13 Epilogue

Infanticide in the Shadows of the Modern State

e fi nal chapter of habitual infanticide in Japan makes for a story full of surprises and contradictions. Aft er the Meiji Restoration of 1868, infanticide briefl y stood at the center of attention but then largely dropped from view. Abortion was designated a crime, but annual convictions remained in the hundreds at a time when late-term abortions and infanticides alone numbered in the tens of thousands. Imperial Japan had a rapidly growing population and an increasingly modern state apparatus, but contrary to its reputation, before the 1930s the modern state was less pronatalist than many domains of the Tokugawa system. Although it concerned itself far less with the reproductive lives of its subjects, the national government nonetheless presided over a surge in fertility and the fi nal retreat of infanticide, decades aft er the last targeted countermeasures had been fashioned and abandoned.

It is true that in the fi rst Meiji years, some opponents of infanticide justifi ed their policies in pronatalist terms, and one proclamation even urged that “the way of propagating the population should be spread.”1 A few years later, however, the disappearance of all such schemes left the new nation-state without anything comparable to the childrearing subsidies and pregnancy surveillance of the late Tokugawa period. Imperial Japan had its panegyrists of population growth and motherhood.

Its government may have portrayed large families as a service to nation and emperor.2 Beyond rhetoric, however, Imperial Japan did not invest in pronatalist policies between 1880 and 1937.3 Infanticide lost its prominence in public discourse, producing little more than reminiscences, piecemeal private philanthropic eff orts, and a few more votive tablets with baby-killers shedding their humanity.4

Amid this new lack of attention, infanticide persisted for several decades. Its scale was now smaller than during the eighteenth century but remained none-208

17/04/13 3:54 PM

Epilogue 209

theless vast. Th

en, over the 1910s and 1920s, a time with no notable new eradication policies, the number of infanticides fell by nearly nine-tenths and fertility rates rose to new heights. During the 1930s, infanticide was no longer a signifi cant demographic factor, but even as Imperial Japan marched into total war, hundreds and probably thousands of newborns died at the hands of their parents every year.

Pronatalist policies returned during the war. Th

eir target was no longer infanti-

cide, but securing the welfare of mothers and children and encouraging women to

“give birth and multiply” ( umeyo fuyaseyo) the emperor’s subjects. Modern contraception had made some inroads before the war, as had a movement for decriminalizing abortion.5 While it was being fought, World War II halted these trends, but its aft ermath released their full potential. In the new political climate of the late 1940s, birth control advocates secured the eff ective legalization of abortion. Th

e new law triggered a boom in what was now called the “artifi cial interruption of pregnancies.” In the 1950s, therefore, infanticide as a family planning tool was fi nally becoming history in Japan.6

I N FA N T I C I D E I N T H E 1 8 7 0 s

e infanticide prevention schemes of the early 1870s did not disappear because Japan’s newborns were now universally safe from their parents. Looking for the traces of infanticide is relatively diffi

cult for these years, but not impossible. Th

household registers, introduced in 1872, cannot be probed for their secrets, locked away as they are out of privacy concerns.7 We therefore have to draw our tentative conclusions from published statistical tables. According to these fi gures, many districts reported sex ratios at birth between 110 and 130 in the decade aft er 1873

(Map 13). If we assume that for every three girls, two boys were killed at birth, a sex ratio of 110 implies that 10 percent of births were followed by infanticides; the proportion rises to 43 percent for a sex ratio of 130.8 Infanticides were probably more frequent at this time than we can derive from skewed sex ratios alone. An inconspicuous sex ratio does not prove the absence of infanticide; girls and boys may simply have been killed in roughly equal numbers. Beginning in 1886, wildly implausible stillbirth rates are on record even for districts whose sex ratios of live births would have given no cause for suspicion.

It is possible that some areas experienced a resurgence rather than a persistence of infanticide. With the disappearance of childrearing subsidies, poor families were now once again left to their own devices. Even the symbolic importance of childrearing may have suff ered, since both the child-protection policies and the constant stream of exhortations ceased. Although the diff erent format of the data—village samples versus comprehensive district-level fi gures—makes this a tentative comparison, sex ratios at birth seem to have risen in many parts of Eastern Japan aft er the collapse of the old regimes.

17/04/13 3:54 PM

map 13. Sex ratio of births by district, 1873–1882. (sources: Based on numbers in Naimushō and Naikaku Tōkeikyoku, * Kokusei chōsa izen. * Up to 1878, the number of births is reconstructed from age structure data; from 1879, it is based on reported births.)

17/04/13 3:54 PM

Epilogue 211

e geography of infanticide between the 1870s and the 1890s nevertheless suggests that decades of moral suasion, charity, and surveillance during the Edo period had a lasting eff ect. It is notable that the greatest sex ratio distortions in the 1870s are mostly found in areas that had no eff ective population policy during the Tokugawa period. Many generations of Tosa lords banned infanticide, but none introduced countermeasures that represented a real commitment of resources.

Around 1880, each of the seven districts of the former domain had sex ratios at birth above 110. In the eighteenth century, Mimasaka province (today, the north of Okayama prefecture) was more notorious for infanticide than its southern neighbors, Bizen and Bitchū. Around 1880, however, aft er decades of countermeasures in Mimasaka, the most distorted sex ratios now clustered in the southern reaches of Okayama prefecture. In the provinces of Tango, Inaba, Hōki, and Izumo on the Japan Sea coast, initiatives against infanticide were small and came late in the Edo period. Around 1880, the region was among the more dangerous parts of Japan for baby girls. In the Northeast, some areas with a long history of childrearing policies, such as southern Akita, also had many missing girls. Th e distortions were

even greater, however, in the northern half of Akita and some of the lands once ruled by the Tsugaru and Nanbu clans, where no such policies existed during the Edo period.

