Heian Court Aristocracy

Japanese Marriage Institutions in The Heian Period

Author(s): William H. McCullough

Source: Harvard Journal of Asiatic Studies , 1967, Vol. 27 (1967), pp. 103-167

Published by: Harvard-Yenching Institute

Stable URL: https://www.jstor.org/stable/2718385

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

is collaborating with JSTOR to digitize, preserve and extend access to Harvard Journal of Asiatic Studies

This content downloaded from �������������54.84.104.155 on Wed, 02 Jun 2021 18:37:54 UTC�������������

All use subject to https://about.jstor.org/terms

JAPANESE MARRIAGE INSTITUTIONS

IN THE HEIAN PERIOD

WILLIAM H. McCULLOUGH STANFORD UNIVERSITY

For in what stupid age or nation Was marriage ever out of fashion ?

Hudibras, Pt. ii, Canto ii

INTRODUCTION

N k i rARRIAGE institutions constitute one of the dominant strands in the fabric of every human society, shaping the

family, determining relationships among individuals, affect- ing the distribution of property, and operating in countless other ways to help make a society and its civilization the unique entities they inevitably are. Common sense and brief reflection upon our own so- ciety confirm the crucial role assigned to these institutions by sociolo- gists, and there is no need to labor the obvious conclusion that one's understanding of an alien society must always be distorted as long as the nature of its marriage forms remains obscure. It is disconcerting, then, to find that most Western studies of the Heian period either have ignored the marriage institutions of the time altogether, or, per- haps worse, have assumed tacitly that those institutions were gener- ally identical with the ones familiar from more recent periods of Japa- nese history. The consequent failures in understanding and missed insights have tended to obscure the workings of Heian society and to make it seem even more remote and mysterious than it actually is.

A notable exception in this regard is Ivan Morris' recent work on Heian court society, The World of the Shining Prince (New York, 1964). In his chapter, "The Women of Heian and Their Relations with Men" (pp. 199-250), Professor Morris breaks new scholarly ground by devoting several pages to a description of Heian marriage

institutions and ceremonies. Although his brief account does not attempt to do more than give a bird's-eye view of the subject, it is by far the most detailed description published in a Western language to date, and may be taken hopefully as a sign of awakening interest among Western scholars in this vital area.' Unfortunately, Professor

103

This content downloaded from �������������54.84.104.155 on Wed, 02 Jun 2021 18:37:54 UTC�������������

All use subject to https://about.jstor.org/terms

104 WILLIAM H. McCULLOUGH

Morris' description is neither full enough nor factually accurate enough to provide a satisfactory introduction for the critical reader.

Part of the responsibility for Western neglect of Heian marriage

institutions must be borne by Japanese scholars, who themselves have

often misunderstood or failed to appreciate fully their significant fea-

tures. Publication of the works of the late Takamure Itsue iMORA has advanced our knowledge of Heian marriage practices immeasur-

ably, however, and there is increasing reflection of this development

in recent Japanese studies. Takamure's works,2 taken together with

those of Nakayama,3 Nakagawa,4 Yanagida,5 and Ariga,6 provide a

sound base from which any study of Heian marriage must begin, and

it is a pleasure to acknowledge my heavy indebtedness to them in the

pages that follow.

My own researches stem not from sociological interests, but from

the practical problem of attempting to understand Heian texts in

which references to marriage practices pose serious problems of inter- pretation. The principal purpose of this paper, therefore, will be to

provide a reasonably complete description of these practices, espe-

cially in their institutional aspects. The description will be limited to the courtier class at Kyoto during the tenth, eleventh, and twelfth cen- turies, since the bulk of the pertinent documentation for the period lies in those centuries and is restricted almost exclusively to the

aristocracy at the capital. For reasons of space, descriptions of mar- riage ceremonies and of the institutions of imperial marriage must be

deferred to a later occasion. It may be useful at the outset to provide some general definition of

what is meant in this discussion by the term "marriage." For present purposes, perhaps the most useful definition is that of Edward Wester- marck, who has said that marriage is "a relation of one or more men to

one or more women which is recognized by custom or law and involves

certain rights and duties both in the case of the parties entering the

union and in the case of children born of it."' This sums up what seem

to be the minimum and necessary conditions for the existence of mar- riage as an institution in Heian society: i) the physical relationship

between a man and woman, continuing normally over an extended

period of time and resulting in children; 2) the recognition of the re-

lationship by the society as an accepted mode of behavior; and 3) the acknowledgement of family responsibilities (however minimal) by

This content downloaded from �������������54.84.104.155 on Wed, 02 Jun 2021 18:37:54 UTC�������������

All use subject to https://about.jstor.org/terms

JAPANESE MARRIAGE INSTITUTIONS 105

both members of the marriage.8 It will be assumed that whenever all three of these conditions are met the result is marriage, but that the absence of any one of them is sufficient to deny that a relationship is marital. This means in concrete terms that it will not be necessary to

deal with any relationship that is ephemeral or condemned by the so- ciety-e.g., rape, incestuous unions, casual liaisons, etc.-but that relationships that fulfill the conditions set forth in the definition must be accounted for, regardless of whether or not they fit preconceived notions concerning the location of marriage, the frequency of relations

between the husband and wife, the, nature of the relationship between parents and children, or any other factor.

MARITAL RESIDENCE

In the marriage institutions of any society, one of the most critical features is the marital residence rule, which determines where a cou-

ple lives after marriage. Sociologists and anthropologists have identi- fied three principal modes of marital residence among societies gener-

ally: 1) virilocal, in which the man and wife take up residence near or at the house of the man's parents; 2) uxorilocal, in which, they reside at the house of the woman's parents; and 3) neolocal, in which they

occupy an independent house of their own.9 A fourth, less common, type is duolocal residence,, in which spouses live separately, with the husband visiting his wife but not living with her.10 It will be the task of this section to show that marriage residence among aristocrats in the

Heian period was consistently either uxorilocal, duolocal, or neolocal, and, equally important, that it was never virilocal.

The earlier history of marriage residence rules in Japan is not at all

clear. As we go back into the Nara period (710-784) and before, the sources of information become very meager and appear at times to be

contradictory. The surviving parts of the Yoro Code (718) relating to

marriage and the family1" point generally to the institutions of a Chinese kind of patriarchal Grossefamilie, according to which a bride

was brought to her husband's family home and lived there permanent- ly with him in a very subordinate position., The poems of the Man%- yoshui (ca. 759), on the other hand, suggest that duolocal residence

was common, with the husband visiting his wife's house by night,

having relations with her, and departing the following morning. Scholars who are acquainted with the sources and the period seem to

This content downloaded from �������������54.84.104.155 on Wed, 02 Jun 2021 18:37:54 UTC�������������

All use subject to https://about.jstor.org/terms

106 WILLIAM H. McCULLOUGH

be generally agreed that the legal provisions of the code were not

wholly representative of social realities,12 but there is no agreement on just how great the divergence might have been. Some would prefer to believe that marriage in the Nara period followed the forms suggested

by the legal code in being essentially virilocal, but that couples began their married life with the visiting relationship reflected in the poetry. This view assumes that after an initial period of duolocal residence the

bride moved to her husband's house, or, more particularly, to the house of his family.'3 Other scholars believe that the marriage and family provisions of the code were largely window dressing concocted by impressionable sinophiles, and that the normal marriage in the Nara period was of the duolocal type.'4 Still others take an intermedi-

ate position, asserting that permanent virilocal marriage was practiced by the upper classes in advanced areas of Japan, while duolocal mar- riage was the rule for the lower classes and the inhabitants of more

backward areas.'5 These views, and other modifications of them, are based not only on the provisions of the code and on contemporary

literature, but also on eighth-century family registers (koseki PU) and corvee rolls (keicho tf 0*). More refined analysis of the latter docu- ments may lead ultimately to a consensus of opinion.

Nara marriage institutions remain obscure because no extant eighth-

century prose literature deals with marriage in sufficient detail to per- mit reliable interpretations of the official documents, or to give sub- stance to suggestions implicit in the poetry. In the early centuries of the Heian period, however, the body of relevant prose works begins

to grow, and by the tenth century, literary and historical sources be-

come plentiful enough to supply a generally clear outline of the chief features of contemporary marriage practices among the aristocratic classes at Kyoto. From that time until the end of the twelfth century,

the documentation increases steadily in variety, quantity and reliabil- ity. During these later centuries it is possible to identify three princi-

pal types of marital residence: duolocal, uxorilocal, and neolocal. The precise nature of each varies from century to century, and there is an overall drift leading ultimately to the dominance of neolocal residence in the last part of the period, but throughout there is an almost total

absence of virilocal marriages.'6 This point deserves repeated stress- ing, since a failure to comprehend it and interpret it correctly has

proved a frequent stumbling block in studies of Heian court society and culture.

This content downloaded from �������������54.84.104.155 on Wed, 02 Jun 2021 18:37:54 UTC�������������

All use subject to https://about.jstor.org/terms

JAPANESE MARRIAGE INSTITUTIONS 107

Although tenth-century Japanese literature contains innumerable allusions to what seems to be duolocal residence, the references are

often so brief and ambiguous that they provide little concrete informa- tion about the institutions involved."7 The first detailed account of such a marriage occurs in Kagero nikki, a diary written late in the cen-

tury by one of the wives of Fujiwara Kaneie * (929-99o) .18 In this illuminating work the author describes her marriage, beginning with its inception in 954 and continuing to what appears to have been its

end twenty years later, when relations between her and Kaneie had

practically ceased. It is abundantly clear that the marriage remained a

visiting relationship from start to finish. The author's vagueness some- times makes it difficult to determine precisely where in Kyoto she was

living, but there is never any doubt that she always remained apart from Kaneie. For most of the marriage she seems to have stayed in her

mother's house "beside the cavalry grounds in the First Ward,"'9 i.e., near the intersection of Ichijo and Nishinotain avenues in the north-

eastern part of the capital. In later years she lived in two other houses

for short periods: one of uncertain ownership near her husband's residence,20 and the other a house owned by her father just outside

the eastern limits of the city, south of Konoe avenue.21 Although Kaneie visited these houses occasionally during the daytime, his visits usually occurred at night and were followed by an early morning de- parture on the following day. The frequency of the visits varied from time to time and tended to decrease through the years, but the author of the diary apparently continued to be recognized as Kaneie's wife, as is evidenced by the clothing she supplied for him and the correspond-

ence that was maintained between them. Kaneie, moreover, never

ceased to recognize his obligations toward their son, Michitsuna Saf (955-1020). He participated in the boy's training for court activities and obtained appointment to an appropriate official post for him after he had come of age.22

Michitsuna's mother was a secondary wife. This was the typical status of a woman involved in a duolocal marriage during the Heian period.- It is possible, however, that in the tenth century it was not uncommon for a principal wife to marry duolocally also. Despite the

existence of Kagero nikki and other sources, there is nothing to show

that Kaneie ever lived with Tokihime "13V (d. 980), the woman who is usually considered his main wife,23 and in Genji monogatari, which is

This content downloaded from �������������54.84.104.155 on Wed, 02 Jun 2021 18:37:54 UTC�������������

All use subject to https://about.jstor.org/terms

108 WILLIAM H. McCULLOUGH

thought to reflect generally the institutions of Kaneie's century,24 there is some further evidence along these lines. Genji's marriage to Aoi, for instance, remains duolocal throughout, and To no Chujo's marriage to the Minister of the Right's daughter is of a similar type. Duolocal residence of this kind may seem extraordinarily tenuous to

modern eyes, and, indeed, the lack of common residence would ap- parently be taken by some sociologists as sufficient grounds for deny- ing the existence of any marital relationship at all.25 Nevertheless, the

evidence that duolocal residence was a normal practice in the early

Heian period is too compelling to be dismissed. It may be noted, moreover, that the viability of such marriage institutions is corrobo-

rated in part by the evidence of similar practices found among primi- tive tribes in Malaya; New Guinea, Sumatra, Assam, Formosa, and India.26

Although duolocal residence may have occurred in marriages to principal wives during the tenth century, it is by no means the only

kind that appears in the literature. Uxorilocal marriage, with continu- ous residence of the man at the house of his wife's parents, was also probably known,27 and it is this type that is thought to have become in

the following century the expected residence arrangement for a man

and his principal wife, if she were the designated heir to her parents' house.28 There are a number of literary and historical instances of such marriages in the first three-quarters of the century, but perhaps the best documented is the marriage of Kaneie's son, Michinaga a (966-1028), to Rinshi {T- (964-1053), a daughter of Minamoto Masanobu OM (920-993).

Michinaga married Rinshi at the beginning of 988,29 when he was still in his early twenties. For some time thereafter the marriage re-

mained a visiting relationship of the sort maintained by Kaneie and Michitsuna's mother,30 but by 991 at the latest Michinaga had taken

up permanent residence at his father-in-law's Tsuchimikado ?i_PP4 mansion.3" Although Michinaga acquired several other houses by purchase or as the chieftain of the Fujiwara clan,32 his main residence continued to be at the Tsuchimikado mansion for the rest of his mar- ried life. This can be demonstrated in several ways. The sources for

the period, for instance, refer frequently to the mansion as Michinaga's

house,33 meaning not that he was its owner-title to the house prob- ably did not come to him until he rebuilt it following a fire in ioi634-

This content downloaded from �������������54.84.104.155 on Wed, 02 Jun 2021 18:37:54 UTC�������������

All use subject to https://about.jstor.org/terms

JAPANESE MARRIAGE INSTITUTIONS 109

but that he resided there. Moreover, Michinaga remained at Tsuchi-

mikado even after he took holy orders in 1019, for it was on the eastern grounds of the mansion that his H`joji " temple buildings were

erected35 and it was in the temple's Muryojuin 1 Hall that he

lived until his death in 1o28.36 The best evidence that the Tsuchi-

mikado mansion was Michinaga's principal residence is found, how-

ever, in his own diary, Mido kampaku ki. The diary covers the years

from 998 to io2i (with omissions) and it is apparent from its entries that throughout this period Tsuchimikado was the house where

Michinaga spent most of his time. There are more than three hundred

entries in the diary that connect Michinaga with his various houses,

but none can be interpreted to show him spending any extended

length of time away from the Tsuchimikado mansion, except when the mansion burned in ioi6 and was under reconstruction.37 There are, on the other hand, numerous indications that Tsuchimikado was his

customary residence. He explains during a drought in 1000, for in-

stance, that he has been forced to spend several days at his Biwa house

while the well at the Tsuchimikado mansion was being cleared out,38

and when he speaks of "returning home" from the palace, or from

another house, or from a trip, it always seems to be to the Tsuchi- mikado house.39

At some time after the marriage of Michinaga and Rinshi, Rinshi's

father and mother moved to another house nearby,40 leaving the cou- ple as the principal occupants of the Tsuchimikado mansion. Despite

the move, Michinaga's diary makes it clear that he and his wife main-

tained close ties with her parents through frequent visits to their

house, occasionally spending several days at a time there.41 The re-

moval of the parents to another house after a period of common resi-

dence with their daughter and son-in-law seems to have been fairly

common in uxorilocal marriages and may have even been the normal

practice. In these cases it is difficult to form a clear notion of how long

the common residence ordinarily lasted, but it must have varied ac-

cording to the circumstances of the family, the convenience of the

house, the age of the couple, and other such factors. When the parents

moved to another house they also took with them their remaining un- married children, and if there was a daughter among them, she might also marry uxorilocally. An illustration of this can be seen in the mar-

riages of the daughters of Fujiwara Akisue &* (1055-1123). Akisue's

This content downloaded from �������������54.84.104.155 on Wed, 02 Jun 2021 18:37:54 UTC�������������

All use subject to https://about.jstor.org/terms

110 WILLIAM H. McCULLOUGH

eldest daughter married Fujiwara Munemichi ,; (1071-1120) at some time before 109342 and appears to have lived uxorilocally with him in her father's Hachija house.43 By 1093 another of Akisue's

daughters had married Fujiwara Akimasa XA (1074-1136) and was perhaps also living at the Hachija house. At any rate, when Akisue

moved to his newly constructed Takamatsu A- house in that year, he took Akimasa with him, apparently leaving Munemichi and his

wife to occupy the Hachijo house by themselves.44 Akisue's daughter was soon divorced from Akimasa and married to Fujiwara Nakazane

44' (1064-1122), who had taken up uxorilocal residence with her at the Takamatsu house by io96.45 Not long after that date, Akisue left Nakazane and his daughter at the Takamatsu house46 and moved with his other children to another large establishment in Rokujo,47 where

he appears to have been joined for a time by Munemichi and his wife,48 and where two of his younger daughters also entered uxorilocal mar-

riages: one to Fujiwara Saneyuki tIi (io8o-i162), and the other to Minamoto Masasada Ot (1094-1162). The marriage to Saneyuki had taken place by 1103 or 1104, the date of the birth of the couple's eldest son.49 Saneyuki and his wife probably lived with Akisue in the Rokujo establishment from the outset of the marriage, but the earliest clear evidence of their residence there does not occur until 1112.50 They continued to live in Akisue's house for at least another six years after that date,51 apparently sharing the establishment with Masasada and his wife, who had been married in 1 108.52 As in Saneyuki's case,

there is no concrete evidence to show that Masasada was living in the Rokujo establishment during the early years of his marriage, but we

can probably assume that he was, since he was definitely living there

in 1119,53 and his continued residence in the establishment in subse- quent years can be traced to as late as 1152.54 A few years after Masa- sada's marriage, Akisue seems to have moved to another house that he

had constructed in Shichijo,55 leaving his daughters and sons-in-law

as the sole occupants of the Rokujo establishment. Many scholars, both in Japan and the West, have sought to deny,

either explicitly or by implication, that uxorilocal marriage existed in

the Heian period. In their view, after the conclusion of the marriage ceremonies at the house of the bride, the newly wed couple might re- main in the bride's house for a longer or shorter period, but always moved eventually to the husband's house, or more particularly, to the

This content downloaded from �������������54.84.104.155 on Wed, 02 Jun 2021 18:37:54 UTC�������������

All use subject to https://about.jstor.org/terms

JAPANESE MARRIAGE INSTITUTIONS 111

house of his parents-the type of marriage called "matri-patrilocal" by Murdock,56 or, to substitute the terminology employed here, "uxori-virilocal." Although the earliest scholar to give explicit formu-

lation to this view was perhaps the Edo antiquarian and writer Tada Nanrei XE1* (1698-1750) ,7 its currency in modern times is attributable mainly to the works of Yanagida Kunio, and especially to his "Mukoiri kb." In this essay Yanagida sought to show by inference from recent and modern ethnographical materials that Heian marriage

institutions did not differ essentially from later virilocal practices, except that the marriage rites took place at the bride's house and were

followed sometimes by a period of residence there. There are no

sources from the Heian period that can tell us whether marriages of

this type were ever practiced among the lower classes of the time, but in so far as the aristocracy is concerned, it is clear that Yanagida's

thesis is contrary to the facts established by contemporary sources. While not a single historical or literary example of the kind of mar- riage described by Yanagida has been adduced from the period to support his hypothesis, there are innumerable references to uxorilocal

residence in the contemporary sources, and there is more than one

historical case like that of Michinaga and Rinshi, in which the per- manency of such residence can be demonstrated beyond all reasonable doubt. The evidence on this point seems irrefutable, especially when viewed in the light of other features of the period which would tend to corroborate the existence of uxorilocal residence, such as female in- heritance of houses and maternal responsibility for children.58 Yet inexplicably, as Jenaga Saburo g7JC-I has pointed out,59 Yana- gida's conclusions have been accepted without criticism by some of

the most eminent Japanese scholars.60 And they have been perpetu- ated in the West, one regrets to add, by Ivan Morris, who has further confused the issue by asserting, "If the husband was a muko [in a mar- riage to a principal wife], he invariably moved to his wife's house and

lived there. . . . The muko system was already widely practiced in Heian Japan. The main raison d'etre of the muko, of course, was to ensure the continuity of the family where there were no suitable sons."61 The implication here that the term muko was reserved in the

Heian period for a man who was adopted into his wife's family to con- tinue the family line represents an opinion that, so far as I am aware, is unique with Professor Morris. The word muko, of course, could be

This content downloaded from �������������54.84.104.155 on Wed, 02 Jun 2021 18:37:54 UTC�������������

All use subject to https://about.jstor.org/terms

112 WILLIAM H. McCULLOUGH

used of any son-in-law, and, indeed, it appears in the most common

term for marriage in the eleventh century, mukotori, "taking of a son-

in-law."62 Quite to the contrary of Professor Morris' assertion, the

adoption of a man into his wife's family was by no means widely prac-

ticed in Heian Japan, and, in fact, I have myself neither found nor

seen reference to a single historical or literary instance of it. Takamure

(1953), p. 643, states the opinion that such adoptions did not take

place in this period at all. When a man took up residence at his wife's

house, as Michinaga did, he retained membership in his own clan, and in no sense did he become his father-in-law's family heir. In this con-

nection it is interesting to note that one does occasionally find cases in

the Heian period in which a daughterless man adopted a girl in order

to obtain a son-in-law. This happened most frequently perhaps in mar- riages to the emperor, but it also occurred in ordinary court families.63

Ariga Kizaemon has attempted to support Yanagida's thesis con- cerning Heian marriage on the generally sound sociological principle

that descent reckoning tends to conform to residence rules: i.e., in a

society with virilocal residence, descent reckoning tends to be patri-

lineal, and in uxorilocal societies, matrilineal.64 Ariga contends that since Japanese society has always been patrilineal as far back as records carry us, it is difficult to accept the existence of uxorilocal

marriage at any time in Japanese history.65 In addition to substituting theory for the evidence of sources, Ariga's view also implies that a patrilineal, uxorilocal society is a sociological impossibility. But this is not at all the case, despite the observed tendency of residence rules and descent reckoning to parallel each other. Robert F. Murphy has

observed the existence of patrilineal, uxorilocal institutions among the Munduruciu Indians on the upper Tapajos River in Central Brazil,66 and similar combinations are also reported for the Atjeh people in Northern Sumatra,67 the Yukaghir of northeastern Siberia,68 the

Karens of Burma,69 the Bare'e Toradja in the Celebes,70 and the Yurok Indians of Northern California.7" If the relative rarity of patrilineal,

uxorilocal institutions is due, as Murphy suggests, to the inevitably transitional nature of the antagonistic patrilineal-uxorilocal combina-

tion, this would accord well with the marriage practices of Heian court society, which were in a process of more or less continuous

change from at least the tenth century on, shifting toward the virilocal institutions of the Kamakura period and after.

