The stars and the state: astronomy, astrology, and the politics of natural knowledge in early medieval Japan

PDF

Download

Open document

Flip pages

Download a page range

Download transcript

Copy asset link

Request this asset

Transcript (if available)

Content

THE STARS AND THE STATE:
ASTRONOMY, ASTROLOGY, AND THE POLITICS OF NATURAL KNOWLEDGE
IN EARLY MEDIEVAL JAPAN

by

Kristina Mairi Buhrman

A Dissertation Presented to the
FACULTY OF THE USC GRADUATE SCHOOL
UNIVERSITY OF SOUTHERN CALIFORNIA
In Partial Fulfillment of the
Requirements for the Degree
DOCTOR OF PHILOSOPHY
(HISTORY)

August 2012

Copyright 2012 Kristina Mairi Buhrman

ii

To my nephew, Noah Robert Deever,
who was briefly enthralled by the idea of the International Date Line
when I explained it in 2010

iii

ACKNOWLEDGEMENTS

Very few projects ever feel completed, and as I contemplate the next stage in my research,
it is a little difficult to see this as the endpoint—or even way-station—on this journey.
Nevertheless, at this moment of completing my dissertation, I would like to take a
moment to acknowledge some of the aid that I have received along the way.
The research for this dissertation was begun under a Mitsubishi/Fulbright
Dissertation Research Fellowship in 2007-2008, for which I am deeply grateful. I would
also like to add my appreciation for the assistance of the staff of the Fulbright office in
Japan who provided invaluable assistance while I was there. Further research in the fall
of 2010 was funded by the Center for Religion and Civic Culture at the University of
Southern California, which generously gave me the opportunity to return to Japan to
examine more materials relating to religion and astronomy. Finally, the Needham
Research Institute in Cambridge, UK provided me with an Andrew W. Mellon
Fellowship, access to an excellent library, and a place to work overlooking the gardens
while I edited this manuscript in 2012.
Many individuals have helped me along the way. My committee, Joan Piggott,
Deborah Harkness, and Lori Meeks most prominent among them. Professor Deborah
Harkness of the Department of History helped me figure out how to write a history of
sciences that did not fit the Western or modern mold. Professor Lori Meeks was not only
an endless source of support for my research and teaching, but also helped me navigate
iv
the Japanese hospital system in 2007, for which she has my eternal gratitude. Thanks
must especially go to my chair, Professor Joan Piggott, who not only patiently bore my
incessant questions about Kamakura-period society before I even began my degree
program, but also bore my dedication to what turned into a project of intellectual history.
The Department of History at the University of Southern California has been
overwhelmingly supportive of my research as a unit, but I would like to take the time to
extend special thanks to Professor Emerita Charlotte Furth, who (perhaps unintentionally)
introduced me to the history of science in East Asia; Professor Josh Goldstein, who was
an excellent teaching supervisor and a particular source of encouragement and
conversation (as were Professors Brett Sheehan, Terry Seip, Azade-Ayse Rorlich, and
Professor Emeritus Jack Wills); and Professor Emerita Carole Shammas, who taught me
quantitative methods and gently steered me away from adding complicated statistical
analysis to my work when inappropriate. I also wish to thank Grace Ryu of the East
Asian Studies Center, who also was a source of cheerful support. The students and
faculty of the University of Southern California are generally due my appreciation, and I
only refrain from listing more names here in the interests of space.
At the University of Tokyo Historiographical Institute, Professors Kato Tomoyasu
(now at Meiji University) and Yoshida Sanae welcomed me into their seminars. Professor
Kato Tomoyasu also graciously supervised me as my project changed dramatically in
2007 and 2008, and I owe him my warmest thanks. Professor Endo Motoo not only
introduced me to revisionist views of the conflict of the Insei Period, but also agreed to
supervise me on my second trip to the Institute in 2010. Professor Endo Motoki
v
entertained discussions of the Nakahara lineage. Many other faculty at the Institute
patiently answered my questions about guchūreki almanacs and paper reuse at court.
Ōhashi Akiko was also invaluable emotional support during my time at the Institute when
I was sick or missing my family, and she put up with my random observations about
Japanese idioms and tendency to show pictures of my nephews to all and sundry. Also in
Japan, Professor Hayashi Makoto of Aichi Gakuen University both encouraged my
research and gave me invaluable advice and contacts, and I am deeply grateful to
Yamashita Katsuaki for twice (once in New York, and once in Japan) taking time out of a
very busy schedule to show me materials and answer questions.
Many other scholars and teachers have crossed my path and aided me in this
project, almost too many to list, but I would like to extend special thanks to Lucia Dolce
of SOAS; and Michael Como and Bernard Faure of Columbia University who, along with
the Center for Japanese Religions, allowed me on very short notice to attend the 2009
Onmyōdō symposium there. To these and others, I wish to give thanks and plead their
continued indulgence.
Finally, I am deeply grateful for the support of my friends and family through this
project. My parents, Robert and Karen, have been nothing but highly supportive of my
desire to do research (even when the topic, I think, bored them). They prodded me to
return to work when I have been stuck, and provided a sympathetic ear and sounding
board for ideas many a time. I would also like to thank my siblings and siblings-in-law,
my brother John and his wife Joslin, and my sister Susannah and her husband Aaron.
John and Joslin did their best to encourage me to relax now and then, while Su talked me
vi
through some writing blocks and edited some early drafts of these chapters. Without their
support and assistance I would not have been able to finish this work.
At this point I would like to express my appreciation and love for my nephews
(and newest niece), Noah and Jamie Deever, and Robbie and Abigail Buhrman, who all
help me to remember what the most important things in life are.

vii
TABLE OF CONTENTS

Dedication ii

Acknowledgements iii

Abstract viii

Preface xi

Introduction 1

Chapter One: Written Above: Authority and Control in the
Practice of Tenmon ( ) in Japan 16

Chapter Two: Leaks in the System: Access and Authority in the
Changing Practice of Tenmon 64

Chapter Three: “Many Secret Texts:” The Origins of Variation and
Lineage Traditions in Japanese Calendrical
Astronomy ( ) 109

Chapter Four: Invisible Presences: Buddhist Calendarists at the
Mid-Heian Court 183

Chapter Five: Competing Calendars: Eclipse Prediction, Shōshō,
and the Establishment of Buddhist Calendrics as an
Independent Tradition in Mid-Heian Debates 218

Chapter Six: “Deathless Texts:” Innovation, Conservatism, and
the Chinese Classics in the Supervision of
Calendrical Knowledge 283

Conclusion: Fragmentation and the Ideal of Unity in Medieval
Japan 332

Bibliography 355

viii

ABSTRACT

This dissertation examines the social factors involved in the practices of observational
astrology ( Ch. tianwen, Jp. tenmon) and calendrical astronomy ( Ch. lifa, Jp.
rekihō) at the Japanese court. The production and monopolization of astrological and
astronomical knowledge had, from the time of the Han Dynasty in China, been part of the
state bureaucracy and one of the signs of legitimate rule. In the seventh century, Japan
too had imported and implemented these state sciences of the Chinese-style imperium.
However, by the twelfth century, while state control of astronomical knowledge
continued to operate at a surface level, within the Japanese court bureaucracy dissent and
debate reigned. A number of lineages and factions cooperated or competed over
astronomical and astrological facts, which resulted in a situation where there was no
unified “truth” about the stars accepted by the majority of elite members of the court. The
political fragmentation and factionalism that characterized the early medieval Japanese
state was also to be found in knowledge about the natural world circulating at court.
The major reason for this fragmentation of knowledge was the diversity of the
population that produced this same knowledge, a population that did not share either a
common identity or definition of practice. Astrological and astronomical knowledge was
no longer produced solely by the technical bureaucrats whose offices had been
established in the eighth-century Chinese-style law codes ( Jp. ritsuryō)—instead,
these officials contested with other legitimate but non-official purveyors of natural
ix
knowledge: Buddhist monks and court scholars and mathematicians prominent among
them. Furthermore, the statements of fact produced by all three of these factions were
subject to critique and revision by members of the top echelon of the court bureaucracy,
the elite nobility. Clearly there were no independent professional fields of astrology or
astronomy in late classical or early medieval Japan.
As a result, specialists of astrology and astronomy employed a number of
strategies to ensure a receptive audience for their work, at least among some members of
the court. Many entered into client-patron relationships with the top level of the nobility,
wherein knowledge and technical skill were traded for economic and social rewards. Two
groups in particular, the members of the Bureau of Onmyō ( Jp. Onmyō-ryō) and
Buddhist monks, cultivated an aura of supernatural power and ritual efficacy. While the
primary goal of this strategy might not have been debates over the stars at court, the use
of this capital is clearly documented in the historical sources. Therefore, the social history
of debates about astrology and astronomy in the Heian (794–1192) court provides
valuable insight into the rise and social perception of the onmyōji ( ), a group of
specialists in divination, exorcism, and apotropaic ritual who loom large in the Japanese
cultural imagination.
In examining the social history of astrology and astronomy in Japan late classical
and early medieval periods—how debates first arose then came to shape the very
practices of astrology and astronomy themselves—this dissertation also demonstrates the
vitality and political importance of these fields in the eighth through thirteenth centuries.
In contrast to previous scholarship on the history of science in pre-modern Japan, this
x
dissertation shows that astrology and astronomy were hardly stagnant during this period.
It becomes clear, therefore, that the pursuit of natural knowledge in Japan, while it did
not develop along expected Western or Chinese trajectories, was still an active part of the
intellectual world in pre-modern Japan. Pre-modern Japan’s “failure” to follow either of
these paths was not in fact stagnation or devolution, but a separate trajectory shaped by
the political and social realities of the early medieval period.

xi

PREFACE

This dissertation is both distant from the project I initially intended and intimately related
to it. While I was living in Yokohama during the winter of 2004-2005, a span of years
beset by typhoons, earthquakes, and train derailments, I became preoccupied with
questions of how states respond to disasters besetting their population. I was already
studying religious and political rituals in pre-modern Japan, and so I turned my attention
to how the court during the Heian Period (792 – 1192), and court and shogunate during
the Kamakura Period (1192 – 1333) succored its people and dealt with anxiety during an
age rife with natural disasters and ominous phenomena. Through analyzing the response
to disasters in the historical record, I hoped to chart the history of the relationship of the
Japanese elite with the general population.
When I began my dissertation research in 2007, however, I quickly discovered
that the historical record was not necessarily reliable in its descriptions of natural
phenomena. Records were not typically kept about nature for their own sake, but for
political or ritual purposes. Information about natural phenomena, including disasters,
was often preserved only insofar as it was implicated in the history of court procedure
and protocol. Where I had hoped to compare disasters that elicited a state-level response
(religious or charitable) to those that did not, I realized it was hopeless—usually, only a
response was ever record. The fact the event was recorded in the first place was
dependant on a bureaucratic, divinatory, or ritual procedure that generated the document.
xii
Attempting to determine how disasters were interpreted through ritual responses
came to appear as foolish as relying on the historical record for a complete depiction of
natural disasters. Severe disasters initiated massive and diverse responses. For example,
in the case of a severe drought, food would be distributed, sutras chanted, criminals
pardoned, shrines cleaned up and offerings made to the deities therein. No single source,
furthermore, described a complete set of responses to any single disaster—the
incompleteness of sources became abundantly clear when sources were read against each
other. If I was to investigate how the Japanese court saw itself in relation to nature and
the population, I would need another approach.
I turned then to a question I had asked my supervisor at the Historiographical
Institute, Professor Kato Tomoyasu, during the 2003 USC Kambun Summer Workshop.
In what was one of my first experiences reading primary sources from the Heian Period
(794 – 1192), we were assigned a section of the Nihon kiryaku ( “Abbreviated
Chronicles of Japan”) about termites invading the palace. The consensus among the
participants had been that this was clearly a bad omen—even though the text was silent
as to what the event had meant, if anything. I haltingly asked Professor Kato after that
day’s session whether everyone at court would have known what the termites had
meant—if there had been a set meaning for this phenomena. He had thought there was a
commonly accepted interpretation, but also could not tell me precisely what it was. When
my uncertainty about the meaning of phenomena resurfaced in 2007, I once again asked
Professor Kato how the members of the Japanese court, or the Japanese population at
large, would have interpreted such an event. He advised me to start looking for my
xiii
answer in records of divination, and recommended in passing that an examination of a
particular form of divination based on astronomical phenomena, the tenmon no missō, or
secret memorial of observational astrology, might provide some answers.
As a result, I found myself, as many historians of natural philosophy have in the
past, starting from astronomy and astrology. Along the way, I uncovered a great deal of
material that had been ignored by previous treatments of the history of science and court
culture in Japan. I also soon began to talk with historians of science about how pre-
modern Japan might be brought into that context. In the historical sources, I found a great
deal of uncertainty and doubt about the meaning of phenomena—uncertainty that
mirrored my own. The historical authors I read were convinced that there was meaning to
be found in natural phenomena. I had thought so myself, only I had searched for
historical or social revelation in their actions in search of that meaning. Finding meaning
in natural phenomena, moreover, was difficult; perhaps the search was even beyond
human ken. The authors of the sources I studied and I all agreed that our sources were
incomplete and sometimes unreliable. Only I was dealing with the problems of the
archive, while they were navigating conflicting authorities and issues of interpersonal
mistrust. I found that these authors therefore seeking out ways in which to uncover
reliable information about the messages that the natural world provided. I focused, in turn,
on tracing and analyzing this process of finding meaning, and I have used it herein to
reveal key aspects of mental life during the late classical and early medieval periods.
And so I present a dissertation on how members of Japan’s late Heian court
sought to find true and reliable information about the sun and moon, stars and planets,
xiv
that moved above them. The Japanese court, like the Chinese state before it, attempted to
harmonize with these movements, in their actions as well as in their time-keeping.
Questions about reliable information about nature, therefore, were embedded in the very
daily fabric of Japanese life during the Heian period. In telling the story of the doubtful
and conflicted nature of knowledge in the late Heian court, I am also presenting a story of
shifting interpersonal relationships, the nature of trust, and of textual difficulties and
mathematical incompatibilities. All the while I have kept looking back on my original
project, and I hope that this dissertation will stand as a first step in helping modern
readers like myself finally approach the issue of how the people of Heian Japan saw the
world around them, both natural and social.
1
INTRODUCTION

In the late Heian period (also known as the Insei period, 1087–1192), the Japanese
court periodically became embroiled in disputes about astronomy. Most notable among
these controversies were debates over competing eclipse predictions and about the
astronomical calendar that shaped public and private life. The historical sources also
reveal debates concerning the identification of ominous astronomical phenomena, and the
interpretation of the same. By the middle of the thirteenth century, however, such debates
fade from the record. This dissertation looks at the social and political factors that
contributed to the debates of the twelfth century and shows how these twelfth century
debates derived from a tradition of earlier debates. The debates from the twelfth century
and before all had roots in practical and institutional contradictions that fed uncertainty
and doubt concerning expert information, particularly such information derived from or
regarding the stars, in late Heian Japan.
Some of these contradictions date even from the initial implementation of state
astronomy and astrology in the early stages of Japan’s Chinese-style bureaucratic state
( Jp. ritsuryō kokka). Others derived from difficulties in translating the
prescriptions of classical Chinese astronomical texts into practice, particularly in a land
distant geographically from the locations known to the original observer-authors in China.
Still other difficulties resulted from social change at the Japanese court over the four
centuries of the Heian Period (794–1192). It was a desire for accurate knowledge about
2
the natural world in the face of mistrust and contradiction that drove the debates over
knowledge in the late Heian court, even as the content of the debates and the resulting
decisions, the agreed-upon “facts” themselves, were shaped by competition among rivals
amid the growing importance of client-patron relationships at the court.
The debates of the Heian Period, particularly in its last century, epitomized
court-wide participation in the production of official knowledge. Such production of
state-determined knowledge about the world was very much a project in the Chinese
imperial mode performed in the service of a Chinese-style state, even in the context of
twelfth-century Japan. The debates, however, show how contingent and fragile the
authority of state-produced knowledge could be. When debates about astronomy and
astrology fade from the historical record in the thirteenth century, it should not be seen as
the result of a growing confidence in technical specialists or because the court abandoned
the project. After all, the production of official facts about the stars continued in the royal
court through the seventeenth century—even if, from the seventeenth century onward, the
Tokugawa shogunate took a more than equal role in that production.
1
Instead, the end of
the period of debate in the thirteenth century reveals a growing fragmentation of both
polity and knowledge in high and late medieval Japan. In the later medieval period,

1
In many respects, the Tokugawa shogunate’s role was larger, as in official projects to classify regional
botanical information in massive surveys. On this natural history project, see Frederico Marcon, "The
Names of Nature: The Development of Natural History in Japan, 1600-1900" (Columbia University, 2007).
The Tokugawa shogunate’s participation in astronomical projects is mentioned briefly in the
afterword to this dissertation. The standard English-language history can be found in Shigeru Nakayama, A
History of Japanese Astronomy: Chinese Background and Western Impact (Cambridge, Mass.: Harvard
University Press, 1969). Some of Nakayama’s discussion is badly outdated, however, when considered in
light of recent scholarship on the institutional history of the early modern shogunate and the continued role
and influence of the royal court in this time.

3
although certainty and unity were still desired, no single institution had the authority or
alliances to impose a decision or resolution to controversies over natural facts.

The involvement of the Japanese state in the production of knowledge from and
about the celestial realm derives from the role such phenomena played in the early
Chinese state. In Chinese correlative cosmology, all aspects of the world—morality,
personal health, politics, and the natural world—were thought to resonate and influence
each other.
2
In particular, the famous Mandate of Heaven ( Ch. tianming, Jp.
tenmei), which legitimized centralized rule, was tied to the heavens above. This
star-consciousness has been sometimes called China’s “first religion,”
3
and star watching
and prediction became a part of the centralized Chinese imperial bureaucracy. Through
the concept of “disaster philosophy” ( Ch. zaiyi sixiang, Jp. saii shisō) as
systematized by Dong Zhongshu
4
during the Han Dynasty (206 BCE – 220 CE), the
observation and prediction of phenomena in the sky served not only as a symbol of the
throne’s direct line to the heavens (or Heaven), but also as a useful tool, through which
the ministers at court could evaluate current policy and choose the correct line of action.
5

2
John Henderson, The Development and Decline of Chinese Cosmology (New York: Columbia
University Press, 1986).. The connection between human and natural phenomena is known technically as
(Ch. tianren ganyang shui, Jp. tennin kan’ō setsu)

3
David W. Pankenier, "Early Chinese Astronomy and Cosmology: The "Mandate of Heaven" As
Epiphany" (Stanford University, 1983).

4
, 179 – 104 BCE

5
Ministers could also use these philosophies to critique the throne or push for reform. For more details on
Dong Zhongshu and the development of this philosophy, see Sarah A. Queen, From Chronicle to Canon:
4
“Disaster philosophy” stated that when politics went against the will or way of Heaven,
nature itself would respond with the appearance of ominous astronomical, meteorological,
and botanical phenomena, as well as with disasters and rebellion. Finally, the ruling
dynasty’s Mandate of Heaven would be rescinded and bestowed upon a new dynasty. By
monitoring nature, and by learning what was normal in contrast to what was abnormal,
the court could stave off potential dynastic change before it started.
Elements of this worldview can be found before Dong Zhongshu, and it
persisted long after the fall of the Han Dynasty in China. Dong Zhongshu was
systemizing ideas found in older guides to political activity such as the Monthly
Ordinances ( Ch. Yue ling, Jp. Getsu ryō) of the Book of Rites ( Ch. Liji, Jp.
Reiki)
6
that mandated the seasons of state activity be harmonized with the meteorological
seasons. The continued political importance of natural signs in Chinese political history
has been shown in Ming and Qing China as well.
7
It can be found well outside of China
itself, in other states that grew up in the “shadow of the Han,” and that adopted and
adapted Chinese tradition as they took form—in other words, the states of the Chinese

The Hermeneutics of the Spring and Autumn, According to Tung Chung-Shu (Cambridge: Cambridge
University Press, 1996).

6
While the text is also referred to as the Rites of Zhou ( Ch. Zhou li, Jp. Shūrei), the text as it exists
today was reworked in the Han Dynasty.

7
For the Ming, see Sarah Schneewind, A Tale of Two Melons: Emperor and Subject in Ming China
(Indianapolis: Hackett Pub. Co., 2006). For the Qing, see Mark. Elvin, "Who Was Responsible for the
Weather? Moral Meteorology in Late Imperial China," Osiris, no. 13 (1998). The “disasters” discussed by
Elvin include solar eclipses.

5
cultural sphere.
8
The importance given to natural signs did not mean, however, that the
court accepted any given interpretation of natural phenomena. In China, a “disaster”
could be attributed to faults of the emperor, his ministers, or the local population, or any
combination of these.
9
Ministers and the throne were also both aware of the potential for
the “spin” of such phenomena.
10
For this reason, state-supported and controlled
monitoring of natural phenomena was a political necessity. Without it, the state would
not only be potentially unaware of pending danger, but would not be able to answer or
mitigate criticism from without based on such “disaster philosophy.” In China, such
monitoring was assigned to the Astronomy Bureau. The Bureau had a monopoly (at least
on paper) on the observation and interpretation of the night sky, as well as on turning
astronomical and seasonal phenomena into an annual astronomical calendar, by which the
activities of the entire realm could be harmonized with the cosmos.
In the eighth century, Japan, too, instituted its own version of the Astronomy
Bureau. Japan’s Bureau of Onmyō ( Jp. Onmyō-ryō)
11
differed in many respects

8
The role of this philosophy in non-Chinese states has been less studied. Even when the role of saii shisō
in Japanese history is discussed, its importance tends to be downplayed, or the philosophy itself is limited
to an early period of Japanese history, seen as outmoded even by the middle of the Heian Period. For an
influential paper that argues along those lines, see Matsumoto Takuya, "Ritsuryō kokka ni okeru saii shisō:
sono seiji hihan no yōso no bunseki," in Kodai ōken to saigi, ed. Mayuzumi Hiromichi (Tokyo: Yoshikawa
kōbunkan, 1990). By contrast, newer work such as that by Mori Shinnosuke (for example,
Mori Shinnosuke, "Sekkan Inseiki kizoku shakai ni okeru matsudaikan: Saii shisō ya unmeiron to no
kanren kara," Nihon shisōshi kenkyū, no. 40 (2008)) argues that Chinese ideals continued to have rhetorical
force through the end of the Heian Period, even if such ideals did not necessarily dominate the Japanese
political scene.

9
See Elvin, "Who Was Responsible for the Weather? Moral Meteorology in Late Imperial China."

10
This is a major theme in Schneewind, A Tale of Two Melons: Emperor and Subject in Ming China.

11
The characters or yinyang in Chinese can be read a number of ways in Japanese. While when
referring to the pair of yin and yang they are typically read as in’yō in Japanese, when referring to the set of
6
from its Tang model: yet within it, too, divisions for the regular practice of astronomical
observation and interpretation ( Ch. tianwen, Jp. tenmon), as well as astronomical
prediction and calendar production ( Ch. lifa, Jp. rekihō) were instituted. The texts
and methods used in Japan were imported either directly from China or via Korean states
as intermediaries. Nonetheless, once instituted in Japan, the practices of tenmon and
rekihō began to diverge from their Chinese models.
This divergence marks an important moment of independence in Japanese
intellectual and cultural history, as well as an important stage in the history of how
astronomical phenomena were treated in Japan. And yet, the history of astronomical
observation and prediction in Japan has tended to ignore the Heian Period. There are a
number of reasons for this. Firstly, the Mandate of Heaven and Confucian thought are
supposed to have been supplanted by the tenth century, as native and Buddhist ideals of
kingship took over—thus, the history of continued Chinese practice has been
de-emphasized. Secondly, histories of science in Japan have focused on

divinatory and ritual practices known as the “Way of Yin and Yang” or , the modern Japanese
pronunciation is Onmyōdō. However, as Suzuki Ikkei has forcefully pointed out, historically the
pronunciation of the name of the official Bureau was not Onmyō-ryō, but On’yō-ryō—the nasal
assimilation of the yō character to myō is only attested from the medieval period onwards. He argued for
distinguishing On’yōdō and Onmyōdō as a way of distinguishing the official practices from the later cult or
religion. Cf. Suzuki Ikkei, "’On’yōdo’ and ‘Onmyōdō’: Reflections on the Lineage and System of the ‘Way
of Yin and Yang’," in Onmyōdō: The Way of Yin and Yang (Columbia Center for Japanese Religions,
Columbia University 2009)..
While I find Suzuki’s arguments convincing, there is a certain amount of historical continuity
and fidelity lost in enforcing too strict a distinction between the Bureau and the popular religious practices
that sprung up around it—an at-least social connection between official timekeeping and divination and
ritual practices that I hope the following chapters make clear. For that reason, and also due to the fact that
the onmyō pronunciation is slightly more prevalent in English-language sources due to the dual influences
of Murayama Shū’ichi and popular culture, in this dissertation I will be using the term Onmyō-ryō in place
of the linguistically more accurate On’yō-ryō.

7
modernization—specifically how Japan adopted advanced scientific practices from either
China or the West.
12
For these authors, a history of how Japan continued to use, and
change, Chinese practice that was “outdated” was not a subject of interest. Likewise,
although the history of astronomical observation and interpretation are intimately tied to
developments in Japanese religious history—particularly the rise of a distinct population
of diviners and exorcists ( Jp. onmyōji)—the state-level production of facts about
astronomical phenomena has not been of particular interest to these scholars either,
focused as they tend to be on medieval personal religion.
13
Finally, there is a lack of
surviving treatises or texts dedicated to theoretical issues of the nature of the stars or
technical affairs. For these reasons, it has been possible even recently to assert

12
Nakayama, A History of Japanese Astronomy: Chinese Background and Western Impact,
Sugimoto Masayoshi and David L. Swain, Science and Culture in Traditional Japan, A.D. 600-1854
(Cambridge, Mass.: MIT Press, 1978).

13
This is not to say that there have not been discussions of tenmon and rekihō (often referred to as
Jp. rekidō) in Onmyōdō studies. For example, there is a chapter on the two fields in Yamashita Katsuaki,
Heian jidai no shūkyō bunka to Onmyōdō (Tokyo: Iwata shoin, 1996). However, here as elsewhere, the
focus of the study is on Onmyōdō and onmyōji as practices and practitioners, and tenmon and rekihō
explained insofar as necessary to flesh out some of the “other duties” of onmyōji employed by the Bureau
at court. Furthermore, the practices of tenmon and rekihō are generally assumed to be static, and so any
discussion of forces for change, which is the heart of the current dissertation, tends to be omitted.
Hosoi Hiroshi has undertaken the closest scholarship to the present study, as he focuses on the
history of science and Onmyōdō-related topics, particularly in recent articles. (See especially Hosoi Hiroshi,
"Tenmondō to rekidō: Kodai ni okeru seiritsu no haikei to sono yakuwari," in Onmyōdō no kōgi, ed.
Hayashi Makoto and Koike Jun'ichi (Kyoto: Sagano shoin, 2002), Hosoi Hiroshi and Minezaki Ryōichi,
"Rikkokushi mishūroku no nisshoku to kokushi," Kwassui ronbunshū: ippan kyōiku, ningen kankei gakka,
ongaku gakubu hen, no. 43 (2000), ———, "'Nihon tenmon shiryō' mishūroku no nisshoku to kiroku:
Rikkokushi shūryō ikō senroppyaku-nen izen," Kwassui ronbunshū: bungakubu ningen kankei gakka,
ongaku gakubu hen, no. 45 (2002).) However, in these Hosoi does not discuss social forces involved in
developments in tenmon and rekihō practice. Hosoi’s focus is more positivist—he sees change as largely
driven by a desire for more accurate results. (This is clearly articulated in Hosoi Hiroshi, "Tenmondō to
rekidō: Kodai ni okeru seiritsu no haikei to sono yakuwari.") This leads him to ignore shifts in practice that
do not result in more accurate, more “scientific” practices in either tenmon or rekihō.
As this dissertation takes a far more social-constructivist view of tenmon and rekihō practice, I
will not be addressing Hosoi’s work on a point-by-point basis here.

8
(mistakenly) that in Heian Japan “[s]cience for its own sake does not exist, nor do
abstractions. The Heian mind was quite incurious about the natural world.”
14

Indeed, any examination of the surviving historical record concerning debates
about astronomy shows this statement to have been in error. As the following analysis of
debates about astronomical interpretation and prediction in official histories and court
diaries will show, Heian minds were in fact desperately seeking accurate and reliable
information about the world around them. This quest was both politically and privately
motivated, as astronomical calendars (particularly in the form of annotated almanacs or
Jp. guchūreki) and divination based on astronomical phenomena served as a
guide to appropriate action, as well as a warning of problems to come, at both state and
personal levels. This dissertation will show how the desire for such information
outstripped the capabilities of the official organization of the state, which meant that
individual experts from both within the Bureau of Onmyō and without began to respond
to new demand. The increase in the number of recognized experts itself then sparked
conflict as individuals came to different conclusions about phenomena. The search for
information that was “certain” and indisputable, ironically, sowed doubt and mistrust at
the highest levels of the court. The attempts of these courtier-bureaucrats to overcome
doubt resulted in the production of more doubt. Each measure taken to ensure certainty
seems to have spawned more debate.

14
Marian Ury, "Chinese Learning and Intellectual Life," in Cambridge History of Japan, ed. Donald
Shivley and William McCullough (Cambridge: Cambridge University Press, 1999), 349.

9
The history of astronomical investigation in Heian Japan, shows not only how
important this effort was to the Japanese state, but also details about interpersonal
relationships and what they meant in terms of which “facts” were accepted as true and
reliable, and which were not.
15
It has been held that there was no epistemology in
classical or early medieval Japan—and in terms of widely circulating philosophical
works concerning the nature of truth, this is correct. However, members of the Heian
court were desperately searching for information they could rely on as “true” and
“accurate.”
In this respect, there was a search for something like a native epistemology—and
perhaps even an epistemological crisis—during this period. It is therefore possible to do
what Peter Dear has referred to as “epistemography,” or the historical study of
epistemologies.
16
Such work is in fact essential to understanding how changes in
Japanese society, and specifically in the interpersonal relationships between ruling and
technical members of the Japanese bureaucratic court, affected the way that the court and
its information-getting practices functioned during the early medieval transition. This
work is also necessary to understand records of natural phenomena as they have been
preserved in the historical record, as these records are neither purely political

15
This study, therefore, takes inspiration from Steven Shapin, A Social History of Truth: Civility and
Science in Seventeenth-Century England (Chicago: University of Chicago Press, 1994)., in looking for the
cultural qualities that made certain facts more acceptable to both experts and the establishment.

16
Peter Dear, "Science Studies as Epistemography," in The One Culture?: A Conversation About Science,
ed. Jay A. and H. M. Collins Labinger (Chicago: University of Chicago Press, 2001).

10
fabrications
17
nor naïve reflections of physical reality.
18
Instead, such records are the
traces of a technical process of investigating nature that was documented by individuals,
which documents were in turn read, critiqued, and edited by others. The social history of
astronomical information in Japan, therefore, is a necessary step for reconstructing the
environmental history of Japan.
Although a society can be said to have knowledge of something so long as one
part of the society has that knowledge, this only works if the source of that knowledge is
trusted.
19
The evidence of doubt and conflict in Heian Japan brings into question
assumptions that have been made previously about ideology and mentalité. In studies on
such topics it has been assumed that most or all of Japanese society assented to and
understood the elements of cosmology under discussion.
20
The history of doubt in

17
This is the position taken by Herbert Plutschow in his analysis of “ominous” phenomena in historical
narratives concerning the political exile and death of Sugawara Michizane, preceding his eventual
posthumous pardon and deification. However, his best evidence—accounts of solar eclipses in the
historical record that astronomically could never have occurred—is based on a misunderstanding of how
these chronicles were compiled. Chronicle histories were compiled using eclipse predictions and not
observations, and therefore an erroneous eclipse prediction does not prove to be the political fabrication of
an omen that never existed in nature. For Plutschow’s analysis, see Herbert Plutschow, "Ideology and
Historiography: The Case of Sugawara No Michizane in the Nihongiryaku, Fusô Ryakki and Gukanshô"
(paper presented at the Historiography and Japanese Consciousness of Values and Norms 2001).

18
This position tends to be taken by environmental historians. The primary scholar on the environmental
history of Heian and medieval Japan, William Wayne Farris, does not go to extremes on this issue (see
William Wayne Farris, Japan's Medieval Population: Famine, Fertility, and Warfare in a Transformative
Age (Honolulu: University of Hawai'i Press, 2006), but especially ———, "Famine, Climate, and Faming
in Japan, 670-1100," in Heian Japan, Centers and Peripheries, ed. Mikhael et al. Adolphson (Honolulu:
University of Hawai'i Press, 2007). for his discussion of the sort of material that is left out of the historical
sources), yet it is my opinion that Farris does not engage sufficiently with the constructed nature of records
of natural phenomena.

19
See Shapin, A Social History of Truth: Civility and Science in Seventeenth-Century England, especially
xxv-xxvi and 3-42.

20
See the introduction to Bernard Frank, Kata-Imi Et Kata-Tagae: Etude Sur Les Interdits De Direction
À L'époque Heian, Bulletin De La Maison Franco-Japonaise (Tokyo: Isseidō, 1958). Such a position—that
the cosmology was necessarily intelligible to all members of the cultural elite, is also mentioned explicitly
11
Japan—of doubt even about the correct day and month—forces us to re-evaluate
depictions of medieval thought and culture, and to re-emphasize the role of secret
traditions, or hiden ( ), in Japanese history.
21
Once the court no longer had the
authority to enforce a single interpretation or depiction of the world—a bed that the court
had made for itself—the medieval Japanese individual had a number of traditions from
which to solicit information, which could then be accepted or rejected. How authoritative
knowledge about the stars broke down in Japan, therefore, provides clues to how
religious and political plurality could thrive, even without a single unifying cosmology.
22

So this dissertation looks at the history of astronomical observation and
interpretation, or tenmon, and of calendrical production and astronomical prediction, or
rekihō, in Japan. It traces the development of a “Japanese-style” of originally Chinese

in Edward H. Schafer, Pacing the Void: T'ang Approaches to the Stars (Berkeley: University of California
Press, 1977), although for the Chinese context

21
For the best treatment of hiden in its original context of esoteric Buddhism, see Jacqueline Ilyse Stone,
Original Enlightenment and the Transformation of Medieval Japanese Buddhism (Honolulu, Hawaii:
University of Hawai'i Press, 1999). The influence of hiden on medieval Japanese literary traditions can be
found in Susan Blakeley Klein, Allegories of Desire: Esoteric Literary Commentaries of Medieval Japan,
Harvard-Yenching Institute Monograph Series (Cambridge, Mass.: Harvard University Press, 2002).
Despite these two studies, more work on the influence and role of the concept of hiden and related concepts
in other aspects of Japanese culture remains necessary.

22
While this is not a major facet of this dissertation’s argument, I hope that the discussion in this
dissertation might destabilize some assumptions about the underlying and universal influence of esoteric
Buddhist cosmology in the mental life of pre-modern Japan, particular as was famously formulated by
Kuroda Toshio on kenmitsu synchronic religion. I do not intend to deny the influence of esoteric Buddhist
schools in the culture of medieval Japan, which was clearly important and pervasive. Instead, I question
whether there was much social consensus at the level of the underlying cosmology. I also question how
systematic any cosmology was at a larger, societal level (leaving aside the question of systematic
cosmologies within individual scholarly works or schools of thought). I hope that this dissertation shows
some of the ways that practices and models could be built out of tradition and accretion—showing that a
systematic underpinning logic need not be necessary to explain all aspects of medieval thought, even as it
is clear that logics of metaphor and identification did have a role to play in medieval Japanese religious
culture.

12
practices, to their final fragmentation under institutional constraints, social shifts that
changed interpersonal relationships at court, and new trends in textual interpretation. It is
divided into two parts: the first, consisting of two chapters, covers a brief history of
tenmon; the second, consisting of four chapters, covers developments in the Japanese
interpretation of the Xuanming li system of Chinese calendrical astronomy.
Specifically, the first chapter, “Written Above: Authority and Control in the
Practice of Tenmon in Classical Japan,” shows that received images of the astrological
expert as having unquestionable authority and influence over the Japanese population are,
in fact, incorrect. Historical records show that astronomical interpretation were at times
heavily shaped by demands of the powerful nobles who served as gatekeepers and
monitors for the production of this knowledge. This chapter describes the early history of
state control over the identification and interpretation of astronomical phenomena, and
shows how an institutionalized regime of secrecy in fact laid the stage for both court
interference in the work of its “trusted” experts, and for the fragmentation of
astronomical observation in later ages.
The second chapter, “Leaks in the System: Access and Authority in the Changing
Practice of Tenmon” analyzes how changes in how documents were received and
forwarded at court, leading to the development of private archives of “secret”
astronomical knowledge, which the members of the high aristocracy could use to critique
the work of tenmon experts, even shape it. In response to this interference, tenmon
specialists sought new sources of authority that they themselves could control, resulting
in the creation of the image of the infallible astrologer described in the first chapter.
13
The second part of this dissertation begins with the third chapter, “‘Many Secret
Texts:’ The Origins of Lineage Traditions and Variation in Japanese Calendrical
Astronomy.” This chapter covers the early history of calendrics in Japan up through the
end of the tenth century. It demonstrates that traditions of practice associated with
particular lineages of experts can be traced back to the beginning of the Heian Period (the
ninth century at the latest), well before previously assumed. Furthermore, the
development of these traditions was driven by both demands from the court for standards
of “correct” calendrical production and eclipse prediction, and by the strategies utilized
by individuals contesting with each other for primacy within the Bureau of Onmyō. But it
was the tenth century that saw the true beginnings of a “Japanese-style” of Chinese
calendrical astronomy. Thus it is evident that the history of calendrical astronomy in
Japan is not nearly as static and rote as previous studies have assumed.
The fourth chapter, “Invisible Presences: Buddhist Calendrics at the Mid-Heian
Court,” introduces what would become a second, non-court, source for calendrical
knowledge in Japan: the technical expertise of Buddhist monks of esoteric sects. This
chapter traces how these monks were brought into the official calendar-making project of
the Heian court, and why they were sought after as collaborators by the young Kamo
lineage of calendarists.
Public recognition of clerical expertise in matters of calendrical astronomy laid
the foundation for the events discussed in the fifth chapter, “Competing Calendars:
Eclipse Prediction, Shōshō, and the Establishment of Buddhist Calendrics as an
Independent Tradition.” This chapter covers the mid-tenth century career of the Shōshō,
14
who was the first monk to publicly criticize the work of official calendarists. Through
such criticism, and bolstered by the support of powerful patrons, he destabilized the
authority of official calendrical production. Much late Heian-period doubt about
calendrical astronomy can be credited to this man.
The sixth chapter, “Deathless Texts: Innovation, Conservatism, and the Chinese
Classics in the Supervision of Calendrical Knowledge,” shows a final tactic that the
Japanese court took in order to find reliable sources of astronomical knowledge. By
prioritizing the Chinese classics, many members of the court hoped to reform politics,
Japanese society, and knowledge of the world. By this time, however, the tradition of
Japanese practice had in and of itself a strong history and authority, and so could only be
modified, not replaced. Ironically, it was insistence on a conservative reading of
outmoded texts of Chinese calendrical astronomy that put into place the final element of
the “Japanese-style” of calendrical astronomy that took shape over the Heian Period.
The afterword, “Fragmentation and the Ideal of Unity in Medieval Japan,”
summarizes the history of calendrics and astronomical interpretation up through the
reform of the seventeenth century. In contrast to previous studies that emphasize isolation
as the primary reason Japan did not implement new developments in astronomy and
astrology from China, in this dissertation it is argued that it was a lack of central authority
that led to fragmentation in time-keeping and astrology, which was only partially
resolved under the mature Tokugawa shogunate of the late seventeenth century. Even at
that late date, although the importation of Chinese systems of calendrical astronomy was
proposed, the long history of an independent tradition of Japanese calendrics favored the
15
adoption of a (ostensibly) natively developed system of calendrical astronomy. The desire
to reform the world and find the single true knowledge, which was suspected to be out
there, predated this time—no one before the seventeenth century, however, had the
authority to enforce a standardized production of astronomical knowledge.
The search for certainty in the Heian Period had led to the cultivation of multiple
lineages that looked to or claimed different and conflicting authorities, and from which
individuals could get “true” information. What lineage an individual supported depended
significantly on interpersonal relationships and institutional history. The fragmentation of
political authority in the Heian Period, this dissertation suggests, was reflected in a
fragmentation of knowledge that persisted until the early modern state.

16
CHAPTER ONE

WRITTEN ABOVE: AUTHORITY AND CONTROL IN THE PRACTICE OF
TENMON ( ) IN CLASSICAL JAPAN

There has been a marked tendency for scholars focusing on pre-modern Japan to look
askance at, or else ignore or downplay, the role of diviners in court society and culture.
When the topic had not been treated as an unanalyzed part of the cultural milieu, the
diviner, particularly the diviner working from astronomical phenomena, has been oft
depicted as a Svengali exerting a great deal of influence on a credulous population. Yet
historical records reveal that the identification and interpretation of astronomical
phenomena was often the subject of doubt and debate. A close analysis of the practice of
observing and recording astronomical omens in pre-modern Japan not only reveals that
common depictions of the absolute authority of the diviner are in error, but also that this
state-run practice had, from its start, internal inconsistencies and weak points that
undermined the authority of the astrologer.
A number of factors contributed to this perception of diviners and their work,
including the legal structure surrounding observational astrology ( Ch. tianwen, Jp.
tenmon), defined as the regular observation, identification, and interpretation of
astronomical phenomena for the purposes of the state. This legal structure defined
tenmon as restricted knowledge and limited its legitimate practice to the officials of the
Bureau of Onmyō ( Jp. Onmyō-ryō), even as it left open loopholes in its program
of secrecy so that higher officials could monitor and intervene in the Bureau’s work.
17
In this chapter, the practice of tenmon and the image of its practitioner are
described, and the limits of state control over this field of knowledge analyzed.

An Instance of Debate
In the spring of 1189, something strange appeared in the night sky. This
phenomenon would launch a debate between the regent and a specialist on celestial
events and the interpretation of the same. This debate reveals much about how
astronomical observation and observation changed over the course of the Heian Period
(794-1192) in Japan, and more importantly, the social and textual forces that shaped the
way in which it changed. The cause of these changes, and the roots of the 1189 debate,
arose in part from the legal establishment of astronomical observation and interpretation
as a state monopoly in the eighth century.
Later chronicles describe that this particular phenomenon as first appearing in
the east, intersecting part of the constellation Ursae majoris.
1
These same sources noted
that from the time it was sighted some people were referring to it as a “comet” ( Ch.
huixing, Jp. suisei or hōkiboshi).
2
By identifying the phenomena as a comet, these

1
Azuma kagami Bunji 5/2/28; Ichidai yōki 5/2/28. The primary editions consulted were
Kokushi Taikei Hensankai, ed. Azuma Kagami, 4 vols., Kokushi Taikei (Tokyo: Yoshikawa
kōbunkan,1978–1979) and Shintō Taikei Hensankai, ed. Ichidai Yōki, 3 vols., Shintō Taikei (Tokyo: Shintō
Taikei Hensankai,2005-2006), respectively.
The Azuma kagami entry describes its position as “north of Doukui ( Jp. Tokai),” which
as an astronomical term could refer to the stars Ursae majoris α, β, γ, or δ; or else to the area delineated by
these stars.

2
Literally “broom star,” (Ch. huixing, Jp. suisei or hōkiboshi) is the modern Japanese generic term
for comet. In terms of historical usage, however, the categories differed at different terms, and suisei did
not necessarily refer to the genera of all comets, as opposed to a comet with a well-defined, single tail. In
18
unnamed individuals were making a political situation and a statement about the likely
future, just as much as they were speaking of a distant astronomical phenomenon. A
comet indicated danger for the dynasty, even the threat of regime change.
3

Astronomical observation of this sort had been tied to political prognostication
since before the Han Dynasty in China.
4
The first Japanese record concerning something
like a comet,
5
describing a phenomena from 634, notes that a “long star” ( Ch.
zhangxing, Jp. nagaboshi) appeared in the southeast, and that “people of the time called it
a comet (suisei).”
6
From that time up through the year 1106, when a comprehensive list

this work, suisei will be translated as “comet,” but this should be read as a more constricted definition than
the modern astronomical one, and not one which necessarily includes all comet-like phenomena.

3
Other comet-like phenomena had similar although distinct meanings. For example, a “spear-star” (
Jp. hokoboshi) would indicate invasion or rebellion. Two instances of “spear-stars” can be found in the
Japanese sources, for 947 and in 948; Nihon kiryaku Tenryaku 1/1/27 and Teishinkō-ki shō Tenryaku
2/1/19 respectively. Interestingly, the corresponding entry in the Nihon kiryaku for the latter records the
phenomena as a “comet” ( ). (The major editions consulted were Kokushi Taikei Hensankai, ed. Nihon
kiryaku, 3 vols., Kokushi taikei (Tokyo: Yoshikawa kōbunkan,1985–1988). and Tōkyō Daigaku
Shiryō hensanjō, ed. Teishinkōki, Dai Nihon kokiroku (Tokyo: Iwanami shōten,1956) respectively.)

4
206 BCE-220 CE.
The earliest elaborated expression of the connection between astronomical phenomena and
political personages and events can be found in Sima Qian’s Records of the Grand Historian ( Ch.
Shiji, Jp. Shiki; as has been noted by others, the title of this work could just as accurately be translated as
Records of the Grand Astrologer), although examples of the political and military interpretation of
astronomical events can be found in commentaries on the Spring and Autumn Chronicles traditions as well.
See David Schaberg, A Patterned Past: Form and Thought in Early Chinese Historiography, Harvard East
Asian Monographs (Cambridge, Mass.: Harvard University Press, 2001).
Despite this evidence for earlier traditions of interpretation, the section on the Heavenly Offices
( Ch. tianguan shu, Jp. tenkansho) in Sima Qian’s Records of the Grand Historian are seen as the
earliest surviving attempt at a systematic synthesis of disparate prognostication traditions.

5
Or so identified as such by Kanda Shigeru. Kanda Shigeru, Nihon tenmon shiryō (Tokyo: Maruzen
kabushiki kaisha, 1935)., 475.

6
Nihon shoki 23, Jōmei 6/8. (The edition consulted was Kokushi Taikei Hensankai, ed. Nihon shoki, 2
vols., Kokushi taikei (Tokyo: Yoshikawa kōbunkan,1979).)

19
of comets was compiled and submitted to court,
7
there had been at least 34 comets, all
tied to a particular calamity.
8
Additional comets had appeared since, even within living
memory. In official reports, members of the Japanese court were reminded again and
again that comets were signs that “sweep away the old and bring forth the new.”
9

7
This document was submitted by Nakahara Morotō (II) in order to divine the 1106 comet through a new
method. This document will be discussed in more detail below. The edition preserved in the surviving
chapter of the Shodō kanmon compilation ( , “Reports from the Various Experts,” attr. Nakahara
Moroyasu, first half of the twelfth century). While the document Shodō kanmon shows some later editorial
notations, the compilation itself does not seem to date from long after 1106. The only surviving text of the
Shodō kanmon is Nakahara Moroyasu (attr.), "Shodō kanmon," in Gunsho ruijū, ed.
Zoku gunsho ruijū kanseikai (Tokyo: Yagi shoten, 1960).

8
According to a contemporaneous usage of the term “comet,” included by its appearance in the Shodō
kanmon list, there had been 34 comets in Japanese history. Some comets were associated with more than
one calamity according to this list.
This number of historical comets for 634-1106 does not match with the accounting compiled by
Kanda Shigeru, who included a number of “guest stars” ( Ch. kexing, Jp. kakusei) and other
phenomena in his selection of records of comets. “Guest stars” are often associated with novae, although
not exclusively, and a number of major supernovae have been associated with Chinese and Japanese
records of these phenomena. Modern dictionary definitions of “guest stars” include both novae and comets
as examples. See the definition of kakusei in Shōgakkan kokugo jiten henshūbu, ed. Nihon kokugo daijiten,
14 vols. (Tokyo: Shogakkan, 2000–2002).
There are also a number of historical instances of the term “comet” to describe phenomena that
do not appear in the Shodō kanmon list, primarily: Hōki 3/12/23 (Shoku Nihongi), Kanpyō 6/2 (Fusō
ryakki), Shōtai 3/7 (Kakumei kanmon), Tenryaku 2/1/19 (a suisei in Nihon kiryaku, but (Ch. gexing,
Jp. hokoboshi) in Teishinkō-ki shō), Ōwa 1/2/27 (Fusō ryakki and Kakumei kanmon), Jōgen 2/2/24 (Nihon
kiryaku, described as comets in two directions), Chōgen 7/8/13 (Sakeiki). (Editions consulted include
Kokushi taikei hensankai, ed. Shoku Nihongi, 2 vols., Kokushi taikei (Tokyo: Yoshikawa kōbunkan, 1979),
Unknown, "Kakumei kanmon," in Gunsho ruijū, ed. Zoku gunsho ruijū kanseikai (Tokyo: Yagi shoten,
1960), Kokushi taikei hensankai, ed. Fusō ryakki; Teiō hennenki, Kokushi taikei (Tokyo: Yoshikawa
kōbunkan, 1929), and Zōhō shiryō taisei kankōkai, ed. Sakeiki, Zōhō shiryō taisei (Kyoto: Rinsen
shoten,1965), among others cited previously.)
The reason for these omissions by Nakahara Morotō (II), the author of this section of the Shodō
kanmon, is unknown. In the case of kakusei or hokoboshi not included in this record (14 instances total,
from the period 642-1054), that these records are consistently not included in the 1106 report is further
evidence that historical usage of suisei does not show the same expansiveness as the modern definition of
comet. On this, there is further discussion in the next chapter.

9
a phrase that appears often in astrological memorials ( Jp. tenmon no missō). A
collection of these is found in the Shodō kanmon compilation cited in the previous two notes.

20
Nineteen days after this new comet’s first appearance on the twenty-eighth of
the second month, supernumerary Palace Kitchen Supervisor
10
Abe Suehiro
11
brought
his report on the phenomena before the regent Kujō Kanezane.
12
Suehiro was a known
astrologer, diviner, and ritualist from a long line of the same, and possessed years of
experience in this capacity of identifying and interpreting celestial phenomena. Kanezane,
on the other hand, at age forty, had been appointed regent four years before. According to
the record in Kanezane’s court dairy, Gyokuyō ( ), Kanezane proceeded to question
Suehiro on the details of the report. Already Suehiro’s report was late by previously
established court standards—less than one hundred years earlier, a delay of twenty days
between an astronomical event and the official report on it was considered “greatly
suspicious.” At that time, the officials involved were forced to submit a formal apology.
13

Kanezane’s critique, however, focused not on Suehiro’s tardiness, but his wording.
In his report Suehiro referred to the astronomical event not as a comet ( ),
literally “brush star,” but instead as a “brush aurora” (or, by analogy, a “comet-like

10
This post was clearly intended in Suehiro’s case as a sinecure, to support him as he worked for the
court on technical and ritual matters. By the middle of the twelfth century there were far too many technical
specialists employed by the court than could be housed within the Bureau of Onmyō, which, even after the
addition of supernumerary officials, only held a legally constrained number of posts.

11
, dates unclear.

12
, 1149 – 1207.

13
Chūyūki Kanji 6/12/30 (1097). (The primary edition consulted was Tōkyō Daigaku Shiryō hensanjō, ed.
Chūyūki, 7 vols., Dai Nihon kokiroku (Tokyo: Iwanami shoten,1993 –)., but for years not yet covered in the
Dai Nihon kokiroku edition Zōhō shiryō taisei kankōkai, ed. Chūyūki, 7 vols., Zōhō shiryō taisei (Kyoto:
Rinsen shoten, 1965) was used.)
While no explicit timeframe for the submission of tenmon reports was mandated by law, a
survey of dated reports on astronomical phenomena indicates that the initial report was expected within 2-3
days. The same phenomena might appear again in later reports, or in combination with others.

21
aurora”— Ch. huiqi, Jp. suiki). This, Kanezane said, was unacceptable. In his
account of this conversation, Kanezane could not help breaking from his accustomed
style of writing to record his opinion. Kanezane, a dedicated scholar of Chinese classics,
typically wrote in Sino-Japanese ( Jp. kanbun), using Chinese grammar—such was
the typical language of government documents in both Japan and China. In this case,
however, Kanezane had to resort to the less formal classical Japanese in order to record
his words to Suehiro. Using grammatical terms that reflected strong emotions, he
demanded that Suehiro change the report: “If it is a comet, call it a comet! If it is a
strange aurora ( Ch. yiqi, Jp. iki), [then] call it an ‘ominous aurora’ ( Ch. yaoqi,
Jp. yōki) or ‘guest aurora’ ( Ch. keqi, Jp. kyakki)! If you use the character
‘comet-like,’ then replace ‘aurora’ with ‘star.’ Such a term [as ‘comet-like aurora’] has
never yet been seen. How could you present a report in this way?”
14

Kanezane’s insistence on the appropriate definition was not merely a question
of linguistic monitoring. The proper interpretation of an astronomical phenomenon, the
proper reading of what this omen meant for the future, depended on what the phenomena
was called.
15
Manuals of astronomical divination were organized by phenomena name,
and by their logic even a slight error in classification might mean the difference between
reading the message from the heavens about the future properly and missing the message

14

Gyokuyō Bunji 5/3/17. (The primary text consulted was
Kokusho sōsho kangyōkai, ed. Gyokuyō, 3 vols. (Tokyo: Meichō kangyōkai,1998).)

15
There are parallels here with the “Rectification of Names,” a Confucian principle that individuals
should conform to their category, and that the categories should adequately represent their contents.

22
entirely. Yet for many phenomena, particularly classes of comets and auroras, there was a
great deal of ambiguity in the various classifications.
16
Categories overlapped and were,
moreover, self-contradictory. Perhaps it was for this reason Suehiro had proposed a new
category of phenomena.
Despite Kanezane’s strong criticism, Suehiro attempted in this audience to
defend his apparent neologism. According to Kanezane, Suehiro argued that an “original
star” ( Ch. benxing, Jp. honsei) had not been present, and therefore the term “comet”
( )—which contains the word “star”—would have been incorrect. Whether Suehiro
was indicating that the cometary core was not distinguishable from the tail, or whether he
was referring to the definition of comets that indicated that they appeared first as “guest
stars” ( Ch. kexing, Jp. kakusei)
17
and only later developed tails, is unclear.
18

It is likewise uncertain from the point of view of modern astronomy precisely
what phenomena had appeared in the sky above Japan. The Chinese terminology for
“comets” shifted over time, as can be seen in a comparison of how “comet” terms and
categories were used in the “Treatises on the Heavens” ( Ch. tianwenzhi, Jp.

16
For a fuller description of this, see chapter two.

17
“Guest star” has become the standard translation of this Chinese term for modern astronomers working
with the Chinese records for references to comets or potential supernovae.

18
Another possibility is can be found in references to a assertion that a comet was originally a fixed star,
or originated from a fixed star, and only later developed a tail. While this assertion has not yet been
identified in Japanese sources, references to “original stars” can also be found in references to meteors in
Japanese sources. Despite this phrasing, it had become standard in Chinese materials to deny that shooting
stars were fixed stars that left the heavens. On the belief that comets originated from planets, see Joseph
Needham, Science and Civilisation in China, vol. 3: Mathematics and the Sciences of the Heavens and the
Earth (Cambridge: Cambridge University Press, 1959), 433.
A list of “types” of comets and their associated planets (as or henxing) can also be found
in the Treatise on Heaven in the fifth century History of the Jin. Jin shu 13.

23
tenmonshi)
19
in the Han History ( Ch. Han shu, 1
st
century CE
20
) and in the Jin
History ( Ch. Jin shu, 636).
21
In many places, the distinction between types of
“comets” was morphological: that is, based on the shape of the astronomical event. The
definition repeated in reports on astronomical phenomena was that a “comet” was a
transient star that developed a single, essentially straight, cometary tail that pointed away
from the sun. This seems plainly to be a comet by modern astronomical definitions.
However, for other “comet” phenomena, particularly something known as Chi You’s
Banner ( Ch. chiyouqi, Jp. shiyūki), the definition is less clear. Although it is
typically considered a type of comet, some historical examples of this phenomena appear
more like a type of aurora borealis.
22
It seems that at the very least, the definition of Chi
You’s Banner shifted over time, if it was ever clear to begin with.

19
Tianwenzhi is often translated as “Treatise on Astronomy.” The contents of these treatises, however,
combined not only a description of the appearance and nature of astronomical bodies and phenomena, but
also a chronology of astronomical phenomena as portents and their associated political results. The
tianwenzhi also do not include the mathematical details for calculation the motions of astronomical bodies
such as the sun, moon and planets, used for the prediction of lunar and solar eclipses—that such material
was to be found in the lülizhi ( Jp. ritsurekishi), the “Treatise on Pitch-pipes and Calendrics”
(which did not always include material related to the pitch-pipes). As a result, “Treatise on Astronomy”
seems inadequate, and the more literal “Treatise on the Heavens” (or “Treatise on Heavenly Patterns”) to
be preferable.

20
While most of the Han shu was compiled by Ban Biao ( , 3-54 CE) and his son Ban Gu ( ,
32-92), parts of the text, including the Treatise on the Heavens, is thought to date from slightly later, during
the rule of Emperor He ( , r. 88-106).

21
The editions of the Chinese dynastic histories cited here and elsewhere are from the 1995 Beijing
edition of the Er shi si shi published by Zhonghua shu ju.

22
Michael Loewe, "The Han View of Comets," Bulletin of the Museum of Far Eastern Antiquities, no. 52
(1980).

24
Even for elements of the typological categorization of comets that are thought
to have become relatively stable by the seventh century,
23
it can be difficult to fully
account for mislabeling or mistaken identification in either the Chinese or Japanese
sources.
24
In 1189, Abe Suehiro certainly thought that the phenomenon visible in the sky
was not a comet. He did allow, however, through the use of the term sui , or
“comet-like,” that this phenomena did have certain features in common with comets.
It should be noted, furthermore, that Suehiro’s attempt at innovation and
Kanezane’s resistance to it was not merely a question of terminology or morphology.
Suehiro’s report served both as an official account and description of the phenomenon
that had appeared in the sky, and as an interpretation of that same phenomenon as an
omen. By the logic of the tenmon report or divination, the true meaning of phenomena
that appeared in the sky could only be determined once the identification was correct.
Without correct classification, the meaning of the event would be unknown. This was the
underlying point of contention between Suehiro and Kanezane. Once the type of
phenomena was identified, and the section of the night sky in which it occurred and other
details were determined, the rest of the report was compiled through looking up

23
Michael Loewe asserts this in his article on the silk manuscript from the Mawangdui tomb. He bases
this conclusion on a comparison of the Mawangdui comet diagrams, the descriptions in the Han shu and Jin
shu, and Ma Duanlin’s 1317 Wenxian tongkao ( ), wherein the major terms used to describe
comets in the Mawangdui manuscripts and the Han Histories are subdivided into two categories. Ibid.

24
While the record under debate dates from before the “maturity” of category types identified by Loewe,
some of the potential difficulty in reconciling Chinese sources with modern astronomy can be found in
dueling articles in Nature attempting to identify the “guest star” of 185 as either a nova or as a comet. See
Y. N. and Y. L. Huang Chin, "Identification of the Guest Star of Ad 185 as a Comet Rather Than a
Supernova," Nature, no. 371 (1994), S. E. Thorsett, "Identification of the Pulsar Psr1509–58 with the
'Guest Star' of Ad 185," Nature, no. 356 (1992).

25
precedents from Chinese history and listing them. Occurrences of the same phenomena in
the past were thus to be the guide to the future.
25

By the twelfth century, the primary sources quoted in such reports were
Tianwen yaolu ( Jp. Tenmon yōroku) and Tiandi ruixiang zhi ( Jp.
Tenchi zuishō shi), both Tang-dynasty compilations of quotations from other sources that
were organized by topic.
26
While identified as Chinese compositions, both texts only
partially survive in Japanese editions, the Tianwen yaolu more completely than Tiandi
ruixiang zhi. No surviving section from either text includes material that would indicate
whether Suehiro would have been able to bring up Chinese precedents to reinforce his
categorization of the sky and provide meaning to the phenomena.
27
If there were no
Chinese cases in his report, it would not have matched the documentary format that had
been established in Japanese court practice over two hundred years earlier for such
documents.
Yet, even if there were Chinese precedent, Kanezane probably would not have
found Suehiro’s argument convincing. In their debate, Kanezane cited Japanese history to

25
The most explicit example of this principle—that the “result” or “consequence” of a phenomena in the
past was the guide to what this phenomena meant for future possibilities—can be seen most clearly in a
1106 report on comets found in the Shodō kanmon ( ) collection, even though it involves Japanese
and not Chinese precedent. In it, every “comet” from Japanese history is listed along with the consequences,
either within the same year or a few years later. A variety of options are thus provided, from famine to
epidemic to the death of a notable. Circumstances—specifically, the death of the ruler Horikawa Tennō in
1107—meant that in retrospect the meaning of the 1106 comet was clear: it presaged “that thing,” as
Horikawa’s death was referred to in a note added later to the report.

26
Nakamura Shōhachi, Nihon Onmyōdō sho no kenkyū: Zōhōban (Tokyo: Kyūko shoin, 2000).

27
The sections dedicated to comets are missing in both. See ibid. Nakamura’s statement about the
Tianwen yaolu was compared against the manuscript copy held by the Historiographical Institute of the
University of Tokyo, Unknown, "Tenmon yōroku," (Tokyo: Tōkyō Daigaku Shiryō hensanjō, n.d.).

26
remind Suehiro that there was already an established standard in place for categorizing
such ambiguous phenomena. He stated: “In [the case of the phenomenon of] the
Genyraku era [1184-1185], there was no ‘original star’ [ ]—yet [this phenomena]
was officially pronounced a comet. [Even] if [you say] now there is no originating star,
why is it that you are not calling it a comet?”
28
This statement implicitly allowed
Suehiro’s point about the 1189 phenomenon not fitting well into the category of comet to
stand while denying the validity of his objection to the standard categorization. Given
precedent for dealing with such phenomena, Kanezane would not accept Suehiro’s
attempt at innovation.
Ultimately, Suehiro capitulated to the regent’s demands and changed his report,
although whether he ever personally accepted the correctness of Kanezane’s
counter-arguments is unclear. Certainly, Suehiro would have had no options except trying
further to convince Kanezane of his position or rewriting his report, if he desired to have
his report officially submitted to the court. The regent stood as the gatekeeper between
Suehiro and the throne: as the regent held the right of document inspection ( Jp.
nairan), nothing Suehiro wrote would have reached the ruler without Kanezane’s
approval and seal. Nor would Suehiro’s observations be accepted into the official record
as an account of what had happened in the sky without Kanezane’s cooperation.
29

28
Gyokuyō Bunji 5/3/17.

29
It should be noted that this was the official document record, which only preserved official reports
submitted to court. Suehiro could have his opinion and findings preserved in non-official sources, such as
court diaries, which were compiled on personal whim and not restricted to those documents that made it
through official documentary procedures.

27
Although Suehiro possessed recognized skill and experience in astronomical observation
and interpretation, Kanezane held much of the power to determine what would be
accepted as truth and what would not stand.
Kanezane was evidently comfortable critiquing the work of Suehiro. Yet
Kanezane did not put forth his own interpretation or identification of the phenomenon;
instead, he demanded that Suehiro work within the established boundaries which he,
Kanezane, preferred. This shows that Suehiro still had a measure of authority in this
exchange that Kanezane lacked—Suehiro was recognized as able to perform
astronomical observation and identification. Kanezane, however, had enough knowledge
(at least in his own estimation) to critique and supervise Suehiro’s work.
Despite his position of power in this exchange, Kanezane did not let his record
of all this end with a mere record of Suehiro’s capitulation. Suehiro’s attempt to create a
new category of astronomical phenomena was apparently to Kanezane one of the worst
types of derelictions of duty, perhaps even something of an attempted plot. He continued
his grumbling even afterwards in his diary entry, noting that there was much in Suehiro’s
report that had remained unexplained, and that it, as well as Suehiro’s behavior, was
“most suspicious, most suspicious.”
30

All of this reveals that the production of facts about events in the sky was the
result of the interaction of multiple types of authority: the authority of the Chinese source
texts for astronomy and astrology; the authority of the observer who had the skill to
discern and categorize phenomena in the sky; the authority of the official documents and

30
Gyokuyō Bunji 5/3/17.

28
unofficial archives that shaped the expression of such phenomena for the state; and the
authority of the elite of the court bureaucracy, who studied and monitored the practice of
astronomical observation and astrological interpretation, and who outranked the
practitioners yet ceded to them public recognition of skills and expertise. Many of these
elements of competing authority were already present in the earliest days of sky-watching
for the state in eighth-century Japan. The particular balance of authority in 1189, however,
as this and the following chapter will show, was the result of changes in the structure and
practices of the Japanese court that had taken place over the eleventh through early
twelfth centuries. Such changes explain why and how some elements of the practice of
astronomical observation and astrological interpretation remained consistent over the
length of the Heian Period, while explaining motivations and opportunities for the
changes that were also occurring during this period.

Early History of Tenmon in Japan: State Technologies and State Control
The practice of tenmon ( , literally “heavenly patterns” or “heavenly
writing,” tianwen in Chinese)
31
was considered an important tool for rulership in Japan

31
In this section, when discussing the Japanese practice, the Japanese pronunciation of tenmon will be
used in the place of the modern Chinese tianwen. The reason for this will become clear in the discussion of
evidence for divergent practice of astronomical observation and interpretation.
The use of the term tenmon instead of the “observational astrology” used in the title of this
chapter is to emphasize the independent nature of this practice, between both modern conceptions of
astronomy (regular and systematic observation of the sky) and the divinatory ends to which such
observations were put. Despite the fact that historians of science have done much in recent years to show
the interconnectedness of the “scientific” astronomy and “superstitious” astrology in early modern Europe
(see Kepler’s horoscopy work, for example), the modern colloquial senses of the term remain very strong,
and for clarity in transmitting the nature of this work, I have decided to avoid using either term as an
English-language equivalent for tenmon/tianwen.

29
from even before the formulization of the ritsuryō state at the beginning of the eighth
century.
32
However, what tenmon was as a practice, and how to perform it, was not well
defined in early state documents. There was precedent for this: the details of tianwen
practice—how to conduct observations, how to select the appropriate interpretation—are
not to be found in the Chinese official histories ( Ch. zhengshi, Jp. seishi) in their
“Treatises on the Heavens.”
33
These Treatises were, furthermore, the very textbooks
assigned to official students of tenmon in the Japanese government. Nevertheless, it is
abundantly clear how important tianwen or tenmon were for both the Chinese and
Japanese states from the sources.
Even from the time of the founding of the Han Dynasty in the third century
BCE, the connection between the stars and rulership had already been long established.
34

This connection would become further formalized under the Han. In the “Book on
Heavenly Officials” ( Ch. Tianguanshu, Jp. Tenkansho),
35
for instance, the
famous historian and court astrologer Sima Qian described a bureaucracy and
architecture among the stars that matched that found in Imperial China. In this, Sima
Qian was not being innovative—merely systematic. Sima Qian was, as his father Sima

32
The importation of Chinese technologies and statecraft, and its role in the development of the Japanese
state, can be found described in detail in Joan R. Piggott, The Emergence of Japanese Kingship (Stanford,
Calif.: Stanford University Press, 1997).

33
(Ch. tianwenzhi, Jp. tenmonshi).

34
Pankenier, "Early Chinese Astronomy and Cosmology: The "Mandate of Heaven" As Epiphany". Aihe
Wang, Cosmology and Political Culture in Early China (Cambridge: Cambridge University Press, 2000)
also covers the early history and development of Chinese cosmologies involving the sky in greater context.

35
A section of the Records of the Grand Historian ( ).

30
Tan had been before him, a member of the Taishi liao ( , Jp. Taishi-ryō, “Bureau
of Divination”), in charge of both written materials as well as astronomical observation,
interpretation, and the construction of the civil calendar. While Sima Qian is more
famous for his work in compiling the historical work Records of the Grand Historian ( Ch. Shiji, Jp. Shiki), his training under his father had involved not only the study of
Chinese classics and philosophy but also the identities and meanings of the stars, which
he described in the “Book on Heavenly Officials.” It followed, according to Sima Qian,
that what happened above in the celestial empire could provide insight concerning what
had happened, or would happen, below.
36
This was justification and mandate for
systematically observing and recording what occurred in the nighttime sky. As the
predictive power of such observation depended on the correct identification of
phenomena—otherwise the correct earthly analogue for the heavenly phenomena would
not be known—tianwen or tenmon also required that its practitioners be extensively
trained in observation and classification techniques. This was the system of beliefs and
practices that was later transmitted to Japan in the seventh century.
Records of astronomical phenomena that are at least mostly reliable from the
point of view of modern astronomy begin to appear in the Japanese sources around first
half of the seventh century.
37
Many of the first records of particular phenomena date

36
On this “resonance” ( Ch. ganying, Jp. kan’ō) aspect of Chinese cosmological thought, see
Henderson, The Development and Decline of Chinese Cosmology.

37
Excluding, for example, interpretations of legendary or mythological events such as Amaterasu’s
seclusion in the heavenly cave as referring to a total solar eclipse.

31
from the period from 628 to 643. The first record of a solar eclipse in the Nihon shoki ( “Chronicles of Japan”) appears in 628.
38
It is followed by the first account of a
comet in 634;
39
of a meteor shower in 637;
40
of a moon-planet conjunction in 640;
41
of
a “guest star” in 642;
42
and of a lunar eclipse in 643.
43
The first record of a planetary
conjunction appears much later, in 692, during the account of Jitō Tennō’s reign.
44
What
these records show is a growing attention to and systematic recording of astronomical
phenomena at the Japanese court, as well as an increase of technical skill, with the
planetary conjunction, the phenomena that took the most skill to correctly identify,
appearing chronologically last.
The timing supports a famous account of the introduction of tenmon to the
court. According to the Chronicles of Japan, tenmon was first taught at court by the
Korean monk Gwalleuk
45
who arrived in the winter of 602 during the reign of Great

38
Nihon shoki Suiko 36/3/2. The next solar eclipse record is from 636 (Jōmei 8/1/1). This and the
following first occurrences were identified from Kanda Shigeru, Nihon Tenmon Shiryō. and confirmed
through a search of the texts of the Six National Histories (Rikkokushi), using the Kokushi taikei editions.

39
Nihon shoki Jōmei 6/autumn 8
th
month.

40
Nihon shoki Jōmei 9/2/23. The first record of a meteor shower identified by Kanda in his Nihon tenmon
shiryō refers to an event from BCE 15. The account, however, is from the medieval Mizukagami ( )
chronicle, and therefore should be best read as anachronistic. See Kanda Shigeru, Nihon tenmon shiryō.

41
Nihon shoki Jōmei 12/2/7.

42
Nihon shoki Kōgyoku 1/7/9.

43
Nihon shoki Kōgyoku 2/5/16.

44
Nihon shoki Jitō 6/7/28.

45
(Jp. Kanroku), dates unknown.

32
King Suiko.
46
He brought with him books on tenmon, geography, calendrical calculation,
and dunjia divination from the continent.
47
Upon receiving a royal order, Gwalleuk then
proceeded to teach others the techniques, passing down to one Ōtomo Takasato the skills
of tenmon.
48
Along with two other named students were assigned other subjects,
Takasato successfully mastered the arts taught him. While there are aspects of this
account that might bring parts of it into question—neither Takasato nor any of the other
two named students of Gwalleuk appear elsewhere in the text—it has never been fully
rejected by scholars. The timing of this account does seem to be approximately correct,
even though accepting the account as written means that it took over twenty-five years
before this new set of skills to was reflected in the record-keeping of the Japanese court.
Certainly, the aspect of the tale concerning the transmission of such skills from
a recent immigrant seems correct. Archeological evidence indicates that the observation,

46
Suiko ōkimi. While Suiko herself is known as a tennō now, as that term was applied
retroactively to all Japanese rulers, historical or legendary, during her reign the term had not yet been
adopted. The term ōkimi like tennō was the same whether the ruler was male or female. As queen in
English can refer to queen-regnant or to queen-consort, while king cannot refer to king-consort instead of
king-regnant, for this reason “king” is adopted for those female ōkimi of the past. See Piggott, The
Emergence of Japanese Kingship for the history of this terminology.

47
(Ch. dunjia, Jp. tonkō, literally “avoiding the yang-wood stem”) was one of the three main
methods of “mathematical astrology.” (See Ho Peng Yoke, Chinese Mathematical Astrology: Reaching out
to the Stars (London: RoutledgeCurzon, 2003) for more information on how this method was employed.)
Since then, the term has been read as referring to “arts of invisibility.” This is how Herman Ooms translates
the term, although he was by no means the first nor will be the last to do so. Herman Ooms, Imperial
Politics and Symbolics in Ancient Japan: The Tenmu Dynasty, 650-800 (Honolulu: University of Hawai'i
Press, 2009), 70.

48
Nihon shoki Suikō 10/winter 10
th
month. Ōtomo sukuri Takasato—sukuri is a kabane
( ) or lineage-rank—does not appear anywhere else in the text.
Ōtomo sukuri Takasato is misidentified as two individuals, Ōtomo Sukuri and Gōsō in Jae-eun
Kang, The Land of Scholars: Two Thousand Years of Korean Confucianism (Paramus, NJ: Homa & Sekey
Books, 2006), 45, and this error is likely to be found elsewhere as well.

33
even veneration, of the nighttime sky had strong continental connections. Some of the
first material evidence is tied to populations originally from the Korean peninsula, since it
is found in tombs associated with immigrant lineages.
49
Instead of accepting the account
as written, it may be more accurate to view the 602 date for transmission as being
somewhat anticipatory. Gwalleuk seems to have remained in Japan for decades after his
reported arrival, as he was promoted to a Buddhist prelate post in 624.
50
Some of the
temporal lag between the account of Gwalleuk as a teacher and the first reliable records
of astronomical phenomena may be due to his students not working for the court until
about that time.
Not long after knowledge of continental astrology and astronomy was available,
it was exploited. The utility of astronomical observation for rulership can be found too in
the subsequent account in the Chronicles of Japan that describes the ruler Tenmu
Tennō.
51
Tenmu came to the throne after rebelling against his nephew and winning the
resulting civil war, the Jinshin War ( , Jp. Jinshin no ran), in 672. He is
described in the Chronicles of Japan as having been expert in tenmon and other arts of
divination.
52
This sort of knowledge is implicitly credited as a source of military

49
For example, the star diagrams found inside the Kitora Kofun tomb.

50
Nihon shoki Suikō 32/4/ (jinjutsu) [17
th
day]. The rank he is given is (Jp. sōjō), and this is
the first time the rank appears in a historical record. Sōjō would be come part of the system of Buddhist
ranks and hierarchies in the ritsuryō system.

51
631–686, r. 673–686.

52
Nihon shoki Tenmu biography (chapter 28).

34
victories in the account of Tenmu’s rise to power, along with the intervention of
superhuman forces.
It is not surprising then that in the legal formulization of the Japanese state that
Tenmu and his heirs undertook,
53
tenmon and other forms of heavenly knowledge
became heavily restricted state secrets.
54
Namely, in the Chinese-style ritsuryō codes that
legally delineated the new Japanese state soon thereafter, the practice of tenmon was both
regularized and regulated.
55
The equivalent of the Chinese Taishi liang was established
as the Onmyō-ryō ( ), or Bureau of Onmyō. Like its Chinese counterpart, the
Bureau of Onmyō handled some forms of divination, the production of the calendar, and
the regular performance of tenmon.
56
Nevertheless, the organization of this new Bureau,
the first recorded mention of which dates to 675,
57
differed from the Tang institution in
that it combined in one government institution several sub-bureaus that were distributed

53
His consort and successor Jitō was chief among them. See Ooms, Imperial Politics and Symbolics in
Ancient Japan: The Tenmu Dynasty, 650-800, throughout.

54
On the political role of the control of techniques housed in the Onmyō-ryō, see ibid.. As noted in the
description above, however, some of these techniques such as tianwen or tenmon had a longer history of
being associated with the Chinese imperial state than with religious Daoism; as a result, referring to such
techniques as Daoist or “Daoissant” (as in Ooms) can be a bit misleading.

55
The following description, as are most descriptions of ritsuryō legislation, is based on the Yōrō
Ritsuryō of 718. While not the first ritsuryō codes to be composed and implemented, it is this set of codes
that survives in fragments (the ritsu) and in an annotated form in two commentaries (the Ryō no gige of 833,
and the Ryō no shūge of 868). Authoritative editions of these commentaries can be found in
Kokushi taikei hensankai, Ryō no gige, Kokushi taikei (Tokyo: Yoshikawa kōbunkan, 1979), ———, Ryō
no shūge, 2 vols., Kokushi taikei (Tokyo: Yoshikawa kōbunkan, 1979).

56
(Ch. yinyang, Jp. onmyō) was also glossed as “divination” in a 868 commentary on the ritsuryō
(Ryō no shūge, Shokuin ryō), further emphasizing the thematic connection between the Bureau of Onmyō
and the Taishi liang, even though the two were structurally distinct.

57
Nihon shoki Tenmu 4/1/ (heigo) [1
st
day]. The students of the Bureau joined the students of the
Bureau of the University ( Jp. Daigaku-ryō) in the New Year’s ceremonies honoring the throne.

35
among disparate departments in the Tang bureaucracy. Furthermore, the Japanese Bureau
operated with a smaller designated staff than its Chinese counterparts.
58
The law codes
mandated the presence of one Instructor of Tenmon ( Jp. tenmon no hakase)
who was in charge of “observing the appearance of heaven,”
59
presenting secret
memorials when abnormalities occurred therein, and instructing students on this
process.
60
While the codes made it clear that material on tenmon was not to circulate too
freely,
61
more details on the duties of the Instructor were not provided.
A more detailed description of tenmon cannot be found in texts predating a
commentary on the ritsuryō codes from shortly before 868.
62
In the Ryō no shūge ( , “Collected Exegesis on the Administrative Codes”), the compiler, thought to be
Koremune Naomoto,
63
includes some information on tenmon under the description of the

58
The most concise description of the differences can be found in Nakayama, A History of Japanese
Astronomy: Chinese Background and Western Impact.

59
The (Jp. keshiki) might also refer to signs indicative of the state of the
Five Phases, or of qi ( Jp. ki) in general. Ryō no gige, Shokuin ryō.

60
Ibid. The law codes designated that there should be a full cohort of ten students of tenmon within the
Bureau.

61
In the Zō ryō (Miscellaneous Administrative Codes) and the Zokutō ritsu (Penal Laws on Theft) as
discussed in below and in Ooms, Imperial Politics and Symbolics in Ancient Japan: The Tenmu Dynasty,
650-800.

62
Textual evidence indicates that the Ryō no shūge was compiled before the promulgation of the
Jōgan-kyakushiki ( ) in 869, which gives a terminus ante quem of 868. Hayakawa Shōhachi, "Ryō
No Shūge," in Kokushi daijiten, ed. Kokushi daijiten henshū iinkai (Tokyo: Yoshikawa kōbunkan,
1979–1997).

63
, dates unknown. Active 877 (Sandai jitsuroku Gangyō 1/12/25) to 907 (according to a
mention in the Gengo hiketsu shō . Tōkyō Daigaku Shiryō hensanjō, Dai Nihon shiryō, 12
vols. (Tokyo: Tōkyō Daigaku Shuppankai, 1968–), vol. 1 part 3, 785-786).

36
duties of the Head of the Bureau of Onmyō.
64
“Tenmon,” he writes, “is the sun, the moon,
the five planets and the twenty-eight lodges.”
65
Later, he provides another definition:
“The ‘Old Text’
66
says: tenmon is solar and lunar eclipses and astral anomalies.”
Naomoto goes on to cite from a number of texts, naming and listing both descriptive and
portentological features of astronomical bodies—primarily those of the planets and major
constellations.
The text is made up of a hodge-podge of information on astronomical bodies,
calendrical mathematics, and the meaning of signs. Naomoto includes as well a short
treatise on clouds, noting that: “heavenly patterns and qi ( Jp. ki) patterns together are
the two things.” Finally he adds a note stating that while the numerical matters of the
calendar are discussed and decided “within,” all things that touch upon tenmon should be
announced to the throne.
67
Notable, too, is the fact that although this commentary on the
law codes describes some of the features of the sky and what they mean, thereby

64
Jp. onmyō no kami.

65
Ryō no shūge, Shokunin-ryō
The “twenty-eight lodges” or “mansions” are twenty-eight constellations that divide the entire
sky (not just the portion visible at night) into twenty eight sections of different length around the zodiac
band.

66
(Jp. Kyūki) which has been identified as one of the lost commentaries on the Ritsuryō that
Naomoto consulted when compiling his text.

67
Ryō no shūge, Shokuin ryō.

37
supplementing the codes with the material (methods for making observations, and even
how often observations should be made) that they themselves left unstated.
68

Even in a fragment of an evaluation from the Tenpyō (729-749) era in the
Shōshōin archive, few clues emerge about the early practice of tenmon. When the
Instructor of Tenmon on the list, Ō Nakafumi,
69
is evaluated, like all the other members
of the Bureau of Onmyō that appear in this document, he is “a person well-skilled in
divination”
70
and his skills in a number of different styles is listed.
71
Tenmon only
appears as one of a type of divination in this list. While this shows how the
widely-known the divinatory efficacy of tenmon was (at least in the first half of the
eighth century), and that tenmon was one of the major skills by which a member of the
Bureau of Onmyō could be evaluated, unfortunately it reveals nothing of how the practice
of tenmon was carried out.
It was in 757 that the first official statement describing how practitioners of
tenmon should be trained appeared. In that year a decree ( Jp. kyaku) was issued
prescribing a education regimen for all students in the Bureau of Onmyō as well as

68
This was a potentially dangerous and illegal undertaking for Naomoto, depending on how it would
have been viewed under the laws forbidding the ownership or study of prognostication texts, discussed
below.