T H E F O R M A L C R I M I NA L I Z AT I O N O F A B O RT I O N

During the early-Meiji fl urry of anti-infanticide policies, infanticide was made punishable as homicide and abortion was designated a crime.9 Although the law code of 1871 punished abortions only if the pregnant woman died as a result of the procedure, the Ministry of Justice replied to inquiries by local offi cials in 1871 and

1872 by ordering a varying number of cane blows for men and women found guilty of providing or procuring an abortion. In 1873, a preliminary new law code stated that abortions would be punished with one hundred days of imprisonment with hard labor.10 Under this law, several hundred people were convicted of abortions every year.

During the same years in which the infanticide prevention schemes withered away, offi

cials at the Ministry of Justice were busy draft ing a new criminal code.

Published in 1880, it went into eff ect in 1882. Like its European models, it included a statute on abortion. Th

is law has received much attention as part of a supposed eff ort of the Meiji government to control the reproduction of its subjects.11 Perhaps the most remarkable feature of the law, however, is how lenient it was by the standards that mattered to its draft ers—the legal systems of the contemporary West—

and how indiff erently it was enforced in subsequent decades.

During the second half of the 1870s, the Ministry of Justice had called on the French legal expert, Gustave Emile Boissonade de Fontarabie, to assist their

212 Epilogue

deliberations on the new law code. Boissonade opened the discussion on how to punish abortion by observing that since Japan had a large population, the punishments could be a little lighter than for example in sparsely settled Russia. If the Japanese offi

cials disagreed with Boissonade’s assessment of the demographic situation, they left no record of their dissent. Given how much the fear and reality of depopulation had animated the infanticide debates only a few decades earlier, this complacency is a monument to the recovery of fertility in the early nineteenth century, thanks to which local populations were now growing in nearly every part of Japan.

With the abolition of the domains, local headcounts may also have mattered much less to the offi

cials of the new age, for whom the size of the entire nation would have been the most pertinent number. Whatever their thoughts on Japan’s demographic situation, the offi

cials at the Ministry of Justice espoused a remark-

ably empathetic view of abortion. Since abortion was a widespread custom, they noted, people did not realize that it was a serious crime. Th is could easily have

been an argument for using the law to reeducate the populace. Instead, the ministry offi

cials saw in the deep popular entrenchment of abortion a reason to avoid severe punishments. Many people who committed abortions, they observed, had no other choice. Th

e ministry offi

cials even had doubts about punishing a husband

harshly for his part in an abortion that caused his wife’s death; since abortions were oft en motivated by poverty, sentencing the widower to a long prison term would leave his already pitiable children in still greater diffi culty.12

Based on such views, the 1880 criminal code stipulated penalties that were lighter than those in most European countries. Th

e German Empire punished

aborting mothers with six months to fi ve years in prison; those who had accepted pay for acting as an accessory to an abortion could be locked up for as many as ten years.13 In Russia, women who had procured an abortion could be exiled to Siberia. Elsewhere in Europe and North America, prison sentences ranged from six months to life. By contrast, Japan’s new penal law prescribed one to six months of prison for women who had abortions, and one to three years of imprisonment for abortionists whose clients died as a result of their procedures.14

Abortion providers and their clients remained reasonably secure even from these punishments. In the late nineteenth century, Japanese women had tens of thousands of abortions every year.15 Th

e annual number of convictions peaked at

712 persons in 1884 and then declined to 381 in 1900, 343 in 1910, 160 in 1920, and a mere 52 in 1927.16 Sentences for abortion were oft en more lenient than the law stipulated. Th

is remained true even aft er the revised penal code of 1907 doubled the maximum prison term for the woman to one year.17 Between 1895 and 1926, the average sentence length was a little over fi ve months in prison, with only 6 percent longer than one year.18 From 1926 to 1930, nearly 83 percent of abortion sentences put the convict on probation.19 Th

is gulf between practice and prosecution was

broadly in line with the situation in Europe. In France, many commentators

Epilogue 213

described abortions as common even within marriage by 1880, but only 1,019

people were tried between 1873 and 1890; of these, most were acquitted.20 In England, there were all of 37 convictions for abortion between 1870 and 1900.21

It has become commonplace to understand the criminalization of abortion as part of a Meiji state project to control women’s reproduction, along with the introduction of licensed prostitution and the promotion of the “good wife, wise mother”

ideal. An illustrious list of scholars has even argued that the criminalization of abortion was undertaken with demographic goals in mind.22 More recently, this last contention has been eff ectively contradicted by Ishizaki Shōko and others.23

at the authorities brought so few cases to trial and punished the unlucky few who were convicted with relatively short sentences certainly cautions against reading too much into the fact that abortion was offi

cially a crime. Since all “civilized”

nations of the late nineteenth century had laws that threatened punishment for abortion, Japan’s eff ort to be accepted into their ranks made it a foregone conclusion that abortion should be designated a criminal off ense.

L I C E N S E D M I DW I V E S A N D R E P R O D U C T I V E

SU RV E I L L A N C E

is is not to say that the government was not interested in the reproductive life of its subjects. To improve the health of the nation and its children, it introduced a licensing system for midwives in 1874. For decades, its administration was left in local hands and had little eff ect on the practice of midwifery. Typically, women who were already working as midwives could obtain temporary licenses without taking an exam. Th

ose who opted to take an exam oft en simply needed to pass an oral test, which meant that illiteracy was no obstacle to a license. Historian Yuki Terazawa observes that while training programs and exams in this period were largely a formality, they allowed the state to identify midwives and to impose on them its symbolic authority.24