This content downloaded from �������������54.84.104.155 on Wed, 02 Jun 2021 18:37:54 UTC�������������

All use subject to https://about.jstor.org/terms

JAPANESE MARRIAGE INSTITUTIONS 113

It is true that the nature of the change in Heian marriage institu-

tions poses difficult problems of analysis from the point of view of general sociological theory. It is usually assumed that a change in the

marriage residence rule is the catalyst that leads toward a change in

descent reckoning, and the implication of patrilineal, uxorilocal insti-

tutions is, therefore, that the society is moving away from patrilineal-

virilocal practices toward their matrilineal-uxorilocal counterparts. The earlier history of descent and marriage in Japan is too obscure to

allow firm opinions concerning the origins of the institutions found in

the Heian period, but perhaps part of the solution to the apparent

contradiction of general sociological experience-matrilineal descent

reckoning does not develop from Heian uxorilocal marriage, of course

-is to be found in the smallness of court society, since special groups

seem often to behave in sociologically irregular ways; and in the great influence exercised upon that society by the strongly patrilineal Chi-

nese civilization. It may also be observed that part of the theoretical

difficulty may arise from the characterization of Heian descent reckon-

ing as patrilineal, when it seems in fact to have been bilateral, but with

a marked patrilineal bias.

Whatever the ultimate answer to this problem may be, it must not ignore the facts established by contemporary sources to achieve theo-

retical consistency, as Ariga has seemingly done. It may be noted in this connection that even if the Yanagida-Ariga position is accepted,

a certain contradiction remains, since, according to Murdock, "matri-

patrilocal residence indicates immediate or proximate derivation from

a matrilocal structure."72 If Japanese descent reckoning has always been patrilineal, as Ariga seems to prefer to believe, then the conflict between patrilineal descent and uxorilocal residence is merely pushed

back in time, and one must explain how the uxorilocal rule developed from its presumed virilocal predecessor and then passed on into uxori-

virilocal marriage without ever affecting the system of descent reck- oning.

At least part of the confusion surrounding the question of uxorilo-

cal residence in the Heian period can be traced to a misunderstanding of the neolocal marriages that were also practiced during this time.

Since these marriages began often with a period of uxorilocal resi- dence, which was followed by removal to an independent house, they are easily confused with the true uxori-virilocal marriages found in

This content downloaded from �������������54.84.104.155 on Wed, 02 Jun 2021 18:37:54 UTC�������������

All use subject to https://about.jstor.org/terms

114 WILLIAM H. McCULLOUGH

rural areas of Japan in more modern times.73 There is a vital difference in quality between uxori-virilocal and neolocal (or "uxori-neolocal") marriage, however, and those scholars who have confused them, or failed to distinguish clearly between them, are as much in error as those who have assumed that one or the other represented the exclusive marriage practice of the Heian period.

It will have been observed that in Michinaga's case a period of duolocal residence preceded the initiation of his permanent residence

at Tsuchimikado. This was apparently a common enough practice in marriages to principal wives in his day, but as the century progressed

the period of duolocal residence seems to have been shortened, until

at length it disappeared altogether. When Michinaga's son, Yorimichi

RA (992-1074), married his principal wife in 1009, their relation- ship apparently remained a visiting one only until a new residence could be prepared for them by Yorimichi's father-in-law.74 By the generation of Yorimichi's son, Morozane MiA (1042-1101), the sources cease to mention these initial periods of duolocal residence, and one is probably to assume that in most cases husbands and their

principal wives took up immediate residence together. There are two general types of neolocal residence in the Heian

period, distinguishable according to whether the house was supplied by the wife's family or by the husband. When a girl did not marry uxorilocally, it seems usually to have been the responsibility of the girl's family to provide the initial residence for her and her husband. The husband normally supplied the residence only if the social or economic status of his wife differed radically from his own,75 or if the

wife had been deprived of her family's support through some mis- fortune such as the death of her parents. However, even in normal

marriages, after a couple had lived for a time in the house of the wife's

parents or in an independent house provided by them, the husband did often supply another, neolocal residence to which he and his wife moved, and as uxorilocal marriages became less and less frequent in the twelfth century the husband's role in providing the marital resi- dence seems to have expanded steadily, coming then to include some- times the initial residence too.

Neolocal marriages in which the husband provided the initial resi- dence are found with particular frequency in Heian fiction, perhaps because they were associated especially with romantic love. Poor girls

This content downloaded from �������������54.84.104.155 on Wed, 02 Jun 2021 18:37:54 UTC�������������

All use subject to https://about.jstor.org/terms

JAPANESE MARRIAGE INSTITUTIONS 115

and girls without proper families were probably never in a position to receive aristocratic husbands in uxorilocal or duolocal marriages,

since a girl of this sort either would not have a house where her mar-

riage could take place-presumably a frequent case among orphaned girls-or the house would not be grand enough for a high-ranking

courtier to reside in or visit as a husband. If a marriage was to take

place under these circumstances, therefore, it was necessary for the husband to supply the house himself.

On the other hand, there was obviously little except love to attract

an aristocratic man to such a marriage. Wealth, influence, prestige,

political advantage- the normal considerations in Heian marriages- would have all been absent from the man's calculations, and the con- temporary literature is undoubtedly correct in depicting these matches

usually as affairs of the heart. In Genji monogatari, for instance, Gen-

ji's putative son Kaoru plans to marry an orphaned daughter of Prince

Hachi and bring her to live at his Sanjo house.76 Although the daugh- ter dies before Kaoru is able to marry her, the novel makes it abun- dantly clear that the marriage would have been a love match if it had

taken place, and there is good reason to consider the prospective mari- tal residence as basically neolocal in type. The house, it is true, had belonged originally to Kaoru's widowed mother and was her normal residence at the time of the proposed match, but it had come to her from her own father77-not from Genji-and she and Genji had never occupied the house together as husband and wife. Moreover, the

original house had burned after Genji's death78 and the place that Prince Hachi's daughter was to occupy had been rebuilt by Kaoru.

The neolocal nature of the house is suggested too when we recall that

true virilocal residence for Kaoru would have meant his father's

Rokujo establishment, where his mother and Genji were married and spent all of their married life together.

Other marriages of a similar type occur, for example, in Utsubo monogatari79 and Ochikubo monogatari,80 and one also sometimes

comes across reverse cases in which neolocal residence resulted from

the marriage of a man to a woman of higher rank than himself.8' It was probably because of the close association of romantic love with marital

residences provided by the husband that the author of Sarashina

nikki A EI H i depicted herself when she was a young, romance- reading girl as yearning for a Genji-like lover who would establish her

This content downloaded from �������������54.84.104.155 on Wed, 02 Jun 2021 18:37:54 UTC�������������

All use subject to https://about.jstor.org/terms

116 WILLIAM H. McCULLOUGH

in a residence of this sort.82 In a story from a little later period we even discover the parents of a wife of proper rank and with an adequate home of her own demanding that their daughter's husband demon- strate his affection for her by taking her into his house.83

Individual circumstances must also often have made it necessary or convenient for the husband to provide the house in the course of an ordinary marriage. When a couple's original residence was destroyed in one of the frequent Kyoto fires, for instance, the responsibility for providing a new house might devolve upon the husband, especially if

the wife's father was dead, retired, or otherwise incapacitated, or if the

husband was of an age and position that made it as easy for him to provide a new house as it was for his father-in-law. A case of this sort is found in the marriage of Fujiwara Morozane to an adopted daughter

of Fujiwara Nobuie MM (1019-1061). The marriage began in 1059 at Nobuie's house,84 but in the month following the nuptials the house

burned85 and Morozane and the rest of the family were apparently forced to move into temporary quarters elsewhere. Morozane and his

wife did not acquire a permanent residence again until three or four

years later, when they moved into a house built by Morozane on

Kazan'in ltiliiR land obtained from his sister Kanshi 14i (1036- 1127) .86 Morozane may have been expected to provide a new residence for himself and his wife after the fire as a matter of course, since he was already seventeen years old at the time of his marriage and held almost the same office and rank as his father-in-law.87 However this may be, any possibility that Nobuie might supply the couple with another house was eliminated by his death in the early part of io6i, just two years after the fire, and it appears actually to have been from about this time that Morozane initiated work on the Kazan'in house.88

Although the evidence is sparse and uncertain, there is also some suggestion that in neolocal marriages involving children of the em-

peror the house was frequently, or perhaps even regularly, supplied

by the husband. At any rate, in the few cases where we can obtain any idea at all of the source of the residence in such marriages, it seems al-

ways to have come from the husband's side, and never from the wife or her family. The marriages of Reizei's son Atsumichi VI (981-1007) to daughters of Fujiwara Michitaka M1* (953_995)89 and Fujiwara Naritoki j41 (941-995),90 of Daigo's son Minamoto Takaakira ?iJJ (914-982) to a daughter of Fujiwara Morosuke,9' and of Sanja's daughter Shishi I1i4 (d. 1048) to Fujiwara Norimichi &A (996-

This content downloaded from �������������54.84.104.155 on Wed, 02 Jun 2021 18:37:54 UTC�������������

All use subject to https://about.jstor.org/terms

JAPANESE MARRIAGE INSTITUTIONS 117

1075)92 all appear to have involved neolocal residence of this type. So too, for instance, did Genji's marriage to Murasaki and Niou's to

Kozeri (Nakanokimi) in Genji monogatari, but in these cases the resi- dence can also be accounted for by the positions of the wives, neither of whom had the resources or family backing for uxorilocal marriage to a high-ranking husband.

There were thus a number of opportunities for the husband to pro-

vide the residence in neolocal marriage during the Heian period, but

if the residence was to be neolocal at the outset of a marriage the pro-

vision of the house was almost always the responsibility of the wife's

family, and later residences for the couple also came frequently from the same source. In particular cases it is usually impossible to deter-

mine why a father chose to establish his daughter in a neolocal resi- dence, instead of having her live uxorilocally in the family house, but it is clear, at least, that the age of the couple was an important

negative consideration. Very young couples were rarely, if ever, al-

lowed to live in independent houses, no doubt for obvious practical reasons of domestic management: a husband and wife who were still

children could scarcely have been expected to run a household by themselves. This limitation was a significant factor in determining the shape of residence patterns, since child marriage was extremely com- mon in the Heian period, the bride often being only eleven or twelve years old and her husband a scant year or two older.93

In the case of a more mature couple, the wife's family might elect to provide a neolocal house for them, either from the beginning of the

marriage94 or after they had spent a period of uxorilocal residence in the family's house.95 Considerations of living space in the house of the wife's family and the availability of alternative housing elsewhere may

have frequently been decisive factors in determining whether a mar- riage would be neolocal or not. We find, at any rate, that when an

older daughter was married and living uxorilocally, a younger daugh-

ter ordinarily was able to have a similar marital residence only if the parents had left the first daughter and her husband and moved to another house, or if the older daughter herself had moved to a neolocal residence;96 otherwise, the younger daughter seems usually to have

married neolocally.97 If space was a determining factor in these cases, as appears likely, one can also infer perhaps that when an elder

daughter married neolocally it was often for similar reasons.

Whatever the origins, neolocal residence sponsored by the wife's

This content downloaded from �������������54.84.104.155 on Wed, 02 Jun 2021 18:37:54 UTC�������������

All use subject to https://about.jstor.org/terms

118 WILLIAM H. McCULLOUGH

family was very common in the eleventh and twelfth centuries. During

much of that time, in fact, the normal rule of residence at the begin-

ning of regular marriages between spouses of equal status was either of

this type or uxorilocal. It may be more meaningful to follow Takamure

in making no essential distinction between these two residence types.

Uxorilocal marriage is usually defined as the custom that "requires

the groom to leave his parental home and live with his bride, either in

the house of her parents or in a dwelling nearby."98 But when we re-

call that in the Heian marriages of which we are speaking both the

wife's family and the husband's usually lived within the limited con- fines of Kyoto (an area about three and one-third miles long and three miles wide), and more particularly in the eastern half of that city, it is

quite possible that from most points of view the ownership of a house

was a more crucial factor in determining the nature of a marriage than its precise location. If one adopts this position-and it seems a per-

suasive one in some ways-then the neolocal marriages sponsored by the wife's family can be treated for the most part simply as variations on a basically uxorilocal mode. Even if such an interpretation is ac-

cepted, however, it is important to bear in mind that the neolocal vari- ation represented an important development that led eventually to the virilocal residence patterns of the Kamakura period. Uxorilocal mar- riage at the house of the wife's family disappears almost entirely in the late twelfth century, it seems, and the independent marital residence becomes nearly universal, with the husband and his family assuming

an ever greater responsibility in providing the house.

HOUSES AND PROPERTY

Closely related to the question of marital residence was the problem of house ownership and succession in the Heian period. In uxorilocal marriages, and in neolocal marriages sponsored by the wife's family,

title to the marital residence was usually passed on to the wife by her

parents.99 Indeed, the feeling that a married woman shoucld be the owner of her own residence appears to have been so strong among the

Kyoto aristocracy that even in the case of a neolocal house provided

by the husband, title to the house was sometimes transferred to the

wife."'0 Since the house that a wife received from her parents was fre- quently property that had belonged originally to her mother or to her

mother's family, succession to marital residences in this period was

This content downloaded from �������������54.84.104.155 on Wed, 02 Jun 2021 18:37:54 UTC�������������

All use subject to https://about.jstor.org/terms

JAPANESE MARRIAGE INSTITUTIONS 119

often clearly matrilineal in character. Moreover, even when the father was the owner of the house, the succession seems normally to have

gone by preference to a daughter, rather than to any of the sons, and if this type of case can be included with the first, it is probably actually

correct to say that succession followed the female line in the majority

of marital residences at this time. In most instances the house passed

down from mother to daughter for only one or two generations before

circumstances intervened to produce a change in ownership. Taka-

mure (1953) has shown, however, that in one case, that of the Onono-

miya 4'2k house, the succession lasted through four generations over a period of nearly two hundred years (see genealogical chart

below) .101

SUCCESSION TO THE ONONOMIYA HOUSE

Fujiwara Saneyori

Naritoshi Yoritada

Sanesuke

Sukeyori Daughter Kaneyori

Daug hter Sukeie|

|DIaug,hter _Yoshizane|

Tadayori Daughter Moroyori DaughterL= Tadamune

Moromitsu Tadamitsu

(Names of persons whose marital residence was at the Ononomiya house are enclosed in squares.)

This content downloaded from �������������54.84.104.155 on Wed, 02 Jun 2021 18:37:54 UTC�������������

All use subject to https://about.jstor.org/terms

120 WILLIAM H. McCULLOUGH

Although there is little that one can add to Takamure's able demon- stration of the Ononomiya succession, a brief recapitulation of the case's main features may be given here, since it is of considerable interest for the general history of Heian marriage. The Ononomiya house, located south of Oimikado and west of Karasumaru in the northeastern part of Kyoto,'02 belonged originally to Emperor Mon- toku's son, Koretaka IR- (844-897), who is said to have won it in a sugoroku game.'03 After Koretaka's death the house passed by some unknown means into the hands of Fujiwara Saneyori 23g (goo- 970) ,104 who bequeathed it to his adopted son and grandson, Sanesu- ke,105 perhaps because Saneyori's daughters had all predeceased him and there were no eligible female heirs.'06 As a husband for his daugh- ter Chifuru ?1:t, Sanesuke brought into the Ononomiya house Fuji- wara Kaneyori *fl (1014-1063), the eldest son of Yorimune (993- 1065) and a grandson of Michinaga. The surviving parts of Sanesuke's diary, Shoyfiki, do not mention the marriage itself, but references to Kaneyori in the last part of the diary leave little room for doubt that he

was married to Chifuru and living in his father-in-law's house. At one point, for instance, Sanesuke explains that although a period of absti- nence (monoimi) has ended for him, only the north gate of his house has been opened, the east gate remaining closed because Kaneyori is observing a period of abstinence too.107 It is obvious from this entry that Kaneyori was living at the Ononomiya house, and when we read another entry in the diary from the following month showing Kane- yori and Sanesuke's daughter riding together in the same carriage on a visit to Kaneyori's sick mother,108 we can conclude without much difficulty that the marriage had in fact occurred. It is confirmed in the

longer texts of Okagami ), which state specifically that Sanesuke's daughter was the wife of Kaneyori.109 Several years before the mar- riage, apparently, Sanesuke had already executed a document convey- ing the Ononomiya house and most of the rest of his real and movable property to Chifuru.1"0 Sanesuke's sons, it is interesting to note, are mentioned in this document only to warn them against interfering in

their sister's property and to divide among them a few agricultural

holdings that had not been assigned previously to Chifuru. Sanesuke lived to be nearly ninety years old, however, and may have survived his own daughter, for in the end she failed to receive legal title to the

Ononomiya house, which passed instead directly from Sanesuke to a

This content downloaded from �������������54.84.104.155 on Wed, 02 Jun 2021 18:37:54 UTC�������������

All use subject to https://about.jstor.org/terms

JAPANESE MARRIAGE INSTITUTIONS 121

daughter of Chifuru and Kaneyori born in 1036 or 1037."' This girl became the wife of Fujiwara Sukeie *WIi (1o36-1o88),1i2 and, like her mother, lived uxorilocally at the Ononomiya house, as we know from sobriquets used of both her and her husband. In her later years, for instance, Sukeie's wife was called "the Ononomiya nun,""13 and

Sukeie himself was known as "the Ononomiya chiunagon" 1-,114 the same title that had been applied earlier to his father-in-law, Kaneyori.1"' Sukeie and his wife in their turn also had a daughter,

who also was married uxorilocally. The daughter's husband was Fuji-

wara Yoshizane tEw (1070-1132) ,116 the fourth son of the Fujiwara clan leader, Morozane. Yoshizane is known to have lived at his father- in-law's house because of his sobriquet, "the Ononomiya dainagon"

4kf E,117 and also because of an entry in Hyakurensi% that speaks of "Minister Yoshizane's Ononomiya house" in recording the burning of

the establishment in 121.118 The house did not actually belong to Yoshizane and his wife, however, but remained in the hands of his mother-in-law, Sukeie's wife, a very long-lived woman who died in 1134

when she was nearly one hundred years old."9 Before her death the mother-in-law passed the deed to the Ononomiya house on to her daughter's daughter (the daughter of Yoshizane and his wife).120 This was done, it must be noted, despite the existence of a number of sons and grandsons. The granddaughter became the wife of the poet Mina- moto Moroyori 0IAl (1o68-1140) ,121 who was called "the Ononomiya dainagon,"'22 and whose residence at Ononomiya is further indicated in Hyakurensho, which speaks of the place in 1145 as "the house of the

late Minister Moroyori.""3 By the time of Moroyori's children, how- ever, in the last half of the twelfth century, marital residence seems to have been shifting toward the virilocal rule, and succession to the

Ononomiya house passed similarly from the matrilineal to the patri- lineal line. Thus Moroyori's son, Moromitsu Mit, inherited the house from his mother'24 and passed it on to his descendants,'25 who became the Ononomiya branch of the Minamoto family.

The history of the Ononomiya house illustrates another interesting

result of Heian marriage institutions. As readers of the contemporary

literature will be readily aware, one of the most common ways of re-

ferring to an individual at that time was by the name of his place of residence. In Genji monogatari, for instance, there are many personal names of this sort: Kiritsubo no koi 1 (Genji's mother), Nijo

This content downloaded from �������������54.84.104.155 on Wed, 02 Jun 2021 18:37:54 UTC�������������

All use subject to https://about.jstor.org/terms

122 WILLIAM H. McCULLOUGH

no ue Z1WE (Murasaki), Momozono no miya fA1R1I (Genji's aunt), Rokujo no miyasudokoro A1i.?iI BJ (Lady Rokujo), Akashi no kimi WRIi (Lady Akashi), and so on. Chronicles and diaries of the time also often refer to great nobles by names taken from their places of

residence. Minamoto Shigenobu AV (922-995), for instance, was called the Rokuja Minister of the Left;126 his elder brother, Masanobu (Michinaga's father-in-law) was called the Ichija Minister of the Left,127 or the Tsuchimikado Minister;128 Minamoto Yasumitsu 4I3Y2

(924-995) was called Momozono;129 and Fujiwara Nagayori 7-g

(932-1010) was called the Third Rank Minister of Yamanoi i14.130 Such instances might be multiplied almost without limit, but of even more interest than their frequency is the way in which they tended to

follow a matrilineal line of succession, passing from son-in-law to son-

in-law as each succeeding daughter in a house married. This is well

illustrated by the Ononomiya name. Ononomiya was used first as a

designation for Prince Koretaka,13' and then passed on with the Ononomiya house itself to Saneyori,132 and later to Sanesuke.133 From

the generation of Sanesuke, it followed a regular line of matrilineal descent through the sons-in-law, Kaneyori, Sukeie, Yoshizane, and Moroyori, all of whom are referred to in contemporary sources by the

Ononomiya name. On the other hand, one finds no trace of the name among the male descendents of these men. Sanesuke himself, for example, had four adopted sons,134 none of whom was ever called Ononomiya, and Yoshizane's son by Sukeie's daughter is listed in the genealogies simply by his official title.'35

The matrilineal descent of residential names in the Heian period

can be demonstrated in a number of other cases,136 and its existence is further corroborated by the discontinuity of such names in the male line. Perhaps the best illustration of this patrilineal diversity of names

-parallel to our own matrilineal diversity of surnames-is found

among the principal scions of the main Fujiwara family. As may be seen in the accompanying genealogical chart, in which each name is followed by the residential designation that was commonly used of a man before his retirement or entrance into holy orders, no Fujiwara heir bore the same designation as his father.

This content downloaded from �������������54.84.104.155 on Wed, 02 Jun 2021 18:37:54 UTC�������������

All use subject to https://about.jstor.org/terms

JAPANESE MARRIAGE INSTITUTIONS 123

Fuyutsugu (Kan'in)

Yoshifusa (Ichijo)

Mototsune (Horikawa)

Tokihira (Hon'in) Tadahira (Koichijo)

Saneyori (Ononomiya) Morosuke (Kuj6)

Koremasa (Ichijo) Kanemichi (Horikawa) Kaneie (Higashisanjo)

Michinaga (Tsuchimikado)

I1 1 Yorimichi (Takakura) Norimichi (Nijo)

Morozane (Kyogoku)

Moromichi (Gonijo)

Tadazane (Kamonoin)

Tadamiclhi (Kujo) Yorinaga (O3imikado)

Toward the end of the twelfth century and at the beginning of the

thirteenth, however, as the inheritance of the marital residence shifted

to the patrilineal line, so also did the accompanying house designa- tions, which thus became some of the best known family names of the

Kamakura period: Kujo, Nijo, Rokujo, Kyogoku, and so on. These later names, therefore, bore a very different significance from their counterparts of the Heian period, and we must distinguish carefully

between the two usages if we are to avoid the mistake of the historians

who refer to Saneyori's descendants as the Ononomiya branch of the Fujiwara family, despite the fact that none of them except Sanesuke is -actually called by that name in contemporary sources.