69
. Dates unknown. Nakafumi had been laicized in 701. Before this laicization his Buddhist name
had been Tōrō ( ). Shoku Nihongi Hōji 1/8/ (Jp. jin’in)

70
Tōkyō Daigaku Shiryō hensanjō, Dai Nihon komonjo, 25 vols. (Tokyo: Tōkyō
Daigaku shuppankai, 1968–1970), vol. 24, 553. more usually refers to efficacy, so it can be
assumed that the evaluator is considering Nakafumi’s divinations either accurate or efficacious in averting
the harm foretold.

71
Specifically, in Nakafumi’s case, the Great Monad ( Ch. , Jp. tai’itsu), Avoiding Wood Stem ( ), Six Yang Waters ( Ch. , Jp. rikujin) systems of mathematical divination; mathematical divination
( Ch. , Jp. sanjutsu); geomancy ( Ch. , Jp. sōchi); and tenmon.
38
students of medicine.
72
Five texts are specifically mandated for students of tenmon: Sima
Qian’s Book on Heavenly Officials; the two Treatises on the Heavens from the Han
History and the Jin History; a text identified as Sanjia buzan ( , Jp. Sanke
bosan), which is now associated with a syncretic chart of constellations and astronomical
phenomena;
73
and a divination guidebook referred to as Han Yang yaoji ( Jp.
Kanyō yōshū).
74
This list of mandated texts has influenced descriptions of tenmon in
Japan since, despite important evidence for change.
75

There are good reasons why more details on the early practice of tenmon in
Japan do not survive in the sources. The same codes that established the Bureau and the
practice of tenmon, however vaguely, also set restrictions on who was permitted to study

72
(Jp. isei). “Medicine” is not necessarily the most accurate or complete translation for this class of
students, as some types of medical treatment were not included in this field.

73
It has been identified with two surviving illustrated fragments dating from the Kamakura period, passed
down through the Wakasugi lineage, which had a hereditary relationship with the Tsuchimikado (formerly
the Abe). See Daitō Bunka Daigaku Tōyō kenkyūjo, ed. Wakasugike monjo "Sanke bosan" no kenkyū
(Tokyo: Daitō Bunka Daigaku Tōyō kenkyūjo,2004) and Murayama Shū'ichi, ed. Onmyōdō kiso shiryō
shūsei (Tokyo: Tōkyō bijutsu,1987).
The “three authors” ( ) of the title refer to three legendary Chinese astronomers, each of
which determined a separate system of constellations and divinatory associations, which were finally
unified in the Tang Dynasty. On the most prominent of the three systems, that attributed to Shi Shen ( ),
see Xunchen and Jacob Kistemaker Sun, The Chinese Sky During the Han: Constellating Stars and
Society (Leiden: Brill, 1987).

74
This text does not appear to survive in either Japan or China, but is identified as a divination manual.
Han Yang was a head of the Bureau of Astronomy under the Jin Dynasty. Ruijū sandai kyaku 5,
Tenpyō-hōji 1/11/9. (The edition consulted was Kokushi Taikei Hensankai, ed. Ruijū sandai kyaku, vol. 13,
Kokushi taikei (Tokyo: Yoshikawa kōbunkan,1979).)

75
For example, see Nakayama, A History of Japanese Astronomy: Chinese Background and Western
Impact. The 757 list, after all, does not include mention of either Tianwen yaolu or Tiandi ruixiang zhi,
which were the two most important texts in the composition of tenmon reports, as mentioned above.

39
or make pronouncements on celestial phenomena. First, the non-official interpretation of
astronomical phenomena was made illegal—the penal laws
76
of the ritsuryō code laid
out penalties for the private practice of astronomical interpretation. The “fabrication” of
stories or tales of marvelous phenomena, or else the “falsification” of interpretations of
omens that would “counter official authority”
77
and “mislead” the population, was to be
punished through exile or official corporal punishment. Buddhist clergy were considered
a particular risk for such practices, judging from the inclusion of the same prohibition in
the Sōni-ryō ( , Administrative Law on Monks and Nuns). Second, the private
possession of materials for the study of astronomy or astrology, along with other types of
“dangerous knowledge,” was forbidden by Article 20 of the Shikisei-ritsu ( , Penal
Laws Concerning Officials).
The private possession of the following items is forbidden:
astronomical instruments,
78
maps of the heavens, [divination
manuals such as] the River Chart [ Ch. Hetu, Jp. Kato] and the
Lo Writing [ Ch. Luoshu, Jp. Rakusho], books predicting the
future, books on military strategy or dealing with the calendar,
manuals of divination techniques
79
such as the Methods of the Great
Monad [ Ch. Taiyi, Jp. Taiitsu] or the Methods of the God of
Thunder [ Ch. Leigong, Jp. Reikō]. The punishment will be one

76
The ritsu ( ) of ritsuryō.

77
As found in the translation of the article 21 of the Zokutō-ritsu ( , Penal Laws on Theft) in Ooms,
Imperial Politics and Symbolics in Ancient Japan: The Tenmu Dynasty, 650-800, 93

78
Precisely what instruments these might have been remains unclear from other sources. The term used
here often refers to an armillary sphere, but need not necessarily indicate one.

79
The translation this is based on is from Ooms, Imperial Politics and Symbolics in Ancient Japan: The
Tenmu Dynasty, 650-800, 92. In it, Ooms uses “prognostication books.” However, both the Taiyi and the
Leigong cited here were specific divination techniques, and so I have adjusted the translation to better
reflect that.

40
year of penal servitude. The prohibition and penalty applies to the
private study of these materials.
80

Furthermore, Article 8 of the Miscellaneous Administrative Laws ( Jp.
Zō-ryō) restricted access to materials that would seem necessary to perform astronomical
observation and divination. It separated out the work of observation and divination,
stating that students of tenmon within the Bureau were not to read texts on divination.
Furthermore, all anomalies were to be reported to the tennō only through the Bureau of
Onmyō—the results were not to be shown to others. The reports of phenomena, once any
material concerning divination was stripped, were to be forwarded to the Ministry of
Central Affairs ( Jp. Nakatsukasa-shō), where they would be stored pending
inclusion in the official histories ( Jp. kokushi).
81
In other words, the law codes of
the eighth-century Japanese state placed limits on who might legally gain and exercise
the skills of astronomical observation and interpretation. Some information (the meaning
of the stars, how this meaning was gleaned) was more tightly controlled than others (that
there was an astronomical sign). It is little surprise, therefore, that details on how the
heavens were to be observed and interpreted are hard to find in eighth-century sources,
even as, at the same time, astronomical events begin to appear in the same sources.

80
Translation of Article 20 of Ritsuryō Shikisei-ritsu based on the one found in ibid., 92. The last
sentence of the article has been omitted, and will be discussed later in the chapter.

81
From the eighth century until the end of the ninth century, six “official histories,” the Rikkokushi (
) were produced. These were the Nihon shoki, Shoku nihongi, Nihon kōki, Shoku Nihon kōki, Nihon
Montoku tennō jitsuroku, and Nihon sandai jitsuroku.
One more, known as Shin kokushi ( “New National History”) was started, but never
completed. Parts of Shin kokushi were included in the compilation of Nihon kiryaku. See Hosoi Hiroshi,
Kodai no tenmon ihen to shisho (Tokyo: Yoshikawa kōbunkan, 2007).

41
Much has been made of how these laws made knowledge of astronomy and
divination a state monopoly, the exercise of which shored up and supported the
cosmological authority of the throne.
82
As in China, the early Japanese state could brook
no rivals when it came to translating the edicts of the cosmos, a major source of the
authority to rule in Chinese ideology.
83

It would certainly seem that the state was successful in keeping such
information secret, judging from the lack of details on practice that survive in the early
historical sources.
84
Yet the debate between Kujō Kanezane and Abe Suehiro shows that
institutionalized control was far from perfect. Even as the regent Kanezane monitored
Suehiro’s performance of tenmon within the political organization of the state, the fact
that Kanezane had the information necessary to even critique Suehiro’s work indicates
that much had changed since the letter of the ritsuryō laws limiting the practice and
dissemination of tenmon divination had been set in the eighth century.

82
Most notably and explicitly in Ooms, Imperial Politics and Symbolics in Ancient Japan: The Tenmu
Dynasty, 650-800.

83
Again, on the development of Chinese political ideology with relation to Heaven, see Wang,
Cosmology and Political Culture in Early China.

84
Hosoi Hiroshi has also been able to use how astronomical information appears in the Nihon kiryaku and
the Rikkokushi (aside from the Nihon shoki) to show how these chronicles were compiled. Using the legal
strictures that mandated that the astronomical information only be kept (and without the accompanying
divination) in the palace archives, he sows that at least through the Shin kokushi ( ) project in the
ninth century, state control of information continued to work as set out in the law codes. (The Shin kokushi
was an aborted attempt at a seventh addition to the Six National Histories that was cannibalized to create
later chapters of the Nihon kiryaku.) Hosoi Hiroshi, Kodai no tenmon ihen to shisho.
While I concur with Hiroshi that the structure of information processing and archiving
described in the law codes above continued to work roughly as described in the law codes, I wish to
emphasize, as I will show in the subsequent paragraphs, that there were early and explicit leaks in this
system. Such leaks allowed for external circulation of information gathered from astronomical observation.
What implications this may have for Hosoi’s textual analysis will have to be a topic for later study.

42

The Image of the Astrologer
While the interaction between Kujō Kanezane and Abe Suehiro, particularly
Kanezane’s own authority within that dispute, does not fit well with the depiction of
tenmon practice that emerges from these early sources, nor with studies of institutional
ideology and control that draw from these same sources,
85
it also does not correspond
with the traditional image of the diviner or astrologer as found in Japanese literature.
Fictional tales have been, in fact, the primary source scholars have used to derive a sense
of the astrologer and the role of astrology in Heian culture.
86
Yet, just as with the legal
codes, relying on these popular tales can be misleading as to what were actual historical
practices or social roles. In these sources, discussed below, the diviner as hero or advisor
is infallibly correct, and does not waver in the face of opposition. His skills are
otherworldly and effortless.
In the account in his court diary, Kanezane shows himself to be an informed
and skeptical reader of Suehiro’s report. Such skepticism does not much accord with
traditional images of a pre-modern, or even pre-“information society,” Japan—as

85
Primarily Ooms, Imperial Politics and Symbolics in Ancient Japan: The Tenmu Dynasty, 650-800..

86
A prime example is the pioneering work by Murayama Shū’ichi on Onmyōdō, especially his major
works Murayama Shū'ichi, Nihon Onmyōdō shi sōsetsu (Tokyo: Hanawa shobō, 1981), ———, Nihon
Onmyōdō shiwa (Osaka: Osaka shoseki, 1987).. This reliance on fictional depictions has been viewed
critically, as in the introduction to Shigeta Shin'ichi, Onmyōji to kizoku shakai (Tokyo: Yoshikawa
kōbunkan, 2004).
Examples of such images will be discussed below, using some depictions of Abe Seimei (
) and Abe Yasuchika ( ).

43
depicted by Ivan Morris and George Samson.
87
Writing about pre-modern Japan in the
first post-war decades, both depicted Japanese court society as irrationally controlled by
“superstition,” under the thrall of those men who could present themselves as part of
what sociologist Norbert Elias might classify as a “priesthood.” That the Heian court had
institutionalized forms of divination, including the regular observation and interpretation
of “heavenly patterns”,
88
was for these two scholars indicative of a cultural force that
retarded progress, and that was a sign of decadence and decay in Japanese elite society. A
tendency on the part of members of the Japanese court to employ divination (that is, to
interpret signs in the natural world) has been used there and elsewhere as an explanation
for why pre-modern Japan “failed” to progress, or to adopt newer developments in
astronomical calculation and prediction from the Chinese continent,
89
let alone develop
and implement innovations in social welfare or practical defense. Samson, in particular,
states that, “Scarcely any action could be taken by a Fujiwara nobleman in his public or

87
See Norbert Elias’ depiction of the power of “priest” (his example is ancient Egypt), as a depiction of
one theorist’s idea of the power relations before an “information society,” as occurred in Europe after the
print revolution, occurred. Norbert Elias (with Peter Ludes), "Knowledge and Power: An Interview by
Peter Ludes," in Society & Knowledge: Contemporary Perspectives in the Sociology of Knowledge &
Science, ed. Nico Stehr and Volker Meja (New Brunswick, NJ: Transaction Publishers, 2005). While this
term does not appear in either Morris or Samson, its usage encapsulates the attitude that these authors take
towards this “superstition” and its promoters.

88
(Ch. tianwen, Jp. tenmon), literally “heavenly patterns,” can refer both to the phenomena being
interpreted and to the act of interpretation.

89
This is a major explanation in Nakayama, A History of Japanese Astronomy: Chinese Background and
Western Impact. This explanation is picked up again by Sugimoto Masayoshi and David L. Swain, Science
and Culture in Traditional Japan, A.D. 600-1854., although there political isolation is given more
explanatory power.

44
private life without consulting the oracles,”
90
and “[e]dicts of very early date call
attention to the excesses of such men, who deceive the common people by their claims to
magic power; and there is no doubt that the prevalence of such superstitions lay like a
heavy cloud over mediaeval life.”
91
In short, the employment of and consultation with
diviners has not been seen as a positive aspect of Japanese court culture, but rather as a
hindrance that suppressed rational thought and inquiry.
Even recent works that refer to this aspect of Japanese cultural history have a
tendency to categorize practices relating to divination or omens using negative terms such
as “superstition.”
92
Not all recent studies do—but even those works that do not use such
negatively-valent terms
93
still tend to argue that it was through control of such ways of
knowing that these specialists in divination found a secure political and social place at
Japanese court.
94
The implication is that decisions concerning what statements were to
be presented and accepted as fact were made by either individual diviners or by the

90
George Bailey Sansom, A History of Japan to 1334 (Stanford, Calif.: Stanford University Press, 1958),
212.

91
Ibid., 214. Emphasis has been added.

92
For example, as in Roy Ron, "Powerful Warriors and Influential Clergy: Interaction and Conflict
between the Kamakura Bakufu and Religious Institutions" (University of Hawai'i, 2003).. For the history of
“superstition” as a negative (and nearly non-descriptive) category, see Dale B. Martin, Inventing
Superstition: From the Hippocratics to the Christians (Cambridge, Mass.: Harvard University Press, 2004).

93
For example, two recent books by Michael Como (Michael Como, Shōtoku: Ethnicity, Ritual, and
Violence in the Japanese Buddhist Tradition (New York: Oxford University Press, 2008), ———, Weaving
and Binding: Immigrant Gods and Female Immortals in Ancient Japan (Honolulu: University of Hawai'i
Press, 2010)) are relatively objective and value-neutral on the subject of divination.

94
See Como, Shōtoku: Ethnicity, Ritual, and Violence in the Japanese Buddhist Tradition and Ooms,
Imperial Politics and Symbolics in Ancient Japan: The Tenmu Dynasty, 650-800.

45
community of diviners. Quibbling about terminology is not part of the image of the
diviner in twentieth-century scholarship.

There are parallels to scholarly opinions about the power and influence of
diviners at the Japanese court to be found in popular Japanese tales about diviners that
date from the early medieval period.
95
For instance, the legends about Abe Seimei,
96
the
famous ancestor of Abe Suehiro, show him as supernaturally infallible.
One such legend involves the sudden abdication of the ruler Kazan Tennō in
986. This abdication, which was the result of plotting on the part of members of the
Fujiwara regental family, was performed at night and in secret, so as not to let the court
or public know. Kazan himself left the palace through a minor gate and in disguise. As he
passed by Seimei’s house, he and his attendants overheard the diviner state from behind
his closed gate that “[t]he heavens foretold His Majesty’s abdication, and now it seems to
have happened.”
97
An invisible spirit that was Seimei’s servant opened the gate and
reported back to his master that, indeed, the tennō had just passed by the house. The
accuracy of Seimei’s words, like his control of the world’s invisible powers, is clearly
meant to appear uncanny and uncannily accurate.

95
A few of the major setsuwa ( ) or “tales” collections that feature divination or onmyōji date from
the late Heian Period—such as Konjaku monogatari shū ( “Tales of Times Now Past”)—while
others date from the Kamakura Period, mostly 13
th
century—such as Uji shūi monogatari (
“Tales from the Uji Counselor”).

96
, 921 – 1005. “Seimei” is now the traditional reading of his personal name, although “Hareaki”
is also attested in older sources.

97
Helen. McCullough, Ōkagami, a Great Mirror: Fujiwara Michinaga (966-1027) and His Times: A
Study and Translation (Princeton: Princeton University Press, 1980), 81.

46
Similar stories about Seimei’s fifth-generation descendant Abe Yasuchika
98

appear in versions of the Tale of the Heike ( Jp. Heike monogatari), where
Yasuchika, upon seeing a comet, foretells Antoku Tennō’s
99
death by drowning at
Dan-no-Ura. In these stories, not only do the diviners discern the truth about things
hidden in the present and future, but they also obtain this knowledge through merely
glancing at the sky. There is no method, in other words, to this art. Likewise there are no
errors.
Yet no such depictions of how knowledge was gleaned from the stars matches
with the description of the argument that Kujō Kanezane had with Abe Suehiro in 1189.
Nowhere in such fictive or scholarly images of the astrologer is there room for a client
that not only demands changes to how the sky is seen and interpreted,
100
but furthermore
successfully does so. However, when looking more carefully at depictions of successful
astrologers in folklore, there are nonetheless hints of the prevalence of disputes. There are
even traces of the accurate predictor’s mirror image: the erroneous diviner.
In the tale of Kazan’s abdication, Abe Seimei has no rivals present. But in
many of the other tales that feature him, Seimei is often opposed by another diviner and
ritualist: most famously, Ashiya Dōman.
101
While Ashiya Dōman most likely did not

98
, 1110 – 1183.

99
, 1178 – 1185.

100
A process that, furthermore, seems to have much in common with taxonomic classification in natural
history.

101
, no known dates.

47
exist,
102
historical accounts of Abe Seimei’s activity show that Seimei did operate as one
of a number of diviners employed by the court. This was true even for astrological
divination, where he was one of two men officially employed for that purpose.
103
In
some versions of the Tale of the Heike story, Seimei’s twelfth-century descendant
Yasuchika also has rivals, and in one version they even prevailed over him for a time:
Yasuchika was exiled for his prediction.
104
Yet even these fictional dissenters—who
function more as plot obstacles than reasonable opponents—do not successfully oppose
an astrologer’s attempt at making astronomical facts to the same extent that Kujō
Kanezane does against Abe Suehiro in 1189.
Therefore, the logical question when faced with the image of the astrologer as
an unquestioned authority (or as an authority that should not be questioned, when he
appears to be temporarily thwarted in literature), is why the figure was depicted in this
way. While it may not be the sole motivation for the depiction, it is clear that such an
image of the astrologer in Japanese literature and popular tales benefited astrologers
when they faced opposition to their predictions. The cultivation of an aura of
reliability—even infallibility—was, after all, employed by tenmon experts in other ways.

102
At least, the Ashiya Dōman in the tales most likely did not exist, but he may be an abstraction of either
a lesser-known historical figure, or else a group of such figures. For more information on individuals or
groups that Dōman might be standing in for, see Shigeta Shin'ichi, Abe no Seimei: Onmyōji tachi no Heian
jidai (Tokyo: Yoshikawa kōbunkan, 2006).

103
As Instructors of Observational Astrology ( Jp. tenmon no hakase). See the section on
astronomical reports ( tenmon missō) in Chikanobu-kyō ki ( , 972-974) for how duties
were split between Seimei and his Nakahara colleague.
A critical edition of Chikanobu-kyō ki can be found in Satō Sōjun et al., ed. "Chikanobu-kyō ki"
no kenkyū (Kyoto: Shibunkaku shuppan,2005).

104
There is no historical record of Yoshichika ever having been exiled. This seems to have been an
invention of the author of the Nagato-bon Heike monogatari.

48

The Practice of Divination and the Production of Infallibility
There is evidence that diviners themselves sought to present themselves as
essentially infallible. An explicit example can be seen in Yasuchika-ason ki ( ), a collection of memorials from 1166 concerning astronomical and seismic events,
that provides one of the most comprehensive collections of such memorials that survives
to this day. All of the memorials collected in this text were authored or otherwise
co-signed by Abe Yoshichika. At the end of this compilation is an evaluation of the
predictions made over the previous year, showing how the astronomical events all
combined to foretell a number of destructive fires in and near the capital at the end of that
year.
105
This appendix reveals the reason for this text’s survival as well as its purpose:
the Yasuchika-ason ki was advertisement and proof of Abe Yasuchika’s preternatural
skill of discernment.
Moreover, according to the text, Yasuchika’s predictive streak of 1166 was not
limited to predicting these fires. In a note appended to a report of the moon eclipsing one
of the stars of the Southern Dipper,
106
Yasuchika is credited with having successfully

105
Yasuchika-ason ki. This text is available in two printed editions, but the most complete is to be found
in the Kurokawa Harumura, ed. Rekidai zanketsu nikki, 35 vols. (Kyoto: Rinsen shoten,1989-1990). series.
Due to the lunisolar year beginning later than the Gregorian or Julian calendar years, these fires
took place early in 1167. The fires occurred on Nin’an 1/12/10, when 20 chō ( ) and over 1000 houses
burned; Nin’an 1/12/22 when the northern part of the capital along with several structures at Enryakuji
were destroyed; Nin’an 1/12/23, when structures at Horikawa and Higashi Sanjō mansions burned. The
latter was the crown prince’s residence, and necessitated his moving to a new residence. The Nin’an
1/12/24 fire is also covered in the Hyakuren shō ( ) chronicle. (The text consulted was
Kokushi taikei hensankai, ed. Hyakuren shō, Kokushi taikei (Tokyo: Yoshikawa kōbunkan, 1979).)

106
otherwise known as the Dipper lodge ( ) it traditionally consists of six stars of the western
constellation Sagittarius: ζ Sgr, λ Sgr, µ Sgr, σ Sgr, τ Sgr, and φ Sgr. The particular star eclipsed was also
49
predicted the illness of the former provincial governor of Chikugo, Taira Iesada.
107

However, a closer look at this particular incident also reveals the inherent weakness to
using classical astronomical divinations to prove one’s accuracy. The method of
composing such memorials always included a range of potential readings for each
phenomenon. The report that “successfully” predicted the illness of the former provincial
governor ran, in full, as follows:
Respectfully announced:

This month on the twenty-third day, hi-no-to tori:
108
at dawn in the
double-hour of the Tiger, the moon eclipsed the second-star of kui
109

of the Southern Dipper.

Respectfully searched. The Tianwen yaolu states: the moon is the
spirit of the great yin, the symbol of a female sovereign. The
Southern Dipper is the court from which the Son of Heaven,
110

through virtuous authority, ensures smooth governance.

Divination states: When the moon invades
111
the Southern Dipper, in
the palace precincts loose words arise and the grand minister is
exiled. The Chart of the Tallies Granted by the Phoenix
112
states: the

named “dipper” (Ch. kui, Jp. kai or sakikage, although the latter Japanese term indicates a forerunner.
Pronounced “Sakikage,” it is also the name given to a Japanese probe sent to study Haley’s Comet in 1985),
and may have been Lambda Sagittarii (λ Sgr).

107
Yasuchika-ason ki Eiman 2/2/28. Taira Iesada ( -1167) would die the next year. Akihiro-ō ki
Nin’an 2/5/28.

108
(Ch. dingyou, Jp. also teiyū), the 34
th
day of the sexagenary day cycle.

109
The star eclipsed was as discussed in note 106 above.

110
(Ch. , Jp. tenshi), a common epithet for the Chinese emperor.

111
(Ch. fan, Jp. han or okosu). A translation closer to other meanings might be “violates” or
“trespasses,” but this verb was typically used when the moon, planet, or other moveable astronomical
phenomena entered into a portion of the sky designated by fixed stars.

112
(Ch. Chifeng fubiao). This text does not appear mentioned anywhere in the Siku quanshu
collection.
50
small people cause a large battle. The wise minister flees. Two years
do not pass.
113
The Three Spirits Record
114
states: When the moon
enters the Southern Dipper, it hides the great wind and rain. The five
grains are hurt by drought, and many people die from famine. Li
Feng’s Mirror
115
states: When the moon enters the Southern
Dipper’s kui dipper, the honored person comes to grief. Thieves enter
the palace. One year does not pass. The Tiandi ruixiang zhi states:
When the moon invades the Southern Dipper, thieves enter the
palace. It also states: Rebellious ministers revise the Son of Heaven’s
laws. The time, if it is soon
116
is thirty days. [Even] if it is distant,
117

three years do not pass. The Yisi-zhan
118
states: When the moon
enters and eclipses [the stars of] the Southern Dipper, the general
falls ill. The noble woman faces misfortune. The period is sixty days.
If far, a year does not pass.

The following anomaly
119
is respectfully announced and respectfully
memorialized.
120

Second year of Eiman [1166], second month, twenty-eighth day.

113
. This is the typical way temporal limits are designated in such reports.

114
This text does not appear mentioned anywhere in the Siku quanshu collection.

115
This text does not appear mentioned anywhere in the Siku quanshu collection.

116
Literally “close”: (Jp. chikai).

117
Literally “far”: (Jp. tooi).

118
(Jp. Isshi-sen). This text, which survives in a Siku Quanshu edition, provides lists of
interpretations for natural phenomena. It is known to have existed in ninth-century Japan as it appears in
the Nihon-koku genzai sho mokuroku ( , Fujiwara Ariyo c. 891; the primary edition
consulted was Fujiwara Sukeyo, "Nihon-koku genzai sho mokuroku," in Zoku gunsho ruijū, ed.
Zoku gunsho ruijū kanseikai (Tokyo: Yagi shoten, 1958).). It often appears in divination reports such as
above.
(Ch. yisi, Jp. isshi or ki-no-to mi) is 42
nd
in the sexagenary cycle, and in Five Phase
theory consists of a combination of wood (the or “stem”) and fire (Snake, the or “branch”), and so
thus can be read as “wood generates fire.”

119
(Ch. bian’yi, Jp. hen’i).

120
(Jp. sō). To present a memorial orally, it is somewhat redundant in the context of the formula used:

51
Junior fifth-rank upper, acting Instructor of Observational Astrology,
Abe Noritoshi
121

Junior fourth-rank upper, attendant
122
and Assistant Head of the
Bureau of Onmyō, Abe Yasuchika.

As the full memorial shows, a number of potential meanings from Chinese
sources are suggested by a single astronomical phenomenon, that of the moon moving
into a specific part of the sky. However, when this report was compiled into the
Yasuchika-ason ki text, a note was appended to the end of the report, stating: “Since the
twenty-second day the former provincial governor of Chikuzen, lay-monk Iesada fell ill
in the head
123
[which fulfills] the divination of ‘a general’s illness.’”
124
This note serves
to draw the reader’s attention to the correctness of the interpretation—it stands as a
confirmation of and comment on the divination.
However, when judged on its own, the memorial itself undermines the
conclusion suggested by the appended note. A number of potential implications for the
phenomena had been listed: drought, famine, and rebellion are also potential readings. It
is only after the fact, only after circumstances determined that “general’s illness” was the
“correct” interpretation out of all of these possible results, that in a perfect example of
circular reasoning Iesada’s illness proves the “correctness” of the interpretation of the
lunar conjunction. The language of the text itself is furthermore obscure and full of

121
, dates unknown.

122
(Jp. ōtoneri).

123
(Jp. nō) More accurately “brain,” but there is some doubt whether the diagnosis of Iesada’s illness
should be that specific.

124
Yasuchika-ason ki Eiman 2/2/28.
52
grammatical constructions atypical of other texts composed in Sino-Japanese ( Jp.
kanbun) from the same time period. This is certainly not the certain and clear
foreknowledge presented in the tales of diviners and their successes.
The indications that Yasuchika found in astrological phenomena throughout
1166 that “foretold” the fires at the end of the year were likewise embedded within a
number of potential associations. Although all of the indications for “fire” were present in
the divinations from the previous year, just as in the case of the “general’s illness,” the
“correct” reading of each report—discarding the other divinations, and emphasizing the
divination that indicated fire—could only be known in retrospect. It would have been
very difficult to prepare for the fires of 1166 in advance, even when forewarned by
Yasuchika’s detailed reports. There were many other potential results according to
Yasuchika’s readings at the time.
This presentation of choice, while it differs from the back-and-forth exchange
famously found in forms of African divination,
125
does allow for a social interaction
between diviner and audience that results in a divination that feels true. In the
archetypical cases of African divination, the correct divination is produced after a long
interpersonal negotiation wherein the diviner and client collaborate to read not only the
signs, but also the social circumstances of the query. In the case of Japanese tenmon
reading, the correct divination is reconstructed after the fact from post-hoc knowledge
and the selection of the most apt prediction from a number of options.

125
See the essays in Philip M. Peek, ed. African Divination Systems: Ways of Knowing (Bloomington:
Indiana University Press,1991), but especially Peek’s introduction to the volume.

53
Despite the impossibility of foretelling anything concrete through this method,
such uncertainty was not as large a handicap as it might appear. The precise future was
not as important in the practice of Japanese tenmon divination, as most reports (although
not the one cited above) included a generic warning against fire, illness, or other
misfortune in their summations. This non-specific danger could be yet avoided through
appropriate Buddhist or Onmyōdō rituals.
126

The appearance of having successfully predicted the future was undeniably
useful for the reputation of tenmon specialists. The fiction of the diviner as
always-accurate, and his words as indisputable, was useful both for establishing the
utility of divination and for establishing the reputation of an individual diviner. However,
as was shown above, this image did not correspond well to divination as it was actually
practiced.
This did not mean that accuracy was completely moot.
127
Furthermore, the
report itself had to appear “correct”—not just in function, but also in form. This is clearly
shown by the interaction between Kujō Kanezane and Abe Suehiro in 1189. Simply
because a variety of results could be indicated by a single phenomenon did not mean that
misidentifying the phenomenon was excusable. Instead of accepting any identification of
the phenomena so long as the resulting divination proved socially true, the top members

126
The typical phrase was . That either Buddhist or Onmyōdō rituals would work can be found
in the section on Jp. tenmon missō in the court manual Shin gishiki ( ). Unknown, "Shin
Gishiki," in Gunsho ruijū, ed. Zoku gunsho ruijū kanseikai (Tokyo: Yagi shoten, 1960).

127
The possibility of nothing happening after an ominous astronomical phenomenon was also not
discounted. In particular, debates periodically sprung up at court over whether subsequent periods of rain
lessened or “extinguished” such the malevolent influence of such phenomena. Some such debates are
referenced in the following chapter.
54
of the Heian court bureaucracy were constantly vigilant concerning the work of their
diviners. These noble bureaucrats were clearly concerned about maintaining the accuracy
of official knowledge. Despite the commonly presented image of the diviner or
astronomical observer as a “priest” that could “control” or “sway” the population through
his manipulation of pronouncements, it was instead the production of knowledge by the
diviner that was monitored, shaped—even, perhaps, “controlled.” The forces influencing
the diviner were the institutions and procedures established at court to ensure the
production of correct and accurate knowledge of the future gained through the stars.

But this was simply not a case of powerful men at the top of the Japanese state
dictating statements of fact made by lower-ranking client-specialists. Any control
exercised by the highest ranking men at court of the practice of astronomical observation
and divination was not and could not become complete due to the nature of the skills
involved: as the high nobility left the mastery of observational skills to their specialists,
and recognized their expertise, so then these observers and diviners could make
statements that their audience and patrons could not challenge. Suehiro’s failure in 1189
was that he attempted to use (what Kanezane thought was, at least) a neologism in his
description of what had appeared in the night sky. Such a neologism in the report
amounted to the invention of a new type of astronomical phenomena. For Kanezane, the
categorization of all possible astronomical phenomena had been set by centuries of
previous practice. However, as shall be shown in some of the examples discussed in the
next chapter, courtiers, including Kanezane himself, apparently hesitated before
55
dismissing the identification of a phenomenon once it had been made by one of these
court specialists, provided that it adhered to the accepted format. Specialists also
developed other strategies to ensure the acceptance of their statements. The text and
presentation of the Yasuchika-ason ki itself can be seen as such a statement, as can
legends about the preternatural accuracy of some diviners.
In short, facts about the stars and what they meant were the result of a
well-defined system of observational and textual practice, subject to social and
institutional negotiations. Statements of fact presented at court were thus subject to
questioning and dismissal. This was not “superstition” that cast a detrimental shadow
over all “medieval Japanese life,” as Sansom would put it. Instead, like all social facts, it
was produced through the practices and interactions of life at the Japanese court. The
facts produced through astronomical observation and divination, in this respect, did not
differ greatly from other types of knowledge produced at court.
128
It did differ, however,
in the degree of attention and monitoring given to explicit issues of its production.
How the nobility gained enough specialized knowledge to influence and
monitor the work of its tenmon specialists has much to do with key loopholes left in the
eighth-century law codes as much as it did with social changes in the middle and late

128
Nor does this differ, necessarily, from the acceptance of “truth” as pronounced by gentleman-scientists
in early modern England, as described in Shapin, A Social History of Truth: Civility and Science in
Seventeenth-Century England.
It should be noted, of course, that this was the acceptance of truth—the actual scientific validity
of statements made by Japanese specialists in astronomical observation and interpretation in the twelfth
century, and statements about vacuums and air-pressure made by members of the Royal Academy of
Science are different epistemological problems. On the distinction between “epistemography” (how things
were determined true , and its history, including the historical study of science) and “epistemology” (of
which the scientific method is normative), see Dear, "Science Studies as Epistemography."

56
Heian Period.
129
That these loopholes exist may in fact be further proof of the
importance that the framers of the Chinese-style bureaucratic state in Japan felt tenmon
held for legitimacy and political control. Due to the limits of the legislation concerning
who and how individuals could access tenmon-specific knowledge, courtiers such as
Kanezane were able to study the source texts and materials and thus pose a well-read
critique of attempts at innovation. If such monitoring was in fact one of the intended
results of such control, then it might be said that the eighth-century founders of the
Japanese ritsuryō state felt that tenmon was too important to be left to the tenmon
specialists.

The Institutional Limits of State Control
Earlier in this chapter, the early system legal control over who could
legitimately study tenmon and produce prognostications based on it was discussed. Yet
there were two major flaws in this attempt at control over the production of this specific
class of knowledge, particularly if it was meant to be absolute, of which the framers of
the ritsuryō system may or may not have been aware.
The first flaw, and initially the more pressing one, was the problem of keeping
the Bureau of Onmyō staffed. The court resorted to laicizing Buddhist monks to staff the
Bureau more than once in the eighth century,
130
including a laicization in 701 which
moved Ō Nakafumi, the Instructor of Tenmon evaluated in the first half of the eighth

129
These social and institutional changes are the subject of the following chapter.

130
See Ooms, Imperial Politics and Symbolics in Ancient Japan: The Tenmu Dynasty, 650-800.

57
century, to secular court life. Records also show that laicizations were ordered in 703
131

and 714
132
for individuals with skill in calendrical calculation and in divination,
respectively. That these were not punishments is made clear in the latter case, when the
former monk was concurrently granted court rank. The fear of personnel shortages within
the Bureau can also be seen in royal decrees mandating more economic support for its
work. Specifically, particular “salary” fields to support students of tenmon were
designated in 757,
133
and cloth—a medium of economic exchange—was also distributed
to students in 771.
134
In the case of first decree, tenmon was described as one of the ways
by which “customs were transformed” (that the population was civilized), and that it was
“essential for the state.”
135
That a decree delineating the texts all students of tenmon in
the Bureau were to master dates from later that same year is unlikely to be a
coincidence.
136

131
Shoku Nihongi Taihō 3/10/ (Jp. kōjutsu).

132
Shoku Nihongi Wadō 7/3/ (Jp. teiyū).

133
Shoku Nihongi Tenpyō 1/8/ (Jp. kigai).

134
Shoku Nihongi Hōki 2/11/ (Jp. heigo).

135
and respectively. Shoku Nihongi Tenpyō 1/8/ (Jp. kigai).

136
Despite these two counter measures against a shortage of staff working on tenmon in the Bureau, one
of the most obvious measures, the addition of new positions, came to the section on tenmon late: although
three “special students” ( Jp. tokugyōsei) were added to the divination ( Jp. onmyō.) section in
730, and two “special students” to calendrics ( Jp. reki). The special students were referred to as
(Jp. onmyō tokugyōsei.) and These special students were referred to as (Jp. reki
tokugyōsei), respectively. At this time, however, none were added to tenmon. (Ryō no shūge, Shokuin ryō,
quoting a directive of the Council of State ( Jp. Dajōkan) dated Tenpyō 2/3/27 (730).)
Furthermore, the first appearance of a Supernumerary Instructor for Tenmon ( Jp.
gon tenmon no hakase) also occurs latter than do supernumerary instructors for other fields: not until 973.
Chikanobu-kyō ki, Tenroku 3/12/6. In this entry, both Abe Seimei and Tōichi Mochitada ( , d. 981
(Keizu sanyō )) are referred to as tenmon no hakase. The first occurrence of the title gon tenmon
58
Both the laicizations and increased support suggest some of the problems that
the court had in keeping an adequate supply of official experts in astronomical
observation and interpretation available. Moreover, the case of the laicizations show that
despite the legal restrictions on knowledge relating to divination and the stars, sometimes
the best experts on such fields were to be found outside of the Bureau of Onmyō.
137
The
court could not afford to eradicate “outside” knowledge completely. This was a problem
in the idealized system of control described in the ritsuryō codes.
138
Ultimately, a second
Instructor of Tenmon was added to the Bureau in an attempt to help keep the number of
available experts steady, although even this proved insufficient, as was proved by
incidents in the eleventh century and discussed in the next chapter.
The second flaw in the control of tenmon-related knowledge can be found in
the text of Article 20 of the Penal Laws Concerning Officials itself, quoted earlier in this
chapter. The very last part of that article eases the restrictions that the law codes placed

no hakase was not until 1002 (Gonki, Chōhō 4/6/15). Zōhō shiryō taisei kankōkai, ed. Gonki, 2 vols., Zōhō
shiryō taisei (Kyoto: Rinsen shoten,1965).
By contrast, the first mention of a Supernumerary Instructor of Onmyō ( Jp. gon
onmyō hakase) is in 858 (Sandai jitsuroku Ten’an 3/9/2), and the first evidence of a Supernumerary
Instructor of Calendrics ( Jp. gon reki no hakase) can be found in a dispute dating to 836 (Shoku
Nihon kōki Shōwa 3/7
th
month), and the first direct mention of a Supernumerary Instructor of Calendrics
appears in 882 (Sandai jitsuroku Gangyō 6/1/7). Kokushi Taikei Hensankai, ed. Nihon kōki, Kokushi taikei
(Tokyo: Yoshikawa kōbunkan,1978), ———, ed. Nihon sandai jitsuroku, 2 vols., Kokushi taikei (Tokyo:
Yoshikawa kōbunkan,1979).

137
Although the Yōrō Ritsuryō codes were not compiled until 718, it is generally assumed that they did
not differ greatly from the Taihō Ritsuryō of 701, which is no longer extant. If this is the case, then the
legal restrictions on astronomical and divinatory knowledge discussed above date from the first laicizations
of monks to the staff the Bureau, the laicization that included Ō Nakafumi.

138
Herman Ooms identifies the control of such knowledge as a major component of the ideological
foundation of the early Japanese state. What the failure of such control means for his argument is a topic
for further consideration. Ooms, Imperial Politics and Symbolics in Ancient Japan: The Tenmu Dynasty,
650-800.

59
on the practice of interpreting astronomical phenomena: “However, apocryphal
interpretations of the Five Classics and the Book of History, as well as prognostications
based on the Analects are exempt.”
139
This was a major exception, and one that calls into
question the entire extent of the system of control that the state placed upon the
production of tenmon knowledge. The type of divination exempted from official control
here is, in fact, the major style of divination preserved in the “Treatises on the Heavens”
found in the Chinese dynastic histories themselves.
In this style of divination, the protasis or omen is the astronomical event, and
the apodosis or future event indicated is a historical event, often taken from
commentaries on the Spring and Autumn Chronicles.
140
This Spring and Autumn
Chronicles was one of the Five Classics referred to in the exception above, meaning that
one of the most significant means used to interpret astronomical events, the use of
Chinese classical texts and apocryphal exegesis of such texts,
141
was not legally
restricted in Japan. When surviving examples of divinations based on astronomical events
produced in Japan are examined,
142
these divinations themselves are mainly composed of

139
Translation taken from Ibid., 92.

140
This style of divination also is prominent in the text of editions of the Spring and Autumn Chronicles
themselves. The use of protasis and apodosis to describe the parts of an omen-meaning pair is taken from
the scholarly literature on Babylonian omen texts.
That there were multiple editions of Spring and Autumn Chronicles circulating at the time, with
different interpretations and slightly different accounts, can be seen in the discussion in Schaberg, A
Patterned Past: Form and Thought in Early Chinese Historiography..

141
On the development of the Chinese apocryphal tradition, see Jack L. Dull, "A Historical Introduction
to the Apocryphal (Chʻan-Wei) Texts of the Han Dynasty" (University of Washington, 1966).

142
As in the example involving the conjunction between the Moon and a star of the Southern Dipper in
1166 from the Yasuchika-ason ki quoted above.

60
quotations from the apocryphal ( Ch. chanwei) tradition based on Chinese
texts—the so-called “weft texts” ( Ch. weishu, Jp isho).
143
In other words, all
surviving examples of official tenmon divination in Japan utilized a method that was not,
in fact, restricted to the Bureau of Onmyō by the ritsuryō codes.
Whether this was always the case for the Japanese practice of tenmon is unclear.
All of the surviving examples of tenmon divination date from the tenth century onwards.
No astronomical event in the Six National Histories ( Jp. Rikkokushi) for the
eighth century or earlier survives with its divination attached—in accordance with the
legal restrictions found in the ritsuryō codes. It is therefore impossible to know for
certain if the practice of tenmon divination in eighth-century Japan matches that from the
tenth century onwards in relying on the Chinese “weft” tradition. As a result, it cannot be
determined whether the predominance of weft-text style interpretation evolved after the
eighth-century as a result of its permissibility according to the law codes, or whether
surviving memorials from the tenth-century onwards reflect instead a continuation of a
tradition of practice that dates from before the eighth century. In the latter case, the
exemption in the law codes for weft-text based interpretation reflects rather than shapes
eighth-century tenmon practice. Whether weft-text based interpretation was an innovation
from the tenth century onward, or whether it had always been the major mode of
astronomical interpretation in Japan, is impossible to determine from surviving sources.

143
This tradition of Chinese texts was often apocryphal interpretations of the classics ( Ch. jing, Jp.
kyō) of the Chinese tradition. See Dull, "A Historical Introduction to the Apocryphal (Chʻan-Wei) Texts of
the Han Dynasty"..
61
In either event, the presence of this exception opened an important legal
loophole for the circulation of tenmon divinatory reports. What it did not open, however,
was any room for the circulation of astronomical charts or tools that might have permitted
more detailed lay observation of the heavens. The ability to circulate interpretative texts
when combined with the restriction of technical training continued to play a significant
role in the history of the practice of tenmon in the service of the state at the Japanese
court. Meanwhile, the separation of the texts and practices of divination from the texts
and practices for observation would establish types of expertise as well. The result was a
division of expertise and authority at court. Only those with training, which originally
meant individuals from the Bureau of Onmyō, were recognized to have the skills to
identify and classify astronomical phenomena.
Despite this expertise gap, the circulation of divination reports opened up a
means by which courtiers and officials could critique the work of their experts. The
image of the diviner as all-influential and unquestionably believed is not only ahistorical,
it was undermined by the very structure of the law codes that established divination as the
business of the state.

Conclusion
Despite a pervasive image of the tenmon specialist or astrologer at the Heian
court as a Svengali or Rasputin-like figure with unquestioned influence and authority,
historical examples show that this image far from reflected historical reality. In fact, the
state controlled the legal production of tenmon knowledge. Furthermore, there were
62
important early loopholes in the system, weakening laws of state secrecy, which opened
up the legal room for a more fine-grained court-level control of the production of tenmon
knowledge. The permissibility of studying tenmon created a situation wherein individual
courtiers could intervene and correct “errors” produced by specialists—at least at the
textual level. While these loopholes date from an early stage in the history of the
Japanese ritsuryō state, it would not be until the Heian courtly state of the tenth through
twelfth centuries that these loopholes would be fully exploited.
As the next chapter will show, the circulation of information about the night
sky, faintly attested to in the eighth century, would only increase in prominence as
changes in the organization of the state, and in the flow of state documents, occurred in
the ninth century. These structural changes would set the stage upon which public debates
about astronomical phenomena could occur. These changes would also lay the
groundwork for the presence of the archive that Kujō Kanezane could call upon to
critique the work of tenmon experts in the twelfth century. In addition, new groups of
interpreters of astronomical phenomena would appear in the next centuries, occupying
space opened up by loopholes in the eighth-century law codes. This increased population
of astrological experts would provide further raw material for courtiers intent upon
critiquing the work of their divination experts.
And yet there was still room for specialist expertise, as the restrictions on the
circulation of maps and charts appear to have remained in force. Kujō Kanezane, and
others like him, never suggested that they could observe and interpret the night sky
63
themselves, even as they reserved the right and ability to critique the divinations of
others.

64
CHAPTER TWO

LEAKS IN THE SYSTEM: ACCESS AND AUTHORITY IN THE CHANGING
PRACTICE OF TENMON

As shown in the previous chapter, the legal control over tenmon practice at the Japanese
court was flawed and incomplete from the start. Legal restrictions on such knowledge
were further undermined by new systems of documentary control at court. When these
new bureaucratic standards combined with archival practices at court that encouraged the
development of household archives, the result was a situation where high-ranking
courtiers had extensive access to previous examples of astronomical divination reports.
Thereupon, courtiers had the ability to critique and further control the production of facts
about the meaning of astronomical events. This corpus and the knowledge gained from it
did not merely exert a conservative pressure on the practice of tenmon at the court—it
also provided the inspiration for the development of a variant off-shoot of tenmon
practice, wherein the texts that provided the meaning of astronomical phenomena were
not the Chinese classics, but from Japanese history.
In contrast to the textual elements of tenmon practice, the knowledge necessary
for the proper identification of astronomical events, as well as the expertise involved in
astronomical observation, did not circulate nearly so far. This allowed for a new strategy
utilized by tenmon specialists when their identifications of astronomical events were
criticized on a textual level. These specialists could now cite the special knowledge
necessary to identify phenomena that were difficult to distinguish from other, related
65
phenomena. The use of this strategy can be seen in the identification of comet subtypes,
which—as comets portend regime change—were a particular area of concern for the
Japanese court.
Furthermore, esoteric techniques and lineage secrets were strategically utilized
by tenmon specialists to enhance the authority of their statements. This strategy was both
a response to doubt about authoritative knowledge about the meaning of astronomical
phenomena, as well as one of the causes for doubt about such knowledge. The strategies
used by tenmon specialists in the face of such pressures further reinforced those pressures.
The presence of conflicting divinations became not at all unusual, even almost expected.
As a result, interpersonal trust and persuasion were the keys to creating and promulgating
the meaning of astronomical phenomena. Such a situation shows the instability in
astronomical knowledge at the Heian court, and makes clear that astrology was not an
inhibitory force on the “progress” of early medieval Japanese society, but instead a site
where the forces that shaped the intellectual history of early medieval Japan played out.

A Normal Sky
Late in 1107,
1
the courtier Fujiwara Munetada
2
received Senior Secretary of
the Council of State Nakahara Morotō
3
for a discussion of court affairs. 1107 had not at

1
December 10, 1107 in the Julian reckoning.

2
, 1062 – 1141.

3
, d. 1130. This Nakahara Morotō, the second of that name in the Heian Period and grandson to
the first, will be referred to in other sections of this chapter as Morotō II.

66
all been a good year for the court: Horikawa Tennō had died young in the summer of that
year after a reign marked by illness. Munetada’s distress at events was both personal and
political. Munetada had held high hopes for Horikawa’s reign that were dashed by the
tennō’s sudden death. Munetada had also had served the former ruler personally as a
royal attendant, and had a close relationship with him. In 1107, he was not therefore a
man to see signs of good fortune around every corner.
On the back of his court diary entry for this particular day, Munetada wrote an
additional note about an astronomical matter Morotō had brought up. Munetada at the
time was a member of the Council of State, although not among the highest ranking
members. He had finally been promoted to Middle Counselor ( Jp. chūnagon) the
year before, but this did not mean that he had the status necessary to influence major
decisions made in the Council. Nor was Munetada of a rank or position that normally
would receive astronomical divination memorials ( Jp. tenmon no missō)
directly. Morotō, however, was a close ally,
4
and although Munetada did not have the
rank to receive astronomical divinations personally (that was limited, officially, to the
regent), Morotō, as Major Secretary to the Council of State ( Jp. daigeki), had
greater access to such materials.
5
Morotō brought the aforementioned information
directly to him.

4
Morotō II’s son, Moroyasu ( d. 1154) would also be an ally of Munetada’s and would take his
father’s place as an advisor and transmitter of information to Munetada after Morotō’s death. This can be
seen in post 1130 entries of Munetada’s court diary Chūyūki

5
The Senior Secretaries, or indeed any members of the Secretariat to the Council of State ( Jp.
Geki), were not directly related to the document chain that had Instructors of Tenmon delivering memorials
to the regent first for approval, then to the Palace Secretariat for conveyance to the throne itself—a process
67
According to Morotō, the Instructor of Tenmon Abe Muneaki
6
had recently
announced the appearance of the “old man star” ( Ch. laoren xing, Jp. rōjin-sei)
in the skies that year. This was a particular auspicious astronomical phenomenon. The
laoren xing is known to Western astronomy as Canopus, the second brightest star visible
from earth after Sirius. It is primarily a southern-hemisphere star, circumpolar when
viewed from locations south of 36 degrees latitudes south. Canopus only becomes visible
in the northern hemisphere, and then under ideal conditions, from points south of about
37 degrees north latitude, and it becomes more easily visible as the observer moves
further south. Both Chang’an and Kyoto are located at about the same latitude (34
degrees north and 35 degrees north, respectively). Thus the rarity of conditions of
sufficient quality to sight Canopus, along with its brilliance, contributed to its attribution
as a sign of benevolent rule.
Munetada, however, was not much inclined to think that the Japanese court in
1107 deserved such a sign of benevolent rule and magnificent future fortunes. He could
not accept it.
7
Munetada and Morotō found room to criticize the sighting of such even an

described later in this chapter. However, reports having made this journey might have then been submitted
to the Council of State, in which case they would have passed through the Secretariat’s—in this case,
Morotō II’s—hands.
More importantly and likely, Morotō II was himself an officially recognized tenmon specialist,
although one who had never held a post within the bureau—he had previously received an official decree
permitting him to make tenmon memorials to the court (a decree known as Jp. tenmon
missō senji). It is almost certainly through this function, and so that he might “check” the work of Muneaki,
that Morotō II had received a copy of the initial report.
Morotō II’s actions here can be seen as strategic—disparaging his composition, Abe Muneaki,
in the presence of a sympathetic and higher-ranking audience.

6
, dates unclear.

7
Chūyūki Kashō 2/intercalary 10/24.
68
auspicious sign. Munetada wrote that Canopus was typically a star of the southern
horizon, and thus it was only when the star rose high in the sky that it should be
considered a fortuitous omen. But in recent days it had remained on the southern horizon,
and therefore its appearance did not count as one. There would be no legitimate
celebration at court—none, at least, that Munetada or Morotō would recognize.
Munetada’s stance in 1107 depended largely upon his personal opinion of the
state of political and cosmological affairs. His beloved ruler had died needlessly young
and the court was not functioning as it should. Therefore he would not accept word of a
good omen, and found technical reasons through which to criticize the divination report.
This critique, moreover, was based on the characteristics of the star
Canopus—knowledge that Munetada doubtless received from Morotō and his personal
archive of past astronomical reports. This knowledge was, in other words, the result of
leaks in the state’s control of astronomical expertise. Munetada’s access to such
knowledge had resulted from the structural weakness in that control that existed from the
very establishment of the Japanese bureaucratic state in the eighth century and had only
strengthened thenceforth. The leaks were archive-based as well as interpersonal.
Whereas Kujō Kanezane depended on his personal archive of
astronomy-related material to critique Abe Suehiro’s definition of the phenomenon of
1189 as a “cometary aurora,” Fujiwara Munetada in 1107 was relying on the expertise of
Nakahara Morotō, an officially recognized expert in astronomy even though he had never
held a position in the Bureau of Onmyō. Both private archives and personal connections
to astronomical experts would shape the practice of astronomical divination over the
69
duration of the Heian Period (794 – 1192) by providing courtiers the access to certain
types of expert knowledge that was necessary for the monitoring and critique of the work
of tenmon experts. With such access, and their higher status at court, the top nobility
could influence the way that phenomena were identified and analyzed. They became
active participants and an influential force in the practice of astronomical observation and
interpretation in Japan.

Leaks in the System
As was laid out in Article 20 of the Penal Codes Concerning Officials, there
was a style of divining the meaning of astronomical phenomena that was not, in fact,
legally restricted to officially sanctioned members of the Bureau of Onmyō. This
exception allowed the first hints for how astronomical phenomena were interpreted in
Japan to appear in the official state record, despite the provision in Article 8 of the
Miscellaneous Administrative Laws mandating that divinations be stripped from tenmon
reports that were to be included as raw material for the Six National Histories ( Jp.
Rikkokushi).
Of all of the Six National Histories, it is well established that the last official
national history, Nihon sandai jitsuroku ( , “The Veritable Record of Three
Reigns of Japan”) differs in character from those that went before: when compared to the
previous four national histories, more omens and ominous natural phenomena appear,
70
and the text is less shy about indicating the causes of such anomalies within its pages.
8
It
is also in The Veritable Record of Three Reigns that the first textual evidence appears for
what sort of divinatory techniques was used on astronomical phenomena.
The first reference is indirect. On the twenty-ninth day of the eighth month of
858 (October 9, 858), a shooting star “red like a blaze,” appeared in the night sky near the
pole star.
9
The head of the Bureau of Onmyō and the Instructor of Tenmon presented a
memorial interpreting its meaning. Then, in an apparently new development, the Office
for the Official Histories ( Jp. Shūkokushi kyoku) summoned the members of
the Bureau and had them compare the memorial to what was written in the histories.
10

This is the first indication of the sort of monitoring of what appeared in astronomical
reports that Kujō Kanezane would perform in 1189, although Kanezane’s actions then
were of a different quality. Clearly, the matter in astronomical memorials was no longer a
well-kept secret between Bureau and throne, if they indeed ever were such.

8
Particularly when compared with the one immediately preceding, Nihon Montoku tennō jitsuroku (
, The Veritable Records of Montoku Tennō of Japan’s Reign). Yamashita Katsuaki attributes this
difference to both cultural trends and the influence of beliefs about omens, curses, and vengeful ghosts ( Jp. onryō or Jp. goryō). Yamashita Katsuaki, "Onmyōdō no seiritsu to jukyō-teki rinen no
suitai," Kodai bunka 59, no. 2 (2007)..
The Nihon shoki must be treated as at least somewhat sui genera in this respect, due to the
sections on legendary tennō and the Age of the Gods ( ).

9
(Ch. Ziwei gong, Jp. Shibi-kyō “Purple Forbidden Palace”), more properly known as
(Ch. Ziwei yuan, Jp. Shibi-en “Purple Forbidden Enclosure”), it consisted of Ursae minor and major, as
well as multiple stars from Draco, Cassiopeia, Boötes, and Cepheus. This part of the sky contained most of
the celestial court. The length of the shooting star is given as about 10 jō ( ). Sandai jitsuroku Ten’an
2/8/29.

10
ibid.

71
The second reference to the text of a divination is from seven years later. At
some point before the first month of the year Jōgan 7 (865),
11
the state’s Buddhist
prelates in the Sōgō ( ) had been ordered to perform rituals in response to an
astronomical divination that the Bureau of Onmyō had presented to the court. While the
actual astronomical event is not described in detail, the resulting prediction was. “In the
previous year, the Bureau of Onmyō memorialized that in the coming year there should
be military uprisings and epidemic disease. Recently, the Instructor of Tenmon
memorialized: that military uprisings should be anticipated.” The Sōgō responded to the
court that such dire events would not take place, provided that the appropriate sutras were
read.
12
Technically, this reference to a divination did not violate the rule set out in
Article 8 of the Miscellaneous Administrative Codes as it did not appear together with its
astronomical referent—this, however, is hair-splitting. The record of 865 is the most
direct and earliest reference to the result of a divination based on an astronomical event
referred to in the official histories. Yet another reference to the results of an astronomical
divination would appear in a royal edict from 867.
13
Clearly, astronomical matters were
being discussed at court, and under such pressure official secrecy could not be expected
to hold.

11
865 is the common-era year in which the lunisolar year Jōgan 7 began. The first day of Jōgan 7 was in
fact January 31, 865. Therefore, the astronomical events and counteracting rituals took place at some point
in 864 or in January 865.

12
Specifically, the Diamond Sutra ( or Vajra prajna paramita sutra). Sandai jitsuroku Jōgan
7/1/4.

13
Sandai jitsuroku Jōgan 9/11/29.

72
The Veritable Records of Three Reigns was the last of the official histories
produced by the Japanese court, and it only covers the period up through 887. Whether
Article 8 of the Miscellaneous Administrative Codes would have been fully abandoned
given enough time is therefore unclear, due to the lack of subsequent official national
histories for tenmon material to appear in. In any event, the ninth century also saw a
major change in the structure of the court, and subsequently in how documents were
relayed and circulated. This change played a major role in the building of private archives
of astronomical and divinatory material over the next two centuries. It is likely that the
new system also played a role in breaching the wall of secrecy around the astronomical
divinations performed by the members of the Bureau of Onmyō.
Around the time of the attempted coup known as the Kusuko Incident ( Jp. Kusuko no hen, 810), the victorious reigning monarch, Saga Tennō,
14
undertook a
reform of court personnel who worked closest to the throne. Specifically he added a
group of male retainers, the Royal Secretariat ( Jp. Kurōdo-dokoro), who were
now to handle many of the responsibilities involving documents and oral communiqués
to the throne that had previously been handled by female attendants ( Jp. naishi).
15

Instead of oral memorials presented directly to the throne or conveyed by female
attendants, communication now had to go through members of the Secretariat.

14
, 786-842, r. 809-823.

15
This extra-legal organization had to sometimes adopt female personas when signing documents, due to
the prescriptions in the legal codes that had established the naishi as those responsible for handling
documents going to and from the throne. Toyoda Masahiro, "Naishi-sen," in Kokushi daijiten, ed.
Kokushi daijiten henshū iinkai (Tokyo: Yoshikawa kōbunkan, 1979–1997).

73
This new official document chain also became the standard operating procedure
for memorials on astronomical phenomena presented by either the Head of the Bureau of
Onmyō or by the Instructor of Tenmon. Although the term for memorials on astronomical
phenomena, tenmon no missō ( ), implies an oral presentation ( Jp. sō),
surviving sources instead indicate that by at least the tenth century, the information was
now conveyed through documents that were passed from the memorializer to the throne
through the Secretariat.
Moreover, the members of the Secretariat now also began to question the
interpreters of astronomical phenomena on any obscure or doubtful points in the
memorial. As the actual interpreter of the astronomical event no longer had direct access
to the throne, it was a Royal Secretary ( Jp. kurōdo) acting as go-between who
presented the information. In order to convey the information correctly, the Royal
Secretary needed to confirm his understanding of the material in the document before
submitting it to the ruler. If the throne or his deputy did not understand or doubted any
parts of the memorial, the Secretary would then be forced to go back to the tenmon
specialist to ask for clarification. It was likely this duty that resulted in members of the
Secretariat copying out sections of these “secret” memorials for later study, the first of
many “leaks” in document secrecy at court.
The first evidence of this information “leak” can be seen in the tenth-century
Chikanobu-kyō ki, a topically organized reference drawn from the personal court diary of
74
Taira no Chikanobu,
16
from the period when he was one of the two heads of the Royal
Secretariat ( Jp. kurōdo no tō). As only one of the two heads of the Secretariat,
and only one member of a rotating staff that served as the intermediary between
memorializer and throne, Chikanobu did not have personal access to all the astronomical
and divinatory information that passed through the court during these years. Nevertheless,
Chikanobu did take the opportunity to copy information from memorials that passed
through his hands while on duty, and a section of the manual based on his diary is
dedicated to them. These cribbed memorials would have been useful for Chikanobu’s
later study of the material, or for educating his heirs, themselves potential future Royal
Secretaries. That Chikanobu did not understand everything that went through his hands
can be seen in his copies of the memorials, which have some mistakes in the official
names of stars mentioned therein.
17

The Royal Secretariat would not pose the only institutional leak in the dike of
secrecy that had been erected around astrological memorials. Another breach opened in
the tenth century along with the rise of Fujiwara regental power. The Fujiwara regency is
most closely and famously associated with the Northern Fujiwara lineage’s marriage
politics. The position of regent, either sesshō ( , technically used when the tennō was
young or otherwise incapacitated) or kanpaku ( , a regent who operated on the behalf
of an adult tennō), was legitimated due to the official relationship between the Fujiwara

16
, 946-1017.

17
Chikanobu-kyō ki, tenmon no missō section. The “spelling” error is noted in Kanda Shigeru, Nihon
Tenmon Shiryō.

75
regent and the throne: stereotypically, the regent was the maternal grandfather of the
tennō. He was therefore deferred to, however nominally, as a result of that senior familial
connection.
18
The particular means through which the Fujiwara regents exercised their
power over the court was through the performance of the act and duty of nairan ( ) or
“document inspection.”
19

In the nairan system, all documents intended for the throne first had to be
approved by the regent or other individual who held the right of nairan. Requests or
memorials could be expedited, or else might languish, at this stage in the document chain,
making the man who held this right exceptionally powerful at court. Once this system
was established, before a divination on an astronomical event could even be presented to
the throne through the Royal Secretariat, it first needed to be read and stamped with the
regent’s seal.
20

18
Regents did not always hold this relationship to the tennō—maternal uncles were also common. As the
position of sesshō and kanpaku became institutionalized, the familial relationship became less important,
and by the Kamakura Period some Fujiwara were appointed to this position even though they held no close
familial link to the current occupant of the throne.

19
For example, this can be shown through the career of Fujiwara Michinaga ( , 966-1028), the
most famous of the Fujiwara “regents.” While definitely the most powerful man at court, Michinaga never
held the title of kanpaku and only became sesshō in 1015, towards the end of his career, long after his rise
to power. Instead, Michinaga ruled as regent while Minister of the Left ( Jp. Sadaijin), and through
his right of nairan. He received the two posts in quick succession, receiving nairan in 995, and becoming
Minister of the Left in 996.

20
On the appearance of the seal on tenmon no missō, see Shin gishiki 4. This early tenth-century text
shows an earlier stage in the development of this process, where the necessary seal belongs to “the senior
minister” ( ).

76
The first detailed record of this process can be found at the beginning of the
eleventh century,
21
although there is evidence that the procedure was in place by the first
half of the tenth century.
22
According to a description of the procedure from the eleventh
century, the regent not only read the report describing and interpreting the astronomical
event, but he also kept a copy of the report in his possession. Presumably this was to
guard against any modifications to the material before it reached the throne. The result
was, however, that regental households now held an archive of tenmon reports that could
be studied, and potentially compared with each new report brought forth by officials from
the Bureau of Onmyō.
So by the end of the tenth century, there were at least two different avenues
through which private archives of tenmon-related knowledge were being built up, as
some courtiers utilized their access to “secret” reports through their positions as members
of the Royal Secretariat, and others obtained access through the regent’s nairan privilege,
either directly or through personal connections to the regent’s household.
23

21
Midō kanpaku-ki, Kankō 4/12/22 (February 1, 1008). Michinaga, as nairan, read and sealed the
memorial. The editions consulted were Tōkyō Daigaku Shiryō hensanjō, ed. Midō kanpaku-ki, 3 vols., Dai
Nihon kokiroku (Tokyo: Iwanami shoten,1952–1954), and Yōmei bunko, ed. Midō kanpaku ki, 5 vols.,
Yōmei sōsho (Kyoto: Shibunkaku shuppan,1983–1984).

22
A few records of astronomical phenomena with enough detail that it can be assumed that the
information originally came from a tenmon specialist can be seen in the abbreviated version of regent
Fujiwara Tadahira’s ( , 880-949) diary, Teishin-kō ki shō. See Jōhei 1/7/7 (931), Tenryaku 2/1/19
(948). Although a record of the moon being visible during the day can be found for Enchō 3/7/7 (925), and
shooting star event is recorded for Tengyō 9/1/2 (946), the level of detail in these two records does not
necessarily indicate that the original source was a tenmon report.

23
Records of astronomical events found in diaries kept by courtiers who also had important roles in
regental households, such as Fujiwara Yukinari’s ( , 972-1028) Gonki also indicate, in their level
of detail, access to and copying from tenmon reports.

77
It should be remembered that this archive of knowledge was only related to
one-half of the practice of tenmon—the act of interpreting the meaning of astronomical
phenomena. There is no indication that maps or charts were also circulating among the
courtiers of the tenth or even eleventh centuries, almost as if Article 20 of the Penal
Codes Concerning Officials was still in effect for these texts. The archive of astronomical
knowledge that was being built up through the two leaks in the official-document chain
was thus primarily related to divination, and the knowledge therein was textual. This is
perhaps the most important factor in shaping the Japanese historical archive relating to
astronomical phenomena, which is almost exclusively textual through the end of the
fifteenth century.
24
Knowledge related to the observation and identification of
phenomena, the important first half of the tenmon practice, remained out of reach of most
courtiers. However, although members of the court might not have been able to acquire
such knowledge directly, particularly in the absence of training materials, there was a
method through which they could access to such expert knowledge through personal
connections, as shall be shown below.

Careers and Patronage

24
Aside from the text that has been identified as a copy of Sanjia buzan that was kept in the archive of a
lineage of tenmon specialists, there are two surviving illustrations of astronomical phenomena. The first,
originally dating to 1106 but recopied into a report of 1118, diagrams the different forms between two
types of “cometary” phenomena, (Ch. beixing, Jp. haisei) and (Ch., Jp. suisei), and thus
consists of images of ideal types, rather than being a visual record of an observed phenomenon. (Shodō
kanmon Eikyū 6/4/1.)
The other surviving diagram, a solar halo included in Fujiwara Teika’s Meigetsuki diary
(Kenpō 5/2/14), is by contrast a representation of an observed event. The edition of Meigetsuki used for this
study was the one-volume Kokusho meichō kankōkai, ed. Meigetsuki (Kokubun meichō kankōkai,1935)
edition, which does not include this drawing, which can be found in Kanda Shigeru, Nihon tenmon shiryō.

78
There was, in fact, one other, though non-institutional, way in which
tenmon-related knowledge circulated in textual form outside of the Bureau of Onmyō.
This was through patronage links that the tenmon experts, who were all of relatively low
rank, forged with top-ranking members of the court bureaucracy. Such relationships were
common from the mid-Heian Period, as economic necessity drove new ways through
which the top nobility interacted with lower-ranking officials in the bureaucracy, officials
who offered expert knowledge and staffing for noble households in return for career
advancement and access to economic opportunities.
Patronage was a major organizing principle during the medieval transition in
Japan. Under the ritsuryō codes of the eighth century, members of the Japanese state and
officially supported religious institutions were to have been economically supported
through stipends from the court, or through the direct assignment of “salary fields” (
Jp. shokuden). In the latter case, which constituted a more direct relationship between
individual (or institution) and agricultural producer, the entire legally extractable
(taxable) portion of the field’s production was to go directly to the grantee. All remained
at least theoretically reassignable by the state, however, and did not exist independently
of the state’s fisc.
Over the next centuries this “support field” system evolved into a mature estate
system ( Jp. shōen-sei), wherein the rights to the produce from a designated area
were divided among a hierarchy of individuals, all with a duty to the estate as a corporate
entity, organized under a high-ranking religious institution or noble proprietor. This
79
hierarchy, therefore, stretched from the individual producer on the estate to patrons in the
capital, forming a new set of links from center to periphery.
However, and unfortunately for the court fisc established in the ritsuryō codes,
many of these estates became increasingly exempt from normal state levies. As a result,
the Japanese court suffered a slowly unfolding economic crisis from the tenth century
onwards. Official stipends went unpaid. As a result of economic necessity, lower-ranking
officials began forging relationships, either temporary or more permanent, with more
economically secure and powerful nobles in return for patronage, which was paid in
official promotions, positions in high-ranking households, or other forms of material
support.
Ironically, this was also a time when provincial producers were increasingly
squeezed by governors sent by the court to administer the affairs of provincial
government.
25
By the year 1000, one way in which the higher-ranking nobles secured
their economic livelihood was through gifts from provincial governors, who paid them
from the “excess” materials they extracted from the provinces, in the hopes of securing
future posts and benefits through their patron. The post of a provincial governor ( Jp.
zuryō, which was more accurately the representative of the provincial government that

25
For more on this situation, see Charlotte von Verschuer, "Life of Commoners in the Provinces: The
Owari No Gebumi of 988 " in Heian Japan, Centers and Peripheries, ed. Mikhael et al. Adolphson
(Honolulu: University of Hawai'i Press, 2007).

80
was actually present in the province)
26
was therefore one of the most valuable “coins” of
the Japanese court patronage system from the late tenth through the twelfth centuries.
The services rendered by members of this lower-ranking group of officials,
who often topped out at the fifth or fourth courtly rank (the middle of the court-rank
spectrum) were not limited to the economic. The ability to provide services or
information was just as important. This can be seen through the relationship between Abe
Yoshimasa
27
and Fujiwara Sanesuke,
28
in the first half of the eleventh century.
Appointed Minister of the Left in 1021, Fujiwara Sanesuke had an influential,
although perhaps stunted, career at court. He was the grandson of a regent, and although
his political career was limited by the rise of Fujiwara Michinaga
29
and Michinaga’s
descendants, Sanesuke still made a name for himself as an expert in court protocol and
affairs, utilizing the archive of documents he inherited from his grandfather, to which he
added to throughout his life. Abe Yoshimasa was the son of the famous Abe Seimei, and
an active member of the Bureau of Onmyō, although he was not Seimei’s only heir. Not
only was his older brother Yoshihira
30
a rival—and in 1018 Yoshihira outranked

26
The zuryō might hold the rank of Provincial Governor ( Jp. kokushu), or might be the Assistant
Provincial Governor ( Jp. kuni no suke), nominally the second-in-command in the provincial
government. If the top post was held by someone who did not actually do a tour in the province itself, it
was often the Assistant Provincial Governor who was the zuryō.

27
, 955? – 1019.

28
, 957 – 1046.

29
, 966 – 1028.

30
, 954? – 1027.

81
Yoshimasa by a full court rank
31
—but also there was at least one other Instructor of
Tenmon from a different lineage to contend with.
32

Although 1019 would turn out to be the last year of Yoshimasa’s life, he can
still worked in that year to establish a patron-client relationship with Sanesuke. Upon
request, Yoshimasa provided Sanesuke with copies of divination reports on astronomical
phenomena.
33
Sanesuke was at the time merely a senior counselor ( Jp.
dainagon) on the Council of State—he held nothing near the authority of the regent who
had the right of nairan. Nor could Sanesuke ever expect to reach such a level. Sanesuke’s
receipt of these reports, then, should be seen as completely outside the official circulation
of documents at the court.
Yoshimasa failed to receive benefits as a result this favor to Sanesuke due as he
died suddenly only two months later.
34
Nevertheless, Sanesuke proceeded to make use of
the “illegally” procured documents he had obtained in this exchange.
35
Later that year,
two months after Yoshimasa’s death, a star was visible in the daytime sky near the

31
Yoshihira was fourth rank to Yoshimasa’s fifth. Shōyūki Kannin 2/10/16. The primary edition
consulted was Tōkyō Daigaku Shiryō hensanjō, ed. Shōyūki, 11 vols., Dai Nihon kokiroku (Tokyo:
Iwanami shoten,1959).

32
After Yoshimasa dies in 1019, there is mention of another Instructor of Tenmon, one Wake Hisakuni
( , dates unknown). The names of other non-Abe tenmon specialists can be found in the historical
record for the early eleventh century.

33
Shōyūki Kannin 3/2/5.

34
Shōyūki Kannin 3/4/28.

35
As these documents used the “apocrypha based on the Chinese classics” method of divining the
meaning of astronomical material, their ownership was not in fact illegal under article 20 of the Penal Code
Concerning Officials, the private circulation of material originally intended for the throne was rather
heavily frowned upon.

82
moon.
36
While Sanesuke and others held suspicions concerning the identity and this
phenomena, which was apparently striking enough to the untrained eye to make news
among the court population, and what it meant, no official report on the implications on
this event could then be made. Yoshimasa was newly dead, his post yet unfilled, and the
other Instructor of Tenmon, Wake Hisakuni, was in Iyo Province on the island of
Shikoku, far from the capital. As a result, all the other senior counselors on the Council
of State began circulating old tenmon reports in an attempt to decode the phenomena.
Thanks to the additional material Sanesuke had received from Yoshimasa, he, too, had
the ability to participate in this emergency measure—even if he ultimately decided not to
exercise it.
37
Such stopgap measures ended up not being necessary, since Abe Yoshihira
was granted official permission to present tenmon reports to the court five days later.
38

Yet that there was such an attempt to create an “amateur” tenmon memorial speaks much
about the knowledge which privately collected reports gave the non-technical nobility.
Even if Sanesuke did not explicitly exercise the ability to compose his own
divinations based on his archive of tenmon-related material, he did make use of this
material to check the products of other diviners. Sanesuke’s own diary, Shōyūki, shows
him to have been a veritable magpie regarding information, and he used it to keep a
gimlet eye upon the workings of a court that he often criticized. For example, in 1014,

36
Shōyūki
Kannin 3/6/4.

37
Sanesuke does not mention that he also sent around old reports, although he now had the ability to do
so. The rest of the entry decries the state of the court and expresses the fear that the state might be punished
for its inability to respond to the admonishments of Heaven.

38
Shōyūki Kannin 3/6/10.

83
Sanesuke referred to old tenmon reports to check on a newly composed report he had
received from the regent. Using his archive, he noted that predictions that had appeared in
previous reports on the same phenomenon—concerning fires, major earthquakes, and the
person of the regent—did not appear in the current one. He sent the report back with the
recommendation that it be amended with such information.
39
After Yoshimasa’s death in
1019, there had been apparently been a few attempts at amateur tenmon reports, as
Sanesuke also critiques one he saw as “exceedingly bad” and “made by an apparently
ignorant person.”
40

Sanesuke’s primary use of his archive to critique instead of to compose
divination reports not only presages Kujō Kanezane’s actions in 1189—it also reflects the
bifurcation of tenmon knowledge at court. In 1019, although Sanesuke describes the
daytime appearance of the moon, Venus, and the sun together in the sky in some detail,
including an estimate of the distances and relative locations of the moon and Venus, he
was not at all confident in identifying the “star” as Venus.
41
The distances Sanesuke
provides are another clue: generally, a distance of more than seven degrees in the sky,
such as Sanesuke indicates, would not have been considered a conjunction at all by
tenmon specialists.
42
Therefore when Yoshihira is quickly appointed to take the place of
his deceased younger brother Yoshimasa, it should not be understood as just a response

39
Shōyūki Chōwa 3/2/9.

40
and respectively. Shōyūki Chōwa 3/5/25.

41
Shōyūki Kannin 3/6/4.

42
See, for example, the distances given for conjunctions listed in Saitō Kuniji, Kokushi kokubun ni
arawareru hoshi no kiroku no kenshō (Tokyo: Yūzankaku shuppan, 1986).
84
to “poorly constructed” divination reports by courtiers with private astronomy archives.
The appointment was largely related to the practical ability that Yoshihira and Yoshimasa
both possessed, through their training both in the Bureau and from their famous father, to
visually identify and categorize astronomical phenomena.

Through patronage relationships, top-ranking members of the Japanese court
obtained access to documents that provided them with a certain amount of critical
knowledge about astronomical matters. Sanesuke’s activities related to the observation
and interpretation of astronomical phenomena in the early eleventh century, however,
show some of the limits of this source of knowledge and authority. These activities,
recorded in his diary, also show that Sanesuke, like other members of the court, was a
critical audience that received the pronouncements of tenmon specialists with
reservations. These courtiers had been building up archives for the express purpose of
critiquing and confirming the statements of divination found in these reports. In other
words, by at least the early eleventh century, there were powerful members of the court
who did not simply accept statements of astronomical and divinatory fact from the
experts they employed, even as they kept employing them
Limits to the personal and private study of tenmon, however, remained. All
surviving evidence indicates that astronomical maps and charts were still officially
proscribed materials and did not circulate at court.
43
In response to such limitations, and

43
As mentioned earlier (note 73 in the previous chapter), the one surviving set of charts was preserved in
the archive a lineage of tenmon experts from the thirteenth century. Thus its private preservation likely has
85
driven by the need for confirmation and checks upon the expert knowledge of the
Instructors of Tenmon, the court turned to expanding the number of tenmon specialists
available for consultation, tentatively in the tenth, but increasingly so in the eleventh
centuries. This constituted a shift from a system where knowledge was authoritative due
to the producer of the knowledge, to a less credulous system where one specialist’s work
could be confirmed or tested against another’s. Such a systematic shift would give rise to
a new problem, however, as officials of the court now had to determine which of many
reports was, in fact, the “correct” one.

The Rise of the Non-Bureau Tenmon Specialists and the Nakahara Lineage
As was shown by the legal support of the eighth century, and again in the early
eleventh century with the death of Abe Yoshimasa, the court had a consistent problem
ensuring an adequate staffing level of tenmon experts within the Bureau of Onmyō. This
problem was only acerbated by policies at court. By Yoshimasa’s day in the early
eleventh century, one response to the problem of ensuring economic support for these
specialists was to appoint these men to additional positions within the provincial
government, where they could make a living off of the land.
44
Joint appointments as
Instructor of Tenmon and a member of provincial government can be traced back as early

more to do with the shift to lineage-based training of tenmon specialists at the court than to relaxations in
the legal strictures surrounding such material.

44
The exploitative activities of members of the Bureau of Onmyō in the provincial government can be
seen in the Owari gebumi petition from the residents of Owari province. One of the members of the
governors’ staff who are charged with abuses is a Student of Tenmon ( Jp. tenmon sei).

86
as the mid-ninth century,
45
and an appointment in the provincial government of Iyo
likely explains Wake Hisakuni’s unavailability during the astronomical event of 1019.
It is therefore unsurprising that the strictures on who could compose and
officially submit a report interpreting astronomical phenomena began to loosen in the
tenth century. New manuals of court protocol, such as Saikyūki ( ) and Shin gishiki
( ), describe just such a loosening.
46
According to these two sources, by the late
tenth century anyone designated by royal order could now submit a tenmon memorial to
the court, although punishment was still prescribed for anyone who dared to compose and
submit such a memorial to the court without official approval. In contrast to outsiders,
Instructors of Tenmon did not need to receive a royal decree in order to submit a
memorial. By the middle of the tenth century, they were no longer required to submit all
memorials to the court through the Head of the Bureau of Onmyō, as had been the
previous procedure.
47
However, this did not limit the production of tenmon memorials
only to the Instructors of Tenmon within the Bureau—Bureau Heads and Assistant Heads

45
Nakatomi Shihi Harutsugu ( , dates unknown) was appointed supernumerary assistant
governor of Settsu in 865, while retaining his position as Instructor of Tenmon. Sandai jitsuroku Jōgan
7/1/27.

46
Saikyūki and Shin gishiki are both early examples of gishikisho ( ), manuals of court procedure,
which began to appear at court in the tenth century. This was likely in response to changes in court
protocols that occurred since the establishment of the ritsuryō state and when the court last collected and
promulgated a set of supplementary law codes ( Jp. kyaku or Jp. shiki). As changes in court
procedure still needed to be recorded for later consultation, the genre of gishikisho appeared. Tokoro Isao,
Kyūtei gishikisho seiritsu shi no saikentō (Tokyo: Kokusho kankōkai, 2001).
This need for documentation has also been credited with driving the rise to prominence of
courtier diaries as a major aspect of court life and business. See the first chapter of Matsuzono Hitoshi,
Ōchō nikki ron (Tokyo: Hōsei daigaku shuppankyoku, 2006). for a summary of major scholarly theories
about the reason for the rise of courtier diaries through that time.

47
As shown by the description of the process described in Chikanobu-kyō ki.

87
of the Bureau who had previously been Instructors of Tenmon continued to submit
memorials.
48

Initially this system for appointing outside experts was probably intended only
as a stop-gap measure—just as the amateur production that was attempted in 1019 when
circumstances made obtaining a memorial from a full Instructor of Tenmon difficult, but
that would not become standard procedure at court. The designation of outside experts
became standard sometime in the eleventh century, and the end result was a situation
where two (or more, when both the position of Instructor of Tenmon and that of the
Supernumerary Instructor were filled) officially recognized specialists were working at
the same time. Additionally, other members of the Bureau of Onmyō with tenmon
experience could also be called upon when needed for the production of official
interpretations of astronomical phenomena. Therefore, at the same time that the
population of individuals at court with access to tenmon-related memorials increased, the
number of specialists who worked in tenmon was also growing. This population growth
led ultimately to competition.
As in the case of increased access to tenmon-related materials, the earliest
evidence for competition between tenmon specialists can be found in Taira Chikanobu’s
diary, Chikanobu-kyō ki. In one of Chikanobu’s entries, Abe Seimei and Tōichi

48
A prime example of one such individual being Abe Yasuchika, whose contribution to memorials while
Assistant Head of the Bureau, was discussed above. In Yasuchika’s case, he was more often co-author than
primary author of the memorials in the collection bearing his name. The actual division of labor between
Yasuchika and the other author remains unclear.

