monarch, who objected, “The Precious Consort has lived continuously in the recesses of the palace. How could she know of Kuochung’s rebellious plots?” Kao replied that, “The Precious Consort is truly innocent. However, although the troops have already killed Kuo-chung, she is still intimately close to you. How could you personally venture to be at ease ? … If the troops and officers are at ease, then you will be at ease!”60 Hsüan-tsung admitted the logic of Kao’s argument, went to Yang Kuei-fei and revealed her 60  TCTC Chih-te 至德 1/5 after i-wei; cf. ALSSC C. 8b, 9a.

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

472

Levy

fate. He granted her dying wish that she be strangled before a Buddhist chapel. Thirty-seven years of age, she walked to her execution grounds with Kao Li-shih. Her death put an end to the troop insurrection, and the soldiers after viewing her corpse swore eternal allegiance to the T’ang. The poignant way in which she died, a sacrificial pawn caught in a ruthless struggle for supremacy, has endeared her to countless generations of her countrymen and earned Yang Kuei-fei a permanent place among China’s immortals.

Appendix One

Translation of the Biography of Yang Kuei-fei in the “Old T’ang History” (Chiu T’angshu), po-na ed., ch. 51. f. 11b–13a. 11b. Hsüan-tsung’s 玄宗 Precious Consort (kuei-fei 貴妃) Yang’s 楊 great-greatgrandfather Ling-pen 令本 was a Prefect (tz’u-shih 剌史) of Chin 金 Prefecture. Her father Hsüan-yen 玄琰 was in the Finance Bureau (ssu-hu 司戶) of Shu 蜀 Prefecture. The consort was orphaned in her childhood and reared by her paternal uncle Hsüanchiao 玄璬, of the Bureau of Works (shih-ts’ao 士曹) of Ho-nan 河南 Metropolitan Prefecture. At the beginning of the K’ai-yüan 開元 era, because Favored Consort (hui-fei 惠如) Wu 武 was specially favored by the emperor, Empress Wang 王 was dismissed. The Favored Consort died in the twenty-fourth year (of K’ai-yüan). The emperor grieved and felt compassion for her over a long period of time. There was no one among the several thousand in the women’s quarters who could please him. Someone reported to the throne that the beauty of Hsüan-yen’s daughter was superlative, and that she should be allowed to see him. The consort was clothed in Taoist robes at this time, and was called T’ai-chen 太真. After she went to see him, Hsüan-tsung was greatly pleased. In less than a year, her treatment was on par with that of Favored Consort Wu. T’ai-chen’s carriage and sensuous figure were most captivating. She was skilled in singing and dancing and expert in music. Her talent and wit surpassed the norm. Her every charming glance fascinated the emperor, and her movements accorded with his feelings. She was called his wife in the palace, and her treatment there was really the same as that of an empress. She had three elder sisters, all talented and beautiful. Hsüan-tsung enfeoffed them simultaneously with titles as ladies of counties. The eldest was called his wife’s eldest sister and enfeoffed with the County of Han 韓; his wife’s third sister was enfeoffed with the County of Kuo 虢, and his wife’s eighth sister was enfeoffed with the County of Ch’in 秦. They all were favorites of the emperor and went in and out of the side-apartments of the palace. Their influence spread throughout the empire. The consort’s father Hsüan-yen was successively presented (with the

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

The Career Of Yang Kuei-fei

473

titles of honorary) Commander-in-Chief (t’ai-wei 太尉) and Duke of Ch’i County; her mother was enfeoffed as Lady of Liang 梁 County; and her paternal uncle Hsüan-kuei 玄珪 (was appointed) President of the Board of Imperial Banquets (kuang-lu ch’ing 光祿卿). As for her second cousins of the same surname, Hsien 銛 became President of the Board of Ceremonial Towards Foreigners (hung-lu ch’ing 鴻臚卿) and Ch’i 錡 became Censor (shih-yü shih 侍御史) and married Favored Consort Wu’s daughter the Princess of T’ai-hua 太華. Her treatment surpassed that of the other princesses because her mother had been loved (by the emperor). She was presented with a mansion connected to the forbidden palace. Everytime the five families (consisting of) the three Ladies of Han, Kuo and Ch’in, Hsien and Ch’i had requests, the prefectures and districts obeyed these with the majesty accorded imperial mandates. There were gifts sent from the four regions. Their gates were like market-places. In the seventh month of the fifth year (of T’ien-pao), the Precious Consort was sent back to Yang Hsien’s home because of a minor fault. The emperor had still not eaten by noon, thinking of her. Kao Li-shih ascertained the imperial intentions. He requested that he be sent to the home of the Precious Consort, and that he supply her with curtains, accessories and food provisions. Altogether he loaded more than one hundred cartloads. 12a. The emperor divided his imperial meal in order to send it (to her). He found fault with everything, burst into anger and beat and flogged those around him. Li-shih humbly memorialized the throne and requested that the Precious Consort he welcomed back to her (palace) residence. That evening, the gate of An-hsing 安興 Ward was opened and she entered the Inner Palace. The consort kneeled to the ground and acknowledged her guilt. The emperor joyfully consoled and comforted her. On the following day, (the Ladies of) Han and Kuo brought forth food and the emperor ordered a musical performance which lasted all day. Those around him were lavishly given presents Her favored treatment after this became increasingly marked. The yearly stipend of cosmetic money for the three ladies of Han, Kuo and Ch’in was one thousand strings of cash. Hsien was conferred (the title of) third ranking Imperial Pillar of the State (shang-chu-kuo 上柱國) and halberds were placed at (the gate of) his mansion. The mansions of the five families, namely those of the (three) sisters, the elder brother and the younger brother, were connected and comparable to the palace (in their splendor). They illumined the capital with their carts, horses, servants and chariots, and boasted to one another of their eminence. Every time they constructed a building, the expenses for it exceeded an estimated ten million cash. When they saw that someone else’s (mansion) was built on a grander scale than theirs, they destroyed their own (buildings) and rebuilt. Their architectural projects were carried on night and day. Hsüantsung made equal bestowals to them and (people of) the four corners gave similar

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

474

Levy

contributions to each of the five families. There was an unceasing line of eunuchs (bearing gifts for them). From the K’ai-yüan (713–42) era onwards, none of the powerful and eminent were comparable to the Yang clan in influence and splendor. Everytime Hsüan-tsung went on a journey, the Precious Consort invariably accompanied him. When she rode on horseback, Kao Li-shih held the reins and handed over the whip (to her). There were about seven hundred people in the palace who supplied weaving and embroidery for the Precious Consort’s home, while several hundred more did carving and molding for her. The prefects of Yang 楊 and I 益 Prefectures and of Ling-piao 嶺表 (the Kwangtung and Fukien regions) sought skilled workers to make rare objects and unusual clothing for them to offer to the Precious Consort. Those who made these congratulatory offerings were elevated to distinguished positions above the rank and file. Hsüan-tsung visited the Palace of Splendor and Purity (hua-ch’ing kung 華淸宮) every year in the tenth month and was followed in attendance by the five families, consisting of Kuo-chung and the sisters. Each family formed one unit and each wore one color of clothing. When the five families joined ranks, they glittered like the brilliance of a hundred flowers. When they set out, the adornments they lost and the slippers they dropped made tinkling sounds, whilst (their hair ornaments of) pearls and kingfisherblue sparkled and filled the road with their fragrance. Kuo-chung had illicit relations with (the Lady of) Kuo County and did not mind being criticized as the “virile fox”. Everytime they entered the court, they rode side by side on the road and did not bother to pull down their curtains. Every third audience there was a celebration, when at the drumbeats of the fifth (nightwatch) they waited for the clepsydra (to indicate dawn and the beginning of the audience); the streets overflowed with their elegance and ornamentation, while candles lit up (the scene) like daylight. The princes of the ten residences and their numerous descendants in arranging marriages all relied on (the Ladies of) Han and Kuo for introductions. 12b. They first had to bribe them with a thousand strings of cash and then petition the throne. Everyone of these petitions was approved by the emperor. In the ninth year of T’ien-pao (750), the Precious Consort again disobeyed the emperor’s wish and was sent back to an outside mansion. At that time Chi Wen was on good terms with the eunuchs. He entered and memorialized the throne, saying: “Your wife’s knowledge is not extensive, and she has disobeyed the imperial sentiments. However, the Precious Consort has been a recipient of your grace for a long time. Why do you begrudge a one-mat space in the palace from becoming her execution ground? How can you bear to exhibit her shame to the outside?” The emperor then ordered the eunuch Chang T’ao-kuang 張韜光 to present her with the imperial meal. The consort approached T’ao-kuang and in tears memorialized the emperor through him, saying: “This concubine has displeased the imperial countenance and deserves death tenthousandfold for her crime. Apart from her clothing, everything has been bestowed by the grace of His Majesty. There is nothing which she retains as her own. However, her

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

The Career Of Yang Kuei-fei

475

hair and her skin came from her parents”. She then drew out a knife, cut off one lock of hair and gave it to him to be offered up to the emperor. When Hsüan-tsung saw it, he was startled and felt compassion. He then sent Li-shih to summon her back. Kuo-chung held important ministerial rank and simultaneously commanded the Legateship of Chien-nan 劍南. His influence was such that he became more and more overbearing. On the full moon of the first month and tenth year (Feb. 15, 751), the five households of the Yang family went out looking for pleasure. They raced toward the gates of the west market with the Princess of Kuang-p’ing 廣平 and her attendants. A slave of the Yang clan flourished his whip and it touched the clothing of the princess. She fell from her horse and when (her husband) the imperial son-in-law (fu-ma 駙馬) Ch’eng Ch’ang-i 程昌裔 helped her up, he too was struck several times. The princess memorialized the emperor about this in tears. His majesty ordered that the slave of the Yang clan be killed and that Ch’ang-i be suspended from office. Kuo-chung’s two sons Ch’o 昢 and Hsüan 暄 and the consort’s younger brother Chien 鑑 all married princesses. This one Yang clan married two princesses and two countesses. A private ancestral temple was established for the Precious Consort’s father and grandfather. Hsüan-tsung personally composed the inscriptions for their temple monuments. Hsüan-kuei 玄挂 was successively raised (in rank) until he came to be President of the Board of War (ping-pu shang-shu 兵部尙書). During the T’ien-pao era (742–56), the Fan-yang 范陽 Legate An Lu-shan 安祿山 greatly established his merit along the frontiers. His majesty deeply favored him. When Lu-shan came to court, the emperor ordered the Precious Consort and her sisters to form a family compact with him. Lu-shan served the Precious Consort as he would his mother. In every banquet, she showered him with gifts. When Lu-shan’s rebellion was exposed, his summons to war enumerated the crimes of Kuo-chung. His Ho-pei bandits staged their uprising. Hsüan-tsung appointed the imperial Heir Apparent as Commander-in-Chief of the Imperial Troops and Cavalry (t’ien-hsia ping-ma yuanshuai 天下兵馬元帥) and directed him to supervise and settle military and national operations. Kuo-chung was terrified, and the Yangs assembled together and wept. The Precious Consort held a clump of dirt in her mouth and pleaded before the emperor. 13a. He therefore did not abdicate. When T’ung-kuan lost its defenses, (the Yangs) followed the imperial procession to Ma-wei 馬嵬. Ch’en Hsüan-li 陳玄禮, the Grand General of the Forbidden (Palace) Armies (chin-chün ta-chiang 禁軍大將), secretly advised the Heir Apparent to execute Kuo-chung and his sons. And still the four armies refused to disperse. Hsüan-tsung sent Li-shih to make a formal inquiry and (the military) replied: “The cause of (this) banditry still exists”. They referred (in this way) to the Precious Consort. Li-shih again memorialized and the emperor could not help himself. He gave orders to the consort and she was consequently strangled to death at a Buddhist shrine. She was thirty-eight years old at the time. She was buried alongside the road west of the courier station.

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

476

Levy

When the Exalted Emperor (shang-huang 上皇, Hsüan-tsung’s title after his abdication) returned from Shu 蜀, he ordered his eunuchs to offer up sacrifices and libations to her. He ordered by imperial mandate that she be reburied. Li Kuei 李癸, the VicePresident of the Board of Rites (li-pu shih-lang 禮部侍郎), said, “The officers and soldiers of the Dragon Militant (Army) (lung-wu 龍武) executed Kuo-chung because he betrayed the nation and incited rebellion. If we now rebury the late consort, I fear that the officers and soldiers will become suspicious and apprehensive”. It was not proper to carry out the burial rites and therefore the emperor (officially) desisted. The Exalted Emperor secretly ordered his eunuchs to change her burial to another place. When she was originally buried, she had been wrapped in purple matting. Her flesh and skin had already decayed but her perfume sachet still remained. A eunuch offered it up to the Exalted Emperor and he looked at it, grieving and lamenting. He then ordered that her likeness be drawn in a side-room. He looked at it day and night. When Kuo-chung was executed at Ma-wei, the Lady of Kuo County heard of the trouble and fled on horseback to Ch’en-ts’ang 陳倉 (District). The District Magistrate (hsien-ling 縣令) Hsüeh Ching-hsien 薛景仙 led men and officials in pursuit of her. She fled into a bamboo forest and there first killed her son P’ei Hui 裴薇 and one daughter. P’ei Jou 裴柔, the wife of Kuo-chung, said to her: “Please do me the favor of extinguishing my life.” She then pierced her to death and cut her own throat but did not die. The district official carried her to the jail where she was incarcerated. She still was able to say to the official: “Are you a loyalist or a rebel?” “I am both”, he replied. She died when her blood congealed and reached her throat, and was then buried beyond the suburbs. The Lady of Han County’s husband was Ts’ui Hsiang 崔垧, Assistant-Director of the Imperial Library (mi-shu shao-chien 秘書少監); their daughter had become a consort of Emperor Tai-tsung 代宗. (The Lady of) Kuo County’s son P’ei Hui had married the Princess of Yen-an 延安, a daughter of Emperor Tai-tsung. (The Lady’s) daughter had married a son of (Hsüan-tsung’s elder brother) Jang-ti 讓帝. The Lady of Ch’in County’s husband Liu Ch’eng 柳澄 had died previously; their son Chün 鈞 married the Princess (hsien-chu 縣主) of Ch’ang-ch’ing 長淸. Chang’s younger brother T’an 潭 married the Princess of Ho-cheng 和政, daughter of Emperor Su-tsung.



Appendix Two

Translation of the Biography of Yang Kuei-fei in the “New T’ang History” (Hsin T’angshu), po-na ed., ch. 76. f. 16a–18a. The Precious Consort (kuei-fei 貴妃) of Hsüan-tsung 玄宗 was a fourth generation descendant of (Yang) Wang, Governor (t’ung-shou 通守) of Liang Commandery 梁郡 in

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

The Career Of Yang Kuei-fei

477

the Sui Dynasty. He moved his residence to P’u Prefecture 蒲州 and (his descendants) became natives of Yung-lo 永樂. (Yang Kuei- fei) was orphaned in her childhood and raised by her paternal uncle. She first became the consort of the Prince of Shou 壽王. Favored Consort (hui-fei 惠妃) Wu 武 died in the twenty- fourth year of K’ai-yüan 開元 (ca. 736), and there was no one (else) in the rear palace who could satisfy the emperor. Someone said that consort (Yang’s) carriage was naturally elegant, and that it would be proper for her to serve in the imperial palace. She was consequently ordered to enter the forbidden palace (grounds). (The Emperor) marvelled at her and (pretended that) it was the personal wish of the consort (to enter his harem). She was appointed as a Taoist nun and entitled T’ai-chen 太真. Furthermore, while T’ai-chen secured the imperial favor, (the emperor) betrothed the daughter of Wei Chao-hsün 韋照訓 for the Prince of Shou. (Yang T’ai-chen) was skilled in singing and dancing and quite proficient in music. She was, moreover, wise, contemplative, quick-witted and bright, and was instantly aware of the emperor’s desires. The emperor was greatly delighted, and consequently monopolized her (company) and gave banquets (for her alone). She was called his wife in the palace, and the propriety and treatment (she received) was the same as that of an empress. At the beginning of the T’ien-pao 天寶 era (ca. 742), she was formally appointed Precious Consort (kuei-fei). Her father Hsüan-yen 玄琰 was posthumously presented with the titles of (Honorary) Commander-in-Chief (t’ai-wei 太尉) and Duke of Ch’i County (Ch’i-kuo kung 齊園公), and her paternal uncle Hsüan-kuei 玄珪 was elevated to be President of the Board of Imperial Banquets (kuang-lu ch’ing 光祿卿). As for her elder cousins, Hsien 銛 became President of the Board of Ceremonial Towards Foreigners (hung-lu ch’ing 鴻臚卿), while Ch’i 錡 became Censor (shih-yü-shih 侍御史) and married the Princess of T’ai-hua 太華. Borne by the Favored Consort (née Wu), this princess was the most favored by the Emperor. And Chao 剑 also became more and more distinguished. Chao was Kuo-chung 國忠. Three of her sisters were all beautiful and talented. The emperor called them his wife’s sisters, and enfeoffed them as Ladies of the three Counties of Han 韓, Kuo 虢, and Ch’in 秦. They went (freely) in and out of the palace compartments and shook the empire with the resounding echoes of their grace and favor. Everytime the ladies entered according to their rank, even those with the eminence of princesses yielded to them; none dared to be seated (before they took their places). Those in the Boards and Courts and in the prefectures and districts respectfully obeyed the requests (of the Yang sisters) and flurried about them, exceeding the services stipulated by imperial mandate. The four regions offered wealth in order to get closer to them. Their gates were like market-places. Because the two Princesses of Chien-p’ing 建平 and Hsinch’eng 信成 conflicted with the consort’s family, (the emperor) took back the gifts and

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

478

Levy

enfeoffements (which they had received) and the imperial son-in-law ( fu-ma 酣馬) Tu-ku Ming 獨孤明 lost his official position. On another day, the consort was sent back to Hsien’s 銛 mansion because of an infraction. By noon, the emperor still had not eaten. He flogged and vented his anger on those around him. Wishing to ascertain the imperial intentions, Kao Li-shih then said that (the Court of) the Supervision of Agriculture (ssu-neng 司農) in charge of palace provisions had sent to the consort’s house over one hundred cartloads of wine and food. The emperor immediately divided the imperial meal and sent it to her. Li-shih understood the imperial sentiments, and that night he requested that the consort be summoned. When they unlocked the gate of the An-hsing 安興 Ward, she quickly entered on her mount. Upon seeing the emperor, the consort fell to the ground and confessed her faults. The emperor forgave her and soothed, comforted, and was very cordial towards her. When his wife’s sisters presented music and entertainment at an imperial banquet on the following day, on the spur of the moment the emperor bestowed an incalculable number of gifts on those around him. From this time on, (the Yang family) increasingly received the imperial favor. He presented his wife’s sisters with a million cash yearly for their cosmetic expenses. Hsien was made Exalted Pillar of the State (shang chu-kuo 上柱國) and halberds were placed at his gate. The mansions of the five families of his wife’s sisters, Ch’i, and Kuo-chung were connected and comparable in lustre to the imperial palace. The expenses for one building were sometimes ten million strings of cash. If they saw a mansion which was superior (to theirs), they immediately destroyed and rebuilt, boasting to one another of their preeminence and extravagance. Their construction projects were carried on unceasingly. Rare gems and offerings which the emperor obtained were allotted and presented to them, and the procession of envoys along the road was unbroken. The five families were treated identically. Everytime the consort followed the emperor on his journeys and was mounted, Li-shih held the reins and whip. About a thousand people were needed to fulfill the official duties of embroidery and ornamentation and to carve gold and jade. There were rare clothes (for her) and mysterious curios which were tranformed as if they were supernatural. The four corners competed in offering her the rare and the precious. These startled one’s eyes and ears. The Ling-nan 嶺南 Legate (chieh-tu shih 節度使) Chang Chiu-chang 張九章 and the Chief Administrator (chang-shih 長史), of Kuang-ling 廣陵 commandery, Wang Chi 王冀, consequently advanced the furthest because of what they had offered up (to the Yangs). Chiu-chang received the (honorary) rank of (Grandee of the Board of Imperial Banquets with) a silver (seal) and blue (seal cords yin-ch’ing chieh 銀青階), and Chi was selected to be Vice-President of the Board of Revenue (hu-pu shih-lang 戶部侍郎). This became the universal fashion. The consort was fond of lichees and always wanted them sent fresh to her. The emperor therefore established mounted couriers to carry

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

The Career Of Yang Kuei-fei

479

them back. They traversed several thousand li 里 but the taste (of the lichees) had not yet changed by the time they reached the capital. The consort again committed an infraction in the ninth year of T’ien-pao (740) and was returned to a mansion beyond (the royal palace). Kuo-chung schemed with Chu Wen 朱温, who therefore saw the emperor and said: “This woman deserves death for her transgressions and obstinacy. Why should you be concerned that the width of one mat in the palace forms an area for quartering her? Why exhibit her disgrace to the outside?” The emperor was moved. Halting his meal, he ordered the eunuch Chang T’ao-kuang 張韜光 to present it to her. Through the intermediary of T’ao-kuang the consort apologised to the Emperor, saying: “I should be executed ten-thousand fold for my crimes. However, apart from my skin and hair, everything of mine has been bestowed by the emperor. If I die now, there will be no way for me to express my gratitude to him.” She drew out a knife, cut off one strand of her hair and presented it to him, saying: “Let these be my parting words.” The emperor saw this and, startled and pained, hurriedly summoned her to return. Her treatment was as before. The emperor then further visited the mansions of (the Lady of) Ch’in County and Kuo-chung, and presented both families with one hundred million cash. Since Kuo-chung was already commanding Chien-nan 劍南 in absentia, the emperor, when visiting the Palace of Splendor and Purity 華淸宮 during the tenth month in every (year), was followed by the carts and couriers of the five households. The families were differentiated into units and each unit wore one color. When the five units happend to join together, they were like the streams and valleys with their ten thousand flowers, forming a glorious brocade. Kuo-chung led them with a Chien-nan flag token. Their scattered hairpins and fallen slippers made tinkling sounds, whilst pearls lay in profusion along the road. Their perfumes were perceived for several tens of li. The consort’s household had a horse-race with the slaves of the Princess of Kuang-ning 廣寧 on the night of the full moon in the first month of the tenth year (Feb., 741), riding towards the walled gates of the city. Due to the wielding of the whips and the shouting, the princess fell off her horse, but the slave got away. When she had visited the emperor in tears, he proclaimed that the slave of the Yang clan be killed and that (her husband) the imperial son-in-law Ch’eng Ch’ang-i 程昌裔 be reduced in rank. When Kuo-chung was in charge of the administration, his son Ch’o 础 married the Princess of Wan-ch’un 萬春, Hsüan 暄 married the Princess of Yen-ho 延和, while his younger brother Chien 鑑 married the Princess of Ch’eng-jung 承榮. It was imperially proclaimed, moreover, that a family temple be established for Hsüan-yen 玄琰. The emperor personally wrote his tomb inscription. Since Hsien and (the Lady of) Ch’in County had died earlier, (the Ladies of) Han and Kuo and Kuo-chung were eminent for the longest period of time.

