Why Leave Nara?

: Kammu and the Transfer of the Capital

Author(s): Ronald P. Toby
Source: Monumenta Nipponica , Autumn, 1985, Vol. 40, No. 3 (Autumn, 1985), pp. 331-
347
Published by: Sophia University

Stable URL: https://www.jstor.org/stable/2384764

JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide
range of content in a trusted digital archive. We use information technology and tools to increase productivity and
facilitate new forms of scholarship. For more information about JSTOR, please contact [email protected].
Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at
https://about.jstor.org/terms

Sophia University is collaborating with JSTOR to digitize, preserve and extend access to
Monumenta Nipponica

This content downloaded from

158.195.4.7 on Mon, 20 Feb 2023 11:38:24 UTC
All use subject to https://about.jstor.org/terms
 Why Leave Nara?

Kammu and the Transfer of the Capital

by RONALD P. TOBY

IN the Fifth Month of Enryaku 3 (784) the recently enthroned Emperor

Kammu ido, r. 781-806, ordered the abandonment of the capital city of
Heijo (Nara), seat of Japan's imperial government for all but five years
since 710, in favor of a new capital to be constructed at Nagaoka, about twenty
miles to the northwest. Ten years later, before the Nagaoka capital could be
completed, but after a substantial amount of construction had already been
done, Kammu rejected the new site, too, and ordered the capital moved once
more. This new capital city, in a valley near the southern end of Lake Biwa,
was named Heian-kyo, 'Peaceful Capital', and remained the seat of Japanese
emperors until early 1869, when the Meiji emperor moved east to Edo, which
was renamed Tokyo (Eastern Capital).
Prior to the establishment of HeijO in 710, there had been no permanent
capital in Japan,' only a series of ever-grander 'palace complexes', courts that
moved with changes of reign. More than a dozen such transfers of the court
have been identified for the period from 593 to 710.2 HeijO is regarded as mark-
ing the maturation of the archaic Yamato state into the sinicized early imperial
state of ancient Japan.3 As the seat of the 'imperial state', the capital may have
housed as many as 10,000 government officials, and as the central cathedral of
a national Buddhism, the landscape was dotted with temples large and small.

THE AUTHOR is Associate Professor of history at this time, and it may be fallacious to
and East Asian studies, University of Illinois, speak of a 'Japan' at all until the eighth cen-
Urbana-Champaign. He wishes to thank the tury. Cornelius J. Kiley, 'State and Dynasty
two anonymous readers of the present article in Archaic Yamato', in JAS, 33:1 (1973), pp.
for their valuable suggestions. 25-49.
1 For simplicity I use the fiction that from 2 J. Edward Kidder, Early Buddhist Japan,
the time of Keitai Ad;, r. 507-531, the Yamato Praeger, New York, 1972, p. 60.
kings, to use the terminology of Cornelius J. 3 Kiley proposes this distinction in 'State
Kiley, were kings of 'Japan', although there and Dynasty'.
were probably still competing regional 'kings'

This content downloaded from

158.195.4.7 on Mon, 20 Feb 2023 11:38:24 UTC
All use subject to https://about.jstor.org/terms
 332 Monumenta Nipponica, 40:3

The best estimate of the population of the capital and its immediate surround-
ings in the late eighth century is 200,000 people.4
The abandonment of the Heijo capital at Nara has yet to be fully explained.
The most common explanation for the move is the desire of the imperial house
(Kammu) and the bureaucracy (nobility) to escape the baneful influence of the
Nara-centered Buddhist church, which had accrued so much economic and
political power 'that . .. the authority of the Imperial House was, or seemed to
be, endangered.'5 The function of Buddhism as 'protector of the state', some
scholars argue, inevitably drew the clergy into politics, and '. . . the monks' im-
mense political success more than anything else contributed to the declining
prestige of Buddhism and forced the Nara bureaucrats to transfer the capital
site.'6 Others also see this reason as the principal motivation to abandon Nara,
but adduce as well the rationale offered by Kammu himself in 788, that the
new capital district was 'convenient by water and by land',7 an important con-
sideration in both the prosecution of the frontier wars in the northeast, and
the pursuit of communications with the southwest and with the continent.
It may be satisfying to suppose that the growing power of the Buddhist in-
stitution, infringing on the imperial prerogative in the Dokyo L60 incident in
the 760s, motivated Kammu to abandon the old capital, and Ross Bender
leaves little doubt that the Dokyo affair represented serious conflicts between
Shinto legitimating principles and Buddhist principles.8 But the new capital
was not much less Buddhist oriented, although perhaps somewhat more world-
ly, and Kammu's own relations with younger clerics, especially Saicho ?,& 9
do not suggest strong hostility to Buddhism on the emperor's part. The revul-
sion against Dokyo, and the reaction against Nara Buddhist institutions may
be part of the reason for the move of the capital. But surely there are other
reasons to explain the uprooting of an urban complex approximating 200,000
souls in population-about 3Wo to 4Wo of the total population at the time -

4 Sawada Goichi jRM-2I-, Nara-cho Jidai and Practice (A Study of a Shinto-Buddhist

Minsei Keizai no Suteki Kenkya . Syncretic School in Contemporary Japan),
Agj~ettS~t, Fuzanbo, 1927, pp. 276-83. Buddhist Books International, Los Angeles,
See also Yazaki Takeo, Social Change and the 1982, p. 13.
City in Japan, From the Earliest Times 7 Shoku Nihongi [SN] Aft H ZVIdLd(Kokushi
through the Industrial Revolution, Japan Taikei #2), Yoshikawa Kobunkan, 1966, p.
Publications, Tokyo, 1968, p. 34, and Paul 531. See also Yazaki, p. 47.
Wheatley & Thomas See, From Court to 8 For a fascinating analysis, see Ross
Capital: A Tentative Interpretation of the Bender, 'The Hachiman Cult and the DOkyo
Japanese Urban Tradition, University of Incident', in MN 34:2 (1966), pp. 125-54.
Chicago Press, 1978, p. 141. 9 Sansom, p. 118.
5 G. B. Sansom, A History of Japan to 10 Sawada's estimate, p. 182, of the popula-
1334, Stanford U.P., 1958, p. 9. The same tion of 'free people' (ryomin Jt>K) for the
basic explanation is also offered in John W. entire country at mid-century at between 5.6
Hall, Japan From Prehistory to Modern and 6 million, is the most widely accepted. For
Times, Delacorte, New York, 1970, p. 61, and a brief explanation of his methodology, see
Conrad Totman, Japan Before Perry, Aoki Kazuo -bInt, Nara no Miyako iJ4ojD
University of California Press, 1981, p. 26. A, Chao Koronsha, 1965, pp. 12-15.
6 Minoru Kiyota, Gedatsukai: Its Theory