88
Mochinobu are recorded as having forwarded reports to the throne on consecutive days,
49

although other sources imply that the two Instructors of Tenmon had divided up some of
the responsibilities of observation and reporting, and switched off regularly.
50

Clearer evidence of debate can be found in the documents preserved in the
twelfth-century collection Shodō kanmon ( , “Reports from the Various Fields”),
compiled by a member of the Nakahara lineage that descended from Tōichi Mochinobu.
51

While it is thought that this work was once more comprehensive, covering official reports
and memorials on a number of subjects, only two chapters survive to the present day: a
fragment of a debate concerning the meaning of omens at Iwashimizu Hachiman Shrine,
south of Kyoto; and a collection of documents from debates about the identification of
comets and comet-like phenomena.
The documents collected sometimes reflect both sides of a dispute, and
sometimes only one side. In most of the debates members of Abe Seimei’s
lineage—which was beginning to develop a hereditary association with the Bureau of
Onmyō, mathematical divination, and propitiatory rituals—appear arguing against
members of the Nakahara lineage descended from Mochinobu. Although Mochinobu had

49
Chikanobu-kyō ki Tenroku 3/12/6.

50
For example, only some of the known astronomical phenomena for 1182 appear in the Yōwa ni-nen ki,
a diary kept by a junior member of the Abe lineage who had probably been assigned part of the work of
observing. See the discussion and critical edition of this source in Yamashita Katsuaki, Heian jidai no
shūkyō bunka to Onmyōdō.

51
It is traditionally attributed to Nakahara Moroyasu ( ). It is clearly the product of someone
who had access to the Nakahara archive, as many of the Nakahara-authored documents appear in draft (
Jp. an) form, and were recycled for multiple submissions to court.

89
been recognized as a member of the Bureau,
52
the Nakahara family had become more
closely associated with the Secretariat of the Council of State ( Jp. Geki); they also
became known as experts on Japanese precedent and legal protocols.
53
Despite the fact
that one later member of the Nakahara lineage is known to have also been a student of
tenmon within the Bureau of Onmyō,
54
the first Tōichi (whose descendants would later
become the Nakahara) who received permission to submit a memorial interpreting an
astronomical event, Tōichi Ryōsuke,
55
does not appear to have had any experience in the
Bureau before receiving his royal mandate.
56
Therefore it seems likely that the
individuals who were selected by decree to participate in the production of official
tenmon memorials may have been, like the monks laicized in the eighth century, recruited
as a direct result of the court’s desire to employ skilled individuals from outside of the
formal court bureaucracy and system of training.
57
From the tenth century through the

52
Nakahara family records are unclear as to whether he ever held a position in the Bureau, or if the
Instructor of Tenmon title was more descriptive of his function than his role.

53
Beginning in the late Heian Period, the Nakahara also became notable as authors of texts on court
protocol and precedent. See, for example, the Morotō nenjū gyōji ( , “Nakahara Morotō’s
Calendar of Annual Court Performance”), which was compiled by the Nakahara Morotō II ( ) who
will appear in the late eleventh and early twelfth century examples below.

54
Nakahara Morotō I ( , d. 1062).

55
, dates unknown

56
In 930. Ruijū fusenshō 9.1212. Keizai zasshisha, ed. Ruijū sandaikyaku. Ruijū fusenshō. Zoku sajōshō,
Kokushi taikei (Tokyo: Keizai zasshisha,1900).

57
It is worth noting that the memorials included in Shodō kanmon that predate the career of Nakahara
Morotō I, who did have a stint as a student officially in the Bureau, do not precisely match the formatting
conventions found in tenmon-related materials composed by Abe or Kamo individuals found in other
sources. After Morotō I’s career, however, Nakahara memorials do not significantly differ in format from
those produced by members of the Abe lineage.

90
twelfth, individuals from this Nakahara lineage often appear in the historical sources as
tenmon specialists.
Of all of the lineages at court, excluding those with a historical association with
the Bureau of Onmyō, the Nakahara may have been the best placed to perform the duties
of astronomical observation and interpretation. While initially they may not have had
access to the maps and diagrams held within the Bureau of Onmyō before the tenth
century, when one of their members briefly became a student in the Bureau, as a lineage
they obtained access to written descriptions of the night sky and charts. As a scholarly
lineage, they had already been familiar with at least three-fifths of the texts mandated by
the decree of 757 for the training of new specialists in tenmon: the Records of the Grand
Historian, the Han History, and the Jin History were all works studied by scholars who
needed to be familiar with Chinese precedents—particularly those lineages that
specialized in legal precedents. The “Treatises on the Heavens” found in each of these
works include written descriptions of the constellations—but these texts were lacking any
diagrams or charts. In other words, as referenced earlier, despite the restrictions placed on
the circulation and private possession and study of charts and maps of the stars, there
were ways by which dedicated individuals with access to alternative sources of
authoritative information could learn what they needed for astronomical divination.
The Nakahara also had an advantage as they began to participate in the official
practice of tenmon. The two texts that played the most significant role in the composition
of memorials divining the implications of astronomical events, the Tianwen yaolu and the
Tiandi ruixiang zhi, had never been restricted in their circulation among court officials.
91
This, despite their apparent resemblance to prescribed texts listed in Article 20 of the
Penal Codes Concerning Officials. Indeed, the first appearance of a citation from Tiandi
ruixiang zhi appears not in a cribbed copy of a memorial on an astronomical event, but
instead in a notation written on a legal decision.
58
Despite the Nakahara’s relative
position as newcomers to the practice of tenmon, in terms of background knowledge, they
do not seem to have been at much of a disadvantage at all.
Nevertheless, the Nakahara still overcame some significant handicaps as they
rose to prominence as secondary tenmon specialists, working alongside Instructors of
Tenmon who had been trained within the Bureau. Some disadvantage, for example, can
be assumed to have faced a newcomer forced to jump from the description of
astronomical phenomena as written in texts to identifying such phenomena in the night
sky itself. Such problems are not explicitly discussed in the sources, but an echo can be
found in Fujiwara Sanesuke’s uncertainty over whether the star he saw during the day in
1019 was, in fact, Venus. Moving from even detailed written descriptions to visual
recognition of phenomena was a non-trivial step
Additional difficulties were posed by the fact that the source texts, the “Book
on Heavenly Officials,” and the two “Treatises on the Heavens,” found in the official
Chinese histories, were at times unclear and self-contradictory in their descriptions of
astronomical phenomena. This was particularly true for cometary phenomena, which
likely explains the survival of the chapter on cometary debates from the Shodō kanmon
collection, as well as the prominence of debates over comets and comet-like phenomena

58
Referenced in Nakamura Shōhachi, Nihon Onmyōdō sho no kenkyū: Zōhōban.

92
in the historical record. In the “Treatise on the Heavens” in the Jin History, for example,
the text describes a comet-like phenomena known as Chi You’s Banner as “a type of
comet, with a curving tail like the image of a banner,”
59
as “a single line of a red
cloud,”
60
and as “yellow above and white below,”
61
all in quick succession. How to
reconcile these descriptions, however, is left unstated.
Given the variation presented in descriptions of defining characteristics for
phenomena, then, it is little surprise that in the chronological section of the “Treatise on
the Heavens” there are cases where a single phenomena is referred to variously as a
“white aurora,” “comet,” “Chi You’s Banner,” and “long star.”
62
The potential for such
confusion only added to the amount of variation possible in the classification of
astronomical phenomena that was the first step of tenmon. It is little surprising, in light of
such murkiness, that in 1195 Abe Suehiro would have attempted to introduce a new
category for analysis (no matter how suspicious the regent Kujō Kanezane might have
seen his efforts).
63
Presented with a number of equally authoritative and contradictory
definitions for phenomena from officially designated texts, the only possible consistency
in identifying phenomena would have been found in the training of new tenmon
personnel by experienced ones. Such training had been initially established by the

59
Jin shu 12.

60
ibid.

61
ibid.

62
, , , and , respectively. Jin shu 13.

63
The details of the incident and debate between Suehiro and Kanezane are discussed in the beginning of
the preceding chapter.
93
ritsuryō codes as the primary duty of the Instructor of Tenmon in the Bureau of Onmyō.
Later, as tenmon within the Bureau became the specialty of specific lineages, training
moved from the Bureau itself to the education and training of children within these
lineage households. In the correct identification of astronomical phenomena, outsiders
would have been at a disadvantage.
The Nakahara (originally Tōichi) lineage overcame this disadvantage when
Nakahara Morotō I, in the first half of the tenth century, joined the Bureau as a Special
Student of Astronomy.
64
Morotō was then able to take the training he received from the
Bureau and pass it along to his own sons. Once Morotō I established a program of
training in tenmon within the Nakahara lineage, the lineage regularly produced
individuals with the experience-based knowledge necessary to observe and correctly
identify astronomical phenomena, and a second consistent source of specialists on
tenmon became available to the court.
65
This increased population of experts had a
previously unidentified impact on the practice of astronomical observation and divination
at the Japanese court, as competition between a number of experts drove attempts at
lineage differentiation and innovation—such as Abe Suehiro’s comet neologism of 1189.

64
Jp. tenmon tokugyōsei.

65
At the moment, it does not seem as if this new lineage was particularly long-lived. The last Nakahara
yet identified that participated in tenmon debates was Morotō I’s great-grandson, Moroyasu, in the twelfth
century. Why the Nakahara lineage of tenmon (apparently) died out after four generations is a topic I intend
to pursue in the future.

94
The Effects of a Marketplace for Astronomical Interpretation
The availability of multiple specialists in tenmon, after all, could work against
any individual practitioner. This can be seen in the case of a hailstorm that occurred
during the summer of 1096.
66

At that time, Instructor of Tenmon Abe Chikamune noted in his memorial that
such summer hail should be considered a “normal anomaly” ( Jp. tsune no hen).
This conclusion, however, did not fit with the immediate reaction of the nobility to the
storm when it occurred. In his diary entry describing the hailstorm, royal intimate ( Jp. denjōbito) Fujiwara Munetada described the event as an ominous occurrence,
even before Chikamune had a chance to present an official interpretation through a
memorial. Munetada was hardly the only member of court with a pre-established
understanding of the negative meaning of this storm: even before the storm had ended,
people had rushed to the palace in order to protect the ruler.
67
The ruler at the time of the
hailstorm of 1096 was Horikawa Tennō,
68
who was much beloved by a circle of courtiers
that included Munetada, but a man plagued by illness throughout his reign. Against such
a political background it is not surprising that Chikamune’s memorial was rejected.

66
“Sub-lunar” events such as earthquakes or meteorological phenomena were in China and Japan, as it
was initially in medieval and Renaissance Europe, considered part of the study of the heavens.

67
Chūyūki Eichō 1/5/9.

68
.

95
Instead, Nakahara Morotō I’s grandson, Nakahara Morotō II,
69
was called
upon to present another memorial interpreting the event.
70
Most likely, the individuals
who had ordered this decision had hoped that this new memorial would not contradict the
intuition of top members of the court elite. While the exact contents of this memorial
were not preserved in any of the diaries kept by top courtiers of the time, it seems likely
that Morotō II satisfied his audience. When the Morotō II’s memorial was combined with
an additional method of divination,
71
the implications of the storm were clear and
ominous. Propitiatory offerings to six shrines were performed the following month.
72

The fact that the audience for divinatory interpretations of astronomical
phenomena could call in second or third opinions when the result did not please them
could not but have had a significant effect on divinatory practice. Yet it is difficult to
think that Abe Chikamune, an experienced member of the official tenmon system in place
at the court, could not have read the mood of the court and adjusted his memorial
accordingly. Chikamune’s adherence to a position that would clearly not in accord with
the prevailing interpretation of the event at court can be seen, therefore as principled.

69
.

70
Go-nijō Moromichi ki, Eichō 1/5/18. The edition consulted was Tōkyō Daigaku Shiryō hensanjō, ed.
Go-Nijō Moromichi ki, Dai Nihon kokiroku (Tokyo: Iwanami shoten,1956–1958).

71
Performed on the same day that Morotō II’s memorial was presented—it is likely the court was not
about to leave anything to chance. Ibid.
The additional divination used to back up Morotō II’s interpretation was a combination of
mathematical calculation based on the sexagenary value of the date and time of the event, and a
interpretation of heat-applied cracks on a tortoise shell (plastronmancy). This combination, the konrō no
miura ( ), was considered one of the most authoritative and last-resort methods of determining the
implications of any phenomena or decision. A general description of the procedure, and its history, can be
found in Nishioka Yoshifumi, "Rikujin shikisen to konrō no miura," in Ōken to jingi, ed. Imatani Akira
(Kyoto: Shibunkaku shuppan, 2002).

72
Eichō 1/6/26.
96
Irrespective of Chikamune’s integrity, however, an important institutional fact
was also likely a major contributor to his position: the presence of an archive of already
produced and circulating memorials interpreting such phenomena. Just as Fujiwara
Sanesuke had critiqued a work of divination in 1014 for not matching with previous
examples, and as Kujō Kanezane would critique the work of Chikamune’s descendant
Suehiro in 1189, Chikamune in 1096 operated under the implicit control of an archive
and history of previous divinatory interpretations of phenomena. This control only
afforded limited space for successful innovation in interpreting phenomena. In the right
circumstances, however, room for innovation or modification could be found, as
Nakahara Morotō II no doubt did in his memorial on the same hailstorm. An individual
among a group of tenmon specialists would be well rewarded for finding avenues for
successful differentiation within institutional limits. His audience, too, would be well
satisfied.

Supernatural Skills as Proof of Expertise
The textual constraints shaping the memorials on, and the interpretations of the
meaning of, astronomical phenomena were strong. Even so, there was a method through
which individual and groups of tenmon specialists could successfully set themselves apart
from others. While the archive of tenmon reports held by members of the aristocracy
shaped the presentation of observations and the interpretation of the same, the actual act
of observing and identifying astronomical phenomena was not defined by this textual
tradition. In the absence of authoritative charts or diagrams circulating among the upper
97
echelons of the nobility, the correct identification of phenomena remained within the
control of the tenmon specialist. The only way for the lay audience to check such
statements of identification, absent personal expertise in the matter, was to let individual
specialists debate each other.
In such instances a number of strategies were employed by the tenmon
specialist for the purpose of convincing his audience. One was establishing a reputation
for accuracy of prediction, as seen in Abe Yasuchika in his Yasuchika-ason ki
compilation.
73
As the correct interpretation of the divinatory meaning of any particular
astronomical phenomena depended upon its correct identification—conversely, an
accurate prediction derived from the phenomena would indicate that the initial
identification had been correct. Such a confirmation could only work in retrospect, but
once an individual such as Yasuchika had developed a reputation for accuracy, he could
employ it in future disputes.
The strategy of developing a reputation for accuracy in divination would be
employed with a new twist by one of Abe Yasuchika’s sons, Yasushige, in 1185.
74
His
audience at the time was again Kujō Kanezane, the same skeptical audience his older
brother Suehiro would face four years later. A debate had erupted among tenmon
specialists over how to classify a phenomenon that had appeared in the east on the first

73
Discussed in an earlier part of this chapter.

74
, dates unclear.

98
day of the year.
75
Although the initial popular opinion classified it as an instance of Chi
You’s Banner, all of Abe Yasuchika’s sons disagreed. As the representative of the family,
Abe Yasushige went to Kujō Kanezane to make his case.
76

Kanezane does not seem to have been initially well disposed to the
identification of the phenomenon as a comet—he had himself noted down the argument
for the phenomena being an instance of Chi You’s Banner in his court diary a few days
previously.
77
Kanezane himself was also known as a serious classical scholar, who had
co-founded a group of like-minded young courtiers dedicated to reforming the court and
reviving proper practice.
78
He would not have been an easy audience to convince.
Yet trying to convince Kanezane would have been worth the risk. Although at
the time of the phenomenon of 1185 Kanezane was not yet regent, he was surely
identified as a future power broker at court, even potential regent. That Yasushige made a
specific trip to this courtier to argue his case was likely a strategic move to cultivate a
sympathetic audience for his future career as a tenmon specialist. The ensuing
conversation, as recorded in Kanezane’s Gyokuyō diary, shows a rhetorically complicated
strategy wherein Yasushige attempted to sway Kanezane into accepting the identification
of the phenomenon not as Chi You’s Banner, but as a comet.

75
Its first appearance can be found in Kitsuki Bunji 1/1/1, classified as a Chi You’s Banner.
Zōhō shiryō taisei kankōkai, ed. Kitsuki, 2 vols., Zōhō shiryō taisei (Kyoto: Rinsen shoten,1965).

76
Gyokuyō Bunji 1/1/12.

77
Gyokuyō Bunji 1/1/5.

78
Gomi Fumihiko, Yakudō suru Chūsei, vol. 5, Nihon no rekishi (Tokyo: Shogakkan, 2008).
99
Yasushige opened his the argument for his position with quotations from
classical texts. Instead of using the Records of the Grand Historian or the Han or Jin
Histories, however, he cited a definition of comets found in the Treatise on the Heavens
in the [Liu] Song History ( Ch. Song shu, Jp. Sōsho). Using this definition, a comet
need not have a clear “originating star” ( Jp. honsei), he argued.
Yasushige then proceeded to reveal a bit of family lore about comets. He
reminded Kanezane that a similar debate had erupted regarding an astronomical
phenomenon in 1178, a debate that Kanezane had himself observed and recorded in his
diary.
79
Kanezane had heard at the time that Yasushige’s father Yasuchika, and
Yasushige’s older brother Suehiro, had taken different positions about the phenomenon.
80

However, while Kanezane had recorded that incident as a dispute over whether the
phenomenon was a comet or not,
81
Yasushige informed Kanezane that the point of
contention was on whether the phenomenon was a comet or Chi You’s Banner.
Yasuchika had taken the comet line, while Suehiro insisted that it was a case of Chi
You’s Banner.
Yasushige told Kanezane that to settle the dispute, both Yasuchika and Suehiro
had sworn oaths before Heaven, asking that the one who held the incorrect opinion be
struck down. Then they waited for the cosmos to settle the issue. Suehiro fell rapidly and

79
Gyokuyō Jishō 2/1/18.

80
That a dispute arose between Yasuchika and his son Suehiro is not surprising from the standpoint of
other historical sources. The Yōwa ni-nen ki (1182), a diary kept by Yasushige’s son Yasutada ( ),
records a major row that took place between the Suehiro and Yasuchika previously.

81
The non-comet position was merely recorded as Gyokuyō Jishō 2/1/18.
100
mortally ill, and his father Yasuchika was forced to perform a ritual to save his son’s life.
Fortunately (and further proving that Heaven favored Yasuchika), he succeeded. The
result of this experience was that Yasuchika, and later his sons, had learned to correctly
identify this particular type of phenomena. It was this knowledge, Yasushige informed
Kanezane, that Yasushige was using in 1185.
As authoritative as the intervention of Heaven itself might seem, Kanezane
continued to object. He stood firm on the position that if an “originating star” could not
be determined, by observation or other means, then the phenomenon should be
considered an appearance of Chi You’s Banner and not a comet. Kanezane was likely
basing his position on his knowledge of descriptions of Chi You’s Banner as aurora from
the Chinese classics. Undaunted, Yasushige went on to declare that Chi You’s Banner
and comets were actually indistinguishable to the untrained eye—without his family’s
secret knowledge, obtained by the supernaturally accurate Abe Yasuchika and passed
down to his sons, any attempt at distinguishing the two would fail, no matter what the
textual criteria.
At this point in the record, Kanezane recorded his dissatisfaction with
Yasushige’s argument. Kujō Kanezane was, after all, a man schooled in the Chinese
classics and dedicated to their correct reading. In his 1185 entry, Kanezane also cites as
support for his position a debate that had occurred between Abe Norichika
82
and
Nakahara Morohira
83
in 1056, over whether a comet-like phenomenon with no

82
, dates unclear.

83
, dates unclear.
101
“originating star” was a “guest aurora” or a comet.
84
Kanezane concluded by doubting
that tenmon specialists could be relied upon at all for particular expertise. “Concerning
tenmon specialists,” he wrote, “These practitioners even have trouble with known things.
Can it be that [such] people are not properly trained?”
85

Yet something brought Kanezane around to Yasushige’s point of view. The
1185 precedent—that this ambiguous phenomenon was, in fact, a comet—is the precise
evidence that Kanezane would use in 1189 against Suehiro’s neologism. In 1189,
however, Suehiro was making the same argument against Kanezane that Kanezane had
made against Yasushige in 1185: namely, that without a distinguishable “originating star,”
the proper identification of the phenomena was as not a comet.
What appears to have swayed Kanezane’s opinion over the intervening years
was Yasushige’s reputation for ritual skill and extrasensory knowledge. Yasushige can be
seen cultivating this reputation as a deliberate strategy in the 1185 debate. Immediately
following Yasushige’s attempt to convince Kanezane that the 1185 phenomenon was, in
fact, a comet, Yasushige moves on to what appears to be another topic of conversation,
which also served as his rhetorical gambit. He mentions that he was recently called upon
to perform a ritual
86
aimed at healing a senior prelate named Kōken. Yasushige
performed the ritual a number of times, and before each performance, Yasushige received

84
The particular debate can be found in the Shodō kanmon collection.

85
Gyokuyō Bunji 1/1/12.

86
The ritual in question was the Taizan fukunsai ( ), an Onmyōdō ritual directed at the deity
of Mt. Tai, the Eastern Sacred Peak in China.

102
a copy of the ceremonial text in a dream to use in the ritual. He showed a copy of the text
he had received in the dream to Kanezane, who remarked that it was indeed a rare thing
“in this degenerate age.”
87
At the end of the entry, Kanezane concludes with the
comment that Yasushige was evidently a man with supernatural ability.
In the years between 1185 and 1189, despite his initial reluctance, Kanezane
came to adopt Yasushige’s identification of the 1185 phenomenon as a comet as his own.
Yasushige’s careful burnishing of his reputation, through even the retelling of prophetic
dreams, played a clear role in the Kanezane’s final acceptance of Yasushige’s opinion as
evident fact. Kanezane’s ultimate acceptance of Yasushige’s argument would prove to be
an obstacle to Yasushige’s older brother Suehiro’s attempt at a neologism in 1189.
Even though Suehiro was accused of attempting to introduce “unprecedented”
terminology into the practice of tenmon in 1189, Yasushige, along with his father
Yasuchika,
88
also had been previously modifying and innovating the practice of
identifying astronomical phenomenon, although in a less obvious way textually. By
restricting their modifications to the identification process in of the practice of tenmon,
Yasushige, and other unknown tenmon practitioners like him, would have been able to
make significant changes to the practice of tenmon without stumbling across the barrier

87
, referring to the ages since the time of the Sage Emperors in China, when human behavior and
the natural world had degenerated far past its original, ideal state. Mori Shinnosuke, "Sekkan Inseiki kizoku
shakai ni okeru matsudaikan: Saii shisō ya unmeiron to no kanren kara."

88
At least the posthumous support of his examples. By 1185, Abe Yasuchika was conveniently dead and
thus available for any of his sons to be used as support without their running the risk that Yasuchika might
not have, actually, agreed.

103
presented by the nobility’s archive of past tenmon reports—the same archive that would
trip up Suehiro in 1189.
Despite the success of 1185, there were limits to these attempts. Such
innovations could only find a hospitable home with regards to a specific subset of
astronomical phenomena in the night sky—that is, phenomena that were already difficult
to distinguish from similar ones. The distinction between a “guest star” and a “comet”
depended on the presence of a tail visible to the naked eye; and the distinction between a
“comet” and “Chi You’s Banner” depended on even less precise criteria, which
furthermore shifted depending on the reference text used. Modern reconstructions of
records of astronomical phenomena that appear in the Japanese historical record show
how generally reliable records of planetary position and conjunction are.
89
The debates
about the identity of comets and comet-like phenomena provide clear examples of the
way that different forms of authority and ways of eliciting belief interacted to create
“facts” about something that all members of the court agreed had appeared in the sky for
those cases where astronomical reality was less of a constraint. The practice of tenmon
was clearly splitting into two parts: identification, over which the experts had some
control, and interpretation, over which they held little.

Splitting Tenmon, Multiple Authorities
The historical evidence for the practice of tenmon from the eighth through the
twelfth centuries shows that what was originally and ideally a unified three-stage

89
See Saitō Kuniji, Kokushi kokubun ni arawareru hoshi no kiroku no kenshō., throughout.
104
process—observation, identification, and interpretation of the night sky—had begun to
split by the end of the twelfth century. The rise of a textual archive gave courtiers who
otherwise would have not been trained in astronomical observation or interpretation the
models and experience that allowed them to critique the performance of specialists
working with what might have otherwise been an unassailable and obscure art. However,
the lack of experience in the bodily practices of observation, as well as the lack of access
to star-maps or charts, created a wall separating the observation and identification parts of
tenmon practice. As a result, surviving critiques of astronomical memorials are limited to
the text of the memorials, and not to the eyes or intelligence of the individual tenmon
specialist. When the classification of a phenomenon was questioned, these critiques
explicitly rely upon observations or identifications made by other tenmon specialists.
In fact, the very method of determining the divinatory meaning of astronomical
phenomena through tenmon—that is, through the recognized practice of citing from
recognized Chinese textual authorities—came to be challenged during the same period.
Buddhist astrologers working in the system of Sukuyōdō ( )
90
added a new
element to the meaning of events in the sky for at least some courtiers—the relationship
of the constellation that the event occurred in with regard to an individual’s date and time

90
This system of Sino-Indian horoscopy, with Ptolemaic elements, is discussed in greater detail in
chapter four. For a general introduction, see Yano Michio, Mikkyō senseijutsu: Sukuyōdō to Indo
senseijutsu (Tokyo: Tōkyō bijutsu, 1986). and the collection of articles on the subject in Momo Hiroyuki,
Rekihō no kenkyū, 2 vols. (Tokyo: Shibunkaku shuppan, 1990).

105
of birth changed the meaning, according to these monks.
91
Buddhist monks had already
made themselves available for interpreting other events in an individual’s life, so
consulting a monk to determine the personal implications of an astronomical
phenomenon was no far stretch. As this style of interpretation tended to be very personal,
involved with the very birth of an individual, it could work in parallel with tenmon-style
divination, which relied on Chinese histories and apocryphal texts for determining the
fates of entire populations and courts.
The style of interpretation in tenmon—the collection and citation of previous
occurrences of the phenomena in ancient Chinese history as a guide for a new
occurrence’s implications for the future—also opened up a variant style of interpretation.
This style became a hallmark of the Nakahara lineage, which was already deeply
involved in the collection and deployment of precedents from the Japanese court’s own
history. The clearest and earliest example of this mode dates from 1106, when a dramatic
comet appeared in the night sky during Horikawa Tennō’s final illness. In addition to the
usual tenmon memorials Nakahara Morotō II also submitted a list of every comet that had
appeared in the Japanese historical record, associated with a historical event that he
attributed to each comet.
92
The appeal of this variant method of divination can be seen in
how Morotō’s memorial was re-used, submitted again a few years later. Morotō II’s son

91
Mitsuhashi Tadashi, Heian jidai no shūkyō to shinkō girei (Tokyo: Zoku gunsho ruijū kanseikai, 2000)..
Mitsuhashi argues, however, in his study that this displaced tenmon-style divination, which is clearly not
the case from historical sources. Both systems continued in tandem.

92
Shodō kanmon.

106
Moroyasu employed the same technique when consulting with other members of court.
93

Although divergent from classical tenmon practice, this new method logically grew from
it, and was practiced in a wide geographical region by the thirteenth century.
94

Meanwhile, however, the textual form of mainline tenmon divination practiced at court
remained constrained by the forces of the archive, creating the situation where a
divination about a comet produced in the eleventh century, and one produced in the
seventeenth century, essentially resemble each other in almost every respect.
95

Conclusion
Demand for accurate information on the meaning of phenomena that appeared
in the sky drove both the textual conservatism that hampered would-be innovators like
Abe Suehiro in 1189, as well as the development of new forms of astronomical divination
which took their inspiration from court tenmon practices while drawing upon Japanese
texts. It is clear from these developments, and the general courtly attention to heavenly
phenomena, that observational astrology was a major concern at court and such
divination highly valued. Yet this divination did not, and could not, have the constraining
influence on “medieval Japanese thought” of which it has been accused by previous

93
For example, with Fujiwara Munetada in Chūyūki.

94
This style of interpretation was used multiple times in the Kamakura chronicle Azuma kagami.

95
This can be seen by comparing surviving and fragmentary tenmon no missō reports on comets. Many
are collected in Kanda Shigeru, Nihon tenmon shiryō. and Osaki Shōji, Kinsei nihon tenmon shiryō (Tokyo:
Hara shobō, 1994).

107
scholars.
96
There were simply too many alternative methods of interpreting the meaning
of phenomena, even alternatives to be found among the divinations made by tenmon
specialists, for this to be so.
While the divinatory facts of astronomical phenomena multiplied in an
intellectual world where variation became the norm, the authority of the tenmon specialist
in the identification of phenomena remained strong through at least the thirteenth century.
In 1226, the poet and low-ranking but ambitious courtier Fujiwara Teika
97
heard of a
three-planet conjunction. He followed up upon this news by consulting Abe Yasutoshi,
98

who had inherited the training and position of Instructor of Tenmon from the Abe tenmon
specialists featured in this chapter. Yasutoshi informed Teika that while there had indeed
been two conjunctions, between Mars and Saturn, and between Mars and Jupiter, that
Saturn and Jupiter were too distant from each other to count as a “three-planet
conjunction.” Chastened, even though he was no slouch in the observation of the sky
himself,
99
Teika recorded this lesson in his diary.
100

In this one respect, then, the tenmon specialist retained part of his authority.
This authority was tied to the least textual and most experiential part of the practice of

96
See the analysis of how divination—particularly astrology—is portrayed in the work of Morris and
Sansom in the previous chapter.

97
, 1162 – 1241.

98
, dates unclear.

99
This was noted by astronomer Saitō Kuniji the introduction of his analysis of Teika’s court diary,
Meigetsuki. Saitō Kuniji, Teika "Meigetsuki" no tenmon kiroku: Kotenmongaku ni yoru kaishaku (Tokyo:
Keiyūsha, 1990).

100
Meigetsuki Karoku 2/11/12.
108
tenmon, the expertise that the specialist had acquired and refined through his training and
continued practice: the recognition of his ability to correctly identify and classify
astronomical phenomena. This least textual aspect of tenmon practice in Japan also, as
shown by the debates about comet-identification, is the site where at least some glimpses
of a tenmon practice that was not precisely Chinese tianwen, but rather a Japanese-style
of observation and classification can be found. Accurately tracing the history of
non-textual knowledge from a historical corpus that is primarily textual, however, is
difficult. This topic still requires a great deal of further analysis and study.
Finding evidence for a Japanese-style of Chinese astronomical practice is easier
when the less-embodied practice of astronomical calculation is considered. The
development of a distinctively Japanese tradition for the predictive calculation of
astronomical and cosmological phenomena based on Chinese models will be the subject
of the following chapters.

109
CHAPTER THREE

“MANY SECRET TEXTS:” THE ORIGINS OF LINEAGE TRADITIONS AND
VARIATION IN JAPANESE CALENDRICAL ASTRONOMY ( )

It has long been assumed that the Japanese practice of Chinese calendrical astronomy was
essentially static from the adoption of the Xuanming li ( , Jp. Senmyō-reki) system
of calendrical astronomy
1
in 864 until its abandonment in 1682.
2
Yet, when the dramatic
social changes that occurred in Japan and the transformation of the structure of the
Japanese court are considered, such an assumption becomes problematic. In fact, from
the ninth through twelfth centuries, a number of changes were made to the Japanese
implementation of the Xuanming li system of calendrical astronomy, amounting to what
might be called a “Japanese-style” of Chinese calendrical astronomy.
The changes involved in the Japanese tradition of using the Xuanming li that
discussed in this section of the dissertation were not modifications to the text of the
Xuanming li canon itself.
3
Instead, they were reinterpretations of certain rules for
rounding or other adjustments, or the re-introduction of cycles that the author of the

1
An official system of calendrical astronomy created for and adopted by the Tang Dynasty, first
implemented in China in 822.

2
This assumption can be found in Nakayama, A History of Japanese Astronomy: Chinese Background
and Western Impact, and Sugimoto Masayoshi and David L. Swain, Science and Culture in Traditional
Japan, A.D. 600-1854, and it underlies the analysis of the Xuanming li’s predictive accuracy in John Steele,
Observations and Predictions of Eclipses by Early Astronomers, Archimedes (Boston: Kluwer Academic
Publishers, 2000). There, Steele uses Japanese eclipse predictions as if they were pure examples of rote
calculation based on the Xuanming li—an assumption that this dissertation will show is unfounded.

3
Given the poor survival of only late complete manuscripts, the possibility of such changes cannot be
discounted

110
Xuanming li had abandoned. The changes amounted to a tradition of post-calculation
modifications to the calculated astronomical calendar, as it were. These changes to the
implementation of the Xuanming li did not appear ex novo, but were the result of public
debates at the Japanese court—debates that belie the image of the Japanese court as
uninterested or unconcerned with technical knowledge or natural phenomena.
4

This chapter introduces a twelfth-century debate as evidence of the court’s
prolonged involvement in the production of a “correct” or “accurate” calendar, even as it
shows that the modern concept of what an accurate astronomical calendar would be was
not the same ideal that motivated the Japanese court, even as some common ground can
be found. The chapter then continues by showing how variant interpretations of the
proper implementation of the Xuanming li first appeared as lineages began to organize
within the Bureau of Onmyō ( Jp. Onmyō-ryō). Along the way, the strategies
used by competing Bureau lineages are analyzed—both the successful strategies and the
more precarious. The history of lineage competition within the Bureau in the ninth and
tenth centuries laid the ground for the appearance of non-Bureau lineages of calendrical
experts, which is the subject of the next chapter.

Figuring the Winter Solstice: Japanese Standards for Calendrical Practice
In the middle of the tenth month of 1156, members of Japan’s Council of State
( Jp. Dajōkan) met not once, but twice, to discuss a serious matter. Earlier that
year, a long-simmering succession dispute had broken out in armed conflict in the streets

4
This contradicts the assertion in Ury, "Chinese Learning and Intellectual Life."
111
of the capital, and had been only recently suppressed.
5
Death in the fighting, punishment,
and exile had shook the population of the Council of State, which had only recently been
brought up to full strength by a number of quick appointments and promotions. However,
the old and new members of the Council were not there to discuss rewards, punishments,
or the resumption of normal ceremonies and affairs. Instead, they had gathered in
response to a petition submitted to the Council by a member of the Bureau of the
University ( Jp. Daigaku-ryō), Instructor of Mathematics
6
Miyoshi Yukiyasu.
7

Yukiyasu was bringing their attention to a grave error that he feared was to about to be
committed by the state: the winter solstice was scheduled that year to fall on the first day
of the eleventh lunar month. According to the cosmological philosophy current at the
time, it should have fallen on any day except that day.
Throughout the Heian Period (794-1192), the Japanese court repeatedly showed
its active concern about ensuring a “correct” civil calendar.
8
What this “correctness”

5
The 1156 Hōgen Disturbance ( Jp. Hōgen no ran). There were two succession disputes
involved—that of the throne, between retired tennō Sūtoku and Tōba, Sūtoku’s father at least in name; and
that of leadership of the Fujiwara regental house, between Fujiwara Yorinaga and his older brother
Fujiwara Tadamichi. This skirmish is generally considered a precursor to the Genpei Wars that ended the
Heian Period.

6
Jp. san no hakase.

7
, dates unknown. Active 1156-1191. In genealogies he is listed as the grandfather of Miyoshi
Nagahira ( , 1168-1244), who is known for his support of the Kamakura Bakufu. However,
genealogies of the Miyoshi lineage are all of late composition, and are generally considered suspect by
scholars. See Tokoro Isao, "Miyoshi uji," in Kokushi daijiten, ed. Kokushi daijiten henshū iinkai (Tokyo:
Yoshikawa kōbunkan, 1979–1997) and Iinuma Kenji, "Miyoshi Nagahira," in Kokushi daijiten, ed.
Kokushi daijiten henshū iinkai (Tokyo: Yoshikawa kōbunkan, 1979–1997)..

8
It should be noted that no term for “correctness” or “accuracy” appears in these debates. However, the
use of terms for “error” reveals that there was a concept of “correctness” which motivated both the court
and its calendarists. What was “correct” or “accurate” needs to be examined through the arguments and
decisions made by the court.
112
meant for the high courtiers and members of the Council of State, however, differs from
modern astronomers’ evaluations of the practice of Japanese calendrical astronomy.
9
Yet
previous scholarship has focused on the use of one system of calendrical astronomy for
over eight hundred years as revealing a lack of concern that the Japanese court had for
astronomical prediction. Or else they point to such continuity as indicative of a lack of
ability among Japanese specialists to master, or even understand, Chinese-style
calendrical astronomy. After all, the discrepancies between the astronomical constants
used in the Tang-dynasty Xuanming li ( Jp. Senmyō-reki, “Extending Brightness
Calendrical Astronomy System”), devised in 822, and the actual astronomical orbital
periods meant that, as these discrepancies compounded year after year, the Xuanming li
system could not have continued to produce results that accurately reflected actual
astronomical events such as eclipses or the timing of the equinoxes and solstices.
10
In

In general in the following discussion, I will be using “correctness” when discussing the
standards used by the Japanese court and calendarists to evaluate calendrical production and prediction,
reserving “accuracy” for evaluations from a modern standpoint.

9
“Calendrical astronomy” will be my standard translation of (Ch. lifa, Jp. rekihō), although at
other times “calendrics” will be used. Nathan Sivin has argued for translating when it refers to the
product that determines the civil year as “ephemeris,” and when it refers to the system as “mathematical
astronomy.” Nathan Sivin, Granting the Seasons: The Chinese Astronomical Reform of 1280, with a Study
of Its Many Dimensions and a Translation of Its Records: Sources and Studies in the History
of Mathematics and Physical Sciences (New York: Springer, 2009), 38-40.
While such translations rightly emphasize the astronomical components of time-keeping in the
Chinese-style luni-solar system, where the months and years are more closely linked to astronomical
phenomena than they are in the west, such emphasis comes at the cost of de-emphasizing the purpose of the
systems, which was, after all, the determination of the calendar (through measures of astronomical time).

10
In fact, the “error” of the seventeenth century is generally given as being two days between the winter
solstice as depicted in the calendar and the astronomical winter solstice. Nakayama, A History of Japanese
Astronomy: Chinese Background and Western Impact, 119. However, as Nakayama notes, this discrepancy
“caused little inconvenience or confusion in daily life” and would not have had a major agricultural impact.
Nevertheless, this does not prevent Nakayama from presenting the continued use of the Xuanming li as
“backwards.”
113
China there had been twenty-four distinct systems of calendrical astronomy implemented
and replaced by 1684, the year that the Japanese state finally abandoned the Xuanming li
for official use.
11
According to these scholars, the Japanese court, for all intents and
purposes, must have been ignoring the inevitably increasing rate of prediction
failure—therefore, this counterintuitive situation where the Japanese court was apparently
ignoring “reality” requires explanation.
12
This has been identified as Japan’s “failure” to
develop science, an assumption which itself presumes that the modern scientific method
is somehow natural or inevitable. The assumption of these scholars is clearly that under

11
The adoption of the 1684 Jōkyō-reki ( ) calendrical system composed by Shibukawa Shunkai
( , 1639-1715) was a joint project between the Tokugawa Bakufu and the court in Kyoto, as
neither had the exclusive legal authority necessary to force its adoption.

12
It is worth questioning the depiction of the Xuanming li predictions as inevitably becoming less and
less accurate, and noticeably so. John Steele has charted the accuracy of eclipse predictions in Steele,
Observations and Predictions of Eclipses by Early Astronomers, 226-227, and it shows an increasing
average error between solar eclipse as predicted using the Xuanming li system and the modern calculation
of the eclipse, using data taken from predictions in the Japanese record. The type of error also changes over
time, as eclipse predictions using the Xuanming li were generally late before the fifteenth century, and
generally early thereafter. However, there are three major problems with the assumptions Steele is using to
evaluate the Xuanming li system’s accuracy. First, as this and the following chapters will show, the
Japanese calendarists using the Xuanming li adjusted the system over time, sometimes in an ad-hoc fashion.
This means that there are significant barriers to using the Japanese eclipse record as a proxy for the results
of the no-longer completely extant Xuanming li. Second, systems other than the Xuanming li were used for
eclipse prediction, as will be discussed in chapters four and five.
Third, and most importantly, Steele’s chart shows a wide range in accuracy depending on the
prediction. Although Steele’s graph shows an upward trend in error, the noise in the chart could have made
it very difficult to discern the pattern. Examples would the years around 1400, when one eclipse prediction
was exactly on the mark, three were an hour or less too late, two were an hour or less too early, three
between two and three hours too late, and one about two and a half hours too early. (———, Observations
and Predictions of Eclipses by Early Astronomers, 226). It would have proved difficult for a Japanese
observer to notice a pattern in these range of errors, and decide to blame the system rather than the
practitioner.

114
normal circumstances, the divergence between prediction and reality should have
prompted a re-evaluation of the predictive system being used.
13

Explanations have been plenty. One proposed cause is that the Japanese court had
turned away from “reality” under the influence of esoteric Buddhist mysticism
14
—yet
1156 was long after esoteric Buddhism became prominent at the Japanese court, and still
the Council of State was concerned about astronomical “accuracy.”
15
There are clearly
problems with this explanation. Therefore, in order to explain the discrepancy between a
concern for astronomical accuracy and a failure to achieve it (by adopting a new
calendrical system), some scholars have argued that Japan was not aware of, or else could
not access, newer Chinese developments in astronomy due to economic and political
isolation from the continent.
16
But this argument is severely weakened by recent
scholarship that has illuminated Japan’s continuing trade and communication with the
continent—isolation can no longer be assumed for much of Japanese history.
17
The

13
This understanding of the motivation for change, which I find as an assumption in the works discussed
subsequently, is of course different from the model of scientific revolution found in Thomas S Kuhn, The
Structure of Scientific Revolutions (Chicago: University of Chicago Press, 1970).
However, it should be noted, that the switch of one system of calendrical astronomy to another
in China, particularly as most systems built upon each other, and often shared constants in common, as was
the case with the Xuanming li and its immediate predecessors. Xin Tang shu, Lülizhi.

14
Nakayama, A History of Japanese Astronomy: Chinese Background and Western Impact.

15
Of course, the Japanese court’s definition of “accuracy” and the modern astronomical definition
differed, as shall be shown below.

16
The “isolation” argument is the primary model found in Sugimoto Masayoshi and David L. Swain,
Science and Culture in Traditional Japan, A.D. 600-1854, where contact and isolation alternative advance
or hinder “progress” in the major sciences of pre-modern Japan.

17
One of the most important studies for this subject for economic and trade links is Charlotte von
Verschuer, Le Commerce Extérieur Du Japon Des Origines Au Xvie Siècle (Paris: Maisonneuve & Larose,
1988), which was revised and translated in 2006 as ———, Across the Perilous Sea: Japanese Trade with
115
apparent lack of “scientific progress” has also been blamed on the rise of lineages that
monopolized the practice of calendrical astronomy and the (presumed) decline in
standards of education and production that must have resulted
18
—yet even these scholars
often admit that nepotism need not necessarily indicate a decline in performance over
generations.
19
This explanation, too, is insufficient. Furthermore, as the events of 1156
show, it cannot be assumed that the Japanese court ceased to care about the calendar. If
the behavior of the court and its specialists in calendrical astronomy fail to match up to
modern assumptions, then it is perhaps these assumptions that must be re-evaluated.
A close reading of court debates and controversies about the calendar shows n
that not only was there a prevalent and lasting concern on the part of the Japanese
authorities to produce a “correct” official calendar, and concerted effort towards ensuring
that “correctness,” but also that these efforts created a particular tradition of interpretation
and implementation based on the Chinese Xuanming li calendrical system. In other words,
a new Japanese style of Chinese calendrical astronomy emerged. Questions of the

China and Korea from the Seventh to the Sixteenth Centuries, ed. Cornell East Asia Series (Ithaca, NY:
East Asia Program, Cornell University, 2006).

18
By both Nakayama, A History of Japanese Astronomy: Chinese Background and Western Impact. and
Sugimoto Masayoshi and David L. Swain, Science and Culture in Traditional Japan, A.D. 600-1854; but
also by Hosoi Hiroshi, "Onmyō-ryō to tenmon rekigaku kyōiku," Tenmongaku-shi kenkyū-kai shūroku, no.
2 (2008).

19
Nakayama, A History of Japanese Astronomy: Chinese Background and Western Impact. specifically
states that nepotism need not be any barrier to progress or good practices in a discussion of the Shibukawa
lineage among the Tokugawa Bakufu’s Tenmonkata ( ).
In fact, the Heian court often relied upon systematic education within lineages to ensure a
population of trained experts, particularly once the official system of education and training declined in the
tenth century. See Robert Borgen, Sugawara No Michizane and the Early Heian Court, Harvard East Asian
Monographs (Cambridge, Mass.: Harvard University Press, 1986) for more information on how the training
within the Sugawara lineage produced a population of educated scholars in the Chinese classics.

116
astronomical significance of eclipses that did not take place over Japan were considered,
as was the need for extra-systematic adjustments to bring the mathematically
incommensurate cycles of the sun, moon, and the earth’s rotation into a coherent system
for marking time. This distinctly Japanese version of the calendrical system was forged
through court debates, through the innovations and arguments of calendarists, and the
decisions of the senior nobles of the Council of State. At the same time, court debates
provided a forum for calendarists to promote their careers and the reputations of their
lineages.
In 1156, the final piece of this new Japanese interpretation was put into place by
the Council of State over the objections of the official calendarists it employed. At this
time it was decided that every nineteen years, and only every nineteen years, the winter
solstice should fall on the first day of the eleventh month.
In this chapter, as well as in chapters five and six, the history of public debates in
Japan concerning issues of calendrical astronomy is traced from the ninth through the
thirteenth centuries. Through these debates, lineages of official astronomers formed and
wrested control of official pronouncements of calendrical fact from rivals in the Bureau
of Onmyō. As these lineages attempted to differentiate themselves from each other, a
Japanese-style interpretation of Chinese calendrical astronomy emerged. As controversies
continued, new groups of experts in calendrics also emerged. These “outside”
20
experts

20
These new groups of individuals and lineages with recognized expertise on calendrical matters, or
matters related to calendrics, existed outside of the legal structures that had previously established
calendrics ( Ch. lifa, Jp. rekihō) as a field. They therefore stood as “non-official” in the sense that they
were not legally members of the official structure, but they were recognized and employed by the Council
117
initially sought to correct the calculation of the Japanese year. Later, as the victorious
lineage of calendrical astronomers within the Bureau of Onmyō became more secure,
these experts used public criticism of the calendar as a method for self-promotion the
world of court patronage.
These various lineages of experts came to be organized around particular
canonical texts,
21
and formed schools of interpretation centered on them. All debates
relied on the authority of texts, although early debates regarding the calendar also
sometimes included statements involving the technical details of methods of calculative
astronomy. Eventually this came to mean that the primary authority cited in the debates
was the authority of the interpretive tradition of a text that a particular lineage possessed.
In 1156, it was the Miyoshi combination of court history with a reading of the
Zhou bi suan jing ( Jp. Sūhi san kyō, “The Arithmetical Classic of the Gnomon
and the Circular Paths of Heaven”) that prevailed over the official calendarists’ emphasis
on the results of calculations derived from their tradition on the Xuanming li. It was
explicitly determined that from this time forward the court’s calendarists were to
calculate the calendar using the Xuanming li system—then adjust the results of such
calculations until they were “correct” according to standards developed over the history
of debates over calendrics at the Japanese court.

of State and the high nobility as checks upon the production of officially-designated calendarists, and in
that sense might be considered “official.”

21
In the case of the Buddhist “lineage” or school, often referred to as Sukuyōdō ( ), while the
lineage claimed certain Buddhist texts, and they are associated with one particular set of texts, the evidence
discussed in chapters four and five shows that the individuals of this lineage may have been less unified in
their use of texts than has been previously assumed.

118
This is not to say that the authority of the position of officially-designated
calendarists in the court bureaucracy was not itself a factor in determining what sorts of
calendars were “correct.” In fact, in some cases, it seems that it was only due to an
individual holding such a position of official calendarist that awarded them the final word
in a debate. While the pronouncements of the individuals holding these positions might
not be taken as indisputable truth,
22
only they could legitimately produce the official
civil calendar. As a result, outsiders who wished to effect changes or corrections to the
official civil calendar had to rely on the power of the court, usually through the Council
of State, to force the calendarists to adjust their work. Their opinions were placed before
the Council due to relationships with powerful patrons within the Council itself.
Although the lineages of official calendarists failed to establish a monopoly on
statements of calendrical or astronomical truth, they like other lineages, also utilized the
politics of patronage. They used strategic alliances with outsiders, and built a reputation
for skill in calendrical and ritual affairs. Thus they developed patron-client relationships
in order to secure their lineages’ monopoly on the production of the official year.
This monopoly, however, was threatened by external pressures as well as internal
ones. By the mid-eleventh century, even though the bureaucratic position of the
remaining calendarist lineage at court was by then secure, the product it produced, the
official calendar, was newly challenged by the appearance and circulation of unofficial
calendars. And by the thirteenth century, despite a nominal court monopoly on the

22
This is a major point of difference between these experts at the Japanese court and Norbert Elias’
“priesthood” of the pre-information society, which could control truth as they wished. Elias, "Knowledge
and Power: An Interview by Peter Ludes."

119
definition of time, it was unofficial calendars, produced in the provinces, which came to
dominate astronomical timekeeping for most of Japan. What this fracturing of the official
definition of time meant for the legitimacy of the Japanese court has something to say as
to why the Japanese court was concerned about calendrical “correctness” in the first
place.

The Difficulties of the Chinese-Style Year
Defining what made a “correct” Chinese-style luni-solar calendar, however,
was not a simple task. A Chinese-style luni-solar calendar attempted to reconcile a
number of astronomical cycles—the solar year and the lunar phases most
prominently—using an interlocking set of approximate cyclic periods. Most of these
cycles, including the year, lunar month, and longer cycles such as the 19-year Rule Cycle
for intercalation, were approximations. This meant that over time the discrepancy
between the actual astronomical event and the calendrical cycle would compound,
creating a perceivable “gap” between the calendar and the astronomical world it was
meant to represent.
Nevertheless, that a calendar be “correct” was the lynchpin not only of its
usefulness for agricultural and civic ends, but the essential element of its political
significance. This political significance was even more important in motivating continual
improvements of the Chinese-style calendar.
23
The correspondence between

23
That agricultural need was not a driving force for continued calendrical reform in China, as agricultural
requirements were in reality quite simply handled by the length of the solar year, is argued in Sivin,
120
astronomical phenomena and the calendrical year was one of the ideological
underpinnings of the Son of Heaven’s right to rule in China. The association of an
accurate calendar with the Mandate of Heaven dates to before the Qin Dynasty (221 –
206 B.C.E), and can be found explicitly in the classic text known as the Book of
Documents ( Ch. Shang shu, Jp. Shō sho). According to this text, the duty of the
ruler is to observe heaven and determine time for the people.
24
The phrase
(Ch. guanxiong shoushi, Jp. kanshō juji) referred to this ideal and would subsequently
become a by-word for legitimate rule.
25
Yet the complicated demands made upon a
Chinese-style calendar meant that ensuring such “correctness” was a constant struggle.
The unending nature of the effort to improve the calendar was incorporated into
the very ideology that defined the duty of a new dynasty once it received the Mandate of
Heaven. According to Sima Qian’s The Records of the Grand Historian ( Ch. Shiji,
Jp. Shiki), every dynasty needed to undertake a program of calendrical reform when it
took power. This would serve as a concrete symbol of its new status as the legitimate
recipient of the Mandate of Heaven. Furthermore, by the second half of the Han Dynasty
(206 BCE – 220 CE), reforms to the calendrical system also could occur at any point

Granting the Seasons: The Chinese Astronomical Reform of 1280, with a Study of Its Many Dimensions
and a Translation of Its Records: , 41.

24
Shang shu, Yao dian. (The text consulted was the Siku quanshu edition.)

25
The name of the Yuan Dynasty system of calendrical astronomy, Shoushi li ( , 1281) also comes
from this phrase.

121
during a dynasty: thus the Mandate of Heaven could also be symbolically renewed
(without the risk of regime change).
26

The justification generally used to introduce each reform was that the old
system of calendrical astronomy was no longer accurately predicting astronomical
phenomena.
27
By the end of the Han Dynasty, the highest test of the accuracy of a
system of calendrical astronomy was the prediction of solar eclipses, which depended on
the combination of several cycles based on a relatively accurate mathematical model of
the sun-earth-moon system.
28
The system of regular reform in China has been hailed as a
sign of China’s early accomplishments in astronomy.
29

The Chinese calendrical system produced what is known as a luni-solar year,
which typically consisted of twelve lunar months of either 29 or 30 days. The addition of
a leap month approximately every three years kept the calendrical months and the

26
See the debates on the calendrical reform found in Dull, "A Historical Introduction to the Apocryphal
(Chʻan-Wei) Texts of the Han Dynasty". See also Kiyosi Yabuuti (Yabuuchi Kiyoshi), "The Calendar
Reforms in the Han Dynasties and Ideas in Their Background," Archives internationales d'histoire des
sciences, no. 24 (1974).

27
This can be shown through an examination of the prefaces to the description of each calendrical section
in the Treatise on Pitchpipes and Calendrics ( ) in the Xin Tang shu.

28
Steele, Observations and Predictions of Eclipses by Early Astronomers, 177.

29
For this reason, some prominent scholars have called for translating (Ch. li, Jp. reki) as
“mathematical astronomy,” in order to better highlight China’s prominence in this area. However, it should
be noted that the goal of both astronomical calculation and the reform of systems of mathematical
astronomy was actually the determination and dissemination of official, cosmological, time. For this reason,
the current dissertation uses the translation “calendrics” and “calendrical astronomy.”
The political motivation behind reforms was also identified early, so that calendrical reform in
China was long recognized as not an apolitical, “purely scientific,” endeavor. Yabuuti, "The Calendar
Reforms in the Han Dynasties and Ideas in Their Background.".

122
seasons roughly in line.
30
The indivisibility of the various astronomical units, however,
made this task not as simple as it might initially seem.
For instance, just as in the modern Western calendar, the day was an indivisible
unit. If the start of a new year is defined by the return of the earth a particular spot in its
orbit,
31
then it should be recognized that this moment need not occur at the beginning of
a day. In fact, it rarely does so, as the period of the earth’s orbit cannot be divided neatly
by 24-hour units. This is the meaning hidden in the fractional day in the length of our
solar year, roughly 365.242 days. Therefore, while it might be said that the astronomical
year does not begin until this moment of return, the civil year begins with the day that
roughly corresponds with it—days are not split between years.
Despite the fact that Chinese astronomy did not have at its center a heliocentric
model of the universe, their systems of calendrical astronomy recognized that the
beginning of an astronomical period such as the solar year might begin at any time of day,
not just at the day’s beginning, however defined. For civil purposes, however, splitting a
day between years simply would not do. Much of the complexity of the Chinese system
involved adjusting the relationship of the day to the astronomical event (which occurred
at a specific time) to create the most “accurate” fit.

30
The need for intercalary, or “leap” months was recognized by many civilizations, apparently
independently, and can be found also in Ancient Greek and Jewish calendars. The methods of intercalation,
however, differ from system to system—nevertheless, the roughly 1-in-3 years occurrence of an intercalary
month can be found in all of these systems.
The Islamic calendar, being not luni-solar but lunar, has its months “untethered” from the
seasons. It is for this reason that the dates of Ramadan vary so widely over the years.

31
Or, from a geocentric point of view, the moment at which the sun returns to the same position relative
to the angle of the ecliptic and the background stars.

123
In the modern Western calendar, only the unit of the day is indivisible, or
cannot be divided between astronomical periods. In other words, a day may not begin in
one year and end in another. As a result, the fractional day in the length of the earth’s
orbit around the sun is incorporated into the calendar by adding a full day roughly once
every four years—this is our “leap year” system. By contrast, the “leap” system in
Chinese-style luni-solar calendars focused on bringing not just days and the year into
mathematical correspondence, but the days, the lunar orbit, and the solar year. Other
important cosmological cycles were also incorporated.
In the Chinese-style luni-solar calendar, unlike in the Julian or Gregorian
calendars of the West, the correspondence between months and lunations (the period
between new moon and the next new moon) was preserved. As the average lunation is
29.531 days,
32
the luni-solar year usually consisted of twelve months, with an
approximately equal number of 29-day months (“small”) and 30-day months (“large”).
Again, days were indivisible, just as in the West, so although the average lunar orbit was
not divisible by whole days, Chinese-style lunar or civil months were an approximation
of astronomical months (lunations) counted in full days.
33
However, a year of twelve
lunar months, each with an average length of 29.5 days, resulted in a shortfall of

32
Actual lunation periods from lunation to lunation differ due to a number of gravitational factors—there
is no constant period of lunation. The Chinese-style calendar systems under discussion here used values for
an average lunation period instead of attempting to fully model the sun-moon-earth system, which is
technically quite difficult. See Nathan Sivin, "Cosmos and Computation in Early Chinese Mathematical
Astronomy," T'oung pao 55, no. 1–3 (1969).

33
Short and long months initially alternated. See Christopher Cullen, Astronomy and Mathematics in
Ancient China: The Zhou Bi Suan Jing, Needham Research Institute Studies (Cambridge: Cambridge
University Press, 1996) for the historical background of the early Chinese calendars.

124
approximately 18 days between that twelve-month year and the solar year of roughly
365.242 days.
34
In the Chinese-style system, lunations were also considered indivisible,
so a year could not contain only part of a lunar cycle. Instead of dividing lunar months
between years in order to preserve the relationship between the seasons and the lunar
months, the system kept a rougher sort of correspondence between the lunar months and
solar years through a system of “leap” or intercalary months that were inserted regularly
when the discrepancy between lunar month and solar season grew too large.
The first regular system of intercalation in China, dating to as far back as to the
Warring States Period (465 – 221 B.C.E), was based on the mechanical repetition of a
nineteen-year cycle, known as the Rule Cycle ( Ch. zhang, Jp. shō). This cycle, also
known in the West as the Metonic Cycle, is the shortest period of time in which the
number of days in lunar months and the number of days in solar years roughly
correspond: the number of days in 19 years and in 235 months are largely equivalent.
35

The Rule Cycle also marked a cosmological cycle, a “zero-hour,” one of a
number of “millennial” elements in the Chinese calendrical astronomy system. Every
nineteen years, as the correspondence between a day’s place in the solar year and its
place in a lunar month reoccurred, was seen as an expression of cosmological return. This
cycle and its near-religious significance were recorded not only in the Han histories,
36

34
The length of time from vernal equinox to vernal equinox.

35
This cycle is only, however, a rough correspondence and is not exact, as will be discussed below.

36
The Han shu ( Jp. Kan sho, “Han History”) and Hou Han shu ( , Jp. Kō Kan sho, “Latter
Han History”).

125
but also in texts of mathematics such as The Arithmetical Classic of the Gnomon and the
Circular Paths of Heaven. It was not the only such cycle typical of Chinese systems of
calendrical astronomy,
37
but it became the most prominent one in Japanese court culture.
The Rule Cycle first appears in Japanese historical sources with great pomp in the
year of 794, when the co-incidence of the winter solstice and the first day of the eleventh
month in the official calendar
38
—the astronomical start of each year, even if the civil new
year would not be for another two months
39
—was noted. As this also corresponded both
with a new capital and with its founder Kanmu Tennō’s
40
need for symbols of cosmic
legitimation,
41
it was observed with great ceremony. The celebration not only rewarded
loyal ministers with promotions, but also featured songs of praise for the (obviously)
enlightened rule that saw this cosmic sign of the millennium come to pass. From the year
794 until the fifteenth century, the reoccurrence of the winter solstice and the new moon
of the eleventh month every nineteen years became a regular event for the Japanese

37
An overview of some of the major cycles can be found in Sivin, "Cosmos and Computation in Early
Chinese Mathematical Astronomy."

38
Called the “eleventh month,” this lunar month need not necessarily be the eleventh lunar month in a
year—an intercalary month, which was technically uncounted for naming purposes. For example, if it was
placed after the fourth lunar month, it would be designated “intercalary fourth month” ( Jp. uruu
shigatsu). The subsequent month would then be called the “fifth [lunar] month” even though it would be
the sixth, in order.

39
Astronomical calculations used in producing the calendar were all began from the winter solstice and
the eleventh month, as can be seen in surviving draft calculations from the sixteenth century. Unknown,
"Kan'ei nenkan reki suisankōtō," in Yōmei bunko (Tokyo: Tōkyō Daigaku Shiryō hensanjo, 17th c.).

40
737 – 806, r. 781 – 806.

41
Kanmu Tennō’s reign in these years was marked by natural disasters attributed to victims of his rise to
power, as well as by ruinous wars in the north and the cost of building not one, but two capitals in quick
succession.

126
monarchy, and the observance of the Rule Cycle was an important part of the regular
system of court ceremony.
42
The celebration of the Rule Cycle continued despite the fact
that the system of calendrical astronomy used by the Japanese court from the
ninth-century onward no longer included the Rule Cycle, having abandoned it in favor of
cycles that more accurately approximated astronomical periods.
From before even the middle of the Tang Dynasty (618 – 907) the Rule Cycle had
been dropped from the mathematics of calendrical systems in China.
43
More precise
44

values for lunar and solar periods had necessitated its abandonment, as these new values
could not be easily incorporated into a 19-year cycle. Furthermore, as is known by
modern astronomers, the Rule Cycle (or Metonic Cycle as it is known to the West) is
merely encapsulates an approximate correspondence between the lunar and solar cycles.
If more precise values for the lengths of lunations and solar years are used, the 19-year
cycle breaks down.
45
If one attempts to preserve both the Rule Cycle and these new
values for the lengths of years and months, the result is a growing gap between the

42
Momo Hiroyuki, "Sakutan tōji," in Kaitei yūshoku kojitsu jiten, ed. Katō Tejiro et al. (Tokyo: Kisho
kankōkai, 1976).

43
The technical term for this development in the history of Chinese calendrical astronomy, at least in
Japanese scholarship, is (Jp. hashō). Calendrical systems that do not include the Rule Cycle are said
to be hashō. Yabuuchi Kiyoshi, Chūgoku no tenmon rekihō (Tokyo: Heibonsha, 1969)..

44
More precise values did not necessarily mean that the values were more accurate. Throughout the Han
Dynasty, there was a general trend towards using values in calendrical systems that were carried out to
further and further decimal places (or, to put it more accurately, which were fractions with larger and larger
denominators). However, such precision did not necessarily mean that the resulting value was any more
accurate, particularly when compared to modern astronomical values. See Sivin, "Cosmos and
Computation in Early Chinese Mathematical Astronomy."

45
Nineteen years is approximately .087 days less than 235 lunar months. As a result, over time, strict
adherence to the Rule Cycle would result in either a shortening of the proper value of days in an
accumulation of lunar months, or a lengthening of the proper value of days of an accumulation of years.

127
predictions that the Rule Cycle and the calendrical system produces. At first, this gap
would not amount to a full day and would be ignorable; but as time passed and the
discrepancy compounded, either the Rule Cycle or the value of the length of months and
years would have to give. This is the problem that Japan’s calendarists faced in the Heian
Period.
The Rule Cycle was not part of the tables and instructions for calculations in the
Xuanming li system as it had been composed in China. In Japan, therefore, many of the
occurrences and celebrations of the winter solstice-new moon conjunction that marked
the start of each new Rule Cycle had to have been the result of adjustments made to the
calendar year after it was calculated. This was an ad-hoc addition to the Xuanming li
system, applied by brute force post calculation, in order to preserve a politically
significant astronomical “constant.”
The “re-introduction” of the Rule Cycle to the Xuanming li was only one of the
ways in which Japanese practice diverged from the original Chinese system. This style of
calendrical astronomy emerged slowly over time, as debates took place at court between
factions, and new court decisions were added to practice. The result of the 1156 debate
was only the latest major addition to an idea of “correct” practice which had evolved over
two and a half centuries, and which had its origin even earlier.

The court’s reason for allowing debates to determine “correct” practice had
much to do with the uncertainty that surrounded matters involving calendrical astronomy.
Even as the yearly calendar was something that everyone at court was expected to know
128
and follow, very few had the mathematical skills and astronomical knowledge necessary
to predict what the calendar for the next year would or should be before it was
promulgated. The Chinese calendrical system, with its multiple cycles and rules intended
to preserve cosmological harmony, was complex enough that few in either China or
Japan likely knew exactly what the next official year would be like in advance—how
many months, how many days—except for the calendarists.
46
As debates between
calendarists show, sometimes not even these experts were absolutely certain of the way
the next year should take shape. However, the inherent complexity of the system that
created a luni-solar year was not the only reason for why it was difficult for untrained
members of the Japanese court to themselves predict or determine what a “correct” year
would look like, let alone effectively supervise the work of calendarists to ensure the
production of the official calendar.
In the record of the debate of 1156, members of the Council of State expressed
their uncertainty concerning many of the finer points of the debate between Miyoshi
Yukiyasu and the official calendarists. For example, Controller
47
Taira Nobunori
48
noted
in his diary that the debate was hard to follow, not only due to the complexity of the issue

46
Some attempts at perpetual calendars were made, as for example can be found in the Shokuin ryō
section of Ryō no shūge, but they would not have proved particularly accurate when compared to actual
court calendars, particularly in the face of the adjustments and debates described below.

47
Jp. benkan, a position in the Japanese ritsuryō bureaucracy involved in handling and transmitting
the decisions made by the senior nobility.

48
, 1112 – 1187.

129
but because many “secret texts” and methods were being cited.
49
These “secret texts”
were the concretization of familial expertise, which had been itself created by the
demands from the court that had driven the development of a Japan-specific
implementation of the Xuanming li calendrical system.
50
Over the past three hundred
years, demands from the Council of State for the production of a “correct” calendar had
contributed to the creation of “secret texts.”
By proposing to change the year of 1156 in mid-course by fiat,
51
the Council of
State added yet another component to the system, not to be found within the text of the
Xuanming li. This increased the amount of components not found in the text of the
Xuanming li in the Japanese implementation of that calendrical system.
52

This corpus of non-textual practice was exactly what Nobunori was complaining
about with his talk of “secret texts.” Yet these very “secret texts” were the result of

49
Hyōhanki Hōgen 1/10/18 and 1/10/20. The edition consulted was Zōhō shiryō taisei kankōkai, ed.
Hyōhanki, 5 vols., Zōhō shiryō taisei (Kyoto: Rinsen shoten,1965).

50
In this way, the “secret texts” referred to by Nobunori resemble the “real,” or content-driven, aspects of
hiden ( ) or “secret transmissions” found in the esoteric Buddhism of the Tendai School. Despite the
social use of these transmissions, it had been argued that they should be taken seriously as a mode of
criticism. On hiden, see Stone, Original Enlightenment and the Transformation of Medieval Japanese
Buddhism.

51
While changing the accounting of the year while the year was in progress would have only added to the
uncertainty about the length and make-up of each year, in effect informing the population that what had
been previously a fact about time was now to be considered incorrect, such an act by the Council of State
was not completely unprecedented. This is known from a note attached to the seventh month of 936 in
Shoku Nihon kōki, discussed below. That incident, however, did not occur as close to the dates being
changed as was the case in 1156—in this respect, 1156 might be seen as indeed being unprecedented. The
drastic nature of the 1156 adjustment can be seen in how it was used as precedent for later calendrical
shifts.

52
In this case, a new rule that if a winter solstice/first-day of the eleventh lunar month conjunction should
appear before the 19-year cycle began again, that the calendar should be adjusted to prevent it. In other
words, this too was a rule that was applied after the initial calculation of the calendrical year.

130
mathematically incompatible demands placed upon the production of a “correct”
astronomical calendar after the introduction of the Xuanming li system in 862.

Importing Chinese Calendrics into a Japanese Court
The set of extra-textual demands (later reflected in “secret texts”) on the
production of the Japanese calendar using Chinese mathematical canons for calendrical
astronomy took time to develop. Despite the fact that its final rule was only implemented
in the twelfth century (in the court debate discussed above), Japanese divergence from the
Chinese textual ideal began long before.
It is difficult to know exactly when individuals in Japan began to calculate their
own luni-solar calendar. It is assumed that the earliest yearly calendars that were used in
Japan had been imported from China or Korea. Early artifacts, such as the Inariyama
Tumulus Sword, have inscribed upon them elements of Chinese-style date
keeping—specifically the sexagenary year cycle. Yet the corpus historical dates found in
chronicles compiled later, for instance the Nihon shoki ( , “Chronicles of
Japan”) and Shoku Nihongi ( , “Continued Chronicles of Japan”) for the period
leading up to the early eighth century, do not correspond to any single known calendrical
system. This became the subject of attention by scholars in Japan as early as the
eighteenth century.
53
Some of these inconsistencies might merely be the result of
eighth-century editorial interference. Others, however, seem to have been the attempt

53
See the discussion in the first section of Hosoi Hiroshi, Kodai no tenmon ihen to shisho.. The fact that
these dates don’t match with known Chinese systems led some early Kokubungaku scholars to identify
them as proof as a pre-Chinese calendrical system, native to Japan.

131
retroactively to get years that were calculated under a previous system to conform to the
system currently in use.
54

The first references to calendrical system use in the historical record are
themselves perplexing. The first appearance of a named calendrical system in the
historical record is found in the entries concerning the reign of Jitō Tennō, who
succeeded her husband to become ruler of Japan in 687. According to the Chronicles of
Japan, in the eleventh month of the fourth year of Jitō’s reign (690), a royal decree went
out that two calendrical systems, the Yuanjia li ( Jp. Genka-reki) and Yifeng li
( Jp. Gihō-reki), were to be used in tandem. The use of two systems in tandem, if
this account is accurate, would be the earliest example of a Japanese practice deliberately
diverging from the continental model—in China only one system of calendrical
astronomy, perforce, was employed at any one time. How this system may have worked

54
Hosoi finds evidence for this in the Heian-period Seiji yōryaku ( ), where the sexagenary
count applied to dates for precedents from the Six National Histories ( ) have been changed to
accord with the current calendrical system. Hosoi traces such corrections back to the lectures on the
Nihongi held as a court event once every thirty years during the ninth century, and these lectures’
connection to the Bureau of the University ( Jp. Daigaku-ryō). Ibid.
It should be noted, however, that the type of work necessary to adjust the counts is more
closely associated with the skills of the Instructors of Mathematics ( Jp. san no hakase) in the
Bureau of the University than with those Instructors of other fields who were more likely to participate in
the Nihongi lectures themselves—particularly given the need in calendrics to manipulate large numbers.
The author of the Seiji yōryaku, Koremune Tadasuke ( , d. 1009?), was known as a legal expert
(an Instructor of Laws, Jp. myōbō hakase), which would not naturally seem to include
mathematical ability, so it seems logical to assume that such information came from elsewhere. This may
indicate that either there was more participation by Instructors of Mathematics in other activities such as
the Nihongi lectures, or that the mathematical skills necessary to calculate a calendar were more widely
distributed than previously assumed. (On estimations of the numeracy of the Heian population, see Yoshio
Mikami, A History of Japanese Mathematics (Chicago: Open Court Pub. Co., 1914).. Unfortunately, there
have been no more recent studies.)

132
has been the subject of a great deal of controversy,
55
even confusion. As was pointed out
in a text from the eighteenth century, due to the different divisors used in each system,
combining the two systems mathematically was patently unfeasible.
56

Even if both systems were never actually used in some sort of combination for
the calculation of any single year, the very name of the latter system already indicates a
divergence from Chinese practice. The “Yifeng li” (Gihō-reki) was the Goguryeo
57

appellation for the Tang Dynasty Linde li ( Jp. Rintoku-reki). In a ninth-century
of Chinese books known to have been present in the capital of Heian, both the “Yifeng li”
and the Linde li are listed as separate texts, with a different number of volumes given for
each.
58
This implies that the “Yifeng li” and Linde li editions at the Heian court differed
in some crucial features. Given the evidence for the existence of alternative versions of
calendrical texts from China as early as the seventh century, the assumption that the
Koreans or Japanese merely adopted the Chinese calculations without question or
modification seems hard to support.
59

55
See the history of scholarly debate on this topic in the first section of Hosoi Hiroshi, Kodai no tenmon
ihen to shisho.

56
The divisor for the Yuanjia li was 752, and that for the Yifeng li was 1340. Rekihō shinsho ( )
Hōreki era (1751-1764).

57
One of the Three Kingdoms of early Korea, it was the northernmost, stretching from the north of the
peninsula to north of the Yalu River. 32 BCE – 668 CE.

58
Hosoi Hiroshi, Kodai no tenmon ihen to shisho.. The catalog in question is Nihon koku genzai sho
mokuroku ( ), compiled by Fujiwara Sukeyo ( , 847-897) around 891.

59
The possibility of alternate texts after the ninth century is also strong. While the original text of the
Xuanming li, used in Japan from 862 to 1684, does not survive, fragments in other sources do: a partial
description in the Xin Tang shu ( , “New Tang History”); a more detailed description in a edition of
the Koryeo sa ( , 1451) (Daitō Bunka Daigaku Tōyō kenkyūjo, ed. Kōraishi rekishi Senmyōreki no
kenkyū (Tokyo: Daitō Bunka Daigaku Tōyō kenkyūjo,1998).); and manuals from Japan’s early modern
133
When calendar-making was instituted as an official duty of the Bureau of
Onmyō ( Jp. Onmyō-ryō) in the ritsuryō law codes of the eighth century, further
divergences from Chinese practice were institutionalized. While both the Tang and
Japanese bureaucracies had a single Instructor of Calendrics ( Jp. reki no hakase),
in other respects the bureaus were different. In Japan, the Instructor of Calendrics was
also responsible for calculating the calendar for each year, while in the Tang-dynasty
Astronomy Bureau ( ) there were two Technicians who had this responsibility.
60

Japan’s section on calendrics was also comparatively understaffed, with ten students
compared to the forty-one mandated in the Tang.
The texts studied also differed. It is well known that in 757 the Japanese court
mandated a course of study focusing on the Treatises on Pitchpipes and Calendrics ( Ch. lülizhi, Jp. ritsureki shi) from the Han and Jin histories ( Ch. Han shu, Jp.
Kan sho; and Ch. Jin shu, Jp. Shin sho, respectively) along with the Nine Chapters
on the Mathematical Arts ( Ch. Jiuzhang suanshu, Jp. Kyūshō sanjutsu), and
the previously mentioned Arithmetical Classic of the Gnomon and the Circular Paths of
Heaven, as well as other mathematical and cosmological texts.
61
What is left out of these
typical depictions of calendrical study in Japan is that by the time of the adoption of the

period, which were themselves reconstructions. There is room in the omissions of each version for variation
between the texts to exist—for a more detailed analysis, comparative study will be required.

60
Nakayama, A History of Japanese Astronomy: Chinese Background and Western Impact, 19.

61
Ruijū sandai kyaku 17. Cited, for example, in Nakayama, A History of Japanese Astronomy: Chinese
Background and Western Impact, 72.

134
Xuanming li system in 862, the aforementioned texts had been officially abandoned—at
least for calendrics.
62
This is attested in a petition from an Instructor of Calendrics to
introduce texts to supplement the Xuanming li for study in 877. In response to this
petition, the court allowed the re-adoption of earlier texts, provided that they were only
used for instruction and not for calculating the calendar.
63
Although this is the last
detailed official decree on the training of official calendarists, that the mandated training
program changed even once indicates that it may have been changed at other times as
well—in other words, the depiction of the field as essentially static requires revision.
The irony that would appear later years is that although this decree mandated
that the Arithmetical Classic of the Gnomon and the Circular Paths of Heaven was not to
be used for calculating the calendar, that particular text would nonetheless be cited as
authoritative for evaluating the calendar of 1156. Through that decision, it would be
re-incorporated into the calendar-making process from that year forward. Such a
requirement was a divergence from prescribed Chinese calendrical practice in the canon
of the Xuanming li—yet this was a divergence built upon Chinese classical texts. Such a
situation embodies the complicated relationship between Chinese tradition and
contemporaneous Chinese practice in the East Asian sphere.
64
Similarly, it should be

62
There is evidence to suggest that some of these texts remained objects of study for experts in
mathematics—in particular, the Jiuzhang suanshu and Zhoubi suanjing were assigned texts for the study of
both calendrics and mathematics at court. See chapter six for more discussion.

63
Ruijū sandai kyaku 17, Gangyō 1/7/22.

64
There are marked similarities with the Yi Choseon Korean court’s preservation of “Chinese culture”
after the establishment of the Mongol Qing Dynasty, although the situation described above lacks the moral
example found in the Korean case. On the use of reign names to signify a continuity with the Ming (Han)
past in opposition to the Qing (Manchu) present in the East Asian political sphere, see JaHyun Kim
135
noted that the structural difference between the Tang Astronomy Bureau and the Japanese
Bureau of Onmyō was a matter of difference despite familial resemblance, rather than of
radical divergence. Such slight differences, however, could have significant
consequences in practice. These differences would also lead to conflict not only within
the Bureau, but also between the Bureau of Onmyō and other sections of the court
bureaucracy.

Starting Debates: Interference and Independence
The debates that arose between members of the Bureau of Onmyō derived, in
large part, from the incompatibilities within the system of calendrical astronomy that
Japan inherited from Tang China, as well as to the very structure of the division for
calendrics within the Bureau of Onmyō, as the Council of State had modified it. As
shown above, the Japanese practice of Chinese calendrical astronomy differed
structurally from its Tang model from its inception: in the size of the population of
calendarists at court as mandated by administrative laws and in terms of the texts used by
calendarists at the end of the eighth century. Japanese practice continued to diverge under
the pressure of changes made to the Bureau of Onmyō and competition between
calendarists.
Much of the early eighth-century history of the calendarists in the Bureau of
Onmyō is hard to trace. Instructors of Calendrics were among the lowest ranking officials

Haboush, "Contesting Chinese Time, Nationalizing Temporal Space: Temporal Inscription in Late Chosŏn
Korea," in Time, Temporality, and Imperial Transition: East Asia from Ming to Qing, ed. Lynn A. Struve
(Honolulu, Hawaii: Association for Asian Studies, 2005).

136
in the court bureaucracy—legally, the associated court rank was the Seventh Rank, the
second lowest—and their absence from the early official histories can be explained by
this low status. Even so, there is evidence that the court had periodic difficulties keeping
the Bureau properly staffed.
65
The court eventually established a second, supernumerary
( Jp. gon) Instructor of Calendrics position in response to this problem. Soon, however,
it would be the very availability of a second Instructor of Calendrics that gave rise to
public debate over calendars.
The earliest explicit indication of public need for a second Instructor dates from
833. A note in the official history Shoku Nihon kōki ( , “Continued Later
Chronicles of Japan”) for the sixth day of the twelfth month
66
notes that the official
calendar was presented to the court on this day, over a month late. Earlier that year the
Instructor of Calendrics Tokinotatai Kiyohama had died. According to the official history,
as no one in the Bureau could take over the work, the court called in one Ōkasuga
Yoshimune from Ōmi Province
67
to produce the official calendar.
68
Yoshimune went on
to found a lineage of calendarists in the Bureau that would be prominent and active
through the rest of the ninth and tenth centuries.

65
See Ooms, Imperial Politics and Symbolics in Ancient Japan: The Tenmu Dynasty, 650-800, passim.

66
Which would place the actual date of the note to 834, as Tenchō 10 extended into the next year by the
Julian reckoning. The events described in this account, however, date to earlier in the year and so referring
to the year 833 as remains correct in that respect.

67
Around Lake Biwa, not far from the capital of Heian (Kyoto).

68
Shoku Nihon kōki Tenchō 10/12/ (Jp. boshi).

137
The statement that this outsider was necessary because there had been no
members of the Bureau with the required skill, however, is brought into question by a
record from just three years later. In 836, there is a note at the beginning of the section for
the seventh lunar month in the same chronicle. It states that in the official calendar
originally distributed the previous year, the seventh month had been a “small” month of
29 days. This designation of the length of that month was contradicted, however, by the
calendar in the Seven Luminaries Ephemerides ( Jp. shichiyō-reki)
69
that had
been presented to the court on the first day of the year. In that astronomical calendar the
seventh month was a “large” month of 30 days. The note prefacing the seventh month in
the Continued Later Chronicles of Japan states that the decision was made in 836 to use
the designation of months as determined in the Seven Luminaries Ephemerides. The
difference between the two calendars, the civil one presented in the eleventh month of
835 and the astronomical one presented in the first month of the following year, is
attributed to a difference of opinion between two Instructors of Calendrics.
70
This

69
The shichi-yō reki was a prediction of the position of the sun, moon, and five planets against the
backdrop of the stars of the ecliptic for the coming year, calculated by the calendarists of the Bureau of
Onmyō, and presented to the throne (by legislation in the Engi-shiki), as part of the New Years’ ceremonies
each year. It could be used, at least in theory, to identify astronomical phenomena in the sky (by ruling-in
or out particular planets.) (Yamashita Katsuaki, Personal conversation, October 11, 2010.) It could also be
used, again in theory, to predict potential danger posed by the position of the Five Planets over the coming
year. Such was the purpose of horoscopes devised by sukuyōji (Buddhist astrologer-monks) starting from
the late tenth century. See chapter four for more information on that form of astrology.
The relevant legislation about the shichiyō-reki can be found in an annotated edition in
Shintō taikei hensankai, ed. Engi shiki, 2 vols., Shintō taikei (Tokyo: Shintō taikei hensankai,1991–1993),
and an English translation is available in Felicia Bock, Classical Learning and Taoist Practices in Early
Japan: With a Translation of Books Xvi and Xx of the Engi-Shiki, Occasional Paper (Arizona State
University. Center for Asian Studies) (Tempe, Ariz.: Center for Asian Studies, Arizona State University,
1985).

