The Birth of Japanese Historiography 2020031875, 9780367407315, 9780367647629, 9780367809591

Table of contents :
Cover
Half Title
Series Page
Title Page
Copyright Page
Table of Contents
Acknowledgements
Abbreviations
Illustrations
Conventions
Introduction
Notes
References
Chapter 1 The medium: The mechanics of writing history
Wa on the peninsula
Notes
References
Chapter 2 Words and their meanings
Cultural issues
Notes
References
Chapter 3 Mythical beginnings and the beginning of mythology
Hints to the “Ur-Mythology”
Izumo myths
The filtering of myths
Hyūga as the birthplace of the imperial family
Notes
References
Chapter 4 From kings to heavenly sovereigns
Notes
References
Chapter 5 The beginnings of writing a history
History under Keitai and Kinmei
Suiko era historiography
Early Nara historiography
Notes
References
Chapter 6 The Tenmu-Jitō historiographical project
A new historiographical project
Putting Suiko on the Throne
Kojiki and Nihon shoki
Jitō’s Involvement
Creation of Nihon shoki
Jinmu
Chūai
Yūryaku
Keitai
Ankan-Senka
Notes
References
Conclusion
References
Index

Citation preview

The Birth of Japanese Historiography

As the first book in English on the origins of Japanese historiography, using both archaeological and textual data, this book examines the connection between ancient Japan and the Korean kingdom of Paekche and how tutors from the kingdom of Paekche helped to lay the foundation for a literate culture in Japan. Illustrating how tutors from the kingdom of Paekche taught Chinese writing to the Japanese court through the prism of this highly civilized culture, the book goes on to argue that Paekche tutors guided the early Japanese court through writing, recording family history, and ultimately an early history of the ruling family. As the Japanese began to create their own history, they relied on Paekche histories as a model. Triangulating textual data from Kojiki, Nihon shoki, and Sendai kuji hongi, the author here demonstrates that various aspects of early king genealogies and later events were manipulated. Offering new theories about the Japanese ruling family, it is posited that Emperor Jitō had her committee put Jingū in power, and Suiko on the throne in place of original male rulers to enhance images of strong, female rulers, as she envisioned herself. The Birth of Japanese Historiography will be a valuable resource for students and scholars of Japanese history, historiography, and linguistics. John R. Bentley is a professor of Japanese at Northern Illinois University. He is the author of ABC Dictionary of Ancient Japanese Phonograms, An Anthology of Kokugaku Scholars, and The Authenticity of Sendai Kuji Hongi.

Routledge Studies in the Early History of Asia

1. Imperial Tombs in Tang China, 618-907 The politics of paradise Tonia Eckfeld 2. Elite Theatre in Ming China, 1368-1644 Grant Guangren Shen 3. Marco Polo’s China A Venetian in the realm of Khubilai Khan Stephen G Haw 4. The Diary of a Manchu Soldier in Seventeenth-Century China “My service in the army”, by Dzengeo Introduction, Translation and Notes by Nicola Di Cosmo 5. Past Human Migrations in East Asia Matching archaeology, linguistics and genetics Edited by Alicia Sanchez-Mazas, Roger Blench, Malcolm D. Ross, Ilia Peiros and Marie Lin 6. Rethinking the Prehistory of Japan Language, genes and civilisation Ann Kumar 7. Ancient Chinese Encyclopedia of Technology Jun Wenren 8. Women and the Literary World in Early Modern China, 1580-1700 Daria Berg 9. Asian Expansions The historical experiences of polity expansion in Asia Edited by Geoff Wade 10. The Emergence of Civilizational Consciousness in Early China History Word by Word Uffe Bergeton 11. Cultural Astronomy of the Japanese Archipelago Exploring the Japanese Skyscape Akira Goto 12. The Birth of Japanese Historiography John R. Bentley

The Birth of Japanese Historiography

John R. Bentley

First published 2021 by Routledge 2 Park Square, Milton Park, Abingdon, Oxon OX14 4RN and by Routledge 52 Vanderbilt Avenue, New York, NY 10017 Routledge is an imprint of the Taylor & Francis Group, an informa business © 2021 John R. Bentley The right of John R. Bentley to be identified as author of this work has been asserted by him in accordance with sections 77 and 78 of the Copyright, Designs and Patents Act 1988. All rights reserved. No part of this book may be reprinted or reproduced or utilised in any form or by any electronic, mechanical, or other means, now known or hereafter invented, including photocopying and recording, or in any information storage or retrieval system, without permission in writing from the publishers. Trademark notice: Product or corporate names may be trademarks or registered trademarks, and are used only for identification and explanation without intent to infringe. British Library Cataloguing-in-Publication Data A catalogue record for this book is available from the British Library Library of Congress Cataloging-in-Publication Data Names: Bentley, John R., author. Title: The birth of Japanese historiography/John R. Bentley. Description: Abingdon, Oxon; New York: Routledge, 2021. | Includes bibliographical references and index. | Contents: The Medium: the Mechanics of Writing History – Words and Their Meaning – Mythical Beginnings and Beginning of Myth – From Kings to Heavenly Sovereigns – The Beginnings of Writing a History – The Tenmu-Jitō Historiographical Project. | Summary: “As the first book in English on the origins of Japanese historiography, using both archaeological and textual data, this book examines the connection between ancient Japan and the Korean kingdom of Paekche and how tutors from the kingdom of Kara helped to lay the foundation for a literate culture in Japan. Illustrating how tutors from the kingdoms of Kara and Paekche taught Chinese writing to the Japanese court through the prism of the highly civilized culture of Paekche, the book goes on to argue that Paekche tutors guided the early Japanese court through writing, recording family history, and ultimately an early history of the ruling family. As the Japanese began to create their own history, they relied on Paekche histories as a model. Triangulating textual data from Kojiki, Nihon shoki, and Sendai kuji hongi, the author here demonstrates that various aspects of early king genealogies and later events were manipulated. Offering new theories about the Japanese ruling family, it is posited that Emperor Jitō had her committee put Jingū in power, and Suiko on the throne in place of original male rulers to create images of strong, female rulers, as she envisioned herself. The Birth of Japanese Historiography will be a valuable resource for students and scholars of Japanese history, historiography and linguistics” – Provided by publisher. Identifiers: LCCN 2020031875 | ISBN 9780367407315 (hardback) | ISBN 9780367647629 (paperback) | ISBN 9780367809591 (ebook) Subjects: LCSH: Japan–Historiography. | Genealogy–Historiography. Classification: LCC DS834.7 .B46 2021 | DDC 952.0072–dc23 LC record available at https://lccn.loc.gov/2020031875 ISBN: 978-0-367-40731-5 (hbk) ISBN: 978-0-367-80959-1 (ebk) Typeset in Times New Roman by Deanta Global Publishing Services, Chennai, India

Contents

Acknowledgements Abbreviations Illustrations Conventions

vi vii viii x

Introduction

1

1

The medium: The mechanics of writing history

9

2

Words and their meanings

44

3

Mythical beginnings and the beginning of mythology

72

4

From kings to heavenly sovereigns

95

5

The beginnings of writing a history

116

6

The Tenmu-Jitō historiographical project

152

Conclusion Index

197 202

Acknowledgements

This manuscript began as a research project in the spring of 2014 while I was on sabbatical. I am grateful to the Department of World Languages and Cultures at Northern Illinois University for its support, especially to Katharina Barbe, the then chair of the department, and Linda Saborio, director of undergraduate studies, who were supportive and enthusiastic about my research. I am grateful to Kuniko Yamada McVey at the Harvard-Yenching Library who arranged for me to have access to the stacks. I am grateful to both Ross Bender and Jonathan Best who read earlier drafts of the work and gave helpful comments. I also wish to express my gratitude to the perceptive comments of two readers. They have spared me much embarrassment. Any errors that remain are my responsibility. I am also grateful to the following individuals who took the time to answer a variety of questions over the years that I formulated the ideas and theories that became this monograph: Bjarke Frellesvig, Marc Miyake, Robert Ramsey, Kerri Russell, Leon Serafim, Aoki Shūhei, Rumiko Shinzato Simonds, and John Whitman. I am thankful for the continued love and support of my wife, Chiemi.

Abbreviations

EMC L LH LHC LMC NA OC OJ PART QD

Early Middle Chinese low tone/pitch accent low-high tone/pitch accent Later Han Chinese Late Middle Chinese not available Old Chinese Old Japanese particle queen dowager

Illustrations

Figures 1.1 1.2 1.3 1.4 2.1 3.1 3.2 3.3 3.4 3.5 4.1 5.1 5.2 5.3 5.4 5.5 6.1 6.2

Commanderies on the Korean Peninsula Timeline of Classical Chinese Nara Palace wooden slip. Kyujanggak Institute for Korean Studies Nara era wooden slip. Kyujanggak Institute for Korean Studies Korean peninsula in 108 BCE Deities and descendants Offspring of the creation deities Grains produced Kumaso versus Hyūga: common thinking Kumaso versus Hyūga: new proposal Two dynasties Wa genealogy seen in Song shu Two competing lines Hierarchy of early records Sample kings’ length of lives Creating overlapping lives of kings Putting Suiko on the throne Source material for Kojiki/Kujiki/Nihon shoki

10 24 30 31 54 76 76 79 90 91 105 124 125 130 140 141 161 187

Tables 1.1 1.2 3.1 4.1 4.2 4.3 4.4 4.5 4.6 5.1

Possible Japonic words Reconstruction of *Dźɨn King’s name List of events in mythology Spelling of certain names of rulers Core names of rulers Yamato Neko versus Tarashi Hiko Comparison of rulers’ names Expansion in Paekche Royal lineage Comparison of Paekche royal succession Basic template of a record

16 21 75 102 102 103 104 110 111 127

Illustrations 5.2 5.3 5.4 5.5 5.6 5.7 6.1 6.2 6.3 6.4 6.5 6.6 6.7 6.8 6.9 6.10 6.11 6.12 6.13 6.14 6.15 6.16

Comparison of peninsular data Names of Yamato rulers in Fudoki Overview of early records in Shoki Chūai and Jingū records combined Ages of six rulers Events in Shoki taken from Shiji Changes Tenmu initiated Suiko’s actions in Shoki Suiko’s actions in Kujiki Comparison of Suiko and Wu Zhao Basic information of recent rulers Basic information of earlier rulers Revision of royal chronology Ancient template for a dynastic record Word usage regarding the capital Distribution of ways to record children Distribution of marriage Events in Yūryaku’s reign Ankan and Senka’s reigns Possible manipulation of records S-initial graphs in Shoki EMC versus LMC

ix 127 136 139 140 142 146 155 163 163 165 165 166 167 176 178 179 180 183 186 187 191 192

Conventions

This monograph is an attempt to merge linguistics with history to answer questions about the origins of Japanese historiography. This exercise necessitates an analysis of primary texts from China, the three kingdoms of the Korean peninsula, and Japan. I have represented the original text in traditional Chinese characters. When representing or reconstructing phonograms in Chinese, Korean, or Japanese, I have relied on the innovative Chinese reconstructions of Baxter and Sagart (2014) for Old Chinese, Schuessler (2007) for Later Han Chinese, and Pulleyblank (1991) for Early Middle Chinese. In quotations of the original Chinese, a square (□) represents a character that is illegible. In the English representation, illegible characters are written with a capital X. All translations are my own, unless specified otherwise. I have included the original text so that the reader can check any word or term. I have represented Old Japanese in a modified Hepburn system (for those conventions, see Bentley 2013:10–11). When glosses of Japanese words come from the Heian era or later, I have elected to write /p/ as /f/. Middle Korean words are written according to Yale Romanization, but well-known place names are written in McCune-Reischauer. The names of deities in the Japanese mythologies are written in standard Hepburn Japanese. I transcribe phonograms in lowercase letters, but semantograms (kungana) appear in capital letters. Thus, 意美 omi but 臣 OMI. There have been times when the pitch accent of (Old) Japanese is relevant to a discussion and is noted. H represents “high,” while L is “low.” Thus, HLH is the pitch accent for a trisyllabic word: “high-low-high.” When talking about early kings, I have used the names as found in Nihon shoki for Paekche kings, with the modern names given in parentheses. I have used the original name for quasi-historical kings, and Sinified posthumously awarded the names of Japanese “emperors,” i.e., Jinmu, Jitō, etc., for legendary and historical names. I have done this because both names tend to be quite long in Japanese, while the names from Ōjin (Homuda) to Keitai (Wohodo) are short. When referring to Japan or the Korean peninsula as a geographical location, I have used the modern labels, but when referring to specific polities, I have called Japan “Wa” when dealing with pre-fifth century and Yamato after that period. I have given lesser-known peninsular place names in both the Japanese rendition and Korean. Korean Romanization has followed the McCune-Reischauer system.

Conventions

xi

References Baxter, William H. and Laurent Sagart. 2014. Old Chinese—A New Reconstruction. Oxford: Oxford University Press. Bentley, John R. 2013. Tamakatsuma: A Window into the Scholarship of Motoori Norinaga. Ithaca: Cornell East Asia Series. Pulleyblank, E. G. 1991. Lexicon of Reconstructed Pronunciation in Early Middle Chinese, Late Middle Chinese, and Early Mandarin. Vancouver: University of British Columbia Press. Schuessler, Axel. 2007. ABC Etymological dictionary of Old Chinese. Honolulu: University of Hawaii Press.

Introduction

The term “historiography” originally meant the writing of history and as such signifies what we term “historian” today. Modern definitions of historiography have evolved from earlier economic or political analyses into more “human” realms such as social and cultural history. Breisach (2008:3) notes that the task of the historian is to “trace the ways in which people in Western culture have reflected on the past and what these reflections have told them about human life as it passes continuously from past to present to future.” However, as far as early Japanese historiography is concerned, there clearly was more at work than simple reflection on a single past or a stream of time from the past to their present. It is apparent that early Japanese viewed their history through the fractured prism of “family history.” Evidence also strongly suggests that later a plurality of parallel yet competing pasts were subjected to reflection, review, and then, likely through coercion, revision. It is illuminating to reexamine the actual process of writing history in ancient Japan, as an elucidation of the process also sheds light on what the early Japanese deemed important, significant enough to manipulate and conflate. One of the distinct differences of the present monograph compared to almost any other that has been written on Japanese historiography is the focus on the mechanics of writing and the inclusion of evidence gleaned from historical and philological linguistics. This adds an important window through which historians can get a better view of the early process of historiography in Japan. Historical linguistics attempts to recover earlier remnants of a language, a language that underpins a civilization. Culture, in its many forms, written, performed, or sung, is transmitted through language. Understanding the language of earlier periods of Japanese has the power to unlock a deeper perception of the culture. Linguistics also makes it possible for scholars to elucidate critical elements found in the historiographical process. As a variety of written documents exist in Japan, linguistics married to philology is potentially a valuable tool, one that is seldom used across fields in Asia. For example, in ancient Japan, we encounter the concept of kotodama, “the spiritual power in words,” as early as Man’yōshū, a poetic anthology whose inception may be dated to around 720–730 CE. It is possible that this concept of kotodama evolved from the earlier belief in the power inherent in curses and spells. This power is evident in “spell songs” (waza uta) strategically placed in

2 Introduction the final third of Nihon shoki (720 CE), where certain types of songs portended an evil or fortuitous event. Thus, the choice of words in specific situations was believed to be potent or prophetic. It is not difficult to connect this evidence to the cultural proclivity for the use of elongated, auspicious words (as epithets) inserted in the names of people or places in ancient Japan. Auspicious words used to describe place names, such as Yamato (Japan) being called Toyo Ashihara Chiiho Aki no Mizuho “abundant-reed-plain-1500 gem-like ears of rice in the autumn (harvest),” had the power, people believed, to not only praise the prosperity of the land but also encourage the continuation of that prosperity. This belief in the power of certain words was later grafted into writing, allowing that power to survive longer than a single utterance. Wang Haicheng (2014:21–52) makes a compelling case that writing was used in many early civilizations not only for administration but also for legitimizing the rulers of the state. In its most fundamental form, this was accomplished by preserving “king lists,” outlining a lineage of rulers that created “the state as an imagined political community” (2014:301) projected back in time. These royal names could then be associated by the community with existing physical artifacts such as temples, pyramids, or keyhole-shaped tumuli. Thus, we can view historical writing as a rather incestuous endeavor, recording history as well as generating the vehicle to preserve itself. Typically, Japanese academia has been less interested in historiography as a field, being more focused on the intersecting realms of the origins and development of Japanese culture, mythology, historical records, and the nation/state. Thus, from a Western perspective, research into the birth of historiography in Japan is a fertile field in which many Japanese scholars have plowed sections, but there has been little comprehensive work. Little has been written on this broader topic in the English language, aside from studies focused on the two earliest works, Kojiki and Nihon shoki.1 This statement is not intended to be critical or negative, as starting with these two works is quite natural. Japanese history seems to begin with these two titles, but scholarly work here simply scratches the surface of a richer story to be told about how historiography came about in Japan. According to its preface, Kojiki (古事記 “a record of ancient events”) was presented to Empress Genmei (r. 707–715) on the 28th day of the first month of 712. Ō no Yasumaro (d. 723), the author of the preface, notes that the empress commanded him to abridge and record the work of what I call “the Hieda no Are project” on the 18th day of the ninth month of 711. Thus, according to his own record, Yasumaro completed his task in roughly four months. This short span of time, compared to the nearly four decades required to complete Nihon shoki, suggests that much of the work of compilation was already completed, and Yasumaro’s role may have consisted of little more than organizational and scribal work. In other words, Yasumaro simply edited an already completed manuscript. According to a terse announcement preserved in Shoku Nihongi (続日本紀 “Chronicles of Japan, continued”), Nihon shoki was presented to the throne by Prince Toneri (676–735), the fifth son of the influential ruler known as Tenmu (r. 672–686), sometime on or before the fifth month of 720. Lacking a preface

Introduction 3 or document of presentation, the circumstances surrounding its compilation are much murkier. This lack of clarity in the records has long intrigued people in both Japan and the West. Why does the first official record, containing the name of the state (nihon), appear without fanfare or any account about the process, and yet the smaller, more insular record of Kojiki, which makes no appearance in the state histories, survives with an ostentatious preface? One is forced to conclude that the historiographical process that gave birth to both Kojiki and Nihon shoki was much more complex and turbulent than we currently appreciate. This veil over these events naturally leads to curiosity. The birth of their own historical records has intrigued the Japanese as far back as records exist. The Japanese court sponsored a number of lectures regarding the text of Nihon shoki during the Nara and early Heian eras. Quotes from the minutes of these lectures were compiled into a text called Shaku Nihongi (ca. 1293).2 These lectures took the format of questions and answers between a learned individual, called “the scholar,” and members of the audience who were generally courtiers. The following series of questions and answers is illustrative: QUESTION: What documents were available and formed the basis for the compila-

tion [Nihon shoki]? The scholar explains that some people believe Kojiki was the basic work; others state that it was Sendai kuji hongi. If the compilers [of Nihon shoki] used Kojiki as the basic material for this work, then we must address the many problematic differences existing between the two texts [making that theory untenable]. Kojiki simply establishes the meaning of words and sets up the origin of things, without giving any thought to literary beauty. Thus, during the compilation of this work, the compilers would have been obliged to modify many passages. (Onoda 1986:7)

ANSWER:

Within a century or so, the educated courtiers no longer had access to the institutional memory about the fundamental source material for Nihon shoki. Without knowing what the source material was, knowledge regarding the origins of this highly influential historical work was doomed to slip into the fog of time. People were rightly intrigued by Nihon shoki, with its many versions of a story, while Kojiki is a linear narration. Regarding the many variant quotes (J. issho, 一書) in the first two books of Nihon shoki, the following exchange appears in Shaku Nihongi: QUESTION:

The comments in this record are not commentary from historical works. There are many quotes from variant works (一書) or other traditions (或説). What are these? ANSWER: Previous scholars have denied what some have said about these comments lacking an explanatory character. Regardless, there are many comments, inserted with the words “another work says,” “another theory is,” “a tradition says,” or “it is also said that….” The reason [for these comments]

4 Introduction is because there were families in the past who crafted records in the ancient language as they enjoyed [learning about] events from the ancient period, and [that resulted in] quite a number of these works. When this record (Nihon shoki) was compiled, the compilers did not want to insert these other traditions into the text, but they found they could not discard them. Thus, they added these quotes to the text. (Onoda 1986:11) Umezawa Isezō (1976:457) finds it difficult to accept this theory about families crafting earlier works, but he notes that this quote at least preserves a tradition underscoring the basis for Tenmu’s claim preserved in Kojiki’s preface that records in possession of various families had been corrupted by the inclusion of facts the throne deemed untrue. As events surrounding the historiographical project that resulted in Nihon shoki are shrouded in mystery, the only clue scholars have for the impetus of the project that resulted in either Kojiki or Nihon shoki is found in Yasumaro’s preface and the text of Nihon shoki. A multitude of scholars have mined the preface of Kojiki looking for clues, but David Lurie (2001:270) cautions that Yasumaro’s preface sets up a “fictional construct,” suggesting that the information within the preface must be read and used with extreme care—a shot across the bow of readers and scholars who tend to believe a variety of facts provided by Yasumaro. The key here is to keep in mind that we may figuratively be reading Alice in Wonderland and not Finnegans Wake. What makes research into the origins of Japanese historiography somewhat difficult is the two-dimensional nature of these surviving data. Historians have worked mainly with these two semi-detailed sources (Kojiki and Nihon shoki), but while some have moved forward, there has been a constant drumbeat by others trying to demonstrate that X or Y in either record is not reliable and should be downplayed or ignored. The thickening scholarly skepticism has, at times, had the unintended consequence of casting doubt on the overall value of both works. Wakai (2006:289–90) sounds a warning, noting that many scholars place differing degrees of value on Kojiki and Nihon shoki, depending on their skepticism of the worth of these two works. He sees these scholars putting distance between their own research and these two works but believes that both works should not be discarded as completely unworthy of historical inquiry. Kobayashi (2006:324) takes a step further, cautioning us, “We cannot say that traditions (found in Kojiki and Nihon shoki) are historical facts, but this author has confidence in the power of the traditions in Kojiki and Nihon shoki.” In other words, a balanced approach to these works has more to offer that is positive than negative. My research proposes to resolve this difficulty by triangulating these data with the addition of a third text, Sendai kuji hongi (hereafter Kujiki). Condemned as a forgery in the Edo Period, and then later feebly resuscitated as “derivative,” much of the supposed scholarly work on the authenticity of Kujiki has been cursory, haphazard, and unscientific. In a nutshell, scholars have tended to get

Introduction 5 hung up on the problem of seeing Kujiki as a Heian era creation, but if Kanda (1992) and Bentley (2006:44–48, 67–70) are correct, then Imibe Hironari relied on an untitled manuscript of Kujiki in the process of compiling Kogo shūi (807), and not the other way around. This should cause us to reexamine the dating of Kujiki. While I have argued that Kujiki is still derivative in nature, I posit that the period of inception for this important text is contemporaneous with the period of compilation of both Kojiki and Nihon shoki, the work done under the authority of Minister of the Left Isonokami Maro (640–717). The work was completed as a draft, and with the death of Maro, this draft was shelved, not to see the light of day until the early Heian era (Bentley 2006:117–19). Access to these three works grants a better glimpse not only into the past but also into the process of historiography, especially since there are several unique sections in Kujiki that provide important additional information. The search for the origins of historiography in early Japan is entangled in several different but related issues: an oral tradition versus a written one, a native tradition versus one decked out in Chinese garb, and the fractured feature of primitive chronology versus a manipulated chronology.3 One purpose of this monograph is to untangle these issues to understand the earlier elements of historiography and examine how foreign literate cultures (primarily from Paekche and China) were adapted into the Japanese historiographical process and then reengineered later for political and social purposes, creating the early histories that resulted in the histories that have survived until today. The process for this search is understandably somewhat tenuous, as our main evidence is textual. Texts say whatever the author wants, but the hope is that by triangulating our data, we increase the possibility of screening out layers of manipulative noise. Here Kobayashi (2006:85) has a word of wisdom for us because many scholars argue for the “creation” or “fabrication” of material in Kojiki and Nihon shoki with a tendency to shut the door to pathways within traditions that may lead to answers to a variety of questions. I interpret this warning not to mean that we trust everything in these records, but rather that we should spend adequate time using internal evidence to tease out what the ancient historiographers believed, regardless of whether it was historical fact or not. Our modern low tolerance for things not historical should not be allowed to become an unyielding standard upon which we judge the ancient historians to whom the idea of “truth” or “fact” was likely different from our own. My method is to use comparative textual analysis to find vestiges of data, mainly genealogical, to reconstruct the earlier state of histories. By trying to uncover an earlier template, I will demonstrate how early Japanese historiography was deeply indebted to the philosophy, skill, and technology of the kingdom of Paekche. Many have suggested and even argued for such a link, but the inclination continues to be that scholars see China as the great teacher and Japan as the student. This monograph argues that such a conclusion is not true until at least the beginning of the seventh century. This work is constructed in the following chapters.

6 Introduction Chapter 1, “The medium: the mechanics of writing history,” starts from the very beginning, illustrating how the techne of writing came to Yamato. Evidence is provided showing clearly that the kingdoms of Koguryŏ and Paekche accepted the Chinese writing system in toto, and then manipulated it to fit characteristics of their own language. That early Japan did the same is neither coincidence nor accident. Peninsular tutors taught and guided Yamato students in this process. With Chinese writing came Chinese philosophy, politics, and record keeping. This discussion is set in a broader geopolitical framework, as none of these kingdoms acted in a vacuum. Chapter 2, “Words and their meanings,” surveys the introduction of writing to the Yamato elite. With this cultural tool came extended Chinese cultural and political philosophies, but these entered Yamato through a Paekche filter. Cultural power, through the written word and writing technology, came into Yamato in a two-pronged wave: one through shared contact by the Wa on the peninsula and the other through direct importation from Paekche. This chapter introduces Paekche and Silla records that outline the mythological beginning of their histories, and it is natural to segue into Japanese mythology, as all the early Japanese records begin with the mythological foundations of the ruling family. Finally, cultural issues are addressed regarding the importation of these cultural tools. Chapter 3, “Mythical beginnings and the beginning of mythology,” surveys Japanese mythology as contained in Kojiki, Nihon shoki, and Kujiki, compared with the foundation stories in Okinawa. This chapter attempts to peel back the layers of political thinking to uncover an “ur-mythology” that originally formed the core of the mythology of the powerful families in Izumo and Yamato. Chapter 4, “From kings to heavenly sovereigns,” outlines the process from the earliest attempts at historiography, starting from “king lists” to a simple chronology of rulers and then eventually to the creation of an unbroken succession of rulers that extended back in time beyond the institutional memory of the court. The main point is that serious historiography began in the reign of Yūryaku (Wake Take) when oral traditions were set to writing. An attempt is made to further the work of historians like Mizuno (1954) to determine which rulers in the genealogy are historical and which are either fabricated or legendary. One conclusion is that competing dynasties ruled the Yamato basin during the late fifth into the sixth centuries. This chapter also concludes that the manipulation of the age of the early kings, and the plagiaristic quoting of Chinese works in the histories, is not something the Japanese invented, but was a technique learned from their Paekche tutors. Chapter 5, “The beginnings of writing a history,” argues that the Keitai-Kinmei era saw the first actual history of the Yamato court with a genealogy of rulers going back six generations. This was the first record to connect the ruling lineage back to a mythological figure. I introduce the three Paekche histories quoted in Nihon shoki, as evidence of a Paekche template used by the Yamato court. The simple historiographical template is preserved in the skeleton record in the dynastic section of Kujiki. This chapter then introduces the thorny subject of historiography in the Suiko era. Using hints from Nihon shoki, it is argued that historiography

Introduction 7 passed from a courtly function to one controlled by a few peninsular families, who retained the technology for writing. It is also during this season that the Yamato court decided a unified foundational myth was necessary. The historiographical work of this period was never finished but remained in draft form. Chapter 6, “The Tenmu-Jitō historiographical project,” outlines the beginning of what later becomes Kojiki and Nihon shoki. While Tenmu ordered the beginning of the process, I argue in this chapter that the stamp of Jitō was what set this project apart from previous ones. Jitō wanted the realm to view her court as controlling an empire, a mirror image of a female sovereign who anciently had the peninsula under her umbrella; thus, the court elevated Jingū to the status of a military queen, much like the sun goddess thwarting her unruly brother. The dissonance between the history as portrayed in Nihon shoki and the description of Suiko’s era in Suishu is resolved by concluding that the Suishu is the more accurate, and the Shoki version is heavily manipulated. In other words, Jitō wanted a female sovereign on the throne and had the history revised. Ultimately Jitō’s vision was to create a mirror image of the ruling family in the High Plain of Heaven, with herself as the sun goddess. The conclusion ties this together, arguing that greater appreciation is needed for the guidance and influence that earlier kingdoms, mainly Kara and Paekche, had in laying the cultural and technological foundation for Yamato to become literate, keep its own written records, and then begin to write its own history.

Notes 1 Exceptions include Barnes (1987, 2012) and Piggott (1997). 2 I have followed Onoda (1986:22) in seeing Shaku Nihongi as being completed between 1286 and 1301; he also adheres to the work by Akamatsu Toshihide, who posits a date around 1293 to be the most appropriate. 3 This refers to oral traditions with a vague, fluid chronology being later mated to a “constructed” narrative complete with a detailed chronology.

References All works are published in Tokyo, unless otherwise noted. Barnes, Gina. 1987. “The Role of the Be in the Formation of the Yamato State.” In Elizabeth M. Brumfiel and Timothy K. Earle, eds. Specialization, Exchange, and Complex Societies. London: Cambridge University Press, pp. 86–101. Barnes, Gina. 2012. “The Emergence of Political Rulership and the State in Early Japan.” In Karl Friday, ed. Japan Emerging: Premodern History to 1850. Boulder: Westview Press, pp. 77–88. Bentley, John R. 2006. The Authenticity of Sendai kuji hongi: A New Examination of Texts, with a Translation and Commentary. Leiden: Brill. Breisach, Ernest. 2008. Historiography: Ancient, Medieval and Modern. Chicago: University of Chicago Press. Kanda, Hideo. 1992. “Sendai kuji hongi o Megutte.” In Nakamura Hirotoshi, Aoki Shūhei, Sugano Masao, and Yamazaki Masayuki, eds. Umezawa Isezō Sensei tsuitō kiki ronshū. Zoku Gunsho Ruijū Kanseikai, pp. 3–8. Kobayashi, Toshio. 2006. Nihon kodai kokka keiseishi-kō. Kōsō Shobō.

8

Introduction

Lurie, David B. 2001. The Origins of Writing in Early Japan: From the 1st to the 8th century C.E. Columbia University Ph.D. Dissertation. Mizuno, Yū. 1954. Zōtei nihon kodai ōchō shiron josetsu. Omiyayama Shoten. Onoda, Mitsuo. 1986. Shaku Nihongi. Vol. 5 of Koten Chūshaku-hen of Shintō Taikei. Shintō Taikei Hensankai. Piggott, Joan R. 1997. The Emergence of Japanese Kingship. Stanford: Stanford University Press. Umezawa, Isezō. 1976. Zoku Kiki hihan: Kojiki oyobi Nihon shoki no bunkenteki sōgo kankei no kyūmei. Sōbunsha. Wakai, Toshiaki. 2006. Kiki shinwa no genzō to Yamato ōken no kigen. Contained in Vol. 27 of Nihon shoki Kenkyū. Hanawa Shobō. Wang, Haicheng. 2014. Writing and the Ancient State: Early China in Comparative Perspective. New York: Cambridge University Press.

1

The medium The mechanics of writing history

The term “historiography” naturally implies that some historical narrative is being written on some physical material through a specific linguistic medium. This is such a basic assumption that scholars generally gloss over it or provide little more than a sentence or two. However, as Lurie (2011:4) points out, there has been such an overzealous search for the origins of phonographic writing in Japan that other dominant forms of writing have been ignored or underappreciated, resulting in a skewed understanding of how writing sprouted and then grew in Japan. It is helpful to retrace these steps to examine how a small group of literate inhabitants on the Japanese archipelago wrote a century before the era when the work on Kojiki and Nihon shoki began. To do that, we need to go back several centuries earlier in time. Kobayashi (2010:20–23) has argued that with the establishment of the four Han commanderies on the Korean peninsula by Emperor Wu around 108 BCE Chinese culture began to seep systematically into the cultures of the various native polities on the peninsula. One of these key components of Chinese culture was written communication, or the ability to read and write.1 A number of scholars have previously suggested that Wa-speaking (or para-Japonic) people inhabited both the southern part of the Korean peninsula and the archipelago at a transitional time in history (cf. Unger 2005, Vovin 2007, Bentley 2008a), based on archaeological or linguistic evidence. The most fundamental data are tantalizing tidbits from Chinese records.2 Thus, we are not required to see immigrants from other polities as the only catalyst for burgeoning literacy in Yamato. Kwon (2013:82–85) has argued that Han China originally set up an “outer vassal” (外臣) system with Old Chosŏn during the reign of Emperor Hui (r. 195–188 BCE), where responsibilities of defense and security were entrusted by the Han Court to Old Chosŏn. When Old Chosŏn’s belligerence and insolence became too much to stomach, Emperor Wu (r. 156–87 BCE) invaded Old Chosŏn and destroyed it. Wu then established four commanderies (Chinese colonies) in the conquered territory: Zhenfan, Lintun, Xuantu, and Lelang. The precise locations of these are a matter of debate, but scholars agree that Lelang, the most important of the four commanderies, was located near modern Pyongyang. Excavations at T’osŏng-ni confirm that the Lelang administrative center was in that area (Eckert et al. 1990:14). Kim (2012:18) believes that the Zhenfan Commandery was

10

The medium

located north of the Han River in Hwanghae; the center of Lintun was in South Hamgyong, and Xuantu was near the middle reaches of the Yalu River basin. Byington (2013:294) has provided a helpful map that has informed the map in Figure 1.1. The cultural center of the Chinese-controlled peninsula was clearly Lelang, a large city at the time. Part two of the geographical section in Hanshu describes Lelang thus: “The Lelang Commandery has 62,812 households, with a population of 406,748 people” [樂浪郡，戶六萬二千八百一十二，口四十萬六千七百四十八]. In contrast, the Xuantu Commandery had 45,006 households and a population of 221,845. Nothing appears in Hanshu about the size of the commanderies at either Zhenfan or Lintun, suggesting that these were of relatively minor importance.

Figure 1.1 Commanderies on the Korean Peninsula. This map is based on “File: Atlas of South Korea.png.” Wikimedia Commons, the free media repository. https://co mmons.wikimedia.org/wiki/Atlast_of_South_Korea (accessed June 28, 2017).

The medium 11 Not many years after Emperor Wu’s passing, the Han Court reorganized these commanderies, abolishing Zhenfan and Lintun, and relocating Xuantu. The territory of Zhenfan was folded into the territory under the jurisdiction of Lelang. Sometime between 215 and 220 CE, the territory originally controlled by Zhenfan was reorganized and put under the jurisdiction of a newly established commandery called Daifang. As Eckert et al. (1990:14) have argued, the commanderies exerted a profound influence on the surrounding indigenous population, as “neighboring territories coveted the fruits of the highly advanced Chinese culture.” This fact is demonstrated by chieftains in the Samhan (三韓 “Three Han”) areas willingly accepting Chinese office and rank, with its attending physical symbols, such as official seals, ribbons, and banners. Scholars have excavated these objects in burial areas, attesting to the stratification of the elite. To underscore this pervasive Chinese cultural influence, Ri (2005:33–34) notes that the area under the control of the Lelang Commandery was divided into 25 districts, and the names of 23 of these districts have been discovered on fūdei “clay seal” inscriptions. Fūdei are balls of clay placed on twine used to tie up bamboo boxes or clay containers that hold important documents. An official then firmly presses an official stamp or seal into the clay, which then hardens, preventing unauthorized opening of the contents. These “documents” sealed by fūdei are concrete proof that written communication occurred between the commanderies and district offices in deeper locales on the peninsula. Hou Hanshu includes the following report, In the twentieth year of jian-wu (44 CE) an individual from *Liam-sie [which land] belongs to the Kara (Han) Territory, [named] *Sɑ-maɁ-de and others presented tribute at Lelang. (Emperor) Huang-Wu enfeoffed *Sɑ-maɁ-de and made him Lord of the *Liam-sie Settlement within Han China, Envoy attached to the Lelang Commandery, summoned to court at each of the four seasons [建武二十年，韓人廉斯人蘇馬諟等詣樂浪貢獻。光武封蘇馬諟 為漢廉斯邑君，使屬樂浪郡，四時朝謁]. Scholars believe that the territory of *Liam-sie (廉斯) came to be part of Shinhan (K. Chinhan). The Kara Federation sprang from this territory, and Ri (2005:35) reports that from the Taho-ri excavation site in Ch’angwŏn, five different types of brushes have been excavated. These are peculiar brushes, with bristles on both ends. From this site, archaeologists also found a small iron dagger with a ring on the handle. Workers used this dagger to carefully carve off thin layers of wood on wooden tablets, thus removing written mistakes or erasing previous messages written in ink. This excavation site is dated around 100 BCE.3 As this site is not far from modern Pusan, it is easy to imagine this area being a crossing point for people going and coming from Kyūshū, suggesting that writing (or at least written artifacts) also had opportunities to cross. Hanshu contains this terse report, “The people of Wa exist in the middle of the ocean from Lelang, divided and forming 100 small states. It is said

12 The medium they appear at [the Commandery] and present tribute at regular intervals” [樂浪海中有倭人，分為百餘國，以歲時來獻見云]. Kimura (1998:22–23) argues that this line is based on a report from about 50 BCE, because Lelang is mentioned, and not Daifang. He interprets the first line to mean that the people of Wa exist somewhere in the ocean, reached by passing through the governmental channels of Lelang and then traveling by ship. He then addresses the brief information about Wa found in Shanhaijing, noting that Wa is subsumed under the jurisdiction of Greater Yan, and not mentioned in relation to Lelang, demonstrating that this information predates the establishment of the commandery in 108 BCE (Kimura 1998:24). We know that around 275 BCE, Yan occupied an area approximately covering the modern districts of Hebei and Liaoning in China, with its capital in what is now Beijing. Thus, it appears from this description in Shanhaijing that the territory of *ɡɑp (鉅J. Kō, K. Kae) was located south of Yan, and south of that point was Wa. One is struck by the fact that the ocean is not mentioned in this description, giving the strong impression that Wa is south of Yan on the peninsula. I return to this issue later in relation to the Kwanggaet’o Stele. Hou Hanshu further records that sometime after the commanderies were established, “Roughly 30 territories (of Wa) sent translators and envoys to Han (漢)” [使驛通於漢者三十許國]. Again, the underlying meaning is that Wa envoys went to the Han Court by passing through the proper channels at the commandery. It is difficult to ignore Wa as an ethnonym, describing an ethnic group occupying part of the peninsula and extending to the Japanese islands, at least as far as Kyūshū. This idea that translators were members of the envoy’s party strongly suggests that there already were bilingual people in the area. It is clear from a number of accounts that the people of the various Han and Wa adopted Chinese titles and ranks. Consider that Hou Hanshu records that in 57 CE, an envoy from the Wa territory of Na appeared at court: “The envoy designated himself ‘grandee’ (大夫)” [建武中元二年，倭奴國奉貢朝賀，使人自稱大夫]. In addition, there are a few Chinese offices mentioned in the Wa section of Weizhi, such as grandee and grand general (大率). This demonstrates some familiarity with Chinese rank and prestige. Kobayashi (2010:21–22) has also argued that there must have been a number of officials residing in the territory of *Itɔ (Ito) in Wa who could read and write Chinese because of the appearance in the record of terms such as “document” (文書), “proclamation” (詔書), “document of receipt” (上表), and “announcement” (檄). These terms assume there is someone on the receiving end who can read and write Chinese. However, we should be cautious, as Lurie (2011:75) suggests because the record of the Wa in Weizhi “contains no references to writing (a significant omission, given the attention it receives as a marker of cultural attainment in such treatises on peripheral barbarians).” Thus, we must conclude that the Chinese observers who visited the Wa did not find a literate society among the various polities dwelling together under the umbrella term of Wa. Regarding some of the literary terms noted above, Lurie (2011:77) responds that “a passing reference (such as this) so in keeping with the ideal of

The medium

13

tributary relations need not be taken literally. Even if there were such a manifest, it was not necessarily composed in Yamatai.” Lurie would rather see the queen (Himiko) and her federation completely beyond the reaches of any kind of Chinese literary power, relying on literate people on the peninsula, at the Daifang Commandery. While this statement is fundamentally sound, I find this skepticism rather unbalanced, however. It is easy to sweep aside everything that cannot be accounted for with concrete evidence, but in this case, where evidence is so hard to obtain, I fail to see why more faith cannot be placed in the Chinese documents as we have them. I would argue that Han (韓) families who served in various positions within the commandery, and had thus become literate, had sons and grandsons who traveled south in search of opportunities to engage economically with Mimana, Kara, and the Wa, looking to financially leverage their skills in Chinese literacy. Granted, having a denizen who can pilot a ship does not therefore make that locale a shipbuilder. Having a few literate officials need not (nor does not) equal a literate society. Consider further the fact that Han Chinese families had fled to Old Chosŏn by the tens of thousands during the Qin-Han Transition (Kwon 2013:84). Centuries later when the political orbit of the Eastern Han dynasty was in decay, Han Chinese living in the Lelang area began to escape into the southern territories (Kwon 2013:93), meaning that Han Chinese were now entering the territory of the Three Han. This better explains why brushes and other writing equipment are being excavated from Han territory. Surely, there were Chinese who were bilingual, as well as some indigenous people who had become literate in Chinese. Uchida (1996) brings up an excellent example when he compares a Paekche missive sent to the Northern Wei court in 472 with the missive sent to the Song court in 478 by Yamato’s King Muɑ (Bu). Both missives share an interesting characteristic in that the texts of these two have relied heavily on word usages found in Jinshu. Uchida (1996:96–106) uncovers 92 phrases in the Paekche missive that have come from Jinshu. The Wa missive has 41 phrases from Jinshu, a smaller number, but only because the Wa missive is shorter than the one from Paekche. Uchida concludes that someone of Chinese descent helped write the missive for the Paekche kingdom, and then with the destruction of the Paekche capital of Hansŏng in 475, this individual fled to Wa, where his services were again employed. We must acknowledge that having literate people at court had the potential to catalyze a larger group of literate people in Wa, so the logical question is why this did not occur at this early point in time. I believe two pivotal events prevented this from happening at this early date. The first is the instability of the society in Wa. With the death of Himiko (ca. 248 CE), the federation fell into war and chaos. After an unspecified period of war, Toyo becomes queen. It is possible that later with the death of Toyo the same sequence of events reoccurred. There is even the possibility that intermittent civil strife was an issue for at least another century. Perhaps people with the ability to read and write became unwanted targets and put to death.

14 The medium The second event relates to the first: strife and war on the peninsula. Koguryŏ destroyed the Lelang Commandery in 313 CE and the Daifang Commandery in 314 CE. The Three Han (馬韓Mahan, 辰韓Shinhan, 弁韓Benhan) then began to gel into larger units: Paekche, Silla, Mimana-Kara. It is highly probable that with conflict and chaos on the peninsula, there was an influx of immigration toward Wa territory. These events suggest that the Wa were not in an environment where literacy could spread beyond a few peninsular officials. Unger (2005:100) postulates that para-Japonic speakers on the peninsula who had survived wars with Koguryŏ took their skills and made their way to Wa on the Japanese islands. These military skills would have allowed these immigrants to wield great social influence. I believe the Wa on the peninsula were closely integrated with the Mimana-Kara Federation, allowing the creation of what I call “the Wa sphere,” bridging the Wa on the peninsula with the Wa in Kyūshū. Park (2018:139–59) demonstrates through archaeological data that the Wa were closely connected to the Kara states (his Kaya) in the third through the fifth centuries. In other words, the connection with Paekche comes later, and scholars should not look exclusively to Paekche for early cultural exchange. Park (2018:174) concludes, Although there had been continuous political interchange between Paekche and Wa from the late fourth century, considering the peninsular-produced artifacts in the Japanese archipelago and the archipelago-produced artifacts in the Korean peninsula, the major player in the interchange … was clearly Kaya before the fifth century. I argue that Yamato’s unsettled situation begins to shift when Paekche requests diplomatic ties with the Wa. While Paekche is credited in the myth-histories as the teachers of Chinese to the Wa (cf. Bentley 2001a), it seems clear that these Paekche teachers were not starting from scratch. What does seem plausible is that the Paekche taught the small circle of people in Yamato (those in the capital areas) how to process Chinese in a manner that they had developed, or possibly had learned from teachers at the commandery. Ri (2005:40–44) introduces evidence, such as the style of writing, abbreviated characters, as well as terminology, that clearly came from Koguryŏ and was transmitted to Paekche and Silla. He claims that Koguryŏ was a pipeline of literacy for the kingdoms to the south. However, it is also plausible that literate Han Chinese from the collapsing commanderies fled into the neighboring territories and were employed as teachers. If true, then we could say that Lelang supplied these teachers to Koguryŏ, while Daifang supplied teachers to what would later become Paekche and Mimana-Kara. It also cannot be overemphasized that Paekche at this time (late fourth to early fifth centuries) was a powerful, cultured kingdom. After the destruction of Lelang and Daifang, it is strange to relate that Koguryŏ simply decided to support the remnants of these commanderies, instead of replacing these with a truly native organization. This suggests that the connection with Chinese culture and technology was highly valued. As Yeo (2013:199–200) argues, Koguryŏ was interested in expanding its territory and made the fateful decision to try conquering

The medium

15

the Former Yan territory in the Liaodong region. This left Koguryŏ’s southern border open to attack by an aggressively expanding Paekche. The destruction of Daifang at the hands of Koguryŏ allowed Paekche to consolidate its power among the Mahan polities, organize a formidable army, and become a powerful kingdom within less than a century. Kwon (2008:71–78) notes that the excavation of the P’ungnap Walled Site in the southern section of Seoul has put to rest the debate about where Paekche’s capital of Hansŏng had been located. He reveals that the work of excavation has uncovered remains of a wall that was 11 meters (36 feet) high, 43 meters (141 feet) wide at the base, and likely extended 3.5 kilometers (2.2 miles) around the capital. To construct an enormous wall on such a scale would have required a great number of laborers. Excavation of a variety of artifacts demonstrates “the magnificence of the aristocratic culture.” Thus, Paekche had the cultural and technological prowess to be an important influence on the ruling elite in Wa.

Wa on the peninsula It is helpful to take a slight detour to consider the situation of the nebulous ethnic group called “the Wa” during these formative centuries. The designation Wa (倭) is vague, and while many scholars make the hasty assumption that Wa = Yamato = Japan, this is a tenuous association at best. The greatest hindrance is the nature of the label. Wa is an ethnonym, but other labels like Yamatai / Yamato are toponyms (geographic labels). I believe that the ethnic group labeled Wa in the various records referred to a group of people who occupied an area from the southern area of the Korean peninsula down into the Japanese islands (Bentley 2008a). A brief examination of the 12 place names of Benhan is illuminating. Weizhi records these 12 names, minus the added name of Benshin (弁辰), which is tacked on to the beginning of each of these names: Mie-liai-mie-toŋ (彌離彌凍国), Tsiap-dɑ (接塗), Kɔdzi-mie-toŋ (古資彌凍), Kɔ-dzuin-dze (古淳是), Pɑn-lɑ (半路), Gɔk-nɑ (樂奴), Mie-Ɂɑ-ja-ma (彌烏邪馬), Kɑm-lɑ (甘路), Ko-ja (狗邪), Tso-dzou (走漕馬), An-ja (安邪), and Dok-liɑ (瀆盧). As the record gives no clue to what the semantics of these place names were, it is difficult to tease out the language, but on the surface, at least, the language appears to be Japonic in nature. Consider Table 1.1. In just these 12 place names, a pattern is visible. Several places include the suffix ra or na, which I believe means “land.” The alternation between r and n is a feature in several languages, including it being a very old feature of IndoEuropean.4 It is interesting that kokuna may be composed of koku “Koguryŏ,” a form that appears as a gloss in Nihon shoki, plus land. Clearly, there is not a Koguryŏ-administered land that far south this early, but it is possible that immigrants from the north settled in this area, resembling modern names like “Little Italy.” It is also very plausible that the Shoki identification of koku with Koguryŏ comes from the fact that this place name was later conquered by Koguryŏ, and originally was not related to that kingdom. I also theorize that yama and ya are related to the Japanese “mountain” and “valley.” While these data are quite tenuous, this one piece of evidence buttresses

16

The medium

Table 1.1 Possible Japonic words Later Han

Transcription

Possible meaning

Mie-liai-mie-toŋ Tsiap-dɑ Kɔ-dzi-mie-toŋ Kɔ-dzuin-dze Pɑn-lɑ Gɔk-nɑ Mie-Ɂɑ-ja-ma Kɑm-lɑ Ko-ja Tso-dzou An-ja Dok-liɑ

me-ri-me-tomu si-pu-ta ko-di-me-tomu ko-di-nze pa-no-ra koku-na me-a-ya-ma kamu-ra ko-ya so-zo a-ja or ano-ja toku-ra

Meri-woman-wed Paralyzed-field Kudi-woman-wed Shallow stream of Kudi Land of Pa Koku-land Woman-foot-mount Divine-land Basket-valley Ten-X Foot-valley Destination-land

what has been said above regarding the Wa still inhabiting part of the peninsula and being related to the Wa of the Japanese islands. Another important piece of evidence is found in the inscription on the Kwanggaet’o Stele (dated 414 CE). Parts of this inscription have been damaged by erosion and human mistreatment. Wang (1984) has examined not only several surviving rubbings but also the stele itself. I provide a partial translation, which follows his analysis of the inscription. This translation covers the section of the inscription that outlines the military achievements of King Kwanggaet’o, known in life as Eiraku (永楽), or Wanglek in the late fourth century. In the seventeenth generation from [King Toru] we come to the reign of Grand King “Opening Great Tracts of Land, [ruling in] Harmony and Peace, the Handsome King buried on the Hill [overlooking] the Country.”5 He ascended the throne in his eighteenth year. His title of respect was Grand King Wanglek6 [遝至十七世孫國罡上廣開土境平安好太王二九登柞号 為永樂太王・・・]. The words of the inscription say that in the fifth year of Wanglek’s reign, a yiwei year (395), the king went himself at the head of his army to attack Pie-le [to rescue hostages] that Pie-le refused to return. He crossed Mount Pə, and with the mountains behind him, he arrived at the upper reaches of the Jam River. Here he destroyed three villages and 600 or 700 encampments of the Pie-le, capturing innumerable oxen, horses, and sheep. The king passed through Siɑŋ-bɨɛŋ District and headed east. He reached the walled cities of X-li, Lɨk, and Pək-phuŋ. The king had his men prepare a hunting competition. Hunting as they rode through the area, the king returned to the capital [其詞曰永楽五年歳在乙未王以碑麗不歸□人 躬率住討過富山負山至鹽水上破其三部洛六七百營牛馬羣羊不可稱 數於是旋駕因過襄平道東來□城力城北豐王備猟遊觀土境田猟而還].

The medium 17 From ancient times Paekche and Silla have been vassal states of our country, and they have presented tribute at our court, but from the xinmao year (391) onward the Wa came, crossing the sea, and destroyed Paekche, XX, and Silla. The Wa made these [three states] vassals, so in the sixth year of the king’s reign, a bingshen year (396), the king himself led a naval armada and conquered Paekche. Our great army reached the southern border of Paekche and closed in on the capital city of Paekche, having captured the walled cities of [50+ cities listed]. The bandits [Paekche army] would not surrender to the righteous army [Koguryŏ] but dared to counterattack. Incensed at this, the king led his army across the A-li River, sending his forward troops to put pressure on the [Paekche] capital. The soldiers of [the Paekche] flank fled back to their holes. Our advance on the capital caused the king of Paekche great distress, so he presented 1000 men and women, and 1000 bolts of fine cloth as tribute. These were returned to the king with an oath, “Henceforth and forever we will be your bondservants.” The grand king granted an amnesty and pardoned the king to teach him discipline, and then recorded the sincerity of the allegiance of the king. Then he bestowed on the king 58 walled cities and 700 villages (that had been captured). The king also took the younger brother of the Paekche king, as well as ten ministers (as hostages) and returned to the [Koguryŏ] capital. In the eighth year, a wuxu year (398), the king dispatched a battalion to Bak-dzin-thɑ Valley on the border to observe [Paekche]. Eventually they plundered and kidnapped 300 men and women from Mak-sie-lɑ and Ka-thɑ-lɑ Valley. From this time on Paekche finally presented tribute and became obedient [百殘新羅舊是屬民由來朝貢 而倭以辛卯年來渡海破百殘□□新羅以爲臣民以六年丙申王躬率水軍 討伐殘國軍至窠南攻取…其國城賊不服氣敢出迎戰王威赫怒渡阿利水 遣刺迫城橫兵帰穴就便國城百殘王困逼獻出男女生白一千人細布千匝 歸王自誓從今以後永為奴客太王恩赦先迷之御錄其後順之誠於是得五 十八城村七百將殘王弟並大臣十人旋師還都八年戊戌教遣偏師觀帛慎 土谷因便抄得莫新羅城加太羅谷男女三百餘人自此以來朝貢論事]. In the ninth year, a jihai year (399), Paekche broke its oath and allied itself with the Wa. The king toured the area of Bɨɛŋ-naŋ, when an envoy sent by Silla appeared and told the king, “The people of Wa are flooding into the borders of our land, destroying the walls and moats of our cities, and subjugating our people. The king of Silla pledges his allegiance to Koguryŏ, awaiting the king’s instructions.” Later the grand king had mercy on Silla and praised their loyalty. He sent a special envoy to Silla to relay his plans. In the tenth year, a gengzi year, the king dispatched 50,000 infantry and cavalry to aid Silla. The people of Wa had filled the area between the walled cities of Nəm-kɨɑ and Silla. As soon as our army arrived, the Wa bandits began to retreat. Our army attacked the retreating Wa and drove them as far as the walled city of Dzioŋ-bat [belonging to] Nim-na Ka-lɑ (Mimana-Kara). 7 It quickly surrendered. This walled city was handed over to Silla for them to defend.8 Our army attacked the walled cities of Silla and Jam, filled with the

18 The medium Wa, and the Wa bandits suffered a great loss. Ninety percent of the Silla people in the walled city refused to go with the Wa, [so after the Wa were driven out] the Silla army was given responsibility for the defense of the city. The walled city of Silla… [九年己亥百殘違誓與倭和通王巡下平穰而新羅遣 使白王云倭人滿其國境潰破城池以奴客為民歸王請命太王恩後稱其忠 誠特遣使還告以□計十年庚子教遣步騎五萬往救新羅從男居城至新羅 城倭滿其中官兵方至倭賊退自倭背急追至任那加羅從拔城城即歸服安 羅人戍兵拔新羅城塩城倭滿倭潰城内十九盡拒隋倭安羅人戍兵新羅城 □□其□□□□□□□言]. XXXX… [T]he remaining Wa bandits fled. Our army occupied the walled city of X, and it was handed over to Silla to defend. Previously the king of Silla did not pay tribute or follow the commands of our court. … Grand King Kwanggaet’o to the king of Silla … finally appeared at court with tribute. In the fourteenth year (404), a jiachen year, the Wa were underhanded and invaded the borders of Daifang, joining forces with Paekche, [the two forces attacking] the walled city of Dzak. … Our king took command of the army and led a counteroffensive. He departed from Bɨɛŋ-naŋ … and the lead troops engaged the enemy. The king stopped the advance of the enemy, inserting himself into their ranks, cutting them down. The Wa bandits suffered another loss. Countless numbers of the enemy were cut down [□□且□□□□□□ □□□□□□□□□□□□□□□□□□辭□□□出□□□□□□□ 殘倭潰逃抜□城安羅人戍兵昔新羅寐錦未有身來論事□□□□広國岡上 廣開土境好太王□□□□寐錦□□僕句□□□□朝貢十四年甲辰而倭 不軌侵入帶方界和通殘兵□石城□連船□□□王躬率住討從平穰□□ □鋒相遇王幢要截盪刺倭寇潰敗斬殺無數]. In the seventeenth year, a dingwei year (407), the king dispatched 50,000 infantry and cavalry. … The army surrounded the enemy on all four sides and engaged them in battle. We annihilated the enemy’s army and the spoils numbered 10,000 pieces of armor, and countless numbers of provisions and weapons. During the campaign, our army captured the walled cities of Sa-ko, Lio, Guan-dio. … In the twentieth year, a gengxu year (410), the eastern Puyŏ had originally been a vassal state of King Tsho-mu, but later they rebelled and ceased presenting tribute at our court. The king took command of the army and went to attack eastern Puyŏ. When the army reached the capital of Puyŏ fear spread throughout the city and they surrendered [十七年丁未 教遣步騎五萬□□□□□□□□王師四方合戰斬殺湯盡所獲鎧鉀一萬 餘領軍資器械不可称數還破沙溝城婁城還住城□□□□□□那□城廿 年庚戌東夫餘舊是鄒牟王屬民中叛不貢王躬率往討軍到餘城而餘城國 駢服□□□□□□那□□]. Regardless of how one may wish to interpret the information here, clearly the writers of the stele inscription have engaged in exaggeration and fabrication. Having said that, there are still facts that we can corroborate. The record in Samguk sagi preserves information that the Wa had harassed Koguryŏ repeatedly for decades. The stele preserves the fact that the Wa were active on the peninsula. Based on

The medium 19 shared excavated artifacts, Park (2018:142) argues that the Wa invaded Silla and Koguryŏ in conjunction with the military might of Kŭmgwan Kaya (金官伽倻), which may have been the nucleus of the Kara Federation. Historically, we know that Paekche and Koguryŏ were often at odds with each other. We also know that Paekche had a connection with Yamato, initially through the Mimana (Imna) Federation, which Park (2018:154) argues was located in present-day Hadong District, thus sandwiched between Paekche and Great Kara (大加羅). Is it too far-fetched to believe that Paekche opened diplomatic ties with Yamato just before the beginning of the fifth century, because Yamato’s cousins, the Wa on the peninsula, had demonstrated that they could offer a counterweight to aggressions by Koguryŏ and Silla? Accepting that Paekche earlier had diplomatic ties with Wa on the peninsula allows a smoother transition to opening diplomatic ties with Wa (= Yamato) of the Japanese islands, as well as explaining why Paekche would have reached out to Yamato at all. If Yamato never had any interest in the peninsula, aside from gaining passage through the land to the commandery, then why would Paekche reach out to them? Would it not make more sense to have Paekche reach out to Puyŏ or another polity north of Koguryŏ instead? I thus contend that as Paekche absorbed Mimana to create a buffer with Silla, Paekche also turned its eyes for help to the Wa living on Kyūshū and other parts of the Japanese archipelago. This connection is facilitated by the existence of a figurative bridge between the Wa on the peninsula and those in Kyūshū. This bridge between the insular and peninsular Wa was created by people of Wa traveling back and forth between these two points. Park (2018:148–50) has also contended that Silla was an ally of a rather short-lived Yamato dynasty, perhaps Ingyō’s, but Silla withdrew after it had successfully fought off Koguryŏ. With the advent of the Yūryaku dynasty, Paekche filled the void left by Silla. Having both kingdoms reach out to Yamato suggests a strong connection with the peninsula and the archipelago. Yamato’s ready willingness to cement ties with Paekche would have been informed by those Wa who had dealings with the Paekche. Through these Wa intermediaries, Paekchean culture and technology would flow more readily to Yamato. This vital connection is glossed over by Kojiki and Nihon shoki, partially because of nationalistic feelings, but perhaps more directly because the Wa were instruments in this transition, giving rise to the subsequent perception that Yamato was more important in the equation than Paekche may have originally felt. What is clear is that Paekche was the catalyst for the expansion of literacy among a small circle of people at the Yamato court, the foundation having likely been laid by Kara and Silla, before and during the Ingyō Dynasty. Many have speculated over the years that Paekche taught the Japanese how to write, as the legend in Kojiki and Nihon shoki illustrates, but there is a question few ask: what is being taught? As I (Bentley 2001) have argued, the writing system that early Japan adopted was a hybrid system taught by members of the educated echelon of the kingdom of Paekche to a small group of elites in Yamato. The general scholarly consensus seems to be that there were two competing systems in

20 The medium place: writing in what we can broadly call Classical Chinese (or Literary Chinese, including kundoku), and writing in a system much like the Japanese phonographic system called man’yōgana. However, this is an overly generic categorization, and the actual system as preserved on surviving documents is more homogeneous. It is helpful to reconsider some of these basic data. Below are six examples found on artifacts preserved over the course of roughly three centuries, with translations and explanations of the texts. These examples demonstrate the Chinese-based system that was already in place throughout much of East Asia. The first two examples come from Weizhi. The second set of documents is inscriptions from two swords: the Inariyama Tumulus sword and the Eta Funayama Tumulus Sword. The Inariyama sword was discovered during the excavation of the tomb located in Gyōda of Sakitama City in Saitama Prefecture. In 1978 when the sword was being cleaned, a technician noted a design on the heavily rusted sword blade. Further investigation revealed a gold inlaid inscription. The Eta Funayama sword was excavated from a tomb in the Nagomi Machi area of Kikusui in Kumamoto Prefecture in 1873. Another inscription is found on a mirror, known as the Suda Hachiman Shrine Mirror because it was kept at the Suda Hachiman Shrine located in Hashimoto City in Wakayama Prefecture. The final document is an inscription on an artifact called the Seven-Branched Sword, which historically was kept as a treasure by the Isonokami Shrine located in Nara. The first two examples provide a benchmark, demonstrating the default system (Chinese) in the third century, and the system that surrounding kingdoms would have imitated. The first quote is from a subsection titled “Han” in the “Eastern Barbarians” chapter (Fascicle 30) of Sanguozhi: (i) The *Dźɨn9 King rules the land of *nGyɑt-ke and has the title of *Gin-ṭeh or they [the people] lengthen [the name] and call him by the title *Gin-wunkhianɁ-ke-pouh-Ɂɑn-ja-tsuk-ke-bun-gin-liɑi-ńe-pu-liaj-kou-ya-dzin-ke-liem [辰王治月支國、臣智或加優呼臣雲遣支報安邪踧支濆臣離兒不例拘邪 秦支廉之號].10 This first example nicely demonstrates how a writer could insert a long string of Chinese characters employed as phonograms into an otherwise Chinese text, and the Chinese-educated reader would naturally analyze these as representing individual sounds and not semantic strings.11 What I find interesting is that the longer name is likely a string of attributive clauses, much like Native American chiefs who had names such as “He who guards the gate of sunset.” This type of naming is also true of Ainu. To demonstrate the attributive nature of this name, I provide an analysis of the constituents of the name. The following analysis of this name is based on the assumptions in Bentley (2008:3), where I noted, “The possibility that the Wa had a literate person who transcribed these toponyms and titles for the Chinese envoys needs to be seriously entertained.”12 How much more so in relation to the titles and toponyms on the peninsula.

The medium

21

I present the following as a tentative, initial attempt at an interpretation of this title, demonstrating that it actually preserves a strain of a language that is related to the later languages of Japanese and Ryūkyūan.13 In the analysis that follows, I give (i) the orthography followed by (ii) the Later Han Chinese reconstruction (LHC), then (iii) a reconstruction of the pre-Old Japanese form, (iv) a grammatical analysis, and (v) finally an English rendition (Table 1.2 Reconstruction of *Dźɨn King’s name). Thus, I propose that this title is composed of three attributive clauses: “The father who shines as a seer, a type (example) of the ruling seat, a drill that is excellent (though) dull (because of his age?).” This analysis is based on several considerations. As this polity is in close proximity to the Chinese commanderies, it is reasonable to assume that there are some Chinese loans included. *Gin (臣) “seer,” I believe, is one of these loans. “Father” anya is preserved as aya in southern Ryūkyūan languages like Ishigaki, Uehara on Iriomote, and Kohama Island. It is surely a related form of oya “parent” found in the central dialect. I posit that the form suke “rule” is the infinitive of a variant of siku, demonstrating a phenomenon of high vowel backing found in the southern dialect of Tsushima: suzuku “drop” instead of sizuku, or sutadari “trickle down” instead of sitatari.14 I interpret bunki to be a form of “sit”’ in a nominalized form, meaning “seat,” *bu- > wi-. I also interpret kerem to be “drill,” under the assumption that the current word kiri “awl, drill” which belongs to accent class 2.4 originally belonged to accent class 2.5. Vovin (2008:141–156) and others have argued that this accent class is rare in that it obtained its register through the loss of a final nasal (-m). The second example comes from a description of an area on the Japanese islands, and it appears in the subsection 倭 “Wa” from the same chapter: (ii) “Continuing south you arrive at Yamatə, which is where the queen has her capital” [南至邪馬臺國、女王之所都]. Here the capital of the queen’s federation is represented as 邪馬臺, *yamatə, clearly representing Yamato.