This content downloaded from

158.195.4.7 on Mon, 20 Feb 2023 11:38:24 UTC
All use subject to https://about.jstor.org/terms
 TOBY: Why Leave Nara? 333

and to justify the vast expense of replacing the hundreds of official buildings
and thousands of official domiciles that were to be abandoned. Even Yazaki
Takeo's restatement of Kammu's point, that communications with the rest of
the Japan and with continental Asia were superior from the Yodo River
watershed than from the Nara Plain, seems insufficient. Relations with the
continent and with western Japan had been more troubled in the late seventh
and the early eighth centuries than they were in Kammu's own day.
The establishment of a permanent royal city in the early eighth century,
however, was not a random occurrence, nor the result of royal whim. Rather,
the founding of such a city was the culmination of more than a century of
evolution in both the Yamato kingship and its governmental trappings, and in
the physical embodiment of government, that is, the buildings that housed
king and court. This gradual transition from 'court to capital' was a manifesta-
tion of transformations in the polity at large, in Japanese society, and in the
structure of the competing elements of the Yamato lineage group.1" If the foun-
ding of the 'first permanent capital' in 710 was an expression of such basic
forces in Japanese society, then perhaps there are equally basic elements of
social and political dynamics that may help to explain the more momentous
decision to abandon Nara after three-quarters of a century. If Nara was to be a
permanent capital, then why did it prove to be temporary? Or, to phrase the
question in different terms, was Heijo the 'first permanent capital', or merely
the penultimate 'court'?
These conclusions must be regarded as tentative, but however much the
founding of Heijo in 710 may have been based on the expectation that it was to
be a permanent capital, its abandonment in 784 seems to partake of the
dynamics of the movement from one court, Heijo, to another at Nagaoka, and
only thence, in 794, to what was to be a truly permanent capital.
The conduct of the quintessential Nara-period monarch, Shomu Tf , r. 724-
748, reinforces the vision of Heijo as an elaborate court rather than a settled
capital. Shomu often absented himself from Heijo, and spent the years from
740 to 745 elsewhere, while a debate raged among the nobility as to where the
'court' and 'capital' should be located. The emerging Yamato imperium was
beset by frequent rebellions in the 730s; in 737 a smallpox epidemic wracked
the home provinces, taking the lives of many leading nobles, including the
heads of all four branches of the Fujiwara clan. 12 Shortly thereafter, in the fall
of 740, Fujiwara no Hirotsugu -W gMP, the Assistant Vice-Governor of
Dazaifu (Dazai Shoni J rose in rebellion in northern Kyushu, forcing

11 Wheatly & See present one view of this

Fujiwara, Heijo, Kuni-no-miya, Naniwa,
transition. and Nagaoka proved transitory 'courts' or im-
It may be more accurate to take the view permanent 'capitals' because of the unresolv-
that the intention to have a 'permanent' ed social and political tensions discussed
capital emerged by the mid- to late-seventh cen- below.
tury, as evidenced by the archeologoical re- 12 SN, 12, pp. 145-46. This is the first record-
mains of city street plans in the Fujiwara ed smallpox epidemic in Japanese history.
capital complex (Kidder, pp. 44 & 66), but that

This content downloaded from

158.195.4.7 on Mon, 20 Feb 2023 11:38:24 UTC
All use subject to https://about.jstor.org/terms
 334 Monumenta Nipponica, 40:3

the court to mobilize 17,000 troops from the provinces of the Tokai, Tosan,
San'yo, San'in, and Nankai regions.13 In the midst of this civil unrest, Shomu
ordered the court moved from Heijo because he 'wished to change the hearts'
of the people.14 On Tempyo 12(740).11.30, after notifying his general in the
field that he was leaving Heijo and telling him 'not to be alarmed', Shomu left
the city.15 By month's end the monarch had arrived at a 'temporary court'
(tongu A1I9) at a place called Seki-no-miya in Iga province. On Tempyo
12.12.6, Shomu commanded his Minister of the Right, Tachibana no Moroe 4A
:5 to start construction on a promising site that the minister had found in the
Sawara district of Yamashiro province, 'because he planned to move the
capital (sento AtF) there.'16 By the 15th, the monarch had already procee
to the site, Kuni-no-miya, and 'there began to construct a capital.'17
On New Year's Day of Tempyo 13 (741), 'The Emperor (tenno 3kr) held
court in Kuni-no-miya for the first time; the palace compound was not yet com-
pleted, so they surrounded it with curtains.'18 The location of the capital
remained an unresolved issue, and the court spent much of the following four
years roaming the home provinces in search of a seat, not deciding until 745 to
return to Heijo, and then only after much debate.
It is noteworthy that among Shomu's acts in this five-year sojourn in the
wilderness are some of the most important of the entire Nara period-the
decree establishing a network of provincial temples and nunneries was issued
from Kuni-no-miya, as was the decree legalizing the permanent private posses-
sion of newly opened rice fields, the legal basis for shoen -T landholdings.19
The location of the capital continued to be an active issue and remained in
doubt through much of 744. On the 15th day of the First Month the monarch
left Kuni-no-miya for Naniwa-no-miya,20 at the head of Osaka Bay, and in a
poll of the nobility taken on the 1st day of the intercalary First Month a slim
majority favored Naniwa over Kuni as the site for the capital. A poll of people
in the marketplace (ichi 4i) showed that 'all wanted to make Kuni-no-miya the
capital, but there was one who desired Naniwa, and one who desired Heijo.'21
It was only after a series of earthquakes in mid-745 that the issue was decided
in favor of 'Heijo-no-miya'('the Heijo court'), whither Shomu returned on
Tempyo 17(745).5. 11. The imperial arsenal caught up with the court only at
the year's end, moving from Kuni-no-miya to Heijo on the 15th day of the
Twelfth Month.22
That Shomu's departure from Nara and his vacillation over where to locate
were not merely brief imperial progresses to commodious vacation spots or
other temporary quarters may be seen by comparison with the much briefer

13 SN, 13, pp. 158-59. 8 SN, 14, p. 163.

14 SN, 13, p. 159. 19 SN, 14, pp. 163-64 & 174.
15 SN, 13, p. 160. 20 SN, 15, p. 176.
16 SN, 13, p. 162. 21 SN, 15, p. 176.
17 Hajimete kyoto o tsukuru , SN, 22 SN, 16, pp. 182-83 & 185.
13, p. 162.