This content downloaded from �������������54.84.104.155 on Wed, 02 Jun 2021 18:37:54 UTC�������������

All use subject to https://about.jstor.org/terms

124 WILLIAM H. McCULLOUGH

The house was perhaps the most important part of a woman's

belongings in the Heian period, but it was by no means her only pos-

session. She was also able to inherit and own all other types of prop-

erty, including rights in agricultural land. There is even some sugges-

tion that her claim to the inheritance of parental private property was

stronger than a man's, since it seems to have been not at all unusual

for a father to bequeath his entire estate to a daughter, even though he

also had sons whom one might have expected to share in the inheri-

tance. Fujiwara Tamemitsu Al (942-992), for instance, had two sons living at the time of his death,'37 but it is reported in Eiga monogatari that he left his elaborate Ichijo house and all his other property to his

third daughter.'38 One can find a number of similar cases elsewhere,

both historical and fictional,'39 and there is no suggestion that they

were considered out of the ordinary in any way. The property of

Heian women, moreover, remained theirs even after marriage, as we

have seen. This was no doubt a matter of major importance in a society

where polygyny was the rule, divorce was unrestricted, and the

economic responsibility of a man for his family was dependent upon

conscience or convenience. The independently owned property of a

woman provided her and her children with a certain degree of security against the loss of her husband and compensated to some extent for

the looseness observed in many Heian marriage ties. The property of a woman's family played a direct, and perhaps

sometimes a dominant, role in the economy of her marriage even be-

fore she inherited it. The ability of a woman and her family to meet the material requirements of a husband was, in fact, one of the major

factors affecting the woman's marriage prospects. This is stated with admirable succinctness in a speech delivered by Nakayori's mother-

in-law in Utsubo monogatari:

Nowadays when a man thinks of taking a wife he wants to know whether her father and mother are living, whether she has a house and is able to launder and sew, and whether she will reward his men and take care of his horses and oxen. No matter how beautiful the girl is and no matter how refined and charming, if the man discovers her living a miserable life in some wretched, desolate hut, he will think, "How squalid! What a trial it would be to live there!" And he will never go near her.'40

The reference to laundering and sewing in this passage is note-

worthy, since the supplying of clothes for the husband seems to have

This content downloaded from �������������54.84.104.155 on Wed, 02 Jun 2021 18:37:54 UTC�������������

All use subject to https://about.jstor.org/terms

JAPANESE MARRIAGE INSTITUTIONS 125

been regarded as one of the more important economic functions ful- filled by the wife and her family. The elaborate robes of the Heian

nobility were perhaps the most costly of contemporary consumer

goods, and the attention lavished upon them has few parallels even

among the most dedicated modern devotees of fashion. When a wife

provided her husband with a costume, therefore, it constituted a

major form of support, and this was still true even when the husband

supplied the cloth himself, since the tailoring required skill and a large

investment of labor. The importance attached to this task may be

gauged perhaps by the great frequency with which it is mentioned in

all varieties of sources from the tenth to the twelfth centuries. It may,

indeed, have been the wife's principal economic contribution to her

husband in some duolocal and neolocal marriages, as appears to have

been the case, for instance, in the marriage of Kaneie and the author of Kagero nikki. In other marriages, however, the provision of clothes

clearly constituted only one part of the general economic support

given a husband by his wife and her family. For example, Michinaga's mother-in-law (Rinshi's mother) is said to have provided him with clothes regularly twice a year until her death,'14 but in addition to this, of course, he was also living in his wife's house and no doubt

being supported in many other ways by her family. Moreover, the heroine of Ochikubo monogatari in the earlier part of the story spends

most of her time sewing for her sisters' husbands, who we later learn were also dependent upon their father-in-law for general support.'42

The husband, therefore, almost always had an economic stake in

the marriage, although its precise nature might vary considerably

according to circumstances. In a neolocal marriage at the husband's

house, the wife's economic contribution might be limited to the sew-

ing of clothes, but in uxorilocal marriage the husband was perhaps

largely or even entirely supported for a time by his wife's family. In

the latter case it is difficult to determine how long the support ordi-

narily lasted, but it must have depended largely upon the relative

circumstances of the husband and his wife's family. It is perhaps legit- imate to infer, both on the basis of the limited evidence available and from our knowledge of the general conditions of Heian economic life,

that a husband assumed an ever greater economic responsibility for his wife and family as his office and rank rose and his income increased. Such, for instance, seems to have been the case with Michinaga, who

This content downloaded from �������������54.84.104.155 on Wed, 02 Jun 2021 18:37:54 UTC�������������

All use subject to https://about.jstor.org/terms

126 WILLIAM H. McCULLOUGH

in the later years of his marriage expended great sums of money on his wife's residences.

While economic considerations were important in determining a man's choice of wives, conversely they might also become the occasion

for the dissolution of a marriage. A number of early examples might

be cited to illustrate this point,'43 but perhaps the frankest expression

of it occurs somewhat later in Konjaku monogatari shiu + E W13, in a story entitled "How the Daughter of the Vice-Minister of Central

Affairs Became the Maidservant of a District Official in Omi."''44 This is the story of a girl and her husband who were living together in her

parents' house, but were forced to separate after the death of the

parents removed the couple's source of support. The girl says to her

husband, "While my parents were alive we were able to manage in one way or another, but now I shall no longer be able to take care of

you. It won't do for you to be shabbily attired at court. You must act

in your own best interests." The husband delays a while, but his clothes become more and more disreputable, and finally he leaves the

house. Eventually the wife becomes a maid in the service of a district

official in Omi, where she and her ex-husband, now the governor of

the province, meet again and become lovers without recognizing each other. When their identities are finally established, she dies from the shock.145

The economic circumstances of a marriage were affected not only by the wealth of the wife and her family, but also by the influence that a father-in-law could wield to advance the career of his son-in-law.

This seems to be what is meant when Michinaga is made to remark in

Eiga monogatari that "a man's wife makes him what he is,"146 and it is undoubtedly the case that such considerations loomed largely when a man was seeking a spouse. In Ochikubo monogatari, for instance,

Michiyori's former wet nurse rejects Ochikubo as a wife for him be-

cause she has no parents on whom he can rely. The nurse wants Michiyori to marry a girl whose high-ranking father can look after

him properly, the rather clear implication being that the father-in-law would see to the advancement of his son-in-law's career.'47 Reflections of similar sentiments and examples of the consequences of the influence or lack of influence of a father-in-law occur repeatedly throughout Heian literature, and even in a late derivative of the Heian novel like

Shinobine monogatari U 43F I0* (13th c.?) the success of the hero,

This content downloaded from �������������54.84.104.155 on Wed, 02 Jun 2021 18:37:54 UTC�������������

All use subject to https://about.jstor.org/terms

JAPANESE MARRIAGE INSTITUTIONS 127

Kintsune, is attributed not to his high-ranking father, but to the influence of his father-in-law.148 This, like the husband's economic

dependence on his wife's family, was unquestionably another effect of

the uxorilocal tendency in Heian marriage, illustrating again the close

bond that existed between a husband and his affinal relatives. The

husband, of course, often inherited wealth from his own family and

obtained official advancement through the intercession of his own

father, but it is fair to say that in many cases a man's official and

economic position seems to have been determined more by his father-

in-law than by his father.

REGULATION OF MARRIAGE

The Chinese-inspired legal code of the eighth century made a rigid

legal and ceremonial distinction between wives and concubines, speci- fying among other things that a man might have only one wife, but

placing no restriction upon the number of his concubines.149 Polygy- nous marriage continued to be practiced in the Heian period, but the

distinction between wife and concubine seems to have been consider- ably blurred, and it sometimes becomes difficult to establish any status

differences at all among.the female spouses. The three main wives of

Kaneie, the two wives of Yiigiri (Kumoi and Ochiba), and the four wives of Morozane's son, Tsunezane, were all referred to and treated more or less equally as principal wives. Furthermore, Takamure has shown that all the sons of the Fujiwara leaders from Kamatari to Kaneie received the same initial appointments to court rank (Junior

Lower Fifth), although the several sons of each father were by different

mothers.150 This was contrary to the codes, which provided that in the case of the upper nobility only the heir by a man's legal wife was to re-

ceive the Junior Lower Fifth Rank as his initial appointment. Taka- mure's inference-that there was in fact no distinction in status

among the wives of the Fujiwara leaders-may be open to question,

but the evidence she has marshalled does suggest that such distinc- tions as may have existed were far less pronounced than the code would lead one to expect.

There are other indications of a more general equality among

Heian wives than the legal code allowed for, especially in the earlier part of the period, but it is also true that in most polygynous mar-

riages of the time, and perhaps in all of them, a main wife does seem to

This content downloaded from �������������54.84.104.155 on Wed, 02 Jun 2021 18:37:54 UTC�������������

All use subject to https://about.jstor.org/terms

128 WILLIAM H. McCULLOUGH

have been recognized, and that the other wives were in some sense secondary to her. The main wife was distinguished in three chief

ways: i) by the titles applied to her (kitanokata, kitanomandokoro,

mukaime, chakusai A-A); 2) by her husband's residence with her, if a neolocal or uxorilocal marriage was involved; 3) and by the higher

official positions occupied by her sons. Takamure contends that the recognition of a main wife was essentially ex post facto, resting upon the choice of her son as her husband's heir or upon the husband's affection for her.'5' This may have been true in some cases, but since the first woman formally married by a man almost always became his principal wife, particularly in neolocal and uxorilocal marriages, the

status of the first wife was identifiable, in prospect at least, from the

very beginning of her marriage. She might be divorced eventually and replaced by another woman, but until that time she had a strong customary claim to the position of principal wife. Reference to the

main line of the Fujiwara family demonstrates this very clearly. If we begin with Tadahira in the early tenth century, we find that he

had sons by two wives, Minamoto Kishi hFT- (875-925) ,152 a daughter of Emperor Uda, and Minamoto Shoshi EI3T, a daughter of Yoshiari

NM (845-897). Kishi was probably the earliest of Tadahira's two wives, since she bore him his first son, Saneyori, in goo or 9O1,'53 when he was about twenty, and since it was not until eight years later that Shoshi bore his second son, Morosuke.154 There is only one oblique reference to Shoshi in the surviving parts of Tadahira's diary,155 but Kishi is mentioned frequently,'56 and on two occasions, in inter- linear notes added possibly by Saneyori, she is referred to as the kita- nokata,'57 a term reserved normally for the principal wife.158 There can be little doubt, therefore, that she was regarded as Tadahira's main

wife, and this is confirmed by her son's pre-eminence in office among his brothers and by his position as clan chieftain.'59 Saneyori's mar-

riage to his main wife, a daughter of his uncle Tokihira Jf4` (871- gog), seems to have been his first also, since it took place shortly after his coming-of-age ceremony in 915, when he was fourteen or fifteen years old.160 This wife was the mother of all his sons,'6' and since the

second of these, Yoritada &T, eventually became sessho and clan chieftain,162 she may be safely regarded as Saneyori's principal spouse,

even though she died relatively early (in 933)163 and Saneyori had other wives after her.164 After Yoritada, political power and the clan

This content downloaded from �������������54.84.104.155 on Wed, 02 Jun 2021 18:37:54 UTC�������������

All use subject to https://about.jstor.org/terms

JAPANESE MARRIAGE INSTITUTIONS 129

chieftainship passed back to the line of his uncle Morosuke, whose sons had come to occupy a dominant position at the court, thanks

largely to the influence of their sister, Anshi +&'T (927-964), the mother of Emperors Reizei and En'yul. The three eldest sons of Moro-

suke-Koremasa 'fRU'f (924-972), Kanemichi *A (925-977), and Kaneie-may be regarded as his principal heirs, since they surpassed their brothers in office, rank, and influence. Each held the offices of sessho and daijo daijin,'65 and each had his turn as clan chieftain.166 All three of these sons were born of the same mother, FuLjiwara Seishi

T4 (d. 943) whose marriage to Morosuke when he was fourteen or fifteen years old makes it almost certain that she was his first wife.168

Morosuke had other wives who bore him sons, and two of them, the

imperial princesses Gashi 3T (910-954) and Koshi MUfT (d. 957), were of much better birth than Seishi, who came from a family of rela-

tively low-ranking officials. At least one of the two, Gashi, was mar- ried to Morosuke during Seishi's lifetime,169 but the positions of

Seishi's sons make it clear that it was she who was regarded as Morosuke's principal spouse.

After the death of Kaneie in ggo, the office of kampaku and the clan chieftainship passed to his eldest son, Michitaka,'70 who was followed in these or equivalent positions by his two younger brothers, Michi- kane AX (961-995)171 and Michinaga.172All three of these men were Kaneie's sons by Tokihime.173 Kaneie had married Tokihime by the time he was twenty-three or -four years old,'74 and although it is pos- sible that he had had other wives previous to her, she preceded by at least two or three years the author of Kagero- nikki,75 the only other woman by whom he had a recognized son. Tokihime's superior posi- tion among Kaneie's wives is suggested by the pre-eminent position of

her sons-Kaneie's son by the author of Kager3 nikki (Michitsuna) never rose above the office of dainagon'76-and also by her listing in

Sompi bummyaku (Pt. 2, p. 203) as Kaneie's kitanomandokoro, a term equivalent in use to kitanokata when applied to the wives of sessho or kampaku.

From the generation of Kaneie's sons, political power and the clan

chieftainship descended through Michinaga and his heirs. Michinaga's two wives, Rinshi and Meishi, were both of high birth177 and both bore

sons,178 but Rinshi, his first wife179 and the mother of his eldest son, Yorimichi,180 was unmistakably the main wife. Her son was Michi-

This content downloaded from �������������54.84.104.155 on Wed, 02 Jun 2021 18:37:54 UTC�������������

All use subject to https://about.jstor.org/terms

130 WILLIAM H. McCULLOUGH

naga's heir, receiving the offices of sessho and clan chieftain directly after him in 1017,181 and it was at the house of Rinshi's parents that Michinaga had his permanent residence. Furthermore, Fujiwara Sane-

suke's diary specifically identifies Meishi as a "secondary wife" (sho/sai

a), in contrast with Rinshi, who is called kitanokata.182 After Yori- michi the positions of kampaku and clan chieftain passed first to the

second of Rinshi's sons, Norimichi,183 and then to Yorimnichi's son, Morozane.184 Yorimichi married his first wife, Takahime, a daughter

of the learned imperial prince Tomohira, in 1009.185 She remained his principal spouse throughout his life, despite her failure to bear chil-

dren'86 and despite the existence of other wives who presented him

with both sons and daughters.'87 Perhaps as a result of Takahime 's childlessness, Emperor Sanjo suggested in 1015 that Yorimichi take

another wife, and proposed his own daughter, Shishi,'88 but this plan

was given up when Yorimichi fell ill soon thereafter,'89 and Yori-

michi's heir came eventually from a secondary wife, Fujiwara Kishi g T. Although Kishi bore five sons and a daughter to Yorimichi, she never replaced Takahime as his main wife, and, indeed, Yorimichi is said to have kept his relations with Kishi secret for some time through

fear of Takahime's jealousy.'90 Takahime had no sons whose careers could serve to mark her position among Yorimichi's wives, but there can be little question, nevertheless, that she was his principal wife. She is the only one of the wives mentioned in Michinaga's diary,' which is undoubtedly a reflection of her prominence; she was referred to throughout her long life as Yorimichi's kitanomandokoro,'92 a title never used of his other two wives; and she was the only woman with

whom Yorimichi ever lived over an extended period of time.'93

Kishi's son, Morozane,- succeeded his father as kampaku and clan

chieftain in 1O75. 14 His main wife was Reishi FXqF (1040-1114), an adopted daughter of Fujiwara Nobuie whose true father was Mina-

moto Morofusa Mi?g (1008-1077). Reishi's superior position among Morozane's numerous wives can scarcely be doubted, since she was

the wife with whom he lived,'95 the mother of his heir, Moromichi,196 and the only one of his wives who is referred to in contemporary

sources as his kitanomandokoro.'97 Two years before Morozane's mar- riage to Reishi in 1059, however, he had had a son by a daughter of Minamoto Sadanari t (1002-1044),198 and it is possible, therefore,

that Morozane's marriages constitute an exception to the general rule

This content downloaded from �������������54.84.104.155 on Wed, 02 Jun 2021 18:37:54 UTC�������������

All use subject to https://about.jstor.org/terms

JAPANESE MARRIAGE INSTITUTIONS 131

that the first wife was the main wife. On the other hand, it should be

noted that Morozane's relationship with Sadanari's daughter was pos-

sibly nothing more than a casual liaison, involving none of the cus-

tomary ceremonies and conditions associated with formal marriage.

There are several circumstances which would agree with this inter-

pretation: i) the relationship seems to have been of very brief dura-

tion, for there is no mention of Sadanari's daughter in contemporary

sources after the birth of her son; 2) the son became a Buddhist

priest,i"' which might be taken to mean that he did not have Moro-

zane's support in public life; 3) the mother was a lady in waiting at the

court of Empress Kanshi (Morozane's adopted daughter)200 and thus

belonged to a class in which the casual liaison was a commonplace;20i and 4) Morozane was especially likely to be involved in such a rela-

tionship, since he was famous from his youth for his numerous love affairs.202

Morozane's son Moromichi is known to have been married to, or to have had children by, five women, daughters respectively of Fujiwara

Toshiie {tg (1019-1082), Fujiwara Nobunaga 4>k (1022-1094),203

Fujiwara Yoshitsuna AMij , Taira Sadatsune A , and Minamoto Yoritsuna , (1025-1097). Since he married Toshiie's daughter

Zenshi k (io6o-ii5o) in 1076,204 when he was only thirteen or fourteen years old, and since she bore his eldest son, Tadazane, less than three years later,205 it seems probable on the face of it that she was his first wife, and this is borne out by what is known about his

other marriages. His children by the daughters of Yoshitsuna, Sada-

tsune, and Yoritsuna were all born from two to twenty years after the

birth of Tadazane,206 suggesting that they were the issue of later mar- riages, and Nobunaga's daughter is known definitely to have been a

later wife.207 The daughters of Yoshitsuna, Sadatsune, and Yoritsuna

were at best only secondary wives,208 but Zenshi and Nobunaga's daughter were both at different times principal spouses of Moromichi.

Zenshi's position is indicated by her husband's residence with her

during the five or so years following their marriage in 1076209 and by

her son's succession to Moromichi's offices after his death in 1Ogg.210 Relations between Moromichi and Zenshi deteriorated, however, 21 and the marriage ended in the kind of informal divorce that appears to

have been characteristic of the Heian period.212 The date of the estrangement seems to have been somewhere around 1o82,253 but

This content downloaded from �������������54.84.104.155 on Wed, 02 Jun 2021 18:37:54 UTC�������������

All use subject to https://about.jstor.org/terms

132 WILLIAM H. McCULLOUGH

Moromichi's marriage to Nobunaga's daughter may not have occurred

until about 1094.214 In that year, at any rate, she is referred to specifi- cally as Moromichi's kitanohata,215 and whenever this term, or kita-

nomandokoro, is used thereafter it always appears to refer to her.

Although Moromichi never seems to have had any children by Nobu-

naga's daughter, he apparently remained married to her until the end

of his life, living with her and calling her his kitanokata,216 and it was

she, not Zenshi, who attended to the Buddhist services for him after

his death.217

Like his father, Tadazane also had two main wives, Minamoto Nin-

shi 4{T- (io66-i 152), whom he divorced, and Minamoto Shishi Oi1i (1070-1149). Ninshi, the daughter of a high court official named

Toshifusa, married Tadazane in 1o89,218 when he was ten or eleven

years old and she was twenty-one or -two.219 During the several years

that the marriage lasted, Tadazane lived with her, usually at her

father's house,220 and she was regarded as his main wife, as we know

from Ima kagami +, which refers to her as Tadazane's kitanokata.221 But Ninshi was, in effect, childless, since all of her children died in

infancy,222 and perhaps as a result of this Tadazane eventually left her for her cousin Shishi,223 a daughter of Akifusa W (1037-1094).