In 1899, the Home Ministry issued a new ordinance on midwives. It permitted existing licensees to continue their work but made entry into the profession contingent on passing a much more demanding written exam. Th e new midwives

were to be literate, versed in modern hygiene and obstetric knowledge, and subordinate to that higher rung of the medical profession, male physicians. Yet it took years for these new professionals to make any great inroads. Only in 1914 did their number begin to exceed that of the oft en illiterate midwives licensed under the old system. Outside these two categories, unlicensed midwives continued to ply their trade in considerable numbers. For a variety of reasons, many women still preferred the old or unlicensed midwives to their modern competitors.25

e state did try to turn licensed midwives into its agents on the ground.26

Although such attempts preceded the 1899 ordinance, they were more likely to

214 Epilogue

succeed with post-1899 licensees, who could read and write and had received a patriotic education. Licensed midwives were expected to report live births to the municipality if the parents failed to do so themselves. In cases of stillbirths and miscarriages aft er the third month, they were to issue a certifi cate with detailed information on the fetus, the parents, and the course of the pregnancy. If the circumstances of the stillbirth were suspicious, at least one prefecture required them to withhold the certifi cate and instead inform the police.27 Th e fact that so few

suspicious stillbirths were actually prosecuted, however, suggests that midwives rarely informed on their clients. Nonetheless, their information gathering furnished the raw material for increasingly elaborate stillbirth statistics, which every prefecture published from 1886. In many prefectures, the number of stillbirths rose from such unrealistically low to such extraordinarily high numbers as to suggest that it took time for midwives and parents to respond to the offi cial

reporting requirement, but by about 1901 that process was substantially complete.28

With the exception of short-lived eff orts in Chiba and Miyazaki during the 1870s, the Meiji state neither monitored nor counted the pregnancies of its subjects.29 It did encourage licensed midwives to maintain detailed charts on the pregnancy, the birth, and its aft ermath;30 however, such charts did not amount to pregnancy surveillance for two reasons. First, midwives did not submit their charts to the authorities unless the police demanded them as evidence in a specifi c investigation.31 Many new midwives, moreover, did not adopt the offi cial format

for their notes, and traditional midwives, who were oft en illiterate, typically did not maintain charts at all. Second, no pregnant woman was forced to employ the services of a new midwife. It was not before the early 1920s that they delivered even half of Japan’s babies.32 Some women continued to give birth without assistance.33 Even among women who called in professional assistance, traditional midwives long retained their popularity. Unlike the new midwives, who told clients to lie down during delivery and tried to impose other newfangled practices, traditional midwives did not challenge established belief systems. Old midwives typically had far more experience in delivering children, and their fees, oft en paid in kind, were more aff ordable. Into the 1920s, they continued to practice in large numbers.34

e stillbirth statistics that this system collected show how far Japan was from reproductive surveillance. As we have seen repeatedly in this book, reported stillbirths showed unusual sex ratios and were much more frequent in many areas of Japan than in the statistics of any other nation. Th

e only plausible explanation is

that midwives and parents reported infanticides and abortions as stillbirths, a practice that observers noted as early as the 1790s and that is attested in oral histories as late as the 1950s.35 If we apply to each prefecture the methodology that Chapter 7 used to establish the number of fi ctitious stillbirths for Eastern Japan, the levels shown on Map 14 emerge for the fi rst decade of the twentieth century.

map 14. Infanticides and late-term abortions as percentage of reported live births by prefecture, 1900–1909. (source: Naimushō and Naikaku Tōkeikyoku, eds., * Kokusei chōsa* *izen, * analyzed with the methodology outlined in Chapter 7.)

216 Epilogue

Based on the assumptions detailed in Chapter 7 (and esp. p. 322n 42), more than 15 percent of fetuses that would otherwise have been liveborn in Ibaraki between 1900 and 1909 were aborted or killed at birth. Th

e same rate exceeded 10 percent

in Saitama and Chiba and approached 9 percent in Tochigi and Tottori. In spite of these continuities with the late Edo and early Meiji periods, some regions once infamous for infanticide, such as Kōchi (Tosa) and Miyazaki (Hyūga), showed only small anomalies in their stillbirth rates by the early twentieth century.

Although more numerous in some regions than in others, these traces of the destructions of fetuses and newborns were widely dispersed. In only seven of Japan’s forty-six prefectures does the estimated proportion fall below 1 percent.

Even in those seven prefectures, the paucity of evidence need not be seen to indicate the rarity of abortion and infanticide; parents and midwives may simply have left such interventions unreported altogether rather than hiding them in plain sight as stillbirths or neonatal deaths.

In those prefectures where high stillbirth rates are on record, no alternative explanation is immediately persuasive. In Japan, pregnant women oft en performed hard physical labor to the very day of their delivery, which may have raised their stillbirth risk over that of women elsewhere. However, stillbirth rates in Sendai’s Edo-period pregnancy surveillance system were, with certain exceptions, not particularly high by international standards (see Table 7 in Chapter 10). Women in early-twentieth-century Yamaguchi and Toyama presumably worked as hard as those in other prefectures, but had stillbirth rates in line with those of European populations.36 Given the continuities between high stillbirth rates and the earlier geography of infanticide, the burden of proof is on showing that women’s work practices diff ered systematically between diff erent parts of Japan, and to an extent that could account for stillbirth rates that in a few districts were ten times as high as European and Taiwanese rates.

Syphilis was widespread in early-twentieth-century Japan, and syphilitic women are at far higher risk of stillbirth than seronegative women.37 However, even the notoriously high prevalence of the disease in early-twentieth-century Japan could not have raised the stillbirth rate far beyond 5 percent even in the worst-aff ected prefectures.38 My estimates above allow for more than 6 percent of children to be stillborn from natural causes. While there are other diseases that lead to frequent fetal deaths, it is not clear that their geography follows the pronounced regional diff erences in reported stillbirths.

Nor does the rapid decline in the stillbirth rate aft er 1910 point to an obvious alternative explanation (Figure 19). Calculated by the assumptions outlined in Chapter 7, the number of false stillbirths remained above 70,000 between 1901 and 1909. Except for minor upticks—the largest of them in 1918, a year notable for rice riots and diffi

cult household budgets—it declined continuously thereaft er. Aft er 1926, it never exceeded 10,000 cases nor 1 percent of all births.

