C O N T I N U I T Y, C HA N G E , A N D D I F F U S I O N
Eastern Japan’s culture of infanticide lasted longer than the modern reproductive system of reliable contraceptives and safe abortions has so far endured in any part of the world. In some areas, it spanned more than two centuries, assuming its distinctive characteristics in the late seventeenth century and persisting into the age of cinema and motorcars. Infanticide was rooted in the most fundamental worldviews of those who employed it to plan their families. Springing from particular understandings of humanity, death, and immortality, it expressed ideals of responsible parenthood and obligations to elders, of how families should seek economic security and the respect of the community.
When these understandings and ideals changed, so did the reproductive culture they once sustained. Th
e people of Eastern Japan did not rearrange their
worldviews spontaneously; rather, they did so gradually and in response to prodigious eff orts on the part of governments and private individuals. Some of these men argued against infanticide even though they seem to have shared some of its premises—most importantly, that infants were not fully formed human beings.
Th
eir opposition to infanticide was oft en, if by no means always or exclusively, motivated by the problems declining populations caused in a society whose institutions were designed for stability, not for growth or contraction. Th e fi scal
health of a domain, however, did not make for very persuasive rhetoric among villagers. Opposition to infanticide was therefore predominantly framed in moral terms. Genuine outrage no doubt played its role, but it is possible that some who argued against infanticide on ethical grounds convinced as much themselves as 232
Conclusion 233
their intended audience that newborn children could not be killed and that raising many children need not lead to poverty.
A second theme in this conceptual rearrangement concerned the social meaning of reproductive choices. In the culture of infanticide, excessive childrearing had been considered irresponsible toward both the chosen children and the larger community. When public policy began to contest infanticide, it marked numerous progeny as a social contribution rather than as a sign of foolish self-indulgence.
From the 1790s, the growing perception of a foreign threat drew infanticide into a discussion about national security, and from the 1820s, the rise of neo-Shinto Nativism introduced the wishes of gods into the discussion.
The concepts that favored or negated infanticide did not live silently in individual minds, but were propagated and reinforced in conversations. Some of these exchanges connected people across considerable distances, and many concerned questions of policy. In this sense, they may have amounted to a genuine public sphere. Administrators looked across their borders to see how other domains combated infanticide.1 So did their subjects. When they learned from friends, travelers, or business partners that other lords invested in the protection of infants, they came to expect that their own ruler would live up to the same standards of benevolent governance. Th
e texts and images of the moral suasion eff ort drew on
one another and elaborated shared themes; the busy networks of concerned samurai, priests, headmen, and village doctors that read and copied these materials could therefore project the thoughts of a townsman of Mito into every village of Sendai within a matter of months, or confront women in Tsuyama with an image devised by a retainer of Kurobane more than four hundred miles away.
Most conversations about rearing or thinning children, however, were local.
Th
e web of discourses that made infanticide seem unproblematic, advantageous, and oft en necessary was spun between neighbors, friends, and relatives. Such exchanges could sometimes cross long distances, but far more frequently they were bound by the limited physical spaces in which most people lived their lives.
Infanticide was therefore rendered normative or outrageous by the balance of an overwhelmingly local conversation.
Some of the best evidence for this process can be found in the stillbirth statistics of the 1890s. Elevated stillbirths tended to cluster. Th ese clusters—Tosa, Bizen,
and Bitchū, the mountainous spine of northern Honshu, the stretch of land between eastern Kazusa and southern Hitachi, and a few others—had only one conspicuous commonality: their historical records bear clear traces of infanticide in the Edo period or in the 1870s. Most of them did not experience long-term pregnancy surveillance and childrearing subsidies. Th
e clusters oft en crossed the
prefectural boundaries of the 1890s, so that it is unlikely that local policies at the time these statistics were collected were a major infl uence. In the Kantō, high stillbirth rates radiated outward from Yamabe district on the Pacifi c coast of Chiba,
234 Conclusion
where close to half of all babies were reported stillborn in 1890 (see Map 17). If a traveler started in Yamabe and headed northeast, southwest, or north, he would pass through one to three districts with stillbirth rates above 30 percent, then another few districts with rates above 20 percent, before reaching a larger outer ring whose stillbirth rates were still above 10 percent. Yamabe and its neighbors were not otherwise exceptional. Th
ey were, as far as we can glean from the causes of
death of its people, not exceptionally affl
icted with syphilis or other diseases that
cause stillbirths. Th
ey were neither the center of a particular industry nor of a religious movement, not notably prosperous but also not especially poor.