Table 1.2 Reconstruction of *Dźɨn King’s name LHC p-OJ Gram Engl LHC p-OJ Gram Engl

臣雲遣支報安邪 *ɡin-wun-khianɁ-ke-pouh-Ɂɑn-ja *gin-wun-kangepo-anya seer-ACC-shine-ATT-father Father who shines as a seer 兒不例拘邪秦支廉 *ńe-pu-liaj-kou-ja-dzin-ke-lem *nepuria-koyazi-nə-kerem dull-INF-excellent-DV-drill The drill that is dull-excellent

踧支濆臣離 *tsuk-ke-bun-gin-liɑi *suke-bunki-nə-riai Rule-INF-seat-DV-type A type of a ruling-seat

22

The medium

The next two examples are from inscriptions found on the blades of swords, likely dating from the same fifth-century period, as both contain the same name of the king. First the Eta Funayama tumulus sword inscription: (iii) During the reign of Great King *Wak-XXX-rwo who ruled all under heaven, a man named *Mu-ri-tei who served in the Bureau of Directors used a great iron kettle and during the eighth month forged together [the iron] and made a long sword four shaku in length. The metal was tempered 80 times, and the blade brandished 90 times. It has a three-inch well-shaped blade. The person who girds himself with this sword will have long life, his posterity will be great in number, and the owner will obtain the favor X (of the king? of heaven?). He will not lose that which he governs. The name of the swordsmith is *I-ta-wai (or *I-ta-ka), and the author of the inscription is *Tjɑŋ-Ɂan [治天下獲□□□鹵大王世、奉事典曹人、名无利弖、八 月中、用大鉄釜、并四尺廷刀、八十練九十振、三寸上好刊刀、服此 刀者、長寿子孫洋々得□恩也。不失其所統。作刀者、名伊太和 (加) 、書者張安也].15 The second example of a sword inscription is the Inariyama Tumulus sword inscription, which contains 115 characters inscribed proportionately on both sides of the blade16: (iv) [obverse] Recorded during the seventh month of a xinhai year. *Wa Wa-ka OMI,17 whose distant ancestor was named *O-po-pi-kuai, whose son was named *Ta-ka-ri Swo-ku-nei, whose son was named *Te-yo-ka-ri Wa-ka, whose son was named *Ta-ka-pai-si Wa-ka, whose son was named *Ta-sakui Wa-ka, whose son was named *Pa-nde Pi-(kuai), [inverse] whose son was named *Ka-sai Pai-ya, whose son is named *Wa Wa-ka OMI. Through the generations, we have served the court down to the present as Chiefs of the sword bearers. During the [reign of the] court of Great King *Wak-ka Ta-kye-rwo, when he resided in the *Se-kui Palace, I assisted him in governing everything under heaven. I had this hundred-times-tempered sharp sword forged and record this as testimony of my service [[obverse] 辛亥年七月中 記乎獲居臣上祖名意富比垝其児多加利足尼其児名弖已加利獲居其児 名多加披次獲居其児名多沙鬼獲居其児名半弖比□ [inverse] 其児名加 差披余其児名乎獲居臣世々為杖刀人首奉事来至今獲加多支鹵大王寺 在斯鬼宮時吾左治天下令作此百練利刀記吾奉事根原也]. The fifth example is the inscription on the Suda Hachiman Shrine mirror, and I freely admit that the interpretation of the inscription is somewhat tenuous:

The medium 23 (v) On the tenth day18 of the eighth month of a gui-wei year, during the harvest of the Great King, when the king’s younger brother, the prince, resided in the *O-se-sa-ka Palace, *Se-ma pondered on a long life,19 and dispatched two people, *Kəi-NA ATAPI and *Kəm-tsu-ri, a person of Wai, and had this mirror made from 200 han of the finest copper [ 癸未年八月日十大王年男 弟王在意柴沙加宮時斯麻念長寿遣開中費直穢人今州利二人等取白上 同二百旱作此竟20 ]. One of the names in this inscription, 開中費直 has been difficult to decipher. Based on the following considerations, I believe this represents someone from Paekche. The title Atapi (費直) is believed to be a continental precursor for the Japanese version of 直. Notice that a mokkan excavated from the Asuka Ike dig contains the name 大＿費直伊多 Ō-X ATAPI Ita. I have interpreted the first two characters, 開中, as Kena (Keyna). The name of a general sent to fight on the peninsula appears in Nihon shoki, known as Kena Omi of Afumi. As this name does not appear in Kojiki, I believe he is from a peninsular family that immigrated to Yamato. Reading 中 as na appears in the beginning of the Tenmu record, where the name 天渟中 has a gloss “This (渟中) is to be read nu-na” [渟中、此云農難]. The sixth and final example is the inscription on the Seven-Branched Sword sent from the Paekche court to open diplomatic relations with Yamato: (vi) [obverse] On the sixteenth day (an auspicious day) of the eleventh (twelfth?) month of the fourth year of Tai (he? shi?), at high noon we forged this hundred-times-tempered iron sword with seven branches. With this sword, you may avoid the ravages of war. We bestow [this sword that is worthy for all] ministers and princes. It was crafted (at XXXX).21 [inverse] From the time of previous ages, there has never been a sword like this. As Heir to the Paekche throne, I was born with a wise, mystic power.22 Thus, I caused this sword to be carefully crafted for the King of Wa. May it be handed down for future generations [ [obverse] 泰□四年十□月十六日丙午正陽、造百錬銕七支 刀、出辟百兵、宜供供侯王□□□□作 [inverse] 先世以来、未有此刀 、百濟王世子、奇生聖音、故為倭王旨造、伝示後世23]. First, it is helpful to place these six texts within a concrete chronological framework. The first two examples from Sanguozhi date to around the end of the third century. The sixth example from Paekche very likely dates to the year 369,24 which corresponds to the fourth year of Taihe—based on the Eastern Jin calendar—and matches closely with the record in Nihon shoki, when it is recorded that

24

The medium

Paekche presents a Seven-Branched Sword to the court.25 This is dated the 52nd year of Jingū (252), and then if we delete the 120-year (backdated) manipulation, we are left with the date of 372. The three-year gap at first blush seems anomalous, but it is possible that there was a three-year gap between the forging of the sword (and the imprinting of its date) and the actual presentation of the sword at the Yamato court. Then, as now, diplomacy can take time. It should be noted, however, that Shinzō (2005:187) argues that the sword was forged in Eastern Jin when Paekche opened diplomatic relations in 372. The court then backdated the inscription to 369. The Eta Funayama text has no year designation, but the sword is contemporaneous with the Inariyama Tumulus sword and its inscription. There are several theories as to which year the date xinhai corresponds. The two theories with the greatest support from scholars are for the year 471 or 531. There is a strong consensus that the Great King corresponds to Yūryaku as noted in Nihon shoki. If we delete the place where he was born from his name (Hatsuse), then his recorded name is Wakatake. It is plausible that the name on the sword is an older, infinitival form of later waka takeru, “the brave young man.” Nihon shoki has him reign from the years 457 to 479. The other option, 531, would correspond to Keitai (r. 507–531) or Ankan (r. 540–571), though some, like Mizuno Yū (1954, 1968:142–47), suggest that Ankan’s reign be recalibrated, starting in 528 to 535. The major obstacle to this theory is the disparity in the name: Wohodo (for Keitai) or Hirokuni Oshitake Kanahi (for Ankan). Based on these considerations the date of 471 is the most appropriate. Most scholars connect the Suda Hachiman mirror inscription to one of two dates: 443 or 503. If the personal name Sema in the inscription above is the same person named sema recorded in the fourth year of Buretsu in Nihon shoki, which corresponds to King Muryŏng of Paekche (r. 501–523), then the date 503 is more secure. The major difficulty with this association is the lack of any royal designation (王) in the inscription; however, the sloppiness of the inscription may suggest more than just malformed and incorrect characters—it might also suggest a dropped character or more due to poor judgment regarding spatial limitations on the face of the mirror. Lurie (2001:158) has already noted, “[T]he designer did not take into account the left-right reversal of the casting process.” Even if this evidence is set aside as too speculative, Lurie (2011:379) later notes the scholarly consensus is that the date corresponds to 503. Creating a timeline makes it easier to visualize the continuity of Classical Chinese throughout the “Wa sphere” (Figure 1.2). Sanguozhi

285

Seven-Branched Sword

369

Figure 1.2 Timeline of Classical Chinese

Eta Funayama / Inariyama

Suda Hachiman

471

503

The medium

25

Several points should be noted. First, each of these six texts is written in Classical Chinese, not surprising since the first two come from the Jin Court, but the other texts cited demonstrate that this tradition of writing in Classical Chinese continues across East Asia into the fifth and sixth centuries in Japan. Second, within these syntactically ordered Chinese texts, phonographic spellings are inserted in almost all cases.26 This provides evidence that each text is intended to be read in Chinese. This is especially true since the Seven-Branched Sword is sent from non-Chinese country A to non-Chinese country B, and Classical Chinese was the lingua franca of the time. In relation to the inscription on the Seven-Branched Sword, Yoshida (2001:52–53) has argued against the theory for reading 奇生 as a phonographic rendering of a name by stating that putting phonographic script in a diplomatic text intended for a foreign party went against diplomatic protocol set up according to Chinese custom.27 It would have been considered arrogant to expect the recipient of the sword to read a foreign name, and that is why the text is written in Chinese and contains so many auspicious phrases. Third, as Lurie (2001:142–43) has pointed out, it is highly significant that the name of the great king is represented with the same set of phonograms on both sword inscriptions (Eta Funayama and Inariyama), reminiscent of the somewhat consistent spellings found in Sanguozhi (cf. Bentley 2001). These data lead us to conclude that there is really only one form of writing at work here, at least until perhaps the middle of the sixth century: Classical Chinese. How that is unpacked in the mind of the readers of each country is a different matter, but for the time being it is difficult to argue against the fact that we are dealing with Chinese as the linguistic medium. There have been some, like Hong (1988:98, 1994:258), who have argued based on the work of Miller and Murayama (1979) that the Inariyama Tumulus sword inscription is actually written in idu (吏讀) “clerk readings” or a form of early Paekche-ized Chinese. Lee and Ramsey (2000:54) note that idu “began as subtle Koreanizations of Literary Chinese, with a few Korean elements added to Chinese-language texts.” This appears to be what Hong means, as Miller and Murayama state that the inscription contains “Koreanisms … blatant Old Paekche Koreanisms” (1979:422, 424). A careful study of Chinese, as well as recent research, however, casts doubt on these earlier conclusions. One of these supposed “Koreanism” is found in the date at the beginning of the inscription, 七月中, which they translate as “in the seventh month,” a usage that also appears in other peninsular inscriptions, as well as a few times in Nihon shoki. However, a careful examination of texts written in Classical Chinese reveals that this is actually a Chinese usage and those on the peninsula simply adopted it.28 Consider the following passage from the Gaozu Annals Chapter of Shiji, the first of China’s official histories (ca. 100 BC): In the middle of the eleventh month, Hsiang Yü as expected led the troops of the feudal lords west, intending to enter the Pass, but the gates to the Pass were closed…. In the middle of the twelfth month, he finally

26

The medium reached Hsi (translation from Nienhauser 1994, 2:39–40) [十一月中， 項羽果率諸侯兵西，欲入關，關門閉…十二月中，遂至戲].

We have two examples of this usage in the space of one paragraph. Having said that, there are several illustrative examples where people on the peninsula took a Chinese grammatical function and altered one point to fit their own writing, and this later was transmitted to Japan. For example, take the grammar point concerning the use of 之 after a verb to represent an object, often a referent equivalent to English “it,” “him,” or “her.” The following is an example from the first chapter of The Analects: 先王之道斯為美，小大由之。 The Way of the Former Kings was beautiful in this: in small things and great they followed it. 有所不行，知和而和，不以禮節之，亦不可行也。 If there is something that does not go right, one should recognize the principle of harmony, then it will be harmonious. But if it is not moderated by ritual itself, it still will not go right. (Brooks and Brooks 1998:191) Because this marker had a strong syntactic tendency to appear at the end of clauses and sentences, peninsular writers began to use the graph to mark the end of a sentence, much like the Western use of a period. Inukai (2014:126–27) mentions two examples that I quote here. The monument marker for the rebuilding of the Namsansin Fortress begins with this sentence: 辛亥年二月廿六日南山 新城作節如法以作後三年崩破者罪教事為聞教令誓事之。This introductory sentence announces the vow the builders and laborers made to receive divine punishment if their work were to crumble to the ground within three years from the (start?) date, the 26th day of the second month of 591 CE. Notice that 之 follows the nominalizer 事, which is not a verb. The use of 之 here can only mean “end of sentence,” not a referent in a predicate. The other example is from a mokkan excavated from the Morinouchi excavation site in Shiga Prefecture: 椋□伝之, where Kura X (Atahi?) relays his report. His report then follows. Here 之cannot be “it,” because nothing has been mentioned yet that could be referred. Again, it functions as a form of punctuation. Inukai (2014:129) also demonstrates that the graph 鎰 underwent a semantic change. In China, this graph represented a measure of the weight of the metal, being between 20 and 24 liang of metal, roughly one pound today. When the usage of this graph was transmitted to the peninsula, it underwent a semantic change and came to mean “lock,” as one 鎰 of metal was required to forge a lock. The use of this graph was later imported into Japan, and the semantics expanded further so that now this graph is used to mean both “lock” and “key.” Based on

The medium

27

the expanding semantics, the direction of the loan is thus China → Paekche → Yamato. Next, Inukai (2014:130–31) introduces a wooden tablet excavated from the Sanri Temple ruins near the tombs of the Paekche kings. On that tablet is the designation 漢城下部對徳疏加鹵, a name: *Sio-ka-luo, eleventh rank, the lower section of the capital. Inukai interprets this name (疏加鹵) to be sugaru, but it is possible this reading changed after the family immigrated to Japan at the fall of Paekche to the combined armies of Tang and Silla. He notes that a tag with the name 須我流 Sugaru was excavated from a Nara site in the ancient Afumi Province. He believes that a group of Paekche immigrants had lived in Afumi. Notice also the variant writing of 鹵 has 夕 inside the box instead of ※. This variant graph is also found on both the Inariyama Tumulus sword and the Eta Funayama sword, again showing the influence of Paekche writing. Finally, Ri (2005:42–43) argues persuasively that the character椋, read kura “storehouse,” found above in the mokkan from Shiga, originally came from Koguryŏ through Paekche. This character appears in an inscription on the north wall of a tomb in Hwangnam. Ri mentions an interesting line in the “Koguryŏ” section in Weizhi, “Koguryŏ does not have large storehouses, but each house has its own small storage shed. These are called *bu-kɨaŋ” [無大倉庫，家家自有小倉，名之為桴京]. He then introduces an older theory by Inaba Iwakichi (1876–1940) that the Koguryǒ literati took this Chinese account and melded the two phonetic symbols together: 桴+京=椋 and used this new graph to refer to “storehouse,” because the character 京 also has the meaning of “storehouse.” This idea about graphic evolution may be true, but Nienhauser (1994:58, note 287) include the following comment about the line 家京下 “below the storehouse of the residence” found in Shiji in the note, “京 ‘storehouse’ unusual usage.” The graph may simply be a fusion of the two graphs, as Inaba thought. At any rate, the Japanese reading of kura was attached, likely because of the semantics. It is also plausible that the actual word kura “storehouse” is itself a Paekche loan. This rather large cache of excavated evidence demonstrates that the peninsular kingdoms, particularly Koguryŏ and Paekche, were educated in Classical Chinese and used it for communication and official record keeping a century or more before Yamato. The “Paekche Annals” in Samguk sagi quote from “an old record,” stating that during the reign of King Kŭnch’ogo (r. 346–375), the writing was introduced to the Paekche court by a literate person believed to be of Chinese extraction (Best 2006:258, n69), and the preservation of official records began at that time.29 Then in 384, a monk apparently from Central Asia introduces Buddhism to Paekche (Best 2006:79). Is it a coincidence that the order of acquiring an advanced literary culture in Yamato mirrors that in Paekche, at least according to a peninsular tradition? First writing is introduced, and then next Buddhism. This point is important, as Suishu mentions that because Buddhism was introduced into Japan, the Japanese became literate, suggesting a historical order that is reverse of that in the native records mentioned above.

28

The medium

Best (2006:78) speculates, “Certainly it is reasonable to assume that immigrants from the former Chinese commanderies to the north formed the main body of clerks and teachers who introduced Chinese literacy culture to Paekche.” Best (ibid.) even believes that the individual(s) sent by Paekche to the Yamato court was actually someone of Chinese extraction. He (2006:79) further argues, “In the long term, however, Buddhism provided a more effective incentive for the spread of literacy among the kingdom’s elite than any other aspect of adopted Chinese culture.” This method of cultural transfer is imitated again in Yamato, as Paekche immigrants30 in Japan likely were the earliest scribes at the Yamato court after the major movement toward unification had been completed, and then monks from Paekche monks came to the court with the importation of Buddhism. The key question we should address is: do we believe it is a coincidence that Yamato and Paekche have similar traditions regarding the origin of writing, or is the early Japanese record simply a rewriting of what happened in Paekche? Cheng (1988) has provided other archaeological evidence that strengthens the view that Paekche was the teacher of Literary Chinese to Yamato. She has examined a variety of excavated or discovered inscriptions and notes two noncontroversial trends. Both Koguryŏ and Paekche tend to abbreviate certain characters when they write, while Silla tends to combine two characters into one. One example of abbreviation is the graph 卩 to represent 部, found in instances such as左卩 and 後卩, abbreviations for 左部 “Left Circuit” and 後部 “Rear Circuit” designations of rank within the Paekche bureaucracy, where the capital was divided into five circuits (部, see Cheng 1988:50). This same abbreviation appears in an inscription found on the heavily damaged Okadayama Tomb Number One Sword. This sword was discovered in a burial mound within a group of tombs called the Okadayama Kofun-gun in Matsue City in Shimane Prefecture. The inscription is各田卩臣□□□素□, and it is believed that the first four characters actually represent the name 額田部臣 NUKATABE OMI; thus, the first and third characters are abbreviated. Another example is the abbreviation of the character 岡 “hill” as 𦊆. This abbreviation appears on both Koguryŏ and Paekche inscriptions and appears in an inscription from a bronze plate at Hōryūji that contains the phrase 𦊆王寺 Kataokaōji “The Kataoka Royal Temple.” Cheng (1988:52) notes the inscription includes the name of the Ōhara family, who were Paekchean by birth. Interestingly, the Silla tradition of combining two characters, such as writing 水 “water” and 田 “agricultural field” as one character (畓) to denote a rice field (that requires water 水+田) is a practice according to Cheng (1988:65) that is not found in Koguryŏ, Paekche, or Yamato. Lee Seung-Jae (2014:153–54) refutes this statement by arguing that a graph like 畠 “a cultivated field” is a peninsular invention and represents a dry, or “white” field (白+田). The Old Korean reading may have been *patak, resulting in later Middle Korean pàth, which entered Yamato as patake “field.” It seems plausible that both Paekche and Silla combined two graphs into one.

The medium

29

Over the past decade, other evidence has appeared in South Korea, with excavations from territory controlled originally by either Paekche or Silla. A variety of bamboo slips have been excavated from several tombs at the Laju Pok-am site, in an area historically controlled by Paekche. One bamboo slip excavated from this site has the date 庚午年, and based upon other evidence from the excavation site, scholars feel certain that this represents the date 610. Another slip found in tomb number two has the following text (example and explanation from Hirakawa 2014:58, 60): Line one: 兄將除公丁　婦中口二　小口四 Line two: □兄定文丁　妹中口一　定 Interestingly, the graph 定 “determined” is written larger than the rest of the inscription, apparently used much like a paid stamp in modern times. A similar usage is found in documents in the Shōsōin (Lee YH 2014:60). Another slip excavated from the Puyŏ Kuari dig site, near the ancient Paekche capital of Sabi (fl. 538–660), contains the following: 太公西美前部赤米二々. The first part is likely a name, but the usage of 赤米 “red rice” is also found on bamboo slips excavated from both Fujiwara Capital (fl. 694–710) and Nara Capital digs. Lee SJ (2014:159–610) brings up an exciting new discovery of a Paekche-ized graphic form also found in Nara excavations. It is the graph 石 minus the top stroke ¯. In cursive, this graph appears as 々, and Japanese scholars have misinterpreted this to be an abbreviation of 口 “mouth”; however, as Lee argues, this is actually a Paekche innovation, a counter for dry measurements. Lee SJ (2014:160) mentions the following example excavated from territory controlled by Paekche: 父送塩二々 … (from the Puyŏ Nŭngsan-ni Temple Site wooden tablet no. 10) “Father sends two seom of salt…” From the excavation site north of the dairi of the Nara Palace, number 477 (see Figure 1.3) is another example. The online “Mokkan database” has the text as 折櫃八合輦籠二口, but Lee insists that this actually is 祈樻八合輦籠二々, translated as “a prayer chest (in capacity of) eight hop, a palanquin (in volume of) two seom.” I alter this rendition to “a prayer chest (in capacity) of eight gopu and a palanquin (in volume of) two semu.” Lee notes that this graph appears to have been used as a counter for objects made of wood. Another assumed Paekche innovation is the counter 了. Lee SJ (2014:167) provides this example from a bamboo slip excavated from territory originally belonging to Paekche: … 處二了 (Kyŏngju Anapchi Wooden Tablet No. 19) “…prescribing two ryang of (medicine)…”

30

The medium

Figure 1.3 Nara Palace wooden slip. Kyujanggak Institute for Korean Studies. Courtesy of Lee, Seung-Jae, 2014. “Old Korean writing on wooden tables and its implications for Old Japanese writing.” Seoul Journal of Korean Studies, No. 27.2, pp. 151–185.

Compare this example with a long fragment from the Nara capital, dated the 17th day of the fourth month of 764 (Figure 1.4): 請縄参拾了 (Mokkan no. 11861), meaning, “We beg for 30 X of rope.” While the true reading of this measurement is unclear in either Paekche or Japanese, it is difficult to dispute that both countries are using the same counter. Finally, Lee SJ (2014:169) argues that the well-known repetition mark (々) is a Paekche innovation that was passed along to the Japanese. He cites the following example:

The medium

31

Figure 1.4 Nara era wooden slip. Kyujanggak Institute for Korean Studies. Courtesy of Lee, Seung-Jae, 2014. “Old Korean writing on wooden tables and its implications for Old Japanese writing.” Seoul Journal of Korean Studies, No. 27.2, pp. 151–185.

… 凡作形丶〻中了具 (Puyŏ Nŭngsan-ni Temple Site Wooden Tablet No. 7) “generally, being already prepared in the middle of forming shapes” There are many examples of this usage on mokkan from Japan, but an example from the ruins of the Fujiwara Palace (fl. 694–710) contains this fragment (mokkan no. 157): 須〻支 susuki “a type of fish, perch.” The evidence overwhelmingly demonstrates that many types of Paekche writing were transmitted to Japan. As Chinese is a graphic writing system where stroke order and proportions are critical in correct writing, it is interesting to find examples of writing practice

32

The medium

in Yamato. Wang (2014:276) mentions a fragment of writing practice in China dating from 869. It is thus fascinating that evidence of writing practice exists in Yamato from an earlier period than this. From the ruins of the Fujiwara Palace, workers have excavated a number of examples of writing practice preserved on wooden slips. Mokkan number 223 has the following text 而而而而而而, but some of the cursive characters here have the same style of writing as that found on the “King Sinhung of Silla border marker” at Changyeng (dated 561). Another mokkan (number 30) excavated from the Fujiwara Palace ruins clearly shows a person practicing writing 大 “big, great,” 道 “way,” and 實 “bear fruit,” where the wooden slip has 大大有大大大 大大大大道道道道道道道道道實實實實實實實實實實. The intended meaning is that “the great way (Buddhism) bears fruit,” but the writer has repeated characters to practice writing these. As far as taking Classical Chinese and altering the order of the constituents is concerned, consider what appears to be one of the earliest forms of writing on the peninsula, called sekichey (誓記體 “vow-writing”); the actual document is a small stone engraved with the following inscription: Sixteenth day of the sixth month of a renshen year (552 or 612 CE), the two of us together record this pledge. We vow before Heaven that from this time forward for three years we will maintain the way of loyalty. We vow that there will be no negligence, and if there are errors, we pledge to receive great punishment from Heaven. Even if the country should be uneasy and society in great chaos, we vow to gladly perform. Also, we separately made a great pledge last year, a xin-wei year (551 or 611 CE), on the twenty-second day of the seventh month. We pledge for three years to obtain (learning) of Shijing, Shangshu, Liji, Chuan, and Lun [壬申年六月十六日 二人并誓記 天前誓 今自三年以後 忠道執持 過失无誓 若此事失 天大罪得誓 若国不安大乱世 可容行誓之　又別先辛未年七月廿二日大誓 詩尚書礼 伝倫得誓三年]. This inscription reminds the reader of the pledge mentioned on the monument of the rebuilding of the Namsansin Fortress in relation to the workers vowing to receive retribution if their work should crumble within three years. What is interesting in this vow above is that many phrases have been altered according to what must have been the vernacular of the writers. Instead of grammatically correct Chinese 誓1无2過3失4, we have 過3失4无2誓1, which fits the syntax of Korean (Silla?), a verb final language. Notice how often the verb誓 “vow, pledge” comes at the end of these sentences. Another example is the phrase 忠道執持, where the verb 執持 “take and hold” is sentence final. This manipulation of Chinese to fit the vernacular syntax is also seen in a few older toponyms altered during the reign of King Kyŏngdŏk in 747 CE, as recorded in the geographical sections of Samguk sagi: 牛見県 “cow-see district” is altered to 目牛県 “look at-cow district.” Inaoka Kōji (1976:214) argued rather early on that Ō no Yasumaro’s choice to use a mixed system—Chinese and phonographs—in recording Kojiki was

The medium

33

not invented or developed by Yasumaro, but simply reflected his choice to use a system that already existed. The evidence as provided so far confirms what Inaoka originally argued. This system existed prior to Yasumaro. We can thus conclude that the evidence is demonstratively persuasive that Paekche instructed people at the Yamato court in the art of reading and writing.31 The key question is what Paekche taught to the students at the Yamato court? That Literary Chinese was taught is a foregone conclusion, but is that all that was taught? In an earlier paper, I concluded that Paekche had taught Yamato how to use Chinese characters as phonograms to represent the sounds of their own language (2001a:70), and Miyake (2006:200) argued that the Paekche tutors helped “create a new script for the Japanese incorporating elements from several sources.” There is a bigger picture that tends to be overlooked, however. For Paekchean teachers to educate the Yamato elite in writing their language in phonograms, they first had to be taught how to use the characters as they were originally intended: Classical Chinese. As noted above, the teachers from Paekche did not have to start from scratch. It is very probable that there were some, albeit a small number, who could already read and write Classical Chinese. The Paekche teachers built on this knowledge. Here I argue that Paekchean teachers employed a three-step educational process: (i) educate students in the knowledge of reading and writing basic Classical Chinese; (ii) instruct these students how to use certain semantically narrow characters as phonograms to represent Yamato names of people and places; (iii) teach these students how to convert Chinese into Paekchean. I will go so far as to claim that those at the Yamato court were educated in the language of Paekche, because Paekche was viewed as a more advanced civilization than Yamato, analogous to the early Hawaiians learning to write in English script, so they could represent their own language, which had no writing system. In learning to write English, they also learned the language. At a later date, the students in Yamato were taught how to convert Classical Chinese into Japanese, a process termed kundoku, or “reading by gloss.” To keep the Japanese version of kundoku clearly distinct within this discussion about Paekche, I will call the Paekche method initially taught to the group of elites in the Yamato capital kumanari, named after Ungjin 熊津, the capital in Paekche (fl. 475–538).32 This naming practice is much like the warriors and citizens of sixteenth-century Japan calling matchlock rifles tanegashima, because the Portuguese had introduced the weapons through the island of Tanegashima. My argument operates on the theory that teaching kumanari, the Paekche method of converting Chinese into the language of Paekche, to the people of Yamato was rather easy and logical, because the languages of Paekche and Yamato had similar syntactic structures.33 It stands to reason that the initial steps regarding the teaching of kumanari involved making basic associations with Chinese characters that had direct equivalents to Paekche/Japanese nouns and verbs. Lurie (2011:176) provides an apt example, using part of the first sentence to The Analects: 學而時習之 “To learn (something) and at times practice it” 不亦說乎 “is this not pleasant?” He notes that certain Chinese words corresponded cleanly with Japanese

34 The medium manabu

narapu

words: 學 “study” and 習 “practice.” Thus, rather simple sentences were easily converted into Paekchean (and then later into Japanese), as long as the student remembered that syntactically Chinese had the verb come after the subject but before the direct object, while Paekchean and Old Japanese (as well as the modern languages) has the verb at the end. Consider another example from the “Xia Annals” in Shiji: 予陸行乘車，水行乘舟 “When I travel on land I ride in a carriage; when I travel by water I ride in a boat.” This would be processed as Old Japanese ware kuga yuku noru kuruma; midu yuku noru pune, or something akin to ungrammatical English, “I land go ride carriage; (I) water go ride boat.” Below are two examples of how the Paekche likely did the same thing, based on textual evidence from Nihon shoki and Man’yōshū. First, a quote from Paekche sinch’an: King Mata’s behavior was unbecoming, and he oppressed the masses. The people joined forces and disposed of the king. King Munei ascended the throne. His real name is King Sema, and he is the son of King Konki. Thus, King Mata was a brother by a different mother to Munei. Konki went to Wa, and when he had reached an island by Tukusi, his wife gave birth to Prince Sema. He sent him home from the island and did not go to the metropolis. Since he was born on an island, he received the name of Sema. Now there is an island in the Sea of Kakara called Nirimu Sema. This is where the king was born. That is why the people of Paekche called the island “Lord Island” [末多王無道、暴虐百姓。國人共除。武寧王立。諱斯麻王。是琨支王 子之子。則末多王異母兄也。琨支向倭。時至筑紫嶋、生斯麻王。自 嶋還送、不至於京、産於嶋。故因名焉。今各羅海中有主嶋。王所産 嶋。故百濟人號爲主嶋]. As noted above, while we know little about the language of Paekche, it is probable that the languages of Yamato and Paekche were not closely related genetically; however, ancient people lacking any linguistic training would have still noticed that the word order of the two languages was similar. Thus, the first three sentences in the above quote likely were processed in Paekchean in the following manner:34 Mata KONKICI mudou + (hoy) + gerund, poyk PART tilpwo + infinitive verb + past. Narak poyk guo di + (hoy) + past. Mune KONKICI syel-past. Another example is the first cryptic poem in Man’yōshū, poem #9 in the anthology: 莫囂圓隣之大相七兄爪謁氣 吾瀬子之 射立為兼 五可新何本. Vovin (2002:458) originally speculated that the poem initially was written in Korean (what I interpret to be Paekchean), but other than the first two stanzas, the rest was lost, aside from a Japanese translation. I provide his transliteration of the first two lines, and a translation (Vovin 2017:39). 莫囂圓隣 NACOkʌ-s tʌrari

之大相七兄爪謁氣 thi-ta-PO-n-[i]-isy-a=ca mut-ke

The medium

35

After [I] looked up at the evening moon, [I] did ask: As Vovin notes (2017:47–52), part of the difficulty here is that some of these graphs are logograms, while others are phonograms or kungana. It seems that 莫 is a loan for 暮 “evening.” He notes that “evening” in Late Middle Korean would be nacwo but should have a final –h, so the complex graph 囂, LHC hɨɑu, appears to have been used to represent the final –h. The graph 相 appears to be used as a verb “look.” The complicated trigraph 七兄爪 appears to have been selected to represent a suffix attached to the verb “look.” While some of this reconstruction is conjecture, I believe it well demonstrates that the literate Paekche officials composed and read their prose in kumanari style, with a mixture of Paekche and Chinese words, the verbs inflected in the Paekche vernacular. We can conclude that kumanari processed Chinese by inverting the order of some constituents to make it easier to read in the vernacular. This fact is clearly noted above in the discussion on sekichey. Thus, I here conclude that the practice of kumanari laid the foundation for the later development called kundoku. The other technique I believe that came from Paekche is the use of small reading marks, such as kaeriten “inversion marks” or glosses, to aid the reader in the proper order of reading of Chinese texts. In the tradition that has been preserved in Japan, the two most common marks are レ, which is an inversion mark, and 一二三, numbers that indicate which constituents should be read in which order. The following example from the beginning of the preface of Kojiki demonstrates this: 無レ名無レ爲、誰知二其形一　“It had no name, and no action, so who could know its shape?” An English rendition is straightforward, as English syntax is closer to that of Chinese than Japanese. The complexity comes from that simple sentence with its unassuming gloss being the baseline to reconstruct the Old Japanese sentence of na mo naku siwaza mo nakereba tare ka sono katati wo siramu. Underlined elements need to be supplied by the reader according to context. When a sentence was so complex that the two categories of marks above were exhausted, then the reader had several other sets available: 上中下 “upper, middle, bottom,” which subset functioned within the clauses marked by 一二三. If more marks were needed, then the alphanumeric characters甲乙丙 were inserted into clauses marked with上中下. An illustrative example appears in the first half of the Tenmu record in Nihon shoki: 則且下發二五百軍一、塞中鈴鹿山道上　rendered in English as “Thereupon they raised an army of 500 troops and blocked the mountain road into Suzuka.” Another difficulty naturally arose when Japanese lacked a corresponding word for specialized Chinese (or Paekchean) vocabulary, or when the sentences became overly long and complex. Regarding the first issue with advanced or technological terminology, consider this simple sentence: 宛，大郡之都也 “Yuan, it is the capital of a great commandery.” Regarding the place name Yuan (宛), the Yamato student would likely have pronounced it as yemu, an approximation of the Chinese pronunciation. In the late sixth century when the complexities of Chinese laws or culture were poorly understood by the Japanese, there would have been no Japanese word for 大郡 “great commandery.” Here the elite in Yamato likely used the Paekche term35 since the precursor polity that resulted in Paekche had

36

The medium

existed south of the Lelang and Daifang commanderies for years. These examples illustrate that Literary Chinese was introduced to the people of Yamato through a Paekche filter. Thus, once the Yamato student became proficient in using kumanari, it was rather easy to convert this method into a Japanese version: kundoku. The student’s brain simply swapped out Paekche verbs and nouns and replaced these with Japanese equivalents, conjugating declinable words in Japanese. The next step in making Literary Chinese more readable as Japanese was to insert phonetic hints into the text. Lurie (2011:176) notes that the particle 而 found in the quote from The Analects (學而時習之) above could be interpreted as the gerund te, but it is not clear when that step took place. Nevertheless, these Chinese linguistic markers provided linguistic hints as to how one could insert nonsemantic (grammatical) information into the text to aid in reading and interpreting. The same type of phonetic script is found in the peninsular system known as kugyŏl (口訣) “mouth-cut marks.” The earliest surviving examples of this system appear in the seventh century. Lee and Ramsey (2000:52) provide the following example (phonetic script is double-underlined): 天地之間萬物之中厓　唯人伊　最貴爲尼　 at subject does, and so…. In the multitude of the myriad things amidst heaven and earth (at that place), Man (he) is the most noble (and so). This example demonstrates that not only the subject marker but also particles and verbal markers were inserted in Chinese sentences to aid in reading and recitation. As time goes on, these phonetic markers undergo a process of simplification; examples are 伊 becoming イ, 多 becoming タ, and 加 becoming カ. This evidence leads Lee and Ramsey (2000:52–53) to acknowledge that “there are many who believe, in Japan as well as Korea, that the beginnings of katakana, and the orthographic principles they represent, should be traced at least in part to earlier practices on the Korean peninsula.”36 As noted above, linguistic items that resisted conversion into Chinese were personal names and toponyms. This is seen clearly in the names found on a variety of inscriptions. Consider the following names found in the genealogy of Shōtoku Taishi preserved in a fragment from Jōgūki:37 “This Prince Opopodo married NAKAsiwa (NO MIKOTO) and gave birth to Prince Wopi [此意富富等王　娶中斯和命　生兒乎非王].” Here most of the names are in phonograms, though the wife’s name is a mixture of logograms (中 NAKA) and phonograms. The rest of this genealogy continues as: He [Prince Wopi] married Kuru Pime NO MIKOTO, the daughter of Izimura KIMI, the provincial governor of Mugetu Province, and she gave birth to Prince Usi, and [he later] married [the daughter of] Great King Ikumuneri Hiko, and fathered Ipatuku Wake, who fathered Ipati Wake, who fathered Ipakori Wake, who fathered Mawakake, who fathered Akapati KIMI, who

The medium

37

fathered Wopati KIMI, who married Anani Pime, the founder of the Yonu Omi, and she gave birth to Tunumusi KIMI. [Prince Wopopodo] married Puri Hime, the younger sister of Tunumusi KIMI [娶牟義都國造　名伊 自牟良君女子　名久留比彌 [賣]命　生兒汗斯王　娶伊久牟尼利比 古大王　生兒伊波都久和希　兒伊波智和希　兒伊波己里和氣　兒 麻和加介　兒阿加波智君　兒乎波智君　娶余奴臣祖　名阿那爾比 彌　生兒都奴牟斯君　妹布利比彌命也]. Each of these names is written in phonograms, and other than a few noble titles attached in logograms (臣 OMI and 君 KIMI), the majority of this genealogy is preserved in phonogrammatic script. This demonstrates that the movement to convert names to kungana appears to have not taken hold until the seventh century. There is also the issue revolving around the dating of the movement to attach titles to these names, and how these were represented, but I address that issue in Chapter 2. Before concluding, it is important to note that a number of scholars have assumed for years that there was a connection with Wa and Paekche. The evidence, however, has been somewhat speculative and while it has been clear for several decades that there was a line of communication and trade between Paekche and the Wa, it has been difficult to tease out three different data points regarding this contact: (1) Paekche, (2) Peninsular Wa, and (3) Insular Wa. There have even been theories about Paekche (or some other kingdom) invading and colonizing the Japanese islands. The “horserider theory” has been attacked from several angles, most persuasively from the view of archaeology. There simply is little evidence of an invasion. The archaeological evidence, as we have it, points to a gradual change, consistent with consecutive waves of immigration, not something abrupt as an invasion. What perhaps has not been given sufficient attention is the possibility that there was an influx of immigration on such a scale that it left a heavy cultural mark, analogous to an event like “the Beatles invasion.” In conclusion, it is manifest that there was a relationship between Wa and Paekche, but there is no evidence that Wa was a satellite state of Paekche. I have provided evidence to show that Literary Chinese and the ability to communicate by the written letter came from China via the peninsula. While admitting the belief that there was extremely limited exposure to this literary tool in the beginning in Yamato, I have argued that the movement toward gaining the ability to write one’s history took time and close supervision by a more literate group. This is where I have postulated that Paekche appeared. It is important to see the process in Yamato mirroring that in Paekche, partially because Paekche had undergone a similar experience a century prior. Thus, language is exported as a tool: intercourse with other states becomes possible, as well as the recording of one’s lineage and the chronicling of memorable events in a simple fashion. With the advent of Buddhism as a cultural icon with a literary foundation in scripture, a watershed event takes place. With this, Literary Chinese becomes more than a tool, but a creative endeavor. At this point, it becomes possible to record something approaching history, embedded with stories and legends. It is critical to remember that none of this occurs in a vacuum but in imitation of the events that happened in

38

The medium

Paekche. Thus, history, legend, and poetry are composed in imitation of Paekche. That three different Paekche historical chronicles are quoted in Nihon shoki is no coincidence. What has yet to be fleshed out is the possibility that the arranging and telling of Yamato’s ancient stories also mirrored that in Paekche.

Notes 1 Lee Sungsi (2013:172) argues that based on excavated data, Chinese writing and its technology had trickled down into the Samhan area (the three Han) by the first century BCE. The “three Han” are Mahan (馬韓 K. Mahan), Shinhan (辰韓 K. Chinhan), and Benhan (弁韓 K. Pyŏnhan). 2 Weizhi contains this line under the heading of “The Biography of the Han,” “Kara (Han) is located south of the Daifang Commandery, extending to the sea on the east and west, bordering the Wa in the south” [韓在帶方之南、東西以海為限、南與倭接]. This strongly suggests the Wa were on the southern tip of the peninsula. It is also interesting that the fifth-century work, Shui jing zhu, makes this comment, “Commentary on the High Ping River. Its water issues from the northwest of the Ping River. It flows immediately to the Wa Fortress in the north. Perhaps the people of the Wa territory moved along it” [高平川水注之、水出西北平川、東流逕倭城北、蓋倭地人徙之]. 3 Lee Sungsi (2013:172) notes that by the first century BC, “written documents were utilized in the exchanges between Han chiefs and Lelang Commandery.… .We may surmise further that the exchanges between Lelang and the two Han chiefs in the historical records … may have been conducted through the written medium.” 4 See Levin 2002:338. It is also a feature of Korean and Japanese. 5 This is an English rendition of the long title 國罡上廣開土境平安好太王. 6 I wonder if this name, wanglek, is intentional, as phonetically it sounds like Chinese “King Oak.” 7 Here 任那加羅 appears to be a single toponym. I consider this to be a federation and refer to it as the Mimana-Kara Federation. 8 This phrase (安羅人戍兵) appears three times on the stele, usually interpreted as “The Alla (or Anra) troops….” Wang (1984:230) argues that 安 functions here as a verb, and 羅人 is an abbreviation for Silla people, a usage also found in Samguk sagi. This interpretation makes better sense within the larger context of the inscription. 9 There is a rather remote chance that this usage of 辰 is very old, analogous to the use of 倭 Wa, based on Old Chinese. If true, then here 辰 would be OC *dər- (Baxter and Sagart 2014:181, 255), and theoretically could represent Tara, a polity mentioned in both Paekche quotes and Japanese accounts mentioned in Nihon shoki. 10 All pre-seventh-century phonetic script is reconstructed according to Schuessler’s Later Han Chinese or Middle Chinese (2009). The asterisks alert the reader that these are reconstructed (and approximate) readings for the individual characters. 11 Takeda (1994) has a concise outline of previous scholarship regarding this name, but he notes there is a tendency by some scholars simply to claim that Chen Shou butchered this section (1994:4). 12 This statement follows the reasoning as laid out in Kotani (1999:43–44). 13 Klaproth theorized back in 1823 that the language of Japanese may have been spoken on the peninsula: “Formerly the southern half of Korea was inhabited by a people named Chan (Han), which consisted of the tribes of Mahan, Pianhan, and Winhan, which collectively were called Samhan ‘the three Han.’ These tribes seem to have spoken a language different from that of Korean and to have shared similarities with the Japanese in education, customs, and practices” (1823:334, original in German). 14 It is also plausible that the original vowel was high and back, and all examples of /i/ show fronting.

The medium

39

15 I have had reference to Tōno regarding the interpretation of this inscription (2004), but I have not followed all his emendations. Mizutani (2006) has some helpful explanations about Tōno’s readings. 16 Because of the complex phonology of some of the Later Han reconstructions, I have converted these names to a simplified Early-Old Japanese Romanization (based on Schuessler) to make it easier for the reader to digest. Words in capitals represent a logogram instead of ongana. For certain terminology, I have followed Murayama and Miller (1979). 17 The reading tradition of this name is Wo Wake, but as Miyake (2003) believes, based on an innovative reconstruction of Later Han Chinese, the character 居 is not read ke here but ka. This is not to say that *waka did not change to wake later, especially if –i was later affixed, much as *saka “wine” becomes sake with –i affixed. 18 The text has 日十大王年, which seems garbled. The solution could be as simple as the date being inverted, and according to Occam’s razor I have followed this theory. There have been some who have seen 日十 as an imprecise casting for 曰, and then try to get 曰十 to be the name of the great king (大王), but other phonograms on the mirror match other artifacts, so Occam’s razor again suggests that the simplest solution is inversion of the date. 19 The inscription here seems to have the string: 斯麻念長奉 “*Se-ma pondered on his long service,” which does not make much sense. I have followed Mabuchi (1999:19), who follows the much earlier theory of Fukuyama Toshio in seeing the deformed character 奉 as either a corruption or poor casting of 寿, resulting in my rendition of “pondered on a long life.” 20 It seems clear that the individuals involved in the inscription and the casting of the mirror were either illiterate, poorly trained, or both, as Lurie (2001:160) argues, “reading difficulties are … caused by the inherently clumsy nature of its writing itself. Riddled with puzzling malformed characters and several undoubted mistakes, this inscription is indicative of either a low level of ability on the part of its composer and caster, or of a lack of concern of legibility on their part.” 21 Kayamoto Morito viewed the sword and made a handwritten copy of the inscription. He copied down these last five graphs as dots, instead of blank spaces as many have (cf. Yoshida 2001:19), and based on what people believe they see, some have argued that the name of the swordsmith belongs here, but others have argued that a name should not appear here, before the mentioning of the royal person on the inverse. Yoshida (2001:39–40) notes Miyazaki Ichisada’s theory that the final clause of the obverse should have been another auspicious phrase, and postulates that the final characters are 永年大吉祥 “an auspicious omen of many years (to come) and great fortune.” I have tentatively followed Takaguchi (1995:20) who believes the place name where the sword was forged belongs here. 22 Many, including Lurie (2011:86), interpret the characters 奇生 as representing the name of Crown Prince Kusu (later the 14th king of Paekche, King Kusu [r. 375–384]), but the phonological match of the two characters is too loose to be tenable. Nihon shoki records his name as 貴須 *kui-siu, but the two characters here, 奇生, are *kɨe-seŋ, and the phonology is too far apart to be relatable, as only the first consonant of each character match. I have thus interpreted this as a Chinese phrase instead of a phonetic transcription. 23 For the inscription here, I have relied on the work of Yoshida (2001), who has used the comparative work of Watanabe Kimiko (1981). I have also been persuaded by Shinzō (2005) in several instances. 24 The text has the 16th day as a bingwu day, which does not match the calendar, but many scholars now believe bingwu was used as an auspicious designator, a “good day” to fashion a sword (cf. Yoshida 2001:28–29, Shinzō 205:178). 25 Niizuma (1967:323–24) argued that the difficult to decipher regnal year of Taihe (泰和) on the Seven-Branched Sword is Taihe (太和), a Wei regnal year, which is 230 in the

40

26 27 28

29

30 31 32

33

34

35 36 37

The medium Western calendar. This dating would thus place Jingū in the realm of the world of Himiko, making the connection between the two individuals reasonable, but this theory is too fantastic to believe, considering that Paekche, as a state, did not yet exist. I hedge because of so many lacunae on the Seven-Branched Sword. Takaguchi (1995:18) also argues that putting in the real name of the ruler would show that the country is a tributary state, something that is very doubtful Paekche would have tolerated. Inukai (2014:126) argues that on the peninsula 中 with a month came to have a different chronological time frame. While 二月中 meant “during the second month (the whole month),” peninsular usages meant “from the previous month through the second month,” but I find this explanation less than persuasive. It is interesting that Samguki sagi records that “an old record notes that from the foundation of Paekche characters (Chinese) had not been used to keep records.” Cheng (1988:47–48) believes that this tradition is not reliable, and it seems probable that Paekche was familiar with Chinese writing a short time before the reign of Kŭnch’ogo, especially since the Chinese established diplomatic relations with Paekche in 372. The term “Paekche immigrant” does not preclude the individuals being either of Chinese or native Paekchean extraction. I view “Paekche” as the conduit for literacy, assuming the ethnicity of the individuals was rather fluid. It is important to remember that many have already argued this point: Bentley (2001a, 2004), Best (2006:78), Lurie (2011:200–01), Ichi (2014:34–35), Lee SJ (2014). Here I have simply gathered a large amount of information to buttress this theory. After the destruction of the Hansŏng capital by Koguryŏ in 475, the Paekche royal family moved south and established their capital in Ungjin (熊津 “bear ford”) on the Kŭm (Bear) River. Kumanari is the ancient Paekchean reading for “bear river”; see Bentley (2000:425, 428). A dialectal version of this reading is also preserved in one of the ancient Silla poems, in a collection called “Songs of the Flying Dragons to Heaven” (3, 15). There it is written kwomanoro (MR komanara). Vovin (2005:108–40) and Unger (2009:156) make a robust argument that the languages of Koguryŏ and Paekche were related to the language of Old Korean, to which Silla is also a member. Silla is the ancestor of Middle and Modern Korean, and all three have a word order structure very similar to Japanese. To make this visually decipherable, I have left Paekche phonograms in lowercase, and Paekche logograms in CAPITALS. Chinese words are in small caps. Syntactic markers are attached with a plus sign +. Particles are in bold. Verbs inserted to create a native sentence appear in parentheses. It is not clear what the Paekche term would have been, but it may have been koni kopori “great commandery” or even koni soko “great fortress,” based on glosses from Nihon shoki. Lee SJ (2014:170–71) reaches the same conclusion, but he tries to strengthen it by appealing to the use of ト based on 止 as found in some Korean evidence. 上宮記, a lost record that appears to have been a history much like Nihon shoki. The quoted fragment appears in Shaku Nihongi (cf. Onoda 1986:311–12).

References All works are published in Tokyo unless otherwise noted. Baxter, William H and Laurent Sagart. 2014. Old Chinese—A New Reconstruction. Oxford: Oxford University Press. Bentley, John R. 2000. “New Look at Paekche and Korean: Data from Nihon shoki.” Language Research, 36(2): 417–443.

The medium

41

Bentley, John R. 2001. “The Origin of Man’yōgana.” Bulletin of the School of Oriental and African Studies, 64(1): 59–73. Bentley, John R. 2004. “Kudara to Kodai Nihon no bunka kōryū.” In Tanaka Takaaki, ed. Kōsaku suru kodai—Nihon, Chūgoku, Chōsen hantō. Bunseisha: Waseda Daigaku Kodai Bungaku Hikaku Bungaku kenkyūjo, pp. 124–134. Bentley, John R. 2008. “The Search for the Language of Yamatai.” Japanese Language and Literature, 42: 1–43. Best, Jonathan W. 2006. A History of the Early Korean Kingdom of Paekche. Cambridge: Harvard University Asia Center. Brooks, E Bruce and A Taeko Brooks. 1998. The Original Analects: Sayings of Confucius and his Successors. New York: Columbia University Press. Byington, Mark E., ed. 2013. The Han Commanderies in Early Korean History. Cambridge: Korea Institute, Harvard University. Cheng, Chomyo. 1988. “Chosen sankoku to kodai Nihon no moji.” In Naoki Kōjirō sensei koki kinenkai, ed. Kodaishi ronshū. Hanawa Shobō, pp. 45–66. Eckert, Carter, Ki-Baik Yi, Young Ick Lew, Michael Robinson and Edward W Wagner. 1990. Korea Old and New: A History. Seoul: Published for the Korea Institute, Harvard University, by Ilchokak. Hirakawa, Minami. 2014. Kodai Nihon to kodai Chōsen no moji bunka kōryū. Taishūkan Shoten. Hong, Wontack. 1988. Relationship between Korean and Japan in the Early Period: Paekche and Yamato Wa. Seoul: Ilsimsa. Hong, Wontack. 1994. Paekche of Korea and the Origin of Yamato Japan. Seoul: Kudara International. Ichi, Hiroki. 2014. “Miyako no naka no moji bunka.” In Minami Hirakawa, ed. Kodai Nihon to kodai Chōsen no moji bunka kōryū. Taishūkan Shoten, pp. 30–51. Inaoka, Kōji. 1976. Man′yō hyōkiron. Hanawa Shobō. Inukai, Takashi. 2014. “Kodai Nitchō ni okeru gengo hyōki.” In Minami Hirakawa, ed. Kodai Nihon to kodai Chōsen no moji bunka kōryū. Taishūkan Shoten, pp. 124–145. Kim, Jinwung. 2012. A History of Korea: From ‘Land of the Morning Calm’ to States in Conflict. Bloomington: Indiana University Press. Kimura, Makoto. 1998. “Wajin no tōjō to Higashi Ajia.” In Kunio Hirano, ed. Kodai o kangaeru: Yamataikoku. Yoshikawa Kōbunkan, pp. 21–45. Klaproth, Julius. 1823. Asia Polyglotta. Paris: A. Schubart. Kobayashi, Toshio. 2010. Nihon kokugō no rekishi. Yoshikawa Kōbunkan. Kotani, Hiroyasu. 1999. Jōdai bungaku to mokkan no kenkyū (Research into ancient Japanese literature and mokkan). Tokyo: Izumi Shoin. Kwon, Oh Young. 2008. “The Influence of Recent Archaeological Discoveries on the Research of Paekche History.” In Mark E. Byington, ed. Early Korea, Vol. 1. Early Korea Project, Harvard University, pp. 65–112. Kwon, Oh Young. 2013. “The History of Lelang Commandery.” In Byington, ed. The Han Commanderies in Early Korean History, pp. 81–96. Lee, Iksop and S. Robert Ramsey. 2000. The Korean Language. Albany: State University of New York Press. Lee, Seung-Jae. 2014. “Old Korean Writing on Wooden Tablets and Its Implications for Old Japanese Writing.” Seoul Journal of Korean Studies, 27(2): 151–185. Lee, Sungsi. 2013. “The Samhan, Ye, and Wa in the Time of the Lelang and Daifang Commanderies.” In Byington, ed. The Han Commanderies in Early Korean History, pp. 165–190.

42

The medium

Lee, Yong-Hyeon. 2014. “Kinnen hakken sareta Kankoku kodai moji shiryō no gaiyō.” In Minami Hirakawa, ed. Kodai Nihon to kodai Chōsen no moji bunka kōryū. Taishūkan Shoten, pp. 52–64. Levin, Saul. 2002. Semitic and Indo-European: Volume II: Comparative Morphology, Syntax and Phonetics. Amsterdam: John Benjamins Publishing. Lurie, David B. 2001. The Origins of Writing in Early Japan: From the 1st to the 8th Century C.E. Columbia University Ph.D. Dissertation. Lurie, David B. 2011. Realms of Literacy: Early Japan and the History of Writing. Cambridge: Harvard University Asia Center. Mabuchi, Kazuo. 1999. Kodai Nihongo no sugata. Musashino Shoin. Miller, Roy Andrew and Murayama Shichirō. 1979. “The Inariyama Tumulus Sword Inscription.” The Journal of Japanese Studies, 5(2), 405–438. Miyake, Marc Hideo. 2003. “Philological Evidence for *e and *o in Pre-Old Japanese.” Diachronica, 20(1), 83–137. Miyake, Marc Hideo. 2006. “Kana’s Korean Origins.” In Françoise Bottéro and Redouane Djamouri, eds. Ecriture chinoise: données, usages et Representations. Paris: CRLAO, pp. 185–205. Mizuno, Yū. 1954. Zōtei nihon kodai ōchō shiron josetsu. Omiyayama Shoten. Mizuno, Yū. 1968. Nihon kokka no seiritsu. Kodansha. Mizutani, Chiaki. 2006. Eta Funayama kofun daitōmei no seijiteki haikei. Contained in Vol. 27 of Nihon shoki kenkyū. Hanawa Shobō, pp. 239–256. Nienhauser, William H, ed. 1994. The Grand Scribe’s Record. Bloomington: Indiana University Press. Niizuma, Toshihisa. 1967. Yamato Yamataikoku. Shingetsusha. Onoda, Mitsuo. 1986. Shaku Nihongi. Vol. 5 of Koten Chūshaku-hen of Shintō Taikei. Shintō Taikei Hensankai. Park, Cheun Soo. 2018. “Kaya, Silla, and Wa: Relationships and Their Historical Backgrounds.” In Byington, Mark E., Ken’ichi Sasaki, and Martin T. Bale, eds. Early Korea-Japan Interactions. Early Korea Project, Korea Institute. Cambridge, MA: Harvard University, pp. 133–181. Ri, Songshi (Lee, Sung-shi). 2005. Kodai Chōsen no moji bunka. Contained in Hirakawa, Minami, Kodai Nihon—moji no kita michi. Taishūkan Shoten, pp. 32–65. Schuessler, Axel. 2009. Minimal Old Chinese and Later Han Chinese: a Companion to Grammata Serica Recensa. Honolulu: University of Hawaii Press. Shinzō, Masamichi. 2005. Shichishitō meibun no saikōsatsu. Contained in Vol. 26 of Nihon shoki Kenkyū. Hanawa Shobō, pp. 173–192. Takaguchi, Keizō. 1995. “Isonokami Jinja shichishitō meibun no kaishaku.” Kodaigaku kenkyū, 133: 17–20. Takeda, Yukio. 1994. “Sankan shakai ni okeru Chin-ō to sinji.” Chōsen bunka kenkyū, 3:1–24. Tōno, Haruyuki. 2004. Nihon kodai kinsekibun no kenkyū. Iwanami Shoten. Uchida, Kiyoshi. 1996. “Kudara, Wa no jōhyōbun no genten ni tsuite.” Higashi Ajia no kodai bunka, 86: 93–121. Unger, J Marshall. 2005. “When was Korean First Spoken in Southeastern Korean?” The Journal of Inner and East Asian Studies, 2(2): 87–105. Unger, J Marshall. 2009. The Role of Contact in the Origins of the Japanese and Korean Languages. Honolulu: University of Hawaii Press.

The medium

43

Vovin, Alexander. 2002. “An Old Korean Text in the ‘Manyoshu’.” In Fabrice Cavoto, ed. The Linguist’s Linguist: a Collection of Papers in Honour of Alexis Manaster Ramer. Muenchen: Lincom Europa, pp. 455–460. Vovin, Alexander. 2005. “Koguryŏ and Paekche: Different languages or dialects of Old Korean?” Journal of Inner and East Asian Studies, 2(2): 108–140. Vovin, Alexander. 2007. “Cin-Han and Silla Words in Chinese Transcription.” In SangOak Lee, Choon-Yon Park and James H Yoon, eds. Promenades in Language: to Honor Professor Chin-Woo Kim. Seoul: Hankookmunhwasa, pp. 603–628. Vovin, Alexander. 2008. “Proto-Japanese beyond the accent system.” In Bjarke Frellesvig and John Whitman, eds. Proto-Japanese: Issues and Prospects. Amsterdam: John Benjamin Publishing, pp. 141–156. Vovin, Alexander. 2017. Man’yōshū, Book One. Leiden: Brill. Wang, Haicheng. 2014. Writing and the Ancient State: Early China in Comparative Perspective. New York: Cambridge University Press. Wang, Jianqun. 1984. Shirizu rekishi kenkyū: Kōtaiōhi no kenkyū. Yukonsha. Watanabe, Kimiko. 1981. “Shichishitō meibun no kaishaku o megutte.” In Higashi Ajia no sekai ni okeru Nihon kodaishi kōza, Vol. 3. Gakuseisha, pp. 148–175. Yeo, Hokyu. 2013. “The Fall of the Lelang and Daifang Commanderies and Its Aftermath.” In Mark E Byington, ed. The Han Commanderies in Early Korean History. Cambridge: Korea Institute, Harvard University, pp. 191–216. Yoshida, Akira. 2001. Shichishitō no nazo o toku: yonseiki kōhan no Kudara to Wa. Shin Nippon Shuppansha. Yoshida, Takashi. 1997. Nihon no tanjō. Iwanami Shoten.

2

Words and their meanings

At the outset, an important distinction needs to be made about literate versus preliterate societies, as this will underpin several assumptions made in this chapter. Nissen et al. (1993:19) provide an important baseline about early Babylonian society when they say, From the point of view of society in general, we should rid ourselves of the notion that the emergence of writing marked the beginning of an entirely new era. The preliterate means of administration … demonstrate that society had not only come to master similar administrative problems during previous periods; they also demonstrate that organizational structures were more or less the same in earlier periods. In regard to this, they note that “a whole series of precursors did in fact exist, although not in the form of the presumed written documents containing more primitive symbols, but evidently more in the sense of primitive ways of mastering problems of information storage” (ibid). This argument calls to mind a description of the Wa in the Chinese account before Chinese script was introduced to the society on the archipelago. Suishu records, “The Wa [originally] had no writing, but only carved on wood and tied knots in rope.” While this description may simply be a marker of a barbaric culture, it is still true that the Chinese are saying that the Wa had some other method to conduct business before the importation of writing. Because of the close affinity between cultural and technological artifacts on the south and southwestern areas of the peninsula and those in Kyūshū (cf. Rhee et al. 2007:417–19), we can argue that a preliterate system for business and information storage was still functioning in Wa. It is important to remember that other preliterate societies have proven adept at manipulating a fair amount of informational data via rather simple symbols. Nissen et al. (1993:11–15) note that in Babylon numbers were recorded with clay tokens that could be molded into the shape of the object being counted. Later, to preserve the integrity of the system and prevent counterfeiting, society invented “sealed bullae.” Here, “the contents were thus secured from unauthorized access” by being encased in wet clay and then sealed (1993:12) is reminiscent of the use

Words and their meanings 45 of fūdei on the peninsula. Another method was “numerical tablets,” where a scribe (or accountant) made “round or oblong depressions along the edges of flattened, approximately rectangular tables made of gypsum” (1993:13). Thus, ancient Babylon used the material at hand to record data analogous to the Japanese, who may have used rope, shells, or wood. It is thus reasonable to argue that about the time writing was introduced to the elite in the government of the archipelago that Wa society was already sophisticated, albeit working within a primitive system. As noted above, it is also certain that some of the systems in place in Wa society had antecedents from the peninsula, suggesting that the transition to writing in Chinese may not have been as arduous as is often imagined. There are also some scholars who adhere to the belief that writing developed from (religious) ritual. In a study on governmental tallies in China, Falkenhausen (2005:98–99) has argued that there is religious significance to the writing when found on tallies and that these symbols of authority also were a “formalized ritual medium of communication.” He continues that the writing by superiors, even in government, is “an offshoot of earlier uses of writing as a means of communicating with the supernatural sphere” (2005:102). Wang (2014:210), however, painstakingly advances the other possibility that writing emerged from the arena of administration of land, people, and trade. This viewpoint better equates with what the Chinese records symbolically report: the people relayed information via wooden devices and knots on a rope until writing was introduced. This description is more administrative than ritualistic. This transfer of power (cultural and technological) from the peninsula to the islands comes through many channels. I argue that what some call Sinitic cultural power entered the Japanese islands in a two-pronged process. One was through assimilation, as Wa inhabitants on the peninsula were in contact with Chinese descendants of the commanderies, as well as inhabitants of Mahan and other polities that had originally operated within the sphere of influence of the commandery. The other prong is a direct link with the later kingdom of Paekche. The absence of a unified entity in China after the destruction of the commanderies in the peninsula allowed Paekche and Koguryŏ to flourish. With this growth came territorial expansion, creating centuries of conflict between these two neighbors. Best (2006:84) notes that prior to the rise of King Kwanggaet’o, both Koguryŏ and Paekche had engaged in conflict, but neither could gain an offensive advantage. However, after the rise of Kwanggaet’o, “the balance shifted decidedly in favor of Koguryŏ.” It invaded and ultimately destroyed the Paekche capital at Hansŏng in 475, forcing Paekche to relocate its capital farther south in Unjing. As Best (2006:102) argues, however, this crisis forced Paekche to change its “political structure and cultural character of the royal court,” a change that allowed the kingdom to reinvent itself, which reinvigorated Paekche society. Later, with two ruling centers in China, Paekche and Koguryŏ were able to keep each other at bay, though Silla in the east began to become aggressive, eventually destroying the Mimana-Kara Federation in 562. With the unification of China under the Sui Empire in 581, Paekche’s king, Widŏk, tried to get in the good graces of the Sui with the hope that the Sui would dominate Koguryŏ. It is clear

46

Words and their meanings

that Paekche needed Yamato as an ally to counterbalance Koguryŏ to the north, a necessity that pushed Paekche to open a diplomatic dialog with Yamato. As argued earlier, I believe that the teachers from Paekche were instrumental in helping a small core of the elite at the Yamato court learn to read and write. This fact is represented allegorically in the legend in Nihon shoki, where it notes that in the 16th year of Ōjin, Wani came to court (from Paekche). He was made tutor to heir to the throne Uji Wakairatsuko, and the prince learned various texts from Wani. There was nothing he did not understand. This so-called Wani was the founder of Fumi Obito. The name of the Fumi Obito is written in Nihon shoki as 書首 “the document Obito,” but other records have the name as 文首 “cultured Obito.” Connecting this story to Prince Uji Wakairatsuko is an editorial ploy by a later compiler, considering this story is lacking in Kojiki. The idea that Ōjin wants the prince to ascend the throne appears in both Kojiki and Nihon shoki. After the death of Ōjin, Nintoku and Prince Uji withdraw their intention to rule, deferring to the other. When a resolution to this stalemate does not appear to be forthcoming, the prince takes his own life, allowing (or forcing) Nintoku to ascend the throne. The Nihon shoki compiler constructed this story to portray a Confucian air of “self-sacrifice” by the prince to a domestic readership—perhaps to illustrate that having become literate and educated in Chinese culture, the prince had changed from a barbaric warrior to a cultivated gentleman. Naturally, this small group of literate individuals at the Yamato court would have been taught to read and write within the framework of the basics of Chinese education. The Ōjin record in Kojiki relates that Wani Kishi came to court, bringing with him The Analects in ten chapters, and one volume of Qian zi wen.1 Notice what these two titles symbolically tell the reader. The first title is for reading (a primer), and the second is a textbook to help students write. Thus, one group of historiographers is presenting a picture depicting a scene where in the reign of Ōjin, the court not only learned to write Chinese, but they also became educated in Chinese culture, starting with Confucius. On the other hand, a different group of historiographers in charge of the source documents that formed the basis of Nihon shoki describes the account of Wani coming to Yamato as recorded in the 16th year of Ōjin but makes no mention of the specific materials he brought. Shoki simply says, “Wani came to court. He was made tutor to heir to the throne Uji Wakairatsuko, who learned various texts from Wani.” Comparing the two accounts, it is evident that Kojiki’s purpose was to record this event with little fanfare, while Nihon shoki is less interested in the details than the cultural impact of the event. The Shoki complier also wanted to make clear that while Yamato acknowledged the cultural superiority of Paekche, one of their royalty, Prince Uji was able to master this vast knowledge with little difficulty. Here, one point is pertinent to note regarding this transfer of knowledge taking place in the reign of Ōjin. Samguk sagi records that the Paekche court became

Words and their meanings 47 literate during the reign of King Kŭnch’ogo (r. 346–375). It is difficult not to argue that the Japanese historians at one point aligned their chronology with the Paekche royal calendar so that Ōjin coincided with King Asin (r. 392–405) in an attempt to say that while Paekche was educated earlier than Yamato (otherwise, how could they be the teachers?).2 The narrative demonstrates that chronological distance was relatively short, only a few decades, but the truth seems to be that about a century or more passed before literacy was able to take root in Yamato. It would appear that while the Yamato court is acknowledging an intellectual debt to Paekche, they are trying to shrink the chronological distance (and decrease the weight of the debt) as much as possible, in an attempt to avoid appearing too provincial. In addition, the Nihon shoki line about the quick-learning prince, noting, “There was nothing he did not understand,” illustrates that the people in Yamato were quick studies, painting a picture that they were as sharp as the people in Paekche. I admit that the account in Kojiki is difficult to take at face value, but rather than focusing on the veracity of the titles in the record, the “ideology” behind the account is of greater interest. The title of the primer, Qian zi wen, as found in Kojiki is quite suggestive. Some have argued that this recorded account is a fabrication because the date of inception of Qian zi wen in the early sixth century is later than the supposed story of Ōjin. However, Ogawa (Ogawa and Kida 2013:385–94) proposes that the text of Qian may have originated from earlier stages of creation, implying that the dating of the early sixth century is not as pivotal as some have argued. While it is not my intention to debate the merits of these theories, it is insightful that the compiler of Kojiki selected a title that early learners of Chinese would no doubt recognize as a textbook. That fact may be used to underscore the legendary quality of this story since it is logical for new students of Chinese on the continent to start with this text. Furthermore, the existence of this title also suggests that at some point in the Asuka period, students studied classical Chinese through textbooks like Qian zi wen. A brief look into this work is enlightening, for the content gives the student more than just a textbook for writing. Qian zi wen starts with the quatrain: 天地玄黄 / 宇宙洪荒 / 日月盈昃 / 辰宿列張 “Heaven and earth, respectively dark and yellow. Space and time, respectively vast and incoherent. The sun and moon—waxing and waning. The polar star and the constellations—taunt and strung out.” Thus, students absorb fundamental Chinese cosmology and theology while they learn the basic characters for “heaven” and “earth.” The primer teaches writing as well as the pregnant meaning behind the various characters. The mentioning of the title The Analects is also important because a teacher can multitask with this work: one can learn to read Chinese while learning Confucian values. Thus, while we may admit that this title in Kojiki is a later interpolation, it is also possible to deduce that the underlying information the early historiographers had regarding the material originally brought to Yamato consisted of “primers and Confucian texts.” The Analects is a collection of recorded sayings by Confucius and later written down by his followers. The key element of these sayings is the Confucian belief that good government is dependent on the moral

48

Words and their meanings

integrity of its citizens. The foundation of this integrity was the idea of ren 仁 “goodness,” which initially was demonstrated to one’s parents, and ultimately to those in power. The ruler was to be guided by ren, and Confucius believed that a moral society could not exist unless all citizens lived their lives based on this concept. Let us assume the following, as Kojiki relays that the students in this Paekche School were taught from the Confucian classics. These new students would have learned about two central tenets of Confucianism: ren 仁 “goodness” and li 礼 “ritual,” which leads to proper behavior. Consider some pithy statements the early students likely memorized, “The Master said, It is best to dwell in ren. If he chooses not to abide in ren, how will he get to be known” (4:1)?3 Also, “The Master said, The gentleman likes virtue; the little man likes partiality. The gentleman likes justice; the little man likes mercy” (4:11). Finally, “The Master said, The gentleman concentrates on right; the little man concentrates on advantage” (4:16). It is reasonable to conclude that these were foreign concepts to the early Japanese, but The Analects contained enough information for a student to begin to learn basic standards that underpinned Chinese civilization, and indirectly how one should view history and its proper recording. These basic principles form the foundation of Chinese historiography, principles that we may assume Paekche also observed. It is interesting that as students studied The Analects, they would have come across quotes such as, “The Shū (書) says, ‘Be ye filial, only filial, be friendly toward your brothers, and you will contribute to the government’” (2:21). Or The Shū says, “When Gāu-dzūng was in the mourning hut, for three years he did not speak.” What does this mean? The master said, Why just Gāu-dzūng? All men of old were like this. When the ruler passed away, the hundred officials continued in office. (14:40) Or “Yao said, Oh, you Shùn! Heaven’s order of succession, upon your person comes to rest / Unto the Mean do you hold fast! / And within the Four Seas vast / Heaven’s favor long will last” (20:1). Here, the student encounters quotes from Shujing (書経, “the Shū” or 尚書 Shangshu (尚書), which is now Englished as “The Book of Historical Documents,” “Venerated Documents,” or just “The Documents.” Sadly, the origins of this fragmented and damaged collection of works are poorly understood, but recent scholarship places the origin of this collection between 1000 BCE and 600 BCE (cf. Feldherr and Hardy 2011:391). These venerated documents were actually transcriptions of speeches held at court about government and how to succeed as a nation, a society, a community, and a family. The three Shujing quotes from The Analects above are illuminating. The first demonstrates the foundation for good government: love of parents and people around you. The second quote in essence says that when the ruler dies and people go into mourning, the government does not (should not) cease functioning. Thus, the officials continue

Words and their meanings 49 to go to work and the apparatus of government endures the transition to a new leader. The final quote appears to be a fragment from Shujing. The full quote talks about the emperor regulating government and taking personal responsibility for things. In part it says, [The Emperor] attended to weights and measures, examined rules and standards, restored disused offices, and a government of the Four Quarters put them in practice withal. He revived extinguished states, continued interrupted successions, promoted subjects in seclusion, and the people Under Heaven gave their hearts to him. (Brooks and Brooks 1998:192) It is mere conjecture to address what the students at the Yamato court, and thus connected in some way with assisting in ruling the realm, would have felt when reading about “good government.” We may be forgiven for believing that the Yamato students would have been encouraged to make a connection between these Chinese speeches and their own sovereign who issued edicts or liturgies. Nonetheless, it seems safe to say that at a basic level the discerning student would have realized the importance of creating and preserving documents of essential utterances. Essential information, preserved in oral traditions, would have included king lists and the genealogy of powerful families. Based on the research of Wang (2014:21–28, 44–45), who studied writing and its importance in preserving early states as found in ancient documents from China, Egypt, Sumeria, and Amerind (Inca and Maya), civilizations, I contend that ancient Japanese family histories were canonized in oral traditions, embroidered in a fabric that bridged two genealogies: a king list and a family list, where powerful families anthologized the heroic actions of their ancestors within a vague framework of the reigns of specific kings. Astute students would have realized that these traditions were excellent candidates for preservation via the new medium of writing. Furthermore, reading accounts of people quoting an ancient document, and then asking a sage about it, naturally would have impressed students who had no other access to information than through one’s own memory, or through someone appointed the receptacle of institutional memory. A written document gave those who could process it access to information others would not have. The astute student would see advantages to having things written down. A written document created access to information that previously was limited to those with institutional memory. Writing also increased the integrity and security of the information, because once written down, it created a precedent that was somewhat resistant to manipulation. One additional work that students would become familiar with as they studied and digested Literary Chinese was the great history, Sima Qian’s Shiji, rendered in English as “The Grand Scribe’s Record.” Most scholars believe it was completed around 100 BCE. The first 12 volumes are known as benji (本紀) “basic annals” and contain the essence of a millennium of Chinese history. Thus, Shiji is the book of Chinese history. It is vast in its scope, starting from the beginning

50 Words and their meanings with the earliest time of the five emperors and concluding with Emperor Xiao Wu (r. 140–87 BCE) of the Han dynasty. It became the model history for all histories going forward. Containing 525,000 characters, according to Nienhausen (1994, 1:v), the work comprises 130 volumes, with the basic annals, chronological tables (ten volumes), treatises (eight volumes), and hereditary house records (30 volumes), and it concludes with 70 volumes of memoirs. Reading Shiji, the Yamato students would have learned how chronicles are constructed, with royal reigns and their genealogy framed within a chronology. The students would have learned how genealogy and stories could be intertwined, and surely, they would have made connections with their own oral traditions that did much the same. Eventually, the Yamato students would have been exposed to Paekche attempts at historiography, which mirrored the Chinese tradition. In imitation of these two great civilizations, the Yamato court would eventually want to create a history of its own. The catalyst for this may have been the reaction of reading accounts of Wa people and events as portrayed (sometimes fragmentally or unflatteringly) in Chinese and Paekche histories. As Albert Dien (2011:509) points out, Chinese historiography was based on the Confucian principle of tradition. Because the sages of the past were never to be surpassed, a record of the past became a record of precedent. This tradition motivated succeeding dynasties to go back and record the reigns and actions of previous dynasties. This fact has resulted in the view that Chinese history was “written by officials for officials” (from Étienne Balazs, as quoted in Dien 2011:510). Looking at history from a genealogical reference point, however, Wang (2014:45) argues that there is a widespread tradition of early kings being embedded in the skeleton of a dynastic history “constitut[ing] the history of legitimate kingship.” What will later influence Japanese historiography is the other side of this coin: “Sima Qian believes that legitimate kingship is bestowed by Heaven” (ibid). The existence of a list of kings is the simplest form of history, but later historians can leverage its existence to legitimize the institution of the king. Thus, the Chinese models reinforced the importance of king lists and their legitimacy to the ruling family. In contrast to the orderly and stratified world in Shiji, by the seventh century, the Yamato student would have also been exposed to the well-known text of Chunqiu “Spring and Autumn Annals,”4 a reading of which makes it clear that these Confucian ideals, especially li, were important to stability in the world but also that history as a vehicle was more malleable than one might believe. This provided room for more than just historical fact; there was also room for anecdotes and aporia.5 While the process is still foggy, we can postulate that Paekche interpreted this according to their own cultural condition, leading to the conclusion that there was room in national history for the inclusion of oral traditions that dealt with “history” further back than anyone living could remember or anything recorded—so far back in time, in fact, that oral history intersected with the divine beginnings of the state. The source material for the histories contained in Samguk sagi are difficult to date and even more difficult to trust, especially in relation to Paekche, as Silla was

Words and their meanings 51 the compiler, and an unknown amount of violence was likely done when compiling the record, either in changing events deemed untrue or unworthy or in simply ignoring other accounts. Notice that the compiler claims that Silla, the last of the three kingdoms to emerge culturally, was established before either Koguryŏ or Paekche. It is helpful to turn to Chinese accounts of the origins of these kingdoms to understand what the early peninsular myths were and what historiographical sleight of hand these stories may have contributed to Yamato and the view of its own origins. In other words, rather than following a common trend of seeing the early Yamato historians twisting a certain fact or decapitating a different event, I argue that it is more plausible that Yamato historians simply followed their Paekche exemplars in using peninsular historiographical ruses in fabricating convenient aspects of their own dynastic history. Lunheng is an important work from around 84 CE (cf. Forke 1907, 1:9), which contains information in the form of essays on a variety of natural phenomena. It records a foundation myth that has influenced the entire peninsula. Quoting from Forke (1907), with my own additions in parentheses, Lunheng records: A servant girl of the king of T’o-li (OC thâk-rai, LH thɑk-liai) of the northern Yi was with child. The king wanted to kill her. The girl said by way of apology:—“A vapour, big as an egg, descended from heaven, and made me enceinte (pregnant).” Afterwards, she was delivered of a child, which she threw away into a pig-sty. The pigs sniffed at it, but it did not perish. Then it was removed again to the horse stable, in order that the horses should kill it, but the horses also only sniffed at it, and it did not die. The king thereupon imagined that the child would become a sovereign, and therefore ordered the mother to take it back, and had it nursed by his slaves. The boy received the name of Tung Ming (OC tôŋ-mraŋ, LH toŋ-mɨaŋ). He was employed as a shepherd for cattle and horses. As he was an excellent archer, the king got afraid, that he might deprive him of his kingdom, and therefore wished to kill him. Tung Ming went southward to the Yen-hu River (OC ɁamɁhlâɁ, LH Ɂɨam-hɑ), where with his bow he shot fish and turtles in the water. They formed a floating bridge, enabling Tung Ming to cross. Then the fish and turtles separated again so, that the troops pursuing him could not follow. Subsequently he became king of Fu-yü (OC pa-la, LH puɑ-jɑ, K. Puyŏ). Among the northern Yi there is a kingdom of Fu-yü [北夷橐離國王侍婢 有娠，王欲殺之。婢對曰：「有氣大如雞子，從天而下，我故有娠。 」後產子，捐於豬溷中，豬以口氣噓之，不死；復徙置馬欄中，欲使 馬藉殺之，馬復以口氣噓之，不死。王疑以為天子，令其母收取，奴畜之，名東明，令牧牛馬。東明善射，王恐奪其國也，欲殺之。東明 走，南至掩淲水，以弓擊水，魚鱉浮為橋，東明得渡。魚鱉解散，追 兵不得渡。因都王夫餘，故北夷有夫餘國焉。]. When Tung Ming’s mother first became pregnant, she perceived a vapour descending from heaven, and, when she threw the newly born away, pigs and horses sniffed at him. After he had grown up, the king desired to kill him, but the fish and turtles, which he had shot, formed a floating bridge.