70
Shoku Nihon kōki Shōwa 3/7
th
month.

138
notation is the first historical hint of the presence of a second Instructor of Calendrics,
appearing even before the first explicit mention of a Supernumerary Instructor of
Calendrics in 882.
71

That a second calendarist was present in the Bureau only three years after
Yoshimune had been appointed to the position, ostensibly because there was “no one
else,” makes the initial record of 833 somewhat suspicious. It is hard to credit the
suggestion that although there was no one in the Bureau with the ability to work on the
calendar in 833, not only had a student had finished his training and been promoted only
three years later, but this individual would have also gained the confidence to challenge
Ōkasuga Yoshimune, by then established in his position. Such a sequence of events is
possible, but unlikely. Yoshimune’s appointment in 833, then, might more accurately be
said to reveal a case of the court meddling within the Bureau by appointing an outsider
over the trainees already there. Court meddling would occur more explicitly in 877,
although over the subsequent centuries the Bureau, by forming a unified front, would
successfully resist the court on issues of appointment, if not on issues of how the calendar
should be produced or on how to predict eclipses.
72

Despite court meddling, the independence of the Bureau grew during the ninth
century. The growing authority of the Bureau during the period after Yoshimune can be

71
Sandai jitsuroku Gangyō 6/1/7. The Supernumerary Instructor was Ōkasuga Ujinushi ( ,
dates unknown.)

72
The ability of the Bureau to resist further meddling on matters of appointment is discussed in
Takada Yoshihito, "Heian jidai kinō kanjin ni okeru kagyō no keishō: Tenmondō o chūshin toshite,"
Kokugakuin zasshi 109, no. 11 (2008), although Takada’s study covers a much latter time period.

139
seen in his descendant Ōkasuga Shinnomaro’s petition for official adoption of the
Xuanming li calendrical system in 862. The text describing the new system had arrived in
Japan a few years previously, courtesy of a merchant from the state of Parhae, north of
the Korean peninsula
73
—but this was not the only or primary difference between the
adoption of the Xuanming li and how other, previous, calendrical systems had been
adopted by the Japanese state. Whereas all previous decrees mandating the adoption of
new calendrical system present the decision as coming directly from the ruling tennō,
Shinnomaro’s petition presents the push for the Xuanming li as his own decision. The
historical record likewise grants credit for this decision to Shinnomaro. Under his own
initiative, Shinnomaro had tested the system over the previous two years, and he then
reported to the court that he found that it more accurately predicted astronomical events.
74

The court and tennō ratified his recommendation.
That the technical experts of the Bureau might, on their own initiative, test and
then recommend the adoption of a new system reveals a great deal of confidence on their
part. It is especially striking in light of the low official status of the staff of the Bureau in
general and of the Instructor of Calendrics in particular. This growing confidence on the
part of the Bureau, however, was not always appreciated by the court, as revealed by the
court’s attitude regarding the Bureau’s “failure” to predict a solar eclipse in 877.

73
Parhae (698–926) was a successor state to Goguryeo north of the Korean Peninsula, and co-existed
with Unified Silla, often uneasily.

74
Ruijū sandai kyaku 17, Jōgan 3/6/16.

140
Nighttime Eclipses and a Failure to Inform
The growth of the Bureau’s authority and independence did not always go
smoothly in the ninth century. Compared to the successful exercising of independence
and initiative in 862 shown in Ōkasuga Shinnomaro’s petition to adopt the Xuanming li
system, the “nighttime eclipse” debate of 877 shows the Bureau in a much weaker
position. In 877 the Heian court effectively denied the Bureau of Onmyō’s ability to
determine, based on its own experience and expertise, what was relevant to calendrics
and what was not. This case reveals the tension between the expertise of the technical
officials within the Bureau of Onmyō, and the authority that the Council of State held to
determine what astronomical facts were true and relevant. It also exemplifies the
Council’s ability to demand a certain type of performance from its specialists in the
Bureau.
By all accounts, news of the Bureau of Onmyō’s “failure” recording the eclipse of
877 broke suddenly. Specifically, according to Nihon sandai jitsuroku ( ,
“True Account of Three Reigns of Japan”), in the third month of 877, the Instructor of
Calendrics, Iehara Kuniyoshi,
75
was called before the court for interrogation regarding
his failure to announce an upcoming solar eclipse. Kuniyoshi both acknowledged that
there had been an eclipse predicted by the Bureau’s calculations of the calendar for that
year and defended the Bureau of Onmyō’s decision not to report it formally to the court.
His argument was that the eclipse was predicted to occur at night. According to

75
Kuniyoshi appears in the text as , missing part of his name. Sandai jitsuroku Gangyō
1/4/1.

141
instructions left by previous Instructors of Calendrics, and preserved in writing on the
walls of the offices of the Bureau of Onmyō, nighttime eclipses were not to be officially
announced. The unstated logic of this position was that an eclipse that would not be
visible to the population of the capital required no countermeasures. Thus it did not need
to be formally acknowledged.
The debate that this statement launched questioned whether or not to officially
prepare for and observe countermeasures for a solar eclipse that would occur at
night—what would be, essentially, a non-observable eclipse. The debate took place
primarily between instructors of the Chinese classics in the Bureau of the University and
scholars of court history, the latter group including some members of the Council of State.
The court concluded that even a solar eclipse that would occur at night (and thus over
some other land)
76
still counted as a dangerous sign, and therefore countermeasures,
including shutting down the court for one day, were required. A solar eclipse, the
victorious side noted, was the obstruction of the astronomical embodiment of Yang (
Jp. yō). As the ruler was the embodiment of Yang on earth, the eclipse indicated dire
events to be contravened at all costs.
What is notable about this case is that although Iehara Kuniyoshi attempted to
bring calendrical expertise and Bureau tradition to his defense, the court roundly rejected
the attempt. The court’s position was that Bureau of Onmyō precedent (particularly when

76
The recognition of a geographical component to the visibility of eclipses is sometimes present in the
Japanese historical record, and sometimes not. When a predicted eclipse is not observed in Kamakura in
1246, the record states that perhaps it was an eclipse “of some other land” ( ). Azuma kagami Kangen
4/1/1.

142
written on the wall of a building in the Bureau compound, as Kuniyoshi revealed this rule
was) was not to be considered a valid or authoritative measure.
77
In the face of this
criticism by the Council of State in 877, members the Bureau realized that in future
attempts to direct the production of calendrical facts they would need to find more
authoritative texts to support their claims, or else to raise the public estimation of the
expertise calendrical experts within the Bureau had acquired through years of experience
calculating the astronomical year.
The first two recorded public disputes over the calendar in the ninth century, in
836 and 877, do not seem to have been particularly fierce. Certainly, little indication of
the intra-Bureau conflict to come was on display. Neither controversy involved the public
reputation of individual calendarists, nor saw them competing with each other. Although
two calendarists were involved in the 836 calendar discrepancy, neither is named; nor are
their arguments presented. The dispute in this case might not have even have taken place
before the Council of State—the record merely explains for future readers potential
discrepancies between dates for that calendar year. In 877, the two sides in the debate do
not both come from the Bureau of Onmyō. Instead, the Bureau, represented by Iehara
Kuniyoshi, was forced to defend its position against criticism by the court.
By the tenth century, however, the personalities and personal positions of
individual calendarists would appear before court in a new way. The ability of
calendarists to determine their own practice and argue for their own interpretations would
be enhanced, not through professionalization and independence from court interference,

77
Sandai jitsuroku Gangyō 1/4/1.

143
but by utilizing personal relationships with powerful members of the Council of State.
Such alliances would transform the way that the court thought of solar eclipses.

The Ōkasuga and Katsuragi Lineages, and Court Factionalism
The rise of patron-client alliances between technical experts in calendrics and
members of the Council of State developed, perhaps ironically, from growing dissent
among individual calendarists. In the tenth century, the question of nighttime solar
eclipses emerged once again at court, this time in a dispute that occurred between two
Instructors of Calendrics from the Bureau of Onmyō. According to the Nihon kiryaku ( , “Abbreviated Chronicles of Japan”), two Instructors of Calendrics publicly
debated an eclipse prediction that had been made for the first day of the new year 918
(Engi 18).
78
The debate had taken place between Ōkasuga Hironori
79
(of the same
lineage as Yoshimune and Shinnomaro) and Katsuragi Munekimi
80
on the preceding
twenty-seventh day of the twelfth month. While this may not have been the first debate to
take place between the two men, it is the first time such a debate shows up in the
historical record. This suggests that while there might have been simmering tensions
within the Bureau of Onmyō, these had now spilled over and the issue could no longer be
resolved in-house. It is not surprising that it is not until the tenth century that the first
attested debates between calendarists appear: as all communication from Instructors

78
February 14, 918.

79
, dates unknown.

80
, dates unknown.

144
within the Bureau to the court were to go through the Head of the Bureau of Onmyō,
there was an institutional barrier to such disputes reaching the public.
81

In 918 the Council of State made the decision to support the position of
Katsuragi Munekimi, who insisted that there would be a solar eclipse—in contrast to
Hironori, who argued that one would not be visible (echoing, perhaps, Kuniyoshi in the
previous century). This resulted in the shutting down of court and the cessation of court
ceremony on the ritually significant first day of the year.
82

Later that same year, however, the question of how to handle a solar eclipse
predicted for the first day of the new year arose once more, this time for the following
year. Now it was Katsuragi Munekimi who argued against closing down court in
response to the eclipse. Supported by scholars who specialized in classic Chinese texts,
his argument was that as the eclipse would take place at night, there was no need to take
the propitiatory step of canceling the court’s ceremonies. The position of Ōkasuga
Hironori, Munekimi’s opponent, is less clear from the sources. Nevertheless, once again
Munekimi’s position was affirmed: an official decree was issued endorsing Munekimi’s
arguments and establishing the precedent from 919 onwards that the Japanese court

81
The procedure is described in the Engi shiki ( , “Procedures of the Engi Era) compiled in 921,
and so would have presumably been in effect during this dispute of 918.

82
Nihon kiryaku Engi 17/12/27 and 18/1/1. Note that the Teishin-kō ki shō only mentions the 18/1/1
shutdown in observance of the eclipse prediction, and not the debate itself.
It may be possible that the sources for the Engi 18 and Engi 19 debates were garbled in the
Nihon kiryaku compilation process and that there was no debate in 918. Against such a position, however,
is the fact that the sources describe Katsuragi Munekimi has having held different opinions regarding the
visibility of the eclipse of 918 and of 919. As a result, it seems best to treat the Nihon kiryaku entries as
reliable, and consider the lack of a record of the debate in Teishin-kō ki shō as having resulted from the
process of editing and abstraction that particular source went through.

145
would not officially observe nighttime solar eclipses as astrologically or ritually
significant.
83

The difference between the 877 and late 918 debates is dramatic. In 877, the
Bureau of Onmyō presented a unified front supporting the position that solar eclipses
predicted to take place at night should not be officially recognized.
84
On the other side of
the issue, arguing against the Bureau, were the assembled scholars of Chinese classics.
By contrast, in 918 the debate featured two Instructors of Calendrics taking opposing
sides in public, in contrast to 877. This time, also in contrast, the cited scholars of
Chinese classics supported the position of ignoring nighttime solar eclipses. The Council
of State was also divided in 918, with members presenting arguments both for and
against shutting down the court in response to an invisible eclipse. No one organ of the
court bureaucracy was unified in the 918 debate, in stark contrast to the 877 situation.
Whether or not the Bureau of Onmyō was indeed unified in 877 therefore
should be examined more closely. It may seem as if there must have been some
intra-Bureau conflict in the case of the 877 “nighttime eclipse” debate. As the prediction

83
Teishin-kō ki shō Engi 18/12/28. This decree was also copied in the court ritual manual Saikyūki (
). Why Hironori lost the debate to Munekimi is not fully clear, but the powerful Minister of the Left,
Fujiwara Tadahira ( , 880 – 949), in his diary entry as preserved in Teishin-kō ki shō faulted
Hironori for relying on “drafts” from scholars, and not referring to the official texts:

84
The question of whether such eclipses should be counted for calculation purposes would appear in
subsequent debates.
“Nighttime eclipses” are eclipses that, according to the prediction, would occur at night. In
modern astronomical terms, they can be said to be eclipses predicted to occur over a different part of the
globe where it was currently day. The concept of a spherical earth with one half illuminated at any time is
nowhere to be found in Chinese or Japanese astronomers. Even so, the possibility that such eclipses might
occur in a different part of the world when they were invisible in one location can be found in Japanese
sources. This uncertainty was one reason why the question of “nighttime eclipses” was so difficult to
resolve.

146
of eclipses is thought to have been beyond the skill of most members of the Heian court,
the Council of State would have presumably needed to receive some inside information
about the fact that there was an eclipse prediction that had not been made public. The fact
that the Bureau had failed to make a public announcement of the eclipse prediction had
been the entire point of the interrogation of Iehara Kuniyoshi. However, no other
calendarist or member of the Bureau of Onmyō is credited in the sources with having
brought the issue to the Council of State’s attention. Instead it seems more likely that this
information came from the calendar itself, as eclipse predictions were indeed recorded on
the official almanacs presented to the court on the first day of the eleventh month each
year for the upcoming year. Eclipse predictions on the official almanacs were, moreover,
not the final word on eclipses as far as the court was concerned. Tenth century court
procedure mandated that the eclipse prediction be officially presented to the court once
more eight days before the eclipse was to occur.
85
Only then would the court be shut
down. It was this second prediction that the Council of State was demanding, despite the
objections of the Bureau of Onmyō. Therefore, there is no need to assume dissention
within the ranks of the Bureau in 877, although dissent may have been hidden there.
The debate in 918 was significantly different. Whereas in 877 the Council of
State was largely unified against the Bureau of Onmyō, in 918 the two Instructors of
Calendrics and the Council of State were twice found on opposing sides. Exactly how
these sides corresponded to political factions at court is unclear, but is worth considering.

85
Therefore it can be said that by the tenth century that the eclipse predicted the year before in the official
almanacs were no longer considered authoritative predictions, as they all required reaffirmation closer to
the predicted day of the eclipse. The procedure described can be found in Engi shiki 20.

147
There were good reasons why some members of the Council of State would have
preferred not to consider the 919 new year’s eclipse a real eclipse if the chance to
discount it arose—while others may have been better served by using the eclipse as a
political statement. Certainly, members of the Council of State held personal opinions on
the matter. Although Fujiwara Tadahira,
86
the Minister of the Left and a ranking member
of the Council of State in 918, does not seem to have joined in the dispute, indications
from his diary indicate that he was more sympathetic to the position in support of
ignoring the eclipse and holding the new years’ ceremonies at court.
87

Tadahira’s position in favor of holding ceremonies in the event of an ominous, but
likely non-visible, eclipse is understandable, particularly in light of the symbolic and
apotropaic roles such new-year ceremonies had. For example, the new year began with a
pre-dawn ritual dedicated to the Four Directions and the stars of the Big Dipper, rituals
that stem from the same source as rituals found in Chinese religious Daoism.
88
This
ritual was performed by the tennō or a close proxy. It was followed by court-wide
ceremonies that reaffirmed court hierarchies and the loyalty of the officials to the tennō.
Abandoning such performances two years in a row as the result of inconvenient invisible
astronomical events may have been hard for Tadahira to bear. Nevertheless, a balance
needed to be struck between observing the necessary court ceremonies and protecting the

86
, 880 – 949.

87
Judging by his dismissal of Ōkasuga Hironori’s argument. Teishin-kō ki shō Engi 18/12/28.

88
Particularly Lingbao Daoism, in rituals to protect the adept entering the mountains.

148
health of the tennō from the potential malignant effects of exposure to a solar eclipse.
89

This meant that Tadahira needed to be convinced of the non-visibility of the eclipse. The
success of Tadahira’s career depended, after all, on both the health of the ruling tennō, his
relative through the tennō’s maternal line, and the smooth reaffirmation of court structure
and loyalties. In this case, Tadahira decided in favor of the innocuousness of invisible,
nighttime eclipses; and in the end the Council of State sided with him.
In the records of the debates of 918, each side in these debates was associated
with an individual Instructors of Calendrics. It is likely that each Instructor had a link
with a political faction at court. Katsuragi Munekimi seems to have had one such
connection with Fujiwara no Tadahira, at least in late 918, judging by the record in
Tadahira’s court diary. It was just such a social link between expert and high courtier that
would be exploited by Katsuragi Munekimi’s successor, Katsuragi Shigetsune,
90
as he
continued his predecessor’s opposition against Ōkasuga Hironori.

Katsuragi Shigetsune and the Implications of Invisible Eclipses
The career of Katsuragi Shigetsune illustrates the importance of successful
eclipse prediction on burnishing a personal reputation and building a (at least
temporarily) successful career as a calendarist in the mid-Heian court. Shigetsune first

89
That exposure was the primary concern can be seen in the accounts of successful esoteric Buddhist
anti-eclipse rituals, which call in clouds. See the discussion of the example in the Asabashō ritual manual in
chapter five.

90
, dates unknown

149
appears as the junior Instructor of Calendrics in 936.
91
Earlier, Ōkasuga Hironori had
been the junior Instructor to Munekimi’s senior Instructor—but by 936, Hironori had
been promoted to the senior position. It is the same year that Shigetsune first appears in
the historical record, assuming the (now-vacant) junior post.
Katsuragi Shigetsune’s first appearance, however, is not as that of a compliant
subordinate to his senior, Hironori. According to the Abbreviated Chronicles of Japan, on
the eleventh day of the tenth month of Jōhei 6 (936), Shigetsune petitioned the Council of
State for the issuing of a decree ( Jp. kanpu) ordering that Ōkasuga Hironori’s
“erroneous” ( Jp. ayamari) calendar for the next year be put aside. In its place,
Shigetsune wanted his own draft of the calendar adopted. While 936 saw the first time
Shigetsune would directly petition the Council of State against the higher-ranking
Hironori, it would not be the last —Shigetsune would continue to present his opinions
directly higher ranking members of the court bureaucracy, bypassing the Bureau of

91
The gon or Supernumerary position of Instructor of Calendrics can be seen as junior to the codal
Instructor of Calendrics position. That gon was seen as junior to shō ( ) positions can be seen in
descriptions of careers in the Council of State found in the Kugyō bunin compendium—there, gon
dainagon are described as being promoted to dainagon as the next step in their careers.
(Kokushi taikei hensankai, ed. Kokushi bunin, 6 vols., Kokushi taikei (Tokyo: Yoshikawa kōbunkan,
1977–1979).)
It is unclear, given that the Katsuragi lineage was apparently no significant enough to be
included in later genealogical compilations, exactly what Shigetsune’s relationship to Katsuragi Munekimi
was—even if a genetic relationship can be assumed. His appearance after Munekimi vanishes from the
historical record, however, implies that he may have been Munekimi’s heir, a son or some other younger
relative, particularly in his appearance as an Instructor of Calendrics in the Bureau. As shown by the
repeated appearance of Ōkasuga lineage members in the position, the association of certain lineages with
expert positions in the Bureau had already been established. (And, in fact, can be traced back to the very
establishment of the Bureau—see two books by Michael Como on immigrant lineages: Como, Shōtoku:
Ethnicity, Ritual, and Violence in the Japanese Buddhist Tradition, and ———, Weaving and Binding:
Immigrant Gods and Female Immortals in Ancient Japan. Interestingly for Como’s argument for earlier
ages, neither the Ōkasuga nor Katsuragi had a continental origin, for both claimed local native ancestry.
Shintō taikei hensankai, ed. Shinsen shōjiroku, Shintō taikei (Tokyo: Shintō taikei hensankai,1981).)

150
Onmyō hierarchy multiple times over the next few years. It appears that his methods
were not much appreciated, at least not by all of Shigetsune’s fellow Bureau members.
There may well have been cause for Shigetsune to go outside of the Bureau
with his concerns. Later chronicles would remember 936 as the year of the “Calendarists’
Error” ( Jp. rekike no shitsu), when the winter solstice fell on the second day of
the eleventh civil month, rather than on the first.
92
As 936 was the first year of a Rule
Cycle,
93
this could be seen as a major violation of the system—despite the fact that the
cycle was no longer one that was included in the mathematics of the system of
calendrical astronomy.
94
At this time, regular observance of the Rule Cycle by the
calendrists of the Bureau was over 130 years old. From one standpoint, something
untoward was clearly going on within the Bureau—even if this something might have
been as well-intentioned as merely letting the results of calculations stand as the
instructions in the calendrical canon dictated.
95
Nevertheless, if Shigetsune had spoken
out against the Rule Cycle lapse in 936, the Bureau had not heeded his protest.
96

92
As, for example, in the record of this year in Nihon kiryaku. Jōhei 6/11/1 and 6/11/2. The Nihon
kiryaku entries, as the compilation of this chronicle is thought to date from the mid- to late-eleventh
century, is one of the earlier examples of such a characterization of the “deviation” from the Rule Cycle.
See, for example, the Hyakurenshō entry for Eishō 5/9/28 (1050).

93
The ninth Rule Cycle since its formal adoption as a feature of court ceremonial by Kanmu Tennō in
794.

94
See the section “The Difficulties of the Chinese-Style Year” above.

95
936 pre-dates most of the identifiable adjustments to the Xuanming li system made by Japanese
calendarists. (See Uchida Masao, Nihon rekijitsu genten (Tokyo: Yūzankaku shuppan, 1992) for a list of
known adjustements) Shigetsune would initiate one such adjustment in 938.

96
Shigetsune might have just as easily been on the other side of the issue—the point here is that there was
dissatisfaction with the Bureau of Onmyō’s calendarists at court, even though such dissatisfaction took
151
Almost exactly one year later, on the second day of the tenth month of Jōhei 7
(937), another dispute between Hironori and Shigetsune broke out in public. According to
the Abbreviated Chronicles of Japan, both calendarists were called before the Minister of
the Right and questioned about the calendar for the following year (Jōhei 8, later Tengyō
1, 938). The details of the debate or their positions are not recorded—what is clear in the
record, however, is that they did not agree.
97
In response to the debate, the court charged
the Dazaifu ( ) offices in Kyūshū with forwarding Chinese calendars for the
current and upcoming years, presumably obtained from Chinese traders who were legally
required to disembark there.
98
Unfortunately how the court used the calendars with
regards to the dispute is unknown: the arrival of the calendars from the Dazaifu in 937 is
the last historical record of the controversy.
What the presence of the Chinese calendars in the court case over the calendar
for 938 makes clear is that although the Japanese were well on their way towards
developing an independent tradition, one that differed in some ways from the practice of
calendrics as it was performed in China, Chinese performance still remained authoritative.
The facts of the calendrical year as had been determined by the Chinese state on the
continent were still considered relevant for settling matters of dispute between Japanese

years before it was explicitly recorded in the sources. At the very least, Shigetsune seems to have taken
advantage of this discontent.

97
Nihon kiryaku Jōhei 7/10/2. Even though the names of the two Instructors are not given in this entry,
their identities are clear from the temporal context.

98
Nihon kiryaku Jōhei 7/10/13.
For information on the role of the Dazaifu for the Heian state, the best source is Bruce Batten,
Gateway to Japan: Hakata in War and Peace, 500-1300 (Honolulu: University of Hawai'i Press, 2006).

152
calendarists. This was despite the fact that the Chinese states in 938 were utilizing a
different system of calendrical astronomy than the Xuanming li still current in Japan.
99

In the following year (938), Shigetsune once again brought a protest against
Ōkasuga Hironori’s proposed calendar before the Council of State. Even though that year
had the court dealing with other pressing crises such as the rebellion of Taira
Masakado
100
in the east and Fujiwara Sumitomo’s
101
pirate raids in the south,
Shigetsune found a receptive audience for his arguments. While the Abbreviated
Chronicles of Japan does not provide details about Shigetsune’s or Hironori’s arguments,
the twelfth-century Honchō seiki chronicle ( , “Chronicle of the Ages of Our
Realm”) does—it also provides information as to how this debate played out at court.
According to the Hochō seiki text, six days after Shigetsune’s petition, on the seventeenth
day of the tenth month, both Shigetsune and Hironori were summoned before the Council
of State to testify before it.
102

99
The year 938 was well into the period of the Five Dynasties and Ten Kingdoms, and so there were a
number of states on the Chinese continent that the Dazaifu might have received its calendar from. The state
of Wuyue ( Jp. Goetsu) in particular is listed as a destination for Japanese monks traveling to China in
the middle of the tenth century. See the account (perhaps legendary) of Nichien in the next chapter.

100
, d. 940. For a good account of Masakado’s life and rebellion, see Karl Friday, The First
Samurai: The Life and Legend of the Warrior Rebel Taira Masakado (Hoboken, NJ: John Wiley & Sons,
2008).

101
, d. 941.

102
The following discussion is based on the entry in Honchō seiki for Jōhei 6/10/17.

153
Hironori spoke first, stating that his calculations showed that the length of the
luni-solar year that began in 939 was to be 383 days.
103
The first month would be short,
and the months should have a long-short alternation thereafter. The intercalary month
would be set after the seventh month. Shigetsune, on the other hand, stated that he had
determined the length of the year as one day longer, at 384 days. This meant that
Shigetsune was stating that the new moon for the first month would occur one day earlier
than Hironori had determined. Shigetsune had his months starting as long in the first
month, with a short-long alternation thereafter. Like Hironori, Shigetsune had determined
that the intercalary month should follow the seventh month.
The debate seems, on the surface, to have primarily concerned the length of the
months. On further questioning, however, it was revealed that Shigetsune and Hironori
were also disagreeing on the issue of on what day should the current year end. Because
the lengths of luni-solar years varied from year to year, time was also marked with a
sexagenary day-cycle
104
—as the sexagenary cycle was always constant, unlike the rest of
the calendar, it provided an important reference through which days could be counted.
105

103
Again, the luni-solar year in both Japan and China typically began in mid-January or in February by
the Western reckoning.

104
This cycle, which could also be applied to years, is often referred to as “the Chinese zodiac.” It
consisted of a combination of 10 “heavenly stems” ( Ch. gan, Jp. kan) and 12 “earthly branches” ( Ch.
zhi, Jp. shi). The “heavenly stems” were each associated with one of the Five Phases, and the twelve
“earthly branches” with an animal. The combination of each of the phases with an animal resulted in 60
combinations.
The original reason for a 60-day cycle as opposed to a 120-day cycle was that the 10-day and
12-day cycles were originally independent, but run at the same time. Mathematically, the cycle then
became not combinatory, but instead of the Least Common Multiple for both 10 and 12, or 60. On the early
history of the sexagenary cycle, see Wang, Cosmology and Political Culture in Early China.

105
The “conjunction” between the sexagenary cycles and the year and month cycles—when the first day
of a month would reoccur with the same sexagenary value, was one of the cosmological points of return in
154
Shigetsune assigned one day, the fortieth of the sexagenary cycle,
106
to the last day of the
twelfth month of Tengyō 1. In contrast, Hironori assigned the same day as the first of
Tengyō 2.
The critical difference between Shigetsune and Hironori’s calculations arose
from an eclipse prediction. The difference between them involved the status of eclipses
that were predicted to occur at night. When an eclipse appeared in the calculations
determining the months and length of a year, the Xuanming li system that was in use in
Japan called for an “advancement-retreat” ( Ch. jintui, Jp. shintai) adjustment. This
was a method of adjusting how the moment of conjunction between the sun and moon
(the moment of the new moon) was included in the calendar. In months without solar
eclipses, this conjunction was determined using averaged values of the length of the
lunation.
107
The exact moment of the new moon, after all, was nearly impossible to
observe.
108
However, such was not the case when a solar eclipse occurred—as a solar
eclipse only occurs at conjunction, it could be used to determine the precise time when
the new moon occurred. This was the logic behind the “advancement-retreat” adjustment,

the Chinese calendrical system. See Sivin, "Cosmos and Computation in Early Chinese Mathematical
Astronomy."
The sexagenary value for days was not merely useful for mathematical placeholding, however.
Each sexagenary position, or each stem or branch position, also provided information for temporal
divination (hemerology) useful for making decisions on what activities would be good or bad for each day.
See Michael Strickmann, Chinese Magical Medicine (Stanford, Calif.: Stanford University Press, 2002),
69-70.

106
Ch. guimao, Jp. mizu-no-to u or kibō.

107
The Xuanming li system also used averaged values for solar periods, the twelve “solar months” (
Ch. zhongqi, Jp. chūki) that divided the year. This will be discussed in more detail below.

108
As this would be the moment at which the moon was in line with the sun—thereby at daytime, and
only visible when the moon would obscure the sun, as in the case of a solar eclipse.

155
which used predicted eclipses to determine a more accurate moment of
conjunction—although, of course, it presumed that the eclipse predictions thus used were
accurate. When a solar eclipse was predicted using calendrical calculations, this
prediction could be used to adjust the calendar to reflect the actual moment of
conjunction. The result was a more precise value for the timing of the new moon that
could then be used in calculating the months of the year than would normally result from
calculations based on the averaged length of a lunation. If the eclipse occurred late in the
day, the month was “rounded up” and begun instead on the following that. This was
intended to better keep the apparent phase of the moon and the day of the month in
accordance.
Much of the Xuanming li system, like other calendrical systems of its era, was
calculated using average values for astronomical periods in the place of precisely
determined astronomical periods. This was largely a matter of utility, as incorporating
either the changes in the speed of the sun’s apparent motion across the sky, or the even
more complex variability found in lunar periods, would have made the mathematics of
the system more unwieldy.
109
Therefore the months and years of the civil calendar were
calculated using average values for lunations and solar months ( Jp. chūsetsu-ki).
Solar eclipse predictions, however, were calculated using a different set of cycles, one
that included values for solar months ( Ch. zhengzhong, Jp. shōchū) that more

109
The variation in the sun’s apparent speed was incorporated in the Shoushi li ( Jp. Juji-reki) of
the Yuan Dynasty (Sivin, Granting the Seasons: The Chinese Astronomical Reform of 1280, with a Study of
Its Many Dimensions and a Translation of Its Records: ), as well as in Shibukawa Shunkai’s
1684 Jōkyō-reki ( ).

156
accurately reflected the variation in the sun’s apparent speed.
110
This meant that eclipse
predictions could also be calculated separately from the rest of the calendar.
111
That they
could be calculated separately is evidenced in the few drafts of calendrical and eclipse
calculations that survive to the current day.
112
This did not mean, however, that eclipse
predictions had no effect on the calculation of the civil year. Through the system of
“advancement-retreat” adjustment, a predicted solar eclipse would either push forward or
pull back the start of the month to better correspond with the precise moment of the new
moon.
Katsuragi Shigetsune and Ōkasuga Hironori differed on what the rules about
eclipses predicted to occur during the night meant for this type of adjustment. In terms of
court procedure, it has been decided through the precedent of the new year of 919
(decided in 918) that eclipses predicted to occur at night would not be observed in court
protocol. Shigetsune was taking this principle and extending it, applying it to calculations

110
Sivin, "Cosmos and Computation in Early Chinese Mathematical Astronomy." The averaged values
were known as (Ch. dingzhong, Jp. teichū).

111
This will be a factor in later disputes involving solar eclipses, covered in chapters five and six.

112
Kan’ei nen-kan reki suisan kōtō ( , “Calendrical Drafts from the Kan’ei Era,
1624-1645”), a collection of fragments and rough drafts of calendars from the first half of the seventeenth
century. Preserved in the Yōmei Bunko archive (Konoe family). Unknown, "Kan'ei nenkan reki suisankōtō
."
Unfortunately drafts of calendars from earlier periods do not survive, most likely due to the
value paper had for re-use. That these drafts survive at all from the seventeenth century attests to the
growing interest on the part of courtiers in the early modern period in such topics as how calendars were
made. These drafts were likely originally preserved by a calendrist lineage for study and teaching—how
they became incorporated into the Konoe family archive is unclear.
Although from a much later period than the years under discussion here, these
seventeenth-century drafts utilized the same basic Xuanming li system as was used in the tenth century, and
so can provide clues as to calendrical practice in earlier periods.

157
of the “correct” position of astronomical bodies. That which did not count for ceremony
was not also to count for mathematics
Hironori took what appears to be the more conservative position: that even if a
solar eclipse was calculated to occur at night, the rule of adjustment still applied.
Shigetsune, on the other hand, was reconciling court procedure and astronomical models.
He could also bolster his argument by relying on court documents from the 918 debate.
113

These documents cited by Shigetsune, which stated that nighttime solar eclipses could
not be confirmed by observation ( Jp. fuken), helped him carry the day. A decision
was made to issue an official decree in favor of Shigetsune’s calendar, and it was issued
on the nineteenth of the tenth month of 938.
The official decision in favor of Shigetsune did not resolve the issue, however.
On the day of the official presentation of the calendar for the next year, the first day of
the eleventh month of 938, the Bureau of Onmyō failed to carry out its part in the
performance. The minister in charge of the ceremony, one Taira Koremochi,
114
had
arrived as scheduled and began the proceedings. But when called upon to bring forth the
calendar, a junior clerk in the Bureau of Onmyō
115
reported that the two calendars had
still not yet been brought into agreement. As no calendar that reconciled the two had been
completed, he said, it was impossible to present the official calendar. In other words,

113
Shigetsune’s arguments also included the Bureau of Onmyō precedent of not announce eclipses that
dated from 836, even though that precedent had been rejected when Iehara Kuniyoshi had attempted to use
it in the dispute of 877.

114
, dates unknown.

115
Nakatomi Yoshimasa ( ), dates unknown.

158
within the Bureau, and despite an official court order, Hironori and Shigetsune’s debate
had continued.
The Bureau insisted that they were not ignoring the order: instead, they
defended themselves by saying that the time that had elapsed between the official decree
and the day of presentation had not been enough to acquire the necessary materials for
the production of the formal, official almanac. In particular, they noted the difficulty of
acquiring the vermillion ink used for annotations. Koremochi sent his report of the matter
to the regent Fujiwara Tadahira for a decision, even as he noted that Tadahira was
indisposed that day and could not settle the matter swiftly. The Bureau also attached their
statement to Koremochi’s report.
116

Two days later, Koremochi and the other court officials who had been present
at the failed presentation reported to Tadahira and the Council of State. The Council was
not sympathetic to the Bureau’s excuses, given that since the nineteenth day of the
previous month, when the order to use Shigetsune’s draft had gone out, eleven days had
elapsed in which the Bureau could have completed the drafting and copying of the
calendar.
117
It was hard not to see the plaint of insufficient time as an excuse, since four
days later the clerks who had protested they could not present the calendar were
identified as those very men within the Bureau who had received Shigetsune’s draft with

116
Honchō seiki Jōhei 6/11/1.

117
Honchō seiki Jōhei 6/11/3.
This incident gives a good sense of how much time the Council of State thought it would take
to create a formal copy of the calendar for presentation. It is a far shorter time frame than is found in the
schedule of work for calendar production found in book 20 of the Engi shiki. For an English translation of
the relevant section, see Bock, Classical Learning and Taoist Practices in Early Japan: With a Translation
of Books Xvi and Xx of the Engi-Shiki.

159
instructions to adopt it.
118
If these men were obstructing the distribution of Shigetsune’s
calendar, then it seems that Hironori had more allies in the Bureau, or Shigetsune fewer
friends, than initially might be supposed. The technical dispute between two calendarists
may have concealed struggles between factions within the Bureau that fell beneath the
notice of the historical sources.
Two additional days later, the Bureau of Onmyō was made to issue an official
apology for their “wrongdoing.” The sense was that further obstructions to the
promulgation of Shigetsune’s calendar would not be tolerated.
119
The official
presentation of the calendar was duly rescheduled for the first day of the twelfth
month,
120
and the administrative staff of the Bureau who had been accused of obstructing
the distribution of Shigetsune’s calendar was forced to attend. The record of the
presentation in the chronicle carries with it a note reminding the reader of the
unforgivable dereliction of duty that had necessitated such rescheduling. Katsuragi
Shigetsune of course, attended his delayed moment of triumph. But Ōkasuga Hironori did
not.
121

118
Specifically and notably Nakatomi Yoshimasa. Honchō seiki Jōhei 6/11/7.

119
Honchō seiki Jōhei 6/11/9.

120
Following a precedent set in Tenchō 10 (833), when the sudden death of Tokinoatai Kiyohama
motivated the court to call in Ōkasuga Yoshimune at the last minute. This case is discussed earlier in this
chapter

121
Honchō seiki Jōhei 6/12/1.

160
Katsuragi Shigetsune built on this success in further petitions to the court that
year and the next. He presented his personal prediction of a lunar eclipse for the fifteenth
day of the twelfth month of Tengyō 1 not long after his calendar was forcibly accepted by
the court.
122
The record of this prediction—subsequently confirmed by observations of
an actual lunar eclipse—is described as having been in an individual report made by
Shigetsune to the Council of State ( Jp. kanshin). It was thus not an official
prediction made by the Bureau of Onmyō through the Ministry of Palace Affairs, which
was the standard bureaucratic procedure for announcing an eclipse prediction.
123

In his report, Shigetsune describes the subtle and mysterious skill necessary to
produce an eclipse prediction. When the actual lunar eclipse was observed just as
Shigetsune had predicted, down to the time and magnitude of the eclipse, the account
records the Council of State praising him as a proven expert of his field.
124
That
Shigetsune made this prediction both publicly and as an individual, not as a member of
the Bureau of Onmyō through standard procedure, speaks of continuing tensions between
himself and other members of the Bureau, as well as of his strategy of utilizing direct
lines of communication with the Council of State.
The following year yet another dispute between Shigetsune and the Bureau
surfaced. On the last day of the sixth month of Tengyō 2 (939), Shigetsune’s predictions

122
Or January 7, 939 in the Julian calendar. While Tengyō 1 had begun on February 2, 938, it would end
the following year (by Julian reckoning), on January 22, 939.

123
See again the procedure set out in Engi shiki 20.

124
Honchō seiki Tengyō 1/12/15.

161
and the “laziness” of other members of the Bureau of Onmyō were once again court
business. According to the account, Shigetsune had earlier presented a prediction of a
solar eclipse for the first day of the seventh month of Tengyō 2.
125
Although “days had
passed,” still the Bureau of Onmyō persisted in not presenting an official announcement
of the upcoming eclipse through the Ministry of Palace Affairs. This account clearly
shows that Shigetsune had once gone over the heads of the senior members of the Bureau
of Onmyō to have his case heard directly by members of the Council of State.
126

Shigetsune was heard, and believed by the Council, despite the fact that there were
clearly members of the Bureau that vehemently did not agree with his position. Once
again grumbling about the Bureau’s dereliction of duty, the regent Tadahira declared that
the court should shut down on the day of the predicted eclipse.
127

At this point, however, things began to go wrong for Shigetsune. The
Abbreviated Chronicles of Japan reports that the solar eclipse did not occur at the
predicted time. There was doubt whether it occurred at all, the Abbreviated Chronicles
reports: “Some persons said there was no eclipse.”
128

According to modern calculations this eclipse, the eclipse of July 19, 939 was
indeed visible in Japan as a partial eclipse. However, it began at 6:23 PM local time and

125
Or July 19, 939 in the Julian calendar.

126
At this time, Katsuragi Shigetsune was still the junior Supernumerary Instructor of Calendrics (
Jp. gon reki no hakase). Ōkasuga Hironori outranked him.

127
Honchō seiki Tengyō 2/6/29.

128
Nihon kiryaku Tengyō 2/7/1.

162
ended at 7:02 PM, when the sun set.
129
As a partial eclipse, with only 36% of the sun
occluded at sunset and occurring low in the sky, it is difficult to say for certain whether
anyone successfully observed this eclipse or not. The Abbreviated Chronicles of Japan
seems to indicate that some of those who looked did not sight any eclipse. This chronicle
was compiled largely from the records of the Palace Secretariat or the Secretariat of the
Council of State.
130
The observation and divination reports of tenmon specialists were,
by law, to have copies archived with the Palace Secretariat for future use.
131
The tenmon
specialists, likewise, were charged with observing predicted eclipses. If the failure to
sight the eclipse came from the Palace Secretariat records, then that indicates that it may
have been a member of the Bureau of Onmyō itself, one of the Instructors of Tenmon,
who noted the eclipse’s failure to occur. This would not be surprising, given the previous
evidence of discord within the Bureau as far as Shigetsune’s arguments were concerned.
The eclipse of 939 would have been harder to confirm visually than an eclipse that
occurred when the sun was high in the sky. Therefore it is not even necessary to assume
foul play on the part of observers within the Bureau to find the root of “Some persons
said there was no eclipse.” Anti-Shigetsune sentiment might have been enough to color

129
Information on the eclipse was taken from NASA, "Nasa Eclipse Web Site,"
http://eclipse.gsfc.nasa.gov.

130
See Hosoi Hiroshi, Kodai no tenmon ihen to shisho. Hosoi emphasizes the Naiki ( , Palace
Secretariat) archives over the Geki ( , Secretariat of the Council of State) in the compilation process,
although his conclusion and reasoning has been disputed (Yamashita Katsuaki, "Shohyō Hosoi Hiroshi cho
"Kodai no tenmon ihen to shisho"," Nihonshi kenkyū, no. 555 (2008).).

131
Ritsuryō, Zatsu ryō. See also chapter one.

163
observers’ perceptions of a solar eclipse that would have been difficult to discern even for
the most impartial observer.
It is therefore possible that Shigetsune’s opponents within the Bureau of Onmyō
finally found an audience for their position. Although modern astronomical calculations
indicate that Shigetsune’s prediction of an eclipse was more accurate than the Bureau’s
insistence on its lack, the position against Shigetsune was the one that found its way into
the historical record. Shigetsune’s opponents finally found an audience once the
Abbreviated Chronicles of Japan was compiled.
It is difficult to determine what Tadahira himself might have thought—his
surviving court diary is silent on this issue.
132
Tadahira seems to have placed a certain
amount of trust in Shigetsune’s predictions, judging both by the praise heaped upon
Shigetsune by the Council of State in early 939,
133
and by Tadahira’s swift reaction to
Shigetsune’s prediction of the solar eclipse in the summer of that year. As for Shigetsune,
he disappears from the historical record thereafter. His triumphs of 938 and early 939,
along with his notoriety at court, appear to have been short-lived.
This was not merely a quiet return to duty: historical record makes it clear, in fact,
that Shigetsune’s career did not last much longer. In Tengyō 4 (941), Shigetsune’s
long-time adversary Ōkasuga Hironori petitioned that Student of Calendrics ( Jp.
reki-shō) Kamo Yasunori be allowed to participate in the production of the next year’s

132
There is no surviving record from his diary on this matter, despite surviving entries from this period.
See Teishin-kō ki shō for Tengyō 2.

133
In response to his lunar eclipse prediction. See the previous discussion.

164
calendar. Despite the lack of any prior public attestation to Kamo Yasunori’s career or
experience, permission was granted.
134
Not only does this 941 event mark the beginning
of the Kamo lineage’s participation in the production of the official calendar,
135
it also
marks the end of any trace of the Katsuragi lineage’s presence at court.
136
Perhaps
Shigetsune was as combative as he was because he saw his lineage being pushed out of
the Bureau of Onmyō, or because he was the last member of a previously successful line.
In any event, Katsuragi Shigetsune’s career marks the beginning of a new stage in
the relationship between members of the Bureau of Onmyō and the court. Shigetsune’s
strategy of taking his arguments to powerful members of the court bureaucracy directly
marks the full appearance of patronage politics in the business of calendrics at court. This
strategy would be used repeatedly by other Instructors of Calendrics in the future. Direct
communication with high-ranking members of court would also become a means through
which men with an interest or skill in calendrics, but who were outside of the Bureau of
Onmyō, could participate in the official calendrical astronomy project. Personal
relationships with high-ranking members smoothed the way, but so did demonstrations of
skill that established personal reputations for expertise and technical ability. Personal
connections and individual reputations would play a major role in court decision-making

134
Fusenshō Tengyō 4/7/17. Found in Dai Nihon shiryō vol. 1,part 7, 936.

135
As noted in the introduction, the Kamo lineage is generally the only lineage associated with pre-Edo
calendrical astronomy in Japan. As this chapter has shown, that association is historically dubious,
particularly for the early Heian Period.

136
The reasons for Shigetsune’s disappearance from the historical record are unknown, and may have had
as much to do with mortality as any potential disgrace. The timing, however, is striking; as is the
disappearance of any attestation of Katsuragi lineage participation in court life.

165
when debates over eclipse predictions became more frequent in the eleventh and twelfth
centuries. In this respect, Shigetsune was a trailblazer.
Shigetsune also left a legacy within the Japanese tradition of calendrical
calculation. His victory in the debate of 938 established a precedent that future calendrists
were forced to observe: the removal of “advancement-retreat” adjustments when a
calculated solar eclipse was predicted to occur at night. Eclipses that could not be
confirmed by observation, and that had no effect on court ceremony, were no longer to be
observed in the mathematical calculation of the state of the universe. This was an
addition not found in the Chinese corpus, although based off the logic found within it.
137

Katsuragi Shigetsune and Ōkasuga Hironori’s dispute regarding the implications
of “invisible eclipses” for the math of calendrics speaks of the dedication of both men to
individual visions of what the correct calculation of the yearly calendar meant. It also
attests to the variation possible within even a single system of calendrical astronomy.
Such variation has previously been discounted—calendrical technicians in China were
supposedly put to pains to run the calculations for the calendar exactly as written.
138

What the debates between Shigetsune and Hironori show is how much variation could
exist even within practice performed “as written.” Variant interpretations and
implementations of techniques would appear again in future debates, while Shigetsune’s

137
As no complete text of the Xuanming li canon survives, whether this was an innovation from the
original Xuanming li practice is ultimately unclear. In the historical account, however, Shigetsune did not
cite the Xuanming li in presenting his case against using the “advancement-retreat” adjustment. This would
seem to indicate that Shigetsune’s rule was not in the text to cite. In either case, Shigetsune was proposing
an innovation to existing Japanese calendrical practice from the ninth and earlier in the tenth centuries.

138
Sivin, Granting the Seasons: The Chinese Astronomical Reform of 1280, with a Study of Its Many
Dimensions and a Translation of Its Records: , 39.

166
strategy of making direct patron-client links with members of the Council of State—a
strategy that did indeed succeeded for a time—would be repeated time and again. These
later “heirs” of Shigetsune, moreover, would broaden the variety of the displays of skill
and service they used to elicit such links. As a result, their careers would be more stable
and even more successful than Shigetsune had realized in his own attempts.

The Ōkasuga and the Rise of the Kamo Lineage of Calendrists
As the figure of Katsuragi Shigetsune has been forgotten by historians of
Japanese astronomy or the Bureau of Onmyō, so too has the role of the Ōkasuga lineage
in the establishment of the Kamo been unacknowledged. Many historians begin their
studies of calendrics in Japan with the career of Kamo Yasunori,
139
as if he appeared
fully formed as a calendarist unparalleled. Most studies on the calendar in Japan, if they
focus at all on personalities,
140
emphasize the famous Onmyōdō ( ) lineage of the
Kamo—even if the start of Kamo participation in the official calendar project was not
necessarily smooth. That the Kamo lineage got its start at all was due to Ōkasuga
patronage. The Kamo can be said to have modeled themselves after the Ōkasuga in many
respects. Furthermore, when debates broke out between the Kamo and the Ōkasuga, it
was the Ōkasuga victory that would help shape Kamo strategies when they sought to
preserve their newly found position within the Bureau of Onmyō.

139
, 917 – 977.

140
As a prime example, Watanabe Toshio, Nihon no koyomi (Tokyo: Yūzankaku, 1993). does not include
many individuals for the pre-Jōkyō calendar in its pages.

167
Exactly why Ōkasuga Hironori selected Kamo Yasunori for his partner in the
calendar-making process of 941 is unclear. Later legends featuring Yasunori emphasize
his preternatural abilities of discernment, showing that at some point in his career (or,
more likely, posthumously) he developed a fearsome personal reputation. This reputation
should not necessarily be read backwards to the start of his career, yet such
interpretations can be found in studies of Onmyōdō. It is even uncertain whether Kamo
Yasunori had a family background or training in one of the branches of activity within
the Bureau of Onmyō before he was tapped by Ōkasuga Hironori for work in
calendrics.
141
Surviving records in later chronicles, as well as evidence from the
signatures on surviving almanacs,
142
indicate that what Hironori saw in this Kamo
individual was his potential to become a helpmeet or placeholder for continued Ōkasuga
dominance within the Bureau in general and within calendrics in particular. The irony is
that the Ōkasuga are largely forgotten while the Kamo became famous.

141
The Sonpi bunmyaku genealogy, compiled in the Muromachi era, state that Yasunori’s younger
brothers Yasuaki and Yasutō had been Instructor of Composition ( Jp. monjō hakase) in the
Bureau of the University and Supernumerary Instructor of Onmyō ( Jp. gon onmyō hakase),
respectively. Even later traditions would assign Yasunori’s father Tadayuki the post of either Instructor of
Onmyō or as onmyōji ( ) for his career. All of these would indicate some family training in related
arts. However, it should be noted that there are no contemporaneous records of these men’s careers, and all
of these sources date from at least 400 years after Kamo Yasunori’s time.
Medieval genealogies did invent family traditions predating the famous “founders” of
Onmyōdō lineages—Abe Seimei ( , 921 – 1005) became, for example, a descendant of Kibi
Makibi ( , 695 – 775)—making the statements about the early history of the Kamo family hard to
judge. At the least, it can be said that these traditions say something about the importance of family history
and precedent in the later medieval period in Japan.

142
A survey of all the surviving signatures on official almanacs ( Jp. guchūreki) can be found in
Atsuya Kazuo, ed. Guchūreki o chūshin to suru koyomi shiryō no shūsei to sono shiryōgaku-teki kenkyū
(Tokyo: Tōkyō Daigaku Shiryō hensanjo,2008)., along with a survey of all surviving almanacs conducted
over several decades.

168
Nevertheless, the Kamo did not replace the Ōkasuga quickly. Contrary to
retrospective accounts, Kamo dominance of both Instructor of Calendrics positions does
not begin before the second half of the eleventh century.
143
For much of Kamo
Yasunori’s career, he served as Instructor of Calendrics jointly with a member of a more
established lineage from the Bureau. At first this lineage was the Ōkasuga; in his later
years, the Kamo would work jointly with the Nakatomi
144
lineage that had been known
in earlier sources as the Iehara.
Yasunori’s career largely dates from before many courtier diaries survive in
complete or near-complete form, but some accounts of his activities have been preserved
in some excerpts that were compiled in manuals of court ritual and procedure. His
participation in some early debates reveals details about how the Kamo came to
prominence, and how rocky that road initially was. The first public debate that Yasunori
participated in was against Ōkasuga Masumitsu,
145
and it took place in the fourth year of
Tenryaku (950). Clues to the matter of the debate were preserved in two manuals of court
protocol that were compiled later, the late tenth-century Saikyūki ( ) and the early

143
While the almanac for 1019 shows only a single signatory, and that signatory was a Kamo, this does
not yet indicate Kamo dominance. As will be shown in the following chapter, there was an invisible class
of Buddhist contributor to some of these single-signatory almanacs, who, according to precedent, could not
sign the official calendar and publicly declare it his work. These collaborators were brought in by the Kamo
to secure their position in the Bureau, much as the Kamo had been brought in by the Ōkasuga for the same
purpose. Just as in the case with the Kamo and Ōkasuga, these collaborators and the Kamo also struggled
over the “correct” year and who had the authority to produce it. Until this collaboration broke down by the
middle of the eleventh century, the Kamo cannot be said to have controlled calendrics within the Bureau.

144
Also known as Ōnakatomi ( ) in later years.

145
, dates unknown.

169
eleventh-century Hokuzan-shō ( ).
146
In both cases the record was not preserved in
accounts of the formal presentation of the official calendar, but in sections describing
various types of extraordinary decisions in which a member of the Council of State or its
attendant functionaries might be called upon to participate.
147

The details that can be reconstructed are as follows. On the seventeenth day of
the tenth month, Minister of the Left Fujiwara Saneyori
148
called in both Instructors of
Calendrics to explain their different positions ( Jp. iron) concerning the next year’s
calendar.
149
The two men were called to testify at the assembled senior nobles’ jin no
sadame deliberations ( , “Guardhouse Deliberations”). Both Yasunori and Masumitsu,
being of comparatively low rank, were seated outside, in the garden of the jin structure.
The issue under consideration was on what day to start the fifth month of the
following year. Yasunori had placed the start of the month on the fifty-eighth day of the
sexagenary cycle,
150
while Masumitsu had placed it one day later, on the fifty-ninth day

146
The editions consulted were the annotated Shintō taikei edition (Shintō taikei hensankai, ed.
Hokuzan-shō, Shintō taikei (Tokyo: Shintō taikei hensankai,1992).) and a published photographic copy of
the Sonkeikaku manuscript (Sonkeikaku bunko, ed. Hokuzan shō, 3 vols., Sonkeikaku bunko zenpon eiin
shūsei (Tokyo: Yagi shoten,1995–1996).).

147
Particularly the Secretariat of the Council of State ( Jp. Geki) and the Controllers’ Office (
Jp. Benkan).

148
, 900 – 970.

149
Saikyūki Rinji 2. The precise date of the questioning is not preserved in the Hokuzan-shō entry,
although the latter is otherwise more detailed.

150
(Ch. xinyou, Jp. ka-no-to tori or shin’yū).

170
of the cycle.
151
In his arguments Yasunori emphasized that he had performed his
calculations precisely as mandated by the text of the Xuanming li calendrical system. Yet
he lost the debate to Masumitsu, who cited familial knowledge, passed down among the
Ōkasuga lineage from the time of the adoption of the Xuanming li system itself by
Shinnomaro. Masumitsu cited extra-textual knowledge of how to implement the
Xuanming li, information that was not included within the Xuanming li canon itself. And
relying on family history, Masumitsu won.
Masumitsu argued for the relevance of an rule he referred to as kaishōkaku ( ). Unfortunately, exactly what was meant by this term is unclear. The phrase does
not appear in any major calendrical work preserved in the Siku quanshu ( , “The
Complete Books of the Four Imperial Repositories”) collection of Chinese texts, nor in
any of the Chinese dynastic histories. The characters used seem to indicate that it may be
an adjustment to the results of calculations called for when a cycle matches up or
coincides ( Jp. kai/awaseru) with another.
152
Given other errors in characters used in

151
(Ch. renxu, Jp. mizu-no-e inu or jinjutsu). Hokuzan-shō 4, Shui zatsuji ge (“On calling experts to
answer questions or be tested”).
The Hokuzan-shō gives the sexagenary day selected by Yasunori as hi-no-to tori ( ) and by
Masumitsu as tsuchi-no-e inu ( ). These two days, the 34
th
and 35
th
days of the sexagenary cycle,
respectively, do not match with what is known about the fifth month of Tenryaku 5: that it was, in fact,
mizu-no-e inu. As Masumitsu was accorded the victory, Yasunori’s suggestion has been modified to
correspond to it. It appears that the Hokuzan-shō text is in error. (Both annotated editions and an original
manuscript were consulted.)
A ka-no-to tori sexagenary value may also help explain the name of the rule that Masumitsu
invokes, the otherwise obscure kaishōkaku ( )—the rule mandating an era name change for
ka-no-tori years was known as kakuryō ( Ch. geling).

152
This character also appears in the names of what Nathan Sivin calls “concordance” cycles. Sivin,
"Cosmos and Computation in Early Chinese Mathematical Astronomy."

171
the Hokuzan-shō account, the shō ( ) character might perhaps refer to a homonym, such
as shō ( ), the Rule Cycle.
153
No matter what the content or nature of the kaishōkaku
rule, Masumitsu’s knowledge of this element of calendrics did not come from the text of
the Xuanming li itself, but from his family traditions—Masumitsu claimed a lineage
relationship to Ōkasuga Shinnomaro, who had first tested the Xuanming li system for
Japan and then petitioned for its adoption. Yasunori’s claim to textual fidelity could
hardly be expected to withstand such arguments.

Yet the situation in 950 was already markedly different from 939, when
Katsuragi Shigetsune failed to persuade the public about his eclipse prediction. While
Ōkasuga Masumitsu won the Council of State’s decision in his favor in the 950 debate,
Kamo Yasunori’s career does not appear to have been much damaged. He appears again
as Instructor of Calendrics in 952, when he requests that a promotion intended for him be
granted to his father Tadayuki instead (Yasunori now outranked his father by a full
degree, fifth rank to his father’s sixth).
154
Fifth rank was high by Bureau of Onmyō
standards—only the post of Head of the Bureau of Onmyō required such a high rank,

153
Another error in the text can be found in the sexagenary-cycle designation of the days, discussed in
footnote 151 above.
It should be noted that as far as the Hokuzan-shō text is concerned, the purpose of the account
was to relay the procedure for calling in experts to testify before the council of state, not to record details
about calendrical mathematics. Such errors in technical terminology, therefore, are unsurprising.

154
Honchō monzui 6, Tenryaku 6/4/27. The petition is also preserved in the Chōya gunsai collection of
document models. The petition describes Yasunori’s filial feelings, as well as praises the learning and
dedication of Tadayuki, who had mastered things both “Japanese and Chinese” ( ).
(Kokushi taikei hensankai, ed. Chōya gunsai, Kokushi taikei (Tokyo: Yoshikawa kōbunkan,1964),
Ōsone Shōsuke et al., ed. Honchō monzui, Shin Nihon koten bungaku taikei (Tokyo: Iwanami
shoten,1992).)

172
according to the law codes of the eighth century.
155
Therefore, Yasunori had outstripped
the requirements for his position by a full rank.
156

Yasunori also did not remain an Instructor of Calendrics for much longer. In
960, Yasunori is described as an Instructor of Tenmon when he is called in for a
consultation on prohibited directions, indicating that he had switched positions by that
time.
157
Later tradition states that Yasunori had already become Head of the Bureau of
Onmyō when he was appointed as Instructor of Tenmon.
158
That Yasunori did indeed
serve as an Instructor of Tenmon and performed the duties assigned to that position can
be seen in entries in the Abbreviated Chronicles of Japan for the year 962—he presented
memorials on astronomical phenomena in the seventh and eighth months of that year.
159

A divination memorial ascribed to Yasunori also survives, although the history of its
transmission is unclear.
Later in 962, as well as in the years following, Yasunori appears frequently as
an active expert-on-call for advice on calendrical and directional prohibitions, and he
continued to present further astronomical memorials. He also appears as a ritualist.
160

155
Ryō no shūge, Kan’i ryō.

156
Instructors of Calendrics only had to be of sixth rank. Ibid.

157
Saikyūki (Kujōke-bon), section on moving the tennō’s residence. Dai Nihon shiryō vol. 1, part 10,
785-6.

158
Fusō ryakki Tentoku 4/4/28.

159
Nihon kiryaku Ōwa 2/7/10 and 2/8/4.

160
As a ritualist: Nihon kiryaku Ōwa 2/8/18, Kiu nikki Ōwa 3/7/23, Saikyūki Rinji 2 on changing the era
name ( Jp. nengō), Kōhō 1/8/21. As an expert: Ōwa yon-nen kakuryō kanmon (Ōwa 2/12/22), Chōya
gunsai 15 (Tenroku 4/5/26), Shōyūki Jian 3/9/2 (citing a document from Ten’en 2/8/10), Chōya gunsai 15
(Ten’en 2/9/7). As an Instructor of Tenmon: Nihon kiryaku Kōhō 3/1/9.
173
These activities burnished his personal reputation and drew the necessary attention from
the higher nobility for promotions within the court bureaucracy. Around 969 he was
promoted to Head of the Accounting Bureau ( Jp. shukei no kami), a position
with higher official rank than the position of Head of the Bureau of Onmyō, although
likewise a position that could well utilize a man with recognized mathematical talents.
161

Yasunori’s consulting activities continued, and continued to serve him, and before he
died in 977, he is reported to have reached the junior fourth rank, lower-grade.
162
While
the fourth rank was still a full rank below the level required to become a member of the
Council of State, it was far higher than any rank obtained by Yasunori’s predecessors in
the Bureau less than a century earlier.
Yasunori’s activity as a consultant and ritualist consisted of public displays of
skill, similar to the ones performed by Katsuragi Shigetsune in the first half of the tenth
century. Even so, there was a marked contrast between the strategies of the two men.
Whereas Shigetsune limited his displays of skill to work within the field of calendrics
within which he had specialized, Yasunori’s activity was more wide-ranging, and this
was reflected in the broader scope of his official career. Yasunori’s was also ultimately
the more successful strategy.

161
The dating of Kamo Yasunori’s appointment to the Bureau of Accounting ( Jp. Shukei-ryō) is
given as the second year of Anna (969) by the Kamakura-era encyclopedia Nichūreki (in the section on
people who have been heads of two Bureaus). Sonkeikaku bunko, ed. Nichūreki, 3 vols., Sonkeikaku bunko
zenpon eiin shūsei (Tokyo: Yagi shoten, 1997–1998).
Yasunori also appears as Head of the Bureau of Accounting when he submitted a report on the
applicability of a calendrical prohibition against construction in 973. Chōya gunsai 15, Tenroku 4/5/26.

162
“Kanrei” in the Yōmei bunko archive, a list of precedents for the highest ranks people achieved in their
official careers. Cited in Dai Nihon shiryō vol. 1, part 16, 135.

174
Yasunori’s major activity as an expert was to consult with the high nobility on
issues of shifting temporal and directional prohibitions, a subject not included in the
mathematical canons of the Xuanming li, but related to them—the calculations used to
determine such “taboos” were directly based on the annual calendar, and many were
included in the official almanac which the Instructors of Calendrics presented each
year.
163
Yasunori’s final promotion was also tied to calendrics, as it was a promotion
granted by the court in honor of the beginning of a new Rule Cycle in 974.
164
By
presenting himself as an expert on all temporal matters, not just the prediction of solar
eclipses, Yasunori insulated himself and his career from the effects of any one particular
failed prediction or lost debate. This explains his continued success despite having been
forced to admit the inferiority of his knowledge in the face of Ōkasuga tradition in 950.
Yasunori could still present himself as a useful expert in other, related matters. Displays
of expertise in a clutch of subjects and the performance of a variety of services to the
high nobility was a successful social strategy that Yasunori bequeathed to his sons as they
sought to continue the place of the Kamo lineage within the Bureau of Onmyō.
After Yasunori’s promotion, his son Kamo Mitsuyoshi
165
continued the Kamo
presence within the Bureau of Onmyō by assuming the Instructor of Calendrics position.
However, he did not hold the position alone. When Mitsuyoshi first appears as an

163
While the article concerns the early modern period, not the Heian, there is an overview of some of the
hemerological aspects of the yearly calendar in John Breen, "Inside Tokugawa Religion: Stars, Planets, and
the Calendar-as-Method," Culture and Cosmos 10, no. 1–2 (2006).

164
This is if later tradition is correct. “Kanrei,” Dai Nihon shiryō vol. 1, pt. 16, 135.. The date given for
the promotion in honor of the start of the new Rule Cycle on Ten’en 2/11/1 is Ten’en 2/11/16.

165
, d. 1015.

175
Instructor of Calendrics in 974, it is in the junior position to Ōkasuga Masumitsu, who
was at that time concurrently Head of the Bureau of Onmyō.
166
In an eclipse prediction
from the following year, however, Mitsuyoshi is not only listed as the senior Instructor of
Calendrics, he is described as “Acting Instructor of Calendrics” ( Jp. gyō reki
no hakase)—as the holder of junior fifth-rank, lower grade, he was already of to high a
rank for his post.
167
Like his father, Mitsuyoshi actively worked as a consultant and
ritualist to the throne and high nobility,
168
to the great benefit of his career. In 986
Mitsuyoshi received a benefit that his father had not, an additional appointment as the
assistant governor of the province of Bitchū (the western part of present-day Okayama
Prefecture). By 997 he had been promoted out of the Bureau of Onmyō, to become head
of the Palace Kitchen Bureau ( Jp. Ōi-ryō).
169

While the position of Head of the Palace Kitchen Bureau was a sign of
Mitsuyoshi’s career success (it, like that of Head of the Accounting Bureau, was a
position of higher rank than any within the Bureau of Onmyō), it was his position in
provincial government that would have proved the greater reward. Although there is no
evidence that Mitsuyoshi went to Bitchū as part of his duties as assistant governor, or to

166
The record of the 974 presentation of the calendar is lost, but the involvement of both Mitsuyoshi and
Masumitsu was recorded in documents presented to court in a debate in 1240. Heikoki Ninji 1.
(Zōhō shiryō taisei kankōkai, ed. Heikoki, 2 vols., Zōhō shiryō taisei (Kyoto: Rinsen shoten,1965).)

167
Chōya gunsai 15, Ten’en 3/6/23.

168
Mitsuyoshi’s appearances as consultant: Shōyūki Tengen 5/3/5, Tengen 5/5/10. As a ritualist: Shōyūki
Tengen 5/5/7, Kanna 1/4/18.

169
Midō kanpaku-ki Chōtoku 4, Yōmei bunko text. His signature with rank and position is preserved at
the end of this volume of the almanac Fujiwara Michinaga used to write his diary on. The presentation date
is given as Chōtoku 3/11/1 (997).

176
Harima when he became Supernumerary Assistant Governor there, his posts as a member
of the provincial governors’ offices gave him access to significant economic
opportunities. Not all appointed governors or assistant governors spent much time in their
provinces—their posts were in these cases sinecures, granted as sources of income to
support or reward them as they continued their duties in the capital. That Kamo
Mitsuyoshi received such appointments in provincial government is a prominent sign of
the regard his patrons held for him.
There were reasons to pay Mitsuyoshi even after his promotion beyond the
Bureau of Onmyō. While Kamo Yasunori’s exact involvement in calendrics from the
period after he was promoted out of the position of the Instructor of Calendrics is unclear,
his son Mitsuyoshi is known to have continued to work on the calendar in an official
capacity even after being promoted out of the Bureau. Copies of official almanacs with
signatures of those involved begin to survive sporadically for the years after 987, making
it possible to trace who received credit for a calendar for these surviving years.
Mitsuyoshi appears in surviving almanacs for the years 997 to 999 as the senior member
of a team of calendarists, even after his promotion out of the Bureau of Onmyō. Thus, by
997 at the latest, it can be said that recognized expertise in calendrics and participation in
the creation of the official calendar of the realm had become partially separated from the
official post of Instructor of Calendrics.
Despite such separation of expertise and post, Mitsuyoshi strove to ensure the
careers of his sons after him within the Bureau of Onmyō and in the position of Instructor
of Calendrics. The expertise of the Kamo family, in both ritual and consulting roles, was
177
still tied to the subjects handled by the Bureau of Onmyō. Therefore, experience as a
member of that office could only help legitimate the expertise of members of the Kamo
lineage—expertise that the Kamo lineage employed for promotions both within and
outside of the Bureau of Onmyō. Unfortunately for the Kamo, the Bureau of Onmyō,
while not controlled by any single lineage at this time,
170
was still dominated by
individuals who came from lineages with a history in the Bureau. Kamo Yasunori had
demonstrated the risks a newcomer took within the Bureau when he lost the debate with
Ōkasuga Masumichi in 950, who by contrast could draw upon unwritten traditions
handed down over generations of official calendarists. The Kamo did not have yet the
strength to dominate calendrics on their own.
During the careers of Kamo Mitsuyoshi and his son Morimichi
171
as calendarists,
both of them appear working in tandem with members of other lineages from within the
Bureau of Onmyō. In his earliest surviving almanac, that for Kanna 3 (987), Kamo
Mitsuyoshi shares the credit as senior Instructor of Calendrics with one Ōkasuga
Yoshinori,
172
then Instructor of Onmyō.
173
Yoshinori appears to have become the
Ōkasuga lineage heir, as he appears as Instructor of Calendrics in later almanacs. It was,
however, a brief career, as his place was taken by Instructor of Calendrics Ōkasuga

170
A point emphasized by Shigeta Shin’ichi. Shigeta Shin'ichi, Onmyōji to kizoku shakai.

171
, d. 1030.

172
, dates unknown.

173
In the Kanna san-nen guchūreki. Atsuya Kazuo, ed. Guchūreki o chūshin to suru koyomi shiryō no
shūsei to sono shiryōgaku-teki kenkyū, 1.

178
Yoshitane
174
in official almanacs for 998 to 1000.
175
After the presentation of the
almanac for 1000 late in Chōhō 1 (999), however, Yoshitane too disappears from the
historical record. At this point the Ōkasuga lineage also fades from both the historical
record and the top positions within the Bureau of Onmyō.
176

At this time, Kamo Mitsuyoshi was succeeded in turn by his son Morimichi, in
time for the production of the official calendar for 1004.
177
Morimichi did not himself
see Kamo dominance in calendrics at the start of his career, however. By 1004, although
Morimichi held the post of senior Instructor of Calendrics, he shared the duty of creating
the official almanac with Nakatomi (later Ōnakatomi) Yoshimasa.
178
While the
Nakatomi had not been prominent in the division of calendrics in the Bureau for over a
century, they were a lineage with long-standing ties to the Bureau of Onmyō, and with a
connection with the post—after all, the Iehara Kuniyoshi featured in the 877 controversy
over unannounced eclipses had been a Nakatomi. Yoshimasa’s appearance in the
Morimichi team from 1003 to 1011 is a sign of how the Kamo had not yet achieved a
monopoly on the production of the calendar.

174
, dates unknown.

175
The characters in these two names are distinct enough to make a copyist mistake in the 997 almanac
unlikely, although the possibility of a single individual who changed his name cannot be ruled out entirely.

176
As the holders of the top posts of the Bureau of Onmyō are mentioned in historical sources more
frequently than the Bureau’s specialists, the absence of Ōkasuga lineage members can be confirmed for
many of these years.

177
Chōhō 5/Kahō 1.

178
, dates unknown.

179
Along with pursuing careers as outside ritualists and consultants, the Kamo dealt
with their parvenu handicap in two ways. Their debt to the Ōkasuga lineage was well
known, and they themselves acknowledged it—at least internally. Even after the Ōkasuga
lineage had vanished from Bureau history, they kept the Ōkasuga lineage as textual
ancestors, according to prefaces in Kamo-owned manuals of calendrical matters.
179

Recognizing this debt, at least as long as the Ōkasuga were in living memory, and
learning the Ōkasuga traditions (or at least pretending to do so), the Kamo sought to
shore up their newly pre-eminent positions within the Bureau of Onmyō with the
authority of a more established lineage. In later centuries, the Kamo would eventually
present themselves as the only inheritors of calendrical knowledge within the Bureau, and
employ their “secret texts” and traditions freely,
180
but arriving at such a point took a
long time.
The second way in which the Kamo sought to overcome potential rivals in the
Bureau appears in the signatures on the official almanac in 1018. Here, Kamo
Morimichi’s name appears alone. But this does not mean that Morimichi necessarily
worked alone—it should be remembered that for the past thirty years the Kamo had been
forced to collaborate with another lineage within the Bureau. However, just as Ōkasuga
Hironori may have selected Kamo Yasunori as a collaborator in order to preserve an

179
The preface of an edition of the Daitō Onmyō sho ( ) dating from the Tokugawa Period
contains the information that this text was passed down from Yasunori, who had received permission to
copy it form the Ōkasuga, its owners. How the text then became copied into a samurai lineage as a military
manual, however, is left unstated. Daitō Onmyō reki sho Shimazu-bon. unknown, "Shimazu-ke denrai
gunjutsu sho," in Shimazu-ke monjo (Tokyo: Tōkyō Daigaku Shiryō hensanjo, n.d.).

180
See the discussion in chapter six on hiden as employed by the Kamo.

180
Ōkasuga dominance in the field, by selecting a junior collaborator of indeterminate status,
so too did Kamo Mitsuyoshi when he sought to secure his sons’ future careers in the
Bureau. Mitsuyoshi selected a more strategic partner than Hironori, however, and perhaps
it was for this reason that his Kamo lineage was ultimately more successful.
Specifically, when looking for a new collaborator, Kamo Mitsuyoshi selected a
Buddhist monk named Jinsō,
181
a man who could not hold an official position in the
Bureau of Onmyō and thus posed no risk to Kamo heirs in the future. Mitsuyoshi’s son
Morimichi later selected a collaborator from the same Buddhist lineage. This strategy
benefited the Kamo lineage’s ability to maintain their positions within the Bureau. At the
same time, however, it destabilized the official control of the production of the calendar
at court, as well as undermined views of the reliability of the traditions of calendarist
lineages among the high nobility. This destabilization, and the rise of Buddhist
calendarists, is the subject of the following chapter.

Conclusion
The Chinese-style calendrical system, with its technical and ideological demands,
was a difficult and complex model for the practice of mathematical astronomy. This
difficulty, along with inherent instability within the system, however, left room for
individual calendarists to contest against each other, as they produced different results
based upon their understanding of the rules of the system and the expectations of the
court, as well as derived from their family training. To bolster their position in these

181
, dates unknown.

181
debates, calendarists began reaching out to members of the Council of State to form
patron-client relationships, where the member of the Council of State gave the calendarist
political backing for his position, while the calendarist provided the member of the
Council of State first with technical information and expertise on the subject of calendrics,
and later with other ritual and technical performances. This was in effect a recognition of
the expertise members of the Bureau of Onmyō had on technical matters, even as the
Council of State reserved the authority to enforce its own standards of practice. This is
shown by its support of Shigetsune in 936, as well as by the official punishment the
Bureau faced in 877. The Bureau of Onmyō was given the authority of expertise, while
the Council of State reserved the right to determine what would be “correct.”
Such a strategy was effective as long as both calendarists and members of the
Council of State were concerned with the production of a “correct” calendar, although
among individuals the definition of what such “correctness” meant could vary widely.
These definitions depended sometimes on feelings of trust gained from personal
relationships with experts, as shall be shown in chapters five and six. Sometimes it
depended on matters of technical interpretation as well. Although Katsuragi Shigetsune
and Ōkasuga Hironori differed on what the mathematical implications of non-visible
eclipses should be, decrees made by the court and the resulting precedent meant that by
the tenth century Japanese court consensus was that, apparently contrary to the text of the
Xuanming li, solar eclipses that were calculated to occur but that would not be visible
from Japan, would not in fact be treated as solar eclipses as far as the mathematical
calculations of the calendrical system was concerned. This was a divergence from the
182
letter of the Chinese “law”, and an element of technical knowledge that was preserved
within the Bureau of Onmyō in the “secret texts” that would be introduced during later
debates such as the winter solstice controversy of 1156.
In fact, the development of a “Japanese-style” interpretation of the Xuanming li
canon would lead to further calendrical controversies and disputes once a new lineage of
recognized calendrical experts would appear in the eleventh century. This new lineage of
experts, in turn, gained their recognition in part from their clerical status and knowledge
of esoteric Buddhism. However, and ironically, the largest part of this public recognition
of their calendrical expertise came from the efforts of the Kamo lineage to secure its own
place in the Bureau of Onmyō and control of calendrical knowledge.

183
CHAPTER FOUR

INVISIBLE PRESENCES: BUDDHIST CALENDRICS AT THE HEIAN COURT

From the eighth through the ninth centuries, the Japanese court exhibited consistent
concern over whether it had a reliable population of technical experts in calendrical
astronomy. Such a population was necessary for both the production of the yearly
luni-solar calendar and the prediction of astronomical events, such as lunar and solar
eclipses, that were thought to present a tangible danger to the throne and court. This
concern led to the establishment of a second, junior Instructor of Calendrics post.
1

However, the establishment of a second Instructor of Calendrics within the
Bureau of Onmyō set the stage for increased conflict and further public debates over the
“correct” production of the calendar and prediction of eclipses. While dissent may have
been appreciated by members of the court nobility who disagreed the official stance on
calendars—for they could perhaps find new experts more congenial to their own point of
view—it did not much help the cause of producing authoritative and unquestioned facts
about time.
2
Furthermore, as the case of Katsuragi Shigetsune
3
shows, such debates

1
In addition to the Instructor of Calendrics ( Jp. reki no hakase) mandated by the
eighth-century law codes, a Supernumerary Instructor of Calendrics ( Jp. gon reki no hakase)
was created. In practice, the latter was the junior position, as well as extra-codal ( Jp. ryōgai) or not
part of the ritsuryō ( ) codes which had established the bureaucratic court and delineated its functions.

2
An issue that will be explored in more detail in chapters five and six of this dissertation.

3
, dates unknown, active 936 – 939.

184
could be both advantageous and dangerous for an individual’s career.
4
Therefore, even
though it can be seen as to the court’s advantage to encourage debate among its
experts—in order to better determine the “truth”—such debates were not always to the
advantage of the experts themselves.
5

In the end of the tenth century and the beginning of the eleventh, the Kamo
lineage was newly ascendant within the calendrical division of the Bureau of Onmyō ( Jp. Onmyō-ryō), but not yet dominant or unopposed. That the Kamo needed allies
can be seen in the signatures on surviving copies of almanacs (annotated calendars)
6

from the late tenth century—while a Kamo name is always present senior position, the
name of an older Bureau lineage, such as Ōkasuga or Nakatomi, also appears, showing
continued collaboration. To continue to rely on such lineages, however, was to risk being
displaced by them. Displacement is what happened to the Ōkasuga lineage by the very
Kamo that had been initially sponsored by them. Therefore it was desirable that the

4
As discussed in the preceding chapter.

5
The court’s fondness for determining “truth” by employing different methods simultaneously can be
seen in the type of state divination known as konrō no miura ( ). In this method of divination, used
to determine the causes or implications of national-level disasters or omens, both a mathematical method of
astrology —the Six Yang Water divination method ( Ch. liuren shizhan, Jp. rikujin shikisen),
performed by members of the Bureau of Onmyō—and a method of plastronmancy—the reading of
heat-applied cracks on a tortoise’s plastron (lower shell), performed by officials in the Jingikan (
“Council of Kami Affairs”)—were employed simultaneously. There was no cosmological connection
between these two methods, yet when they agreed with each other, the result was considered reliable.
For more information on konrō no miura see Nishioka Yoshifumi, "Rikujin shikisen to konrō
no miura."

6
Jp. guchūreki. These were annotated copies of the calendar, where the astronomical predictions
of the official state calendar were supplemented with cosmological, temporal, and medical guides for living
in harmony with the universe and the flow of time itself. This was the form in which the official calendar
for the coming year was presented to the court on the first day of the eleventh month.

185
Kamo seek out other collaborators, individuals who would not be able to replace them at
court.
The Kamo found such collaborators among members of the Buddhist clergy.
Strands of esoteric Buddhism from China encouraged the study of astronomy and
astrology for the purpose of ritual and healing. While the story of how Buddhist clergy in
Japan came to practice calendrical astronomy or horoscope astrology, in violation of
eighth-century legal prohibitions, is shrouded in legend (and, in fact, tied to the Kamo
lineage founder Yasunori
7
himself), a closer examination of references to astronomy
shows how pervasive such knowledge was among a certain type of Buddhist monastic.
These individuals had the necessary skill-set and the ability to learn; furthermore, as
members of the Buddhist clergy, they were prevented from holding most positions within
the Japanese court bureaucracy, and especially from holding positions within the Bureau
of Onmyō. Buddhist monks were even prohibited from signing any official calendars to
which they may have contributed.
It seemed as if the Kamo had found the perfect partners as they consolidated
their position within the Bureau. However, the end result was that collaboration with the
Kamo legitimated the calendrical knowledge of Buddhist monks not just for horoscopy,
ritual day selection and healing, but for matters of state timekeeping as well. From the
history of collaboration and sponsorship in the late tenth and early eleventh centuries, the
Kamo reinforced the recognized expertise of the so-called Sukuyōdō ( ) monks of

7
. See the previous chapter for more on what is known about Kamo Yasunori’s career.

186
the Tendai and Shingon schools, and so generated the lineage that would produce their
most persistant opponents in debates over calendars and eclipses from the mid-eleventh
through the thirteenth centuries.
This chapter explores some of the murky background behind these astrological
monks and re-evaluates the role of astronomical knowledge in Heian Buddhism. The first
part reconstructs the initial stage of the collaboration between the Kamo and Buddhist
monks, while the second part questions the origins and sources of knowledge about
calendrical astronomy among Japan’s Buddhist clergy. By so tracing this history, this
chapter explores the benefits that the Kamo and the monks both gained from this
collaboration, and sets the stage for the conflict between official Kamo calendarists and
Buddhist astrologer-monks that is the topic of the following chapter.

Searching for the Origins of Buddhist Calendrics in Japan
Despite an official monopoly on calendar production and astronomical
divination, the Bureau of Onmyō in the Japanese court bureaucracy was not the only
spring from which knowledge of the stars and time flowed. Buddhism, particularly
esoteric Buddhism, also produced individuals with knowledge of and experience with
how to calculate astronomical calendars. Such expertise within the population of the
sangha (Buddhist monastic community) had already been recognized by the Tang court
in China, where Buddhist monks had been employed in the creation of official systems of
187
calendrical astronomy.
8
Awareness of this expertise among Buddhist monks also existed
in Japan, and sets the scene for the first record of Buddhist participation in the calculation
of the official calendar.
9

On the eighth day of the seventh month of 1015, Kamo Morimichi,
10
the
grandson of Yasunori and the eldest son of Mitsuyoshi,
11
petitioned that the monk Jintō
12

be permitted to collaborate in the production of the next year’s civil calendar. In his
petition, Morimichi drew both on Jintō’s reputation at court and historical precedent. The
petition apparently mentioned that Jintō’s master, Jinsō,
13
had worked together with
Morimichi’s father, Mitsuyoshi, in the past.
14

In 1015 Jintō was not unknown to either Fujiwara Sanesuke or Fujiwra
Michinaga,
15
two of most influential men in the court bureaucracy at that time.
16
Jintō

8
For example, the famous astronomer-monk Yixing ( 683 – ?), who authored the Taiyen li (
Jp. Taien-reki), and whose other activity as an observer and calculator of astronomical events can be seen
in the Treatises of Heaven in the Xin Tang shu ( Jp. Shin tōsho, “New Tang History”).

9
To clarify, this is not the record of the earliest participation, which is attested to in retrospective sources.
Instead, this is the earliest surviving record of such participation. The meaning of this distinction is part of
my argument later in the chapter about the implications for the paucity of sources attesting to Buddhist
collaboration on the official calendar.

10
.

11
.

12
dates unknown.

13
dates unknown.

14
Shōyūki Chōwa 4/7/8. The petition itself does not survive, but an account of it appears in Fujiwara
Sanesuke’s ( 957 – 1046) diary Shōyūki. Sanesuke received reports concerning this petition and
its answer over two days from the Head of the Royal Secretaries Office ( Jp. kurōdo no to), who
served as the liaison between the Royal Secretaries Office and the Council of State.