in 60

es,

nticid

infa 40

s an 30

tion

bor 20

a of 10

berm 0

rths

l bi

of al

tageen 3%

perc

a as

ticiannfd i 1%

s an

tion 0%

orab

definite infanticides, from reported deaths in days 6 to 30 after birth definite infanticides, from reported deaths in the first five days after birth definite infanticides, from stillbirths in months 8 to 11

definite infanticides, from stillbirths in months 4 to 7

children who died in late-term abortions or infanticides, from stillbirths in months 8 to 11

children who died in abortions or infanticides, from stillbirths in months 4 to 7

figure 19. Minimum number (above) and incidence (below) of infanticides and abortions in the second and third term of pregnancy from reported stillbirths and neonatal deaths, all of Japan, 1899–1940. (“Defi nite infanticides” are based on sex ratio distortions of stillbirths and neonatal deaths [lower end of 95 percent confi dence intervals] and the assumption that half of infanticides were strictly sex-selective. Since the hinoeuma year of 1906 motivated parents to misstate the birth year of their children, live births, stillbirths, and neonatal deaths were distributed evenly over 1905–1907 for the purposes of this analysis. Th

ese fi gures exclude Okinawa. source: Naimushō and Naikaku Tōkeikyoku, eds., * Kokusei chōsa izen. *)

218 Epilogue

e 1.6 million cases of infanticide or abortion in the fi rst four decades of the twentieth century are to be understood as a minimum, not a best estimate. First-trimester abortions would have escaped the stillbirth statistics entirely, since very few stillbirths were reported before the fourth month of the pregnancy.39 In addition, women who gave birth without the assistance of a midwife, or employed a midwife who did not insist on fi ling a stillbirth report, could have killed or aborted their children without aff ecting the statistics on which the estimates in this chapter are based. One small trace of these otherwise uncounted infanticides and abortions is the fact that in some prefectures and years, stillbirths and live births add up to suspiciously high sex ratios.40

L O N E VO I C E S

Although not the subject of a major public conversation, the scale of stillbirths did attract the attention of some contemporaries.41 Prominent among them was one of the pioneers of statistics in Japan, Kure Ayatoshi (1851–1918). In an article, a book, and a number of lectures between the late 1890s and 1907, Kure pointed out that Japan’s national stillbirth rate was three times that of European countries. Based on comparisons with Europe, he believed that at least half of Japan’s stillbirths were concealed abortions and calculated by how much the population would grow every year if the stillbirth rate could be reduced by that proportion.42 Kure saw the leading nations of the world locked in a struggle in which a large population was a key advantage. With Germany increasing by close to a million people every year, and Russia by twice that, Kure believed that Japan could not aff ord to lose so many subjects to fetal deaths.

Kure was not alone in his wish for Japan’s population to grow even faster. Like the various nations of the West, Japan had its share of academics, journalists, administrators, soldiers, and politicians who took an interest in demographic questions. A statistical system of growing sophistication fed their opinions with a steady stream of data, while the contending worldviews of various Western thinkers leavened their debates. Among these men and, especially in the incipient birth control movement, women, it is possible to fi nd representatives of most plausible as well as of some implausible positions.43

Although the temporary triumph of wartime pronatalism has drawn much attention to its antecedents, the advocates of more rapid population growth never enjoyed uncontested control of public opinion or government policy between the 1870s and the late 1930s. Th

e government did occasionally hinder the spread of

contraceptive knowledge and in 1930 issued regulations that banned certain contraceptives as dangerous to the health of women, but advertisements continued to appear, various devices were openly sold in pharmacies, and new birth control clinics opened.44 Th

e desirability of both population growth and contraception

Epilogue 219

continued to animate debates. In 1927, a government-sponsored Commission on Population and Food Problems concluded that “establishing fundamental measures against Japan’s population growth is an urgent task of the present moment”; two years later, its recommendations included the “creation of facilities suitable for giving medical advice on marriage, births, and contraception.”45 In 1928 a member of the Imperial Diet even submitted a draft bill for the partial legalization of abortions.46 In spite of his eff orts, abortion remained a crime, and in the 1930s several leading birth control activists were arrested for assisting in pregnancy terminations.47

In contrast to the diverging views on birth control, there was throughout the early twentieth century broad agreement that Japan’s death rates, and especially infant and child mortality, were far too high.48 Before the late 1930s, population policy, such as it was, therefore mostly concerned itself with public health and hygiene.

While Japan’s high death rates received much attention, its abnormal stillbirth rate was apparently not widely discussed. When the Home Ministry launched a Health and Hygiene Survey Offi

ce (Hoken Eisei Chōsakai) in 1916, it justifi ed the

new institution with the prediction that Japan, suff ering high and rising death rates, would soon follow the recent European trend of declining birth rates, with worrisome eff ects for its demographic vigor. Th

e statement made no reference to

Japan’s high stillbirth rates.49 When, the following year, the Health and Hygiene Survey Offi

ce established its eight research foci—infants, toddlers, and school children; tuberculosis; venereal diseases; leprosy; mental illness; the necessities of life; public health in villages; and statistics—stillbirths, abortions, and infanticides again did not make the list.50

Even among the minority of observers who expressed concern about the stillbirth rate, the persistence of infanticide was not a prominent issue.51 Kure Ayatoshi implied that most of the excess stillbirths were concealed abortions. Even the scale of abortions was not widely known. In 1903, a prefectural offi

cial by the name of Ogawa

Hayashi confessed in a letter that the fi rst time he had heard that stillbirths were so frequent was at one of Kure’s lectures. Th

is is particularly remarkable given that

Ogawa worked for Ibaraki prefecture, which at the time had a stillbirth rate of 19

percent, easily Japan’s highest. Even with the benefi t of Kure’s insights, it took Ogawa several steps of statistical reasoning to arrive at the tentative conclusion that many stillbirths were really concealed abortions. He also considered a physical weakness of either parent, sexual misconduct, syphilis, and a lack of hygiene as possible causes.