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

480

Levy

However, (the Lady of) Kuo County committed incest with Kuo-chung. This was extremely well known among the people, but they were not ashamed of it. Everytime they entered (the palace) to visit (the emperor), they rode side by side on the road, accompanied by one hundred cavalry of eunuch attendants and by ladies-in-waiting. The sky was illuminated by innumerable candles, giving the appearance of daylight. The entire area was filled with ornamentation and adornment. They did not put up curtains and screens. People of the time called him the “virile fox” (hsiung-hu 雄狐). Imperial descendants who were to be betrothed were always sure to make their requests through (the Ladies of) Han and Kuo first, and all invariably had to pay as much as several hundreds and thousands of cash as a gratuity. Previously, when An Lu-shan 安祿山 had acquired merit on the frontiers, the emperor had favored him and ordered his wife’s sisters to agree to form a fraternal pact with him. Lu-shan served the consort as he would his own mother. When he came to court, he always gave banquets and entertainments in order to please her. When Lu-shan rebelled, he took as his slogan the execution of Kuo-chung. Moreover, he indicated in his words the crimes of the consort and her sisters. The emperor wished to appoint the Heir Apparent as Commander-in-Chief and then abdicate the throne. All of the Yangs were greatly frightened and wept at court. Kuo- chung entered and informed the consort. She held a clump of dirt in the mouth and requested death. The emperor’s intentions were obstructed and he therefore desisted. When the emperor reached Ma-wei 馬嵬 on his tour westwards, Ch’en Hsüan-li 陳玄禮 and others executed Kuo-chung for the sake of the empire. Kuo-chung had already died, but the troops would not dissolve. The emperor despatched Li-shih to inquire into the cause, and he said: “The root of the disaster still exists.” The emperor could not but part with the consort and take leave of her. She was strangled at the foot of a shrine on the roadside. They wrapped up the corpse with purple cushions and buried it by the side of the road. She was thirty-eight years old. When the emperor arrived from Shu 蜀 and passed her place (of burial), he ordered sacrifices to her. He further proclaimed that her burial place be changed. Li Kuei 李揆, the Vice-President of the Board of Rites (li-pu shih-lang 禮部侍郎), said: “The Dragon-Militant soldiers and officers killed Kuo-chung on behalf of the empire because he turned against Your Majesty and accelerated rebellion. If we now give the consort an imperial burial, I fear that they will become uneasy and naturally suspicious.” The emperor therefore desisted, but secretly despatched eunuchs to prepare inner and outer coffins for her imperial burial elsewhere. When they opened the grave, her old sachet was still there. A eunuch then offered it up to the emperor. He grieved and wept as he looked at it and ordered an artist to draw a picture of the consort in a separate temple. He went there day and night and invariably was distressed and sobbed because of her. During the difficulty at Ma-wei, (the Lady of) Kuo County and Kuo-chung’s wife P’ei Jou 裴柔 and others fled to Ch’en-t’sang 陳倉. The District Magistrate (hsien-ling

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

The Career Of Yang Kuei-fei

481

縣令) led officials in pursuit. Thinking they were bandits, (the women) abandoned

their horses and fled to the forest. (The Lady of) Kuo County first killed her two sons. (P’ei) Jou said: “I beg for my death.” (The Lady) immediately assembled her daughters and stabbed them to death. She cut her own throat but not fatally. An official carried her to the jail, where she asked him: “Are you a loyalist or are you a rebel?” “I am both”, he replied. She then died and was buried outside of the eastern suburbs of Ch’en-ts’ang.

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

The Flight from the Capital and the Death of Precious Consort Yang Paul W. Kroll Every student of the T’ang—indeed almost every student of any period of traditional China—knows in general outline the unfortunate story of Li Lungchi’s 李隆基 (pht. Hsüan Tsung 玄宗, r. 712–56) furtive and hasty flight from the capital city of Ch’ang-an 長安 in July of 756, and his forced execution en route of his beloved lady, Precious Consort Yang (Yang kuei-fei 楊貴妃. The full tale of these events, narrated in moving, and sometimes pathetic, detail, is recounted in Ssu-ma Kuang’s 司馬光 (1019–1086) Tzu-chih t’ung-chien 資治通鑑. Surprisingly, however, a complete English translation of that memorable account has never been published. Most of our histories summarize it in a sentence or two; and even the most specialized and prestigious Englishlanguage portrayal of that era spares only a paragraph to review the incidents.1 Unsettling as it is to contemplate, it may well be that the majority of students derive their most detailed impression of these events from Po Chü-i’s 白居易 (772–846) romantic poem, “The Song of Everlasting Remorse” (Ch’ ang hen ko 長恢歌), in the pleasant but inadequate translation of Witter Bynner.2 The rendering below aims to supply this deficiency. I hope it may be of some use both to scholars in the field and to students who have not the ability to read the original text contained in the Tzu-chih t’ung-chien. In addition to the actual narrative, it also includes the k’ao-i 考異 sections, or “Examination of Differences,”3 of Ssu-ma Kuang, as well as the extensive and often extremely informative notes of the thirteenth-century commentator Hu San-hsing 胡三省 (1230–1302).4 Some readers may, however, wish to content themselves Source: “The Flight from the Capital and the Death of Precious Consort Yang,” T’ang Studies 3 (1985): 25–53. 1  The Cambridge History of China, Vol. 3: Sui and T’ang China, Part I, ed. D. Twitchett (Cambridge, 1979), 460. 2  As reprinted in the widely used textbook Anthology of Chinese Literature from Early Times to the Fourteenth Century, ed. C. Birch (New York, 1965), 266–69. 3  Headed by the rubric “varia” in the translation. 4  Hu’s notes, like the k’ao-i sections, are indented for easy identification in this version and are placed exactly where they occur in the Chinese text, even when this means breaking in midsentence a line of the main text.

© koninklijke brill nv, leiden, ���9 | doi:��.��63/9789004380158_016

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

flight from the capital & death of precious consort yang

483

by perusing only, or first of all, the main text by itself, so as to take in uninterruptedly the compelling course of the narrative. The account of the events of 10–15 July 756 is one of the finest pieces of writing, and certainly one of the most exciting segments, in the entire Tzu-chih t’ung-chien. It is, indeed, as much “literature” as it is “history,” if one employs the narrowly defined academic terms of the present day. The major figure of the narrative is of course Li Lung-chi himself, referred to in the text as “His Highness” (shang 上). The depiction of the monarch is drawn in careful and affecting fashion. He appears, almost, a sinitic Richard ii—a master of sweet and dignified language, whose human weaknesses at last undermine his divinely sanctioned authority. Through the pages translated below he is seen striving to maintain, even in the worst circumstances, the formal and demanding role allotted him. But the performance, for all its undeniable grandeur, is flawed; and this is betrayed to the reader by a few brief but precisely placed and tellingly worded comments of the narrator. The first rent in Li Lung-chi’s accustomed cloak of self-confidence is exposed when the beacon fires of the T’ung Barrier are not lit on the evening of 10 July, and we are told “Now for the first time His Highness was frightened”—a chilling sentence. Later we see him proclaiming to the populace of Ch’ang-an his intention to join the battle against the rebels himself, but “Of those who heard this, no one at all believed it.” During the first day of his flight from the capital, the accepted order of things begins to disintegrate swiftly: officials entrusted personally with important tasks run off, the members of the imperial family in the accompanying retinue are reduced to eating with their hands coarse and scanty meals, and the thearch himself listens patiently and penitentially to an accusatory harangue addressed to his face. The mantle of rule is already slipping from his shoulders. On the fateful second day of the journey it falls completely. At the Ma-wei post-station the soldiers in the entourage, after murdering the despised chief minister Yang Kuo-chung 楊國忠 (cousin of Yang kuei-fei), refuse to regroup and move on, unless the monarch orders the execution of the Precious Consort herself. Upon hearing this demand, he turns and goes back within the gate—and there, we are told, for a long time “he stood, leaning on his staff, his head bowed.” In this plain but supremely moving sentence the once majestic emperor is revealed a weary and beaten old man— no longer sovereign but now wholly human, and wholly powerless to prevent the death of his lady. It is an unforgettable moment and a wonderful narrative stroke. Let no one say we are dealing here with mere political history. This is wordcraft and narrative art at a high level. I hope that the translation captures something of the texture and precision of the language in which the account is written, as well as the sheer force of the events described.

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

484

Kroll

At the time our excerpt begins, the strategic pass known as the T’ung Barrier (T’ung kuan 潼關), about 75 miles east of Ch’ang-an and protecting access to the capital from the central plains, has just been taken by the rebel forces of An Lu-shan 安祿山. For some time previously the T’ung Barrier, a vritually unassailable defensive lodgement, had been held by the imperial army under the command of Ko-shu Han 哥舒翰 But, owing to various court intrigues and, particularly, the growing mistrust and jealousy of the prepotent Yang Kuochung, Ko-shu Han had been foolishly commanded by the court to abandon his defensive posture on 5 July and to seek a direct confrontation with the enemy. On 9 July a major battle took place which, predictably, resulted in the utter discomfiture and rout of the royal forces. The next day, the 10th, the rebels took control of the pass, and the road to Ch’ang-an lay open. No word of the debacle had yet reached the court. It is at this point that we join the narrative.5

[10 JULY]

On this day, [someone] from under [Ko-shu] Han’s standard arrived to report the urgency [of the situation]. His Highness did not make time to summon him to audience; he merely despatched Li Fu-te 李福德 and some others, leading the forces of the Pastoral Inspectorate (chien-mu 監牧),6 to proceed to the T’ung Barrier. When sunset came, the fires of peace and tranquility were not in effect. « The [T’ang] liu tien [唐]六典 (Sixfold Institutes of T’ang): “The beacon watches effected in T’ang garrisons and strongholds are usually separated one from another by 30 li. Every day, at the beginning of evening, each is to be set smoking by a single torch. This is referred to as ‘the fire of peace and tranquility’.” At the time [of which our passage speaks], the defending forces had already been overrun and there was no longer anyone to light the fires. » Now for the first time His Highness was frightened.

5  Tzu-chih t’ung-chien (Hong Kong, 1976), 218.6970–74. 6  This was the service in charge of the crown’s livestock, most important of which were of course the imperial horse herds. See Robert des Rotours, Traité des fonctionnaires et traité de l’armée (1948; rev. ed., San Francisco, 1974), 396–403, also 884–904. The emperor is ordering to the T’ung Barrier the soldiers who usually guard the pastures and keeps in the suburbs near Ch’ang-an. Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

flight from the capital & death of precious consort yang



485

[11 JULY]

On the day I-draconian (jen-ch’en 壬辰) he summoned the Mactatory Ministers7 to give him counsel. Yang Kuo-chung had himself taken control [some time previously] of the Chien-nan 劍南 area;8 upon hearing of An Lu-shan’s revolt he had immediately ordered the Deputy Commissioner (fu-shih 副使) Ts’ui Yüan 崔圓 to prepare secretly a cache of supplies that would be ready, should an emergency force one to take refuge there. Now when events had [actually] come to this point, he came forward to tout the strategy of gracing Shu with the [imperial] Presence.9 His Highness assented.

[12 JULY]

On the day J-ophidian (kuei-szu 癸巳), Kuo-chung convened the hundredfold officials at the Hall of the Levée. Looking alarmed and anxious, with tears flowing, he asked to have their strategies or schemes. They each of them stammered and did not respond. Kuo-chung said, “That Lu-shan has given the appearance of revolt has been reported for a decade now. But His Highness would not believe it. The event of this present day is not the fault of the Mactatory Ministers!” The honor guard stood down. « When the levée is concluded, those who are in the Standing Honorguard of the Triple Paladins of the Left and Right (tso-yu san-wei li-chang 左右三衛立仗) all relax and stand down. » 7  That is, the tsai hsiang 宰相, of which there were two at this time—Yang Kuo-chung and Wei Chien-su 韋見素 (687–762). Tsai hsiang is usually given the functional translation “chief minister,” but the first word of the title indicates the ritual slaughtering and sacrifice of animals (“mactation”) anciently one of the responsibilities of the king’s highest minister. I do not believe we gain anything in our efforts to understand more exactly T’ang culture by obliterating these historical and linguistic nuances in translation. 8  The “circuit” (tao 道) of Chien-nan corresponded approximately to the present province of Szechwan. The name itself means “South of Sword,” denoting the area south of the important mountain pass called “Sword Gate” (Chien-men” 劍門), located in present-day Chien-ko 劍閣 district in northern Szechwan. Yang Kuo-chung had begun his political career as a local official in Chien-nan and continued thereafter to maintain close ties with the area. Late in 751 he had himself appointed “Measuring and Ruling Commissioner” (chieh-tu shih 節度使, functionally military governor) of the region and began building up a private territorial base there. 9  “To grace with the Presence” renders hsing 幸, denoting any visitation or royal progress to an area by the sovereign, even if it be—as here—in flight. Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

486

Kroll

The gentlefolk and the peasants fled hastily in confused panic, going who knows where. The markets and lanes became drear and desolate. Kuo-chung sent [the Lady of] Han 韓 and [the Lady of] Kuo 虢10 into the palace to encourage His Highness to make entrance into Shu 蜀.11

[13 JULY]

On the day A-equine (chia-wu 甲午), there were no more than ten or twenty percent of the hundredfold officials who attended the levée. His Highness held sway at the Loft of Zealous Administration (Ch’in-cheng lou 勤政樓),12 whence he handed down a rescript stating he would personally join the battle.13 Of those who heard this, no one at all believed it. He appointed the Governor of the Capital Municipality (Ching-chao yin 京兆尹), Wei Fang-chin 魏方進, as Great Officer, Notary to the Autocrat (yü-shih ta-fu 御史大夫.)14 and concurrently Commissioner for Disposing the Gests (chih-tun shih 置頓使),15 The Lieutenant Governor (shao-yin 少尹) of the Capital Municipality, Ts’ui Kuangyüan 崔光遠 of Ling-ch’ang 靈昌, was made Governor and fulfilled the duties 10  Elder sisters of Yang kuei-fei who had been ennobled as “Ladies” (fu-jen 夫人) of the Han and Kuo principalities, respectively, and had long enjoyed the monarch’s favor and confidence. A third sister, who will be mentioned later, was given the title Lady of the Principality of Ch’in (秦國夫人). All had free access to the imperial palace. Rumor had it that Yang Kuo-chung and the Lady of Kuo had indulged in an extended (and, according to strict social mores, incestuous) affair. 11  That is, to Chien-nan. Shu was the classical name for the area of what is now central and western Szechwan. Pa 巴, often coupled with Shu, referred to the eastern portion of the province. 12  This multi-storied building was not located on the grounds of the Palace of Great Luminosity (Ta-ming kung 大明宮), north of Ch’ang-an, where the emperor usually resided, but rather stood in the compound of the Palace of Ascendant Felicity (Hsing-ch’ing kung 興慶宫), near the city’s Eastern Market. The Hsing-ch’ing quarter ( fang 坊) had been the place where Li Lung-chi had dwelt from 701 till he came to the throne in 712. After his accession he formally designated the complex of buildings in this place a “palace,” some years later had the whole compound renovated and expanded, and held there throughout his reign many ceremonies, expecially those of a more personal nature. 13  The verb used is cheng 征 literally “engage in a corrective (i.e. punitive) expedition.” 14  Functionally: censor. 15  An irregular and temporary position, with responsibility for preparing the stoppingplaces on the route of the emperor’s flight. “Gest” is a useful word, listed as obsolete by the o.e.d. but deserving of revival, that indicates “the various stages of a journey, esp. of a royal progress.”

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

flight from the capital & death of precious consort yang

487

of Abiding Warden (liu-shou 留守) of the Western Capital. The Army Leader (chiang-chün 將軍) Pien Ling-ch’eng 邊令誠 had in his keeping the keys and cotters of the palace gateways. The Great Commissioner for Measuring and Ruling in Chien-nan (chieh-tu ta-shih 節度大使),16 Chiao 璬, Prince of Ying (潁王),17 was entrusted and employed to proceed ahead to the stronghold and contrive the provisioning of stored supplies along the route. On this day His Highness shifted the honor guard to the Northern Interior (pei nei 北内). « When the T’ang made Ch’ang-an the Metropolis, the Palace of Grand Culmination (T’ai-chi kung 太極宮)18 was regarded as the Western Interior, the Palace of Great Luminosity as the Eastern Interior, and the Palace of Ascendant Felicity as the Southern Interior. The Northern Interior ought, then, to have been within the Gate of the Dusky Warrior (Hsüan-wu men 玄武門).19 But if one speaks of it in terms of a groundview, then, since the Palace of Ascendant Felicity is to the south and the Palace of Great Luminosity is to the north, and the honor guard was shifted from the former and returned to the latter, one may therefore refer to the Palace of Great Luminosity as the Northern Interior. VARIA: HsingShu chi 幸蜀記 (Record of the Progress to Shu)20 says: “His Highness despatched the Commissioner Penetralian (chung-shih 中使) Ts’ao Hsien 曹仙 in charge of a thousand men to beat the drums outside of the Gate of Vernal Luminosity (Ch’un-ming men 春明門).21 He also commanded

16  A titular position. 17  He was the thirteenth son of Hsüan Tsung. 18  The T’ai-chi kung was located In the “palace city” (kung ch’ eng 宮城) at the extreme north-central part of Ch’ang-an. 19  The north-central gate of the city. 20  This text is listed in Hsin T’ang shu 新唐書 (hereafter HTS) (Peking, 1975), 58.1468, as being preserved in the Sung imperial library. It is there described as a work in one chüan and its author is named as Sung Chü 宋巨, an attribution that Ssu-ma Kuang accepts. But there is some question about the authorship; see E. G. Pulleyblank, “The Tzyjyh Tongjiann Kaoyih and the Sources of the History of the Period 730–763” (hereafter cited as “Sources”), Bulletin of the School of Oriental and African Studies 13.2 (1950), 461–62. In any event, the book usually receives disdainful criticism from Ssu-ma Kuang, whenever he cites its version of events. 21  The east-central gate of the city, close by the Hsing-ch’ing kung. Thus, if the drums announcing the monarch’s imminent arrival were beaten outside this gate, the populace would believe that the emperor was about to ride eastward—to the T’ung Barrier, as he had stated earlier that day.