15
966 – 1028.
188
had already built a reputation as a ritualist and diviner. He had spectacularly
demonstrated his skill in horoscopy earlier that year, allegedly predicting the leg ailment
that had struck Michinaga that summer.
17
Jintō also had a history with the Kamo lineage,
collaborating with Kamo Morimichi’s father, Mitsuyoshi, on a star-related ritual in
1010.
18
In other words, Jintō had already publicly demonstrated his ritual and
mathematical skills by the time that Morimichi petitioned for his inclusion in the
calendar-making project in 1015. The regent Michinaga granted Morimichi’s petition
with the following proviso: “Any future participation [by Jintō] should also be requested
by petition [at the time that it is desired].”
19

16
Michinaga was at this time regent in all but name, and Sanesuke, although at a much lower rank of
Major Counselor ( Jp. dainagon), was the heir to the Ononomiya ( ) lineage of court
protocol, and consulted on issues of court protocol and history.

17
Shōyuki 4/7/16. Although this report appears in Sanesuke’s diary after the petition, this is when
Sanesuke became aware of the reported prediction, through an interlocutor who had the facts of the
matter—Sanesuke had likely elicited the information due to Jintō’s appointment to the calendar project for
the year. The prediction had been that Michinaga would face that year the risk of head, eye, or leg ailments.
Michinaga’s collapse occurred on the nineteenth day of the previous month (Chōwa 4/intercalary 6/19).

18
Midō kanpaku-ki Kankō 7/5/19. Jintō was employed to perform the honmyō-kyō ( ), while
Mitsuyoshi performed the honmyō-sai ( ).
This incident is one of the earliest occurrences of an “inner/outer” ritual set, where Buddhist
monks and onmyōji were employed in tandem for a protective ceremony. (The “inner/outer” terminology is
taken from the distinction of naiten and gaiten , which designate Buddhist and non-Buddhist
knowledge respectively. Such terminology appears in the Azuma kagami.) As in many cases, the relevant
distinction in the Heian ritual world was between Buddhist and non-Buddhist ceremonies (or ceremonies
where Buddhist clergy were forbidden, as in certain shrine rituals), even though many “inner-outer” ritual
sets drew upon the same Chinese sources and held a strong religious Daoist influence in common.

19
Shōyūki Chōwa 4/7/9. Michinaga’s reported words are . While usually fukunin
( ) referred to a second appointment to the same or similar post, here it refers to Jintō’s future
participation in the calendar project. As a monk, Jintō could not hold a post in the court bureaucracy.
At this time, Fujiwara Michinaga did not hold either of the titles associated with the Fujiwara
regency, neither kanpaku ( ) nor sesshō ( ). However, since 996, he had effectively function as
regent, as he had the right of document inspection ( Jp. nairan), as well as the highest-ranking post on
189
There were no provisions within the ritsuryō law codes themselves for anyone
other than the Instructors of Calendrics ( Jp. reki no hakase) within the Bureau of
Onmyō to work on the calendar. What Moromichi was requesting was an official decree,
a zōreki-sen ( ), that would designate an outsider as an ad-hoc collaborator.
20
A
similar system was also established for astronomical divination ( Ch. tianwen, Jp.
tenmon) around the same time.
21
In both cases, the system was initially intended to
permit official participation in restricted fields by men who had once held the relevant
position but had since been promoted out of the post.
22
A decree that both permitted and
demanded their return to such duties would be granted in those cases when it was felt that
the skills of these ex-Instructors were required at court.
This system was then later expanded to include those individuals with the
relevant skills who had never held the relevant post for the duty.
23
For instance, Ōkasuga

the Council of State ( Jp. Dajōkan), Minister of the Left ( Jp. sadaijin). (Michinaga would
indeed receive the title of sesshō the following year.)

20
In the twelfth century, this official decree would be used to designate Kamo participants in the official
calendar-making project who had never held an official calendarist post within the Bureau. This reference
can be found Iin keizu ( , “Genealogies of Medical and Onmyōdō Lineages”), a Kamo-lineage
genealogy of the fourteenth century. (Murayama Shū'ichi, ed. Onmyōdō kiso shiryō shūsei.)

21
The first preserved tenmon missō-sen ( ) in the Ruijū fusenshō collection dates to 1019, and
was granted to Abe Yoshihira ( ) after his younger brother’s death (Chōtoku 1/10/17, see chapter
two). Another, granted to Nakahara Morotō I ( ), is referred to in the Edo-period Jige kaden
genealogical text, dates to 1030 (Chōgen 3). Masamune Atsuo, ed. Jige kaden, 6 vols. (Tokyo: Gendai
shichōsha,1978).

22
Cases such as Kamo Mitsuyoshi’s continued participation in the production of the official calendar
after [his promotion out of the Bureau of Onmyō.

23
This system is notably different from the forced laicization and incorporation employed by the
Japanese court when establishing the bureaucratic system in the late seventh and early eighth centuries. See
Ooms, Imperial Politics and Symbolics in Ancient Japan: The Tenmu Dynasty, 650-800. For a similar
190
Hironori had sought a similar decree when he brought Kamo Yasunori into the field of
calendrics in 941.
24
Participation in either calendrical production or astronomical
interpretation at court (at least at this time) still required official sanction. This legal
requirement would hold strong throughout the next few centuries, even as “unofficial”
means of participation (as consultant, or the producer of “private” calendars) proliferated,
as shown by the prominence of such decrees in genealogies of expert lineages.
25

This petition of 1015 is the first direct evidence of Buddhist participation in the
production of the official calendar, and yet it references and earlier history which can
only be found in much later documents. It is to the tracing of this history that this chapter
now turns.

Invisible Presences
According to the 1015 petition, Jintō was not the first Buddhist monk to
contribute to the production of the official Japanese calendar: Morimichi’s father had
worked together with Jintō’s master at some point in years previous. Tracing the pre-Jintō

system employed in the field of tenmon, see the discussion towards the end of chapter two involving the
Nakahara lineage.

24
Fusenshō Tengyō 4/7/17. Dai Nihon shiryō vol. 1, part 7, 936.

25
See Iin keizu and Sonpi bunmyaku for notations indicating who had received such decrees. These
notations tend to mark individuals of the Kamo or Abe lineages that began participating in court astronomy
or divination before they officially “graduated” to the appropriate post within the Bureau of Onmyō.

191
history of Buddhist participation in official calendrics, however, proves quite difficult for
a number of reasons.
26

The first such reason is institutional. The separation of the court bureaucracy
and the Buddhist clergy dates back to the beginning of the Japanese bureaucratic state.
The Sōni ryō ( , “Adminsitrative Laws on Monks and Nuns”) in the eighth-century
ritsuryō codes not only mandated the official registration and administration of all
Buddhist clergy in Japan, it also limited their lawful activity to activities that took place
within temple walls—primarily, to the ritual performance of Buddhism. Astronomy,
astrology and divination were strictly forbidden. Monks and nuns were excluded from
secular positions in the state, although an official clerical administration ( Jp. Sōgō)
was established to direct native Buddhist affairs. Buddhist affairs were also monitored by
the the Genbaryō ( ), which was a state agency officially in charge of foreign
“guests.” The clerical administration was thus made officially subordinate to the secular
one. Furthermore, the law codes mandated that any individual on official business dress
in secular style.
27

26
The reason for the difficulty is not merely because of the gaps in the historical record, or the relatively
few almanacs that survive from the late tenth and early eleventh centuries (Atsuya Kazuo, ed. Guchūreki o
chūshin to suru koyomi shiryō no shūsei to sono shiryōgaku-teki kenkyū. presents a catalog of all known
official almanacs, and copies of such almanacs, to survive.) The structure of the Japanese court bureaucracy,
from the time of its establishment in the eighth-century law codes through to the eleventh century and
beyond, served to exclude Buddhist monks from certain aspects of court life—at least officially. While the
history of the participation of Buddhist monks in the official calendar might be public knowledge—as both
Fujiwara Michinaga and Sanesuke let Morimichi’s claim to precedent pass without comment—it could not
be reflected within the official documents of state.

27
Sachiko Takeda, "Roads in the Tennō-Centered Polity," in Capital and Countryside in Japan,
300-1180: Japanese Historians in English, ed. Joan Piggott (Ithaca, NY: East Asia Program, Cornell
University, 2006). While monks and nuns might hold very important unofficial roles at court during both
192
With few exceptions, this official exclusion of monks and nuns from the
official business of the court bureaucracy continued into the Heian Period, even as the
ritual role of the Buddhist clergy in court life strengthened.
28
Manuals of court ritual
compiled in the tenth and eleventh centuries state that monks were not to be allowed into
court between the end of the Buddhist Later Seven Day Ritual ( Jp. mishūe) in
the first month and the successful completion of the first Kamo Festival ( ) in the
second month.
29
While Buddhist clergy were indeed involved in the worship of kami
deities, both those enshrined on temple grounds and in sutra-reading and other
ceremonies held at kami shrines during emergencies, Buddhist monks were still officially
excluded from much kami-worship performed at court.
30
Given all this it is not

the Nara (712 – 792) and Heian (794 – 1192) periods, it was only during Kōken/Shōtoku’s second reign (as
Shōtoku) in the second half of the eighth century that monks were appointed to official positions within the
court bureaucracy.

28
On the strengthening of this ritual role, and how the development of Japanese schools of esoteric
Buddhism figured into it, see Ryū'ichi Abe, The Weaving of Mantra: Kūkai and the Construction of
Esoteric Buddhist Discourse (New York: Columbia University Press, 1999).
Abe does not much address, however, what the continued exclusion of Buddhist clergy from
some realms of Japanese court life means for his view of esoteric Buddhism as ascendant early in the Heian
Period.

29
This restriction shows up in several manuals, but early and notably in Fujiwara Sanesuke’s Ononomiya
nenjū gyōji ( “The Ononomiya Schedule of Annual Court Events”). Fujiwara Sanesuke,
"Ononomiya Nenjū Gyōji," in Gunsho ruijū (Tokyo: Yagi shoten, 1960).

30
Not only were they not included, but court officials involved in upcoming offerings at major shrines (or
in Jp. shinji), were expected to abstain from meeting Buddhist clergy. This equated contact with
Buddhist monks with drawing blood (receiving acupuncture was also forbidden), and handling dead
animals.
Japanese religion in the late classical and medieval periods is typically described as syncretic
between kami-worship and Buddhism ( Jp. shinbutsu shūgō). See Mark and Fabio Rambelli
Teeuwen, ed. Buddhas and Kami in Japan: Honji Suijaku as a Combinatory Paradigm (London:
RoutledgeCurzon,2003), Yoshie Akio, Shinbutsu Shūgō (Tokyo: Iwanami shoten, 1996). However, aside
from the famous “Buddhist exclusion” at the Kamo and Ise Shrines, what this separation of Buddhist clergy
from officials involved in shinji means for Japanese religion has yet to be adequately addressed in
scholarship.
193
surprising that Buddhist monks, even if they received official permission to work on the
astronomical calculations that produced that year’s calendar, were not permitted to place
their signatures on the resulting almanac.
31

The legal position of Buddhist clergy in the Japanese bureaucracy was not the
only reason, however, for their apparent “invisibility” in the production of the official
calendar: the established form of the annotated calendar as an official document with da
fixed signatory style also excluded monks. In all surviving almanacs up through the end
of the sixteenth century with signatures, the only signatures that appear are of individuals
with official posts in the secular court bureaucracy. The formal requirements of official
court documents exerted a significant influence on how such documents were presented
at court, and under whose name, regardless of the actual conditions of authorship. For
example, in the case of the official almanac, although there are multiple known incidents
when the presentation of the official almanac was delayed for one reason or another, all
surviving almanacs have as their date of presentation the first day of the eleventh month.
This was the designated date of presentation in court procedure, and it was reflected in
the format of the document, regardless of when the official calendar was actually
presented at court. Other types of documents show signatory-imposture, where the
official using the document format is forced, by diplomatic protocol, to adopt a fictional

31
That this was the case was cited in Shunki Chōryaku 2/11/27. (The edition consulted was Zōhō shiryō
taisei kankōkai, ed. Shunki, Zōhō shiryō taisei (Kyoto: Rinsen shoten,1965).)

194
persona for his signature.
32
As a result, it is not surprising that no calendars survive with
a monk’s signature present. It is striking, nevertheless, that the almanacs for 1018 to 1020
only bear a single signature: that of Instructor of Calendrics Kamo Morimichi. While it is
possible that Morimichi worked alone on calculating the calendars for these three years,
all other surviving almanacs bear two or more signatures, marking this as highly
unusual.
33
As these years are all soon after Morimichi’s request to collaborate with the
monk Jintō, it seems as if Jintō may have been a silent partner for these years.
Making a final confirmation of Jintō’s participation more difficult for the years
1018 to 1020. In contrast to 1015, there are no records of Morimichi petitioning for
Jintō’s continued participation, despite the fact that such a re-petition was mandated by
Fujiwara Michinaga’s decision. The absence of further documentary evidence, however,
may indicate more about diary-keeping practices at court during that time than about
Morimichi’s discarding of Jintō as a collaborator. While Fujiwara Sanesuke included the
initial petition of 1015 in his court diary, the Shōyūki, the regent Fujiwara Michinaga did

32
A prime example of this is the naishi-sen ( “Head of Palace Attendants Decree”). Originally,
under the ritsuryō bureaucratic system, some forms of direct communication to and from the throne were
designated as the responsibility of the all-female Naishi-dokoro ( “Palace Attendants Office”).
After the Kusuko-no-hen ( ) attempted coup of 810, however, many of the duties of these female
officials were gradually assigned to the newly established, and extra-legal ( Jp. ryōgai),
Kurōdo-dokoro ( “Royal Secretaries Office”), which was staffed by men. These men continued the
use of the naishi-sen, but they had to sign such documents in the persona of women. Thus the legal
bureaucratic procedure was continued in appearance, even as it was being circumvented. See
Toyoda Masahiro, "Naishi-sen.".

33
Exceptions would be almanacs copied out by kokugaku scholars in the early modern period, as for
some of these the signatory section was abbreviated. See Atsuya Kazuo, ed. Guchūreki o chūshin to suru
koyomi shiryō no shūsei to sono shiryōgaku-teki kenkyū.

195
not record the petition in his.
34
The reason for this lack is simple: court diaries are biased
towards novelty. While a first instance in court procedure was important for establishing
precedent, subsequent repetitions of the same did not have the same impact. Court diaries
were kept largely, but not exclusively, to provide information for courtiers or their
descendants for determining court procedure, and as a source for legal precedent and
court history.
35
For Sanesuke, the story of clerical participation in the production of the
official calendar was still new and valuable information, as shown by his inclusion of the
Jinsō precedent in his diary. By contrast, the regent Michinaga, who had presumably been
more aware of the precedent or already held documents on the subject, did not make a
record of either the petition or of his response to it. Tracing the early history of Buddhist
participation in the official Japanese calendar requires more detective work.
Further complicating matters, the separation of Buddhist clergy from official
courtly functions extended even to the rewards for accurate calendar calculation. For
example, when both Kamo Morimichi and Jintō correctly predicted a solar eclipse in
1021, the official records of the court only preserved the occasion of Morimichi’s reward:
when the Nihon kiryaku ( “Abbreviated Chronicles of Japan”) was compiled
based on these records, Morimichi is the only calendarist mentioned.
36
Yet Morimichi is

34
See Midō kanpaku-ki Chōwa 4/7/7 and Chōwa 4/7/8. Michinaga had no diary entry for Chōwa 4/7/9.

35
The most recent and complete study of the purpose and uses of courtier diaries is Matsuzono Hitoshi,
Ōchō nikki ron. In the first chapters, Matsuzono explains the history and scholarly debate as to why courtier
diaries arose as a genre—despite this debate, the utility of the courtier diaries as a tool for court careers is
well attested in the remainder of his monograph, as well as in his earlier work, ———, Nikki no ie: Chūsei
kokka no kiroku soshiki (Tokyo: Yoshikawa kōbunkan, 1997).

36
Nihon kiryaku Jian 1/7/1.The compilation of the Nihon kiryaku is described in Hosoi Hiroshi, Kodai no
tenmon ihen to shisho.
196
known to have collaborated with Jintō before—so it is not surprising to find that courtier
diaries preserve Jintō’s reward for also contributing to this successful eclipse prediction.
While Kamo Morimichi is awarded with cloth goods at court, Jintō had to receive his
reward in the courtyard of the regent’s residence.
37
Although this courtyard was still very
much a “public” space, it was not considered an official space of state. Likewise, while
Jintō’s efforts were known to the court and even publicly recognized, they were not,
however, recorded in official documents of state. In a fashion similar to the concepts of
omote (official statements) and ura (the hidden political conditions) in early modern and
modern Japan,
38
Buddhist monks were both included and excluded from the practice of
official state calendrics.
This distinction between “official” and “public,” even if its strongest expression
was only on paper as it were, characterizes the relationship between official and
non-official calendarist lineages in both the Heian Period and beyond. Jintō’s
participation in both the calendar project of 1015 and the eclipse prediction of 1021 was
known to the members of court, but could not be acknowledged in the formal records or
spaces of bureaucratic government. Yet the account of Jintō’s reward reveals the type of
evidence which can be used to uncover more details of the participation of Buddhist
monks in calendrical astronomy at court—to trace the history of Buddhist involvement in

37
Sakeiki Jian 1/7/1. Jintō’s name appears as Jin’en ( ) in some published editions.

38
This concept has been considered a “key term” for understanding Japan, but its most recent full
development as a method of analysis can be found in Luke Shepard Roberts, Performing the Great Peace:
Political Space and Open Secrets in Tokugawa Japan (Honolulu: University of Hawai'i Press, 2012).

197
calendrics requires paying close attention to references in court diaries and other
“non-official” sources.

Although in court diaries, Jintō’s master Jinsō’s participation in the production
of the official calendar is only referred to as a precedent, more details can be found in a
petition from 1105 from the Kamo lineage. This petition, a history of all special
promotions related to work on the calendar, was submitted to court during a dispute
between the Kamo and the Abe regarding whether the Abe could share in such rewards.
Jinsō’s participation had in fact been preserved in records that the Kamo lineage kept on
historical figures who had been rewarded for their participation in calendar production,
and thus brought up again as the Kamo presented a history of calendrical production and
reward at the Heian court.
39

While it might seem unusual that the Kamo lineage preserved records that
acknowledged the past participation and skill of individuals who did not belong to the
lineage, when examined in the light of the politics of record preservation at court it
should not be surprising. After the ninth century, the history and precedents of court were
not solely, or even primarily, preserved in a central location. Instead, various lineages
privately kept court diaries and copies of official documents, and used the knowledge

39
Information from these records were employed in a petition collected in Chōya gunsai 15 Chōji 2/2/21.
This petition was in fact intended to preserve Kamo-lineage ownership of calendrics and its attendant
rewards from a claim to the same by the Abe lineage in the twelfth century.

198
therein for prestige and competition at court.
40
Such distributed knowledge meant that
trying to whitewash a rival out of history was a dangerous gambit—the deceit could be
discovered in the records of another house, and the credibility of the petition thus lost.
Therefore, preserving information on the participation of Buddhist monks on the official
calendar was to the Kamo lineage’s benefit. Furthermore, since under the laws of the
realm Buddhist monks could not hold the positions of the secular officials of the Bureau
of Onmyō, such acknowledgment posed little actual danger to the posts held by the Kamo
members. Failing to recognize the history of participation by Buddhist monks, then,
could have cost the Kamo family dearly, while recognizing it would have cost them little.
Unfortunately, beyond the reference to his participation, there are few details
concerning Jinsō’s work on the official calendar in the 1105 petition. However, some
facts can still be deduced. The petition concerns rewards given out for participation on
the calendar in the first year of a Rule Cycle ( Ch. zhang, Jp. shō),
41
but no dates are
attached for the entries on the three monks noted. Jinsō is the first monk of the three,
implying primacy—perhaps even that he was the first monk to have worked on an official
calendar at all. The reward he received was the post of director ( Jp. bettō) of
Saidaiji ( ) temple in Nara. Promoting a monk in this way mirrored the official

40
For more on this aspect of knowledge and politics in the Heian court, see Matsuzono Hitoshi, Nikki no
ie: Chūsei kokka no kiroku soshiki.

41
The Rule Cycle (known in the west as the Metonic Cycle) is a cycle of 19 years, wherein the number of
days in the year and the number of days in lunations (235) roughly coincide. It was observed in Japan when
the winter solstice fell on the first day of the eleventh lunar month ( Jp. sakutan tōji). See chapter
three for more details.

199
rank promotions that were given out to official calendarists at such events, while
preserving the nominal separation of Buddhist clergy and official bureaucratic posts.
It is from this reward that the approximate years when Jinsō may have
collaborated with the Kamo can be sought. No record of Jinsō’s appointment appears in
the historical record, although documents from the Tōji ( ) temple identify Jinsō as
the director of Saidaiji in 1000.
42
It is known that he was replaced in 1004.
43
If Rule
Cycles are counted off from the first year of the first Rule Cycle in Japan in 794, then the
only two years that fall within Kamo Mitsuyoshi’s known career as a calendarist
(974-1000) and before Jinsō first appears as director of Saidaiji (1000) are 974 and 993.
Ōkasuga Masumitsu seems to have been an active calendarist in 974, which would
indicate that there was no lack of official calendrical personnel in 974, which is a major
factor that would have inhibited Jinsō’s inclusion. It seems more likely, therefore, that
Jinsō participated in the production of the calendar of 993 rather than that of 974,
although both remain possible.
It is more difficult to determine the duration of this collaboration. As only the
single reward is listed, it is impossible to say whether this was the only year Jinsō
participated in the production of the official calendar, or whether he was a long-term
collaborator with Mitsuyoshi. Nor is much known about Jintō’s calendars, although there
are a few more details available for Jintō’s career. Jintō received permission to work on
the calendar in 1015, and he is also listed in the same 1105 document as having received

42
Chōhō 2. Tōji monjo kō ( ) 28, Dai Nihon shiryō vol. 2, part 21, 298-299.

43
Midō kanpaku-ki Kankō 1/3/7; Gonki Kankō 1/3/7.

200
a special appointment as a reward. Just as Jinsō had been before him, Jintō was awarded
the position of administrator of Saidaiji.
44
However, in Jintō’s case, another source,
again from the Tōji archives, gives a date for this appointment: the second day of the
third month of the second year of the Manju era (1025).
45
This appointment is rather
delayed from the beginning of the last Rule Cycle in 1022, indicating either that a lot of
time passed before Jintō was rewarded for his calendar work, or else that the Saidaiji
appointment is better considered as general reward for his efforts. In either event, the
evidence is strong that Jintō continued to serve as a collaborator with the Kamo in the
production of the official calendar well after 1015. He also began to consult on
calendrical issues, presenting an opinion on whether it was mathematically and
hemerologically
46
necessary to change the era name ( Jp. nengō) in 1020.
47

44
Chōya gunsai 15 Chōji 2/2/21.
What significance Saidaiji might have for the history of Sukuyōdō in Japan, as it is not
traditionally considered a center of such activity, is a topic for further study.

45
Tōji monjo kō ( ) 28, Dai Nihon shiryō vol. 2, part 21, 298-299.

46
“Hemerology” is divination based on the calendar. On this use of the word, see Marc Kalinowski,
"Diviners and Astrologers under the Eastern Zhou: Transmitted Texts and Recent Archeological
Discoveries," in Early Chinese Religion, ed. John Lagerwey and Marc Kalinowski (Leiden: Brill, 2009),
341-396.

47
Nihon kiryaku Kannin 4/10/13. Changing nengō served both as a symbolic renewal of the mandate of
heaven, as well as a way of magically avoiding bad luck. Popular religious tales, particularly from China,
show that name changes served as a way of tricking the universe, which like any good bureaucracy kept
track of individuals and states through identifying tags.
Ross Bender has published an article on nengō changes, although he misidentifies them in the
title as “calendrical changes.” Although both nengō changes and calendrical changes (at least for the Han)
served as renewals of the Mandate of Heaven, era-names did not change the means of calculating the days
and months of the year—nengō only applied to years. Ross Bender, "Changing the Calendar: Royal
Political Theology and the Suppression of the Tachibana Naramaro Conspiracy of 757," Japanese Journal
of Religious Studies 37, no. 2 (2010).

201
There is plenty of evidence, therefore, that at least a few Buddhist monks
collaborated quietly on the calendar from the end of the tenth century through the
opening decades of the eleventh century. Previous scholarship has put the end of the
collaboration between the Kamo and Buddhist monks as 1038,
48
but it may be more
accurate to say that 1038 saw a spectacular breakdown in the collaboration between one
specific monk and a member of the Kamo calendarist lineage.
49
There are indications
that Buddhist monks continued to work on the calendar, at least sporadically, until
1092.
50
It was only when conflict arose, however, that calendarist monks are explicitly
listed in the sources. In other cases, they remained officially invisible if publicly
acknowledged—just as the Kamo had initially intended. For these invisible presences, it
seems that the rewards for participating in the production of the official calendar was
worth any official disavowal.

Sino-Indian Astrology and the Futian li ( “Tallies of Heaven Calendrical
System”)
The reason why the Kamo would select Buddhist monks in particular as
collaborators cannot be merely boiled down to their legally marginal status within the

48
Yamashita Katsuaki, Heian jidai no shūkyō bunka to Onmyōdō.

49
This topic will be covered in the next chapter.

50
Nōsan ( d. 1094) appears in Go-nijō Moromichi ki as having been rewarded for his work on the
calendar, despite the fact that the calendarists of the Bureau had yet to receive their “payment.” Go-nijō
Moromichi ki Kanji 6/12/30.
Nōsan’s death is recorded, along with a biography, in Hōryūji bettō shidai. Dai Nihon shiryō
vol. 3, part 3, 382-3.

202
“official” world of the court bureaucracy. Jintō had displayed skill in horoscopy before
his selection as a contributor to the official calendar and such skill made him a
knowledgeable collaborator for calednrical astronomy. But how Jintō had come to
employ such skill, and how it came to be publicly acknowledged, still requires
examination. To do this requires looking once more at legends involving Kamo Yasunori,
and the role that astrological and calendrical knowledge played in esoteric Buddhism.
Knowledge of astronomy is intimately tied with the ritual and astrological field
known as Sukuyōdō ( “The Way of Lodges and Luminaries”). Sukuyōdō is a
word that appears both in the Tale of Genji
51
and in studies of Japanese astronomy,
52
but
has remained somewhat obscure. The word also appears elsewhere is discussions of
Japanese religion and divination-related beliefs.
53
It is often said to be intertwined with
Onmyōdō
54
and has even been confused with it. There was a technical component,
involving divination, day-selection and astronomical calculation, as well as a
performative aspect, a set of rituals dedicated to the stars.
55

51
The first appearance is in the Kiritsubo chapter. The term is somewhat misleadingly translated in
Royall Tyler’s most recent edition of the Tale of Genji, where all of the diviners consulted are called
“physiognomists.” See Royall Tyler, The Tale of Genji (New York: Viking, 2001), 12-13,

52
Sugimoto Masayoshi and David L. Swain, Science and Culture in Traditional Japan, A.D. 600-1854,
74 and 125-127.

53
Allan G. Grappard, "Religious Practices," in Cambridge History of Japan, ed. Donald Shivley and
William McCullough (Cambridge: Cambridge University Press, 1999)., 553-554.

54
Sugimoto Masayoshi and David L. Swain, Science and Culture in Traditional Japan, A.D. 600-1854,
125-127.

55
While sukuyōji ( ) did not hold a monopoly on Buddhist star-related rituals, sukuyōji, when they
are employed in ritual endeavors, are often found in the historical sources performing star-offerings. See
Osan buruiki ( “Records of Royal Births, Arranged by Type”) for this pattern, Unknown,
"Osan Buruiki," in Zoku gunsho ruijū, ed. Zoku gunsho ruijū kanseikai (Tokyo: Yagi shoten, 1958).
203
Two major texts have been seen as central for the technical aspect of Sukuyōdō.
The first work is a sutra, the Sukuyō-kyō ( Ch. Xiuyao jing, “Sutra on
Constellations and Luminaries”) from which the field takes its name.
56
Presented as a
translation of a Sanskrit sutra, it is actually the combination of two attempts to translate
methods of Indian astrology into Chinese, and into a Chinese astronomical system. The
sutra was later incorporated into the Taishō canon, although that version has errors that
were introduced at a late date in its history.
57
The second text, the Futian li ( Jp.
Futen-reki), was a system of calendrical astronomy.
58
The Futian li, composed around
722, was never officially used by the Tang Dynasty (618–907), but was rather created
unofficially outside of the court, most likely by a Buddhist author, in order to reconcile
some elements of Chinese and Indian systems of mathematical astronomy.
59
The Futian
li does not survive in full in either China or Japan, but a description of it survives in the

56
The full title of the sutra is (Ch. Wenshushili pusa ji
zhuxiansuo shui jixiong shiri shan’e xiuyao jing). Yano Michio emphasizes this work as central to
Sukuyōdō. Yano Michio, Mikkyō senseijutsu: Sukuyōdō to Indo senseijutsu.

57
Ibid., introduction.

58
Its role in Sukuyōdō was emphasized by Momo Hiroyuki. Momo Hiroyuki, "Sukuyōdō to sukuyō
kanmon," in Onmyōdō sōsho, ed. Murayama Shū'ichi (Tokyo: Meicho shuppan, 1991), 367-388.
Momo and Yano do not necessarily contradict each other, but they are focusing on different
areas. Yano is looking at Sukuyōdō as something that developed in China, while Momo is focusing on the
documentary traces of Sukuyōdō divination in Japan.

59
The nature of the combination is discussed in Shigeru Nakayama, "The Position of the Futian Calendar
on the History of East / West Intercourse of Astronomy " in History of Oriental Astronomy. Proceedings of
an International Astronomical Union Colloquium, No. 91, New Delhi, India, November 13-16, 1985, ed. G.
Swarup et al. (Cambridge: Cambridge University Press, 1987).
Although never adopted by the Tang, its appearance in the Xin Wudai shi indicates that it was
at least studied by some successor states.

204
Xin Wudai shi ( Jp. Shin Godai shi, “New History of the Five Dynasties”)
60
and
in a Japanese manuscript about calendrics from the sixteenth century.
61
From these texts,
and from calculations attributed to the Futian li system, some of it has been reconstructed.
In these two texts, Buddhist monks would have found instructions on how to
“locate” elements of Greco-Indian astronomy and astrology: including the days of the
week (associated with classical planets), the moon’s location among the nakshatra or
“lodges”,
62
and where the sun was in the zodiac.
63
Other texts, not yet recognized, may
have played a role as well. Combined with further material in the sutra, such information
provided guides to ritual performance and to illness and healing. Monks found therein the
raw material for a product they could offer to lay patrons—including knowledge about
the cosmos and the individual’s location within it. Such information was desirable to
members of court desired, and thus could be used as cultural capital when monks went
searching for sponsors and clients. Such skills could promote a Buddhist monastic’s
career as well.

60
Compiled by Ouyang Xiu, 1053.

61
Held by Tenri University. I have not yet the opportunity to examine this manuscript personally

62
The 27 constellations around the ecliptic. As the moon’s sidereal orbit (appearance in its starting
position against the background stars) takes about 27 days, each nakshatra is “inhabited” by the moon for
each day. For this reason, nakshatra is often translated as “lunar lodge” although Christopher Cullen has
argued against such a translation. Christopher Cullen, "Translating *Sukh/Xiu and
*Lhah/She—‘Lunar Lodges’, or Just Plain ‘Lodges’?," East Asian Science Technology & Medicine, no. 33
(2011).

63
The first two are found in the Xiuyao jing. The last is not preserved in any surviving text of the Futian li,
but is presumed to have been included as such calculations appear in two horoscopes drawn up using the
Futian li: Sukuyō unmei kanmon (Unknown, "Sukuyō unmei kanmon," in Zoku gunsho ruijū, ed.
Zoku gunsho ruijū kanseikai (Tokyo: Yagi shoten, 1958).), and in the archive of Rokujō Ariyasu. (———,
"Sukuyō onunroku," in Rokujō Ariyasu-shi shozō monjo (Tokyo: Tōkyō Daigaku Shiryō hensanjo, n.d.).).

205
While the Sukuyō-kyō is thought to have been brought to Japan by Kūkai
64
or
one of his followers, the history and first use of the Futian li in Japan is unclear.
Influentially, a document from the mid-eleventh century strongly ties the first importation
of the Futian li to the Kamo lineage. This document has been greatly exploited by
scholars working on the topic of Sukuyōdō. Specifically, in 1053,
65
this document was
produced the provincial government at the Dazaifu in Kyūshū for the temple Ōuradera
( )
66
in order to confirm the succession of a new abbot ( Jp. zasu). The
document also reaffirmed the temple’s right to select its own leaders, and reiterated the
temple’s history of service to the state.
The document retells the foundation legend of the temple. Ōuradera had been
found in or near the grounds of the Dazaifu in the middle of the tenth century by the
monk Nichien,
67
who did so at the behest of the regent Fujiwara Morosuke.
68
Nichien’s
career as described in the document begins with his trip to “Tang” China (actually to the
state of Wuye, 907-978), to study in the Tiantai mountains. When word of his planned
trip spread, he was summoned to court, where he received an official order that had been
requested by “Head of the Bureau of Accounting and Instructor of the Three Fields”

64
774 – 835.

65
The document itself is undated, but Takeuchi Rizō identified it as dating from 1053 (Tengi 1). The last
date mentioned in the document itself is Eishō 6/2/19 (1051). Heian ibun #4623. Takeuchi Rizō, ed. Heian
ibun, 14 vols. (Tokyo: Tōkyōdō shuppan, 1963–1976).

66
While this temple may have originally been known as Daihoji (the Sino-Japanese reading of the
characters in its name), the area in which it was located is now referred to as Ōuradera.

67
dates unknown.

68
909 – 960.

206
Kamo Yasunori. According to the Dazaifu document, Yasunori had petitioned the court
thus:
Instructors of other fields all study deathless texts and all study them
diligently; however, calendrics preserves the principle of reform, and new
techniques follow from changes in the observations of the heavens. As a
result, the Chinese change their calendar once every rotation of the Big
Dipper.
69
Since the Xuanming li canon
70
was imported and implemented
in the first year of Jōgan [859],
71
approximately 130 years have passed.
[Meanwhile] the reformed calendrical systems of Great Tang have not
been transmitted [to Japan] by anyone. The Monk Ambassador Nichien is
the disciple of deceased Precepts Master Jinkan,
72
so it will be easy for
him to inquire after and study calendrical methods. It would be fitting for
Nichien to receive a royal decree ordering him to study and bring back
new calendrical methods to Japan.
73

Having received this order, Nichien is reported to have left for the “Wu capital,”
where he received permission to study at their Astronomical Bureau. There, he mastered
the Futian li calendrical system. When he returned in 957, he brought back to Japan not
only his newly-acquired training but also the text of the Futian li itself. He also reportedly

69
Much of the text seems badly garbled from the point of view of the history of calendrics, and there are
numerous errors. For example, Xuanming li ( ) is written as . This sentence is a translation
of , which appears garbled and is a little obscure. This reading is based on
the fact that the Big Dipper was also used to mark time, and that each year in a 12-year branch cycle ( Jp.
shi, “Chinese Zodiac”) was assigned one of the stars in the Big Dipper. If the document author is attempt to
assert that calendrical reform occurs in China roughly every twelve years, however, the estimated rate of
calendrical reform is far too frequent. Another meaning may have been intended, but is difficult to
determine from the text as it is written.

70
Written in error as , as mentioned in the previous footnote.

71
The date given is incorrect. While the Xuanming li system was adopted in the Jōgan Era, it was the
fourth year of the era (862).

72
dates unclear

73
Heian ibun #4623.

207
carried with him over a thousand other texts, both religious and secular.
74
Nichien is then
credited with transmitting both the text of the Futian li and the knowledge of how to
implement it to Yasunori. The Buddhist texts went to temples, and the other material
went to the Ōe lineage.
75
Although Nichien was subsequently showered with honors at
court, and was even offered a spot in the clerical administration (Sōgō), he declined it all,
in favor of a contemplative existence. Upon the request of Fujiwara Morosuke, however,
he did move to Kyūshū to found the temple there.
This document has been noted as a key source for the history of Sukuyōdō and
for Buddhist calendrical astronomy in Japan. Yet there are several doubtful elements in
its story. The Futian li, which Nichien was supposed to have transmitted to Kamo
Yasunori, is only known to have been used by Buddhist monks for horoscopy and for
making some eclipse predictions.
76
Even though study of the Futian li has been credited
with markedly improving eclipse predictions by members of the Bureau of Onmyō in the
twelfth century,
77
there is no evidence it was owned or studied by the Kamo at that time.

74
Nichien has been most famous in Japanese historiography for his connection to the Futian li, and so
these other texts—referenced in the petition—have not been much noticed. However, their appearance, and
the appearance of the Ōe lineage, indicates that this connection was as important for this foundation legend
as the Futian li itself.

75
Notably the texts and lore that he learned about the Zhou yi ( Jp. Shū eki, “Book of Changes”), as
well as the record he kept of his stay in China.

76
An eclipse prediction preserved on En’ō ni-nen guchūreki ( , 1240), on the reverse of
the Maeda-bon edition of the Gōdan shō. Both the Futian li and the Xuanming li are cited in an eclipse
prediction preserved on the reverse of the Kiu hō nikki (Unknown, "Kiu hō nikki," (Tokyo: Tōkyō Daigaku
Shiryō hensanjo, n.d.).).

77
Saitō Kuniji, Kotenmongaku no michi (Tokyo: Hara shobō, 1990), 116.

208
In fact, so far the only identified users of the Futian li system have been Buddhist monks,
which makes Yasunori’s request very peculiar in retrospect.
78

There are other elements of this 1053 document that appear suspicious. There
are miswritten characters and temporal errors throughout. The role of Kamo Yasunori in
the tale serves as a particular problem. Notably, his title is anachronistic,
79
and the timing
of the reported request itself seems unusual, coming only three years after he lost a public
debate to Ōkasuga Masumitsu. Furthemore, despite having reportedly received the Futian
li tradition from Nichien, Yasunori never petitioned for the adoption of the system nor,
evidently, ever employed it.
80
Yasunori was reputed to have applied Buddhist calendrical
for hemerology (he was in fact critiqued by later medieval Buddhist commentaries for

78
The only indication I have yet found that it was ever used by the Kamo lineage is a legend that was
initially recorded by Ichijō Kaneyoshi in the fifteenth century, and preserved as a citation in an
eighteenth-century text: Tanigawa Kotosuga ( ), Nihon shoki tsushō. Kojima Noriyuki, ed. Nihon
shoki tsūshō (Kyoto: Rinsen shoten, 1978).
It is difficult to take Kaneyoshi’s assertion too seriously: although records from the late middle
ages also presented other calendrical canons that the Kamo lineage was supposed to use, but although the
presence of these texts can be attested to in Japan, their consistent use cannot.
For example, the Qintian li ( Jp. Kinten-reki) which is described as having been used in
conjunction with the Xuanming li in Bunki 1 (1501). (See Kunaichō Shoryōbu, ed. Kujō-ke rekisei kiroku,
4 vols., Zushoryō sōkan (Tokyo: Kunaichō shoryōbu,1989–1999).) The Qintian li, used under the Northern
Song from 956 to 963, was studied and utilized in a court debate by one Kiyohara Noritaka ( ).
However, mathematical reconstruction of the calendar in Japan finds that it primarily accords with the
Xuanming li system (with only 8% deviating in any way), and not any other system. Uchida Masao, Nihon
rekijitsu genten.

79
Kamo Yasunori is listed as Head of the Bureau of Accounting and “Instructor of the Three Fields.”
While the ahistorical use of Yasunori’s title of Head of the Bureau of Accounting may be simply ascribed
as a retrospective application of his highest position, then is no other evidence that Kamo Yasunori ever
served as Instructor of Onmyō ( Jp. onmyō hakase), along with his historically attested Instructor
of Calendrics ( Jp. reki no hakase) and of Tenmon ( Jp. tenmon hakase).

80
Aside from the aforementioned eclipse predictions, there is absolutely no evidence of the Futian li
having been used to calculate the Japanese calendar. As Uchida Masao has shown, the Japanese calendar
from 862 to 1683 basically accords with the Xuanming li system as it has been reconstructed with only
some notable exceptions, which can be mostly explained through manual adjustment to the calendar after it
was calculated using the Xuanming li. Uchida Masao, Nihon rekijitsu genten.

209
having done so incorrectly) but these were all from the Sutra on Constellations and
Luminaries text.
81
Yasunori indeed may have had a connection to Buddhist calendrics,
but there is no evidence it was through the Futian li.
While 1053 is more than a century before Yasunori’s first appearance in medieval
tale literature as a charismatic ritual genius,
82
it may be that the Ōuradera temple was
attempting to associate its founder’s skill in calendrics with Yasunori’s growing aura as
the preternaturally-skilled founder of the Kamo lineage. Assigning the credit for the
origins of Futian li-based calculations in Japan to Kamo Yasunori is also an unusual
depiction of power relations, given the conflicts that would come between Kamo
calendarists and Buddhist monks over correct calendrical practice.
In fact, there was no need for Kamo Yasunori to request that a Buddhist monk be
directed to study calendrical astronomy in China. Even if official request from the court
of Japan would have had some weight with Chinese authorities, it was as if there was a
lack of knowledge of calendrical astronomy among Buddhist monks in Japan. The truth
was that Japanese monks had been concerned about correct calendrical calculation since
the eighth century.

Needing to Know the Stars and Days

81
Nakajima Wakako, "Genji monogatari no Dōkyō, Onmyōdō, Sukuyōdō " in Genji monogatari kenkyū
shūsei, ed. Suzuki Hideo and Ii Haruki Masuda Shigeo (1998), 314-316.

82
One of the teachers of the even more famous Abe Seimei. Tales involving both men can be found in
English translation in Yoshiko Kurata Dykstra, The Konjaku Tales. Japanese Section: From a Medieval
Japanese Collection, 3 vols., Intercultural Research Institute Monograph Series (Osaka: Intercultural
Research Institute, Kansai Gaidai University Publication, 1998–2003).

210
The interest of Japanese Buddhist monks in issues of calendrics and horoscopy
dates to well before Nichien and is tied to the growth of esoteric Buddhism in Japan. The
founder of the Shingon school of Buddhism, Kūkai, is credited with the introduction of
the Sutra on Constellations and Luminaries as well as the days of the seven-day week to
Japan.
83
During his time in China, the Tendai monk Ennin
84
took care to purchase a
calendar there.
85
In subsequent years, the monk Enchin
86
in his Sasa gimon (
“Record of Unsettled Doubts”) listed questions that he hoped would be answered through
discussions with Chinese monks and the study of Chinese texts. Among these points are
details of “Western” (Indian) calendrical methods and concepts.
87
Therefore, well before
Nichien is reputed to have gone to China in the middle of the tenth century, Japanese
Buddhist monks had been pursuing these same problems. Given this history, there is no
reason to assign the initial impetus for the pursuit of Sino-Indian methods for calendrical

83
The days of a seven-day week, with the planetary associations found in the Latin tradition, was
transmitted along the Silk Road through the Sogdian kingdoms, finally making it to Tang China. Instead of
serving as a marker for time in Japan or China, or to designate a holy day of rest, the seven-day weeks was
primarily used for divination purposes: for example, indicating which malevolent deity was responsible for
causing illness. Yano Michio, "Bukkyō kyōten no naka no koyomi, Sukuyō-Kyō," in Onmyōdō sōsho, ed.
Murayama Shū'ichi (Tokyo: Meicho shuppan, 1991).

84
794 – 864.

85
Edwin Reischauer, Ennin's Travels in China (New York: Ronald Press Company, 1955), 126.

86
814 – 891.

87
For example, in the first part of the second half of the text is dedicated to questions about Indian
calendars referenced in a commentary on the Mahavairocana Sutra. This commentary, Yixing’s Dari-jing
shu ( Jp. Dainichi-kyō so), included astronomical references such as to the moon’s apparently
speeding up and slowing down. Enchin questioned what this reference referred to, as well as to what a
“small month” meant for this system of astronomy. Enchin, "Sasa gimon," in Dai Nihon Bukkyō zensho, ed.
Bussho kankōkai (Tokyo: Meicho Fukyūkai, 1978).
On Yixing’s commentary, see Yano Michio, Hoshi uranai no bunka kōryūshi (Tokyo: Keisō
shobō, 2004).

211
astronomy to the figure of Kamo Yasunori. In fact, the Futian li text is attested in Japan
from the ninth century.
88

The connection between Chinese-style calendrics and esoteric Buddhism, in fact,
originated in China.
89
Even though Buddhist monks in China seem to have adjusted to
the Chinese civil year, using it to guide their activities and even utilizing popular
almanacs,
90
there were reasons for monks or temples to perform calendrical calculations
themselves.
91
Furthermore, increasing interaction with Buddhist travelers and texts from
India during the Tang Dynasty further encouraged interest in Indian calendrical
calculations in China and in Chinese Buddhism.
92
As new technical and medical texts
were imported and translated, they contributed to the continuing development of esoteric

88
The text appears in the Nihon-koku genzai sho mokuroku catalog (Fujiwara Sukeyo, c. 891).

89
Astronomy/astrology and Buddhism are also linked in Tibetan Buddhism, but this link is often
attributed solely to the Indian tradition of astronomy, despite Chinese-style features in modern Tibetan
astrology. Further work on the subject is needed.

90
Almanacs preserved in the Dunhuang Caves, for example, conform to what is known of the official
civil calendar used at the time. Many of them, even so, were unofficially produced. For more information
on these almanacs see Nishizawa Yūsō, Tonkō rekigaku sōron: Tonkō guchūrekijitsu shūsei (Tokyo:
Nishizawa Yūsō, 2004).

91
One such use for calendrics was to calculate the dates of the Three Ages of the Dharma—forecasting
the period of the End of the Dharma ( Jp. mappō) might provide millennial hopes as well as useful
forewarning. On the apocalyptic trend in post-Han Buddhism, see E. Zurcher, "‘Prince Moonlight:’
Messianism and Eschatology in Early Medieval Chinese Buddhism," T’oung Pao 68, no. 1/3 (1982).
For the culture of preparing for the End of the Dharma in Japan, see D. Max Moerman, "The
Archeology of Anxiety: An Underground History of Heian Religion," in Heian Japan, Centers and
Peripheries, ed. Mikhael Adolphson et al. (Honolulu: University of Hawai'i Press, 2007).

92
For the history of Sino-Indian interaction and trade in the Buddhist period (including information on
Brahmin astronomers) see Tansen Sen, Buddhism, Diplomacy, and Trade: The Realignment of Sino-Indian
Relations, 600-1400 (Honolulu: University of Hawai'i Press, 2003).

212
Buddhism. Later, such texts would be categorized as belonging to “mixed esoteric
Buddhism” ( Jp. zōmitsu).
93

The ability to correctly determine the “real” day and date via the correct positions
of astronomical and astrological bodies had been a concern not only for the Chinese
imperial state and the polities that had modeled themselves on it. In both China and India,
celestial information held a great deal of relevance for issues of personal health and even
an individual’s ultimate fate.
94
There had been a long-standing tradition in China,
furthermore, of almanacs and manuals based on the temporally-driven interaction of yin
and yang and the Five Phases, wherein authors provided prescriptions involving the
appropriate timing of all sorts of activities, from medical to social, to political and
economic.
95
This context shaped how Indian astrological medicine was received and
translated, and influenced Sino-Indian astronomy and astrology as it developed among
Buddhist communities in Japan.

93
The categorization zōmitsu comes from the Japanese Shingon tradition, and should not be interpreted as
an indication of any fundamental ritual or doctrinal divergence from junmitsu ( “pure esotericism”)
texts, which were categorized by the Shingon tradition as more orthodox. Neither does this distinction
indicate an evolutionary, or de-evolutionary, relationship between the two text classifications, which
largely developed at the same time.
As the texts involving astrology and calendrical methods for star-related health rituals are
generally identified by this tradition and subsequent Buddhist scholarship as zōmitsu, I have employed the
term while recognizing it as ahistorical.
For criticism of the term zōmitsu and some history of its use, see Charles D. Orzech, "The
Trouble with Tantra in China: Reflections on Method and History," in Transformations and Transfer of
Tantra in Asia and Beyond, ed. István Keul (New York: De Gruyter, 2012).

94
The connection, at least from India, stretches even further west. A part of the horoscopy tradition
transmitted to Japan that appears calculated from the Futian li includes elements from Ptolemaic astrology.
More details can be found in Yano Michio, Hoshi uranai no bunka kōryūshi.

95
What Michel Strickmann calls a “specialized literature of practical, operational demonology” (referring
to li shu ). Strickmann, Chinese Magical Medicine, 69.

213
The determination of the correct astronomical definition for each day, whether
with regards to asterism and planet (the days of the “week”) or merely by the position of
the moon, therefore became embedded within the texts of “mixed” esoteric Buddhism.
For example, the correct healing ritual for a health problem was selected through
knowing the actual constellation associated with the day.
96
“Knowing” the correct
astronomical information for the day, however, remained difficult. It was made even
more difficult due to the incommensurability between Chinese and Indian astronomy. For
example, although the Indian nakshatra were translated directly into the Chinese xiu (
Jp. shuku), as both referred to a set of 27 (Indian) or 28 (Chinese) constellations along the
ecliptic which could be used to define the position of the sun, moon, or planet in the sky,
the translation resulted in both confusion and astronomical inaccuracy. Nakshatra and xiu
were not astronomically equivalent.
In fact, the system for figuring out in which xiu the moon was located for each
day of the month was written twice into the Sutra on Lodges and Luminaries—once in a
chart, and once instructing the reader to find an “Indian astronomer” from whom to
request the information—the two halves of this sutra contradict each other in this way.
The reason for the different instructions was simple: it depended on whether the Chinese
xiu were used in place of the Indian nakshatra, or whether merely the Chinese names for
the xiu were reassiened to the nakshatra. The Chinese xiu were of widely varying size,

96
As in Yixing’s Qiyao xingchen bie jingfa ( “[Healing] Methods Determined by
Luminary and Asterism,” Taishō 1309. Nihon Daizōkyō hensankai, ed. Nihon Daizōkyō, 49 vols. (Tokyo:
Nihon Daizōkyō hensankai,1914–1921).). Evidence from Dunhuang shows how important rituals for the
lay community, including healing rituals, could be for a Buddhist organization.

214
for they had never been used for calculating time astronomically.
97
The Indian nakshatra,
by contrast, were of all equal size,
98
and thus for each day of the month, the moon was to
be found in a new lodge.
99
The table in the sutra, although it uses the names for the
Chinese xiu and not the Indian nakshatra, is accurate for the nakshatra concept. The
instruction directing the reader to an “Indian astronomer,” on the other hand, is accurate
for the Chinese xiu and their divergent sizes.
100

In other words, there was much for Japanese monks like Enchin to hold doubts
about while seeking answers for Chinese Buddhist texts involving the stars.

Spreading the Dharma of the Stars
Once esoteric Buddhism was well-established in Japan, with a sufficient supply of
texts and regular attention to issues involving calendrics, horoscopy performed by
Buddhist monks appeared. Although historical sources, once again, are thin for this early
period, horoscopy’s use in Japan’s elite circles seems to have been established around the
time when Jinsō was first collaborating with Kamo Mitsuyoshi. The earliest textual
mention of a Buddhist horoscope dates to 999.
101
By the tenth century seeking out copies

97
Admittedly, this is a somewhat controversial statement. However, no reconstruction of the Chinese
night sky for any time period will provide an era wherein these constellations are all to be found of equal
size, as would be necessary for astronomical day-keeping.

98
At least at the time of the composition of the sutra.

99
The lunar sidereal orbit, or its motion against the backdrop of stars, is roughly 27.5 days (shorter than a
lunation period—the difference has to do with the movement of the earth-moon system around the sun).

100
A more detailed account of this part of the sutra, along with the textual history of the composition of
the sutra, can be found in Yano Michio, Mikkyō senseijutsu: Sukuyōdō to Indo senseijutsu.

101
Gonki Chōhō 1/10/16.
215
of one’s own horoscope had become standard annual practice among the higher-ranking
circles at court.
102

Consider, for instance, the horoscope from 999 for one Fujiwara Yukinari.
103

This horoscope was in fact composed by Jinsō, who, as has already been shown, was also
a collaborator on the official calendar. Yukinari apparently received the horoscope
unsolicited, although he was pleasantly impressed by its detail and apparent accuracy.
104

That Yukinari, a low-ranking but well connected member of court,
105
received such a gift
shows that Buddhist monks were utilizing the same sort of displays of skill to promote
their reputation that as secular Kamo lineage was doing.
106
Horoscopy was a way for
monks to build a reputation for themselves and to drive demand for their health-related
and propitiatory rituals.

102
Momo Hiroyuki, "Sukuyōdō to Sukuyō kanmon.", and Yamashita Katsuaki, Heian jidai no shūkyō
bunka to Onmyōdō.. Surviving records seem to indicate that it was mostly the highest level of court that
had regular access to horoscopes, although others sought them out for special occasions. Receiving or
seeking out horoscopes ( Jp. sukuyō kanmon) do not seem to have been the sort of event always
recorded in courtier diaries.
Fujiwara Sanesuke’s Shōyūki entry for Chōhō 1/12/27 records that he received his own
horoscope from a different monk, one Gakuen ( dates unknown). This indicates that the practice was
spread among esoteric Buddhist monks beyond the Jinsō-Jintō lineage.

103
972 – 1028.

104
Gonki Chōhō 1/10/16.

105
Fujiwara Yukinari at this time was a provincial governor. However, he was soon to become the head
of Prince Atsuyasu’s ( ) household (in 1001) and had a friendly relationship with Ichijō Tennō. If
Jinsō was intending to make himself and his skills known to officials with close relationships with a
potential heir to the throne, Yukinari may have been a good target to select.

106
Discussed in the previous chapter.

216
There were, in other words, good reasons for esoteric Buddhist monks to develop
skills in calendrical calculation, and to promote and utilize such skills in public—reasons
that need not be tied to a legendary relationship between Kamo Yasunori and the monk
Nichien. Before the Jinsō-Mitsuyoshi collaboration, however, even these specialist
monks would not have been able to publicly participate in, perhaps not even to comment
on, issues of state-level calendrical astronomy. Participation in the production of the
official calendar project still required official permission.
107
Through the precedent of
such collaboration, that would soon change.

Conclusion
It is difficult to say precisely who reached out to whom first in forging a
partnership between Bureau of Onmyō and Buddhist establishment. Buddhist monks had
long-standing reasons to develop and promote their ability in calendrics—reasons which
predated the supposed 957 importation of the Futian li by Nichien. These reasons were
deeply tied to esoteric Buddhism, as shown above. By the end of the tenth century, there
is evidence for several monks having cultivated a public reputation for horoscopy—a
field related to calendrical astronomy or that at least utilized overlapping
skills—apparently independently of any Kamo involvement or encouragement. For these
monks, participation in the production of the official calendar may have appealed as
another method for cultivating their personal reputations, and the reputations of their
lineage and school, among elite circles at court.

107
As discussed earlier in this chapter.

217
The Kamo lineage, on the other hand, was looking for collaborators who would
not displace them from their new position in the Bureau, yet who also held the
prerequisite skill to aid them in the production of the official calendar. Buddhist monks,
who could not hold posts within the Bureau of Onmyō according to both law and custom,
were a non-threatening choice.
108
Both sides certainly benefited from the initial
relationship. The general silence of the historical sources, with only a few references to
official participation in the production of the state calendar through 1092, shows a system
of occasional collaboration that worked smoothly and as it should, at least from the Kamo
point-of-view—that is to say, beneath the notice of official court documents.
Yet in the next few decades of the eleventh century, some Buddhist calendarist
monks would become suddenly more prominent in the historical sources. This
prominence is directly related to conflict, as disputes about the proper practice of
calendrical astronomy sparked controversy between patrons of individual calendarists at
the highest levels of court. Even as the collaboration between the Kamo calendarists and
Jintō had, to all appearances, worked smoothly, in one of Jintō’s disciples the Kamo
calendarists would find a persistent challenger. As a monk, this new challenger could not
threaten the Kamo’s place in the court bureaucracy. Nonetheless, his public critique of
the official calendar would undermine the authority of the Kamo lineage, and of the
Bureau of Onmyō as a whole, as providers of astronomical truth.

108
The laicization of monks to staff the Bureau is covered both in chapter two, and in Ooms, Imperial
Politics and Symbolics in Ancient Japan: The Tenmu Dynasty, 650-800. No documented Buddhist monk
kept his clerical identity while serving in the Bureau of Onmyō.

218

CHAPTER FIVE

COMPETING CALENDARS: ECLIPSE PREDICTION, SHŌSHŌ, AND THE
ESTABLISHMENT OF BUDDHIST CALENDRICS AS AN INDEPENDENT
TRADITION

By the opening decades of the eleventh century, internal conflict and public debates had
shaped a Japanese tradition of calendrical astronomy based on the Xuanming li, but
differing from canonical Chinese practice. Such conflict also drove the Kamo lineage of
the Bureau of Onmyō ( Jp. Onmyō-ryō) to seek collaborators from outside the
Bureau. Buddhist monks from esoteric schools seemed, at first, the perfect collaborators:
Buddhist monks could not hold the positions that the Kamo were fighting for, and due to
the history of esoteric Buddhism in China, some already had the background knowledge
and skills to collaborate. The Kamo never intended for their Budhist collaborators to
overshadow them, and the vague history of early Buddhist collaboration with the Kamo
lineage is proof that this scheme initially worked as the Kamo had hoped.
However, that would soon change as the eleventh century progressed. The
public recognition of Buddhist lineages’ skills in calendrical astronomy emboldened
some individuals—most notably and first, the monk Shōshō
1
—who then began to
challenge the official Kamo calendarists directly. As a result, debates about eclipse
prediction and the correct calendar spilled outside of Bureau walls, and any method of

1
dates unclear. He is also listed in sources as and , both homonyms.
219
controlling dissent from inside the Bureau no longer had traction. As the Kamo
consolidated their position and authority within the Bureau of Onmyō, they were
damaging the authority of the position of Instructor of Calendrics outside of the Bureau.
Meanwhile individuals like Shōshō, who were neither trained within the
Bureau nor beholden to its traditions, further complicated the scope of Japanese practice
by introducing new methods or re-introducing abandoned ones to the work of calendrical
astronomy. This resulted in a wealth of alternative predictions—particularly concerning
solar eclipses—which ended up destabilizing general confidence in any individual
astronomical prediction, no matter what the source. The elite members of the Japanese
court came to view any prediction with a jaundiced eye, as they turned to new ways of
evaluating the credibility and reliability of their specialists.

Watching the Skies
On the morning of the first day of the second month of the second year of the
Kahō era (1095),
2
the courtier Fujiwara Munetada
3
was one of a number of men at court
anxiously watching the skies.
4
There had been an eclipse predicted for that day, and the
news had brought the court to a standstill. The regent and his father
5
had both entered

2
March 9, 1095 in the Julian reckoning.

3
, 1062 – 1141.

4
The following account is based on Fujiwara Munetada’s diary, Chūyūki Kahō 2/2/1.

5
Fujiwara Moromichi ( , 1062–1099) and Fujiwara Morozane ( , 1042–1101)
respectively.

220
into seclusion with ten high-ranking attendants and others to practice austerities,
6
and all
official royal activity was to shut down for the next five days.
7

Such was an extreme reaction, but well within standard operating procedure for
the court when a solar eclipse was predicted to be visible. The sun was the symbol of
yang and thus corresponded to the ruler, the embodiment of yang on earth.
8
An
obstruction of the sun’s rays by an eclipse was potentially dangerous for the ruler, who
needed to be protectively enveloped and shielded from such effects.
9
The risk extended
beyond the throne to other individuals as well. So the person of the tennō was protected
by keeping him indoors, lowering the blinds, and canceling court business for the day.
10

The prediction for the eclipse of 1095 had been announced in advance by the
calendarists in the Bureau of Onmyō ( Jp. Onmyō-ryō), as required by law.
11
As

6
ibid.

7
ibid.

8
This association, drawn from Treatises on the Heavens in the Chinese Dynastic Histories, was the
theoretical justification for courtly anti-eclipse measures. This reasoning was cited the most explicitly in
arguments during the controversy over nighttime eclipses in 877 (Sandai jitsuroku Jōgan 17/11/1).

9
The protective “wrapping” of the ruler is discussed in the first chapter of Kuroda Hideo, Ō no shintai, Ō
no shōzō (Tokyo: Heibonsha, 1993).
Initially, solar eclipses that were predicted to occur at night or that were otherwise not visible in
Japan were also considered dangerous. However, at the beginning of the tenth century the court had finally
determined that “nighttime” solar eclipses were not dangerous, and therefore shifted to not observing
countermeasures for such eclipses. The process leading up to this decision (of 918 for effect in 919) is
described in chapter three, and some late objects will be covered below.
The decision to ignore the ritual and health consequences of “nighttime” solar eclipses resulted
in a change in the mathematics of calendrical astronomy, as also described in chapter three.

10
Such actions were also taken in response to the appearance of comets, which in the Treatises on the
Heavens indicated revolution, and in Japanese interpretation (Shodō kanmon Kashō 1) indicated disaster
among the highest levels of society. Chūyūki Kashō 1/1/4.

11
Chūyūki Kahō 2/2/1. Although the first time that the eclipse
prediction appears in the Chūyūki is on the day of the eclipse prediction, the preplanning required to shut
221
the court had protected the tennō before the eclipse started by preventing any exposure to
the eclipse, the danger posed by the event had been already avoided. Furthermore, as a
precaution, both an event in honor of Confucius ( Jp. sekiten)
12
and usual offerings
to Ōharano shrine (for the Jp. Ōharano-sai) had been cancelled. All of this
turned out to be ritually unnecessary: in fact, the sky had been cloudy that morning, with
rain the night before. The cloudy skies, obscuring the eclipse, also should have eliminated
any threat—yet members of the court still waited. When finally the clouds parted around
noon and the sun appeared round and bright, people in the street reacted with shouts of
celebration.
13
The protocol of silence at court was lifted,
14
and people returned to their
daily lives.
Yet, as Munetada knew, any failure to sight the eclipse did not mean that the
eclipse did not occur. It has been suggested by previous scholarship that the Japanese
court’s failure to adopt any new system of calendrical astronomy for so long despite some
dramatic failures to accurately predict eclipses was due to the influence of esoteric

the court down to observe monoimi ( ) indicates that the announcement had been made earlier, quite
possibly the eight-days previously required by the Engi shiki codes. (Engi shiki 20.)
While the word (Jp. rekidō) could in theory refer to any sort of calendarist, at this time it
exclusively refers to calendarists associated with the Bureau of Onmyō—the Kamo lineage.

12
On the sekiten observances, see Borgen, Sugawara No Michizane and the Early Heian Court.

13
Chūyūki Kahō 2/2/1.

14
ibid. Onsō ( ) was the announcing of hours at the court. During times of mourning
and during monoimi it would be suspended—in the latter case, it may have symbolized the throne’s
acceptance of any cosmological censure.

222
Buddhism.
15
Yet although the influence of esoteric Buddhism in Japanese culture from
the Heian Period (794 – 1192) onward was indeed strong, the influence of esoteric
Buddhism did not lead the Japanese court to reject perceptions of the natural reality
around them, nor to disregard predictions of eclipses. The power of esoteric ritual was
employed to prevent harm from the eclipse, not to prevent the eclipse itself, nor to deny
the reality of eclipses.
The actual influence of esoteric Buddhism on eclipse prediction and calendrics
in tenth and twelfth-century Japan was both more subtle and more in line with variations
in predictive methods for eclipses found between the Katsuragi and Ōkasuga lineages, (as
well as between the Ōkasuga and the Kamo lineages), in the previous century.
16
The
debates between official and Buddhist calendarists from 1038 to 1248 not only reveal the
presence of two different schools of calendrical astronomy present in Japan—they also
show that alternate methods of performing calendrical calculations were possible. The
early debates, which took place in the first half of the eleventh century, also shaped how
the “Japanese-style” of calendrical astronomy promoted by specialist lineages then
emerging in Japan would be viewed and accepted by the court. And these debates reveal
to historians—just as they did to the lay audience of courtiers at the time—how uncertain
Japanese knowledge of astronomical events could be.

15
Particularly Nakayama, A History of Japanese Astronomy: Chinese Background and Western Impact,
74–76.

16
The discussion and evidence for such differences between lineage or individual’s
methods—particularly on the application or elimination of post-calculation rules—is covered in chapter
three, as is the history of the Katsuragi and Ōkasuga lineages

223
When the eclipse of the second month 1095 was not sighted due to cloud-cover,
Munetada did not think that this was because the eclipse itself had not occurred or had
been prevented through supernatural means. As he relates in his diary, the Bureau of
Onmyō
17
had announced that the partial eclipse was to take place early: to start before
dawn, and finish before mid-morning.
18
Munetada explicitly noted in his court diary that
the clearing of the skies took place after the eclipse was scheduled to end, so he was
clearly aware of the eclipse’s predicted timeframe.
19
The jubilation and praise that went
up from the crowds after the clouds broke, therefore, was not interpreted by Munetada as
due to the fact that the eclipse did not occur, or because it was thought to have been
prevented through divine virtue. The celebration was because the danger was over.
The limits accepted by Munetada on the effects of Buddhist ritual can be seen
in his diary. When, in his account of the day’s events, Munetada praises the supernatural
efficacy of the prayers of the Buddhist prelate Ryōshin,
20
Munetada explicitly credited

17
Here referred to by a “Tang name” ( Jp. tōmyō) of Sitiantai ( , Jp. Shitendai), which was
one of the names of the Bureau of Astronomy in the Chinese bureaucracy. In other places this term would
be used to refer to the tenmon specialists within the Bureau of Onmyō—however, as the context in this
Chūyūki entry clearly refers to the calculated prediction of an eclipse and not to an observation of the same,
either Munetada seems to be using the term in an atypical fashion, to refer to the entire Bureau.

18
12 and a half of 16 parts of the sun obscured, or 78%. The start of the eclipse was given as during the
double-hour of the Tiger, or between 3 AM and 5 AM, and the end during the double-hour of the Dragon,
or between 7 AM and 9 AM.

19
ibid. Noon was indeed well after the eclipse had ended in Japan: from
Kyoto, the eclipse would have been visible from sunrise at 6:12 AM until 6:56 AM, and was at its
maximum eclipse in Japan (65% of the sun eclipsed) at sunrise. NASA, "Nasa Eclipse Web Site."

20

224
the presence of the clouds and rain to Ryōshin’s efforts.
21
For Munetada and other
grateful members of the court, Ryōshin did not prevent the eclipse—which, unsighted,
may or may not have occurred. Ryōshin did, however, successfully provide another layer
of shielding between the tennō and the dangerous sky.
Therefore, it can be said that the Japanese courtiers remained in doubt over
whether or not the eclipse had actually occurred. Given this, it is then odd that talk of the
accuracy or inaccuracy of eclipse predictions appears nowhere in Munetada’s 1095
account, even though such concerns were raised with regard to other predictions in the
same court diary. What this entry makes clear, therefore, is that the Japanese court was
not particularly concerned with confirming eclipse predictions. This is contrary to
assumptions made by most historians of astronomy in pre-modern Japan, who assume
that accurate predictions of eclipses were, in fact, the main aim of calendrical
astronomy.
22
Instead, avoiding danger was the primary goal. The accuracy of eclipse

21
Chūyūki Kahō 2/2/1. The
connection between anti-eclipse rituals and Buddhist rituals for praying for rain has, to my knowledge, not
yet been examined. Yet the prevalence of clouds and rain as the result of esoteric anti-eclipse rituals in
Japanese Buddhist sources indicates a connection worth noting. Interestingly, clouds and rain appear as the
result of a performance of the Shijōkō-hō ( ), a ritual dedicated to Tejaprabhā Buddha (“Great
Blazing Perfect Light” ), as it appears in the Tendai Asabashō ( ) manual—in the
Japanese historical record, this ritual does not have a strong connection to rain-production.
Bussho kankōkai, Asabashō, 7 vols., Dai Nihon Bukkyō zensho (Tokyo: Bussho kankōkai, 1912–1914).

22
Therefore, the “problem” identified by both Nakayama, A History of Japanese Astronomy: Chinese
Background and Western Impact, and Sugimoto Masayoshi and David L. Swain, Science and Culture in
Traditional Japan, A.D. 600-1854—primarily, how could the Japanese have kept using the Xuanming li
system as time rendered it less and less accurate—is not a “problem” to be explained at all. The goal of
calendrical astronomy in Japan was not an increasingly accurate depiction of reality. Instead, the goal was
the avoidance of danger and the coordination of the state. The “eclipse problem,” therefore, while a major
concern, as this chapter will show, was not the be-all and end-all of Japanese calendrical practice.

225
predictions was only necessary as far as it aided that goal.
23
Unnecessary precautions
were undesirable, and the reliability of predictions were an issue on that account. But in
the absence of a dissenting voice, any somewhat trustworthy eclipse prediction would
stand.
This did not mean that the population was completely unconcerned about the
accuracy of eclipse predictions. Despite all precaution and preparation, the members of
court kept watching the sky well after the time-period predicted for the eclipse by the
officials in the Bureau of Onmyō. This attention is the symptom of a general mistrust in
the accuracy of eclipse predictions, and of a desire to ensure that no harm come to the
state. The court had become painfully aware of the potential for error in these predictions
through a number of major eclipse debates between calendarists from the Bureau of
Onmyō and outside experts over the past century. In particular, such debates centered
around visibility and time of the eclipse. Even though the solar eclipse prediction for
1095 had not been controversial—or at least, there is no record of any debate—still the
court watched the sky long after the eclipse had been predicted to end. Obviously, the
high members of court did not behave as if they fully trusted the prediction of their
experts.

23
I differ from Michael Loewe, in this regard, as for why predictable (at least in theory) solar eclipses did
not become “non-ominous.” Loewe assumes that under normal circumstances they would become
non-ominous, and therefore has solar eclipse predictions unreliable until court ritual becomes fossilized in
China—thereby freezing the “negative” meaning of solar eclipses for future generations. Michael Loewe,
"China," in Divination and Oracles, ed. Michael Loewe and Carmen Blacker et al. (London: Allen &
Unwin, 1981). I do not share in the assumption that something predictable would become non-ominous,
and therefore see no paradox requiring explanation.

226
The eclipse debates which destabilized elite trust in specialists in calendrical
astronomy flourished then fade away in a period of roughly 250 years spanning over the
eleventh and thirteenth centuries.
24
The largest factor in this activity was the rise of new
groups of recognized experts in calendrical astronomy, groups not housed within the
Bureau of Onmyō but still employed by the court in a different capacity.
25
As these
newer lineages clashed with older established ones, they most often did so over eclipse
prediction.
There were reasons for those with calendrical expertise to debate over eclipse
predictions. Significant benefits awaited the winner of an eclipse debate in eleventh and
twelfth century Japan, albeit these rewards were more for the winner’s reputation than his
purse.
26
The selection of eclipse predictions as the primary topic for debates—although
not the sole topic
27
—among calendarists had a long history. This was due to the
importance of solar eclipse predictions both for the court audience concerned about
danger, as well as due to the primary place eclipse prediction held as the most difficult

24
Noted previously by Saitō Kuniji, although he does not provide any theories as to why this might have
been the case. Saitō Kuniji, Kotenmongaku no michi., 102-116.

25
While it is possible that this historical pattern might be attributable in part to new trends in
recordkeeping at court as well as the increased survival of some categories of court documents, the
appearance of eclipse debates in the eleventh century—well after the first shifts in court archival
patterns—and fading away in the thirteenth century—which does not correspond with any major
documentary patterns—indicates that if archival practices were a factor, it was a minor one.

26
Material rewards could also be provided for accurate eclipse predictions, as was in the case for Jintō
discussed in the previous chapter, but records of such “payments” are rare.

27
The proper length of months and of the year remained issues of potential debate, as shown by the
controversy over the official calendar in 1039 discussed below.

227
performance and ultimate standard for Chinese-style calendrical astronomy.
28
Eclipse
predictions served not only as a significant warning of potential danger at court, but also
served as a way to prove expertise and skill in calendrics. Lunar and solar eclipses had
the advantage of being readily recognizable even to non-expert observers—provided
conditions were good and that the eclipse was of sufficient magnitude.
29
This public
nature of a visible eclipse, however, could prove a double-edged sword. The potential
disadvantage that eclipse prediction held for calendrical astronomers and their predictions
was that on a day with clear skies, the success or failure of their predictions would be laid
bare before a large audience, in a way that the precise dating of the winter solstice—also
confirmable by observation, but only expert observation—was not.
30

The belief that the prediction of solar eclipses was the most exacting and best
test of a calendrical system, as well as the most difficult astronomical phenomena to
predict, is perhaps best expressed in the statement of a thirteenth-century Chinese
astronomer:
The exactitude of an astronomical system stands and falls on its
treatment of eclipses. In this computational art exactitude is hard to
come by. There is always uncertainty about whether the (predicted)
time is early or late, and whether the (forecast) immersion [or
percentage of the sun eclipsed] is too shallow or too deep. If exact

28
This can be seen in part through the role that eclipse predictions played in the build up to the official
adoption of the Jōkyō reki ( ) of Shibukawa Shunkai ( 1639 – 1715) in 1684.

29
The difficulty of observing eclipses in suboptimal conditions is a major factor in the events discussed in
the following chapter.

30
As noted in Nakayama, A History of Japanese Astronomy: Chinese Background and Western Impact,
124, the changes in the length of the shadow of the gnomon at winter solstice are sufficiently difficult to
observe through pure observation that an interpolation method was used.

228
agreement (with the phenomena) be the goal, there can be no room
for chance.
31

Japanese observers of the eleventh and twelfth centuries would have agreed
with this passage generally, but they used eclipse predictions to evaluate the work of
individuals, and rarely to test systems. As the debates of the tenth century had shown, it
was possible for different individuals to derive different results from calculations based
on the same Xuanming li system, it would be expected that they would differ if they
came from different traditions of how to read this complex text. Such variation was not
only possible, but clearly attested. Furthermore, additions and ad-hoc adjustments were
made to the raw results post-claculation—another potential source of variation in the
results. The calculation of calendrical system was hardly “inviolable, not to be tinkered
with or dismembered by technicians,” as some historians have claimed, at least for
Japan.
32

Given this variability, it made less sense for inaccurate eclipse predictions to be
seen as a failure of the system itself than as an individual’s lapse. Conversely, the

31
Li Chien, quoted in Yuan shi ( , Jp. Gen shi, “History of the Yuan”) 53. The translation is by
Nathan Sivin, cited in Steele, Observations and Predictions of Eclipses by Early Astronomers, 175. The
material in parenthesis was added by Sivin, whereas the material in square brackets has added by the
myself in order to define a technical term (“immersion” which refers to the amount of the sun or moon
which is eclipsed.)

32
Sivin, "Cosmos and Computation in Early Chinese Mathematical Astronomy," 2. Later in the article he
does admit the possibility of modification: “The astronomical Bureau could violate the spirit of this precept
in emergencies by adopting a new technique ‘on a provisional basis’” and he gives an example (———,
"Cosmos and Computation in Early Chinese Mathematical Astronomy," 59).
The Japanese case, as shown in the previous chapter and will be elaborated below, was more
extensive than would be accounted for by Sivin’s statement or exception: the modifications were multiple
and employed for centuries. This is a problem for those, like Steele (———, Observations and Predictions
of Eclipses by Early Astronomers), who would use Japanese eclipse predictions to test the Chinese
Xuanming li system.

229
difficulty of predicting solar eclipses, along with the importance of such predictions for
the health and safety of the court, made eclipse predictions the ideal medium for displays
of skill by competing calendarists at court.
33
The consequences of such competition,
along with the inherent difficulties of prediction solar eclipses—particularly when using a
system devised for observers thousands of kilometers away
34
—created general mistrust
of eclipse predictions at court, even if the importance of such predictions meant that even
doubted experts could not be fully dispensed with. It was within this context that
Buddhist calendarists and the official calendarists of the Kamo lineage first cooperated,
then competed.
The actions taken by Munetada and other members of the court in the spring of
1095 were not independent or unprecedented. Their vigiliant monitoring of the sky when
under threat of an eclipse was prompted by a history of fierce debate among individuals
from independent but intertwined fields who all claimed authority and expertise about
calendrical astronomy. These debates had their origins in the disputes of the tenth century
that had occurred between different calendarist lineages as they formed within the Bureau
of Onmyō. The results of these debates had initiated additions to the practice of the
calculation of the calendar using the Xuanming li system of astronomy. These additions
were largely unwritten, but circulated among the members of the Bureau of Onmyō.

33
It was also a potentially dangerous strategy, as shown by the failure of Katsuragi Shigetsune in his solar
eclipse prediction in the tenth century.

34
The visibility of a solar eclipse depends on the observer’s relationship to the moon’s shadow. The
relative size of the moon to the earth means that even on the portion of the earth facing the moon, only a
small portion see the eclipse as a total or annular eclipse, and the rest see a partial eclipse, with the amount
visible decreasing with the distance from the path of the eclipse. Kyoto is time-zones away from Chang’an
(Xian) in China, so this distance would also have an effect.

230
Ironically, it was the actions taken by the early heads of the Kamo lineage to sidestep
intra-Bureau disputes—specifically, the utilization of outside experts as
collaborators—that would launch the debates of the eleventh century. Specifically, it was
the incorporation of mathematically-skilled Buddhist monks into the production of
calendrical knowledge that destabilized the authority of official calendrical astronomy.
The effect that esoteric Buddhist monks had on the practice of Japanese
calendrics was not to drive the attention of the court away from the potential for error in
eclipse prediction. Instead, their efforts focused the attention of the court on eclipse
prediction, and on the potential in other, non-official, sources for valid astronomical
knowledge. Esoteric Buddhism did in fact greatly influence on the practice of calendrical
astronomy at the Japanese court, but this was not the influence that many scholars have
previously assumed.

A Dissenting Astrologer: The Buddhist Monk Shōshō
The historical sources show that Buddhist monks had been collaborators with
official calendarists of the Kamo lineage before the end of the tenth century. Even so, the
exact details of this participation remain difficult to pinpoint. Although Jintō
35
seems to
have participated in the production of the official calendar—or at least in eclipse
prediction—into the early 1020s, precisely when he stopped and whether he was replaced

35
dates unknown. His career is discussed in the previous chapter.

231
by one of his disciples or another monk is unclear.
36
Historical sources are largely silent
about calendar-making monks in this early period.
By contrast, the Buddhist monk Shōshō of Kōfukuji ( ) temple in Nara
leaps out of the records as one of the most influential characters in the history of
calendrical astronomy in Japan. Specifically, Shōshō made arguments about the
importance of the invisible in calendrical calculations, thus reviving points that had been
in dispute between the Ōkasuga and Katsuragi in the tenth century. His career also marks
a turning point in the role of Buddhist astrologer-monks in court life. Now, instead of
collaborating quietly with official calendarists, post-Shōshō, Buddhist monks appear in
the historical record as challengers to the Kamo. Such debates between these two groups
led directly to the uncertainty and doubt shown by Fujiwara Munetada as he watched for
the eclipse of 1095.

Shōshō’s Activity as an Astrologer
In previous scholarship the split between the Kamo calendarists and the
Buddhist astrologers ( sukuyōji) has been dated to a calendrical dispute between
Shōshō and Kamo Michihira
37
in 1038. Signs of this upcoming conflict, however, can be

36
A monk named Ryūshō ( ) is listed in the 1105 petition by the Kamo lineage preserved in Chōya
gunsai 15 (discussed in the previous chapter), but he has largely been ignored as a historical Buddhist
astrologer by historians. See, for example, Yamashita Katsuaki, Heian jidai no shūkyō bunka to Onmyōdō.,
323. A reference to a Ryūshō of Kōfukuji ( ) temple can be found in a record of the Yuima-e lecture
of 1038 (Dai Nihon shiryō vol. 2, part 29, 186), but more details are not forthcoming. Whether or not
“Ryūshō”’s activity would have predated or postdated the controversy with Shōshō described below, is
likewise unrecoverable.

37
, d. 1083.
232
found in the earliest accounts of Shōshō’s career as an astrologer and consultant on
calendrical matters. His first appearance as a consultant on astronomical matters reveals
clues to how his interpretation of the proper calculation and interpretation of the cosmos
differed from that of the Kamo calendarists.
Much about the early portion of Shōshō’s career is vague. By the time he appears
in the historical record, he had already drawn the attention of some very prominent
officials, who are employing him as a ritualist and consultant.
38
Indeed, sometime before
1027, he had obtained the trust and ear of then Minister of the Right, Fujiwara Sanesuke.
As a result, when a lunar eclipse occurred as predicted by Instructor of Calendrics Kamo
Morimichi (Jintō’s collaborator), Sanesuke asked Shōshō about its implications for
Sanesuke himself.
39
Shōshō’s first appearance on the public stage was thus tied to
Buddhist horoscopy.
As a minister of the court, Sanesuke was in a yin position to the ruler’s yang.
40

The moon symbolized yin among astronomical bodies, and by that logic, phenomena
involving the moon presented a potential omen concerning ministers of the court as much

38
This ignores for the time being Shōshō’s appearance in the Sekidera engi ( ), a temple
foundation narrative. In it, he is recorded as having a dream related to the future revival of Sekidera in 1021.
If this tale in any way reflects Shōshō’s career, he was a resident of a temple in Omi at the beginning of the
1020s. Considering that he was consulting with Fujiwara Sanesuke in the capital in 1027, presumably he
would have had to relocate to a location closer to Kyoto by that time. The Sekidera engi’s historical
reliability, in any event, has yet to be established.

39
Shōyūki Manju 4/3/15. The eclipse occurred on Manju 4/3/14, and is also recorded in Shōyūki.

40
On the flexible nature of yin and yang roles, it is possibly best explained in Angela Zito, Of Body &
Brush: Grand Sacrifice as Text/Performance in Eighteenth-Century China (Chicago: University of Chicago
Press, 1997). Although her analysis is on material from a much later time period, the logic of yin and yang
that she describes comes from a much earlier time, and is thus useful for deciphering the multi-layered
meaning of such cosmological symbolism.