We may therefore read the concentric circles as evidence for the importance of social networks in permitting infanticide. People were more likely to kill their babies if they knew people who did the same. Networks of acquaintance, marriage, and gossip transcended the village level but were nonetheless constrained by physical distance. By the late nineteenth century, Yamabe found itself at the center of a large concentration of people who practiced infanticide, so that its residents were more likely than those of other districts to have a critical mass of baby-killing acquaintances.
Historically, Yamabe lay at the edges of Eastern Japan’s culture of infanticide. In the eighteenth century, *mabiki * was most oft en reported in areas farther north, especially the provinces of Hitachi, Shimotsuke, and Mutsu. Th at Yamabe had
become the center of what remained of that culture by the 1890s is suggestive about the processes by which infanticide receded. Around Edo-Tokyo and at the southern tip of Chiba’s Bōsō peninsula, infanticide had probably never been quite as widespread, which would explain the gradual diminution of stillbirths as one traveled south and west from Yamabe. In the north, however, the low rates were achieved in a hundred years of campaigns by domains, intendancies, and private individuals, which had succeeded in reducing the incidence of infanticide and as a result created a larger and larger area in which most people did not know a critical mass of others who openly killed their newborns.
Th
at reproductive cultures live in local conversations is no new insight. Low fertility in Europe spread from one adjoining district to another, traveling fast through areas of shared faith and tongue and stopping for sometimes more than a generation at linguistic or confessional boundaries.2 Europe’s modern fertility transition started in France but did not reach the areas of the country where Breton, Corsican, or German predominated until decades aft er speakers of French and Occitan had adopted a small-family norm. Low fertility crossed the political border into the Francophone parts of Belgium with little delay, but then stalled for decades at the linguistic boundary between Walloons and Flemings. It took fully sixty years (ca. 1860 to 1920) for the fertility decline to cover the hundred miles between the Francophone Belgian arrondissement of Hainaut and Flemish-speaking Maasiek.3 In Japan, too, the diminution of infanticide is not the only
map 17. Stillbirth rates by district in the 1890s. (District-level stillbirth fi gures do not survive for every year; this map is therefore a composite of 1890 [Chiba], 1891 [Tochigi], 1893 [Ibaraki], and 1892 [all other prefectures]. Figures for Saitama are for the whole prefecture, since district-level fi gures were published only from 1905. source: Meiji *nenkan fuken tōkeisho shūsei. *)
236 Conclusion
example for how social context shapes a reproductive culture. Among the respondents to a 1950 survey who endorsed abortion unconditionally, the second most frequently stated reason for their acceptance was that “everyone is doing it.”4
E A S T E R N JA PA N I N WO R L D D E M O G R A P H I C H I S T O RY
Th
ere is reason to think that the scale and patterns of infanticide in Eastern Japan were far from exceptional. In most corners of the world, the historical or ethnographic record is full of references to habitual infanticide.5 In some of these settings, the fragmented references that survive suggest motives and patterns very similar to those we have studied in this book. Twins were at particular risk of infanticide in many societies beyond Japan.6 Horoscopes, calendars, and oracles informed infant selection in Madagascar and ancient China, among the Khonds of Orissa and the protagonists of Greek myth.7 Around 1100, one scholar in the southeast of the Song Empire reported that “the commoners of the Yue and E regions customarily want to rear only two boys and one girl. Newborn babies aft er that will be killed.”8 Another described the childrearing ways of the same people in terms that could have equally applied to eighteenth-century Japan: “If they have many sons, they kill their boys, if they have many daughters, they kill their girls, and call this long-established custom ‘thinning out children’ ( hao zi).”9 Alas, too few sources seem to survive from Yue and E’s reproductive culture for *Hao Zi * to become a book title anytime soon. Most other settings where infanticide was frequent, too, lack Tokugawa Japan’s abundance of demographic data, and few approach the richness of Japan’s textual record.10 Eastern Japan’s culture of infanticide is therefore exceptional in at least one way: the density of its surviving documentation.