This must have occurred somewhere around 1097, for although Nin-

shi and Tadazane were still definitely living together in 1095224 and Ninshi's second daughter was born in 1097, Tadazane had already

had a daughter by Shishi in 1095 or 1o96,225 and by 1102 at the latest he was living with Shishi226 and she was being referred to as his

wife.227 Shishi had been in the service of Tadazane's grandmother, Reishi, and had become a favorite of ex-Emperor Shirakawa, by whom

she had had a son in 1092.228 According to Eiga monogatari,229 the ex- emperor ceased to have any communication with her after the birth of

the child, and her father, Akifusa, seems to have disowned her. Ima

kagami, on the other hand, reports as hearsay that Tadazane had be-

come infatuated with Shishi after catching sight of her at his grand-

mother's house, and had finally won her as his wife by persuading

Reishi to intercede on his behalf with Shirakawa.230 Regardless of how the marriage began, there can be no doubt that Shishi became Tada-

zane' s main wife. His residence with her,23' the references to her as his kitanomandokoro,232 and her son's position as his heir233 all support this view. Her marriage to Tadazane, one should note, represents the

This content downloaded from �������������54.84.104.155 on Wed, 02 Jun 2021 18:37:54 UTC�������������

All use subject to https://about.jstor.org/terms

JAPANESE MARRIAGE INSTITUTIONS 133

case of a secondary wife (as she was when the marriage began) replac- ing another woman as her husband's main wife, contradicting Profes-

sor Morris' opinion (p. 224) that Tone course of action . .. was cate-

gorically forbidden: to let a secondary consort take the place of the

principal wife." Considerably later in Tadazane's life, when he was

past forty and Shishi was nearly fifty, he had another son, Yorinaga,

by a daughter of Fujiwara Morizane *X but there is nothing to sug- gest that Yorinaga's mother ever succeeded in displacing Shishi as

Tadazane's main wife. On the contrary, the absence of any clear

reference to her in contemporary sources suggests that the author of

Sompi bummyaku was correct in listing her as Tadazane's concubine

(shol&) .234 The marriages of Shishi's son, Tadamichi, may be taken as a final

illustration of the general rule that the first wife of a Heian man

normally became his principal consort. When Tadamichi was about

fourteen years old, it was arranged for him to marry a daughter of ex-

Emperor Shirakawa,235 but the wedding was postponed,236 and al-

though it continued to be discussed until Tadamichi was eighteen or so,237 it apparently never took place. During this time, Tadamichi had

a son, the priest Eshin EMS (1114-1171), by a daughter of the pro- vincial governor Fujiwara Motonobu 9f{ ,238 but his relationship with Motonobu's daughter was apparently brief and informal,239 and it was not until 11 i8 that he formally married for the first time. The bride

was Fujiwara Sashi + (io8-i 155) ,240 a daughter of Munemichi -A (1071-1120), who was a brother of Tadamichi's grandmother, Zenshi. After the marriage Tadamichi resided permanently with

Sashi,241 a circumstance indicating clearly that she was his main wife,

and her position never changed, although she bore only one daugh-

ter242 and Tadamichi's male heirs were sons by other women. Sashi

was the only one of Tadamichi's wives to be called kitanomandokoro,243

and when she was granted the status of an empress (jugo )k1) in 1149, the imperial edict identified her specifically as e"the main wife of

the regent" (setsuroku no chakushitsu i2t*t).244 Tadamichi had children by other women, but these were all the issue of what were un-

doubtedly much later marriages, and there is nothing to indicate that

any of the mothers ever approached Soshi in status. The earliest of

them was probably Fujiwara Shinshi I,T- (d. 1178), who bore Tada- michi's eldest son and immediate heir, Motozane AX (1143-1i66),

This content downloaded from �������������54.84.104.155 on Wed, 02 Jun 2021 18:37:54 UTC�������������

All use subject to https://about.jstor.org/terms

134 WILLIAM H. McCULLOUGH

when Tadamichi was in his mid-forties. Shinshi's sister, Shunshi fS :F, was also the mother of a son by Tadamichi, born two years after Motozane,245 and a nyobo named Kaga b1IR (1124-1156) bore four other sons, the first in 1149,246 when Tadamichi was past fifty.247

Virtually nothing is known about any of these three women. This in itself is probably indicative of their secondary positions, and the few

available facts further confirm Soshi's superior status. Unlike Soshi, none of the three is ever identified in the sources as Tadamichi's wife,

but only as the mother of his children,248 and Sompi bummyaku makes Kaga's inferior position explicit by calling her "the serving maid

(kanjo '9*) of the Lord of Hosshoji 1f t4t1 [Tadamichi]."249 Although there was, of course, no legal requirement that the first

formally married wife of a man should be his principal wife, these

cases in the main line of the Fujiwara family and parallel cases from other families indicate that this was, in fact, the custom. There were

undoubtedly exceptions to the rule, but under the weight of such a

well-established tradition the husband's discretion in recognizing a

woman as his main wife could scarcely have been so wide as Takamure suggests. In practice one can assume that the woman first married by a man was his principal wife, and, conversely, that a principal wife was also her husband's first wife, unless divorce or death had intervened to remove a predecessor.

As in most polygynous societies, the number of wives actually kept by a Heian aristocrat at one time was very small, usually no more than two or three, and monogamous unions were probably in the majority. Under the conditions of duolocal marriage, it was comparatively sim-

ple for a man to maintain marital relationships with several women simultaneously, but even a noted gallant like Kaneie seems to have

restricted himself to three wives, and most men appear to have agreed

with the mother in Ochikubo monogatari who told her son, "To keep many ladies is to suffer many troubles."250 In neolocal and uxorilocal marriage it was much more difficult to arrange multiple marriages, and very few men could manage more than two wives at a time.251 A husband was normally expected to live with his main wife; if he had a secondary wife, his relationship with her would usually be of the duolocal type in a separate establishment.252 In the case of neolocal

marriages in which the house was supplied by the husband and the

wife was dependent upon him, a secondary wife might occasionally be

This content downloaded from �������������54.84.104.155 on Wed, 02 Jun 2021 18:37:54 UTC�������������

All use subject to https://about.jstor.org/terms

JAPANESE MARRIAGE INSTITUTIONS 135

brought to live in the same house with the main wife, but this situation seems to have been rare, and certainly there is no historical case to

match Genji's menage at the Rokujo palace. On the other hand,

divorce or death of a spouse might lead to multiple marriages for both

men and women,253 and for the men casual liaisons with the nyobo class

at the several imperial and princely courts served to further diversify

their relations with the opposite sex and also to increase the number

of their children. Such liaisons sometimes developed into relatively

stable relationships and the women would then become true secondary

wives, but in most cases the bond was ephemeral. In Genji monogatari

Uma no Kami is speaking of this kind of affair when he says, "A man

may amuse himself well enough by trifling from time to time with

some lady at the Court; will get what pleasure he can out of it while

he is with her and not trouble his head about what goes on when he is

not there."254 One also finds a nyobo serving as a temporary wife for a

married man while he is away on appointment in the provinces,255 but

it is impossible to tell how common this was.

Other important determinants of Heian marriage were the customs

concerning incest, class endogamy, and divorce, all of which were

relatively simple and may be described very briefly.

The only unions that seem to have been regarded as definitely and unequivocally incestuous were those between parents and children and between brothers and sisters. There are no known examples of such marriages from this period, and there is positive evidence to

show that they were considered unthinkable. In Genji monogatari, for

example, as Tamakazura's supposed father, Genji, is pressing his suit

upon her, she laments that whatever defects her real father might

have, he at least would not make love to her.256 Yugiri, who thinks he is Tamakazura's half-brother, is restricted by incest taboos in his rela- tions with her, but when her true paternity is revealed, the restraints upon him are removed257 and fall instead upon Tamakazura's suitors Kashiwagi and Kabai, who learn now for the first time that she is their

half-sister.258 The rule against parent-child marriages does not seem

to have extended with equal force to unions between stepparents and

stepchildren, but such unions were still disapproved. Fujitsubo looked back upon the relationship she had with her stepson Genji "as some- thing horrible," but she allowed it to continue and became pregnant as a result of it.259 The manner in which Yugiri's guilty love for his

This content downloaded from �������������54.84.104.155 on Wed, 02 Jun 2021 18:37:54 UTC�������������

All use subject to https://about.jstor.org/terms

136 WILLIAM H. McCULLOUGH

stepmother, Murasaki, is portrayed also indicates that such relation- ships were considered wrong, but not completely out of the ques- tion,260 and this is suggested again in the simultaneous marriage of

ex-Emperor Kazan to a mother and her daughter.261 Except for parent-child and sibling marriages, however, there seem

to have been no limitations placed upon unions between relatives. Men openly married their aunts, nieces, and cousins, and there is no suggestion that these were considered undesirable matches in any

way. Examples of the following types of marriage are found:

i) With a paternal aunt: Fujiwara Michikane and Hanshi M.j-.262 2) With a maternal aunt: Minamoto Tsuneyori and his mother's

sister.263

3) With a niece (brother's daughter): Fujiwara Saneyoshi tAE

(1097-1157) and a daughter of Michisue A (1o9o-1128).264 4) With a niece (sister's daughter): Minamoto Takaakira and a

daughter of Gashi.265

5) With a first cousin (daughter of a paternal uncle): Fujiwara

Atsutada #ULP, (906-943) and a daughter of Nakahira.266 6) With a first cousin (daughter of a paternal aunt): Fujiwara Moro-

suke and Koshi 1-T (929-957), a daughter of Onshi S (885- 954). 267

7) With a first cousin (daughter of a maternal uncle): Minamoto

Toshikata MR (96o-1 027) and a daughter ofTadagimi .L.a(d. 969) .268 This type of marriage was especially frequent in the imperial family.

I have not succeeded in finding a case of marriage with the daughter

of a maternal aunt, but this is probably accidental, since the mothers of daughters are seldom listed in genealogies and the relationships are difficult to trace. Among several possible instances of such marriages, one may cite the case of Fujiwara Tsunezane, who married a daughter

of Fujiwara Kinzane, the husband of a sister of Tsunezane's mother.269 It is not known, however, whether Tsunezane's wife was actually a daughter of his mother's sister.

At the time of the promulgation of the court legal codes in the

eighth century, a strict and detailed system of class endogamy was established, and although the specifics of the system were no longer

operative in the tenth century, the notion of class-restricted marriage remained very strong.270 For the Heian aristocrat this meant that he

virtually never married outside the relatively small circle of families

This content downloaded from �������������54.84.104.155 on Wed, 02 Jun 2021 18:37:54 UTC�������������

All use subject to https://about.jstor.org/terms

JAPANESE MARRIAGE INSTITUTIONS 137

that were permitted to hold court rank and office, and that even with- in this circle his main wife tended to come from a family of approxi-

mately the same standing as his own. There seem to have been no legal penalties imposed against marriages that violated the general practice of class endogamy, but the man's interest in making a mar- riage that would best further his career and the woman's interest in

obtaining protection for both herself and her children tended together to enforce the custom effectively. Moreover, since good birth was prized for itself and a child of an unequal union was socially dis- advantaged compared to the position of one of his parents, a further impetus toward marriage among class equals was created.

Nevertheless, it must always be kept in mind that such restrictions

as did exist were not legal but customary, and that in practice numer-

ous exceptions to them occurred, at least between the classes within the court circle.271 This may be illustrated by referring again to the

marriages of the Fujiwara leaders discussed above. Most of the mar- riages followed the principle of class endogamy, with the husband and

wife coming from families that were roughly equal in social status: Tadahira's wife, Kishi, was the daughter of an emperor; Saneyori's

was a cousin; Michinaga's was the daughter of a Minister of the Left (sadaijin) and a granddaughter of an imperial prince; Yorimichi's was the daughter of an imperial prince; Morozane's was the daughter of a Minister of the Right (udaijin) and the granddaughter of an imperial prince; Moromichi's first wife was the daughter of a Minister of the Right and his second the daughter of an Acting Great Counselor (gondainagon) and the adopted daughter of a Chancellor (daijo daijin); Tadazane's two wives were both granddaughters of an imperial prince

and daughters respectively of a Minister of the Right and Minister of the Left; and Tadamichi's wife was the daughter of an Acting Great

Counselor. On the other hand, Morosuke's wife, Seishi, came from a definitely inferior family. The highest office held by her father Tsune-

kuni N$ 1 was a provincial governorship, and his highest rank was the Junior Fifth. The position of her grandfather, Arisada 4-0 (827- 873), as a Lesser Captain of the Imperial Guards (konoe no shosho 1z !iJ .t-) with Junior Fourth Rank272 was only slightly better. Tokihime, Kaneie's wife, was also of less illustrious birth than her husband, but

the difference in this case was not so marked as between Seishi and

Morosuke. Tokihime's father, Nakamasa ItE, was Commissioner of

This content downloaded from �������������54.84.104.155 on Wed, 02 Jun 2021 18:37:54 UTC�������������

All use subject to https://about.jstor.org/terms

138 WILLIAM H. McCULLOUGH

the Left City (sakyo no daibu AJ*:) and held the Junior Fourth Rank,273 which placed him somewhat above Seishi's father, but still excluded him from the circle of upper officials (those of the Third Rank and above). The position of Tokihime's grandfather, Yamakage

iii - (824-888), may have made her a more socially suitable wife for Kaneie, since Yamakage was a Middle Counselor (chuinagon) with Junior Second Rank274 and was included therefore in the higher class of officials.

In view of these and many other cases of a similar nature, it is dif-

ficult to agree with those who maintain that class endogamy was strictly observed at this time within the court circle. This is apparently

the opinion of Ivan Morris, for instance, who says that in Genji mo- nogatari "Genji is never able to take Murasaki as his principal wife, because, though her father is an imperial prince, her mother is the daughter of a provincial official."-275 Murasaki's mother, of course,

was not the daughter of a provincial official, but of a Great Counselor (dainagon),276 and as such she was well qualified by birth to be the parent of Genji's wife, or even of an imperial consort. Even assuming, however, that Murasaki's mother was of the class Professor Morris thought her to be, one still cannot find in Genji monogatari itself any evidence to support the view he has stated, and contemporary practice seems to disprove it conclusively. To illustrate this point more con- cretely, the marriage of Minamoto Takaakira and Morosuke's third daughter may be cited,277 since in the class relationship of the spouses involved it closely resembles Professor Morris' view of Murasaki and Genji. Takaakira, who is thought by some to have been one of the models for the fictional Genji,278 was, like him, the son of an emperor

(Daigo), while his wife's mother, Seishi, was the daughter of a pro- vincial governor, and her father, though high in rank, was actually inferior in birth to Murasaki's princely parent. Despite Seishi's in- ferior birth, the position of her daughter as Takaakira's main wife is as well established as can be hoped for in the tenth century (the age of Genji, it will be remembered): she bore her husband's three eldest sons;279 she is referred to as his kitanokata in Okagami (p. 125); and although she died in childbirth and was replaced as Takaakira's kitanokata by a younger sister whose mother came of imperial stock,280 it was one of the sons of Seishi's daughter who rose ultimately to the highest office.281

This content downloaded from �������������54.84.104.155 on Wed, 02 Jun 2021 18:37:54 UTC�������������

All use subject to https://about.jstor.org/terms

JAPANESE MARRIAGE INSTITUTIONS 139

Divorce as it appears in the available sources for the last three cen- turies of the Heian period was vague and informal to a degree that often makes it difficult to specify the time of a divorce, or, indeed, even to say whether or not it actually occurred. It may be that divorce was infrequent enough and its details unremarkable or unpleasant enough to create a false picture of its simplicity in the sources, but if we re- strict ourselves to what is actually known about it in this period, di-

vorce was no more complicated than the practices reported for the Zufni Indians by Ruth Benedict (p. 98). Despite the detailed provi- sions in the court legal code relating to the conditions and procedures

of divorce,282 one finds in practice no trace of any legal or customary

step, such as the issuance of a bill of divorcement, that served to mark clearly the end of a marriage. Instead, the man or woman who wished to terminate a marital tie simply ceased to have relations with the

other spouse, and no additional act of any kind appears to have been required to make the divorce effective. The ease and informality of

Heian divorce were related no doubt to many factors, but among the most important must have been the comparative independence of each

spouse from the other. As we have seen, the wife's property was sep- arately held and often separately administered, and in the early years of the marriage, at least, support for her and her children was usually provided by her own family. This meant that in the event of divorce there were few problems of property settlement, and since the divorced woman usually kept her children, it also meant that severe family dis- locations and economic hardships were not a necessary complement of broken marriages. The wife, moreover, retained membership in her own clan even after marriage, continuing to worship its divinities, to bear its name, and to be involved in its affairs, and divorce, therefore, required no change in her clan status even if her husband came from

a different clan.

The precise form taken by divorce in this period depended, of course, upon the type of marriage involved. It was most vague in the case of duolocal marriages. The visits of the husband in such marriages

might become very infrequent or even cease altogether for long periods of time, but since it was always possible that they might be resumed

with no preliminaries other than the consent of the wife, one cannot be certain that a marriage has actually ended until one of the spouses has taken holy orders or died, or the wife has remarried. Kaneie's

This content downloaded from �������������54.84.104.155 on Wed, 02 Jun 2021 18:37:54 UTC�������������

All use subject to https://about.jstor.org/terms

140 WILLIAM H. McCULLOUGH

relations with the mother of Michitsuna reveal the uncertainty of di-

vorce in duolocal marriages. At the beginning of the marriage in 954, Kaneie's visits were regular and frequent,288 but by the fall of the same

year Michitsuna's mother was already complaining of his absences,284

and in 956 his relations with "the lady in the alley" brought the mar-

riage almost to the breaking point.285 During the next year and more,

Kaneie paid only occasional visits to Michitsuna's mother, but in 957 he seems to have begun seeing her more frequently and a somewhat

closer relationship between them was resumed.286 At one point in 958,

Michitsuna's mother was apparently urged by someone to marry an-

other man,287 implying that Kaneie's unfaithfulness was sufficient ground for a clearcut divorce. She did not heed this advice, however,

and her marriage continued on its rocky course for several years more.

Finally in the summer of 973 Kaneie paid his last visit to her,288 and

thereafter the marital tie seems to have been definitely severed, for there is nothing in Kagero3 nikki or elsewhere to suggest that it was ever resumed.

In neolocal and uxorilocal marriages, divorce is more readily iden- tifiable, since it involved the departure of one of the spouses from the

common residence, but even in such cases an element of uncertainty may have remained if the man sought to continue his first marriage as

a visiting relationship. Fujiwara Asateru 40 (951-995), for in- stance, is said to have tried occasionally to visit his beautiful young first wife after he had left her to marry an ugly but wealthy widow, and

he was prevented from doing so only by the refusal of his ox drivers to

take him to her house.289 In most instances of uxorilocal marriage,

nevertheless, the departure of the man from his wife's house may be taken as a definite sign that the marriage was at an end. Heian sources

yield a number of examples of this kind of divorce, including the two

cases already cited above,290 and it is likely, therefore, that they were fairly common. It is noteworthy that contemporary texts reflect the uxorilocal customs of the time when they speak of divorce as the hus-

band leaving (izu, saru, etc.) his wife's house, an expression which

contrasts sharply with later centuries when a divorced wife was said to

have been sent back (oidasu, hima woyaru) to her parental home. Neo-

local marriages in which the house belonged to the wife or her family differed very little in matters of divorce from uxorilocal marriages, but

when the husband owned the house, divorce seems to have been rel-

This content downloaded from �������������54.84.104.155 on Wed, 02 Jun 2021 18:37:54 UTC�������������

All use subject to https://about.jstor.org/terms

JAPANESE MARRIAGE INSTITUTIONS 141

atively infrequent, perhaps because of the wife's greater dependence upon her husband. When divorce did occur under these circum- stances, it was necessary for the wife to leave her husband's house, as Prince Atsumichi's wife did when he brought Izumi Shikibu home

with him,291 and as Higekuro's first wife did in Genji monogatari when her husband married Tamakazura.292

Professor Morris has spoken of "the irrevocability of the principal marriage" in Heian society,293 and it may be worthwhile, therefore, to note finally that both historical and literary sources demonstrate on the contrary that the main wife could be and often was divorced. The

examples of divorce in uxorilocal marriage, such as those of Tadazane

and Moromichi, all involved main wives, and Higekuro's first wife was also his principal spouse until she was replaced by Tamakazura.294

CHILDREN AND THE FAMILY

In the aristocratic society of the Heian period, it is necessary to dis- tinguish carefully between the clan (Uji)295 and the co-residential family or household (ie). The clan at this time was a patrilineal in- stitution whose members bore a common name, worshipped common deities, held certain properties in common, claimed a common de- scent, were frequently buried together, and in some cases occupied hereditary offices and occupations. Its functions were primarily cere- monial and political in nature, and it was related only tangentially to the family life of an individual. The household, on the other hand, was, as we have seen, oriented strongly toward the female line, and it was within its context that the most intimate ties with kinsmen were formed. The physical location of the family belonged usually to the female kin and unless its ownership reverted to the clan chieftain, it passed down as a rule in the matrilineal line. The sons of a family usually left home at a young age to enter an uxorilocal or neolocal marriage elsewhere, leaving their sisters behind to receive husbands in the family house or in a house supplied by the family. Thus it was one of the female siblings usually who formed the link of continuity in the household.

Under these arrangements it was natural perhaps that the primary responsibility for the support and rearing of children should fall upon the mother and her family. This was most clearly the case in duolocal marriages, where the children lived ordinarily with their mother in

This content downloaded from �������������54.84.104.155 on Wed, 02 Jun 2021 18:37:54 UTC�������������

All use subject to https://about.jstor.org/terms

142 WILLIAM H. McCULLOUGH

her own house and were seldom if ever seen in the father's residence.296 But even in neolocal and uxorilocal marriages, where the father's role in the upbringing of his children was undoubtedly much greater than in duolocal marriages, the view that children were mainly the respon-

sibility of the wife and her family seems to have still prevailed. This is suggested in a number of ways, but perhaps most forcefully by the

disposition made of children in the event of divorce and by the prom- inence of the maternal grandparents in the child-rearing process.

Children of divorced parents continued to bear the clan name of their

father, and a son might even become his father's heir, but such chil- dren almost always continued to reside with their mother and to rely

upon her and her family for their support. Koichijoin 41'- iUI (994- io5i), for instance, left his first wife, Fujiwara Enshi WY- (d. 1019) and married a daughter of Michinaga, but his children by Enshi re-

mained with her at her father's Horikawa mansion and were never transferred to Kochijoin's new establishment.297 Another example from about the same time shows a man leaving his wife because of bad relations with his father-in-law and unable to take either his wife or children with him.298

In such cases, however, it was often the maternal grandparents, in- stead of the mother, who actually bore the burden of supporting and raising the children, and even if the dissolution of a marriage was brought about by the death of the wife, her children seem sometimes, at least, to have remained in the maternal home under the care of the grandparents.299 This reflected the important role that maternal grandparents usually played in the upbringing of their grandchildren, a role so large that it often seems to have relegated the parents to a subordinate position. Since children were frequently born in the houses of their maternal grandparents800 and spent all or part of their childhood with them, it was natural that the maternal grandparents should figure prominently in their lives. It was very often the maternal

grandparents who made the arrangements for the births of their daughter's children, commissioning prayers for her safety, securing a

parturition house for her if that were necessary, and otherwise attend- ing to her pre-natal requirements."'i After the delivery it was usually they also who made the arrangements for the first ceremony held to celebrate the birth of a child,302 and they were intimately involved

thereafter in most of'the important events and ceremonies in the child's

This content downloaded from �������������54.84.104.155 on Wed, 02 Jun 2021 18:37:54 UTC�������������

All use subject to https://about.jstor.org/terms

JAPANESE MARRIAGE INSTITUTIONS 143

life. This was in marked contrast with the paternal grandparents, who

were often only remotely concerned with their grandchildren, except on certain ceremonial occasions.M03 Even the parents themselves might

have less to do with a child's upbringing than the maternal grand- parents, since marriages frequently took place at such an early age

that a couple's first child was sometimes born while the husband and wife were little more than children themselves and scarcely capable of looking after their progeny.304

An instance of a birth and another of the death of a child will serve

to illustrate the relative positions occupied by parents and grand-

parents with regard to their children and grandchildren. The birth was that of Seishi, Fujiwara Tadamichi's only child by

his wife Sashi. As we have already seen above (p. 133) ,Tadamichi mar- ried Soshi in 1 1 18 and took up residence with her in her father's Gojo

house. Soshi's father, Munemichi, continued to live in the Gojo house with his daughter and her husband for two years,305 but moved to an- other house in Kujo) sometime in the first half of 11 0o, and in the

seventh month of that same year died there.306 Munemichi's illness coincided with the middle months of his:daughter's first pregnancy,307 but before the illness became too severe he had already started to

make preparations for his grandchild's birth. As the disease pro- gressed and Munemichi was incapacitated, however, it became neces- sary to decide who should take charge of the preparations. Mune- michi's father-in-law, Fujiwara Akisue, asked Tadamichi's father, Tadazane, what should be done, but Tadazane had no opinion, and Akisue referred the question to Munemichi's nephew, Munetada, the author of Chuiyuki. Munetada said that he thought Tadazane should take charge, but we learn that his advice was not accepted and that in the end Tadamichi, the prospective father, assumed responsibility for

the arrangements himself.308 It is apparent from this case that the burden of preparing for a child's birth-supplying all the necessary clothes, implements, and gifts, arranging for a parturition house, and so on-devolved naturally upon the maternal grandfather, but that in

his absence the paternal grandfather or the father could serve.