233
as it might also concern the women of the back palace.
41
Such correspondences
circulating in memorials interpreting astronomical events at court; Sanesuke himself, as a
sometime student of such texts, would have been well aware of the implications.
42
There
were also, however, additional factors that made this particular lunar eclipse a worrisome
one for Sanesuke.
Even before the lunar eclipse took place on the fourteenth day of the third
month, Sanesuke had been informed that the eclipse was due to take place in what
Western astrologers often refer to as the First House, the part of the sky corresponding to
an individual’s lifespan and fate. The Twelve Houses of Western horoscopy had been
incorporated into Sino-Indian horoscopy at some point in China, as shown through the
two surviving horoscopes that are preserved in Japan.
43
Sanesuke was apparently aware
of the meaning of the First House at least at this time, even though references to the
Twelve Houses do not appear much in Japanese-language sources. Far more common
were references to birth constellations ( Jp. honmyō-shuku), parts of the sky that
corresponded particularly to an individual as determined by the date and time of that
individual’s birth.
44

41
This rationale was explicitly stated in tenmon divination reports on lunar eclipses. See the example
preserved in Chōya gunsai 15: Tenmondō.

42
As shown in chapter two, Fujiwara Sanesuke had a history of acquiring and studying tenmon
memorials.

43
See footnote 63 in chapter four.

44
There is no other reference to the First House in courtier diaries indexed in the Historiographical
Institute database (accessible from http://www.hi.u-tokyo.ac.jp/index.html).
Much more common was use of the Ch. xiu, Jp. shuku—particularly the honmyō-shuku—to
determine which individuals would be particularly effected by an event. As this term shows up far more
234
Due to this worrying location of the event, Sanesuke took precautionary
measures. To start with, he took care not to observe the eclipse himself.
45
Afterwards,
and after Kamo Morimichi had been rewarded for predicting the previous evening’s
eclipse, Sanesuke summoned Shōshō to consult with him about the possible dangers
posed by this eclipse.
At this point, Shōshō informed Sanesuke that although the eclipse began in the
House that would correspond to Sanesuke’s life, the moon had moved to another House
before the eclipse ended. This movement changed the meaning of the eclipse for
Sanesuke’s health: “this can be said to reduce
46
[the effect].”
47
Shōshō also informed
Sanesuke how the presence of Jupiter impacted the effect of the eclipse.
48

In this case, Shōshō demonstrated to a potential patron that he was able to
interpret the implications of an astronomical event for an individual by using information
laid out in astrology texts written and translated by Buddhist monks working in China.
He also showed his specialized skills in calendrics, as he was able to speak with

frequently in the context of lunar eclipses (for example, in 1031, when a lunar eclipse occurs in the birth
constellation of the reigning Go-Ichigō Tennō, Shōyūki Chōgen 4/7/17) the possibility that the 1027
reference refers to a honmyō-shuku rather than honmyō-kurai ( , “Location of Fate” or First House)
should be considered. However, given Sanesuke’s consistant use of honmyō-kurai over the two days, and
the timing of the lunar eclipse as it occurred also seems to fit the honmyō-kurai reading more than that of
honmyō-shuku, so at least initially I have interpreted the entry in this way.

45
In this case, (Jp. mizu) refers to Sanesuke not observing the eclipse,
not to an eclipse that was not visible ( Jp. arawarazu), either due to a failed prediction, or cloud cover
or some other obstruction. Shōyūki Manju 4/3/14.

46
Literally “it resembles a reduction.” Shōyūki Manju 4/3/15.

47
Shōshō then elaborated, saying that by this logic, the effects of the eclipse on the House in which it
ended would be amplified. ibid.

48
Jupiter appears here in the text as Jp. mokusei. Ibid.

235
confidence about where in the night sky the eclipse began and ended. While it remains
possible that this likewise was determined through the force of observation, Shōhsō had
determined the position of the start and end of the lunar eclipse through calculation.
49
As
mentioned in the previous chapter, remarks in Enchin’s Sasa gimon ( Record
of Doubts) show that the movement of the moon and its varying speed across the sky was
a matter of particular concern for Buddhist astronomical texts.
50
What precise method he
used, however, is unclear.
51

Furthermore, Sanesuke showed faith in Shōshō’s abilities. Judging by surviving
diary entries, Sanesuke had no personal skill in either astronomical observation or
calculation, despite his interest in such matters and desire to see that the interpretation
and prediction of astronomical events were performed accurately. Sanesuke’s personal
knowledge of astronomy was largely textual, based on archived materials such as a
collection of tenmon reports he had received from his retainer-client ( Jp. keishi),
the Instructor of Tenmon, Abe Yoshimasa.
52
Yet that tenmon-related material was

49
This becomes clear once the references to Houses are analyzed. The moon remained in the same xiu for
the eclipse ( or Root). As an eclipse of long duration, however, it appeared to move across the sky as the
earth rotated—to a great enough degree that it can be said to have switched houses.
However, the eclipse began after the moon had already risen in Japan—Shōshō either had a
faulty (or unique) interpretation of the Houses (as the First House is below the eastern horizon), or he had
made a calculation error. As places further west would have seen the eclipse lower in the sky, or not seen
the start of the lunar eclipse, it can be said that Shōshō’s statements about the lunar eclipse were accurate
for methods designed for the continent of Asia. Which all formulized systems of calendrical astronomy
were.

50
See chapter four.

51
As shown below, there is evidence that Shōshō may have been more involved with the Xuanming li
calendrical system than the Futian li.

52
. See chapter two for more information on this relationship.
236
limited to the interpretation of astronomical phenomena at the level of state-wide
influence—for personal implications, it was necessary for Sanesuke to consult a Buddhist
astrologer.
53
And Sanesuke was apparently impressed by Shōshō.
Subsequently, as far as Sanesuke was concerned, evidence of Shōshō’s ability in
astronomical calculation only continued to accumulate. Shōshō provided another
demonstration of his skill with regards to a lunar eclipse later that same year. Although
the court’s official calendarists had not announced a lunar eclipse for the sixteenth day of
the ninth month of 1027, somehow Sanesuke had received word that one might occur. He
noted the apparent dereliction of official duty in his court diary. Was it an error on their
part, he wondered. To investigate the matter, he summoned Shōshō, who had clearly
made himself available to answer such calls from the Minister of the Right.
54

Shōshō hypothesized that the eclipse was not announced officially since it would
occur during the day according to the calculations of calendrical astronomy—that is, the
lunar eclipse began and finished while the moon had not yet risen, and was thus invisible
to an observer in Japan. According to Shōshō’s calculations, the eclipse would have
begun about an hour after noon and end before sunset.
55
This second lunar eclipse of

53
This is the point made by Mitsuhashi Tadashi in his study of the influence of esoteric Buddhism on
Heian courtly religion. Mitsuhashi Tadashi, Heian jidai no shūkyō to shinkō Girei.

54
Although the entire account is written as the entry for the sixteenth day of the ninth month (Shōyūki
Manju 4/9/16), the entry itself says that Sanesuke consulted Shōshō “the following day” ( Jp.
yokujitsu). As a result, the entry was composed at least in part on the seventeenth of the month or later, and
may have in fact been composed over multiple days.

55
A total lunar eclipse, beginning at the double-hour of the Horse, 3 koku ( ) 34 bun ( ); maximum
eclipse at the double-hour of the Sheep, 8 bun; and the end of the eclipse in the double-hour of the Rooster,
237
1027 was a daytime lunar eclipse, the equivalent of the nighttime solar eclipses that had
caused such controversy in the previous century. The mathematical calculations of
calendrical astronomy had predicted the eclipse, but it was equally clear that the eclipse
would not be visible from Japan’s capital. The level of detail in this report attests to
Shōshō’s ability to run the calendrical calculations necessary for eclipse prediction. He
had evident skills—even if,the accuracy of his work, this time, could not be verifiable
through observation.
Even though Shōshō provided Sanesuke with a rationale for why the court’s
official calendarists would not have officially announced the eclipse, he did not share
their opinion about the innocuousness of invisible eclipses. The removal of nighttime
solar eclipses from the mathematics of the Xuanming li as performed in Japan, which had
occurred in the previous century, had resulted from the extension of a logic that argued
that eclipses that could not be seen had no effect.
56
It was this logic of a connection
between invisibility and harmlessness that lay behind the ritual countermeasures
employed by the court: blinds lowered, the tennō kept indoors, and monks enjoined to
pray in clouds to cover the sky. Although the “second” lunar eclipse of 1027 is the first
time this logic is explicitly applied to lunar eclipses in Japan, it is clear that the rule was
thought to apply not only to the sun but also to the moon. Yet Shōshō was arguing for the
dangerous significance of even invisible eclipses.

at 1 koku 2 bun. For the times set for sunrise and sunset through the year, see Bock, Classical Learning and
Taoist Practices in Early Japan: With a Translation of Books XVI and XX of the Engi-Shiki.

56
See the discussion in chapter three.

238
This position clearly influenced Sanesuke’s subsequent reaction to this
astronomical “non-event.” In response to a follow-up report from Shōshō on the
implications of the “invisible” lunar eclipse, Sanesuke commissioned a nine-day,
two-monk recitation of the Humane Kings Sutra ( Ch. Renwang jing, Jp.
Ninnō-kyō) on behalf of his wife and daughter.
57
In contrast to the position held by the
Bureau of Onmyō and validated through court precedent, Shōshō’s position was that even
if an eclipse was not visible, it still existed. And if an eclipse appeared in calculations,
then its effects could be known. The irony was that this position was opposite to that for
which Buddhist monks were rewarded when clouds obscured troubling parts of the sky,
as would Ryōshin be in 1095.
58
It is difficult to say how widespread the position held by
Shōshō was. Most monks who worked in horoscopy or calendrics were also ritualists, as
Shōshō himself was. Whether he, or any other such monk, felt any tension between these
two positions on invisibility and influence, or its lack, is hard to say. If Shōshō was aware
of a logical difficulty, he did not speak of it.
The fact is, there was also a logic of practice present in Shōshō’s position of
counting the influence of astronomical events that were not visible, however—a logic
that was drawn from the practice of Buddhist horoscopy itself. There were many invisible
elements in Sino-Indian astronomy and horoscopy. Sino-Indian astronomy dealt

57
. Shōyūki Manju 4/9/21.
Sanesuke was particularly protective of his daughter, and apparently spared no expense on her
health or other needs. This concern is well documented in his diary. See Shigeta Shin'ichi, Kaguyahime no
kekkon: Nikki ga kataru Heian himegimi no endan jijō (Tokyo: PHP Kenkyūjo, 2008), which uses
Sanesuke’s diary account of his daughter’s marriage as its central case.

58
See the description at the beginning of this chapter.

239
systematically with things that could not be seen, not only determining the location of
where the planets would be in the daytime sky or under the horizon,
59
but also with two
additional “planets” that were added to Western horoscopy in India: Rahu ( Jp.
rago) and Ketu ( Jp. keito).
Horoscopes drawn up by Buddhist monks in Japan take much from the Western
tradition of astrology developed in Greece, down to the graphic depiction of the cosmos
and astronomical bodies in a circular diagram.
60
Such a diagram invariably includes all
of the constellations of the zodiac
61
(and ecliptic, from Chinese astronomy
62
), even
though only a portion of this sky is visible at night at any time of the year. All of the
planets are represented, even those that are invisible due to their position in the daytime
sky or because they were below the horizon at night. The houses of Ptolemaic astrology
divide the entire sky, both above and below, into twelve equal sections beginning at the
eastern horizon. These unobservable houses and hidden planets are still thought to exert a
significant influence on an individual’s life and fate, and so things that are not visible

59
As shown by their locations on the two surviving birth horoscopes. See footnote 63 in the previous
chapter.

60
Greek horoscopes were circular. Medieval European horoscopes, as well as Hellenistic and modern
Indian examples, use a square format where the twelve houses are represented around a center square, in a
formate that resembles a subdivided 3 x 3 magic square. The resemblance to the Chinese magic square in
form has led Johannes Thomann to suggest that this format was a reverse transmission over the Silk Road.
Johannes Thomann, "Square Horoscope Diagrams in Middle Eastern Astrology and Chinese Cosmological
Diagrams: Were These Designs Transmitted through the Silk Road?," in The Journey of Maps and Images
on the Silk Road, ed. eds Phillipe Forêt and Andreas Kaplony (Leiden: Brill, 2008).

61
Specifically, the Greco-Roman zodiac.

62
That is, the 28 xiu of Chinese astronomy, which were incorporated through a translation into the Indian
nakshatras. See the previous chapter.

240
must still be known. The invisible could not be discounted in Western astrology, let alone
in its offshoot, Sino-Indian astrology as it was practiced in Japan.
63

Rahu and Ketu, which were added in Indian astronomy, also had to be accounted
for, even though, unlike the classical planets, they were never visible. These two “planets”
are in fact the lunar nodes: the points at which the lunar orbit intersects the ecliptic (the
sun’s apparent path in the sky). The moon’s orbit, viewed from the earth, is offset by
about five degrees from the ecliptic. As a result, the sun and moon are on different “paths”
that cross at two points—the lunar nodes. Over time, these nodes “move” across the night
sky, as the points where the lunar orbit and ecliptic intersect shift. It is only when the
conjunction of the sun and moon, or the new moon, occurs very near one of these two
points that a solar eclipse is visible from earth.
64
The movement of the nodes, and the
limiting factor of the tilted path of the moon (the “white path” Ch. baidao Jp.
hakudō in Chinese astronomy), is one reason solar eclipses are so hard to predict. The
conceptualization of these points in the lunar orbit as “planets” with regular motions and
a periodicity that could be calculated would have been a valuable innovation for the
prediction of solar eclipses bythe Indian astronomers who included them in their system

63
The qualification here is due to the lack of surviving horoscopes in the Chinese context. Further
examination of surviving “fate calculation” texts from the Chinese tradition should determine how
influential the horoscopy tradition might have been in Japan. Nevertheless, its practice by Buddhist monks
in Japan indicates strongly that it had also been present in China.

64
The earth’s shadow, or umbra, is much larger relative to the moon’s size than the moon’s umbra is
relative to the earth’s size. As a result, lunar eclipses are not only visible to a larger part of the earth’s
surface at once when they occur, they can also occur when the moon is about twice the distance from these
notes in the sky than would permit the viewing of even a partial solar eclipse. As a result, although solar
eclipses occur slightly more often than lunar eclipses, they seem less frequent, because they are visible to
fewer people.

241
of astronomy.
65
Present in astronomical calculations, they also became part of the system
of horoscopy, and were also personified as malevolent deities. As deities, they were
incorporated into the Sino-Indian astrological system compiled by Buddhist monks in
China during the Tang Dynasty. Through this method, they arrived in Japan.
It is therefore not particularly surprising that, even though there was a strong
tendency in both Japanese courtly practice and the ritual logic of its esoteric Buddhist
monks to consider invisible astronomical phenomena as non-events, Shōshō should have
taken a different point of view. It seems natural that Shōshō, in determining for Fujiwara
Sanesuke what the implications of astronomical events was using this system of
Sino-Indian astrology, would include in his divinations the impact of events that he knew
from his calculations would occur or had occurred, even if they would be never be visible
to the eye. He gives the effects of such non-observables, such as the “second” lunar
eclipse of 1027, the same weight in his advice as events that could be observed, such as
the first lunar eclipse that year.
Shōshō’s role as a consultant to Fujiwara Sanesuke not only shows him of
providing displays of skill and advice to attract powerful patrons, but it also presages
Shōshō’s subsequent visible role in public debates on calendrical astronomy. Shōshō was
not hesitant in providing a different interpretation of the “second” lunar eclipse of 1027,

65
A further convenience is provided by the fact that Rahu and Ketu are always directly opposite each
other, so calculating the position of one means that the position of the other can be easily derived without
further effort. It is probably for this reason that Rahu has retained astrological significance in multiple
cultures, as in modern Thailand, while Ketu is now more obscure. Such relative obscurity is reflected in the
entries for keitosei ( ) and ragoshō ( ) in Shōgakkan kokugo jiten henshūbu, ed. Nihon kokugo
daijiten.

242
even though the Bureau of Onmyō had already, through its silence, declared that eclipse a
non-event. The gap between Shōshō’s and the Kamo calendarists’ understanding of the
cosmos, particularly on the relevance of the invisible (and therefore of the tradition of
Japanese calendrical practice,)would play the pivotal role in the first public dispute to
occur between these two camps.

The Dispute over the Solar Eclipse of 1028
While the “breakdown” of the alliance between the Kamo calendarists and
Buddhist monks has been typically dated to 1038 by scholars,
66
clear signs of conflict
appear imuch earlier. Indications of the coming conflict can be found in the historical
record of the 1027 lunar eclipses, but it is the controversy over the 1028 solar eclipse that
was the most telling sign of things to come. The lunar eclipses of 1027 showed that
Shōshō had an understanding of calendrics similar to that of the official calendarists of
the Bureau of Onmyō, but yet with significant differences on some points (such as
visibility) that placed him in a separate tradition than that exemplified by Kamo
Morimichi, the head Instructor of Calendrics at the time. These differences are only
inferable from his statements about the lunar eclipses, however. Shōshō had not yet
launched an outright attack on Morimichi, even as he discussed with Sanesuke how he
and Morimichi differed. In 1028, by contrast, such differences would be made public
during a dispute over a solar eclipse prediction.

66
This is the position presented in an early paper by Yamashita Katsuaki, reprinted in his 1996 volume,
and subsequent scholars such as Nakajima Wakako have also adopted it. Yamashita Katsuaki, Heian jidai
no shūkyō bunka to Onmyōdō., 322-324; Nakajima Wakako, "Genji monogatari no Dōkyō, Onmyōdō,
Sukuyōdō ".
243
During the previous year Shōshō had, through conversations and
correspondence with Fujiwara Sanesuke, positioned himself as an expert who could be
consulted about astronomical predictions made by others. Now, in 1028, Shōshō
presented himself to the court as a whole as someone with the knowledge and expertise to
critique the work of the official calendarists of the Bureau of Onmyō. He did this by
inserting himself into a court debate over an eclipse prediction, which had been sparked
by an anonymous attack on the senior calendarist of the Bureau of Onmyō.
The controversy of 1028 began with an unattributed critique on the official
designation of an upcoming solar eclipse as a nighttime eclipse. On the first day of the
third month of the first year of the Chōgen era,
67
Minamoto Tsuneyori
68
wrote in his
diary that he had received a report on an eclipse predicted for that day. According to this
report, it was to be a partial solar eclipse (8/15
th
or about 53% eclipsed). It would start at
approximately 3 AM,
69
reach maximum eclipse in the middle of the next double-hour or
around 5 AM,
70
and end around 7 AM.
71
This time frame put most of the eclipse at night.
However, the very end of this predicted eclipse would have taken place as the sun rose,
and might even have been visible once the sun rose above the surrounding mountains,

67
March 29, 1028 in the Julian reckoning.

68
985 – 1039.

69
Double-hour of the Tiger, first koku. The bun would have presumably been present in a report at this
level of detail—and they do appear for the other times given in the report. They are just missing here.
Sakeiki Chōgen 1/3/1.

70
Double-hour of the Rabbit, 1 koku 47 bun.

71
Double-hour of the Dragon, 2 koku 61 bun.

244
assuming optimal conditions for observation.
72
But this eclipse had apparently not been
announced in advance, nor was it noted in the official almanac for that year.
73
As a result
the court was not scheduled to go into seclusion. Nor, Tsuneyori heard it said, did anyone
know of the eclipse at all.
Tsuneyori in 1028,was one of two heads of the Royal Secretariat ( Jp.
kurōdo no tō), and he was thus in charge of relaying messages between the throne and
other organs of state. His low rank, however, meant that he was not in a position to

72
Sunrise at this time of year in Kyoto occurs a little before 6 AM. That the end of the eclipse would
occur after sunrise should have been knowable to members of the court reading the document, as, at least
on paper, the opening and closing of the gates were set by sunrise and sunset, with times determined by law
in the past. The 921 Engi shiki ( ) manual of court procedure defines sunrise during the third solar
month (in contrast to the lunar months) as occurring between in the double-hour of the Rabbit at 3 koku 7
bun (for the first day of the solar month, or Ch. qingming jieqi, Jp. seimei setsuki) to in the
double-hour of the Rabbit at 1 koku 7 bun (for the last day of the third solar month, or the last day of
Ch. guyu zhongqi, Jp. kokū chūki). Engi shiki 16, and Bock, Classical Learning and Taoist Practices
in Early Japan: With a Translation of Books XVI and XX of the Engi-Shiki.

73
Sakeiki Chōgen 1/3/1. The use of both Jp. shirusu and Jp. mōsu
may indicate both writing and oral announcement, although in combination the separate meanings are less
clear. The use of the phrase to indicate hearsay ( Jp. unnun) seems to indicate that Tsuneyori is
attributing this information to Senior Secretary to the Council of State ( Jp. Daigeki) Kiyohara
Yoritaka ( ), who had brought the initial report to Tsuneyori.
Interestingly, the attribution of the knowledge that the eclipse was neither officially announced
( ) nor officially noted ( ) to hearsay would seem to indicate that Tsuneyori had not himself
personal access to the official almanac. Tsuneyori’s almanac, if he had one, was likely a copy that he either
made himself or had acquired from a low-ranking member of the Bureau of Onmyō. By this time in the
eleventh century, the official distribution of almanacs to the offices of the court had broken
down—although such distribution was not intended for personal use. The only official almanac
presentations at this time was the presentation to the throne (and by extension the whole court) on the first
day of the eleventh month, and then subsequently to the crown prince, the royal consorts, and those granted
equivalent status ( Jp. junsangō).
The almanacs of lower-ranking officials, for example the almanac for 1199 owned by and used
by Shirakawa Narisuke and used for his diary, could be heavily abbreviated and omit notations that would
be found in the official almanacs. (Unknown, "Narisuke-Ō ki [Gūchūreki]," in Tanaka Yuzuru kyūzō
tenseki komonjo (Tokyo: Tōkyō Daigaku Shiryō hensanjo, 1199). photographic copy of a manuscript from
the Tanaka Yuzuru archive. The original is currently held by the National Museum of Japanese History
(Rekihaku).)
Therefore, it is very possible that Tsuneyori would have not had the ability to confirm or deny
whether an eclipse was officially included in the almanac for the current year using his own personal
archive.

245
suddenly order and enforce eclipse countermeasures. That the message came to
Tsuneyori directly and at the last minute indicates that individuals concerned about the
unannounced appearance of the eclipse were exhausting all possible avenues in hopes of
warning the throne. Tsuneyori certainly made sure that this warning was heard.
Those above Tsuneyori ordered an investigation into the affair. When
questioned about his apparent lapse in duties, Instructor of Calendrics Kamo Morimichi
stated that this eclipse, according to his calculations,
74
was not only a nighttime eclipse,
but also of small magnitude. It was for this reason that it had not been officially noted: it
would not be visible, and whatever effect it could have exerted, were it visible (which it
would not be), would have been minor. By the time Morimichi had answered the inquiry,
however, Tsuneyori had received another eclipse prediction that brought Morimichi’s
statement into question. Tsuneyori also copied this new report into his diary, and added
that “both explanations are different, although they come from men ‘on the same
path.’”
75

This second report came to Tsuenyori from the monk Shōshō. Unfortunately,
Shōshō’s prediction is garbled in Tsuneyori’s account. Even so, the text as it survives
appears to present it as a minor eclipse, just as Morimichi had stated, with only about

74
Literally (Ch. li, Jp. ri or kotowari), the nature of things. Sakeiki Chōgen 1/3/1.
75
ibid.
Although (Jp. dō or michi) in modern usage is thought to refer to the field, in the eleventh
century it referred just as frequently to the group of practitioners of a field as to the abstracted subject
matter. That the author of the prediction that Kiyohara Yoritaka had forwarded (perhaps Yoritaka himself)
and Shōshō were considered of the “same path” is quite interesting, particularly given Yoritaka’s activity as
an expert in Onmyōdō-related texts at about the same time. Yoritaka’s role in hemerology debates is
touched upon in the next chapter.

246
22% of the sun’s face eclipsed.
76
Shōshō predicted the beginning of the eclipse to be
later than the first, anonymous, prediction had, although it was within the same
double-hour.
77
His prediction also had the eclipse ending earlier, about 13 minutes after
sunrise.
78
This eclipse would have been barely visible to an observer in the capital,
whereas the eclipse in the first report Tsuneyori had received would have been much
easier to spot. In the first prediction, when the sun had risen high enough in the sky to be
seen, a careful observer may have been able to confirm the eclipse through observation as
the moon continued to move away from the sun. According to Shōshō’s prediction, the
eclipse would be over before it could be clearly observed.
79

There is no record whether this eclipse was actually sighted by anyone,
although it was treated by some members of the court as it if had been. Shōshō’s report

76
Approdimately 3 and 1/3 parts of 15. What I read here as the second part of the fraction is in fact
written in Jp. warigaki—as an inline annotation—in current print editions of Sakeiki. However,
such a formatting forces a reading of << >> or “15/UNKNOWNths eclipse
([approximately] three thirds).” As three thirds is a whole, this should have been noted as a total eclipse,
and thus appear as or . It seems then that either the formatting as warigaki or else the
numbers are in error. As the warigaki are confused elsewhere in this entry—normally the number of bun in
a time phrase would be put into warigaki and here they are not—it seems the most reasonable course of
action to assume that there was at least a copyist error concerning the formatting of the warigaki in the
transmission of the Sakeiki manuscripts.

77
Double-hour of the Tiger, 7 koku 83 bun.

78
About 11/50ths of an hour later. Shōshō gives sunrise in his report as in the double-hour of the Rabbit,
3 koku 26 bun. The end of the eclipse in his prediction occurred in the double-hour of the Rabbit, 3 koku 37
bun. This leaves a difference of 11 bun between sunrise and the end of the eclipse. The maximum eclipse,
according to Shōshō, would occur in the double-hour of the Rabbit, 1 koku 46 bun, or at roughly the same
time as in the first prediction (double-hour of the Rabbit, 1 koku 47 bun).

79
In fact, NASA’s Solar Eclipse Calculator predicts that in Kyoto, this eclipse would have begun at 5:52
AM, or after sunrise. It would have reached maximum eclipse, with 44.6% of the sun obscured, at 6:43 AM,
and ended at 7:50 AM. This eclipse was thus, at least under good conditions, far more observable than
Shōshō’s prediction would have made it out to be. NASA, "Nasa Eclipse Web Site."

247
included his analysis of the implications this eclipse held for members of the court,
80

although, as shown above, Shōshō did not hold that an eclipse needed to be visible to
have its effects felt.
81
Whether an eclipse could be confirmed by observation or not was
not a concern of his. The danger from any predicted eclipse, even one that would not be
visible, was real and present.
The following day, as was customary for an eclipse that had been confirmed
through observation, the Instructor of Tenmon Abe Tokichika
82
presented an
interpretation of the eclipse to the regent,
83
wherein he advised that the ruler and
Ministers of the Right and Left take extra precautions.
84
While it was typical for
Instructors of Tenmon to attempt to confirm eclipses themselves, as part of their duties
along with the regular observation of the night sky, it is also known that they sometimes
used the observations made by others—particularly when the event was visible only
outside of the capital.
85
Writing a divination report based on a prediction was not much
different from relying on the observations of others—in both cases, it required trust in the

80
Sakeiki Chōgen 1/3/2.

81
Shōyūki Manju 4/9/16. See the previous section.

82
, dates unlcear.

83
At this time the regent was Fujiwara Yorimichi ( 992 – 1074), Fujiwara Michinaga’s son.

84
Sakeiki Chōgen 1/3/2.

85
Standard observations were made twice a night, once at sunset and again before sunrise, judging from
surviving astronomical records. The use of observations made by others is discussed in Hosoi Hiroshi,
"Tenmondō to rekidō: Kodai ni okeru seiritsu no haikei to sono yakuwari."

248
skills of another. Therefore, the presence of this divination report is not conclusive as to
whether the disputed eclipse had been actually observed.
Tsuneyori himself might not have been deeply concerned about whether or not
the eclipse had been sighted—that there had been two predictions that an eclipse would
occur seems to have been enough for him to become concerned about the state of court
affairs. Fortunately, from Tsuneyori’s point of view, it had rained heavily on the day of
the eclipse, reducing any potential effects. The first report that Tsuneyori had seen had
stated that if it was overcast within five days of an eclipse, then the eclipse was not
ominous.
86
Tsuneyori apparently agreed with this view, for when he discussed Shōshō’s
report with Kiyohara Yoritaka eight days later, Tsuneyori criticized Shōshō’s work
because his report and the follow-up horoscope
87
did not include any mention of the
rain’s effect on reducing the eclipse’s malevolent impact.
88
Instead, Shōshō was treating
the effects of the eclipse just as if no rain had occurred at all. “How can something light
be treated as something grave?” Tsuneyori grumbled in his court diary.
89
Yet the stance
that rain would do nothing to reduce the effects of an eclipse would not at all have been
out of character for Shōshō, as shown in his treatment of the daytime lunar eclipse of

86
Sakeiki Chōgen 1/3/1.

87
Jp. myōkan.

88
In his discussion of Sukuyōdō at the Heian court, Mitsuhashi actually attributes the statement about the
rain to Shōshō’s eclipse prediction. I believe this to be mistaken, although the entry is garbled enough that
both readings seem possible. However, Mitsuhashi does not address Tsuneyori’s entry for the tenth of the
month, where Shōshō is criticized for not mention the effects of the rain. It is my current opinion that this
lends weight to reading the 3/1 entry as assigning the statement about the rain to the first report.
Mitsuhashi Tadashi, Heian jidai no shūkyō to shinkō girei.

89
Sakeiki Chōgen 1/3/10.

249
1027. It is not surprising that Shōshō would discount the reduction of an eclipses effect
when rain appeared. Elements that eliminated visibility—time, mountains, the horizon, or
in this case, clouds and rain—simply were not factors that were not included in Shōshō’s
calculations, neither for predictions of astronomical phenomena nor for deducing their
effects.
Despite this uncertainty over whether the Japanese court thought the eclipse
occurred or not, later chronicles would determine a winner among the three predictions. It
was the prediction that had passed through the hands of Kiyohara Yoritaka,
90
the
anonymous first prediction that Minamoto Tsuneyori saw, that formed the basis for the
record in the Abbreviated Chronicles of Japan:
Third month, first day…. In the double-hour of the Rabbit, there was a
8/15ths solar eclipse. The calendarists did not note it; as a result the
Ministry of Palace Affairs ( Jp. Nakatsukasashō) did not
announce it. The court did not take measures.
91

Kiyohara Yoritaka had been the Senior Secretary to the Council of State ( Jp.
Daigeki) at this time, and thus in charge of the official archives of the court that were
used in the compilation of the Abbreviated Chronicles of Japan.
92
For determining the

90
For this reason, some scholars have assumed that Kiyohara Yoritaka himself made the prediction.
(Saitō Kuniji, Kotenmongaku no michi, for example.) That does not seem to be clearly indicated in the
sources, which merely identify Yoritaka as the man who had passed the prediction along to Tsuneyori. It
may be said, at least, that the text is somewhat ambiguous.

91
Nihon kiryaku Chōgen 1/3/1.

92
Even though Hosoi Hiroshi asserts that the Nihon kiryaku was largely compiled from Palace Secretariat
( Jp. Naiki) records (Hosoi Hiroshi, Kodai no tenmon ihen to shisho.), particularly for the later period,
the records of the Secretariat of the Council of State ( Jp. Geki) were also used. Yamashita Katsuaki,
"Shohyō Hosoi Hiroshi cho "Kodai no tenmon ihen to shisho"."

250
winner in a dispute about knowledge for future generations, it helped to control “the
press.”

A final interesting aspect of this debate is the common ground that can be
detected between Shōshō’s prediction and the calculations that led Kamo Morimichi to
disregard this eclipse as a “nighttime” eclipse of little to no impact. A prediction of 3 and
1/3 parts of 15 does not contradict Morimichi’s conclusion that the sun would be “only
very slightly eclipsed,”
93
and that it was therefore insignificant. The calculation that the
eclipse would only be visible for its last few minutes right after sunrise also corresponds
with Morimichi’s conclusion that the 1028 eclipse was a “nighttime eclipse.” In other
words, while it is by no means certain, it is very probable that Shōshō and Morimichi
were using the same calculative methods. This has significant implications for the
practice of Sukuyōdō as it has been assumed by previous scholarship, and the nature of
the collaboration between Buddhist monks and official calendarists.
94

To elaborate, the prevailing assumption among scholars, after the rediscovery
of the story of Nichien’s sojourn in China, has been that Buddhist calendarists primarily
utilized the Futian li system. Certainly, as described in previous sections, it is the Futian li
system that was used to calculate the two horoscopes that survive from medieval Japan.
There are late medieval instances where the Futian li is cited in the calculation eclipse
predictions. Therefore, when Kamo calendarists improved their ratio of accurate to

93
Sakeiki Chōgen 1/3/1.

94
For more details, see the discussion in chapter four.

251
inaccurate eclipse predictions in the twelfth century, some scholars have attributed this to
their adoption of some aspects of the Futian li system for eclipse prediction.
95

However, there are some potential problems with this view. In 1028 Shōshō does
not seem to have been using the Futian li, which had very distinctive mathematics for its
calculations of the sun’s apparent motion.
96
While not enough of the Futian li system
survives to easily determine the differences between Futian li and Xuanming li systems
of eclipse prediction, the close correspondence between Shōshō’s and Kamo Morimichi’s
predictions indicate that Shōshō may well have been using the Xuanming li system.
97

There is a link through which Shōshō may have obtained access to the Xuanming
li methods. Buddhist monks working on the official calendar, if they were to do any
significant calculation work, would have needed access to the canons of the Xuanming li
to complete it. As a result, the access of at least some monks to the Xuanming li system
can be assumed. Since Shōshō was the disciple of Jintō, who had worked in collaboration
with the Kamo on the official calendar earlier that century, Shōshō may well have
obtained access to Bureau of Onmyō texts through his master. This may explain some of

95
Saitō Kuniji, Kotenmongaku no michi., 112.

96
It has been described as approximating parabolic. It would have to approximate the function, as
Chinese mathematics at the time did not have quadratic equations. Nakayama, "The Position of the Futian
Calendar on the History of East / West Intercourse of Astronomy ". See especially the commentary added
to the end of the paper.

97
The Futian li and the Xuanming li had different underlying models for the sun’s apparent motion in the
sky in their tables, and both systems used very different epochs (starting points) for their calculations.
Nevertheless, it remains a distant possibility was that both Shōshō and Morimichi were using
the same Futian li system. However, this seems unlikely—the surviving eclipse predictions using Futian li
are from a later period. Reconstruction of the likely calculations used by Morimichi and Shōshō employing
what is known of both systems, if it is possible, may be the only way to determine the system employed by
them both for sure.

252
the spread of Xuanming li methods to religious institutions in the provinces by the late
thirteenth century, a historical process that has not been much examined. Much as Royal
Secretaries and regents were leaks in a chain of documentary secrecy that allowed for the
circulation of knowledge about tenmon, so too may the monks who worked with the
Kamo in producing the official calendar have been vectors for the spread of calendrical
knowledge.
Furthermore, the major distinction between Shōshō’s and Morimichi’s positions
was not at the level of the calculation of the eclipse. These two men differed on the
significance of visibility in calendrics. Shōshō seems to have been using the same
Xuanming li system of the official calendarists when he began critiquing their work.
However, in 1028 as later, his assumptions about certain elements (particularly related to
visibility) deviated from the tradition of practice that had developed within the Bureau of
Onmyō over the previous century. Shōshō was not practicing in the “Japanese-style” of
official calendrical astronomy.

The Non-Consequences of Error
Another new aspect in the eclipse dispute of 1028, particularly when compared
to those of 918-919 or 938-939, was that this dispute over the visibility of the eclipse had
no effect on the production of the calendar. This indicates that Shōshō was not working
on the official calendar, at least not at that time. He also failed to use this eclipse
prediction to argue that the calendar for that year was incorrect, unlike the Bureau of
253
Onmyō’s own Katsuragi Munekimi or Katsuragi Shigetsune had done a century earlier.
98

Indeed, it is very likely that Shōshō ran his eclipse predictions separately from a
calculation of the entire astronomical calendar. That such a method was possible is
attested by sources from the end of the medieval period.
99
As a result, despite Shōshō’s
attack on the relationship between visibility and effect, the principles established in the
previous debates still held firm for the production of the calendar—at least within the
Bureau.
There were no rewards given out and no victor declared (the Abbreviated
Chronicles of Japan notwithstanding) in 1028. No punishments were meted out either:
there is no record of Kamo Morimichi or the Bureau of Onmyō being forced to officially
apologize for putting the court in danger through their lassitude, as had been the case in
877. This further shows how the critiques of Shōshō and others did not necessarily reach
deep into the issue of calendrical production. Neither did Morimichi personally seem to
suffer any ill effects as the result of his “error.” Perhaps he had his previous reputation to
bank on. Like his father, Morimichi had made a name for himself in the consulting
sidelines of Bureau-related business. He appears in diaries primarily as a diviner,
specializing in a type of mathematical astrology known as the Yang Six Waters method

98
See chapter three.

99
This can be seen in the drafts of calendrical calculations preserved in Unknown, "Kan'ei nenkan reki
suisankōtō." There eclipse calculations are clearly figured indepenedently of the main calendar. The
essential independence of the eclipse cycles is also covered in Sivin, "Cosmos and Computation in Early
Chinese Mathematical Astronomy."

254
( Ch. liuren shizahn, Jp. rikujin shikisen).
100
The reputation he had developed
through his performance of this particular skill might have helped him weather any
incidental criticism. And while Morimichi may not have won over everyone who
mattered (certainly he had failed to convince archivist), he had previously been rewarded
for accurate eclipse predictions: once for the “first” lunar eclipse of 1027, and once for a
solar eclipse in 1021. In the 1021 case, even if he had to share the credit in the court of
public opinion with Jintō, officially he received all the credit.
101
Therefore, Morimichi
must have had some previous reputational capital to use to defend himself against
criticism in 1028.
Following the 1028 eclipse dispute, Morimichi continued to be called upon by the
regent and even Fujiwara Sanesuke, who would prove to be Shōshō’s biggest patron.
Later the same year, Morimichi was summoned to perform a Yang Six Waters divination
to determination of the implications of a haunting affecting the residence of one of the
royal consorts.
102
The following year, 1029, saw a summons from Sanesuke to consult
on the issue of picking the most auspicious day for ceremonies relating to the impending

100
Significant examples of Morimichi’s activitiy can be found for 1027 (Shōyūki Manju 4/9/7) and 1028
(Shōyūki Chōgen 1/10/4). In both cases, Morimichi is credited with performing the divination by himself.
Morimichi does not appear often as a ritualist, with the exception of when Morimichi was called upon to
perform a Seven-Shallows Purification Rite ( Jp. nanase oharae) in 1024. Shōyūki Manju
1/20/28.
For more information on the Yang Six Waters method of divination, see Ho Peng Yoke,
Chinese Mathematical Astrology: Reaching out to the Stars and Kosaka Shinji, Abe no Seimei sen Senji
ryakketsu to Onmyōdō (Tokyo: Kyūko shoin, 2004).

101
See the discussion in chapter four about the distribution of rewards. Sakeiki Jian 1/7/1 compared to
Nihon kiryaku Jian 1/7/1.

102
Sakeiki Chōgen 1/10/4.

255
marriage of Sanesuke’s only daughter.
103
Kamo Morimichi did not disappear from the
historical record after failing to correctly predict a solar eclipse, as had been the unhappy
fate of Katsuragi Shigetsune in the previous century.
104

Nor did the 1028 eclipse debate harm Shōshō’s career. Sanesuke continued to rely
upon Shōshō for astrological advice and predictions relating to personal health and
danger. Later in 1028, Sanesuke employed Shōshō to calculate the fate of his adopted son
and heir Sukehira, who had been suffering from a serious illness.
105
Thus, despite failing
to achieve victory, Shōshō’s participation in the 1028 dispute had succeeded in making
other members of court aware of him and his skills.
106
Not every member of court
received a good impression of him for his pains, however. Certainly Minamoto Tsuneyori
did not have a positive view of Shōshō’s expertise, but his opinion did not apparently
matter for Shōshō’s success at court. Despite naysayers like Tsuneyori, Shōshō
nonetheless continued to receive benefits form his relationship to Sanesuke, as shown by
his inclusion in the rituals performed in the fall of 1028 to ensure the queen-consort’s
107

successful pregnancy.
108

103
Shōyūki Chōgen 2/9/10. See also Shigeta Shin'ichi, Kaguyahime no kekkon: Nikki ga kataru Heian
himegimi no endan jijō.

104
See chapter three.

105
Shōyūki Chōgen 1/9/29.

106
The Chōgen 1/3/1 entry is the first time that Shōshō appears in the Sakeiki diary.

107
Jp. chūgū.

108
It was an offering to the lunar lodges and luminaries ( Jp. sukuyō-kyō), related to the practice
of Sino-Indian astrology. The record is only found in Sanesuke’s diary, indicating perhaps that Sanesuke
was keeping track of his client’s career. Shōyūki Chōgen 1/7/1.

256
Kamo Morimichi died in 1030.
109
According to an entry in an index for Fujiwara
Sanesuke’s court diary,
110
Shōshō was ordered to present an official calendar later that
same year—timing that seems hardly to have been a coincidence.
111
Within the Bureau
of Onmyō, Morimichi’s son Michihira had taken his father’s place, but Shōshō’s
cooperation was elicited by the court, possibly at the suggestion of Sanesuke, to ensure
that the calculation of the official calendar proceeded smoothly. There were apparently
no problems for a few years,
112
and precisely how many calendars Shōshō participated in
is unknown. But in 1038, the relationship between Michihira and Shōshō would decline
spectacularly, and involve the entire court in a case of competing calendars.

Competing for the Official Calendar, 1038-1039
The controversy that erupted in 1038 between Buddhist monks’ practice of
calendrical astronomy and that of the official Kamo calendarists started out quietly, yet
the consequences would be felt for over a century afterwards.
There is no surviving record of the presentation of the official calendar for
1039. The calendar—the designation of the months and days for the coming year—was

109
He died Chōgen 3/3/4.

110
Not all of Fujiwara Sanesuke’s Shōyūki survives, but an index he made of his diary entries does
survive for some of the years that are now lost from the main text. This subset of the Shōyūki corpus is
known as Shōki mokuroku. This text can be found included in the Dai Nihon kokiroku edition.
(Tōkyō Daigaku Shiryō hensanjō, ed. Shōyūki.)

111
Shōki mokuroku 13, Chōgen 3/7/29.

112
Although Shōshō may have been replaced by Ryūshō in 1031. See the discussion above about Ryūshō
as a largely unknown figure.

257
presented in the form of an almanac, which was an annotated calendar complete with
hemerological
113
and astronomical information. While diary entries from courtiers
survive for the designated day (the first day of the eleventh month of the second year of
Chōryaku, or 1038) none mention anything about the presentation of the calendar.
However, this was not uncommon for the eleventh century or afterwards. While the
presentation of the calendar on the first year of a new Rule Cycle was a grand event with
a banquet and awards, in most years the presentation of the calendar was routine business.
Even though there would be “no question” of canceling or postponing the presentation in
normal, or even abnormal years,
114
in most years the tennō himself did not even attend
the ceremony.
115

The number of almanacs that were officially produced dwindled from the
tenth-century onwards as well.
116
At the same time, however, private demand for
almanacs was increasing. As a direct consequence of the gap between official supply and
popular demand, the high nobility obtained their copies of the calendar privately from
members of the Bureau of Onmyō, or else borrowed and copied those of others. This
situpation prompted the appearance of “bespoke” almanacs, custom tailored to suit the
needs of the noble recipient. For example, extra blank lines might be inserted between

113
Divination based on time-keeping or temporal elements.

114
Words used when the cancellation was proposed in the fourteenth century, after the Northern Court
had been forced out of Kyoto.

115
Momo Hiroyuki, "Sakutan tōji."

116
See Yamashita Katsuaki, Heian jidai no shūkyō bunka to Onmyōdō.

258
day-entries for convenient diary writing.
117
Different sets of divinatory notes could be
added or requested for these almanacs as well.
Such variation, however, had its limits. Court ceremony still had to be regulated
and schedules rectified, and the almanacs in the palace and in the Bureau of Onmyō
remained available for consultation should confusion or controversies arise. Despite the
presence of a variety of supplementary divinatory notations on almanacs—including
material taken from Sino-Indian astrology—in the eleventh century the basic set of
divinatory notes remained stable, based on those found in the official almanac presented
to the throne itself ( Jp. mi-guchūreki). Furthermore, the day and month had to
be consistent across the court for court business to function properly. This need for
standardization, and the existence of standards, would prove one of the biggest stumbling
blocks for Shōshō in 1038 as he attempted to displace the official calendar. Shōshō, as
previously shown, was a man already known to have “non-standard” views on
astronomical events.

Initially, it seemed as if there would be no controversy at all about the calendar
for 1039. The annual presentation of the next year’s calendar in the eleventh month of
1038 was performed without a hitch.
118
From Fujiwara Sukefusa’s
119
standpoint, as

117
See the variety of blank lines in the catalog of surviving almanacs, ranging from none to five, found in
Atsuya Kazuo, ed. Guchūreki o chūshin to suru koyomi shiryō no shūsei to sono shiryōgaku-teki kenkyū.

118
As indicated by the lack of comment about the presentation in the sources.

119
1007 – 1057.

259
revealed in his diary Shunki, the problem with the calendar for the next year only became
clear near the end of the eleventh month, on the twenty-seventh day.
Sukefusa was the eldest son of Sanesuke’s adopted son and heir, and the scion of
two lineages renowned for their scholarship.
120
He had been made one of the heads of
the Royal Secretariat that very year. It was in that role that he was called in by the regent
and made aware of the trouble, so that he could convey this information to others. If
Sukefusa had any thoughts or knowledge of any issue with the calendar before the end of
the eleventh month, he had kept it to himself.
121

In his court diary entry for the twenty-seventh day of the eleventh lunar month,
Sukefusa writes that after informing him about matters of provisioning for the upcoming
Kamo Festival,
122
Regent Fujiwara Yorimichi turned to other business:
Then [the Regent] also said: the calendar made by Instructor of
Calendrics [Kamo] Michihira, in the length of the months and in the
various almanac notations, differs greatly from the calendar made by

120
Sukefusa was not only the eldest son of the adopted heir of Sanesuke, and thus heir to the Ononomiya
school of court protocol, but through his mother he could trace his lineage back to Sugawara Michizane.
Ryūfuku Yoshitomo, "Fujiwara No Sukefusa," in Kokushi Daijiten, ed. Kokushi daijiten henshū iinkai
(Tokyo: Yoshikawa kōbunkan, 1992–1997).

121
At least according to what records survive. Unfortunately, the surviving diaries and court records for
1038-1039 are rather fragmentary, with Sukefusa’s Shunki as the most complete source. (See the charts in
Minakawa Kan'ichi, "Kiroku nenpyō, kiroku mokuroku," in Kokushi daijiten, ed.
Kokushi daijiten henshū iinkai (Tokyo: Yoshikawa kōbunkan, 1992–1997).) However, Shunki too is rather
fragmentary for these years, which are among the first to survive from this source. Chōryaku 2 (1038) has
entries for the 8
th
through 12
th
months that survive in some form, while Chōryaku 3 (1039) is limited to the
period from the tenth month through the intercalary twelfth month. Fortunately, the date for the official
presentation of the calendar is present in both years.

122
Kamo rinji-sai. While in name this procession and offering to the Kamo Shrines was
“extraordinary” or “adhoc” (rinji), it had become by the tenth century an annual feature of religious
observances at court. The name rinji-sai was preserved in order to distinguish it procedurally from the
Kamo Festival ( Kamo matsuri) of the fourth lunar month (now May), wherein the Kamo Royal
Priestess ( Kamo sai-in) was the central figure.

260
Shōshō. Shōshō had received an official decree permitting him to work
on the calendar. As a monk, he cannot add his name as a signatory.
[But] regarding the coming year, he submitted [a calendar] which
differs greatly from Michihira’s. Shōshō and Michihira had previously
disagreed, and therefore [Michihira] said he would not collaborate
with Shōshō. Now, Shōshō is distributing his calendar here and there.
It differs [from the official calendar]. This is a highly suspicious
thing.
123

Yorimichi then told Sukefusa that Michihira should be summoned and asked to
give a statement, which Sukefusa was to forward to both Yorimichi and Sukefusa’s own
grandfather Sanesuke, at that time still Minister of the Right and a major figure at court.
Shōshō would be questioned orally. Both statements then should be presented to the
throne, and the tennō’s decision implemented.
124

The importance of an accurate calendar for the state made this a potentially
serious issue. The calendar, after all, represented the throne’s direct connection to Heaven
(or the heavens), and if a calendar came from another source, this would negate its
function as a symbol of legitimate rulership. It could even represent a claim to legitimate
rule from another quarter. This was the entire rationale predicating laws against unofficial
calendars in China, and it had motivated the monopolistic control by the court on the
regulation of time in Japan’s early bureaucratic state.
125
Curiously, however, this point
was not raised in the discussion of the 1038 and 1039 disputes over the calendar.

123
Shunki Chōryaku 2/11/27.

124
This section is full of status-related information, showing the multiple uses of the verb Jp. ohosu
to indicate relaying information, receiving orders, and implementing them.

125
See the discussion in chapter three, and also Ooms, Imperial Politics and Symbolics in Ancient Japan:
The Tenmu Dynasty, 650-800. For China, see Richard J. Smith, Chinese Almanacs (Oxford: Oxford
University Press, 1992).

261
Certainly by this time unofficially copied calendars were already circulating around court
without criticism or comment, which perhaps weakened the sense that Shōshō’s
unofficial calendar was an attack on the legitimacy of the court’s monopoly on the
regulation of time. Perhaps this very weakening of official calendrical control had even
made Shōshō bolder in distributing his variant, unofficial calendar to interested parties at
court.
Nevertheless, there were practicalities to consider. The entire realm was at least
nominally synchronized using the official calendar. Although by the eleventh century tax
goods from the provinces no longer arrived at court on legally prescribed days, this fact
did not negate the need for a single unified calendar to synchronize affairs within the
court. Even if temporal synchronization between capital and province was less
necessary,
126
the need for a standardized accounting of days had been steadily increasing
at court.
To elaborate, between the ninth and the twelfth centuries, the number of court
events that were included in manuals of court procedure increased markedly.
127
The vast
majority of these events were assigned to particular days on (at least in theory) which

126
Such synchronization had always been tricky. See for example the scheduling of realm-wide shrine
offerings in the Engi shiki—offerings to certain deities were to be made on prescribed days, and yet
offerings for the most important of these shrines were to be presented to representatives to these shrines at
the palace on this day as well. The travel time—often many days—to get the offering to the deity itself was
not factored into the genesis of this system.

127
This can be seen through comparing the number of court events included in the Engi shiki set of court
procedures (921) to those found in the Saikyūki manual or the Ononomiya nenjū gyōji (Fujiwara Sanesuke),
both from the early eleventh century.

262
they were to be carried out.
128
If members of the court did not agree on what day of the
month, or even what month it was, then the court would be unable to function
smoothly.
129
This was the reason that a royal decree deciding the dispute between
Michihira and Shōshō was necessary.
Yet it is curious too that there is no record of official punishment befalling
Shōshō in this case. The unofficial distribution of his calendar had the potential to be seen
as an act of lese majeste—and while no surviving law against the private distribution of
calendars or almanacs exists for Heian-period Japan, the existence of such laws in China
is well attested.
130
And yet Shōshō’s career would survive this, as shown by his
continued activity in calendrical affairs afterwards.
131
In fact, his unofficial almanac for
the third year of Chōryaku may have even enhanced his reputation, at least in some
quarters. Apparently, Shōshō’s almanac held a great deal of appeal for the top members

128
There was during the same period a simultaneous increase in the number of days of official or
personal seclusion ( Jp. mono-imi). The days for mono-imi varied from year to year and person to
person, and tended to be trigged by an event (such as exposure to ritual pollution, Jp. kegare), or as
determined by a divination (using mathematical astronomy). As a result, the number of events delayed or
cancelled due to either mono-imi or kegare directly increased through the duration of the Heian Period.

129
That the importance of the annual cycle of events to the court’s image of itself continued beyond the
Heian Period can be seen in the fourteenth-century war tale, Taiheiki ( “A Record of Great Peace”).
In chapter 24 of this work, there is a list of all the court events to be held in a year, along with the notation
that none of these had in fact been held due to the civil war raging between the competing Northern and
Southern courts. This chapter explicitly states that the result of this symbolic disorder is environmental and
economic collapse, with famine and poor harvest ensuing from the breakdown in the cosmic and political
order. Kristina Buhrman, "Conceptions of the Court and Ritual in Medieval Japan: Examples from the
Taiheiki," in Redefining Identities in Asia: 9th Graduate Student Conference (Harvard East Asia Society,
Cambridge MA 2006).

130
Smith, Chinese Almanacs. Such laws would also appear on the books in Japan in the late seventeenth
century, as the Tokugawa Bakufu began establishing control over the production of calendars. See Peter
Kornicki, The Book in Japan: A Cultural History from the Beginnings to the Nineteenth Century (Leiden:
Brill, 1998), 353-357.

131
For example, in the solar eclipse debate of Kōhei 2 (1059).

263
of the court bureaucracy, even if the regent Yorimichi was opposed to it, as shall be
shown below.
132

The fact that Shōshō had significant support is proven by the length of time it
took for the official decree to appear reaffirming Michihira’s official calendar. According
to later chronicles, this decree was the results of deliberations of the Council of State in
the summer of the following year.
133
However, even that official decree did not halt the
influence of Shōshō’s almanac, as Sukefusa’s own diary makes clear. Although Sukefusa
wrote his diary according to Michihira’s calendar—which, after all, had been presented
as the official calendar of the realm and reaffirmed as such the following year—Sukefusa
still praised Shōshō’s work as a more accurate reflection of the natural world.
Sukefusa himself had obtained a copy of Shōshō’s almanac at some point and
began to compare predictions from both of the calendars to astronomical events in the
natural world. As discussed in the third chapter, the months and the days of the month in
the Chinese-style luni-solar calendar did not merely mark time—these elements of the
luni-solar year also functioned as predictions of the phase of the moon or the rough
position of the sun (as in the case of the 24 solar periods). Therefore, even in years

132
According to Sukefusa’s diary, it seems that Yorimichi was the first to take action against Shōshō’s
unofficial almanac, and his efforts to obtain a “Chinese” calendar to prove the accuracy of Michihira’s
calendar also indicates that he was not a partisan of Shōshō’s work.
Why it was that Shōshō’s calendar would have found such a receptive audience—one that
stretched beyond the Sanesuke-Sukefusa household—remains murky at this point, and an interesting
question for further study.

133
Hyakuren shō Chōryaku 3/5/23. Unfortunately details of the deliberations do not survive, merely the
final decision and decree.

264
without solar or lunar eclipses, the calendar could—in theory—be checked against
natural phenomena. This was often done using the phases of the moon.
On the night of the last day of the Chinese-style lunar month ( Jp. misoka), the
moon should be invisible. On the first day of the month, after the new moon ( Jp.
tsuitachi), a faint sliver of moon was to be seen. The middle night of the month had a full
moon, whether the fifteenth, fourteenth or sixteenth.
134
Adjustments to the calendar, as
well as the variation in the speed of the lunar orbit, meant that this was not always the
case, but it was at least theoretically possible to confirm the days of the month through
astronomical observation. It is unclear how many members of the Heian court performed
such tests, let alone how many possessed the visual acuity to distinguish a second-day
moon from a third-day moon. Very few records of this sort of effort survive—Sukefusa’s
test of 1039 remains a rare documented example of such an effort.
135
No matter the
difficulty of the project, Sukefusa certainly thought that his eyes, and those of others that
he knew, were enough to prove Shōshō correct.
The tenth lunar month of 1039 was the month that differed the most between
Shōshō’s and Michihira’s calendars, and it was to the days of this month that Sukefusa
paid the closest attention.

134
The note in Royal Tyler’s translation of the Tale of Genji on the moon of the fifteenth day is therefore
not fully accurate. There is the possibility that the full moon was indicated in popular usage with the phrase
“fifteenth day,” just as “thirtieth day” served as a shorthand referent for the last day of the month during the
Nara Period, even for those months that held only 29 days. The best discussion of this can be found in
Yuasa Yoshimi, Koyomi to tenmon no kodai chūsei shi (Tokyo: Yoshikawa kōbunkan, 2009).

135
The only one I ahave yet found the eleventh century.

265
Tenth month, first day… clear…. Today, the light of the moon at dawn
did not appear, or so was the word from all those
136
who looked. This
day is Shōshō’s last day [of the ninth lunar] month—it accords
greatly.
137

Second day… cloudy…. Today was cloudy, so whether the moon
appeared or not is unknown. However, yesterday when watching for
the moon, it did not appear. As it was Shōshō’s last day [of the ninth
lunar month], perhaps for that reason it did not appear.
138

Third day… clear, with a white rain now and then…. When the sun’s
light descended, just a short while later, a sliver of moon appeared
about 2 jō [ ] above the mountains in the western sky. Its thinness
was like that of a fishhook, not at all like a [typical] third-day moon.
It’s said that it first appeared today. Today is Shōshō’s second day [of
the month]. Does this not fulfill his explanation? It greatly accords
[with his calendar].
139

While this may seem a disinterested experiment on Sukefusa’s part, later entries, however,
make it clear that Sukefusa had been receiving some help with his observations: on the
fourth day, Shōshō arrived at Sukefusa’s residence to consult with him, and informed
Sukefusa that “today the moon appeared, resembling the third-day moon, rising high.”
140

While Sukefusa’s diary is extant for the period from the tenth month through
the end of the year (the intercalary twelfth month),
141
nowhere else in that period does

136
Literally “ten-thousand men.” This is clearly hyperbole, but the implication is clear that Sukefusa
thought that many of the people at court had observed the night sky and thought it confirmed Shōshō’s
prediction.

137
Shunki Chōryaku 3/10/1.

138
Shunki Chōryaku 3/10/2.

139
Shunki Chōryaku 3/10/3.

140
Shunki Chōryaku 3/10/4.

141
Selections from the Shunki dated to earlier in Chōryaku 3 also survive, although only in excerpted
form. The full context of the diary for these earlier months has been lost.

266
such concentrated attention to the astronomical definition of days appear. As previously
mentioned, the differences between Shōshō’s almanac and Michihira’s calendar were
greatest in the tenth month. Yet the actual divergence between the two almanacs, aside
from the divinatory notations,
142
only amounted to a single day. It is likely, considering
previous instances of diverging calendars, that the months largely matched up with each
other, and with only a few months showing differences—as was the case here, mostly in
the tenth month. The remaining months probably essentially agreed with each other
across both calendars.
Furthermore, it is likely that in the tenth month of 1039 Shōshō himself was
making a particular case for his calendar’s correctness. In previous disputes over the
calendar, the tenth month had been the time when debates were fiercest.
143
Since the
official presentation of the almanac was set for the first day of the eleventh month, the
vigor of disputes in the tenth month can be seen as individuals making the case for their
own interpretation of calendrics before the presentation of the next year’s calendar.
Although by the tenth month the official calendar should have been largely completed, at
least according to a schedule put into law in the tenth century,
144
Shōshō was probably
making a last-ditch attempt to get himself included in the official calendar-making project
once again. The timing matches well with Katsuragi Shigetsune’s successful attempt in

142
Referenced by Yorimichi in Shunki Chōryaku 2/11/27.

143
See the dates of debates discussed in chapter three.

144
Engi shiki 16.

267
937. But if Shōshō was to get a calendar for 1040 distributed after losing the Council of
State’s official approval earlier in 1039, he would have had to make his case quickly.
Notably, however, there is no evidence indicating that Shōshō was included in the
calendar-making project for 1040, nor for any year thereafter.
145
As for why this may
have been the case, the records of Sukefusa’s conversations with the regent are telling. As
shown in his diary entries for the tenth month, Sukefusa found Shōshō’s calendar more
accurate that Michihira’s. It should be remembered, however, that Sukefusa was of the
same lineage (though adopted) as Sanesuke, who was a major patron of Shōshō. If
Sukefusa represents Shōshō’s partisans among the members of the court, however,
Yorimichi stood as Michihira’s supporter, and a very powerful one at that.
The relationship of Michihira and Yorimichi before 1038 is not especially clear,
but in his words and actions Yorimichi stood as a major supporter of Michihira’s calendar
in these court debates. In early 1040,
146
while Sukefusa had once again been called to the
regent’s residence on another matter, Yorimichi took the time to inform Sukefusa of the
ongoing status of his own, Yorimichi’s, investigations concerning the two competing
calendars. Yorimichi’s own research left no ambiguity as to what he thought the truth of
the matter was.

145
The official presentation of the calendar is mentioned without any details by Sukefusa in his diary.
Shunki Chōryaku 3/11/1.
Although, as noted in the previous chapter, the last case of a monk participating in the
production of an official calendar seems not to have been before 1038, as previously assumed
(Yamashita Katsuaki, Heian jidai no shūkyō bunka to Onmyōdō.), but instead a second disciple of Jintō’s,
Nōsan, in 1092.

146
As previously mentioned, the Chinese-style luni-solar year began in what would be late-Janaury or
early February by the Julian reckoning. As a result, the end of the third year of the Chōryaku era and the
beginning of the fourth, discussed in this section, both correspond to the Julian year of 1040.

268
The regent said, “The other day I had a Chinese [lit. Tang] calendar
brought. I’ve heard that Goryeo [lit. Silla] uses the Chinese calendar.
Therefore this past summer, I privately sent to the Commander of
the Dazaifu [in Kyūshū] [for one]. Today it arrived. It is a printed
calendar. Furthermore, this calendar corresponds exactly to
Michihira’s calendar. In the lengths of months there is not one
difference. This is a rare thing. Previously we had the experts of the
myriad fields all estimate and made our decisions [based on
that—now we have stronger proof]. However, although we already
knew truth from falsehood [as the result of the earlier
deliberations],
147
my heart is [now] lighter. For this calendar
matches completely—it accords greatly. In times new and old there
has been nothing like this. [Michihira] should be rewarded, he
should surely be rewarded. Shōshō’s [words] have no use. He
should repent, he should surely repent. All the nobles will be
surprised, but that is to be expected. Tomorrow, present this
Chinese calendar [to the court].”

Afterwards, I heard from Tsunenari that the presentation was
completed.
148

Yorimichi’s defense of Michihira was strong, strong enough to indicate that
Yorimichi might have been utilizing Michihira as an expert on calendrical and
hemerological affairs, much as Sanesuke had utilized Shōshō in earlier years. Michihira
had received an award personally from Yorimichi for correctly predicting a lunar eclipse
in 1031,
149
and Michihira is also known to have served the court as a diviner, just as his
father had. Although no court diary by Yorimichi survives, were there one, it would not
be surprising to find accounts of divinations that Yorimichi asked Michihira to perform

147
This statement about already knowing “truth from falsehood” refers to the deliberations of earlier in
the year that reaffirmed Michihira’s calendar. Hyakuren shō Chōryaku 3/5/23.

148
Shunki Chōyraku 3/intercalary 12/23.

149
Sakeiki Chōgen 4/7/17.

269
for his household, or for the Fujiwara lineage as a whole, as well as less formal occasions
where Michihira was consulted for advice.
It is not Yorimichi’s support of Michihira, but rather Yorimichi’s method of
confirming Michihira’s calendar of 1038, however, that is of particular interest. Where
Sukefusa and other courtiers supporting Shōshō cited the authority of
observation—somewhat ironically, given Shōshō’s position on the unimportance of
visibility—Yorimichi reached abroad to find the authority to bolster Michihira’s side.
Perhaps he did so because observations of nature would not have supported Michihira’s
calendar over Shōshō’s—certainly, Yorimichi knew that he had precedent to support his
actions, as Chinese calendars had been brought to court after a calendrical dispute in
938.
150
That he used Chinese and Korean calendars to “prove” that Michihira’s calendar
was the correct one says volumes about the relative epistemological power of both
observation and Chinese practice at this time. However, the applicability of Chinese
calendars for evaluating the work of Japanese calendrics would be reversed later, within
less than a century.
151

That Yorimichi was using here the authority of a foreign state’s calendars to
support his client Michihira’s position was not an uncomplicated stance for him to take.
Over the past one hundred years, Japanese practice had diverged from the Chinese, both
by changing—in the inclusion of the invisible eclipse rule and in adding provisions to
preserve the Rule Cycle in a calendrical system that did not originally include it—and by

150
See chapter three.

151
See the discussion in the middle of the next chapter.

270
not adopting new calendrical systems as they were developed in China. In fact, Sung
China at this time was using a different system to calculate its official calendars. Goryeo,
however, was still using the Xuanming li, just as Japan was, but this made Goryeo
practice different from that of China (contrary to Yoshimichi’s statement). Which type of
calendar was found to support Michihira’s own calendar is not clear from the sources, but
the underlying system used to calculate that foreign calendar would have been different,
depending on whether it came from China or Korea. If based on a different system of
calendrical astronomy—that is, if the calendar in fact came from China—what this
foreign calendar would have to say about Michihira’s ability to calculate the Xuanming li
would be limited. Even if it was a calendar from Korea calculated using the Xuanming li
system, its applicability to issues concerning a Xuanming li calendar calculated in the
Japanese style is doubtful. Yorimichi’s actions in 1038 and 1039 do not show him
dealing practically, to seek out other examples of calculations made using the same
system in order to check Michihira’s ability at mathematics. Rather, they demonstrate the
the continent as a source of authoritative knowledge, even if it was by no means the only
one. The authority of Chinese knowledge would play a major role in debates later in the
eleventh century and into the twelfth.
152
For now, it is important to note that these
calendars were authoritative in spite of evidence to the contrary from observation. The
phases of the moon, after all, supported Shōshō’s work.
153

152
This will be discussed in the next chapter.

153
The 1038-1039 case also proves a significant problem for those who would find backwardness or
devolution in the history Japanese astronomy through the Japanese court’s failure to import and implement
a newer system of Chinese calendrical astronomy. For this position, see Sugimoto Masayoshi and David L.
271
Yorimichi’s use of foreign calendars also went against a major ongoing shift in
concepts at court concerning the relationship between Chinese precedent and Japanese
practice. Although Chinese precedent could, and would, continue to be used throughout
Japan’s medieval period when it suited the arguments of one side, the authority of
Japanese precedent, even when it contradicted Chinese precedent, was growing. In fact,
when Chinese and Korean calendars were brought into a court debate a decade later, they
were not seen as authoritative.
154
Yorimichi’s partially successful deployment of them in
1039 probably had more to do with his position at the head of the court bureaucracy as
regent and his desire to support Kamo Michihira, than with absolute standards of proof.
This incident also shows that the position of official calendarist, that of
Instructor of Calendrics within the Bureau of Onmyō, did not necessarily grant
individuals the authority to persuade members of the court that their production of facts
about astronomy and about time were correct. Instead, some members of the court,
including Fujiwara Sukefusa, trusted their own eyes and the words of an outsider
calendarist over official statements of fact from the Bureau. Shōshō’s credibility drew
much from his personal relationship with Fujiwara Sanesuke, the elder statesman of
Sukefusa’s lineage—however, it also drew upon the past decades of publicly known but
officially unacknowledged Buddhist collaboration on the official calendar.
155
This
combined with mistrust of the technical specialists within the Bureau of

Swain, Science and Culture in Traditional Japan, A.D. 600-1854.. The assumption that a new system from
China would accord more accurately with natural phenomena is clearly not supported by this case.

154
See the discussion of the dispute of 1048 in the next chapter.

155
On the history of Buddhist calendrics in the mid-Heian period, see the previous chapter.

272
Onmyō—mistrust that dated from the internecine conflict between calendarists in the
previous century—to further destabilize the authority of the official calendar. Conflict
between official calendarists and Buddhist monks only amplified this mistrust, as was
shown in a debate over a solar eclipse the next year.

“Victory Will Depend on Observation”—The Solar Eclipse of 1041
Despite losing his bid to promote his calendar as the correct and official
calendar of the realm, Shōshō remained apparently undaunted. Yorimichi’s opposition
did not in any way end his career, as Shōshō continued his public activities as a
calendarist by issuing a solar eclipse prediction the following year. The resulting debate
shows the vast trust in his abilities that Shōshō commanded from his patrons, as well as
the mistrust he elicited from Michihira’s supporters. This splitting of court opinion only
fed more distrust and dissatisfaction over calendrical astronomy at the Heian court, which
ultimately motivated new quests for certainty in knowledge that would take shape in the
next century.
156

At the very end of the following year, the first of the era Chōkyō
(1040-1041),
157
Shōshō managed to get an eclipse prediction before the regent. On the
twenty-eighth day of the twelfth month,
158
Fujiwara Sukefusa was once again in

156
For more on the search for certainty, see chapter six.

157
The Japanese luni-solar year begins in late January or early Feburary by the Julian reckoning. The
standard practice is to refer to the Japanese year by the western year that it began in. However, as this story
took place in 1041 by the Julian reckoning, both years are included here.

158
February 13, 1041 in the Julian reckoning.

273
attendance upon Yorimichi when deliberations began on reports from both Kamo
Michihira and Shōshō regarding an upcoming solar eclipse. The eclipse was predicted to
occur on the first day of the next year, just two days away.
159

Shōshō had used Minamoto Tsunenari,
160
an close associate of Sukefusa’s and
a man close to appointment to the Council of State, to get his report before Yorimichi on
the previous day. It predicted that an annular solar eclipse would occur on the first day of
the year. However, the details of the eclipse, as Sukefusa reports them, indicate that most
of this eclipse would occur before dawn, making it primarily a “nighttime eclipse.”
161

Sukefusa does not provide a more detailed accounting of the temporal aspects of the
prediction, but it is clear that only if the eclipse would continue to the very end of the
timeframe Sukefusa indicated would it be visible at all in the capital.
162
If this eclipse
was going to be visible, only the faintest trace of it would be observable, and only from a
good vantage point without too much interference from mountains or buildings. The
near-unobservability of the eclipse was reflected in Michihira’s position. Just as his father
had argued concerning the solar eclipse of 1028, Michihira state that such liminal eclipses

159
Despite being a debate about an eclipse prediction, this incident is not analyzed by Saitō Kuniji in his
1990 study, for unknown reasons. Cf. Saitō Kuniji, Teika "Meigetsuki" no tenmon kiroku: Kotenmongaku
ni yoru kaishaku.

160
1009 – 1066. He would be made a member of the Council of State ( Jp. Sangi) in 1048.
A descendant of Daito Tennō through Minamoto Shigemitsu, he was not at all closely related to Minamoto
Tsuneyori, the author of the Sakeiki.

161
It was to begin during the double-hour of the Ox (1 AM to 3 AM), and end during the double-hour of
the Rabbit (5 AM to 7 AM).

162
Sunrise in Kyoto in February takes place a little before 7 AM.

274
were not to be counted. (Of course, as the debates of 1027 and 1028 show, visibility had
not been previously shown to be one of Shōshō’s major concerns.)
Fujiwara Sanesuke, although already in the last years of his life, took it upon
himself to question Michihira about Shōshō’s position. Tsunenari was to have been in
charge of relaying this information to Yorimichi, but as Tsunenari was ill at the time,
Sukefusa had been entrusted with it. Sukefusa summarized Michihira’s report in his
diary:
According to the correct interpretation of the main classic [the
Xuanming li], there will be no eclipse. Perhaps just at dawn there
will appear something that can be said to be a small eclipse. That
limited portion accords with Shōshō’s report, but it will not be visible.
Therefore, it is not noted in the official calendar.
163

Yorimichi, perhaps not kindly disposed to Shōshō as a result of the previous
controversy, was not hesitant to criticize Shōshō outright. Even so, he was willing to at
least grant Shōshō the possibility of being right:
The regent said: “What Shōshō has presented cannot be trusted.
However, let the visibility of the eclipse determine victory. Now,
according to Michihira’s report, the eclipse shall not be visible. Let the
decision be determined by the will of the sun.”
164

The next day, Sukefusa announced Michihira’s position again, but selected his
words carefully, in order to sway support towards Shōshō’s position. He reported that
Michihira said that there would be an eclipse on the first day of the year, that it would not

163
Shunki Chōkyō 1/12/28.

164
Ibid.

275
be visible, for “there will be a nighttime eclipse”
165
—this was twisting Michihira’s
statement in order to arose alarm. Sukefusa’s report of Michihira’s statement launched a
court debate on whether to cancel court business and the new years’ observances for the
next day—ultimately, a rehashing of the debates of 917 and 918, revisiting the
century-old precedent that eclipses occuring at night would not require countermeasures.
Concern was also raised about the potential negative effects this possible eclipse might
have on the annual offerings to the royal tombs, already underway. This shows the extent
of Shōshō’s influence. The court ordered Michihira to restate when night would end the
next day, and what part of the eclipse might be visible.
Sukefusa returned to Yorimichi with a new statement from Michihira, which
was a reiteration of the same position: that the eclipse would not be visible. Sukefusa
added his own opinion of things to Yorimichi—that Michihira had not provided a
detailed answer. Sukefusa further stirred the pot by generously providing Yorimichi with
a history of the impediments that had affected offerings to the royal tombs, information
he had received directly from his grandfather Fujiwara Sanesuke.
Meanwhile, Sanesuke had also summoned Michihira to his own residence to
press him further on details of the eclipse. Michihira, once again, stuck to his original
position. Upon stronger questioning, he was forced to admit that a slight, very slight,
eclipse might be visible when the sun appeared above the eastern hills, but that this was

165
Shunki Chōkyō 1/12/29. The rest of this account is taken from the entry.

276
not “true visibility” ( Jp. shōgen). The ring of an annular eclipse would not be
visible. To say otherwise showed a “lack of the knowledge of things.”
166

Daytime was defined as being from the double-hour of the Tiger to the
double-hour of the Rooster. Whether the eclipse would be visible or not, Michihira was
saying, was determined by “knowledge” and did not, in fact, require a decision from the
court: the proper understanding of the texts of calendrics determined that there would be
no visible eclipse. These “texts” included the history of the Bureau of Onmyō’s
experience with the Xuanming li system. Michihira was making the same argument
against Shōshō’s practice of calendrics that Ōkasuga Masumitsu had made against
Michihira’s ancestor, Kamo Yasunori, in 950—that there were not only proper texts for
the calculation of the astronomical calendar, but that there was also a proper way to
interpret these texts.
167
This knowledge had been passed down within a lineage, and it
was knowledge that an outsider such as Shōshō did not have. Shōshō’s insistence on the
validity of “nighttime eclipses” reflects Michihira’s point rather well: Shōshō was
ignoring a long history of practice (that of ignoring nighttime eclipses) that had been
passed down within the Bureau of Onmyō, shaped by decisions made by the Council of
State.
Neither Ōkasuga Masumitsu nor Kamo Michihira argued that an outsider could
never obtain such knowledge. Instead they were arguing that their understanding was
based on this lineage tradition, and was therefore correct—that if their opponent had a

166
ibid.

167
See the discussion in chapter three.

277
different position, then their opponent was clearly in the wrong. This position is more
subtle than the idea of “inheritance rights” that previous scholars had identified for Kamo
dominance within the Bureau of Onmyō. While it may be true that Kamo Michihira
believed that only a Kamo could become Instructor of Calendrics, he was not arguing that
only a Kamo could correctly predict the sky. His father had correctly predicted an eclipse
jointly with Jintō in 1021, after all. Instead, Michihira was arguing that he had the
“correct” understanding of calendrical methodology through the training he had received
as a Kamo lineage member from within the Bureau of Onmyō. If Shōshō differed from
him, then Shōshō was by definition not showing correct understanding. In this case
Michihira was arguing for a conception of knowledge within the Bureau of Onmyō
similar to the concept of hiden ( “secret transmission”) in both medieval Buddhism
and the performing arts. The connection between these concepts of knowledge became
more explicit in the following century.
168

After a short pause to prepare himself for the next day, Sukefusa again reported
to Yorimichi and relayed what he had heard, bringing to him a new report from
Sanesuke’s questioning of Michihira. Yorimichi nonetheless remained staunch in his
support of Michihira. Citing precedent, Yorimichi argued that none of the reports

168
See the account of the eclipse debate of 1129 in the next chapter.
A most study of hiden in the performing arts is Maki Isaka Morinaga, Secrecy in Japanese
Arts: "Secret Transmission" As a Mode of Knowledge (New York: Palgrave Macmillan, 2005), although it
does not really address the pre-modern situation. The best discussion of the origins of hiden in Japanese
intellectual and religious culture remains Stone, Original Enlightenment and the Transformation of
Medieval Japanese Buddhism.

278
suggested that court not convene.
169
The topic at hand was the prediction of the eclipse,
which, in either account, would mostly take place at night. Shōshō may have predicted
the eclipse correctly, but he did not have “correct” knowledge of court procedure.
According to the precedent of Tadahira’s decision at the beginning of the tenth century,
there was no need to cancel court activities in the event of a nighttime eclipse.
Yorimichi told Sukefusa that he should relay that decision—and exactly
that—to the throne. Perhaps Yorimichi was beginning to question Sukefusa’s handling of
words in light of Sukefusa’s, and Sanesuke’s, continuing support of Shōshō’s position.
By this time, it was night. Sukefusa went back to the palace to report on the
matters that had been debated that day. He sent word of Yorimichi’s decision to Sanesuke,
and began serving his assigned part in the new years’ observances at court. According to
court diaries, the day dawned dark and raining. There is no record of anyone sighting the
eclipse.

In this incident, which took place over two days in 1041, it is clear that the
dispute between Kamo Michihira and Shōshō only continued as long and as fiercely as it
did due to the support of their political patrons: Fujiwara Yorimichi on Michihira’s side,
and Fujiwara Sanesuke and Sukefusa on Shōshō’s behalf. That the decision ultimately
came down on Michihira’s side has as much to do with Yorimichi’s outranking Sanesuke
at court as it does with any argument that Michihira made.

169
Presumably this was beyond the scope of Shōshō’s report, which would have primarily spoken of the
eclipse and the danger it caused, and not recommend a specific line of court action (with precedents) to
take.

279
Although Yorimichi had stated that observation would determine who had the
correct answer, a decision still had to be made in advance. This was particularly true
when an eclipse held the potential of ruining the cosmologically vital first rituals of a new
year. While anti-eclipse rituals could be performed after the fact to mitigate the effects of
the eclipse, the goal of eclipse prediction was to stave off such effects before they could
take hold at all. As a result, despite Yorimichi’s hedge about confirmation, whether or not
the eclipse would be sighted or not was ultimately irrelevant to the debate—a decision
had to be made before the eclipse would even appear.
Yorimichi’s position at court secured Michihira’s victory in this debate, but it
was the strength of Sanesuke’s reputation and political seniority that allowed Shōshō’s
argument for the importance of this eclipse to go as far as it did, and for the debate to last
as long as it did. The decision about whether or not the new years’ events should be
canceled was forcibly prolonged up through the very last minute—the evening of the last
day of the year. Sanesuke and Sukefusa did all that they could to ensure that Shōshō’s
position was heard. It is very likely that they believed in Shōshō’s veracity as well,
particularly as Sanesuke consistently consulted Shōshō for information on astronomical
and astrological maters. Truth, at the Japanese court, was as much an interpersonal affair
as one involving the correct interpretation of texts, observation of precedents, or
confirmation from external reality.
Shōshō continued to be active in court debates on calendrics. He was consulted
by the Council of State during a calendrical dispute in 1050,
170
and he once again

170
Hyakuren-shō Eishō 5/9/28.
280
presented an eclipse prediction in opposition to Kamo Michihira in 1059.
171
That time,
however, the eclipse Shōshō predicted was reportedly confirmed.

Conclusion
Shōshō left no identified disciples,
172
and he is not known to have become
particularly famous as a ritualist. Instead his legacy was both more pervasive and more
insidious. Specifically, the period after 1041 saw a marked increase in the number of
debates at court about calendrics. Foreign calendars were solicited by the court again in
1048, and although the purpose was left unstated, the importation was undoubtedly
inspired by the 1038-1039 controversy. New monastic specialists began to appear in
public debates over calendrical astronomy at court: Zōmyō
173
and Nōsan, both fellow
disciples of Shōshō’s master Jintō, and they were followed by Myōsan
174
and Shinsan,
175

Nōsan’s students. Debates over eclipses occurred again in 1106, 1108, 1112, 1129 and
1146.
176
There may have been even more—there is a relative paucity of historical
sources for the end of the eleventh century. From the point of view of modern astronomy,

171
Jūsandai yōryaku Kōhei 2/1/1. Kondō Heijō, Jūsandai yōryaku, Shiseki Shūran (Tokyo: Kondō Heijō,
1883).

172
Yamashita Katsuaki, Heian jidai no shūkyō bunka to Onmyōdō.

173
dates unknown.

174
dates unclear.

175
dates unclear..

176
Chūyūki Kashō 1/12/1; Chūyūki Tennin 1/5/1; Chūyūki Ten’ei 3/9/1; Chūyūki Daiji 4/9/1 and
Chōshōki Daiji 4/9/1; and Taiki Kyūan 2/5/1, respectively. (The edition of the Taiki consulted was
Zōhō shiryō taisei kankōkai, ed. Taiki, 3 vols., Zōhō shiryō taisei (Kyoto: Rinsen shoten,1965).)