52 Words and their meanings According to heaven’s fate he was not to die, therefore he was saved from pigs and horses. As he was predestinated to become king of Fu-yü, the fish and turtles formed a bridge to help him6 [東明之母初妊時，見氣從天 下 。及 生 ，棄 之 ，豬 馬 以 氣 吁 之 而 生 之 。長 大 ，王 欲 殺 之 ，以 弓 擊 水 ，魚 鱉 為 橋 。天 命 不 當 死 ，故 有 豬 馬 之 救 ；命當都王夫餘，故有魚鱉為橋之助也。]. Suishu records the following about the origins of Paekche: The ancestors of Paekche came from the land of Koguryŏ. There was a maid who served the king of Koguryŏ. One day, out of the blue, she became pregnant. When the king tried to kill her, she said, “There was an object shaped like the egg of a chicken. It came toward (and touched) me, and that sensation caused me to become pregnant.” Hearing this, the king let her live. Later she gave birth to a son. The king dumped the baby in the toilet but for a long time the baby would not die. With this the king thought that the child had a strange, divine power, so he commanded that the child be raised. The child was called Toŋ-mɨaŋ (Tongmyŏng). As Toŋ-mɨaŋ grew, the king grew envious and hated him. Fearing that he would be killed, Toŋ-mɨaŋ fled from the presence of the king and came to the Ɂɨam River (淹水). The people of Puɑjɑ all served Toŋ-mɨaŋ [百濟之先，出自高麗國。其國王有一侍婢，忽懷 孕，王欲殺之，婢云：「有物狀如雞子，來感於我，故有娠也。」王 舍之。後遂生一男，棄之廁溷，久而不死，以爲神，命養之，名曰東 明。及長，高麗王忌之，東明懼，逃至淹水，夫餘人共奉之。]. There was a descendant of Toŋ-mɨaŋ called Gu-thə, who was deeply benevolent and trustworthy. He first established his kingdom in the land under the control of the Han commandery of Daifang. Later Han Governor Gongsun Du of the Liaodong family married off his daughter to Gu-thə. The kingdom of Gu-thə gradually prospered and became a strong kingdom among the eastern barbarians. Because there had been one hundred families in the beginning who immigrated across the waters the kingdom was named Paekche (hundred-ferry) [東明之後，有仇臺者，篤於仁信，始立其國於 帶方故地。漢遼東太守公孫度以女妻之，漸以昌盛，為東夷強國。初 以百家濟海，因號百濟。]. It is well known that the foundation myths of Koguryŏ and Paekche are essentially the same: a child is born in a miraculous way, and when in danger as an adult, he flees south and arrives at a great body of water, where something mystical happens. This strengthens the theory that the people of Koguryŏ and Paekche both had Puyŏ origins. Samguk sagi notes that the founder of Paekche was a son of Chumong (Trio-moŋ). The other important point to notice is that these Suishu accounts read much like the story in Kojiki and Nihon shoki, where we have a mythological story, and then a descendant establishes the kingdom. It is a classic example of a “Genesis” story. Did Yamato just happen to use the same motif, or did the early Yamato

Words and their meanings 53 historiographers mimic the story in Paekche? A third possibility is the immigrants from the peninsula had a common origin story. Finally, notice the great difference with the beginnings of Silla, again from Suishu: The land of Silla is located to the southeast of Koguryŏ. During the Han Dynasty, its territory was part of the Lelang Commandery, called Se-la (斯羅). When General Gɑn Khuɡɨam (Guan Qiujian) of Wei attacked Koguryŏ and destroyed it [in 244 CE], people of the defeated Koguryŏ territory fled to the land of Ɂouk-tsiɑ (Woju 沃沮). Later they returned to their homeland, but some of the people who had stayed there [in Woju] formed a kingdom, Silla. That is why there is a mixture of people from China, Koguryŏ, and Paekche in [Silla]…. The ancestors of the Silla kings are of Paekche lineage, having fled from Paekche, crossing the sea into Silla. They later became kings of Silla [新羅國，在高麗東南，居漢 時樂浪之地，或稱斯羅。魏將毌丘儉討高麗，破之，奔沃沮。其後 復歸故國，留者遂爲新羅焉。故其人雜有華夏、高麗、百濟之屬 … 其王本百濟人，自海逃入新羅，遂王其國。]. Clearly, the Chinese record of the origins of Silla is closer to a factual account. The Chinese account gives one the impression that Silla originally was a land formed of immigrants and refugees and thus did not have an indigenous foundation myth; the version in Samguk sagi contains an egg motif for the birth of a boy who later would become king, likely brought by immigrants from their homeland. This story may have influenced one version of the Susanoo legend about his coming to Japan from Silla. One element that needs further attention is the connection the ancient Wa had to the peninsula. The following description appears in Hou Hanshu: There are three types of Kara (韓 Han): The first is called Mra-kara (Mahan); the second is called Dən-kara (J: Shinhan, K: Chinhan); and the third is called Brans-kara (J: Benhan, K: Pyŏnhan). To the west of Mra-kara are 54 territories. It borders Lelang on the north, and Wa to the south. To the east of Dənkara are 12 territories. These border Hwat-mrak (J: Wai-baku, K: Ye-maek) on the north. Brans-kara is located south of Dən-kara and is composed of 12 territories. It borders Wa on the south. Of these 78 territories, Paekche is perhaps one of these. The large territories have over 10,000 households, while the small territories have several 1000 households. Each territory is encompassed by mountains and seas. The area is over 4000 li square. The borders extend to the sea on the east and west. These territories all anciently were part of the land of Dən. Mra-kara is the largest [of the territories], and together with their related groups they established the king of Dən, with the capital in ŋwat-ke (or Muk-ke).7 He became the king over all the land of the three Kara (Han). The ancestors of the kings of the various domains are all from the same ethnicity as Mra-kara. … In the beginning when King TunɁ of the Kingdom

54

Words and their meanings of Tiɑu-sian was defeated by General Wei Man, King TunɁ led several thousands of the remaining people and they fled to the sea [crossed it], attacking [and later] defeating Mra-kara. TunɁ made himself king of Kara (Han). The descendants of TunɁ later died out, so the people of Mra-kara put one of their own on the throne as king of Dən [韓有三種、一曰馬韓，二曰辰韓，三 曰弁辰。馬韓在西，有五十四國，其北與樂浪，南與倭接。辰韓在東 ，十有二國，其北與濊貊接。弁辰在辰韓之南，亦十有二 國，其南亦 與倭接。凡七十八國，伯濟是其一國焉。大者萬餘戶，小者數千家， 各在山海閒，地合方四千餘里，東西以海為限，皆古之辰國也。馬韓 最大，共立其種為辰王，都月支國，盡王三韓之地。其諸國王先皆是 馬韓種人焉。…初，朝鮮王準為衛滿所破，乃將其餘眾數千人走入海 ，攻馬韓，破之，自立為韓王。準後滅絕，馬韓人復自立為辰王。].

Based on this description, Figure 2.1 shows what the southern half of the peninsula may have looked like before the breakup of the Dən State. Since the Wa are never described as having been part of that state, it is reasonable to expect that

Figure 2.1 Korean peninsula in 108 BCE. This map is based on “File: Korean Peninsula topographic map.png.” Wikimedia Commons, the free media repository. https://commons.wikimedia.org/w/index.php?title=File: Korean_Peninsula _topographic_map.png&oldid=49608774 (accessed Jun 7, 2017).

Words and their meanings 55 when the Chinese encroached on the peninsula in the former Han period (110–108 BCE), the two independent states of Dən and Wa were forced further south down the peninsula. According to the chronology preserved in the Chinese records, General Wei Man was sent to Old Tiɑu-sian (Old Chosŏn) to protect its borders around 195 BCE. He attacked TunɁ and set himself up as king. Thus, it would be around this time (195 BCE) that King TunɁ headed south, apparently by boat, to invade the coast of Mra-kara. In the meantime, Wei Man set himself up as king, and, in the end, the destruction of his grandson leads to the establishment of the commanderies in the peninsula in 108 BCE. It is reasonable to assume that because the commandery system provided a period of peace, the local kingdoms within the sphere of influence of Lelang began to become powerful. Hou Hanshu records, Near the end of the reign of Emperor Ling (fl. 156–189 CE) both Kara (Han) and Hwat (J: Wai, K: Ye) became powerful, and the commandery could no longer control them. The farmers were thrown into great confusion, and there were many people who fled to the land of Kara (Han) [靈帝末，韓、濊並盛，郡縣不能制，百姓苦亂，多流亡入韓者]. It is important to remember that Hou Hanshu records that during the reign of Emperor Ling (and his predecessor, Huan), the land of the Wa was also in great chaos. The following line appears in Liangshu, “During the Guang-he period (178–184 CE) of Emperor Ling, the land of the Wa was in commotion” [靈帝光和中　倭國亂]. If the land of the Wa truly did straddle the Tsushima Straits, then it is possible that with the decline in power of the Lelang Commandery, Hwat and Mra-kara pushed their boundaries outward to expand their territory, which may have upset the peace not only in Lelang but also in Wa. This would have forced the brothers and sisters of the Peninsular Wa to move south, until they came to the Tsushima Straits. The purpose of this rather long discussion is to demonstrate that the Wa and their neighbors were closely linked, both in the events surrounding them, and the cultural influence that flowed through the peninsula. It is plausible that the mythical story about miraculous pregnancies or births from eggs is an idea originally inherited from China. The “Annals of Yin” in Shiji records: As for Hsieh of Yin, his mother was called Chien-ti. She was a daughter of the Yu-Sung Clan and the secondary wife of Emperor K’u. [One day when] she and two other women went out to bathe, she saw a black bird lay an egg. Chien-ti took it and swallowed it. For this reason, she became pregnant and gave birth to Hsie. (Nienhauser 1994, 1:41) A similar story appears in Shiji at the beginning of the Qin dynasty (cf. Nienhauser 1994, 1:87). The ancient Chinese anthology Shijing contains this line found at the

56 Words and their meanings beginning of song 303, “Heaven bade the black bird / to come down and bear the Shang / Who dwelt in the lands of Yin so wide” (Waley 1960:275). Thus, a black bird comes down, lays eggs, and gives birth to the Shang. This motif then entered the peninsula and was carried by immigrants to the islands. Compare this with the auspicious story in the Nintoku reign as recorded in Kojiki, where the great ruler has traveled to Hime Island or “Princess Island” to hold a banquet. There they notice that a wild goose has laid an egg. This is considered an auspicious event because wild geese migrate southward for the winter, but do not lay eggs during this journey. The interpretation is that this is auspicious because the natural world is signaling that Great Ruler Nintoku is a virtuous ruler. There is more going on here than just a superficial “auspicious event,” however. It is possible that this story was not invented for Kojiki but was borrowed and placed here, with a change in emphasis. Wild geese are foreign to Yamato, so this story is seen as a metaphor for the miraculous “birth from an egg” story coming to Wa from a foreign territory. This same story was then degraded even further when the historians included it in Nihon shoki, as a wild goose has simply laid an egg on the Mamuta Dike, which then results in the same exchange of poetry as found in Kojiki. It is thus reasonable to assume that because the Wa anciently were influenced by Chinese civilization on the peninsula, they also embraced a variety of Chinese cultural artifacts. I noted in the previous chapter why the most striking artifact, writing, appears much later than most would assume.

Cultural issues Davis (1992:9–10) believes that Koguryŏ and Paekche began keeping historical records at roughly the same time in the fourth century. She believes that the scribes and record keepers were either people from China or descendants of Chinese ancestry connected to the colonies. A description of early literate history by Byington (2016:10) generally concurs with this description. As Yamato students tried to come to grips with foreign cultural ideals of historiography, one of the first issues would have been linguistic. If these students were to record their own history, an originally orally transmitted tradition, the court and its tutors needed to find solutions at the outset for two pressing problems. One concerned the format of recording the mainstay of their oral traditions: lineage. The other would have been the appropriate representation in the Chinese of native political and religious terminology or nomenclature. The first problem was straightforward in its resolution, as genealogy tended to be a simple recitation of a lineage of a father to son. The uncomplicated formula was 上祖名 … 其児名 … 其児名 “Distant ancestor … whose son was named … whose son was named…,” starting with the founding ancestor and moving forward as many generations as were preserved in memory. The lineage found on the Inariyama Tumulus Sword is a textbook example: *Wa Wa-ka OMI, whose distant ancestor was named *O-po-pi-kuai, whose son was named *Ta-ka-ri Swo-ku-nei, whose son was named *Te-yo-ka-ri Wa-ka, whose

Words and their meanings 57 son was named *Ta-ka-pai-si Wa-ka, whose son was named *Ta-sa-kui Wa-ka, whose son was named *Pa-nde Pi-(kuai), whose son was named *Ka-sai Pai-ya, whose son is named *Wa Wa-ka OMI. It is important to understand that Wa Waka Omi appears to have memorized his genealogy to seven generations (including himself). Is this evidence of a Japanese tradition where people of status could recite their direct line back seven generations? If so, does this help answer the question about why Japan’s mythology starts with the lineage of seven generations of deities? If powerful families knew their lineage back seven generations, then the founding deities would naturally be expected to adhere to the same principle. It is clear Yamato imitated a linguistic pattern found in China. From a twelfthcentury BCE tomb at Anyang in China, three bronze ceremonial ge-blades have been excavated, each blade containing an engraved genealogy (Wang 2014, plate iv). Two blades have genealogies of seven generations, while one has six generations. The format of these three blades is a variation on the same theme, based on a simple format: (i) 大祖曰X1, 祖曰X2 … “Great grandfather named X1, grandfather named X2…”; (ii) 祖曰X1, 父曰X2… “Grandfather named X1, father named X2…”; and (iii) 大兄曰 X1, 兄曰X2… “Great older brother named X1, older brother named X2 .”8 Notice how Wa Waka Omi’s genealogy fits nicely into this unsophisticated format: 其児名X “His son was named X.” A separate issue is identifying some of the other terminology that we believe to be calques or approximations of the Japanese in Chinese. Two important phrases appear on the Inariyama sword inscription: 上祖 “(distant) ancestor” and 大王 “great king.” Both usages are found in Chinese texts, but with different semantics. In classical Chinese, the usage of 大王 is a formal expression, much like English “Your Majesty.” The other usage, 上祖, also appears in numerous Chinese sources, but it generally functions as a title of respect toward one’s father (or the patriarch of the family). Based on this linguistic fact, I believe that these usages with altered semantics came through Paekche and not directly from China. It is plausible that Paekche reinterpreted both compounds: the one evolving from a simple title, 王 “king,” to the addition of 大 “great” without any concern for the usage in Chinese. The other appears to have evolved from “honorable (上) (grand)father (祖)” to “distant (in time) father” (= ancestor). The problem is determining, if possible, what the Old Japanese equivalents were for these Chinese words. It is here that we encounter an issue Ō no Yasumaro noticed and recorded in his preface in Kojiki: Anciently words and their meanings were both pristine. Laying out sentences and constructing phrases [in Japanese], it is difficult to convert these into characters. Relying solely on Chinese pronunciation, the characters do not capture the meaning of the [Japanese] words. Stringing sentences together completely in phonograms makes the sentences long [and taxing] [上古之時。言意並 朴。敷文構句。於字即難。已因訓述者。詞不逮心。全以音連者。事 趣更長].

58

Words and their meanings

Lurie (2001:303–5) argues that we should not put too much emphasis on Yasumaro’s preface because it is primarily a construct. Recall also that Umezawa (1988:101) argued decades earlier that even if we allow that the description in Yasumaro’s preface is different from reality (or is a construct), it still exists as a living reality within his ideology. Thus, we should be careful when comparing historical reality with the ideological reality of Yasumaro as we examine the preface. It is clear that Yasumaro is writing for the royal audience, related to Tenmu and Jitō, and it behooves us not to confuse ideology with history. Lurie (2001:302, n. 84) cogently argues that the crux of this section of the preface is the juxtaposition of 訓 with 音. The Shinpukuji manuscript glosses 訓 yomi “reading” and 音 kowe “sound.” It is intriguing that Yasumaro goes to such lengths to delineate the problem, and then essentially opts for the first choice, with a twist. He will write the majority of the text according to the reading with “sound” glosses attached here and there. In other words, he is saying that the courtly tradition is to write in kundoku, but there are certain places where this system engenders confusion, either because some characters have more than one reading or interpretation or because some Japanese lexical items resist conversion to a single Chinese character. Perhaps, in an attempt to dispel the confusion Tenmu’s court may have noticed in the records held by various families, Yasumaro decided to bring into sharper focus ambiguous parts of the story by providing “phonetic clues.” This fundamental problem continues into the future, where scribes begin to gloss words in the text (i.e., all texts from Nara and on). It is these glosses that provide a glimpse into the struggle the Japanese had in interpreting certain Chinese compounds, especially in a religious or mythical context. As an example, the compound 大王 “great king” appears numerous times in Nihon shoki. According to the glosses in the “Kunaichō shoryōbu” handwritten manuscript, these two characters are consistently glossed as ofokimi “great lord” (though in most cases the character 大 has no gloss because ofo is the only possible reading). In the second book of Nihon shoki, the “Kunaichō shoryōbu” manuscript glosses 上祖 as tofotu oya “distant father.” The second book of the autograph manuscript of Urabe Kanekata’s Nihon shoki also has tofotu oya. Oya is generally interpreted as “parent,” but it is instructive to remember that in many languages of the Ryūkyūs, oya < *aya means “father” (cf. Bentley 2008b:262). This understanding helps us connect the character 祖 back to its Chinese meaning of “grandfather.” There has been much debate about when the title of 大王 was pushed aside, and the great ruler was reinvented as 天皇 “heavenly sovereign.” A number of scholars have argued that this title came to be used around the time of Tenmu or Jitō. Recent evidence from excavations in the Asuka area lead many to conclude that the title originated during the reign of Tenmu.9 It is interesting that from a look at the 54,250 mokkan slips in the Mokkan database, only three slips contain this title from the earliest cache, excavated from the Fujiwara Palace ruins (fl. 694–710). There is the following fragment: □比古天皇□ Xpiko SUMERA MIKOTO X, which appears to be the name of “Heavenly Sovereign X-piko.” There is only one ruler with piko in his name in this era, and that is Shōmu

Words and their meanings 59 (r. 724–749), but the date is almost 15 years in the future from the time the Fujiwara Palace was abandoned. The other possibility is that this fragment is not a name but some other word. We should note that the final character, 皇, is not clear and the character may be something else. Another slip from the Fujiwara Palace ruins has this fragment: 天皇聚□, “Heavenly Sovereign gathers….” Excavators discovered this fragment buried in a layer of earth containing three other wooden fragments containing the dates 庚午年 (670), 丙子年 (676), and 丁丑年 (677). This strongly suggests that this title of 天皇 should be dated to around the time of Tenmu. It is clear from the edicts that these two graphs were read sumera mikoto “the august ruling existence,” as that reading is preserved in the liturgies and edicts. It is also clear that there were religious and ritualistic words that were difficult to distill into Chinese. As an example, in the story when Susanoo returns to heaven to face his sister, Amaterasu, the narrative describes her putting on a warrior-like performance to display her strength and determination to defend her territory, quoted below. Phonographic script is underlined (I have deleted Yasumaro’s annotative notes, aside from glosses): 即解御髮。纒御美豆羅而。乃於左右御美豆羅。 亦於御鬘。亦於左右 御手各纒持八尺勾璁之五百津之美須麻流之珠而曾毘良迩者肩千入之 靭【訓入云能理下效此】附五百入之靭。亦所取佩伊都之竹鞆而。弓 腹振立而。堅庭者。於向股蹈那豆美如沫雪蹶散而。伊都之男建【訓 建云多祁夫】。 I have translated the passage as: She immediately undid her hair, and tied it up like a man [with the hair looped on both sides of the head]. She wrapped long threads of numerous eight saka long jewels in both loops with both hands, in the vine [that tied up her hair], while on her armor-clad back she wore a quiver with 1000 arrows, and on the front of her armor she wore a quiver with 500 arrows. She then attached a powerful bamboo arm protector, swung her bow, and stamped on the ground in her courtyard till she sank up to her thighs, where she then kicked and scattered the dirt like powder snow, thus demonstrating her masculine power. (cf. Philippi 1968:74–75, Heldt 2014:19–20) Words such as midura “hair tied up in loops in masculine fashion,” itu “powerful, having an almost magical power,” and PUMInadumi “stamp and sink into” apparently did not have sufficiently tight semantic fits with Chinese characters in the eyes of the compiler. However, I will argue that these powerful words were written phonetically to make recitation easier. There is evidence that the text of Kojiki was intended to be recited aloud. The gloss nori for the character 入 is strange, unless one realizes that the phrase 千入 is to be read tinori, demonstrating elision of the longer ti no iri “contains 1000 of X (arrows).” Such examples

60 Words and their meanings demonstrate that the kataribe were present when this story was written down, as the text preserves an oral quality. These data imply that the oral traditions of the mythology and legends, which were preserved as performances, were a challenge for the court translators to convert into Chinese. Even with the practice of kundoku, certain words simply could not be accurately represented with Chinese. This was true of names, at least in the beginning. As mentioned in Chapter 1, the names in genealogies continued to be recorded mainly in phonograms well into the seventh century. The example of Nukatabe on the Okadayama Number One Tomb Sword inscription (各田卩臣) is an exception, and it likely represents someone from the peninsula. Scholars have dated these tombs to the late sixth century. The sword was discovered in 1915, and at the time the sword was still whole, but due to poor preservation, half of the blade has been lost to corrosion. The name Nukatabe appears in Izumo fudoki.10 One noteworthy change is the evolution of the title for the ruler. As Hirakawa (2008:27–28) points out, a sword excavated from the Inaridai tumuli group contains this fragment: [obverse] 王賜□□敬安? [inverse] 此廷□□□□. The burial goods suggest a date earlier than the supposed date of 471 for the sword excavated from the Inariyama tumulus. Thus, the Inaridai sword contains the character 王 “king,” while the somewhat later inscription on the Inariyama tumulus sword has 大王 “great king.” Linguistically, something has happened during this interval. One other clue is that this inscription carries no date, while three other inscriptions found on excavated swords do. Hirakawa believes that this sword was one of a group of swords with a simple but auspicious inscription crafted as gifts for “the king to bestow” (王賜) upon some of his subjects. Hirakawa (2008:29) also points out that the first two graphs are slightly larger in size and deeper in engraving, giving the impression of a sense of bolding or emphasizing the fact that “the king bestows” this sword on the recipient. It is interesting that the name of the king is not included, as that would have been expected. Two reasons seem plausible to explain the lack of the king’s name. First, the standard title appears to have been 倭国王 “king of Wa,” so the designation of “Wa” may have been dropped because the swords were for a domestic audience. Second, the name of the king was dropped for the same reason; the assumption being that his subjects knew the king’s name.11 Based on textual evidence, it is possible that the single graph 王 was anciently read kimi “lord.” Within several decades of the mid-fifth century, our textual data suggest that the prefix opo “great” was added to create opokimi “great king.”12 This last reading is found in the record in Suishu, “Therefore (he is) called *A-pəj-kɛj-mji (阿輩雞彌).” This is a somewhat sloppy rendition of opokemi. Considering that both /o/ and / po/ have nonhigh, unrounded vowels, the Chinese equivalents still seem to argue for opo-. As Miyake (2003b:118) argued, if 雞彌 is a fairly accurate rendition, it would represent an older kemi. If it is a pejorative usage (with “chicken”), it then represents kimi. If an accurate rendition, it lends credence to the theory of vowel raising that occurred in Japanese, where an older *e rose to i. One title interrelated to kimi is kisaki LHL “queen,” which so far lacks a convincing etymology. Here I argue that the etymology is kisa + kimi, and the final

Words and their meanings 61 syllable elided, resulting in kisaki. Vovin (2008:149–50) demonstrates that certain accent classes, mainly LH-L and LHL-L, go back to an older form with final –m. If the high vowel in mi devoiced leaving –m, this may have been dropped by analogy with other words. If this is correct, this title suggests that queens were being elevated to a status equal with the great king. The addition of the prefix opo “great” suggests that socially Yamato had kimi over certain smaller territories, and these smaller territories were all subject to the authority of the great unifier, the opokimi. I theorize that the king we know as Yūryaku (common name as Waka Takeru, or Muɑ in the Chinese records) was the first to take this designation. Another problematic word was the designation for territory. The Chinese had two different designations: 邦 OC *prôŋ “country, state” and 国 OC *kwək “state, fief.” The early distinction was that *prôŋ referred to large entities, while *kwək referred to smaller units, or smaller parts of a larger whole.13 My theory is that Wa likely called itself *sema, a word that entered the Ryūkyūs and became “settlement, village.” It is thus possible that the later designation kuni is a loan, perhaps from Paekche. As Mabuchi (1999:393–94) points out, in Old Japanese the usages of kuni in poetry point to an area a person has control over, so it can loosely be labeled “domain.” Recall that Suishu recorded, “The land has 120 *kun-ni, and these are much like Chinese Provincial Governors.” I argue later that this word is derived from Paekche koni “large” (cf. Bentley 2000:425–26), which then underwent vowel raising to kuni (see Chapter 4). The flow of political and cultural technology went from China (or the commanderies, which were colonies of China) to Koguryŏ and Paekche and then on to Yamato. Duthie (2014:22–40) has nicely laid out a concise outline of how Sinitic culture (within what he calls the Sinoscript sphere) came to permeate the peninsula and the Japanese islands, arguing that “the ideal of Sinic imperial rule that permeated Sinic texts … became the ideal of government throughout the East Asian region” (2014:22). Thus, as empires in China treated surrounding smaller countries as tributary or barbaric states, these same smaller entities treated their own neighbors according to the same political yardstick. Duthie (2014:28) also maintains that being a tributary state of China did not mean that the polity could not aspire to greater things. Koguryŏ is a case in point where the stele inscription shows that King Kwanggaet’o considered Paekche and Silla to be tributary states of his “empire.” While much of the textual material regarding Paekche has been destroyed, it would not be a surprise if Paekche considered Yamato to be a state lower in prestige than themselves. Nihon shoki records tribute coming from both Paekche and Silla, and even Koguryŏ, but this is likely a later development projected back into the past. Fourth-century Paekche was highly advanced in technology compared to Yamato, and any continental visitor to the islands would have recognized that. However, Paekche needed a powerful ally, and they were willing to nurture Yamato in the hopes of winning greater military support. It is interesting to examine two administrative designations imported from Paekche. The first is well known: OJ kopori “district” (Bentley 2000:424–25). This word remains in Japanese with kopori becoming modern kōri (< kofori
tərəsi. The Paekche word later gives us Middle Korean stuy “belt” (cf. Bentley 2000:424). This new form in Old Japanese may have referred to a royal sash that marked the individual in power. If true, then Kojiki has been more sensitive to the underlying semantics than Nihon shoki. This may provide evidence that Ō no Yasumaro’s family was of peninsular origin. In a related vein, another supposed title for the ruler was yamato neko. The first ruler with this title embedded in his name is Kōrei, the seventh ruler of Japan. His name is recorded in Kojiki as 大倭根子日子賦斗迩命 Opo Yamato Neko Piko putoni. In Nihon shoki, this name is transcribed as大日本根子彦太瓊天皇, with the same reading as Kojiki. Is it a coincidence that the sixth ruler, Kōan, has the name Yamato Tarashi Hiko? It is as if the historiographers are saying that tarasi piko is a title older than yamato neko. I postulate that yamato neko is a contraction of yamato no ye “older brother of Yamato.” Later ko “heir, son” was added when the semantics had blurred. Tokumitsu (1978:53–54), relying on the earlier work of Mizuno (1954), divides the names of the rulers into various chronological groups. He categorizes the names of the rulers after Ankan into six groups, based on different elements, such as ame “heavenly,” firo “wide,” take “brave,” and tarasi fiko “prince of

64

Words and their meanings

the sash.” Thus, (i) fifth century, for the Nintoku lineage of rulers; (ii) end of the sixth century into the seventh century, rulers previous to Suiko, including Sujin, Ankan, Kinmei, and Yōmei; (iii) after death of Saimei in 661, Jomei and Kōgyoku lineage names; (iv) during Jitō era, Tenchi and Tenmu lineage names; (v) Taihō time, Jitō lineage; and (vi) during the reign of Genshō (fl. 715–724), the representation of yamato having been altered from 倭 to 日本. This last change is interesting, as it argues that the change away from traditional representations, like 倭 Wa and 倭国 “state of Wa,” occurred during the reign of Genshō. This dating receives support from excavated mokkan. A search of all mokkan from the ruins of the Fujiwara Palace reveals 14 examples of 倭 but not one secure example of 日本. It should be pointed out that there are very few examples of 日本 on mokkan, making it difficult to date this example. However, as examples exist in Nihon shoki but not in Kojiki, using Genshō as a line of demarcation is reasonable. Wada Atsumu (1975:67–69) argues that these names should be grouped into three categories: (i) those from the beginning to Chūai (mythical), (ii) Ōjin down to Keitai (legendary), and (iii) Ankan onward (historical). He notes that from Ōjin to Keitai, the rulers have their real names recorded, while after Ankan, their names change to a Japanese stylized, embellished appellation, likely instituted at the death of Ankan due to a new ritual where the dead ruler was memorialized by being given a grandiose name. Group (i) names were reconstructed later, based on the rituals observed from the Ankan period on. Thus, he suggests that the names of some rulers before Ōjin belong to a mythical grouping that may not have been included in any structured chronology, other than perhaps a simple lineage of father and son. Three scholars from Yale University penned an article in which they offered a dynamic argument that state formation, when viewed historically, clearly is often coordinated with trying to tame the puzzle of identifying individuals. This taming is most often accomplished by granting (or fostering) certain groups permanent titles or surnames (Scott et al. 2002). In the Wa sphere, based on later information in Kojiki and Nihon shoki, it appears that status at court and within the orbit of court influence was based on the surname of the individual. In the second month of the ninth year of Tenji (670), family registers are created and a census is taken, allowing authorities to police people to prevent robbery and vagrancy. The problem of vagrancy inferred that the census had determined where a family resided, creating a zone from which they were not allowed to freely move, at least on bamboo/paper. This apparatus assumes that there are written records, both as census records and as other records that could be used to inspect documents to make sure people are who they represent and live where the record claims they should. Returning briefly to names, in earlier Japan it is clear that groups tended to be “categorized” and then known by where they lived (were based), or their occupation/skill. Surnames based on toponyms include Katsuraki, Ki, Yamashiro, and Yamato, while occupational surnames include Agatanusi (regional chieftains), Mononobe (makers of weapons), and Umakapibe (horse-tenders). This naming practice based on occupation is quite widespread throughout the world, as

Words and their meanings 65 common English surnames attest: Smith (smyð “one who works iron or other metals”), Miller (myl- “to grind grain into a flour”), Cooper (couper “one who makes or repairs wooden vessels”), Taylor (tailor “to make clothes”), and Weaver (wefan “to create fabric by interlacing yarn or other threads”). Kojiki and Nihon shoki preserve what must have been a much older system where the immigrants had a type of surname called kabane (written 姓 “patronym”). A number of these were titles of occupation and contained a mysterious element, be 部 at the end. First, we should point out that the character 部 being read be is a bit of an oddity,14 because its earlier reading would have been *bo. I theorize that be is the Paekche reading, being either a Sino-Paekche pronunciation or perhaps a Paekche kun (K. hun) rendering of the character.15 It is interesting that the surviving fragment of the Chinese encyclopedic work, Hanyuan (ca. 660),16 contains the following note in the section on Paekche, “One 部 consists of 500 soldiers.” This suggests that the 部 system in Paekche was a military unit, perhaps some kind of militia. One can imagine that these units then had associated groups of skilled workers who provided armor, weapons, and other commodities. It is possible that this system was introduced to Japan, and over time, the military component weakened, and these were gradually recategorized as industrial groups. An example is the Kume (来目), a group of archers, and their attending Kumebe (来目部). There are also the Mononobe 物部 and the Ōtomobe 大伴部, two of the oldest military groups in Japan. Other important titles are those given to specific groups of people based on certain skills or industrial ability. Below I list the prominent titles, with suggestions for etymologies.17 For our own intents and purposes, the most important groups were kataribe and pubito. Kataribe 語部 are a group of people who had committed to memory myth, legend, and history in a lyrical format. In Izumo fudoki, the name 語臣猪麻呂 Katari OMI Wimaro appears. There are also a number of wooden slips (mokkan) that contain the title kataribe. Katari is the nominalized form (a noun) of the verb katar- “to tell, to relate.” The other title, pubito 史, is created from fusing pumi “writing” with pito “person,” pumipito > pumbito > pubito. Martin (1987:417) believes that the word pumi is a Chinese reading of the graph 文, which he reconstructs as earlier *m(b)ywen. Part of the problem with this graphic connection is that as far back as we can go, the character 文 is *m- initial. Granted, in the language of Late Middle Chinese (ca. 600), the reading was mbun, but this is too late a date to believe the word was imported then. Baxter and Sagart (2014:287) reconstruct 文 as *mə[n], where the brackets mean the nasal may go back to an *-r. For the present, I see a semantic connection with the characters 文 and 書, not a phonetic one. An interesting title is atapi, because it appears to have come from Paekche as a two-character title, and the Japanese later shortened it. An example of the name and title is found in Nihon shoki with 加不至費直, read kaputi ATAPI (modern Kawachi Atai). Compare this with a name found in the Suda Hachiman mirror inscription: 開中費直, which as noted previously, I tentatively read keNA ATAPI. The date in the inscription suggests that the mirror originated around 443 or perhaps 503, sometime around the sixth century. The usage of 費直 is

66 Words and their meanings curious, but it also appears in a name found in Gangōji garan engi, Yamato Aya Ō Atahi (山東漢大費直). However, evidence suggests the Japanese shortened this kabane to one graph. The following line appears in Shoki in the reign of Kōtoku (r. 645–654), “During this year (650), Aya Yamaguti Atahi Ōkuti (漢山口直大口) received the imperial decree.” Compare this with the same name found in the inscription of the Virūpākṣa statue at Hōryūji, “Yamaguti Ōkuti Atahi…” (山口大口費), where we have variation between 直 and 費. Finally, Shoku Nihongi records this interesting event from the third month of 767, “Our ‘noble title’ is recorded as Ōsi Atahi (凡直) in the Kanoe Uma (670) Register; however, all registers have the character 費 attached.” This evidence suggests that the shorter or abbreviated version of 直 (glossed as atahi) that appears in Shoki was more economical. Another title is Mononobe 物部, where mono has several meanings: thing, utensil, or supernatural power. Perhaps the name of this group originated from these men being in charge of weapons and their deployment, with mono interpreted as “utensils of war.” A group in charge of the military was the Ōtomo 大伴. The etymology likely is opo “great” and tomo “arm pad.” Tomo specifically is an arm pad that archers wore to protect their wrists from the recoil of the bowstring. The Ōtomo were employed as officers over the archers. Two important service groups are the Murazi and the Miyatuko. The Murazi 連 are called “service nobility” by Barnes (1987:87). The word is usually parsed as mura and uzi. I suggest the word comes from mura-no-usi “lord of the group,” though “lord of the village” is also a possibility. Murazi were managers of service groups. The noble title Miyatuko 造 represented groups that performed a “function, service” for the court (Barnes ibid). The etymology comes from miya “palace” tu “of” and ko “heir, son”; thus, these were important people of the palace. Another group is the Pata (later Pada) 秦. This group oversaw silk cloth and weaving. It appears that the name originated from pata “loom,” but the second consonant became voiced to make a distinction between the machine and the people who were skilled in weaving. The title Nakatomi 中臣 signified people in charge of liturgies. Their name is a fusion of naka-tu-omi “servants between (the ruler and the people).” The middle vowel elided, creating nakatomi. The group name Imibe 忌部 is composed of the infinitive of im- “abhor” and be. This group, along with the Nakatomi, had a specific duty in relation to the liturgies. The group Urabe 卜部 were employed as diviners at court, with the name deriving from ura “divination” and be. The appellation obito 首 appears to have originally been a rank, before it became a noble title. A line in Kojiki from the story surrounding Susanoo reads, “I will have you serve in my palace as obito” [汝者任我宮之首]. The commonly accepted etymology is that this is an elided form of opobito “great person” (cf. Kōnoshi and Yamaguchi 1997:73, n9), but this etymology is historically and phonologically unlikely. One is forced to answer why a related word, opotomo, did not elide to otomo at this time. That the title obito means “someone in charge” seems beyond doubt, but this etymology is phonologically suspicious. Interestingly, the Atsuta manuscript of Nihon shoki glosses the following sentence found in the

Words and their meanings 67 seventh month of the 40th year of Keikō, 村之無長、邑之勿首, as fure ni fitogo no kami naku, mura ni ofuto nasi “In the hamlets they have no leader of the children of men, and in the villages they have no head.” As the word fure “village” is of Paekche origin (cf. Bentley 2000:436–37), it is possible that ofuto (首) is also of Paekche origin, and this word was corrupted to obito by Yamato speakers. Perhaps the etymology is related to the semantics of 首 “head,” as this word originally designated someone in charge of a group of be. The final title is Wosa, or modern Osa. The graphs 曰佐 in Nihon shoki provide the phonographic spelling of wosa, and we find the meaning in Nihon shoki where the graphs 譯語 “translate language” are glossed wosa. I once wondered if this was related to Manchu ulasa “tell, know” (Bentley 2000:436). These families appear to have been originally bilingual, or at least proficient enough to function as interpreters. Interestingly, the palace where Great King Bidatsu ruled was located in Osada, originally written as 乎娑陀 or wosada. Nihon shoki records the place name as譯語田 “translation paddy.” The etymology is wosa no ta “the rice field of translation,” though that may not be the original semantics. It seems this may have been a place where a group of bilingual families resided. It is helpful to look at the origins of these patronyms and titles by examining their lineages as found in Shinsen shōjiroku, a register of families and their genealogies from the Nara era (though the work was completed in 815). The register divided the families at court into three groups: (i) families connected to the ruling deity; (ii) those connected to other, nonroyal deities; and (iii) non-Japanese families. Some Paekche families included in the register claim an original Chinese origin, likely highlighting an earlier connection with the Han commanderies. The following “industrial” or “skilled” families are listed as coming from Paekche or Mimana: Paekche Masuta Fubito [789],18 Ofowoka Miyatuko [790], Kudara Tebito [870], Ofogata Fubito [881], Funado Fubito [882], Ofofara Fubito [883], Sonobe Obito [884], Mitami Obito [885], Takano Miyatuko [886], Asukabe Miyatuko [887], Miike Miyatuko [888], Nakano Miyatuko [889], Mano Miyatuko [890], Sugitani Miyatuko [891], Ayafito [896], Mitami Obito [923], Ibe Miyatuko [924], Uno Obito [949], Fata Miyatuko [950], Komoku Miyatuko [951], Sonobito Obito [950], Fayasi Fubito [977], Winabe Obito [978], Muko Obito [979], Fara Obito [980], Mino Miyatuko [981], Tuki Wosa [1028], Kure Patori Miyatuko [1036], Unu Miyatuko [1037], Asukabe Miyatuko [1038], Asukabe Miyatuko [1039], Kami Wosa [1041], Sinota Obito [1060], Torosi Miyatuko [1061] Mimana Ofoti Obito [810], Simidu Obito [811], Firata Obito [960], Ofotomo Miyatuko [961], Toyotu Miyatuko [988]

68

Words and their meanings

It should not surprise the reader that many of these families come from Paekche. It is also interesting that the Pata family has a diverse family tree. According to Shinsen shōjiroku, there are 14 families under the name of Pata Imiki. There are three families called Patabito, two called Pata Miyatuko, and one family each under the name of Pata Sukune, Pata Kimi, and Pata Suguri. Of these 22 families under the Pata umbrella, one claims to descend from a native Japanese deity, while another claims to descend from a Paekche family. The remaining 20 families all claim descent from the imperial line of China, suggesting these families were originally from the commanderies. A number of families called Pumi Sukune or Pumi Imiki claim descent from the royal family of Han China. There are also three families from Paekche and four from Koguryŏ. It would seem that literate individuals who came from the continent were in some instances given the title of pubito. Thus, we have come full circle. Immigrants to the Yamato court were put in government service and given charge over literacy and technology, receiving titles reflecting these responsibilities. Because of this, a variety of vocabulary in early government and literacy are related to Paekche or China (often through a Paekche filter). This fact is crucial for a better understanding of how historiography sprouted, grew, and flourished in early Japan.

Notes 1 千字文 “A Thousand Character Classic.” This is a poem, 1000 characters in length that does not repeat a single character. It was used as a writing textbook to educate children in writing Chinese. Most believe this title to be the product of a command from Emperor Wu of the Liang era (502–549), but we do not have to believe that this actual title was brought to Japan at this time. Likely, the early historiographers were simply using a title they were familiar with when all they might have had was information that some material was brought from Paekche, which matches the account in Nihon shoki in the 16th year of Ōjin, 王仁来之。則太子…習諸典籍於王仁, “Wani came to court. So, the Heir to the throne … learned from Wani from various books.” 2 Sakamoto (1994:49–50) has earlier argued that the compilers of Nihon shoki aligned the reign of Jingū with the calendar of the Wei as found in Weizhi. 3 All quotes from The Analects are from Brooks and Brooks (1998), but I have standardized their Romanization. Numbering is according to the Harvard Yenching system, which Brooks and Brooks use. 4 Dating this earliest of Chinese histories is difficult, because of the dearth of early commentaries, but most scholars argue that while they do not accept Mencius’ view that Confucius compiled this history, it is quite plausible that it was compiled during the later years of Lu (perhaps between 300 BCE and 249 BCE). 5 See Schaberg (2001:300–12); also Chen (2010). 6 Forke 1907, 1:175. 7 Some manuscripts of Hou Hanshu have 目支 instead of 月支. 8 The originals are written in what is called Large Seal script, which dates from around the eleventh century BCE until the seventh century BCE. My Chinese text is a modernized form of the original. 9 For example, Yoshida (1997:123–45) and Ōyama (2001:132–34). Miura (2008:4–5) mentions the excavation in January 1999 of 33 fuhonsen (富本銭) coins from the Asuka-ike ruins dig. These coins were found in a layer of Earth where pre-700

Words and their meanings 69

10 11 12

13 14

15 16 17 18

materials had been found, including a mokkan with the date 丁亥年, which scholars believe points to the year 687, or the year right after Tenmu’s death. Add to that the date of the mokkan with the writing 天皇聚, which was discovered in a layer of soil in the Fujiwara Palace ruins with two other dated fragments with the dates 670 (Tenji 9) and 676 (Tenmu 5); it is fairly clear that the title 天皇 began in the reign of Tenmu. Old Japanese has a doublet for the word “forehead,” nuka and pitapi. Nuka dies out in the Heian era. I believe that nuka is a word from the peninsula. Hirakawa (2008:28–29) believes the second reason to be the more likely of the two. The catalyst for this expanded title may have been Koguryŏ, which had ranks that were built on a tertiary system: grand X, great X, minor X: 太大兄・大兄・小兄 ranks two, three, and four, or 太大使者・大使者・小使者 ranks seven, eight, and nine (as quoted in Zhoushu). An annotational note attached to a section in the Chinese work, Zhouli “Rites of Zhou,” notes, “Great territories are called 邦 and smaller ones are called 国. Smaller units located within 邦 are also called 国.” [大曰邦、小曰國。邦之所居亦曰國]. Gina Barnes (1987:86, 2012:86) attaches a note for the character 部, “pronounced ‘bay,’” perhaps to prevent the reader from reading the graph as bee. However, the actual pronunciation was closer to be- as in English “better.” It may have even been mbe as in even better, said quickly. Tsuda (1947:33) claims that be was a Paekche word imported into Japan, but he offers no supporting evidence. The manuscript for Hanyuan has been lost, aside from volume 30 that is kept at Dazaifu in Fukuoka Prefecture. Kiley (1973:29) notes that “many old uji titles” had “Korean etymologies.” He does not, however, provide any concrete proof to support this statement. Numbers in brackets refer to the numbers in Tanaka (1996).

References All works are published in Tokyo, unless otherwise noted. Baxter, William H and Laurent Sagart. 2014. Old Chinese—A New Reconstruction. Oxford: Oxford University Press. Barnes, Gina. 1987. “The Role of the Be in the Formation of the Yamato State.” In Elizabeth M Brumfiel and Timothy K Earle, eds. Specialization, Exchange, and Complex Societies. London: Cambridge University Press, pp. 86–101. Barnes, Gina. 2012. “The Emergence of Political Rulership and the State in Early Japan.” In Karl Friday, ed. Japan Emerging: Premodern History to 1850. Boulder: Westview Press, pp. 77–88. Bentley, John R. 2000. “New Look at Paekche and Korean: Data from Nihon shoki.” Language Research, 36(2): 417–443. Bentley, John R. 2008a. “The Search for the Language of Yamatai.” Japanese Language and Literature, 42: 1–43. Bentley, John R. 2008b. A Linguistic History of the Forgotten Islands: A Reconstruction of the Proto-Language of the Southern Ryūkyūs. London: Global Oriental. Best, Jonathan W. 2006. A History of the Early Korean Kingdom of Paekche. Cambridge: Harvard University Asia Center. Brooks, E Bruce and A Taeko Brooks. 1998. The Original Analects: Sayings of Confucius and his Successors. New York: Columbia University Press. Byington, Mark E, ed. 2016. The History and Archaeology of the Koguryŏ Kingdom. Cambridge: Korea Institute, Harvard University.

70

Words and their meanings

Chen, Jack W. 2010. “Blank Spaces and Secret Histories: Questions of Historiographic Epistemology in Medieval China.” Journal of Asian Studies, 69(4): 1071–1091. Davis, Sherrill McCullough. 1992. Views of Paekche History: A Comparative Look at Traditional Sources. Harvard University Dissertation. Dien, Albert E. 2011. “Historiography of the Six Dynasties Period (220–581).” In Andrew Feldherr and Grant Hardy, eds. The Oxford History of Historical Writing, Vol. 1. London: Oxford University Press, pp. 509–534. Duthie, Torquil. 2014. Man’yōshū and the Imperial Imagination in Early Japan. Leiden: Brill. Falkenhausen, Lothar von. 2005. “The E Jun Qi Metal Tallies.” In Martin Kern, ed. Text and Ritual in Early China. Seattle: University of Washington Press, pp. 79–123. Feldherr, Andrew and Grant Hardy, eds. 2011. The Oxford History of Historical Writing, Vol. 1. Beginnings to AD 600. Oxford: Oxford University Press. Forke, Alfred. 1907. Lun-hêng: Philosophical Essays of Wang Ch’ung. Parts 1 and 2. New Yuork: Paragon. Reprinted in 1962. Heldt, Gustav. 2014. The Kojiki: An Account of Ancient Matters. New York: Columbia University Press. Hirakawa, Minami. 2008. Nihon no genzō. Vol. 2 of Nihon no rekishi. Shōgakkan. Kiley, Cornelius J. 1973. “State and Dynasty in Archaic Yamato.” Journal of Asian Studies, 33(1): 25–49. Kōnoshi, Takamitsu and Yamaguchi Yoshinori, eds. 1997. Kojiki. Vol. 1 of Shinpen Nihon koten bungaku zenshū. Shōgakkan. Lurie, David B. 2001. The Origins of Writing in Early Japan: From the 1st to the 8th Century C.E. Columbia University Ph.D. Dissertation. Mabuchi, Kazuo. 1999. Kodai Nihongo no sugata. Musashino Shoin. Martin, Samuel E. 1987. The Japanese Language through Time. New Haven: Yale University Press. Miura, Sukeyuki. 2008. Kojiki o yomu. Yoshikawa Kōbunkan. Miyake, Marc Hideo. 2003b. “Philological Evidence for *e and *o in Pre-Old Japanese.” Diachronica, 20(1): 83–137. Mizuno, Yū. 1954. Zōtei nihon kodai ōchō shiron josetsu. Omiyayama Shoten. Nienhausen, William H, ed. 1994. The Grand Scribe’s Record. Bloomington: Indiana University Press. Nissen, Hans J, Peter Damerow and Robert K Englund, 1993. Archaic Bookkeeping: Early Writing and Techniques of Economic Administration in the Ancient Near East, Paul Larsen, trans. Chicago: University of Chicago Press. Ogawa, Tamaki and Kida Akiyoshi. 2013. Senjimon. Iwanami Bunkō. Ōyama, Seiichi. 2001. Shōtoku Taishi to Nihonjin. Fūbaisha. Philippi, Donald L. 1968. Kojiki. University of Tokyo Press. Rhee, Song-na, C Melvin Aikens, Sung-rak Choi and Hyuk-jin Ro. 2007. “Korean Contributions to Agriculture, Technology, and State Formation in Japan: Archaeology and History of an Epochal Thousand Years, 400 B.C.–A.D. 600.” Asian Perspectives, 46(2): 404–459. Sakamoto, Yoshitane. 1994. “Higashi Ajia kara mita Nakoku to Yamataikoku.” In Oda Fujio and Tamura Enchō, eds. Nakoku no shuto suku okamoto iseki: Nakoku kara Yamataikoku e. Yoshikawa Kōbunkan, pp. 42–54. Schaberg, David. 2001. A Patterned Past: Form and thought in Early Chinese Historiography. Cambridge: Harvard University Asia Center.

Words and their meanings 71 Scott, James C, John Tehranian and Jeremy Mathias. 2002. “The Production of Legal Identities Proper to States: The Case of the Permament Family Surname.” Comparative Studies in Society and History, 44(1): 4–44. Tanaka, Takashi. 1996. Shinsen shōjiroku no kenkyū. Vol. 9 of Tanaka Takashi sakushū. Kokusho Kankōkai. Tokumitsu, Kyūya. 1978. Teiki no hihyō-teki kenkyū. Kasama Shoin. Tsuda, Sōkichi. 1947. Nihon jōdai-shi no kenkyū. Iwanami Shoten. Umezawa, Isezō. 1988. Kojiki to Nihon shoki no seiritsu. Yoshikawa Kōbunkan. Vovin, Alexander. 2008. “Proto-Japanese Beyond the Accent System.” In Bjarke Frellesvig and John Whitman, eds. Proto-Japanese: Issues and Prospects. Amsterdam: John Benjamin Publishing, pp. 141–156. Wada, Atsumu. 1975. “Kōtō denshō kirokuka no shojōken—gen-Teiki / gen-Kuji no hensan.” In Ueda, ed. Nihon kodai bunka no tankyū—moji, pp. 41–79. Waley, Arthur. 1960. The Book of Songs. New York: Grove Press. Wang, Haicheng, 2014. Writing and the Ancient State: Early China in Comparative Perspective. New York: Cambridge University Press. Weiner, ESC and JA Simpson, eds. 1971. The Compact Edition of the Oxford English Dictionary. Oxford: Clarendon Press. Yoshida, Takashi. 1997. Nihon no tanjō. Iwanami Shoten.

3

Mythical beginnings and the beginning of mythology

Our three core texts begin with what on the surface appears to be essentially the same mythological story. The compilers of Nihon shoki and Kujiki have been heavily influenced by Chinese cosmological ideology, as found in works such as Huainanzi. Kojiki, on the other hand, has been influenced to a lesser degree by Chinese philosophy, as it appears that the initial text avoids a Chinese-influenced description of the beginning of creation, referencing yin and yang. There is a tendency for most readers, even scholars, to view these three texts as fundamentally exhibiting the same starting point, but this is not necessarily true. To set a benchmark, it is instructive to review briefly how each of these three texts begins its story: Nihon shoki: Long ago when heaven and earth had not yet split apart, and yin and yang were not separated, chaotic and inchoate, like a hen’s egg, the boundless expanse of water contained sprouts [signs of life]. That which was pure and bright spread out to form heaven; that which was heavy and turbid remained to form earth. It is easy for that which is pure and subtle to converge but difficult for the heavy and turbid to congeal. Therefore heaven was completed first; earth was fixed afterward. After this event, a spirit-like sage was produced.1 The beginning of Nihon shoki’s mythology is based on a line in Huainanzi, “Heaven and earth had not yet split apart; yin and yang had not yet been carved out; the four seasons had not yet differentiated; the myriad things had not yet been generated” (Major 2010:86). It is clear that the compilers spliced Chinese ideology into a native tradition. Thus, Nihon shoki begins its narrative with the beginning of a “Genesis” type motif: heaven and earth had not yet been created. Ōbayashi (1987) persuasively argues that this complex path to creation results in two different levels or spheres. The first is a basic, natural level where creation settles down and falls into order. This results in the legend dealing with the birth of the islands. The second is the world of humans, which also becomes orderly, represented by Ninigi as ruler descending to the islands. Notice, however, that while the narrative mentions that a deity is produced, this deity does not

Mythical beginnings 73 receive a name. A few lines later, Kuni Toko Tati is spontaneously created in a space between heaven and earth, seemingly insinuating that he acts as the bridge between the two spheres. Next, consider the beginning of Kujiki. Sendai kuji hongi: Anciently, the original qi of heaven and earth remained undivided, resembling the egg of a chicken, with light and dark substances intermixed. Later, the lighter qi gradually floated upward, thinned out, and formed the heavens. The coarser floating elements were heavier, so these sank downward, coagulated, and formed the earth. The so-called “floating arrangement of the land” refers to this opening and dividing of the substances. This event is analogous to a fish swimming and floating on the surface of the water. At that time, heaven was created first, and the earth formed afterward. Then there appeared one deity in the High Plain of Heaven who is called Ame no Yuzuruhi Ame no Sakiri Kuni Yuzuruhi Kuni Sakiri. Kujiki also takes a cue from Huainanzi, but instead of quoting whole sentences like Nihon shoki, Kujiki has selected a phrase here and a term there. The method of the quotation is different, but Kujiki still follows the same philosophical path as Nihon shoki. Notice, however, that Kujiki begins with information not seen anywhere else. While Nihon shoki announces a nameless deity in the beginning, Kujiki names the first deity: Ame no Yuzuruhi Ame no Sakiri Kuni Yuzuruhi Kuni Sakiri, a deity seen in no other Asuka or early Nara era work. Regardless of the characters selected to represent the name, based on the phonology of Old Japanese, I believe the name means “heaven-spirit-that-transfers-(power) heave n-truly-mists-up-earth-spirit-that-transfers-(power) earth-truly-mists-up.” If this analysis of the name is correct, the name signifies that this entity embodies the power to obscure things: it has a spiritual power to obscure the heavenly realm from those below, before heaven and earth are separated into different spheres, where the next step, the creation of deities and humanity, proceeds. Kujiki thus positions this deity as a transitional figure in creation. Finally, consider Kojiki’s beginning mythology: When heaven and earth first appeared animated, the name of a deity that appeared in the High Plain of Heaven was Ame no Minakanushi, then Takami Musubi, then Kamumi Musubi. These three deities disappeared as individual deities. Next when the land [earth] was young, it was like oil floating on the ocean, drifting about like a jellyfish. The name of the deity who appeared from within something like a reed shoot sprouting forth was Umashi Ashikabi Hikoji, then Ame Toko Tati. These two deities also disappeared as individual deities. The five deities above are special heavenly deities. Kojiki starts with the assumption that heaven and earth have already been created. This starting point is less sophisticated than Nihon shoki or Kujiki, suggesting

74

Mythical beginnings

that Kojiki’s mythical narrative is the most primitive of the three. This basic cosmology resembles other primitive cultures, such as those on the Asian mainland, where the creation of humanity occurs after the great flood. It is also clear from even a cursory reading of these three texts that the character and stories of these myths are diverse. Therefore, due to the multifaceted nature of these myths, we are forced to conclude that a variety of myths have been interwoven over time into one rather choppy narrative. Even if we accept that these narratives have the same goal, it is worth exploring why several of these works start from different points. As Mizoguchi (1992:44) points out, not only does Nihon shoki ignore various aspects of the mythical story in Kojiki, but Kogo shūi and Kujiki also ignore these.2 What is fascinating is that after this introduction noted above, both Nihon shoki and Kujiki (as well as Kogo shūi) introduce the seven generations of deities, ending with Izanagi and Izanami. Kojiki also introduces seven generations of deities, but it is curious that five “special heavenly deities” are inserted at the beginning of what appears to have been a different strain of the mythological story. Mizoguchi (1992:46–49) believes that the proto-typical story of Ninigi descending to the islands described him as the offspring of Takami Musubi, but Tenmu’s court wanted Amaterasu to be the supreme deity and was determined to graft her into the genealogy. To placate those groups that had Takami Musubi as their supreme deity, he being a member of the musubi group, the Tenmu-era compiler(s) of Kojiki took the musubi group and grafted them into the story, ahead of the seven generations, but then let them disappear. Later, the compiler(s) relegated Takami Musubi to a subservient role with Ninigi, the offspring of both Amaterasu and Takami Musubi. This realignment of the beginning of the story demonstrates that manipulation of the myths apparently continued up until the start of the compilation of the three individual works, which hints at the existence of several varied myths revered by several groups at court. Compare the order of events within the myths, contrasting Kojiki, Nihon shoki (only the orthodox story), and Kujiki (NA is not applicable) . Table 3.1 illustrates several things, the most fascinating being that while the orthodox mythology in Shoki leaves much out that appears in Kojiki (acknowledging that many of these other events appear in the variant notes in Shoki), the basic order of major events in the story is the same between the two. This strongly suggests that a common framework operated behind these two works before the work of compilation. Kujiki is the odd man out here, as it not only contains more elements than either Kojiki or Nihon shoki but also has the middle constituents in a different order, likely demonstrating that Izumo had several competing versions of its myths. First, it is helpful to look at parts of the mythology all three have in common. One common strain is that of a two-step process of creating the islands and then the deities. Okada (1970:101, 110) argues that the origins of the creation myth’s current dichotomy of (1) creating the islands, and then (2) the birth of the deities can be traced back to a more primitive mythic strain as found in the local legends in the Ryūkyūs and parts of Kyūshū, where a specific island is born, and then the

Mythical beginnings 75 Table 3.1 List of events in mythology Event

Kojiki

Nihon shoki

Kujiki

Creation Birth of islands/deities Yomi Events in Takamahara Nigi Hayahi descends Events in Izumo Suga Palace Inaba rabbit Susanoo’s trip to the Netherlands Susanoo’s genealogy Ceding authority Preparation for Ninigi Saruta and Uzume Ninigi comes to Tsukushi Ninigi marries Hook story Events in Tsukushi Jinmu reigns in Hyūga

NA 1 2 3 NA 4 5 6 7 NA 8 9 10 11 12 13 14 15

1 2 NA 3 NA 4 5 NA NA NA NA 6 NA 7 8 9 10 11

1 2 3 4 5 8 9 11 10 12 7 6 13 14 15 16 17 18

deities give birth to the inhabitants. Later, he argues, the Yamato historiographers changed the second element from giving birth to the masses to that of the deities. Ōbayashi (1973b:84) has argued that the myths in the beginning are an amalgamation of earlier traditions.3 He believes the Izanagi/Izanami motif came from Southeast Asia and/or Polynesia. Okada (1970:101–02) notes that after Izanagi has had a role in the creation myths, he then appears again only in Nihon shoki in the record of Richū, where he has become the deity of Awaji Island. It is logical to argue that the role of Izanagi was originally that of a local deity, who gave birth to Awaji Island, and then later promoted by a powerful family to a higher role, that of procreating the entire archipelago with Izanami. Takioto (2005:27–29) argues that one reason there are several strains of the creation myth is due to different periods of inception of these myths. He concludes that the Kojiki version is the newest of the strains. His evidence is based on the fact that the order of constituents coming into existence does not follow that of the stories preserved in Nihon shoki, as well as the fact that little islands are created from the foam after the main islands are created in Nihon shoki. In Kojiki, these islands are given a place in the order of birth, being born before that of Tsukushi (Kyūshū), as if to demote the importance of Kyūshū in the myths. I counterargue that later manipulation has altered parts of the myths contained in Kojiki, where a primitive narrative later came to include newer elements, so this ordering does not necessarily reflect a chronological feature. Ōbayashi (1973a:10) argued earlier that the Japanese mythological story as preserved in the early records (Kojiki and Nihon shoki) is comprised of various and sundry elements, demonstrating its eclectic nature. We should also remember

76

Mythical beginnings

that the first items the Yamato court wanted to be recorded when they became literate were their genealogies. It is possible that many of the genealogies of the powerful families began with a mythological element, likely a tradition inherited from their peninsular cousins. The following illustrates a theory about this genealogy-mythology, and how the early historiographers tried to create a more homogenous mythology. Figure 3.1 shows a sample of the powerful families at court, and their divine ancestor, while Figure 3.2 illustrates a later attempt by court historians to order these in an organized story within the realm of a heavenly “family.” It is not difficult to imagine historiographers taking the ancestors of each powerful family at court, including the royal family, and elevating these to the status of deities, or at least semi-deities present when the creation of Yamato takes place in the mythology. When humanity is created in the mythology, the historiographers then have each ancestor descend from heaven and establish their own powerful family. This grafting of ancestors into the heavenly family may have looked like that shown in Figure 3.2. The point I wish to make in Figure 3.2 is not that these families necessarily claim the same deity in the mythical past but that the historiographers desired to make it clear that these families descended from a different deity than the ruling family. In this way, the royal family was able to generate distance between themselves and the other families that lent military and industrial support to the court. This attempt to create ranks of power and prestige in the realm of the mythological past resulted in an ordering and reorganization of what originally had been a nebulous mythology. It also mirrored what was occurring in society.