Only aft er reviewing the concentration of stillbirths in districts whose inhabitants

“thought any new pregnancy superfl uous aft er two or three children” did Ogawa decide that abortion was aft er all a major cause for the high stillbirth rate.52

Other offi

cials must have known that infanticide was still common. Newspapers carried reports about individual cases of murdered newborns with great

220 Epilogue

regularity.53 In some village communities at least, infanticide was still committed more or less openly, without any sense of social deviance. Th is is at least the

impression that the many oral histories about the early twentieth century create in their specifi c stories about individual instances of infanticide. At that time, Japan’s history of analyzing, debating, and excoriating infanticide was still in living memory. Infanticide may have sat uneasily with the idea that Japan was now a civilized nation, but these memories should have kept it only a quick associative step away for an offi

cial reviewing a suspicious stillbirth statistic. Surely, many administrators simply did not think that this issue merited a confrontation with the population. Th

e occasional administrator or policeman who tried to crack down on infanticide did so at considerable cost to his position in local society. In a letter from the 1920s, a friend of the eminent rural sociologist and historian Ono Takeo described the situation in a village in Yamanashi as follows. It had no police offi

cer, so a forestry offi

cial based in the village was in charge of law enforcement.

If such an offi

cer, whose term in the village was typically short, aspired to popularity, he would “pretend not to hear anything about the villagers’ infanticides. . . .

Naturally,” the letter continues, “babies were few in such years. In years when an unpopular offi

cer resided in the village, babies would conversely be numerous.”

Apparently, employees of the municipality were aware of this dynamic, which they observed with “wry smiles.”54

at abortions and infanticides were no longer a policy priority by the early twentieth century is evident in the fruitless eff orts of the aforementioned employee of Ibaraki prefecture, Ogawa Hayashi. So moved was he by Kure Ayatoshi’s stillbirth lecture that he advocated better policing and a raft of welfare policies to keep more fetuses alive. To set the process in motion, he submitted a lengthy memorandum to his superiors and a separate proposal to Ibaraki’s conference of city mayors and district chiefs. In Ibaraki, he conjectured, “abortion” served the same economic desires that in France motivated a fashion for “avoiding pregnancies by artifi cial means.” For this and other reasons, he believed that an entrenched custom like abortion could not be easily stamped out by police force alone.55

Rather, Ogawa thought, nothing less than a major eff ort by both private individuals and the state was required. Such an eff ort would have many facets, including visits by police offi

cers to the homes of pregnant women, better state control of midwives, educating primary school teachers about the problem, improving welfare institutions, and helping the poor with exemptions from local taxes, reduced school fees, and childrearing subsidies.

In a passage that shows how much offi

cial concern over infanticide and abor-

tions had diminished since the Edo period, Ogawa conceded that it was diffi cult to

argue that childrearing subsidies should be a priority use of state funds. Perhaps it was for this reason that he urged the mayors and district chiefs to fi nance childrearing subsidies from village, town, and district budgets. Th is would have

Epilogue 221

allowed a minority of jurisdictions to adopt the program even if it failed to win the acceptance of a majority throughout the prefecture. Unfortunately for Ogawa and the fetuses of Ibaraki, it seems that the lower levels of the administration had no more wish for an additional burden than the prefectural government. When Ogawa reported his eff orts to Kure three and a half years later, none of his proposals had been implemented yet. If they ever were, no documents from Ogawa’s system survive in Ibaraki’s prefectural archives.

A few years earlier, neighboring Chiba prefecture had demonstrated that better policing alone could in fact bring down the stillbirth rate. Another offi cial who

had audited Kure’s lectures, an employee of Chiba’s police department named Ōkubo Fukukichi, complained to Kure that he seemed unaware of Chiba’s success.

In his 1903 letter, he reported that the stillbirth rate of Chiba had been declining in recent years, all thanks to a prefectural ordinance issued in the fall of 1898: “In cases of stillbirth and neonatal deaths within the fi rst three days, the police must be informed before the body is buried or cremated. Th

ose who violate this ordi-

nance will be punished by one to ten days in prison or a fi ne of 5 sen to 1 yen.”56

Five sen (0.05 yen) was a trivial sum even in 1898, and 1 yen was no longer a ruinous fi ne, but presumably because the ordinance was accompanied by more determined policing, Chiba’s stillbirth rate fell by over a third in its aft ermath, from 21 percent in 1899 to 13 percent in 1900. In Sanbu district (an amalgamation of old Yamabe and Musha), the eff ect was even more impressive: the proportion of stillbirths plunged from 46 percent in 1897 to 16 percent in 1900. Nor did people continue as before and simply stop reporting their infanticides and abortions as stillbirths. Th

e total number of births in Sanbu was unchanged between those two years, which meant that live births increased by 80 percent.57

Kōchi prefecture in Shikoku was another partial success story. In 1888, 18 percent of its children were reported as stillborn, a rate exceeded only in Chiba.

Over the next ten years, however, that rate dropped by half. Kure thought that the credit should go to a private initiative, the prefectural Childrearing Society (Ikujikai). Founded in 1884 by men who had been active in earlier organizations of the 1870s, it was able to collect 70,000 yen from private and offi cial sources, which it

used for raising orphans as well as assisting poor parents. In Kōchi, such subsidies were a novel institution. Its Edo-period government, Tosa domain, had only aided parents of twins and triplets, and otherwise contented itself with a long series of infanticide bans. It is open to question whether it was the Ikujikai or some other factor, such as more determined policing, that reduced Tosa’s stillbirth rate at a time when no such tendencies were evident in most other parts of Japan. At the very least, the activities of the Ikujikai and the offi

cial support it received do show

that in Kōchi, philanthropists and offi

cialdom identifi ed infanticide as a problem.

eir eff orts were unusual in a country in which the eradication of infanticide and abortion had generally ceased to be a policy priority.58

222 Epilogue

T H E R E T R E AT O F I N FA N T I C I D E

I N T H E TA I S H Ō P E R I O D

Between 1908 and 1927, my estimates for the national rate of infanticides and late-term abortions drop from 4.3 to 0.45 percent, a reduction of nine-tenths. Remarkably, this decline occurred at a time when the state and private eff orts placed comparatively few obstacles in the path of parents determined to kill their newborns.59 Th

e question of how and why infanticides diminished so rapidly in this period, which roughly coincided with the reign of the Taishō emperor, deserves a great deal more analysis than is possible in this chapter. We can speculate that the intrusion of state power played a role, in the form of new midwives and more insistent policemen, and that modern contraception began to off er an alternative means of family planning. Yet there are diffi

culties with both these explanations.