A similar order of responsibilities can be observed in the funeral arrangements that were made for a five-year-old (by Japanese count) daughter of Fujiwara Muneyoshi who died in 1114. Muneyoshi mar- ried a daughter of Fujiwara Tametaka in 1 i08 and took up residence

This content downloaded from �������������54.84.104.155 on Wed, 02 Jun 2021 18:37:54 UTC�������������

All use subject to https://about.jstor.org/terms

144 WILLIAM H. McCULLOUGH

with her at Tametaka's Shichijo house.309 Several years later Mune-

yoshi and his family moved to a house that had been provided for them

by Tametaka near the Nakamikado residence of Muneyoshi's father, Munetada,310 and it was while he was living there that Muneyoshi's

daughter died.31' In the Chkiyilki entry for the day following her death,

Munetada, her paternal grandfather, notes that he has been forced to

take charge of her funeral arrangements because the father, Mune-

yoshi, is away from the capital taking medicinal baths, and the mater-

nal grandfather, Tametaka, is involved in court religious affairs that

make it necessary for him to avoid the sort of ritual defilement caused by contact with the dead.312 Munetada at this time was actually living next door to Muneyoshi, but it is clear from his own account that it

was only because of the special circumstances described that he be- came involved in the funeral arrangements for his granddaughter, and

that if the maternal grandfather had been able to assume the responsi- bility it would have been his instead.

The literary and historical documents of the Heian period abound

in instances illustrative of the close ties that normally existed between

maternal grandparents and their grandchildren. The intimate scenes

of Genji with his grandchildren by the Lady of Akashi's daughter

come quickly to mind, but one might also cite other well-known cases involving, for instance, Yugiri and his mother's parents, or Kaneie and the children of his daughter Ch6shi.313 The relations between a

child and his paternal grandparents, on the other hand, seem often to have remained very distant, at least until after the child's coming-of- age ceremony had been performed. The evidence suggests in fact that

a child might have been several years old when his paternal grand-

parents met him for the first time. Kagero nikki's failure, for instance, to mention any meetings between Michitsuna and his paternal grand-

father, Morosuke, is probably to be interpreted as meaning that the

two never saw each other, for it is unlikely that Michitsuna's mother would have passed over such an occasion in silence. Although Moro- suke lived for five years after Michitsuna's birth in 955, he and his

grandson appear thus to have remained complete strangers to each

other, and even if some unrecorded meeting between them did occur,

it is obvious that their relationship could not have been a very intimate

one. In the same way, Genji and the children of his son Y-ugiri re- mained unknown to each other for several years, and to the end of

This content downloaded from �������������54.84.104.155 on Wed, 02 Jun 2021 18:37:54 UTC�������������

All use subject to https://about.jstor.org/terms

JAPANESE MARRIAGE INSTITUTIONS 145

Genji's life there is never any significant contact between them. The contrast between a child's relationship with his maternal grandparents on the one hand and his paternal grandparents on the other appears

vividly in the passage where Genji remarks that he has never seen

Yulgiri's children, for as he says this he is fondling a newborn infant of his daughter.314 The infrequency of meetings between children and their paternal grandparents is reflected also in the Heian diaries, where children appear with their paternal grandparents much less

often than with their maternal relatives, and even then usually under special circumstances. We learn, for instance, from Michinaga's diary that he saw the children of his son Norimichi for the first time when

the eldest of them was already two years old,315 and another two years pass before he mentions seeing them again at their joint hakama-

donning ceremony, which was held at his own mansion.316 Paternal grandparents did not always wait so long after the birth of their grand-

children to see them, of course,317 but by way of comparison, one notes that Norimichi's eldest child had lived with her maternal grandparents and presumably had seen them frequently until at least 1015, when the house they were occupying burned.318 Moreover, when Nori- michi's first son was born in ioi8, the year of the hakama-donning ceremony mentioned above, it appears to have been the maternal grandparents who were present and in charge of the birth arrange- ments.319

There may have been occasions when special circumstances led to

close relations between a child and his paternal grandparents, as a few fictional cases from Heian literature would suggest,320 but such occa- sions were sufficiently rare, at any rate, to have left few traces in the historical sources.321 The nature of the Heian household explains

readily why this should have been so. The nucleus of the household in

this period was a mother and her children. In the case of duolocal

marriage, the father did not become a member of the household, since he always remained simply a visitor there, but the mother's own mother was usually included for as long as she lived, and the mother's father and siblings might also be present, according to the type of marriage in which each was involved, or became involved. In the duo- local household, therefore, one might find, in addition to the mother and her children, either or both of the children's maternal grand- parents, their maternal aunts and uncles, their maternal cousins, and

This content downloaded from �������������54.84.104.155 on Wed, 02 Jun 2021 18:37:54 UTC�������������

All use subject to https://about.jstor.org/terms

146 WILLIAM H. McCULLOUGH

possibly even more distantly related maternal kinsmen. In uxorilocal

marriages the household remained essentially the same, except that the father also became a member of it. If a couple's children also mar-

ried uxorilocally, the sons would leave the household to marry else-

where and husbands would be brought in for the daughters. When a

daughter bore children in such a household, the children might thus find themselves living not only with both of their parents and both of

their maternal grandparents, but also possibly with a maternal aunt

and her husband and children, and any of the maternal uncles who

remained unmarried. In neolocal households there were no grand-

parents present, but a married sister of the wife might be included,

together with her husband and children. It is important to note,

moreover, that neolocal households tended usually to become uxori-

local in the generation of the original couple's children. It will have been observed that in each of these household types

paternal kinsmen were almost entirely absent, except that fathers and

sons dwelt together in neolocal and uxorilocal marriages and brothers might continue living together even after their nuptials if they mar-

ried duolocally. Since the clan was patrilineal and marriage did not affect clan membership, this resulted in closely related clansmen liv- ing apart from each other in households that might be composed of members belonging to more than one clan. A father and his son and his son's son all bore the same clan name, but only the father and his son might share a common residence, and that usually only until the son's marriage. In the household, on the other hand, the father and his children might belong to one clan, the mother and her father to

another, the mother's mother to still another, and the husbands of the

daughters and their children to still others.322 The household, in short, was essentially matrilineal and failed to coincide with patri-

lineal clan groupings.

SUMMARY

In the preceding pages an attempt has been made to describe the

marriage institutions of the courtier class at Kyoto during the tenth,

eleventh, and twelfth centuries in terms of the location of the marital residence,, the inheritance of houses and property, the chief regula- tions governing the selection and divorce of spouses, and the nature of the marital household\. The strong uxorilocal and duolocal tendencies

This content downloaded from �������������54.84.104.155 on Wed, 02 Jun 2021 18:37:54 UTC�������������

All use subject to https://about.jstor.org/terms

JAPAN+ESE MARRIAGE INSTITUTIONS 147

observable in the tenth and eleventh centuries are perhaps particularly

noteworthy, since they were undoubtedly connected with the promi- nence of the female line in the inheritance of houses and property, and also with the pronounced matrilineal bias of the household. But neo-

local residence represented the wave of the future, and it became increasingly the norm in the twelfth century. The rules governing

marriage were customary in kind, and they seem to have had little relation to the provisions of the court's legal code, with which they

were often in conflict. Polygyny was freely but prudently practiced, with definite status distinctions among the wives; the choice of princi-

pal wives was restricted to the courtier class, and usually, though not always, to equals within that class; incest restraints were limited to the closest degrees of blood relationship; and divorce was, in form at any rate, easy and uncomplicated.

NOTES

1 Scattered references to one or another aspect of Heian marriage may also be found, e.g., in H. Weipert, "Japanische Familien- und Erbrecht,"MDGNVO, v(1889-1892).83-

140; W. G. Aston, "The Family and Relationships in Ancient Japan," Transactions and Proceedings of the Japan Society, London, II(1892-1893).160-176; Kojiro Iwasaki, Das

japanische Eherecht (Leipzig, 1904); T. Nakajima in the article on Japanese and Korean marriage in Hastings' Encyclopedia of Religion and Ethics; Robert Briffault, The Mothers (New York, 1927), I, 369; Arno Box, Das japanische Familiesystem, Sammlung oriental- istischer Arbeiten, No. 6 (1941); and Andre Gonthier, "Le statut de la femme dans l'em-

pire du Japon," Recueils de la Societe' Jean Bodin tour l'histoire comparative des institutions, XI(1959).295-297. All of these contain the gravest mistakes in fact and interpretation.

Frederic Joilon des Longrais, in his treatment of marriage in Japan in L'est et l'ouest

(Paris and Tokyo, 1958), pp. 304-333, reveals greater understanding of Heian marriage practices than any other Western writer on the subject, but his account is too brief and imprecise to provide a clear picture of the institutions and ceremonies involved.

2 See especially her Shoseikon no kenkyu f4-47cQ0i 6JF (Tokyo, 1953) and Nihon kon'in shi EJ *J,W-fi (Tokyo, 1963).

Nakayama Tara Ni hi i J, Nikon kon'in shi Li * (Tokyo, 1928). 4 Nakagawa Zennosuke +t1 )I|J.Z j, "Kon'inshi gaisetsu" -Lfl ftX, Kazoku

seido zenshui A- f "_ eds. Hozumi Shigeto a and Nakagawa Zen- nosuke (Tokyo, 1937-1938), "Shiron-hen," I, 1-49.

5 Yanagida Kunio 1P9W , Kon'in no hanashi CD -M (Tokyo, 1950). See especially "Mukoiri ko" WA;# pp. 233-304.

6 Ariga Kizaemon I* 2 t3 Nihon kon'in shiron EJ 4tEt r_$ (Tokyo, 1948).

This content downloaded from �������������54.84.104.155 on Wed, 02 Jun 2021 18:37:54 UTC�������������

All use subject to https://about.jstor.org/terms

148 WILLIAM H. McCULLOUGH

7 The History of Human Marriage, 5th ed. (New York, 1922), I, 26. 8 George Murdock's definition of marriage (Social Structure [New York, 1965], p. l),

with its insistence upon common residence, seems too narrow to be useful in dealing

with Heian court society. As will become clear later, residences were frequently separate

in Heian marriage.

9 The older term, CCpatrilocal,"' and its counterpart, "matrilocal," have been largely

replaced in recent sociological and anthropological writings by "virilocal" and "uxori-

local." See Raymond Firth, We, The Tikopia (New York, 1936), p. 596; a letter from

Leonhard Adam in Man, XLVII(1948).12; and Ebman R. Service, Primitive Social Organi-

zation (New York, 1962), p. 30. For a general discussion of residence rules, see Murdock, pp. 16-17 and 201-218.

10 Kathleen Gough describes this residence type in "Nayar: Central Kerala," Matri-

lineal Kinship, eds. David Schneider and Kathleen Gough (Berkeley, 1961), p. 335. 1 The law on marriage (kokonritsu )li 4M4:) in the Yoro Code has been lost, but

numerous scattered quotations from it in Rya no shkuge Hoso shiyo sho Jjj tJx p and elsewhere have been collected by Ishiwara Masaaki iIE flJi (d. 1821) and Takikawa Masajiro gaS)IIi and may be found printed together in Ritsu'itsubun

4A , in Kokushi taikei * rev. ed. (Tokyo, 1929-1964), XXII, 111-120. The household statute (koryoi '*4), which survives in Ryo no shlige (Kokushi taikei, XXIII, 259-342) and Ryo no gige JW (Kokushi taikei, XXII, 91-106), also contains a number of important references to marriage. For a factual account of the marriage pro-

visions of the code and their revisions, see Miura Hiroyuki 5 r-f, Hoseishi no kenkyiu (D gt (Tokyo, 1958), pp. 336-413.

12Hora Tomio inn&& and Tamaki Hajime i , "Kekkon, ren'ai, sei" p j '?, Iwanami koza Nihon rekishi EJ *II El * P (Tokyo, 1962-1964), XXIII, 186-187; Nakayama, pp. 548-549; Takamure (1963), p. 83.

13 Ishii Ryosuke l: I1 /, Hoseishi jll$J A, in Taikei Nihon shi sosho j * Jk,Z*, IV (Tokyo, 1964), 85; Takikawa Masajiro, Nihon hkiseishi ;iA 1)h (Tokyo, 1959), p. 193.

14 The position taken by Takamure (1953), pp. 215ff.

15 Ishimoda Sho j5j*JEiF, "Nara jidai nomin no kon'in keitai ni kan-suru ichiko- satsu" I R rj l-'4$, Rekishigaku kenkyul h

IX (1939), 900-933, 1126-1143. 16 Among hundreds of known examples, it is difficult to identify any non-imperial

Heian marriage as definitely virilocal, but as J. A. Barnes has pointed out, there are

always likely to be exceptions to and infractions of any rule of marital residence ("Mar-

riage and Residential Continuity," American Anthropologist, LXII[1960].853).

17 Duolocal residence appears, for instance, in Ochikubo monogatari XM 4f, where it is apparent that the Kurodo no Shasha ; k1XJ' lives apart from his wife,

the third daughter of the Chiinagon q3t M (Nihon koten bungaku taikei E1 *j h iZ4$kWi [cited hereafter as NKBT], eds. Takagi Ichinosuke iI 1Z et al. [Tokyo, 1957-], XIII, 141; Wilfrid Whitehouse, Ochikubo Monogatari [Kobe, 19351, p.

112). Other instances appear in Ise monogatari # , NKBT, IX, 114; Yamato monogatari *Fq*I , NKBT, IX, 237; and Heichli monogatari N4jj,, NKBT, LXXVII, 62-66.

This content downloaded from �������������54.84.104.155 on Wed, 02 Jun 2021 18:37:54 UTC�������������

All use subject to https://about.jstor.org/terms

JAPANESE MARRIAGE INSTITUTIONS 149

18 Kaneie's birth date has been calculated from his age at death according to Japanese reckoning; it is possible that 991 is the correct date. Most of the birth dates supplied

below have been similarly derived and contain the same margin of error. Kagero nikki has been translated into English by Edward Seidensticker under the title The Gossamer

Years (Tokyo, 1964). As Seidensticker points out (p. 7) the work is not actually a diary (nikki), but a combined autobiography-diary.

19 Seidensticker, p. 60; Kagero nikki, NKBT, xx, 155. 20 Seidensticker, p. 63; Kagero nikki, p. i6o. 21 Seidensticker, pp. 144-145; Kagero nikki, pp. 291-292. 22 Seidensticker, pp. 92, 164; KagerU nikki, pp. 207, 322. 23 See the references to her in Kagero nikki, pp. 119, 122, 153.

24 Hisamatsu Sen'ichi N 1 et al., Nihon bungaku shi El *VORk, rev. ed. (Tokyo, 1964), Chuiko, p. 318.

25 Murdock, p. 1.

26 On Malaya, see Willystine Goodsell, A History of Marriage and the Family, rev. ed. (New York, 1934), p. 25, and Westermarck, I, 38; on New Guinea, Ruth Benedict, Patterns of Culture (New York, 1946), p. 125; on Sumatra and Assam, Westermarck, i, 38-39; on Formosa, George Candidius, A Short Account of Formosa in the Indies, in Awnsham Churchill, A Collection of Voyages and Travels, I (London, 1752), pp. 476-477; and on India, Gough, p. 335.

27 Its existence may be inferred, for instance, from a statement in Eiga monogatari

I4_."j;: "Since the wife of the newly appointed Middle Counselor, Michiyori RRi [971-995], lived at Yamanoi, her husband was called the Middle Counselor of Yamanoi"

(in Nihon koten zensho El J* #, eds. Sasaki Nobutsuna { {[fW3 et al. [Tokyo, 1953-1958], I, 264).

28 Uxorilocal marriage occurs frequently among the primitive peoples of the world, but from the point of view of Japan it may be worthwhile to note especially the uxorilocal practices reported for some of the Formosan aborigines. In a letter dated in August, 1715, collected in Lettres difiantes et curieuses e.crites des missions etrangeres, xviii (Paris,

1781), 441-442, Father de Mailla has the following to say about the aborigines in the northern part of Formosa: "Leurs mariages n'ont rien de si barbare [as their costume]. On n'achete point les femmes comme a la Chine, & on n'a nul egard au bien qu'on peut avoir de part & d'autre, comme il arrive communement en Europe: les peres & les meres [sic] n'y entrent presque pour rien. Lorsqu'un jeune homme veut se marier & qu'il a trouve une fille qui lui agree, il va plusieursjours de suite avec une instrument de musique a sa porte: si la fille en est contente, elle fort & vajoindre celui qui la recherche;

ils conviennent ensemble de leurs articles. Ensuite ils en donnent avis a leurs peres & a leurs meres. Ceux-ci preparent le festin des noces, qui se fait dans la maison de la fille, ou le jeune homme reste sans retourner d'sormais chez son pere. Des-lors le jeune homme regarde la maison de son beau-pere comme la sienne propre, il en est le soutien; & la maison de son propre pere n'est plus a son egard, que ce qu'elle est a l'egard des filles en Europe qui quittent la maison paternelle pour aller demeurer avec leur epoux. Aussi ne mettent-ils point leur bonheur a avoir des enfans [sic] males, ils n'aspirent qu'a avoir des filles, lesquelles leur procurent des gendres qui deviennent l'appui de leur vieillesse." See also Suzuki Tadasu T Taiwan banjinfiuzoku shi j g J<,

This content downloaded from �������������54.84.104.155 on Wed, 02 Jun 2021 18:37:54 UTC�������������

All use subject to https://about.jstor.org/terms

150 WILLIAM H. McCULLOUGH

i6,A~j (Taipei, 1932), pp. 64-67, and Okada Ken j 13 , Mikai shakai ni okeru kazoku X4 UflL t * J h i f - (Tokyo, 1942), pp. 278-290. Joiuon des Longrais (1958), p. 313, says that in Kogury6 (Korea) "l'usage demeura longtemps pour le mari

et la femme d'habiter dans la maison maternelle jusqu'a ce qu'ils aient eleve leurs

enfants."

29 Fujiwara Yorinaga (1 120-1156), Konki t in Shinko Gunsho ruijiu U441 ValMa (Tokyo, 1938-1939), XXII, 754 (the entry for Kyuian, 4th year, 7th month, 3rd day, or, in the abbreviated form that will be used hereafter, Kyulan 4:7:3).

30 Eiga monogatari, I, 241.

31 According to Shkfkaisho 49W40, in Kojitsu sosho , 2nd rev. ed. (Tokyo, 1956-1957), XXII, 400, the mansion occupied a large area south of Tsuchimikado and

west of Ky6goku. It was also known both as the Jotomon _.-_f*P3 mansion and the Ky6goku jjg mansion. Masanobu's ownership of the estate is attested in Shoyiuki

/J4,4jRjj, the diary of Fujiwara Sanesuke X (957-1046), who notes under a date in ioi6 that the former lord of Michinaga's Tsuchimikado residence was Masanobu (in

Shiryo taisei &t4* J;,k [Tokyo, 1934-19441, I-III, Pt. 2, p. 102 [Ch6wa 5:6:2]). Eiga monogatari refers to Masanobu as the "Tsuchimikado minister" (I, 239, 245, 246, etc.)

and mentions the presence of his family at the mansion (e.g., I, 242), indicating that he actually lived there himself. Michinaga's residence at the Tsuchimikado mansion in 991

is established in a diary entry for that year preserved in Ingo sadame burui ki K- VA)t g^N (as quoted in Dai Nihon shiryo }K El Z Vi, Pt. 2, I, 831). The name of the diary in which the entry is found is given as Goshoki /$>jj., which was perhaps an

alternative title for Shkyulki. The latter's author, Sanesuke, was sometimes called the

"Nochi no Ononomiya" `* +J'WX, a name whose orthography might have led naturally to the title of the diary, just as Ononomiya ,'J>;. and Sanesuke's position as udaijin iyC > were the source of the title ShoyFuki 'J' di7 . Present Shoyuzki texts lack entries for the year of the Goshkhi entry, which is dated Sh6ryaku 2:11 :3. Takamure

(1953), p. 405, and Fujiki Kunihiko #S*njZ,,j, Heian jidai no kizoku no seikatsu 2Ti "tL{T3Azii 41 (Tokyo, 1960), p. 209, state the opinion that Michinaga moved into the Tsuchimikado mansion about the time of the birth of Rinshi's first child in late

988, but there does not seem to be any documentary evidence for this view.

32 Minamoto Tsuneyori ,M3 (985-1039) notes in his diary, Sakeiki ZEfTU,, in Shiryo taisei, iv, 15 (Ch6wa 5:3:23), that Michinaga's Takakura j house was

purchased from the widow of Takashina Narita Va. The house was located south of Tsuchimikado and west of Takakura (Shkfkaisko, p. 400) and was thus next door to

Rinshi's Takatsukasa j. PI property (see below, n. 40). Michinaga's Biwa ;Jt{4 and Higashisanj6 J3Eif houses were Fujiwara properties, the first having been the resi-

dence of both Mototsune A (836-89i) and Nakahira 44i2 (875-945), and the latter of Michinaga's father Kaneie. On Mototsune's and Nakahira's residence in the

Biwa mansion, see Fujiwara Morosuke 011*f (908-960), Kyu!reki ,JLf, in Dai Nihon kokiroku X * (Tokyo, 1952- ), p. 1l9 (Tengyo 7:12:11); Fujiwara

Tadahira , ,* (880-949), TeishinkU ki A in Dai Nihon kokiroku, p. 72 (Engi 20:4:23); and Shiikaishk, p. 400. The source of Michinaga's Nijo estate is unknown.

33 See, e.g., Nihon kiryaku EJ ;*, in Kokushi taikei, X-Xi, Pt. 2, p. 173 (Shoryaku 3:4:27); Fujiwara Yukinari ffJ1A (971-1027), Gonki in Shiryo taisei, xxxv, 113 (Choho 2:2:25).