Key Ancestor

Nigi Hayahi

Oshihi

Koyane

Family

Mononobe

˜tomo

Nakatomi

Honoakari

Owari

Ninigi

Ruling family

Figure 3.1 Deities and descendants.

Deity A

Ninigi

Deities XYZ

Nigi Hayahi

Oshihi

Figure 3.2 Offspring of the creation deities.

Koyane

Honoakari

Mythical beginnings 77

Hints to the “Ur-Mythology” Wakai (2006:294–306) outlines a compelling narrative of what the original, local myths looked like. What follows is a condensed version of his argument. With the creation of heaven and earth, a deity appeared in heaven, Kuni Toko Tati. His descendants are Izanagi and Izanami, who descend to an island (variously Awaji or Onogoro) in Ōsaka Bay. There they create several islands, constituting the realm of “Yamato.” Izanami dies when she gives birth to the fire deity. Izanagi travels to Yomi “the afterlife,” which is a depiction of a graveyard. Izanami going into the palace of the deity of Yomi, with Izanagi following, represents Izanagi entering the rock-hewn room of a burial tomb and seeing the corpses. Shocked by the scene of corpses and the sense of defilement, he flees to Ahagihara located on the shores of Hakata Bay in Kyūshū (originally called Hyūga). There the sun and the moon are mystically created. The sun is the embodiment of good, and the moon is the embodiment of evil, thus explaining why these two heavenly bodies separately rule the day (the sun) and the night (the moon). In heaven is the supreme deity, Takami Musubi, a son of Kuni Toko Tati, who sends down his son to rule the land. That son is Ninigi, who descends to a mountaintop in Kyūshū (Hyūga). He marries the daughter of a local deity, and she gives birth to three sons. One son is known as “the master of the mountains,” while the other son is known as “the master of the sea.” From one of these sons is born the great king of Yamato, who fights his way eastward to unify the land. Much of this information is gleaned from Nihon shoki, juxtaposed with Kojiki. For example, we know from a song in the record of Nintoku in Kojiki that Onogoro and Awaji islands are off the coast of Naniwa (see KJK 83). Also, Shoki records in the record of Richū that while the great king sojourned on Awaji Island, the deity Izanagi possessed an officiator and announced his anger at the scent of blood. This demonstrates that Izanagi was originally a procreative deity in a local myth, likely from the area of Naniwa. Wakai (2006:298–302) argues that when Izumo and its traditions were grafted into the Yamato canon, the Bureau of History (my term) was obliged to alter the original myths in an attempt to create harmony. The powerful Izumo deity, Kamu Musubi, was added to the pantheon and set off against Takami Musubi. Susanoo, a hero of Izumo legends, was stripped from the Izumo pantheon and inserted in the central myths as a younger child born to Izanagi and Izanami. This addition altered the original bisymmetry of the myth, necessitating the replacement of the moon deity by Susanoo as a foil to the good sister, the sun deity.4 Wakai (2006:306) concludes that Susanoo was grafted into the myths as a political move to show the people that Izumo was like a younger brother and that Yamato was the benevolent older sister. To help untangle the various myths and create individual groups that can be better analyzed, it is helpful to investigate the myths of Izumo. First, there are a

78

Mythical beginnings

few similarities in Japan related to a myth wherein a deity creates an island (or a territory) and then populates it. This has survived in the creation myth as a representation of Onogoro Island or Awaji Island. I postulate that a variation of this myth existed earlier in scattered locales around the archipelago. Izumo preserves what is perhaps the most famous of these stories. Kojiki and Nihon shoki both include the skeleton of a story where a deity (Ōkuninushi or Ōnamuji) creates the land with the help of Sukuna Bikona. There appears to be a “great” versus “minor” element here, as represented in the names opo “great” versus suku “small.” I believe there is more at work here, however. My theory is that the original legend from Izumo also had a dual, male and female, procreator: Ōnamuji and Sukuna Bikona. I postulate that the name Sukuna Bikona actually referred to a female, composed of these elements: suku “small” na “land,” piku “pull,” and womina “woman.” When strung together as a name, sukuna no piku womina undergoes changes of elision and then prenasalization: sukuna n pikuwomina > nbikwonna > bikwona. As evidence for this, we should remember that Izumo fudoki contains the legend of “land pulling” in the story of the Ou District. A deity called Yatukamidu Omituno is credited with pulling the land together, thus increasing the size of the territory. If the analysis of Sukuna Bikona’s name is accurate, then she is the foundation for the land-pulling theme seen in Fudoki. Thus, Izumo also preserved a tradition of a male–female duo procreating the land. The male was designated “great” and the female “small.” Yamato later (under Empress Jitō?) took this motif and flipped it on its head, making the sun a female goddess who is great and the moon deity a male who is small. The creation myths of the Ryūkyū Islands are also very interesting, preserving what I believe is a much earlier version that had existed on the islands. The Ryūkyūan kingdom produced two histories, Chūzan seikan (中山世鑑, 1650), the first official history, and Chūzan seifu (中山世譜, 1701). The first was written in Japanese, and the second is a Chinese version of the first. According to Chūzan seikan, a deity in the heavenly palace named Amamiku receives a command from the Sovereign of Heaven (天帝) to go down and create islands, because while there already exists a dwelling for the deities, there are no islands/villages for humanity. After taking an excursion down to observe conditions, Amamiku returns to the heavenly palace and petitions for materials to create a dwelling place. The Sovereign of Heaven bestows material on him, including the earth, rocks, grass, and trees. With these materials, he creates villages without number.5 First is the Asu Grove by Kunigami, next the Kanahi Thicket of Nakijin, then Chinen Forest, and so forth. Millennia pass, but humanity does not appear; Amamiku petitions the Lord of Heaven for “human seed” to populate the islands, and the Lord of Heaven grants him a son and daughter of his own. At last, the female gives birth to three boys and two girls. The oldest son is the founding ancestor of the kings, titled Tenson “heavenly grandson.” The second son is the founding ancestor of the various noble families. The third son is the ancestor of the commoners. The oldest daughter is the ancestor of the priestesses who serve the court, and the second daughter is the ancestor of the lower priestesses.

Mythical beginnings  79 The myths as contained in Chūzan seifu show minor variations. Here, a male and a female deity create and populate the islands.6 The name of the deity, Amamiku, is also preserved in the dialect of the people living on Hamahiga Island off the coast of Okinawa (part of the municipal city of Uruma), where the name is Amamichū. One possible etymology is ama “heaven” mikura “august child,” where *kura (or *kora) changes to ko and then the vowel raises to ku (cf. Bentley 2008a:17, 2008b:251). However, -chū in the Hamahiga form is clearly chū “person,” so it is possible that we have two competing forms: “heavenly august child” versus “heavenly august person.” Chūzan seikan goes on to say that after humanity is established, two deities of protection appear. Together they are labeled Kimimamono. The deity of heaven is called Obotukagura, while the deity of the sea is called Giraikanai. Interestingly, Chūzan seifu has the label of these two deities reversed, with Giraikanai the deity of heaven, and the deity of the sea as Obotukagura. Chūzan seikan continues with its record, noting that in the beginning humanity lives peaceably in holes, without envy or malice. However, they lack any knowledge of agriculture. They simply eat the fruits of trees. They do not know how to make fire; they simply drink the blood of animals and eat the flesh. So Amamiku returns to heaven and petitions to receive seeds for the five grains, and this is the first time that humanity harvests barley, foxtail millet, soybeans, and common millet on Kudaka Island. Rice is planted in an area by the capital city. In general, a common strain is apparent: there is a central deity in heaven, two procreator deities creating the land (islands), and the offspring connected to the deities. After the land is populated, the deities grant grain to the inhabitants to provide sustenance to the people. Kojiki, Kujiki, and Nihon shoki all contain related myths about grain being developed in the islands (created from the body of a deity) to feed the people. These depictions demonstrate the centrality of agriculture. Figure 3.3 lists the grains produced. It is interesting that all four records have rice, foxtail millet, and soybeans in common. We can conclude that the immigrants to the Japanese islands brought with them both a tradition of divine creation, of land and humanity, and a divine blessing of agriculture. That all four records contain “rice” demonstrates the centrality of the grain to the islanders. Grain

Kojiki

Kujiki

Shoki

Chūzan

⢖ “foxtail millet” 㯩 “common millet” ✄ “rice” ኱㇋ “soybeans” 㯏 “barley” ᑠ㇋ “red beans” ⛳ “barnyard millet”

‫ۑ‬ ‫܉‬ ‫ۑ‬ ‫ۑ‬ ‫ۑ‬ ‫ۑ‬ ‫܉‬

‫ۑ‬ ‫ۑ‬ ‫ۑ‬ ‫ۑ‬ ‫ۑ‬ ‫ۑ‬ ‫܉‬

‫ۑ‬ ‫܉‬ ‫ۑ‬ ‫ۑ‬ ‫܉‬ ‫ۑ‬ ‫ۑ‬

‫ۑ‬ ‫ۑ‬ ‫ۑ‬ ‫ۑ‬ ‫ۑ‬ ‫܉‬ ‫܉‬

Figure 3.3  Grains produced.

80 Mythical beginnings In an argument about the ancient people not having the concept of a “unified state,” Kobayashi (2006:76) maintains that ancient Japan saw the authority to legitimize a state as a status granted through Chinese recognition; this legitimization went back in an unbroken succession to Queen Himiko and her Yamato state, or so the court argued. It is not my intention to argue for or against a location for Yamatai versus Yamato, but it is reasonable to assume that later rulers grafted their dynasty into that of previous reigns, which likely had a tradition linking them back to the dynasty of Himiko and her successors, whether there was a historical link or not. It is interesting, therefore, that Himiko’s name does not appear in any early, indigenous record: Kojiki, Kujiki, or Nihon shoki. What seems plausible is that Himiko and the stories of her federation were distilled into one hazy tradition. Evidence for this appears in Nihon shoki, where it records: With this, Izanaki and Izanami considered together, “We have given birth to the land of Ohoyasima, and the mountains, rivers, grasses, and trees. Should we not now give birth to the ruler of all that is beneath heaven?” With this, they gave birth to the deity of the sun, named Ohohirume Muti. 大日孁貴 is read opopirume no muti. Another record says that she is called Amaterasu Ohomikami. A different manuscript says that she is called Amaterasu Ohohirume. This child was of a lustrous brilliance, and her radiance shined throughout the six quarters of the realm. Therefore, the two deities were pleased, and said, “Though we have had numerous children, none have been as wonderful or mysterious as this child. She should not remain in this land for long; naturally, she should be sent back to heaven in haste to rule over heavenly matters.” At this time, the distance between heaven and earth was not very great. That is why they were able to send her up the pillar of heaven. It is interesting that this name, Hirume, appears in the orthodox version of the mythology, suggesting that the title Amaterasu Ōmikami is nothing more than a variant tradition, one followed by Kojiki and Kujiki. The meaning of the name firume is a matter of debate, but it is possible it means “woman of the sun.” Jōgū Shōtoku Teisetsu contains a number of earlier spellings where the graph 弥 is read me (instead of mi), showing vestiges of a graphic tradition where the Sino-Japanese reading of an earlier graph had not yet undergone vowel raising (e > i). It is thus very probable that there was an older reading tradition at court where Himiko’s name 卑弥呼 was read in Japanese as pimeko. It is also possible that the court concocted its own etymology for this name as “sun-woman-child.” This may have been later reconstructed as piru “sun at noonday.” We should not overlook the fact that the Shoki compiler attached the graphs 日孁 “sun-divine woman” to the older reading of piru-me, demonstrating that there were some at court who wanted to connect this deity to the sun and, by extension, back to Himiko in the ancient past.

Izumo myths The role of the myths of Izumo has been the subject of a number of studies. Kadowaki (1976) outlines the complex and layered history of the Izumo region

Mythical beginnings 81 and notes that as Izumo has experienced a number of invasions from neighboring areas, the Izumo myths are rather eclectic and should be understood within this context (1976:212). Briefly sometime in the fourth century, chieftains of the Kibi domain took control of the area of Izumo.7 By the fifth century, however, the Ou kingdom came to power in the midsection of the Izumo region, having overthrown the Kibi. The Kibi may have still controlled other parts of what we later call Izumo, but an important part, Ou, had established itself as a kingdom. Near the end of the fifth century, as the Yamato domain grew in military power, Kibi found itself under pressure, its influence weakening in and around the Izumo area. This allowed the Ou Kingdom to expand its territory, using the surrounding mountains as a strategic shield, and the natural ports to create a transportation and communication system. With this the king of Ou was able to expand militarily, also extending his cultural influence. With the complete withdrawal of Kibi, Ou expanded its territory further. The Ou Kingdom appears to have used its sea routes to make contact with domains in Koshi and points farther north as well as establishing contact with chieftains in Kyūshū. While the timeline is not clear, at some point Yamato makes contact with Izumo, and later it becomes a satellite of Yamato. It is interesting that the story found in the seventh month of the 60th year of Sujin in Nihon shoki appears to reflect an event within the history of Yamato subjugating Izumo and taking control of Izumo’s “sacred treasures” (神宝). Kadowaki (1976: 62–63) connects these sacred treasures to the rites and festivals of the Kitsuki Shrine. It is noteworthy that the legend in the Sujin record contains a story about a child who speaks words that are difficult to understand; these appear to have some connection to Izumo and its relationship with Yamato. The original text has 玉菨鎭石。出雲人祭、眞種之甘美鏡。押羽振、甘美御神、 底寶御寶主。山河之水泳御魂。靜挂甘美御神、底寶御寶主也。 While this is written in Chinese, the order of the constituents mirrors Japanese, perhaps attesting to the age of this segment. Some scholars agree that these words spoken by the child are meant to reflect divine words spoken through a medium.8 Context would suggest that the deity of Izumo is asking that the “sacred treasures” pilfered by Yamato, and which rightfully belong to Izumo, be returned. My translation (interpretation) of this passage is: There is a rock like a jewel sunken beneath the water plants. It is the perfectly (shaped) beautiful mirror worshipped by the people of Izumo. It is the mirror of the wonderful deity who exudes power. The treasure of the bottom of the pond, the actual shape of the treasure. The spirit of the cleansing water of the mountain river. The sunken mirror of the splendid deity. The treasure of the bottom of the pond, the actual shape of the treasure. The meaning is difficult to tease out, as many of these phrases are simply auspicious or mystical epithets strung together to describe the beauty or power of the word “treasure.” “Rock” in the first line is the usual association of a jewel, which then refers to treasure. The idea of being sunken in a pond, or underwater, may

82 Mythical beginnings refer to the jewel no longer being accessible, having been confiscated by Yamato. The word 玉菨 is glossed as tamamo “gem weed.” This is the only secure reading in these words because a note has been appended, which says 菨 “floating heart” is to be read mo. The image of “gem weed” appears often in Man’yōshū, but the intended image is that this weed floats on the surface of water; it is not sunken underwater. It is possible that this counterimage demonstrates that something is not right. It is also possible that this character is used as a loan for 翣 “feather decorations for a casket,” creating a subliminal reference to the sadness that occurred when Yamato confiscated this treasure. The symbolism appears to be that the treasure of Izumo needs to be returned, and, by extension, Izumo needs to be liberated from Yamato’s hegemony. I submit that these words are loaded with symbolism, much of which the Nihon shoki compilers may not have fully realized. The interesting fact hiding behind this story is, I believe, that anciently Yamato had heard stories of a sacred power in Izumo and wanted to control it. This is further reflected in the negative portrayal of most of the Izumo myths within Kojiki, where Izumo is either a place of banishment (for Susanoo) or subjugation by representatives of the sun goddess. Those sections Yamato deemed neutral appear to have been left intact. Thus, this is a rewriting of the myths by the historiographers at court. Kadowaki (1976:213) believes that after being subjugated, Izumo sent kataribe to the Yamato court. He argues that the change and distortion of the original Izumo myths not only happened at the Yamato court but also took place back in Izumo (ibid). Likely a core myth belonging to the Ou kingdom was later expanded to incorporate myths in surrounding areas of the larger territory of Izumo. It is thus interesting that there were elements of the Izumo myths that the Yamato court simply ignored, likely because there was no place in the Yamato mythical storyline where some of these could be grafted. An example is the creation myth in Izumo, known as the “land pulling myth.” This is preserved in Izumo fudoki, the beginning quoted below: The reason it is called Ou is because Yatuka Midu Omitunu no Mikoto who pulled the land together decreed, “The land of Izumo where the eight-fold clouds rise is because it is a young land like fine-woven cloth. I have created small pieces of land in the beginning. Now let us sew these together.” He looked over at Misaki in Silla, a land white like mulberry trousers, to see if there was any left-over land and said, “There is left over land.” He took in his hand a large spade, broad like the breast of a maiden, and cut away the land, as if he were prying open the gills of a large fish and sliced the land as if he were waving a banner of bundled ears of pampas grass. He threw a great rope woven of three strong strands and tied the pieces together, and as one would haul in the black stalks of frostbitten grass, he hauled in the rope back and forth. And like the slow movement of a ship on the river little by little, “Come here, land! Come here, land!” he said. The land that he hauled in and sewed together is now land from the tip of Kodu to the cape of Yahonikiduki. Thus, the strongly driven wooden stake is

Mythical beginnings 83 the border of the land of Iwami and Izumo; it is this, called Mount Sahime. The rope that he held in his hand and used to pull in the land is Nagahama of Sono. This process of carving out parts of various neighboring lands is then repeated three more times, searching several other areas around the Izumo coast. This myth ends with the words, “He decreed, ‘I have finished pulling the land together.’ He stuck his staff in the ground at the sacred forest in Ou and said, ‘Owe (meaning finished).’ Thus, this area was called Ou.” Another interesting fact is that Susanoo, in most mythological accounts, is on the losing side of an attempt to take control of Yamato (or the power invested in the sun goddess). The fact that many strains of the Susanoo myth relate that he was banished to the upper reaches of the Hi River again shows that Izumo was despised. Kojiki states, “Thus Susanoo was cast out and descended to the upper reaches of the Hi River in the land of Izumo, in a land called Torikami.” Izumo fudoki notes, “Mount Torikami, it is 35 li southeast of the district office (of Nita District). It is on the border of the lands of Hahagi and Izumo.” Regarding the river Hi (anciently spelled Pii), Izumo fudoki notes, “Pii River is located 57 paces due west from the district office (of Ōhara District). It flows west to the village of Tagi in Izumo District.” Kadowaki (1976:23) mentions the following line from Izumo fudoki: The great Izumo River has its origin from Mount Torikami located on the border of the lands of Hahagi and Izumo. It flows through the village of Yokota in Nita District and passes through the four villages of Yokota, Mitokoro, Misawa, and Fuse and goes through Hikinu Village on the border of Ōhara District. At last it then passes through the four villages of Kisugi, Hi, Yasiro, and Kamufara, and then it goes through Tagi Village on the border of Izumo District, passing through the two villages of Kafuti and Izumo, flowing north, where it bends west. There it passes through the two villages of Inu and Kizuki, emptying into the ocean at Kamudo. This is what society calls the lower reaches of Hi River. He concludes that the lower reaches of the Hi River refers to the great Izumo River. In the story of the eight-headed serpent, Susanoo sees chopsticks floating downstream and realizes that people are living upstream. He then finds a couple weeping, their last remaining daughter sitting between them. They inform Susanoo that a serpent comes from Koshi to devour one daughter each year. Kadowaki finds it difficult to reconcile these disparate places. He believes the place the mythologists had in mind is where Hi River and Ushio River come together. This may appear to be a fussy point, but he goes on to show that this area is close to the border of modern Ōhara and Ihishi districts. In this district is the village Susa. Izumo fudoki records: Susa Village, it is 19 li due west of the district office. Kamu Susanoo decreed, “This land is a small land, but it is a proper place to call my land. Thus,

84

Mythical beginnings my name should not be given to the trees and rocks.” Thus, he pacified his spirit and deposited his spiritual power here. He immediately established the Greater and Lesser Susa rice plots. That is why it is called Susa.

Kadowaki (1976:25) concludes that in its more primitive form, Susanoo, or “the male from Susa,” originally was not a deity of nature, as he is grouped with the sun deity and the moon deity in Kojiki and Nihon shoki but was a localized, heroic deity. It is even more probable that the original version described a legendary figure who was a chieftain from the area, someone who became a hero to the local population. Also, the myth about Susanoo slaying the serpent is simply an amalgamation of legends put together to glorify a warrior-king of the area. It is probable that the court expanded a primitive story of a mighty, young chieftain rescuing a maiden from the captivity of a neighboring group, marrying her and bringing peace to the region. As Ōbayashi (1973a:167–70) pointed out decades ago, many scholars, including W.G. Aston (1841–1911) and Shiratori Kurakichi (1865–1942), have noted that the myth of Susanoo slaying the eight-headed serpent is a mangled version of the well-known legend of Perseus and Andromeda. As the legend spread throughout the Middle East and Africa, the monster in question changed form to something multiheaded. This confusion may have originated from contamination with the story of Perseus slaying the Gorgon Medusa, whose hair was composed of many serpents before he came back and found Andromeda chained to the rock by the sea. One motif common with this legend and the version found in many locales in Asia is the discovery of an iron weapon in the body of the serpent after the hero has slain it. Aoki (1971:53) has extrapolated that as the area in question in Izumo was known for producing iron, this legend may have been popular in Izumo around the Kofun era (250–600 CE).9 Thus, producing the sword from inside the serpent is a metaphor for digging iron ore from the ground. It appears the local Izumo myth of Susanoo slaying a monster and marrying the daughter of some other chieftain was inserted into the Yamato mythological story with only minor changes. This likely was done to placate those at court with Izumo ties since it is clear that other parts of the Izumo myths were treated poorly by the historians. For example, Susanoo presenting the sword found in the serpent’s tail to Amaterasu is a thinly veiled symbolization of Izumo’s military submission to Yamato’s superior authority. This completes the subjugation of Izumo, taking their hero and having him transfer his military symbol to those in power in Yamato. Another story has Yamato Takeru (a symbol of the power of Yamato) travel to Izumo to bring peace. He makes a fake sword out of wood, and while he and Izumo Takeru (a representative of the powerful chieftains of Izumo) go to the Hi River to bathe, he swaps the real sword for a fake one. Yamato Takeru says that as a sign of friendship they should exchange swords. After they have dressed and girded their swords of friendship, Yamato Takeru says, “Let’s cross swords.” He draws his sword, but Izumo Takeru cannot draw his, because there is no blade. Yamato Takeru then slays Izumo Takeru. Here it is clear that Yamato Takeru has beaten his foe, not because of superior strength but because of superior intellect.

Mythical beginnings 85

The filtering of myths As demonstrated above, over time the court took a variety of heterogeneous myths with intersecting commonalities and massaged these into a set of narratives that portrayed the Yamato court holding overall authority and the outlying territories having offered their allegiance to the court. This point is important because the later historiographers use the same attitude in revising the court history. A discussion of this process necessitates an understanding that violence was done to the original myths. Naturally, the most violent method would be to suppress a myth, excluding it from the official record. As evidence for this, Miyake (1984:5–6) mentions four mythological episodes recorded in the “historical” sections of Nihon shoki that do not appear in Kojiki nor in the two books of Nihon shoki, specifically dealing with mythology. These appear below in order of appearance, quoted directly from Nihon shoki. (i) Kenzō, third year, spring, second month, first day. Abe Omi Kotosiro received the imperial command, and the court dispatched him to Mimana as an envoy. At this time, the moon deity possessed a medium and spoke to Kotosiro, “My ancestor, Takami Musubi, was involved in the meritorious service of forming both heaven and earth. Take land with their inhabitants and present these to the moon deity. If you present this to me as requested, there will be a reward of good fortune and rejoicing.” (ii) Kinmei, sixteenth year, spring, second month. Prince Yosiau of Paekche dispatched Prince Kuwei (Prince Kuwei is the younger brother of King Witoku) to (the Yamato court to) report, “King Mei has been killed by the enemy.” … Presently Soga Omi [Iname] enquired, “King Mei was very knowledgeable about the logic of Heaven and Earth, and his name was known throughout the four quarters. He preserved peace for many years, governing the western frontier countries. Though he wished to serve the divine ruler for a thousand years, even ten thousand reigns, he suddenly departed for a far off world and cannot return, like flowing water never is the same. We had no idea that he would be resting peacefully in a tomb [so quickly]. How painful and cruel. How sorrowful and lamentable. There will not be a dry eye among sentient beings. What sin has been committed to bring down such a punishment? What policy will now be followed to bring your country peace?” Kuwei answered, “I am but a fool and do not understand profound policy. How could I comprehend the origin of punishment or the survival of a nation?” Minister Soga said, “Anciently, during the reign of Divine Ruler Ōhatuse [Yūryaku], your country was invaded by Koguryŏ, and the situation of your country was as perilous as a pile of eggs. At that time, the Divine Ruler commanded the patriarch of the office of the deities of heaven and earth to inquire of the deities as to what policy to follow. The deities possessed a medium, instructing us, ‘If you entreat the deity who founded the country, and go to

86 Mythical beginnings the assistance of the ruler about to be destroyed, you will surely bring peace to that land, and the people will be subdued.’” “With these words, the Divine Ruler entreated the deities, sent military help and your country obtained peace. If we analyze the origins of this, we find that this is the deity who descended from heaven and founded the country when heaven and earth were divided, and vegetation still had the power of speech. We have heard reports that your country no longer worships this deity. If you repent of your carelessness, repair the shrines, and worship the spirit of the deities, your country will prosper. Take care not to forget.” (iii) Kōtoku, fifth year of Hakuchi, second month. At last, they (the envoy and his entourage from Yamato) arrived in the Tang metropolis and had an audience with the son of heaven. With this Guo Zhang Quan, Overseer of the Gates of the East Palace, asked about the geography of Yamato, and the names of the original deities of the land. All questions were answered accordingly. (iv) Tenmu, first year, seventh month. Before this time when the army had gathered at the well of Kanazuna, Taketi Agatanusi Kome, District Governor of the district of Taketi, found that he had become mute and could not speak. After three days, a deity possessed him and spoke, “I am Koto Sironusi, the deity who dwells in Taketi Shrine. I am also the deity Iku Mitama who dwells in Musa Shrine.” He then spoke the will of the deities, “Take horses and various weapons to the burial tumulus of Divine Ruler Kamu Yamato Ihare Hiko.” It was further revealed, “We stood in front of and behind the royal grandchild [Tenmu] and escorted him to Fufa. Even now, we stand in the midst of the royal army for protection. An army approaches from the western road. Be watchful.” Several comments are in order. First, two of these stories have a connection with the continent (ii) and (iii). We should be careful not to take the narrative at face value, as clearly some manipulation was likely done. In addition, there is a possibility that the connection with the peninsula is nothing more than a vestige of a link with immigrants who came across the straits to the islands of Japan. Second, the myths in (i) and (ii) hint at a story of the creation or founding of the state differing from that currently preserved in Kojiki, Kujiki, or Nihon shoki. Miyake (1984:6) comments: If the myth [in (i)] is from the era of Kenzō, then there is a possibility that a version of the creation of heaven and earth was in existence before the Kinmei court came to power that differed from that in Kojiki and Nihon shoki. Furthermore, the concept of “the deity who founded the country” appears to be a rather nebulous myth, likely brought to Yamato by immigrants from the peninsula. It should be remembered that the founder of both Koguryŏ and Paekche is someone whom the story relates had proclaimed, “I am a grandson of the deity of the river, a son of the sun.” Instead of evaluating the Shoki version as if Yamato were claiming common mythology with the peninsula because of

Mythical beginnings 87 a Yamato connection, I believe the Shoki compilers recorded the actual legend as it was but altered the emphasis. The original would have been that the immigrants from the peninsula brought with them the story of a land founded by a deity. It is possible that as centuries passed, the name of the founding deity became confused or lost. Soga Iname’s claim that “we have heard reports that your country no longer worships this deity” is a trick of the brush by the compilers to put the onus of not remembering who the founding deity was back on Paekche. Any hint of this connection to the Yamato mythology was erased when the official mythology was put together for inclusion in Kojiki and Nihon shoki. It is telling, however, that the usage of 自天降 “descended from heaven” in quote (ii) is also used exclusively in the mythological story of Nigi Hayahi, the divine hero of Kujiki. He does not assist in the creation of the land but that event may have been washed away as the myth went through the editorial process. Kujiki does not mention a connection with Paekche, but it still makes the claim that Susanoo had a connection with Silla. While this connection also appears in a variant comment in Nihon shoki, Kujiki goes much further in not only unabashedly claiming a Silla connection with Susanoo but also insinuating Silla blood in the lineage of the first three rulers of Yamato (Jinmu, Suizei, and Annei; cf. Bentley 2006:79–80). The myth mentioned in (ii) preserves evidence that an earlier strain of a creation mythology bridged the peninsula and the islands. As Miyake (1984:7) points out, by the time the court was sending envoys to Tang China (from 630 CE on), the Yamato court appears to have settled on a lineage of deities connected with courtiers who served at court. He goes on to note that the appearance of Jinmu (Ihare Hiko) in the fourth quote demonstrates that the basic structure of the Jinmu legend was already in place and predates the creation of either Kojiki or Nihon shoki. Here the existence of Kujiki allows us to examine a critically important historiographical ploy. The line about presenting weapons to the tomb of Jinmu illustrates that the compilers are trying to weave a common thread between Jinmu’s success in uniting the country (Jinmu versus the villain Nagasune Hiko) and Tenmu’s attempt to do the same (Tenmu versus the rebellious Prince Ōtomo). I earlier argued (Bentley 2006:93): Kujiki travels a path the other records do not. It divides Jinmu’s life into two sections: pre-ascension and post-ascension. This is reminiscent of Nihon shoki’s division of Tenmu’s record into two halves—the Jinshin Disturbance (pre-ascension) and his reign after his coronation. Also, Jinmu’s military and strategic leadership is demonstrated in the pre-ascension, but his ability to rule is captured in the post-ascension record. These characteristics are closely connected to the construction of Tenmu’s record, and I have asserted that the Kujiki author-editor rearranged the original story of Jinmu to mirror the downfall of Prince Ōtomo and justify the rise of Tenmu and praise his reign (Bentley 2006:94). My core contention is that Nihon shoki later incorporated this in its final version and not the other way around.

88 Mythical beginnings An analysis of all the available myths demonstrates that the early Bureau of History selected a core set of myths that were critical to the ideology of the ruling families. These include a connection with a heavenly deity who came down and created the islands, allowing the creation of the inhabitants of the islands. This suggests a fundamental narrative that the people were originally immigrants to the islands. Next, the ruling families plugged their genealogies into this heavenly pedigree. This myth likely is related to the same miraculous event that underpinned the Koguryŏ and Paekche ruling families. Another core myth outlined a ritual at court tied to solar eclipses. Here the myth had the sun hide in a cave because of some offense, and the officiators would invoke their power to solve (and purify evil), thus coaxing the sun back out. This simple event was later expanded to include elements of spirit worship present at court, where the rulers believed they could call upon the solar orb in the sky because of their lineage. It is interesting that in the myths, the land of Yamato is euphemistically called Ōyasima Kuni (大八島国) “the great land of eight islands” in Kojiki. Nihon shoki uses the same term but represents it with different characters as 大八洲国. Tsuda (1948, 1:342) argues that the name referred to the territory governed by the ruling family in Yamato. He goes on to mention that there are variations of the names of the islands or territories included in this number “eight” as found in Kojiki, the orthodox version in Nihon shoki, as well as the variants; however, never once is the peninsula included in this count.10 Notice, however, in a variant tradition we find that the deity of heaven commands Izanagi and Izanami with the words, “(Below) is the land of Toyo Ashihara no Chiiho Aki no Mizuho. You will go there and rule over it.” This euphemistic name is written as 豊葦原千五百秋瑞穂 “abundant-reed-plain-1500-gemlike ears of rice in the autumn (harvest).” The name refers to the islands as a blessed land that has an abundance of rice production in autumns that are long and prosperous. This likely is a newer, expanded version of the older Ashihara no Nakatsukuni 葦原中国 “reed-plain-middle-land” found in Kojiki. I believe this newer title is a product of immigrants to the archipelago after the introduction of rice cultivation. Miyake (1984:176–77) has postulated that by looking at the kuni yuzuri “land-transfer” myth as well as the “eight-headed monster” myth, Ashihara no Nakatsukuni was originally located in Izumo (Miyake 1984:248). He finds that the various versions of the “land-transfer” myth can be organized into two groups: (i) Takami Musubi is the ruler, and Futsunushi is dispatched as the envoy with Takemi Kazuchi as an assistant. This is the older version of the myth; (ii) Amaterasu is the ruler, with Takemi Kazuchi as envoy and Futsunushi as an assistant. This is the newer version, dating from around the middle of the seventh century. It is possible to conclude that chieftains in Izumo had preserved their traditions from an earlier time than Yamato; thus Yamato felt obliged to rely on the Izumo traditions to strengthen and fill out their own legends. In the end, Yamato subjugates Izumo, but influence from Izumo culture continued to permeate the court and survived in the mythology when Kojiki and Nihon shoki were compiled. Finally, I conclude that these complex myths preserve a simple fact about East Asian political contact: Paekche and China sought access

Mythical beginnings 89 to the Yamato court through Hakata, while Silla originally sought access through Izumo until Paekche was destroyed in 660.

Hyūga as the birthplace of the imperial family Many have wondered why the ruling family has its origins in a mountainous area far removed from the capital basin, in a remote area called Hyūga. Anciently known as Himuka (Hyūga) 日向 “facing the sun,” the records of Kojiki, Kujiki, and Nihon shoki portray this land far west of Yamato as the birthplace of the imperial family. Kojiki’s story is the least embellished: Amaterasu and Takami Musubi together declare the land of Yamato to be pacified and command their son, Ame no Oshiho Mimi, to descend and rule the land. Oshiho Mimi has a son, Ninigi, during the interval of the pacification of the land, and decides to defer to him. Ninigi descends from heaven with five other deities who represent powerful families at court. He lands on the floating bridge of heaven, and then instead of descending to Yamato, which had just been pacified, he descends to Mount Takachiho in Hyūga in Tsukushi. Ninigi scouts an appropriate place to build his palace, and when he finally finds it, he declares: This place faces directly toward the land of Kara (Han), a land you arrive at, straight through the cape of Kasasa, a land the morning sun directly pierces, and the evening sun shines down on. Therefore, this is a very blessed land [此地者向韓國。眞來通笠紗之御前而。朝日之直刺國。夕日之日照國 也。故此地甚吉地詔而]. As is worthy of the offspring of the sun goddess, Ninigi marries the most beautiful woman in the land and gives birth to several sons through a trial of fire. One of these, Hiko Hohodemi, gives birth to Fukiaezu, who is the father of Jinmu. This family is all born in Himuka. It is instructive that the story in Kojiki records that this land faces directly toward the Korean peninsula, making the current connection with Miyazaki Prefecture (anciently known as Hyūga) improbable,11 as it sits on the eastern coast of Kyūshū. Most scholars try to solve this problem by working retroactively, noting that the land of Kumaso seems to be situated around modern Kumamoto, and we know Hyūga was down the eastern seaboard of Kyūshū. This forces scholars to downplay the words of Ninigi calling Hyūga a blessed land that faces Kara, a reference to the benefits of technology and commerce exchange with the peninsula, as well as diplomatic contact with the kingdoms across the straits. A careful reading of the story in Kojiki lends evidence to Hakata being the location of the original Himuka. Kojiki has Jinmu leave Himuka and travel to the USA and then to the Okada Palace in Tsukushi. While this location is unclear, Hakata is the best port to leave before heading through the Inland Sea. It is plausible that the original story had Jinmu leave from Okada Palace to conquer Yamato, and later historians found they needed to tweak the story, because over time the name Himuka had migrated south. Tsuda (1948, 1:159–60) argued that as the

90

Mythical beginnings

imperial family claimed to have originated from the sun goddess, a birthplace worthy of that was needed, thus somewhere in the west was selected, a place that figuratively represents the sun, as the sun sets in the west. I have created two maps: Figure 3.4 illustrates the current thinking, and Figure 3.5 demonstrates how I believe the ancient myths originally envisioned the island of Tsukushi (later Kyūshū). Thus, the mythological story as found in our three records postdates parts of the underlying story. Tsuda (1948, 1:160) admits that the name himuka refers to a broad area, but he follows the foundation story found in a fragment from the lost Hyūga fudoki, which places Hyūga where it is historically located. One way out of this conundrum is to expand our horizons and see the term himuka as originating from *pe-muka “face-west.” Thus, it is easy to associate the word Himuka to a large area, that being much of Kyūshū, which later came to refer to the remote area now known as Miyazaki and Kagoshima. Maruyama (1955:330–31) believes that the court simply placed the foundation myth in a locale that was assimilated into Yamato’s sphere near the end of the process, and then merely inverted the course. However, it is difficult to understand why the court would do that. Why locate the birthplace of the royal family in a locale that pledged its allegiance to the court at such a late date?

Figure 3.4 Kumaso versus Hyūga: common thinking. This map was based on “File: Topographic Map of Kyushu.png.” Wikimedia Commons, the free media repository. https://commons.wikimedia.org/wiki/Atlas_of_South_Korea (accessed June 7, 2017).

Mythical beginnings 91

Figure 3.5 Kumaso versus Hyūga: new proposal. This map was based on “File: Topographic Map of Kyushu.png.” Wikimedia Commons, the free media repository. https:// commons.wikimedia.org/wiki/Atlas_of_South_Korea (accessed June 7, 2017).

While Kojiki has Ninigi declare that the location is auspicious because it faces the peninsula, the orthodox story in Nihon shoki relates that Ninigi descends on a mountain in So of Himuka. He then travels and resides in Kasasa of Ada, which most scholars locate in modern Kagoshima. The two variations in Shoki simply say he descended on Mount Takachiho in Himuka. None of these episodes includes a reason why Ninigi resides there, far from Yamato, nor do they orient the reader as to where Himuka is in relation to Kyūshū. The story in Kujiki is reminiscent of that in Kojiki: The Imperial grandchild went and resided there. He proclaimed, “This land faces the country of Kara (Han). You come straight to the cape of Kasasa. The morning sun shines directly onto this land, and it is a land which is even illuminated by the evening sun.” Thus, he labeled this “good land.” (Bentley 2006:229–30) One plausible theory is that the skeleton of this story is based on a legend prevalent in the immigrant culture, where any land on the western coast facing the peninsula was deemed “good land,” seeing the peninsula as representing an excellent repository of culture and technology. I envision this story following the immigrants up

92

Mythical beginnings

to Yamato, where it became part of their mythological fabric. Yamato may have later reinterpreted the connection with the peninsula as a sign of prestige and economic prowess. However, it is significant that none of the versions preserved in Nihon shoki make mention of this. The only mention of Kara (Han) in the mythological stories in Shoki is in connection with Susanoo. On the other hand, Kara (Han) is only mentioned in Kojiki in connection with Ninigi. Kujiki is thus the sole record of the three to include both strains, Ninigi and Susanoo. The final issue here is the reasoning behind the historians’ decision to have Jinmu leave Himuka and sail toward Yamato. Kojiki gives this simple motivation: “‘Where should we dwell so that we may rule all under heaven in peace? I think we should go eastward.’ And they immediately departed from Himuka.” Shoki provides a more elaborate reason: Moreover I inquired of the old man, Shihotsutsu, and he told me, “There is a beautiful land in the east encircled by green mountains. It is there that someone descended from heaven in the Heavenly Rock Boat.” I believe we can expand the great work in that land to extensively consolidate everything under heaven. Would it not become the center of the six-paired directions? And perhaps the person who descended from heaven was Nigi Hayahi. Shall we not go there and build our capital? Interestingly, Shoki makes this cursory statement about the central figure in the mythical story in Kujiki. The genealogy contained in Kujiki declares that Ninigi had an older brother, Nigi Hayahi, who was given the political power and descended to Yamato in a Heavenly Rock Boat. Ninigi was granted military power and descended to Takachiho in Himuka. Jinmu departed Himuka, defeated Nagasune Hiko, and established his capital in Yamato; he then gave the audience to Nigi Hayahi and took control of the two branches of power—political and military—and combined them. This is the story the historians behind Kujiki relate. It makes one wonder if the ancient court preserved a tradition of two strains of “Wa”: one in Kyūshū and one in Yamato. By having Jinmu combine the two, the court would be saying, “We have united the islands, including both the military and religious power.” In the Shiki minutes is the following question and answer: Question: There are two designations of Wa, a northern and a southern Wa. What does this mean? Answer: The theory from [the scholar of the earlier] Engi lectures stated that the northern Wa refers to this (our) land, while the southern Wa refers to the Queen’s (Himiko) territory. … [However] there is no evidence for this theory. (Kuroita 1932:185) Most scholars focus on the phrase “there is no evidence for this theory,” and ignore the fact that there was a tradition of two different Wa spheres in Japan,

Mythical beginnings 93 regardless that the court could find no evidence for the tradition in the early Heian era. If anciently there were two branches of the Wa (Himiko in the west, and an immigrant ruler in the east), then the historians are using Jinmu as a unifier, someone who united these two factions into one. The key conclusion is that the historiographers had many data points they tried to reorder to bring some uniformity to the mythical chaos. This meant that a great deal of violence was required to make the myths run smoothly in a linear fashion.

Notes 1 This translation has been informed by the translations in Major et al. (2010:86, 114, 276, 536). 2 Ōbayashi (1973b:65–66) says much the same thing but disregards Kujiki in his analysis because he believes the text to be derivative and thus of little value. He states, “Kuji hongi’s mythological set up should be viewed as a product of compilation in a later era, so we cannot attach any importance to it.” It is interesting, however, that if Kujiki’s mythical story is so strikingly different from Kojiki and Nihon shoki, then how is it derivative, and from what sources did it take its material? Is it not more balanced and equitable to admit that Kujiki has preserved variant data that have been lost? 3 Research into foreign origins of many aspects of Japanese myths was argued in detail in the nineteenth century by a few European scholars before Japanese scholars became involved in such work (Ōbayashi 1973a:10). 4 This change in the mythology was argued much earlier by Takagi (1973:15). 5 The word in the original is 嶋 “island,” but as I have argued (Bentley 2008a:14, 17), the language of the Ryūkyūs originally had two etyma: *sema “island,” and *sima “village.” I believe that by the time of the recording of these traditions, these two forms had merged, becoming one and the same in the language of Shuri. Thus, I believe Amamiku is creating villages, not the actual islands. 6 Taken from Ifa (2000:388–89). 7 The following is a synopsis of Kadowaki (1976:90–110). 8 Cf. Kojima et al. 2002, 1:293, n20. 9 The spread of her postulated dates is 300–700 CE, but scholars have refined these dates according to a series of newly excavated evidence, which I have noted above in parentheses. 10 Tsuda (1948, 1:342–43) argues that ya “eight” in the name takes on a political significance in this story. I would argue that ya was a magical number connected to the ruling family in Yamato. 11 Anciently, the land of Hyūga included what later was divided into Hyūga, Ōsumi, and Satsuma, or modern Miyazaki and Kagoshima Prefectures.

References All works are published in Tokyo unless otherwise noted. Aoki, Michiko Yamaguchi. 1971. Izumo fudoki. Sophia University. Bentley, John R. 2006. The Authenticity of Sendai kuji hongi: A New Examination of Texts, with a Translation and Commentary. Leiden: Brill. Bentley, John R. 2008a. “The Search for the Language of Yamatai.” Japanese Language and Literature, 42: 1–43. Bentley, John R. 2008b. A Linguistic History of the Forgotten Islands: A Reconstruction of the Proto-Language of the Southern Ryūkyūs. London: Global Oriental.

94

Mythical beginnings

Ifa, Fuyū. 2000. Ko-Ryūkyū. Iwanami Bunkō. Kadowaki, Teiji. 1976. Izumo no kodaishi. NHK Books. Kobayashi, Toshio. 2006. Nihon kodai kokka keiseishi-kō. Kōsō Shobō. Kojima, Noriyuki, Kōjirō Naoki, Kazutami Nishimiya, Susumu Kuranak and Masamori Mōri, eds. 2002. Nihon shoki. Shōgakkan. Kuroita, Katsumi, ed. 1932. Nihongi shiki. Vol. 8 of Kokushi Taikei. Yoshikawa Kōbunkan. Major, John S. 2010. The Huainanzi: A Guide to the Theory and Practice of Government in Early Han China. New York: Columbia University Press. Maruyama, Jirō. 1955. Nihon shoki no kenkyū. Yoshikawa Kōbunkan. Miyake, Kazuo. 1984. Kiki shinwa no seiritsu. Yoshikawa Kōbunkan. Mizoguchi, Mutsuko. 1992. “Kaiseki shinwa no iden to Kojiki no hensan ito.” In Nakamura, Hirotoshi, Aoki Shūhei, Sugano Masao, and Yamazaki Masayuki, eds. Umezawa Isezō Sensei tsuitō kiki ronshū. Zoku Gunsho Ruijū Kanseikai, pp. 29–49. Ōbayashi, Taryō. 1973a. Nihon shinwa no kigen. Kadokawa Shoten. Ōbayashi, Taryō. 1973b. Yottsu no jindai. Contained in Vol. 7 of Nihon shoki Kenkyū. Hanawa Shobō, pp. 63–89. Ōbayashi, Taryō. 1987. Tenjin to nishu no tochi no konton. Contained in Vol. 16 of Nihon shoki Kenkyū. Hanawa Shobō, pp. 31–42. Okada, Seishi. 1970. Kodai ōken no saishi to shinwa. Hanawa Shobō. Takagi, Toshio. 1973. Nihon shinwa densetsu no kenkyū, Vol. 1. Tōyō Bunkō. Takioto, Yoshiyuki. 2005. Kamigami to kodaishi no nazo o toku: Kojiki to Nihon shoki. Seishun Shuppansha. Tsuda, Sōkichi. 1948. Nihon koten no kenkyū, 2 Vols. Iwanami Shoten. Wakai, Toshiaki. 2006. Kiki shinwa no genzō to Yamato ōken no kigen. Contained in Vol. 27 of Nihon shoki Kenkyū. Hanawa Shobō, pp. 289–315.

4

From kings to heavenly sovereigns

A discussion about Yamato historical periods requires prudence. The historian Kobayashi Toshio (2006:174) has argued that while there are varying opinions about the royal lineage of the rulers, how advanced each administration was, or the level of sophistication and development of “the state,” all scholars agree that the center of power was in Yamato. Other than this narrow consensus, there is little concrete evidence to point historians in the proper direction toward a more tangible description of the state of earlier Japanese history. Much of the somewhat reliable evidence about earlier periods in Japanese history centers on archaeological data and information contained in foreign records from China, Paekche, and Silla. With this caveat about our data, it is still reasonable to claim that a peek into the early seventh century is enlightening, as most believe this was a time when historiography began to diversify. As noted earlier, the Chinese work Suishu (656) provides some interesting material on Japan and its rulers. Suishu records: In the twentieth year of Kai-huang (600), Wa had a king. His cognomen was *a-mwai (阿毎), and his agnomen was *Ta-ri-sɨ-pji-kuo (多利思比孤), and he was therefore called *a-puai-kiei-mjie (阿輩鶏彌). He sent an envoy to Our Capital. When We asked our officials to question them on their customs, (the Wa envoy) replied, “The king of Wa makes Heaven his elder brother, and the sun his younger brother. Before the sun rises, he goes out and listens to government, sitting on his throne. When day breaks he stops his work and declares that his younger brother will take over from there.” … The king’s wife is called *kiei-mjie (鶏彌). The Palace of the Queen employs 600–700 women. The Crown Prince is called *ri[wa]-ka-mjie-tapjwət-ri (利[和]歌彌多弗利), but he has no fortress. … The Wa (originally) had no writing, but only carved on wood and tied knots in rope. They revere Buddhism. They requested Buddhism from Paekche, and (when granted) that was the first time they had writing. They know about divination and have had mediums from ancient times. … Both Silla and Paekche treat Wa as a superior country, claiming it has many rare treasures, and they pay their respects by sending envoy missions regularly to the Wa court. … In the following year (fourth year of Tai-ye [608]) the Emperor dispatched the literary scholar,

96

From kings to heavenly sovereigns Bei Qing, to Wa as envoy. His group crossed through Paekche and reached Bamboo Island. Facing south to the land of Tanra, they crossed over to *tuosje-ma (Tsushima), which is far off in the great sea. Again, heading east they reached *ik-kie (Iki), then reached *tuk-sje (Tsukushi). Continuing east they reached the land of *dzjen-wang.

Considering this account describes the era when Kojiki, Kujiki, and Nihon shoki claim Suiko, a female, was on the throne, it is perplexing that there is no mention of a female sovereign (女王). Furthermore, notice what this unassuming entry from Suishu records: “In the third year of Tai-ye (607) the king (of Wa), *Ta-ri-sɨ-pji-kuo, sent an envoy who presented tribute at our court” [大業三年、其王多利思比孤遣使朝貢]. It is interesting that here the king is again identified as Tarishi Hiko, clearly the name of a male king. The veracity of this account is strengthened by the fact that the Sui court sent Bei Qing to observe the Yamato court, and his record makes no mention of a female sovereign. The record notes that the cognomen of the ruling family is ame “heaven.” It is important to note that Suiko’s grandfather, Kinmei, is the first recorded ruler with the element ame in his name: Ame Kunioshi Haruki Hironiwa. Every ruler after Suiko, from Jomei (r. 629–641) to Jitō has this element in his or her name. Strikingly, the name of Suiko, Toyo Mike Kashikiya Hime, does not include this element. The Sui record also notes that the ruler’s agnomen is Tarishi Hiko, which presents historians with another conundrum. Neither Suiko nor Shōtoku Taishi has this designation. Jomei, the ruler after Suiko, is called Tarashihi Hironuka, and Kōgyoku is known as Tarashi Hime. The four earlier (legendary) rulers with “Tarashi Hiko” in their names are Kōan, Keikō, Seimu, and Chūai, though this last ruler’s name is Tarashi Nakatsu Hiko, “Tarashi Hiko who is in the middle.” Finally, the Sui record mentions that the sovereign is called *a-pəj-kɛj-mji (阿輩鶏彌), which I believe represents opo-kemi and with later vowel raising this would have resulted in opo-kimi “great lord” (cf. Miyake 2003b). The next interesting piece of information is the line about the crown prince. The Chinese record that he was called 利歌彌多弗利 *ri-ka-mji-ta-put-ri, but Old Japanese does not allow r- initial morphemes in word initial position, so it is reasonable to assume that 利 is a copyist’s error for 和,1 resulting in the name *wa-ka mji-ta-put-ri or Waka Mitaputori (modern Waka Mita Futori). Notice that Sushun is called Waka Sazaki, and Bidatsu is Waka Futo Tama Shiki. Again, Shōtoku Taishi, the crown prince under Suiko, lacks this “flowery title.” It is possible that the crown prince at the time was labeled “younger Bidatsu” (Waka Futori). The critical point I wish to make is that no matter how we dissect the evidence, the problem remains, because if this designation does indeed point to Shōtoku Taishi, then we have a new problem, as the lineage is at variance—the current lineage shows he was not the son of Bidatsu but of Yōmei. It is fortuitous that the Sui have left a record of their contact with Yamato, but historians must address several places where information is at variance with Kojiki and Nihon shoki. Some Japanese scholars believe that this account is just a description of customs and not of a specific reign (cf. Masumura 1988:130).

From kings to heavenly sovereigns

97

Other scholars believe that the Chinese botched this account2 or it is an outright fabrication.3 Along these same lines, Piggott (1997:349, n. 62) mentions Kadowaki Teiji’s earlier argument that because the Suishu mentions a male king, Prince Shōtoku must have been the actual ruler. To this deduction she responds, “I do not find Kadowaki’s argument persuasive, however, and see no reason why later Nihon shoki compilers would have or could have attempted such a fraud.” Wang (1994:28) notes that the Yamato court reopened diplomatic channels with the Sui in 600, abandoning the traditional, subservient title of 倭王 “king of Wa,” and instead used the king’s personal title, Ō-kimi-ame-tarashi-hiko, “rendered phonetically in Chinese characters, to display his independent authority and dignity.” Wang further argues that the only reason the Sui court accepted this new posture was because the Sui court interpreted the title as a personal name and missed the cue that this was a politically motivated change in decorum. Wang skirts the problem with Suiko being a female ruler but consistently refers to the ruler in Yamato as “he.” To his credit, Tsuda (1948, 2:139) argued that the account is factual, but the facts were not recorded in any Japanese record, creating the discrepancy. I would rather suggest that our records—Kojiki, Kujiki, and Nihon shoki—have engaged in significant manipulation of their documentary material.4 Ōyama (2009:104–08) argues that the important narrative about the king of Wa is based on an eyewitness account and demonstrates that the ruler was a male. His reasoning hearkens back to Hanzawa (2002:308), who argued that the basis for the record in Suishu in relation to Wa (Yamato) was more than just an interview by the Sui court of the Wa envoy. The Chinese court dispatched Ambassador Bei Qing and his entourage to Yamato on a fact-finding mission, and their observations then became the basis for the report. As an example, Suishu records the following about Yamato society in the year 607: Among their customs, everyone convicted of murder, armed robbery, or rape is put to death. People found guilty of theft make restitution according to the value of the items stolen, and if they cannot make restitution, they are made slaves. Other crimes are punished according to their severity, some by banishment, and others by caning. The record in Suishu is very detailed, and as far as we can tell from other evidence, quite accurate. Regarding the strict society Suishu describes, consider the following piece of supporting evidence. Kujiki preserves an edict recorded in the 28th year of Suiko not found in Nihon shoki. It reads: Disloyalty denotes anyone who does not show respect to the Sovereign or the Queen. Being unfilial describes anyone who does not show respect to his or her deceased father and mother. If someone tries to hide disloyal, unfilial people by not reporting them to the authorities, then he becomes an accomplice,

98

From kings to heavenly sovereigns and will be dealt as severe a punishment as the offender [奉爲君后謂不忠者 亦奉爲考妣稱不孝者若不舉達而隱之者同處其罪重科刑法也]. (Bentley 2006:359)5

This edict lends evidence about the severity of society at the time. Thus, we should not rashly sweep aside this evidence in Suishu. The account in the Sui record forces us to conclude that the picture of Suiko’s court portrayed in Kojiki and Nihon shoki is not as accurate as we would have hoped, even for a record from the early seventh century. Consider a second Suiko edict preserved in Kujiki not found in Nihon shoki. In the 27th year of Suiko, Kujiki records: The court issued a decree, stating, “In seeking for subjects who faithfully serve their Sovereign, we find that such a subject in reality is like a person who reveres and loves both his parents. This is because our father is Heaven. Thus, following the will of Heaven is known as filial piety. Again, the Sovereign is the sun. Following the will of the Sovereign is called loyalty. The Queen is the moon and the mother. When one follows her will, that person becomes her subject” [制曰夫覔事君竭忠之臣者實在崇優二 親之子也何者夫父者天也故順天謂孝矣復君者日也故從君謂矣其后者 月也亦母也故順此謂臣]. (Bentley 2006:357) It is intriguing that both Suiko edicts mention “the sovereign and his wife (the queen).” Naturally, this is set up as a precedent that mirrors society, because the ideal is to teach that the throne matches the life of the people: there is a father and mother, just as there is a king and a queen. The disparity in Kujiki between the edict and the record portraying a widowed queen on the throne suggests that either the edict was added later to the record or the edict was original, and the actual occupant of the throne was altered later. Whatever the relationship, Kujiki’s record says very little about a supposed female sovereign, setting up a disconnected scenario. The record of Suiko comments that when Suiko was 39 years of age, “which is the fifth year of Great King Hatsuse, eleventh month, the Great King was murdered by Minister Soga Umako Sukune” (Bentley 2006:354). It then follows with this line, “The throne was empty.” Once the historian has placed Suiko on the throne, the remainder of the Suiko record deals with Shōtoku Taishi, as if he is the actual ruler. This depiction reminds the reader of the sleight of the hand found at the beginning of Kojiki, where the musubi line of five deities is inserted at the beginning of the mythological record, and then it is mysteriously forgotten. In the fourth month of the first year of Suiko, Kujiki records: “The court established Prince Umayato Toyoto Mimi as Heir to the throne and made him Regent over the country (録摂政), having control over the myriad minutiae of the government.” In other words, if you take Suiko out of the picture, very little changes, as it is clear that a male ruler is running the government. This denuded depiction found in Kujiki reflects the account in Suishu. I return to this issue when dealing with the

From kings to heavenly sovereigns

99

historiographical manipulation of Kojiki in Chapter 6. My purpose for bringing it up here is to argue that scholars should give the Suishu evidence fair weight, and not brush it aside because it varies from the Japanese native accounts. It is important to have as accurate a view of history as possible to determine how the early Yamato historians dealt with recording and/or reconstructing their own history. Evidence as we have it suggests that by the Tenmu era, the Bureau of History categorized the various records into two subgroups: “events surrounding the royal family” and “people and events outside the ruling family.” This dichotomy is retroactively reflected in the first mention of a historiographical movement in the 28th year of Suiko when the record states that Shōtoku Taishi and Soga Umako worked together to compile a series of records, Tennōki, “a record of the heavenly sovereigns,”6 and Kokki, “a record of the provinces (provincial rulers).” In the 10th year of Tenmu (681 CE), the great ruler orders the compilation of a record, mentioning Teiki and “The various accounts of ancient events.” Of course, Yasumaro’s preface to Kojiki contains the title Teiki and Honji or Kuji. Here we have three different accounts, and while it is difficult to discern how much of this is accurate or part of a construct, it is clear there was an underlying tradition of a dichotomy within the perception of historiography. To those at the Yamato court, “history” was bifurcated into the elevated and the common. I agree with Kōnoshi (1983:45–47) that rather than speculating about what Teiki or Kuji contained or lacked (or even avoided), it is more productive to glean information from the texts that have survived. In the next two chapters, I address the creation of a history, demonstrating an evolution from a limited record of rulers, or the traditional “king lists,” to an unbroken succession of rulers who claim divine lineage. There are several scholars who have argued that Yūryaku was the first historical king,7 based on evidence such as the first poem quoted in Man’yōshū being attributed to Yūryaku (cf. Kishi 2003:186–87).8 Umezawa (1968:124–25) has a chart listing every source quoted in Nihon shoki. Of the 18 unique sources quoted in Nihon shoki, 11 are quoted first in Yūryaku or thereafter. The emphasis suggests that the greatest amount of surviving source material originated at or after the time of Yūryaku. This evidence should not be interpreted to mean that these sources appeared during this period, but to me this signifies that there was an old tradition that historiography began around the time of the reign of Yūryaku. Supporting evidence for this tradition also exists in inscriptions from excavated artifacts. Yūryaku happens to be the oldest identifiable ruler on any kind of excavated artifact (Inariyama Tumulus sword and Eta Funayama Tumulus sword inscriptions). Also, the Inariyama Tumulus sword inscription mentions that the Great King resided in the *Se-kui Palace, which later comes to be known as “the Shiki Palace.” Yūryaku’s palace is recorded as having been in Shiki agata in Yamato. It is also important to remember that Yūryaku’s assistant, Wa Waka, knew his own lineage back six generations, so it makes sense that Yūryaku and his court had a memorialized lineage back at least six generations. Until other evidence appears, it seems certain at this point that Yūryaku may have been the first ruler to designate himself “great king,” a title I believe was

100 From kings to heavenly sovereigns created in imitation of the practice on the peninsula. Notice that on the Kwanggaet’o Stele appellations such as 大朱留王 “Great Suru (Chuyu) King” and 永楽太王 “Grand King Eiraku (Yŏngak)” appear. In the 1802-character inscription, there is one example of 大王 “great king,” nine of 太王 “grand king,” and 17 of 王 “king, prince.” Most examples of 王 deal with events surrounding the royal authority or court, while 大王 and 太王 refer directly to the person of the king. The fragmented inscription from the Koguryŏ Epitaph of Mutoru (dated around 391–413 CE) contains one example each of the appellations 太王 and 大王, and the title 聖王 “sagacious king” twice. The Chungwŏn (中原) Koguryŏ border marker inscription (dated 449 CE) contains the title 高麗大王 “Great King of Koryŏ.” Such usages do not appear on Paekche artifacts but perhaps only because of the dearth of surviving evidence. Piggott (1997:54–59) illustrates how Yūryaku’s realm was quite vast, with a sword excavated in Kyūshū in the west and one in Musashi in the east, both containing the great king’s name.9 While Yūryaku surely is not the first actual king in Yamato, it seems highly plausible that Yūryaku’s court marks a turning point, a watershed where the capital and his court reached a level of cultural power previously unknown in the islands. There is debate about whether this constituted a Yamato state,10 but what is of critical importance to us here is the fact that after the establishment of diplomatic ties with the peninsula during a previous dynasty, Yūryaku and his court were able to enjoy the fruits of engagement with advanced cultures such as that found in Paekche and China. For the purpose of this research, I use Yūryaku as a time marker, the first court to make an organized attempt at compiling a history of the ruling family. The next step is a look into the royal lineage preserved in Kojiki and Nihon shoki because as I have already argued, genealogy formed the starting point of any history in Yamato. Ōishi (1975:82) has argued that the rulers from Jinmu to Kaika are fabrications,11 inserted into the record during the reign of Tenmu. Kobayashi (1994:206) believes that the “eight rulers lacking a history” (kesshi hachidai 欠史八代, Suizei to Kaika) were inserted into the record as a prelude for the true ruler, Sujin. He also argues that historiographers created these eight “place holder” rulers to allow women from agatanushi families to preserve their genealogical ties to the ruling family. This theory dovetails nicely with my own argument that powerful families had orally preserved their family history, including accounts of their work in the service of specific rulers. This evidence suggests that originally there was no royal family, the right to rule rotating among a few powerful families. Deciphering which is a historical ruler and which is a fabricated one is difficult. Using the death dates that appear in Kojiki in small script, which some have argued were added later to the text by Yasumaro (cf. Nishimiya 1993), Mizuno (1968:100–16) believes that the four names (Seinei, Kenzō, Ninken, and Buretsu) between Yūryaku and Keitai are creations, inserted into the record in place of a female ruler, Princess Ihidoyo (see also Mizuno 1954:78–81). Each of these four rulers has a thin record, aside from legendary stories or episodes gleaned from continental records. There is also a scholarly consensus that the insertion

From kings to heavenly sovereigns

101

of the image of a violent and gluttonous Paekche king in the reign of Buretsu, who is also portrayed as an evil ruler, is meant as a Paekche mirror image. This strengthens the common theory that Buretsu is nothing more than a fabricated, “transitional” ruler. Regardless of what we deem factual or fabricated, it is clear the records as handed down to us are not reliable, even when it comes to basic information, such as where a specific ruler resided. As an example, it is true that the Inariyama tumulus sword mentions Great King Waka Takeru (Yūryaku) reigning from the palace in Shiki, but Kojiki labels his palace Asakura, while both Kujiki and Nihon shoki make his palace Asakura in Hatsuse. Scholars agree that Asakura was anciently in the limits of the modern city of Sakurai in Ōsaka; tradition states that anciently that area was called Shiki. Consider, however, that the inscription on the Suda Hachiman mirror mentions the *O-se-sa-ka Palace, which later becomes the place name Oshisaka. The name Oshisaka appears in the Jinmu record of Kojiki, but it is a place where the enemies of Jinmu are hiding. The Kōgyoku record in Nihon shoki mentions that the court exhumes Jomei’s body and reinters it in a tomb in Oshisaka. Oshisaka is also located in Sakurai City, thus close to Shiki, but neither Kojiki nor Nihon shoki make any mention of an Oshisaka Palace. This omission (or revision) is even more startling if the date on the inscription on the mirror is 503 CE. It would seem that the conversion of simple biographical facts into a lyrical account by the kataribe meant the embellishment of those facts. The other, stronger possibility is the manipulation of facts later by the Bureau of History. One innovative way that scholars have tried to sift through these king lists to filter out historical and fabricated rulers is to examine their names. Maruyama (1955:19) originally argued that the earliest records (like Teiki) contained basic information, including the name of the ruler. Based on the work of Mizuno, Inoue (1967:267) examines the names of the various rulers and claims that Jinmu to Kaika are fabrications (or at least spruced up individuals who may only have been names in a legendary storyline). Tokumitsu (1978:49) categorizes the rulers from Jinmu to Suiko into five groups based on the composition of their names: (A) Jinmu to Kaika, (B) Sujin and Suinin, (C) Keikō to Nintoku, (D) Ōjin to Buretsu, and (E) Keitai to Suiko. Based on the work of Mizuno, Inoue, and others, Tokumitsu (ibid) argues that the names of these first nine rulers in (A) are later creations, as most of their names include elements of the names from Asuka rulers Jomei through Shōmu. In addition, those from Jinmu through Keikō demonstrate a straight, linear “father to son” succession, something difficult to believe, whereas many of the rulers after this period follow a lateral succession: “brother to brother then son.”12 It is interesting to note that even some of the names of the rulers across our three records appear to be in flux. The names of several rulers exhibit variation among the historical accounts. Consider the six names in Table 4.1. Table 4.1 demonstrates that some form of manipulation existed even during the early Nara era (see Kobayashi 2006:185). In spite of this, we can still argue that there is a core name here, and compilers (or families) have embellished the name over time. Research by Tsuda (1948, 1:80) finds that the core or basic name

102

From kings to heavenly sovereigns

Table 4.1 Spelling of certain names of rulers

Ōjin Seinei

Kojiki

Nihon shoki

Kujiki

Homuda Wake Shiraka Ōyamato Neko

Homuta Shiraka Takehiro Kunioshi Waka Yamato Neko Woke (弘計)

Same as Shoki Same as Shoki

Kenzō

Woke (袁祁) no Ihasu Wake Ninken Ohoke Oke Ankan Hirokuni Oshitake Kanahi Magari Ōe Hirokuni Oshitake Kanahi Sushun Hatsusebe Waka Sazaki Hatsusebe

Woke (雄計) Same as Shoki Same as Kojiki Same as Shoki

Table 4.2 Core names of rulers Three syllables

Name

Five syllables

Name

Wousu Sazaki Wakugo Anapo Shiraka Woke* Ohoke Wohodo Kanahi Haruki

Yamato Takeru Nintoku Ingyō Ankō Seinei Kenzō Ninken Keitai Ankan Kinmei

Ihare Hiko Mimaki Hiko Ikume Hiko Oshiro Wake Tarasi Hiko Nakatu Hiko Tarashi Hime Homuda Wake Izaho Wake Mizuha Wake Waka Takeru Ihidoyo* Waka Sazaki Ositate*

Jinmu Sujin Suinin Keikō Seimu Chūai Jingū Ōjin Richū Hanzei Yūryaku Princess Iidoyo Buretsu Senka

of most of these rulers had a nucleus of three or five syllables. Taking a clue from this observation, Table 4.2 lists the names of major rulers from Jinmu down to Kinmei. As can be seen in this table, most of these rulers, minus the “eight rulers lacking a history,” have names built around a core element. Only three names have two or four syllables, an aberration marked with an asterisk. To these basic names, the kataribe attached an appellation of respect or royalty. I theorize that there were two competing appellations for the king who ruled over the realm. The older appellation was tarasi piko. This appellation appears in the names of the following rulers: Kōan (6), Keikō (12), Seimu (13), Chūai (14), Jingū (15), Jomei (34), and Kōgyoku~Saimei (35). Later, after the death of Tenmu, the appellation yamato neko, a title connected to the importance of the palace in Yamato, was pulled out of legendary stories and attached to the names of a number of rulers. It appears in the names of the following rulers: Kōrei (7), Kōgen (8), Kaika (10), Genmei (43), and Genshō (44).13 Table 4.3 illustrates how

From kings to heavenly sovereigns

103

Table 4.3 Yamato Neko versus Tarashi Hiko # 6 7 8 10 12 13 14 15 34 35 43 44

Yamato Neko Kōrei Kōgen Kaika

Genmei Genshō

Tarashi Hiko Kōan

Keikō Seimu Chūai Jingū Jomei Kōgyoku~Saimei

these two titles appear if arranged in chronological order based on our records. One can thus argue that two competing factions were at work here in the Asuka and early Nara eras: Jomei ~ Kōgyoku versus Genmei ~ Genshō. Analyzing the data in Table 4.3, one can conclude that different dynasties are represented by these appellations, manipulated by the historiographers through coding these terms into the names of the rulers. It is as if two different groups at court wanted the fame of their ancestors represented by having the names of these royal ancestors glorified with dynasties demarcated by their later names. The appearance of Kōan results in the founding of the Tarashi Hiko dynasty, which does not actually appear until the reign of Keikō through Jingū, only to appear later with the coming of Jomei (the actual progression is Jomei back into time). This dynasty takes credit for subjugating the Kumaso and the barbarians in the east. Chūai conquers the peninsula, bringing Silla into subjugation, creating a satellite ally in Paekche. This mangled legendary story is later rewritten with Jingū as the heroine. Jitō’s family, which included the blood of the violent Soga, was represented by the Yamato Neko name. The Bureau of History after Jitō wanted the historical record to show that the Jitō-Genmei-Genshō dynasty was founded by Kōrei and continued until Kaika. Genmei was a younger sister of Jitō and granddaughter of Kōgyoku. This family lays claim to control of the dynasty, preserving a blood connection to Keitai, the founder of the current dynasty. Next consider Table 4.4, where I line up the first 31 rulers, from Jinmu to Keitai (with the addition of Yamato Takeru, Kinashi Karu, and Princess Ihidoyo), categorizing their names according to one of four possible etymological criteria: Tarashi Hiko, Yamato Neko, names based on toponyms, and actual (birth?) names. Of these 31 names, 5 belong to the Tarashi Hiko group, and 4 to Yamato Neko, while 12 are toponyms and 8 use the real name. The striking anomaly is Buretsu, whose recorded name is a fusion of Nintoku’s real name and the toponym in Yūryaku’s name, providing more proof that he is a fabrication.