This content downloaded from

158.195.4.7 on Mon, 20 Feb 2023 11:38:24 UTC
All use subject to https://about.jstor.org/terms
 TOBY: Why Leave Nara? 335

visits he had made earlier in his reign to Kii prov

Naniwa-no-miya in the spring of 734,24 or to Yo
mer of 736.25 These earlier departures were de
Hi, and also in the captions to Man'yoshu 7
visits, as 'imperial progresses' (gyoko 4W)', while the moves of the 740s were
termed 'moving the capital (sento)', and were attended by the appointment of
advance scouts for site selection, commissioners of capital construction, and
the like.
The progresses of the 720s and 730s, moreover, were the occasion of no
debate, at least as recorded in the Shoku Nihongi, and the poets of the
Man'yoshu celebrated them as joyful events, as glorious, legitimate elements
in imperial behavior. Of the progress to Yoshino in 736, Yamabe no Akahito
mLyj,2i& wrote:26

Our Sovereigns from the age of the gods

Have often come to hold court
Here at the Palace of Yoshinu-
Because of the beauty of the mountains and river.

Note the contrast between Akahito's expectation of a prompt return to Nara

and his ringing endorsement of Yoshino, and Tanabe no Sakimaro's fflAj
belief, expressed in his 'Lament for the Decay of the Old City of Nara', in the
740s that Shomu had left Nara forever:27

... This Imperial City of Nara was here founded,

That hence the heirs born to the Throne
Might rule the under-heaven in endless succession
Down through the ages.

The courtiers of eighty clans

Having built their mansions in rows,
I thought its Great Palace would flourish
As long as heaven and earth endured-
O Nara, city of my abiding trust!
But because the times are new,
All have gone-led by their Sovereign....

Envoys

Now that with the change of the times

Nara is become

23 SN, 9, p. 102. 25 SN, 12, p. 140; MYS, 6, #1005, in NKBT, 5,

24 SN, 11, p. 133; Man'yoshI [Mys] Aft,
p. 169.
6, #997-1002, in NKBT 5, Iwanami, 1957-1962,
26 NGS, p. 196; MYS, 6, #1005, in NKBT, 5, p.
p. 167. See also the Nihon Gakujutsu 168.
Shinkokai [NGS] translation of Man'yosha, 27 NGS, pp. 228-29; MYS, 6, #1047-49, in
Columbia U.P., 1965, p. 195, #1001. NKBT, 5, pp. 188-90.

This content downloaded from

158.195.4.7 on Mon, 20 Feb 2023 11:38:24 UTC
All use subject to https://about.jstor.org/terms
 336 Monumenta Nipponica, 40:3

An Imperial City that was,

The grass grows rank in the streets.

Since Nara, the Imperial City,

So long familiar to me,
Is now falling to decay,
Whenever I go outdoors,
Bitterer grows my grief.

Yet Sakimaro was sufficiently in tune with Shomu's whims that, for all his
grief at the demise of Nara, he could with equal emotion celebrate the new
'capital'28 at Kuni-no-miya, for 'Many are the provinces / And unnumbered
the dwelling-places of men,' where 'Well did our sovereign . .. / . . . establish
after his august desire / The great imperial abode,' which '. . . shall ever re-
main / The imperial abode for a hundred ages.'29 Indeed, Sakimaro proclaim-
ed, the Futagi Palace 'shall remain the great imperial abode / Immutable for
all ages,' and only 'Should the waters of the Izumi River / Ever cease to flow, /
Then the great imperial abode- / Then only-might suffer change.'30 When
Shomu pulled up stakes once more and tried Naniwa as his capital, it suddenly
became 'the Ajifu Palace [at Naniwa], on which I [Sakimaro] never tire to
look. 31
Thus even Nara in its glory was susceptible to abandonment and decay.
Within a span of less than five years ShOmu had 'established' two new
'capitals' that were to be the 'great Imperial abode immutable for all ages',32
in provinces outside Yamato, in violation of Kakinomoto no Hitomaro's X-I4;
AFM injunction that it was 'in Yamato that spreads to the sky' that emperors
'ruled the realm under heaven, each following each like generations of the
spruce.'33 Hitomaro had not been able to comprehend why Tenji XKl, r. 662-
671, would 'abandon Yamato, and cross the fair blue hills of Nara'34 for
another province, just as Shomu was doing now, nearly a century later.
What is at issue here, however, is not the reasons for Shomu's five-year
search for a new capital, a search that ended up where it started, in HeijO.
Rather, two points are fundamental. First, HeijO was somewhat less an 'im-
mutable' capital than we are accustomed to think, but still retained some
characteristics of a court site that had to compete with other candidates in the
capital district. A second, corollary point is that the emperor and the court

28 Kuni no aratashiki miyako o homuru#1055,

rfi in NKBT, 5, p. 190.
33 Ian Hideo Levy, tr., Ten Thousand
29 NGS, p. 230; MYS, 6, #1050-52, in NKBT,
Leaves: A Translation of the Man'yoshk,
5, pp. 188-90. Princeton U.P., 1981, 1, p. 53; MYS, 1, #29, in
30 NGS, p. 231; MYS, 6, #1053-54, in NKBT,
NKBT, 4, pp. 26-27; NGS, p. 27.
5, p. 190. 34 Yamato o okite / aoniyoshi / Narayama
31 MYS, 6, #1062, in NKBT, 5, p. 192. o koe f y-tIrsJ24 . MYS, 1, #29, in
32 ... momoyo ni mo / kawaru mashijiki /
NKBT, 4, pp. 26-27.
ohomiyadokoro Cik fJ'MT' )-iW!. MYS, 6,