Despite the huge achievements of several generations of scholars, our knowledge of the world’s demographic history consists of little pools of light in a vast dusky forest. For the centuries before the twentieth, reliable records have as yet been analyzed only for a mere handful of societies—mostly in Europe, its overseas settlements, and Japan, as well as smaller parts of Europe’s colonial empires, Qing China, and very recently also Joseon Korea. As new sources are discovered and new techniques invented, new islands of knowledge will emerge. For many times and places, however, population history will always be an art of plausible interpolation. If it is possible to prove that infanticide was very frequent in one large and complex society, it becomes a more plausible explanation for others whose sparser sources leave more room for speculation.
Th
e potential for the history of the Edo period to infl uence our understanding of demographic systems with less ample surviving documentation is evident in the work of Walter Scheidel, an authority on ancient Greek and Roman demography.
His views deserve to be quoted at length, for they show how fertile the encounter of East Asian historical demography with that of the ancient Mediterranean has
Conclusion 237
already been, and hint at what light Eastern Japan’s history of infanticide may yet shine on distant times and places.
Th
ere is no good reason to believe that in the Roman period, the population of the empire practised forms of family limitation that led to demographic contraction.
However, the view that natural fertility prevailed in most segments of that population invites a word of caution. Quantitative evidence from early modern China . . .
shows that even a pre-transitional population may accommodate a certain amount of family limitation. . . . In China, [infanticide] . . . was not so much a desperate last resort in times of crises than an accepted means of long-term fertility control. Th e
same is true of Tokugawa Japan. In classical antiquity, the inhabitants of ancient Egypt were singled out for raising all their children. By implication, Greeks and Romans did not. Th
us, as far as the availability of post-natal family limitation is concerned, the core areas of the Greco-Roman world may well have been more akin to East Asia than to early modern Europe. As usual, ancient statistics that would permit us to test this assumption do not exist. Th
us, we are left with an intractable
problem: although the Egyptian fertility schedule and comparative evidence from later periods of Mediterranean history support the notion that natural fertility was common throughout the Roman world, the mere existence of the East Asian model lends a measure of credibility to the notion (which could otherwise have been no more than free-fl oating speculation) * * that the inhabitants of Roman Italy or Greek cities may have actively engaged in family limitation. . . . However that may be, what is emphatically not implied by the East Asian or any other scenario is that such forms of family limitation in pre-transitional societies would result in demographic contraction. Between 1700 and 1850, when the relevant fertility data were recorded, the Chinese population grew by some 160 per cent.11
In Eastern Japan, of course, infanticide was a major reason for a long and deep contraction of the population. Th
ere exist important parallels between infanticide
in Tokugawa Japan and in the ancient Mediterranean world. As Scheidel observes, Greek and Roman authors thought it noteworthy that certain foreign peoples—Egyptians, Etruscans, Jews, Britons, and Germans—supposedly raised all their children, rather in the way that the inhabitants of eighteenth-century Japan commented on what regions did and did not practice infanticide.12 In the second century bce, Polybius claimed that his fellow Greeks reared “only one or two of their children, so as to leave them in affl
uence,” a statement that is almost
identical with some descriptions of Eastern Japan’s low-fertility norm.13 In a phrase that suggests that infant selection was routine in his day, Socrates used the decision of whether a newborn child was “worth raising” as a metaphor for examining the merits of an idea.14 Th
ese parallels do not mean that infanticide was
consistently rampant in Greece and Rome. One of the most important lessons from the history of infanticide in Japan is that demographic systems and the attitudes that underpin them can vary greatly over relatively small spaces, and can at times change rapidly.
238 Conclusion
F E RT I L I T Y A N D M O D E R N I T Y
Eastern Japan’s fertility rise does more than remind us of the changeability of demographic regimes. It raises questions about the relationship between modernity and fertility.15 Social scientists have long considered modernization a reliable recipe for low fertility.16 A view that reverses causality and states that family planning is a precondition for successful modernization has also been infl uential. It has long had many proponents among international development professionals, and for good reason.17 It has also been a staple of family planning campaigns, whose posters show large families huddling in traditional poverty and contrast them with wise parents of fewer children surrounded by the accoutrements of modern prosperity, sometimes before a backdrop of futuristic cities.