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

488

Kroll

that the stables and corrals22 and the ricks of fodder be set ablaze. Smoke and flames kindled the sky. When His Highness was about to mount his horse, Yang Kuo-chung offered remonstrance, to wit: ‘One ought prudently to safeguard the Ancestral Vault.23 It is improper to set off recklessly.’ Wei Chien-su strenuously opposed [this view], to wit: ‘The power of the outlaws presses near and the people’s hearts are not firm. Your Majesty can do nothing other than leave here and escape the infidels! Kuo-chung is in covert communication with the outlaws, and his words are not to be heeded.’ Back and forth to the number of four times [went the argument], before His Highness at last followed Chien-su’s opinion. Wei Fang-chin was raised to Great Officer, Notary to the Autocrat, and fulfilled the duties of Commissioner Cognizor of the Gests on the Road Ahead (ch’ien-lu chih-tun shih 前路知頓使).” Note that when the simurghbelled chassis was about to leave, the outlaws had invested the T’ung Barrier and men’s hearts were already ill at ease. Surely [the emperor] would not have had drums beaten and fodder set ablaze, to alarm them even more! Kuo-chung had long nourished the design of gracing Shu with the Presence, and Chien-su was after all his adherent. Surely there would never have been this factious dispute before His Highness. Very likely this is owing to Sung Chü’s desire to restore merit to Chien-su, but the events [as narrated by him] are close to calumny. We do not accept them. » When night had fallen, Ch’en Hsüan-li 陳玄禮, Great Army Leader of the Dragon’s Might (lung-wu ta chiang-chün 龍武大將軍),24 was called on to marshal and draw up side-by-side the Six Armies.25 He bestowed cash and silks

22  Housing that portion of the imperial horse herds near the capital. 23  The Ancestral Vault (tsung t’iao 宗祧) was where the commemorative tablets of the distant ancestors of the reigning monarch were kept, when those of more recent antecedents filled all the allotted places in the Ancestral Shrine (tsung miao 宗廟). The tablets of the Shrine would have been transported with the emperor on his retreat from the capital, but the precipitate nature of his departure might have made difficult the collecting of the more numerous tablets housed in the Vault. 24  On the creation and history of the Army of the Dragon’s Might, see des Rotours, Fonctionnaires, 562, n. 2. 25  The “Six Armies” refers to the capital-based soldiers of the imperial guard. The use of the phrase here seems anachronistic, since at this time only four of the six armies that came to be called by the term were in existence (viz. the Armies of the Forest of Plumes on the Left and on the Right [tso, yu yü-lin chün 左右羽林軍] and the Armies of the Dragon’s Might on the Left and on the Right. The Armies of the Divine Might on the Left and on the

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

flight from the capital & death of precious consort yang

489

lavishly and selected more than 900 horses [for his troops] from the stables and corrals. Of those outside [the palace grounds] no one knew anything of this.

[14 JULY]

On the day B-ovine (i-wei 乙未), in the false dawn,26 His Highness, isolated with the Precious Consort, her sisters, the Illustrious Heirs, the consorts, princesses, Illustrious Grandsons, Yang Kuo-chung, Wei Chien-su, Wei Fang-chin, and Ch’en Hsüan-li, as well as the eunuchs and palace ladies who were closest to him and most in his confidence,27 went out [from the city] through the gate of Protracted Autumn (Yen-ch’iu men 延秋門). « The Gate of Protracted Autumn was the western gate of the Forbidden Preserve28 in Ch’ang-an. In the Yung lu 雍錄 (Register of Yung)29 by Ch’eng Ta-ch’ang 程大昌30 is contained “Reference Charts of the Crucial Locales of Han and T’ang,” [showing that] the northeast section of the T’ang Forbidden Preserve encompassed the former city walls of Han Ch’ang-an. The T’ang later changed [the name of] the Night-is-Young Palace (Wei-yang kung 未央宮) to the Basilica of Comprehensive Light Right (tso, yu shen-wu chün 左右神武軍), which made up the remaining two numbers, were not created till 757). Des Rotours, Fonctionnaires, 567, n. 3. 26   Chiu T’ang shu 舊唐書 (hereafter ChTS) (Peking, 1975), 9.232, contributes the dramatically satisfying fact that a “drizzling rain was falling, which wet and soaked” the riders. 27  Although this may read like a long list, it describes in fact a very small retinue, when compared with the hordes of souls, including thousands of officials and servants, that were wont to accompany the ruler on, for instance, his periodic removals to the Eastern Capital at Lo-yang or even his more frequent excursions to the nearby Floriate Clear Palace (Huach’ing kung 華清宮 by the hot springs at Black-horse Mountain (Li shan 驪山). Note, though, that the escorting troops, far from being the “small cavalry contingent” they are often described as, actually numbered about a thousand men (Ch’en Hsüan-li having selected 900 horses for them from the stables). 28  The royal park, located directly north of the city walls and west of the Ta-ming kung enceinte. 29  Yung is the classical appellation for the Shensi area, including the site of Ch’ang-an. For critical comments on this useful book, see Ssu-k’u ch’üan-shu tsung-mu t’i-yao chi ssuk’u wei-shou shu-mu chin-hui shu-mu 四庫全書總目提要及四庫未收書目禁燬書目 (rpt., Taipei, 1978), ii, 1513–14; and Fonctionnaires, xcii–xciv; also the entry on “Yung lu” in A Sung Bibliography (Bibliographie des Sung), initiated by E. Balazs, ed. Y. Hervouet (Hong Kong, 1978), 152. 30  Dates: 1123–1195.

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

490

Kroll

(T’ung-kuang tien 通光殿); its western exit was in fact the Gate of Protracted Autumn. VARIA: The Record of the Progress to Shu reads “On the day C-simian (ping-shen 丙申),31 the hundredfold officials even yet attended the levée.” But that took place on the B-ovine (i-wei 乙未) day.32 Sung Chü is in error. » Abandoning every one of the consorts, princesses, and Illustrious Grandsons who resided outside [the palace grounds], they set off. As His Highness was passing the Depository on the Left (tso tsang 左藏),33 Yang Kuo-chung requested that it be set afire, saying, “Let it not come under the outlaws’ control!” His Highness blanched and replied, “If the outlaws, upon their arrival, do not gain it, they will be certain instead to appropriate [what they need] from the common folk. Better to give it them, and let them not oppress my children heavily.” « History records that Hsüan Tsung had the words of a sovereign of men. » On this day there were still some among the hundredfold officials who came into the levée. When they reached the gates of the palace, they heard still the sound of the clepsydra, and the Standing Honor-guard of the Triple Paladins was justly formal. « The conventions of assembly at the T’ang levée: The [men of the] Triple Paladins stand above as buffers [between the emperor and the officials], divided into five honor-guards, termed the Pentamerous Paladins Within the Forecourt (ya-nei wu-wei 衙内五衛). The first [group] is called the Rendering Deference Honor-guard (kung-feng chang 供奉仗), made up of the Paladins on the Left and on the Right. The second is called the Proximate Honor-guard (ch’in chang 親仗), made up of the Proximate Paladins. The third is called the Guerdoned Honor-guard (hsün chang 勳仗), made up of the Guerdoned Paladins. The fourth is called the Winged Honor-guard (i chang 翊仗), made up of the Winged Paladins. The fifth is called the Honor-guard of Unassigned Hands (san-shou chang 散手仗), made up of [reserves of] the Proximate, Guerdoned, and Winged

31  Equivalent to 15 July. 32  Equivalent to 14 July. 33  One of the two large supply depots of the capital. On its organization, see Ta T’ang liu-tien 大唐六典 (Tokyo, 1974), 20.15b–17b. As we shall see, it was later to be set afire by rampaging citizens, anyway.

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

flight from the capital & death of precious consort yang

491

Paladins.34 At level light,35 when the “transmitting of the strokes”36 is concluded, the Inner Gate [to the palace] is opened. The hundredfold officials enter and stand in ranks. The Illustrious Thearch then ascends the autocrat’s throne. One of the Army Leaders of the Golden Apotropaions (chin-wu chiang-chün 金吾將軍)37 proclaims that the inner and outer sidehalls, to left and right, are tranquil and settled. With the Chamberlains to Communicate Affairs (t’ung-shih she-jen 通事舍人)38 assisting them, the Mactatory Ministers and the officials of the Two Bureaus39 perform the double salutation and ascend the basilica. The Internuncio of the Interior (nei yeh-che 内謁者)40 upon receiving the Directive, calls up the honor guard; the Army Leaders of the Forest of Plumes on the Left and on the Right (tso, yu yü-lin chiang-chün 左右羽林將軍)41 collate their wooden tallies.42 General entrance is then made from the east and west vestibules. At the conclusion of the levée, the Illustrious Thearch walks through the portal [leading] to the antechamber to the east, after which the honor guard is released. The files of honor guards, both inside and outside [the 34  See also Fonctionnaires, 503–06, esp. 504, n. 3. 35  “Level light” (p’ing ming 平明) is sunrise. 36  “Transmitting of the strokes” (ch’uan tien 傳點) refers to the activity of the functionary charged with striking the “cloud-clapper” (yün-pan 雲版), to assemble the court officials to attention outside the door of the audience hall. The “cloud-clapper” was a gong of percussive bronze, fashioned in the shape of a cloud, and suspended from a small scaffold; see the illustration reproduced from the 17th-century San-ts’ai t’u-hui 三才圖會, in Morohashi, 42235.519. 37  See Ta T’ang liu-tien, 25.10a/b. There were two Army Leaders of the Golden Apotropaions, apparently responsible on a rotating basis for security in the two side-halls of the basilica. 38   Fonctionnaires, 187–88. There were sixteen men who held this position, with responsibility for presenting officials and petitioners who had business in the audience hall. 39  Two Bureaus (liang sheng 兩省) are the Bureau of Texts from the Penetralia (chung-shu sheng 中書省, functionally the imperial secretariat) and the Bureau Beneath the Gates (men-hsia sheng 門下省, functionally the chancellery). 40   Fonctionnaires, 243, notes that there were a dozen such officials, whose duties were restricted to audiences held by the empress. But here the title likely refers to the senior of the Chamberlains to Communicate Affairs. The T’ang t’ung-shih she-jen assumed the charges held by the Internuncios of Han and Six Dynasties times. Hu San-hsing is probably using the older title for its dignified effect, despite its obsolescence and the risk of confusion with the differently oriented nei yeh-che of the T’ang. 41  These officers were in charge, inter alia, of guarding the stairs leading to the throne itself. Fonctionnaires, 557. 42  The tallies had to match, verifying that the officers on duty were indeed who they were supposed to be.

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

492

Kroll

hall], dispersed by [the time of] the seventh notch.43 Usually the participants [in the levée] assembled at sunup and dispersed by the sixth notch.44 » When the gates were opened, the palace women came chaotically out. Both within and without [the palace] was tumult and turmoil. No one knew where His Highness had gone. Thereupon royal lords and gentlefolk fled away into hiding, in every direction. Simple peasants from the mountain valleys vied to enter the palace adyta, as well as the mansion-houses of the royal lords, to seize and plunder their gold and valuables; some rode asses up into the basilica. Further, the Depository on the Left, [called] the Greatly Replete Treasury (Ta-ying k’u 大盈庫) was set afire. Ts’ui Kuang-yüan and Pien Ling-ch’eng45 led their men to put down the fire. Then they mustered men to sustain the officials of the [capital] Archive and the [twin] Townships,46 detailing [soldiers] to protect them (i.e. the administrative officials). More than ten people were killed before things settled down somewhat. [Ts’ui] Kuang-yüan then despatched his son eastward, to seek audience with [An] Lu-shan; [Pien] Ling-ch’eng, for his part, offered up the keys and cotters [of the palace] in tribute to him. When His Highness had passed over the Easeful Bridge (Pien ch’iao 便橋),47 Yang Kuo-chung sent someone to set the bridge afire. His Highness said, “The gentlefolk and the plebs will each seek survival by fleeing from the outlaws. How could we sever their road!?” Kao Li-shih 高力士,48 Attendant Inspector of the Inner Court (nei shih-chien 内侍監) stayed back to effect the smothering and squelching [of the fire], and then came on again. 43   K’o 刻 refers to the notches on a sundial or the graduations of a clepsydra. In either case, one hundred k’o made up a full 24 hours; thus, each k’o was roughly 14.4 minutes long (hence the modern use of the word to signify a quarter of an hour). The “seventh notch” would be about one hour and forty minutes past sunrise. 44  That is, the audience normally lasted slightly less than an hour and a half. 45  Who had been made warden and keymaster, respectively, of the city; see above. 46  As an administrative unit, Ch’ang-an was referred to in toto as the Archive of the Capital Municipality (Ching-chao fu 京兆府; prior to 716 it had been called the Commandery [chün 郡] of the Capital Municipality), with the chief officer being a Governor (yin 尹). It was further subdivided into two townships (hsien 縣)—the township of Ch’ang-an, comprising the western half of the city, and the township of Wan-nien 萬年, comprising the eastern half; each township was headed by a Commandant (ling 令). 47  The Easeful Bridge spanned the Wei river about twelve li west of the city. 48  The eunuch Kao Li-shih (d. 762) was perhaps Li Lung-chi’s most trusted—and certainly most enduring—confidant, having served him faithfully since before his accession to the throne.

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

flight from the capital & death of precious consort yang

493

« Hsüan Tsung was the first to establish the [position of] Attendant Inspector of the Inner Court; he ranked it in the Third Grade (san-p’in 三品), and had Kao Li-shih, along with Yüan Ssu-i 遠思藝 take it up.49 » His Highness despatched the eunuch Wang Lo-ch’ing 王洛卿 to proceed ahead and inform and notify the [nearest] commandery and district to arrange for his stop. At the time of the [mid-day] meal, they reached the Palace for Holding Worthies in View (Wang-hsien kung 望賢宮), at Hsien-yang 咸陽. « Hsien-yang district lay 40 li west of the capital city; the Palace for Holding Worthies in View was in the eastern part of the district. » [Wang] Lo-ch’ing and the district Commandant had both absconded. When the Commissionier Penetralian summoned and hailed, there were no clerks or commoners who responded. The sun was approaching the midpoint, and His Highness had still not eaten. Yang Kuo-chung offered up some western cakes (hu ping 胡餅) from the marketplace. « “Western cakes” are the steamed cakes (ch’eng ping 蒸餅) of the present day. Kao Ssu-sun 高似孫50 says: “ ‘ Western cakes’ speaks of putting [the seeds of] ‘western hemp’ (hu ma 胡麻, i.e. sesame)51 on it.” Ts’ui Hung’s 崔鴻 Ch’ien Chao lu 前趙錄 (Register of the Former Chao)52 [says]: “As Shih Hu’s 石虎53 ineffable name was [homophonous with] hu 胡, one says 49  This new post was created earlier in 756. Its placement in the upper Third Grade of the nine-grade bureaucratic hierarchy (on the fifth rung of the 30-step ladder), usually reserved for heads of Bureaus, was in fact rather scandalous. Yüan Ssu-i did not prove as worthy of it as Kao, for, as we shall see, he deserted the emperor, ultimately to offer his services to An Lu-shan. 50  Dates: ca. 1160–1220. He was the author of numerous works, extant among which are the Yen lu 剡錄, Shih lüeh 史略, Tzu lüeh 子略, and Wei lüeh 緯略, for which see entries in A Sung Bibliography, 146–147, 208, 215, 312–13. 51  On this and the other names for sesame, the classic discussion is Berthold Laufer, SinoIranica: Chinese Contributions to the History of Civilization in Ancient Iran, With Special Reference to the History of Cultivated Plants and Products (Chicago, 1919), 288–96. 52  The Ch’ien Chao lu cited here is the first of the sixteen sections of Ts’ui Hung’s (fl. ca. 520) Shih-liu kuo ch’un-ch’iu 十六國春秋. But I do not find this quotation in either the “Ch’ien Chao lu” or the “Hou Chao lu” 後趙錄 (where one would more naturally expect it to be) chapters of the extant version of the work. 53  Third monarch of the Later Chao dynasty (r. 335–49).

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

494

Kroll

‘hemp cakes’ (ma ping 麻餅) instead of ‘western cakes’.” Hsiang-su tsa-chi 緗素雜記 (Assorted Records on Blond Plainsilk)54 says: “There are hawk-

ers of western cakes who, not knowing the [proper] name by which they are designated, have altered the name to ‘stove cakes’ (lu ping 爐餅).” It is due to the fact that they are what westerners (hu jen 胡人) relish that one calls them “western cakes.” » Thereupon the commoners vied with each other to offer up their coarse rice, mixed together with wheat and beans.55 The Illustrious Grandsons vied with each other56 to eat it, in cupped hands. In a moment it was gone, and they still had not eaten their fill. « VARIA: The T’ang li 唐暦 (Almanac of T’ang):57 “When they reached the halt at [the Palace for] Holding Worthies in View, the autocrat’s horse was sickly. His Highness said, ‘Slaughter this horse; strip the wood from the quarters of the mobile palace.58 [With the wood, make a fire to] boil the horse and eat it.’ But the assembled retinue could not bring themselves to eat of it.” The Record of the Progress to Shu [says]: “When they reached the Palace for Holding Worthies in View, all followers in the entourage were hungry. His Highness went into the palace [compound] and rested beneath a tree, disconsolate, as though he were reflecting on having forsaken the Realm Within the Seas.59 Kao Li-shih sensed this and consequently enfolded His Highness’ feet; while giving voice to sobs, he vented his own counsel. His Highness then became calm.” Su Tsung

54  I have been unable to find an independent reference to this title. 55  We are to understand that they had previously been ashamed to come forward and offer their humble fare to the sovereign but, following Yang Kuo-chung’s example, now hastened to do so. 56  Just as the commoners “vie” (cheng 爭) to proffer the rough food, so the royal descendants “vie” to scoop it up: they have already fallen to the level of the general populace. The verbal identity emphasizes the present lack of social distinctions between the two groups. 57   H TS, 58.1460 records that a copy of this work by Liu Fang 柳芳 (fl. 760), in 40 chüan, was preserved in the Sung imperial library. It is no longer extant but much of it was used in compiling the HTS. Apparently it owed much to Kao Li-shih’s personal reminiscences, and it may indeed be behind some of the more private views of Hsüan Tsung revealed In the account before us. See Pulleyblank, “Sources,” 459–60, for further comments. 58  “Mobile palace” (hsing kung 行宮) denotes any temporary dwelling place or rest-house for the emperor, as the Wang-hsien kung had become on this day. 59  That is, the empire.

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

flight from the capital & death of precious consort yang

495

shih-lu 肅宗實錄 (The Veritable Register of Su Tsung):60 “Yang Kuo-chung himself went into the marketplace and filled the sleeves of his jacket with western cakes, which he offered up to the High Illustrious One.” T’ien-pao luan-li chi 天寶亂離記 (Record of Disorder and Disjunction during [the era of] Heavenly Trasure):61 “On the eleventh day of the sixth month62 the Great Equipage, on its progress to Shu, reached the Palace for Holding Worthies in View. The officials and clerks [of the district] had run off into hiding. Only in the dark of twilight were there some of the commonfolk who came forward by slow degrees. His Highness personally asked of them, ‘Is there any food in your households? Do not choose out the fine from the coarse; just bring it forward directly.’ Old and young thereupon competed to carry forth and raise up pots of broth in which were mixed bran and rice, and presented them before His Highness. First to be supplied were the men at arms, then the [ladies of the] Six Palaces63 as well as the Illustrious Grandsons, down to the others, each one of them eating out of cupped hands. In a moment’s time it was all gone, and they still had not eaten their fill. There was a dearth of vessels and utensils and there also were no englobed tapers.64 Those who were following the Equipage pillowed themselves in disarray for slumber and rest, elders and youngsters with no separation or division, trusting the moonlight to enter the portalled courtyard.65 His Highness, along with the [ladies of the] Six Palaces, and the likes of the Illustrious Grandsons did accord 60  This work was in 30 chüan, written during the reign of Li Yü 李豫 (pht. Tai Tsung 代宗, r. 762–79), but is no longer extant. See Pulleyblank, “Sources,” 458. It was Su Tsung (i.e. Li Heng 李亨) who replaced Hsüan Tsung as sovereign on 12 August 756. During the reign of Su Tsung, Li Lung-chi had till his death the title “High Illustrious One” (shang-huang 上皇), by which he is referred to in this quotation. 61   H TS, 58.1468 records this text (under the title T’ien-pao luan-li hsi-hsing chi 天寶亂離 西幸記 [Record of the Progress Westward amid the Disorder and Disjunction during Heavenly Treasure]) as being preserved in the Sung imperial library. It notes it as a book in one chüan and attributes its authorship to Wen Yü 溫畬, who also composed a text called Hsü ting ming lu 續定命錄 (HTS, 59.1542). The T’ien-pao luan-li chi appears to have been extant as late as the early fourteenth century, for it is also listed in the bibliographic monograph of Sung shih 宋史 (Peking, 1976), 203.5112. 62  Equiv. 12 July 756. 63  “Six Palaces” (liu kung 六宮) denotes, by way of classical precedent, the sleeping quarters of the emperor’s chief consort and concubines. See Chou li chu-shu 周禮注疏, 7.9a (Shihsan-ching chu-shu ed.); Li chi cheng-i 禮記正義, 61.7b (Shih-san-ching chu-shu). 64   Kang-chu 釭燭, bowl-shaped lanterns. 65  That is, they slept in the open air. Note that in this account the first night was evidently passed here.