This content downloaded from

158.195.4.7 on Mon, 20 Feb 2023 11:38:24 UTC
All use subject to https://about.jstor.org/terms
 TOBY: Why Leave Nara? 337

were not confined to a single capital city, as they would become in the Heian
period. The strongholds of provincially based nobles could compete with Nara
in attempting to attract the court thither because these nobles had power bases
outside the 'capital', or rather, the court where the emperor happened to
reside.
Similarly, it seems likely that the move in 784-794 to abandon the Nara Plain
as a capital district was based on constants in royal and national politics at the
time-perhaps the same sorts of issues that had led to the establishment of Hei-
jo and the designation of the Nara Plain as capital district seventy-five years
earlier, or to Shomu's repeated 'transfers of the capital' in the 740s. Conflicts
within the Yamato royal group, among the various royal lineages, and within
the nobility who were their allies, were among the major determinants of
changes in royal venue prior to the founding of Heijo, as will be seen below.
It is noteworthy that the establishment of the permanent Heian-kyo at the
end of the eighth century correlates closely with the resolution of the last major
conflict until the thirteenth and fourteenth centuries35 between competing lines
within the royal lineage group over the scarcest resource for which its members
could compete: the unique 'high office' of the Yamato kingship, which had
evolved into the office of the 'emperor' of 'Nihon'.36
Prior to the foundation of the capital at Nara as the first 'permanent' capital
city, the 'capital' had generally been located at the court of the reigning king of
Yamato. This propensity to pick up and move the 'capital' has often been ex-
plained by reference to Shinto conceptions of the polluting quality of death,
and the especially polluting effects associated with the death of a king. The
death of a king polluted his residence, it is argued, and required the transfer of
the court and capital to a new, ritually pure location.37 But this view appears to
accept the simplistic historiographical tradition of a unitary 'imperial' line
descended from the gods and surviving to the present day. Native ideas of
pollution may have played a role in the compulsion to move the court/capital,
but this is an insufficient explanation, since there is not a one-to-one cor-
respondence of 'imperial' reigns and seats of the court. If we are to explain
these early changes of royal venue, it is now clear, from the work of Cornelius
J. Kiley in particular, that we must attend more closely to the nature of the
'royal' class, its social structure, its marriage practices, and its child-rearing
customs.
First, as Kiley shows, the 'old dynasty' (pre-Keitai Wf*, r. 507-531) archaic

35 On the dynastic schism of the fourteenth 'at the top' of a social organization, which is
century, see H. Paul Varley, Imperial Restora- 'unique and non-duplicating', such that 'the
tion in Medievel Japan, Columbia U.P., 1971. scarcity of the resource [represented by the
36 I am using the term 'high office' here in office] magnifies the conflict situation that cen-
the sense defined by Jack Goody, in the In- tres around the transfer of the office from one
troduction to Succession to High Office (Cam- individual to another.'
bridge Papers in Social Anthropology, 4), 37 Sansom, p. 82.
Cambridge U.P., 1966, p. 2. That is, an office

This content downloaded from

158.195.4.7 on Mon, 20 Feb 2023 11:38:24 UTC
All use subject to https://about.jstor.org/terms
338 Monumenta Nipponica, 40:3

Yamato polity was characterized by a corporate class of potential royals,

lineages who were charismatically or genealogically qualified to compete for
the throne. In what Kiley calls a 'multilineal dynastic regime', a number of
mutually endogamous royal lineages were competent to produce either kings
or queens, and 'kingship and queenship circulated among several family
groups.'38 Under such circumstances it is not surprising that the venue of the
'court/capital' shifted to the center of power of the lineage that happened to
control the kingship at the time.
With the accession of Keitai in 507, there seems to have been a shift in the
nature of the royal group, and in its rules for identifying potential successors
to kingship (Keitai's was probably a new dynasty). By the mid-sixth century
there emerged a single patrilineal group, all descended from Keitai, who were
uniquely qualified to provide royal successors. The appearance of such a
lineage, however, had as its correlate the elimination of formerly competing
royal lines, or their relegation to a second tier in the royal/noble corporate enti-
ty, where they were judged competent to produce only queens, but not kings.
A third tier, the regional or territorially based nobility, was considered compe-
tent to produce mothers of kings, but not to produce queens, nor to compete
for the throne itself.
This change in organization minimized the potential for conflict among
distinct royal lineages over the throne, it is true, but it gave rise to a conflict
among royal sons and grandsons, and their marriage relatives, for there still
remained a plurality of eligibles, and there were no fixed rules of priority in suc-
cession. Thus, as Kiley observes, 'All rulers following Keitai were his
patrilineal descendants, but the royal group itself retained its multilineal
character,'39 for the royal clan could not afford to narrow the pool of eligibles
so far as to risk the absence of a qualified successor. Kiley posits a court govern-
ed by 'segmentary factions' drawn from the hierarchically differentiated sub-
groups of the royal and noble status group.
Since royal and noble males were polygamous, and their marriages seem to
have been predominantly duolocal or uxorilocal in character, royal heirs were
raised in the homes of their maternal relatives, and were therefore surrounded
by factions that included a large component of consort-clan members.40 The

38 Kiley, p. 38. social and chronological range.

39 Kiley, p. 46. Dana Morris, 'Peasant Family in Eighth
40 William McCullough, 'Japanese Mar- Century Japan', unpublished presentation at
riage Institutions in the Heian Period', in the Colloquium of the Center of Japanese
HJAS, 27 (1967), pp. 103-67, discusses mar- Studies, Berkeley (March 1980), finds evidence
riage practices among the Heian nobility, of duolocal marriage and matrilocal child-rear-
while Wakita Haruko, 'Marriage and Proper- ing among the eighth-century peasantry,
ty in Premodern Japan from the Perspective although in most cases this was only a stage in
of Women's History', in JJs, 10:1 (1984), pp. the family cycle, later shifting to virilocal mar-
riage.
73-99, addresses the subject across a wider

This content downloaded from

158.195.4.7 on Mon, 20 Feb 2023 11:38:24 UTC
All use subject to https://about.jstor.org/terms
 TOBY: Why Leave Nara? 339