This content downloaded from �������������54.84.104.155 on Wed, 02 Jun 2021 18:37:54 UTC�������������

All use subject to https://about.jstor.org/terms

JAPANESE MARRIAGE INSTITUTIONS 151

34 By the time of the fire, the mansion itself, or a part of it may actually have belonged

to Rinshi's daughter Shashi W (991-1074), who was a consort of Emperor Ichija. This is suggested by a passage describing the fire in Eiga monogatari (Ii, i96), where the

mansion is referred to as Shashi's property (omiya no goryo no miya ZkQ)jifMQJA t). The same passage, however, also implies that Tsuchimikado was still the usual

residence of Michinaga and Rinshi, since it attributes the great loss of movable property

in the fire to the absence of Michinaga, Rinshi, and Shashi from the house on the night

of the disaster. Michinaga probably acquired title to Tsuchimikado as a result of the reconstruction which was carried out under his direction and at his expense in io16-

1017 (see the numerous references to the reconstruction in Michinaga's diary Midi

kampaku ki f-J0Ik pf, in Dai Nihon kokiroku, III, 70-155 [Chowa 5:7:25-Kannin 2:4:17), and also Shoyuzki, Pt. 2, pp. 196-200 [Kannin 2:6:20-29] and Eiga monogatari,

II, 196, 245). Shashi continued to occupy Tsuchimikado for a number of years after the death of Michinaga and then passed the property on to her granddaughter and niece,

Shashi t:i (1026-1105), a consort of Go Reizei (on the transfer of the property, see

Eiga monogatari, iV, 142). Ultimate ownership of Tsuchimikado, however, seems to have

remained in the hands of the Fujiwara clan chieftain after Michinaga, for in 1114 we find

the deed to the mansion in the hands of Tadazane (Fujiwara Munetada , [1032-

1141], Chfuyuki 4f3j-GjjE in Shiryo taisei, VIII-XIV, Pt. 4, p. 315 [Eikyul 2:6:11]). 35 Shoyiuki, Pt. 2, p. 269 (Kannin 3:7:17); Eiga monogatari, II, 259, 265-268.

36 Shoyiuki, Pt. 3, p. 150 (Manju 4:11:25); 152 (Manju 4:12:4). 37 III, 69 (Chowa 5:7:21), notes the burning of the mansion, and there are a multitude

of entries having to do with its reconstruction during the next year. Michinaga was able

to move back into his house in 1017 (III, 156 [Kannin 1:4:27], 169 [Kannin 1:7:4]).

38 I, ii6 (Chaha .2:10o:29), 117 (Choho 2:11 :7) . 39 See, e.g., I, 192 (Kanko 3:9:7); 230 (Kanko5 3:8:14); II, 133 (Chowa 1:1:14); 214

(Chowa 2:4:2).

40 Called the Ichijo-f residence. The earliest mention of Masanobu's living there occurs in Shkyuzki under a date a few months before his death in 993 (in Dai Nihon kokiroku, 3, 257 [Shoryaku 4:1:24]). If the house is actually to be identified with Rinshi's Takatsukasa residence, as Takamure (1953), p. 400, suggests, it was located next door

to the Tsuchimikado mansion (Shkikaishk, p. 400). Even if not next door, however, the two houses would have been close to each other, since they were both in Ichijo.

41 Mido kampaku ki, III, 70 (Chowa 5:7:25).

42 Her eldest son by Munemichi, Nobumichi ff, died in 1120 at the age of 29 by Japanese count (Chyiiii, Pt. 5, p. 256 [Hoan 1:1o:21]), which would place his birth in 1092 or 1093.

43 Munetada mentions attending a music and poetry party at Munemichi's Hachij6 ,Ar residence in 1094 (Chkiyuiki, Pt. 1,p. 141 [Kanji 8:int.3:17]),and we learn from the

preface to some of the poems of Minamoto Shunrai fw# (1055-1129) that Akisue also lived in Hachij6 at one time (Samboku kika shii * in Shinko Gunsho ruiju, Xi, 582, 625-626). Because of Munemichi's marital connection with Akisue, it seems

plausible to conclude that the reference in both cases is to the same Hachijo house.

44 Chizyuki, Pt. 7, p. 262 (Kanji 7:3:26), notes that Akisue and Akimasa moved to the new house together, but makes no mention of Munemichi.

This content downloaded from �������������54.84.104.155 on Wed, 02 Jun 2021 18:37:54 UTC�������������

All use subject to https://about.jstor.org/terms

152 WILLIAM H. McCULLOUGH

45 Chuiyki, Pt. 1, p. 397 (Eicho 1:12:2), and Sompi bummyaku ?tLhIgF, in Kokushi taikei, LVIII-LX, Pt. 2, p. 381.

46 The house eventually became the property of the daughter, as we learn from an

entry in Hyakurensho iff tJ noting its burning in 112o (in Kokushi taikei, XI, 51 [Hoan 1:3:6]), but it was still referred to as Akisue's house as late as 1102 (Chuyiuki, Pt. 2, p. 189 [Kowa 4:6:31). Hyakurensho also refers to Nakazane as "the Takamatsu dainagon" (loc. cit.), suggesting that he continued to live in the house until at least 1115,

the year of his appointment to the office of gondainagon C > (Kugyo bunin ^4bjpp ;Mjj, in Kokushi taikei, LIII-LVI, Pt. 1, p. 380).

47 Chuiyuiki, Pt. 3, p. 332 (Tennin 1:2:25), refers to the Rokujo house as Akisue's

residence in 1097, and numerous references in contemporary sources confirm this for

subsequent years. See, e.g., Fujiwara Atsumitsu t)t (1063-1144), Kakinomoto eigu ki ti$;>{,:F,e, in Shinko Gunsho ruijui, XIII, 53, and Fujiwara Tadazane 07 (1o8- 162), Denreki, , in Dai Nihon kokiroku, Pt. 1, p. 33 (Kowa 2:11:3).

18 Munetada records a visit to Munemichi at his Rokujo residence in 1104 (Chuiyiiki, Pt. 2, p. 362 [Choji 1:7:5]). In i1o6 Munemichi moved to another house on or near

Gojobomon (ibid., Pt. 3, p. 78 [Ch6ji 2:12:19]) and his wife presumably went with him,

since she later came to be known as "the Bomon ;WJjH nun" (Fujiwara Michinori X [d. 116o], Honcho seiki *tJ1k, in Kokushi taikei, IX, 728 [Kytian 6:9:20]).

49 Named Kinnori f He died in 116o at the age of 58 by Japanese count (Kugyo bunin, Pt. 1, p. 447).

50 Chuyuki, Pt. 4, p. 16i (Ten'ei 3:4:13).

51 Chuyiuki, Pt. 4, p. 346 (Eikyui 2:8:27), refers again to Saneyuki's Rokujo residence

in 1114, and he is known to have held utaawase contests there in 1i16 and iii8. On the

latter, see Rokujo saisho no ie no utaawase * r $ Rg r*, in Shinko Gunsho ruiju, viii, 520-523, and Uhyoe no kami Saneyuki no ie no utaawase ti * A in Hagitani Boku VkC;& , Heiancho utaawase taisei ZS fg')(Ji3 (Tokyo, 1957-1965), vi, 1753-1756. The location of the Uhyie no kami . . . utaawase can be

established by comparing Minamoto Shunrai's poem (No. 29) with the same poem as it appears in Shunrai's Samboku kika shui (p. 630), where it is prefaced by the explanation

that it was composed on the occasion of an utaawase contest held at Saneyuki's Rokujo house.

52 Chuiyiuki, Pt. 3, p. 406 (Tennin 1:10:15). The Rokujo establishment would have

been large enough to accommodate both couples, since it seems to have consisted of at

least two distinct residences, one on Higashinotoin and the other on Karasumaru. Cf.

Chuyuiki, Pt. 4, p. 161 (Ten'ei 3:5:13); Kokon chomonjiiu in Kokushi taikei, XIX, 10o; Kakinomoto eigu ki, p. 53; and Denreki, as quoted in Dai Nihon shiryo, Pt. 3, XIII, 115.

53 Minamoto Morotoki OIff iI (1077-1136), Chushkki kaE, in Shiryo taisei, VI- VII, Pt. 1, p. 16o (Gen'ei 2:9:3).

54 Morotoki mentions going to Masasada's Rokujo residence in 1135, referring to the

residence by its alternative name, Nakanoin ql{p (of. cit., Pt. 2, p. 262 [Hoen 1:4:17]). On the identification of Nakanoin with Akisue's Rokujo establishment, see Chuiyiuki, Pt.

3, p. 332 (Tennin 1:2:25). Fujiwara Yorinaga ( 120-1156) refers to Masasada's Rokuj6

residence twice in his diary, once in 1151 and again in 1152 (Taiki f, in Shirya

This content downloaded from �������������54.84.104.155 on Wed, 02 Jun 2021 18:37:54 UTC�������������

All use subject to https://about.jstor.org/terms

JAPANESE MARRIAGE INSTITUTIONS 153

taikan 91 *kk [Tokyo, 1898-1899], pp. 346 [Nimpei 1:2:28] and 351 [Nimpei 2:1:26]).

55 Chiuyiuki, Pt. 4, p. 61 (Ten'ei 2:7:27), notes Akisue's move to the "main [or original] Shichijo house" (moto no Shichijo-tei *-I?1+) in 1 1 1 1. The house is perhaps identi-

fied in this way to distinguish it from the Shichijo house that seems to have been oc-

cupied by Fujiwara Tsunezane fli (1068-1131) and another of Akisue's daughters after their marriage in 1105. On the marriage and residence, see Chizyuiki, Pt. 3, pp. 50

(Choji 2:7:1 1) and 352 (Tennin 1:5:2o), and Denreki, Pt. 2, p. 141 (Kasho 1:6:3). Note

that in the first Chuyuki entry "Tsunesada" u:?t should read "Tsunezane," as refer-

ence to Denreki and the Chulyiuki text given in Dai Nihon shiryo, Pt. 3, xvI, 128, will show.

Tsunesada (iioo-1156) was Tsunezane's son by an earlier wife. Evidence of Akisue's

residence at the Shichijo house is found in the prefaces to some of his own poems. See,

e.g., Rokujo suri no daibu no shu * i {I0kA;, in Shinko Gunsho ruiju, xi,

286, 290. 56 Op. cit., p. 17.

57 Nanreishi *jA-, in Nihon zuihitsu taisei EI tfRki, Series 1 (Tokyo, 1927-1928), IX, 322.

58 See below, pp. 118-123, 141-146.

59 Review of Shoseikon no kenkyiu, SZ, LXII(1953).685.

60 Nakagawa, p. 30; Nakayama, p. 655; Ikeda Kikan E1Jl , Heiancho no seikatsu

to bungaku 2-'Z Q)4Li < C* (Tokyo, 1964), p. 101. 61 Op. cit., p. 219, n. 1.

62 On the frequency of the appearance of the word mukotori in Heian texts, see Takamure (1953), pp. 388ff.

6i See, e.g., the marriage of Fujiwara Morozane to the adopted daughter of Fujiwara Nobuie (below, p. 130).

64 Murdock, pp. 59 and 194. 65 Nihon kon'in shiron, p. 8. 66 "Matrilocality and Patrilineality in Munduruciu Society," American Anthropologist

LVIII(1956) .414-434.

67 Edwin M. Loeb, Sumatra: Its History and People (Vienna, 1935), pp. 220-221. 68 Vladimir Jl'ich Jochelson, The Yukaghir and the Yukaghirized Tungus, [Publications

of the Jesup North Pacific Expedition, xiii, Pt. 1 (Leiden and New York, 191O) ], p. 92. 69 Edwin M. Loeb and Jan 0. M. Broek, "Social Organization and the Longhouse in

Southeast Asia," American Anthropologist, XLIX(1947).417.

70 Loeb and Broek, p. 420.

71 A. L. Kroeber and T. T. Waterman, "Yurok Marriages," in A. L. Kroeber, The Nature of Culture (Chicago, 1952), pp. 250-251, report that about one-quarter of the

marriages among the patrilineal Yurok in 1gog were uxorilocal. 72 Op. cit., p. 330. 73 For numerous examples of the later practice, see Yanagida's "Mukoiri ko" and

Ariga, pp. 9-1o. 74 Eiga monogatari, ii, 87-88.

75 An interesting parallel to this is found among the patrilineal, uxorilocal Atjeh

people of Sumatra. According to Loeb, p. 237, virilocal marriage takes place among the

This content downloaded from �������������54.84.104.155 on Wed, 02 Jun 2021 18:37:54 UTC�������������

All use subject to https://about.jstor.org/terms

154 WILLIAM H. McCULLOUGH

Achehnese only when the woman's family is very inferior in rank or social position to

that of the man.

76 Genji monogatari, NKBT, xIv-xvlll, Pt. 4, p. 434; Arthur Waley, The Tale of Genji

(New York, 1935), p. 872.

77 Genji monogatari, Pt. 4, p. 81 (Waley does not translate this maki).

78 Genji monogatari, Pt. 4, p. 375; Waley, p. 837. 79 Kanemasa's marriage to the orphaned daughter of Toshikage. On the marriage and

residence, see Utsubo monogatari, NKBT, x-xII, Pt. 1, pp. 96-101. 80 Ochikubo's marriage to Michiyori may be included in this category, since her

mother was dead and she was treated as an orphan by her family. On the residence, see

Ochikubo monogatari, pp. 119-121, 124, 139; Whitehouse, pp. 86-89, 93, 110. 81 See, for example, the story of the udoneri who steals the dainagon's daughter in

Yamato monogatari, pp. 326-327.

82 Sarashina nikki, NKBT, xx, 503. 83 "Haizumi," in Tsutsumi chiunagon monogatari 4-" r NKBT, xiii,

418-419.

84 Taira Sadaie 5E (fl. 1058-1062), Koheiki * =, in Shinko Gunsho ruiju, XIX, 569 (Kahei 2:3:20); Eiga monogatari, IV, 170.

85 Hyakurensho, p. 26 (Kohei 2:4:13).

86 On the construction of the house, see Fujiwara Koremichi f (1093-1165), Taikai hishoz -k @ ;,# in Shinko Gunsho ruiju, XXI, 413. On the date of the move and the source of the land, see Hikunsho JIJ{4t, as cited in Oshima Takeyoshi k;kMjt (fl. 1704), Yamashiro meisho shi IIjJA,tAj=, in Kyoto sosho At:* (Kyoto, 1914-1917), I, 127, and Hyakurensho, p. 28 (Kohei 6:7:3).

87 Kugyo bunin, Pt. 1, p. 312. 88 Taikai hisho, p. 413, says that Morozane built the house with income he received

from his first year as naidaijin N ,k a post to which he was appointed in 1o6o. A similar case of a husband providing the residence for himself and his wife after the

burning of his father-in-law's house is probably to be found in the marriage of Fujiwara Norimichi and a daughter of Fujiwara Kinta. See below, notes 299 and 318.

89 On the marriage and residence, see Eiga monogatari, I, 272. Since Michitaka's

daughter left the residence after she and Atsumichi were divorced (Okagami, p. 179), and since Atsumichi used the same place in his marriage to Naritoki's daughter, it seems

reasonable to infer that Atsumichi provided the residence for the marriage. Notice also that Naritoki's daughter finally left Atsumichi to return to her father's Koichijo house

(Okagami, p. loo; Izumi shikibu nikki, NKB[, xx, 445). 90 Eiga monogatari, II, 102. 91 The type of residence involved in this marriage is suggested in Kagero nikki, when

the author notes the transfer of Takaakira's wife to her own home after the exile of her husband and the burning of his house (p. 179; Seidensticker, p. 75).

92 Eiga monogatari, III, 249-250.

93 Among many examples of such marriages, a few typical cases may be cited: 1) Fuji-

wara Nagaie O (1005-1064) and a daughter of Fujiwara Yukinari, who married when Nagaie was 12 and Yukinari's daughter was 1o or 1i; 2) Fujiwara Tsunezane and a daughter of Fujiwara Kinzane 4 (1053-1107), who married when Tsunezane was 13

This content downloaded from �������������54.84.104.155 on Wed, 02 Jun 2021 18:37:54 UTC�������������

All use subject to https://about.jstor.org/terms

JAPANESE MARRIAGE INSTITUTIONS 155

or 14 and Kinzane's daughter was i o or 11; and 3) a certain kurodo no taifu 2Ji kjq and a daughter of Taira Nobunori f (l112-1187), who married when the kurodo no taifu was 13 or 14 and Nobunori's daughter was 11 or 12. On the marriage and ages of Nagaie and Yukinari's daughter, see Eiga monogatari, III, 241-242; Mido kampaku ki, i, 157 (Kanka 2:8:2o); and Sakeiki, p. 65 (Kannin 2:3:13). On the second marriage and the

ages of the husband and wife, see Fujiwara Tamefusa 14 (1049-1115), Tamefusa-kyo ki g1l-"=, as quoted in Koji ruien f (Tokyo, 1931-1936), XLI, 244. On the kurodo no taifu and Nobunori's daughter, see Taira Nobunori, Heihanki JE;- , in Shiryo taisei, xv-XVIII, Pt. 3, p. 19 (Hagen 3:2:9).

94 As in the case of Yorimichi's marriage to his principal wife, Takahime 3 (995- 1087), a daughter of the imperial prince Tomohira 2Ji (964-1009). See above, p. 130. Yorimichi was about 17 and Takahime about 14 at the time of their marriage.

95 An instance of this type of residence ("uxori-neolocal") is seen in the marriage of

Fujiwara Muneyoshi ';bb (1085-1170) and a daughter of Fujiwara Tametaka (1070-1130). See below, p. 143. Muneyoshi was about 29 at the time of the marriage.

96 See the case of Fujiwara Akisue's daughters described above, pp. log-iio. 97 An example of neolocal residence for a younger daughter is found in the marriage of

Fujiwara Munetada's second daughter to Minamoto Noritoshi { (d. 1149). Mune-

tada's eldest daughter married her uncle Munesuke 7gX (1077-1162) in 1ogg and lived with him uxorilocally; at first, presumably, at her father's Karasumaru , .yJ1 house in Gojo, and then, after that place burned in 1103, at his newly constructed Nakamikado

rpI1JP9 house. The second daughter married Noritoshi in 1114 and lived with him neolocally in a house that her father had reconstructed for them near Horikawa and

Gojabamon. On Munesuke's marriage and residence, see Chuzyiuki, Pt. 2, pp. 140 (Jotoku 2:12:16), 143 (Jotoku 2:12:28), 298 (Kowa 5:11:16); Pt. 3, pp. 17 (Choji 2:2:20), 20 (Choji 2:2:28), 6i (Choji 2:10:3), 104 (Kasha 1:3:10); and Pt. 4, p. lo (Ten'ei 2:1:19). On the marriage and residence of Noritoshi, see Chuzyulki, Pt. 4, pp. 338 (Eikyui 2:8:10), 340 (Eikyul 2:8:14), 382 (Eikyui 2:12:13).

98 Murdock, p. i6.

99 In addition to the examples mentioned below and in the preceding discussion, note also the numerous deeds of sale from the Heian period that show female ownership of houses in the capital. See, e.g., Takeuchi Riza 3 ed, Heian ibun " (Tokyo, 1947-1960), Komonjo-hen, Nos. 955, 1245, 1294, 1823, and 1832.

100 As in the marriage of Fujiwara Morozane and Fujiwara Reishi J+f (1040-1114). After Morozane had become clan chieftain and inherited the clan houses, he gave Reishi

the deeds to his two principal residences, Kayanoin -10 K and Kyagoku. See Chui- yflki, Pt. 4, p. 315 (Eikyul 2:6:11).

101 Pp. 730-738.

102 Sha2kaish1, p. 401.

103 Kokon chomonjif, p. 241.

104 Shakaisha, p. 401.

105 Sanesuke's residence at Ononomiya is attested in many sources: see, e.g., Sakeiki, p. 336 (Chogen 5:4:4).

106 Sompi bummyaku, Pt. 2, p. 4, lists three daughters of Saneyori, two of whom (the consorts of Murakami and Suzaku) are known to have died before their long-lived father. The third cannot be identified.

This content downloaded from �������������54.84.104.155 on Wed, 02 Jun 2021 18:37:54 UTC�������������

All use subject to https://about.jstor.org/terms

156 WILLIAM H. McCULLOUGH

107 Pt. 3, p. 233 (Chogen 4:2:9). For other indications of Kaneyori's residence at the Ononomiya house, see Pt. 3, pp. 243 (Chogen 4:3:4) and 307 (Chogen 5:10:11).

108 Pt. 3, p. 252 (Chagen 4:3:24).

109 Okagami, ed. Matsumura Hiroji t t# t,4 Iwanami bunko (Tokyo, 1964), p. 67. One must note, however, that this information is lacking in the shorter 0Okagami texts,

which are thought to represent the work's oldest form. Cf. the Okagami edition in NKBT

(the edition cited hereafter unless otherwise specified), XXI, 89.

110 Shoyiuki, in Shiryo taisei (the edition cited hereafter), Pt. 2, p. 303 (Kannin 3:12:9). This must have occurred before Chifuru's marriage, since the entry describing Sane- suke's disposition of his property occurs under a date in lo2o, when Kaneyori would have been only five or six years old.

111 A detailed statement of the line of inheritance of the Ononomiya house is found in

Fujiwara Teika Zg (1162-1241), Meigetsuki EJ! )]Re (Tokyo, 1g11-1912), I, 432 (Genkyui 2:7:29).

112 Chulyuki, Pt. 1, p. 115 (Chosh6 3:10:21), calls her Sukeie's wife.The identification

of Sukeie's wife as a daughter of Sanesuke in Choshilki, Pt. 2, p. 224 (Chosho 3:10:22), is undoubtedly mistaken, since it would leave Sukeie's residence at Ononomiya unex- plained. It would also make a father of Sanesuke when he was nearly eighty years old.

113 Chufyuki, Pt. 2, p. 224 (Chosho 3:10:22).

114 See, e.g., Chutyiuki, Pt. 1, p. 15 (Kanji 2:7:28), and Minamoto Tsunenobu g1r? (1016-1097), Sotsuki jWipe,, in Shiryo taisei, v, 150 (Kanji 2:8:1).

115 Eiga monogatari, iv, 251, and Chuzyulki, Pt. 3, p. 155 (Kasho 1:12:13). 116 Sompi bummyaku, Pt. 1, p. 216, lists Sukeie's daughter as the mother of Yoshi-

zane's son Tadayori 4T,. 117 Ima kagami, pp. 225, 228; Sompi bummyaku, Pt. 1, p. 216. 118 p. 52 (Hoan 2:3:10).