Women were not forced to employ new midwives; policing strategies did not obviously change between 1900 and 1910; modern contraception did not come into wide use until the 1950s. Understandings of what it meant to kill an infant, however, may have been diff erent among the new generations of women who reached their childbearing years in a world of motion pictures and compulsory education, empire and electricity.

New midwives came to assist the majority of births probably sometime in the early 1920s. Th

ey were typically young and educated. Even before they began their professional training, they had received years of schooling that stressed subjects’

duty to the state and instilled an ideal of what it meant to be modern and civilized.

Infanticides presumably had no place in their worldview. Th e midwife schools

would have reinforced these ideas and explicitly enjoined students not to assist in infanticides and abortions. New midwives were also asked to inform on colleagues.60

In all likelihood, the temptation to break these rules was small. New midwives probably rarely received requests for an abortion or infanticide, since families hoping to procure either would surely have turned to a trusted traditional midwife rather than to a young outsider. By the time the retirement of midwives operating under old licenses resulted in regions in which the only birth assistants were new midwives, the reputation of new midwives for not assisting in abortions and infanticides must have been well established. Before the onset of wartime rationing, no pregnant woman was required by law or circumstance to employ the services of a new midwife. Although the new obstetrics had gone a long way in convincing women to lie down during delivery, some would still have remembered the old kneeling position that made it possible to give birth unassisted.

Nonetheless, the very presence of new midwives in village society may have played a part in changing views of infanticide and abortion. An image of modernity with their bicycles, economic independence, and advanced education, midwives had an

Epilogue 223

eff ective platform for infl uencing the views of younger women in particular, many of whom regarded them as role models. On balance, then, the increasing prominence of new midwives may have played a limited role in the rapid decline of infanticide. Th

is seems to be borne out by the statistics; in the 1920s, there was a slight negative correlation between the per capita number of licensed midwives and the stillbirth rate in the prefectures of Japan.61

Policemen, too, could be moderately eff ective foot soldiers in the fi ght against infanticide. As we have seen, offi

cers did sometimes, depending on their beat and

their personal inclinations, try to use their authority to convince pregnant women and their families to raise the unborn child. While such eff orts oft en fell fl at, their cumulative eff ects could have been considerable. A single investigation into a suspected infanticide could scare a hundred neighbors for years to come. In an interview she gave in the 1970s, Sakai Misao, a woman in a fi shing village in Ehime, explained the survival of her youngest sister, born in the 1920s, with this dynamic:

“In my grandmother’s time [childbearing years ca. 1870s–1890s], there were far fewer children in an average family because women practiced infanticide and did abortions on themselves. My mother [childbearing years ca. 1893–1925] heard about a woman who killed her own infant and was later questioned by the police

[probably around 1920]. For this reason, my mother decided not to kill her lastborn, my youngest sister.”62 It is notable that this account only refers to an investigation, not to an actual conviction for infanticide. Th

is kind of police pressure

probably worked mostly on the timid, or on those who were ambivalent about whether to raise another child. Th

e limits of policing are evident in the case of a

woman who was apprehended in rural Shizuoka prefecture in 1928; although the police found that over the previous twenty years she had performed about a thousand abortions, a term that may have really been a cipher for infanticides, they made no eff ort to charge her former clients.63

e tail end of infanticide in Japan overlapped with the incipient rise of modern contraception as an alternative form of family planning. Th e army began issuing

condoms to soldiers during the Russo-Japanese War (1904–1905) and by 1931

instructed its men to carry their “hygiene matchboxes” when leaving base.64 As the name implies, the emphasis was on the prevention of infection, not of pregnancy, and condoms long retained their association with commercial sex. Around 1914, books on birth control ( sanji seigen or sanji chōsetsu), a term that entered Japanese at this time, began to appear, as did articles fi rst in middlebrow journals and then in women’s magazines.65 By the late 1920s, Japan produced most kinds of contraceptive devices then known, and advertisements for them appeared with great frequency.66 More than sixty birth control clinics existed in Tokyo alone.67 In a newspaper article in 1928, the director of the Bureau of Statistics interpreted the (minuscule) decline in the national birth rate as evidence for the spread of birth control.68 In the great cities, which embraced contraception earlier, total fertility

224 Epilogue

rates had been falling for some time, and between 1925 and 1930 the national rates also entered a gentle decline.

At the same time, contraception was still a suffi

ciently novel idea that maga-

zines debated which methods were reliable and which were not.69 Reading about contraception did not necessarily lead to its immediate adoption. Sakai Misao recalled that the fi rst indistinct reports of this new world reached her remote fi shing village in Shikoku around 1925. “When I was about fi ft een or sixteen years old, I remember hearing from my eldest sister something about birth control and the name of Margaret Sanger. My sister liked to read and may have read a newspaper article about it.”70 Sakai Misao went on to marry and bore fi ve children; like everyone else in her community, she claimed, she never used contraception.

Chronology and geography create diffi

culties for an account that links contra-

ceptives to the decline of infanticides and late-term abortions. Contraception spread rather slowly. If a postwar newspaper survey is anything to go by, fewer than 10 percent of couples used birth control in 1937.71 Modern contraception was still rare in 1925, by which time fi ctitious stillbirths amounted to less than seven births in a thousand. Birth control made its greatest early inroads in the big cities, and not in the areas that had previously had the highest incidence of infanticide such as Ibaraki, Saitama, and Tottori. Not only were the major cities home to most birth control clinics, but the only prefectures with fertility declines of more than half a child between 1910 and 1925 were Osaka, Hyōgo, Kyoto, and Tokyo with their large urban centers (Map 15). For the rest of Japan, the early days of modern contraception can only be a supplemental explanation for the diminution of feigned stillbirths around 1920, except for one intriguing possibility. Even where modern contraception was not used, its theoretical availability may have weakened the case for infanticide. If unwanted pregnancies could be avoided by other means, it was harder for parents to justify infanticides as an unfortunate necessity.