Jomei (34) 1
r. 629-641 Takara Kotoku (36)
Kogyoku (35) r. 642-645 r. 645-655

Saimei (37) r. 655-661 n

Arima no Oji

Soga no Ochihime Naka no Oe 1 Oama no Oji

O Tenji (38) 9 ? Temmu (40)
r. 661-671 1 r. 672-686
I Shiki no Oji Ki no Tokihime
Otomno no Oji d. 716

Jito (41) r.Koun719)672

r. 686-697,
Toneri Shinno
Kusakabe no Oji Abe no Ojo
CZ 4 q q Gemmyo (43)
4 ' r. 707-715

CZ E ; Karu no Oji
CZ - Mommiu (42)
= cZ Hidaka no Ojo r. 697-707
>'> Gensh6 (44)
E~ 1- ell (4 Obito nlo Oji Fujiwara

rr 724-749

;3 0 r. 715-724 Sh~S n Shuto45

Abe no Ojo (4)AnL 4) hoain
Koken (46).
r.749-758 Oi no O
Khonokn (49) KawatsuJunnin (47) oya
764-770 758-764 (rebelled, 764)
I ) | ~~~~~~~~~~~~~(deposed)

? , r + + ~~~~~~~~~~~~~~Shir akabe )no0 ?Takano no Niikas Hikamni no

ce o Osabe Shinno
(deposed as crown Fujiwara no Ryoshi
Yamanobe no O (Momokawa's daughter)
Fujiwara no Omnuro Kammiu (50) ALi(;
(Tanetsugu's daughter)l r. 781-806

z () Heiei(51)Fujiwara no Kusuko
Heizei (51) * O Tanetsugu's Saa(2Saga (52) Tachibana no ,Junna (53)
daughter
X T r. 806-809 T (rebelled, 810) r. 809-823 Kachishi r. 823-833

-0 Takaoka ShinnoI
> > (deposed as crown Aho Shinno Fujiwara
S = prince, 810) no Junshi
CZ ce () i iNimmyo (54) Fujiwara no
Q~ r. 833-850 Takushi
2 Ariwara nor Ariwa
. Narihira Yukihira Mo Hanshi no Ojo
= ~~~~~~~~~~~Montoku (55) K6kO (58) HasinOo
gO O 1? ~~~~~~~~~~ ~~r. 850-858 T r. 884-887 1

:, v ~~~~~~~~~~Seiwa (56) 1 riUda (59)

Q r. 859-876 T r. 887-897

Yozei (57)
r. 876-884 All succeeding emperors

A male 0 female AO reigning monarch

Figures in parentheses indicate the traditional

numbering of monarchs.

Royal Genealogy, Mid-Seventh to Ninth Centuries

This content downloaded from

158.195.4.7 on Mon, 20 Feb 2023 11:38:24 UTC
All use subject to https://about.jstor.org/terms
 340 Monumenta Nipponica, 40:3

death of a king often gave rise to a succession struggle, but when the successor
finally emerged, the venue of the court moved as much to be where he resided,
as to avoid the pollution attendant upon the death of his predecessor. Other-
wise if pollution were the principal issue, it would be difficult to explain the
occasional re-use of former royal palace sites.41
Such lineage conflicts over the succession-conflicts, in Kiley's terms, be-
tween segmentary factions centered on rival candidate royals-occurred often
in the first two centuries of the Keitai dynasty, sometimes involving bloodshed,
as in the 580s, the 640s, and the 670s. With inadequate rules of exclusion to
eliminate an excessive number of candidates for the throne, succession
disputes were virtually inevitable. Increasingly strategies were designed to
assure patrilineal succession, and to minimize the potential for succession
disputes. Two strategies were particularly important prior to the late eighth
century. The exclusion from candidacy of all but the eldest son, designated the
oe AR, of any heir-producing consort constituted a form of dynastic shedding
that simultaneously reduced the number of candidates to a smaller, more
manageable, and less volatile pool of eligibles, and enhanced the solidarity of
each group of uterine siblings by eliminating rivalry between uterine
brothers.42 In a parallel strategy, a stake-holder reign by a female monarch
assured that her successor would be an heir who was the lineal descendant of
her late husband and, of course, herself as well.43
The last major fraternal conflict in the late Archaic period was the Jinshin
Disturbance (Jinshin no Ran fE1IQ)L) in 672, when Prince Oama Xt-1RA
wrested the succession from the line of his brother, Prince Naka F4+ (Tenji, r.
661-671, leader of the Taika coup in 645). Oama unseated his nephew,
Emperor Kobun L r. 671-672, took the throne himself, and established his
court at Kiyomihara in the following year, reigning as Temmu. When Temmu
died in 686, his widow and niece, Sarara m A, succeeded him as stakeholder for
their son, Prince Kusakabe A But the prince died in 689 without ascending
the throne, and Sarara, as Emperor Jito t~JL,44 reigned until Kusakabe's son,
41 In the endleaf of Kodai Kokka no Seirit- Chapter 3. Hurst's view that female monarchs
su -tMM~o)0Ua, Chfio Koronsha, 1965, were primarily stakeholders whose main func-
Naoki Kojiro 6 lists the monarchs tion was to guarantee the identity of their suc-
from Kimmei to Mommu, showing, for exam- cessors has been challenged; see, for example,
ple, that Jomei's By first palace site, Asuka Patricia Tsurumi, 'The Male Present versus
Okamoto, was used twenty-six years later by the Female Past: Historians and Japan's An-
Saimei AjM, and again sixteen years after that cient Female Emperors', in Bulletin of Con-
by Temmu, while Saimei also used Kogyoku's cerned Asian Scholars, 14:4 (1982), pp. 71-75.
By Asuka no Itabuki site in 655 after a lapse 44 It is worth noting, in the context of the
of twelve years. debate over the nature of female occupation
42 Kiley, p. 47. of the Yamato kingship, that the name Jito
43 On the role of female monarchs of this can be glossed 'maintaining the lineage' or
period, see G. Cameron Hurst, Insei: Ab- 'maintaining the succession'. It is probably
dicated Sovereigns in the Politics of Late also true that this posthumous name was given
Heian Japan, 1086-1185, Columbia U.P., by male officials.