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

496

Kroll

thereat with the [proper] distinctions.”66 Note that it was on the ninth day [of the month] that His Highness [began] the progress to Shu; Wen Yü says it was on the eleventh day, which is wrong.67 As for the remainder [of this passage], one may impartially adopt it. » His Highness acknowledged, in each case, their uprightness,68 and he consoled and reassured them. Everyone in the throng was weeping, and His Highness too wiped away tears. There was an elderly patriarch, Kuo Tsung-chin 郭從謹, who then put forward an utterance, saying, “That Lu-shan embraced and laid up dire intentions was assuredly not a matter of only a single day. And yet, when there were those who came directly to the palace pylons to report of his schemes, Your Majesty time and again punished them, thus bringing it about that he has been able to carry out his treacherous revolt to the fullest degree and resulting in Your Majesty’s exodus and flight. Presumably it was in regard of situations of just this sort that the Former Kings made an effort to draw out and call upon the loyal and the well-born, so as to broaden their own insight and perception. Your vassal still recalls that Sung Ching 宋璟,69 when he was a Minister of State, numerous times put forward candid utterances, which the subcelestial realm relied on for tranquility and peace. But ever since his fall, those vassals who have resided at court have regarded such utterances as ineffable, only ingratiating themselves through flattery and blandishments. Owing to this, Your Majesty has not in any instance been able to be cognizant of affairs outside the gates and pylons [of the palace]. Even your subjects in the weedy countryside have recognized certainly that this day was coming for a long time; it is just that the nine layers [of the palace barriers] were forbidding and inaccessible,

66  The members of the royal family and their intimate companions did keep to the requisite social segregations in their sleeping arrangements, unlike those who followed in the retinue’s train. In other words, the breakdown of social order had extended to all but the highest level of the encampment. 67  Ssu-ma Kuang is himself confused here, for the departure from Ch’ang-an was actually made, as the dates incorporated in his main text indicate, on the thirteenth day of the month (equiv. 14 July 756). Wen Yü is, however, still in error by two days—but two days early rather than late. 68  The “uprightness” of those who were enduring with him the hardships of the flight. 69  Sung Ching (663–737), much-celebrated and right-minded statesman, had an exceptional career, beginning in the time of Empress Wu. His years of greatest power were 717–20, and he was traditionally credited (along with Yao Ch’ung 姚崇 [651–721]) for directing the effective policies of the first years of Hsüan Tsung’s rule, which so prospered the state. Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

flight from the capital & death of precious consort yang

497

and there was no route for one’s trifling, trivial sentiments70 to ascend and break through. If matters had not reached the pass they have now, how would I, gazing upon Your Majesty’s face, find myself able to make this indictment?” His Highness said, “This is due to Our dim-sightedness, and Our regrets have nowhere else to extend,”71 and, with a consoling injunction, sent him away. Presently the Magister of Repasts (shang shih 尚食)72 arrived, bearing the autocrat’s viands. « “Magister” (shang 尚) is “Presider” (chu 主). Of the officials who presided over the autocrat’s viands, there were Stewards to the Autocrat ( feng yü 奉御)73 and Unswerving Aldermen (chih chang 直長).74 » His Highness decreed that they be bestowed first on the officials in his retinue. Only afterward did he eat. He then commanded the men of the armies to disperse and call on thorp and village in search of food. They rejoined at the hour of the ox75 and, with everyone gathered together, moved on. When the night was almost half over, they at last reached Chin-ch’eng 金城. « The district of Chin-ch’eng was attached to the Capital Municipality. Originally [called] Shih-p’ing 始平 district, its name was changed to Chinch’eng when Chung Tsung 中宗, in the second year of Spectacular Dragon (Ching-lung 景龍),76 came to this place to send off the Princess Chinch’eng who was being surrendered to the Tibetans (T’u-fan 吐蕃),77 It is located 45 li west of the Capital City. »

70   Ch’ü-ch’ü chih hsin 區區之心, “nugacious notions”—this is polite, and here heavily ironic, self-depreciation. 71  That is, I have only myself to blame. 72  Although this title, strictly speaking, applies to the official in charge of foodstuffs for the palace ladies (Fonctionnaires, 269–70), here it is clearly being used as an abbreviated designation of the official in charge of the shang-shih chü 尚食局 (Office of the Magistery of Repasts), the food service for the emperor. 73  There were two Stewards to the Autocrat supervising the Office of the Magistery of Repasts. Fonctionnaires, 226; Ta T’ang liu-tien, 11.8a. 74  There were five holders of this title, who assisted the Stewards. Fonctionnaires, 226; Ta T’ang liu-tien, 11.8b. 75  The double-hour equivalent to 1–3 p.m. 76  Equiv. 708 AD. 77  On this incident of international marriage diplomacy, see ChTS, 196A.5226–28, and HTS, 216A.6081. The Chin-ch’eng Princess, who was sent off in 708 to be the consort of the Tibetan king, died in 739. Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

498

Kroll

The district Commandant, for his part, had absconded and the commoners of the district, extricating their own selves, had, each one of them, taken flight. All of their vessels and utensils for food and drink remained, and the [army] officers and soldiers confiscated them for themselves. At this time, many of those who were following [the emperor] absconded and Yüan Ssu-i, Attendant Inspector of the Inner Court,78 likewise slipped away. There were no lanterns in the post-station. People pillowed themselves on each other in disarray for slumber, there being no way anymore to discriminate the base from the noble. Wang Ssu-li 王思禮79 arrived, having come from the T’ung Barrier, and it was then learned that Ko-shu Han had been taken prisoner.80 Ssu-li was made Measuring and Ruling Commissioner for Ho-hsi 河西 and Lung-yu 龍右 and was forthwith commanded to proceed to the stronghold there,81 to collect and muster the scattered soldiers in anticipation of a punitive expedition eastward.82

[15 JULY]

On the day C-simian (ping-shen 丙申), they reached the Ma-wei 馬嵬 post-station. « The Chiang-yü t’u 疆域圖 (Charts of the Border Regions) by a person of the Chin 金83 [shows that] the Ma-wei post-station was located in the [current] Hsing-p’ing 興平 district of the Capital Municipality.84 »

78  See n. 49. 79  Wang Ssu-li had been in command of the veteran troops who formed the vanguard in the disastrous battle of 9 July before the T’ung Barrier. See Tzu-chih t’ung-chien, 218.6968. 80  On 10 July, after the T’ung Barrier had been overrun by the rebels, Ko-shu Han attempted to rally his remaining forces at the Kuan-hsi 關西 post-station, but he was taken prisoner by a cavalry detachment under the allied general Huo-pa Kuei-jen 火拔歸仁 and escorted under guard to An Lu-shan in Lo-yang. Tzu-chih t’ung-chien, 218.6969. 81  Lung-yu, roughly corresponding to the area of the modern province of Kansu along with eastern Tsinghai, was one of the 15 governmental “circuits” in existence at this time. Hohsi designated generally this same area. The stronghold referred to was located near present-day Wu-wei 武威 district in Kansu, about 140 miles north and slightly west of Lanchow. 82  That is, for an eventual counterattack against the rebels. 83  The Jurchen dynasty that ruled North China from 1115 to 1234. 84  Modern geographers place it 25 li west of the present city of Hsing-p’ing, in Shensi. Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

flight from the capital & death of precious consort yang

499

The officers and troops were famished and weary, every one of them testy and exasperated. Ch’en Hüan-li85 considered the [entire] disaster to be due to Yang Kuo-chung [and his misguided policies] and wished to strike him down. Accordingly, Li Fu-kuo 李輔國, eunuch to the Eastern Palace,86 communicated this to the Grand Heir.87 The Grand Heir had not yet made up his mind when, as chance would have it, about twenty of the Tibetan envoys88 checked Kuochung’s horse and were complaining to him over the lack of food. Kuo-chung had not even answered, when the soldiers cried out, “Kuo-chung is plotting revolt with the foreign recreants!” Someone let fly an arrow, which struck his saddle. Kuo-chung fled within the western gate. « The western gate of the Ma-wei post-station. » The soldiers pursued and slew him, hacking and butchering his limbs and body. They raised his head on a spear, outside the gate of the post-station; in addition, they slew his son Hsüan 暄, who had been Esquire-Attendant of the Department of the Census (hu-pu shih-lang 戶部侍郎),89 as well as the Lady of the Principality of Han and the Lady of the Principality of Ch’in. Wei Fangchin, Great Officer, Notary to the Autocrat,90 spoke: “How do you presume to murder a high minister !?” The mob then slew him too. Hearing the disturbance, Wei Chien-su went out; he was beaten by the mutinous men at arms till the blood flowed from his cranium onto the ground. [Someone in] the mob then called out, “Let us not injure the Lord Minister Wei!” They spared him and he was able to escape.91 85  Who had been put in command of the armies by Hsüan Tsung; see above. 86  The Eastern Palace was of course the residence of the crown prince (the “Grand Heir,” t’ai-tzu 太子), Li Heng. 87  Ch’ en Hsüan-li’s plan was carried to Li Heng since, obviously, Hsüan Tsung himself would never have agreed to the execution of Yang Kuo-chung and Ch’en would therefore certainly need the backing of the heir-apparent to carry out the assassination. If Li Heng were in fact to support Ch’en in this, it would clearly have to be followed by the crown prince’s immediate assumption of the throne. 88  Who were part of the imperial contingent. 89  The Department of the Census, sometimes referred to paraphrastically as the “Ministry of Finance,” controlled all matters of taxation. There were two Esquire-Attendants, functionally vice-presidents of the department. Fonctionnaires, 71–72. 90   Governor of the Capital Municipality, he had been given this office and that of Commissioner for Disposing the Gests two days earlier; see above. 91  Wei was the other—and quite malleable—Mactatory Minister at this time; see n. 7. His lack of any real influence was surely the main reason his life was not exacted by the soldiers. Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

500

Kroll

The soldiers then surrounded the post-station. His Highness, hearing the clamor and uproar, asked what was taking place outside. His attendants and acolytes answered that Kuo-chung had rebelled. His Highness took up his staff and put on his sandals,92 and went out the post-station gate. He consoled and reassured93 the soldiers, and then commanded them to draw up into ranks; but the soldiers would not respond to him. His Highness had Kao Li-shih question them, and [Ch’en] Hsüan-li answered, “Since Kuo-chung was plotting revolt, the Precious Consort is no longer fit to render service.94 Let Your Majesty pare away your kindness [to her] and give the ordinances justly.” His Highness said, “We will Ourself dispose it.” He passed in through the gate. And he stood, leaning on his staff, his head bowed. After a time, Wei O 韋諤, Intendant Overseer (ssu-lu 司錄) of the Capital Municipality, came forward and spoke, saying: « Coadjutor Militant Intendant Overseer of the Archive of the Capital Municipality (Ching-chao fu ssu-lu ts’an-chün 京兆府司錄參軍) was [a position] in the upper Seventh Grade, Primary (cheng ch’i-p’in shang 正七品上).95 At the beginning of Martial Virtue (Wu-te 武德, 618–27), the [title] Chief Actuary (chu po 主簿) in an “Isle” (chou 州) was altered to Coadjutor Militant Intendant of Affairs (lu-shih ts’an-chün 錄事參軍); he held sway over the rectifying of wrongs and omissions, and he controlled the seals and tallies. In the initial year of Opened Prime (K’ai-yüan 開元, 713–42), this was altered to Intendant Overseer.96 »

92  He had been resting. This is a nice dramatic touch. 93  Earlier, we recall, the monarch had successfully “consoled and reassured” (wei lao 慰滎) the officials and members of the royal family traveling with him. But the result of the identically described action this time is quite different. Shortly (after Yang kuei-fei’s death) the phrase will be repeated a third time, again with the soldiers as object, and at that time the monarch’s actions will once more be allowed to produce the desired effect. 94  The argument is that her close familial connection to Yang Kuo-chung implicates her in his supposed “revolt” and requires her punishment. 95  Of the bureaucratic hierarchy. There were two such officials for the Capital Muncipality; Fonctionnaires, 689. 96  Both this note of Hu San-hsing and the next are, one must admit, rather exasperatingly placed. We have to wonder whether Hu was not consciously—and mischievously—trying to draw out the suspense of the narrative by interrupting it at these tense moments.

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

flight from the capital & death of precious consort yang

501

“Now, when the mob is enraged, it is difficult to oppose them. « Quoting the words of Tzu-ch’an 子產 of Cheng 鄭, in the Tso chuan

左傳.97 »

“Security and peril lie in a notch of the gnomon’s shadow.98 I would that Your Majesty decide quickly.” And he kowtowed till blood flowed. His Highness said, “The Precious Consort has always dwelt deep in the palace. Whereby could she know of Kuo-chung’s plot for revolt?” Said Kao Li-shih, “In truth, the Precious Consort is blameless. Yet, with the officers and troops having already slain Kuochung, how can Your Majesty presume even yourself safe [from them], if the Precious Consort remains by your side? I would that Your Majesty consider this carefully. If the officers and troops be placated, Your Majesty will be safe.” His Highness at last gave the orders for Kao Li-shih to lead the Precious Consort to the Buddha Hall,99 where he put her to death by strangling her. The corpse, carried on a litter, was deposited in the courtyard of the poststation. [Ch’en] Hsüan-li and some others were summoned in to look upon it. Hsüan-li and the others removed then their helmets and loosened their armor; they touched their heads to the ground and implored pardon. His Highness consoled and reassured them, and he commanded them to make known [the deed] and proclaim it to the soldiers. Hsüan-li and the others all cried out “Long life!” and, performing a double salutation, exited. Only now did they set in order the companies and files and make plans to move onward. [Wei] O was the son of Chien-su. Kuo-chung’s wife, P’ei Jou 裴柔 « P’ei Jou, who had formerly been a gleemaiden in Shu. » and his young son Hsi 晞, along with the Lady of the Principality of Kuo100 and the Lady’s son P’ei Hui 裴徽, all fled to Ch’en-ts’ang 陳倉.101 The district 97  See Legge, The Chinese Classics, vol. 5: The Ch’un Ts’ew, with the Tso Chuen (1872; rpt., Hong Kong, 1960), 444, line 11; tr. 448. 98  That is, there is scant time to mull the decision. One “notch” of the sundial was not quite fifteen minutes (see n. 43). 99   Fo-t’ang 佛堂, the building within which the Buddha images stood, for worship. May we imagine that Lady Yang was allowed to die here so that she might end this life while gazing on a representation of Amitābha and invoking, with her last thoughts, rebirth in his Western Paradise? 100  See n. 10 above. 101  About 100 miles west of Ch’ang-an; present-day Pao-chi 寶雞 district, in Shensi.

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

502

Kroll

Commandant Hsüeh Ching-hsien 薛景仙, at the head of his bailiffs, hunted down and apprehended them and took their lives. Later this day, as the text goes on to record, the crown prince Li Heng was persuaded by his advisers not to journey on into Shu with the emperor, but rather to establish his own headquarters in the northwest, whence to gather loyalist forces to combat the rebels. Refusing Hsüan Tsung’s offer to resign the throne to him immediately, Li Heng and a portion of the entourage then turned to follow a different track. It was decided presently that Li Heng would occupy Ling-wu 靈武,102 the seat of the Shuo-fang 朔方 circuit, as his base. He reached Ling-wu on 9 August. Three days afterward he accepted the urgings of his supporters and declared himself sovereign, inaugurating on the same day the new reign-period of “Ultimate Virtue” (Chih-te 至德). Meanwhile, Li Lung-chi with a diminished retinue continued on to the southwest, finally arriving in Ch’eng-tu 成都 on 28 August. Only 1,300 followers (bureaucrats and soldiers, some of whom had joined the progress along the route) remained with him. His glory and his might had long since departed. On 10 September he was informed by messengers sent from Ling-wu of his son’s assumption of the throne four weeks earlier. On 14 September he published a rescript, acknowledging his emeritus status. Though he would live on for six years yet, he would no longer figure in affairs of state. An era had ended. 102  Present-day Yin-ch’uan 銀川, in Ningsia.

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

Foreign Policy P. A. Herbert In contrast to later periods of her history, China in the T’ang Dynasty was receptive to foreign influence and welcomed foreign trade provided it remained within the bounds of government control. The attitude to foreigners expressed in the first half of the T’ang reflects the self-confidence of the Chinese. The Confucian concept of civilization limited the term culture to Chinese culture. China was the centre of the world and the bastion of civilization. There was no concept of race or nationality in the modern sense in T’ang China. Individual foreigners, if they adopted the Chinese way of life, were accepted as Chinese and occasionally even received the highest possible social recognition: a post in the Chinese civil bureaucracy. Similarly, foreign states were judged to be civilized or barbarian according to the degree to which they modelled their political and social systems upon the moral principles of Confucianism. The only form of foreign relations acceptable to the T’ang, and indeed to China throughout the imperial era, was one which reflected China’s assumption of her cultural superiority over her neighbours. If a neighbouring state wished to maintain friendly relations with China, she had to send tribute to the imperial court to acknowledge acceptance of Chinese suzerainty. The Chinese emperor in return confirmed the foreign ruler in his position, conferring on him an appropriate Chinese title, similar to that of a high provincial official within the Chinese state. Like a Chinese official, the foreign ruler received a set of half tallies, the corresponding halves of which were retained at the Chinese court. Foreign envoys were required to produce one of these half tallies as their credentials when they came to court.1 With the exception of concluding the wars in Korea, inherited from the Sui, the T’ang rulers were reluctant to undertake large scale military campaigns to advance foreign policy. Rather they relied on diplomatic activity, cloaked in Confucian moral terms, to pursue their ends in relations with their neighbours. If a tributary state defied the orders of one of the T’ang emperors or infringed the Confucian moral code, the emperor would castigate its ruler in terms similar to those used by a father admonishing his son. Source: “Foreign Policy,” in P. A. Herbert, Under the Brilliant Emperor: Imperial Authority in T’ang China as seen in the Writings of Chang Chiu-ling, Canberra: Faculty of Asian Studies, Australian National University, 1978, 66–86. 1  E. H. Schafer, The Golden Peaches of Samarkand, pp. 26–27. © koninklijke brill nv, leiden, ���9 | doi:��.��63/9789004380158_017

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

504

Herbert

Chang Chiu-ling was a firm believer in the moral and cultural superiority of China over her neighbours and in the power of the ‘superior’ form of Chinese culture to impress itself upon less civilized peoples and gain their ready acceptance. He constantly advocated diplomatic, rather than military, solutions to the problems of foreign relations and opposed any initiative from military men or glorification of their achievements. In his eyes, generals should take orders from the court, while foreign policy should be strictly in the hands of civil court advisers. Chang’s orthodox Confucian attitude to foreign relations was no doubt based to some extent on personal experience. The indigenous people of his native Ling-nan had a lifestyle based on a primitive form of agriculture which was not incompatible with the ways of the Chinese. To integrate them into the Chinese system of provincial administration and train them to follow the Confucian moral way of life was comparatively easy. In conducting relations with China’s northern neighbours, Chang sought to follow the same methods which, since the Ch’in and Han Dynasties, had proved so successful south of the Yangtze. The policy of moral suasion and acculturization was doomed to failure in the north. The people of the northern steppes had developed a pastoral nomadic way of life suited to their environment, while the tribes of the area which is now northern Korea and Manchuria were hunters and fishermen, in whose lives agriculture played a subordinate part. Confucianism recognised only one lifestyle, that based on intensive cultivation, as civilized, all other ways of life being condemned as barbaric. Geographical conditions forced the Chinese constantly to face up to the problem of relations between themselves and the ‘northern barbarians’. There was no sharply divided frontier between the lands suitable for Chinesestyle intensive agriculture and the pastoral activities of the nomads. In early times, the Chinese tried to mark the frontier by constructing walls to keep the barbarians out and their own people in. But even the Great Wall system was no real defence, given the wavering balance of power. The border lands were thus a source of constant conflict, coming first under the control of a powerful Chinese dynasty and then acknowledging the suzerainty of a strong nomadic group. In the T’ang, the Chinese sought to control these often disputed areas indirectly, by encouraging the growth of small buffer states, tied to the Chinese by tributary status. China called upon these states to supplement her own forces in time of war with the powerful, outer tribes.2 2  The question of the balance of power between the Chinese and the nomads and of transition areas is treated in O. Lattimore, Inner Asian Frontiers of China, passim.

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

Foreign Policy

505

The ‘barbarian’ form of warfare was based upon lightning raids and equally sudden withdrawals. Permanent defences in the form of fortifications after the Chinese pattern had no place in the traditional fighting techniques of the northern nomads.3 The Chinese had therefore to equip themselves with highly mobile cavalry forces. Ironically, they were dependent upon their northern neighbours to supplement and improve their stock of horses. At first the Chinese relied upon tribute gifts and upon private trade to replenish the numbers of horses in their state pasturelands, but in 727, as part of a treaty between the Chinese and the Turks, a regular official system of marketing was set up by which foreign traders obtained Chinese silk in exchange for horses.4 During the early years of Hsüan-tsung’s reign, the main threats to China’s border came from the Tibetans and the Turks, in the north and north-west. The Tibetans sent a series of missions to the Chinese court requesting a settlement of the frontier between them, but their proposals were constantly rejected by the Chinese because of the highhanded attitude of the Tibetans towards the etiquette demanded of tributary states. A series of Tibetan raids on China’s western borders ended in 730, after internal dissension and Chinese military power had chastened Tibetan pride. The Tibetans again sued for peace and Hsüan-tsung finally sent a mission to treat with them. In return, a tribute mission was sent to Hsüan-tsung, thus sealing the submission of Tibet to China. The border question was finally settled in 733.5 At the beginning of Hsüan-tsung’s reign, the Turks were weak and divided. However, in 720, under their new khan, P’i-ch’ieh, and his brilliant adviser, T’un-yü-ku, they defeated the Chinese at Liang-chou6 and won back their prestige among the northern nomads. In 721, P’i-ch’ieh sued for peace. The Chinese urged him to make a sincere submission and warned him that they would take the measures necessary to protect themselves against his possible treachery.7 In 725, when preparations were being made for the feng-shan sacrifices at Mount T’ai, the Chinese invited a Turkish delegation to observe the great dynastic sacrifice, hoping thus to awe the Turks into submission.8 Friendship between China and the Turks was finally sealed in 727 by the tribute mission and establishment of the official horse trade mentioned above. 3  When in 716 P’i-ch’ieh, khan of the Turks, wished to build fortifications, his adviser T’un-yüku argued against this strategy as being alien to the Turkish tradition; see TCTC 211, p. 6722. 4  T CTC 213, p. 6779; CTS 194A, p. 16b. 5  T CTC 212–213, passim. 6  T CTC 212, p. 6743. 7  T CTC 212, p. 6744. 8  T CTC 212, pp. 6764–6765.