In the areas with the greatest falls in feigned stillbirths, fertility increased considerably between 1910 and 1925 (Figure 20). Ibaraki, for example, reduced its rate of fi ctitious stillbirths from 10 percent of births to 1.3 percent; its fertility rate rose by almost one child per woman during the same years. In most prefectures, the fertility increase was about twice or thrice greater than the decline in the stillbirth rate alone can explain. Th

is may mean that the stillbirth reports greatly understate the actual number of infanticides and abortions in 1910. Th ere are other possible

proximate reasons for why fertility rose, but much more research is needed to show that any of them is persuasive.72

e geography of stillbirths off ers some indication of the forces at work. In Map 16, which averages the published fi gures for 1925 and 1930 to improve the sample size for each district, high stillbirth rates still clustered as they had in the 1890s, but at a much reduced level (cf. Map 11 in Chapter 10), with no district reporting more than 12 percent of its children stillborn by 1930. If newfangled

map 15. Change in the total fertility rate by prefecture, ca. 1910–1925. (sources: Based on Hutterite Indices in Takahashi, “Meiji-ki no chiiki-betsu shusshōryoku,” for 1908–1912; for 1925, Naikaku Tōkeikyoku, eds., *Kokusei chōsa igo. *)

25 -1%

0-19

Osaka

Hyǀgo

, 191tea -2%

Tokyo

irth

Iwate

Kyoto

stilb 3%

e th

Akita

Miyagi

Nara

-4%

change

Fukushima

Tochigi

-5%

Chiba

-6%

Ibaraki

Saitama

-7%

-1.5

-1.0

-0.5

0.0

0.5

1.0

1.5

2.0

change in the total fertility rate, 1910-1925

figure 20. Change in fertility and stillbirth rates in the prefectures of Japan, ca. 1910–1925.

Bubble sizes refl ect the stillbirth rate in 1910. (sources: Takahashi, “Meiji-ki no chiikibetsu shusshōryoku”; Naimushō and Naikaku Tōkeikyoku, eds., *Kokusei chō sa izen; * and Naikaku Tō keikyoku, eds., *Kokusei chō sa igo. *)

Epilogue 227

forms of birth control or inquisitive policemen drove infanticide to the margins, we would expect an altogether diff erent spatial pattern. Modern contraception and early-term abortions were much more readily available in the large cities than in the countryside; if modern contraception accounts for the drop in reported stillbirth rates, there should have been a penumbra of low-stillbirth districts around the major cities. Instead, we fi nd many districts with stillbirth rates above 6 percent in the environs of Tokyo and Osaka, and to a lesser extent also in those of Kyoto and Kobe. Since the intensity of policing depended on the varying enthusiasm of individual offi

cers and their commanders, we would expect its ef-

fects to be more pronounced in some prefectures than in others. On Map 16, however, areas with frequent stillbirths spill over prefectural boundaries. Individual clusters of such districts straddle Shimane and Tottori, Ibaraki and Chiba, Kyoto and Nara, as well as the area where Gunma, Tochigi, and Saitama meet. If, however, a slow but steady erosion of the old understandings that enabled infanticide—shared, reinforced, and reshaped in the conversations of neighbors and friends—explains its retreat in these years, we would expect its geography to look much like Map 16.

Even in remote villages, people may have found it increasingly diffi cult to

openly justify infanticide and abortions as its inhabitants took pride in their country’s new status as a seat of empire. Already in the eighteenth century, critics had connected infanticide with the past and the peripheries. In the Meiji period, this association was made with still greater emphasis. It is possible that in the 1910s and 1920s imagining oneself as part of an old, peripheral world was particularly distasteful to many citizens of Japan.73

It is also possible that most people gave up infanticides and abortions not just because they found them abhorrent, embarrassing, or dangerously unlawful, but simply because they actively wanted more children. Th

e valorization of childbirth

as a contribution to the community, whose origins in the 1790s we surveyed in Chapter 11, may have reached new levels in a Japan whose farmers and laborers of-ten felt an intense concern for the strength of the nation. If so, a large brood would have not only given a couple a sense that they were doing their patriotic duty, but also raised their social status. Th

e cult of domesticity, which told women to aspire

to becoming “good wives and wise mothers” ( ryōsai kenbo), * * may have fostered a larger ideal family size. Such ideals could have worked their way into the minds of young men and women by a number of paths. Magazines and motion pictures celebrated forms of femininity that did not feature a recourse to abortion or infanticide. New midwives, policemen, and schools could all be eff ective conduits for disseminating ideals nurtured in government ministries. So could the associations of ladies, young men, and young women that existed in every village.74

If fertility rose because domesticity and patriotism made high fertility more desirable for individual parents, that process would be yet another variation on the

map 16. Stillbirth rate by district, average of 1925 and 1930. (source: Naikaku Tō keikyoku, eds., *Kokusei chō sa igo. *)

Epilogue 229

theme of this book, the interplay of discourse and demography. As pleasing as this explanation would be in this context, it is not without diffi culties. As we saw on

Map 15, fertility actually fell in Japan’s new industrial centers—Tokyo, the Kansai, and northern Kyushu—over the years in question. Were the more rural prefectures, and especially Eastern Japan, somehow more receptive to the kind of ideals that would have inspired a higher fertility norm? Even in the countryside, we should not overestimate the extent to which any patriotic urge to procreate determined attitudes toward childrearing. As one reformer recollected about the situation in a sericultural village in Yamanashi in about 1933, “Because [the villagers]

earned money from silkworms, they referred to them with the honorifi c *o-ko-san. *

Children they would call ‘hungry ghost’ ( gaki) or ‘snotnose’ ( hanattarashi).”75

SU B C U LT U R E S O F I N FA N T I C I D E

I N T H E 1 9 3 0 s A N D 1 9 4 0 s

In some pockets, infanticides remained common as late the as the 1930s. In 1940, two researchers obtained the vital statistics of Shimazu, a municipality of 2,800

inhabitants on the shores of Lake Kasumigaura in Ibaraki, for the years 1930–1939.