This content downloaded from

158.195.4.7 on Mon, 20 Feb 2023 11:38:24 UTC
All use subject to https://about.jstor.org/terms
 TOBY: Why Leave Nara? 341

Prince Karu -finl, could succeed to the throne in

r. 697-707.
For seventy-five years after the death of Temmu, all male successors to the
throne were patrilineally descended from him. The only monarchs who were
not descended from Temmu, but rather from Tenji, on the male line, were two
female monarchs, Temmu's widow Jito and his daughter-in-law Gemmei vui-M,
r. 707-715, both of whom, as daughters of Tenji, were Temmu's nieces. But
more importantly, both women were also mothers of male candidates who
were patrilineal descendants of Temmu. Jito's interests were best served by
holding the throne for her son by Temmu, Prince Kusakabe. When that
hapless young man died in 689, her grandson Prince Karu, 683-707, became
the candidate of choice, and he succeeded Jito as Mommu. Gemmei, too, reign-
ed principally as a stakeholder, abdicating in 715 in favor of her daughter
by Kusakabe, GenshO id, r. 715-724, who reigned until Mommu's son was
old enough to ascend the throne as Shomu, r. 724-749. The establishment of
the Heijo 'capital' at Nara, then, was accomplished under the reigns of the
Temmu lineage and its consorts, who monopolized the imperial office until
770.
In Nara, it was Temmu's great-grandson Obito -6- (Emperor ShOmu, r. 724-
749) who built the edifice of state-sponsored Buddhism, centered on the cult of
Vairocana and of the emperor, in Todaiji. Before him, it was Temmu himself
who was responsible for implementing the centralized, bureaucratic Chinese-
style state envisioned in the Taika Reform program of 645-646, despite a
quarter-century of control of the throne by his older brother Tenji, master-
mind of the coup, either through stakeholders (Kotoku A r. 645-655; Saimei
AM, r. 655-661) or directly (661-67 1).
Temmu's patrilineal descendants did not fare well, however, beyond the mid-
dle of the eighth century. Shomu's only son had died young,45 leaving only his
daughters Abe FrE and Inoue t+. The devoutly Buddhist ShOmu, who had
declared himself 'the servant of the Three Treasures' of Buddhism, however,
found himself trapped by conflicting desires. On the one hand he wished to ab-
dicate and devote himself to the pursuit of religious progress by taking
religious vows, but on the other he also wanted to preserve the monarchy in his
own line, and his only son had died in 728, less than a year old. Rather than
yield the throne to his cousin, Prince Oi Jki0, he chose to 'violate the "im-
mutable law" of father to son succession,'46 and passed the throne to his
daughter Abe, who succeeded him as Emperor Koken -' , r. 749-758. Koken
was induced to abdicate in 758 in favor of her distant cousin Oi, a patrilineal
grandson of Temmu (Junnin rAh, r. 758-764, d. 765), but he was deposed

45 Inoue Mitsusada t-A, Nihon Kodai Iwanami, 1965, p. 235.

46 Hurst, p. 44.
Kokka no Kenkyai H- ?- N M X X fi ,

This content downloaded from

158.195.4.7 on Mon, 20 Feb 2023 11:38:24 UTC
All use subject to https://about.jstor.org/terms
 342 Monumenta Nipponica, 40:3

when his principal supporter, Fujiwara no Nakamaro P rose in

rebellion against the continued control of the affairs of state by the retired
Koken and her intimate adviser, the Buddhist cleric DOkyo.47 Koken then
assumed the throne once more, taking the name Shotoku I'm for this, her se-
cond reign as emperor, 764-770, and for the next seven years left the direction
of state affairs largely to Dokyo.
Shotoku's death without issue left DOkyo discredited and without protec-
tors, and he was banished to a rural temple in Shimotsuke province. Despite
the survival of several patrilineal male descendants of Temmu, however,
Shotoku had been compelled to leave a testament naming her distant cousin,
the 62-year-old Prince Shirakabe Aim-not patrilineally descended from Tem-
mu, but from Tenji-as crown prince and successor. When he succeeded as
Emperor Konin ft, r. 770-781, the Temmu lineage lost control of the throne
permanently. Like himself, all seventy-five of Konin's successors-eighty, if
we include the five emperors of the Northern Court, 1331-1392, who are not
officially recognized by modern historiography-are descended from Tenji,
whose line was in eclipse through the glory of the Nara period dominated by
the Temmu line.
There was a flurry of activity in the Temmu line once more when Konin
named as his successor Prince Osabe fSpiFi, his young son by a daughter of
Shomu, but when Osabe's mother was accused of plotting against his father,
the crown prince was deposed, and mother and son were sent to prison, where
they both died a few years later.
With the Temmu line out of the way, Konin's supporters at last turned
somewhat reluctantly to the aged emperor's eldest son, Prince Yamabe LL1%,
who, unlike Osabe, was a mature man of thirty-five at the time. Also unlike
Osabe, Yamabe had neither patrilineal nor matrilineal connections with the
Temmu line, for his mother was from a family descended from Paekche -a
immigrants. His mother's humble, alien antecedents apparently caused
resistance to Yamabe's selection, but he had supporters among the Fujiwara
who had placed K6nin on the throne and his selection as crown prince stood.
When Konin abdicated in 781, Prince Yamabe succeeded as Emperor Kammu,
to reign for twenty-five years. This sire of all succeeding Japanese emperors
would bring more changes to the ancient imperial court, however, than merely
the reversion to the Tenji lineage and the total eclipse of the Temmu lineage
that had been ascendant for a century.
In particular, it was Kammu who, in 784, scarcely two years after the death
of his father, moved the capital once more-actually twice-away from the

47 It has been suggested, although without speaks of DOkyo as the emperor's 'reputed
clear proof, that the emperor and her priestly paramour'; Aoki, p. 478, doubts the charge,
adviser were romantically involved as well, while Ueda Masaaki fHI M, Nihon no Jotei
but this seems like the sort of charge that Con- H V&k%-, KOdansha, 1973, pp. 197-98, re-
fucian historiographers might naturally level jects it. See also Bender's discussion, p. 139.
at the last member of a dynasty. Hurst, p. 45,