104

From kings to heavenly sovereigns

Table 4.4 Comparison of rulers’ names Ruler Jinmu Suizei Annei Itoku Kōshō Kōan Kōrei Kōgen Kaika Sujin Suinin Keikō Yamato Takeru Seimu Chūai Jingū Ōjin Nintoku Richū Hanzei Ingyō Kinashi Karu Ankō Yūryaku Seinei Ihidoyo Kenzō Ninken Buretsu Keitai

Tarashi Hiko

○

Yamato Neko

Toponym

Actual

Ihare Nunakawa Shiki Yamato Mima ○ ○ ○

○

Mimaki Ikume Yamato

○ ○ ○

Homuda

Asatsuma Karu Anaho ○

Hatsuse

Sazaki Izaho Mizuha

Waka Takeru Ihidoyo Woke Oke (Ōke) Sazaki Wohodo

Based on these data, I reach the following tentative conclusions: there was a group of legendary rulers, named after their base of operations, i.e., the toponym that represented their sphere of influence. During Yūryaku’s reign when his court initially attempted to compile a written history, I theorize that oral history made mention of two types of kings: (a) those connected to Yūryaku’s lineage and (b) those connected to a group with whom Yūryaku’s family had strenuous relations, likely because this group had been grafted into the ruling family through marriage (cf. Barnes 2007:159–60). I believe the oral record of the rulers of two earlier polities, Saho and Miwa (cf. Barnes 2007:153), was part of the Yūryaku tradition. The Tarashi Hiko and Yamato Neko groups were inserted at a much later stage of the historiographical process. My theory here is that an original oral history of two camps, Yūryaku’s family, and another family that had shared governance at certain times but was sidelined after Yūryaku came to power, was sausaged into a

From kings to heavenly sovereigns Y°ryaku Dynasty

105

Other Dynasty

Homuta

Sazaki

Izaho

Asatsuma

Mizuha

Kinashi

Waka Takeru

Ihidoyo

Figure 4.1 Two dynasties.

linear genealogy by later historiographers. Figure 4.1 illustrates how the original may have looked. Another interesting wrinkle in the historical narrative is the existence of two rulers who have the same title, “the first heavenly sovereign to reign.” According to the account in Nihon shoki, Jinmu receives the title of 始馭天下之天皇, while Sujin receives the title 御肇國天皇, both glossed as fatu kuni sirasu “the first to rule the land.” Ienaga et al. (1986, 1:581) argue that the title originally belonged to Sujin, who was the founder of the Yamato dynasty, but the historiographers created a variation of this title to elevate Jinmu and glorify his initial work of establishing the Yamato sphere. Historians then changed this title so that it reflected the realm as a nation and not just certain regions. Hence, Sujin is the first to rule over a nation-state, like Yamato and its satellites (國), but Jinmu is the first to rule the empire, or all under heaven (天下). Takioto (2005:87) argues that neither ruler as currently portrayed is worthy of the title of the founder of a state, but by splicing together the acts and events of these two rulers, one gets “a perfect ruler.” Thus, he concludes that later historians split a legendary, traditional founder of Yamato into two personae, Jinmu, the bridge between the divine and the human, and Sujin, the actual, foundational ruler. There is an interesting edict found in the eighth month of the fifth year of Jitō (691), “The court issued an edict to the heads of 18 families … to submit records of their ancestors’ burials.” There is some issue with the record, whether the archetype is 墓記 “burial records” or 纂記 “collected record of events.” Ienaga et al. (1986, 2:510, n19) argue that the archetype is the former. Of the 18 groups mentioned, two are worthy of note: Isonokami (originally the Mononobe) and Ōtomo. These two families were at the height of their military power during the Keitai-Kinmei dynasty, and their influence later weakened under the expansive battering of the power-hungry Soga. Most scholars have argued that the records that formed the basis for later historiography originated sometime between Keitai and Kinmei,14 so it is rather certain that the legends and stories of these military families would have been included in these early attempts at historiography. Thus, having Jitō order these groups to submit their family records presents us with several possible interpretations. One interpretation is to see the Jitō court censuring events or dictating truth. Another possibility has the court revisiting Teiki and

106

From kings to heavenly sovereigns

Kuji, as these two records may have carried a tradition that these two powerful military families influenced the creation of the early history. One area people tend to ignore or downplay is the influence of the peninsula on the early historiographers. It is clear that the compilers of Nihon shoki had access to several different Paekche histories, as these are quoted by title in the record. Careful textual work suggests that these three records were much like Chinese official histories in that they recorded the reigns of dynasties in a chronological framework. The first record mentioned in Shoki is Paekcheki “Record of Paekche,” and this work contained, at minimum, the records of kings Ahwa (Asin), Chikchi (Chŏnji), and Kaero.15 It is plausible this record ended with the reign of king Kaero, who was killed by Koguryŏ forces when the capital Hansŏng was overthrown. The next record mentioned is Paekche sinch’an “New Compilation of Paekche,” and this record includes the reigns of Kaero to Muryŏng (r. 501–523). I assume that this record began when the Paekche court moved their capital to Unjing after the destruction of Hansŏng. The inclusion of “new” in the title may refer to the establishment of a new capital, and thus a new history. This record began with the downfall of Kaero to provide historical context. The final record quoted is Paekche pon’gi “Paekche Basic Annals,” which contained at least the reigns of Muryŏng, Sŏng (r. 523–554), and Widŏk (r. 554– 598). All three Paekche records are constructed around a chronological framework using the sexagenary calendar, but each record demonstrates increasing sophistication in its calendar. As an example, the oldest of the three, Paekcheki, only quotes the year and season in question. Paekche pon’gi, however, quotes the year, the month, and the day. This rise in sophistication possibly points to one of three situations: (i) the older records were a listing of events by year and as the skill of record keeping evolved, more accurate dates were recorded; (ii) the older records contain reconstructions of orally preserved events that only included years and seasons; (iii) the most likely scenario is that we have a mixture of the (i) and (ii). One important conclusion that stands out is that as discussed in Chapter 2, the early elite of Yamato were exposed to Paekche writing and techniques, so it is reasonable to expect that Paekche historiographical techniques were also employed as a model, based on the basic elements of the sexagenary calendar system. Below are two examples from these Paekche records, illustrating the basic format. First, from Paekcheki: 百濟記云、壬午年、新羅不奉貴國。貴國遣沙至比跪令討之。新羅人 莊飾美女二人、迎誘於津。沙至比跪、受其美女、反伐加羅國。加羅 國王己本旱岐、及兒百久至・阿首至・國沙利・伊羅麻酒・爾汶至等 、將其人民、來奔百濟。百濟厚遇之。加羅國王妹既殿至、向大倭啓 云、天皇遣沙至比跪、以討新羅。而納新羅美女、捨而不討。反滅我 國。兄弟人民、皆爲流沈。不任憂思。故、以來啓。天皇大怒、即遣 木羅斤資、領兵衆來集加羅、復其社稷。

From kings to heavenly sovereigns

107

Paekcheki says: In a renwu year (382 CE) Silla did not present tribute to the honorable country (貴国). The honorable country then dispatched Sachi Hiko to punish (Silla). Silla dressed up two women beautifully and sent them to the port to greet Sachi Hiko. Sachi Hiko received the two women, and instead (of attacking Silla) attacked Kara. The king of Kara, Kopo Kanki, took his sons, Pak-kuti, Asuti, Kok-sari, Iramasu, Ni-monti and others and led their people and fled to Paekche. Paekche treated them warmly. The younger sister of the king of Kara, Ke-denti, came over to Great Yamato (大倭) and declared, “Great Ruler, you dispatched Sachi Hiko to punish Silla, but Sachi Hiko has married beautiful women from Silla and halted his attack, and has instead destroyed my country. My brothers and people are all wanderers now. I cannot endure the thoughts of anxiety. And that is why I have come here to make this report.” The Great Ruler was incensed and immediately dispatched Mokra Konsi, who went at the head of troops and assembled in Kara and restored the altars of soil and grain (quoted in the 62nd year of Jingū). I have underlined phrases lifted from Chinese works, underscoring the fact that the harsh criticism leveled against Nihon shoki for creating some of its text by pilfering phrases from Chinese classics actually was a “skill” learned from the native historiographers’ Paekche teachers and their peninsular source documents. The second example comes from Paekche pon’gi: “In the winter, tenth month, Nasol Tokmon, Nasol Kema and others returned from Yamato (日本) and remarked that there was no report (from the Yamato Court) about Kafuchi Atahi, Enasi, or Mato.” (fifth year, tenth month of Kinmei) [百濟本記云、冬十月、奈 率得文・奈率奇麻等、還自日本曰、所奏河内直・移那斯・麻都等事、無 報勅也。] It is important to note that the years and months in the Paekche record match those in Nihon shoki, demonstrating that the Yamato historiographers aligned their native calendar with that of Paekche and its records. The Paekche calendar appears to have been a gold standard for the early Yamato historiographers to anchor their native record.16 This is nowhere more visible than at the death of Keitai. Nihon shoki records: In the second month of the 25th year of his reign, the great ruler’s illness became severe, and on the seventh day he passed away in the Iware Tamaho Palace. He was 82 years at the time. In the winter, twelfth month, he was interred in the Aino Tomb. Another manuscript says that the ruler passed away in a kinoe tora year (534), the 28th year of his reign. The reason that we record here that he passed away in a kanoto i year (531), the 25th of his reign, is because we have had reference to the text in Paekche pon’gi. That text says, “In the third month of a xin-hai year (531), the army of Paekche advanced and reached the land of Ara and constructed the Kottok Fortress.

108

From kings to heavenly sovereigns During this month the king of Koguryŏ, King An, was assassinated. We have also heard reports that in Yamato (日本) the Great Ruler and his heir apparent both passed away.” According to this account, the year kanoto i represents his 25th year. People who investigate this may later figure out which date is correct [廿五年春二月、天皇病甚。〇丁未、天皇崩于磐余玉穗宮。 時年八十二。〇冬十二月丙申朔庚子、葬于藍野陵。或本云、天皇、 廿八年歳次甲寅崩。而此云廿五年歳次辛亥崩者、取百濟本記爲文。 其文云、太歳辛亥三月、軍進至于安羅、營乞乇城。是月、高麗弑其 王安。又聞、日本天皇及太子皇子、倶崩薨。由此而言、辛亥之歳、 當廿五年矣。後勘校者、知之也。].

This passage illustrates the prestige that the Paekche histories and the attending calendar enjoyed at the Yamato court. One reason for that prestige is the simple fact that the histories and the technology of the historiography of Paekche predated those in Yamato. These data also allow us to reconstruct the template that the early historiographers in Yamato had at their disposal. The Yamato historiographers in the beginning may have found the Chinese histories too complex, too sophisticated, or too unconnected to be of much use as a template. It is more reasonable to see the heavy hand of peninsular historians working in the Bureau of History, relying on records from their native land. Add to that the fact that the Paekche records contained detailed accounts of events in Yamato, events the historians at court could corroborate. It is helpful to take a slight detour and look at the founding story of Paekche. While “the Annals of Paekche” in Samguk sagi preserve three different strains of the foundation story of Paekche, the story contained in Book 81 of Suishu is the most reliable. Rather than quote from Samguk sagi’s abbreviated version, I quote a representative section from Suishu. The ancestors of Paekche came from the land of Koguryŏ. There was a maid who served the king of Koguryŏ. One day, unexpectedly, she became pregnant. When the king tried to kill her, she said, “There was an object shaped like the egg of a chicken. It came toward (and touched me), and that sensation caused me to become pregnant.” Hearing this, the king let her live. Later she gave birth to a son. The king dumped the baby in the toilet but for a long time the baby would not die. With this the king thought that the child had a strange, divine power, so he commanded that the child be raised. The child was called Toŋ-mɨaŋ (Tongmyŏng). As Toŋ- mɨaŋ grew, the king grew envious and hated him. Fearing that he would be killed, Toŋ-mɨaŋ fled from the presence of the king and came to the Ɂɨam River (淹水). The people of Puɑ-jɑ served Toŋ-mɨaŋ. Kiu-tai (K: Kut’ae仇台) was a descendant of Toŋ-mɨaŋ, and he was sincere in his benevolence and trustworthiness, and he first established his kingdom in territory under the control of the Daifang Commandery. Kung-sun Tu, Governor of the Han Commandery at Liaodong, gave his daughter in marriage to Kiu-tai. At last the kingdom prospered and became a powerful kingdom among (the states of) the Eastern Barbarians.

From kings to heavenly sovereigns

109

In the beginning, it was said that one hundred households crossed the sea (to get to where they settled), so they called the land Paekche “one hundred crossings.” The royal house has had ten or twenty reigns, and a representative of the Chinese court recorded the previous history in detail. In the beginning of the kai-huang era (581–600) the Puɑ-jɑ King Chang sent an envoy with tribute to our court, and we bestowed on Chang the title of Senior Palatine,17 Duke of the Daifang Commandery, King of Paekche [百濟之先，出自高麗國。其 國王有一侍婢，忽懷孕，王欲殺之，婢云：「有物狀如雞子，來感於 我，故有娠也。」王舍之。後遂生一男，棄之廁溷，久而不死，以爲 神 ，命 養 之 ，名 曰 東 明 。及 長 ，高 麗 王 忌 之 ，東 明 懼 ， 逃 至 淹 水 ，夫 餘 人 共 奉 之 。東 明 之 後 ，有 仇 台 者 ，篤 於 仁 信 ，始 立 其 國 于 帶 方 故 地 。漢 遼 東 太 守 公 孫 度 以 女 妻 之 ，漸 以 昌 盛 ，爲 東 夷 強 國 。初 以 百 家 濟 海 ，因 號 百 濟 。曆 十 餘 代 ，代 臣 中 國 ，前 史 載 之 詳 矣 。開 皇 初 ， 其王餘昌遣使貢方物，拜昌爲上開府、帶方郡公、百濟王。]. The territorial domain of the kingdom is 450 li from east to west, and over 900 li from north to south. Its southern boundary borders Silla, and its northern boundary prevents incursions by Koguryŏ. The capital is called Wi-bat Fortress (K: Kŏbal-sŏng 居抜城). … The position of Chief Clerk is rotated once every three years. The Capital Provinces have five circuits, each circuit has five highways, and the court officials have residences along these highways. Each of the five provinces has one Intendant, and two assistants. Each province is divided into ten districts, and each district has a District Governor. The residents (of these districts) are a mixture of people from Silla, Koryŏ, and Wa, and there are also some Chinese there. … In the middle month of each season, the king worships the deities of heaven and the Five Emperors. The people have set up an altar to the founding king, Kiu-tai, in the capital city and worship him four times a year [其國東西四 百五十里，南北九百余里，南接新羅，北拒高麗。其都曰居拔城 … 長史三年一交代。畿內爲五部，部有五巷，士人倨焉。五方各有方領 一人，方佐貳之。方有十郡，郡有將。其人雜有新羅、高麗、倭等， 亦有中國人…每以四仲之月，王祭天及五帝之神。立其始祖仇台廟于 國城，歲四祠之。國西南人島居者十五所，皆有城邑。]. As observed earlier, it is noteworthy that the miraculous beginnings of the rulers of Koguryŏ and Paekche are based on a related story found in the first-century Chinese work, Lunheng. Best (2006:15) has argued that while Paekche had several different, though related, strains of their foundation myths, it is the third, with Kut’ae as the king, that Suishu notes is the founding king. This version appears to be closer to the truth, though all three versions have a shared commonality regarding immigrant origins: “the monarchs of Paekche were not indigenous to the area of southwestern Korea where they ruled, but rather they descended directly or indirectly from Fu-yü” (ibid 2006:20). It is impossible to know what the earlier Paekche histories contained regarding the miraculous beginnings, but it seems safe to conclude that Paekche’s royal

110

From kings to heavenly sovereigns

Table 4.5 Expansion in Paekche Royal lineage SSR #

Family name

Ancestor

Generations

779 780 781 790 857 871 878 887 890 976 1029 1034 1035 1039

Yamato Ason Kudara Ason Kudara Kimi Ōoka Miyatsuko Sugano Ason Hirotsu Muraji Memu Uji Asukabe Miyatsuko Mano Miyatuko Hiroi Muraji Kawachi Muraji Okahara Muraji Hayashi Muraji Asukabe Miyatsuko

King Muryŏng Prince Hye Prince Mun Yen Konan Yencho King Kwisu King Kŭn’gusu King Piryu King Piyu King Ch’ogo King Piryu Prince Umt’ae Kwisu Chichong King Chikchi King Malta

18 from King Tomo 30 from King Tomo 24 from King Tomo 12 from King Sokko 10 from King Tomo NA NA NA NA NA Son of King Tomo Son of King Chinso NA NA

line claimed a divine origin through the connection with the foundational myth of Puyŏ (Best’s Fu-yü). The early Japanese historiographers could relate to this story, as there were various stories within their own traditions that connected the ruling family to a deity. To mere mortals, the birth of the race via deity was also considered miraculous or mysterious. Another historiographical strategy the Japanese adapted from the Paekche model is the creation of a history of royal succession from father to eldest son (cf. Best 2006:108). This is the pattern in the beginning of the records but becomes messy when historical records come into play. A third critical technique Yamato learned was extending the ages of the rulers by antedating events. It is well known that the reigns of the early rulers in the Korean annals, Samguk sagi, are also exaggerated. Best (2006:17–20) argues that the exaggerated length of royal succession is not a product of Samguk sagi but appears to have existed centuries earlier. I provide data from Shinsen shōjiroku (SSR) to support Best’s assertion; the following data allow a small window for scholars to peer into Paekche traditions preserved in eighth-century Japan. Table 4.5 lists the Paekche immigrant family, the purported ancestor, and the number of generations between this ancestor and a specific Paekche king. To add one more link to this chain of expanded Paekche royal longevity, Shoku Nihongi records in the seventh month of the ninth year of Enryaku (790) the death of Tu Sukune Mamiti, who was of Paekche lineage. Shoku Nihongi records: The (record of the) lineage of (Tu Sukune) Mamiti and others notes they descend from King Kwisu of Paekche. Counting from the founding of the kingdom of Paekche, Kwisu is the sixteenth ruler. In the beginning, Great King Tomo, the founder of Paekche, established a country in the region of Puyŏ that the spirit sent down by the (heavenly) deity caused him to govern.

From kings to heavenly sovereigns

111

The heavenly sovereign bestowed on him a book of prophecy, giving him authority to rule over the various lands of Kara, and called him king. After that, we come to the reign of King Kun’sukwo (Kŭnch’ogo), and he had affection for the distant court of the emperor (of Yamato), and for the first time established diplomatic relations with our honorable country. … Later, in the reign of Emperor Ōjin, who ruled from the Karushima Toyo Akira Palace, the emperor dispatched Arata Wake, the distant founder of the Kamitsukeno, and dispatched him to Paekche to request the dispatch of scholars. The ruler of Paekche, King Kwisu, showed affection and diligently followed the request, and from among his relatives selected a grandson, Prince Sinswon, and sent him to the Yamato court with the envoy. … During the reign of Nintoku, who ruled from the Naniwa Takatsu Palace, Prince Ara, the oldest son of Prince Sinswon, was made a close retainer. The son of Sinswon is Lord Hayyang, and his son is Lord Woceng. Lord Woceng had three sons, the oldest being Misan, the second Sin’i, and the youngest Maro. (Aoki et al. 1989, 5:469–71) To illustrate this tradition of early Paekche royal succession, Table 4.6 outlines the kings found in Samguk sagi (SS) and their succession versus what we know from early Japanese sources. There is agreement that the son of King Tomo—traditionally known as Onjo— is the founder of Paekche. Based on two different but related representations, his original name is reconstructed as 温祚 *Ɂuən-dza versus 陰太 *Ɂɨm-thɑs. Combining these two pieces of information, I reconstruct the Paekche name as *Ɂəm-ta. This disparity in his name suggests that he was a legendary figure, and different eras recorded his name in different but related ways. Another interesting disparity concerns the length of the royal lineage. The ancient Japanese tradition says that Kwisu was the tenth in line from Tomo, but Samguk sagi claims he was 14 generations removed, an addition of four generations. Notice, however, that Shoku Nihongi claims he is 16 generations from King Tomo. Muryŏng is a 25th descendant of Tomo, but the record in Shinsen Table 4.6 Comparison of Paekche royal succession King

SS claim

Name in Japan

Japanese source claim

Onjo Ch’ogo Piryu Kŭn’soko Kŭn’gusu Chinsa Chikchi Piyu Tongsŏng Muryŏng

Son of Tomo, 1st king 5th king from Tomo 11th king 13th king 14th king 16th king 18th king 20th king 24th king 25th king

Umt’ae Ch’ogo Pilyu Sokko Kwisu Chinsa Chikchi Piyu Malta Muryŏng

Son of King Tomo None None None 10th from King Tomo None None None None 18th from King Tomo

112

From kings to heavenly sovereigns

shōjiroku only claims 18 generations, a subtraction of seven generations. The first Paekche king mentioned in Nihon shoki is Soko (written 肖古), the 13th king according to tradition. It is believed that the king mentioned in Suishu, 余句 “Ko of Puyŏ,” is the same king, especially since the character 句 was read kou at the end of the Later Han era (and 古 was read ko). If we can trust this, then it appears that Paekche was founded perhaps two generations earlier than Soko, resulting in the fabrication of roughly nine or ten generations of earlier kings. Thus, while Samguk sagi appears to contain greater expansion than what is hinted at in Japanese sources, evidence from Shoku Nihongi demonstrates that expansion of the Paekche royal line was underway earlier than the Nara era. As Best (2006:432–33) argues, some scholars believe that Onjo’s brother, Putlyu (K. Piryu 沸流), is really a mirror image of the later 11th king Pilyu. He also notes that because of the similarity of the names of the Kings Soko (K: Ch’ogo, 5th king) versus Kin-soko (K: Kŭnch’ogo, 13th king), and Kusu (6th king) versus Kin-kusu (K. Kŭn’gusu, 14th king), one of these is an addition based on the other (2006:428–29). It is important to notice that both Kŭnch’ogo and Kŭn’gusu appear in Nihon shoki without the later-added appellation of kŭn 近 “closer to the present” (something akin to the English addition of “junior”). It is possible that this manipulation occurred in later versions of Paekche’s histories, and that while these individuals are not included in Nihon shoki (perhaps because the Japanese historiographers did not have these copies with them when putting earlier versions of Yamato’s history together), it seems reasonable that the court would have known about this manipulation. This manipulative trick reminds us of the creation of Buretsu, whose name is Waka Sazaki “younger Sazaki,” and who was created to punctuate the end of the genealogy of Nintoku (whose name is Sazaki); this is justified in the records by stating that Buretsu had no posterity. Is it too difficult to believe that having Buretsu fabricated based on Nintoku’s name was a strategy learned from a related type of Paekche manipulation? It is reasonable to expect that as time passed, the Yamato court sensed the decay of the high orbit of Paekche.18 Sometime, in the beginning of the seventh century, the Yamato court adjusted its sights, deciding to send people to China to learn directly from that advanced civilization. Understanding that the Chinese viewed their ruler as “a son of Heaven,” and based on their own traditions, it is possible that the Yamato court began to shift (or modernize) the title of the king, where the now outdated title of opo-kimi “great lord” needed to be either upgraded or discarded and a new title created. I argue that during the dynasty of the husband and wife team of Jomei (r. 629–64) and Kōgyoku/Saimei (r. 642–645/655–661), the government turned toward a more nativist trend (something we could label “Shintō”) and away from Buddhism and discarded the title of “great lord.” I theorize that they took a word from the vocabulary of the native religion. A number of ancient liturgies include the word sume-mima “the divine grandson,” a title reserved for Ninigi, or whomever the legend portrayed as the divine entity that descended to the archipelago and gave birth to the Japanese populace. After the royal family had grafted their lineage into that of Ninigi’s, the door blew open to appropriate the title for themselves, adding the suffix –rama, a suffix found mainly

From kings to heavenly sovereigns

113

in the edicts, to the word sume. I have earlier argued that –rama is a “focus suffix,” drawing attention to the noun to which it is affixed (Bentley 2001b:120–21). Over time, this was shortened to –ra and resulted in the innovative title of sumera. This allowed the Yamato court to face a new direction, where they could claim to be the divine descendants of Ninigi. The reason this did not occur until Tenmu and Jitō’s era was because of the invasion and destruction of Paekche in 660. In conclusion, Paekche was not only a teacher of writing and record keeping for the elite in Yamato, but also Paekche histories were a touchstone for the very history that Yamato compiled. The historians, many of whom must have been from the peninsula, early on learned a variety of historiographical techniques from their peninsular teachers. While Yamato had a crude chronology of early events, these events were ultimately lined up and corroborated according to the royal chronology of the kingdom of Paekche. In essence, early Japanese historiography was a twin son of different mothers,19 Paekche and Yamato.

Notes 1 Hanyuan (660) quotes this line, apparently from Suishu, 王長子号 和哥彌多弗利。華言太子, “The king’s oldest son is called *Wa-ka-mji-ta-put-ri. This is a flowery title for the crown prince.” Here we clearly have 和 and not 利. 2 Kawamoto (2004:68) consistently calls the ruler of Japan 倭王 “king of Wa,” thus skirting the issue. When quoting from the Suishu record where it says, “The king of Wa is the younger brother of heaven,” Enomoto writes, “Shouldn’t it be ‘the younger sister?’” Enomoto (2008) claims the account was based on fragmentary information and cannot be trusted. 3 Motoori Norinaga is the first to declare the Sui account to be untrue (in Kyojū gaigen, Ōkubo and Ōno 1976, 8:41-43). Sakamoto Tarō (1979:59–61) believed that this account may have recorded an unofficial mission, or the entire account is fabricated. What is lacking is any credible reason the Sui court would fabricate this account, especially considering the linguistic fragments match what we know about the history of early Old Japanese. 4 Yoshida (1997:87) maintains that rather than blaming the Sui, it makes more sense to see the Nihon shoki compilers as intentionally ignoring this rather unflattering account. What scholars tend to overlook, however, is that there has been no good explanation for why the compilers ignored the account of an entire system of rule that fails to match that in Nihon shoki. 5 I have modified my earlier translation, as I had inadvertently ignored the parallelism in this passage, with 君后 “the sovereign and the queen” versus 孝妣 “one’s deceased father and mother.” 6 One reason scholars doubt the veracity of this account is because, as Yoshida (1997:14245) points out, the label tennō (天皇) was not used until around the time of Tenmu, some 60 years after the 620 date in Suiko. However, it is also possible that the record originally was titled something like 倭王記 and a later historian updated the title. 7 Here “historical” is narrowly defined as an individual about which something anciently was written about based on eyewitness accounts. 8 Takabayashi (1976:127) splits the record of the rulers in Kojiki into two halves, the dividing line being Yūryaku-Kenzō. 9 Shiraishi (1999:149–51) goes further, arguing that Yūryaku envisioned himself as a great ruler over a large realm, including the kingdoms of Paekche, Silla, and the Kara Federation. 10 Compare Barnes (1988:265–67) and Piggott (1997:65).

114

From kings to heavenly sovereigns

11 Mizuno (1954:149–75) was one of the first to argue this point. Tsuda (1948, 1:261, 303–304) would counterargue that no events are recorded, because these eight reigns came from Teiki, which lacked any story. However, this lack of storyline should not necessarily mean these rulers did not exist. 12 William George Aston noticed this back in 1889, “Early Japanese Historiography.” 13 Kishimoto (2013) has argued for dual kingship at an earlier time in Yamato. It is possible that these two competing names were gleaned from this ancient system. 14 Cf. Tsuda (1948:1, 46–49). 15 The dates of these three are tenuous. Based on Samguk sagi, these three are: Asin (r. 392–405), Chŏnji (r. 405–420), and Kaero (r.455–475). 16 Mishina (1966) argued this point when he demonstrated that the three year hiatus between the death of Keitai and the ascension of Ankan was due to a reliance on the Paekche calendar. 17 I have used Best’s translation of 上開府 as “senior palatine” (2006:343), with the understanding that this is an abbreviation of the full title of 上開府儀同三司 “Senior Palatine with Dignities Equal to Those of the Three Ministers.” 18 One trend seen repeatedly within Paekche’s political circles is dissention and intrigue. This eventually leads to the destruction of the kingdom. 19 Cf. Dan Fogelberg and Tim Weisberg’s 1978 album, Twin Sons of Different Mothers.

References All works are published in Tokyo, unless otherwise noted. Aoki, Kazuo, Inaoka Kōji, Sasayama Haruo and Shirafuji Noriyuki, eds. 1989–2000. Shoku Nihongi. SNKBT, Vols. 12–16. Iwanami Shoten. Aston, William George. 1889. “Early Japanese History.” Transactions of the Asiatic Society of Japan, 1889-90: 212–248. Barnes, Gina. 1988. Protohistoric Yamato. Ann Arbor: University of Michigan Center for Japanese Studies. Barnes, Gina. 2007. State Formation in Japan: Emergence of a 4th-Century Ruling Elite. London: Routledge. Bentley, John R. 2001. A Descriptive Grammar of Early Old Japanese. Leiden: Brill. Bentley, John R. 2006. The Authenticity of Sendai kuji hongi: A New Examination of Texts, with a Translation and Commentary. Leiden: Brill. Best, Jonathan W. 2006. A History of the Early Korean kingdom of Paekche. Cambridge: Harvard University Asia Center. Enomoto, Jun’ichi. 2008. “Zuisho Wakokuden no shiryōteki seikaku ni tsuite.”Arena, 5: 186–191. Hanzawa, Eiichi. 2002. Shōtoku Taishi hōō Waōron. Contained in Vol. 24 of Nihon shoki kenkyū, Hanawa Shobō, pp. 299–378. Ienaga, Saburō, Inoue Mitsusada, Ōno Susumu and Sakamoto Tarō, eds. 1986. Nihon shoki. NKBT, Vols. 67–68. Iwanami Shoten. Inoue, Mitsusada. 1967. Nihon no rekishi: Shinwa kara rekishi e, Vol. 1. Chūō Kōronsha. Kawamoto, Yoshiaki. 2004. “Zuisho Wakokuden to Nihon shoki Suikoki no kijutsu o megutte: kentōshi oboegaki.” Shien, 141: 53–77. Kishi, Toshio. 2003. “Kyodai kofun o tsukuru.” In Ōtuka Hatsushige, ed. Shiwa Nihon no kodai, Vol. 4. Sakuhinsha, pp. 186–197. Kishimoto, Naofumi. 2013. “Dual Kingship in the Kofun Period as Seen from the Keyhole Tombs.” Urban Scope, 4: 1–21. Translated by Joseph Ryan. Kobayashi, Toshio. 1994. Kodai ōken to agata-agatanushisei no kenkyū. Yoshikawa Kōbunkan.

From kings to heavenly sovereigns

115

Kobayashi, Toshio. 2006. Nihon kodai kokka keiseishi-kō. Kōsō Shobō. Kōnoshi, Takamitsu. 1983. Kojiki no tassei: sono ronri to hōhō. Tokyo Daigaku Shuppan. Maruyama, Jirō. 1955. Nihon shoki no kenkyū. Yoshikawa Kōbunkan. Masumura, Hiroshi. 1988. Kentōshi no kenkyū. Kyōto: Dōhōsha. Mishina, Shōei. 1966. “Kinen shinkō.” Contained in Naka Michiyo and Mishina Shōei, Zōho jōsei nenki-kō. Nara: Yōtokusha, pp. 189–246. Miyake, Marc Hideo. 2003b. “Philological Evidence for *e and *o in Pre-Old Japanese.” Diachronica, 20(1): 83–137. Mizuno, Yū. 1954. Zōtei nihon kodai ōchō shiron josetsu. Omiyayama Shoten. Mizuno, Yū. 1968. Nihon kokka no seiritsu. Kodansha. Nishimiya, Kazutami. 1993. “Kojiki no kitei--seichō o chūshin toshite.” Kokugo to Kokubungaku, 4: 1–14. Ōishi, Yoshiki. 1975. Nihon ōken no seiritsu. Hanawa Shobō. Ōkubo, Tadashi and Susumu Ōno. 1976. Motoori Norinaga zenshū. Chikuma Shobō. Ōyama, Seiichi. 2009. Tenson kōrin no yume: Fujiwara Fubito nado no purojekuto. NHK Books. Piggott, Joan R. 1997. The Emergence of Japanese Kingship. Stanford: Stanford University Press. Sakamoto, Tarō. 1979. Shōtoku Taishi. Yoshikawa Kōbunkan. Shiraishi, Tarō. 1999. Kofun to Yamato seiken. Bunshun Shinsho. Takabayashi, Miyuki. 1976. Nihon shoki hennenkō. Contained in Vol. 9 of NSK. Hanawa Shobō, pp. 91–166. Takioto, Yoshiyuki. 2005. Kamigami to kodaishi no nazo o toku: Kojiki to Nihon shoki. Seishun Shuppansha. Tokumitsu, Kyūya. 1978. Teiki no hihyō-teki kenkyū. Kasama Shoin. Tsuda, Sōkichi. 1948. Nihon koten no kenkyū, 2 Vols. Iwanami Shoten. Umezawa, Isezō. 1968. Kiki hihan: Kojiki oyobi Nihon shoki no seiritsu ni kansuru kenkyū. Sōbunsha. Wang, Zhenping. 1994. “Speaking with a Forked Tongue: Diplomatic Correspondence between China and Japan, 238–604 A.D.” Journal of the American Oriental Society, 114(1): 23–32. Yoshida, Takashi. 1997. Nihon no tanjō. Iwanami Shoten.

5

The beginnings of writing a history

Neither Nihon shoki nor Kujiki had a preface, as the current preface found with the text of Kujiki is clearly a later addition (cf. Bentley 2006:21, 40–41). This means that Kojiki is the only work to contain a preface. Thus, there is very little external evidence for the historiographical process or the source documents of any of the three works. What is quite abundant is internal evidence, which some scholars have mined with varying degrees of success. Tsuda (1948, 1:47) argues that the basic material from which Kojiki was compiled originated around the time of Kinmei. Umezawa (1976:367–68) provides support for this theory, arguing that a somewhat cryptic note found in the second year of Kinmei has more importance than just its reference to the proper reading of names in a genealogical list. The second year of Kinmei records that the great king had selected five women to be royal consorts. Their names are then noted with a list of the children they bore. Within this listing of mothers and offspring appear a number of notes recording variations of the graphic rendition of each name. After two rather long notes, the compiler attaches a final editorial note: There are many ancient characters in the record Teiō hongi, and the people who abridged [this record] often changed. When later people read through and copied down the record they purposely altered (modernized?) some characters (readings). The manuscript has already passed through many hands, and confusion has crept into the texts. The order of birth has been lost, and older and younger brothers are no longer together. We have now considered and examined things from the past and present and have returned the genealogy to its correct state. When it was difficult to know the true order, we followed one text and recorded the variants in a note. All other (issues) follow this example. It is this note that Umezawa highlights, and the information we can glean from this description will serve as our departure point to postulate what may have been the framework for the creation of the first history in Yamato. Several points require clarification. First, the title Teiō hongi (帝王本紀) demonstrates that there were actual documents in existence prior to Tenmu’s day that the compilers relied on, though I contend that the original title of this work may have been something like

The beginnings of writing a history 117 Wa kimi hongi (倭王本紀), modeled after the benji (本紀) of Shiji, or the pon’gi of Paekche records. Second, the existence of this title (and likely other records) having passed through a long process of textual transmission and abridgement attests to the age of these records. Third, we should interpret 古字 “ancient characters” mentioned above from a linguistic perspective, referring to characters with older readings. This label does not point to the characters themselves being ancient, something akin to pre-Han dynasty glyphic renditions. Thus, as the older readings were lost or misunderstood, errors crept in regarding the reading of the names. I have already mentioned Teisetsu, which notes that 彌, usually transcribing mi, earlier had represented me. Scholars are quick to point out that this phraseology (“There are many ancient characters….”) was lifted from Hanshu xuli, which is true, but I agree with Kojima et al. (2002, 2:366, n4) that this wording was used because the historiographers saw the sad state of the available manuscripts in a similar light to that found described in Hanshu xuli. Having the editor of the Kinmei volume of Nihon shoki lift this phrase from a Chinese source does not invalidate the state of the manuscripts that the editor was trying to describe. Finally, the line “All other (issues) follow this example” [他皆効此] is an editorial comment that likely appeared at the beginning of an early manuscript. Notice that in the very beginning of the “Divine Age” section of Nihon shoki, when the editors attach a note about a distinction between the graphic usage of 尊 and 命, both read mikoto, they end the note with the phrase, “All examples onward follow this case” [下皆倣此]. The reader should remember that this example of下皆倣此 appears in numerous Chinese texts, while the first example is harder to find. I assert that the usage of 他皆 is a calque, based on Old Japanese poka pa mwina “all other examples.” This usage is likely a product from an age before Yamato historians had sufficiently mastered classical Chinese, which would allow them to create literary-quality Chinese. Umezawa (1968:303) argues, like Hirata Atsutane a century earlier, that this note in the second year of Kinmei functions much like “explanatory notes” and should rightfully appear at the beginning of the work. This strongly suggests that the Shoki compiler of the Kinmei record lifted this section from the beginning of an older record that was at hand as a reference. In a later work, Umezawa (1976:367–68) revisits this quote, noting that the title 帝王本紀 easily could be shortened to 帝紀 Teiki and create a connection with the preface to Kojiki and the tenth-year entry in the Tenmu record in Shoki. This evidence suggests that a history was compiled (or at least initiated) in the reign of Kinmei. A second piece of evidence supporting this claim was first brought to scholarly attention by Tsuda Sōkichi (1948, 1:80–82), who noted that the names of the rulers travel the following path: long, auspicious names ⇒ simple, basic names ⇒ long, auspicious names. The logic here is that the earlier names of the kings were simple, and later became longer, with the inclusion of auspicious elements. At some later point, historians in the Bureau of History employed a trick where they antedated or created a mirror image of names, only further back in time, in imitation to what Paekche earlier historians had done. This evidence demonstrates

118

The beginnings of writing a history

that some form of manipulation had occurred. While many scholars agree with the idea of manipulation, it is critical to take a further step to tease out what the historiographers in the bureau used as their chronological template. As I have argued, the institutional memory of the Yamato court appears to have included at least two groups of rulers: those with a written history, and those with a nebulous, legendary history. Using the structure of these names, a dividing line comes into focus. That dividing line is the rather hazy period of Ōjin~Keitai.1 The names of the rulers before Ōjin are long, packed with auspicious elements. Simple, basic names begin with Ōjin and continue until Keitai, and afterward these names revert to more sophisticated appellations. Tsuda concluded that the basic names were original, and the earlier sophisticated names were a product of a later era when historiographers embellished the names of the rulers, influenced by continental philosophies. A logical conclusion many reach is that the rulers before Ōjin are fabrications, but this judgment is hasty. We will return to this point later. For the moment, I suggest we look to the reign of Keitai as a possible catalyst for the beginning of serious historiography in Yamato. Ōishi (1975:54–55) has argued that the work of putting together a comprehensive history occurred in six stages: (i) in the reign of Kinmei, a record of the death dates of the ancient rulers was created; (ii) the compilation of a history up to the reign of Keitai; (iii) the compilation of the record of the reigns of Kinmei and Bidatsu; (iv) the compilation of a history up to the reign of Suiko; (v) the work of gathering and certifying records during the reign of Tenmu; (vi) the compilation of Kojiki and Nihon shoki. Ōishi argues that Kojiki contains the death dates of 15 rulers, because it is a compilation of vestiges of one of the oldest genealogies from ancient Yamato (1975:55). I find many of the basics of Ōishi’s theory well formulated but prefer to simplify the argument and push certain stages further back in time. As a working hypothesis, I postulate that historiography in Japan went through three “great stages.” The first is the era of inception in the Keitai~Kinmei era. The second occurred during what we call the Suiko era. The third and final stage began with Tenmu and concluded with the presentation of Nihon shoki in 720 CE.

History under Keitai and Kinmei As argued above, it thus appears that one early historiographical project started with the reign of Great King Kinmei. I here argue that the court of Kinmei put the finishing touches on a historiographical project started under the auspices of Keitai. This era appears to have been a type of golden age in Japanese history. Carefully sifting through data preserved in Nihon shoki, I put forth the following historical scenario as a working hypothesis. The court of Asatsuma (Ingyō, ca. 450 CE) had systematized the use of peninsular (Kara and Silla) scribes at court, many of whom could trace their lineage back to Chinese families connected to the commanderies. Park (2018:148–50) demonstrates through excavated artifacts such as gilded helmets, armor, and dragon-pattern gilt-bronze saddle frames found in both royal Silla tombs and large

The beginnings of writing a history 119 burial mounds in Ōsaka that during the time that correlates with the Ingyō dynasty Yamato had close diplomatic ties to Silla. Thus, we can conclude that Silla not only sent prized golden items but also artisans and scribes. These literate individuals laid the groundwork for later developments in literacy at court. At this time, these scribes were employed in creating documents for foreign entities, such as the Chinese court, and the kingdoms on the peninsula. I believe that the powerful families connected to the royal court also used these scribes to record lineage and events of earlier ancestors. Along these lines, Barnes (2012:85) wonders if powerful “court clans … weren’t more powerful at court than the nominal kings.” If this is true, I believe that royal dynastic lineage and family history was not yet committed to writing. Instead there were competing genealogies and family histories. I also believe that the court employed foreign scribes to create or imitate inscriptions found in China to cast domestically made mirrors. These mirrors, with their attending inscriptions, were then bequeathed to chieftains throughout the realm. These gifts became symbols of authority. Here the mirror was a type of investiture, and the writing on the mirrors could be used as a sign of royal authority; in return, by accepting the mirror, the recipient demonstrated his accountability to the court. This practice was done in imitation of mirrors granted to vassal states by the Chinese court (cf. Wang 2014:183). As Lurie (2011:62) notes, we are asking the wrong question when we wonder if the artisan who made these mirrors was literate, or if these mirrors hinted at literacy at the court. At this stage of the development of Wa society, literacy was not the issue. It was the power of the written word, as vaguely understood by those in authority—the representation of Chinese culture and influence as endowed in the script—that was the driving force for the imitation of Chinese mirrors and their inscriptions. Here it is important to see the interaction and influence of the polities on the peninsula in shaping these developments in Wa. Woo (2018:188–208) has argued, based on archaeological evidence, that states in the Kaya (my Kara) Federation and territories southwest of the Paekche capital were intersections for peninsular culture and technology with Wa. Artifacts have been excavated from tombs on both the peninsula and the archipelago, demonstrating a two-way flow of technology and culture. Six keyhole-shaped tombs have been identified around the modern Yŏngsan River basin, with another seven located to the south and west of the basin. Woo (2018:202–03) concludes that these tombs were built for Wa descendants recognized as Paekche officials. This mirrors the situation in Wa, where Yamato recognized local chieftains with title and symbols, allowing these people to be buried in large tumuli. This porous sphere of culture and technology allowed cross-pollination of the southern part of the peninsula, Kyūshū, and eastward up the archipelago. I theorize that the court of King Waka Takeru (Yūryaku), a faint depiction of which is preserved in Songshu, ushered in a progressive and enlightened era, a time of peace and prosperity. Iwasaki (1990:151–58) has demonstrated that the spread of giant keyhole tumuli was not due to military might but a desire by the various chieftains to form an alliance with the cultured Yamato king. By the end of the fourth century or early fifth century, this servitude relationship was firmly

120

The beginnings of writing a history

in place. I follow Kobayashi (2006:76–77), who criticizes those who see Yamato as an alliance of chieftains, or a federation of smaller units. He says, “It is unclear with whom they allied, or with whom they were hostile.” He would rather see this complex relationship with Yamato and its periphery as “servitude built around the personality of the great king.” This belief is strengthened by archaeological data that demonstrate there was a generous flow of culture and technology in the Wa sphere. With the possibility of warfare minimized, the resulting period of peace allowed greater development of the power structure in the Yamato Basin. Built around the charisma of this great leader, the court instituted changes and expanded the role of foreign scribes. It is during this time that the court began to employ foreign scribes to compile a record of genealogy and history surrounding the ruling family. Kobayashi (2006:52) sees a change in the structure of royal authority during the reign of Yūryaku. He envisions an internal change from an older, dual system of “king-hime” to a newer, one-dimensional system of “great king-queen.” The older system was based on magical power, where the hime acted as a medium between the spirits and the king, the temporal administrator. Through a sacred marriage of the king and the medium, the king was able to arrogate the magical power of the woman and rule the realm. The newer system, however, put the king and the queen on equal footing in ruling the realm. Kobayashi also sees a shift from a feudal system to a court-organized be system. It is highly likely that the be system had been introduced by Paekche (or Kara) at an earlier date, but evidence suggests that during the reign of Waka Takeru, this system was reinforced. The court of Great King Waka Takeru employed scribes from Paekche, Kara, and Koguryŏ tasked with writing and recording diplomatic materials exchanged between his court and the kingdoms on the peninsula as well as the court in China. As missives and documents slowly accrued at court, it became clear that preserving these documents as a hedge against later treason or disloyalty was not the only value these had. Preserving these documents enlarged the institutional memory of the court, thus providing a timeless account of the king’s achievements and glory. Instead of just visualizing one king’s reign symbolically encapsulated in the image of a huge burial tumulus, I envision that Great King Waka Takeru perceived the value and power in the written word as a vehicle to cement his legacy for coming generations. While I do not believe the great king was literate, I do believe he was enlightened to see that a record of his achievements would allow him and his reign to become immortal through a written account, a triumph vastly exceeding the capacity and memory of the kataribe, who memorialized previous reigns in vignette-like performances. At some point, the great king invited the kataribe, “living repositories of the oral traditions,” of the central land and the various satellite territories to appear at court and recite their oral traditions so that scribes could write these down and create a record to be archived. No doubt, the kataribe may have found this trend profane or even dangerous, as it clearly would have been a threat to their purpose of existence. It seems fair to claim that the kataribe would not have recited everything they had, likely keeping back the most sacred traditions of their territories.

The beginnings of writing a history 121 At any rate, over time Great King Waka Takeru was able to gather a portion of the oral traditions to be committed to writing. With this, he instituted the establishment of a repository of these records, what later would become what I provisionally call the Bureau of History. As an example, it is plausible that Great King Waka Takeru wanted to codify the stories he had heard from his father about his grandfather, known as Yamato Takeru, from whom he was named Waka Takeru, “the younger Takeru.” The missive King Bu (Muɑ) sent to the Song court makes brief mention of these stories: From ancient times my fathers have put on their own helmets and armor and have trodden across the mountains, forded rivers without time to rest. In the east we have subjugated 55 territories of hairy people; in the west we have subjugated 66 territories of barbaric masses, and across the northern sea we have subjugated 95 territories of people. Kobayashi (2006:47–48) notes that there are shadows of three traditions embedded here: “in the east” alludes to Yamato Takeru’s subjugation of Azuma; “in the west” is a reference to Yamato Takeru (later rewritten as Keikō) subjugating the Kumaso; finally, “across the northern sea” suggests the story of Chūai’s military adventures on the peninsula (later rewritten as Jingū). It is plausible that the kataribe grouped several heroic accounts under the umbrella of Yamato Takeru, and then later the Bureau of History disassociated these to fill out the record. The numbering of the territories taken in these conquests is nothing more than a mathematical monument to demonstrate the education of the ruler, as 55, 66, and 95 all have the greatest common related factor, 1.2 I suggest these numbers were added (adjusted) by the peninsular scribe(s) to make the king sound more sophisticated. Having established the Bureau of History, Great King Waka Takeru needed a way to protect his future, necessitating a strong heir. From among his sons, he made Prince Shiraka his heir. There is a tendency to take the characters of his name, 白髪 “white hair” in both Kojiki and Nihon shoki, at face value, but I argue that scholars should consider that the name contains a reference to the peninsular kingdom of Silla (sira). We should also not ignore the tradition which claims that Prince Shiraka was born of a peninsular woman, known as Kara Hime (written variously as 訶良比売・韓比売), “the princess from Kara” (Kara = peninsula). This again suggests that there was a continuing bridge between Great King Waka Takeru’s court and the kingdoms on the peninsula. At the death of Waka Takeru, Prince Shiraka inherits the throne, and we later come to know him as Seinei. The records claim he never married and had no children. I find this difficult to believe. It seems more plausible that he had two children, namely, Oke and Woke. This genealogy is later decoupled in a crude attempt to portray Yūryaku as a wicked ruler. I postulate that Shiraka ruled only a short time, because a coup overthrew the dynasty, instigated by a new group from the Afumi region northwest of Waka Takeru’s capital. That new group was led by a great chieftain named Wohodo, later known as Keitai. He overthrows Shiraka and establishes a new dynasty.3 He

122

The beginnings of writing a history

establishes a harem with a rather large number of wives. Two sons are born from a woman named Princess Menoko: Prince Magari (Ankan) and Prince Takada (Senka). A different wife, Princess Tashiraka gives birth to a son, likely named simply Haruki (Kinmei).4 With Wohodo’s success in dethroning Shiraka, he declares himself the ruler of the realm. He establishes Magari as his heir. After the death of Wohodo, Magari becomes king, and establishes his brother, Takada, as heir to the throne. Haruki rebels against his half-brothers, raises an army, and cuts them down; he then places himself on the throne and reigns over the country. This scenario explains the note found at the end of Keitai’s reign, quoted above. Shoki records that Keitai passed away early in the 25th year of his reign. A note from the compiler sheds some light on the historiographical conundrum of the day: One manuscript says, “The Divine Ruler (Wohodo) divinely departed in the twenty-eighth year of his reign, Kinoe Tora (534).” The text above says the twenty-fifth year (Kanoto I) because it was based upon the text in Paekche pon’gi. The Paekche text says that in the third month of the Kanoto I year, our army advanced to Ara and camped at Koltak Castle. During this month, the people of Koguryŏ put their king, An, to death. It was also heard that the Divine Ruler divinely departed and the Heir to the throne of Yamato passed away. That is why the year of Kanoto I fits the twenty-fifth year. Readers of later generations will perhaps be able to discern which date is correct. First, this quote makes it clear that the traditional calendar of the source material was altered according to events recorded in a Paekche record. Kujiki still contains these events under the 28th year of Keitai (Bentley 2006:341). This again demonstrates that Yamato placed great faith in the chronology found in Paekche records. Having the great king and his heir to the throne pass away around the same time suggests that they were either defeated in battle or simply cut down. I suggest that after Keitai had died, Haruki moved quickly to overthrow King Magari (Ankan) and put to death his half-brother and heir, Prince Takada (Senka). Mizuno Yū (1968:150) believes that Princess Tashiraka, who records note, is a sister of Buretsu, is a fictional character inserted simply to create a connection with the lineage of the virtuous king, Nintoku. Thus, Keitai’s real queen was Princess Menoko, a fact that helps explain why Kinmei was passed over as heir to the throne, and the great king established Ankan instead. It is possible that Kinmei was not even related to Ankan and Senka, but later historiographers grafted him into the lineage, using his mother as the link between the virtuous reign of Nintoku and this new reign of Keitai. For the time being, I believe that Wohodo’s dynasty was strong enough to survive the machinations of brotherly intrigue. Thus, Haruki took Wohodo’s throne in a rather seamless coup. I postulate that Ō no Yasumaro was aware that the circumstances surrounding the era of Keitai-Ankan-Senka-Kinmei were messy, so he grafted the new

The beginnings of writing a history 123 genealogy into the old one without editorial note. In the record of Buretsu in Kojiki, Yasumaro dryly records: The Divine Ruler had already divinely departed. He had no princes (sons) who could inherit the sun throne. Thus, Prince Wohodo, a descendant of Divine Ruler Homuda (Ōjin) in the fifth generation was caused to come to the capital from Closer Afumi Province, and to marry Princess Tashiraka; he received authority to reign over everything under heaven. I propose that a history of the state was started under the auspices of Great King Wohodo, using the peninsular scribes and their knowledge and technology, coupled with the records housed in the Bureau of History established by Waka Takeru. One of the first tasks Haruki undertook once on the throne would have been to canonize the genealogy of Wohodo. It is interesting to examine the lineage of Wohodo as contained in a variant tradition recorded in the fragment of Jōgūki, quoted in Shaku Nihongi (cf. Onoda 1986:312). According to that record, we have the following six-generational lineage: Homutu Wake ⇒ Waka Nuke Futamata ⇒ Ohohodo ⇒ Wohi ⇒ Usi ⇒ Wohodo. Judging by the archaic nature of the orthography, clearly the lineage is older than that found in Kojiki or Nihon shoki. However, there are still several glaring problems with this lineage. Consider that two of these supposed names are nothing more than placeholders. In Old Japanese, Wohi (wopwi 乎非王) means “nephew.” The other supposed name, Usi (usi 汙斯王), means something akin to “esquire.” It seems to me that Wohodo’s actual father was Ohohodo “the great Hodo,” and he named his son and heir, Wohodo “the minor Hodo.”5 Thus, the first history of the state may have had Great King Wohodo’s lineage back six generations, likely starting with Wohodo ← Ohohodo ← Homutu Wake ← X3 ← X4 ← Iniwe ← Hikoyu Musumi.6 I theorize that in imitation of the Koguryŏ-Paekche mythology of the ruling family claiming a miraculous birth from a woman connected to a deity, the court tweaked Wohodo’s genealogy so that it was grafted into one of the Musubi “creative” deities.7 Interestingly, the etymology of this name appears to be piko “prince” yu “from” musumi “the binding deity”; thus, it is not a personal name as much as a single designation for someone claiming to be of divine lineage. The next important piece of information the historiographers would need to include would be the death dates and burial places of significant individuals. The recorders in the Bureau of History would have chronicled this information from previous dynasties, likely written down from the lips of the kataribe. After this, the historiographers would then record the genealogy of powerful scions at court and their descendants. Influenced by the Chinese legal system brought to Yamato by Paekche teachers, the court likely had established a law where a royal heir could not be more than five generations removed from the present. Based on this, later historiographers updated this genealogy: they noticed that Homutu Wake’s name was close to another ruler, Homuda Wake (Ōjin), and then they inserted two nondescript relatives into the genealogy and deleted the names of those previous to Homutu Wake.8

124

The beginnings of writing a history

It is possible that the historiographers put together a history of the realm in a rather transparently descriptive manner, detailing the violent overthrow of the Waka Takeru dynasty by Wohodo, backed by the powerful Mononobe composed of men from the Okinaga territory. This account may have described the turmoil in heroic terms, or the narrative may have described Great King Waka Takeru’s heir, Shiraka, as a despot (or a usurping outsider from the peninsula) in need of removal. While this is conjecture, it does seem highly probable that much of the continental ideology undergirding the later works of Kojiki, Kujiki, and Nihon shoki was not present in the initial years of the birth of Japanese historiography. In addition, it is fair to say that while peninsular scribes were the workhorses of the project, there must have been a committee of native Yamato nobility who oversaw the project and guided the work in a specific direction. In fact, the Bureau may have been overseen by Yamato sons with peninsular fathers, thus being literate, and empowered to oversee the actual work. If my theory is fairly accurate, then it stands to reason that after Haruki (Kinmei) overthrew his two half-brothers and the kingdom was pacified, his court revisited the historiographical project of the Wohodo court. It is even possible that the Kinmei court may have felt it necessary to create legitimacy after violently overthrowing King Magari by portraying several previous dynasties as existing in a kind of linear order, and then being overthrown, much like the narrative in so many overthrown dynastic kingdoms on the Korean peninsula. This theory operates on the assumption that Kinmei’s court had institutional memory of two or three previous dynasties: the Ōjin~Nintoku court, the Ingyō court, and the Ankō~Yūryaku court. My argument is partially based on the record in Songshu, as it lists the relationships of the five kings as follows: (i) when *Tsɑm (賛王) died, his younger brother, *Mie (彌王), inherited the throne; (ii) *Tʂɛi (濟王), king of Wa, sent an envoy (unclear what the relation here is with *Mie); (iii) when King of Wa *Tʂɛi died, his heir (世子), *Xəŋ (興王), dispatched an envoy with tribute; (iv) *Xəŋ died and his younger brother, *Muɑ (武王), became king of Wa. Based on this description, it is possible to graphically represent the following relationship using Figure 5.1. The dotted line between Kinashi Karu, ˜JIN

NINTOKU

Kinashi Karu INGY˜ ?

ANK˜

Y°RYAKU

Figure 5.1 Wa genealogy seen in Song shu.

The beginnings of writing a history 125 who is a placeholder, and Ingyō signifies that the record in Songshu lacks enough evidence to establish a link. Is this a break in the lineage or did the Song court have insufficient data to specify the relationship? Kishimoto (2013:9–12) believes that this break shows that the kings were not from a single lineage but that there were two ruling lineages. Based on a diachronic analysis of construction techniques and attributes of keyhole tombs, Kishimoto postulates that the current identification of imperial mausoleums is incorrect. According to the dating of the tombs, and the dates in Nihon shoki, Kishimoto (2013:16) reassigns identities to these tombs. Based on this information, he posits that there were two parallel lineages, as noted below (see Figure 5.2). I assume that the original historiographical project was completed under the auspices of the Kinmei court as Kinmei established a line of kings that continued to the present (regardless of whenever “present” was for each project). Also, starting with Kinmei allowed later historiographers to skirt the issue of how Kinmei came to power. It is plausible that the peninsular teachers of historiography may have used Chinese models to demonstrate how a story could be fudged, or certain facts suppressed. For example, it is unclear how much the Yamato court understood regarding the ideological differences between the violent Qin dynasty and the successive Han dynasty, where the Qin were Legalists, their government centralized and oppressive, while the Han were devoted to Daoism and displayed a general sense of leniency (in comparison to the Qin). Based on knowledge of how power was transferred in violent upheavals in China, it is possible that Kinmei and later Yamato courts were sensitive to how their history was portrayed. This suggests that in the beginning of the creation of the historical narrative the overthrow of Ankan was given a simple, factual account, while later iterations of the history reconstituted Ankan and Senka and grafted both into a line of continuing rulers to create a portrayal of an ideal royal family that had a mandate from the ruling deity to reign over Yamato in peace. If there were upheavals, these came from outside the royal family. The existence of a great king, called Waka Takeru “the young fierce one,” could not be ignored by later courts or their historians, as institutional memory preserved the fact that the establishment of their own archives was indebted to him. It is easy to imagine that there were stories of Great King Waka Takeru and the military prowess of his ancestors, especially in creating a kind of political fabric of great chieftains exerting some influence on the peninsula. These legends were easy to manipulate as evidence of an earlier empire the current rulers wished Line #1

Line #2

˜jin (Homuda) Rich° Hanzei Ichinobe Y°ryaku

Nintoku (Sazaki) Ingy˛ Kinashi Karu Shiraka

Figure 5.2 Two competing lines.