281
official calendarists began to improve their predictions in these debates around 1108,
possibly due to competition, and possibly due to their having learned a thing or two from
the Buddhist calendarists.
177
However, as Shōshō’s career shows, such improvement
need not be due to the Kamo calendarists having been convinced the superiority of the
Futian li system for predicting eclipses by Buddhist challengers such as Shōshō.
178

Shōshō was almost using the Xuanming li system when he disagreed with the official
calendarists—he was not, however, using the same tradition of practice that they were, as
Shōshō refused to ignore the impact of non-observable phenomena. From this one
difference, major disputes arose.
Another factor in these debates, however, was social as well as technical. Later
Buddhist monks learned from Shōshō’s example, and they too presented eclipse
predictions and opinions on calendrical matters at court through patrons who were
high-ranking members of the Council of State, or were otherwise well connected.
Courtiers during the late eleventh and twelfth centuries were also painfully aware of
controversy and uncertainty about calendrical matters, particularly eclipse prediction, in a
way they had not been before. By that time, there were no institutional buffers or bonds
of collaboration working to suppress expressions of divergent opinions on calendrical
practice before it caused public incidents at court. The position of Instructor of Calendrics
in the Bureau of Onmyō also proved defense against critique, nor did it afford any

177
Saitō Kuniji, Kotenmongaku no michi, 109-112.

178
As discussed above, it is not at all certain that the Kamo ever employed the Futian li for eclipse
prediction, although that possibility is an interesting subject for future study—if it can indeed be
determined from mathematical reconstruction and surviving records.

282
eleventh-century calendarists unassailable authority. Uncertainty and doubt about the
products of official calendrical astronomy only grew after attacks by Buddhist monks.
Such uncertainty and doubt led to the vigilance about the sky during the late
Heian Period that is reflected in Fujiwara Munetada’s experience during the overcast
eclipse of 1095. It also initiated a closing of ranks and reaffirmation of secrcy by the
Kamo calendarists as their tradition was attact. Finally, it contributed to the rise of a final
group of experts on calendrical matters that were not tied to Buddhist institutions or to
the Bureau of Onmyō—a group that hoped to restore certainty, even as their presence
destabilized it further. This last group to add to the culture of debate about eclipses during
the Heian Period, and that effected the final addition to the “Japanese style” of Chinese
calendrical astronomy, were scholars of Chinese texts.

283
CHAPTER SIX

“DEATHLESS TEXTS:” INNOVATION AND CONSERVATISM IN THE
SUPERVISION OF CALENDRICAL KNOWLEDGE

The last major group of experts to participate in the development of the “Japanese-style”
of Chinese calendrical practice were those specialists in mathematics found in the
Japanese court’s Bureau of the University ( ). They had long been used as on-call
advisors for mathematical matters, but their appearance in calendrical debates from the
middle of the twelfth century through the middle of the thirteenth was the result both of
new employment opportunities for ambitious technical specialists, and of a new
classicism current at the late Heian court. This new classicism was also reflected in new
strands of skepticism, as the study and employment of information in classical Chinese
texts were picked up by “young turk” factions in court politics. Both trends, in textual
classicism and household employment, led to a re-evaluation of classical Chinese texts in
debates at court. In response to this challenge, the Kamo lineage of official calendarists
began to close ranks and classify their “Japanese-style” practice as a “secret
transmission”—as the private property of the Kamo lineage.
Therefore, even as the new mathematical calendarists managed to effect a final
major addition to the calculation of calendrical astronomy as practiced at the Japanese
court, the Kamo’s successful claim to secrecy and ownership meant that the elements and
methods that composed the technical work of official calendar production at court were
no longer openly debated as the circulation of this form of knowledge closed down. As
284
regionally produced calendars appeared, independent from the centralized control of
either the court or the Kamo lineage in Kyoto, the result was the fragmented and
contradictory system of state-level time-keeping that characterizes medieval Japan.

Disorder Above and Below
Both the increased esteem for of classical Chinese texts and the need for expert
consultants in the late Heian period were the result of increased anxiety and mistrust
among courtier society. An example of such mistrust can be seen in the uncertainty over a
predicted eclipse in the middle of the twelfth century.
Early on the first day of the second year of the Kōji era (1143),
1
Inner Palace
Minister
2
Fujiwara Yorinaga
3
received a visitor from the palace. The royal secretary
Mitsufusa
4
had come to inform Yorinaga that the New Year’s Banquet ( Jp. sechie)
and Obeisance to the Retired Tennō ( Jp. In no hairei) were to be held that day as
usual. There had been some debate whether these ceremonies would be held at all, as
word had been circulating that a solar eclipse was expected to occur around sunrise that
morning. But the sun had risen and “appeared as normal, as if there was no eclipse.”
5

Subsequently “early that morning, people were sent to report that the sun had risen, and

1
January 18, 1143 in the Julian reckoning.

2
Jp. naidaijin.

3
1120 – 1156.

4
dates unclear.

5
Taiki Kōji 2/1/1.

285
after the sun had cleared the mountains and as soon as nothing obscured it, then the clear
skies and [full] sun revealed there was indeed no eclipse.”
6
Court business was to
proceed as was customary. But Yorinaga was unconvinced. A younger son with great
ambitions and unusual opinions,
7
he drew upon both his own discontent with the current
state of political affairs and his deep study and respect for Chinese learning to conclude
that things were not as clearly settled as Mitsufusa and his superiors implied. “I have
doubts,” he wrote.
8

Yorinaga had both political and technical reasons to mistrust the dominant
faction at court. He was deeply conscious of the fact that the affairs of the world were
being mishandled. He could discern this simply observing his own family: his older
half-brother Tadamichi
9
was in conflict with their father Tadazane,
10
who had been
trying both to reinstate himself at court and to promote his younger son Yorinaga’s career.
Tadamichi had come to power when Tadazane had been pushed out of his posts as head
of the Fujiwara clan (i.e. as the Jp. tōshi no chōja) and as nairan ( )
11
as

6
ibid.

7
Yorinaga has often been considered a shocking author. The Retired Tennō Hanazono, reading
Yorinaga’s court diary in the late thirteenth century, noted his own astonishment at some of the items
Yorinaga thought it appropriate to record in such a document. This story is related in the introduction to
Gomi Fumihiko, Shomotsu no chūseishi (Tokyo: Misuzu shobō, 2003).

8
Another possible translation of this phrase is “I inquired [after the matter].” In either
interpretation, however, Yorinaga was expressing reservations about the information. Taiki Kōji 2/1/1.

9
1097 – 1164.

10
1078 – 1162.

11
The right of document inspection. For more information, see chapter two.

286
the result of his political and personal conflicts with the powerful retired tennō
Shirakawa.
12
Tadazane’s political troubles had been due to back-palace politics: it was
the result of Tadazane’s reluctance to introduce his daughter Taishi
13
into the palace as
one of Toba Tennō’s
14
consorts. Not only could this be seen as an unreasonable demand
as made by Shirakawa—a man whose retirement from rule and to monastic orders should
have, according to Chinese ideals of statecraft, prevented or tempered his right to
interfere with court events
15
—but Tadazane’s reluctance was further strengthened by
pervasive rumors of dubious affairs occurring in the back palace itself.
16
Now, in 1143,
even well after Shirakawa’s death, things at court were still not in order. Tadamichi was
in direct conflict with his father and interfering with Yorinaga’s own rise at court. Father

12
1053 – 1129 (r. as tennō 1073 – 1087).

13
1095 – 1156.

14
1103 – 1156, (r. as tennō 1107 – 1123). Toba was Shirakwa’s grandson, which futher
complicated the sexual politics described in note 16 below.

15
Of course, during this late part of the Heian Period, known as Insei ( ), the leadership, influence,
and control exercised by retired and monastic monarchs on court politics, such “interference” and
“meddling” was to become the norm. It was not universally accepted, as even a casual perusal of Fujiwara
Munetada’s Chūyūki or Kujō Kanezane’s Gyokuyō will show. For more on the rise of Insei and how it
functioned, the best work in English remains G. Cameron Hurst, Insei: Abdicated Sovereins in the Politics
of Late Heian Japan, 1086-1185 (New York: Columbia University Press, 1976), although more
information about the interactions between the In (retired tennō) and other power blocs can also be found in
Mikhael Adolphson, Gates of Power: Monks, Courtiers and Warriors in Premodern Japan (Honolulu:
University of Hawai'i Press, 2000).

16
The reason for Tadazane’s reluctance is often credited to Tadazane’s distaste for Shōshi ( , 1101 –
1145) the senior consort, who was reputed to have been involved in several sexual affairs. It was said that
her son attributed to Toba, the later Sutoku Tennō ( 1119 – 1164, r. 1123 – 1142) was not the son
of Toba, but the son of Shirakawa himself. According to this rumor, this is why Toba did not favor Sutoku
and while Sutoku was a retired tennō pushed both Sutoku out of court politics, and Sutoku’s own son out of
the succession. This succession dispute is the precipitating cause fo the Hōgen Disturbance of 1156 (
Jp. Hōgen no ran).

287
and son were at odds, and sexual desires were interfering with the business of politics.
Things were clearly out of joint in what most, at this time, called “these latter days.”
17

Like other young men of the Fujiwara regents’ line who found themselves
blocked from political advancement, Yorinaga turned to scholarship.
18
In his own
writings, Yorinaga was explicit in promoting his own reputation for the study of the
Chinese classics, asserting that he had exhibited prodigious skill in the field from a young
age,
19
and he continued the pursuit of Chinese learning throughout his short life.
However, this scholarship does not seem to have won Yorinaga many fans at
court, and certainly few fans among modern historians. In medieval tales, Yorinaga was
depicted was having studied Onmyōdō ( ) and attempted to summon spirits, only
to fail in this attempt.
20
And among modern historians, Cameron Hurst’s comments from
his study on politics of the late Heian period are not atypical. Hurst argues that Yorinaga
had little political influence of his own at court as a result of his scholarship:

17
The account of the intra-Fujiwara conflict described here deliberately goes against the typical depiction
of events, which tends to paint Tadazane and Yorinaga as the villains, or at least as making unreasonable
demands. One of the factors given for the Hōgen Disturbance is often described as Tadazane’s
“unreasonable favor” ( Jp. katayori no ai) for Yorinaga over his elder son Tadamichi. Yorinaga,
who died on the losing side of the Hōgen Disturbance, is traditionally depicted as a villain, although
modern scholars have sometimes sought to modify this image using Yorinaga’s own writings to depict his
point of view. Hashimoto Yoshihiko, Fujiwara no Yorinaga (Tokyo: Yoshikawa kōbunkan, 1964)., 1-4.
A fuller accounting of the political issues leading up to the Hōgen Disturbance can be found in
Hurst, Insei: Abdicated Sovereins in the Politics of Late Heian Japan, 1086-1185, 125-154. An English
translation of a fictional account of this civil war is William Ritchie Wilson, Hōgen Monogatari; Tale of
the Disorder in Hōgen, Monumenta Nipponica Monographs (Tokyo: Sophia University Press, 1971).

18
Early examples include Fujiwara Sanesuke ( 957 – 1046) in the late tenth century, and Kujō
Kanezane ( 1149 – 1207) in the late twelfth century. Both men, and their desire for knowledge
and use of information as a tool, appear in the second chapter of this dissertation.

19
Hashimoto Yoshihiko, Fujiwara no Yorinaga., 14-17.

20
This tale is quoted in Nishioka Yoshifumi, "Rikujin shikisen to konrō no miura."
288
He was indeed a very learned man, but he appears to have been too
scholarly and antiquarian for his time. As a Confucian scholar, he
was fond of quoting from the Spring and Autumn Annals and of
reading the old codes. Preaching frugality and condemning vulgar
displays of wealth […] he went against the current of the times and
appears to have made a number of enemies.
21

This should not be taken to mean, however, that Yorinaga did not employ his
scholarship to practical ends. For example, Yorinaga knew from his study of Chinese
classics that there might have been a mistake on the part of the calendarists. Some of his
doubt concerning this eclipse was surely due to the debates between calendarists and
mathematicians leading up to that day, a debate that, according to Yorinaga, had “sullied
the reputation of the calendarists.”
22
A more explicitly stated basis for Yorinaga’s doubts,
however, was rooted in distant Chinese antiquity. As he wrote: “During the ages of the
Spring and Autumn [period] there were many calendrical errors. How many more [must
there be] now, in these latter days?”
23

Despite his concern that court procedure was not being followed correctly (for
if there had been an eclipse, the events about which Mitsufusa had informed Yorinaga
should have been canceled, in order to avoid any harm coming to the tennō) Yorinaga
continued with his personal new year’s observances.
24
He took care, however, even in

21
Hurst, Insei: Abdicated Sovereins in the Politics of Late Heian Japan, 1086-1185, 174.

22
Taiki Kōji 2/1/1. Unfortunately, the details and contents of this debate were left out by
Yorinaga, and as his is the primary account for this event, much remains unclear.

23
(Jp. masse), literally “last generations.” Ibid.

24
Jp. shi shihōhai, a private performance of the type of ritual that the tennō performed
before dawn on the first day of the year. In this ritual the four directions and the stars of the Big Dipper
were honored. Ibid.

289
his own personal diary, to cite a familial precedent for performing personal new year’s
rituals even during an eclipse. Yorinaga then sent out inquiries to determine if anyone
had personally seen the eclipse. In the following days,
25
he heard from both Fujiwara
Tomonori
26
and Fujiwara Narisuke
27
that when the sun had risen, they had seen it to be
slightly eclipsed. Another correspondent, Fujiwara Michinori,
28
however, stated that
there had been no eclipse at all. Yorinaga wrote in his account that “In this world, for
genius there are only these three men.”
29
Yet, as Yorinaga took care to note in his court
diary, even these three geniuses could not agree. In his search for certainty using trusted
men, Yorinaga found none.

25
Jp. gojitsu. Although this information is found in the same entry as that that describes the
eclipse and the first day’s ceremonies (Taiki Kōji 2/1/1), this information is explicitly described as arriving
after the fact, and having been included in the entry later.

26
dates unknown.

27
dates unknown. Although not a major historical figure, Narisuke appears with Tomonori
often in the Taiki, and often in relationship to Chinese scholarly activities. For example, in a discussion of
the virtues of the Taiping yulan ( ) encyclopedia later in 1143, Narisuke and Tomonori both praise
the work. Taiki Kōji 2/9/29.
Narisuke was also a member of the same “circle” studying the Chinese classics as Yorinaga.
Although Narisuke too was a descendant of the Fujiwara regents’ lineage, he was at the same time a
member of the retired tennō’s court. Hashimoto Yoshihiko, Fujiwara no Yorinaga., 41-43.

28
1106 – 1160. Michinori, more famously known by his lay monastic name Shinzei ( ), he
was a scholar and courtier deeply involved in factional disputes of the mid-12
th
century at court. Allied with
Taira Kiyomori ( 1118 – 1181), Michinori died in the Heiji Disturbance of 1160 ( Jp.
Heiji no ran). Famously he is depicted in the Heiji monogatari ( ) as having hid himself in a hole
in the ground when Minamoto forces attacked his residence. Marisa Chalitpatanangune, "Heiji
Monogatari : A Study and Annotated Translation of the Oldest Text" (University of California, Berkeley,
1987).

29
Taiki Kōji 2/1/1.

290
The case of the 1143 eclipse shows how much uncertainty there was
concerning the products of calendrical astronomy during the late Heian period. It had
been over a hundred years since the Buddhist monk Shōshō
30
began critiquing the
astronomical predictions of official court calendarists from a position outside the Bureau
of Onmyō ( Jp. Onmyō-ryō). So while the critique of the work of the calendarists
of the Bureau was nothing new—it had been a feature of court debates about the
applicability of solar eclipses predicted to occur at night beginning in 877
31
—the past
century had seen far more frequent and public debates about eclipse predictions and the
appropriate length of lunar months. These public disputes had unsettled the court’s
confidence in eclipse predictions in notable ways.
32
As this and previous chapters show,
one way that the court dealt with this lack of certainty was by consulting experts in
different fields. Ironically, the result of such consultation was often further debates.
Yet another strategy used by the court was to return to the Chinese classics for
certainty—this was the very strategy Yorinaga used in 1143. But this practice, too, led to
further debates. Chinese texts were authoritative, but they shared that authority with other
sources of knowledge. Sometimes they contradicted these other souces of authoritative
knowledge or other Chinese texts. The status of Chinese texts and the information they
contained in the eleventh and twelfth centuries was therefore a complicated one: the

30
dates unknown.

31
See chapter three for more details.

32
See the description of the eclipse of 1095 at the beginning of chapter five for more information.

291
Chinese texts and Chinese practice may have kept their authority, but this authority was
not uncontested.

Uncertainty and Authority
The example of 1143 shows clearly the uncertainty about and mistrust of expert
knowledge that circulated within the Japanese court, particularly among the junior and
less politically-favored members. How such men sought to reduce or eliminate
uncertainty reveals a great deal about which sources of knowledge were trusted, and
which sources were doubted, in the Japanese elite culture of the time. Sources show that
other people were not always considered trustworthy. Observation itself was often
doubted. Instead, authoritative sources of past knowledge, both from Chinese classics and
from Japanese court precedent, were sought.
It is telling that during a dispute between technical experts, Fujiwara Yorinaga
turned to men like himself who specialized in the study of the Chinese classics for
authoritative information. His impulse exemplifies a trend in scholarship and political
reform efforts found in the Japanese court of the late Heian period. While it may have
been “against the current of the times”
33
to attempt to enforce austerity at court or to
apply knowledge gleaned from the Five Classics to the reform of court society, Yorinaga
was not at all alone in such attempts. For men such as him, the lessons of Chinese
antiquity were in fact the key to reforming the court and setting the world to rights.

33
Hurst, Insei: Abdicated Sovereins in the Politics of Late Heian Japan, 1086-1185, 174.

292
These Japanese courtiers were participating in a much larger and ongoing trend
in East Asian scholarship and politics, one that dated back to Confucius and his program
for political reform during the Warring States period in China—and would stretch
forward in time to the Neo-Confucianism movement and beyond. The effort to improve
the world by reviving antiquity continued after Confucius in China and was later adopted
by reformers in states in the Chinese cultural sphere on China’s periphery. For some of
these men in Japan, all of the relevant information for political reform was to be found in
the works of the Zhou li ( Jp. Shūrei, “The Rites of Zhou”) and in the texts and
commentaries on the Chunqiu chronicles ( Jp. Shunshū, “Spring and Autumn
Annals”), as well as in a few other ancient texts. This focus on the ancient over the new
explains why Japanese courtiers did not often seek out new scholarship or works from
China, despite increasing trade with the continent during the twelfth century. Instead,
they sought out printed editions of these old texts, as well as encyclopedic works such as
the Taiping yulan ( Jp. Taihei goran “Imperial Readings of the Taiping Era”),
which despite its compilation at the Song Chinese court in the late tenth century, was
comprised mostly of quotations from or commentaries on the classics.
34

And yet it would be a mistake to think that individuals who attempted to reform
the Japanese state through scholarship would limit themselves to the study of Chinese
classics. The study of Japanese history and precedent was also a major component of the

34
Ivo Smits, "China as Classic Text: Chinese Books and Twelfth-Century Japanese Collectors," in Tools
of Culture: Japan's Cultural, Intellectual, Medical, and Technological Contacts in East Asia, 1000-1500s,
ed. Andrew Goble et al. (Ann Arbor, Mich.: Association for Asian Studies, 2009).

293
administrative life of members of the court bureaucracy from the tenth century
onwards.
35
While Yorinaga and others like him may have been called unusual for their
reverence of Chinese texts, this did not mean that they diverged from the court
mainstream by ignoring or devaluing Japanese precedent or history. For although
Yorinaga himself has not always been considered an orthodox master of court ceremony,
he avidly studied the subject and participated in court debates on proper procedure.
36
He
identified himself with an unorthodox but still respected lineage of scholarship on court
protocol.
37
This too was part of his mission to reform the court. Specifically, it was not
enough to know Japanese history and precedent—one also had to determine and follow
the proper precedent from all precedents recorded in the historical corpus This was an
urgent mission for these young men of the court, as evidenced by a rhetoric of cosmic
disorder and decline in their court diaries—signs, according to them, that the proper
procedures from Japan’s past were no longer being followed. Lessons from the past that
would remedy the world were not only to be found in Chinese texts, but also in Japanese

35
A summary of the development of this tendency can be found in Thomas Conlan, From Sovereign to
Symbol: An Age of Ritual Determinism in Fourteenth-Century Japan (Oxford: Oxford University Press,
2011), 14-31.

36
Hashimoto Yoshihiko, Fujiwara no Yorinaga., 51-58 has details on some of these activities.

37
Yorinaga saw himself as heir to Ōe Masafusa ( 1041 – 1111), whom he had studied under.
Ibid.
While Ōe Masafusa’s manual of court procedure, the Goke shidai ( ) is considered one
of the three key manuals of the Heian Period and is thus an authoritative reference for modern scholars, Ōe
Masafusa’s scholarly activities were not always approved of by more mainline Fujiwara scholars in the
same subject. See, for example, the ambivalent depiction of Ōe Masafusa in Fujiwara Munetada’s (
) Chūyūki court diary, throughout.

294
history. The challenge was then what to do if these sources conflicted, as frequently
happened.
Yorinaga’s confidence in the reliability of old precedents furthermore explains
why he turned to men and texts he trusted when other methods failed to resolve lingering
uncertainty. Unfortuately, many eclipse predictions were hard to confirm or disprove
through observation, even when observers were watching.
That turned out to be the case in 1143, for example. Before the new year,
Miyoshi Yukiyasu
38
had argued that the eclipse that the calendarists had predicted for
that day would not, in fact, be visible. This launched the uncertainty that Yorinaga seized
upon. The court as a whole clearly sided with Yukiyasu’s position (although it, too,
continued to watch for the eclipse that morning—just to be safe), and the court reinforced
and supported Yukiyasu’s statement with announcements not long after dawn that the
eclipse had not been sighted. And while from the standpoint of modern astronomical
reconstruction, Yukiyasu was indeed correct—the eclipse was not visible from the capital
of Heian
39
—at least two of the “men of genius” that Fujiwara Yorinaga consulted
reported that they saw a very faint eclipse as the sun rose. Of the men Yorinaga consulted
and trusted, only Fujiwara Michinori said that he did saw no eclipse. Yorinaga did not
distinguish between the expertise or observations of these three men. Although Michinori

38

39
From Kyoto, the moon remained about 0.18 degrees from the sun during the time of the eclipse.
Saitō Kuniji, Kotenmongaku no michi, 113. An eclipse did occur that morning (of January 17, 1143),
however, on the other side of the international date line, and the eclipse was visible from Alaska. NASA
Five Millenium Solar Eclipse Catalog #07471. NASA, "Nasa Eclipse Web Site."

295
is more famous among historians, Narisuke had served as a teacher and inspiration for
Yorinaga, and was just as respected by him as Michinori was, if not more.
40
Yorinaga’s
ambivalence in this case shows that distinguishing between trusted sources of
information, let alone confirming or disproving an eclipse by observation, could be quite
difficult.
Furthermore, the observation of eclipses, particularly in overcast skies, or
around sunrise or sunset, was not an easy task. Circumstances often made confirmation
difficult. While an eclipse prediction might be confirmed through a clear sighting of an
eclipse,
41
the Japanese court well knew that not sighting an eclipse did not mean that it
did not occur.
42
Complicating circumstances might not necessarily be limited to the
meteorological (obscuring cloud cover) or the geographical (buildings and landscape
obscuring the view). They might also be mental—that is to say, there might well have
been preconceptions held by the observers that largely shaped how they saw the world. It
seems that most of the time meteorological obstructions sufficed to cloud the issue of
whether an eclipse occurred or not. Indeed, from 1145 to 1230, historical sources reveal
that at least fifteen predicted eclipses—a significant portion of all recorded eclipses from
the period—could not be observed due to overcast skies or rain.
43
One further eclipse

40
Hashimoto Yoshihiko, Fujiwara no Yorinaga, 39.

41
See for example the praise Katsuragi Shigetsune received in the tenth century for making an accurate
lunar eclipse prediction, discussed in chapter three.

42
See for example the comments Fujiwara Munetada makes concerning the eclipse of 1095, discussed at
the beginning of chapter five.

43
Honchō seiki Kyūan 1/4/1 (1145); Taiki Kyūan 1/6/1 and Honchō seiki Kyūan 1/6/1 (1145); Taiki
Kyūan 4/4/1 (1148); Honchō seiki Kyūan 5/3/1 (1149); Hyōhanki Hōgen 3/3/1 (1158); Gyokuyō Kaō 2/7/1
296
prediction, this one subject to a public dispute in 1228, could not be confirmed or
disproven for unstated reasons.
44
From a ritual perspective, the malevolent effects of an
eclipse that was obscured by clouds or rain was nullified, and therefore it could be
counted as an eclipse that did not occur—such difficulties in observation were not
necessarily undesirable for a court seeking to avoid eclipses. But difficulties in
observation did not help produce certainty about eclipse predictions or predictors.
While clouds might not prevent the confirmation of an eclipse—one solar
eclipse was confirmed despite clouds in the sky in 1160
45
—clear skies could not ensure
that observations would be astronomically accurate. For example, in 1174, although
modern astronomical calculations prove that a solar eclipse began before sunset, the
author of a Japanese source insisted that no eclipse had occurred that day.
46

Furthermore, there was an awareness at court that eclipses might be visible in
one place and not in another. An eclipse that was not sighted in 1246 was noted as being

(1170); Gyokuyō Jōan 1/1/1 (1171); Gyokuyō Jishō 1/9/1 (1177); Gyokuyō Gunji 3/8/1 (1187); Gyokuyō
Shōji 1/1/1 (1199); Azuma kagami Kenpō 4/2/1 (1216); Kanchūki Jō’ō 1/4/1 (1222); Azuma kagami Karoku
1/8/1 (1225); Azuma kagami Antei 2/12/1 (1228); Meigetsuki Kangi 2/4/1 (1230).
(The edition of Kanchūki consulted was Shiryō taisei kankōkai, ed. Kanchūki, 3 vols., Shiryō
taisei (Tokyo: Naigai Shoseki,1936).. While a more recent critical edition is currently being published, it
does not yet cover the entire Kanchūki text.)

44
Hyakuren shō Antei 2/6/1 (1228).

45
Sankaiki Eiryaku 1/8/1. (Zōhō shiryō taisei kankōkai, ed. Sankaiki, 3 vols., Zōhō shiryō taisei (Kyoto:
Rinsen shoten,1965).)
Eclipses could be definitively sighted even when the skies were overcast, but the lack of a
sighting on a cloudy or partly-cloudy day tended not to provide authoritative statements of calendrical error.
For example, once his Midō kanpaku-ki, Fujiwara Michinaga softens his query over whether the
calendarists had made a mistake, even though the sun through the clouds “occasionally appeared round”
( Jp. rin, implying that it was not eclipsed) throughout the day. Midō kanpaku-ki Chōwa 2/2/1.

46
Gyokuyō Jōan 4/11/1. This particular eclipse is analyzed in Saitō Kuniji, Kotenmongaku no michi, 114.

297
“perhaps an eclipse [visible in] another state.”
47
While that phrase is obscure, there are
indications that at least some members of the court were aware of how time differences
between places might affect visibility. There is, for instance, a 1086 report of the sighting
of an eclipse in Ōmi Province, north and east of the capital, even though it had not been
visible in the capital. This sighting was considered potentially valid, as the eclipse was
scheduled to occur before dawn in the capital, and its sighting at dawn in Ōmi would
have been “within the limits of the ‘time difference.’”
48
Fujiwara Yorinaga also accepted
that time and place affected visibility. In his 1143 account, he states that he had kept an
eye on the skies for a bit longer than the eclipse prediction had required since “for many
recent eclipses the time of the eclipse [must be] moved one or two hours later.”
49
Along
with another account of eclipse monitoring from 1095,
50
these incidents show that the
difficulty of accurately sighting and confirming an eclipse prediction was common
knowledge at court.
An additional factor was that the confirmation of eclipse predictions was
almost beside the point for the purposes of court ritual. As previously discussed in
chapters three and five, the decision of whether or not to shut down the court in order to
protect the throne could not be made after the eclipse itself actually occurred—the

47
Azuma kagami Kangen 4/1/1.

48
Go-nijō Moromichi ki Kanji 3/11/1. That this knowledge was not universally held can
be seen in a different courtier’s 1106 dismissal of the argument of a Kamo-lineage calendarist that all
eclipse predictions calculated from Chinese calendrical texts were, in fact, a bit too early for Japan. (See
Chūyūki Kashō 1/12/1)

49
Taiki Kōji 2/1/1.

50
See the discussion of the account from the Chūyūki at the beginning of the previous chapter.

298
protective power of the rituals and of enclosing the tennō depended on these actions
being undertaken before the eclipse occurred. At the very latest, such a decision needed
to be made the day before.
51
It was certainly better to be safe than sorry, and a
conservative attitude towards eclipse predictions allowed the court to function smoothly
and with greater surety.
That attitude did nothing to promote certainty, however, nor did it leave much
room whereby eclipse predictions might be made more reliable—by coming from proven
calendarists, for example. While the confirmation of an eclipse prediction might have a
positive effect on the reputation of the individual who made that prediction, and that
improved reputation might be drawn upon on future attempts, the historical records of
debates show that the successful prediction of one eclipse was no guarantee that future
eclipse predictions would be accepted favorably by the court. A single successful, and
validated, eclipse prediction was insufficient for that purpose. Furthermore, no
calendarist on record achieved a perfect or even consistent record of success in solar
eclipse predictions. As it was impossible for technical reasons to guarantee
always-accurate eclipse predictions, calendarists cultivated other means of building a
reputation for skill and expertise. They consulted on matters requiring technical

51
These procedures of shutting down the court were essentially the same as mono-imi ( ritual
seclusion) at court, and mono-imi always began before sunrise on the first day of the period of seclusion.
On mono-imi and nighttime in Heian culture, see Noguchi Kyōko, "'Yoruka' no jidai: Mono-imi sanrō ni
miru Heian kizoku bunka no yoru," Kodai bunka 59, no. 1 (2007).
That the debates could stretch until the night before a predicted eclipse can be seen in the 1041
eclipse debates discussed in the previous chapter.

299
knowledge, undertook successful ritual performances, and even encouraged the
perception of a supernatural aura.

Such strategies benefited the producers of calendrical knowledge, but did not
much help the doubting audience evaluating their work. For the Council of State and
other members of the high nobility—including the retired tennō who became active in
political decision-making from the beginning of the twelfth century onward—who sought
to select the “truth” from alternative scenarios presented by competing specialists, some
means of evaluation was desired. Interpersonal relationships of trust—based on a
personal history with the particular calendarist—did help, but it was insufficient for the
late Heian court as a whole. The desire for certainty was the origin of the Council of
State’s emphasis on proper procedure in calendrical production. It was due to this need
for certainty that the revival of Chinese classical scholarship began to influence Japanese
calendrical practice.
The authority that Chinese classical scholarship gave those seeking to critique
or evaluate the work of calendarists can be seen in Yorinaga’s own observations and in
his solicitation of other scholars’ observations in 1143. The Chinese sources for
calendrical astronomy, too, held something akin to the textual authority of the literary and
political classics, and these special texts began to play a significant role in Japanese
calendrics in the twelfth century. Despite Ōkasuga Shinnomaro’s plaint in his petition to
adopt the Xuanming li in 862—to the effect that classics related to calendrics were not
300
“deathless and unchanging”
52
—when these texts were interpreted by the Instructors of
Mathematics ( Jp. san no hakase), they became constant and consistent sources
of authoritative knowledge, as will be shown below. This new stream of authoritative
knowledge then interacted with the authority of Japanese precedent, and that of the
lineage of calendarists itself, in the Japanese court’s production of reliable facts from
calendrical astronomy. The newly rejuvenated authority of the Chinese text itself
reinforced tendencies towards conservatism in Japanese calendrical practice. As a result,
the re-evaluation of classical mathematical texts related to calendrical astronomy by
classical specialists resulted in the forceful re-incorporation of the nineteen-year Rule
Cycle ( Ch. zhang, Jp. shō), despite its previous abandonment by the very tradition of
Chinese calendrical astronomy that had produced the Xuanming li adopted by the
Japanese court in 864 and used since.

Mathematics at the Japanese Court
That texts of mathematics, combined with the study of mathematics, were also
part of the classical Chinese tradition in Japan has not been much noted in previous
scholarship. Yet it was this esteem for the subject that allowed experts of Chinese
mathematics to appear as new calendrical experts in twelfth-century Japan. Studies
concerning scholarship utilizing the Chinese classics at the Japanese court often exclude
mathematics, a separation similar to the “absolute taboo” that once kept technological

52
See the discussion of the importation of the Xuanming li in chapter three.

301
and scientific works out of the study of Chinese intellectual history.
53
Yet mathematics
was one of the subjects taught at the Bureau of the University ( Jp. Daigaku-ryō),
and therefore part of the Japanese court bureaucracy. Furthermore, the study of
mathematics was one of the approaches to the study of the Yijing ( Jp. Eki-kyō, the
“Book of Changes”), showing that a strict separation of mathematics from classicism was
not always possible or necessary.
54

Although Yorinaga and other men like him in the late Heian period are
considered unusually devoted students of Chinese classics, respect for Chinese learning
was deeply embedded in Japanese court life. Even as the authority of Japanese legal
precedent and a growing sense of difference between the continental and insular courts
meant that Chinese precedent or knowledge was not always deferred to, the esteem in
which Chinese texts were held did not lessen. This esteem was only reinforced by the
ritual role that Chinese classical texts played in court life. Males born to noble families at
court performed a ceremonial reading of the Xiao jing ( Jp. Kō kyō, “The Classic of
Filial Piety”) at a young age.
55
Even if these young boys might not have understood what
their tutor was having them repeat, the ceremony reinforced the idea that The Classic of
Filial Piety, and Chinese texts in general, were texts to be revered. Moreover, exposure to

53
Benjamin Elman, On Their Own Terms: Science in China, 1550-1900 (Cambridge, Mass.: Harvard
University Press, 2005), introduction.

54
Fujiwara Michinori’s work on the mathematics of the Yijing will be mentioned in slightly more detail
below.

55
This was the ceremony known as dokusho no hajime ( ).

302
Chinese texts in ritual settings continued once these young men joined the court, through
ceremonies such as the annual Ceremony in Honor of Confucius ( Jp. sekiten) and
its attendant lectures. Even though the ability of the average court official to intelligently
discuss matters covered in the texts lectured on during this ceremony may be
doubted
56
—and there was a definite tendency to see experts in such texts as stuffy men
out of touch with current affairs
57
—repeated exposure through ceremony only reinforced
the aura surrounding Chinese texts. While a member of the high noblily might not do as
Yorinaga and his compatriots did and devote significant time to the study of the Chinese
classics, the view that these were texts of truth and value was strong.
The importance of Chinese texts to Japanese politics was institutionalized in
the Bureau of the University, which was charged with teaching a number of subjects
focused on Chinese texts. When the Bureau of the University and the corresponding
exam system was established in eighth-century Japan, they were not intended to establish

56
Sei Shōnagon ( 966? – 1025?), a female attendant of a royal consort and a famous literata,
notes in her Pillow Book that she does not much know about what is covered in the sekiten. How much of
this was an affected pose, and how much might be attributed to her role as a woman at court, is
unknown—therefore, how typical this ignorance and attitude might have been among courtiers cannot be
determined. As shown by the case of Murasaki Shikibu, mentioned in the footnote below, being female did
not disqualify one from knowledge of Chiense scholarship. Ivan Morris, ed. The Pillow Book of Sei
Shōnagon (New York: Columbia University Press,1991), vol. 2, 110 (note 612).

57
See, for example, the depiction of the instructors from the Bureau of the University who taught Yugiri
in the Tale of Genji. The author of the Tale of Genji, Murasaki Shikibu ( dates unknown), however,
was herself a noted scholar of Chinese history, so the mockery seems more directed at the type of
individuals that such instructors were, rather than to the content of their expertise itself.
Sei Shōnagon herself wrote a rather glowing description of experts in the Chinese classics: “I
need hardly say how splendid I find a learned Doctor of Literature. He may be of lowly appearance, and of
course he is of low rank, but the world at large regards him as an impressive figure. As an Imperial Tutor,
he is consulted about all sorts of special matters, and he is free to approach the most eminent members of
the Emperor’s family. When he has composed one of his prayers for the Emperor or the introduction to
some poem, he becomes the object of universal praise.” Ivan Morris, ed. The Pillow Book of Sei Shōnagon,
111.

303
a meritocracy within the court bureaucracy—the promotions for successful examinees
were not sufficient to guarantee such a result.
58
Instead, the Bureau of the University and
the exams seem to have been intended to provide the court with a population of technical
bureaucrats trained in the Chinese texts.
59
Despite the prevalence of literary studies in
subsequent depictions of the Bureau of the University in the early Heian court,
60

mathematics as a subject fit neatly into this mandate.
The Bureau of the University not only trained specialists in the canon of
Chinese classical texts. Instructors within the Bureau taught Chinese pronunciation and
calligraphy as well. The major fields of study, however, were Chinese classical literature
and history ( Jp. Kidendō),
61
the Five Classics ( Jp. Myōgyōdō),
62
the
legal codes of both Tang China and Japan ( Jp. Myōbōdō), and mathematics ( Jp. Sandō). Even as the Bureau of the University declined as a separate and
independent institution, and the exam process became factionalized,
63
these fields and

58
Furthermore, the exam system in Tang China was similarly not intended to create a meritocracy that
was, in any event, some centuries coming, even in China. Therefore the view of the Japanese “failure” to
establish a meritocratic bureaucracy like the Chinese is plainly anachronistic. For an example of this point
of view, see Sansom, A History of Japan to 1334.

59
This point is made incidentally in .Aritomi Jun'ya, "Hyakushō buiku to ritsuryō kokka: Jukyōteki
ideorogii seisaku o chūshin ni," Shigaku zasshi 112, no. 7 (2003).

60
Particularly Borgen, Sugawara No Michizane and the Early Heian Court.

61
The precise content taught by Instructors of Kiden ( Jp. kiden hakase), as well as the
personnel in this field, changed the most during the classical period of Japanese history of any of the
Daigaku-ryō sections. See Momo Hiroyuki, "Kidendō," in Kokushi daijiten, ed.
Kokushi daijiten henshū iinkai (Tokyo: Yoshikawa kōbunkan, 1992–1997).

62
See Suzuki Rie, "Myōgyōdō," in Kokushi daijiten, ed. Kokushi daijiten henshū iinkai (Tokyo:
Yoshikawa kōbunkan, 1992–1997) for more information on myōgyōdō.

63
The factionalism is covered in Borgen, Sugawara No Michizane and the Early Heian Court.
304
their instructors continued to exist as a population of experts at court. Reports from
Myōbōdō legal experts to the Council of State on points of legal controversy can be
found up through the fourteenth century,
64
as can reports from Myōgyōdō experts on the
classics, and Kidendō history specialists.
65
Even though studies of the Bureau of the
University underemphasize the field of mathematics,
66
Sandō should not be thought of as
different from these other fields. It too survived the “privatization” of University fields in
the hands of specialist lineages. Chinese works of mathematics that individuals in Sandō
studied were part of the same collection of authoritative texts that had the aura of
continental authority and truth; even if, as technical knowledge, they appear less often in
surviving historical sources.
Nine texts were mandated for study for Students of Mathematics ( Jp.
sanshō) in the Bureau of the University, although some of these texts are no longer extant
and not much is known about them.
67
Along with the famous Jiuzhang suan (

64
For example, in 1371. Dai Nihon shiryō ser. 6, vol. 34, 162-166.

65
Gukanki Eitoku 1/1/22 (1381). The reports from myōgyōdō and kidendō specialists were on the issue of
whether or not to change the era name that year ( Jp. kaigen).
(This text is also known as Goshinjin’in kanpaku-ki, and while the latter title has been viewed
by some scholars as more correct, the name Gukanki remains more cited. Tōkyō Daigaku Shiryō hensanjō,
ed. Goshinjin'in kanpaku-ki, 4 vols., Dai Nihon kokiroku (Tokyo: Iwanami shoten,1999–2008).)

66
For example, in Borgen, Sugawara No Michizane and the Early Heian Court.
Studies on the Bureau of the University, like Borgen’s, have tended to come from scholars
focused on Japanese literature—as a result, not only the mathematics field of the Bureau of the University,
but also the legal field has been underemphasized. It is more difficult to study these fields from the
standpoint of the surviving sources: for those focusing on the students of instructors of the more literary
fields, history and classical texts, essay exams from these students have been preserved in Heian compendia
of Sino-Japanese ( Jp. kanbun) texts, such as Honchō monzui. The work of Instructors of Mathematics
and Instructors of Myōbō, however, has to be reconstructed from surviving reports to the Council of State.

67
Mikami, A History of Japanese Mathematics.
305
Jp. Kyūshō sanjutsu, “Nine Chapters on the Mathematical Art”),
68
this list of texts
included the Zhou bi suan jing ( Jp. Sūhi san kyō, “The Arithmetical Classic of
the Gnomon and the Circular Paths of Heaven”), a work dedicated to the mathematics
necessary for calendrical astronomy.
69
The importance this second work had for
Japanese mathematics training, superior to even the more famous Nine Chapters on the
Mathematical Art, is shown by a regulation issued in 731 to the effect that Students of
Mathematics who had not mastered the Arithmetical Classic of the Gnomon and the
Circular Paths of Heaven were to remain students until they had done so.
70
As the texts
for Students of Calendrics ( Jp. rekishō) were not set until 757,
71
it can be said that
in Japan, the Arithmetical Classic of the Gnomon was a mathematical text first, and only
later an astronomical one.
It is therefore unsurprising to see Instructors of Mathematics, who were masters
of these texts, called in to evaluate the arguments of Instructors of Calendrics when they
conflicted with each other. The first textual evidence for this can be found in a debate

68
This text was compiled in stages over several centuries, reaching its current form during the Han
Dynasty. For a translation and commentary of this text, see John N. Crossley et al., ed. The Nine Chapters
on the Mathematical Art : Companion and Commentary (Oxford: Oxford University Press,1999).
A study and translation of another of the required texts, the Sun Zi suan jing ( Jp.
Sonshi sankyō) can be found in Lay Yong and Tian Se Ang Lam, Fleeting Footsteps: Tracing the
Conception of Arithmetic and Algebra in Ancient China (River City, NJ: World Scientific, 1992).

69
A translation and study of the Zhou bi suan jing can be found in Cullen, Astronomy and Mathematics in
Ancient China: The Zhou Bi Suan Jing.

70
Shoku Nihongi Tenpyō 3/3/ (Jp. itsubō).

71
The texts for Students of Calendrics also included the Nine Chapters on the Mathematical Art and a
text known as “Six Chapters” ( Ch. Liu zhang, Jp. Rokushō), about which little is known. On the role
of texts in the education of Students of Calendrics, see chapter three.

306
over the length of months for the upcoming year of 951, which occured between
Instructors of Calendrics Ōkasuga Masumitsu and Kamo Yasunori in 950.
72
So when
Instructors of Mathematics begin to participate in calendrical debates in the twelfth
century, it was not their participation that was unprecedented. Instead what was new was
their initiative to monitor calendrical production themselves and to alert patrons and
potential patrons at court to their opinion, before any request from above arrived. Such
initiative, however, did arise from an older tradition: mathematical experts had been one
group through which the court had monitored the production of calendrical astronomy.
And in the late Heian period, mathematical experts took up on this duty themselves,
without further prompting and with great enthusiasm.

In examining why the Japanese court would employ one group of experts to
monitor the performance of another, it is important to note that calendrical and
mathematical knowledge was not limited to members of the Bureau of Onmyō, Buddhist
astrologer-monks, or Instructors of Mathematics. But lest it be assumed that the court
could have simply done away with its troublesome experts, it should be noted that such
knowledge was rare—although, as shown below, it did exist in other spheres. Instead of
providing yet another source of calendrical knowledge, however, the examples below
attest more to a sense of familiarity that members of the court could draw upon when they
were forced to decide between statements from mistrusted and conflicting experts. This

72
The record of the debate in the Hokuzan shō manual of court protocol is, in fact, classified as an
instance of “calling in Confucian experts to the jin no sadame deliberations.” See chapter three for more.

307
familiary helps explain why it too so long before specialist expertise was finally deferred
to—however contingently.
As early as the ninth century, simplified rules for calendrical calculation were
circulating among other scholars at court. Koremune Naomoto’s
73
“personal”
commentary on the ritsuryō codes, Ryō no shūge ( “Collected Exegesis on the
Administrative Codes”) includes detailed descriptions of calendrical calculations,
including a description of how the Rule Cycle might effect the placement of intercalary
months over nineteen years.
For the current calendar, it can be approximated using the Yellow
Therarch’s methods of settling the days of the calendar…. The “Old
Text” states: [As for] calendrical numbers, nineteen years make one
Rule Cycle. In the third year, there is an intercalary ninth month. In
the sixth year, there is an intercalary sixth month. In the ninth year,
there is an intercalary third month. In the eleventh year, there is an
intercalary eleventh month. In the fourteenth year, there is an
intercalary eighth month. In the seventeenth year, there is an
intercalary fifth month. In the nineteenth year, there is an intercalary
month.
74

Naomoto also added an explanation for why intercalary months were necessary:
[If] no intercalary months are inserted, then before three years are up,
there will be a lag of one month. Instead of the first month, it will be
the second month. Before nine years have passed [the first month]
will become the third month. In other words, spring will be [called]
summer. Before seventeen years have passed, the lag will become six
months. It will be as if spring has become autumn.
75

73
, dates unknown. Active 877 – 907.

74
Ryō no shūge Shokuin ryō.

75
Ibid.

308
Lists of intercalary months can also be found in later encyclopedias, such as the
early fourteenth-century Nichūreki ( ). But Naomoto in the early ninth century
went further than later non-calendarist authors in explaining why intercalary months were
necessary and where to place them. He included explanations of the mathematical
complexity of the calendar that are both detailed and slightly garbled. His citations show
the level of technical detail that Naomoto though appropriate and necessary for a legal
expert to know about calendrical astronomy.
In the middle of the twelfth century, Fujiwara Michinori likewise thought
mathematical skill necessary for an educated member of the court bureaucracy.
76

Nevertheless, it seems that Naomoto and Michinori were both unusual in this
respect—even if between the two men they encompass the entire Heian period, and thus
attest to a lasting interest in mathematics. For instance, in a debate over the applicability
of having the winter solstice fall on the first day of the eleventh lunar month in 1156,
Taira Nobunori
77
expressed difficulty following the arguments. He attributed this
difficulty to the “secret texts” and methods being cited by both sides, but he also
confessed to a lack of familiarity with the vocabulary and techniques required for
large-number calculations.
78
When, as in 950, a technically detailed debate was

76
Michinori is famous in the history of Japanese mathematics as the author of Keishizan ( ), a now
lost text that is thought to have involved the mathematics of the Yijing. Mikami, A History of Japanese
Mathematics, 17.

77
, 1112-1187.

78
Hyōhanki Hōgen 1/10/18 and 1/10/20.
The question of the “numeracy” (numerical or mathematical literacy) of the Japanese elite
during the Heian Period is one that, to the knowledge of the author, has yet to be addressed adequately,
309
described by a non-specialist observer, the technical terms became garbled and the
mathematical aspects confused.
79
It is unsurprising, therefore, to find the ministers of the
Council of State and other political elites of the court asking one group of experts to
evaluate the work of another group when the Council members did not have the skills to
do so themselves.

There was indeed one field, however, in which the elite members of the
Council of State and advisors to the throne and other political elites were expert. This was
the study of Japanese precedent. While these elites might not, like Yorinaga, study the
Chinese texts themselves for the relevant knowledge, they did study records kept by their
lineages and other court records as guides to proper decision-making and court
functioning. This was the major reason why Japanese court precedent became such an
important part of debates between calendarists from the middle of the Heian Period
onwards—on this point, the political elites were active experts and could decide things
themselves. When precedent and technical knowledge collided, or Japanese precedent
and Chinese text conflicted, it was then that decision-making became complicated.

The Inapplicability of Foreign Calendars: The Calendrical Dispute of 1050 and
Resistance to the Authority of Chinese Practice

even by historians of mathematics or of mensuration (weights and measures), even though assumptions are
made. See Mikami, A History of Japanese Mathematics.

79
See the discussion of the 950 debate as recorded in the Hokuzan shō ( ) in chapter three.

310
A good example of the problems that arose when Japanese precedent and
Chinese classics were both referenced to conflicting ends comes from the middle of the
eleventh century. The case of 1050 shows that even while Chinese classics were
respected, more recent Chinese texts were not necessarily prioritized when a solid
Japanese precedent was available. The relative authority of these two sources of
knowledge about the proper procedure from the past was less clear-cut than a simple turn
to classicism might explain.
As recounted in the previous chapter, during the calendrical debates of
1038-1039, the regent Fujiwara Yorimichi had imported almanacs from the continent in
order to support the position taken by the official calendarist Kamo Michihira.
80

According to Yorimichi, these “Chinese” calendars supported Michihira’s work in that
the continental almanacs and Michihira’s “match[ed] completely.”
81
This should not be
taken to mean, however, that Yorimichi necessarily recognized the absolute authority of
calendrical knowledge as produced by China or Korea, as was proven to be the case a
little over a decade later.
In 1039, Yorimichi had been following an old precedent, as almanacs from the
continent had been consulted during a calendrical controversy in 937.
82
Calendrical
almanacs were imported from China and Korea a few more times in the subsequent
decade. From 1041 to 1050 even though there are no details concerning about calendrical

80

81
Shunki Chōryaku 3/12/28.

82
See chapter three.

311
debates that occurred, a hint of some of the uncertainty that persisted can be found in the
year 1048. In that year the Daizaifu ( ) in Kyūshū forwarded almanacs from the
continent to the court twice: first from Goryeo in the early summer of the year,
83
then
once again from Song China towards the end of the same year.
84

It is not clear why these almanacs were imported, but according to records in
the Hyakuren-shō ( , late thirteenth century) and Fusō ryakki ( c. 1100)

85
chronicles, the Korean calendar differed from the official Japanese calendar of that
year in the length of the twelfth month.
86
The key point is that in 1048, just as earlier in
1039, foreign calendars were being consulted to help adjudicate a dispute at court.
In just two years, however, the appropriateness of such evidence would be
judged differently. In 1050 foreign calendars would be cited once again in a public
dispute at court over the calendar for the current year, but at that point hey were used to
attack the work of Kamo Michihira. Specifically, on the twenty-eighth day of the ninth
month of the fifth year of Eishō (1050), the members of the Council of State requested

83
Hyakuren shō Eishō 3/5/2 and Fusō ryakki Eishō 3/5/.
Once again, Goryeo is referred to as “Silla” ( Jp. Shiragi) here. In this case, “Silla” stands
for a polity on the Korean Peninsula, much as “Tang” ( Jp. Tō, Kara) could stand for any Chinese
dynasty.

84
Fusō ryakki Eishō 3/11/16.

85
This text is traditionally attributed to the monk Kōen ( 1074? – 1164?) of Enryakuji temple (
) during the latter part of the reign of Horikawa Tennō (1079 – 1107, r. 1087 – 1107). The chronicle
includes events up through the year 1094, and so is thought to date from not long afterwards.

86
The entry found in the Fusō ryakki is slightly more detailed: it states that although the Korean and
Japanese calendars differed in the length of the twelfth month, otherwise “there [was] no difference.”
Fusō ryakki Eishō 3/5/2.

312
reports from various experts of calendrics.
87
The precipitating cause for this was a report
from the senior monk
88
Zōmyō,
89
in which he argued that the placement of the
intercalary month after the eleventh month for that year was incorrect.
90
Zōmyō was
calling for the adjustment of the calendrical year already in progress in order to fix Kamo
Michihira’s “error.”
A Song calendar was submitted to the court as evidence for Zōmyō’s position.
Michihira, however, argued that “despite precedents” for comparing calendars across
states,
91
the throne should not use “foreign explanations” to evaluate the work of its
calendarists. This statement may seem ironic, coming from the very Michihira whose
position had been bolstered by foreign calendars just eleven years earlier.
92
However,
this need not be an about-face on his part. For all accounts and purposes it appears that
the decision to consult continental calendars in 1038-1039 seems to have been made by
the regent at the time, Fujiwara Yorimichi—not Michihira himself. A closer examination
of the circumstances of 1050 reveals the logic in Michihira’s position. The key in his
protest is that the Japanese court should not rely on “foreign explanations.” Why this

87
Hyakuren-shō Eishō 5/9/28.

88
He held the title of Jp. daihōshi.

89
dates unknown.

90
Hyakuren-shō Eishō 5/9/28 and Fusō ryakki 5/11/1.

91
Hyakuren-shō Eishō 5/9/28. The grammar here is a little garbled, and
makes more sense when read in context with the quotation from the Fusō ryakki below as

92
See the discussion in chapter five about the 1038-1039 calendar controversy.

313
should be the case is revealed in other arguments Michihira made for placing the
intercalary month after the eleventh month. Michihira also cited the “Calendarists’ Error”
( Jp. rekike no shitsu) of 936 in support of his position. In 936, the official
calendar had not put the winter solstice on the first day of the eleventh month, even
though 936 marked the first year of a new Rule Cycle. Michihira was arguing against
making the same mistake twice. This, despite the fact that such an argument went against
the unadulterated calculations from the Xuanming li canon for 1050. Once again, a
tradition shaped by court demands was defended by a member of the Bureau of Onmyō
against a literalist reading of the relevant text by an outsider.
Specifically, as 1050 was the first year of a new Rule Cycle, the winter solstice
should have fallen on the first day of the eleventh month. However, as discussed in the
third chapter, the Xuanming li system used to calculate the Japanese calendar did not
include the Rule Cycle in its mathematical cycles.
93
Zōmyō was arguing for placing the
intercalary month where it would have fallen according to the Xuanming li calculations.
But, in order to preserve the Rule Cycle, Michihira had reassigned the intercalary month
to a position after the eleventh month, post-calculation. If the intercalary month were
placed before the eleventh month, as the intercalary tenth month, then the winter solstice
would have occurred in this month. In order to preserve the Rule Cycle, the earliest
possible position for the intercalary month was after the eleventh month.

93
The Rule Cycle had been abandoned in Chinese calendrical astronomy during the Tang Period as
incompatible values for lunar and solar periods were adopted. Yabuuchi, Chūgoku no tenmon rekihō.

314
In arguing against “foreign explanations” and in citing a calendrical incident
from Japan’s history, Michihira was making the same essential argument that he had
made against Shōshō. To wit: that correct Japanese practice as determined by court
precedent—the same precedent that established violating the Rule Cycle as an
“error”—had priority over other considerations. Zōmyō, on the other hand, saw error in
Michihira’s post-calculation movement of the intercalary month. Instructors of
Mathematics also played a role in this debate, but it was consultative only. When the
court solicited reports from other experts, particularly experts in mathematics, they sided
with Zōmyō’s position.
94
Zōmyō and his allies were all arguing for employing the
mathematics of the Xuanming li as it was written, disregarding the tradition of Japanese
practice that had built up within the Bureau of Onmyō.
In this respect, the events of 1050 echo the calendrical debate of 950, in which
Ōkasuga Masumitsu argued that he was in possession of the correct interpretation of how
to calculate the calendar using the Xuanming li, in contrast with Kamo Yasunori’s
argument that he had calculated the Xuanming li “as written.” In both 950 and 1050, the
side that argued for an emphasis on the history of Japanese practice won the debate.
Kamo Michihira, in 1050, used Japanese court precedent to bolster his case, just as
Ōkasuga Masumitsu and Katsuragi Shigetsune had in the previous century. That was a
strategy that fit well within contemporaneous court culture. Zōmyō and the mathematics

94
Fusō ryakki Eishō 5/11/1. The Hyakuren-shō (Eishō5/9/28) also reports that the monk Shōshō was
consulted, but whether he agreed with Zōmyō or not is left uncertain in the text. Yamashita Katsuaki has
seen this incident as a case where Zōmyō was acting as a proxy for Shōshō’s attack on the official calendar.
Zōmyō was himself, however, a famous sukuyōji monk (Nichūreki 13), and so he most likely had the skills
to run calendrical astronomy calculations himself. Yamashita Katsuaki, Heian jidai no shūkyō bunka to
Onmyōdō.

315
specialists both were relying on the text as written, but this did not overcome the example
provided by Japanese precedent. The legitimacy of the tradition Japanese practice would
be further reinforced in another debate about intercalary months in 1129.

Legitimacy of Hidden Traditions: The Intercalary Month Debate of 1129
In 1129, the legitimacy of Japanese precedent over fidelity to Chinese texts
would appear again, only this time with a new twist: in the arguments of 1129, Japanese
precedent and practice of calendrical astronomy became the private, “secret” property of
the Kamo lineage of calendarists. This argument was not only accepted by the court but
was also part of the rationale they relied upon in their desition supporting the Kamo
position in the 1129 debate.
The intercalary month controversy of 1129 is similar to the 1050 debate in
some respects and radically different in others. Once again, a Buddhist monk with
experience in calendrical calculations submitted a critique of the current year’s calendar
to the court. Grand Monk Ryūsan
95
petitioned, saying that the official calendar would
become inaccurate after the seventh month of that year. It had a short (29-day) intercalary
month after the seventh month, then a long (30-day) eighth month. Ryūsan’s calculations
put the intercalary month, still short, after the eighth month. He alerted the court so that
they might fix what was a grave “calendrical error”.
96

95
, dates unclear

96
Chūyūki Daiji 4/6/1.

316
On the first day of the sixth month, the Council of State was summoned to
deliberate the matter. Although this sort of decision had previously been handled by a
meeting of the Council of State for a consultative a jin no sadame ( ) decision, this
time “according to new protocol,” the meeting was to be held at the residence of the
retired tennō.
97
Prior to the meeting, Kamo calendarists had been questioned twice.
98

Fujiwara Munetada,
99
a member of the Council of State, thought that the calendarists’
response was largely pursuasive, but that there were two points that should be
investigated further. These were issues of astronomical conjunction and of Japanese
precedent.
The calendarists had addressed Ryūsan’s critique of their placement of the eighth
month by saying that on the first day of the eighth month, as they had determined it, the
sun and moon would be in conjunction
100
—in other words, that the day they had selected
for the new moon that would start the lunar month was absolutely correct. The
calendarists’ second point drew upon Japanese precedent. The intercalary month of the
fourth year in a Rule Cycle, which 1123 was, was always placed after the ninth, sixth, or

97
Ibid. The reason for this “new protocol” ( Jp. shingi) likely has to do with how Shirakawa was
expressing power over the court. See the opening of this chapter for how Shirakawa pushed Fujiwara
Tadazane out of the regent position. In 1129, Toba was also newly retired.

98
Kamo Ieyoshi ( ), and his sons Munenori ( ) and Yasuyoshi ( ) were the calendarists
questioned. In the rest of the account from the Chūyūki, Ieyoshi responds for the rest of the Kamo
calendarists, as the senior active Kamo at the time.

99
, 1062 – 1141.

100
Chūyūki Daiji 4/6/1.

317
seventh month. Not since the start of Rule Cycle counting in Japan had there been an
intercalary eighth month in the fourth year of the nineteen-year Rule Cycle.
101

Before the Council, Ryūsan was given a chance to respond to the Kamo
calendarists’ defenses. He initially dismissed the comment about the conjunction, but
upon being pressed, he admitted that he did not know much about “conjunctions” or
theories about them, and was basing his critique on mathematical calculations.
102

Likewise Ryūsan emphasized the results of his calculations over whether or not any
precedent for the insertion of an intercalary eighth month in the first four years of a Rule
Cycle. Ryūsan was emphasizing the results of his calculations using the Xuanming li, the
acceptance of which depended on the court’s general opinion of his skill and his
familiarity with the texts. Meanwhile the Kamo were employing precedent in support of
their own mathematical work, a much stronger argument for the court culture of the time.
In his final response to Ryūsan, Kamo Ieyoshi, speaking for the calendarists,
stated that the matter of “conjunction” was something that was a secret teaching of the
calendarists and he was unfortunately unable to say anything more on the subject. He also
re-emphasized the unprecedented nature of an intercalary eighth month at this early point
in the Rule Cycle. Then the decision was turned over to deliberations of the Council of
State.

101
Ibid. The text of the Chūyūki is a little garbled on this point ( ), but when
taken in context with the later section ( ) it seems clear that the is an error for
.

102
ibid.

318
According to Fujiwara Munetada’s own Chūyūki diary record, the first comment
from a member of the Council stated that while Ryūsan’s arguments would be applicable
for matters of Buddhist horoscopy ( Jp. sukuyō no koto), on matters of making
calendars, the calendarists should be deferred to. Others suggested bringing in experts in
mathematics, or even masters of Chinese history and classics (kiden and myōgyō), to
evaluate Ieyoshi’s and Ryūsan’s arguments. The final decision, however, matched the
position made in session by Fujiwara Munetada, the author of the account, himself:
I deliberated and said: For this matter, as I do not know the details and
the fields are all difficult, [I] rely on the [information] given
previously.

Every field has texts. However, there are always secret explanations
and orally transmitted [teachings].
103
In calendrics should there not
also be secret explanations? As Ieyoshi has already said that there are
secret teachings, then the explanation of the calendarists should be
used.
104

It still took several hours before a decision was reached, with arguments going
back and forth.
105
Finally, the court came to agreement. The decision to support
Ieyoshi’s position and the current calendar was a validation, and the first explicit one, of
the traditions and history of practice of Japanese calendrics that had been accumulated

103
Literally, kuden ( )—a term often found in descriptions of secret transmissions in the Tendai
school of Buddhism. For more, see Stone, Original Enlightenment and the Transformation of Medieval
Japanese Buddhism.

104
Chūyūki Daiji 4/6/1.

105
According to Fujiwara Munetada, the meeting began in the double-hour of the Snake (between
roughly 9 am and 11 am) and debate was still fierce in the third-part of the double-hour of the Sheep (a
little after 2 pm). Ibid.

319
among the lineages of experts that staffed the Bureau. As in 1050, Japanese practice was
prioritized over explanations that relied strictly on Chinese texts. Just as his ancestor
Michihira had done in 1050, Ieyoshi, the Kamo calendarist, called upon a history of
Japanese precedent to bolster his case. That he could utilize such history swung the
opinion of the court his way, as shown in the words of the observer Fujiwara Munetada.
Although the Ryūsan’s expertise was acknowledged by the Council of State, they would
not prioritize his calculations over the history of practice within the Bureau of Onmyō. It
would seem that the independence of the Kamo lineage of calendarists was secure, and
that their absolute authority concerning matters relating to the calendar had finally been
established.
However, debates from later in the twelfth century show that not to have been the
case. Although court mathematics experts played a role in the preceding debates, they
were called in as consultants to evaluate the arguments of the Kamo and Buddhist
calendarists. When mathematics experts began to involve themselves in calendrical
affairs in order to perform their own displays of skill, they would use the authority of
Chinese texts and their own expertise in court precedent to effect the last major
modification to the Japanese practice of Chinese calendrical astronomy.

The Classic of the Gnomon: The Miyoshi Lineage and the Rule Cycle
It was in the context of both an increased regard for knowledge gleaned from
classical Chinese texts and increased secrecy among the Kamo lineage of calendarists
that the Miyoshi lineage finally took the stage as a recognized group of calendrical
320
experts in their own right. The influence of their appearance would have enduring
consequences for the Japanese practice of calendrical astronomy.
The last modification to the Xuanming li calendrical system that made up
Japanese practice was implemented following the protest of Miyoshi Yukiyasu over the
calendar for the year 1156. Yukiyasu had not been much noted in previous sources—his
participation in the 1156 debate marks the first time he appears in some sources
106
—yet
his activity resulted in major changes in the relative status of Instructors of Mathematics
among the various experts at court.
The first such change occurred when he critiqued the eclipse prediction of 1143
discussed earlier in the chapter. Not many details survive in the sources. Fujiwara
Yorinaga, in his description of the controversy, does not even mention Yukiyasu’s
name.
107
The year 1143, however, still marked the first time that an Instructor of
Mathematics, or other mathematics specialist, petitioned the Council of State about a
matter of calendrical astronomy on his own initiative.
108
This was not because of a lack
of experience or skill—Instructors of Mathematics had been called in to consult on
calendrical debates in 950 and 1050, and their involvement is suggested in sources for the
controversy of 1123. The reason why a specialist in Chinese knowledge would choose to
participate in the arena of eclipse prediction at this late date, when official calendarists

106
Specifically, the Hyōhanki courtier diary.

107
See Taiki Kōji 2/1/1. Yukiyasu’s name appears in the entry in the Hyakuren-shō for Kōji 2/1/1.

108
Saitō Kuniji also identifies 1143 as the addition of mathematics experts in eclipse debates, although
their presence in such debates as consultants can be attested earlier. Saitō Kuniji, Kotenmongaku no michi..
He does not speculate, however, as to why these individuals participated in these debates, or what they
hoped to gain from their participation.

321
and Buddhist monks had been contesting publically over calendrical matters for over a
century, likely had to do with Yukiyasu’s own personality and intentions. But it also
owed much to structural changes in court society and client-patron relationships.
The aim of experts in mathematics like Yukiyasu who began to initiate
calendrical debates in the middle of the twelfth century was similar to that of Buddhist
monks, and to the members of the Bureau of Onmyō before them. Like these earlier
participants, the mathematics specialists sought to publicly display their skill in order to
elicit patronage from elite members of court. The types of position Miyoshi
mathematicians like Yukiyasu were trying for were, however, different. Although
positions in provincial government remained a source of income in this time period, such
positions were increasingly being monopolized by the military retainers of the most
powerful households.
109
Therefore scholars like Miyoshi Yukiyasu sought newly
formalized positions within high-status households or with the retired tennō’s court (
Jp. In no chō), as these positions rose in status and competition over them increased.
110

Furthermore, positions within these household organizations, tied to blocks of corporate
estates ( Jp. shōen), were becoming more economically secure than positions within
than the court bureaucracy, due to a decline in tax receipts as more and more economic

109
The Kamo did continue to hold governorships and assistant governorships, however, as shown by their
signatures on surviving almanacs. See the database in Atsuya Kazuo, ed. Guchūreki o chūshin to suru
koyomi shiryō no shūsei to sono shiryōgaku-teki kenkyū.

110
On the In no chō, see Hurst, Insei: Abdicated Sovereins in the Politics of Late Heian Japan,
1086-1185, and Michio Maki, In no kinshin no kenkyū (Tokyo: Zoku gunsho ruijū kanseikai, 2001). On
positions within powerful courtier households, see Higuchi Kentaro, "Sekkanke seisho shitsuji no seiritsu
to tenkai," Shigaku zasshi 116, no. 2 (2007).

322
and agricultural activity was incorporated into the estate system. These organizations also
needed men who were expert in Chinese classics and precedent, as these household
“courts” modeled themselves on the Japanese bureaucratic court and developed
increasingly complicated administrative structures.
111
Displays of public skill in a
Chinese classical art, such as mathematics, was one way that specialists in mathematics
could secure the necessary financial support to supplement the nominal income they
received for working in the court bureaucracy.
After Miyoshi Yukiyasu’s first appearance in calendrical affairs in 1143, the
success of which appears to have been inconclusive,
112
he continued his efforts by
submitting a major petition thirteen years later. Specifically, when the official calendar
for 1156 was presented, it had the winter solstice falling on the first day of the eleventh
month. The year 1156, however, was one that fell in the middle of a Rule Cycle—it had
been only eleven years since the beginning of the cycle in 1145.
113
The debate that
ensued concerned whether or not a conjunction of the new moon (first day of the lunar
month) and the winter solstice could occur on a year in the middle of a Rule Cycle, or
whether such a conjunction could only appear at the beginning of a Rule Cycle. In this

111
The use of specialists, not as “housemen” (hereditary retainers) but as clients (who might serve a
number of households, on an as-needed basis, is covered in Higuchi Kentaro, "Sekkanke seisho shitsuji no
seiritsu to tenkai.".

112
As Fujiwara Yorinaga’s account shows, there was some doubt as to whether the contested eclipse
actually occurred or not.

113
The celebration of the beginning of this Rule Cycle ( Jp. sakutan tōji) can be found in both
Fujiwara Yorinaga’s court diary and the Honchō seiki chronicle (Kokushi Taikei Hensankai, ed. Honchō
Seiki, Kokushi Taikei (Tokyo: Yoshikawa kōbunkan,1999).). Taiki Kyūan 1/11/1 and Honchō seiki Kyūan
1/11/1.

323
matter, the “Calendarists’ Error” precedent that had been cited explicitly by Kamo
Michihira in 1050, and implicitly by Kamo Ieyoshi in 1123, was silent. In short, there
was no precedent of such an event in the history of Japanese calendrical production. The
Kamo were at a disadvantage—precedent had been a major factor in their recent court
debates,
114
and this situation was unprecedented.
There were also good reasons in 1156 to seek out providential omens to praise
the current ruler.
115
Go-Shirakawa Tennō was new to the throne, and his selection had
not been uncontroversial—in fact, fighting had broken out in the capital earlier that year,
as conflict over the succession both to the throne and to the Fujiwara regency turn into
open combat.
116
Even if there had been an unwritten rule within the Kamo lineage to
avoid non-Rule Cycle occurrences of the winter solstice falling on the first day of the
eleventh month, there was clearly also sufficient political motivation to ignore just such a
rule among the Kamo calendarists of 1156, and to emphasize the good omen that
mathematical calculations provided.
117
Indeed, the providential meaning that such a

114
See the discussion in the earlier parts of this chapter.

115
This debate and the circumstances around it are discussed in the beginning of chapter three of this
dissertation.

116
See the discussion and notes at the beginning of chapter three, which also discusses this particular
controversy and its background.
Political conflict continued, even after this particular skirmish had been suppressed, as would
be proven by the resumption of armed conflict just four years later: the fighting was later known as the
Heiji Disturbance ( Jp. Heiji no ran) of 1160.

117
The existence of a political motive for the forging of the calendar in this way does not necessarily
imply mathematical foul play. It would have required impressive foresight as the calendar for 1156 was
calculated in summer and fall of 1155. This does correspond with Go-Shirakawa’s accession to the throne
( 1127 – 1192, r. 1155 – 1158), so it is not impossible. Forseeing that the Hōgen Disturbance would
occur and that Go-Shirakawa would need even more legitimation, however, seems frankly unlikely.

324
conjunction had was the very rationale used by members of the Council of State who
wanted to leave the calendar of 1156 as it was: Go-Shirakawa, the new tennō, would
receive on that day a symbol of the approval of the cosmos as the figurehead of a
symbolically renewed dynasty.
118

Miyoshi Yukiyasu’s contrary position was, however, the one that prevailed.
Instead of promoting the cosmic legitimation of the new tennō, Miyoshi’s position
emphasized the continuity of Go-Shirakawa’s reign with the past, and particularly with
the reign of the founder of the capital of Heian, Kanmu,
119
under whose auspices the first
Rule Cycle was instituted. While this echoes the “tradition” arguments used by the
Kamo previously, it was Yukiyasu’s mastery of the Classic of the Gnomon text of
mathematical astronomy that decided the issue. He emphasized that a Rule Cycle, as
defined in this ancient text, had a winter solstice on the first day of the eleventh month
only on the first year of the cycle.
120
Ultimately this, and the desire for continuity, carried
the day among the members of the Council of State, and so a decree was issued ordering
that the calendar year already in progress be changed to conform to an ancient
cosmological cycle.
The calendrical debate of 1156 discussed above shows the authority that
mastery of a Chinese text could give a scholar at court; even as the calendrical debate of

118
As based on sentiments quoted in Hyōhanki Hōgen 1/10/18 and 1/10/20.

119
737 – 806, r. 781 – 806.

120
This does not come out as clearly in the Hyōhanki entries as it does in the surviving petitions of 1163,
which are discussed below.

325
1123 makes explicit the recognized authority of Japanese calendrical traditions at court.
In 1156, however, arguments based on classical Chinese texts won over arguments that
referred to Japanese practice. The distinction here was that there was no publicly
available precedent for the situation that the Kamo, in making the official calendar, had
attempted to achieve: the cosmological return represented by the beginning of a Rule
Cycle, realized eight years too early.
121

While the Kamo lineage could employ secrecy and intra-lineage transmission
to support their authority, they could not control how the texts they used were interpreted
by others. Insofar as the texts that they were using were available to others outside of the
Bureau of Onmyō—the Xuanming li text itself was certainly circulating among Buddhist
monks and others, and the Classic of the Gnomon was as much “owned” by lineages of
mathematics specialists—their practices when based on such texts were subject to
potential critique. The interpretation of such texts needed to become secret in order to
become secure. When there was historical precedent to support the Kamo position,
reference to “secret traditions” could win the day. When there was no such precedent,
textual fidelity could be, and was, prioritized. This complicated weighing of authorities
exemplifies the unsettled relationship between Japanese history and Chinese classics and
techniques at the Heian court.

121
The Rule Cycle was celebrated as the Miyoshi intended it to be on the first day of the eleventh month
of Chōkan 2 (1164). Masayori ki Chōkan 2/11/1. (Minamoto Masayori, "Masayori Ki," in Shōkōkan monjo
(Tokyo: Tōkyō Daigaku Shiryō hensanjo, 12th c.).)

326
Once they had become recognized experts in calendrical affairs through
Yukiyasu’s efforts, the Miyoshi became standard participants in the ongoing debates that
occurred at court. Mathematics specialists are known to have been involved in debates
about solar eclipses in 1171, 1187, 1199, and may have been among the “myriad experts”
involved in one such debate in 1227.
122
The new authority of the Miyoshi lineage can
also be seen in a petition and report of 1163 preserved in the Saionji family archive.
123

Specifically, on the fifteenth day of the ninth month of the first year of Chōkan, the Kamo
calendarists petitioned the Council of State asking for clarification on the Rule Cycle.
They explicitly stated that the mathematics of the calendar for the next year, the first of a
new Rule Cycle, did not result in the winter solstice falling on the first day of the
eleventh month. Instead of fixing this “in house,” they were petitioning the court. Clearly
1156 had shaken the confidence of the Kamo specialists in the Bureau of Onmyō.
In response, the Miyoshi Instructor of Mathematics submitted a long report, a
mix of quotations from the Arithmetical Classic of the Gnomon and examples from
Japanese history. This Miyoshi Instructor does not address the question of how the
mathematics of the Xuanming li might be reconciled with the demands of court precedent.
Instead he holds forth on the astronomical truth preserved in the Arithmetical Classic of
the Gnomon, in effect insisting that the cycles of the universe are to be found in the Rule
Cycle, even though they might contradict the mathematics of the Xuanming li.

122
Gyokuyō Jōan 1/1/1 (1171); Gyokuyō Bunji 3/8/1 (1187); Gyokuyō Shōji 1/1/1 (1199); and Minkeiki
Antei 1/6/1 (Tōkyō Daigaku Shiryō hensanjō, ed. Minkeiki, 6 vols., Dai Nihon Kokiroku (Tokyo: Iwanami
shoten,1975–2007).) and Azuma kagami Antei 1/6/1 (1227).

123
Multiple, "Chōkan sakutan tōji kanmon," in Saionji-ke kiroku (Tokyo: Tōkyō Daigaku Shiryō
hensanjo, 1925 (copy of 1163 ms.)).

327
While the response to this petition is not recorded, in the following year the
winter solstice did indeed fall on the first day of the eleventh month, and it was
oppropriately celebrated with a ceremony and promotions.
124
The Kamo calendarists
clearly found some way to reconcile the demands placed on this by the Miyoshi scholars
of Chinese mathematical classics and the expectations of the court.

Fragmentation and Silence
Ironically, the appearance of the last major group of calendrical experts also
marked the beginning of the end of debates over calendrical astronomy at court. The
re-validation of a conservative interpretation of the Rule Cycle, as counted from 794, in
the debate of 1156 was the last major addition to Japanese calendrical practice as
determined through a debate at court. Debates over the visibility of solar eclipses
continued until at least 1228, and a debate over a lunar eclipse occurred in 1248,
125
but
active disputes over calendrical facts disappear from the record by the middle of the
Kamakura Period (1192-1331). The astronomer Saitō Kuniji has said that the reason for
this “burning out” of public activity in calendrical astronomy is mysterious,
126
but
some causes can be deduced from surviving historical sources.
Firstly, the validation of calendrics and Buddhist astrology (Sukuyōdō) as
separate fields with separate methods in 1123 provides one reason why debates would die

124
Masayori-ki Chōkan 2/11/1 and Chōkan 2/11/16.

125
Hyakuren-shō Hōji 2/5/15.

126
Saitō Kuniji, Kotenmongaku no michi, 116. Saitō is only concerned with solar eclipse debates,
however, and not disputes over the length of months. He likewise misses the 1248 lunar eclipse debate.

328
down. As the fields of mathematics, calendrical astronomy, and Buddhist astrology
continued to separate (or in the case of Buddhist astrology, re-separate) from each other,
the idea that expertise in one field was applicable to another weakened. The growth of
“secret transmissions” ( Jp. hiden) in Japanese scholarly, religious, and artistic
fields—and the recognition of the legitimacy of such transmissions, as can be seen in the
1123 debate—only reinforced this separation. That debates over eclipse predictions
continued into the thirteenth century does not show the weakness of “secret transmissions”
as the legitimation of practice. Instead, the continuation of eclipse debates shows that
once knowledge of eclipses was associated with one field of study, such association was
not easily erased.
Secondly, the benefit of displays of skill such as eclipse prediction was no
longer as strong as it once had been. When one Kiyohara Noritaka
127
imported and
attempted to use a new system of Chinese calendrical astronomy, the Qintian li (
Jp. Kinten-reki),
128
in 1228, he was a younger son in a family famous for classical
scholarship. Noritaka would become famous, but not through eclipse prediction or
service at court. The day of his predicted eclipse was cloudy, and so his prediction could
not be confirmed.
129
Even so, Noritaka did not attempt this again. It seems that he
realized the 1228 eclipse prediction was, by that time, a relic of older ways of seeking

127
.

128
Southern Song, first implemented in 956. Sivin, Granting the Seasons: The Chinese Astronomical
Reform of 1280, with a Study of Its Many Dimensions and a Translation of Its Records: .

129
Hyakuren-shō Antei 2/6/1.

329
patronage at court. Service positions within powerful households as retainers or scholars
were becoming hereditary as well,
130
making any attempt to rise through court society
through impressing potential patrons with displays of skill a less successful. Noritaka,
instead, went east to tutor the Hōjō regents of the Kamakura shoguns, and developed a
school of classical Chinese scholarship and pronunciation.
131
Without hope of such
rewards, there was less reason to make the effort to develop a reputation through
predictions of eclipses or critiques of calendrical production. Patrons, too, seem to have
been less receptive to such attempts, as records of eclipse predictions after the middle of
the twelfth century become less and less detailed, failing even to identify the individuals
involved in the debate.
Finally, the court’s monopoly over the production of calendrical facts itself
weakened. The last debate over a solar eclipse in 1228 corresponds with the appearance
of the oldest surviving almanac written in the kana syllabary.
132
While this particular
almanac may have simply been a translation of one of the official almanacs—which was
written in Sino-Japanese, like most official documents—the appearance of kana almanacs
signified that the reign of the official guchūreki almanac as the single authoritative source
of time in Japan was coming to its end.

130
Higuchi Kentaro, "Sekkanke seisho shitsuji no seiritsu to tenkai."

131
His annotations of Chinese classics are preserved in the Kanazawa bunko, an archive based on the
holdings of a junior branch of the Hōjō lineage.

132
Karoku ni-nen kana-reki, found on the reverse of the Kujōke-bon copy of the Chūyūki. Atsuya Kazuo,
ed. Guchūreki o chūshin to suru koyomi shiryō no shūsei to sono shiryōgaku-teki kenkyū.

330
Indeed, regional calendars began to appear soon afterwards, and these circulated
widely. These calendars and almanacs were produced outside the capital to satisfy the
demands of regional elites and institutions. Despite being calculated from the same
Xuanming li instructions, they could and did differ from the calendar still produced and
presented to court every year in the capital. Such deviations are hardly surprising, given
that the Kamo lineage had by this time classified all modifications and accretions to the
practice of Chinese calendrical as “secret transmissions” held only by that lineage.
The oldest extant regional calendar, which dates from 1332,
133
was produced in
eastern Japan, on the Izu Peninsula. Its Mishima-goyomi ( also Jp. Mishima-reki),
which was produced at the Mishima Shrine ( Jp. Mishima Taisha), one of the
three major shrines honored by the Kamakura Bakufu. The Mishima calendar was widely
used in the east in the later medieval period,
134
and may have developed in response to a
need for calendrical knowledge at the Kamakura Bakufu itself. The proof of its
independence is attested to by how it periodically differed from the calendar produced in
the capital. The earliest historical record of this difference dates to 1374, when the monk
Gidō Shūshin,
135
during his travels in eastern Japan, discovered that what was called the

133
The earliest extant fragment of a calendar that is thought to be a Mishima-goyomi is from a
Fudō-myōō statue in Ibaraki’s Shōkenji temple ( ). Okada Yoshirō, "Koyomi no shurui," in Koyomi
o shiru jiten, ed. Okada Yoshirō et al. (Tokyo: Tōkyōdō shuppan, 2006), 133.

134
It was the official calendar of the Tokugawa Bakufu in Edo.

135
1325 – 1388.

331
fourth day of the third month in the calendar he had brought from the capital was the
third day of the third month locally. The calendars were one day off from each other.
136

Variation between calendars was eventually to be found even in the capital itself,
as independently produced calendars differed from those produced for use at court.
137

While the ideal of a unified state with unified time persisted, as shown by complaints
about such variations, no one institution had the ability to enforce this ideal in reality.
Fragmentation, even at the level of time, had become the new reality.

136
Kūge nichiyō kōfu ryakushū ( ). This source is cited in Okada Yoshirō, "Koyomi
no shurui."

137
One such example can be dated to 1490, where different historical sources exhibit conflicting
information about the length of months of that year. Dai Nihon shiryō vol. 8 part 23, 224.