This content downloaded from

158.195.4.7 on Mon, 20 Feb 2023 11:38:24 UTC
All use subject to https://about.jstor.org/terms
 TOBY: Why Leave Nara? 343

power base of the Temmu line in Yamato to the center of Yamashiro province,
a stronghold of the Paekche immigrants, where his own maternal relatives
were entrenched. Heian-kyo and its early society were not much less Bud-
dhist than Nara and its society, and in Yamashiro, he could conduct the affairs
of state without interference from remnants of the Temmu line, the same peo-
ple who had put forth Prince Osabe a decade earlier.
The foregoing suggests that the Nara 'capital' city of Heijo partakes heavily
of the character of a 'court' for the Temmu dynasty. The Tenji line maintained
a strong identification with Yamashiro province even through a century of con-
centration of political power wielded by the Temmu line in the Yamato basin,
and it also possessed a sufficiently strong power base in Yamashiro to effect the
abandonment of the Temmu line's seat of power.48
One of the important foci for the separate geographic identification of the
two lines was, not surprisingly, the ancestral tombs. The Asuka/Nara area was
the burial area of the Temmu lineage. The Fujiwara palace constructed by
Temmu's widow Jito has even been described as a mandala emanating from
Temmu's burial mound, in which the axial north-south street of the palace
complex described a direct line from the palace itself to the tomb.49 The
Heijo capital looked south a short distance across the Nara Plain upon this
same mandala.
Heian-kyo, by contrast, is not only far from HeijO, but is also conveniently
near Tenji's old palace site at Otsu, and even closer to the site of Tenji's tomb,
the burial mound in Yamashina, just across Higashiyama from modern-day
Kyoto. When, after escaping the grasp of the Temmu line in the Nara Plain,
Kammu finally settled on the site of Heian-kyo, he was returning to the focus
of the Tenji line's ritual base, as well as being near the power base of his mater-
nal relatives.
That the issue between the two capital districts remained in flux is suggested
by the attempt of the eldest of Kammu's sons, Heizei +I, r. 806-809, after his
abdication to retake the throne in league with Fujiwara no Kusuko *An,

48 There is an emerging consensus in Japan

Sasayama Haruo MALUH , in 'Heian Shoki
that the move from Nara and the ultimate no Seiji Kaikaku' Act M X I v P in
choice of Heian-kyo hinged on the resolution Iwanami Koza Nihon Rekishi e- it A kH EMz
of the latent lineage conflict between descen- Ai, 3 (1976), p. 237, argues, 'Emperor Kammu
dants of Temmu and Tenji. Kitayama Shigeo had a strong sense of resistance to the Temmu
2LLLPA, Nihon Kodai Seiji Shi no Kenkya H lineage of emperors, but also was strongly con-
t~tti~i~teF~t Iwanami, 1959, p. 479, scious that his own imperial line had foreign
regards Kammu's lineage loyalty, his antecedents.' Sasayama regards the move as
'nostalgia for the works of his great-grand- an attempt by the Kammu-led Tenji lineage
father Emperor Tenji', as the primary reason and its allies to find an alternative capital to
for leaving Nara. the Temmu lineage's capital at Nara, and
Hayakawa Shohachi Y)iII JA, RitsuryO sought it in the ancient base of the Tenji
Kokka R Shogakukan, 1974, p. 347, lineage in Yamashiro, where immigrant
sees the need to break free of Temmu lineage's lineages were firmly rooted.
power base in the Nara area as the most impor- 49 Kidder, p. 67.
tant reason for the move.

This content downloaded from

158.195.4.7 on Mon, 20 Feb 2023 11:38:24 UTC
All use subject to https://about.jstor.org/terms
344 Monumenta Nipponica, 40:3

and move the capital back to Nara.50 The attempt failed; Heizei's children
were drummed out of the dynastic succession pool, and his grandchildren were
created commoners with the surnames Ariwara W-P and Oe XfL.
So the monarchy remained in Heian-kyo. Temmu had won the battle in
the Jinshin Disturbance in 672, perhaps, but Tenji's descendants won the war.
For the next millenium the monarchy had court and capital in the shadow of
Tenji's tomb.
Lineage identification with particular sites and particular provinces was a
powerful factor in political dynamics, lineage conflict, and choices of the loca-
tion of the court. The elegist for the Temmu line, Kakinomoto no Hitomaro,
surely recognized this. His 'Poem written ... when he passed the ruined
capital at Omi' is a a clear challenge to the legitimacy of the Tenji line for hav-
ing taken the 'capital' out of Yamato province. At the same time, the poem
asserts that the Yamato basin-not coincidentally the seat of the Temmu
lineage's power, and the site of its palaces and tombs-was the only legitimate
location for the 'capital':5 1
Since the reign of the Master of the Sun
at Kashiwara by Unebi Mountain,
where the maidens
wear strands of jewels,
all gods who have been born
have ruled the realm under heaven,
each following each
like generations of the spruce,
in Yamato
that spreads to the sky.

What was in his mind

that he would leave it
and cross beyond the hills of Nara,
beautiful in blue earth?

50 'Heizei' and 'Heijo' are written with the

placed a large statue of a warrior at the sum-
same characters, MAR. mit of Higashiyama to guard the capital, and
51 Translation from Levy, pp. 53-55; MYS,prayed, 'If, in the ages to come, there is an at-
1, #2841, in NKBT, 4, p. 26. tempt to build the capital in some other pro-
Heike Monogatari +T[1oi reports that vince, may thou be the guardian god of this
Kammu guarded against the abandonment of capital and prevent it.' This reported episode
Heiankyo, echoing Hitomaro's defense of is part of Heike Monogatari's condemnation
Yamato. Kammu sets Yamato province down of Taira Kiyomori +-KA for moving the
as but one of many that had housed the capital to Fukuhara in 1180.
capital, saying, 'From of old the emperors The Tale of Heike, tr. Hiroshi Kitagawa &
from generation to generation built their Bruce Tsuchida, University of Tokyo Press,
capitals in many different provinces, but none 1975, pp. 290-95, esp. p. 292.
were the equal of Kyoto [in Yamashiro].' He

This content downloaded from

158.195.4.7 on Mon, 20 Feb 2023 11:38:24 UTC
All use subject to https://about.jstor.org/terms
 TOBY: Why Leave Nara? 345

Though a barbarous place

at the far reach of the heavens,
here in the land of Omi
where the waters race on stone,
at the Otsu Palace
in Sasanami
by the rippling waves,
the Emperor, divine Prince,
ruled the realm under heaven....