126

The beginnings of writing a history

to recapture. It is also possible that legend credited Waka Takeru with being the father of poetry, as institutional memory likely made it clear that writing, and by extension literature, took off during his reign. In this way, later courts romanticized Waka Takeru by attaching primitive songs to his name. Wohodo likely envisioned himself as a great chieftain, a leader whose domain extended across the sea to the peninsula. Later historiographers would have wanted to preserve this memory, as a unified court needed heroic stories from the past to create a sense of national identity and pride. The primitive source material for Kojiki does not appear to have had much in the way of a chronology. This does not mean that such material was nonexistent9; it simply meant that Hieda no Are and Yasumaro did not rely on those sources with a chronology. The three Paekche histories quoted in Nihon shoki all demonstrate that a chronology with varying levels of sophistication existed, and such a historical framework would naturally have been taught to the historians at the court of Yamato. However, there also appear to have been other sources based on oral traditions that lacked a strict chronology. Oral traditions by their very nature are fluid and tale-like. The pertinent question is why Kojiki relied on sources that lacked any dates. Several reasons for this choice are plausible. One, there was a backlash against a foreign model in support of a time-honored model based on oral traditions, where the storyline is more fluid, and the traditional system was deemed critical as a counterweight to the project that eventually spawned Nihon shoki. Two, sources that lacked a concrete chronology were older, being less sophisticated (and probably less colored by influence from Paekche and Chinese ideology), and the very antiquity of these may have been a selling point, i.e., Nihon shoki is newer, and non-traditional. Liberal, avant-garde projects often run into conservative obstacles. It is certain that the oral traditions contained genealogy-infused stories. It is fair to assume that when these were committed to writing, they were recorded in toto. If there was an attempt to filter out stories from truly historical events like birth, marriage, children, death, and burial, this took place after the initial step of recording the traditions committed to memory by the kataribe. As an example, consider the contents of the dynastic records of Keitai and Kinmei found in Book Nine of Kujiki (see Table 5.1). The records of these two reigns contain little aside from a pre-ascension story, and a few facts set in a basic chronology. Contrast this evidence with the fuller accounts in Nihon shoki, with the number of year-month entries found in Nihon shoki and Kujiki: Keitai: 8 in Kujiki, but 37 in Nihon shoki; Kinmei: 8 in Kujiki, but 79 in Nihon shoki. As an example of content, I provide Table 5.2 to outline how many year-month entries each record contains related to affairs on the peninsula. Aside from Kinmei’s deathbed wish that Mimana be reestablished, the records of these two important kings in Kujiki say nothing about the peninsula or foreign affairs. On the other hand, the Keitai record in Nihon shoki spends almost half its entries on the peninsula, and Kinmei spends three-quarters on affairs surrounding Mimana, Paekche, Silla, and Koguryŏ. In Tamakatsuma, Motoori Norinaga

The beginnings of writing a history 127 Table 5.1 Basic template of a record Date (年・月) Keitai Pre-ascension 1.1 1.2 1.3 2.10 5.10 8.1 25.2 25.12 Kinmei Pre-ascension 1.12 2.1 2.7 3.2 15.1 32.4 32.5 32.9

Kujiki Genealogy, story of finding Keitai in the provinces Court invites him in Keitai ascends the throne Invited to establish an empress, sets up eight consorts Buretsu buried Capital removed to Yamashiro, Tutuki Keitai asks Prince Magari to assist him Keitai falls ill and dies He is buried Genealogy Crown prince becomes king. Established Dowager Kinmei establishes a queen Capital moved to Kanasasi Five consorts established Nuna Kura made Crown Prince Kinmei falls ill. Begs for an attack on Silla to reestablish Mimana. He dies He is temporarily interred He is buried

Table 5.2 Comparison of peninsular data Ruler

# in Kujiki (Percent of total) # in Nihon shoki (Percent of total)

Keitai Kinmei

0 (0%) 1 (12%)

18 (49%) 58 (73%)

blasted Nihon shoki’s Kinmei record, saying, “The record is wholly about this event dealing with foreign procedures and the total sum of these fifteen pages is nothing but a waste of time. What reason is there for recording the completely bothersome, drawn out event?” (Bentley 2013:255). Compare these results with that found in Kojiki. The Kojiki record of Keitai makes no mention of the peninsula but does record the revolt of Iwai. Kujiki, on the other hand, ignores the revolt. The entry for Kinmei contains a long list of children, but nothing else. It is interesting that Kujiki has one mention of Mimana that Kojiki lacks, and Kojki has one mention of Iwai’s rebellion that Kujiki lacks. This is quite a startling omission in Kujiki, especially since Mononobe Arakahi is the principle actor in putting down the rebellion by Iwai. If Kujiki was compiled to glorify the Mononobe, the compiler missed a stellar opportunity here. It is also important to note which events both Kojiki and Kujiki ignore. Neither mentions the importation of Buddhism in the reign of Kinmei. The death of the

128

The beginnings of writing a history

powerful Soga Minister, Iname, goes unannounced in either record. In addition, neither record gives any hint to the existence of the supposed dream of Kinmei, where a man is located who found two wolves bloody from fighting, a supposed allegory about the fate of Ankan and Senka. As noted above, it seems fair to assume that the Keitai court inherited (took control of) the archives of the previous administration. Among these records, there surely were fragments or seedlings of stories in the institutional memory that later may have been embellished. Military stories were likely handed down from several generations back, but evidence suggests that the federation of states grouped around the ruler in Yamato was held together by economic and cultural powers, and not the fear of military force. Iwasaki Takuya (1990:151–58) has argued that an investigation into the burial goods of burial tumuli as these spread from the center of Japan into the periphery strongly suggests that the teisetsu that Yamato chieftains brought these satellite territories under their umbrella through military force is not persuasive. The more cogent argument is that Yamato was such an economic and cultural magnet that the chieftains of these satellite territories petitioned Yamato to join their quasi federation. This may have become a necessity as the link between Wa polities in western Japan lost contact with Kara polities because of Koguryŏ and Silla incursions. If this line of thinking is accurate, that would mean that the militaristic stories of Yamato Takeru in western and eastern Japan are highly embellished versions of what were actually smaller clashes. The situation on the peninsula is a different matter, however. There is weightier evidence that the military incursions into the peninsula by the Wa are factual,10 but it is still a difficult and fuzzy matter to connect Wa with Yamato. As far as the textual evidence is concerned, it seems clear that there were families in Kyūshū in the service of Yamato who directed military and diplomatic affairs, specifically in relation to the kingdoms on the peninsula. This is demonstrated by the rebellion of Iwai. Kojiki has a terse comment under the reign of Keitai: During this reign, Tsukushi Kimi Iwai did not follow the commands of the divine ruler and there were many instances of his going against the imperial order. Thus, the court dispatched two people, Mononobe Arakahi Ōmuraji and Ōtomo Kanamura Muraji and they put Iwai to death. The Keitai record in Nihon shoki has a more detailed account. Iwai desires to rebel against the court but is hesitant. After several years of indecision, Silla gets wind of Iwai’s desire, and offers him a bribe if he will move to prevent Kena Omi’s armada and army from reaching their shores. With this, Iwai moves to shut down the sea lanes and prevent Kena Omi from crossing the channel. Shoki records that Iwai sent a message to Kena Omi, saying, “You have now become the messenger. We once were compatriots, who rubbed shoulders and elbows, and ate the same food from the same pot. You cannot suddenly appear as the messenger and have me offer my allegiance to you.” This account strongly suggests that Iwai and his

The beginnings of writing a history 129 family had been intimately involved in military and diplomatic affairs on behalf of the court. It is plausible that under Great King Waka Takeru or one of his predecessors, strong military families in Kyūshū had offered their allegiance and expertise to the Yamato court. Kobayashi (2006:51–52) outlines a three-step process he believes occurred during the Yūryaku era: (1) strong families tied to the court fall into decline; (2) military families, such as the Mononobe and Ōtomo become the brains of the court, creating an imbalance at court with the military near the top; (3) powerful families who had charge of diplomatic matters are removed, and the court takes direct control of foreign affairs. Evidence suggests that during the time of Yūryaku, satellite chieftains who had pledged allegiance were granted title and authority. I believe that these newly designated territories were given a peninsular title, *koni, representing an institution established in imitation of a system in Paekche. This word is seen in the Chinese work, Zhoushu, where the king of Paekche is called 鞬吉支 *kjɐn-kit-tsje and the queen is called 於陸 *Ɂuo-ljuk. These are close to the Japanese renderings found as interlinear glosses in Nihon shoki, where 大王 “great king” is glossed konikisi and 后 “queen” is glossed oruku. With vowel raising that occurred around the late sixth century (Miyake 2003:132), an older *koni became kuni. I believe the etymology of this is “large (tract of land).” This may seem like a vague label to use, but consider that the English word country comes from postclassical Latin contrata meaning, “that which lies opposite or fronting the view, the landscape spread out before one” (Weiner and Simpson 1971:1078). Or how the word district comes from the verb distrain, meaning “the territory within which the lord may distrain” (Weiner and Simpson 1971:535). Distrain means “to control by force.” These chieftains placed in charge of these newly established territories were later labeled kuni no miyatuko “men of the palace (who control) kuni.” It is clear from the titles mentioned in the 28th year of Suiko (620) that these territories were controlled by families that kept their own records. The record in Suiko mentions the title in this context: 是歳、皇太子嶋大臣共議之、録天皇記及 國記、臣連伴造國造百八十部幷公民等本記 “During this year the Heir (Shōtoku) and Great Minister Shima (Umako) discussed things together, and recorded Tennōki and Kokki, along with the Original Records of the Omi, Muraji, Tomo no Miyatsuko, Kuni no Miyatsuko, and the 180 be, along with a record of the subjects.” Tsuda (1966:23–24) argued that Tennōki “record of the heavenly sovereign” was modeled on the Chinese benji or annals, while Kokki “record of the land” was modeled on the Chinese biographies, but there is no way to know if this is accurate. Enoki (1975:123) persuasively reasoned that since the title tennō appeared around the era of Tenmu, this cannot be the true title here. Thus, the original title may have been Tenki (天記), creating a dichotomy between “heaven” and “earth,” or 天 and 国. It is interesting, however, to find both the titles 國記 and 國造本記 in this account. It is tempting to interpret國記 Kokki as a record of the state and its people, as Sakamoto (1979:164) does, but it is highly doubtful that at this stage ancient Japan saw its ruling family and the state as two separate entities. It is

130

The beginnings of writing a history

much more likely that this work was intended to be a compilation of the various territories with local legends and information, including the lineage and stories regarding the ruling chieftains of each territory. For the time being, I assume that 國造本記 Kokuzō hongi is much like that found in Kujiki: a list of the territories and the genealogy of the governors with related stories. From this list of titles, I construct Figure 5.3, which represents a historiographical ideal that likely was canonized in the Kinmei court. This suggests that the ruling family had direct jurisdiction over the various powerful chieftains, who in turn had authority over the Tomo no Miyatsuko (Hanzō) and the industrial groups (the be). The ruling family also had some oversight over the various kuni, which were then given semiautonomous control by the local Kuni no Miyatsuko. Based on the evidence presented in this and the previous chapter, I postulate that the Kinmei era record, Wa kimi hongi (倭王本紀), contained 12 volumes, modeled after the 12 benji of Shiji. As a working hypothesis, I believe these 12 benji were divided among the following legendary rulers and kings: Hikoyu (including the mysterious origins and founding of the dynasty), Sujin, Suinin, Yamato Takeru (outlining battles with Koguryŏ and the Kumaso), Ōjin, Nintoku, Ingyō, Ankan, Yūryaku, Kenzō, Keitai, and Kinmei. The reader may be forgiven for the expected reaction that it appears that I have pulled these names out of a hat, but the selection of individuals I postulate were in the first history of the Yamato court is based on thorough research conducted on the text of Kojiki by Kimura Ryūshi (1992:219–231). Kimura examined the fabric of the story in Kojiki and conducted a statistical analysis of the actors and agents in each sentence, as well as the number of times certain “emotive” verbs are used. This research is based on the assumption that the original oral traditions placed emphasis on the importance of various stories by embedding certain emotive verbs to create literary tension. Kimura arranged a list of ten emotive verbs and then calculated how many times one or more of these verbs appear in a story with a specific actor. He found that the top ten actors statistically are Yamato Takeru Tenn˝-ki

Kokki

Kokuz˝ hongi The Subjects

Figure 5.3 Hierarchy of early records.

Original Records

Slave Records

Hanz˝ hongi

180 Be hongi

The beginnings of writing a history 131 (170 times), Yūryaku (134), Izanagi (130), Susanoo (110), Nintoku (102), Jinmu (80), Suinin (80), Kenzō (77), Izanami (71), and Ōkuninushi (72). There is no one actor in the storyline that has all ten verbs associated with them, but Yamato Takeru has the most, eight. This suggests that the story surrounding Yamato Takeru was popular and had become highly stylized. Kimura then investigated how much text of the record is devoted to each actor, calculated by the number of characters used in each story. The top five actors are Izanagi (2740 characters), Nintoku (2101), Yūryaku (1899), Susanoo (1849), and Yamato Takeru (1829). Taking these data, I filtered out certain names. First, I strip out mythological actors, such as Izanagi, Izanami, Susanoo, and Ōkuninushi, from the list, because it seems evident that these were added to the mythological fabric later. Triangulating these remaining data, I have pulled out the names of the rulers of Suinin, Nintoku, and Kenzō. I have included Sujin instead of Jinmu, because I believe Takioto makes a persuasive case that the story of one archetypical ruler was later split into two, creating Jinmu and Sujin. It is difficult to know if the original was Jinmu or Sujin, or neither. For now, I use the name Sujin as the marker for a legendary “founding” king. It also seems clear that this original record was later adjusted and reworked over time. It is impossible to confirm or refute this suggested framework of the original history. I merely suggest one possibility for a history created by the court of Keitai or Kinmei. The other set of records was kept by the various families, depending on their function at court. As a representative example, I quote below two sections of genealogy from the Owari and Mononobe families as preserved in Kujiki. I ask the reader to pay special attention to how main achievements of a specific individual is grafted into the genealogy of the family and set against the backdrop of a specific king the individual served. Owari: The thirteenth generation is Shirizunane. He was made minister during the reign of Great King Homuda [Ōjin] and served in that position. His sister is Ozuna Mawakatobe, and she was betrothed to Prince Ihoki Irihiko, and gave birth to Prince Homuda Mawaka. Next is another sister, Kanada Yanuhime. She was betrothed to Prince Homuda Mawaka and gave birth to three princesses. They are Takaki Irihime, Nakahime, and Irotohime. These three were all set up as ladies to Great King Homuda and gave birth to thirteen children. (Bentley 2006:203) Mononobe: The thirteenth generation is Mononobe Okoshi Muraji, the son of Arayama Ōmuraji. He was made Ōmuraji during the reign of the Great King who ruled from the Shikishima Kanasashi Palace [Kinmei] and worshipped at the great shrine. He married Asahime and Kaharu Hime, the daughters of Yamatoko, the ancestor of the Yuge Muraji. Asahime bore four children, while Kaharu Hime bore three.

132

The beginnings of writing a history Mononobe Nase Muraji is a brother to Okoshi. Mononobe Masara Muraji, the child of Itabi Ōmuraji. Masara Muraji was made Ōmuraji during the reign of Great King Hatsuse Namiki [Buretsu]. He went and worshipped at the great shrine. He took Imoko, the daughter of Suha Atahi, as his wife, and she bore two children. Another brother is Mononobe Me Muraji. He was made Ōmuraji during the reign of Great King Wohodo [Keitai] and worshipped at the great shrine. (Bentley 2006:221–22)

These two quotes demonstrate what basic family lineages looked like, where genealogy was recorded, as well as achievements of specific individuals, and in what reign these had happened. This suggests that while there were a variety of records among powerful families, repeated recording of these events over time gradually began to create a sense of inter-connectedness through the reference of the reigns of specific kings. Thus, I conclude that by the beginning of the Bidatsu era, histories existed at court in a variety of flavors, but likely there was only one “court” history, that being Wa kimi hongi.

Suiko era historiography As noted in the previous section, historiography around the Suiko-Jomei era is crystalized in one statement found in the 28th year of Suiko in Nihon shoki: During this year the Heir and Great Minister Shima discussed things together, and recorded Tennōki and Kokki, along with the Original Record of the Omi, Muraji, Tomo no Miyatuko, Kuni no Miyatuko, and the 180 be, along with a record of the subjects. Many scholars have questioned the veracity of this account, but some believe the essential information to be factual.11 On the one hand, the account is a difficultto-accept attempt to portray both Shōtoku Taishi (the Heir) and Soga Umako (the Great Minister) as actually compiling a two-pronged history of both the ruling family and those families that support the court. It is highly unlikely that either individual would have done any such thing, aside from functioning in a possible role of oversight, if the account is based on fact. Taking a step back and looking at the process from the end toward the beginning, a discerning eye can see the hand of a historian from the early Nara era piecing together this account in the Suiko record; one naturally reads this account and then remembers the outcome of this attempt at historiography appearing at the end of the Kōgyoku record, when the Soga were dethroned: Soga Omi Emishi and others were about to be put to death, so he (Emishi) set fire to Tennōki and Kokki and all their (family) treasures. Fune Obito Esaka quickly took Kokki which was about to be burned and presented it to Naka no Ōe.

The beginnings of writing a history 133 What is fascinating is that according to this account, the only rescued manuscript is Kokki, the impression in the record being that Emishi successfully burned the manuscript of Tennōki. The conclusion the reader must reach is that this royal record was lost. However, it seems to this scholar that this is a historiographical ploy by the Tenmu~Jitō court that allows them to claim, “Any surviving manuscripts of this work are fraudulent, because Emishi burned the only known copy. That is why it was still in his private possession.” Clearly these accounts were manipulated, perhaps because these were potentially politically charged records. Umezawa (1968:297–98) has focused on the special status the Fune Fubito family enjoys in the pages of Nihon shoki. He notes that it is extremely rare for the origin of any peninsular family to be recorded, and yet Nihon shoki mentions the origins of the Fune Fubito. Later, having Nihon shoki record the story of Esaka rescuing the manuscript of Kokki highlights a special connection that this family had with the historiographical process. It is interesting that the origins of the Fune Fubito are also connected, at least according to Nihon shoki, to the Soga. In the 14th year of Kinmei Nihon shoki records, Great Minister Soga Iname Sukune accepted the imperial command and dispatched Wau Chin I to count and record the tribute brought by ship. With this, the court made Wau the overseer of shipping, and he received the title of Fune Fubito (“Fubito of shipping”). This is the founder of the Fune Muraji. Regarding the identity of this Wau, Shinsen shōjiroku records, “Fune Muraji. They are descendants of Chi In, a great-grandchild of Prince Tay A Lang (of Paekche).” Concerning Wau Chin I, the grave marker for the Fune Obito says, “The founder of the Fune Uji was Na Phay Kwo, the son of Wau Chi In” (Takeuchi 1962:964). This line of reasoning leads us to conclude that Soga Iname had some connection with Paekche. It is plausible, though difficult to prove, that the Soga were originally from Kara or Mimana. With a connection to Fune Fubito, the Soga now have oversight of foreign tribute. Notice, however, that the important reason for the appointment was Wau’s ability to record the tribute. Later, in the third year of Bidatsu (574), Nihon shoki mentions, “The court gave a command to Ushi, the younger brother of Fune Fubito Wang Sinni, and changed his name to Tsu Fubito.” Wang Sinni is put in charge of the Naniwa Port (tsu means port). Here two groups from the same family under the supervision of the Soga now control foreign tribute and the entry point (port of Naniwa). The commonality here is that we continue to have several points that lead back to the kingdoms on the peninsula, the source of advanced culture and technology. It is possible that Tenmu’s purported complaint about records in the possession of some families having deviated from the court’s perception of “truth” was aimed at the Fune Fubito family. An interesting theory sees the Fune family under the authority of Soga Umako, compiling a new history of the ruling families, taking liberties that more orthodox circles at court would have found difficult to swallow.

134 The beginnings of writing a history It is even plausible that this record, Tennōki, may have recorded that the Soga were of divine lineage. Notice that in Shinsen shōjiroku there is an entry each for the Soga [276] and the Sogabe [300]. The first family claims to be descendants of Emperor Kōgen, and the latter claims to descend from Emperor Chūai. Kojiki claims that Soga Ishikawa Sukune was a son of the legendary Takeuchi Sukune. The first appearance of the Soga in Nihon shoki is Soga Machi Sukune in the record of Richū, but this may be a later addition. Interestingly, the record in Nihon shoki preserves evidence that points toward the peninsula. Notice that in the record of Yūryaku, we find the appearance of Soga Karako (韓子) Sukune, literally “Soga, the son of a peninsular (father).” Let us return to the crucial question about whether the account in the 28th year of Suiko is a phantom or not. Most scholars fall into one of two polarized camps: the account is true or fabricated. This scholar advocates a more measured approach. As noted in the previous section, I believe that several kinds of records were already in existence prior to the Suiko era, and these have a connection with Fune Fubito and other family groups who had literate members, but neither Shōtoku Taishi nor Umako was directly involved. It is possible that the Fune Fubito family was involved in compiling a new history, under the authority of Soga Umako, but the Nihon shoki compilers did not feel it necessary (or wise) to clearly demarcate the connection between the Soga and the Fune Fubito. Rather, I believe the compilers felt it critical to engage in hagiography, and thus, they added Shōtoku Taishi’s name to give the project royal authority, and make it appear that the court was directly involved. Since Shoki portrays the project being abandoned before completion, the compilers may have not bothered to entertain the fact that Tenmu and others at his court found the aborted project a ruse by the Soga to claim legitimacy to inherit the throne. It may have been at this critical time, with China sending an envoy (Bei Qing) to Yamato, that people at court realized that Yamato needed a foundational story, in imitation of other histories on the continent. I believe that the story of Mimaki Iribiko Iniwe (Sujin) was fleshed out during the Suiko era. Rewriting the story of Mimaki leaving Kyūshū, traveling up the coast in the Inland Sea, invading the Kinai region to subdue the surrounding people, and establishing his capital in Yamato likely was based on a legendary story handed down by a few powerful families, such as the Mononobe and the Kumebe. Thus, the Bureau of History set up Mimaki as the founder of the Yamato court. As I have previously noted (2008a:22), it is possible that the word *mema > mima was a Kara word meaning “iron,” which explains the place name Mimana on the southern tip of the Korean peninsula, and the place where iron was discovered in Japan, Mimasaka > Iron Hill. The pitch accent matches both names (HHx). Thus, Mimaki is not a personal name, but a toponym, likely *memakiy “Iron Fort.” The appended phrase after this, iribikwo, means “the prince who enters,” suggesting someone with authority to enter and rule. Sujin’s name as found in Kojiki poem 22 (or Nihon shoki 18) has his name as Mimaki Iribiko, lacking the phrase iniwe. I believe Mimaki Iribiko to be the original name and is not really a name but a simple appellation in place of a name, “The Prince entering the land from the Iron Fort.”

The beginnings of writing a history 135 If the above description were accurate, then it would follow that the Suiko-era historiographical committee fleshed out the details for the military stories dealing with the subjugation of the Kumaso in the west, the Emishi in the east, and the near-legendary accounts of military adventures on the peninsula. The first two stories with the hero, Yamato Takeru, were dramatically portrayed, where he went to both the west and the east, and put down rebellions by using a magical power of deception. This allowed the court to portray the ancient kings as men endowed with a power that transcended the other great families (bands). Brute military force was less secure than a metaphysical power, so the historians couched these military exploits in new terms—a ruler who was shrewd and possessed a matrilineal power going back to his royal heritage. The military stories of events on the peninsula presented the court with a different challenge, as families had various accounts about battles that took place centuries earlier on battlefields across the sea. There is evidence demonstrating that Silla was not the original villain in these early stories. Kujiki preserves the title of a lost work that sheds a little light on this subject. In the record of Empress Jingū, Kujiki records: Tenth month, third day. The heavenly and earthly deities with the violent spirits were worshipped in a different ship. Also, the good spirits were worshipped in the ship in which the Queen rode. Leading the naval armada out of the port of Wani, the hawsers were loosened, and the Imperial boat set sail for the journey to the land of Silla. The entire chain of events is recorded in The Record of the Subjugation of the Three Han (征服三韓記). (Bentley 2006:302) While Kujiki shows a renewed emphasis on expanding the story surrounding the humiliation of Silla,12 it stops short of providing details, instead sending the reader to an account of the “three Han,” which may have referred to Paekche, Silla, and Koguryŏ, but just as easily could have referred to Paekche, Silla, and Kara. What it does not refer to is an invasion and subjugation of a single kingdom, Silla.13 While the work to produce a reconstructed (or updated) version of the history was aborted at the end of the Suiko era, this does not mean that a draft was destroyed. I believe that the Tenmu era and later historiographers wanted evidence to cast doubt on any surviving manuscripts. Furthermore, even though the process was aborted before fruition, evidence suggests that certain aspects of this new historiographical movement had gained traction. One aspect was the attempt to provide a fuller list of rulers to create a history that extended back to a time before the foundation of Paekche. This point is critical, because it is plausible that most of the recorded history the Yamato court had originally used as source material for their native histories came from Paekche accounts, as that kingdom became literate several centuries before Yamato. There is some evidence preserved in a variety of legends recorded in Fudoki that there were more rulers (or “great kings”) than are currently codified. This suggests that there was a movement to fill out the history by a variety of groups

136

The beginnings of writing a history

that may have submitted the names of their ancestors with a connection to a ruling group in the past. Four of the surviving Fudoki are possible sources for this claim: Izumo, Harima, Hitachi, and Bungo fudoki. First, we can eliminate Bungo fudoki because internal evidence demonstrates it has relied on Nihon shoki for a large portion of its information. Notice that all the rulers mentioned in the text record the names of these rulers exactly as found in Nihon shoki. I also exclude Izumo fudoki because there are only three rulers mentioned (Tenmu, Kinmei, and Keikō), and there is little information there for us to draw any conclusions. That leaves us with Harima fudoki and Hitachi fudoki, and both contain important information. Consider Table 5.3.14 Several items are worthy of note. There are a number of names that match the spelling in Kojiki. Rather than claiming Fudoki relied on Kojiki, it is more reasonable to view the commonality as a connection with older source material both records cited. Notice these divergences: Kojiki never uses 足 to represent tarasi, only 帯. It also seems clear that Yasumaro altered the name of Ninken to align with a pattern he must have seen in the old records, one where father and son, or older and younger brother, were designated as opo “large, older” versus wo “small, younger.” Thus, the older designation of Oke and Woke was altered to Ohoke and Woke. Another interesting point is the inclusion of two individuals as rulers who are not included as rulers in the current imperial chronology. Thus, Prince Uji, who court documents describe as having refused to ascend the throne, yielding his right to reign to his brother, Nintoku, is found in Harima fudoki as having ascended the throne. The other, Ichinohe, never ruled, according to Kojiki and Nihon shoki, because Yūryaku put him to death, fearing him as a successor to the throne. Clearly, this story was rewritten by the historians in the early Asuka period.15 Consider what Harima fudoki mentions about both rulers. Table 5.3 Names of Yamato rulers in Fudoki Name

Harima

Hitachi

Kojiki‖Shoki‖Kujiki

Sujin Suinin Keikō

— — 大帯日子

初国所知美麻貴 伊久米天皇 大足日子

Chūai Jingū Ōjin Nintoku Udi Richū Ichinohe Yūryaku Kenzō Ninken Suiko?

帯中日子 大帯日売命 品太 大雀 宇治天皇 大兄伊射報和気 市辺天皇 大長谷 袁奚 於奚 小治田河原

— — — — — — — — — — —

所知初国之御真木‖御肇国天皇‖NA 伊久米伊理毗古伊佐知‖活目入彦狭茅‖活目帝† 大帯日子淤斯呂和気‖大足彦忍代別‖日本大 足彦忍 代別 帯中日子‖足仲彦‖S 息長大帯日売‖気長足姫‖S 品陀‖誉田‖S 大雀‖大鷦鷯‖S 宇遅能和紀郎子‖菟道稚郎子‖S 伊耶本和気‖去来穂別‖S 市辺之忍歯‖市辺押磐‖磐坂市辺押羽 大長谷若建‖大泊瀬幼武‖S 袁祁‖弘計‖雄計 意富祁‖億計‖S 小治田宮‖小墾田宮‖NA

The beginnings of writing a history 137 (1) Upper Pakowoka, Lower Pakowoka, Nabetsu, and Apukoda: During the reign of Great Ruler Uji, two people, Elder Brother Takanashi and Younger Brother Takanashi who are the distant ancestors of the Uji Muraji, requested and received the land of Yohuto, and opened that land. When they came to their land and were about to plant seed, a messenger was carrying dishes and kitchen wares on an aputo “a pole,” when the pole suddenly broke in two and the goods fell to the ground. Where the cooking pots (nabe) fell was called nabetu… [上筥岡下筥岡魚戸津朸田宇治天皇之 世宇治連等遠祖兄太加奈志弟太加奈志二人請大田村與富等地墾田將 蒔來時廝人以朸荷食具等物於是朸折荷落所以奈閇落處]. (2) The second story is about Great King Yūryaku inviting Prince Ichinohe out for a hunt, and then assassinating him. A servant heard this report, and took the Prince’s two sons, Oke and Woke, and fled to an area in Harima and hid until after Yūryaku had passed away. Thus, the record calls Ichinohe “great ruler” because it was his right to rule. The song and dance where the brothers announce their lineage is also included in Harima fudoki, and because it is a much simpler version, I believe it is the archetype. It goes: Afumi is a land of lakes. Yamato is a land of mountains like a blue barrier. Within those mountains like a blue barrier resided Great King Ichinohe. We, your servants, are his descendants [淡海者 水渟國 倭者 青垣 青垣 山投坐 市邊之天皇 御足末 奴僕良麻]. It is unclear, however, why Prince Uji is designated as “heavenly sovereign” when he declined to rule, where Ichinohe was the rightful heir but was assassinated before he had a chance to reign. This suggests that in some earlier record, there was a Great King Uji who had ascended the throne. It seems plausible that during the Suiko era, when historiography underwent the first great revision, Prince Uji was added, because he was the rightful heir. When historiography underwent the second great revision, during the Tenmu-Jitō era, Great Ruler Uji was deleted in a rewriting of Nintoku’s history to frame the latter as a virtuous ruler. It would make Nintoku appear more virtuous if he reluctantly accepted the right to rule only because his older brother, Uji, had declined the throne, and then took his own life to demonstrate his determination—an expression of how a cultivated person should act. Uji was thus sacrificed to be reincarnated as an example of a self-deprecating Confucian gentleman. The story in (2) above is fascinating. Two children of Prince Ichinohe are taken and hidden in Harima for their protection, but the overall story is unsophisticated. The declaration in Kojiki is more elaborate and worth quoting here: On the hilt of the long sword gird around the waist of our military lord, he has applied red clay. He has attached a red piece of cloth to the string. When he hoists the battle flag for display, he rules over everything under the heavens, like one who cuts the bamboo at the base on the unending mountain peaks and waves the tips in the air. Or like the master who strums the many strings on the koto. That lord is Prince Ichinohe no Oshiha, son of Great King Izaho Wake. We, your

138

The beginnings of writing a history servants, are his descendants [物部之我夫子之取佩於大刀之手上丹畫著 其緒者載赤幡立赤幡見者五十隱山三尾之竹矣。本訶岐苅末押縻魚簀如調八絃琴所治賜天下伊邪本和氣天皇之御子市邊之押齒王之奴末].

The same scene as depicted in Nihon shoki is the most elaborate of the three, quoted here: Prince Oke then stood up and danced. After he finished, the Divine Ruler (Woke) got up. He adjusted his robes and sash and gave the following auspicious words for the newly built edifice (words omitted) … Odate said, “How interesting! Please, let me hear more.” The Divine Ruler then performed the dance known as the Tatsuzu Dance. … He then proclaimed loudly, “This land of Yamato / a land of rustling reeds / is ruled by the younger prince, me!” With this, Odate was greatly surprised, and caused him to repeat what he had said. The Divine Ruler then replied in a louder voice, “They cut down the sacred cedar / of the Furu Shrine / in Isonokami of Yamato / stripped it of its branches / and built the Ichinohe Palace / from which the Prince ruled the Empire / Prince Ame Yorozu Kuni Yorozu Oshiha / We are his children!” [億計王起 僢既了。天皇次起、自整衣帶、爲室壽曰・・・小楯謂之曰、可怜。願 復聞之。天皇遂作殊儛。誥之曰、倭者彼々茅原、淺茅原弟日、僕是 也。小楯由是、深奇異焉。更使唱之。天皇誥之曰、石上振之神榲、於 市邊宮治天下、天萬國萬押磐尊御裔、僕是也。]. Kojiki and Nihon shoki have relied on the same basic story, as both versions demonstrate a connection with the Mononobe and their Isonokami Shrine, also known as the Furu Shrine. Such a connection may reveal that this dynasty used the Mononobe as military support. Whatever the reason, clearly the version in Fudoki is the least embellished, and therefore must be the older. Thus, it appears that later historiographers had Ichinohe murdered to tee up the quixotic story of Oke and Woke becoming refugees, waiting for a heavenly decreed mandate to come back and claim their rightful place on the throne. Evidence from Kujiki suggests that in the Suiko-Jomei era, a template was created, perhaps based on earlier records that formed a foundation around which each royal reign would be constructed. Table 5.4 contains data taken from the dynastic records of the first 27 rulers from Jinmu until Keitai. I have categorized basic data thus: does the record have a first- and second-year entry? Is the queen elevated to the status of Queen Dowager [QD]? When is the capital established? Finally, Table 5.4 lists the number of total entries in each record. Several important deductions are obvious from this table. First, the template seems to have assumed that a typical regnal record had 5- or 6-year entries, including the inaugural year that often contained the announcement that the previous reign’s queen received the title of Queen Dowager, and that a new capital had been selected. The template also included a year entry showing when the court set up the main wife as the new queen, followed by a listing of the royal children. Finally, the year when a son is made heir to the throne and the death of the king

First

○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○

Name

Jinmu Suizei Annei Itoku Kōshō Kōan Kōrei Kōgen Kaika Sujin Suinin Keikō Seimu Chūai Jingū Ōjin Nintoku Richū Hanzei Ingyō Ankō Yūryaku Seinei Kenzō Ninken Buretsu Keitai

○ ○ ○ ○ ╳ ○ ○ ╳ ○ ╳ ○ ○ ○ ╳ ○ ○ ○ ○ ╳ ○ ○ ○ ○ ╳ ╳ ○ ○

Second

Table 5.4 Overview of early records in Shoki

NA 1st 1st 1st 1st 1st 1st 1st 2nd 1st 1st 1st 1st 1st NA NA 1st 1st None None 1st None None None None None None

QD 1st 1st 2nd 2nd 1st 2nd None 4th 2nd 3rd 2nd 4th 1st None 3rd 1st 1st 1st 1st None 1st 1st 1st 1st 1st 1st 5th

Capital No record 2nd 3rd 2nd 29th 26th 2nd 7th 7th 1st 2nd 2nd 48th 1st NA 2nd 2nd 1st 1st 2nd No children 2nd No children No children 1st No children 1st

Children 42nd 25th 11th 22nd 68th 76th 26th 22nd 38th 48th 37th 51st 48th ╳ 3rd 40th 31st 2nd ╳ 23rd ╳ 22nd ╳ ╳ 7th ╳ ╳

Heir 76th 33rd 38th 34th 83rd 102nd 76th 57th 60th 68th 99th 60th 50th 9th 69th 41st 83rd 6th 5th 42nd 3rd 23rd 5th 3rd 7th 8th 25th

Death 5 6 6 4 5 6 5 6 6 7 11 10 5 3 4 4 9 4 2 5 3 4 4 2 3 3 5

Total

The beginnings of writing a history 139

140

The beginnings of writing a history

rounded out the template. As proof of the existence of such a template, consider how splicing together the information from the two records of Chūai-Jingū into one dynastic record erases the gaps between the two (see Table 5.5). At least on the surface it does appear that the late Asuka historiographers took Chūai’s dynastic record and divided it into two reigns, thus allowing Jingū to rule, where the older records did not have her on the throne. The historiographers re-portioned the information from Chūai’s record and spread it out between two reigns. This theory is strengthened further by evidence mentioned later about the legendary leader of the invasion of the peninsula originally being a male. Another interesting pattern is the cessation of awarding the title of Queen Dowager after Ankō, as if Yūryaku was the point where this announcement either began, and was backdated, or simply abolished. Also, while the establishment of an heir to the throne is regulated through most of the dynastic record, one generation before the reign of Yūryaku we find that establishing an heir becomes problematic until the reign of Kinmei. This demonstrates that here the ideologically rigid template met the messy record of documented reality. Based on this template, we may infer that during the Suiko-Jomei era, historians took the pantheon of historical and legendary rulers and rearranged these, working back in time. This allowed the historians to tweak the older chronology. The reign of each ruler was predicated upon the recorded age of each individual.16 Considering the data as outlined in Table 5.5, I envision that it took the historiographers two steps to create a revised version of the history of the state. The first step was to regulate the length of the rulers, according to how long they had lived, graphically represented in Figure 5.4. Based on all three records, it is clear that at least two sets of dates were committed to institutional memory: the date of royal ascension and the death date of the king. My conclusion is that a birth date did not exist in the institutional memory, forcing later historians to arrange this linear information and massage it, taking into consideration the reigns before and after each king, graphically represented in Figure 5.5. This exercise created an overlap, where the date of birth of each king had to be calculated, based on fuzzy math, sandwiching it between the death date of one

Table 5.5 Chūai and Jingū records combined Name

First

Second

QD

Capital

Children

Heir

Death

Total

Chūai Jingū

○ ○

╳ ○

1st NA

None 3rd

1st NA

╳ 3rd

9th 69th

3 4

King A (68 yrs)

King B (70 yrs)

King C (55 yrs)

Figure 5.4 Sample kings’ length of lives.

King D (64 yrs)

King E (52 yrs)

King F (75 yrs)

The beginnings of writing a history 141 King A King B King C King D King E King F

Figure 5.5 Creating overlapping lives of kings.

king and the date of ascension of another. This imprecise method has resulted in various surviving texts providing conflicting information about the age of rulers in the sixth and seventh centuries, as if one could only know the age of a specific ruler in the past by doing some math. As an example of the newer method, where the number of years of a reign are recorded, the end of Jōgū Shōtoku hōō teisetsu contains the following information (Bentley 2002:114): The reign of the Shiki Shima Great King [Kinmei] (He died in the fourth month. His burial tumulus is in Hinokuma Sakaai) was 41 years long. The Osada Great King [Bidatsu] reigned for 14 years (He died in the eighth month of the fourteenth year [585]. His burial tumulus is in Kafuchi, Shinaga). The Ikenohe Great King [Yōmei] reigned for three years. (He died in the fourth month of the third year [587]. Another record says, “His burial tumulus is in Kafuchi, Shinaga Nakawo.”) The Kurahashi Great King [Sushun] reigned for four years. (He died in the eleventh month of the fourth year [592]. In truth, he was assassinated by Minister Shima [Soga Umako]. His burial tumulus is in Kurahashi Woka) The Owarida Great King ruled for 36 years. (She died in the third month of the thirty-sixth year [628]. Her tumulus is in Ōno Woka. Another record says her tumulus is in Kafuchi, Shinaga, Yamada Village.) Here, there are several discrepancies. Teisetsu claims Kinmei reigned 41 years, while Nihon shoki has him reign 32 years. Kojiki does not record either the age or the reign of Kinmei. Yōmei’s reign is three years in Kojiki and Teisetsu but two years in Shoki. This deficit is made up in the following reign of Sushun, where Kojiki and Teisetsu have him reign four years but Shoki has five. Notice that Kojiki has Suiko reign 37 years, where Shoki and Teisetsu agree that she reigned 36 years. This problem with calculating the ages of kings becomes more pronounced the further back in time one went. Compare the recorded ages of the six rulers from Yūryaku to Keitai, represented in Table 5.6.

142

The beginnings of writing a history

Table 5.6 Ages of six rulers King

Kojiki

Shoki

Kujiki

Fusō ryakki

Jinnō Shōtōki

Yūryaku Seinei Kenzō Ninken Buretsu Keitai

124 – 38 – – 43

– A bit – – – 82

124 Blank – – – 82

93 39 48 50 18 82

80 39 48 50 58 80

The interesting point is how none of the first three works agrees on the age of Seinei. Agreement occurs only later, in the Heian era. Kojiki ignores the issue by not listing an age. Shoki uses a vague phrase 若干 “a bit” as an age. Kujiki leaves it unfinished, as if some committee later would come to an agreement and insert an age: 年 __, “age ___.” There is also an almost 40-year difference in the age of Keitai between Kojiki and Nihon shoki, perhaps because one is actually a reign and the other an age. Furthermore, Kojiki lists Kenzō’s age, but Shoki and Kujiki omit it. Both Fusō ryakki and Jinnō shōtōki try to wring some consistency out of the older records by adding a decade to Kenzō’s age as recorded in Kojiki. Some may say that Kujiki took its text from Shoki—it is natural for there to be copying, or image omissions—but this ignores the fact that while Shoki does not include an age for Yūryaku, Kujiki has the same age as Kojiki. It is also interesting that there appears to have been an effort to bring Yūryaku’s age back down to earth, as it is whittled down from 124 to 93 to 80. A severe discrepancy occurred with rulers in the hoary past. Consider that Nihon shoki has Yamato Takeru pass away in the 40th year of Keikō. According to Shoki’s chronology, Chūai, the second son of Yamato Takeru, becomes heir to the throne in the 48th year of Seimu when he was 31 years old, a full 68 years after his father has died. Kujiki, on the other hand, has Chūai born in the 51st year of Keikō’s reign, eliminating 11 years of this manipulation. G.W. Robinson (1955:117–18) aptly noted: The death of Yamatodake [our Yamato Takeru] is related with much detail in the Shoki under Keikō 40, but briefly in the Kuji Hongi [our Kujiki] under Keikō 51. Both works, however, list the prince’s consorts and offspring in Keikō 51. … This is natural enough in the case of the Kuji Hongi, but it seems a curious misplacement in the Shoki. The chronology of the Shoki is further complicated by the last words of the long entry in Keikō 40. … It is true that the brief reference to the death of the prince in the Kuji Hongi reads like a summary of something longer …, but the separation by the Shoki of the death from the list of children by 11 years … remains unnatural and strongly suggests the compilers forgot to move the list when altering the date of the prince’s death and treating it in such great detail.

The beginnings of writing a history 143 After this template was established, historiographers inserted events and stories. Table 5.6 demonstrates that most of the basic events of a specific dynastic record were prescriptive, but the establishment of the heir likely was one of the few pieces of information based on an actual tradition, a conclusion based on the great disparity of the years. If we can begin cautiously to trust the information close to Yūryaku’s reign, it would appear that the tradition was to appoint an heir when the king sensed that his death was imminent. This is not the case with the establishment of an heir from Jinmu down through Nintoku, where the court established the heir roughly between the middle to the last quarter of a king’s reign. An extreme case is Ōjin becoming heir to the throne at the young age of three, but this account was manipulated when historians rewrote the story with Jingū as a quasi-ruler. I have tried to argue that the inconsistencies and misalignments found in Shoki’s chronology were the result of a series of events involving manipulation and insertion of episodes over time and not the result of one great compilatory blunder during the last stage of writing around the time of Tenmu and Jitō.

Early Nara historiography The preface of Kojiki claims that the impetus for a rewriting of earlier histories by the Tenmu court was based on a belief that the earlier documents had deviated from the truth. While we cannot deny this perception outright, it is wiser to see this rewriting of history as one component of a significant restructuring of government during the reign of Tenmu-Jitō. In the beginning of Tenmu’s reign, after quelling the Jinshin Disturbance, I see the Bureau of History engaged in several new techniques. These were influenced by an avant-garde philosophy at the bureau—that of Chinese thinking based on the Chinese worldview. Some Yamato students who had studied abroad in China in previous reigns likely were employed in the bureau, and the innovative and advanced techniques from China would have been exciting to younger historians weary of the older traditional methods introduced from Paekche centuries earlier. The question is what these new historians contributed. Was it a more complex historical philosophy, or did it include actual recording techniques? There is a rather intractable theory that the early Nara historiographers of Nihon shoki gave the early rulers long lives so they could create long chronologies and backdate the advent of Jinmu to 660 BC, but this theory has considerable problems. Early on Mishina (1948:192–93) argued that Kojiki, which has no chronology, already claims that many rulers had lived long lives; some of these rulers in Kojiki live longer than their counterparts in Nihon shoki. He argues that these early rulers likely were described as living long because this gave them a mystical or sage-like quality. Thus, he contends that the process is backward from what is often argued: sage-like, legendary rulers with mystical longevity existed in earlier records, which Kojiki has simply preserved. Nihon shoki took these legendary leaders, tweaked their ages, and then backdated the entire chronology.

144 The beginnings of writing a history Others have argued that these dates were backdated based on the belief in shenxian philosophy, or the Chinese philosophy of “spirits and immortals.” Kobayashi (2011:397), however, argues that the prescribed age of these legendary rulers was not determined by this philosophy. He points out that the rulers before Ōjin only have their ages recorded, which Mishina earlier argued was the older method of recording the lives of the early kings. Historians began recording the number of years of the royal reign around the time of Keitai. This discrepancy in the nature of the data (age versus reign) resulted in the disparity in the death dates of the rulers recorded in Kojiki compared to Nihon shoki. As Kobayashi has indicated, part of the problem appears to be how early Japanese culture interpreted the reigns of their kings versus their actual lives. In Old Japanese, there is the word yo “generation, world,” and this may be etymologically related to yopapi “age.” As Mori (1956:325) suggests, confusion between a reign and a life appears to have come from both characters 世 “generation, world” and 代 “reign, human lifespan,” being read yo in Old Japanese. The Analects record, “The Master said, ‘If there were one who could be King, it would surely be only a generation (世) until everyone was ren’” (Brooks and Brooks 1998:138). In an annotation to this quote, Lunheng notes that “one generation” refers to a 30-year period. This suggests that the ideal royal reign was 30 years in length. There is also the modern theory that the historiographers of the Asuka-Nara era backdated reigns according to the Chinese-inspired belief in the kanoto tori revolution cycle, which supposedly explains why Jinmu’s reign is keyed to a kanoto tori (甲寅) year. In relation to this theory, Kobayashi (2006:228–29) provides a robust argument that the Chinese much earlier had rigorously tried to stamp out this belief, and the knowledge of this philosophy about “a revolution” only came to Japan during the Heian era (794–1185), too late to have been relevant for the compilation of Shoki. He concludes that this ideology is not behind the chronology in Nihon shoki. Satō Hitoshi (1991:3–4) has also argued that the idea of “one revolution” does not fit the thinking of the historiographers of the Asuka or early Nara eras, as it appears in the Heian era and continues until the Edo period. To explain why the kanoto tori year held some special meaning for the early historiographers, Kobayashi (2006:252) begins by reviewing scholarship, which has demonstrated that the early historiographers in Japan relied on the Genka calendar (ca. 602), while later historiographers used the later Giho calendar (676–679). The Genka calendar came to Japan via Paekche sometime in the seventh century, a date marginally supported by the Nihon shoki claim that a Paekche monk named Kanroku brought a calendar in the tenth year of Suiko (602). A number of scholars have demonstrated that the chronology of Jinmu to Ankō was constructed around the Giho calendar, but the record of Yūryaku onward was based on the older Genka calendar.17 As Kobayashi argues, this means that the fundamental issue is that the chronology of Jinmu was manipulated much later, based on the newer calendar, making it impossible to claim that historiographers in the Suiko era or even the Tenmu era would have seen the establishment of Jinmu as having any revolutionary overtones. In order to answer the question why Jinmu begins his conquest of Yamato in the kanoto tori (甲寅) year, Kobayashi (2006:255–58)

The beginnings of writing a history 145 mentions the Chinese calendar from the Warring States (475–221 BCE) period based on the 12-year revolution of Jupiter around the sun. He quotes from the oldest surviving Chinese dictionary-encyclopedia Erya (ca. second century BCE): “The ten bright stages of the sun: The great year 甲 is called efeng (閼逢). … The twelve stages of darkness: The great year 寅 is called shetige (摂提格).” Li Xun (d. 189 CE) added annotation to the text of Erya, and regarding efeng, he said, “The sharp energy of all creation is about to burst forth, but it is prevented from doing so. Thus, this is called efeng.” Shetige refers to Jupiter having a position where there are three stars to the left and right. According to this calendar, kanoto tori is marked as the first year of the cycle. Kobayashi and others argue that the Shoki compilers were following a Chinese idea that the calendar should start in a 甲寅 year, and thus the compilers had Jinmu’s journey to establish himself in Kashiwara begin in that year. The question that naturally follows is why the compilers avoided having Jinmu ascend the throne and begin his reign in this auspicious year. As I have argued (2006:93–98), the compilers of the Jinmu record of Shoki modeled his pre-ascension record on the pre-ascension events of Tenmu, thus indirectly glorifying Tenmu’s bringing peace to the realm. In essence, Tenmu already reigns during the action to quell the disturbance. In this way, Jinmu mirrors Tenmu, who in turn mirrors Jinmu. Major factions at the court believed Tenmu to have been the rightful heir, so his actual date of ascension was less important than the fact that he had taken steps to secure the country. This movement is an act of benevolence to spare the people from further upheaval. Another noteworthy issue is that Jinmu just happens to ascend the throne on the first day of the first month of his reign. Is it coincidence that Jitō also ascended the throne on that very day? How convenient (or symbolic) for the Shoki compilers to have the first and last ruler of the record ascend on this auspicious day. Other rulers who supposedly ascended the throne on the first day of the first month are Ōjin, Richū, and Kenzō. Jitō’s ascension is especially noteworthy, because she delayed her ascension three years after Tenmu’s death, suggesting that the date of her ascension was carefully planned. Hirata (1959:138) believes that Jitō’s actual ascension on the first day of the first month may have been an event that informed the compilers working on the restructuring of the Jinmu record, but he also allows that it is conceivable that there may have been an atmosphere that had Jinmu ascend on the first day of the first month, and then this may have later influenced Empress Jitō to ascend the throne on the first day of the first month. It is well known that the compilers of Nihon shoki have quoted extensively from a variety of Chinese sources, stuffing Chinese quotes into the mouths of Yamato rulers, or padding the thin records of some rulers. However, what is not as clearly understood is the influence some Chinese histories had on the construction of the stories by the Asuka and Nara historiographers. Kobayashi (2006:266) argues that the Shoki compilers found creative license by using events from the stories

146

The beginnings of writing a history

of Kings Wen (文王) and Wu (武王) found in the Zhou record in Shiji. I have listed several events from both reigns and juxtaposed these with those in Shoki to support this theory. Events are listed in order of appearance in Shiji in Table 5.7.18 By inserting these apparently innocuous events in the record, the historiographers were able to pad rather thin storylines. As I argued earlier, the idea of cannibalizing other records for phrases was something Yamato learned from their Paekche teachers. It was a simple extension of this practice at later stages to cannibalize complete events from earlier Chinese records. As the movement to create a new history began in earnest with the edict of Tenmu, there naturally was tension between creating an updated record and not simply retracing the ground of earlier projects. There is a limitation to mining internal evidence in Kojiki and Nihon shoki, but the existence of other evidence is helpful to examine. The existence of Jōgūki offers some hints into what historiography looked like on the cusp of the Tenmu-Jitō movement toward a new history. The manuscripts of Jōgūki have been lost, but it appears from the surviving evidence that this was an early history, much like Kojiki. There are only three surviving quotes from this lost text: two from Shaku Nihongi and one from Shōtoku Taishi heishiden zōkanmon. Based on the order of appearance of quotes found in Shoki, the first quote appears in Book 16 of Shaku Nihongi, in relation to how to read the characters 浮標19 “drift” and 漂蕩 “float and bob,” a line that appears in the earliest part of the mythological story. Jōgūki glosses these as kurage nasu tadayoferi, literally “floating about as a jellyfish.” The second example appears in Book 13 of Shaku Nihongi, where it talks about the origin of Kinmei. Here the elaborate genealogy of Keitai back to Homutu Wake appears. Finally, Shōtoku Taishi heishiden zōkanmon contains a detailed Table 5.7 Events in Shoki taken from Shiji Shiji

Nihon shoki

King Wen: Declines throne because it is a heavy responsibility * King abolishes previous laws about punishing the family for acts by one family member * King quieted the feudal lords and the Barbarians from the Four Quarters were in harmony * Agriculture is the basis of the world * Anciently the court administered laws with banners and boards, for advice and criticism * He did not augment his palace or other buildings or possessions King Wu: Conducts the “Suburban” sacrifice at the Five Altars (五畤)

Nintoku: Declines throne because it is a heavy responsibility Suinin: Stop burying people alive when their master dies. Use clay models in place of people Yūryaku: World is at peace; the Barbarians from the Four Quarters are submissive Sujin: Agriculture is the basis of the world Ingyō: Anciently the court administered laws so that family names were orderly Nintoku: He did not update his palace or buy new robes or clothing Jinmu: Conducts the “Suburban” sacrifice for the heavenly deities and sets up a Festival Altar (霊畤)

The beginnings of writing a history 147 genealogy of Shōtoku Taishi and his sons. This genealogy and that found in Shaku Nihongi are clearly variants when compared to the genealogies preserved in Kojiki and Nihon shoki. In Shaku Nihongi, Book Five, the title Jōgūki is mentioned, but there is no quotation from the record. However, the title appears as evidence within an answer to an interesting question. Focusing on information found in a quote from “another work,” someone asked: “To whom was the title (of 尊 attached to) the august name of Kuni Toko Tati first given? Also, if it was given to a person (as opposed to a deity), where is this title recorded?” Answer: The scholar responds that this title appears in Kana Nihongi, Jōgūki, and a variety of older works; however, there is no mention of who was the first person to receive this title. In the ancient past there is no annotation noting its origin. (Onoda 1986:103) Also, in the beginning section of Yatabe Kinmochi’s Shiki (936 CE), there is this question: Which works should we consult as we prepare to ponder and read this work (Nihon shoki)? The scholar explained that (the audience) should read works such as Sendai kuji hongi, Jōgūki, Kojiki, Yamato hongi (大倭本紀), Kana Nihongi, and others. (Kuroita 1932:190) From this simple overview, one can conclude that Jōguki covered much the same chronological ground as Shoki, starting with the mythical origins of Japan and continuing through the reign of Kinmei, who was the founder of the family from which Shōtoku Taishi was born. It appears to have continued at least as far as the reign of Suiko. This conclusion is inferred from the breadth of questions noted above, as well as the answer that Jōgūki is helpful in preparing the reader to peruse Nihon shoki. Many believe the title Jōgūki “record of the palace of the crown prince” refers to Shōtoku Taishi, since he was known as Prince Jōgū, the Prince of the Crown Prince Palace. However, considering the account in Suishu, where it says that the crown prince was known as Waka Mita Futori, I hesitate to conclude that the title of this lost work refers to Shōtoku Taishi, because several problems tend to be glossed over. First, if the record covers much of the same ground as Nihon shoki and Kojiki, why would it have the equivalent the title of “Shōtoku’s record?” To claim that the record was titled thus because of its authorship is narrowly focused and puts too much stock in the unsubstantiated Suiko claim that Shōtoku Taishi helped compile historical records. Would it not make more sense to claim that this is a record of the ruling family, where the ruler had given the crown prince presiding authority over the compilation of a history? This line of reasoning matches the description in Suishu, where it notes the king goes to court before the sun is up

148

The beginnings of writing a history

and hears a report, but as soon as the sun is up, he transfers power to his younger brother to do the business of the court for the day. My assumption is that “younger brother” refers to the crown prince. If true, then the association of Shōtoku Taishi with this work came later. Takabayashi (1976:131) believes that various stories about heirs to the throne, like Yamato Takeru, are legendary versions of actual rulers. It is plausible that after the reign of Kinmei, the king may have become the figurehead, while the crown prince did the actual day-to-day work. Takabayashi (1976:152) goes on to state that it makes sense to see the crown prince being invested with the actual authority to govern when he is established as heir to the throne. Thus, a ruler did not take the reins to rule when the previous ruler had died but when he became heir to the throne. The existence of Jōgūki thus helps us see a plausible answer to one vexing question: Why does Kojiki end its record with Suiko? A related question is why Kujiki ends its record with the death of Shōtoku Taishi, in the Suiko era? If it is true, as argued above, that Jōgūki ends its record with Shōtoku Taishi and a listing of his genealogy, then we end up with three works that conclude around the same time. As I am arguing, the plausible answer is that the Suiko era was the last full stage of historiography at the Yamato court before the Tenmu-Jitō revision. If recent courts like Jomei, Kōgyoku, and Tenchi were politically charged, many may have deemed it dangerous to engage in historiography beyond the period we know as Suiko. One must remember that there were many alive in Tenmu’s era who had firsthand knowledge about events in previous reigns. Add to this the uncertainty the court endured when Tenmu died. It is difficult to know what Jitō intended once she took the reins of government and waited a sufficient amount of time to mourn Tenmu’s passing. All we have is the record in Nihon shoki, but there are enough problems to make one cautious about writing or rewriting the history, especially when the half-century before Tenmu appears to have been contentious. Finally, the hagiography that occurred in relation to Shōtoku Taishi may have been an attempt to smooth over the fact that the Soga faction assassinated Prince Yamashiro, Shōtoku’s son, and the rightful heir to the throne. By elevating Shōtoku to the level of a Buddhist sage, and connecting him to the historiographical movement, Prince Yamashiro’s assassination was open to revision, allowing Asuka historians to make him a Confucian tragedy instead of a victim of a bloody coup. This is an attractive narrative, but I believe there is more going on here, which I address in the next chapter.

Notes 1 Tsuda (1966:247) concluded that there was a dividing line between the reign of Chūai and Ōjin. 2 Thus, 55 (51111), 66 (2131111), and 95 (51191). 3 Kobayashi (2006:52) sees the change from the Yūryaku dynasty to the Keitai dynasty as being motivated mainly by economic concerns. 4 Two different versions of his name exist. Nihon shoki has the core of his name as Haraki Hiro, and even attaches a reading to make sure people read it correctly, probably

The beginnings of writing a history 149 5 6

7 8

9

10 11

12 13 14 15

16 17 18 19

in an attempt to discard the traditional reading of paruki piro, found in Kojiki, Jōgū teisetsu, and the Gangōji origin record. It is possible that Hodo (podo) means “degree, interval”; thus his father was the “great interval,” and Keitai was the “minor (second) interval.” Hikoyu Musumi is noted as the eldest son of Kaika but based on the element musumi (a variation of musubi) in this name; I argue that this was originally the name of a deity later changed into a person who is the older brother of Sujin. I theorize that originally this deity and Sujin were connected as father–son. Musumi is a dialect variant, demonstrating the same alternation also found in OJ “snake” pemi versus pebi, or still found in modern Japanese “lonely” samisi versus sabisi. My assertion here is that Homutu Wake and Homuda Wake are not the same name, nor dialectal variations of the same name. These are simply look-alikes, just as Dan and Don are superficially similar but unrelated in English: Dan “judge” being Hebrew in origin, and Don originating from the Celtic name Donald “proud chief.” Takabayashi (1976:125–27) argues that many earlier records had two competing systems: (1) remembering the length of each royal reign, and (2) important events keyed to dates in the sexagenary calendar. These could then be tied together with the label “in the reign of King X.” Park (2018:169) believes an alliance of Kŭmgwan Kaya (金官伽倻, our Kimkan Kara) with a Wa group based in Saki (佐紀) banded together to invade Koguryŏ in the fourth century. Tsuda (1966:23–24) believed Tennōki to be a record based on the Chinese benji template, while Kokki was similar to Chinese biographies. Umezawa (1988:35–50) argued that Tennōki was essentially Teiki as found in the Kojiki preface, while Kokki later came to be known as Kuji or Honji. See Chapter 6 for an explanation of this process. One other possibility is that this is not actually a Yamato record, but an account from Paekche, where Paekche claims to have subdued Koguryŏ, Silla, and Kara. In the chart, NA means the information is not available (not recorded). Bolded S means Kujiki has the same name as Nihon shoki. Kujiki data with a † come from Chapter 10 “Kokuzō Hongi.” Kishimoto (2013:15) argues, based on groups of large keyhole tombs that are contemporaneous but have different structures, that Wa/Yamato was ruled over by dual kings, one who controlled the sacred (ritualistic) and the other the secular (administrative) realm. He believes that Ichinohe inherited the administrative throne at the death of Hanzei (Mizuha Wake) and then was assassinated by Yūryaku. He thus believes Ichinohe actually reigned. Kobayashi (2011:397) argues that the recording of the age of each ruler was the older method, later superseded by the newer method of recording the number of years of the king’s reign. See Uchida Masao (1993), Ogawa Kiyohiko (1997), and Arisaka Takamichi (1999). Synopses of events are slightly altered forms from Nienhausen (1994, volume 2). The original should be 浮漂 “to drift on the water,” but Shaku Nihongi actually has 浮標 “a buoy.” This is clearly a scribal mistake.

References All works are published in Tokyo unless otherwise noted. Arisaka, Takamichi. 1999. Kodaishi o toku kagi. Tokyo: Mainichi Shinbunsha. Barnes, Gina. 2012. “The Emergence of Political Rulership and the State in Early Japan.” In Karl Friday, ed. Japan Emerging: Premodern History to 1850. Boulder, CO: Westview Press, pp. 77–88.

150

The beginnings of writing a history

Bentley, John R. 2002. Historiographical Trends in Early Japan. Lewiston: Edwin Mellen Press. Bentley, John R. 2006. The Authenticity of Sendai kuji hongi: A New Examination of Texts, with a Translation and Commentary. Leiden: Brill. Bentley, John R. 2013. Tamakatsuma: A Window into the Scholarship of Motoori Norinaga. Ithaca: Cornell East Asia Series. Brooks, E Bruce and A Taeko Brooks. 1998. The Original Analects: Sayings of Confucius and His Successors. New York: Columbia University Press. Enoki, Eiichi. 1975. “Suiko-chō no ‘Kokki’ ni tsuite.” Nihonshi ronsō, 5: 121–126. Hirata, Toshiharu. 1959. Nihon koten no seiritsu no kenkyū. Nihon Shoin. Iwasaki, Takuya. 1990. Kofun no jidai. Kyōikusha. Kimura, Ryūshi. 1992. “Kojiki sakubun ishiki no ichi sokumen—shinjō o byōsha suru dōshi kara mita baai.” In Nakamura, Hirotoshi, Aoki Shūhei, Sugano Masao, and Yamazaki Masayuki, eds. Umezawa Isezō Sensei tsuitō kiki ronshū. Zoku Gunsho Ruijū Kanseikai, pp. 217–232. Kishimoto, Naofumi. 2013. “Dual Kingship in the Kofun Period as Seen from the Keyhole Tombs.” Urban Scope, 4: 1–21. Translated by Joseph Ryan. Kobayashi, Toshio. 2006. Nihon kodai kokka keiseishi-kō. Kōsō Shobō. Kobayashi, Toshio. 2011. “Tsuda shigaku to kodaishi.” In Shinkawa Tokio and Hayakawa Mannen, eds. Shiryō toshite no Nihon shoki—Tsuda Sōkichi o yominaosu. Benseisha Shuppan, pp. 374–405. Kojima, Noriyuki, Kōjirō Naoki, Kazutami Nishimiya, Susumu Kuranak and Masamori Mōri, eds. 2002. Nihon shoki. Shōgakkan. Kuroita, Katsumi, ed. 1932. Nihongi shiki. Vol. 8 of Kokushi Taikei. Yoshikawa Kōbunkan. Lurie, David B. 2011. Realms of Literacy: Early Japan and the History of Writing. Cambridge: Harvard University Asia Center. Mishina, Shōei. 1948. “Kinen shinkō.” Contained in Naka Michiyo and Mishina Shōei, Zōho jōsei nenki-kō. Nara: Yōtokusha, pp. 189–246. Miyake, Marc Hideo. 2003. “Philological evidence for *e and *o in Pre-Old Japanese.” Diachronica, 20(1): 83–137. Mizuno, Yū. 1968. Nihon kokka no seiritsu. Kodansha. Mori, Kiyohito. 1956. Nihon kinen no kenkyū. Tokyo: Shōchoku Kōkyūjo. Nienhausen, William H, ed. 1994. The Grand Scribe’s Record. Bloomington: Indiana University Press. Ogawa, Kiyohiko. 1997. Kotenmon rekijitsu no kenkyū: tenmongaku de toku rekishi no nazo. Tokyo: Kōseisha. Ōishi, Yoshiki. 1975. Nihon ōken no seiritsu. Hanawa Shobō. Onoda, Mitsuo, ed. 1986. Shaku Nihongi. Vol. 5 of Koten Chūshaku-hen of Shintō Taikei. Shintō Taikei Hensankai. Park, Cheun Soo. 2018. “Kaya, Silla, and Wa: Relationships and Their Historical Backgrounds.” In Mark E Byington, Ken’ichi Sasaki and Martin T Bale, eds. Early Korea-Japan Interactions. Early Korea Project, Korea Institute. Cambridge, MA: Harvard University, pp. 133–181. Robinson, GW. 1955. “The Kuji Hongi: Volumes 7, 8, and 9 Considered as a Draft of the Nihon Shoki.” The Memoirs of the Toyo Bunko, 14: 81–138. Sakamoto, Tarō. 1979. Shōtoku Taishi. Yoshikawa Kōbunkan. Satō, Hitoshi. 1991. Kakumei-kakurei kanmon to kaigen no kenkyū. Tokyo: Satō Hitoshi Chosakushū Kankōkai.

The beginnings of writing a history 151 Takabayashi, Miyuki. 1976. Nihon shoki hennenkō. Contained in Vol. 9 of NSK. Hanawa Shobō, pp. 91–166. Takeuchi, Rizō. 1962. Kodai jinmei jiten. Yoshikawa Kōbunkan. Tsuda, Sōkichi. 1948. Nihon koten no kenkyū, 2 Volus. Iwanami Shoten. Tsuda, Sōkichi . 1966. Kojiki oyobi Nihon shoki no shin kenkyū. In Separate Volume One of Tsuda Sōkichi zenshū. Iwanami Shoten. Uchida, Masao. 1993. Nihon rekijitsu genten. Tokyo: Yūzankaku Shuppan. Umezawa, Isezō. 1968. Kiki hihan: Kojiki oyobi Nihon shoki no seiritsu ni kansuru kenkyū. Sōbunsha. Umezawa, Isezō. 1976. Zoku Kiki hihan: Kojiki oyobi Nihon shoki no bunkenteki sōgo kankei no kyūmei. Sōbunsha. Umezawa, Isezō. 1988. Kojiki to Nihon shoki no seiritsu. Yoshikawa Kōbunkan. Wang, Haicheng. 2014. Writing and the Ancient State: Early China in Comparative Perspective. New York: Cambridge University Press. Weiner, ESC and JA Simpson, eds. 1971. The Compact Edition of the Oxford English Dictionary. Oxford: Clarendon Press. Woo, Jae-Pyoung. 2018. “Interactions between Paekche and Wa in the Third to Sixth Centuries A.D. Based on Patterns of Trade and Exchange.” In Mark E Byington, Ken’ichi Sasaki and Martin T Bale, eds. Early Korea-Japan Interactions. Early Korea Project, Korea Institute. Cambridge, MA: Harvard University, pp. 183–230.

6

The Tenmu-Jitō historiographical project

A new historiographical project There are various pieces of evidence pointing to the inception of a new history during the reigns of Tenmu and Jitō, what I call the “Tenmu Historiographical Project.”1 I use this label in a broad sense, meaning the impetus that resulted in our extant historical records, including Kojiki, Kujiki, and Nihon shoki. Two sources of evidence for this project appear in the preface of Kojiki and the tenth-year edict of Tenmu found in Nihon shoki. The Tenmu edict of the third month of the tenth year lists the following individuals who were involved in the project, equally divided into two groups: princes and representatives of powerful families. There are six princes: Kawashima, Osakabe, Hirose, Takeda, Kuwata, and Mino. The six representatives are (minus their ranks) Kamitsukeno Michiji, Imibe Muraji Kobito, Azumi Muraji Inashiki, Naniwa Kimi Ōkata, Nakatomi Muraji Ōsima, and Heguri Omi Kobito. On the surface, it is reasonable to assume that the six princes would be assigned to work on matters related to the ruling family, including mythology, genealogy, and the record of events surrounding the divine ruler. The other six representatives likely had charge over legends and events connected to powerful families and their locales. Whether we believe that Tennōki and Kokki or Teiki and Kuji were actual titles or simply rubrics in the historiographical fabric, a clear dichotomy is preserved in both cases: the recording of history surrounding the royal family, and that of the powerful families at court who support the ruling family.2 As the Kojiki preface claims to contain an edict from Tenmu, it is helpful to return to that passage and carefully examine it. Lurie (2001:270) argues that the preface is a construct put forth by Yasumaro, an idea I agree with in principle, but it is difficult to imagine that the text was created out of whole cloth.3 There must have been a preexisting narrative upon which Yasumaro based his account in the preface. That narrative came from either Jitō or Genmei, or it may have come from both, Jitō having passed it on to Genmei. Remember that Yasumaro received the command to finish the work of compiling Kojiki from Genmei (r. 707–715), who was the younger sister of Jitō by a different mother. While Jitō and Genmei had different mothers, it is important to remember that the mothers of both Jitō and Genmei were of Soga lineage.

The Tenmu-Jitō historiographical project 153 The putative edict from Tenmu in the preface is as follows: Thereupon the Heavenly Ruler declared, “From what I hear (the texts of) Teiki and Honji in possession of the various families already stray from truth and fact, and include many falsehoods and fabrications. If at this time we do not correct these errors, before many years have passed the significance [of our traditions] will be lost. These [two works] are the warp and the weft [of the fabric] of the state, the great foundation of the royal influence. Thus, I want to have Teiki selected and recorded, and get to the bottom of the errors in Kuji, eradicating the errors and establishing the truth, and leave these for our posterity” [於是天皇詔之。朕聞諸家之所賷。帝紀及 本 辭 。既 違 正 實 。多 加 虚 僞 。當 今 之 時 。不 改 其 失 。未 經 幾年。其旨欲滅。斯乃邦家經緯。王化之鴻基焉。故惟撰録帝紀。討 覈舊辭。削僞定實。欲流後葉。]. To me the key phrase is 朕聞諸家之所賷。帝紀及本辭 “From what I hear (the texts of) Teiki and Honji in possession of the various families….” The verb 賷 is a variant graph for 齎, and means “possess, present,” but also has the meaning of lamentation. It is interesting that Yasumaro selected this bi-semantic character when he could have used the more transparent graphs 有 or 持. Therefore, I conclude that the character 賷 does double duty: the families presented to court records in their possession, making the connection that likely these records were family histories outlining events of members of each family set within the framework of a specific number of kings. However, by using this multilayered verb, Yasumaro can also give the reader the feeling of lamentation from Tenmu, who is saying, “Ah, these records in possession of these families have strayed from the truth.” While it is speculation, a careful reading of Nihon shoki provides the following clues. Kinmei marries Soga Kitashi Hime, starting a tradition where a number of kings and queens have mothers of Soga lineage. The traditional families, Nakatomi and Mononobe, tried to stand up to Soga abuses of power but were afflicted by this formidable family. In 587 CE, a pivotal battle pitted Mononobe Moriya against Soga Umako (d. 626).4 Umako’s forces won and the Mononobe lost their influence at court. They regrouped under the name of Isonokami, granted by the Tenmu court, almost exactly a century after their defeat at the hands of Umako. The Nakatomi sided with the Mononobe when Soga no Iname invited Buddhism to be a force of influence at court. Mononobe Okoshi and Nakatomi Kamako responded, “The kings of our country have always worshipped the 180 heavenly and earthly deities, venerating them in the spring, summer, autumn, and winter. If we turned to worshipping foreign gods, we fear this will incite the native deities to anger.” The court ignores their pleas, but later a plague sweeps through the country, vindicating the Mononobe and Nakatomi, and Kinmei orders the suppression of Buddhism. A few reigns later, Shoki records the following when Yōmei was on the verge of passing away:

154 The Tenmu-Jitō historiographical project Second year, summer, fourth month, second day. The court held the festival of the Great Thanksgiving in Kahakami in Ihare. On this day, the Heavenly Sovereign (Yōmei) fell ill and returned to the palace. The various ministers attended him. The Heavenly Sovereign commanded the various ministers, saying, “We wish to show our devotion to the Three Treasures (Buddhism). Discuss the issue.” The ministers entered the court and discussed the idea. Mononobe Moriya Ōmuraji and Nakatomi Katsumi Muraji were against the notion, saying, “Why should we abandon the deities of our land and pay homage to foreign gods? We have never done such a thing before.” Great Minister Soga Umako Sukune retorted, “We should follow the Imperial decree and lend our assistance. Who dares oppose the Imperial will?” At this time, the younger brother Imperial Prince … entered the inner precincts of the palace, leading the monk of Toyo Province. … Moriya looked askance at this, filled with anger. Oshizakabe Fubito Kekuso ran in and whispered to Moriya, “The various ministers are presently plotting against you, and are about to cut off your retreat.” Hearing this, Moriya immediately retired to Ato … and gathered men together. Nakatomi Katsumi also gathered an army at his residence in an attempt to lend assistance to Moriya. He crafted statues of the Heir to the throne, Prince Hikohito and Prince Takeda, and then cursed the statues, and prayed for their death. Little time had passed when Katsumi realized that it was impossible to succeed, and he retreated to the Mimata Palace. Again, the Mononobe and Nakatomi are standing up to the Soga, but Umako is clearly threatening these two families, and plotting their destruction behind their backs. Moriya is defeated by Umako’s army, and it appears that the Nakatomi escaped destruction, but why is not recorded. What may be closer to the truth is that the Nakatomi also suffered at the hands of the Soga, but one of the compilers of Nihon shoki avoided any clear mention of this, because when the record was finalized, Minister Fujiwara Fubito (659–720), whose father was Nakatomi Kamatari, was in power. Here a simple dichotomy is evident: the Mononobe and Nakatomi represent a more conservative, traditional group at court. The Soga represent a progressive, continental group. I earlier mentioned that Tenmu may have been worried about the records in possession of the Fune Obito, as it may have contained a revision that would allow the Soga to compete for the throne. On the other hand, it is also plausible that two of the family records Tenmu singled out were in possession of the Nakatomi and Mononobe, containing detailed accounts of atrocities by the Soga. Tenmu may have been perturbed to learn that some family histories depicted the Soga in a disgraceful light. While the historical record as we have it demonstrates that the Soga were ruthless, cunning, and power hungry, Tenmu and Jitō may have desired that the historiographical project they were pushing gloss over these acts (or at least strenuously massage the facts). Interestingly, in relation to the

The Tenmu-Jitō historiographical project 155 assassination of Sushun, Kujiki’s record, compiled under the auspices of Minister of the Left Isonokami Maro, contains the harshest criticism of the Soga, making Umako the driver of the assassination, while Kojiki and Nihon shoki try to shift the blame by placing more of the responsibility on Sushun and his careless words.5 This theory helps explain why Kojiki switches from “poem infused story” to simple genealogy after Kenzō. I would argue that Yasumaro is taking the politically safe route by completely ignoring the historical narrative after Kenzō, effectively robbing both the Soga and their enemies of any voice. Notice that only three Soga names appear in Kojiki: Soga Ishikawa Omi, the founder of the Soga family, Soga Iname, and Prince Soga Kura, these last two appearing in the genealogy of Kinmei. Umako does not appear in Kojiki. In contrast, the Mononobe are mentioned only twice and the Nakatomi only once. The impetus for a new project appears to have been at least twofold: a desire to codify the lineage of the royal family (Ōishi 1975:96) and Tenmu and Jitō’s desire to modernize Yamato’s government to deal with the unfolding events on the continent. In his sixth year (667), Tenji, the previous ruler, moved the capital inland to protect against a potential invasion from the continent, spurred on by the destruction of Paekche at the hands of the Tang-Silla alliance (cf. Kojima et al. 2002, 3:271, n10). Tenmu and Jitō were wise enough to push modernization and domestic reforms to deal with the rapidly unfolding events across the straits. Not long after Tenmu ascended the throne in 673, he began to consolidate power. Having put down the rebellion by Prince Ōtomo (named Kōbun in 1870), he began to initiate a radical change at court, masterfully straddling two separate camps: those allied to Buddhism and those allied to the native religion. Table 6.1 lists the dates of changes Tenmu initiated, according to Nihon shoki’s record. The tenth year of Tenmu’s reign is critical. Within less than a year, Tenmu moves to have the government adopt a system based on the Tang Codes, and then issues an edict setting in motion a revision of the history of the state. A month after this edict, he releases 92 articles of prohibition in relation to what people could and could not wear, and what they could not use as decoration. Finally, he reshuffles the deck of the courtly families, issuing new titles, promoting some families, and demoting others, based on the loyalty shown during the Jinshin Disturbance.