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

506

Herbert

The letters and memorial translated in this section refer to China’s relations with her north-eastern neighbours, the Khitan and the state of Po-hai. Po-hai was situated just north of the Korean peninsula and was recognised by the Chinese as a tributary state. It was but newly established, having been created at the end of the seventh century by Tae Choyŏng (Chinese Ta Tso-jung), leader of one of the group of northern Korean peoples known as the Mo-ho who had been scattered after Kao-tsung’s defeat of Koguryŏ in 668. In 696, Tae Choyŏng proved himself an effective and strong leader when aiding a Khitan insurrection, for he held out against the Chinese even after his allies were defeated. He became a magnet for the Mo-ho peoples and for remnants of the Koguryŏ forces. He settled them in territory to the east of Chinese-held Ying-chou, declaring himself king of the state of Chen, and sent envoys to the Turks. When Chung-tsung was restored to the throne in 705, he was anxious that the balance of power on the north-east border should not be tipped against China’s favour. He therefore sent a mission to Tae Choyŏng seeking friendship. Negotiations between China and the new state were somewhat delayed because of Khitan and Turkish insurrections in the area through which their envoys had to pass, but finally, in 713, China gave Tae Choyŏng formal recognition, declaring him to be ‘King of Po-hai Commandery and Governor of Hu-hanchou (Po-hai chün-wang Hu-han-chou tu-tu). When in 719 Tae Choyŏng died, Hsüan-tsung recognised as his successor his son Tae Muye (Chinese Ta Wu-i).9 In 726, King Muye took exception to the fact that the Mo-ho of Hei-shui (‘Black River’, modern Hei-lung-chiang), a group independent of, but related to, the people of Po-hai, had sent a mission to the Chinese court and had been granted certain lands close to his borders. He claimed that the Mo-ho of Hei-shui had previously promised to consult him in the event of negotiations over such lands. The king intended to attack the Mo-ho of Hei-shui, but his brother Munye (Chinese Men-i), who had previously spent some time at the Chinese court as surety for his state’s loyalty to China, and who therefore understood the Chinese attitude to their tributary states, advised against action which would bring down upon Po-hai the irresistible weight of Chinese wrath. Munye was at first appointed one of the commanders of the expedition against the Mo-ho of Hei-shui, but when he persisted in advising caution, he was dismissed and fled to seek refuge in China. He was well received at court. When his brother sent in a request for his execution, he was secretly sent to safety in the far west of Chinese territory. To shield him from his brother, the emperor informed King Muye that he had sent Munye, repentant of his ‘crime’, to Lingnan. By the autumn of 732, King Muye had discovered the truth concerning his 9  C TS 199B, pp. 11b–12a.

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

Foreign Policy

507

brother’s whereabouts. He therefore wrote an indignant letter to Hsüan-tsung accusing him of preaching trust to others while himself lying, and again calling for Munye’s execution.10 King Muye’s accusation placed the Chinese in an awkward position. Their claim to suzerainty over their tribute states was based, in theory, upon the moral superiority of China, which gave the Chinese emperor the right to appoint and direct the rulers of lesser states. That claim to moral superiority was directly challenged by King Muye’s letter. The reply must convince the king that the Chinese emperor had hidden Munye with the highest possible motives, for the general good of King Muye and his state, and thus cause him to become submissive once more. The letter as drafted by Chang (Passage a, i) shows masterly diplomatic talent. Chang succeeds in manipulating the Confucian moral precept of brotherly love to justify the Chinese emperor’s behaviour, thus turning the explanation of a lie into an expression of righteous indignation which places King Muye in the wrong. Nevertheless, the outcome of this letter was far from satisfactory for China. In the ninth month, Po-hai naval forces attacked the city of Teng-chou on the tip of the Shantung Peninsula, killing its prefect. Hsüan-tsung raised troops against his rebellious vassals.11 At first, Po-hai appeared to submit (see Passage a, ii), but in the first month of the following year, Munye himself was sent by the emperor to recruit men in the area around modern Peking, while Kim Saran (Chinese Chin Ssu-lan), the son of the king of the Korean kingdom of Silla, serving at the Chinese court, was sent home to obtain reinforcements for a combined attack on Po-hai. The allies suffered great losses as a result of a heavy fall of snow and returned to their bases unsuccessful. Encouraged by the failure of the Chinese attack, King Muye tried to have Munye assassinated within Chinese territory, but the injury inflicted upon him was not fatal.12 In 734, Po-hai remained intransigent and Chang Chiu-ling, on behalf of the emperor, wrote to the king of Silla, urging him, should an opportunity present itself, to attack Po-hai. For this action, he was promised rich rewards.13 The king of Silla, through his son Kim Saran, requested permission to establish garrisons on the P’ei River14 in order to co-operate with the forces of north-east China in creating a defensive border against Po-hai. The request was granted 10   C TS 199B, pp. 12a–b; TCTC 213, pp. 6774–6775. 11   T CTC 213, p. 6799; CTS 8, p. 18b; HTS 5, p. 10a; CTS 199B, p. 12b. 12   T CTC 213, p. 6800. 13   C CC 8, p. 5b; CTW 284, p. 13a; WYIH 471, p. 1b. 14  For a discussion of the identification of the P’ei River, see K. H. J. Gardiner, The Early History of Korea (Canberra, 1969), p. 16. E. G. Pulleyblank, in The Background of the

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

508

Herbert

in the spring of 736.15 However by the late spring of 736, Po-hai had apparently ceased to be hostile towards China and, to prove her return to the Chinese fold, had seized Turkish envoys who sought an alliance with Po-hai against the Hsi and Khitan, tributaries of China. Chang Chiu-ling, speaking for the emperor (Passage a, iii), gently reprimanded the ruler of Po-hai for such hasty action, which the Chinese feared might upset the delicate balance of power recently created in the area. A further letter, written in the summer of the same year (Passage a, iv) confirms that Po-hai was considered to have made a sincere submission and to have become once more a docile tributary of China. The other two letters and the memorial translated here refer to relations between China and the Khitan and an allied people, the Hsi, who controlled an area to the west of Po-hai, north-east of the Great Wall. The Khitan, descendants of the eastern branch of the Hsien-pei of the Han and Six Dynasties, were led by the formidable K’o-t’u-kan, who had emerged in the early years of Hsüan-tsung’s reign as king-maker and de facto leader. In 720, K’o-t’u-kan, then a popular officer, had been recognised as a threat by Li P’o-ku, who since becoming king of the Khitan two years before,16 had remained a loyal ally of the Chinese. King P’o-ku wished to dismiss K’o-t’u-kan, but K’o-t’u-kan struck first, raising rebel forces against him. King P’o-ku fled to Ying-chou in Chinese territory and K’o-t’u-kan put the king’s brother Li Yü-kan in his place. In order that the Khitan might not lose the support of their powerful neighbour, K’o-t’ukan petitioned Hsüan-tsung’s pardon for his treacherous action. His request was granted.17 When King Yü-kan died, his successor incurred the enmity of K’o-t’u-kan and fled to China, whereupon Li Shao-ku was made king of the Khitan.18 In 726, King Shao-ku, together with his neighbour the king of the Hsi, received confirmation of authority from the T’ang court in the form of Chinese princely titles. To seal their allegiance, both were given Chinese brides, who, although not actual daughters of the emperor, had been raised to the status of princesses.19 Up until 730, there was no serious clash between the Khitan and the Chinese, but in that year, K’o-t’u-kan killed King Shao-ku and submitted to the Turks.

Rebellion of An Lu-shan, p. 115, note 61, states that it should be identified with the Dĕdong River, which runs through Pyongyang. 15   C CC 9, p. 1b; CTW 285, p. 10a; WYYH 471, p. 2a; TFYK 971, p. 11b. 16   T CTC 212, p. 6733. 17   T CTC 212, p. 6734. 18   T HY 96, p. 1718. 19   T CTC 213, p. 6770.

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

Foreign Policy

509

The murdered king’s Chinese widow fled to the Chinese city of Yu-chou.20 The Chinese, seeing the allegiance of the Khitan thus lost to them, now took action. Various generals launched a series of attacks on K’o-t’u-kan’s forces, but without success. It was not until Chang Shou-kuei21 took up the post of Military Governor of Yu-chou that K’o-t’u-kan met his match. In the sixth month of 734, Chang Shou-kuei succeeded in inflicting a great defeat on the Khitan.22 The letter urging the Khitan to submit to China and thus avoid further defeat and bloodshed (Passage (b)) was probably written shortly after Chang Shou-kuei’s victory, in any case before the assassination of the Khitan king and K’o-t’u-kan in the winter of 734–5. The Khitan remained unmoved by diplomatic pleas and military force alike. Hard pressed by Chang Shou-kuei, they resorted to trickery, making a false declaration of their intention to surrender. Chang Shou-kuei’s envoy, Wang Hui, sent to accept their so-called surrender, realised that he had walked into a trap. However, he succeeded in persuading Li Kuo-che, who shared the command of the Khitan forces, to turn against K’o-t’u-kan. One night early in 735, Li and his followers assassinated K’o-t’u-kan, the Khitan king and their partisans.23 Most of the Khitan surrendered, but a few dissidents fled to inaccessible places, from which they harassed the Chinese frontier forces for some time to come.24 Li Kuo-che did not remain long on the throne which he had seized. Some time before mid-summer of 735, he was murdered by his minister Nieh-li,25 who in turn usurped the position of ruler of the Khitan. Wishing to retain Chinese support, Nieh-li appealed to the Chinese concept of the Mandate of Heaven, expressed through popular feeling, which might be withdrawn from a tyrant and be given instead to a man who would conform to the morals of a true king. He informed Hsüan-tsung that he had killed Li Kuo-che to fulfil the wishes of the people, who felt oppressed by his cruelty. The fluid state of affairs on the north-eastern border left little room for Chinese indulgence in moral niceties, so Hsüan-tsung confirmed Nieh-li in his position, while warning that usurping rulers often fall victim to the fate which they have dealt out to their predecessors (see Passage (c)).

20   T HY 96, p. 1718. 21  For biographical information on Chang Shou-kuei, see CTS 103, pp. 4b–6a; HTS 133, pp. 3a–4a. 22   T CTC 214, p. 6807; THY 96, p. 1718. 23   T CTC 214, pp. 6808–6809. 24   T FYK 358, p. 12a. 25  Chang Chiu-ling’s letter to the assassin is dated midsummer.

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

510

Herbert

The Turks were quick to take advantage of the supposed internal weakness of the Khitan. Soon after Nieh-li’s usurpation, they attacked the Hsi and Khitan, but they had miscalculated their chances and met with a sound defeat.26 This attack probably followed the unsuccessful Turkish attempt to make an alliance with Po-hai, mentioned above. Hsüan-tsung wished to follow up the victory of the Hsi and Khitan by a campaign against the Turks, in the hope of completely destroying the power of China’s troublesome northern neighbours. However, Chang Chiu-ling, in the memorial translated here as Passage (d), urged caution. China must at all costs maintain her moral integrity in relations with her neighbours in order to win their confidence and persuade them to act out a friendly and subordinate role. Chang’s suspicion of the motivation of military men is seen clearly in this memorial. In his opinion, generals, instead of remaining the loyal servants of the imperial will, provoked hostilities in order to win glory and fame for themselves. No great campaign was launched against the Turks after the victory of the Hsi and the Khitan. An uneasy peace was maintained on the north-eastern border. When two decades later danger from the north-east did confront the court at Ch’ang-an, it was from within the ranks of the Chinese army, as Chang Chiu-ling had feared, from a ‘barbarian’ general in the Chinese employ, the rebel An Lu-shan. Translations Passage (a) Four Official Letters to the King of Po-hai (Drafted by Chang Chiu-ling for the Emperor Hsüan-tsung) CCC (SPTK edition) 9, pp. 9b–11b. See also: CTW 285, pp. 10b–12b. WYYH 471, pp. 3b–5b. Note: According to my reconstruction of the events alluded to in these letters, as given above, the letters are printed in incorrect order in these sources, the third letter appearing before the second. The translations of the letters are given here in the order of my reconstruction. 26   T CTC 214, p. 6812. Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

511

Foreign Policy

(i) Decree addressed to the Prefect of Hu-han-chou,27 King of Po-hai, Tae Muye You and your brother having quarrelled with each other, [your brother] Munye, in dire straits, came to seek asylum with Us. How could We refuse him refuge! On the contrary, We sent him to the western border. Moreover, let it be said that for your sake We did not fail [to do what was right], but rather took satisfactory action. Why was this? Although your territory is an out-of-the-way land by the sea, your people are practised in Chinese ways. On the subject of brotherly affection and respect, can it be that you require instruction? Blood ties are so close that it is painful to one to sever them oneself. Even if Munye committed a great fault, you should bear with him now that he has reformed. You make your request that he be sent back to the east with the intention of slaughtering him. How can We who instruct the empire in filial piety and affection bear to grant such a request! We are truly concerned over your reputation and behaviour. It is certainly not the case that We have given protection to a fugitive from justice. You do not recognise the mercy of Our state and so turn against virtue.28 If you base [your actions) on that which goes against [propriety], We can have no other recourse. In recent years, We have practised forebearance out of concern for Our own inner territory. If We have so far failed to give marching orders to Our generals, it is because there is a proper time for action. If you are capable of regretting your error and making a sincere submission, you will avert disaster and experience good fortune. Your words appear to be submissive,29 but in your mind you still cling to your misguided intention and so request permission to kill Munye as a prior condition to rendering allegiance. What a way to speak! We look to you to memorialise just to show loyalty and sincerity. If you can give the matter mature consideration, you will simply not be able to bear [to maintain such an attitude]. Now We send eunuch envoys to convey to you Our opinion, setting it out in full word by word. Your envoy Li Chin-yen we have likewise dealt with personally and he understands the entire matter. [Written in] autumn, cold weather. Wishing you and your officials, commanders and people peace. We also send Ts’ui Hsün-i30 with Our mission to you. We speak of but little.

27  Hu-han-chou was the capital city of the state of Po-hai. 28   Pei-te (背德) WYYH gives pei-chen (背联) ‘turn against us’. 29   Ssu-hsün (似順). WYYH gives i-hsün (已順) ‘are already submissive’. 30  Unidentified. Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

512

Herbert

(ii) Decree addressed to the King of Po-hai Commandery, Governor of Huhan-chou, Tae Muye: Many sailors whom you in your stubborn stupidity sent [to attack us] and Chinese who had previously fallen into your hands came to memorialise your sincere submission and to inform Us that you have entirely ceased hostilities. If you can continue to maintain this position for a long period and always keep to your border, you will be rewarded with great good fortune and We will trouble you no more. [Written at] the time of increasing cold. Wishing you and your officials and people peace. We write of but little matter. (iii) Decree addressed to the King of Po-hai Commandery, Governor of Huhan-chou, Tae Muye: We have never heard of one who, not understanding what constitutes revolt and what obedience, and not recognising the portentous significance of whether the people remain in his domain or flee away, was able to control a state. In recent years you turned against virtue and already took steps towards disaster. Recently you were able to feel regret for your error and did not cast aside altogether the demeanour of a vassal. Your delusion had not led you too far astray and you restored your goodness.31 We take note of people’s strong points and overlook their weaknesses. How much more do We greet this your submission with expressions of joy! Is it not proper that the eastern territory should be eternally blessed! Ta Shu-ch’ing32 and others whom you ordered to come to our court We have already dealt with, bestowing upon them all offices and rewards and We believe they are cognizant of the entire matter. The hostages whose return you request We are likewise sending back. Moreover recently We received your memorial saying that the Turks had sent envoys seeking an alliance with the intention of attacking the two barbarians. The Hsi and Khitan have now submitted to Us and the Turks out of a grudge against Us are seeking to mete out their revenge upon the two barbarians, but you would not agree to [be a party to] their action. You cast aside all caution and when the envoys were about to leave, seized and bound them. This was not an act of righteousness, but a surrender to base feelings. How much worse in a ruler! However, We know that you have a loyal heart and your actions are bound to be

31   Mi fu fei-yüan shan yu ho-chia (迷復非遠善又何加). WYYH gives mi fei fu shan chi yu k’o-chia (迷非復善即又可嘉) ‘you realised the error of your deluded behaviour and returned to goodness, bringing us cause for rejoicing’. 32   Shu (戍). WYYH: Mao (茂). CTW: Ch’eng (成).

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

Foreign Policy

513

such that We may approve them. You will always maintain your sincerity and you will continue to be blessed. [Written at] the end of spring. Wishing you, your officials and people peace. We write of but little. (iv) Decree addressed to the Prefect of Hu-han-chou, the King of Po-hai, Tae Muye: You recently made a misguided plan which brought you close to disaster. But when you were not far short of showing complete disregard for moral principle and received word of what was the correct thing to do, you were able to follow it. What wisdom this shows! We ignore men’s faults and take as the criterion of Our judgement things as they actually are. It is manifest that your heart is cleansed of impurities and you are certainly calm in your mind. We reckon that since you are wholly sincere, you will keep the eastern barbarians under constant control. What anxiety will again trouble your sons and grandsons for a hundred generations to come! Recently when your envoys came, We were fully aware of their submission. Likewise your request regarding the guards and hostages We have already granted. Ta Lang-ya and the others who formerly defied the regulations of the state and fled to the south We likewise absolve of their guilt, and allow them to return to the barbarian territory. May it be known to you that all this is Our intention. [Written at] the beginning of summer, gradually increasing heat. Wishing you, your commanders and people peace. We write of but little. Passage (b) Decree Addressed to Chü-lo,33 King of the Khitan, his Officer K’o-t’u-kan and the Prefect of Shu-huo, Yü-hsien34 (Drafted by Chang Chiu-ling on behalf of the Emperor Hsüan-tsung) CCC (SPTK edition) 8, pp. 10b–11b. See also: CTW 285, pp. 1b–2a. WYYH 471, p. 7a–b. 33  Chü-lo (壉埒). THY 96, p. 1718 gives his name as Ch’ü-li (屈列). Compare Passage (c), note 38. 34  I am unable to identify this place or person.