119 Chuzyuki, Pt. 7, p. 115 (Chosho 3:10:21), gives her age at death as ninety-nine by Japanese reckoning.

120 Meigetsuki, I, 432 (Genkyiu 2:7:29). 121 According to Sompi bummyaku, Pt. 3, pp. 486-487, she was the mother of Moro-

yori's son Moromitsu RfilrJ. Taiki, p. 345 (Nimpei 1:2.:23), also identifies Moromitsu as Yoshizane's maternal grandson.

122 Sompi bummyaku, Pt. 3, p. 486. 123 P. 66 (Ten'y6 1:12:11).

124 Meigetsuki, I, 432 (Genkyiu 2:7:29). Nakayama Tadachika ,eg (1131-1195), Sankaiki g1ij@,, in Shiryo taisei, xix-xxi, Pt. 2, p. 28 (Nin'an 2:3:21), and Hyaku- rensho, p. 8o (Nin'an 2:3:21), also identify the house as Moromitsu's, and Ima kagami,

p. 288, says that he was called "the Ononomiya jiju f4 fX,." 125 The house was in the hands of Moromitsu's son Tomochika AV when Teika

became involved in its affairs in 1205 and recorded the line of inheritance in his diary (loc. cit.). Two years later Teika referred to Tomochika as "the Ononomiya shosho /J% " (Pt. 2, p. i6 [Jogen 1:3:28]).

126 Eiga monogatari, I, 285.

127 Mido kampaku ki, II, 117 (Kanka 8:8:23). 128 Eiga monogatari, I, 274.

This content downloaded from �������������54.84.104.155 on Wed, 02 Jun 2021 18:37:54 UTC�������������

All use subject to https://about.jstor.org/terms

JAPANESE MARRIAGE INSTITUTIONS 157

129 Gonki, Pt. 1, p. 102 (Choho 2:1:1). 130 Shoyuki, Pt. 1, p. 182 (Kanko 2:3:18); Eiga monogatari, I, 242. 131 Eiga monogatari, I, 153.

132 Mido kampaku ki, Pt. 1, p. 172 (Kanko 3:1:1). 133 Meigetsuki, I, 432 (Genkyui 2:7:29); Eiga monogatari, I, 251. 134 Sompi bummyaku, Pt. 2, p. 4. 135 Ibid., Pt. 1, p. 216.

136 Oimikado }kIt1fjI, Nijakarasumaru Z ,f,IL, Yamanoi, Takamatsu, and others. See Takamure (1953), pp. 738-749. 137 Sanenobu B -, (964-1001) and Tadanobu ;44 (967-1035). 138 1, 266-267.

139 Sanesuke and his daughterChifuru have already been mentioned. Another historical

case involved Fujiwara Akimitsu 3IM (944-1021) and his daughters (see Eiga mono- gatari, II, 253-254, 290). A fictional case comes from Genji monogatari, where Hitachi no kami (Ukifune's stepfather) is said to have settled all of his lands and possessions ont a daughter (Pt. 5, p. 141; Waley, p. 963). 140 Pt. 1, p. 253.

141 Eiga monogatari, II, 195.

142 pp. 44, 209; Whitehouse, pp. 2-3, 193.

143 See, for example, the "Tsutsuizutsu" no54 rfi story in Ise monogatari, IX, 126 (translated in Frits Vos, A Study of Ise-Monogatari [The Hague, 1957), I, 187-189), and the reason given by the Kurodo no Sh6sh6 in Ochikubo monogatari for abandoning his wife, the third daughter of the Chuinagon (pp. 140-141; Whitehouse, 111-112). 144 NKB[, XXII-XxvI, Pt. 5, pp. 220-223.

145 Of interest also from this point of view is the Konjaku monogatari story (Pt. 4, pp. 445-447) of the man who abandons one of his two wives because her silkworms die and then returns to her when she begins to prosper again.

146 II, 87.

147 pp. 151, 155; Whitehouse, pp. 124, 128.

148 In Zoku Gunsho ruijiu MO (Tokyo, 1923-1930), XVIII, 256. 149 Miura, p. 337.

150 Takamure (1953), pp. 276-280. 151 Takamure (1963), p. 132.

152 Her name is given sometimes as Junshi IIjff7j (Kugyo bunin, Pt. 1, p. 177, and

Fujiwara Michisue JM-* [15th c.], Honcho koin shoun roku $!M Jf j , in Shinko Gunsho ruijui, III, 425) and also as Keishi f (Okagami uragaki Jk_j*, NKBT, XXI, 325).

153 The date is calculated from Saneyori's age in 931,which was 32 by Japanese count (Kugyo bunin, Pt. 1, p. 177). On his parentage, see Okagami, p. 85, and Kugyo bunin, loc. cit.

154 His birth date may be calculated similarly (Kugyo bunin, Pt. 1, p. 180); this tallies

with the unidentified birth mentioned in Tadahira's diary, Teishinko ki, p. 12 (Engi 8:12:17). His mother is identified in Okagami, p. 115, and Kugya bunin, loc. cit.

155 Teishinko ki, p. 70 (Engi 20:2:3). 116 Pp. 65, 68, 73, 77, 89, 94-100, 114.

This content downloaded from �������������54.84.104.155 on Wed, 02 Jun 2021 18:37:54 UTC�������������

All use subject to https://about.jstor.org/terms

158 WILLIAM H. McCULLOUGH

157 Pp. 91, 99. She is also referred to as Tadahira's kitanokata in Ki no Tsurayuki shku

WA 1, in Shinko Gunsho ruijuz, XI, 358. 158 Kitanokata literally means something like "the person on the north side," and its

meaning of "principal wife" is presumed to have arisen from the customary residence of

a man's main wife in the northern wing of the house they occupied together. Even in a

work as early as Genji monogatari, however, the literal significance of the word appears

to be secondary to its meaning of "principal wife." Yuigiri, for instance, speaks of his mother Aoi as haha kita no kata (Pt. 3, p. 277), although Genji and Aoi never lived to-

gether and Aoi, consequently, never had the opportunity to occupy the northern wing of Genji's residence. A survey of the other eighty or so uses of kitanokata in Genji mono-

gatari, and of the word's usage in other Heian works, shows that when the status of the

woman to whom it refers can be identified, she is, almost without exception, the principal

wife of her husband, just as, indeed, dictionaries define the term.

159 Saneyori became sessho 4fik and daijo daijin ki&,k J (Kugyo bunin, Pt. 1, p. 2og), whereas Morosuke's highest office was udaijin (Kugyo bunin, Pt. 1, p. 198). For Saneyori's appointment as clan chieftain, see Honcho seiki, as quoted from Yanagihara-

ke kiroku 5 in Dai Nihon shiryo, Pt. 1, IX, 428. Morosuke never became clan chieftain, a post held normally by the highest ranking clan member and from the

time of Kaneie always by the Fujiwara who was sessho or kampaku j 3, or who held nairan IAJ powers.

160 The date of the coming-of-age ceremony appears in Nihon kiryaku, Pt. 2, p. 18

(Engi 15:1:20), and the date of Saneyori's marriage is established by the birth of his

eldest son, Atsutoshi VVC (918-947), in 918 or 919. The latter date is calculated from Atsutoshi's age at death, which was thirty by Japanese count. Atsutoshi's death date

may be found in Teishinku ki, p. 250 (Tenryaku 1:11:17); the age is given both in Yotsutsuji Yoshinari j4L+j-jj (1329-1402), Kakaisho jifj , in Motoori Toyokai *J", Kimura Masakoto *"MB, and Inoue Yorikuni 4JIZ?# , eds., Kokubun chiushaku zensho gIJil 4 (Tokyo, 1907-1910), p. 357, and also in Gensei t and Koshi 3;t?, Chokusen sakusha burui t (14th c.), in Yamagishi Tokuhei iii g ed., Hachidaishi zenchuf iA/l;, Sf (Tokyo, 1960), III, 161. Sompi bummyaku, Pt. 2, p. 1, gives Atsutoshi's age at death as thirty-six, which

must be wrong, since it would make Saneyori only twelve years old at Atsutoshi's birth.

161 Okagami, p. 91; Sompi bummyaku, Pt. 2, pp. 1-4. 162 Kugyo bunin, Pt. 1, p. 216.

163 Nihon kiryaku, Pt. 2, p. 33 (Shohei 2:1). 164 Nihon kiryaku records the death of a Minamoto wife of Saneyori in 936 (Pt. 2, p.

35 [Shohei 6:4]) and the deaths of two other wives, Fujiwara Noshi ft:T and Fujiwara Ninzenshi f-f, in 964 (Pt. 2, p. 92 [Koho 1:4:11 and an unspecified day in the same month]).

165 Kugyo bunin, Pt. 1, pp. 211, 214, 233. 166 Sompi bummyaku, Pt. 1, pp. 51-55. 167 Kugyo bunin, Pt. 1, pp. 199, 206, 205.

168 Morosuke's coming-of-age ceremony occurred presumably in 923, when he was appointed to the Lower Junior Fifth Rank (Kugyo bunin, Pt. 1, p. 18o), and it must have

been at this time also that he took Seishi as his wife, since it was just about ten months later that Koremasa was born (TIeishinku ki, p. 88 [Encho 2:7:15]).

This content downloaded from �������������54.84.104.155 on Wed, 02 Jun 2021 18:37:54 UTC�������������

All use subject to https://about.jstor.org/terms

JAPANESE MARRIAGE INSTITUTIONS 159

169 Gashi's second son, Tamemitsu t3 , was born in 942, the year before Seishi's death (Kugyo bunin, Pt. 1, p. 21o).

170 Kugyo bunin, Pt. 1, p. 233. 171 Ibid., p. 241.

172 Ibid.

173 Okagami, pp. 175, 203; Okagamz uragaki, p. 375.

174 Calculating from the date of Michitaka's birth (953) and Kaneie's age in 968 (40 yrs.). For the relevant dates and figures, see Kugyo bunin, Pt. 1, pp. 205, 222.

175 Kaneie's marriage to the author of Kagero nikki began in 954, the year following Michitaka's birth.

176 Kugyo bunin, Pt. 1, p. 269.

177 According to Eiga monogatari, I, 185, Meishi's father was Minamoto Takaakira, a son of Emperor Daigo, and she had been adopted by the imperial prince Moriakira

iF1RI (928-986). 178 Sompi bummyaku, Pt. 1, pp. 60-63 (note the corrections in the upper margins). 179 Michinaga married Rinshi early in 988 (see above, p. io8) and Meishi later in the

same year (Eiga monogatari, I, 245-246).

180 Yorimichi and Meishi's son, Yorimune it' ' (992-1065), were born in the same year, but Yorimichi is called Michinaga's eldest son in Kugyo bunin, Pt. 1, p. 252, and is listed as such in Sompi bummyaku, Pt. 1, p. 59.

181 Kugyo bunin, Pt. 1, p. 265.

182 Shoyuiki, Pt. 1, p. 283 (Kanko 9:6:29); Pt. 2, p. 39 (Chowa 4:12:13). 183 Kugyo bunin, Pt. 2, pp. 317, 322. 184 Ibid., p. 333.

185 See above p. 114.

186 Her lack of children is noted in Eiga monogatari, II, 178, and III, 142, and in Jien 8IIN1 (1155-1225), Gukans_ho , in Kokushi taikei, xix, llo. Sompi bummyaku, Pt. 1, p. 6o, mistakenly lists Takahime as the mother of Yorimichi's daughter Kanshi, but Chlzyiuki, Pt. 5, p. 320 (Daiji 2:8:14), says that Kanshi's mother was Kishi (see

below). Kanshi was adopted by Takahime (Koen 1 g [d. 1169], Fuso ryakki 4Ai;* PE, in Kokushi taikei, XII, 291 [Eisho 5:12:21]), as was Genshi 4)-f- (1016-1039), a daughter of Prince Atsuyasu , (999-log) by Takahime's sister (Eiga monogatari, II, 232), and also perhaps another of Kishi's children, the priest Kakuen X (1031- 1098), whose true parentage is found in Chulyiuki, Pt. 2, p. 86 (Jotoku 2:4:16), but is

called Takahime's son in Jimon kUso ki ?; P91 Ajf{ (13th c.), in Zoku Gunsho ruiju, XXVII, jo, p. 6i.

187 In addition to the wife mentioned below, Yorimichi had a son by a woman who was

in service with him, a daughter of Minamoto Norisada ,t (d. 1017). See Eiga mono- gatari, III, 142-143. Her son, Michifusa A X (1025-1044), would have been Yori- michi's heir if he had not died at an early age (Fuso ryakki, p. 287 [Kantoku 1:4:27]).

Yorimichi also had a child by a daughter of Fujiwara Nagayori, but the mother died in

childbirth (Shoyuzki, Pt. 2, p. 32 [Ch6wa 4:11:17]) and nothing more is known of the child.

188 Shoyuiki, Pt. 2, p. 32 (Chowa 4:10:15).

189 Eiga monogatari, II, 177-184. 190 Gukansho, p. 110.

This content downloaded from �������������54.84.104.155 on Wed, 02 Jun 2021 18:37:54 UTC�������������

All use subject to https://about.jstor.org/terms

160 WILLIAM H. McCULLOUGH

191 Mido kampaku ki, Pt. 2, pp. 62, 116, 205; Pt. 3, p. 121. 192 See, e.g., Chufyiuki, Pt. 1, p. 8 (Kanji 1:11:22). 193 Hyakurensho, p. 22 (Kantoku 2:7:4) notes the burning of "the Chancellor's [i.e.,

Yorimichi's] Takakura residence" in 1045. Takahime may also have spent some time

with Yorimichi at his own Kayanoin residence. See Eiga monogatari, IV, 73. 194 Kugyo bunin, Pt. 1, p. 333.

195 The diary of his son Moromichi 1ii (1062-1099), Gonijo Moromichi ki W PE (in Dai Nihon kokiroku), makes this abundantly clear. See, e.g., Vol. III, p. 19 (Kanji 7:2:11). See also above, p. 130, and Chuiyiuki, Pt. 1, p. 226 (Kaho 2:1:10).

196 Moromichi succeeded his father as kampaku and clan chieftain in 1094 (Kugyo bunin, Pt. 1, p. 356).

197 See, e.g., Gonijo Moromichi ki, Pt. 2, p. 320 (Kanji 6:12:22); Denreki, as quoted in Dai Nihon shiryo, Pt. 3, XV, 106; Chkuyulki, Pt. 4, p. 283 (Eikyti 2:3:26); and Tamefusa- kyo ki, as quoted in Dai Nihon shiryo, Pt. 3, II, 279.

190 Minamoto Toshifusa 1>M (1035-1121), Suisaki 4Z jei, in Shiryo taisei, IV, 92 (Shoryaku 4:5:24); Tendai zasu ki J in Zoku Gunsho ruijul, IV, 593.

Morozane is also said to have had a son by a daughter of Taira Yukichika Q, but this supposed son, who later bore the priestly name Kakujitsu ' (1052-1093), was born when Morozane would have been only nine or ten years old, and the relatively late sources in which the relationship is found stated must be in error. For the attribu-

tion of parentage, see Sompi bummyaku, Pt. 1, p. 61, and Shomonzeki fu iP Yia, (17th c.?), in Shinko Gunsho ruijiu, III, 522. For the dates and figures used in calculating

the relative birth dates of Morozane and Kakujitsu, see Sompi bummyaku, loc. cit.; Gonijo Moromichiki,Pt. 2, p. 319 (Kanji 6:12:18); Chkuyiuki mokuroku FfjNj ,S in Shiryo taisei, XIV, 322 (Kowa 1:6:28); Honcho seiki, pp. 306-307 (K6wa 1:6:28); and Kugyo bunin, Pt. 1, p. 361. Note also that Eiga monogatari, iv, 165, calls Morozane's son by Sadanari's daughter his eldest, though this son was born five years after Kakujitsu. Moreover, when Kakujitsu died in 1093 Morozane's son Moromichi mourned for him not as a brother, but as a nephew (Gonijo Moromichi ki, Pt. 2, pp. 319 [Kanji 2:12:18] and 321 [Kanji 2:12:23]).

199 This was Ningen f: (1057-1108), the Enryakuji abbot. 200 Eiga monogatari, IV, 165.

201 Takamure (1953), pp. 520-524.

202 Eiga monogatari, IV, 165.

203 In his case, actually an adopted daughter. According to Chkiyiuki, Pt. 1, p. 116

(Kanji 8:1:10), her true father was Fujiwara Tsunesuke ,lf$ (ioo6-1o81). 204 Eiga monogatari, IV, 239.

205 According to Tamefusa-kyo ki, as quoted in Dai Nihon shiryo, Pt. 3, pp. ll, 838, Tadazane was born in the early part of 1079.

206 lemasa *R (1080-1115), Morimichi's son by the daughter of Yoshitsuna, was thirty-six by Japanese count at his death in 1115 (Kugyo bunin, Pt. 1, p. 380). letaka

Zf (d. 1125), Moromichi's son by Sadatsune's daughter, had his coming-of-age cere- mony in 1098 (Chkiyulki, Pt. 1, p. 403 [Eicho 1:12:26]), and if he was of a normal age for the ceremony at that time, this would place his birth between 1083 and 1087. The latter date is perhaps the more likely, since both his father and his brother Tadamichi had

This content downloaded from �������������54.84.104.155 on Wed, 02 Jun 2021 18:37:54 UTC�������������

All use subject to https://about.jstor.org/terms

JAPANESE MARRIAGE INSTITUTIONS 161

their coming-of-age ceremonies when they were eleven by Japanese count (on Moro-

michi's ceremony, see Kugyo bunin, Pt. 1, p. 336; on Tadazane's, see Goniju Moromichi ki, Pt. 1, p. 179 [Kanji 2:1:21]). Moromichi's daughter by the daughter of Yoritsuna was

ten by Japanese count at her death in 1107 (Chufyiuki, Pt. 3, p. 244 [Kasho 2:8:131), which would place her birth in 1097 or 1098.

207 Yorinaga reports in his diary that in 1146 Tadazane told him, "The late Nijo Lord

[Moromichi] left Lady Ichijo [Zenshi] and became the son-in-law of the Kujo Chancellor [Nobunaga]" (Taiki, p. 167 [Ten'y6 2:12:24]).

208 None of them is even mentioned in Moromichi's diary. Sompi bummyaku (Pt. 1,

p. 408; Pt. 2, p. 173) calls the daughters of Yoshitsuna and Sadatsune ladies in waiting (nyobo) to Moromichi and notes that Yoshitsuna's daughter was the wife of Minamoto

Morotada gIiJ, (1054-1114). Yoritsuna's daughter was also possibly one of Moro-

michi's ladies in waiting, since the mother of her brother Nakamasa l4ljiIE-who was thus perhaps the girl's mother too-was a lady in waiting to Reishi, Moromichi's mother

(Gonijo Moromichi ki, Pt. 2, p. 242 [Kanji 6:4:16]).

209 They lived first at Toshiie's Higashinotoin I R[ house, then for a brief period

at Morozane's Kazan'in V Q house, and then at another of Toshiie's houses in Shijobomon (Suisaki,v,56 [Joho2:l0:26], lo [Shoryaku4:8:15], 150 [Shoryaku5:9:28]).

210 Tadazane became clan chieftain and received nairan powers in the year of his

father's death (Kugyo bunin, Pt. 1, p. 361). 211 Eiga monogatari, iv, 252, is explicit on this point.

212 The passage cited previously from Taiki (n. 207) makes it plain that a divorce was involved. Diarists of the time continued to refer to Zenshi as Moromichi's wife for

perhaps ten years after he had left her, but they eventually ceased to do so as the fact of

the divorce became undeniable. Compare, e.g., Chufyulki, Pt. 1, p. 69 (Kanji 6:2:3) and Pt. 4, pp. 232-233 (Ten'ei 3:12:10). Reference to Denreki, as quoted in Dai Nihon shiryo, Pt. 3, XIII, 383, shows that the title "Ichijodono" used in the later Chuiyiuki entry refers to Zenshi.

213 Eiga monogatari, iv, 252.

214 She does not appear in Eiga monogatari, which covers the main line of the Fujiwara

family in some detail up to about 1092, and the first definite reference to her as Moro- michi's wife occurs in 1094 (Chufyulki, Pt. 1, pp. 115-116 [Kanji 8:1:1o]).

215 Chufyulki, loc. cit.

216 He seems to have moved into the Nijo residence with her in 1094 (Chufyulki, loc. cit.),

but she may not have remained there permanently, since Moromichi speaks of her subse-

quently as "Kujo" AL11X and "Kujo no ue" lI k, presumably in reference to her father's house (GonijoMoromichi ki, III, 219 [Eicho 1:9:11] and 229 [Eicho 1: 10:17]). Moromichi later moved to a new house in Kujo (Chiuyiuki, Pt. 2, p. 66 [Jotoku 1:12:2]), where

Nobunaga's daughter apparently also lived with him, since he consistently refers to her thereafter as kitanokata (Gonijo Moromichi ki, III, 240 [Eicho 1:12:3], 262 [Kowa 1:3:1 ],

264 [Kowa 1:3:8], 276 [Kowa 1:4:21], 282 [K6wa 1:5:14]).

217 Chulyulki, Pt. 2, p. 325 (Choji 1:1:23), mentions that "the kitanomandokoro of the late Lord Nijo" held a Buddhist service for him in 1104; the Chufyfuki usages mentioned

earlier make it necessary to interpret this as a reference to Nobunaga's daughter. More- over, on the anniversary of Moromichi's death in 11o8 "the wife of the late hakuriku

This content downloaded from �������������54.84.104.155 on Wed, 02 Jun 2021 18:37:54 UTC�������������

All use subject to https://about.jstor.org/terms

162 WILLIAM H. McCULLOUGH

PA R" (Moromichi) made a vow to copy out the Lotus Sutra as an offering on behalf of her deceased husband (Go totokunagon gammonshui rf4iL _ CZ, as quoted in Dai Nihon shiryo, Pt. 3, X, 190-191). Although Tadazane notes the anniversary in his diary, he does not mention the vow, as he would surely have done if the wife mentioned

in it had been his mother, Zenshi. See Denreki, Pt. 2, p. 295 (Tennin 1:6:28).