After the move from Nara, and particularly after the revolt of Heizei and
Kusuko, fundamental changes were made in both the succession system of the
imperial house, and in the relationship of the imperial house, its consort clans,
and the nobility, to the capital and the provinces. These changes guaranteed
both that succession to the throne would never again be challenged in the ways
it was by Temmu, KOnin, or Heizei, and that territorial challenges to the
primacy of Heian-kyo as the capital would not arise. For the move from Heijo,
this recent threat to Heian, and the controversy over the locus of the court and
the capital arose from several related factors.
First, the imperial lineage group proved inconveniently prolific, begetting
too many potential successors. Despite efforts at limiting the succession pool
through dynastic shedding techniques such as the Cie system,52 there were still
too many candidates to ensure smooth succession, or to guarantee an emperor
the certainty that a designated heir-apparent would succeed. Indeed, Temmu
himself was not an oe, but deposed his rival Kobun, who was. Furthermore, as
long as the nobility included clans of heir-producing consorts who resided in
territorial bases outside the court-town itself, there remained the danger of a
territorially based challenge to the locus of the royal court, and this, it seems,
is what occurred in the case of the abandonment of Nara.
The solution to the problem of surplus candidates for the succession has
been mentioned briefly in the case of Heizei's progeny. The imperial line
developed a new and more effective technique of dynastic shedding, of barring
unneeded scions from potential candidacy. Prince Oama, and other younger
offspring, titled oMi - X, C XE, or shinnC )iE, were still 'royal', in the sense that
they fully shared in the charisma of the Yamato line. If necessary, they could
succeed, as did Oama no Jfi or Shirakabe no C. They might be named heir ap-
parent, as in the case of Osabe shinnc, or sire heirs to the throne, as did Shiki
XM- no cji, Shirakabe's father. But Heizei's son Takaoka rA was deposed as
crown prince after the rebellion of Heizei and Kusuko, while the children of
his other son, Prince Aho PM, were demoted from royal status and expelled
from the succession pool, their charisma sullied by the new surnames Ariwara
and Oe. This technique of creating new lineages of 'subjects', a strategy one

52 Kiley, p. 47.

This content downloaded from

158.195.4.7 on Mon, 20 Feb 2023 11:38:24 UTC
All use subject to https://about.jstor.org/terms
346 Monumenta Nipponica, 40:3

might call 'secularization of surplus royals', appears to have been a more effec-
tive method of shedding than had been previous techniques. Unlike the earlier
shedding ploys, no one successfully challenged this secularization process,
returning from a surname-bearing lineage group to succeed to the throne, and
as far as we can be certain, only one person, Taira no Masakado Ar d. 940,
openly tried.
That the Heian nobility was aware that the 'granting' of a surname was the
public mark of secularization seems clear from the first chapter of Genji
Monogatari gfi.$M~. Hikaru X, son of the emperor by a woman of relatively
humble family, has been bereft of his mother from birth. He is his father's
favorite as the young man's mother had been the emperor's beloved compa-
nion. As a result of this favoritism, Hikaru becomes the focus of considerable
political jealousy among the maternal relatives of other imperial offspring of
'better' family. The boy's fate is sealed, ironically, by the glowing prognosti-
cation of a physiognamist from Koryo, who predicts that he will 'ascend to
the highest place [i.e., become emperor]. '53 This prediction merely serves to
intensify the suspicions and hostilities of the maternal relatives of other royal
offspring, especially the Minister of the Right, grandfather of the crown
prince, and the emperor is forced to take drastic steps to save Hikaru:

Lacking the support of maternal relatives, the boy would be most insecure as a
prince without court rank, and the emperor could not be sure how long his own
reign would last. As a commoner he could be of great service.... He was so pro-
ficient [at his studies] that it seemed a waste to reduce him to common rank. And
yet-as a prince he would arouse the hostility of those who had cause to fear his
becoming emperor. Summoning an astrologer of the Indian school, the emperor
was pleased to learn that the Indian view coincided [with his own feelings, and
the Koryo physiognamist's prediction]; and so he concluded that the boy should
become a commoner with the name of Minamoto or Genji.54

Hikaru is created a Minamoto to disqualify him from the succession, thus to

protect him from the attacks of other members of the nobility who were in-
terested in promoting their own nephews for the succession. The technique of
secularization, itself indirectly an outgrowth of the conflict over where the
court should locate the capital, therefore served to eliminate surplus can-
didates, minimize succession disputes, and hence reduce the risk that this new
'capital' might itself, like Heijo, prove to have been only a court.
This discussion suggests that we may be too hasty by about a century in
equating the founding of the imperial city of Heijo with the final transition
'from court to capital', hence with the founding of the Japanese imperium.
The behavior of the imperial house, both in Nara and in its departure from the
city, indicates that Heijo-kyo still shared some qualities of a transitory court.

53 The Tale of Genji, tr. Edward in NKBT 14-18, Iwanami, 1958-1963, 14, p. 45.
Seidensticker, Knopf, New York, 1976, 2 54 The Tale of Genji, 1, p. 15; Genji
vols., 1, p. 14. See GenjiMonogatari , Monogatari, in NKBT, 14, pp. 44-45.

This content downloaded from

158.195.4.7 on Mon, 20 Feb 2023 11:38:24 UTC
All use subject to https://about.jstor.org/terms
 TOBY: Why Leave Nara? 347

Shomu's flirtation with other sites in the central prov

sion, as do the conflicts within competing segments of the imperial lineage
group and the nobility that attended both Shomu's and Kammu's moves from
Nara. The transformation of the imperial nobility in the early Heian period
into an exclusively metropolitan elite on the one hand, and the development
of new techniques of dynastic shedding on the other, were not merely methods
to guard against challenges to the primacy of the imperial line. They were
symbolic of new relationships between court, nobility, and country, keys that
succeeded for the first time in locking the 'court' into a true, permanent
'capital', one that would survive for a millenium.

This content downloaded from

158.195.4.7 on Mon, 20 Feb 2023 11:38:24 UTC
All use subject to https://about.jstor.org/terms