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

514

Herbert

Following the Way brings good fortune. This only the wise can construe. Going against the rules of correct behaviour brings disaster. Can fools realise this? In recent years you have recklessly defined your appointed fate and truly you have reared a source of sorrow.35 If you now realise what you have done, it will not be too late in the light of this knowledge to turn disaster into good fortune, to turn defeat into success, to avoid the danger of a hundred deaths and to guarantee the achievement of complete peace. Thus your former defection showed such insolence [but] if you now return to allegiance, what wisdom will you show! All that is necessary is for you to prepare plans based on loyalty and sincerity and to lead your people away from the field of death. Thereafter what more will there be to trouble you? We have never in our dealings with all the barbarians broken a treaty obligation. Moreover, in your case We formerly showed favour. When We hear that you have thus returned to allegiance, We will rejoice and be comforted, firstly because armed conflict will cease and secondly because [correct] relations between ruler and subject will be resumed as before. Among the common people, [the time of] planting will not be missed and herdsmen will lead their animals to seek out the lush pastures and sweet streams, no longer fearing troubles from across the frontier. Moreover, knowing your will, the people will respect it of their own accord, for who does not seek peace? If you maintain this policy forever, you will have no further cause for concern. In your opinion, what dissimulation can there be in these Our words?36 As soon as your horde returns to allegiance, We will allow you to remain in peace. You may discuss a full settlement with Chang Shou-kuei. We will endeavour to follow the wishes of [your] barbarian tribe, that you may follow the fertile lands according to cold and hot seasons, that your people may have no cause to be discontented. [Written at] the end of winter, at the time of severe cold. Wishing you, your officers, army officers, prefects and lower officials and the common people of your horde peace. We send you our letter speaking of these few matters. Passage (c) Decree Addressed to Nieh-li,37 Governor (tu-tu) of the Khitan (Drafted by Chang Chiu-ling on behalf of Emperor Hsüan-tsung) CGC (SPTK edition) 9, pp. 2b–3a. 35   Shih yang huo t’ai (實養禍胎). WYYH gives huo shih tzu i (禍實自胎) ‘truly you have brought disaster upon yourselves’. 36   Ho chia (何假). WYYH gives pu-chia (不假) ‘there can be no dissimulation in…’. 37  Nieh-li (涅禮). CTW gives Ni-li (泥禮). Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

Foreign Policy

515

See also: CTW 285, pp. 2a–3a. WYYH 471, pp. 7b–8a. In the past, Ch’ü-lieh38 and [K’o-Jt’u-kan who were evil and heartless, brought bitter sorrow upon the common people and rebelled against Us. All through the time that they were consolidating their own position, able-bodied men were prevented from carrying out agricultural work and cattle and horses had no chance to graze. Moreover, by joining with the Turks and levying heavy taxes, they caused the tribesmen to groan under oppression, as you are aware. Li Kuo-che, in response to the frustration felt by the people, assassinated the bearers of evil and the chieftains and warriors of all the tribes followed him in returning their allegiance to Us. Already We have ordered that he be rewarded according to the circumstances. Moreover We stated, ‘Since he will bring peace, Kuo-che is to be enfeoffed as king.’ We did this not simply to reward his achievement, but to satisfy the wishes of the people, relying upon him to bring them comfort. We are not aware that of late [Li Kuo-che] has committed any act of maladministration, yet it comes to Our ears that he has been killed, although innocent has been flogged, and again most of the people are uneasy, so that the Mandate has been infringed. Now your barbarian laws have many injustices with regard to rulers. From of old that has been so, as We are aware. But if your kings, as soon as they commit an evil act, are straightway39 killed, is not the position of king fraught with danger? Indeed We fear that if you now become king, your successors will always be insecure and [then] who will wish to be king? Although you are a barbarian, you are a great warrior of your land and you ought likewise to take thought for the future. How can you [merely] select [the course which will] satisfy your immediate ambitions! Now that Kuo-che is dead, you should first take power as Governor and arrange matters for the common people and so restore peace. Otherwise Chang Shou-kuei will forthwith be sent to your state to settle matters differently. You should be given an official position and rewarded and then matters will be settled. [Written in] the middle of summer, at the time of extreme heat. Wishing you, your officers and your people peace. Now We present you with a set of

38  The character lieh (烈) is supplied from CTW. Chü-lo (據埒) in Passage (b) and Chü-lieh are clearly alternative transcriptions of the same name. 39   Ching (徑). WYYH gives ch’ing (輕) ‘for little reason’. Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

516

Herbert

brocade clothes and seven narrow waist40 belts. If matters arise, you should deal with them. This letter contains but little. Passage (d) Memorial Advocating that the Army of the North-east May Not Conduct an Insufficiently Planned Campaign (By Chang Chiu-ling) CCC (SPTK edition) 13, pp. 9a–10a. See also: CTW 288, pp. 15b–16b. We have received the edict circulated by Kao Li-shih [and understand that] the Turkish captives whom Chang Shou-kuei sent have submitted to Your Majesty’s hearing detailed information and also that Liu Ssu-hsien41 is ordered to leave [for the border]. We consider that the northern barbarians are evil and crafty and likewise it is truly dangerous to rely upon their good faith. However, Your Majesty has extended benevolence towards them and felt kindly disposed towards them for many a long year, so that they who have a ferocious and warlike nature have taken a turn away from violence. Examining the matter, it would seem that the trust put in their word has not been unfounded. Why [do I say] this? The prisoners whom Li Chüan42 recently sent back likewise all said that the East submitted to the Centre (the eastern barbarians submitted to China). Some people said that [the northern barbarians] could not be trusted, but right up until now [the situation] has been as [the captives] stated. They showed sincerity towards the state and have not deceived us in any other way. ‘Moreover the Khitan are unpredictable, some against us and some for us. Recently they attacked [the Turks] from the east. Although it was without our orders, as regards relations among the barbarians, they could not be faulted for upsetting a constant order. If we were to follow up their victory by massacring [the Turks], a source of anxiety [for our state] would likewise be destroyed.’ Formerly Your Majesty’s august opinion was along these lines. We often thought over this [matter] and this was certainly not the conclusion which we 40   W YYH omits the character yao (腰) ‘waist’. 41  Unidentified. 42  Unidentified.

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

Foreign Policy

517

drew. Now the Khitan are our allies, and if it should happen that, being victorious, they took advantage of the decline in power of the northern barbarians to destroy the greater half of them, Your Majesty’s plan would be bound to be adopted, that the matter might properly be decided. If however, the situation was not thus and the generals launched an ill-considered campaign, it would only widen the rift [between barbarians and Chinese] and we too would be considered untrustworthy. No greater cause for anxiety for the state could be created than this. I consider that all the generals serving on the border, seeing in their army a source of profit, launch campaigns in order to seek glory and rewards, and give no thought to long-term strategy. This is truly very harmful [to the military policy of the state]. Now Liu Ssu-hsien is setting out in the hope of obtaining letters of submission. [Chang] Shou-kuei should be instructed to make long-term plans and not launch an ill-considered campaign. This should be impressed upon the generals that they may know Your Majesty’s mind. Should it come to the ears of [those who hold power in] the barbarian courts, it will illuminate further Your Majesty’s all-encompassing heavenly mercy. I do not know whether or not this [policy] may be adopted [but] I respectfully set out this memorial for Your Majesty to hear.

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

The An Lu-shan Rebellion and the Origins of Chronic Militarism in Late T’ang China Edwin G. Pulleyblank One* of the common myths about traditional China is that it was a pacifist, or, at least, a very pacific country. Like many such myths it is an exaggerated truth, which has a basis in reality but cannot be taken at face value without grossly distorting the facts of history. Like all states, the Chinese empire was created and maintained by force of arms. This is not to agree with those who portray the civil side of Chinese government and the civilian ideology surrounding it as a mere façade to cloak the brutal realities of power, for the ideal of rule by moral suasion runs deep in traditional China. Yet wen and wu were two sides of the same coin. Even Confucius included arms among the three requirements of government, though less essential than the people’s trust and a sufficiency of food.1 That military arts had a postive role, even from a Confucian point of view, is well expressed in the traditional explanation of the graph of the Chinese character wu 武 as made up of ‘stop’ plus ‘halberd’, implying that the arts of war were for the sake of establishing peace. This saying is attributed to a Viscount of Ch’u who, having won a great victory over the state of Chin, refused to make a mound of corpses of his fallen enemies in order to glorify his victory to later generations. He went on to say that wu, i.e., military virtue, was that which suppressed violence, gathered in arms, protected what was great, established merit, gave peace to the people, harmonized the masses, and propagated wealth. He modestly declined to attribute such virtue to himself.2 The view Source: “The An Lu-shan Rebellion and the Origins of Chronic Militarism in Late T’ang China,” in J. C. Perry and B. L. Smith (eds.), Essays on T’ang Society: The Interplay of Social and Political and Economic Forces, Leiden: Brill, 1976, 33–60. * Detailed references are not provided for all the events of the rebellion which are narrated in chronological order in the Tzu chih t’ung chien. References to the Tzu-chih t’ung-chien (TCTC) are to the punctuated edition published by the Ku chi ch’u pan she, Peking 1956. Page references to the Chiu T’ang shu (CTS) and Hsin T’ang-shu (HTS) are to the edition of the Ssu pu pei yao. 1  Lun-yü 12/7. Cf. A. Waley, The Analects of Confucius, p. 164. 2  Tso-chuan, Duke Hsüan 12th year. Cf. J. Legge, The Ch’un-ts’ew with the Tso-chuan (Vol. V of The Chinese Classics, reprinted by Hong Kong University Press, Hong Kong, 1960), p. 320.

© koninklijke brill nv, leiden, ���9 | doi:��.��63/9789004380158_018

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

The An Lu-shan Rebellion And The Origins Of Chronic Militarism

519

of the positive role of war, expressed in this passage, is very much like that of the modern sociologist Robert Park who defined its social function, in part, as: (1) to extend the area over which it is possible to maintain peace, (2) to create and organize within that area a political power capable of enforcing it.3 Unfortunately, as the Chinese were well aware, war is also a hsiung ch’i, an ill-omened instrument. Few Chinese rulers, any more than those in other countries, showed the kind of modest and rational attitude attributed to the Viscount of Ch’u. Motives of personal aggrandisement and national glory were as besetting in ancient China as elsewhere; and, the cruel realities of slaughter and rapine were all too frequently more in evidence, as the products of wu, than the ideals he upheld. While Confucius and Mencius were exalting the civil and literary virtues over the military, making them the basis of an ideology that was to dominate Chinese thinking for the next 2,000 years, the Chinese states were engaged in constant warfare which grew increasingly bloody and bitter as smaller states were swallowed up by larger ones and as the remaining handful of great states sought to wipe each other out. In the end, as we all know, China was united under the Ch’in and a new era, that of imperial China, was inaugurated. The ultimate result of this prolonged series of wars could well be used to illustrate Robert Park’s definition. The extension of the area within which peace could be maintained and the establishment of a government to maintain peace in that area were the positive achievements of Ch’in, renewed by other dynastic founders that followed. While Chinese culture undoubtedly lost some of its variety and spontaneity by this unification, there was clearly a great stimulus to the economy with consequent increases in wealth, greater sophistication, and a widening of horizons that led to new intellectual and artistic achievements. As the Chinese of the time could well understand, the choice was not between imperial unity and an idyllic, primitive, pacific anarchy but between imperial unity and a multi-state system in a state of chronic war. The creation of the empire did not, of course, mean the end of war, but, apart from lesser or greater breakdowns in the imperial peace that allowed war to erupt once more within the Chinese-speaking area itself, the frontiers now became the main foci of military activities. The T’ang dynasty was heir to the Ch’in unification and the imperial system founded by Ch’in and Han. Of all the major native dynasties, with the possible but doubtful exception of the Han, it was the one in which wu, the military side of the coin, was most prominent and played the biggest role. This is true 3  Robert Park, ‘The social function of war: observations and notes’, American Journal of Sociology, 46 (1941), 551–570.

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

520

Pulleyblank

in both halves of the dynasty, but in very different ways. During the first half T’ang militarism was expressed chiefly in wars beyond the Chinese frontier, in the subjugation of the nomads in Mongolia and Manchuria, in the conquest of Central Asia (reaching at its greatest extent across the Pamirs into what is now Russian Turkestan and Afghanistan), in the exhausting but ultimately sterile conquest of Kao-kou-li, the state which then occupied North Korea, and in wars in Ch’inghai and other western and northwestern areas against the kingdom of Tibet, newly emerging as a power. In the second half, resulting from a rebellion by one of the major frontier generals, the situation was drastically altered. Frontier defence, particularly against Tibet, continued to be important, but it was now no longer in distant Ch’inghai but in eastern Kansu, right on the edge of the Wei valley on the doorstep of the capital. The biggest difference however, was that the interior of the country was militarized. That is, it was garrisoned by professional armies that usurped many of the normal functions of government and, while nominally owing allegiance to the T’ang, were constantly unruly, sometimes virtually independent. Reaction against the chronic militarism of the post An Lu-shan period was undoubtedly a major factor in the developing antimilitarist sentiments and policies in China from the Sung dynasty onwards. To understand T’ang militarism we must first consider the general military problem that faced Ch’in and all its successor empires once China itself had been united and, as a consequence, no more independent sovereign Chinese states existed to combat one another. There were, of course, still frontiers with the inevitable conflicts that have always arisen between peoples on frontiers. Consequently, there was also the continuing tendency to resolve the problem by extending territory to incorporate the source of the conflict into a larger whole. Ch’in and its successor Han both sought to extend the Pax Sinica beyond the territory occupied by Chinese-speaking peoples. The ease and success with which this was carried out varied greatly, however, in different directions. To the south, the incorporation of non-Chinese speaking peoples who were generally at a less sophisticated cultural level and less organized politically than the Chinese, but who were settled and agricultural like the Chinese, continued a process that had been going on for centuries within the heartland of China itself. Expansion in that direction had not always been pacific but, on the whole, had presented less of a military problem than the other type of frontier to the north and northwest where the Chinese came into contact with the nomadic, herding peoples of the steppe. In this direction, the Chinese found they could no longer expand by incorporation, since the steppe economy and way of life were so different from their own that Chinese social and political organization were quite inappropriate. At the same time, the steppe nomads were masters

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

The An Lu-shan Rebellion And The Origins Of Chronic Militarism

521

of a type of warfare, mounted archery, which made them far more formidable than their mere numbers would have warranted. The clash between these two modes of existence along the long border between steppe and sown in East Asia provided much of the dynamics of political change in that part of the world for many centuries.4 The first explicit evidence for Chinese contact with horse-riding nomads from the steppe is towards the end of the fourth century BC.5 This contact quickly wrought profound changes in Chinese methods of warfare and in other matters such as clothing. It also began the long search for a solution to the new frontier problem which it created. Most of the solutions that were tried can be put under three headings, each with a number of subtypes: (a) attempts to establish a permanent boundary which would keep each of the two sides from encroaching on the other; (b) attempts by the Chinese to exterminate, subjugate or at least subdue the nomads; and, (c) attempts by the nomads to do the same to the Chinese. A fourth type of solution was also conceivable, at least in theory, i.e., to incorporate Chinese and nomads into a true world empire, each having an equal share and neither dominating the other. T’ang T’ai-tsung was, I believe, the only Chinese ruler who had something approaching this grandiose design. In practice, his imperialism no doubt belongs under type (b), that is, as Emperor of China, he attempted to subjugate the nomads, as well as other peoples of Central and East Asia such as the oasis city dwellers of the Tarim basin and the Koreans. On the other hand, as is well known, T’ang had its roots both as a state and as a dynasty in Western Wei-Northern Chou, the last of the nonChinese dynasties of the period of division. Western Wei-Northern Chou, as part of its propaganda to be regarded as the legitimate heir of ancient Chou, had adopted a deliberate policy of racial and culture intermingling which undoubtedly contributed greatly to its 4  The classic treatment of the role of the steppe frontier in Chinese history is still that of Owen Lattimore in Inner Asian Frontiers of China, 2nd ed., American Geographical Society, New York, 1951. Despite errors in detail, his insights into the dynamics of interaction between the two ways of life retain their permanent value. 5  I refer to the famous debate at the court of Chao in 307 BC about the adoption of hu clothing, that is, riding breeches and other gear associated with horse-riding. (Shih chi 43, cf. E. Chavannes, Les mémoires historiques de Se-ma Ts’ien, Paris, 1905, v. 5, 69 ff. It has been argued that the Hsien-yün, whose battles with the Chou dynasty in the 8th century BC are referred to in the Book of Odes as well as on some contemporary inscriptions, must have been horse-riding nomads, but there is no proof of this and it is hard to believe that it would have taken another four or five centuries for such a revolutionary method of warfare to be imitated or even explicitly mentioned by the Chinese.

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

522

Pulleyblank

own success, and that of its successors Sui and T’ang in the struggle for supremacy over other parts of China as they existed at that time. As a result of this policy, the T’ang royal family, though Chinese in the male line, had a good deal of Hsien-pi blood in the female line. Moreover, they were still not very far removed from nomadic traditions in T’ai-tsung’s own reign, as is shown, for example, by the atavistic predilections for the Turkish life-style of T’ai-tsung’s eldest son.6 T’ang T’ai-tsung’s conquests of the Turks and of the city states in Central Asia were achieved with great rapidity, and were more extensive and thorough than those of the Han. They reflect, of course, the military effectiveness of the T’ang armies based on the fu-ping system and the warlike tradition inherited from Western Wei-Northern Chou. At the same time, they would hardly have been possible if he had relied on Chinese arms alone. To a considerable degree, they were made possible by alliances with various nomadic groups hostile to the Turks, especially the Nine Surnames or Toquzoghuz, of whom the leading tribe were the Uighurs. We may, in fact, think of the T’ang ruler not as seeking to impose Chinese sovereignty outward from the Middle Kingdom, as the Han emperors had sought to do, but rather as seeking to substitute himself for the Turkish qaghan as lord of the steppe. Where the Hsiung-nu ruler Mao-tun and Han Kao-tsu had agreed to divide the world into two halves, leaving the southern agricultural world for the Han emperor and reserving the lordship over those who draw the bow for the Hsiung-nu,7 T’ai-tsung wanted to look both ways and combine the two dignities in his own person. This aspiration of T’ai-tsung was symbolized in the title T’ien k’o-han ‘Heavenly Qaghan’ which was proferred to him in 630 by the chieftains of the various nomadic tribes.8 Though we may hold legitimate doubts as to whether the offering of this title was wholly spontaneous, the fact that it was adopted by the Chinese emperor was symbolically very important. It established a separate basis of legitimacy for his rule beyond the Great Wall, with its roots in nomad traditions, and was not simply an extension of universalist claims by the Chinese Son of Heaven. Moreover, it had as its corollary the assumption, quite contrary to Chinese traditional attitudes, of the equality of barbarian and

6  On the family background of the T’ang royal family see Ch’en Yin-k’o, T’ang-tai cheng-chih shih shu-lun kao, Chungking 1944 and Shanghai 1947. 7  Shih-chi 110. Cf. B. Watson, Records of the Grand Historian, Columbia University Press, New York and London 1961, Vol. II, pp. 166, 168, 173 ff. 8  T CTC 193, p. 6073, Chen-kuan 4 (630) 3/ wu-ch’en. On the history of this title see Lo Hsiang-lin, ‘T’ang-tai t’ien k’o-han chih-tu k’ao’, Hsin-ya hsüeh-pao, 1 (1957), 209–243.

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

The An Lu-shan Rebellion And The Origins Of Chronic Militarism

523

Chinese as subjects. This was a point of view consciously maintained and expressed by T’ai-tsung. Such an ideal of a Son of Heaven who looked both ways, ruling impartially over the lands on both sides of the frontier between the steppe and the sown, if it could have been maintained and developed, would have held out the prospect of a tremendous ‘extension of the area in which peace could be maintained’ in Park’s phrase. This was, of course, impossible to continue for very long in the East Asian world of the seventh century AD. The incompatibility of the two ways of life was too great for any consciousness of identity to be felt except by a small elite group. Moreover, the economic and cultural preponderance of China was so great that even the ruling aristocracy was inexorably pulled away from the lingering remnants of its nomadic traditions. Nevertheless, there was sufficient reality in the ideal during the early years of the T’ang for it to leave lasting effects, particularly on T’ang military traditions. I shall not go into detail about the fu-ping militia system or other aspects of the military organization of early T’ang and how it changed into the system of great standing armies on the frontiers that existed in the latter part of Hsüan-tsung’s reign, out of which An Lu-shan’s rebellion arose.9 I should like, however, to draw attention to certain features of this development which are connected to what has been said about T’ai-tsung’s imperialism and which are relevant to the way in which the new military situation of the post-rebellion period emerged. For one thing, this conception of the T’ang as in some degree a joint Sinobarbarian polity helped to provide a justification for the employment of nomadic tribesmen, not merely as allies by soldiers incorporated into T’ang armies. From the beginning of the dynasty there were cases of Turkish chieftains, such as A-shih-na She-erh and Ch’i-pi Ho-li, who surrendered to T’ang and became generals in the T’ang armies.10 They brought their own forces with them and served in the wars in Central Asia or Korea under Chinese commanders along with units drawn from the fu-ping militia system or otherwise recruited from within China. The same practice continued under Hsüan-tsung, only by that time the fu-ping system was in decay and it was becoming increasingly difficult to get Chinese recruits for the large semi-permanent frontier armies which had grown out of the older garrison and expeditionary army system as a result of the protracted wars against the Tibetans and the renewed Turkish empire of

9  See E. G. Pulleyblank, The Background of the Rebellion of An Lu-shan, Oxford University Press, London, 1955, chapter 5. 10  CTS 109, HTS 110.