332
CONCLUSION:

FRAGMENTATION AND THE IDEAL OF UNITY IN MEDIEVAL JAPAN

When court debates over the identification, interpretation, and prediction of astronomical
events fade from the Japanese historical record in the thirteenth century, this does not
mean that the practices of astronomical observation and interpretation ( Jp. tenmon)
and calendrical astronomy ( Jp. rekihō) also disappeared at court. Facts about the
meaning of astronomical phenomena and predictions of the behavior of astronomical
bodies were continually produced at court through to the early modern period. The Kamo
lineage of the Bureau of Onmyō, and then the Tsuchimikado branch of the Abe lineage,
continued to present official almanacs that turned astronomical predictions into date and
time for the use of the state even into the Meiji Period.
1
At least in name, theofficial
determination of natural and political facts from celestial observation and prediction at
the Japanese court, based on the link between the astronomical heavens and the Mandate
of Heaven, continued. In its actions, the Japanese tennō-centered court continued to
perform as if it were the center of a Chinese-style state.
As even casual students of Japanese history know, however, this was not the
case. Even though now the Kamakura Period (1192–1331) is thought of as characterized
by an authority split between the royal court in Kyoto and the warrior government of

1
The Tsuchimikado responsibility over this duty was abolished with the Gregorian reform of Meiji 6
(1872).

333
Kamakura—a situation known as the “Kōbu ( ) polity”
2
—after the establishment of
the Muromachi Shogunate and then the descent into the Warring States decentralized
polity, the royal court has generally been discounted as an impoverished and feeble
remnant of the classical past.
3
It is only relatively recently that studies have
re-emphasized the relevance and continued symbolic role of the tennō-centered court
through the later medieval and early modern periods.
4
And the continued production of
both tenmon-style divination on astronomical observations and of “official” state
calendars fits neatly into this view of the continued symbolic performance of rulership at
court. Official facts about nature were continuously produced by the Japanese court in the
name of a centralized state, just as rituals were performed and government posts filled as
economic circumstances allowed.

2
The concept of the Kōbu Polity was introduced to English-language scholarship by G. Cameron Hurst,
"The Kōbu Polity—Court-Bakufu Relations in Kamakura Japan," in Court and Bakufu in Japan: Essays on
Kamakura History, ed. Jeffrey Mass (New Haven: Yale University Press, 1982). It would also later be
adopted by Jeffrey Mass, Yoritomo and the Founding of the First Bakufu: The Origins of Dual Government
in Japan (Stanford, Calif.: Stanford University Press, 1999), and to a degree it appears in the kenmon
theory as used in Adolphson, Gates of Power: Monks, Courtiers and Warriors in Premodern Japan.
Although the term kōbu has not been much adopted in Japanese scholarship, the idea of “two
capitals” has become standard in overviews of the Kamakura-period. See for example Gomi Fumihiko, Kyō
Kamakura no ōken, Nihon no jidaishi (Tokyo: Yoshikawa kōbunkan, 2003).

3
The weakness of the throne is emphasized by those who argue that Ashikaga Yoshimitsu, the third
Ashikaga shogun, had made an unsuccessful attempt to usurp the throne. For a presentation of this theory,
see Imatani Akira, Muromachi no ōken: Ashikaga Yoshimitsu no ōken sandatsu keikaku (Tokyo: Chūō
kōron, 1990).

4
Some early hints of such a theme can be found in the work of Amino Yoshihiko, who noted the links
between the throne and “outcasts” and non-farmer provisoners of the court. Amino Yoshihiko, Nihon
chūsei no hinōgyō-nin to tennō (Tokyo: Iwanami shoten, 1984).
Works in English that re-emphasize the court in the medieval period include Lee Butler,
Emperor and Aristocracy in Japan, 1467-1680: Resilience and Renewal, Harvard East Asia Monographs
(Cambridge, Mass.: Harvard University Press, 2002).

334
The continuation of such production did not mean, however, that such facts
were always accepted by other power-holders in the medieval Japanese state. To the
contrary, the production of facts about astronomical phenomena became largely
decentralized, as new institutions and organizations began producing their own
knowledge about nature.
Some such activity can be credited to the efforts these new power-holders made
toward imitating the structure and trappings of the Japanese court, or even omodelling
themselves into a Chinese-style state. And at the same time that the Japanese court was
being overshadowed by other institutions in economic, military, and even diplomatic
terms, powerful households were modeling themselves after the royal court. This process
began during the Heian Period itself, as the Fujiwara Regents’ lineage slowly expanded
and became bureaucratized, after which posts serving this lineage organization turned
hereditary.
5
The Kamakura shogunate, too, adopted courtly ways. The “courtification” of
Kamakura shogunate culture had its first significant push under the aegis of third
Minamoto shogun, Minamoto Sanetomo,
6
who formed strong links with the royal court
in Heian.
7
It was under Minamoto Sanetomo, for example, that onmyōji were

5
An analysis of this process is summarized in Higuchi Kentaro, "Sekkanke seisho shitsuji no seiritsu to
tenkai."

6
1192 – 1229, r. 1203 – 1229.

7
While Sanetomo can be seen as launching the process of “courtification,” it continued after his
assassination. Works that discuss how the adoption of court practices affected the bureaucratic organization
(or bureaucratization) of the Kamakura shogunate include Andrew Goble, "The Hōjō and Consultative
Government," in Court and Bakufu in Japan: Essays in Kamakura History, ed. Jeffrey Mass (New Haven:
Yale University Press, 1982), and Carl Steenstrup, Hōjō Shigetoki, 1198-1261, and His Role in the History
of Political and Ethical Ideas in Japan, Scandinavian Institute of Asian Studies Monograph Series
(London: Curzon Press, 1979). The latter also includes moral aspects of this cultural trend.

335
permanently established in Kamakura. First to arrive were members of the Miyoshi
lineage and other scholarly families, who were then supplanted by younger sons and
branch lineages of the Abe and Kamo.
8

There are no indications, however, that the Kamakura shogunate produced its
own calendar. It is only in the production of the Mishima-goyomi calendar, which
appeared late in the Kamakura Period at one of the major shrines venerated by the
Kamakura shogunate, that the shogunate’s demand for calendrical knowledge can be seen
to have sparked—or at least promoted—the production of regional calendars. The
appearance of such regional calendars as the Mishima-goyomi can be tied not only to the
weakening economic power and political control by the court bureaucracy (which could
no longer produce and distribute an official calendar to all reaches of Japan), but also to
the increasing need for calendars and almanacs courtly culture became more widespread.
One of the aspects of this culture was the principle individuals and organizations organize
their lives as best they could according to the rhythms of the cosmos depicted in such
almanacs.
And yet, the historical irony is that as personal and public use of Chinese-style
almanacs increased—a use that was predicated on the assumption that the calendar in the
almanac was a standardized, accurate and reliable depiction of cosmological time—it
came to be independently-produced regional calendars that filled that need. Even in the
medieval capital, it was not the official guchūreki almanac that was the standard of

8
This was first covered in Kanazawa Masao, "Kantō tenmon, Onmyōdō seiritsu ni kansuru ikkōsatsu," in
Onmyōdō sōsho, ed. Murayama Shū'ichi (Osaka: Meicho shuppan, 1991–1993)., and the historical model
of this process refined in Akazawa Haruhiko, "Kamakura-ki no kannin onmyōji," Kamakura ibun kenkyū,
no. 21 (2008)..

336
time-keeping, but rather it was the calendar produced and sold by the daikyōshi ( )
guild—the so-called kyō-goyomi ( )—that was more widely used.
9
Nara had a
separate calendar, produced near Kōfukuji by a branch of the Kamo lineage, known as
the “southern capital calendar” ( Jp. nanto goyomi). Through the seventeenth
century, these calendars were all calculated using the Xuanming li system, yet they were
all calculated independently. Copies of calendrical texts survive in archives as far distant
from the capital as Kagoshima Prefecture to the west and Ibaraki Prefecture to the east.
10

The attested differences between regional calendars, such as between the Mishima and
capital calendars—which could amount to a three-day difference, or even a difference in
the beginning of the year, or in the number or season assigned to a lunar month—arise
from this independent production. Variant practice, the resulting conflicting facts, was an
inevitable result of decentralized production of calendars.
As the preceding chapters have shown, the potential for variation in the
production of astronomical calendars based on the Xuanming li system existed almost as
soon as it was imported from Parhae in the ninth century, as concepts and rules within the
system (for example, “advancement-retreat” adjustment in the event of a solar eclipse)
were contested and reinterpreted. Questions of whether adjustments should be made for
astronomical factors that were not visible in Japan—a reasonable doubt, based on the

9
The struggles between the dakyōshi and rivals who received the support of the retired sovereign (the Jp. in no mikyōshi) is summarized in Okada Yoshirō, "Koyomi no shurui," 130-131.

10
The Shimazu and Rokujizōji archives both hold groups of manuscripts related to calendrics. Unknown,
"Shimazu-Ke Denrai Gunjutsu Sho," and Unknown, "Daitō Onmyō Sho," in Rokujizōji seikyō (Tokyo:
Tōkyō Daigaku Shiryō hensanjo, n.d.), respectively.

337
system—drove Japanese divergence. Further room for variation was cleared by the
demands placed on the Xuanming li system to conform to cosmological cycles to which it
was mathematically unsuited, necessitating manual adjustment of its figures. Meanwhile,
debates over calendrical production at the Japanese court established a tradition of
practice that could be called “Japanese-style” calendrical astronomy. Even as outsiders
rejected this tradition of “Japanese-style” calendrical astronomy as being in violation of
the text,
11
the Kamo lineage began to limit who had legitimate access to the details of the
tradition, reading it as “secret transmissions” ( Jp. hiden) and the private possession
of Kamo calendarists. It is unsurprising then that the practice of calendrical astronomy in
the provinces (or even in the capital if produced outside the control of the Kamo lineage)
should diverge from the official calendar. The consistency that does exists can be
attributed to the widespread use of the Xuanming li—despite the continued presence of
the Futian li and other Chinese calendrical systems—and the reality of the astronomical
phenomena that all Japanese calendars were attempting to capture. The Japanese
calendars calculated from the Xuanming li, after all, were still supposed to be an accurate
reflection of observable cosmological phenomena.
Variation and fragmentation are also evident in the practices of identifying and
interpreting astronomical phenomena. The practice of tenmon interpretation was highly
monitored and restricted by the state, as can be seen by controversies in the late Heian

11
Notably the Buddhist monks Shōshō (chapter five) and Zōmyō (chapter six).

338
period.
12
These disputes also show the active role that scholars-nobles played in
monitoring and controlling the practice of tenmon divination. While such control and
textual conservatism explain the remarkable similarity between surviving tenmon
divination reports from the eleventh and seventeenth centuries, other tenmon-like
methods for divination based on astronomical phenomena were not as successfully
controlled. This can be seen in the development of new types of interpretation, as
revealed by Nakahara Morotō II’s 1106 report on the historical impact of comets on
Japanese history.
13
As in a traditional tenmon report, the meaning of the comet was to be
found in historical precedents of similar phenomena. In Morotō’s report, however, the
examples were taken from Japanese history, not from texts from the Spring and Autumn
or Han periods of Chinese history.
And while such traditional tenmon practices continued at court, with all the
restrictions placed on its practice there, demand in the provinces for authoritative
interpretations of spectacular astronomical phenomena did not require such exacting
fidelity. This is shown by a surviving star chart and divination guide preserved in the
archive of a lineage that was subordinate to the Abe.
14
The interpretations of
astronomical phenomena found therein, although derived from Chinese sources, do not
provide much attribution—and what is provided is not from the standard texts of court

12
Particularly the dispute between Abe Suehiro and Kujō Kanezane in the late twelfth century, discussed
in chapters one and two.

13
See chapter two.

14
The Wakasugi Sanke bosan text, held by the Kyoto National Archives. A printed edition can be found
in Daitō Bunka Daigaku Tōyō kenkyūjo, ed. Wakasugike monjo "Sanke bosan" no kenkyū.

339
tenmon. Such a manual would have been useless for composing an official tenmon missō
for reception at court—proven by Abe Suehiro’s difficulties in 1189. Even so, the
existence of such a chart and manual indicates that divination based on such an
abbreviated text seems to have been sufficient to meet the needs of private demand.
Yet such private demand for the interpretation of astronomical phenomena seems
to have been on the wane, particularly when compared to the demand for astronomical
calendars. Surviving accounts of astronomical events from the later Kamakura through
the Sengoku periods, unless taken directly from a source composed by astronomical
observers,
15
are increasingly vague and undetailed. Instead of identifying the planets in
conjunction, the phrase “two-star conjunction” or “three-star conjunction” become more
common. Constellations were rarely identified by name.
16
Comets, auroras, and shooting
stars—phenomena more spectacular and more easily detected by a less-trained
observer—remain prominent in the records.
The reasons for this lack of interest remain unclear, but one reason may be that,
forcibly frozen in documentary styles that rejected innovation, tenmon divination
remained tied to the fossilized Chinese-style state. That it did not respond well to new,

15
For example from the Kahi yōroku ( “Essential Secret Records of the Lineage”) collection,
initiated by the Kamo lineage and taken over by the Tsuchimikado branch of the Abe lineage. (Multiple,
"Kahi yōroku," (Tokyo: Tōkyō Daigaku shiryō hensanjo, 1958 (copy of older manuscript)).) The
astronomical details in this collection generally accord with the types of details found in earlier tenmon
missō copies.
A fuller study of this collection and how it shows continuity or variation from previous practice
remains to be completed, although the land-holding documents on its reversed have been analyzed.
(Atsuya Kazuo, ed. Guchūreki o chūshin to suru koyomi shiryō no shūsei to sono shiryōgaku-teki kenkyū.)
The addressees of the reports do show some variation, including reports addressed to the Ashikaga shoguns
as well as to the throne.

16
This is clearly indicated through a perusal of astronomical records up through 1550 found in
Kanda Shigeru, Nihon tenmon shiryō.
340
and more pressing, individual or regional needs. Another reason may be the lack of
wide-spread training in identifying and classifying astronomical phenomena—a lack
attributable to the only successful limit on astronomical knowledge employed by the
ritsuryō law codes: the control over who could legitimately own or study star maps. The
shift away from detailed records of astronomical phenomena shows clearly that specialist
knowledge of observation and identification, the type of knowledge that the law-codes of
the eighth century had designated as restricted to the Bureau of Onmyō, did not spread
widely even among scholars and elites at court. The desire to learn the meaning of natural
and astronomical phenomena remained strong there, and spread through the provinces as
well, first via elites in the Kamakura shogunate,
17
then through religious institutions.
18

But the textual practices and observational skills of the tenmon specialists at court do not
seem to have made the same journey. While regular observation and interpretation of
astronomical phenomena continued at court, for most courtiers and regional elites the
interpretation of astronomical phenomena became something they demanded of
consulting experts when a spectacular phenomena attracted their attention. Unlike the
high officials of the Heian Period who sought to monitor and correct the practices that
produced astronomical facts on behalf of throne and state, elites of the later medieval
periods seem to have been content to learn the meaning of phenomena only as it related

17
See Akazawa Haruhiko, "Kamakura-ki no kannin onmyōji."

18
Nichiren and other monks of his sect, for example, used natural phenomena in their proselytization
efforts.
Temple records also become a major source of latter records of astronomical phenomena. See
Osaki Shōji, Kinsei Nihon tenmon shiryō.

341
to them. As a practice, therefore, tenmon was less vital to medieval Japanese political and
everyday life than calendrical astronomy.
These changes did not mean, however, that the ideals of tenmon as a guide to
political action, or of an official state-controlled and astronomically-based calendar,
disappeared. To the contrary, their continued existence as ideals, both in Chinese text and
elite imaginations of the ideal state, explains the resurgence in interest concerning both
tenmon practice and the reform and control of calendrical astronomy in the seventeenth
century. There are stories of Sengoku warlords attempting to get astronomical time to
conform to calendrical time, or at least to standardizing it, in the sixteenth century. A
legendary account asserts that around 1582, the Go-Hōjō hegemons of eastern Japan
abolished a regional calendar produced in Ōmiya when it was discovered to have a
different depiction of the days and months of the year from the more prestigious
Mishima-goyomi.
19
The legislative moves by the Tokugawa shogunate to limit the
publication and distribution of calendars in Japan to a limited number of approved guilds
in 1691, therefore, was not an unprecedented event.
20
Neither was the Meiji
government’s program to control the publication of calendars after the Gregorian reform
of 1872, even though the type of calendar was a radical break from the previous system

19
A brief description of the account can be found in Okada Yoshirō, "Koyomi no shurui," 140. The
source that the story comes from, the Hōjō godai-ki, however, not only fails to give a year for the event, but
was written long afterwards. Therefore certain elements of the tale remain subject to doubt.

20
The Tokugawa shogunate’s censorship over calendars is covered in Kornicki, The Book in Japan: A
Cultural History from the Beginnings to the Nineteenth Century, 353-357.

342
of timekeeping.
21
Both reforms tapped into the ideal of time as correctly determined
from astronomical phenomena by a divinely-inspired government, distributed to the
population for their health and benefit. Deviation from this “state time” was a deviation
from the state. What was new in the seventeenth and nineteenth centuries was a level of
centralized power and control to effect such standardization. Such control had not been
exercised consistently by the court since at least the fourteenth century, and likely not
before. After all, unofficial calendars had been approved of by Japanese elites as early as
the eleventh century.
22

The practice of tenmon was also expanded and revived in the seventeenth century,
at least temporarily. The shogunal astronomer who devised a new system of calendrical
astronomy in 1684, Shibukawa Shunkai,
23
was not merely involved in astronomical
observation for the purpose of mathematical astronomy. A student of the then
Tsuchimikado head of the Abe lineage,
24
he was also involved in interpreting the
astrological meaning of astronomical phenomena. Although no divination reports from

21
See Okada Yoshirō, Meiji kaireki: “Toki” no bunmei kaika, (Tokyo: Taishūkan shoten, 1994), and
Steven Tanaka, New Times in Modern Japan (Princeton, NJ: Princeton University Press, 2004), 5-26 for
some of the political implications of this shift for the Meiji government and for society.
Contrary to the radical break described by Tanaka, however, are surviving official calendars
from Ise for the years Meiji 8 through 30 (1874-1897), which re-introduce lunar phase information, only to
slowly phase it out over the next decades. (As shown by a collection of Ise-goyomi held by the Museum of
Kōgakkan University, Mie Prefecture. Unknown, "Ise goyomi," (Ise, Mie Prefecture: Kōgakkan University
Museum, 19-20th cen.).)

22
See, for example, Fujiwara Sanesuke and Fujiwara Sanefusa’s approval of Shōshō’s calendar for 1039.
A difference between this case and the case of the independent Mishima-goyomi, however, is that Shōshō
and his partisans wanted to see his calendar officially adopted by the state.

23
1639 – 1715.

24
One Tsuchimikado Yasutomi ( 1655 – 1717).

343
him survive, instructions for interpreting phenomena can be found in his 1698 work
Tenmon keitō.
25
Although English-language scholarship has characterized Shibukawa
Shunkai as uninterested in divination, and the Tenmon keitō work itself is said to mark a
shift from astrology to astronomy,
26
it is clear that Shunkai’s career had more to do with
the traditions of the Bureau of Onmyō than has previously been supposed.
Although the revival of tenmon in the seventeenth century coincides with the
resumption of state control over the practice of calendrical astronomy, whether tenmon
was as important to the state, the shogunate, and the local daimyo courts as it was to the
Heian-period nobility has yet to be determined. Regionally, observatories were
established, not just in Edo for the Tokugawa shogunate, but also in Sendai and Nagasaki.
Whether the tenmonkata ( ) in charge of the Edo observatory office, or his regional
imitators, regularly produced divination along with observations and astronomical
predictions has not yet been a major topic of study.
27
The roots of these early modern
developments, however, can be found in dissatisfaction with the fragmented condition of

25
A printed version of this text is collected in Furushima Toshio et al., ed. Kinsei kagaku shisō vol. 2
(Tokyo: Iwanami shoten, 1971–1972).

26
See specifically Nakayama, A History of Japanese Astronomy: Chinese Background and Western
Impact.

27
Only recently have works on the organization of Edo-period Onmyōdō become prominent, particularly
Hayashi Makoto, Kinsei Onmyōdō no kenkyū (Tokyo: Yoshikawa kōbunkan, 2005) and Umeda Chihiro,
Kinsei Onmyōdō soshiki no kenkyū (Tokyo: Yoshikawa kōbunkan, 2009). These works touch upon tenmon
and Shibukawa Shunkai insofar as his relationship with the Tsuchimikado lineage organization, which is
their main focus.
Current research on early modern astronomy and related subjects is also being undertaken by
graduate students in the United States as well, particularly Yulia Frumer of Princeton, so more information
about these regional observatories and their astronomers should be forthcoming.

344
medieval practice. And the primary causes of that fragmentation reach into the ways that
the practice of tenmon and of calendrical astronomy changed during the Heian Period.

Previously, the practices of tenmon and calendrical astronomy were assumed to
have been largely static from the time of their importation to Japan from China until the
Jōkyō calendrical reform of 1684.
28
Blame for such stasis has been generally placed at
the feet of the Abe and Kamo lineages, who are said to have seized monopolistic control
over astronomical observation, interpretation, and calculation in the tenth century. As this
dissertation has shown, however, this was not the case. Furthermore, even in relatively
faithful execution of these texts and practices after Chinese models, there was much room
for variation and deviation. Whether such potential for change was realized had much to
do with whether attempted changes were accepted by the court, particularly the Council
of State, which oversaw the production of astronomical facts in Japan.
29
From the
inherent uncertainty in the practices of tenmon and of calendrical astronomy as imported
from China, and the various demands made on these practices by the court, a tradition of
practice specific to the Japanese court began to develop. This tradition need not have

28
Tenmon has largely been ignored by historiography, due to its failure to conform with aspects of the
history of science and its failure to fit into a comfortable location in a religious institution. Despite this,
when tenmon is mentioned, it is primarily considered a faithful reproduction of Chinese practice. See
Shigeru Nakayama, "Characteristics of Chinese Astrology," Isis 57, no. 4 (1966).

29
Therefore, Shigeru Nakayama’s statement that, as far as calendrical astronomy was concerned,
“[r]evisions were made infrequently” (Nakayama, A History of Japanese Astronomy: Chinese Background
and Western Impact, 116), is not necessarily incorrect. That “there was in fact a long period of indifference
to calendar reform,” however, seems inconsistent with recorded discontent with the calendrical situation as
it existed in late medieval Japan, some of which has been covered above. Therefore it seems more accurate
to say that for a long time no single institution, or alliance of institutions, had the authority to effect
calendrical reform until the political changes of the seventeenth century.

345
been unique to Japan, as any state employing Chinese state astronomy might have trod a
similar path. But certainly the circumstances of this tradition’s development are
intimately tied to Japan’s social history. By the twelfth century, these Japanese traditions
had been finaly claimed as the private monopoly of the Abe and Kamo lineages, through
reference to esoteric knowledge and secret transmissions. These claims to knowledge
were readily accepted by the Japanese court.
It is this late Heian development that spawned the image of Abe-Kamo hereditary
control over the Bureau of Onmyō, and read it back in time to the early Heian Period. In
the fifteenth-century text Kuge kongen ( “The Roots of State Affairs”), the
scholar, poet, and regent Ichijō Kaneyoshi
30
described the duties of the Bureau of
Onmyō as “divided” between the Abe and Kamo lineages. According to his work, the
Abe were in charge of tenmon, and the Kamo were in charge of calendrical astronomy.
This dated, he said, from the middle of the tenth-century, when Kamo Yasunori divided
his knowledge between his disciple Abe Seimei and his son Kamo Mitsuyoshi. This story
has been picked up and repeated since the fifteenth century, and can be found even in
relatively recent works of scholarship on the history of Japanese religions.
31

Nevertheless, for all that the traditional “division” of the Bureau of Onmyō
between the Abe and Kamo lineages was supposed to have been effected by Kamo

30
1402 – 1481. An alternate reading of his personal name is “Kanera.” A study of his writings
(largely focused on his poetry) and political activities can be found in Steven D. Carter, Regent Redux: A
Life of the Statesman-Scholar Ichijō Kaneyoshi (Ann Arbor, Mich.: Center for Japanese Studies, the
University of Michigan, 1996).

31
Significantly these formed the primary source used in Murayama Shū'ichi, Nihon Onmyōdō shi sōsetsu,
and ———, Nihon Onmyōdō shiwa.

346
Yasunori in the tenth century, it is not until the twelfth century that explicit reference to
secret transmissions that could not be shared with outsiders
32
—or that were otherwise
inaccessible to them
33
—appears in the arguments of tenmon or calendrics specialists from
these lineages.
34
The presence of non-Abe and non-Kamo members of the Bureau is also
well attested to through the start of the eleventh century.
35
As Jacqueline Stone has
described in her work on the formation of lineages in Tendai Buddhism (specifically kuden, or “oral transmission” lineages), the process of creating a transmission lineage
was simultaneously forward-looking and retrospective.
36
That is, it rewrote the past even
as it created an authoritative system of transmission through which to pass knowledge on
to the future. The kuden for tenmon and calendrical astronomy seem to have worked this
way as well, as the “reading backwards” of Abe and Kamo dominance attests.
Interestingly, the development of lineages within the Bureau of Onmyō appears to
predate the appearance of transmission lineages in esoteric Buddhism—at least as judged
by surviving texts.
37
The legend of the primacy and dominance of the Abe and Kamo

32
See the calendrical debate of 1123, covered in the preceding chapter.

33
For example, Abe Yasushige’s citation of ritual confirmations of the identity of astronomical
phenomena. See chapter two.

34
The use of family tradition by Ōkasuga Masumitsu in 950 (covered in chapter three) is a glaring
exception, and shows that ideas of lineage training and lineage-specific knowledge significantly predates
both Abe and Kamo dominance within the Bureau of Onmyō.

35
See Shigeta Shin'ichi, Onmyōji to kizoku shakai.

36
Stone, Original Enlightenment and the Transformation of Medieval Japanese Buddhism, 106.

37
While the tradition of oral transmission is said to derive from kenmitsu Buddhist training from the late
Heian period, the texts and mature kuden transmissions date from much later. See Ibid, 101 as well as the
rest of her third chapter (97-152).

347
lineages from their very appearance in Bureau of Onmyō activities, as exemplified by the
figures of Abe Seimei and Kamo Yasunori, was a vital part of the legitimacy of their
fifteenth-century descendants who were Ichijō Kaneyoshi’s informants. Kaneyoshi’s
depiction of certain knowledge and texts of the Bureau of Onmyō as the true property of
a few lineages matched the “ownership” of texts and interpretative traditions by both
Buddhist and literary lineages of his day.
38

Yet as the preceding chapters also show—particularly in the discussion of
calendrical astronomy which is better represented in the historical sources of the early
Heian period—the identification of particular lineages with the Bureau of Onmyō and
with the expert practice of Chinese technologies of knowledge predates both the Abe and
Kamo lineages’ association with the Bureau.
39
In the seventh and eighth centuries,
immigrant lineages at court took over specific roles and authority through their
knowledge of continental technologies.
40
In other words, while the dominance of only a
single lineage for any one field and post (for example, the Abe for the Instructor of
Tenmon position and the Kamo for the Instructor of Calendrics position) was new in the

38
For the Buddhist side of things, see again Ibid.. On the esoteric traditions in literature, see Klein,
Allegories of Desire: Esoteric Literary Commentaries of Medieval Japan.

39
The use of the word “technology” here owes much to Piggott, The Emergence of Japanese Kingship.
and Francesca Bray, Technology and Gender: Fabrics of Power in Late Imperial China (Berkeley:
University of California Press, 1997). In the latter work particularly, the “technology” is broadened to
include techniques, practices, as well as material machines.

40
See Como, Shōtoku: Ethnicity, Ritual, and Violence in the Japanese Buddhist Tradition; ———,
Weaving and Binding: Immigrant Gods and Female Immortals in Ancient Japan; and Ooms, Imperial
Politics and Symbolics in Ancient Japan: The Tenmu Dynasty, 650-800.
Interestingly for the depiction of the history of Onmyōdō and related fields in these works, two
of the major lineages of the early Heian Bureau of Onmyō, the Ōkasuga and Katsuragi, are not described as
having continental roots in the early ninth-century Shinsen shōjiroku.

348
twelfth century, it built upon a much older concepts. The ideas of the transmission of
knowledge through a lineage, and of lineage succession to posts within the Japanese court
bureaucracy, can be found even in the early Heian period. This is further evidenced by
the text known as Takahashi ujibumi ( , eighth century), a history of the
Takahashi lineage submitted to the Council of State in support of their suit against the
Azuma lineage over precedence and procedure in the Office of the Palace Kitchen.
Lineage-specific and lineage-controlled knowledge predate their later significance in
medieval esoteric Buddhism and literature. Yet it was medieval Buddhist kuden rituals
and methodologies that played a major role in medieval Japanese literary lineages. It has
been hypothesized that kuden ideals had influence in other intellectual fields as well, but
what influence the kuden transmission might have had on either tenmon or calendrical
astronomy has yet to be uncovered.
41

The traditions that formed the “secret texts” of the Abe and Kamo lineages of the
late Heian Period certainly differed significantly from the material found in medieval
Buddhist kuden transmissions. The interpretations of esoteric texts passed down through
such transmission lineages were deeply symbolic, based on correspondences of meaning
found between textual reference and symbolic referent.
42
In contrast, the traditions of
practice for both tenmon and calendrical astronomy as seen in surviving references during
court debates were tied to identifications of specific astronomical phenomena for tenmon,

41
This does seem a fruitful avenue of study. In particular, a major work of Onmyōdō, the Hoki naiden,
traditionally ascribed to Abe Seimei but assigned by the Abe to the work of an unnamed Shingon monk,
explains aspects of calendrical astronomy—in particular, hemerology—in terms of the esoteric Buddhist
tale tradition. Other influences and confluences likely await further investigation.

42
Definition taken from Klein, Allegories of Desire: Esoteric Literary Commentaries of Medieval Japan.

349
and specific modifications to be performed on calendars calculated using the Xuanming li
for calendrical astronomy. Although the social role that such transmissions played in
legitimating who had “correct” possession of “correct” knowledge was the same in both
the technical fields of tenmon and calendrical astronomy and in Buddhist scholarly
lineages,
43
the content of the transmissions seems to be quite distinct, at least for the
Heian Period. How the growing influence of ritual initiations to knowledge, a practice
that spread from esoteric Buddhism to the arts in Japan, may or may not have influenced
the practices of tenmon or calendrical astronomy in later medieval Japan is another
subject for further study.
The importance of lineages in the production of knowledge in Japan is not merely
an issue of intellectual history. It reflects ways in which political power was organized in
the classical and medieval periods. In his overview of Japanese history, John Whitney
Hall identified pre-ritsuryō and post-ritsuryō periods of “familial authority,” when
particular lineages were pre-eminent in politics, in contrast to the ritsuryō state which
prescribed reliance on legal authority.
44
What the history of the Bureau of Onmyō shows
is that even within the ritsuryō bureaucracy, such “familial authority” remained strong.
While this might seem a violation of Max Weber’s “bureaucracy” ideal type, wherein all

43
A role that might be said to be the function of a secret, which paradoxically has no power if no one else
is aware that the possessor has it. On secrets and secrecy and the problems studying them, see Mark
Teeuwen, "Introduction: Japan's Culture of Secrecy from a Comparative Perspective," in The Culture of
Secrecy in Japanese Religion, ed. Bernhard Scheid and Mark Teeuwen (New York: Routledge, 2006).

44
John Whitney Hall, Government and Local Power in Japan, 500 to 1700; a Study Based on Bizen
Province (Princeton, NJ: Princeton University Press, 1966). See also G. Cameron Hurst, "The Structure of
the Heian Court: Some Thoughts on the Nature of 'Familial Authority' in Heian Japan," in Medieval Japan:
Essays in Institutional History, ed. John Whitney Hall et al. (New Haven: Yale University Press, 1988),
39-59.

350
the authority is held by the organization of the state, such continuity both explains the
resurgence of lineage-based authority in Japan’s medieval period and characterizes the
fragmented nature of authority even in classical Japan. The ritsuryō system added a new
type of authority, the legal authority of state bureaucratic organization; and it also
strengthened this authority. But the older modes of power were not supplanted or
replaced.
45
Besides, as an examination of the modern world shows, admixtures are far
more common than ideal types.
The study of the history of tenmon and calendrics in Japan also raises the question
of how different types of authority mix in practice. The ideal of the ritsuryō state,
expressed in the legal codes of the eighth century, had it that all authority depended on
the proper practice of bureaucracy, the authority of which relied on the faithful execution
of duties to the throne. In practice, however, the state relied upon the authority of older
organizations of kami worship, and newer ideals of Buddhist kingship, as well as upon
connections to networks of local elites, in order to effect its hegemony over the islands.
46

The complex nature of authority can be seen in the production of time itself in the
ritsuryō state. While the calendar is in itself correct and authoritative because it is
produced by the state, at least from the standpoint of the ritsuryō codes, the national
histories of Nara and early Heian Japan show that the “correctness” of the calendar was

45
This is similar to the model of the mature Japanese royal state descripted in Piggott, The Emergence of
Japanese Kingship., where the state fails to fully control appointments all the way down the hierarchy, to
the district ( Jp. gunji) level.

46
The complex nature of Japanese kingship, derived from its long history, is the subject of Ibid..

351
seen as dependent upon its proper production according to systems laid out in Chinese
texts. The production itself was dependent on the skills of a lower-ranking subset of
technical officials within the court bureaucracy. The “correctness” of their products
depended upon their training within the Bureau of Onmyō, and their “correctness”
supervised by the top levels of the court.
That what was defined as “proper production according to systems laid out in
Chinese texts” in Japan diverged from Chinese practice shows a growing Japanese
confidence in their own history of practice and in their own reading of Chinese texts from
the middle of the Heian Period. The re-introduction of the Rule Cycle into the calendrical
system of the Xuanming li was a sort of classicism not found in Chinese calendrical
astronomy. Likewise the production of tenmon reports, while reliant on quotations from
Chinese sources, also drew from the archive of Japanese practice and history. Innovation
was not easily accepted, even if based on Chinese sources, when it deviated from this
history of practice.
47
Therefore, despite the Chinese origin of these technologies, that
there was a Japanese history and “Japanese-style” of practice is clear.
What is also clear is the anxiety over correct performance. During the Heian
Period, the court was monitoring the production of its specialists, and intervening when it
thought necessary. Despite the image of the astrologer or member of the Bureau of
Onmyō (particularly the role of these men as onmyōji, which many of them were) as
something like a Rasputin with amazing influence and control over the Japanese nobility,

47
See for example the difficulty that Abe Yasushige had with Kujō Kanezane, despite his use of the [Liu]
Song History ( Ch. Song shu, Jp. Sō sho) as an authority, described in chapter two.

352
the historical accounts of disputes and debates over the identification and prediction of
astronomical phenomena show that members of the Bureau of Onmyō, and their products,
were viewed with a measure of uncertainty, and even mistrust. There was certainly no
professional independence for specialists in matters of astronomy here.
48
Just as clearly,
such “diviners” did not have control over medieval minds.
Even though the practices of tenmon and calendrical astronomy themselves had
authority derived from their ancient Chinese origin, the same authority did not
necessarily spill over to contemporaneous views of the practitioners. Some of this doubt
certainly derived from the fact of conflicting reports from individual specialists—yet the
cultivation of different groups of specialists for consultation was something that the
Japanese court itself had undertaken in the search for certainty. As specialists vied with
each other to have their position heard and validated, interpersonal relationships became
key. Connections with patrons at court could get a critique of official production heard,
49

and displays of skill in astronomical affairs could reap material rewards.
50
Likewise, a
long-standing relationship with an individual specialist would provide elite members of

48
One of the definitions of professionalization is the group of “professional” specialists’ recognized
authority over their field and members (“gatekeeping”). See Andrew Alcott, The System of the Professions:
An Essay on the Division of Labor (Chicago: University of Chicago Press, 1988) for this definition.
A weaker “semi-professionalism” has been identified for Confucian scholars in Ming China
and for martial arts in pre-modern Japan. See G. Cameron Hurst, Armed Martial Arts of Japan:
Swordsmanship and Archery (New Haven: Yale University Press, 1998).

49
As in the case of the monk Shōshō, covered in chapter five.

50
As shown by the rewards received by Kamo calendarists and monks for correct eclipse predictions, as
well as for calendars presented to court during the first year of a Rule Cycle.
The presence of Abe and Kamo members of the Bureau of Onmyō in provincial government
posts also point to this rewarding, but cannot be untangled for these individuals from their private service in
elite households in ritual and divination matters. Client-patron relationships stretched into many realms of
Japanese court life.

353
court with access to specialized knowledge that they would not otherwise receive. The
nobility could call upon a specialist they had an established relationship with, and who
had proved trustworthy, to evaluate astronomical “facts” they found suspicious or merely
inconvenient. Interpersonal connections—the basis for trust—was key to individuals
searching both for “correct” knowledge of the cosmos and for those who wanted to
ensure the state employed correct knowledge.
It is this interpersonal element of the production of facts about astronomical
phenomena for the state in Japan that explains, in part, why the fragmentation of practice
was allowed to stand. The authoritative production of knowledge, by the late Heian
Period, was a matter of individual production—it depended on the skill of an individual
specialist working in a recognized tradition. The authority was not to be found in the
Bureau of Onmyō itself, or in the fact of official production. To the extent that the texts
were known, such production could be monitored. Even if the text freely circulated,
however, the contents were difficult and obscure, and the growing recognition of
non-textual knowledge passed down in lineages meant that monitoring the performance
of tenmon or of calendrical astronomy by non-specialists became more difficult. A
relationship with a trusted expert, then, was the only way to monitor the production of
“truth.” But which individuals were to be trusted? The answer depended on whose
viewpoint is taken. Without a centralized authority to determine what was true and what
was false, those elites inclined to dissent could search out new specialists they thought
more congenial. The continuity of tenmon divination at court and calendrical production
across Japan using the base system of the Xuanming li attest to the authority of those
354
practices. The facts they produced, however, in their variety and inconsistency, were
anything but authoritative.
It is this situation that explains both the continuity and the variation in Japanese
practices of astronomical observation and prediction for much of the premodern period.
No one questioned the relevance and validity of tenmon divination. After the introduction
of the Xuanming li system in 862, the question of the system’s correctness was not in
doubt—the correctness of the individual implementation was. That there was a “correct”
interpretation or prediction for astronomical phenomena was believed, even as the
particular interpretation or prediction was doubted. The practices were trusted, but not the
practitioners.
As a result, there were no calls for a “revolution” in practice to fix errors
everyone knew were present. For the Japanese court in the Heian Period, and for some
time afterwards, the fault was not in the process but in the individual using that process.

355

BIBLIOGRAPHY

Abe, Ryū'ichi. The Weaving of Mantra: Kūkai and the Construction of Esoteric Buddhist
Discourse. New York: Columbia University Press, 1999.
Adolphson, Mikhael. Gates of Power: Monks, Courtiers and Warriors in Premodern
Japan. Honolulu: University of Hawai'i Press, 2000.
Akazawa Haruhiko. "Kamakura-ki no kannin onmyōji." Kamakura ibun kenkyū no. 21
(2008).
Alcott, Andrew. The System of the Professions: An Essay on the Division of Labor.
Chicago: University of Chicago Press, 1988.
Amino Yoshihiko. Nihon chūsei no hinōgyō-nin to tennō. Tokyo: Iwanami shoten, 1984.
Aritomi Jun'ya. "Hyakushō buiku to ritsuryō kokka: Jukyōteki ideorogii seisaku o
chūshin ni." Shigaku zasshi 112, no. 7 (2003).
Atsuya Kazuo, ed. Guchūreki o chūshin to suru koyomi shiryō no shūsei to sono
shiryōgaku-teki kenkyū. Tokyo: Tōkyō Daigaku Shiryō hensanjo, 2008.
Batten, Bruce. Gateway to Japan: Hakata in War and Peace, 500-1300. Honolulu:
University of Hawai'i Press, 2006.
Bender, Ross. "Changing the Calendar: Royal Political Theology and the Suppression of
the Tachibana Naramaro Conspiracy of 757." Japanese Journal of Religious
Studies 37, no. 2 (2010).
Bock, Felicia. Classical Learning and Taoist Practices in Early Japan: With a
Translation of Books XVI and XX of the Engi-Shiki, Occasional Paper (Arizona
State University. Center for Asian Studies). Tempe, Ariz.: Center for Asian
Studies, Arizona State University, 1985.
Borgen, Robert. Sugawara No Michizane and the Early Heian Court, Harvard East Asian
Monographs. Cambridge, Mass.: Harvard University Press, 1986.
Bray, Francesca. Technology and Gender: Fabrics of Power in Late Imperial China.
Berkeley: University of California Press, 1997.
356
Breen, John. "Inside Tokugawa Religion: Stars, Planets, and the Calendar-as-Method."
Culture and Cosmos 10, no. 1–2 (2006): 63–79.
Buhrman, Kristina. "Conceptions of the Court and Ritual in Medieval Japan: Examples
from the Taiheiki." In Redefining Identities in Asia: 9th Graduate Student
Conference. Harvard East Asia Society, Cambridge MA, 2006.
Bussho kankōkai. Asabashō. 7 vols, Dai Nihon Bukkyō zensho. Tokyo: Bussho kankōkai,
1912–1914.
Butler, Lee. Emperor and Aristocracy in Japan, 1467-1680: Resilience and Renewal,
Harvard East Asia Monographs. Cambridge, Mass.: Harvard University Press,
2002.
Carter, Steven D. Regent Redux: A Life of the Statesman-Scholar Ichijō Kaneyoshi. Ann
Arbor, Mich.: Center for Japanese Studies, the University of Michigan, 1996.
Chalitpatanangune, Marisa "Heiji Monogatari : A Study and Annotated Translation of the
Oldest Text." University of California, Berkeley, 1987.
Chin, Y. N. and Y. L. Huang. "Identification of the Guest Star of Ad 185 as a Comet
Rather Than a Supernova." Nature no. 371 (1994): 398–99.
Como, Michael. Shōtoku: Ethnicity, Ritual, and Violence in the Japanese Buddhist
Tradition. New York: Oxford University Press, 2008.
———. Weaving and Binding: Immigrant Gods and Female Immortals in Ancient Japan.
Honolulu: University of Hawai'i Press, 2010.
Conlan, Thomas. From Sovereign to Symbol: An Age of Ritual Determinism in
Fourteenth-Century Japan Oxford: Oxford University Press, 2011.
Crossley, John N. et al., ed. The Nine Chapters on the Mathematical Art : Companion
and Commentary. Oxford: Oxford University Press, 1999.
Cullen, Christopher. Astronomy and Mathematics in Ancient China: The Zhou Bi Suan
Jing, Needham Research Institute Studies. Cambridge: Cambridge University
Press, 1996.
———. "Translating *Sukh/Xiu and *Lhah/She—‘Lunar Lodges’, or Just Plain
‘Lodges’?" East Asian Science Technology & Medicine no. 33 (2011): 83–96.
Daitō Bunka Daigaku Tōyō kenkyūjo, ed. Kōraishi rekishi senmyōreki no kenkyū. Tokyo:
Daitō Bunka Daigaku Tōyō kenkyūjo, 1998.
357
———, ed. Wakasugike monjo "Sanke bosan" no kenkyū. Tokyo:
Daitō Bunka Daigaku Tōyō kenkyūjo, 2004.
Dear, Peter. "Science Studies as Epistemography." In The One Culture?: A Conversation
About Science, edited by Jay A. and H. M. Collins Labinger. Chicago: University
of Chicago Press, 2001.
Dull, Jack L. "A Historical Introduction to the Apocryphal (Chʻan-Wei) Texts of the Han
Dynasty." University of Washington, 1966.
Dykstra, Yoshiko Kurata. The Konjaku Tales. Japanese Section: From a Medieval
Japanese Collection. 3 vols, Intercultural Research Institute Monograph Series.
Osaka: Intercultural Research Institute, Kansai Gaidai University Publication,
1998–2003.
Elias, Norbert (with Peter Ludes). "Knowledge and Power: An Interview by Peter
Ludes." In Society & Knowledge: Contemporary Perspectives in the Sociology of
Knowledge & Science, edited by Nico Stehr and Volker Meja, 203–42. New
Brunswick, NJ: Transaction Publishers, 2005.
Elman, Benjamin. On Their Own Terms: Science in China, 1550-1900. Cambridge,
Mass.: Harvard University Press, 2005.
Elvin, Mark. "Who Was Responsible for the Weather? Moral Meteorology in Late
Imperial China." Osiris no. 13 (1998): 213-37.
Enchin. "Sasa Gimon." In Dai Nihon Bukkyō zensho, edited by Bussho kankōkai. Tokyo:
Meicho Fukyūkai, 1978.
Farris, William Wayne. "Famine, Climate, and Faming in Japan, 670-1100." In Heian
Japan, Centers and Peripheries, edited by Mikhael et al. Adolphson. Honolulu:
University of Hawai'i Press, 2007.
———. Japan's Medieval Population: Famine, Fertility, and Warfare in a
Transformative Age. Honolulu: University of Hawai'i Press, 2006.
Frank, Bernard. Kata-Imi Et Kata-Tagae: Etude Sur Les Interdits De Direction À
L'époque Heian, Bulletin De La Maison Franco-Japonaise. Tokyo: Isseidō, 1958.
Friday, Karl. The First Samurai: The Life and Legend of the Warrior Rebel Taira
Masakado. Hoboken, NJ: John Wiley & Sons, 2008.
Fujiwara Sanesuke. "Ononomiya nenjū gyōji." In Gunsho ruijū. Tokyo: Yagi shoten,
1960.
358
Fujiwara Sukeyo. "Nihon-koku genzai sho mokuroku." In Zoku gunsho ruijū, edited by
Zoku gunsho ruijū kanseikai. Tokyo: Yagi shoten, 1958.
Furushima Toshio et al., ed. Kinsei kagaku shisō Vol. 2. Tokyo: Iwanami shoten,
1971–1972.
Goble, Andrew. "The Hōjō and Consultative Government." In Court and Bakufu in
Japan: Essays in Kamakura History, edited by Jeffrey Mass. New Haven: Yale
University Press, 1982.
Gomi Fumihiko. Kyō Kamakura no ōken, Nihon No Jidaishi. Tokyo: Yoshikawa
kōbunkan, 2003.
———. Shomotsu no chūseishi. Tokyo: Misuzu shobō, 2003.
———. Yakudō suru chūsei. Vol. 5, Nihon no rekishi. Tokyo: Shogakkan, 2008.
Grappard, Allan G. "Religious Practices." In Cambridge History of Japan, edited by
Donald Shivley and William McCullough. Cambridge: Cambridge University
Press, 1999.
Haboush, JaHyun Kim. "Contesting Chinese Time, Nationalizing Temporal Space:
Temporal Inscription in Late Chosŏn Korea." In Time, Temporality, and Imperial
Transition: East Asia from Ming to Qing, edited by Lynn A. Struve. Honolulu,
Hawaii: Association for Asian Studies, 2005.
Hall, John Whitney. Government and Local Power in Japan, 500 to 1700; a Study Based
on Bizen Province. Princeton, NJ: Princeton University Press, 1966.
Hashimoto Yoshihiko. Fujiwara no Yorinaga. Tokyo: Yoshikawa kōbunkan, 1964.
Hayakawa Shōhachi. "Ryō no shūge." In Kokushi daijiten, edited by
Kokushi daijiten henshū iinkai. Tokyo: Yoshikawa kōbunkan, 1979–1997.
Hayashi Makoto. Kinsei Onmyōdō no kenkyū. Tokyo: Yoshikawa kōbunkan, 2005.
Henderson, John. The Development and Decline of Chinese Cosmology. New York:
Columbia University Press, 1986.
Higuchi Kentaro. "Sekkanke seisho shitsuji no seiritsu to tenkai." Shigaku zasshi 116, no.
2 (2007).
Ho Peng Yoke. Chinese Mathematical Astrology: Reaching out to the Stars. London:
RoutledgeCurzon, 2003.
Hosoi Hiroshi. Kodai no tenmon ihen to shisho. Tokyo: Yoshikawa kōbunkan, 2007.
359
———. "Onmyō-ryō to tenmon rekigaku kyōiku." Tenmongaku-shi kenkyū-kai shūroku
no. 2 (2008).
———. "Tenmondō to rekidō: Kodai ni okeru seiritsu no haikei to sono yakuwari." In
Onmyōdō No Kōgi, edited by Hayashi Makoto and Koike Jun'ichi. Kyoto: Sagano
shoin, 2002.
Hosoi Hiroshi and Minezaki Ryōichi. "'Nihon tenmon shiryō' mishūroku no nisshoku to
kiroku: Rikkokushi shūryō ikō senroppyaku-nen izen." Kwassui ronbunshū:
bungakubu ningen kankei gakka, ongaku gakubu hen no. 45 (2002): 145–66.
———. "Rikkokushi mishūroku no nisshoku to kokushi." Kwassui ronbunshū: ippan
kyōiku, ningen kankei gakka, ongaku gakubu hen no. 43 (2000): 124–12.
Hurst, G. Cameron. Armed Martial Arts of Japan: Swordsmanship and Archery. New
Haven: Yale University Press, 1998.
———. Insei: Abdicated Sovereins in the Politics of Late Heian Japan, 1086-1185. New
York: Columbia University Press, 1976.
———. "The Kōbu Polity—Court-Bakufu Relations in Kamakura Japan." In Court and
Bakufu in Japan: Essays on Kamakura History, edited by Jeffrey Mass. New
Haven: Yale University Press, 1982.
———. "The Structure of the Heian Court: Some Thoughts on the Nature of 'Familial
Authority' in Heian Japan." In Medieval Japan: Essays in Institutional History,
edited by John Whitney Hall et al. New Haven: Yale University Press, 1988.
Iinuma Kenji. "Miyoshi Nagahira." In Kokushi daijiten, edited by
Kokushi daijiten henshū iinkai. Tokyo: Yoshikawa kōbunkan, 1979–1997.
Imatani Akira. Muromachi no ōken: Ashikaga Yoshimitsu no ōken sandatsu keikaku.
Tokyo: Chūō kōron, 1990.
Kalinowski, Marc. "Diviners and Astrologers under the Eastern Zhou: Transmitted Texts
and Recent Archeological Discoveries." In Early Chinese Religion, edited by
John Lagerwey and Marc Kalinowski. Leiden: Brill, 2009.
Kanazawa Masao. "Kantō tenmon, Onmyōdō seiritsu ni kansuru ikkōsatsu." In Onmyōdō
sōsho, edited by Murayama Shū'ichi. Osaka: Meicho shuppan, 1991–1993.
Kanda Shigeru. Nihon tenmon shiryō. Tokyo: Maruzen kabushiki kaisha, 1935.
Kang, Jae-eun. The Land of Scholars: Two Thousand Years of Korean Confucianism.
Paramus, NJ: Homa & Sekey Books, 2006.
360
Keizai zasshisha, ed. Ruijū sandaikyaku. Ruijū fusenshō. Zoku sajōshō, Kokushi taikei.
Tokyo: Keizai zasshisha, 1900.
Klein, Susan Blakeley. Allegories of Desire : Esoteric Literary Commentaries of
Medieval Japan, Harvard-Yenching Institute Monograph Series. Cambridge,
Mass.: Harvard University Press, 2002.
Kojima Noriyuki, ed. Nihon shoki tsūshō. Kyoto: Rinsen shoten, 1978.
Kokushi taikei hensankai, ed. Azuma kagami. 4 vols, Kokushi taikei. Tokyo: Yoshikawa
kōbunkan, 1978–1979.
———, ed. Chōya gunsai, Kokushi taikei. Tokyo: Yoshikawa kōbunkan, 1964.
———, ed. Fusō ryakki; Teiō hennenki, Kokushi taikei. Tokyo: Yoshikawa kōbunkan,
1929.
———, ed. Honchō seiki, Kokushi taikei. Tokyo: Yoshikawa kōbunkan, 1999.
———, ed. Hyakuren shō, Kokushi taikei. Tokyo: Yoshikawa kōbunkan, 1979.
———, ed. Kokushi bunin. 6 vols, Kokushi taikei. Tokyo: Yoshikawa kōbunkan,
1977–1979.
———, ed. Nihon kiryaku. 3 vols, Kokushi taikei. Tokyo: Yoshikawa kōbunkan,
1985–1988.
———, ed. Nihon kōki, Kokushi taikei. Tokyo: Yoshikawa kōbunkan, 1978.
———, ed. Nihon sandai jtsuroku. 2 vols, Kokushi taikei. Tokyo: Yoshikawa kōbunkan,
1979.
———, ed. Nihon shoki. 2 vols, Kokushi taikei. Tokyo: Yoshikawa kōbunkan, 1979.
———, ed. Ruijū sandai kyaku. Vol. 13, Kokushi taikei. Tokyo: Yoshikawa kōbunkan,
1979.
———. Ryō no gige, Kokushi taikei. Tokyo: Yoshikawa kōbunkan, 1979.
———. Ryō no shūge. 2 vols, Kokushi taikei. Tokyo: Yoshikawa kōbunkan, 1979.
———, ed. Shoku Nihongi. 2 vols, Kokushi taikei. Tokyo: Yoshikawa kōbunkan, 1979.
Kokusho meichō kankōkai, ed. Meigetsuki: Kokubun meichō kankōkai, 1935.
Kokusho sōsho kangyōkai, ed. Gyokuyō. 3 vols. Tokyo: Meichō kangyōkai, 1998.
361
Kondō Heijō. Jūsandai yōryaku, Shiseki shūran. Tokyo: Kondō Heijō, 1883.
Kornicki, Peter. The Book in Japan: A Cultural History from the Beginnings to the
Nineteenth Century. Leiden: Brill, 1998.
Kosaka Shinji. Abe no Seimei sen Senji ryakketsu to Onmyōdō. Tokyo: Kyūko shoin,
2004.
Kuhn, Thomas S. The Structure of Scientific Revolutions. Chicago: University of Chicago
Press, 1970.
Kunaichō shoryōbu, ed. Kujō-ke rekisei kiroku. 4 vols, Zushoryō sōkan. Tokyo:
Kunaichō shoryōbu, 1989–1999.
Kuroda Hideo. Ō no shintai, Ō no shōzō. Tokyo: Heibonsha, 1993.
Kurokawa Harumura, ed. Rekidai zanketsu nikki. 35 vols. Kyoto: Rinsen shoten,
1989-1990.
Lam, Lay Yong and Tian Se Ang. Fleeting Footsteps: Tracing the Conception of
Arithmetic and Algebra in Ancient China. River City, NJ: World Scientific, 1992.
Loewe, Michael. "China." In Divination and Oracles, edited by Michael Loewe and
Carmen Blacker et al. London: Allen & Unwin, 1981.
———. "The Han View of Comets." Bulletin of the Museum of Far Eastern Antiquities
no. 52 (1980).
Marcon, Frederico. "The Names of Nature: The Development of Natural History in Japan,
1600-1900." Columbia University, 2007.
Martin, Dale B. Inventing Superstition: From the Hippocratics to the Christians.
Cambridge, Mass.: Harvard University Press, 2004.
Masamune Atsuo, ed. Jige kaden. 6 vols. Tokyo: Gendai Shichōsha, 1978.
Mass, Jeffrey. Yoritomo and the Founding of the First Bakufu: The Origins of Dual
Government in Japan. Stanford, Calif.: Stanford University Press, 1999.
Matsumoto Takuya. "Ritsuryō kokka ni okeru saii shisō: Sono seiji hihan no yōso no
bunseki." In Kodai ōken to saigi, edited by Mayuzumi Hiromichi. Tokyo:
Yoshikawa kōbunkan, 1990.
Matsuzono Hitoshi. Nikki no ie: Chūsei kokka no kiroku soshiki. Tokyo: Yoshikawa
kōbunkan, 1997.
362
———. Ōchō nikki ron. Tokyo: Hōsei daigaku shuppankyoku, 2006.
McCullough, Helen. Ōkagami, a Great Mirror: Fujiwara Michinaga (966-1027) and His
Times: A Study and Translation. Princeton: Princeton University Press, 1980.
Michio Maki. In no kinshin no kenkyū. Tokyo: Zoku Gunsho ruijū kanseikai, 2001.
Mikami, Yoshio. A History of Japanese Mathematics. Chicago: Open Court Pub. Co.,
1914.
Minakawa Kan'ichi. "Kiroku nenpyō, kiroku mokuroku." In Kokushi daijiten, edited by
Kokushi daijiten henshū iinkai. Tokyo: Yoshikawa kōbunkan, 1992–1997.
Minamoto Masayori. "Masayori ki." In Shōkōkan monjo. Tokyo: Tōkyō Daigaku Shiryō
hensanjo, 12th c.
Mitsuhashi Tadashi. Heian jidai no shūkyō to shinkō girei. Tokyo: Zoku gunsho ruijū
kanseikai, 2000.
Moerman, D. Max. "The Archeology of Anxiety: An Underground History of Heian
Religion." In Heian Japan, Centers and Peripheries, edited by Mikhael
Adolphson et al. Honolulu: University of Hawai'i Press, 2007.
Momo Hiroyuki. "Kidendō." In Kokushi daijiten, edited by Kokushi daijiten henshū
iinkai. Tokyo: Yoshikawa kōbunkan, 1992–1997.
———. Rekihō no kenkyū. 2 vols. Tokyo: Shibunkaku shuppan, 1990.
———. "Sakutan Tōji." In Kaitei yūshoku kojitsu jiten, edited by Katō Tejiro et al.
Tokyo: Kisho kankōkai, 1976.
———. "Sukuyōdō to sukuyō kanmon." In Onmyōdō sōsho, edited by Murayama
Shū'ichi. Tokyo: Meicho shuppan, 1991.
Mori Shinnosuke. "Sekkan Inseiki kizoku shakai ni okeru matsudaikan: Saii shisō ya
unmeiron to no kanren kara." Nihon shisōshi kenkyū no. 40 (2008): 18-39.
Morinaga, Maki Isaka. Secrecy in Japanese Arts: "Secret Transmission" As a Mode of
Knowledge. New York: Palgrave Macmillan, 2005.
Morris, Ivan, ed. The Pillow Book of Sei Shōnagon. New York: Columbia University
Press, 1991.
Multiple. "Chōkan sakutan tōji kanmon." In Saionji-ke kiroku. Tokyo: Tōkyō Daigaku
Shiryō hensanjo, 1925 (copy of 1163 ms.).
363
———. "Kahi yōroku." Tokyo: Tōkyō Daigaku shiryō hensanjo, 1958 (copy of older
manuscript).
Murayama Shū'ichi. Nihon Onmyōdō shi sōsetsu. Tokyo: Hanawa shobō, 1981.
———. Nihon Onmyōdō shiwa. Osaka: Osaka shoseki, 1987.
———, ed. Onmyōdō kiso shiryō shūsei. Tokyo: Tōkyō bijutsu, 1987.
Nakahara Moroyasu (attr.). "Shodō kanmon." In Gunsho ruijū, edited by Zoku gunsho
ruijū kanseikai. Tokyo: Yagi shoten, 1960.
Nakajima Wakako. "Genji monogatari no Dōkyō, Onmyōdō, Sukuyōdō " In Genji
monogatari kenkyū shūsei, edited by Suzuki Hideo and Ii Haruki Masuda Shigeo,
1998.
Nakamura Shōhachi. Nihon Onmyōdō sho no kenkyū: Zōhōban. Tokyo: Kyūko shoin,
2000.
Nakayama, Shigeru. A History of Japanese Astronomy: Chinese Background and
Western Impact. Cambridge, Mass.: Harvard University Press, 1969.
———. "Characteristics of Chinese Astrology." Isis 57, no. 4 (1966).
———. "The Position of the Futian Calendar on the History of East / West Intercourse of
Astronomy " In History of Oriental Astronomy. Proceedings of an International
Astronomical Union Colloquium, No. 91, New Delhi, India, November 13-16,
1985., edited by G. Swarup et al. Cambridge: Cambridge University Press, 1987.
NASA. "Nasa Eclipse Web Site." http://eclipse.gsfc.nasa.gov.
Needham, Joseph. Science and Civilisation in China. Vol. 3: Mathematics and the
Sciences of the Heavens and the Earth. Cambridge: Cambridge University Press,
1959.
Nihon Daizōkyō hensankai, ed. Nihon Daizōkyō. 49 vols. Tokyo: Nihon Daizōkyō
hensankai, 1914–1921.
Nishioka Yoshifumi. "Rikujin shikisen to konrō no miura." In Ōken to jingi, edited by
Imatani Akira. Kyoto: Shibunkaku shuppan, 2002.
Nishizawa Yūsō. Tonkō rekigaku sōron: Tonkō guchūrekijitsu shūsei. Tokyo: Nishizawa
Yūsō, 2004.
Noguchi Kyōko. "'Yoruka' no jidai: Mono-imi sanrō ni miru Heian kizoku bunka no
yoru." Kodai bunka 59, no. 1 (2007).
364
Okada Yoshirō. "Koyomi no shurui." In Koyomi o shiru jiten, edited by Okada Yoshirō et
al. Tokyo: Tōkyōdō shuppan, 2006.
———. Meiji kaireki: “Toki” no bunmei kaika,. Tokyo: Taishūkan shoten, 1994.
Ooms, Herman. Imperial Politics and Symbolics in Ancient Japan: The Tenmu Dynasty,
650-800. Honolulu: University of Hawai'i Press, 2009.
Orzech, Charles D. "The Trouble with Tantra in China: Reflections on Method and
History." In Transformations and Transfer of Tantra in Asia and Beyond, edited
by István Keul. New York: De Gruyter, 2012.
Osaki Shōji. Kinsei Nihon tenmon shiryō. Tokyo: Hara shobō, 1994.
Ōsone Shōsuke et al., ed. Honchō monzui, Shin Nihon koten bungaku taikei. Tokyo:
Iwanami shoten, 1992.
Pankenier, David W. "Early Chinese Astronomy and Cosmology: The "Mandate of
Heaven" As Epiphany." Stanford University, 1983.
Peek, Philip M., ed. African Divination Systems: Ways of Knowing. Bloomington:
Indiana University Press, 1991.
Piggott, Joan R. The Emergence of Japanese Kingship. Stanford, Calif.: Stanford
University Press, 1997.
Plutschow, Herbert. "Ideology and Historiography: The Case of Sugawara No Michizane
in the Nihongiryaku, Fusô Ryakki and Gukanshô." Paper presented at the
Historiography and Japanese Consciousness of Values and Norms 2001.
Queen, Sarah A. From Chronicle to Canon: The Hermeneutics of the Spring and Autumn,
According to Tung Chung-Shu. Cambridge: Cambridge University Press, 1996.
Reischauer, Edwin. Ennin's Travels in China. New York: Ronald Press Company, 1955.
Roberts, Luke Shepard. Performing the Great Peace: Political Space and Open Secrets
in Tokugawa Japan. Honolulu: University of Hawai'i Press, 2012.
Ron, Roy. "Powerful Warriors and Influential Clergy: Interaction and Conflict between
the Kamakura Bakufu and Religious Institutions." University of Hawai'i, 2003.
Ryūfuku Yoshitomo. "Fujiwara no Sukefusa." In Kokushi daijiten, edited by
Kokushi daijiten henshū iinkai. Tokyo: Yoshikawa kōbunkan, 1992–1997.
Saitō Kuniji. Kokushi kokubun ni arawareru hoshi no kiroku no kenshō. Tokyo:
Yūzankaku shuppan, 1986.
365
———. Kotenmongaku no michi. Tokyo: Hara shobō, 1990.
———. Teika "Meigetsuki" no tenmon kiroku: Kotenmongaku ni yoru kaishaku. Tokyo:
Keiyūsha, 1990.
Sansom, George Bailey. A History of Japan to 1334. Stanford, Calif.: Stanford University
Press, 1958.
Satō Sōjun et al., ed. "Chikanobu-kyō ki" no kenkyū. Kyoto: Shibunkaku shuppan, 2005.
Schaberg, David. A Patterned Past: Form and Thought in Early Chinese Historiography,
Harvard East Asian Monographs. Cambridge, Mass.: Harvard University Press,
2001.
Schafer, Edward H. Pacing the Void: T'ang Approaches to the Stars. Berkeley:
University of California Press, 1977.
Schneewind, Sarah. A Tale of Two Melons: Emperor and Subject in Ming China.
Indianapolis: Hackett Pub. Co., 2006.
Sen, Tansen. Buddhism, Diplomacy, and Trade: The Realignment of Sino-Indian
Relations, 600-1400. Honolulu: University of Hawai'i Press, 2003.
Shapin, Steven. A Social History of Truth: Civility and Science in Seventeenth-Century
England. Chicago: University of Chicago Press, 1994.
Shigeta Shin'ichi. Abe no Seimei: Onmyōji tachi no Heian jidai. Tokyo: Yoshikawa
kōbunkan, 2006.
———. Kaguyahime no kekkon: Nikki ga kataru Heian himegimi no endan jijō. Tokyo:
PHP Kenkyūjo, 2008.
———. Onmyōji to kizoku shakai. Tokyo: Yoshikawa kōbunkan, 2004.
Shintō taikei hensankai, ed. Engi shiki. 2 vols, Shintō taikei. Tokyo: Shintō taikei
hensankai, 1991–1993.
———, ed. Hokuzan-shō, Shintō taikei. Tokyo: Shintō taikei hensankai, 1992.
———, ed. Ichidai yōki. 3 vols, Shintō taikei. Tokyo: Shintō taikei hensankai,
2005–2006.
———, ed. Shinsen shōjiroku, Shintō taikei. Tokyo: Shintō taikei hensankai, 1981.
Shiryō taisei kankōkai, ed. Kanchūki. 3 vols, Shiryō taisei. Tokyo: Naigai Shoseki, 1936.
366
Shōgakkan kokugo jiten henshūbu, ed. Nihon kokugo daijiten. 14 vols. Tokyo:
Shogakkan, 2000–2002.
Sivin, Nathan. "Cosmos and Computation in Early Chinese Mathematical Astronomy."
T'oung pao 55, no. 1–3 (1969): 1–73.
———. Granting the Seasons: The Chinese Astronomical Reform of 1280, with a Study
of Its Many Dimensions and a Translation of Its Records: Sources
and Studies in the History of Mathematics and Physical Sciences. New York:
Springer, 2009.
Smith, Richard J. Chinese Almanacs. Oxford: Oxford University Press, 1992.
Smits, Ivo. "China as Classic Text: Chinese Books and Twelfth-Century Japanese
Collectors." In Tools of Culture: Japan's Cultural, Intellectual, Medical, and
Technological Contacts in East Asia, 1000-1500s, edited by Andrew Goble et al.
Ann Arbor, Mich.: Association for Asian Studies, 2009.
Sonkeikaku bunko, ed. Hokuzan shō. 3 vols, Sonkeikaku bunko zenpon eiin shūsei.
Tokyo: Yagi shoten, 1995–1996.
———, ed. Nichūreki. 3 vols, Sonkeikaku bunko zenpon eiin shūsei. Tokyo: Yagi shoten,
1997–1998.
Steele, John. Observations and Predictions of Eclipses by Early Astronomers,
Archimedes. Boston: Kluwer Academic Publishers, 2000.
Steenstrup, Carl. Hōjō Shigetoki, 1198-1261, and His Role in the History of Political and
Ethical Ideas in Japan, Scandinavian Institute of Asian Studies Monograph Series.
London: Curzon Press, 1979.
Stone, Jacqueline Ilyse. Original Enlightenment and the Transformation of Medieval
Japanese Buddhism. Honolulu, Hawaii: University of Hawai'i Press, 1999.
Strickmann, Michael. Chinese Magical Medicine Stanford, Calif.: Stanford University
Press, 2002.
Sugimoto Masayoshi and David L. Swain. Science and Culture in Traditional Japan, A.D.
600-1854. Cambridge, Mass.: MIT Press, 1978.
Sun, Xunchen and Jacob Kistemaker. The Chinese Sky During the Han: Constellating
Stars and Society. Leiden: Brill, 1987.
367
Suzuki Ikkei. "’On’yōdo’ and ‘Onmyōdō’: Reflections on the Lineage and System of the
‘Way of Yin and Yang’." In Onmyōdō: The Way of Yin and Yang. Columbia
Center for Japanese Religions, Columbia University, 2009.
Suzuki Rie. "Myōgyōdō." In Kokushi daijiten, edited by Kokushi daijiten henshū iinkai.
Tokyo: Yoshikawa kōbunkan, 1992–1997.
Takada Yoshihito. "Heian jidai kinō kanjin ni okeru kagyō no keishō: Tenmondō o
chūshin toshite." Kokugakuin zasshi 109, no. 11 (2008): 119-37.
Takeda, Sachiko. "Roads in the Tennō-Centered Polity." In Capital and Countryside in
Japan, 300-1180: Japanese Historians in English, edited by Joan Piggott. Ithaca,
NY: East Asia Program, Cornell University, 2006.
Takeuchi Rizō, ed. Heian ibun. 14 vols. Tokyo: Tōkyōdō shuppan, 1963–1976.
Tanaka, Steven. New Times in Modern Japan. Princeton, NJ: Princeton University Press,
2004.
Teeuwen, Mark. "Introduction: Japan's Culture of Secrecy from a Comparative
Perspective." In The Culture of Secrecy in Japanese Religion, edited by Bernhard
Scheid and Mark Teeuwen. New York: Routledge, 2006.
Teeuwen, Mark and Fabio Rambelli, ed. Buddhas and Kami in Japan: Honji Suijaku as a
Combinatory Paradigm. London: RoutledgeCurzon, 2003.
Thomann, Johannes. "Square Horoscope Diagrams in Middle Eastern Astrology and
Chinese Cosmological Diagrams: Were These Designs Transmitted through the
Silk Road?" In The Journey of Maps and Images on the Silk Road, edited by eds
Phillipe Forêt and Andreas Kaplony. Leiden: Brill, 2008.
Thorsett, S. E. "Identification of the Pulsar Psr1509–58 with the 'Guest Star' of Ad 185."
Nature no. 356 (1992): 690–91.
Tokoro Isao. Kyūtei gishikisho seiritsu shi no saikentō. Tokyo: Kokusho kankōkai, 2001.
———. "Miyoshi uji." In Kokushi daijiten, edited by Kokushi daijiten henshū iinkai.
Tokyo: Yoshikawa kōbunkan, 1979–1997.
Tōkyō Daigaku Shiryō hensanjō, ed. Chūyūki. 7 vols, Dai Nihon kokiroku. Tokyo:
Iwanami shoten, 1993 –.
———. Dai Nihon komonjo. 25 vols. Tokyo: Tōkyō Daigaku shuppankai, 1968–1970.
———. Dai Nihon shiryō. 12 vols. Tokyo: Tōkyō Daigaku Shuppankai, 1968–.
368
———, ed. Go-Nijō Moromichi ki, Dai Nihon kokiroku. Tokyo: Iwanami shoten,
1956–1958.
———, ed. Goshinjin'in kanpaku-ki. Edited by Tōkyō Daigaku Shiryō hensanjō. 4 vols,
Dai Nihon kokiroku. Tokyo: Iwanami shoten, 1999–2008.
———, ed. Midō kanpaku-ki. 3 vols, Dai Nihon kokiroku. Tokyo: Iwanami shoten,
1952–1954.
———, ed. Minkeiki. 6 vols, Dai Nihon kokiroku. Tokyo: Iwanami shoten, 1975–2007.
———, ed. Shōyūki. 11 vols, Dai Nihon kokiroku. Tokyo: Iwanami shoten, 1959.
———, ed. Teishinkōki, Dai Nihon kokiroku. Tokyo: Iwanami shōten, 1956.
Toyoda Masahiro. "Naishi-Sen." In Kokushi daijiten, edited by Kokushi daijiten henshū
iinkai. Tokyo: Yoshikawa kōbunkan, 1979–1997.
Tyler, Royall. The Tale of Genji. New York: Viking, 2001.
Uchida Masao. Nihon rekijitsu genten. Tokyo: Yūzankaku shuppan, 1992.
Umeda Chihiro. Kinsei Onmyōdō soshiki no kenkyū. Tokyo: Yoshikawa kōbunkan, 2009.
Unknown. "Daitō Onmyō sho." In Rokujizōji seikyō. Tokyo: Tōkyō Daigaku Shiryō
hensanjo, n.d.
———. "Ise goyomi." Ise, Mie Prefecture: Kōgakkan University Museum, 19-20th cen.
———. "Kakumei kanmon." In Gunsho ruijū, edited by Zoku gunsho ruijū kanseikai.
Tokyo: Yagi shoten, 1960.
———. "Kan'ei nenkan reki suisankōtō." In Yōmei bunko. Tokyo: Tōkyō Daigaku Shiryō
hensanjo, 17th c.
———. "Kiu hō nikki." Tokyo: Tōkyō Daigaku Shiryō hensanjo, n.d.
———. "Narisuke-Ō ki [Gūchūreki]." In Tanaka Yuzuru kyūzō tenseki komonjo Tokyo:
Tōkyō Daigaku Shiryō hensanjo, 1199.
———. "Osan buruiki." In Zoku gunsho ruijū, edited by Zoku gunsho ruijū kanseikai.
Tokyo: Yagi shoten, 1958.
———. "Shimazu-ke denrai gunjutsu sho." In Shimazu-ke monjo. Tokyo: Tōkyō
Daigaku Shiryō hensan-jo, n.d.
369
———. "Shin gishiki." In Gunsho ruijū, edited by Zoku gunsho ruijū kanseikai. Tokyo:
Yagi shoten, 1960.
———. "Sukuyō onunroku." In Rokujō Ariyasu-shi shozō monjo. Tokyo: Tōkyō Daigaku
Shiryō hensanjo, n.d.
———. "Sukuyō unmei kanmon." In Zoku gunsho ruijū, edited by Zoku gunsho ruijū
kanseikai. Tokyo: Yagi shoten, 1958.
———. "Tenmon yōroku." Tokyo: Tōkyō Daigaku Shiryō hensanjō, n.d.
Ury, Marian. "Chinese Learning and Intellectual Life." In Cambridge History of Japan,
edited by Donald Shivley and William McCullough. Cambridge: Cambridge
University Press, 1999.
von Verschuer, Charlotte. Across the Perilous Sea: Japanese Trade with China and
Korea from the Seventh to the Sixteenth Centuries. Edited by Cornell East Asia
Series. Ithaca, NY: East Asia Program, Cornell University, 2006.
———. Le Commerce Extérieur Du Japon Des Origines Au Xvie Siècle. Paris:
Maisonneuve & Larose, 1988.
———. "Life of Commoners in the Provinces: The Owari No Gebumi of 988 " In Heian
Japan, Centers and Peripheries, edited by Mikhael et al. Adolphson. Honolulu:
University of Hawai'i Press, 2007.
Wang, Aihe. Cosmology and Political Culture in Early China. Cambridge: Cambridge
University Press, 2000.
Watanabe Toshio. Nihon no koyomi. Tokyo: Yūzankaku, 1993.
Wilson, William Ritchie. Hōgen Monogatari; Tale of the Disorder in Hōgen, Monumenta
Nipponica Monographs. Tokyo: Sophia University Press, 1971.
Yabuuchi Kiyoshi. Chūgoku no tenmon rekihō. Tokyo: Heibonsha, 1969.
Yabuuti, Kiyosi (Yabuuchi Kiyoshi). "The Calendar Reforms in the Han Dynasties and
Ideas in Their Background." Archives internationales d'histoire des sciences no.
24 (1974): 51–65.
Yamashita Katsuaki. Personal conversation, October 11, 2010.
———. Heian jidai no shūkyō bunka to Onmyōdō. Tokyo: Iwata shoin, 1996.
———. "Onmyōdō no seiritsu to jukyō-teki rinen no suitai." Kodai bunka 59, no. 2
(2007): 177–96.
370
———. "Shohyō Hosoi Hiroshi cho "Kodai no tenmon ihen to shisho"." Nihonshi kenkyū
no. 555 (2008): 66-73.
Yano Michio. "Bukkyō kyōten no naka no koyomi, Sukuyō-kyō." In Onmyōdō sōsho,
edited by Murayama Shū'ichi. Tokyo: Meicho shuppan, 1991.
———. Hoshi uranai no bunka kōryūshi. Tokyo: Keisō shobō, 2004.
———. Mikkyō senseijutsu: Sukuyōdō to Indo senseijutsu. Tokyo: Tōkyō bijutsu, 1986.
Yōmei bunko, ed. Midō kanpaku ki. 5 vols, Yōmei sōsho. Kyoto: Shibunkaku shuppan,
1983–1984.
Yoshie Akio. Shinbutsu shūgō. Tokyo: Iwanami shoten, 1996.
Yuasa Yoshimi. Koyomi to tenmon no kodai chūsei shi. Tokyo: Yoshikawa kōbunkan,
2009.
Zito, Angela. Of Body & Brush: Grand Sacrifice as Text/Performance in
Eighteenth-Century China. Chicago: University of Chicago Press, 1997.
Zōhō shiryō taisei kankōkai, ed. Chūyūki. 7 vols, Zōhō shiryō taisei. Kyoto: Rinsen
shoten, 1965.
———, ed. Gonki. 2 vols, Zōhō shiryō taisei. Kyoto: Rinsen shoten, 1965.
———, ed. Heikoki. 2 vols, Zōhō shiryō taisei. Kyoto: Rinsen shoten, 1965.
———, ed. Hyōhanki. 5 vols, Zōhō shiryō taisei. Kyoto: Rinsen shoten, 1965.
———, ed. Kitsuki. 2 vols, Zōhō shiryō taisei. Kyoto: Rinsen shoten, 1965.
———, ed. Sakeiki, Zōhō shiryō taisei. Kyoto: Rinsen shoten, 1965.
———, ed. Sankaiki. 3 vols, Zōhō shiryō taisei. Kyoto: Rinsen shoten, 1965.
———, ed. Shunki, Zōhō shiryō taisei. Kyoto: Rinsen shoten, 1965.
———, ed. Taiki. 3 vols, Zōhō shiryō taisei. Kyoto: Rinsen shoten, 1965.
Zurcher, E. "‘Prince Moonlight:’ Messianism and Eschatology in Early Medieval
Chinese Buddhism." T’oung Pao 68, no. 1/3 (1982): 1–75.

Asset Metadata

Creator Buhrman, Kristina Mairi (author)

Core Title The stars and the state: astronomy, astrology, and the politics of natural knowledge in early medieval Japan

Contributor Electronically uploaded by the author (provenance)

School College of Letters, Arts and Sciences

Degree Doctor of Philosophy

Degree Program History

Publication Date 07/23/2012

Defense Date 03/22/2012

Publisher University of Southern California (original), University of Southern California. Libraries (digital)

Tag astrology,Astronomy,Buddhism,history of science,Japan,OAI-PMH Harvest,onmyōdō

Language English

Advisor Piggott, Joan R. (committee chair), Harkness, Deborah (committee member), Meeks, Lori R. (committee member)

Creator Email buhrman@usc.edu,kristina.buhrman@gmail.com

Permanent Link (DOI) https://doi.org/10.25549/usctheses-c3-62717

Unique identifier UC11290437

Identifier usctheses-c3-62717 (legacy record id)

Legacy Identifier etd-BuhrmanKri-969.pdf

Dmrecord 62717

Document Type Dissertation

Rights Buhrman, Kristina Mairi

Type texts

Source University of Southern California (contributing entity), University of Southern California Dissertations and Theses (collection)

Access Conditions The author retains rights to his/her dissertation, thesis or other graduate work according to U.S. copyright law. Electronic access is being provided by the USC Libraries in agreement with the a...

Repository Name University of Southern California Digital Library

Repository Location USC Digital Library, University of Southern California, University Park Campus MC 2810, 3434 South Grand Avenue, 2nd Floor, Los Angeles, California 90089-2810, USA

Abstract (if available)

Abstract This dissertation examines the social factors involved in the practices of observational astrology (Ch. tianwen, Jp. tenmon) and calendrical astronomy (Ch. lifa, Jp. rekiho) at the Japanese court. The production and monopolization of astrological and astronomical knowledge had, from the time of the Han Dynasty in China, been part of the state bureaucracy and one of the signs of legitimate rule. In the seventh century, Japan too had imported and implemented these state sciences of the Chinese-style imperium. ❧ However, by the twelfth century, while state control of astronomical knowledge continued to operate at a surface level, within the Japanese court bureaucracy dissent and debate reigned. A number of lineages and factions cooperated or competed over astronomical and astrological facts, which resulted in a situation where there was no unified ""truth"" about the stars accepted by the majority of elite members of the court. The political fragmentation and factionalism that characterized the early medieval Japanese state was also to be found in knowledge about the natural world circulating at court. ❧ The major reason for this fragmentation of knowledge was the diversity of the population that produced this same knowledge, a population that did not share either a common identity or definition of practice. Astrological and astronomical knowledge was no longer produced solely by the technical bureaucrats whose offices had been established in the eighth-century Chinese-style law codes (Jp. ritsuryo)--instead, these officials contested with other legitimate but non-official purveyors of natural knowledge: Buddhist monks and court scholars and mathematicians prominent among them. Furthermore, the statements of fact produced by all three of these factions were subject to critique and revision by members of the top echelon of the court bureaucracy, the elite nobility. Clearly there were no independent professional fields of astrology or astronomy in late classical or early medieval Japan. ❧ As a result, specialists of astrology and astronomy employed a number of strategies to ensure a receptive audience for their work, at least among some members of the court. Many entered into client-patron relationships with the top level of the nobility, wherein knowledge and technical skill were traded for economic and social rewards. Two groups in particular, the members of the Bureau of Onmyo (Jp. Onmyo-ryo) and Buddhist monks, cultivated an aura of supernatural power and ritual efficacy. While the primary goal of this strategy might not have been debates over the stars at court, the use of this capital is clearly documented in the historical sources. Therefore, the social history of debates about astrology and astronomy in the Heian (794-1192) court provides valuable insight into the rise and social perception of the onmyoji, a group of specialists in divination, exorcism, and apotropaic ritual who loom large in the Japanese cultural imagination. ❧ In examining the social history of astrology and astronomy in Japan late classical and early medieval periods--how debates first arose then came to shape the very practices of astrology and astronomy themselves--this dissertation also demonstrates the vitality and political importance of these fields in the eighth through thirteenth centuries. In contrast to previous scholarship on the history of science in pre-modern Japan, this dissertation shows that astrology and astronomy were hardly stagnant during this period. It becomes clear, therefore, that the pursuit of natural knowledge in Japan, while it did not develop along expected Western or Chinese trajectories, was still an active part of the intellectual world in pre-modern Japan. Pre-modern Japan's ""failure"" to follow either of these paths was not in fact stagnation or devolution, but a separate trajectory shaped by the political and social realities of the early medieval period.