Table 6.1 Changes Tenmu initiated Date

Regnal year Event

674–75 676 678 681 681 681

3rd year 5th year 7th year 10th year 10th year 10th year

Encourages change toward a code-based government The court will revere both the native gods and Buddhism Instructions given on criteria for promotion of officials Compile a new set of statutes, and the edict to compile a history 92 articles of prohibition announced New titles given to powerful men and their families

156

The Tenmu-Jitō historiographical project

It is helpful to examine the six names of the nonprincely people appointed to work on the new historiographical project: (i) Kamitsukeno Michiji, from a powerful chieftain family based originally in eastern Japan. (ii) Imibe Kobito, from a family in charge of some liturgies. This family appears to have had access to some technology related to writing. (iii) Azumi Inashiki, a family originally based in Kyūshū, which oversaw a shrine that worshipped the deities of the ocean. They also supplied seafood for the royal table. This family may have had a connection with the peninsula. (iv) Naniwa Ōkata; he was originally named Kusakabe, and his ancestors had the title of Kishi, demonstrating a foreign origin for this family.6 (v) Nakatomi Ōshima, a family that was also in charge of liturgies, specifically surrounding the ritual of purification. This family also likely had access to some technology related to writing. (vi) Heguri Kobito, from a powerful family with close ties to the royal family. We can conclude that these six men were chosen because they were judged to be loyal subjects with the requisite knowledge. It also seems that most, if not all, had some training in documents and the technology of writing. Shoki goes on to record, “Ōshima and Kobito took up brushes themselves and recorded things.” Comparing the two edicts (in Kojiki and Nihon shoki), it is evident that the preface in Kojiki is less sophisticated in its details of the events, other than the purported words of Tenmu. Regarding Kojiki’s preface, Sugano (2012:17–18) identifies four problematic points. The first is Yasumaro’s glossing over who alerted Tenmu to the fact that these family records had strayed from the truth. The assumption here is that families had submitted their records for inspection by the court. Why would it be in the shadows? It should have been public knowledge who had made this accusation. The second problem is that the people who submitted their records are naturally from families of prominence. Why does the court not rectify these problems at this stage instead of publicly announcing that falsehoods exist? The third issue is the basis for this representative concluding that these records had deviated from the truth. The final issue is what standard or touchstone was used to discern between truth and error. If the standard was simply the version of Teiki and Honji in possession of the court judged against that of the various households, then there would have been no need to correct the errors. The court could have simply made copies of their texts and destroyed the copies the families had in their possession. Sugano concludes that this option was not available to Tenmu’s court, leading to the conclusion that the court must not have had a copy of either of these works in their possession. Thus, they had been forced so far to rely on families that not only had compiled the records but also kept these in their possession. This suggests that over time the earlier iteration of the Bureau of History became a hereditary fiefdom for a few families with peninsular connections. The suggestion thus becomes that Tenmu is actually trying to bring the Bureau of History under the umbrella of the government.

The Tenmu-Jitō historiographical project 157 Sugano (2012:24) believes that the Tenmu edict quoted in Kojiki is legitimate. He argues that it is difficult to believe that the court would have turned a blind eye to a document that contained a fabricated edict from Tenmu. He goes on to conclude that this edict originally was not connected to the work of Kojiki, but Yasumaro created a framework with the edict front and center in the preface to give the editorial work of Kojiki greater authority. Sugano quotes a theory from Sakamoto Tarō that Tenmu originally commanded Hieda no Are to put together a history, but because the project did not come to fruition, Tenmu issued the edict of 681 to the committee of six princes and six courtiers. In the end, Sugano believes the edict in the preface of Kojiki to actually point to the action of the tenth year of Tenmu (681). I agree that the edict in Kojiki does not fit well with the parameters of Kojiki, thus allowing both Lurie and Sugano to be correct in principle. However, admitting this point does not water down the fact that a historiographical project began under the watchful eye of the Tenmu court. What is open to debate is what took place. I agree with Sugano that there was a tradition at court that families had possession of a variety of records. There must have been several families, originally with connections to the peninsula that kept the official history, the source documents dating back to the days of Yūryaku, and Keitai, and then Kinmei. After the revision in the Suiko-Jomei era, it is possible that this group of families added a condensed chapter of history with the passing of each king or queen. Each record of the previously deceased ruler was known as Teiki. This record was straightforward and transparent, a work much like the diurnal records of the Chinese court. Instead of being written and compiled by diarists as in China, the Yamato court records were written by literate servants of the court who were eyewitnesses. The problem, as I see it, was in the records of the families, and the unavoidable temptation to embellish or fabricate events and achievements. Consider the following plausible entry from a Mononobe record that Tenmu and Jitō would have found threatening. In the reign of Great King Hatsube (Sushun), Futsu Hime, the younger sister of Moriya, was made a concubine of the Great King. In the fifth year, a mountain boar was presented to the Great King, who pointed to the animal and said, “When will those whom I loathe be cut down like the throat of this boar?” More weapons were prepared and readied than usual. Koteko, the daughter of the Ōtomo, was frustrated with the decline in affection showed to her. She thus sent a servant to Soga Umako Sukune and informed him, “Recently, we have heard that someone presented a wild boar to the Great King. The Great King pointed to the boar and said, ‘When will the person we have in mind be cut down as the throat of this boar?’ Thus, weapons are being prepared in great numbers in the palace.” Hearing this report caused Umako Sukune great bewilderment, so he gathered those close to him to plan the assassination of the Great King. With this, they held counsel regarding who should ascend the throne after the removal of the current king.

158

The Tenmu-Jitō historiographical project

The court would have suppressed an account like this, which directly accuses the Soga of assassinating a sitting ruler. This account also suggests that a powerful family with no authority to the throne is determining who should reign. This kind of censor-centered historiography is typical of many regimes, glossing over, ignoring, or even suppressing events that cast the ruler or a powerful minister in a poor light. The more expansive revision executed on the older histories, as I see it, appears to have been conducted under the umbrella of the Jitō court. The first action was to erect a tribute to the power held by Jitō and her grandmother, Kōgyoku-Saimei.7 If Tenmu was to be portrayed as a reincarnation of Jinmu, then it was only fair to have Jitō mirror a valiant, powerful queen. It is certain that the historians at court knew about an ancient tradition of queens ruling over the state. They knew from Chinese records of accounts about Himiko and Toyo (Iyo). By the Suiko period, names like 卑弥呼 and 台与 would have been read close to our modern readings. Jitō likely wanted the new history to have a precedent with a ruling queen in the past. To create a history of precedence, the Jitō court wanted Jingū to have actually sat on the throne, and to be the “female” ruler who led the invasion of the peninsula with its attack on the villain of Jitō’s era, Silla. Rewriting the story so that Jingū became the “agent of action” alone was insufficient, however. Jitō and her court wanted precedence for a concrete connection with heaven, and an invasion of the peninsula needed more than just a shrewd leader of matrilineal influence; it also required divine intervention. Thus, the story depicts Jingū receiving instructions from a number of deities in relation to the target of her attack. Nihon shoki notes that she communicated with deities from Ise (where Amaterasu was enshrined), most notably the deity Koto Shiro Nushi. Kojiki records that Jingū lined up the ships of her navy stem to stern, and they headed to the peninsula. The account portrays the king of Silla filled with trepidation at the awesome sight of ships filling the sea, and he pledges unending allegiance to the Yamato court. Interestingly, like the dramas of the subjugation of the Kumaso and Emishi, there is no record of any actual fighting. Cultured Yamato rulers vanquish foes through intelligence, not might. This depiction was not wholesale creation but creative manipulation of an older legend. Kujiki abbreviates the details of the invasion, sending the reader to a now lost title, The Subjugation of the Three Han, and one would believe that a volume titled such would have had more information than the terse account found in Kojiki. Kojiki does make passing mention of Silla capitulating and being labeled a land for raising horses, while Paekche is to become a government office. Thus, at least on the surface, two of three Han are mentioned. Many consider the whole episode to be a fabrication, but it is unwise to ignore the following variant quote in Nihon shoki: Then the Heavenly Sovereign [Chūai] heard what had taken place, and his indignation was great, and he put together a large army and decided to obliterate Silla. With this, the Imperial navy reached Silla, and the ships filled the ocean. Seeing this, the people of Silla were so terrified that they knew not

The Tenmu-Jitō historiographical project 159 what to do. They counseled with each other and put the deceased king’s wife to death and offered their apologies for crimes committed [於是、天皇聞之 、重發震忿、大起軍衆、欲頓滅新羅。是以、軍船滿海而詣之。是時 、新羅國人悉懼、不知所如。則相集共議之、殺王妻以謝罪]. This variant quote provides evidence that the original story had 大王 “great king” as the agent of the invasion of Silla, and then later compilers mechanically changed this to 天皇 “divine ruler,” a genderless designation; however, this cannot refer to Jingū, as she is never called 天皇, consistently being labeled 皇后 “empress.” This story apparently was based on a fragmentary knowledge of the ancient attacks on Koguryŏ by the Wa, as portrayed on the Kwanggaet’o stele on the Yalu River. I postulate that the Jitō court wanted to elevate Jingū to the status of a powerful ruler, perhaps to indirectly praise Jitō’s assistance to Tenmu during the Jinshin Disturbance. This revision required several editorial moves: Jingū’s husband (the great king) would need to die prematurely and leave the country to her without an heir, precisely mirroring what Tenmu did to Jitō.8 Thus, the story was rewritten so that Jingū would miraculously become pregnant with Ōjin and not give birth until she returned to Kyūshū. The country she would invade was altered from Koguryŏ to Silla, since Silla was the villain in Jitō’s day, having destroyed Paekche in 660 and Koguryŏ in 668, effectively blocking the more secure land route to China. There is a tendency among modern readers to believe that Jingū was modeled on Himiko, queen of Yamatai, but as she is not described in Weizhi as a military figure,9 this comparison does not hold up to closer scrutiny. The addition of the notes from Weizhi in the Jingū record likely was done at the final stage of compilation of Nihon shoki, and not at the initial stage in the Tenmu-Jitō era. Rather than making the connection between Himiko and Jingū, the inclusion of these quotes simply places Jingū’s reign within a Chinese chronological framework. This allows the association with the Wa attacks on Koguryŏ to better align with Jingū’s timeline, permitting this editorial sleight of hand of altering Koguryŏ to Silla. Hirata (1959:162) earlier argued that the record of Jingū was formulated by the Shoki compilers, saying that the record as found in Kojiki—where Jingū is part of the Chūai story—is the older version. However, there is more going on here than meets the eye. What should not be overlooked is the fact that, as Tanaka (1972:404–411, 421) has demonstrated, over 80% of the details related to Chūai and Jingū in Kojiki appear in the Shoki records of the same rulers. Two other pieces of information in the Chūai record in Kojiki appear in the Ōjin record. The only piece of information in Kojiki not found in Shoki is the fact that Jingū died at the age of 100 and was buried in the Tatanami Tumulus in Saki. Thus, Shoki’s account is clearly a reworking and an expansion of the Kojiki story. Kobayashi (2006:278–80) notes that Jingū’s record is two separate accounts fused into one. From the beginning of her pre-ascension record until the 13th year of her reign is the world of “Okinaga Tarashi Hime.” After that year, nothing is recorded for a quarter of a century until the 39th year. Then, the next 30 years of

160

The Tenmu-Jitō historiographical project

her record is populated with accounts from the peninsula and continent, which Kobayashi labels “the Queen of Wa” account.10 He further observes that while the majority of Nihon shoki is based around a regnal chronology, starting with the year the king ascended the throne, the chronology of Jingū is different, mainly because she never ascends the throne. Interestingly, this variance is also found in the early reigns of Jinmu and Suizei. The designation of the start of a reign is 太歳 “grand year.” This designation appears seven years before Jinmu ascends the throne, likely showing that the historians considered him to have the authority to sweep the rebellious people aside and unify the country under the authority of the sun goddess. Suizei’s designation appears three years after Jinmu’s death, but one year before he ascended the throne. Nihon shoki’s record makes it clear that Tagishi Mimi, half-brother to Suizei, had control of the government. During this one-year interval, Suizei puts his rebellious brother to death; so it appears that the compilers wanted it made clear that the rightful heir had the authority to take this kind of drastic action. Kobayashi (2006:280) believes that this designation was also used to punctuate the end of an era. The final example is Jingū, where it is strange that the designation (太歳) appears in the first year of her regency for her unborn son, Ōjin, following the precedent set earlier in Shoki; however, the designation appears again in the 39th year. It then appears again in the year of her death, the 69th year. The second example appears in the year when the first quotation from Weizhi is found, and these quotes, as well as the diurnal diary quote, span a 28-year period, which covers the lives of both Himiko and Toyo. Based on this evidence, we must conclude that two different records were originally intended, but later dovetailed together, and the compiler forgot to eliminate the second and third “great year” designation. Kobayashi (2006:280–81) makes a compelling case that the compilers did not consider Jingū to be a mirror image of Himiko. In fact, he says, “I understand this to show that the compilers of Shoki viewed the Yamato court as a separate entity from Yamatai-Himiko, and that they did not consider Queen Jingū to be Himiko.” This aligns well with the quote from Shiki, where the early Heian court clearly had a tradition that there were two geographically different lands labeled “Wa.” Tanaka (1972:445) outlines a three-step process for the creation of the Jingū record in Shoki: (i) the original account and her death date; (ii) the addition of information from the peninsula, with quotes from primary and secondary sources; (iii) the final step of adding information from Chinese records that match the chronology of the reconstructed Jingū calendar, and the addition of three “great year” designations. I would argue for the addition of one more step, added before any of the others. That is the change of agent in the legend, swapping out a male ruler and replacing him with Jingū. This step was part of the edict given to both Hieda no Are and the Shoki compilation committee. The only evidence that a male ruler had been the original actor in the legend is preserved in a small variant quote in Shoki, which might have been preserved as a form of resistance against the movement to revise the history.

The Tenmu-Jitō historiographical project 161

Putting Suiko on the Throne A bolder move by Jitō’s historiographical group concerns Suiko. As noted in Chapter 4, the Chinese description in Suishu of the court in 607 fails to make mention, even once, of a female sovereign—the account describes a court with a male ruler and a male crown prince. It is difficult to accept the theories of some that the Suishu account is inaccurate or muddled upon realizing that three Japanese words/ titles are quite accurately represented: 阿毎 ame “heaven,” 阿輩鶏彌 opokimi “great ruler,” and 鶏彌 kimi “ruler.” Furthermore, the 12-grade cap system supposedly implemented at court in 603 is also rather accurately described. It is true that the order of the 5th to the 12th caps are inverted, but the character usage is accurate. This begs the question, how can the record be accurate concerning terms relating to the royal family and the court system, and get the actual ruler wrong? Suishu’s account of the royal family is difficult to square with the record in Shoki, as 607 should be the 15th year of Suiko’s reign. Some have theorized that the male ruler mentioned in Suishu points to Shōtoku Taishi (cf. Hanzawa 2002:303–05), but as noted in Chapter 4, the appellation Tarasi piko is very close to the name for Jomei, whose name is recorded as Okinaga Tarashihi Hironuka, making it difficult to believe this is accidental. The middle part, tarasipi, is strange, and I believe it is a shortened (or garbled) form of tarasi piko. If this is the ruler that the Chinese wrote about, he would have only been 14 years old in 607, according to Nihon shoki’s chronology. Two explanations are possible here: (i) the chronology has been manipulated, making it corrupt; or (ii) Jomei was a young ruler, operating under the shadow of Soga Umako.11 We should not put too much stock in the chronology, as it clearly has been manipulated more than once. One possible scenario is to see Jitō’s historiographers creating a figurative triangular relationship, elevating Suiko, a queen of a sovereign who was widowed at his death, to a position of governing, mirroring Jingū, who reflects Jitō. If this occurred, then the manipulation given in Figure 6.1 became necessary.12 It is plausible, looking at the story in Nihon shoki, that the original history had Umayado (Shōtoku Taishi) ascend the throne, but Umako assassinated him a few years later. This then allowed Umako to put a young Jomei on the throne (who would have been about 14 or 15 at the time). None of this appears in the Suishu, because Jomei was already on the throne when our Chinese eyewitness, Bei Qing, came to court. It would also account for the description that the king came to court early, and when the sun came up, he retired to allow his (older) brother take the Original Bidatsu = Suiko Ymei Sushun Umayado Jomei = K˜gyoku Kgyoku

Manipulated Bidatsu = Suiko Ymei Sushun Suiko Jomei = K˜gyoku Kgyoku

Figure 6.1 Putting Suiko on the throne.

162 The Tenmu-Jitō historiographical project reins. Flipping the “older brother” to “younger brother” would have been done to save face. The Jitō historians may have seen a double benefit in revising the history: rectifying several tragedies and placing Suiko on the throne. In this way, Shōtoku Taishi could rule, though as a regent, but still die an early death allowing his elevation through Buddhism. This act of rewriting also spared Soga Umako the portrayal as being the murderer of members of the royal family.13 A final editorial act would have been to rewrite the actual story about the assassination of Shōtoku Taishi, putting his son, Prince Yamashiro, in his place. The historians then would have adjusted the chronology, and instead of Umako, it would become Soga Iruka (d. 645). Instead of a brutal assassination, the historians wrote the story where Prince Yamashiro and his family committed suicide when they realized they had no way out. Ultimately, instead of the story of Shōtoku Taishi standing up to the brutal, abuse of power by Umako and paying for it with his life, the story became that of Prince Yamashiro standing in the way of the abuses of Iruka, and paying the ultimate sacrifice, effectively punctuating the line of this family. Aside from introductory evidence provided in Chapter 4, there are several other pieces of evidence that support this theory that Suiko was retroactively elevated to the throne. The most glaring is the fact that the story surrounding Toyo Mike Kashikiya Hime (Suiko) in both Kujiki and Nihon shoki is strangely thin and feels much like the “eight rulers lacking a history.” It is thus fascinating that Piggott (1997:79–83) constructs a carefully arranged argument about Suiko and her administration, paired with Shōtoku in a “brilliant political strategy.” She also goes so far as to label Suiko “heavenly heir and polestar monarch,” basing this argument on “her reading of Nihon shoki,” but one cannot help wondering what evidence was gleaned, and how it was weighed, considering how thin the account of Suiko actually is in Nihon shoki. It is difficult to account for the dearth of evidence for any such persona behind the throne in the Shoki record of Suiko. Consider that if we ignore the two paragraphs where Suiko comes to the throne and her lineage is listed, and the final paragraph where she dies, there are precious few instances of Suiko engaging in any meaningful governing. Shōtoku performs much of the heavy lifting of government. Table 6.2 lists each year entry, the event, and its categorization (M: miscellaneous, P: political, R: religious). There are four actions related to Buddhism or the native deities, four vague actions, three of which are recurring events of hunting for medicinal herbs, and six that have political significance. Contrast this information with the same arrangement of information taken from the bare-bones account in Kujiki (Table 6.3). It is indeed stunning that the majority of information from the Suiko record in Kujiki is not found in Nihon shoki. What is of great interest are the three “political” actions that only appear in Kujiki. One is the edict of the 27th year, which includes eight cap ranks. Clearly, the later committee working on Nihon shoki moved part of this information to the record of Kōtoku, when the caps are reorganized in 647. Here, the claim of Kujiki’s fraudulence actually hinders greater work on early historiography, because scholars are forced to ignore evidence in

The Tenmu-Jitō historiographical project 163 Table 6.2 Suiko’s actions in Shoki Year

Event

Category

2nd 8th 9th 11th 13th 14th 15th 16th 19th 20th 20th 22nd 31st 32nd

Issues edict to revere Buddhism Reclaim Mimana, discussion of action plan Resides in Miminashi Palace Suiko shocked at death of Prince Kume Oath to create Buddhist statue; propagate Buddhist scripture Suiko has Shōtoku lecture on Shōmangyō Suiko orders ministers to revere native deities Pardons Imoko, writes to Chinese emperor Hunts for medicinal herbs Poem written in response to Umako’s poem Hunts for medicinal herbs Hunts for medicinal herbs Suiko confers with ministers about attacking Silla Suiko censures Umako and censures unfilial monk

R P M P R R R P M P M M P P

Table 6.3 Suiko’s actions in Kujiki Year

Event

Category

2nd 22nd 27th 28th

Issues edict to revere Buddhism Issues edict to change Yatabe “Muraji” to “Miyatuko” Issues edict about loyalty and duty with caps (ranks) designated Issues edict about being disloyal

R P P P

relation to political events such as the institution of certain cap-ranks being moved between reigns in an attempt to account for differing political winds. Taking Kujiki at face value, we find it difficult to ignore the fact that the real “ruler” here is the regent: Shōtoku. We thus come back around to the initial, vexing question: why is his name not found in the Suishu account? The logical conclusion is that he was not actually ruling as Nihon shoki claims. And yet it is apparent that the Shoki compilers of the Suiko record were keenly aware of the visit of Bei Qing, as his visit is given a fair amount of detail in their record. I believe it is very likely that Suiko was actually the regent during Shōtoku’s reign, acting as a shadow ruler for her uncle, Soga Umako. If Jomei had been a young heir to the throne, then Suiko could have not only protected the young heir but also tried to influence the court with the backing of the powerful Soga. Here I suggest we untangle three threads: (i) the account from the Sui; (ii) the manipulation by Jitō’s Bureau of History; and (iii) the original, native account. First, a look at vestiges of the original, native account is helpful. There is an event in the Shoki record of Yōmei that raises eyebrows: the account of the attempted rape of Suiko by Prince Anahobe.

164

The Tenmu-Jitō historiographical project

The lineage preserved in our three records clearly shows that Prince Anahobe is a half-brother to Suiko. In the first year of Yōmei, Prince Anahobe forces his way into the palace of temporary internment, where Suiko is grieving the death of her husband, Great King Bidatsu; here Prince Anahobe attempts to rape her. This event is doubly heinous, because the prince is accused of trying to rape a woman in mourning. The record also notes that he had planned to become king over the entire realm, which report feels forced, juxtaposed with the attempted rape. Is this not a stereotypical slam where a person is painted with a broad brush as both a power-hungry tyrant and an immoral beast? Then, in the beginning of the record of Sushun, it says that Mononobe Moriya had tried to place Prince Anahobe on the throne as ruler, a historiographical ploy portraying the Mononobe as guilty by association. Under orders from Suiko, Soga Umako leads an armed force and they put Prince Anahobe to death. It is strange that very little appears in Kujiki, which simply records, “At that time, Prince Anahobe rebelled.” I contend that the lineage preserves information to help uncover the truth. Kojiki notes that Prince Anahobe is also known as Sume Irodo (須売伊呂杼), while Shoki records that he is also known as Sumito (住迹). Linguistically, a pattern here allows us to conclude that sume- is the original version of this name. Likely, an older genealogy had his name as 須彌, which in the early seventh century was read sume, a spelling convention also found in Jōgūki and Jōō Teisetsu.14 Later, this spelling came to be read as sumi, and that appears to be the origin of the Nihon shoki name. The underlying meaning of the version in Kojiki is “the ruler who is the younger brother of the same mother.” Prince Anahobe’s older sister by the same mother was wife to the king, and the mother of Shōtoku Taishi. Thus, Kojiki’s genealogy contains the name of the prince, with another appellation, that of his designation as a ruler who was the younger brother to Queen Anahobe, wife of Yōmei. My interpretation of these facts leads me to believe that Prince Anahobe had considered himself the legitimate successor to Yōmei, and with the military backing of Mononobe Moriya, probably believed he would actually reign; however, when he was passed over for Sushun—perhaps installed by Umako—he rebelled. What is difficult to fathom is the need to hatch the story about the attempted rape, unless later historians inserted the story to demonstrate Suiko’s resolve to defend herself in a male-dominated system. I believe Jitō’s historiographers fabricated the accusation of rape to strengthen the image of Suiko, so that when they placed her on the throne, she would have earned the status of an assertive ruler. This image reminds one of Amaterasu dressing up in military garb and acting in a masculine manner to confront her brother, Susanoo, whom she believed had come to rob her of her right to rule. The account also provided a justification for the destruction of the Mononobe at the hands of the Soga. Here I suggest that part of the motivation to rewrite Suiko’s history is the influence of stories out of China concerning Emperor Wu Zhao (r. 690–705), China’s first and the only female sovereign. Known earlier as Wu Zetian (624–705), she first became regent. After the death of her husband, Emperor Gaozong, her two sons ruled in succession, but she removed each after a brief reign, and in 690

The Tenmu-Jitō historiographical project 165 decided to ascend the throne herself to become the first female emperor of China. Her story is one of drama, intrigue, and cruelty, though the degree of this cruelty has been embellished over time as she is often portrayed as “the paradigmatic evil woman” (cf. Rothschild 2008:95). However, it would seem that Jitō’s Bureau of History wanted to piggyback on this story of a powerful, influential female ruler, but flip it so that Suiko attained the same status as the first female sovereign, earlier than China. However, Suiko attains this stature by virtue of her noble (divine) birth and pristine chastity, and not because of her cunning cruelty and subversive sexuality. I also find it difficult to ignore the fact that both Jitō and Wu Zhao ascend the throne in the same year (690). Table 6.4 compares the two women rulers. There must have been a reason that Yasumaro ignored all this when putting together Kojiki, as none of this appears in its pages, with little more than genealogy recorded. However, what is not recorded can be just as important as what is. When Kojiki enters the final phase of its record in Book Three, with only a listing of genealogy after Kenzō, it is important to notice not only what is recorded but also what is missing. Table 6.5, starting with Keitai, lists several pieces of crucial information, and whether present or not in each reign in the Kojiki record. It is Table 6.4 Comparison of Suiko and Wu Zhao Event

Suiko

Wu Zhao

Birth Ascension Regent Buddhism Widow Political change Innovation Promiscuous Cruel

Royal 554 (Jitō 690) Yes (before revision) Supporter Yes Yes Yes No No

Common 690 Yes Supporter Yes Yes Yes Yes Yes

Table 6.5 Basic information of recent rulers Name

Wives

Children

Years ruled

Age

Death date

Keitai Ankan Senka Kinmei Bidatsu Yōmei Sushun Suiko

7 None 2 5 4 3 None NA

19 None 5 25 17 7 None NA

╳ ╳ ╳ ╳ 14 3 4 37

43 ╳ ╳ ╳ ╳ ╳ ╳ ╳

527 535 ╳ ╳ 584 587 592 628

166

The Tenmu-Jitō historiographical project

worthy to note that Keitai is the only ruler to have his age listed, while neither Senka nor Kinmei have anything recorded, including a death date. The fact that Keitai is the only one with an age is even more glaring considering that Kojiki records the ages of Nintoku, Richū, Hanzei, Ingyō, Ankō, Yūryaku, and Kenzō. It is also interesting that only the last four rulers listed mention how many years they ruled. On the surface, this suggests that two sets of records existed. One had the number of years of the reign of the king, while the other had the number of years the king lived. Consider Table 6.6, working backward, listing the same information from Nintoku to Yūryaku. Looking at this from a different angle, clearly something has happened between Yūryaku and Kinmei. The messy reality of the beginning of recorded history is sandwiched between a neatly ordered series of legendary histories and a set of records that more closely resembles actual memory. It is plausible that the kataribe had memorized the ages and death dates of previous rulers. Keeping a record of how many years a king ruled may have been an innovation introduced by the Paekche historiographers. By the time of the Jitō court, it had been almost a century since the muddled era when Soga no Umako was in power. The Jitō court may have assumed it would be relatively easy to peel off Suiko, who was the queen of Bidatsu and make her the court’s first actual female ruler. It is highly likely the Bureau of History in Jitō’s era was aware that Suiko was a relatively young widow much like Jitō. Suiko was about 30 years of age at Bidatsu’s death. Jitō was in her 40th year when Tenmu died. If Suiko had acted as regent when Shōtoku first reigned, there may have been a strong desire from Jitō’s inner circle to flip the roles, mirroring the story of Wu Zhao, and make Shōtoku the regent, and Suiko the actual ruler. However, wedging Suiko in would require some chronological reworking, creating a gap that did not match the record in Suishu. Thus, the Bureau shortened Jomei’s reign. Further proof of this historiographical trick is the fact that Suiko’s reign is credited with great innovations: modernizing the court, instituting a new cap system, and establishing a 17-article set of regulations. Nevertheless, there is also the asymmetry where Suiko’s court could not decide on an heir to the throne, seemingly hamstrung between two candidates. The convoluted story of Suiko on her deathbed giving out cryptic information about who should be heir smacks Table 6.6 Basic information of earlier rulers Name

Wives

Children

Years ruled Age

Nintoku Richū Hanzei Ingyō Ankō Yūryaku

4 1 2 1 1 2

6 3 4 9 None 2

╳ ╳ ╳ ╳ ╳ ╳

83 64 60 78 50 124

Death date 427 432 437 454 ╳ 489

The Tenmu-Jitō historiographical project 167 of historical grudge-holding, perhaps because some later historians resented the fact that the Soga-backed Jomei was the next ruler and no amount of rewriting could erase that fact. The Soga-bloodline through Jitō and Genmei prevented the historians from recording at least some of the truth about the brutality of Soga rule, but the Bureau could not ignore the widely known Soga atrocities at court, so the assassination of Shōtoku at the hands of the Soga was rewritten as the coerced murder of his son, Prince Yamashiro. As penance for the assassination, a prior movement of hagiography regarding Shōtoku was officially sanctioned. The irony is that Shōtoku was added as a coauthor with his assassin, becoming an overseer of the very historical movement the historians a century later were trying to modify. As a hypothesis, I juxtapose a potential “original” history with the Jitō-era revision (Table 6.7). If this line of reasoning is accurate, then Jitō’s Bureau of History promoted two women to a higher position in the new history by tweaking the chronology, and fudging biographical facts. It is important to remember that Tenmu died around the age of 55 years. Granted, Jingū is never considered to have ruled, which may have been one of the political problems the historians faced, causing both projects (Kojiki and Nihon shoki) to grind to a halt. However, the compromise appears to have been that rather than calling Jingū a ruler outright, Nihon shoki simply records her death with the character 崩 “pass away,” a usage reserved for an imperial ruler. After Jingū was given prominence, historiographers in the early Nara era placed her reign back in time to align with the invasions of the peninsula by the Wa. Having tackled this issue, I believe Jitō’s (or later Genshō’s) historical committee felt emboldened to try their hand at giving Suiko her own reign, which also faced stiff political headwinds. The compromise appears to have been to allow Suiko to ascend the throne, but have Shōtoku continue to exercise the actual power, thus satisfying several camps. Furthermore, instead of recording Shōtoku’s assassination, factions originally aligned with his family would have been soothed by the story of a virtuous man dying too soon. This would then allow the budding hagiographical movement access to more court-sanctioned fodder: Shōtoku Taishi was now a holy sage in the realm of Japanese Buddhism. If these many assumptions hold up to further scrutiny, then this scenario would answer the question regarding both Kojiki and Kujiki, concluding their account within the reign of Suiko. It also accounts for the nagging question of why nearly four decades were required to finish the compilation of Shoki. Table 6.7 Revision of royal chronology Ruler

Original

Revision

Yōmei Sushun Shōtoku Suiko

Reigned 586–587, died 587 Reigned 588 until assassinated in 592 Reigned 593 until assassinated in 605 Acted as regent during Shōtoku’s reign Reigned 606 until death in 641

Reigned 586–587, died 587 Reigned 588 until assassinated in 592 Acted as regent during Suiko’s reign Reigned 593 until death in 628

Jomei

Reigned 629 until death in 641

168 The Tenmu-Jitō historiographical project

Kojiki and Nihon shoki I propose that the actual historiographical projects that resulted in Kojiki and Nihon shoki be decoupled and viewed separately. Umezawa (1976:434) notes that the basic fabric of the texts of Kojiki and Nihon shoki demonstrates a marked difference in thinking. Kojiki is composed in a style distinctly reminiscent of the liturgies (norito) and edicts (senmyō), while Nihon shoki is composed in classical Chinese in imitation of the chronicles of Paekche and China. It is important to remember that the Mononobe and Nakatomi families were protectors of the ancient traditions and religion, so it is possible that they were the force behind a historiographical framework that remained loyal to the ancient traditions. It is clear that the ancient texts were written in classical Chinese, but with the introduction of the kumanari method, the court later would have been able to innovate a hybrid script, similar to that used in senmyō announcements; with this innovation, it became possible to give greater emphasis to the spoken word, something the kataribe would have favored. Therefore, Hieda no Are’s focus on the spoken word in Kojiki allowed a break with previous histories. Recall that in the early Heian era Imibe Hironari also placed great emphasis on the spoken word in his version of the mythology preserved in Kogo shūi (cf. Bentley 2002:23). Rather than see Hieda no Are marking a return to an older tradition, I suggest we realign our focus to view Kojiki as a neoclassical attempt to make a break with the continental method and give greater emphasis to the oral quality of the ancient Japanese traditions, a return to the pristine past of traditional record keeping in Yamato, leapfrogging back in time, before Yūryaku established written records. The crucial question often asked and seldom answered satisfactorily is why the historiographical process began near the end of Tenmu’s life but took almost four decades to come to fruition. Yasumaro skims along the surface of this politically explosive issue by simply saying, “However, times changed and his (Tenmu’s) reign ended, and the undertaking still had not come to fruition.” I argue that the original edict by Tenmu as recorded in the tenth year of his reign (681) in Nihon shoki continued a process where the various ancient Japanese records were gathered and arranged into a chronological framework based on an elongated calendar that moved reigns further back into time. Tenmu desired to project the image of an empire under his control. It was this amalgamation of royal history coopting the history of events documented in the records of the powerful families the Tenmu-Jitō court desired to portray as existing harmoniously under the Sunumbrella of the empire of Yamato. With the death of Tenmu in 686, Jitō inherited the project, and with the backing of more conservative, traditional groups she decided to put together a history that was more representative of what the traditional groups considered to be their history. It is crucial to remember that both Kojiki and Nihon shoki travel essentially the same route, with the same rulers. The decoupling I argue for allows us to see two different camps, two different ideologies, both wrestling with the attempt to put forth a new, revised history. When these attempts bogged down with the abdication of Jitō, and the political wrestling to move the capital to Nara, Minister of the Left Isonokami Maro put

The Tenmu-Jitō historiographical project 169 forth a compromise suggestion to meld these two projects, and this resulted in the draft currently known as Sendai kuji hongi (cf. Bentley 2006:116–19). This project in turn stalled when Maro passed away in 717. It is possible that the political headwinds that had delayed both Kojiki and Nihon shoki intensified and capsized the Kujiki project, and the draft was shelved, never to be revisited.

Jitō’s Involvement Mitani Eiichi (1980:98–99) argues that there are hints in Yasumaro’s preface to Kojiki demonstrating that Jitō had been involved in the compilation instead of Tenmu. He claims that the underlying purpose of the preface is to tell the powerful men in the Council of State that while Yasumaro (and Genmei, the Kojiki project’s supporter) was announcing that this project was finished, it should not be construed as a competitor to the unfinished work Tenmu decreed (and which would result in Nihon shoki). To appreciate Mitani’s theory, we need to analyze the beginning of the Kojiki preface, which reveals that the document is a doublelayered text: (a) one layer describing the beginning and history of Japan, and (b) the second layer describing the progress of the Tenmu historiographical project. Consider how this is written and how these two parallel interpretations overlay: 夫混元既凝。氣象未效。無名無爲。誰知其形: (a) In the beginning, the original qi (氣 of the universe) that was in confusion congealed, having no signs of life and no form. It could not be named and had no movement, so who could know its form? (b) In the beginning, the original energy [氣, of the project] that was in confusion came together, but with no signs of life and no form. It could not be called [a success] and had no movement [toward completion], so who could know its [completed] form? 然 乾 坤 初 分 。參 神 作 造 化 之 首 。陰 陽 斯 開 。二 靈 爲 群 品 之祖: (a) However, when heaven and earth separated, the three deities who are the foundation of all creation came into existence. Here yin “female” and yang “male” parted and these two spiritual entities [viz. Izanami and Izanagi] became the progenitors of all creation. (b) However, when [the work recording] heaven and earth were separated, the three deities who are the foundation of all creation came into existence [in one of these works]. Here [the projects of ] yin [Jitō] and yang [Tenmu] parted and these two spiritual (= dead) entities [Jitō and Tenmu now being deified] became the progenitors of all [historical] creation. Thus, out of political confusion spring forth two historiographical projects, meant to be complementary histories not competitors. Later in the preface, when ending his description, Yasumaro adds the phrase, 連柯并穗之瑞。史不絶書 “The court historians will not cease to record the auspicious signs of the one-crown

170 The Tenmu-Jitō historiographical project two trunk tree and the one-ear two stalk rice plant.” This description is juxtaposed with the next line about tribute from foreign countries never ceasing. Mitani (1980:98–99) argues that Yasumaro is signaling that the completion of Kojiki is merely one of two auspicious events but should not be construed as signaling the end of the project to compile a new history. In essence, he declares that Kojiki is one of two auspicious projects; there is no intention to overshadow the other, unfinished project. Mitani then focuses on the line: 敷英風以弘國。重加智海浩瀚 “He [Tenmu] promulgated superior customs and had these extend throughout the realm. That is not all. His superior wisdom was broad like the sea….” Mitani suggests that the phrase 重加 “that is not all” is a signal to the council that Tenmu’s virtue and superior influence on the court has many characteristics, but it also signals that the history of the rulers has many characteristics. Kojiki does not pretend to have exhausted the description—an attempt to calm fears at court that the work would take the place of the official history of the court. Being broad like the sea, there is room for more than just one ship of history. I find Mitani’s analysis compelling, and if accurate, it then helps explain why the completion of Kojiki is ignored in Shoku Nihongi. It is as if Yasumaro is saying through his preface, “We have completed a history that does not claim to be the only history of the court. There is no need to be defensive.” The Council may have responded, “That is acceptable, but do not expect us to accord the work the status of a state history (an account later incorporated into Shoku Nihongi).” This bifurcated preface by Yasumaro appears to have been necessary to quell the throng of voices that opposed Jitō’s major undertaking to rewrite history, especially if it was seen as moving backward (toward oral histories), instead of forward to something more modern like Chinese histories. However, I suggest there was more than just traditional versus progressive factions at loggerheads here. An anti-Tenmu-Jitō faction resisted Jitō’s aggressive attempts to elevate Tenmu’s reign above other previous reigns, and by association place Jingū and Suiko on the throne when they had never ruled as true sovereigns. From the point of view of some factions at court, one particularly offensive movement was that of deification. Sugano (1992) argues that Jitō wanted posterity to see her as an incarnation of Amaterasu, the sun goddess. Notice that her name in Shoki is recorded as 高間原広野姫天皇Takamahara Hirono Hime “The Princess Hirono of the High Plain of Heaven.” This goes starkly against the posthumous name awarded after her death as recorded in Shoku Nihongi, in the 12th month of 703, 大倭根子天之廣野日女尊 “August Princess Heavenly Hirono, Yamato Neko.” Sugano (1992:94) goes on to argue that Jitō wanted her own family to mirror that of Amaterasu’s, with Jitō’s son, Prince Kusakabe, representing Oshiho Mimi and Kusakabe’s son, Monmu, representing Ninigi. It was thus not enough to claim to be direct descendants of the sun goddess: Jitō wanted the court, and the empire at large, to see her family as a physical embodiment of that divine trio. It is thus striking to remember that Emperor Wu Zhao did something similar, proclaiming that she was an embodiment of Maitreya Buddha (McBride 2008:52),

The Tenmu-Jitō historiographical project 171 or that she was a living goddess (Clements 2007:162). There is always a problem, however, when people want an ideology to mirror reality. The problem with creating this image was the clash with what the powerful families already knew. It was public knowledge that Tenmu’s oldest son, Prince Ōtsu, was next in line as heir to the throne, but he stood in the way of Jitō realizing her dream. The account in Nihon shoki has Ōtsu arrested for rebellion, and later committing suicide in the house of Osada, but I agree with Ooms (2009:23), who finds the account in Shoki suspect. Consider that Man’yōshū contains a poem by Ōtsu with the headnote, “When Prince Ōtsu was about to be executed (so Shoki’s portrayal of suicide is an attempt to give him an honorable death), he wrote a poem as he shed tears on the embankment of Lake Iware.” The poem is dated the tenth month of the first year of Shuchō of the Fujiwara Palace (686), but most believe the poem to have been actually composed by someone close to the prince. The prince also supposedly composed a famous poem included in Kaifūsō, lamenting the end of his short life: 金鳥臨西舎 鼓聲催短命 泉路無賓主 此夕離家向

The bird of the golden orb shines on the western hall.15 The sound of the drum hastens my short life onward. The road to the afterlife has no guest nor lord. Tonight I leave home and proceed down that road.

The evidence does not support the assumed fact that Prince Ōtsu planned a rebellion. Like the machinations against the reputation of Prince Anahobe, it seems highly likely that Jitō pushed Ōtsu out of the way by accusing him of rebellion, so her only son Prince Kusakabe could ascend the throne. However, when Kusakabe dies three years later, his young son, Prince Karu, is groomed to ascend the throne when Jitō abdicated in early 697. Regarding Jitō’s desire that society view her family as the embodiment of Amaterasu-Oshiho Mimi-Ninigi, Ooms (2009:36) argues that this mythological lineage was pushed by Jitō to mirror her own attempt to retain control of the throne. This development underscores my theory that Kojiki was compiled with vastly different intentions than Shoki. Umezawa (1988:46–47) argues that Kojiki and Nihon shoki have clearly different methods for “eradicating error and establishing truth.” Kojiki’s methodology is to make a unilateral decision about which of the many traditions was correct and ignore the variations—hiding the truth, if you will—while Nihon shoki’s editorial attitude is, “We will list all the variants, and later generations can determine which is correct.” This unilateral, coercive focus of Kojiki lends credence to the theory advanced by Mitani (1980:92) that Jitō was the driving force behind Kojiki. Umezawa (1988:8, 13) sees Kojiki as a reaction against the historiographical project that later resulted in Nihon shoki, where one or more factions at court did not want their history cemented in a Chinese-style format but wanted something more traditional. Umezawa (1988:4, 60) highlights another fact: comparing the mish-mash text of Nihon shoki with the fluid, almost literary style of the story in Kojiki, one is obliged to admit that all things being equal, it appears that Kojiki took the same source material as Nihon

172 The Tenmu-Jitō historiographical project shoki then slashed and burned through the various older documents to create a smooth flowing narrative. A reader is hard-pressed to believe that there was a host of older records that Kojiki relied on simply by reading through its account. Umezawa (1988:62–67) furthermore believes that Nihon shoki and Kojiki resulted from what he calls “a fracture in Tenmu’s historical concept.” This means that while Tenmu had an idea about cementing a new history, preserving the ancient traditions and genealogies against the onslaught of continental immigrants and their histories, those engaged in the actual work fractured into two camps, obliging Tenmu to create two mirror projects. I believe this account is essentially accurate, but I see Jitō taking control of these two groups and shepherding each through the process to create a text with her imprint on it. Umezawa (1988:71–73) calls our attention to the fact that the supposed edict of Tenmu starts with the emotive words, “We hear [朕聞].” Thus, the great ruler is reacting to a report, likely from someone within his own inner circle about disturbing information: the corruption of ancient traditions and lineage. The difficulty with Umezawa’s theory here is that this supposed edict is being reported by Yasumaro, who we can only view as a member of the Jitō camp. Perhaps it is more reasonable to see the essence of the edict coming from Jitō with Tenmu’s name inserted to give the edict credence. One reason for an adverse reaction against the juggernaut to create a Chinese-style national history was a neoclassical movement that blanched at the perceived corrupting influence of Chinese script. Yasumaro’s preface contains a critical set of sentences that have been subjugated to much speculation: Anciently words and their meanings were both pristine. Laying out sentences and constructing phrases [in Japanese], it is difficult to convert these into characters. Relying solely on Chinese pronunciation, the characters do not capture the meaning of the [Japanese] words. Stringing sentences together completely in phonograms makes the sentences long [and taxing] [然上古 之時、言意並朴。敷文構句、於字即難。已因訓述者、詞不逮心。全 以音連者、事趣更長]. As I have argued earlier, when one realizes that ancient Japanese histories were first composed in classical Chinese, one wonders why this criticism is appearing now, more than a century after the records were first recorded and compiled. I believe this criticism was inspired by the murmuring kataribe, whose very existence was based on oral recitation, in Japanese, of ancient events. In other words, the charge brought to the bench is that composing the ancient histories solely in classical Chinese does not do justice to the native, Japanese words and culture.16 However, as Yasumaro, and other educated courtiers, would have countered, the other option, to use a completely phonographic rendering, makes telling the history too long and laborious and prevents foreign dignitaries from being able to read the record. Yasumaro’s solution is to use a mixture of the two methods, or what we have come to call kundoku. Here, however, I wish to focus on Yasumaro’s judgment of writing in classical Chinese.

The Tenmu-Jitō historiographical project 173 There are several key words here that tell the reader that Yasumaro (and by extension the conservative groups at court, including the kataribe) adhered to the perception that some kind of violence is done to the native traditions when converted into pure Chinese. First, the character 朴 has two readings in Chinese, pò “Chinese hackberry” and pŭ “something in a natural state, pristine.” This word as found in Kojiki has been rendered in English as “so simple” (Chamberlain 1981:13), “unsophisticated” (Philippi 1968:43), “forthright” (Lurie 2001:301), or “simple” (Heldt 2014:4). These renditions appear to be based on the understanding in Japanese of the binome 素朴 “simple, unsophisticated.” In their modern Japanese translation, Kōnoshi and Yamaguchi (1997:25) translate this as 飾り気がなく “unadorned,” while Sugano (2012:364) renders this as 素直 “honest, frank.” Kōnoshi and Yamaguchi’s rendition is closest to what the Chinese actually means. What most scholars appear to ignore is the fact that 朴 refers to something in a natural or pristine state. Lurie (2001:307) argued, “[T]he dominant linguistic contrast in the Preface is not between Chinese and Japanese, but between everyday language and ancient language.” While this is true, I do not think we can ignore the fact that Chinese is the vehicle Yasumaro is describing. Thus, I expand Lurie’s statement to demonstrate not only a contrast between two forms of the language but also the subtextual analogy of pure versus polluted, insinuating that one was seen as inflicting damage on the other. I read Yasumaro as saying that taking the ancient traditions of Japan, with their important lineage, and converting these into a Chinese-style history, with its foreign ideas and embellishments, would be detrimental to the very medium the court wished to preserve. The second key word (phrase) is 不逮 “does not reach up to, does not capture.” Thus, this phrase infers that using Chinese characters in their function as kun does not capture the meaning of all Japanese words, especially concerning Japanese cultural artifacts. It is analogous to going fishing with a hook too small to catch the kind of large fish that will allow you to have some economic freedom. Yasumaro (and by extension the kataribe) apparently feared that taking a native tradition and converting it into classical Chinese would obscure the original meaning, so that a reader would then read the Chinese, and in attempting to convert the sentences back into Japanese, they would get something different. To reduce this to an equation, if you take α and filter it through γ, you end up with β, and α ≠ β. The desired outcome is α = α. I believe that the kataribe had great influence on Yasumaro, because while much of Kojiki is written in a kundoku method, the conjugation of adjectives and verbs is given prominence. People who recite know that the power in a sentence resides in descriptive words like adjectives and verbs. Thus, rather than leave the interpretation of how to read these emotive words to the reader’s imagination, it is critical to provide sufficient hints to aid in proper reading (or recitation). What makes this exercise worthy of note is that Yasumaro is quietly railing against a system firmly anchored in place for several centuries. The Paekche teachers used classical Chinese as a medium of communication and recording, and then they taught the Yamato court how to comprehend that medium through kumanari (later kundoku) so it could be read as Paekchean and then later as Japanese. However, Yasumaro and the kataribe were sensitive to the perceived

174 The Tenmu-Jitō historiographical project imperfections of that complex system, acutely aware that it was prone to misunderstanding. Yasumaro’s answer was to mix two systems: Chinese and phonograms. Yajima (2011:92) quotes research by Ozaki Satoakira, who analyzed the phonogrammatic usages in Kojiki. He notes that Yasumaro relied on a simple group of phonograms to set the phonographic symbols apart from the Chinese semantic usages. He notes two characteristics: (i) the phonograms are almost never used in the text semantically, and (ii) these graphs were selected to be easily recognizable as phonetic clues, making it easier to avoid making a semantic connection. Yajima (2011:92–93) then quotes work by Nishimiya Kazutami, who observed that the phonograms in Kojiki are clearly discernable as having a single function (as phonetic markers), and these phonograms are even distinguished graphically from those used semantically. His example is the use of 餘 “excessive” as a semantogram but 余, the abbreviated version of 餘, is exclusively a phonogram for yo. Thus, Yasumaro made a concerted effort to make his text as transparent as possible to a reader or reciter, given the limitations of the medium. Kojiki thus stands out as a native aspen tree in a forest of Chinese pines. Kojiki tries to preserve its version of history in a way that the reader can mentally recreate when read. Kujiki and Nihon shoki opted to adhere to the standard method of recording history in pure Chinese, though there are areas where Kujiki has mixed parts of the Kojiki text within its Chinese fabric. Nevertheless, if what I have argued above is accurate, it is still critical to remember that our three texts, Kojiki, Kujiki, and Nihon shoki, have a common starting point in that Jitō and her historiographical bureau manipulated the records to create a new history that only mildly resembled previous histories.

Creation of Nihon shoki Ever since the days of Motoori Norinaga, Nihon shoki has been routinely criticized as a Japanese history masquerading in Chinese garb, a slap meant to diminish the value of the work in comparison to Kojiki. A careful reading of either work makes it clear that the influence of Chinese civilization, culture, and philosophy runs deep through both works (Kōnoshi 1995:193–222). The greatest difference, however, is that Nihon shoki (as well as its source material) has engaged in wholesale copying from Chinese sources without naming these. As an example of what the process may have looked like, consider a quote of the thin record of Kōshō: 觀 松 彦 香 殖 稻 天 皇 、大 日 本 彦 耜 友 天 皇 太 子 也 。… 元年春正月丙戌朔甲午、皇太子即天皇位。〇夏四月乙卯朔己未 、尊皇后曰皇太后。 〇秋七月、遷都於掖上。是謂池心宮。◎是年也、太歳丙寅。 廿 九 年 春 正 月 甲 辰 朔 丙 午 、立 世 襲 足 媛 爲 皇 后 。一 云 、 磯 城 縣 主 葉 江 女 渟 名 城 津 媛 。一 云 、倭 國 豐 秋 狹 太 媛 女大井媛也。后生天足彦國押人命・日本足彦國押人天皇。 六十八年春正月丁亥朔庚子、立日本足彦國押人尊、爲皇太子。年廿 。天足彦國押人命、 此和珥臣等始祖也。八十三年秋八月丁巳朔辛酉、天皇崩。

The Tenmu-Jitō historiographical project 175 Heavenly Sovereign Mimatsu Hiko Kaesine was the oldest son of Heavenly Sovereign Ōyamato Hiko Sukitomo…. First year [475 BC], spring, first month, ninth day. The heir to the throne ascended the throne and became heavenly sovereign. Summer, fourth month, fifth day. The heavenly sovereign praised the deceased heavenly sovereign’s queen and called her queen dowager. In the autumn, seventh month, the court removed the capital to Wakinokami. This is known as the Ikegokoro Palace. This was done in the greater year of Hinoto Tora. 29th year, spring, first month, third day. The heavenly sovereign made Yoso Tarashi Hime his queen. Another record states this queen is Nunakitsu Hime, the daughter of Haye, the district leader of Shiki. Another work records that she is Ōi Hime, the daughter of Toyoaki Sada Hime of Yamato Province. The queen gave birth to Ame Tarashi Hiko Kunioshihito and Heavenly Sovereign Yamato Tarashi Hiko Kunioshihito. 68th year, spring, first month, fourteenth day. The heavenly sovereign made Yamato Tarashi Hiko Kunioshihito heir to the throne at the age of 20. Ame Yarashi Hiko Kunioshihito is the first ancestor of the Wani Omi. 83rd year, autumn, eighth month, fifth day. The heavenly sovereign divinely departed. Several potential problems are apparent from this quote. All names have been converted into semantograms, meaning that whoever converted the names is assumed to have known what the individual components of each name meant.17 Another issue is that someone like Kōshō about whom very little is recorded has a full-blown record, even if it is sparse when it comes to details. Yajima (2011:38) notes that Nihon shoki was obligated to convert its history, even back in the beginnings of the hoary past, to a chronological flow, based on a Chinese historiographical idea also of great antiquity: He commanded the brothers Hsi and the brothers Ho, in reverent accordance with their observation of the wide heavens, to calculate and delineate the movements and appearances of the sun, the moon, the stars, and the zodiacal spaces [houses], and so to deliver respectfully the seasons to be observed by the people [乃命羲和、欽若昊天、歷象日月星辰、敬授人時]. (Waltham 1971:4) I have argued throughout that we need to revise this thinking, because the original template came from Paekche, not China. However, by the time of the inception of Nihon shoki, China had become the direct model. One key to understanding the final process of Nihon shoki’s compilation process is to examine Book Four of Nihon shoki, which contains the thin histories of eight rulers. A pattern emerges that we can speculate is a vestige of older records (see Table 6.8). The table lines up these eight rulers (from Suizei to Kaika), and

176

The Tenmu-Jitō historiographical project

Table 6.8 Ancient template for a dynastic record Year/Name

Suizei

Annei

Itoku Kōshō Kōan

Kōrei

Kōgen

Kaika

Pre-ascension First year Next year entry Third entry Year w/ lineage Queen Dowager New capital Death year

○ ○ 2nd 4th 2nd 1st 1st 33rd

○ ○ 2nd 3rd 3rd 1st 2nd 38th

○ ○ 2nd 22nd 2nd 1st 2nd 34th

○ ○ 2nd 36th 2nd 1st None 76th

○ ○ 4th 7th 7th 1st 4th 57th

○ ○ 2nd 7th 7th 2nd 2nd 60th

○ ○ 29th 68th 29th 1st 1st 83rd

○ ○ 2nd 26th 26th 1st 2nd 102nd

notes whether a record has a pre-ascension narrative, and then what the first threeyear entries may contain. It also notes which year-entry contains the genealogy, as well as the establishment of the queen of the previous ruler as queen dowager, the establishment of a new capital, and the year of the ruler’s death. This table demonstrates that even though there was a template that each royal reign was based on, it seems clear that areas with fabricated information could not be systematically plugged into that template by the compiler. The reason for this may be as simple as ineptitude on the part of the compiler, or the basic information of each reign was unsystematic. Consider that 75% of these dynastic records have the second year as the second entry, aside from Kōshō and Kōgen. Kōgen’s second entry as the fourth year is understandable, but Kōshō’s almost three-decade lapse is difficult to fathom. It is possible that a compiler garbled second year into a 29th year, but that is not very plausible, especially since the third entry is another three decades later. Expansion of the chronology seems to be the best answer. The lineage, which rather consistently appears in the year entry after the inaugural year, does appear in the 29th year, showing an attempt to follow the template. My personal theory is that the compiler cut a predetermined 90-year reign into three roughly equal parts. Another interesting fact about Book Four of Shoki is the lack of quotes from Chinese sources. The only glaring anomaly in this book is the record of Suizei, which has a substantial pre-ascension record. Suizei’s pre-ascension records have many quotes from Chinese works, such as Wenxuan and Sanguozhi. Of special notice is this line, after Suizei’s lineage: The Heavenly Sovereign was of a handsome figure and knowledgeable. As a child, he was preeminent in masculine vigor. When he came of age, his appearance was exceptional and complete, and he was skilled in military strategy. His intentions were noble, and he was calm and persevering. The Heavenly Sovereign Kamu Yamato Ihare Hiko divinely departed when he [Suizei] was 48 years old. With the death of his father, Kamu Nunakawa Mimi’s sorrow was without end, because of his sincere and deep filial feelings [天皇風姿・岐 嶷。少有雄拔之氣。及壯容貌魁偉。武藝過人。而志尚・沈毅。至四十 八歳、神日本磐余彦天皇崩。時神渟名川耳尊、孝性純深、悲慕無已].

The Tenmu-Jitō historiographical project 177 The underlined sections underscore descriptions lifted from Chinese sources. In this rather short narrative, there are six descriptions taken from various records, such as Wenxuan, Hou Hanshu, and Weizhi. What is of greater interest is why the compilers felt it necessary to go to such lengths to tell this story about a rebellion among brothers, especially when Nihon shoki claims these early kings all descended father to son. Reflect back on the account in Suishu, where it records that the ruler came to court first, before dawn, and then when the sun came up, he let his younger brother rule. This suggests that a brother-to-brother lineage was traditional in ancient Japan. This pre-ascension story in Suizei appears to be illustrating (and justifying) the reason why a brother would slay a supposed rightful heir. Kojiki also contains this story, but it is lyrical in nature. When Isuke Yori Hime, Jinmu’s consort, finds out that Tagishi Mimi intends to put his brothers to death, she sings two songs, alerting the other brothers to the plot. Nihon shoki takes a different tact. When Suizei and his older brother, Kamuya Mimi, discover that Tagishi Mimi intends to take the throne by force, they decide to cut him down. However, when the opportunity presents itself, Kamuya Mimi is paralyzed with fear, unable to shoot his arrow, so Suizei steps up and puts the rebel to death. This story thus justifies a circumstance where the younger brother should ascend the throne over the older. The fascinating fact is that Jinmu is the youngest of the four sons, but his older brothers all perish in battle, fortuitously allowing Jinmu to overcome a biological obstacle (order of birth) to take the throne because of his valor. Suizei is the third son of Jinmu and gains the right to the throne through his courage and wisdom. This father-to-son succession continues until Richū, who is the oldest son of Nintoku. As noted previously, this father-to-son inheritance is modeled on the Paekche royal tradition (Best 2006:108). The ruler following Richū is Hanzei, Richū’s younger brother by a different mother. We can thus view Nintoku as a demarcation line between the Paekche model of father-to-son succession and the Wa tradition of brother-to-brother succession. Another key to understanding how Nihon shoki came about is to examine word usage. Careful textual analysis of graphic usages in the text allows us to formulate theories on how the actual compilation took place. Based on this evidence, I postulate that subcommittees existed that were tasked with compiling information for various rulers. Fujii Nobuo (1952) published research regarding the written style within the records of the event when a ruler took the throne and moved his capital. His method is to examine the difference in wording, such as 都Y、是謂X宮 “The capital was Y, and this was called the X Palace” versus 遷都Y、是謂X宮 “The capital was moved to Y and this was called the X Palace.” Another variation is 遷都Y、因為X宮号 “The capital was moved to Y and that is why it was called the X Palace.” Based on these varied data, he finds commonalities that fall into four groups, as noted in Table 6.9. Fujii then mentions textual critical work by Harada Toshiaki, Nishimiya Kazutami, Ōta Yoshimaro, and Kōnosu Hayao, who all agree that the books in Shoki can be aligned into eight groups. Fujii says that their textual work matches his own conclusions. Those eight groups are: (i) Books 1 and 2, (ii) Books 3–13, (iii) Books

178

The Tenmu-Jitō historiographical project

Table 6.9 Word usage regarding the capital Group

Books

1 2 3 4

4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 22, 23 14, 15, 16 17, 18, 19, 24, 25, 26, 27 20, 21

14–19, (iv) Books 20 and 21, (v) Books 22 and 23, (vi) Books 24–27, (vii) Books 28 and 29, and (viii) Book 30. Fujii’s research does not deal with the two books of mythology (Books 1 and 2), because these do not describe royal ascension or the capital. Notice that groups (ii) and (v) match Fujii’s group 1. He divides groups (iii) and (vi) into slightly different groups. Finally, group (iv) matches Fujii’s group 4. Yokota Ken’ichi conducted similar research on fundamental wording in Shoki, focusing on how the compilers recorded marriage and the resulting children. This research has the potential to provide greater granularity to our analysis; he notes that there are six basic methods to record the order of children born, and seven ways to record marriage (1966:98–104). First, a look at how the birth of children is recorded is enlightening. The six methods for recording the posterity of a ruler in Kojiki and Nihon shoki are: A numerically, first, second… (第一、第二、第三, with variations: 其一、其二); B a simple listing of the names of the children in order of birth; C attaching the character 次 “next” to each child’s name; D using designations of older, middle, younger (長・仲・少); E using familial designators like older brother, younger sister; F when there are only two children, using the Chinese character 与 “and” between the names. Yokota then charts these results according to the ruler. The method Kojiki employs is almost exclusively C, with an occasional mixture of C and E after Nintoku. This consistency in recording lineage across the record demonstrates that Kojiki was the work of one person. Nihon shoki, on the other hand, shows an interesting distribution, replicated in Table 6.10.18 Yokota notes there are 25 different varieties of recording marriage, according to verb usage, and the status of the wife (queen, consort, concubine, and so on). These complex criteria are more difficult to visualize graphically because of the granularity of these data. However, I have categorized the variety of these different methods, hoping to shed some light on what the source documents contained. Table 6.11 lists each book with lineage and the number of different methods to represent marriage (from one to a maximum of seven). It is clear that some records have more complex methods for recording vital statistics than others do. Combining information from these two sets of data makes it clear that five books in Nihon shoki use very rudimentary recording methods, attesting to the antiquity of the basic source material: Book Three “Jinmu,” Book Eight “Chūai,” Book 14 “Yūryaku,” Book 17 “Keitai,” and Book 18

A B C D E F

○

2

○

3

○ ○ ○

4 5

○

○ ○

6

○ ○ ○

7

○

8

○

10 ○

11

Table 6.10 Distribution of ways to record children

○

12 ○

13

○ ○

○

14 ○

15

○ ○

○

17

○ ○

18

○ ○

○

19

○

○

20

○

○

21 ○

23

○

○

27

○

○

29

The Tenmu-Jitō historiographical project 179

1 2 3 4 5 6 7

x

2

x

3

x

x

4 5

x

6

x

7

x

8

x

10

Table 6.11 Distribution of marriage

x

11

x

12

x

13

x

14 x

15 x

16

x

17 x

18

X

19

x

20 x

21 x

22

x

23

x

25

x

27

x

29

180 The Tenmu-Jitō historiographical project

The Tenmu-Jitō historiographical project 181 “Ankan-Senka.” Perhaps the following conclusion is realistic: Two of these books were rudimentary because of the antiquity of the data, and the remaining three books were created during the second stage (in the Suiko era), based on the basic template from the earlier two books, and remained in a rudimentary state until the final phase began during the Tenmu-Jitō era. I outline this below, in order of appearance.

Jinmu Kojiki does not list any children of Jinmu in the actual record. This information appears later. This suggests that the legendary story of Jinmu existed outside of the historical narrative, and there was no concrete connection to Jinmu and later dynasties. I believe the historiographers in the Suiko era tried to flesh out the story, as they only had a lyrical narrative transmitted in Kataribe versions; this lyrical version survives in Kojiki. The original hero of this story was Hiko Hohodemi. Nihon shoki states that his name was Kamu Yamato Iware Hiko, but his real name was Hiko Hohodemi. Kujiki notes, “He was the fourth son of Hiko Nagisatake Ugayafukiaezu, his real name being Kamu Yamato Ihare Hiko. Also known as Hiko Hohodemi.” Kojiki calls him, “Waka Mikenu, also known as Toyo Mikenu, or Kamu Yamato Iware Hiko.” This provides evidence that various iterations of the story existed in Yamato. As I have earlier argued, a detailed pre-ascension story of Jinmu was created during the early years of the Nara era by the committee under the auspices of Minister Isonokami Maro to celebrate Tenmu’s victory in the Jinshin Disturbance. This reconstituted story was a melding of the lyrical story of Jinmu and the actual battles of Tenmu (cf. Bentley 2006:94–99). It is logical to see this ideological framework coming from Maro as a gesture of gratitude to Tenmu; Tenmu spared his life during the revolt, as Maro was a member of the losing faction (Kōbun).