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

524

Pulleyblank

Kutlug and Mo-ch’o. With the formal abolition in 737 of the practice of sending militiamen and other conscripts back and forth on tours of duty and the enlistment instead of permanent frontier armies the role of non-Chinese elements in the Chinese defence forces became more firmly entrenched. There is no way of telling what proportion of those who composed the frontier armies in 737 were of Chinese origin and what proportion were non-Chinese, but the latter were certainly very numerous. At first, the overall command in these armies was kept in Chinese hands, but in 742 An Lu-shan was given an independent command at P’ing-lu in southern Manchuria and in 747 Li Lin-fu persuaded the emperor to adopt the deliberate policy of using non-Chinese as Military Governors on the grounds that they were better soldiers and had no political ambitions or factional connections at court. It was this disastrous policy more than anything else that set the stage for the rebellion. By 755, An Lu-shan held three military provinces out of ten; his cousin An Ssu-shun held the important Shuo-fang province immediately to the north of the capital; Ko-shu Han, also half Turk half Iranian, held Lung-yu and Ho-hsi; and the two westernmost provinces of Pei-t’ing and An-hsi were also under the command of non-Chinese officers. All these armies, together having a nominal strength of some 500,000 men in 742, contained large elements from Turkish, Khitan and other nomadic tribes that had surrendered to T’ang. And, there is good reason to think that the Chinese in these armies, who had now become permanent frontier soldiers and lost their original attachments to their villages inside the country, were also more or less ‘barbarized’, that is, they had become deeply imbued with the military ethos of their barbarian comrades in arms. Of course, the barbarians themselves were no longer in their pristine tribal state. The mere fact that they had ceased to gain their livelihood by tending herds and had become dependent on rations provided by the Chinese government must have made a big change. By 750, what remained of T’ai-tsung’s combined nomad and Chinese empire? No longer equal partnership (if this had ever really existed) presided over by an emperor with one foot in each camp and an impartial regard for the interests of both. There was still a kind of partnership, but it was based upon a division of roles in which the barbarians acted, to use a common Chinese metaphor, as the ‘claws and teeth’, while the Chinese were the ‘heart and belly’, i.e., the directing mind and the main economic beneficiaries of protection. Unfortunately, the ‘claws and teeth’ showed that they had minds of their own and, instead of protecting the Chinese, decided to move in and try to appropriate or at least share in the spoils. Hence the cancerous growth of militarism within the country which, as the Hsin T’ang-shu says, ended by destroying the dynasty, through the institutions which it had itself set up. The whole story of

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

The An Lu-shan Rebellion And The Origins Of Chronic Militarism

525

how this came about cannot be told here, but let me begin by summarizing the events of the An Lu-shan rebellion itself. Announcing as his pretext that he had received a secret edict from the T’ang emperor to come and get rid of the Chief Minister, Yang Kuo-chung, who was his arch-rival, An Lu-shan began his march south from Fan-yang (present Peking) on the auspicious chia-tzu day in the 11th month of the 14th year of T’ien-pao (December 16, 755). After crossing the Yellow River on an improvised bridge at a point near present Hua-hsien, he entered Lo-yang on January 18, 756, 33 days after setting out. There had only been slight resistance from hastily mustered local troops that were no match for An Lu-shan’s armoured cavalry. It is a moot question whether, if he had pressed on immediately, he could also have penetrated the T’ung-kuan and taken Ch’ang-an, but he stopped to rest his troops and consolidate his political position. He abandoned his pretense of having arisen at the behest of the T’ang emperor and on the first day of the new year (February 5, 756) proclaimed himself emperor of the Great Yen Dynasty. By the time he was ready to advance on the T’ung-kuan he found it defended by Ko-shu Han with an army comprised of Chinese and barbarian troops withdrawn from the northwestern frontier. During the next half year, forces loyal to T’ang began to counterattack and had considerable success in hemming in the rebels, preventing them from expanding their territory, cutting their line of communications from Fan-yang to Lo-yang, and threatening their base at Fan-yang. Even before the proclamation of the new dynasty some of the cities in Hopei that had submitted to An Lu-shan during his first advance had begun to rise against him. Soon, Yen Chen-ch’ing at P’ing-yuan, who before the rebellion broke out had been in secret contact with Yang Kuo-chung and had been making defensive preparations in anticipation of such an eventuality, was at the head of an alliance of T’ang loyalists in Hopei itself.11 Around the southern perimeter of the territory held by the rebels hastily organized defences were stopping any advance 11  The ‘Yen Lu-kung Chen-ch’ing hsing chuang’ by Yin Liang, which still survives (see Ch’üan T’ang-wen 514. 9a), was the primary source for most of what is contained in his biographies in CTS 128 and HTS 153. It preserves many details not found in the official biographies. Though Yen Chen-ch’ing’s appointment as prefect of P’ing-yuan in 753 is attributed to Yang Kuo-chung’s jealous wish to get him away from Ch’ang-an, it seems clear that Yen’s secret preparations against An Lu-shan were done with the knowledge of the Chief Minister. Chien Ang, whom Yen Chen-ch’ing used as a confidential messenger to Ch’angan, was also used by Yang Kuo-chung as a secret agent to spy on An Lu-shan’s house at Ch’ang-an. (See CTS 106. 7a, biography of Yang Kuo-chung, and TCTC 217, pp. 6932–3, T’ien-pao 14 (755) 4/-, with k’ao-i; also E. G. Pulleyblank, ‘An Rokuzan no hanran no seiji teki haikei’, Tōyō gakuhō, 53 (1953), p. 142.).

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

526

Pulleyblank

towards the rich, productive regions of the Yangtze basin. By the second month Li Kuang-pi had come from the Shuo-fang Military Province through Chingching pass into Hopei and was engaging An Lu-shan’s general, Shih Ssu-ming, at Ch’ang-shan. Before long, Li Kuang-pi was joined by the Military Governor of Shuo-fang, Kuo Tzu-i, with a large force and the tide of battle turned definitely against the rebels. At the battle Chia-shan, Li and Kuo severely defeated Shih Ssu-ming and forced him to take refuge in Po-ling (present Pao-ting) where they besieged him. The position of the rebels looked grave. Unfortunately for T’ang, the advantageous situation was thrown away by blunders at Ch’ang-an and at the T’ung-kuan. Since the beginning of the year, Ko-shu Han’s army of some 200,000 men had been sitting idle at the T’ung-kuan. Neither the situation in the army itself, nor the political situation behind it in the capital was conducive to good morale. Ko-shu Han was ill and his deputy T’ien Liang-ch’iu was lacking in authority and left decisions to the commander of foot and the commander of horse who were at odds with one another. Discipline for the ordinary soldiers was strict, rations were meagre, and the heat of the summer was intense. Meanwhile, the officers idled away their time with dancing girls brought from Ch’ang-an.12 What had once been a powerful army of veterans was visibly deteriorating. At the same time, Yang Kuo-chung was becoming more and more nervous at the presence of such a huge military force on his doorstep but not under his direct control. He knew that he was widely blamed for An Lu-shan’s rebellion and was afraid of a coup directed against himself. Some sources say that Ko-shu Han’s subordinates were, in fact, trying to persuade the latter to send a force to Ch’ang-an to oust Yang Kuo-chung.13 Even without such special reasons it is probable that there would have been increasing pressure from the Emperor and the civilian administration on Koshu Han, as the summer wore on, to abandon his defensive posture and advance to retake Lo-yang. Spy reports arrived saying that the passes were lightly held and could easily be assaulted. Ko-shu Han did not believe that an experienced soldier like An Lu-shan would leave such strategic defenses unmanned and suspected a trap, but Yang Kuo-chung prevailed on the Emperor to order an advance in spite of these objections. The result was a complete disaster. The whole army was lost and the way to Ch’ang-an lay unguarded. This occurred in the 6th month of 756. 12  The condition of Ko-shu Han’s army at the T’ung-kuan is vividly described in the words of Kao Shih, who reported it to Hsüan-tsung on his way into exile in Szechwan. (CTS 111.7b, biography of Kao Shih.). 13  TCTC 218, pp. 6965–6, Chih-te 1 (756) 6/ kuei-wei.

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

The An Lu-shan Rebellion And The Origins Of Chronic Militarism

527

The subsequent flight of the emperor, the meeting of his guard at Ma-wei, the assassination of Yang Kuo-chung, and the strangling of Yang Kuei-fei to appease the troops are the most dramatic and celebrated events of the rebellion. After this sacrifice of his favourite, the emperor went on to Szechwan and the Crown Prince went to Ling-wu where he proclaimed himself emperor and gathered his forces to restore the dynasty. The fall of the T’ung-kuan meant not only the loss of the capital but also the collapse of the T’ang counter offensive in Hopei, since the Shuo-fang armies were withdrawn to join Su-tsung. The cities which had been restored to T’ang allegiance fell one by one again into rebel hands. Fortunately for T’ang, the ability of the rebels to profit from their victory was seriously reduced by the illness of the chief rebel An Lu-shan. An Lu-shan’s extraordinary obesity is referred to in many well known anecdotes. He also suffered from boils and skin ulcers and, by the time he became rebel emperor, he was losing his eyesight. It is probable that he suffered from diabetes.14 As a result of his disease he became increasingly irritable and his attendants and ministers went in terror of being flogged, even executed, for small offences. In the end they murdered him, placing his son An Ch’ing-hsü on the throne. Naturally, this weakened the control of the rebel court over the military commanders and detracted from the rebel determination and unity of purpose. Nevertheless, the T’ang recovery took a long time and much costly effort to accomplish. It was well over a year after Su-tsung proclaimed himself emperor that first Ch’ang-an and then Lo-yang were retaken. Moreover, the key to victory was the Uighurs’ participation on the T’ang side, for, while he was gathering his forces at Ling-wu, Su-tsung not only recalled the Chinese armies from Central Asia but also solicited aid from states that had been under Chinese control or, at least, suzerainty. Contingents arrived from Khotan and Ferghana, even from the Arabs who had until recently been the bitter rivals of the Chinese for the overlordship of the trans-Pamirian countries of what is now Russian Turkestan. The presence of these foreign troops in China is interesting, but there is hardly any further mention of them in relation to subsequent events and no reason to think that they played a decisive role. The cooperation of the Uighurs was another matter and great pains were taken to obtain it. A prince of the Li clan was sent as an ambassador along with P’u-ku Huai-en, a general of Uighur extraction serving under Li Kuangpi in the Shuo-fang army. Reversing the usual arrangement, the T’ang prince received a daughter of the Uighur qaghan as his bride. The Uighurs’ first 14  I owe this opinion to Dr. Lu Gwei-djen, who kindly went over the references in the sources to An Lu-shan’s illness for me.

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

528

Pulleyblank

assistance to T’ang was in dealing with some other Turkish tribesmen who had been attached to An Lu-shan’s forces as far as Ch’ang-an but who, for reasons that are obscure, had deserted him there and moved north into the Ordos. In the 7th month of 757 the Qaghan’s son and heir to the throne arrived at Su-tsung’s camp with 4,000 horsemen. Though they were numerically only a small part of the forces assembled by T’ang to attack the rebels in Ch’ang-an, they played a decisive part in the battle, as they were to do on more than one later occasion. The debt to the Uighurs which the T’ang thus incurred was an onerous one in the years that followed, but without their help the outcome of the war might well have been different. The recovery of the two capitals at the end of 757 was looked upon at the time as the symbolic end to the rebellion. The rejoicing as the Emperor and the Retired Emperor returned to their palaces was premature, however, for another five years of fighting lay ahead before the last rebel emperor was killed and the whole empire returned to at least nominal allegiance to T’ang. The failure of the T’ang armies to make an immediate and resolute pursuit of An Ch’ing-hsü after he fled from Lo-yang is at first sight hard to understand. If we can get to the bottom of it, we shall probably have the key to much of the subsequent history. Leaving the reasons aside for the moment, it should be noted that An Ch’ing-hsü was allowed to escape and gather his forces together at Yeh (An-yang) in southern Hopei and it was not until the autumn of 758 that an expeditionary force was mustered against him. Meanwhile, at Fan-yang, Shih Ssu-ming had first transferred his allegiance to T’ang, then revolted again when he discovered envoys from the T’ang court intriguing against him with some of his subordinates. The T’ang expedition against An Ch’ing-hsü was not a unified force under a single command but consisted of armies of nine separate Military Governors coordinated only by a eunuch supervisor. This was symptomatic of deteriorating T’ang control over its armed forces, which was one of the lasting consequences of the rebellion and about which more will be said below. The ill effects of this divided command soon made themselves apparent. The T’ang armies surrounded Yeh with a double circumvallation and triple moat, diverted waters so as to flood the city, and prepared to starve its inhabitants into surrender. Meanwhile, however, Shih Ssu-ming arrived on the scene. After some preliminary skirmishing the two armies gave battle. Shih Ssu-ming was far outnumbered by the combined T’ang forces, but the result of the battle was not in fact determined by clash of arms. Shortly after the fighting started, a violent windstorm suddenly arose, raising so much dust that it was impossible to see. The two sides fled in opposite directions, but, whereas Shih Ssu-ming’s force only went a short distance before reforming, the allies quit the field

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

The An Lu-shan Rebellion And The Origins Of Chronic Militarism

529

entirely scattering in different directions. For the most part, they made their way back to their various provincial headquarters pillaging and looting as they went. Shih Ssu-ming was left in possession of the field, having benefited by a freak of nature but even more by the indiscipline and lack of leadership on the T’ang side. Shih Ssu-ming now proceeded to do away with An Ch’ing-hsü and make himself emperor of Yen. He consolidated his hold on Hopei but was unsuccessful in either pushing south into the Yangtze basin or westward towards Ch’ang-an. Li Kuang-pi, who had been the most successful general in maintaining discipline among his troops after the debacle at An-yang and was now put in charge of defending the approaches to the T’ung kuan, abandoned Lo-yang but defended Ho-yang on the north side of the Yellow River and with this threat from the flank prevented Shih Ssu-ming from attempting a westward advance. Though this was a successful defensive strategy, the emperor and his court eventually became impatient, as they had when Ko-shu Han had been sitting at the T’ung-kuan, and early in 761 (over a year after Li Kuangpi had established himself at Ho-yang) orders were sent to take the offensive against Shih Ssu-ming’s forces at Lo-yang. Battle lines were drawn up at Mangshan. Partly because of the insubordination of his chief lieutenant, P’u-ku Huai-en, Li Kuang-pi was disastrously defeated and once more it looked as if the way to Ch’ang-an might be opened. Unfortunately for the rebels, they too were suffering from dissensions. Shih Ssu-ming’s eldest son, Ch’ao-i, was able and courageous—indeed, he gives one of the most sympathetic impressions of all the rebel emperors—but he was the son of a concubine rather than the principal wife and the relationship between him and his father deteriorated as a result of Shih Ssu-ming’s desire to elevate a younger son, Shih Ch’ao-ch’ing, as his heir. A faction loyal to Shih Ch’ao-i murdered Shih Ssu-ming and made Shih Ch’ao-i emperor in his stead. This led to protracted disorders between the supporters of Shih Ch’ao-i and others at Fan-yang, also putting an end to the immediate menace of an advance on Ch’ang-an. In spite of this setback, Shih Ch’ao-i remained a formidable threat to T’ang for well over a year. In the autumn of 762 T’ang once more sent an embassy to seek assistance of the Uighurs, finding to their consternation that the latter were already preparing to respond to Shih Ch’ao-i’s solicitation for help to the other side. They were persuaded to remain loyal to their earlier alliance with T’ang, however, and sent a large contingent to join an assault on Lo-yang. After a hard fought battle Shih Ch’ao-i was defeated and Lo-yang was retaken. On the previous occasion, when the Uighurs had helped recapture Ch’ang-an and Lo-yang, they had been heavily bribed not to pillage, but this time they were given a free hand. According to the Tzu-chih t’ung-chien, the dead numbered in

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

530

Pulleyblank

tens of thousands and fires burned for several weeks. The T’ang armies joined in the killing and looting in Lo-yang and other cities nearby that had been in rebel hands. Shih Ch’ao-i fled to the northeast, as An Ch’ing-hsü had done earlier, but, unlike him, he was pursued and not given a chance to reestablish himself. After unsuccessful attempts to rally his forces from cities in southern Hopei and make a stand against his pursuers, he was deserted by the remaining rebel generals, who, one by one, offered submission to T’ang. When he finally reached Fan-yang, he found the gates closed against him. The sources differ as to the manner of his death, but the most circumstantial account, adopted by Ssu-ma Kuang, states that, despairing of escape from his pursuers, he hanged himself in a wood. Li Huai-hsien, who had been his governor of Fan-yang and had now gone over to T’ang, obtained his head and sent it to Ch’ang-an. The principal former rebel generals in Hopei who submitted to T’ang after Shih Ch’ao-i’s defeat were allowed to retain their posts. P’u-ku Huai-en, who commanded the pursuit of Shih Ch’ao-i, was later blamed for having permitted this to happen. Since he himself later rebelled, it was suggested that he had been guilty of treachery in recommending to the court that these men be given T’ang titles. It is likely, however, that the real explanation was, as the Tzu-chih t’ung-chien admits, that ‘the court was also sated and worn out by the war and hoped by any means to end its troubles in peace’. After all, exactly the same strategy had been employed in 757 when Shih Ssu-ming had been allowed to surrender to T’ang after An Ch’ing-hsü was driven out of Lo-yang and, since that time, other rebel generals had received offices upon surrendering opportunely to T’ang. The fact that the northeast remained in the hands of men who had once been rebels and who still held control over their own military forces meant that the reestablishment of T’ang authority was only nominal in those regions and a legacy of future trouble was left behind. This is only a brief summary of the central struggle between T’ang and An Lu-shan and his successors. Between 755 and 763 the war spread into a good many side events of greater or lesser magnitude. For example, there was the attempt by a younger son of Hsüan-tsung, Prince Lin, to establish an independent state in the lower Yangtze. Prince Lin was suppressed without much difficulty, but the episode is famous because the poet Li Po, perhaps innocently, was implicated in it.15 Hopei was disturbed by army mutinies at Hsiang-yang in 759 and 760. Though not much is known in detail about the nature of these revolts, they 15  This is described in A. Waley, The poetry and career of Li Po, Allen and Unwin, London, 1950.

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

The An Lu-shan Rebellion And The Origins Of Chronic Militarism

531

no doubt have a bearing on the subsequent fractiousness and independence of the Military Government of this region, about which more will be said presently. The lower Yangtze region was untouched by the war until 760 when Liu Chan, who had been a high ranking officer under the Military Governor of Huai-hsi, stationed at Sung-chou in Honan, was turned into a rebel by a bungled attempt to deprive him of his command, and proceeded to go south into those rich lands. The people of this area suffered even more from the ravages of the P’ing-lu army which was sent down to suppress Liu Chan and which in the process sacked the cities of Yang-chou and Ch’u-chou, killing and robbing thousands of Central Asian merchants who were congregated there. The same regions suffered again in 762 through efforts to collect back taxes for the war years when communications with the northwest had been cut. The harshness of the exactions led to widespread banditry and finally to a full scale popular revolt which was being put down by Li Kuang-pi about the same time that P’u-ku Huai-en was pursuing Shih Ch’ao-i and pacifying Hopei. Finally, one must mention the loss of territory in the west which resulted from the rebellion. The withdrawal of garrisons to fight the rebellion allowed the Tibetans to advance into Ch’inghai and southeastern Kansu. By the time the rebellion ended they were just west of Feng-hsiang, where they posed an immediate threat to the Wei valley and Ch’ang-an itself. In fact, they invaded the Wei valley and briefly occupied Ch’ang-an in 763, forcing the emperor to flee eastwards to Shan-chou and to consider seriously moving the capital permanently to Lo-yang. They were driven out and the emperor returned to Ch’ang-an, but the defence of the western and northwestern approaches to the capital against the Tibetans and the still tribal Tangut peoples, who inhabited the mountainous regions to the southwest, remained a chronic problem for years. In southern Szechwan, also, outlying garrisons were abandoned, in this case to the kingdom of Nan-chao. Nan-chao did not become a military threat to China itself, however, until the middle of the ninth century. After a war lasting over seven years which devastated much of the Yellow River plain and also extended, for shorter periods, into Kuan-chung and the lower Yangtze region, it is not surprising that there should have been lasting effects. We cannot, however, blame the chronic militarism of the next two centuries simply on the war. The civil wars at the beginning of the T’ang had been even more prolonged and widespread, yet the consequences were very different. The founding of T’ang was followed within a few years by the total pacification of the country, demobilization of most of the armies, and the establishment of lasting internal peace. Military activity continued, but it was directed against the Turks and other enemies beyond the frontier.

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

532

Pulleyblank

The most independent of all the Military Governors in the post An Lu-shan period were, of course, the generals who were allowed after surrender to remain in command of their armies and to govern large tracts of land as T’ang military governors.16 The whole of Hopei province was divided between four of them, from north to south: 1) Li Huai-hsien, Military Governor of the Lu-lung army with headquarters at Yu-chou (Peking); 2) Chang Chung-chin, whose name was changed to Li Pao-ch’en, Military Governor of the Ch’eng-te Army, with headquarters at Heng-chou (present Chen-ting); 3) T’ien Ch’eng-ssu, Military Governor of Wei-po, with headquarters at Wei-chou (present Ta-ming); and, 4) Hsüeh Sung, Military Governor of Hsiang-wei, with headquarters at Hsiangchou (present An-yang). In 775 the territory of Hsiang-wei was annexed by T’ien Ch’eng-ssu of Wei-po, after which the three remaining territories were known as the ‘Three Garrisons of Ho-pei’. Though they accepted T’ang titles and owed nominal allegiance to the imperial government in Ch’ang-an, these former rebel generals governed their territories as virtually independent fiefdoms. They appointed their own officials, raised their own armies, collected their own taxes, and endeavoured to establish family dynasties and marriage alliances with one another. The T’ang court sometimes tried to intervene and exert its authority when a governor died or was driven out by his soldiers and the son, or some other successor, put up by the troops and requested the emperor to legitimize his authority, but its successes were limited even on such occasions. Generally speaking, the most the T’ang could achieve was to play off one governor against another and try to prevent them from combining forces to intervene in affairs outside their own province. Early in Te-tsung’s reign, after 781, T’ang managed to split off two new governments from Lu-lung and Ch’eng-te, namely that of I-wu at Ting-chou (modern Pao-ting) and that of I-ch’ang at Ts’ang-chou (near modern Ts’ang-hsien), but they were only slightly more amenable to T’ang control than the other Hopei garrisons. For a short time in the Yüan-ho period (806–820), after the emperor Hsientsung had crushed the independent governors in Honan, the Hopei armies accepted court appointees, but soon these were driven out one after the other and the situation largely reverted to what it had been before, remaining so to the end of the dynasty.

16  The history of the military governors after the An Lu-shan rebellion has been studied by Hino Kaisaburo in ‘Tōdai hanchin no bakko to chinshō’ Tōyō gakuhō, 26 (1939), 503–29; 27 (1939–40), 1–67, 153–212, 311–350. See also his Shina chūsei no gumbatsu, Tokyo, 1942. A more recent study is Wang Shou-nan, T’ang-tai fan-chen yü chung-yang kuan-hsi chih yen-chiu, Taipei, 1969.