Stillbirths there were twice as frequent as the national average and clustered suspiciously in the tenth month of pregnancy. Infant mortality rates rose continuously by birth order, from 13 percent for fi rstborn children to 40 percent for children whose mother had previously given birth fi ve or more times. More than half of all infant deaths were attributed to “congenital weakness,” “congenital deformity,” or

“prematurity” and claimed nearly 13 percent of all infant lives. Nationwide, the proportion was 3.7 percent. When the researchers published these fi ndings in 1943, they judiciously left it to their readers to reach the obvious conclusion that infanticides were still common in Shimazu.76 In 1947, the same two men wrote a follow-up that extended the analysis to fi ve neighboring villages, which also had frequent stillbirths, a particularly dangerous tenth month of pregnancy, and a high incidence of “congenital weakness.” While nationwide only 17 percent of infant deaths occurred in the fi rst fi ve days, that proportion reached 44 percent in one of the villages. Th

is time, the authors stated their conclusion more boldly. Th e cause

of the region’s high rates of infant death and stillbirth was not be found in “medical or biological factors, but entirely [in] evil customs ( rōshū) that arise from social conditions.”77

Statistical traces of infanticides remain even from the fi nal years before abortion was legalized. Between 1947 and 1949, sex ratios for children born to women who had already delivered unusually many children for their age showed notable distortions. Th

e sex ratio for the seventh children of women in their late twenties was 109; for the tenth children of women in their early thirties, it was 116. For children with at least eleven older siblings, living or dead, the sex ratio stood at

230 Epilogue

fully 165. In a pattern that is by now familiar to us, the medical student who drew attention to these fi gures in 1952 avoided calling them evidence of outright infanticide. Instead, he imagined that children, and especially girls, with so many siblings would “not be welcomed” but rather “treated poorly.” If they died shortly after their birth, parents would fi nd it less troublesome to fi le one stillbirth report rather than two separate reports for birth and death.

An alternative explanation is more plausible. Women could only give birth seven times by the age of twenty-fi ve, or ten times by the age of thirty, if their birth intervals averaged less than about eighteen months. We would expect women who still killed some of their newborn children to have just such short birth intervals, since aft er each infanticide they would not be breastfeeding a child and would therefore rapidly return to ovulation. By 1947, they would have been such a minority that at lower birth orders, they were submerged in the large number of women who did their level best to raise all their children. As the last holdouts of the culture of infanticide gave birth to a seventh child in their twenties or a twelft h child in their thirties, however, they outraced most of their contemporaries. Th eir

numbers were now diluted only by women who did not nurse their children, and women who had lost a long succession of infants to illness or accident. It is only at such high parities, therefore, that the traces of the infanticidal minority are so distinct. Even this tip of the iceberg still amounted to more than a thousand missing girls per year in the late 1940s.78

T H E R E T U R N O F P R O NATA L I S T P O L I C I E S A N D T H E

L E G A L I Z AT I O N O F A B O RT I O N

Unequivocally pronatalist policies returned to Japan in the late 1930s. At this time, fertility rates had entered a gentle decline, but based on the experience of the nations of Europe during World War I, Japanese population experts warned that Japan’s current war in China could sharply reduce the number of births. As is well known, expansionist fantasies grew increasingly fevered at this time. Th e population policies they inspired and their slogan “Give birth and multiply” ( umeyo fu yaseyo) are so notorious that they have cast their shadow on the entire Imperial period. Some of the policies were never fully implemented amid the exigencies of war, such as a 1941 plan to raise the birth rate by lowering the average age of marriage, restrict the employment of women in industry, grant loans to newlyweds, and reduce taxes for large families.79 Other measures cost little and were largely symbolic, such as the public recognition in 1940 of about ten thousand couples with ten or more children.80

ose policies that were actually implemented and represented a real commitment of resources were not motivated by demographic expansionism alone. Th ey

protected the welfare of mothers and their children, an eff ort on which diverse

Epilogue 231

political camps could agree. While women’s suff rage had failed to pass the upper house of the Imperial Diet in 1933, its advocates considered the passage of the Bill for the Protection of Mothers and Children (the Boshi hogohō) of 1937 an important victory.81 Coming into eff ect in 1938, the law invited mothers without an able-bodied husband to apply for a subsidy to cover their living expenses.

As the war dragged on and then assumed a new, desperate dimension with Japan’s attack on the United States, civilian life was increasingly patterned by rationing. In 1942, the Ministry of Health and Welfare made pregnant women and mothers with small children eligible for special rations. To obtain them, women were required to present a maternity passbook ( ninsanpu techō). Before such booklets could be issued, in turn, pregnant women needed to present a certifi cate from a licensed midwife or a physician. 82 Th

is was the fi rst time that many women

were eff ectively forced to employ the services of a licensed midwife.83 Th e new

system therefore amounted to a form of pregnancy registration, a version of which continues to operate in Japan to this day. However, even at the height of the war, failure to register only forfeited the extra rations (which would be less urgently needed by a mother not planning to raise the child) and did not carry any penalties.

Although eff orts to legalize fi rst-trimester abortions failed in the 1930s, war and defeat discredited many conservative political agendas, including pronatalism.

is changed the political balance on the abortion question, as did the tremendous diffi

culty of feeding and housing Japan’s population amid the ruins of defeat. Terrible as the war had been, in its course there had been far more Japanese births than Japanese deaths. Th

e population statistics for the early 1940s are problematic, but the census of 1947 returned 10 percent more Japanese than the government had counted in 1938. As soldiers were demobilized and went home to impregnate their wives, a surge in illegal and oft en unsafe abortions came to be seen as a major social problem.84 With the approval of the Occupation authorities, the National Diet eff ectively legalized abortions in 1949, when it amended the Eugenics Law of 1940 to admit “economic reasons” as valid grounds for a legal abortion.85 Abortion remained the single most important form of birth control for the next decade, with the offi

cial number of terminated pregnancies peaking at nearly 1.2 million in 1955.86 Considering that abortionists were so eager to lighten their tax burden that they left many procedures unreported, the actual fi gure was probably twice as high.87 In Eastern Japan, fertility rates soon returned to the low levels of the eighteenth century, and then, along with those of every other part of the country, declined to the fi gures that vex Japan’s policymakers today.