Chūai Yokota’s research (1966:106) illustrates that a look at the record of Kojiki uncovers the fact that the Keikō and Chūai records are the only accounts in Kojiki that mix types B and C in recording the royal offspring. This is anomalous as the rest of the record consistently uses type C. Contamination from type E appears after Nintoku, which I interpret to mean that this is the archetype, and either Hieda no Are or Yasumaro standardized the record to the easier format of C. In format, B is the most basic, a simple listing of children, and suggests that the information in the Keikō and Chūai records are vestiges of an archaic tradition. This again suggests that Nintoku is the upper limit of the institutional memory of the original historiographical group formed during the court of Yūryaku. It is logical to conclude that Chūai and his wife, Okinaga Tarashi Hime (Jingū), were figures in a legend from the powerful Okinaga family based in Afumi. This powerful family may have had stories of a chieftain who participated in incursions against Koguryŏ and Silla. If true, the Suiko historiographers may have rewritten this as a ruler who

182

The Tenmu-Jitō historiographical project

had been instrumental in military affairs on the peninsula. I have argued above that during the Jitō era, Chūai’s story was reconstituted so that Jingū received greater focus, and this shift appears in both Kojiki and Nihon shoki. The compilation process that resulted in Nihon shoki quoting from the Paekche source, Paekcheki, demonstrates an attempt to depict Yamato’s history as contemporary with Paekche. These Paekche quotes pin the events of Jingū to a concrete calendar, that of Paekche. Thus, there was a clear attempt to correlate this rather thin record with events on the peninsula. Ledyard (1975:238–242) argues that most of the events recorded within the Shoki account of Jingū’s invasion of the tip of the peninsula, overrunning Mimana and Silla, are actually based on a Paekche account of Puyŏ defeating the ancient kingdom of Mahan. The directions of the attacking armies demonstrate a northern facing orientation, not a southern one like Yamato. While Ledyard’s argument is quite persuasive, archaeological evidence from the last 20 years demonstrates that it was not actually Puyŏ that was invading Mahan but Koguryŏ who made an incursion into Mahan (and later Paekche) territory (Barnes 2001:35).

Yūryaku There are only three books in Nihon shoki that employ the wording found in category E, matching that in Kojiki. It is natural to assume that the original genealogies went to great lengths to demarcate older and younger brothers or sisters. The fact that genealogies became more removed from a family setting as time marched on, using numbers or rather vague designators, demonstrates that the genealogies were manipulated later as the number of descendants increased. The genealogy preserved in Book Five of Kujiki supports this fact; the genealogies of the Owari and Mononobe therein use the E format to record these names of the children. If my argument that Yūryaku was the first court to commit history to writing is accurate, then this record would have been written early. It is thus interesting that later courts portrayed Yūryaku rather poorly in the records we currently have. Consider Table 6.12, which lists major events in the reign of Yūryaku, and where these appear in our three records. The order of events follows the account in Kojiki, augmented by Nihon shoki. I am inclined to see Kojiki’s story as the older version. Notice that the establishment of royal service groups (代) appears here. These first appear in the reign of Nintoku, and it is easy to imagine that the establishment of these groups was projected back in time to Nintoku, who was the limit of institutional memory in Yūryaku’s era. The presentation of Yūryaku painted by Kojiki is of a romantic military leader who is not terribly sensitive to the feelings of women. The account of Yūryaku killing Prince Ichinohe does not appear in the Yūryaku record, relegated to the Ankō record, so for now I have ignored it. In addition, it is important to realize that some of the violent deeds attributed to Ankō in Kojiki are appropriated by Kujiki and Nihon shoki and applied to Yūryaku in his record. I theorize that the Bureau of History in the Jitō era felt it necessary to employ a Chinese ideology, creating a wicked ruler to portend the destruction of one royal

The Tenmu-Jitō historiographical project 183 Table 6.12 Events in Yūryaku’s reign Event

Kojiki

Birth of children Establishment of industrial groups Story of Hiketa Akaigo The maiden of Yoshino Yūryaku bit by a horsefly The deity Hitokotonushi Love songs from Yūryaku Prince Mayowa put to death Prince Ichinohe put to death Princess Hatabi made queen Yūryaku gets maiden pregnant Ishikawa Tate and Ikehime burned to death Yūryaku kills cook Princess Takuhata commits suicide Yamato attacks Silla Industrial groups come from Kure Koguryŏ destroys Paekche Prince Shiraka made heir Yūryaku’s last words Death of Yūryaku

○ ○ ○ ○ ○ ○ ○

○

Shoki

Kujiki ○

○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○

○ ○ ○

○ ○

branch, and herald the advent of a newer, virtuous branch. Thus, the historians portrayed Buretsu, a grandson of Yūryaku, as extremely wicked, effectively punctuating the end of the Nintoku line, the blame being laid at the feet of Yūryaku and Buretsu. Naturally, this was editorial maneuvering by the committees at the Jitō (and Genmei?) courts and does not represent the older records. Notice that there is very little overlap between the two accounts in Kojiki and Nihon shoki. Kujiki marks the beginning of this restructuring of Yūryaku’s record. Kujiki lays the groundwork in the pre-ascension record, where Yūryaku is responsible for viciously putting Mayowa and Tsubura Omi to death, and then later deceives Prince Ichinohe and kills him. Other cruel acts by Yūryaku do not appear in Kujiki, and later compilers added these to the narrative fabric in Nihon shoki. Again, over time the Bureau of History took a traditional hero of early history and vilified him.

Keitai The record of Keitai also employs a rather rudimentary system to record birth, but marriage is a bit more complex. Like the Yūryaku record, Keitai uses the E system, the archetype. Marriage in the Keitai record is complex, which I ascribe to many families desiring to graft their lineages into the royal line of Keitai, and thus receive legitimacy from the throne for their influence and power. While Kojiki does not portray Yūryaku as an evil ruler, as Kujiki and Nihon shoki do, there

184 The Tenmu-Jitō historiographical project are still vestiges of an editorial movement to create a Chinese space where Keitai could appear on the scene as a virtuous ruler. Thus, the Keitai record in Kojiki starts with the words, “Wohodo, a descendant of Great King Homuda (Ōjin) in the fifth generation, ruled everything under heaven from the Tamaho Palace.” It is thus interesting that the original signpost of institutional memory, Nintoku, is pushed back one generation to Ōjin. An important question to ask is why the historiographical committee felt the need to do this. The impetus appears to be the earlier record, quoted in Jōgūki, which noted that Keitai’s genealogy extended back to a prince called Homutu Wake. Kojiki originally lists Ōjin’s name as Ōtomo Wake, with his variant name being Homuda Wake. This clearly demonstrates that several names existed in earlier records, and the early historiographers simply dumped the names of people from the same era into one bucket labeled Ruler A. This practice is most often seen when dealing with deities, as different but related myths were brought together, and the court wanted various groups to be able to point to their own myths in the records. The easiest solution appears to have been allowing certain actors in the myths to have a variety of names: “A (also known as B or C).” Keitai’s record is mainly genealogy, but it is worthy to note that Yasumaro takes some license when he includes the line, “During this reign, Iwai, Lord of Tsukushi, was disobedient to the command of the divine ruler, and committed many offensive acts. Thus, the throne dispatched Mononobe Arakahi Ōmuraji and Ōtomo Kanamura Muraji and they put Iwai to death.” As I have noted earlier, Kujiki ignores this revolt but mentions Mimana (in the record of Kinmei), while Kojiki ignores the peninsula but mentions Iwai’s revolt. In reality, these two events are interconnected, and the divergence is nothing more than the editorial mandate given to each group. Kojiki describes an empire that includes the peninsula but makes no mention of China. Kujiki and Nihon shoki follow the path with a global view, seeing Yamato as a worthy country that can stand shoulder to shoulder with China. Thus, having influence on the peninsula is important to describe, though Kujiki simply lays the groundwork that Nihon shoki later builds on. It is interesting that while Nihon shoki spends a great amount of time discussing peninsular affairs in the Keitai record, Kujiki is focused on domestic affairs. The Keitai pre-ascension story in Kujiki mirrors that in Nihon shoki, where Buretsu dies without an heir, and the court brings Keitai in from the provinces. Whereas Kujiki makes this a flowing narrative before introducing the first month of the first year of the reign, Nihon shoki starts with a new entry with a specific date (the 21st day of the 12th month). Kujiki devotes most of its Keitai record to his marriage and the children produced. For example, Kujiki’s eighth year entry simply says, “Prince Magari Ōe was ordered to take up residence in the Heir to the throne’s Palace, with the decree. He will assist us in the government, performing benevolence, and supporting us in our faults” (Bentley 2006:342). Nihon shoki, on the other hand, uses its eighth-year entry to explain that the wife of the heir to the throne (Prince Magari) cannot bear any children. The heir grants her a royal service group allowing her name to continue. Notice the

The Tenmu-Jitō historiographical project 185 difference in editorial intent: Kujiki places emphasis on the fact that Prince Magari is the heir and will assist Keitai. Nihon shoki, on the other hand, has Keitai issue a long and elaborate edict praising Prince Magari in the seventh year, including the following: Our beloved son has made known our will to every quarter. May he prosper. Magari Ōe has allowed the light of our customs to shine on the myriad foreign countries. The land of Yamato is tranquil, and everyone will have a reputation throughout. The land of Akizu [Yamato] is prosperous and its admiration is valued in the central districts of the royal region. What should be prized are wise men, who take the greatest delight in doing good. Because of this, our sagacious influence touches people’s hearts in far off places, and the profound merit of the state will last for many years. Truly, this is because of your diligence, and we will have you reside in the palace of the Heir to the throne, helping us, dispensing benevolence, lending me aid and making up for my weaknesses. Kujiki does not make this a royal edict but a simple, sloppy entry. The next entry in Kujiki is the 28th year, where Keitai falls ill and passes away (recorded as the 25th year in Nihon shoki). The Nihon shoki compiler realized the record of Keitai was thin, mainly concerned with family and internal affairs, and added information from Paekche sources to expand it. Notice that the Shoki version of the Keitai record is roughly 6600 characters in length, but Kujiki’s Keitai account is only 1825 characters (roughly a quarter of the length of Shoki’s record).

Ankan-Senka Mizuno (1968:147–48) highlights the fact that the Keitai record in Nihon shoki fails to record the establishment of an heir to the throne. The pre-ascension story of Ankan records, “In the twenty-fifth year, spring, second month, seventh day, Heavenly Sovereign Wohodo established Ōe as Heavenly Sovereign. That is, Heavenly Sovereign Wohodo divinely departed on this very day.”19 Thus, Keitai established Ankan as heir on his deathbed. This account is suspicious and smacks of editorial fudging. Both records of Ankan and Senka remind one of Book Four of Nihon shoki, where a bare dynastic record is dressed up in a thin negligée and inserted into the chronicle. Table 6.13 outlines the main elements of these two reigns across our three records. As this table illustrates, Kojiki recorded bare information, with Kujiki adding a few more details. Kojiki’s editorial stance appears to be that since Ankan’s marriage yielded no children, it was not necessary to record the name of the queen. It is possible, however, that Kojiki is unwittingly admitting that Ankan never married. Within this bare record Nihon shoki has added other information, mainly from the Ōtomo family. The greatest difference between this record and that of Book Four of Shoki is that familial and genealogical data were well set, whereas Book Four was mainly the work of compilers creating nonexistent dynasties from a smattering of information found in various legends.

186

The Tenmu-Jitō historiographical project

Table 6.13 Ankan and Senka’s reigns Event Ankan’s lineage Ankan’s palace named Ankan marries Princess Yamada Royal granary established Ōtomo Kanamura makes declaration about ruler Quarrel about Musashi governor’s position Ankan had no children Ankan’s death and burial tumulus Senka’s lineage Senka’s palace named Senka’s wife and children Diplomatic office at Port of Na refurbished Senka dispatches military to help Mimana Senka’s death date and burial tumulus

Kojiki ○

○ ○ ○ ○

Nihon shoki

Kujiki

○ ○ ○ ○ ○ ○

○ ○ ○

○ ○ ○ ○ ○ ○ ○

○ ○ ○ ○ ○ ○

Many scholars, including Umezawa (1968, 1976), have conducted detailed research on much of Nihon shoki, focused on the many books that contain titles or hints as to where the compilers gathered the information they included, and I will not retrace that information here. It seems clear that the format of putting a historical account together in 30 volumes was influenced by Weishu of Sanguozhi, which also comprises 30 volumes. Nihon shoki also quotes extensively from this work. While the Chinese practice was one book per reign, Nihon shoki breaks this formatting rule with some regularity. Book Four has eight rulers, while Tenmu’s single reign is split and spread over two books: pre-ascension and post-ascension. I believe it is persuasive to argue that those books with more than one ruler were reconstituted during the Tenmu-Jitō historiographical stage. Consider the following speculative information given in Table 6.14. It is clear that if the Tenmu-Jitō courts desired an official historical record worthy of perusal by domestic and foreign audiences, especially foreign envoys, then it was important to adhere as closely to the Chinese model as possible. We know from research into the process of the importing and assimilation of the Tang legal codes that if the glove fit the hand, the Japanese used it almost in toto, but if there were parts that had no Japanese equivalent, the court ignored these. If the Chinese had no equivalent, then the Japanese, likely with Paekche administrative assistance, invented these.20 While the Japanese are known historically as great imitators, this should not mean that they were hamstrung by the process. Likely, the Bureau of History decided to split Tenmu’s record into two equal halves without much distress at marring the model. One of the final steps in the process of compiling Nihon shoki was to flesh out the skeleton of a chronology and regnal calendar created by the committee under the auspices of Great Minister Isonokami Maro. Figure 6.2 shows how I

The Tenmu-Jitō historiographical project 187 Table 6.14 Possible manipulation of records Book Rulers

Notes

4 7 12 13 15 18 21

Created to allow a longer royal chronology The legend of Yamato Takeru rewritten? Lack of material, so both consolidated? Lack of material relegates him here? Reconstituted in place of a real ruler? Confusion in records? Later addition Consolidated to make room for Suiko’s record?

Kesshi hachidai Keikō/Seimu Richū/Hanzei Ingyō/Ankō Seinei/Kenzō/Ninken Ankan/Senka Yōmei/Sushun

Royal Histories

Peninsular Histories

Family Histories

Kojiki Draft Old Nihongi

Kojiki

Nihon shoki

Ur-Kujiki

Kujiki

Figure 6.2 Source material for Kojiki/Kujiki/Nihon shoki.

believe this process took place. Thus, Kujiki relied on much of the same material that created the historical foundation for Kojiki and Nihon shoki, even going so far as to rely on a draft of the unfinished Kojiki to fill out the mythological sections of the story (cf. Bentley 2006:65). I have followed Tomoda (1969) in seeing an older vestige of the Nihon shoki project having existed long enough to have been quoted in the lectures on Nihon shoki. Some call this Wadō Nihongi,21 but Tomoda (1969:85) has labeled this “Old Nihongi.” Some have argued that Wadō Nihongi is a ghost; that quotes from it are simply source material that was confused with the submission year of Kojiki. I have confirmed Tomoda’s theory by using information in Kujiki that has not been examined in the past (cf. Bentley 2006:41, 48–52). While Kujiki has relied on a number of sources, it is clear that this work was a first draft, an attempt to standardize the fundamental recording of each history. Thus, I conclude by saying that the Bureau of History finished “Old Nihongi” around 710 and it languished until the project to compile the draft of Kujiki was finished. With the death of Minister Maro in 717, the committee shelved the Kujiki draft, but later the committee may have felt reenergized and returned to

188

The Tenmu-Jitō historiographical project

work on Nihon shoki; at this point, they cleaned up the compilation, creating the text we have today. Here I introduce a type of forensic research—a mix of textual forensics and linguistics—that sheds new light on the composition of Nihon shoki. Mori Hiromichi published a provocative work in 1991, Kodai on’in to Nihon shoki no seiritsu (Ancient phonology and the origin of Nihon shoki). In this work, he examined the phonology of the phonograms of the annotative notes and poetry throughout (excluding Books 4, 6, 8, 18, 20, 21, 28, 29, and 30, as these have neither poetry nor notes). Based on his analysis he divided the text of Shoki into two chronologically different groups: an older beta group, and a newer alpha group. A cursory examination of his data shows great promise in how he has divided the text and demonstrates that two different strata of material appear to have existed. I point out, however, that there are several weaknesses to his larger theory discussed below. The fact that the text in Nihon shoki was compiled in (at least) two stages is, I believe, a solid fact. Mori’s controversial conclusion is that the alpha section of Shoki was written by two Chinese individuals. To support this theory, he attempts to show that the beta section has ungrammatical usages, while the alpha section does not. In a later work, Mori (2011:82–84) introduces the work of Pak Mihyŏn, who has conducted thorough research into the use of classical Chinese second-person pronouns (汝・爾・儞) in Nihon shoki. He demonstrates that these graphs are not always used grammatically or socio-linguistically correctly. For example, in grammatical Chinese, 汝 is used by a superior to a subordinate. Pak notes that there are two types of ungrammatical usages found in Shoki: (1) where a person in lower status uses 汝 on someone of higher status; (2) where the second-person pronoun has a title of respect added, like 王 “king” or 尊 “your honor.” There are 26 examples of (1) and five of (2) in Nihon shoki, but none of these appears in the beta section of Shoki. There are 31 examples of 爾 “you (higher to a person of lower status)”: 22 of these appear in the beta section and nine in the alpha section. Seven of these alpha usages appear in Book 19, the Kinmei record that is mainly an account of political intrigue between Paekche, Wa, and other kingdoms on the peninsula, and are all used grammatically. Mori then investigates another sociolinguistically sensitive verb, 對 “give an answer (to a superior).” He has found 98 examples of this in Nihon shoki, and the majority of these are grammatically correct. He then focuses on 11 examples where someone in a higher position is responding to someone who is lower. He notes that the verb in general is glossed kotapu, with no explicit direction, so it was easy to make a mistake (2011:85). It is thus interesting that of these 11 mistakes, 8 belong to the beta section, but 3 also appear in the alpha section. Two of these are found in the Kinmei record, and he believes that the mistakes themselves demonstrate that someone may have tampered with (or edited) the text later (and thus corrupted an originally clean narrative). Below are these three examples from the alpha section: (1) 明王對曰、王頭不合受奴手 (Book 19): King Mei responded, “The head of the king should not pass into the hands of a knave.” Said to an enemy about to decapitate him.

The Tenmu-Jitō historiographical project 189 (2) 余昌對曰、諾、即就圖於臣下 (Book 19): Chang of Puyŏ responded, “That is reasonable,” and he immediately discussed this with his subordinates. (3) 山背大兄王等對曰、如卿所噵、其勝必然 (Book 24): Prince Yamashiro Ōe and others responded, “It is exactly as you (pointing to Minister Miwa Fumiya) have said. Surely I would win.” Each of these three examples depicts someone of royal lineage (King Mei, Crown Prince Chang, and Prince Yamashiro Ōe) responding to people who are not of royal lineage; thus, the use of the verb 對 is incorrect. This evidence muddies the water a bit for Mori, who is trying to paint a picture where a native speaker of Chinese wrote the alpha sections of Shoki, and his theory that someone later went back and edited the text, thus introducing the ungrammatical verb, is difficult to swallow. Clearly, his is not the most elegant solution. Mori (2011:86–88) also introduces the work of Doi Miyuki, who notes that there are instances of compound verbs in Shoki that are ungrammatical. Examples of nonstandard Chinese include compound verbs such as 參向 “proceed in the direction to,” 參上 “proceed upward (to the palace),” and 參赴 “proceed (to court).” Doi persuasively argues that these are actually Japanese verbs converted into Chinese. This is not a Japanese innovation, but was done in imitation of actual, grammatically correct Chinese compound verbs like 參見 “to meet,” 參会 “participate in a meeting,” and 參迎 “come to meet.” The only strange usage is found in Book 25 (alpha section), where 晩參 “arrive late” has no Chinese precedent. Mori (2011:88) concludes that a later hand (read a nonnative speaker of Chinese) added this part. I find these arguments less than convincing, however. On the surface, the beta section was compiled by someone who lacked sufficient skill in classical Chinese compared to the alpha section; however, neither requires the compiler be a specific Chinese person. Using statistics, Mori argues that the beta section of poetry and annotative notes used a phonologically mixed group of phonograms, while the alpha section relied on a clean, streamlined set of graphs that he claims represents Middle Chinese. In general, his analysis is correct. For example, when transcribing OJ k- initial syllables, velar-initial graphs (k-, g-, and k‘-) are found mixed with fricatives (h- and ɦ-) in the beta section, but in the alpha section, only velar-initial graphs are used to represent the same velar-initial syllables in OJ (Mori 1991:122). Based on this evidence, Mori concludes in a later work (quoted in 2011:257): During the Jitō court Xu Shouyan and Sa Hongke began to compose the alpha section of Nihon shoki. Xu Shouyan began writing from Book 14, and he finished writing Book 21 and put down his brush. Sa Hongke composed the text for Books 24–27. During the time of the Monmu court, Yamada Fubito Mikata began work on the beta section. From the seventh year of Wadō (714) of the Genmei court Ki Asomi Kiyohito composed Book 30 of the work. At the same time, Miyake Omi Fujimaro went through both sections of the work [both alpha and beta, JRB] and added quotations from Chinese sources, and a few more bits of information to the text. In this manner the thirty books of Nihon shoki were completed and presented to the throne in the fourth year of Yōrō (720 CE).

190 The Tenmu-Jitō historiographical project Mori claims that these two men (Xu and Sa) from Tang China were able to represent accurately the language of Tang China when they wrote the alpha section of Shoki. The court establishes these men as professors of pronunciation in the fifth year of Jitō (691 CE). According to a quote from the lost history known as Nihon seiki (日本世記), Xu Shouyan came to Japan in the 11th month of the 7th year of Saimei (661 CE). By the time he becomes a professor of (Chinese) pronunciation, he had already been in Japan for 30 years. It should also be made clear that Yamada Fubito Mikata was known for his Chinese knowledge, having studied in Silla. Shoku Nihongi records that in 714 CE, the court ordered Ki Asomi Kiyohito and Miyake Omi Fujimaro to compile a national history (國史). Xu and Sa never appear in connection with the compilation of Nihon shoki. Mori spends a great deal of energy and space taking names of people with paper-thin histories and weaving together an intricate explanation for the composition of Nihon shoki. While we duly credit Mori with noticing a difference in two halves of the text, this carefully constructed explanation for the final product titled Nihon shoki is constructed on a sandy foundation, as the evidence in Tables 6.15 and 6.16 will demonstrate. Mori has based his linguistic evidence on the phonology of Middle Chinese as preserved in Nihon shoki, but this is where a careful linguistic examination begins to cast doubt on his conclusions. The most glaring problem with Mori’s use of Middle Chinese is his continued reliance on the outdated reconstructions of Karlgren (1889–1978), whose Chinese reconstructions were cutting-edge in its day (until the 1950s) but have been superseded by the work of others, such as Pulleyblank, Baxter, Sagart, and Schuessler. With the work of these Sinologists readily available, Mori should update his reconstructions. Marc Miyake (2003a:63) provides a helpful example how using Karlgren’s outdated reconstruction leads to incorrect conclusions. He notes the character 許 “allow” used in Shoki to represent ko2, which Karlgren reconstructed as *xi̯ wo, suggests that phonetically this was *ko [kɔ], with a back vowel that was rounded. However, both innovative reconstructions by Pulleyblank, reconstructing 許 as *xɨǝ (1991:348), and Schuessler, as *hɨɑ (2009:52), demonstrate that the vowel was mid-high and unrounded, suggesting OJ *kɨ. Miyake (2003a:65) also notes that Mori’s reconstructions create “a convoluted vowel ‘system’ containing odd diphthongs” such as *ǝe̯ . Returning to the overall hypothesis of Mori that the distinction between the beta and alpha sections was clear and distinct, we find that this evidence is also muddied. The most conspicuous evidence that this is not as clear-cut as Mori claims is seen in the distinction of representing s- initial syllables in OJ. Miyake (2003a:58) notes, “The alpha section scribes could not even transcribe OJ s consistently.” If Mori were correct that a native Chinese speaker was transcribing Old Japanese syllables according to their native phonology, then we would expect a highly consistent representation. Table 6.15 outlines s- initial graphs as found in the poetry in Books 24–27, supposedly transcribed by Sa Hongke, according to Mori. We must conclude that if this was transcribed by one individual from Tang China, he should have consistently transcribed the initial of any of these three

The Tenmu-Jitō historiographical project 191 Table 6.15 S-initial graphs in Shoki OJ

Graph

Fricative

Graph

Affricate

sa

婆

sâ

si

施 斯 始 世 西 栖

śje sje śɨ śjäi siei siei

左 佐 作 旨 志 之 制

tsâ tsâ tsâk tśi tśɨ tśɨ tśjäi

se

syllables, but there is vacillation between a fricative and an affricate in representing all three syllables (sa, si, se). And yet, based on these foggy data, Mori (1991:126) even claims that he can reconstruct what the actual OJ syllables were: sa was *tsɑ, si was *tʃi, and se was *tʃe: all are reconstructed as affricates, but the fact remains, Mori has simply selected one of the two choices in his data. Mori also passes in silence regarding the variation in the vowels. For si, his supposed solo transcriber vacillates between a mid and a high vowel (/i/, /e/, and /ɨ/). While I do not doubt that native or near-native speakers of Chinese (from either China, Paekche, or Koguryŏ) had a hand in the compilation of Nihon shoki, I find the explanation that two people transcribed the alpha sections of Nihon shoki to be unpersuasive given these data. I find Miyake’s conclusion more persuasive (2003a:60): It is true that more inconsistency can be found in the beta section …, but, if Mori were correct, we would expect far less inconsistency in the alpha section. Mistakes in transcription are inevitable, but it is difficult to believe that Chinese scribes capable of distinguishing between *c, *ɕ, and *s in their own language would write 78.3 per cent (18/23) of OJ se … in the alpha section with the “wrong” initial. Finally, while we acknowledge Mori’s contribution of seeing Shoki’s text divided into two different strata, the mistaken assumption that the beta section is based on an older phonology and the alpha section is based on the avant-garde Late Middle Chinese (LMC) is due to a lack of global knowledge, or the ability to see the text in a broader context. Both sections are based on LMC, but the beta section is less consistent in its application than the alpha. Consider Table 6.16, which lines up graphs that have undergone a phonological change between Early Middle Chinese (EMC) and LMC. Yet both values appear in the alpha and beta phonograms in Shoki.22

192

The Tenmu-Jitō historiographical project

Table 6.16 EMC versus LMC Graph

EMC (Examples)

LMC (Examples)

麼 莽 弭 縻 慕 娜 士 倍

ma (12) ma (1) mi (0) mi (0) mo (2) na (0) si (0) pey (6)

ba (27) mo (2) bi (8) bi (2) bo (0) da (9) zi (2) po (7)

This chart makes it clear that two different strata of phonograms are woven together, though in general Nihon shoki relies heavily on LMC, in both the beta and alpha sections. However, Mori’s claim of two different groups should not be ignored simply because his theory about a Chinese writer is not tenable. Based on Mori’s evidence for what is clearly a two-stage compilation of Nihon shoki, I infer that one possible conclusion is as follows: By 710, the basic structure of Nihon shoki was in place. As the text of Kujiki demonstrates, the historiographical committee had already constructed the basic structure for the imperial reigns from Jinmu to Suiko. The compilation committee created by imperial decree in 714 with Ki Asomi Kiyohito and Miyake Omi Fujimaro at the head appears to have filled out the record, and completed the last book for Jitō. At this point, the so-called alpha section of Shoki was tweaked. Notice that the evidence for a rewriting concerns the poetry and the annotative notes. This means that it is impossible to say that the alpha or the beta sections were begun at any specific time. To follow Mori and believe that the beta section was not begun until the reign of Monmu (r. 697–707) means that after Tenmu issued his edict to compile a history, nothing happened at court for almost two decades. This contradicts what Shoki relays about two courtiers going to work right after the edict. Granted, that is likely poetic license, but it is difficult to believe that nothing happened for two decades. The glaring weakness in Mori’s theory is the suggestion that several people simply sat down and began to create a work that textual data clearly demonstrates was based on earlier documents with different subcommittees. It is also unfortunate that Mori appears to have ignored work by Matsuo Yoshiki (1987), who postulates that the text of Nihon shoki contains evidence of Tang colloquialisms throughout the work, spanning both the beta and alpha sections. It is undeniable that the compilers/writers of Nihon shoki had exposure to the Tang language in China. The complex nature of Nihon shoki argues for an intricate, multistage process of compilation based on a set of documents that had undergone accretion of several stages of information earlier. If Shoki had truly been the work of just a few people, it should more closely resemble the homogenous text of Kojiki, and not be the eclectic text we have today.

The Tenmu-Jitō historiographical project 193

Notes 1 This chapter is titled after a term I coined over two decades ago (mentioned in 2006:117). Lurie (2001:270) uses the term “Kojiki project,” defining it thus: “In the discussion below, I refer to the Preface’s narrative of royal command and laborious textual production as the ‘Kojiki project;’ my primary goal here is to understand the rhetorical and ideological functions of that project, not to relate it to the main text of the Kojiki itself.” 2 Yamada (1935:157; 1940:134) and Umezawa (1988:35–37) believe that the various “titles” found in the Kojiki preface—Teiki (帝紀) or Teiō hongi (帝皇日継), and Kuji (旧辞) or Honji (本辞), or Sendai kuji (先代旧辞) are iterations of the same text. Yamada (1935:157) even believes that Teiō hongi is essentially a native Japanese version of the Chinese written Teiki. Needless to say, it is difficult to refute these ideas. I would rather draw attention to the fact that these titles suggest there were a variety of written documents already in existence. 3 There is also the theory that the preface to Kojiki is a later addition, and not written by Yasumaro. This theory appeared as early as 1747, put forth by the brothers Kawamura Hidekai and Hidene in their Kojiki kaidai (cf. Suzuki 1967:2). This theory was followed in 1750 by Tada Yoshitoshi’s Nihon shindai ki; Tada was the teacher of the Kawamura brothers. A few other nineteenth-century scholars argued that because the edict given to Yasumaro regarding Kojiki, or any mention of the work, does not appear in Shoku Nihongi, the preface must be a later addition. However, with the discovery of Yasumaro’s epitaph in early 1979, the majority of these doubts were laid to rest. Regardless, there are those who still consider the preface to be either fraudulent or a later addition; cf. Saijō 1998:99–125, Miura 2008:1–6, Ōwa 2010:439–441, and Bodiford 2013:284–89. 4 Fusō ryakki notes that Umako was 76 years at his death. If true, he would have been born around 551. 5 This may have been one reason that Kujiki was never finished. If the work had yet to be completed when Maro died, and the court of Genshō found the portrayal of the Soga to be too critical, the project may have been shelved. 6 The peninsular title is 吉師 kisi. The individual sent from Paekche to teach the Ōjin court is recorded in Kojiki as Wani Kishi. Kishi is also the 14th level in the Silla bureaucratic government. 7 Originally known as Princess Takara (594–661), she ascended the throne at the death of her husband, Jomei (r. 629–641), and is known as Kōgyoku (r. 642–645). The coup led by her son, Naka no Oe, in 645 that overthrew the Soga, caused her to abdicate the throne. With her abdication, Prince Karu ascends the throne as Kōtoku (r. 645–654). With Karu’s death, Princess Takara reascends the throne, this time known as Saimei (r. 655–661), and she reigns until her death. 8 Viewed from Jitō’s perspective, as the heir Tenmu left was not Jitō’s son, but the son of a different wife. 9 Hirata (1959:164–170) follows the theory that the Shoki compilers inserted a female ruler to account for the existence of Himiko in the Weizhi account, whose reign overlapped with that of Chūai, after the royal chronology was stretched back in time. 10 It needs to be mentioned that these are not uniquely Kobayashi’s ideas. Others have come to similar conclusions; cf. Maruyama 1955:213–220, 273–78; Hirata 1959:165– 180; Tanaka 1972:417–22. It should also be noted that some have argued that the four entries in the record of Jingū (39th, 40th, 43rd, and 66th years), quoting from Weizhi and the Chinese diurnal record, are additions by a later hand (cf. Murakami 1951, Wada 1953), but Tanaka (1972:431–32) persuasively argues that these entries were in the original manuscript. 11 Ōyama (2009:110–11) believes the actual king of Wa at the time the Sui visited was Soga Umako. This is an intriguing idea. It is true that Umako gave audience to the

194

12 13 14 15 16

17

18 19 20 21

22

The Tenmu-Jitō historiographical project

visiting delegation from Silla and Mimana in 601 (cf. Ōyama 2009:110), acting as if he was the head of state. However, I find it difficult to square the large disparity in the names. Morita (2005:54–56) notes that it is clear Umako had a ruling king (Sushun) assassinated, but there is no evidence he wanted to take the throne. Bold denotes the person who ruled. The symbol = shows marriage. A strikeout shows the ruler was assassinated. Jitō’s mother was Ochi no Iratsume, the oldest daughter of Soga no Kura Yamada Maro. Maro is a grandson of Umako. See the explanation in Bentley (2002:111, n 24). The current work of Schuessler (2009:125, 123) posits 彌 as Later Han *mie and 売 as *mɛ. Thus, Kojiki preserves the older rendition (Sume). There was a tradition that a three-legged bird resided in the sun. Thus, 金鳥 is a poetic way to say “the sun.” Its shining on the roof of the western hall demonstrates that the sun is setting, an almost universal image for death. Wittkamp (2018:55) brings up a crucial example, where Kojiki makes a distinction between 生 “give birth” and 成 “come into existence.” When Izanagi and Izanami decide to have intercourse and create the islands, the verb is written in phonograms, 美斗能麻具波比 mitwo no magupapi. It is written this way as “characters such as 生, which have the same meanings as in Chinese but are read in Japanese (umu).” This is a criticism we can also level at Heldt’s (2014) translation of Kojiki. He translates almost all names into English. This choice is based on the misleading and tenuous assumption that we know what many of these names mean. It is true that some names are transparent, while others have etymological stories added to explain their origins. However, many semantic representations are simply attempts to render names that the Nara era compilers may not have accurately understood, and then these representations become fodder for the modern translation. Yokota divides A into three subsets, but I have grouped these as one. He also ignores books that do not contain any information about the birth of royal children (i.e., Jingū, Suiko, and so on). Kujiki makes the same ridiculous claim, but the grammar is not as tight as that in Nihon shoki. By “inventing” it may be more precise to say that the court took an organ or bureau that already existed and reworked it so that it fit within a Chinese-based system. The bureau within which the native, court-centered religion was placed is an example. Ban Nobutomo (1773–1846) first argued for the existence of a manuscript of Wadō Nihongi submitted to the court in Wadō 7 (714), basing that argument on the fact that basic information about Prince Ōtomo was found in records like Kaifūsō and Fusō ryakki that did not appear in Nihon shoki (1907, 4:463–64). Hirata (1959:197–98) argues that the date should be Wadō 5 (712). He concludes, however, that the date Kojiki was submitted to the court (712) and the basic genealogical material that was used to create Nihon shoki were confused in the mind of the compiler of Fusō ryakki (1959:207). EMC data are for the alpha section, and LMC is for the beta section, demonstrating that these two sections are not truly homogenous as Mori claims.

References All works are published in Tokyo unless otherwise noted. Ban, Nobutomo. 1907. Ban Nobutomo zenshū, 5 Vols. Kokusho Kankōkai. Barnes, Gina. 2001. State Formation in Korea: Historical and Archaeological Perspectives. Richmond: Curzon. Bentley, John R. 2002. Historiographical Trends in Early Japan. Lewiston: Edwin Mellen Press.

The Tenmu-Jitō historiographical project 195 Bentley, John R. 2006. The Authenticity of Sendai kuji hongi: A New Examination of Texts, with a Translation and Commentary. Leiden: Brill. Best, Jonathan W. 2006. A History of the Early Korean kingdom of Paekche. Cambridge: Harvard University Asia Center. Bodiford, William M. 2013. “Myth and Counter-Myth in Early Modern Japan.” In Writing Down the Myths. Brepols: Turnhout, Belgium, pp. 277–305. Chamberlain, Basil Hall. 1981. The Kojiki or Records of Ancient Matters. Tuttle. Clements, Jonathan. 2007. Wu: The Chinese Empress who Schemed, Seduced, and Murdered her Way to Become a Living God. Stroud: Sutton Publishing. Fujii, Nobuo. 1952. “Nihon shoki kakukan seiritsu ikkōsatsu.” Ōkurayama ronshū, 1(1): 106–21. Hanzawa, Eiichi. 2002. Shōtoku Taishi hōō Waōron. Contained in Vol. 24 of NSK. Hanawa Shobō, pp. 299–378. Heldt, Gustav. 2014. The Kojiki: An Account of Ancient Matters. New York: Columbia University Press. Hirata, Toshiharu. 1959. Nihon koten no seiritsu no kenkyū. Nihon Shoin. Kobayashi, Toshio. 2006. Nihon kodai kokka keiseishi-kō. Kōsō Shobō. Kojima, Noriyuki, Kōjirō Naoki, Kazutami Nishimiya, Susumu Kuranak and Masamori Mōri, eds. 2002. Nihon shoki. Shōgakkan. Kōnoshi, Takamitsu. 1995. Kojiki: tennō no sekai no monogatari. Kōdansha. Kōnoshi, Takamitsu and Yoshinori Yamaguchi, eds. 1997. Kojiki. Shōgakkan. Ledyard, Gari. 1975. “Galloping along with the Horseriders: Looking for the Founders of Japan.” Journal of Japanese Studies, 1(2): 217–254. Lurie, David B. 2001. The Origins of Writing in Early Japan: From the 1st to the 8th Century C.E. Columbia University Ph.D. Dissertation. Maruyama, Jirō. 1955. Nihon shoki no kenkyū. Yoshikawa Kōbunkan. Matsuo, Yoshiki. 1987. “Nihon shoki to Tō-dai kōgo.” Wakan hikaku bungaku, 3: 1–18. McBride, Richard D. 2008. Domesticating the Dharma: Buddhist Cults and the Hwaŏm Synthesis in Silla Korea. Honolulu: University of Hawaii Press. Mitani, Eiichi. 1980. Kojiki seiritsu no kenkyū. Yūseido Shuppan. Miura, Sukeyuki. 2008. Kojiki o yomu. Yoshikawa Kōbunkan. Miyake, Marc Hideo. 2003a. Old Japanese: A Phonetic Reconstruction. New York: Routledge. Mizuno, Yū. 1968. Nihon kokka no seiritsu. Kodansha. Mori, Hiromichi. 1991. Kodai on’in to Nihon shok no seiritsu. Taishūkan Shoten. Mori, Hiromichi. 2011. Nihon shoki: seiritsu no shinjitsu. Chūō kōron shinsha. Morita, Tei. 2005. Suiko-chō to Shōtoku Taishi. Iwata Shoin. Murakami, Keiichi. 1951. “Nihon shoki henshu ni kansuru shiken (1): Chūgoku shiseki to no kanren ni tsuite.” Tenri daigaku gakuhō, 5: 1–22. Ōishi, Yoshiki. 1975. Nihon ōken no seiritsu. Hanawa Shobō. Ooms, Herman. 2009. Imperial Politics and Symbolics in Ancient Japan: The Tenmu Dynasty, 650–800. Honolulu: University of Hawaii Press. Ōwa, Iwao. 2010. Nihon shoki seiritsu-kō: Tenmu, Tenchi ifu kyōdai-kō. Daiwa Shobō. Ōyama, Seiichi. 2009. Tenson kōrin no yume: Fujiwara Fubito nado no purojekuto. NHK Books. Philippi, Donald L. 1968. Kojiki. University of Tokyo Press. Piggott, Joan R. 1997. The Emergence of Japanese Kingship. Stanford: Stanford University Press. Pulleyblank, EG. 1991. Lexicon of Reconstructed Pronunciation in Early Middle Chinese, Late Middle Chinese, and Early Mandarin. Vancouver: University of British Columbia Press.

196

The Tenmu-Jitō historiographical project

Rothschild, N Harry. 2008. Wu Zhao: China’s Only Female Emperor. New York: Pearson Longman. Saijō, Tsutomu, 1998. Kojiki no mojihō. Kasama Shoin. Schuessler, Axel. 2009. Minimal Old Chinese and Later Han Chinese: A Companion to Grammata Serica Recensa. Honolulu: University of Hawaii Press. Sugano, Masao. 1992. “Tenson kōrin shinwa saisetsu.” In Hirotoshi Nakamura, Aoki Shūhei, Sugano Masao and Yamazaki Masayuki, eds. Umezawa Isezō Sensei tsuitō kiki ronshū. Zoku Gunsho Ruijū Kanseikai, pp. 93–116. Sugano, Masao. 2012. Kojiki—shinwa to tennō o yomitoku. Shin Jinbutsu Ōraisha. Suzuki, Shōzō. 1967. “Kojiki gishosetsu no rekishi to sono igi ni tsuite.” In Rekishi kenkyū, Vol. 5. Ōsaka Kyōiku Daigaku Rekishigaku Kenkyūshitsu, pp. 1–23. Tanaka, Takashi. 1972. “Jingū Kōgō-ki o meguru ki-ki no Shoden: toku ni Jingū Kōgō no seiritsu ni tsuite.” In Jingū Kōgō ronbunshū kankōkai, ed. Jingū Kōgō. Ise: Kōgakkan Daigaku Shuppanbu, pp. 401–445. Tomoda, Kichinosuke. 1969. Nihon shoki seiritsu no kenkyū. Kazama Shobō. Umezawa, Isezō. 1968. Kiki hihan: Kojiki oyobi Nihon shoki no seiritsu ni kansuru kenkyū. Sōbunsha. Umezawa, Isezō. 1976. Zoku Kiki hihan: Kojiki oyobi Nihon shoki no bunkenteki sōgo kankei no kyūmei. Sōbunsha. Umezawa, Isezō. 1988. Kojiki to Nihon shoki no seiritsu. Yoshikawa Kōbunkan. Wada, Hironori. 1953. “Jingū Kōgō-ki no Wa joō chuki ni tsuite.” Shigaku zasshi, 62: 53–65. Waltham, Clae. 1971. Shu Ching: Book of History, a Modernized Edition of the Translations of James Legge. Chicago: H. Regnery Co. Wittkamp, Robert F. 2018. “The Body as a Mode of Conceptualization in the Kojiki Cosmogony.” Tōzai gakujutsu kenkyū kiyō, 51: 47–64. Yajima, Izumi. 2011. Kojiki no moji sekai. Yoshikawa Kōbunkan. Yamada, Yoshio. 1935. Kojiki jobun kōgi. Shiogama-machi, Miyagi Prefecture: Shibahiko Jinja. Yamada,Yoshio. 1940. Kojiki gaisetsu. Chūō Kōronsha. Yokota, Ken’ichi. 1966. “Nihon shoki keifu kisai no shokeishiki.” Nihon shoki kenkyū, 2: 93–129.

Conclusion

The birth of Japanese historiography is shrouded in a fog that is difficult to pierce. The complex processes for early historical writing that I have argued for are speculative, as there is only a thin layer of supporting evidence one can produce. Having admitted that, there is very compelling archaeological evidence that a “Wa sphere” existed, encompassing at least both the southern coast of the Korean peninsula and western Japan in the third and fourth centuries. As Park (2018) has meticulously shown, there was a clear cultural and technological interchange between Wa and Kara (Kaya) from the third until the fifth centuries. Excavations from tombs in areas controlled by Paekche have produced evidence that links Yamato and Paekche both politically and technologically, but only after the fifth century. Park has also demonstrated that there was an active interchange between Yamato and Silla during the fifth and early sixth centuries, which contradicts the narrative in Nihon shoki, but which finds support in Kujiki. Many have argued that Paekche tutors educated Yamato in the art of writing, myself included, but my purpose here has been to demonstrate systematically how the seeds of writing blossomed into record keeping and finally historical writing. A proper view that individuals from the peninsula, particularly Paekche, were instrumental not only in the teaching of writing to the early Yamato court but also in its application will help scholars and students understand how this technology was able to dovetail with recording important facts about people at court—such as family history—and lead to the creation of a historical record. That statement leads to the reasonable conclusion that early strains of Japanese historiography were weaving together traditional Japanese oral histories with written samples of Paekche history. The inescapable fact that keeps rearing its head is that Kara and Paekche exerted a more profound influence on early Japanese culture than has been appreciated so far. I argue that this was a type of mimesis, but instead of imitating China in the beginning, as many seem to believe, I have argued that Yamato imitated Paekche and its entire historiographical apparatus for a century. China remains the great model, but Paekche is a filter through which Yamato saw the world of historical writing. As I have theorized throughout this monograph, I envision the creation of history in Japan as being a multifaceted project that continued with each reign: something Paekche imitated from its own Chinese models. Because the ruler in Yamato

198

Conclusion

shared power with influential families at the early court, necessity would have created a variety of records over time. With the intermarriage of daughters from powerful families into the royal line, records would have emerged from various branches of this ever-growing royal tree. Kojiki and Nihon shoki mention a limited number of titles of earlier works, but internal evidence from Nihon shoki requires us to consider that the number of existing records in the latter years of the Asuka era (ca. 600–710 CE) was greater than anyone now entertains. As I have argued, early historiography in Japan was centered on genealogy and family history. The desire to have one’s genealogy connect to a distant ancestor was an attempt to create a precedent for the authority of the current chieftain, and later families leveraged these distant ancestors, morphing some into deities. This movement initially had the innovative benefit of placing each powerful family on equal footing with the ruling family and its divine ancestor; over time, historiography surrounding the ruling family created a pantheon of deities, casting a heavenly family roughly mirroring society as it existed in the islands, but projected back in time to the High Plain of Heaven. The multiple versions of mythology found in Nihon shoki originated from various families preserving myths that were variations of strains of divine stories found throughout the islands. I have hypothesized that during the years of the first half of the seventh century, there was a diverse assortment of histories in existence. I also argued that Tenmu’s edict lamenting the fabrication and manipulation that appeared in some histories in possession of certain families may have also been a lamentation about the lack of unity, cohesion, or organization of these histories. This appears to have been a pretext for the Tenmu court to exert control over not only the physical organization of the court but also the records that existed. By reorganizing the structure of the court, Tenmu found a reason to argue for a reconstruction of the historical narrative. This standard was later lifted high by Tenmu’s wife, Jitō, after he had passed away. Such a movement to rewrite history and tradition faced resistance. A minor reaction against the Tenmu-Jitō historiographical project appears to have been reinvention of “traditional histories,” based on the perception of some factions that a varied and colorful cache of earlier histories were in danger of being swept away; these oppositional factions feared that the rich histories of Yamato were about to be sausaged into one bland record. Thus, Kojiki was compiled based on an oral narrative and recorded in such a way that it could be recited. This placed Kojiki in opposition to Nihon shoki, which was written in literary Chinese, and thus stood shoulder to shoulder with Chinese and Paekche histories. That three records (Kojiki, Kujiki, and Nihon shoki) travel much the same path demonstrates that Jitō and her committee struggled to create their narrative, but in the end, they were successful in producing the basics of a royal narrative that came to be the voice of the court. As time marched on, various pre-Tenmu records were destroyed or lost. What presently remains are three histories telling roughly the same story; careful scrutiny of these records, however, reveals a small window into the past, allowing us to peer into the process of historical writing in early Japan. My work has tried to climb through this small window to see into the past.

Conclusion

199

This monograph has argued that the beginning stages of historiography in Yamato were an amalgamation of family lineage and traditional history mapped on to a chronology based on Paekche records. This is a natural development due to the reliance on peninsular scribes, mainly people of Chinese descent who came from Kara and Paekche. Many of the “warts” found in Kojiki and Nihon shoki— such as rulers living a long time, the history starting centuries in the past, and families claiming divine lineage—were either vestiges from peninsular traditions or “skills” learned from the peninsular tutors. I have also strenuously argued that the instrument of writing, literary or classical Chinese through a Paekche filter, not only was the method of recording these histories in the beginning but also influenced the type of histories that the court would produce. In spite of the fragmentary excavated evidence on the peninsula and the islands, the stark similarities between writing in Koguryŏ, Paekche, Silla, and Asuka and Nara Japan force one to conclude that peninsular tutors were heavily involved in teaching those at the Yamato court the art of writing. I earlier had observed that by 662 CE, with the destruction of Paekche by the combined Silla-Tang forces, a wave of immigration from Paekche royal and noble families flooded into Japan, and this energized “literacy” in the center of Yamato (Bentley 2012:384). However, there is a tendency to underappreciate this event, perhaps because most scholars do not appreciate that there already was a strong technological connection between the two countries going back a century. While it is difficult for many Japanese to swallow, the undeniable reality is that Paekche was a far more advanced society technologically than Yamato and had much to offer Yamato throughout their long relationship, and yet the picture is still not balanced. I have also argued that based on archaeological findings, where it is clear there were shared cultural and technological exchanges going on in the “Wa sphere,” a foundation for literacy was slowly established over time. The Wa in Japan had an intercultural relationship with Mimana and Kara, energized through ties to immigrants from the peninsula. As events on the peninsula change because of politics and war, Wa’s relationship with Kara later shifts to Silla, and then finally includes Paekche. Paekche tutors thus came in and built on a rudimentary foundation originally laid by literate individuals from Mimana and Kara. Thus, Paekche did not start from scratch in educating the Yamato elite. Literate immigrants were already at court and had offspring who were also literate. The story about Paekche teaching the Japanese to read and write as recorded in Kojiki and Nihon shoki was simply a tradition based on an observation of the end product, not a statement about the overall process. I have also argued that much of the evidence needed to trace the steps back to the beginnings of historiography in Japan has been erased or obscured by several layers of textual manipulation. This argument has relied on linguistic and philological evidence in an attempt to uncover the various archetypes of the ancient “king lists” and the family and royal lineages with their attending histories. The conclusion in almost every chapter is essentially the same: Yamato owed a great debt to peninsular language, models, and technology in creating and maintaining their varied histories. Even when the courts of Tenmu and Jitō were busy

200

Conclusion

rewriting and recreating the official history, the work of literate individuals who could trace their genealogy back to the peninsula continued uninterrupted. Kojiki appears to have been compiled against the backdrop of a nativist movement, and the committee titled it thus the “record of ancient events” to try washing out, or at least diluting, these peninsular influences. Nihon shoki, on the other hand, embraces the peninsular template and models of the earlier histories, even imitating Paekche records’ penchant for quoting liberally from Chinese sources. Some have argued that the beautiful prose that adorns the record of Nihon shoki came from the direct assistance of native Chinese scholars being involved in the final product. Despite its rather clean Chinese prose, the theory that Nihon shoki was written by two Chinese scholars has been shown to be untenable based on linguistic evidence. It is clear that highly educated scribes were involved in the compilation of Nihon shoki, but the vast amount of data and the varied writing styles argue for a much more eclectic method of compilation. This monograph has also highlighted trends showing the compilers sought continental stories to color the native narrative. There was the penchant the Bureau of History had over time to imitate events from the history of Paekche, and then later that of China. One peninsular example is the influence of the stories of a wicked king in Paekche on the fabrication of Great King Buretsu, portrayed as a wicked ruler who was punished with no posterity. Thus, the evil works of Buretsu are rewarded by punctuating the end of his reign with no heir, allowing a fifthgeneration grandson of Great King Ōjin to take the throne. The record is rewritten so that a possible upheaval in Yamato is smoothed over by an ideological change in power: heaven punished an evil ruler by granting him no heir. An example of a model from China concerns the circumstances surrounding Emperor Wu Zhao, which influenced the early Nara historians in restructuring the history of Suiko. I have argued that Suiko was historically a regent when young Jomei was on the throne, operating under the umbrella of the powerful Soga Umako. Historians under Jitō and perhaps later Genmei’s court flipped the two roles, making Shōtoku Taishi the regent and Suiko, the ruler, thus setting a precedent for the politically active ruler Jitō. This manipulation also pushed Jomei’s reign further into the future. Referencing events from Wu Zhao’s life, Suiko is made ruler after the death of her husband, Bidatsu, and the short reigns of two other kings (Yōmei and Sushun). This mirrors the trajectory of Wu Zhao, who ascended the throne to become China’s first female sovereign only after the death of her husband (Gaozong) and two of her sons (Li Zhe and Li Dan), who both briefly reigned. Suiko instituted (via Shōtoku Taishi) changes at court, was a staunch supporter of Buddhism, and acted independently of Umako, all in imitation of the dynamically strong female ruler, Wu Zhao. This alignment of Suiko with the powerful Wu Zhao underscores what I have argued: that Jitō’s court pushed an agenda where the historical narrative was reconfigured, granting female rulers greater power and influence, perhaps in an attempt to recapture what had been the pervasive influence of magical and ritualistic power in earlier society by certain females. While Queen Himiko of the primitive federation of Yamatai is not mentioned specifically by name in any of

Conclusion

201

the three records, the tradition of female authority and influence is difficult to ignore. The most systematized embodiment of this movement appears in Nihon shoki, where the sun goddess, Amaterasu, is the authority in the mythical stories, and three female rulers (Jingū, Suiko, and Jitō) appear at roughly equal intervals in the record, ending with Jitō who styled herself and her children after the family of Amaterasu. From beginning to end, this early movement to create historical records in Yamato was funneled and fueled by immigrant technology, informed by peninsular and continental ideologies, and even buttressed with an innovative vocabulary that was a mixture of native, peninsular, and Chinese terminology. By the beginning of the eighth century, Yamato’s Bureau of History had matured, graduating from the shadow of Paekche, and was creating its own history, albeit still borrowing material from its neighbors. What is fascinating is that with the inception of the next historical record, Shoku Nihongi, the reality of the struggles of society comes into focus, and the awe and mystic of the royal family and their heavenly lineage blends into the background. The creativity and manipulation found in Nihon shoki vanishes, and two centuries of rather dry chronicles begin.

References Bentley, John R. 2012. “Review of David B. Lurie, Realms of Literacy: Early Japan and the History of Writing.” Harvard Journal of Asiatic Studies, 72(2): 381–388. Park, Cheun Soo. 2018. “Kaya, Silla, and Wa: Relationships and Their Historical Backgrounds.” In Mark E Byington, Ken’ichi Sasaki and Martin T. Bale, eds. Early Korea-Japan Interactions. Early Korea Project, Korea Institute. Cambridge, MA: Harvard University Press, pp. 133–181.

Index

alpha vs beta sections of Nihon shoki 188–191 Amaterasu 59, 62, 74, 80, 84, 88–89, 158, 164, 170–71, 201 Anahobe, Prince 163–64, 171 The Analects 26, 33, 36, 46–48, 68, 144 Ankan 24, 63–64, 102, 122, 125, 128, 130 181, 185–86 Asin, King 47, 106 Babylon 44–45 be (部) 28, 65–66 Bei Qing, ambassador 96–97, 134, 161, 163 Bidatsu 67, 96, 118, 132, 133, 141, 161, 164, 166, 200 Bureau of History 77, 88, 99, 101, 103, 108, 117, 121, 123, 134, 143, 156, 163, 165–67, 182–83, 186–87, 200 Buretsu 24, 200–04, 112, 122–23, 132, 183–84, 200 China 5, 9, 11–12, 25–27, 32, 37, 45, 49, 55–57, 61, 68, 87–88, 95, 100, 112, 119–20, 125, 134, 143, 157, 159, 164– 65, 168, 175, 184, 190–92, 197, 200 Chūai 64, 96, 102–04, 121, 134, 140, 142, 158–59, 178, 181–82 Chumong, founder of Paekche 52 Chūzan seifu 78–79 Chūzan seikan 78–79 Commanderies 9–14, 19, 21, 28, 35–36, 45, 52–53, 55, 61, 67–68, 108–09, 118 Confucius 46–48, 68 Daifang 11–15, 18, 36, 52, 108–09 Emishi 132–33, 135, 158 Eta Funayama Sword 20, 22, 24–25, 27, 99

Fudei 11, 45 Fune Obito Esaka 132 Fusō ryakki 142, 193, 194 Genmei 2, 102–03, 152, 167, 169, 183, 189, 200 Genshō 64, 102–03, 167, 193 Han (韓) 11, 38, 53–55, 89, 91–92; see also Kara Han China 9, 11, 68 Hanshu 10–11 Hanshu xuli 117 Hansŏng 13, 15, 40, 45, 106 Hanyuan 65, 69, 113 Harima fudoki 136–37 Hieda no Are 2, 126, 157, 160, 168, 181 Himiko, Queen 13, 40, 80, 92–93, 158–60, 193, 200 historical linguistics 1; see also linguistics historiography 1, 4–5, 9, 48–53, 56, 68, 76, 95, 99, 105, 108, 113, 118, 124–25, 132, 137, 146, 148, 158, 162, 197–99 Hitachi fudoki 136 Hou Hanshu 11–12, 53, 55, 68, 177 Huainanzi 72–73 Hyūga (Himuka) 89–92 Idu 25 Inariyama Tumulus sword 20, 22, 24–25, 27, 56–57, 60, 62, 99, 101 Ingyō 19, 118–19, 124–25, 130, 166 inscriptions 16, 18, 20, 22–25, 27–29, 32, 36, 57, 60–62, 65–66, 99–101, 119 Isonokami Maro 5, 155, 168, 181, 186 Iwai, revolt of 127–28, 184 Izanagi 74–75, 77, 88, 131, 169, 194 Izanami 74–75, 77, 80, 88, 131, 169, 194 Izumo fudoki 60, 65, 78, 82–83

Index  Izumo myths 6, 74, 77–78, 80–84, 88 Izumo Takeru 84 Jingū 7, 24, 40, 8, 102–03, 107, 121, 135, 140, 143, 158–61, 167, 170, 181–82, 201 Jinmu, 75, 78, 89, 92–93, 100–05, 131, 138, 143–45, 158, 160, 177, 181, 192 Jinnō shōtōki 142 Jinshin Disturbance 87, 143, 155, 159, 181 Jinshu 13 Jitō 7, 58, 62, 64, 78, 96, 103, 105, 143, 145–46, 148, 152, 154–55, 157–59, 161–68, 169–74, 182–83, 186, 190, 193–94, 198–201 Jōgū Shōtoku hōō teisetsu 80, 117, 141, 149, 164 Jōgūki 36, 123, 146–48, 164, 184 Jomei 64, 96, 101–03, 112, 132, 138, 140, 148, 157, 161, 163, 166–67, 193, 200 Kaeriten 35 Kara 7, 11, 13–14, 17, 19, 38, 45, 53–55, 89, 91–92, 107, 111, 113, 118–21, 128, 133–35, 149, 197, 199; see also Kŭmgwan Kaya Karlgren, Bernhard 190 Karu, prince 103–04, 124–25, 171, 193 Kataribe 60, 65, 82, 101–02, 120–21, 123, 126, 166, 168, 172–73, 181 Kaya see Kara Keikō 67, 96, 101–04, 121, 136, 142, 181, 187 Keitai x, 6, 24, 64 100–05, 107, 114, 118, 120, 122, 126–28, 130–32, 138, 141–42, 144, 146, 148–49, 157, 165–66, 183–85 Kenzō 85–86, 100, 102, 104, 113, 130–31, 142, 145, 155, 165–66 king lists 2, 6, 49–50, 99, 101, 199 Kinmei 6, 63–64, 85–86, 96, 102, 105, 107, 116–18, 122, 124–28, 130–31, 133, 136, 140–41, 146–48, 153, 155, 157, 166, 184, 188 Koguryŏ 6, 14–15, 17–19, 27–28, 40, 45–46, 51–53, 56, 62, 68–69, 85–86, 88, 100, 106, 108–09, 120, 122–23, 126, 128, 130, 135, 149, 159, 181–83, 191, 199 Kojiki 2–7, 9, 19, 23, 32, 35, 46–48, 52, 56–57, 62–66, 72–75, 77–80, 82–89, 91–93, 96–102, 113, 116, 118, 121, 123–24, 126–28, 130, 134, 136–38, 141–44, 146–49, 152, 155–59, 164–74,

203

177–78, 181–87, 192–94, 198–200; preface 2–4, 35, 57–58, 99, 116–17, 143, 149, 152–53, 156–57, 169–70, 172–73, 193 Kokki 99, 129, 132–33, 149, 152 Korean peninsula 6–7, 9–16, 18–20, 23, 25–28, 32, 36–38, 40, 44–45, 51, 53–56, 60–63, 76, 86–89, 91–92, 100, 103, 106–08, 113, 118–21, 123–29, 133–35, 140, 156–58, 160, 167, 182, 184, 188, 197, 199–201 Kotodama 1 Kuji 99, 106, 149, 152–53, 193 Kujiki 3–6, 72–74, 79–80, 86–87, 89, 91– 93, 96–98, 101, 116, 122, 124, 126–27, 130–31, 135–36, 138, 142, 148–49, 152, 155, 158, 162–64, 167, 169, 174, 182–85, 187, 192–94, 197–98 Kumanari (Ungjin) 33, 35–36, 40, 168, 173 Kumaso 89, 91–92, 103, 121, 130, 135, 158 Kŭmgwan Kaya 19, 149 Kundoku 20, 33, 35–36, 58, 60, 172–73 Kungana x, 35, 37, 62 Kusakabe, Prince 156, 170–71 Kwanggaet’o Stele 12, 16, 18, 45, 100, 159 Kyŏngdŏk, King 32 Kyūshū 11–12, 14, 19, 44, 74–75, 77, 81, 89–92, 100, 119, 128, 134, 156, 159 Lelang 9–14, 36, 53, 55 Liangshu 55 Lineage 2, 6, 37, 53, 56–57, 64, 67, 87–88, 95–96, 99–100, 104, 110–12, 118–19, 122–23, 125, 128–29, 134, 152–53, 155, 162, 164, 171–73, 176–78, 183, 189, 199, 201 linguistics x, 1, 9, 25, 34, 36, 56–57, 60, 62–63, 113, 117, 164, 173, 188, 190, 199–200 Lunheng 51, 109, 144 man’yōgana 20 Man’yōshū 1, 34, 82, 99, 171 Mimaki 102, 104, 134 Mimana 13–14, 17, 19, 45, 67, 85, 126–27, 133–34, 182, 184, 199 Mokkan 23, 26–27, 29–32, 58, 64–65, 69 Mononobe 64–66, 105, 124, 127, 129, 131–32, 134, 138, 153–55, 157, 164, 168, 182, 184

204 Index Mononobe Arakahi 127–28, 184 Mononobe Moriya 153–54, 157, 164 Motoori Norinaga 113, 126, 174 Muryŏng, King 24, 106, 110–11 Nagasune Hiko 87, 92 Names 64–65 Nigi Hayahi 75–76, 87, 92 Nihon shoki 2–7, 9, 15, 19, 23–25, 34–35, 38–40, 46–47, 56–58, 61, 63–68, 72–75, 77–82, 84–89, 91–93, 96–102, 105–07, 112–13, 116–18, 121–29, 132–34, 136, 138, 141–49, 152–56, 158–64, 167–72, 174–92, 198–200 Ninigi 72, 74–77, 89, 91–92, 112–13, 170–71 Nintoku 46, 56, 64, 77, 101–04, 111–12, 122, 124–25, 131, 136–37, 143, 146, 166, 177–78, 181–84 Ō no Yasumaro 2, 4, 32–33, 57–59, 62–63, 99–100, 122–23, 126, 136, 152–53, 155–57, 165, 168–70, 172–74, 181, 184, 193 Ōjin 46–47, 64, 68, 101–02, 104, 111, 118, 123–25, 130–31, 143, 144–45, 159–60, 184, 193, 200 Old Chosŏn 9, 13, 55 Old Japanese x, 21, 30–31, 34–35, 39, 57, 61–63, 69, 73, 96, 117, 123, 144, 190 Ōtsu, Prince 171 Owari 76, 131, 141, 182 P’ungnap Walled Site 15 Paeckche pon’gi 106–07, 122 Paekche x, 5–7, 13–15, 17–19, 23–25, 27– 31, 33–40, 45–48, 50–53, 56–57, 61–62, 65, 67–69, 85–89, 95–96, 100–01, 103, 106–14, 117, 119–20, 122–23, 126, 129, 133, 135, 143, 144, 146, 149, 155, 158–59, 166, 168, 173, 175, 177, 182– 83, 185–86, 188, 191, 193, 197–201 Paekche sinch’an 34, 106 Paekcheki 106–07, 182 Puyŏ 18–19, 29, 51–52, 110, 112, 182, 189 Qian zi wen 46–47 Ryūkyūan 21, 62 Samguk sagi 18, 27, 32, 38, 40, 46, 50, 52–53, 108, 110–12, 114 Sanguozhi 20, 23–25, 176, 186

Sekichey 32, 35 Sendai kuji 193 Sendai kuji hongi see Kujiki Senmyō 168 Seven-Branch Sword 20, 23–25, 39, 40 Shaku Nihongi 3, 7, 123, 146–47, 149 Shangshu 32, 48; see also Shujing Shanhaijing 12 Shiji 25, 27, 34, 49–50, 55, 117, 130, 146 Shijing 32, 55 Shiki 92, 147, 160 Shinsen shōjiroku 67–68, 110–12, 133–34 Shoku Nihongi 2, 66, 100, 111–12, 170, 190, 193, 201 Shōtoku Taishi 36, 80, 96–99, 114, 129, 132, 134, 141, 146–48, 161–64, 166–68, 200 Shujing 48–49 Silla 6, 14, 17–19, 27–29, 32, 38, 40, 45, 50, 53, 61, 82, 87, 89, 95, 103, 07, 109, 113, 118–19, 121, 126, 128, 135, 149, 155, 158–59, 181–83, 190, 193–94, 197, 199 Soga Iname 85, 87, 128, 133, 153, 155 Soga Iruka 162 Soga Umako 98–99, 129, 132–34, 141, 153–55, 157, 161–64, 166, 193–94, 200 Songshu 119, 124–25 spell songs (waza uta) 1 Suda Hachiman Shrine Mirror 20, 22, 24, 65, 101 Suiko 6–7, 64, 96–99, 101, 113, 118, 129, 132, 134–35, 137–38, 140–41, 144, 147–48, 157–58, 161–67, 170, 181, 187, 192, 200–01 Suishu 7, 27, 44, 52–53, 60–63, 95–99, 108–09, 112–13, 147, 161, 163, 166, 177 Sujin 64, 81, 100–102, 104–05, 130–31, 134, 136, 146; see also Mimaki Surnames 64–65 Sushun 96, 141, 155, 157, 161, 164, 194, 200 Takachiho 89, 91–92 Takami Musubi 73–74, 77, 85, 88–89 Tang China 87, 190 Tarasi piko 63, 102, 161 Teiki 99, 101, 105, 114, 117, 149, 152–53, 156–57, 193 Tenmu 2, 4, 7, 23, 35, 58–59, 62, 64, 69, 74, 86–88, 99–100, 102, 113, 116–18, 129, 133–37, 143–46, 148, 152–59, 166–72, 181, 186, 192, 193, 198–99

Index  Tennōki 99, 129, 132–34, 149, 152 Three Han 11, 13–14, 38, 135, 158 Uji Wakairatsuko 46, 133, 136 Ungjin see Kumanari Ur-mythology 6, 77–80 Wa x, 6, 9, 11–14, 15–20, 23–24, 34, 37–38, 44–45, 50, 53–56, 60, 64, 92, 120, 197, 199 Waka Takeru 24, 61, 101–02, 119–22, 123–26, 129; see also Yūryaku Weizhi 12, 15, 20, 27, 38, 63, 68, 159–60, 177, 193 Wenxuan 176–77 Widŏk, King 45, 106 Wu Zetian 164 Wu Zhao 164–66, 170, 200

205

Yamatai 13, 15, 80, 159–60, 200 Yamato x, 2, 6–7, 9, 13–15, 19, 21, 23–24, 27–28, 32–38, 46–47, 49–52, 56–57, 61–64, 66–68, 75–78, 80–92, 95–97, 99–100, 105–08, 110–14, 116–20, 122–26, 128–30, 134–38, 142–46, 148, 155, 157–58, 160, 168, 170, 173, 181–85, 197–201 Yamato neko 63, 102–04, 170 Yamato Takeru 84, 102–04, 121, 128, 130–31, 135, 142, 148, 187 Yōmei 64, 96, 141, 153–54, 161, 163–65, 167, 187, 200 Yūryaku 6, 19, 24, 61, 85, 99–104, 113, 119–21, 124–25, 129–31, 134, 136–37, 139, 140–44, 146, 149, 157, 166, 168, 178, 181–83; see also Waka Takeru