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

The An Lu-shan Rebellion And The Origins Of Chronic Militarism

533

The semi-independence of Hopei after the An Lu-shan rebellion was not merely a matter of a few top generals. If it had been, it probably would not have lasted very long. It seems clear that it was based upon a conscious, strongly held, and broadly based separatist sentiment in the Hopei armies, whose attitude to An Lu-shan and Shih Ssu-ming and their abortive Yen dynasty was very different from that of the T’ang court. This is vividly illustrated by what happened when Chang Hung-ching, the first governor to be appointed from outside since the rebellion, arrived at Yuchou (modern Peking) in 820. ‘The people of Chi, old and young, male and female, all lined the road to watch. The army officers of North of the River were used to enduring heat and cold along with the ordinary soldiers and had no distinctions based on spreading umbrellas or resting in sedan chairs. Hungching had long been wealthy and noble and was unaware of the local customs. When he entered Yen, he was carried in a shoulder litter into the midst of the three armies. The People of Chi were much astonished at it. Because the rebellion of [An] Lu-shan and [Shih] Ssu-ming had begun at Yu-chou, Hung-ching wished to reform popular customs at the source, so he opened up Lu-shan’s tomb and destroyed his inner and outer coffins. The people became still more disaffected. Among his subordinates were Wei Yung and Chang Tsung-hou and several others who frivolously disported themselves with eating and drinking. They were drunk every night and returned home lit by torches filling the streets and bawling and cursing back and forth. The people of Chi were not used to that sort of thing. Moreover Yung and his friends scoffed at the clerks and soldiers, often reviling them as rebellious savages. They said to the soldiers, “Now the world is at peace. You fellows who can draw a bow with a pull of 20 shih are not equal to one who knows a single character”. The troops resented very much being lorded over in this way’.17 In such circumstances, the mutiny which followed is easily understandable. Similar attitudes were found in other parts of Hopei. The first Military Governor of Wei-po, T’ien Ch’eng-ssu, had once set up a temple to the four emperors of Yen—An Lu-shan, An Ch’ing-hsü, Shih Ssu-ming and Shih Ch’ao-i— calling them the Four Sages (Ssu sheng).18 Later, when another member of the T’ien clan who had had a Confucian education was being too subservient to T’ang, his troops demanded that he ‘put into practice the old ways of Ho-shuo’ and in the end deserted him for a military man of non-Chinese origin.19

17  CTS 129.8a, b (biography of Chang Hung-ching). 18  TCTC 224, p. 7222 Ta-li 8 (773) 10/ chia-ch’en. 19  CTS 141.10a (biography of T’ien Pu).

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

534

Pulleyblank

In 763 and the immediately following period a large part of Honan also was occupied by former rebels or by armies that had operated independently for so long that they were difficult to distinguish from the former rebels. In the extreme east was the large military government of P’ing-lu, centred on Ch’ingchou (I-tu), covering roughly what is now Shantung province. The P’ing-lu army originally was stationed in southern Manchuria and had been under An Lu-shan’s command at the outset of the rebellion. It had early declared its loyalty to T’ang, however, and had fought against the rebel forces stationed at Fan-yang. In 757 they sent a contingent across the Gulf of P’o-hai on rafts to the mouth of the Yellow River which fought on the loyalist side in Honan against An Ch’ing-hsü and Shih Ssu-ming. In 760 T’ien Shen-kung led this army south to suppress Liu Chan and, as already noted, was responsible for sacking Yangchou and other commercial cities of that region. In 761 the main force of the P’ing-lu army which had remained in Manchuria, being hard pressed by the rebels in Fan-yang and the Hsi tribes from beyond the frontier, fought their way south and joined up with their colleagues. Their commander, Hou Hsi-i, was confirmed as Military Governor at Ch’ing-chou but also retained the P’ing-lu army. In spite of its record of loyalty to T’ang and opposition to An Lu-shan and his successors, the P’ing-lu army had, because of its isolation, been effectively running its own affairs since the beginning of the rebellion. In fact, it had been the first army to choose its own commander when the Military Governor died in 758. This became a frequent occurrence in other armies in the post-rebellion period and the historian Ssu-ma Kuang blames the emperor for agreeing to it and so establishing a bad precedent.20 But, at that time, given the poor communications that existed, there was very little choice. Once established in eastern Honan (i.e., modern Shantung), the P’ing-lu military governors proceeded to behave just like the former rebels north of the river. Hou Hsi-i, the first of them, was driven out by his troops in 765. Li Chengchi, elevated in his stead, founded a dynasty that lasted for three generations. West of the territory of the P’ing-lu army itself were Ling-hu Chang at Huachou, a former rebel general who had surrendered already in 760, and T’ien Shen-kung at Pien-chou who, as we have seen, had originally been part of the P’ing-lu army. West of them was the Huai-hsi Military Governor Li Chung-ch’en (formerly Tung Ch’in), another P’ing-lu general. Though we know little in detail about the civil administration in the territory of the independent military governors, there are some interesting comments in Li Cheng-chi’s biography which help one to understand why they were able to maintain themselves so well. On the one hand, we are told, ‘His 20  TCTC 220, pp. 7064–6. Historian’s comment at the end of Ch’ien-yüan 1 (758).

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

The An Lu-shan Rebellion And The Origins Of Chronic Militarism

535

government was stern and cruel. Throughout the land no one dared to make unguarded remarks’. On the other hand, ‘He had altogether fifteen prefectures. Since he looked inward (from the sea) towards his fellow warlords, he set up a market for fine horses from P’o-hai (a kingdom in Eastern Manchuria) which continued year after year. His justice was even-handed and taxes were uniform and light. He was highly renowned for his greatness and power.’21 The seriousness of the loss of the northeast to the T’ang government as a source of revenue has been pointed out by Twitchett.22 The converse is that, when the taxes of the northeast were no longer drained off to Ch’ang-an, the burden on the local populace must have been greatly reduced. While the support of the regional armies and their commanders still had to come from the labour of the peasants, even before the rebellion there had been a huge standing army on the northern frontier looking for its supplies to the resources of the hinterland. The demands of the military governors for their own needs in the post-rebellion period were less onerous than the combined demands of the central government and the frontier armies had been before the rebellion. The relative lack of culture and sophistication under the military government at Yu-chou, compared with parts of the country under direct T’ang rule, has been noted above. This was probably typical of the northeast as a whole. While no doubt distressing to the Confucian gentry, this probably helped to lower the cost of government and relieve the burden upon the peasants. The people as a whole were probably quite willing to put up with the rigours of military rule for the sake of lighter taxation and there is little reason to think that they yearned for restoration of T’ang rule. After a series of campaigns in the Yüan-ho period (806–20) Honan was eventually restored to T’ang rule. That is, T’ang appointed military governors and took its share of taxes. The armies remained, however, becoming increasingly lawless and parasitic, so that, as far as the people were concerned, the burden of taxes was redoubled in return for little obvious benefit. It is probably no accident that, when peasant rebellions eventually broke out in the 860s and 70s, they started in Honan. On the other hand, as far as I have been able to discover, the autonomous provinces of Hopei remained relatively free of peasant unrest. In discussing the problem of militarism in the post An Lu-shan period, attention is usually focused on those armies whose commanders had originally come from the northeast frontier and had either served under the rebel emperors or been part of the P’ing-lu army. Generals who had come from other 21  CTS 124.6b. 22  D. C. Twitchett, ‘Provincial autonomy and central finance in Late T’ang’, Asia Major, 11 (1965), 211–32.

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

536

Pulleyblank

sectors and commanded the T’ang’s own forces, however, became almost as difficult to control. We can trace the beginnings of this back to early stages of the rebellion when generals from the northwest were sent to various strategic points in Honan as Military Governors of newly created Provinces, with concurrent civil authority as prefects of the cities which were their main headquarters. For instance, already at the beginning of 756 Lu Chiung, an officer who had served under Ko-shu Han in Lung-yu, was sent to Nan-yang as Prefect and Military Governor to take charge of troops freshly recruited from Ling-nan, Ch’ien-chung, and Hsiang-yang. He maintained a stubborn defence there for over a year before being forced to break out of the siege and retreat to Hsiang-yang.23 Another early appointment was that of Lai T’ien, who had been an officer under the Pei-t’ing Military Governor in Central Asia. At the end of 756 he was made prefect of Hsü-chou. Like Lu Chiung, he organized a heroic defence of the beleaguered town and prevented the rebels from breaking through to the south. He earned the nickname ‘Bite-iron’.24 Though I have found no explicit statement that Lu Chiung and Lai T’ien brought troops with them, there is some evidence that they were accompanied in each case by at least a nucleus of officers. After the retaking of Lo-yang at the end of 757, there was no immediate move to disband these and other armies that had been organized in Honan or to shift them back to the frontiers. Rather, there was consolidation of army bases foreshadowing the situation as it finally emerged in 763. I have already remarked on the fact that there was no hot pursuit of An Ch’ing-hsü, so that he was able to establish himself at Yeh (An-yang) and gather his forces. No doubt this can be partly attributed to a natural relaxation after the main symbolic objectives, the capitals, had been recovered. And certainly, the economic situation in Ch’ang-an was grave and there was dire need of time to recuperate. That in itself, however, hardly accounts for the fact that it was almost a year before a campaign was organized against An Ch’ing-hsü. Nor does it explain why, when the forces were mobilized, there were nine or ten separate commands with no overall authority except that of the eunuch Army Supervisor. The reason given for not appointing an overall commander was that Kuo Tzu-i and Li Kuangpi had shared equally in the T’ang recovery and that neither could be placed over the other. The difficulty of placing Li Kuang-pi, once Kuo Tzu-i’s subordinate but now an independent Military Governor, back under his former chief (with whom he was apparently not on the best of terms) was no doubt an awkward point, but it hardly explains why there had to be nine or ten separate 23  CTS 114.1a ff. Cf. HTS 147. 24  CTS 114.3a ff. Cf. HTS 144.

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

The An Lu-shan Rebellion And The Origins Of Chronic Militarism

537

commands. This arrangement surely reflects, on the one hand, the fact that the Military Governors had been busy consolidating their territorial bases which were coming to be looked on as separate fiefdoms and, on the other, that the court was becoming increasingly mistrustful of all its generals and did not wish to have too much power concentrated into the hands of any one of them, even the conspicuously loyal Kuo Tzu-i. It is difficult to get information on all the armies participating in the siege of An-yang, but we can see what went on from a few examples. Lu Chiung was now Military Governor of Huai-hsi with headquarters at Teng-chou. Li Ssu-yeh, who commanded the Expeditionary Force from the former Central Asian garrisons at Chen-hsi and Pei-t’ing, was stationed at Huai-chou, on the north side of the Yellow River east of Lo-yang. He was made Prefect of Huai-chou, in addition to his military command, which made it easier for him to requisition supplies and manpower for his army’s needs. Wang Ssu-li, who already held the title of Military Governor of Kuan-nei, was stationed at Lu-chou in southeast Ho-tung over which he was given the civil authority as well as being concurrently Military Governor of the region round about. Other commands were centered on Hua-chou, Pien-chou, and Cheng-chou. Li Kuang-pi, Military Governor of Ho-tung, was back at his headquarters at T’ai-yüan. Only Kuo Tzu-i, among the major commanders, whose Shuo-fang armies were encamped somewhere east of Lo-yang, did not receive concurrent civil authority in the region he occupied. Presumably, he relied directly on the resources of the Eastern Capital for provisioning his troops. The role that local inland bases for the armies were coming to play is strikingly illustrated by what happened after the disastrous fiasco at An-yang. The imperial armies made no attempt to reform and regroup as a whole, but each headed as fast as it could back to its provincial headquarters. In the later stages of the war the consolidation of territorially based armies and their increasing alienation and conflict with the central T’ang government went on apace. The history of the Military Government of Eastern Shan-nan with headquarters at Hsiang-yang is both a good illustration and important for the role it played in establishing precedents. Shortly after the An-yang debacle, perhaps in response to it, a revolt broke out in the army at Hsiang-yang.25 This was a particularly serious threat to the T’ang government in Ch’ang-an, since Hsiang-yang controlled the Han river route by which tax goods were sent up from the Yangtze region. Unfortunately, we do not know too much about those who were responsible for the revolt. The surname of the leader K’ang 25  CTS 108.5b (biography of Tu Hung-chien); 138.4a (biography of Wei Lun); TCTC 221, p. 7080, Ch’ien-yüan 2 (759) 8/ i-ssu; p. 7081, 9/chia-wu; p. 7088, 10/chia-tzu.

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

538

Pulleyblank

Ch’u-yüan, who is described as an army officer, suggests that he was a Sogdian and it is quite likely that the garrison of which he was part came originally from one of the western frontier armies. He proclaimed himself Hegemon King of southern Ch’u (Nan Ch’u pa-wang) and moved south to seize Ching-chou (Chiang-ling) and other cities on both sides of the Yangtze. Presumably, he had hopes of either linking up with Shih Ssu-ming or establishing an independent kingdom. He did not last very long. The court was greatly alarmed because of the threat to its tax supplies and within a couple of months Wei Lun, moving down the Han from Shang-chou, had captured K’ang after inducing most of his forces to surrender. Trouble soon broke out again at Hsiang-yang, however. Two other officers, Chang Wei-chin and Ts’ao Chieh (Ts’ao is also a Sogdian surname) killed the prefect and revolted.26 This time the redoubtable Lai T’ien, Lai ‘Bite-iron’, was sent as Prefect and Military Governor and the rebellious officers surrendered to him. They were apparently forgiven and things went more smoothly for a while. Evidently, Lai and the garrison got on well together, for two years later (762), when Lai was summoned to Ch’ang-an to assume a new appointment, he got his men to send a memorial asking for him to stay on. The emperor grudgingly acceded to the request, but his suspicions, and especially those of his eunuch advisers, were aroused. These doubts were further strengthened by accusations coming from the neighbouring military governors as well as the eunuchs that Lai was currying favour with his troops. The situation was compounded when he failed to support the Huai-hsi Military Governor Wang Chung-sheng defending Shen-chou against Shih Ssu-ming’s forces, with the result that the city fell and Wang was captured. It is a significant reflection of the court’s growing impotence in dealing with its generals that, instead of simply dismissing Lai T’ien or again summoning him to Ch’ang-an, the court plotted with one of T’ien’s subordinates to get rid of him. T’ien was appointed Military Governor of Huai-hsi and Honan. Though this was outwardly a promotion, since his territory would be enlarged considerably, he rightly suspected treachery and was greatly alarmed. He sent another request asking to be allowed to stay at his present post at least until the harvest, so that he could provision his troops, and he again got his men to request that he be kept at Hsiang-yang. The emperor was afraid of stirring up trouble and gave in once more, restoring T’ien to his existing office. Meanwhile, the subordinate P’ei Jung who had been plotting with the court had received secret orders to attack him and was coming down the Han river towards Hsiang-yang. 26  CTS 138.4b (Wei Lun); 114.3b (Lai T’ien); TCTC 221, p. 7091, Shang-yüan 1 (760) 3/ keng-yin.

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

The An Lu-shan Rebellion And The Origins Of Chronic Militarism

539

Lai T’ien defeated and captured him, sending him to Ch’ang-an where he was permitted to commit suicide. The court having thus repudiated its own plot against him, Lai T’ien finally submitted and came to Ch’ang-an. He was at first treated with honour. Soon, however, he was accused by the chief eunuch, Ch’eng Yüan-chen, of having made seditious statements and the matter of his failure to relieve Wang Chungsheng was also brought up. At the beginning of 763 he was summarily shorn of his titles and banished to a distant post in southern Szechwan. On the way he was ordered to commit suicide. The treatment meted out to Lai T’ien, a general who had served the dynasty with conspicuous bravery in the early part of the rebellion but had lately showed signs of being hard to control, naturally did not encourage other military governors to leave the security of their own army headquarters and risk their lives by coming to the capital. Lai T’ien’s fate was, indeed, often cited in later times by military governors both as an excuse for refusing to come to court and as an example of the evil machinations of their arch enemies, the eunuchs. Lai T’ien had continued to hold his title as Military Governor of Eastern Shan-nan while he was at Ch’ang-an. After word came of his ruin, the various detachments of the Hsiang-yang army that were garrisoning outlying parts of the province hastened back. A popular officer, Liang Ch’ung-i was acclaimed by the troops and, after killing a rival, successfully demanded that the T’ang court recognize him as Military Governor. He held the post until 781 without ever going to court. The Old T’ang History remarks: ‘Among the various ruffians his territory was the most restricted and his soldiers the fewest. His laws and regulations were the most orderly and his courtesy and demeanour were most respectful. His territory straddled the thoroughfare to the south-east. There were frequent expressions of royal benevolence and so his people knew that they enjoyed the imperial favour. They often urged him to go to court but Ch’ung-i said, “My old commander Lai had performed very meritorious services. In the Shang-yüan period (760–61), because of slanders by the eunuchs, he procrastinated and delayed in responding to the emperor’s summons. When Tai-tsung succeeded to the throne, he went at once, but soon he suffered extreme punishment. Now in my case the rift is wide and the affair is of long standing. How can I face the emperor?” ’27 If the mistrust between Lai T’ien and the court had been an isolated instance, it would no doubt soon have been forgotten. It was typical, however, of the relations between the court and the military at this period. In contrast to T’ang T’ai-tsung, himself a general, who had actively participated in the wars 27  CTS 121.8b.

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

540

Pulleyblank

at the beginning of the dynasty and knew his generals intimately, Su-tsung had been kept in seclusion among his women and eunuchs before the rebellion broke out and still relied to an unfortunate degree on the latter for his contacts with the outside world. Friction and jealousy between eunuchs and generals was constant because of the system of eunuch army supervisors sent to spy on generals and see that they were kept in line. On the other hand, powerful generals were a danger as the rebellion itself had demonstrated. Even the leading generals of the day, to whom the dynasty owed its salvation, were not exempt from suspicion. There is no suggestion in our sources that Kuo Tzu-i was ever guilty of so much as a seditious fancy, but this did not prevent him from being jealously watched. Li Kuang-pi ended his career at Hsü-chou where he had gone in 761 after the defeat at Mang-shan as generalissimo in charge of the armies in Honan. He was not on good terms with the chief eunuch Ch’eng Yüan-chen and, when the Tibetans invaded Ch’ang-an in 763, he procrastinated and did not respond to the emperor’s summons. He died the following year before taking any overtly rebellious steps, but the court was very nervous about his attitude. P’u-ku Huai-en did in fact rebel and, though he is duly blamed as a traitor and villain, the accounts in his biography and in the Tzu-chih t’ung-chien are at pains to show that he had considerable provocation. Having this kind of difficulty with their own generals, it is not surprising that Tai-tsung and his court were unable to extirpate the remaining rebel generals in Hopei and were content to placate them with titles and try to play them off against each other when they became obstreperous. That the maintenance of all these armies spread throughout the country served no useful purpose for the dynasty was obvious to observers at the time. The writer Tu-ku Chi vividly described the waste and extravagance of the military at Ch’ang-an in the 760’s in contrast to the bitter hardships of the common people compelled to support them. He advocated a wholesale demobilization.28 Kuo Tzu-i wanted to move the useless and idle troops from the eastern parts of the country to the northwest where they could have been used against the Tibetans.29 If the court had felt that it had the power to execute such measures, it would no doubt have tried to do so, but clearly it could neither rely on the obedience of its commanders and their troops, nor bring coercion to bear except by bribing other equally unreliable armies. In the circumstances, a policy of temporizing and avoiding confrontations is understandable. When Te-tsung came to the throne in 779, he tried to reverse this 28   Ch’üan T’ang wen 384.20a, cf. HTS 162. 29  CTS 120.10a.

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

The An Lu-shan Rebellion And The Origins Of Chronic Militarism

541

policy and bring the recalcitrant governors to obedience. He roused such a hornet’s nest that he soon reverted to his father’s policy. In the decade from 810 to 820 his grandson Hsien-tsung finally succeeded in breaking the power of the great Honan Governors, even in momentarily bringing Hopei to submission. The armies were not disbanded, however; though the central government had more power over them, they continued to live off the country as parasites without performing any obvious service. The peasant uprisings of Wang Hsien-chih and Huang Ch’ao revealed how useless they had become even for maintaining internal control and at the same time gave them the opportunity once more to become independent of central authority. The culmination of this process was the splitting up of the whole country into independent kingdoms during the Five Dynasties period. It is small wonder that, when the Sung dynasty at last restored the unity of the empire after 960 and succeeded in abolishing the Chieh-tu-shih system once and for all, it displayed a very different attitude than the T’ang had done to military matters of all kinds. Confucianism had always elevated the civil side of government over the military, but the real dominance of the civilian ethic which is often taken to represent Chinese attitudes throughout history dates only from Sung and is, I would guess, to be attributed more to the bad experience of militarism in the T’ang than to the influence of philosophical principles.

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access

Paul W. Kroll - 978-90-04-38015-8 Downloaded from Brill.com11/09/2020 11:07:29AM via free access