218 Goni& Moromichi ki, I, 255 (Kanji 3:1:29); Chufyulki, Pt. 1, p. 23 (Kanji 3:1:29). 219 Ninshi was eighty-seven by Japanese count when she died in 1152 (Heihanki, xv-

xviii, Pt. 1, p. 1o8 [Nimpei 2:3:211). The kind of disparity in ages seen here in the marriage of Ninshi and Tadazane was not uncommon in Heian aristocratic marriages and

is paralleled in other societies where there are natural or artificial restrictions on the

field of marriage candidates. See Robert H. Lowie, Social Organization (New York, 1948),

pp. 94-95. 220 Evidence of their living together may be found in ChuFyulki, Pt. 1, pp. 185-186

(Kanji 8:10:3), 262 (Kaho 2:5:11), 403 (Eicho 1:12:24); Chulgaishoz rp'44p (remarks and reminiscences of Tadazane recorded by Nakahara Moromoto i1jj [1109-1175]),

in Zoku Gunsho ruiji, XI, 884; and Kojidan t -$, in Kokushi taikei, XVIII, 42. 221 In Nihon koten zensho, pp. 194, 292.

222 Ninshi bore two girls (Chulyfuki, Pt. 1, pp. 185-186 [Kanji 8:10:3]; Gonijo Moro- michi ki, III, 246 [Eicho 1:12:24]) and a son (Tamefusa-kyo ki, entry for Kaho 2:7:7, as

quoted in Dai Nihon shiryo, Pt. 3, III, 836). Both of the daughters died a few days after

birth (Chkiyiuki, Pt. 1, p. 187 [Kanji 8:10:8]; Pt. 2, pp. 7-8 [Jotoku 1:1:20]). The son lived for at least six or seven months (Gonijo Moromichi ki, III, 231 [Eicho 1 :1o:26]), but

there is no trace of him after 1096 and he must be presumed to have died.

223 Ima kagami, p. 194, attributes the dissolution of the marriage to Ninshi's lack of

children.

224 Chkfyiiki, Pt. 1, p. 262 (Kaho 2:5:11). 225 Taishi ;-m (1095-1156), Toba's empress. Her age at death is given as 61 in

Taiki, p. 451 (Kyuju 2:12i:16).

226 Denreki, I, 117 (Kowa 4:4:4). 227 Chkuyiuki, Pt. 2, p. 215 (Kowa 4:9:25).

228Ima kagami, pp. 194-195. Eiga monogatari, IV, 228-229, says she served Joto-

mon 'in, i.e., Fujiwara Shoshi V+ (988-1074), but this must be a mistake, since Shishi would have been only three or four years old by the time of Shoshi's death. Shishi's son

by Shirakawa was the priest Kakuho !Ei (1o09-11s3). 229 Loc. cit.

230 Loc. cit.

231 Evidenced frequently in contemporary sources. See, e.g., Denreki, I, 236 (Kowa

5:9:15), and the entries for Ten'ei 1:4:3 and 24, as quoted in Dai Nihon Shiryo, Pt. 3, XI, 108-109.

232 Denreki, entry for Tennin 2.:12:21, as quoted in Dai Nihon shiryo, Pt. 3, X, 762.

233 Her son, Tadamichi e !,fi (1097-1164), succeeded Tadazane as kampaku and clan chieftain in 1121 (Kugyoi bunin, Pt. 1, p. 386).

234 Pt. , p. 131. Note, however, that one text of Sompi bummyaku reads sai a here instead of sho (see the editor's marginal note).

235 Denreki, entry for Ten'ei 2:11:9, in Dai Nihon shiryo, Pt. 3, XII, 137. 236 Ibid.

This content downloaded from �������������54.84.104.155 on Wedf:ffff on Thu, 01 Jan 1976 12:34:56 UTC

All use subject to https://about.jstor.org/terms

JAPANESE MARRIAGE INSTITUTIONS 163

237 Chiuyulki, Pt. 4, p. 333 (Eikyfi 2:8:3); Denreki, entries for Eikyul 2:12:3 and Eikyul

3:2:3o, in Dai Nihon shiryo, Pt. 3, XVI, 2o, 186.

238 Sompi bummyaku, Pt. 1, p. 84. 239 She cannot be traced in contemporary sources.

240 The marriage is described in detail in Chiuyiuki, Pt. 5, pp. 68 (Gen'ei 1:8:8), 86-87

(Gen'ei 1:10:26), 89 (Gen'ei 1:11:2).

241 Chiuyiuki, Pt. 5, pp. 97 (Gen'ei 1:11:26), 104 (Gen'ei 2:2:6), 116-117 (Gen'ei

2:3:9), 123 (Gen'ei 2:4:3), 127 (Gen'ei 2:4:15), 128 (Gen'ei 2:4:19), 197 (Ho)an 1:1: 18); Heihanki, Pt. 1, p. 28 (Kyfian 5:10:16).

242 Seishi .-7' (1122-1182), Sutoku's empress. 243 This term is used of her in Konki, p. 761 (Kyuaan 4:7:20); Heihanki, Pt. 2, p. 7

(Kyuju 2:9:15); and Honcho seiki, p. 704 (Kyulan 5:10:26). She is called kitanokata in

Ima kagami, p. 2o6.

244 Heihanki, Pt. 1, p. 28 (Kyuian 5:10:16).

245 This was Motofusa _ (1145-1230). Shunshi also had other sons by Tadamichi. 246 Kujo Kanezane ? (1149-1207).

247 Sompi bummyaku, Pt. 1, p. 85, also lists the priest Sonchul :,* as Tadamichi's son and says that his mother was a daughter of Minamoto Moritsune f , but nothing

is known about Sonchui, his mother, or her father.

248 Kujo Kanezane, Gyokuyo 3jV (Tokyo, 1906-1907),II, 174 (Jisho 2:8:22); Heihan- ki, Pt. 2, p. 8o (Kyuju 3:2:10); Sompi bummyaku, Pt. 2, p. 466, and Pt. 3, p. 529.

249 Pt. 2, p. 466.

250 Whitehouse, p. 96; Ochikubo monogatari, p. 128. 251 An example of the exception is found in the marriages of FujiwaraTsunezane, who

appears to have had four wives simultaneously. See Takamure (1953), pp. 507-509.

252 As in Michinaga's marriage to Meishi. His visits to her Konoe j house are mentioned frequently in Midi kampaku ki. See, e.g., ii, 1O (Kanko 7:7:19), 133 (Chawa

1:1 :16); III, 48 (Chowa 5:2:28), 129 (Kannin 1:12:14). At the beginning of the marriage he visited her also at the Higashisanjo palace, where she had been adopted by his sister

(Eiga monogatari, I, 245). There is no suggestion anywhere that they ever lived together.

253 There were no restrictions on the remarriage of women, as in later centuries. In

Genji monogatari, e.g., note the remarriage of Ochiba to Yuigiri after her husband's death. Takamure (1953), p. 131, seems to look upon this as a kind of polyandry, but it is

not, of course. Simultaneous marriage of a woman to two or more men did not occur in

Heian society.

254 Waley, p. 32; Genji monogatari, Pt. 1, p. 77. 255 The stepmother of the author of Sarashina nikki. 256 Pt. 2, p. 412; Waley, p. 490.

257 Pt. 3, p. 85; Waley, p. 554.

258 Pt. 3, p. go; Waley, p. 556. Takamure (1953), pp. 532-533, cites the case of Kanemasa in Utsubo monogatari (Pt. 2, p. 464) as an example of marriage between half-

siblings, and on the basis of this she says that such marriages were not prohibited in the Heian period. Even if she is correct in contending, against the opinion of most com-

mentators, that the woman in question was actually Kanemasa's half-sister, the marriage

must be regarded as very exceptional. Corroborative historical examples of such relation-

This content downloaded from �������������54.84.104.155 on Wed, 02 Jun 2021 18:37:54 UTC�������������

All use subject to https://about.jstor.org/terms

164 WILLIAM H. McCULLOUGH

ships do not occur after the beginning of the tenth century, and the evidence of Genji monogatari indicates plainly that they were disapproved socially. Kujo Kanezane con-

demned them even when one of the siblings was an adopted child (Gyokuyo, I, 172

[Shoran 1:12:14]). 259 Pt. 1, pp. 205-206; Waley, pp. 95-96. 260 See, e.g., Pt. 3, pp. 47, 384; Pt. 4, p. 184; Waley, pp. 528, 566, 732. 261 Eiga monogatari, I, 267-268.

262 Sompi bummyaku, Pt. 1, pp. 54-59; Eiga monogatari, I, 235. Kanezane approves this type of marriage in Gyokuyo, I, 172 (Shoan 1 :12:14).

263 Sompi bummyaku, Pt. 1, p. 384; Eiga monogatari, II, 244. 264 Sompi bummyaku, Pt. 1, pp. 146, 148, 177.

265 Sompi bummyaku, Pt. 1, p. 59; Pt. 3, p. 462; Honcho ko-in shiun roku, pp. 426, 428;

Koin keizu & j+J, in Zoku Gunsho ruijuz, v, jo, p. 34. 266 Sompi bummyaku, Pt. 1, p. 47.

267 Sompi bummyaku, Pt. 1, pp. 49-50, 58; Honcho ko-in shoun roku, p. 428. 268 Sompi bummyaku, Pt. 1, pp. 56, 58; Pt. 3, p. 470. In Genji monogatari the Nija no

Otodo (To no Chujo) addresses some remarks to his mother that are usually interpreted as a condemnation of first-cousin marriage of this kind (Pt. 2, p. 293; Waley, pp. 412- 413). The passage is somewhat vague, however, and it may refer not to the blood rela-

tionship of Kumoi and Yulgiri-the couple in question-but to the manner in which they were raised together in the same house (cf. the use of the key term, yukarimutsubi, in Pt. 5, p. 308). Kumoi and Yiigiri are, after all, finally married with the blessing of the Nijo no Otodo and there is no hint of scandal or criticism. In any case, the historical cases of marriage to a maternal uncle's daughter are overwhelming in number.

269 Sompi bummyaku, Pt. 1, pp. 207, 250. 270 For a description of the earlier system and its revisions, see Miura, pp. 349-384. 271 Marriage outside the court circle perhaps never occurred among the higher ranks.

The only example I know of appears in Utsubo monogatari (Pt. 1, pp. 187-188), where a miserly imperial son marries the daughter of a rich merchant. Waley's translation of

Genji monogatari reads at one point (p. 1064), ". . . countless persons in much the same

position as Kaoru himself [a high-ranking court official] had kept girls of the peasant

class," but the text (Pt. 5, p. 308) speaks only of ayashiki musume, "girls of humble

position" (cf. the use of ayashiki in Pt. 1, p. 195, where it is used of Murasaki's grand-

mother). However, a directive of the daijokan k t issued in 874, which continued an earlier proscription of marriage between provincial governors and daughters of local

families, suggests that such alliances may have been a fairly common thing among pro- vincial governors during their tours of duty. For the directive, see Ruiju7 sandai kyaku

laJ5j{-1fi3l;, in Kokushi taikei, xxv, 303. 272 On Tsunekuni's appointments, see Kugyo bunin, Pt. 1, pp. 199, 205; Sompi

bummyaku, Pt. 2, p. 440; and Okagami, p. 132. On Arisada, see Sandai jitsuroku =ft t, in Kokushi taikei, IV, 322 (Jogan 15:3:26).

273 Kugyo bunin, Pt. 1, p. 222; Sompi bummyaku, Pt. 2, p. 297. 274 Kugyo bunin, Pt. 1, p. 147.

275 p. 218. Waley, p. 622, n. 1, says much the same thing. It is perhaps worth noting that there is no reason to suppose with Waley and Morris that Murasaki was not Genji's

This content downloaded from �������������54.84.104.155 on Wed, 02 Jun 2021 18:37:54 UTC�������������

All use subject to https://about.jstor.org/terms

JAPANESE MARRIAGE INSTITUTIONS 165

main wife after Aoi's death. She possessed every discernible attribute of the main wife in

the tenth century and is the only one of Genji's numerous ladies other than Aoi who is

called his kitanokata (Pt. 3, pp. 127, 197). On the use of the latter term in the novel, see

above, n. 158. Waley's assertion that Murasaki was "technically only a chief concubine,

her mother's low rank making it impossible for her to be a kitanokata or 'legitimate con-

sort,' " is mystifying in view of the text's use of this very title to designate Murasaki.

276 Genji monogatari, I, 189. The father of Murasaki's mother was a provincial official

only in the sense that he held the title of azechi V f a purely nominal post asso- ciated with Mutsu and Dewa that was filled by officials holding the office of counselor

(nagon) or above. It would be entirely misleading to lump holders of this title together

with the ordinary provincial official class. It may be noted, moreover, that Genji's own

mother was also the daughter of an azechi dainagon.

277 On the marriage, see Kugyo bunin, Pt. 1, p. 242, and Sompi bummyaku, Pt. 3, p.

462. For the identification of Takaakira's wives, see also Tamai Kosuke 3itJI-, honomine shosho monogatari $i ; I4/T3 (Tokyo, 1960), pp. 132-135.

278 See, e.g., Hisamatsu, Chulko, p. 318. 279 Sompi bummyaku, Pt. 1, p. 58; Pt. 3, pp. 462, 470. 280 Okagami, p. 125; Sompi bummyaku, Pt. 1, p. 59.The second wife is called kitanokata

and okitanokata in Kagero nikki, p. 179, and Eiga monogatari, I, 184-185.

281 Her son Toshikata became gondainagon (Kugyo bunin, Pt. 1, p. 265); the later

wife's son, Tsunefusa , (969-1023), became gonchiunagon (ibid., p. 262). 282 For a discussion of these, see Nakayama, pp. 608-617, and Miura, pp. 398-413. 283 Seidensticker, pp. 34-35; Kagero nikki, pp. log-113. 284 Seidensticker, pp. 35-36; Kagero nikki, pp. 113-114. 285 Seidensticker, pp. 39ff.; Kagero nikki, pp. 119ff. 286 Seidensticker, pp. 42ff.; Kagero nikki, pp. 124ff. 287 Seidensticker, p. 44; Kagero nikki, p. 128. 288 Seidensticker, p. 144; Kagero nikki, p. 291. 289 The story occurs only in the longer texts of Okagami. See Sato Kyu i t

Okagami shokai *kfW (Tokyo, 1940), pp. 356-361. 290 pp. 131-133 (the marriages of Fujiwara Tadazane and Fujiwara Moromichi). 291 See above, n. 89.

292 Pt. 3, pp. 123-136; Waley, pp. 569-577. 293 Morris, p. 221.

294 She is referred to as Higekuro's kitanokata until she is replaced by Tamakazura (Genji monogatari, Pt. 3, pp. 128, 131, etc.), and then as his "former" (hajime no) kitanokata (ibid., p. 322).

295 I follow the customary terminology here. Sociologists might prefer "sib" as a rendering of uji.

296 See the case of Kaneie's son by the author of Kagero nikki.

297 Eiga monogatari, II, 105, 188, 232, 254-255.

298 Shoyiuki, Pt. 3, p. 67 (Manju 2:8:24). The husband was Fujiwara Sadayori tj (995-1045); his father-in-law was Minamoto Narimasa j3W& (d. 1041).

299 As apparently happened, for instance, when Norimichi's first wife, a daughter of

Fujiwara Kinto 5/ff (966-1o41),died. See Eiga monogatari, II, 151-152, 172-173; III,

This content downloaded from �������������54.84.104.155 on Wed, 02 Jun 2021 18:37:54 UTC�������������

All use subject to https://about.jstor.org/terms

166 WILLIAM H. McCULLOUGH

103-113. For the location of Norimichi's marital residence at his father-in-law's Shijo

house, see Shoyiuki, Pt. 1, p. 423 (Chowa 4:4:13). At some undetermined time after their

residence burned in 1015, the grandmother and the children seem to have moved in with

Norimichi at his Nijo mansion (Eiga monogatari, III, 156, 232-233). Another case of

maternal grandparents raising a deceased daughter's children involved Koichijoin's

second marriage to a daughter of Michinaga. See Eiga monogatari, III, 172-175; IV, 72.

One does not find, on the other hand, cases in which the father took charge of his

children by himself while the maternal grandparents were alive and able to provide

support for them.

300 Sometimes the birth took place in a house borrowed from a maternal relative,

family retainer, or other connection. See, e.g., the case of Fujiwara Sanesuke's wife as

recorded in Shoyuzki, Pt. 1, pp. 59-60 (Eikan 3:3:1-2), 66-67 (Eikan 3:4:28). The use of

a borrowed house for parturition was undoubtedly in order to avoid ritual defilement of

the regular residence. See Chluyiiki, Pt. 1, pp. 153-154 (Kaho 1:59:20, 27, 29), and Eiga

monogatari, II, 172.

301 The maternal grandparents' role described here is seen, e.g., in the birth of Fuji-

wara Norimichi's daughter in 1014 (Eiga monogatari, II, 167, 172), and Fujiwara

Tadazane's daughter in 1097 (Goniji Moromichi ki, III, 246 [Eicho 1:12 :24]).

302 The ceremony on the third night. Examples of the maternal grandparents' role in

this ceremony occur in the births of children of Fujiwara Norimichi (cited above), Fujiwara Tadamichi (Chiuyiuki, Pt. 5, p. 250 [Hoan 1:9:1]), and Fujiwara Michinaga

(Eiga monogatari, I, 244-245).

303 The paternal grandparents seem frequently to have been in charge of the ceremony

that occurred on the fifth night after a child's birth (see the instances cited in the pre-

ceding note). Michinaga, moreover, held the hakama-donning ceremony (hakamagi *

;1) for two of his paternal granddaughters at his own house (Midi kampaku ki, III,

i85-186 [Kannin 2:11:91). 304 Eiga monogatari, II, 167, depicts Norimichi as being so childlike that he was dis-

gruntled by his wife's pregnancy.

305 Chiuyulki, Pt. 5, p. 197 (Hoan 1:1:18), shows him still there in 1120. 306 Chulyulki, Pt. 5, pp. 240 (Hoan 1:7:2), 244-245 (Hoan 1:7:22). 307 The chronology and course of the illness are described in Chulyiuki, Pt. 5, pp. 244-

245 (Hoan 1:7:22).

308 These events can be traced in Chulyiuki, Pt. 5, pp. 242 (Hoan 1:7:13) and 246 (Haan 1:7:25).

309 Chulyiuki, Pt. 3, pp. 343 (Tennin 1:4:8), 344 (Tennin 1:4:14). 310 Chiuyulki, Pt. 4, p. 343 (Eikyl 2:8:22).

311 Chayaki, Pt. 4, p. 350 (Eikyul 2:9:11). Munetada reveals here that the girl herself had been staying at her wet nurse's house for some time before her death.

312 Chuyiuki, Pt. 4, p. 351 (Eikyti 2:9:12).Tametaka was in charge of arrangements for a forthcoming imperial visit to the Iwashimizu and Kamo shrines. See Chuiyiiki, Pt. 4,

p. 321 (Eikyul 2:6:29). 313 See, e.g., Eiga monogatari, I, pp. 236ff. Almost any Heian tale or diary will yield

additional examples.

314 Genji monogatari, Pt. 3, p. 283; Waley, p. 640.

This content downloaded from �������������54.84.104.155 on Wed, 02 Jun 2021 18:37:54 UTC�������������

All use subject to https://about.jstor.org/terms

JAPANESE MARRIAGE INSTITUTIONS 167

315 Midi kampaku ki, iII, 66 (Ch6wa 5:6:27) . 316 Mido kampaku ki, III, 185-186 (Kannin 2 :11 :9). 317 When a daughter was born to Fujiwara Tadazane in 1097, the paternal great

grandfather and the paternal grandparents visited the house where the child was born

and saw it about two weeks after its birth (Chiuyuiki, Pt. 1, p. 403 [Eicha 1:12:24]; Pt. 2,

p. 4 [Jotoku 1:1:9]).

315 See above, n. 299. On Norimichi's marriage, see Midi kampaku ki, II, 151 (Chowa

1:4:27). It is not certain where Norimichi and his family lived after 1015, but it was

clearly not with his father, since Michinaga notes that the children involved in the

hakama-donning of 1i18 came to his house for the ceremony (loc. cit.), and Shoyuzki

records their return to their own home on the following day (Pt. 2, p. 216 [Kannin

2:11 :1o]). Norimichi had perhaps already moved to his Nijo house, where he is known

to have lived after his wife's death in 1023 (see, e.g., Eiga monogatari, III, 113, 232-233).

319 Midi kampaku ki, iII, 189 (Kannin 2:12:24); Shoyiuki, Pt. 2, pp. 230 (Kannin

2:12:24), 232 (Kannin 2:12:29), 238 (Kannin 3:2:13).

320 Two instances come from Ochikubo monogatari and Genji monogatari. In the first,

a youthful husband secretly establishes his wife in a neolocal marriage against the wishes

of her parents, and the couple's second son is placed in the care of the husband's mother.

The youth of the couple, the lack of any maternal relatives to help look after their

children, and the fact that the couple's house belonged to the husband's mother prob-

ably account for the role played by the paternal grandparents here. In Genji monogatari,

Kumoi is brought up by her paternal grandparents in order to avoid conflict with her

mother's second husband and his numerous children.

321 The adoption of Fujiwara Tadazane by his paternal grandfather Morozane is an

apparent historical instance, but Sompi bummyaku, Pt. 1, p. 64, notes that the adoption

occurred because of the early death of Tadazane's father Moromichi, and this means that

Tadazane was a fully grown man at the time. The date of Tadazane's adoption implied in

Sompi bummyaku is probably correct, since Moromichi's diary, GonijF Moromichi ki, contains no record of the event. Fujiwara Saneyori adopted several of his patrilineal

grandchildren (Eiga monogatari, i, 158-159; Shoyuki, Pt. 2, pp. 172-173 [Kannin

2:4:1]), but the circumstances involved are not fully known. 322 One of the effects produced by the wife's continued membership in her father's

clan was that she was buried apart from her husband if his clan was different from her

own. For instance, Rinshi, a Minamoto, was buried in the neighborhood of her father's

grave at Koryuji , (Chuigaisho, p. 868; Okagami uragaki, p. 380), while her hus-

band, Michinaga, was buried on the opposite side of the city at Toribeno ,

(Shoyiuki, Pt. 3, p. 156 [Manju 4:12:7]). Similarly, although Minamoto Reishi was adopted into the Fujiwara clan (see above, p. 130), she was eventually interred with the

Minamoto at the clan burial grounds in Kitashirakawa 4t I JII (Chiuyiuki, Pt. 4, p. 298 [Eikyii 2:4:22]; Heihanki, Pt. 1, pp. 217-218 [Kyuju 2:5:20]), while her husband

Morozane was buried in the area of the Fujiwara burial grounds near Uji (Denreki, Pt.

2, p. 45 [Kowa 3:2:21]).

This content downloaded from �������������54.84.104.155 on Wed, 02 Jun 2021 18:37:54 UTC�������������

All use subject to https://about.jstor.org/terms