In Eastern Japan, the arrival of modernity coincided with the retreat of family planning. Over the same decades that Eastern Japan recorded higher and higher fertility, it acquired more and more of the characteristics and institutions that have come to defi ne a modern society. In the course of the nineteenth century, a national consciousness took shape. Economic activity grew ever more diff erentiated, and industrialization was under way by the mid–Meiji period. Th e number
of private schools for commoner children multiplied in the early decades of the century; aft er 1872, they were replaced by a public and compulsory system of primary schools. Beginning in the fi nal years of Tokugawa rule, a boisterous press arose. In the 1870s, the status system was abolished and all citizens declared equally free to choose their profession and rise in the world. Although in Eastern Japan the intensity of state intrusion into the reproductive lives of its subjects actually declined aft er 1870, the state’s desire and capacity to discipline its subjects in other ways probably increased.18 If modernity is defi ned by these and related developments, it was clearly on the march in nineteenth-century Japan.
Th
at fertility in Eastern Japan reached new heights around 1925, a time when it had fallen to replacement level in much of Western Europe, complicates the story of modernity even further. Japan experienced the waves of innovation of the early twentieth century roughly simultaneously with Europe and North America. Th e
Eastern Japan in which men and women who had grown up with three or four siblings reared fi ve, six, or seven children lay in a nation that governed a colonial empire, had introduced universal male suff rage, and prided itself on its victorious conscript armies and its adherence to the gold standard. Most of its children fi nished primary school, and an increasing proportion of its adults left agriculture for factories and offi
ces. Eastern Japan in 1925 crackled with telephone lines, rumbled with hundreds of thousands of motorcars, and in certain corners of its growing cities danced to the tunes of the jazz age. Th
at the rise of this electrifi ed, fast-
moving world coincided, at least briefl y, with nearly unrestrained fertility makes
Conclusion 239
the connections between social and economic development, on the one hand, and fertility, on the other, look less than inevitable.19
A F U T U R E O F M A N Y P O S S I B I L I T I E S
Today, demographers at the United Nations publish projections about fertility rates in the year 2100.20 Some demographic processes, such as population growth, permit moderately confi dent projections a few years into the future, since they depend on existing age structures and are only gradually aff ected by changes in mortality and fertility conditions. Th
ere are by contrast few grounds on which fu-
ture fertility rates can be forecast, other than the extrapolation of a historical narrative.21 Th
e UN projections of fertility are explicitly based on “the past experience of all countries with declining fertility during 1950–2010.” While the UN demographers use sophisticated statistical models to translate these past data series into a probabilistic cloud of future trajectories, the underlying historical narrative is of necessity disarmingly simple: that the past sixty years are a trusty guide to the next ninety because all societies are travelers on the same road from high fertility to low, with the laggards following the same path that the vanguard trod decades earlier.22
Th
e resulting projections, which a wider public tends to read as predictions, have done much to cultivate a sense of complacency about population growth (75 million additional people per year at the time of writing). My powers of foretelling the future are no greater than those of the UN demographers, but I hope that readers of this book will be left with an appreciation of how changeable fertility levels are, and how contingent such change is on worldviews, metaphors, and politics. It is also worth refl ecting on the time scales of the cultural and demographic changes this book has reconstructed. Th
at change can be profound does
not mean that it is necessarily rapid. In parts of Eastern Japan, the culture of low fertility was remarkably persistent in the face of a prodigious eff ort to alter its norms and values. If the history of infanticide in Eastern Japan has anything to teach us about the demographic future, it is that it will be full of surprises, with stubborn trends as well as rapid change in unforeseen directions.
No valid law of demography states that societies must go from high rates of death and birth to low, and in the process describe an arc of equilibrium, rapid growth, and return to homeostasis. Despondency and complacency about the demographic future are therefore equally misplaced. For better or worse, Japan’s population may yet dwindle precipitously, as many of its people today expect.23
Just as in the 1790s, when Eastern Japan’s net reproductive rate was roughly the same as it is today, it is too early to tell. Fertility rates around the world may continue their general decline, but they may equally stagnate or rebound and expose the profound irresponsibility of celebrating the end of the population explosion based on the distant extrapolation of a recent and contingent trend.
240 Conclusion
Th
e experience of Eastern Japan is less a story of linear impersonal forces whose progress is unstoppable than of individual actions and their large intended and unintended demographic consequences. Th
ere was no demographic ambition in
the ingenious idea of tying parishioners to their temples by promising them posthumous deifi cation through the joint eff orts of their descendants and their priests.
Nevertheless, by inviting people to care deeply about the continuity of their line, that idea ultimately became a cornerstone in the eighteenth-century culture of low fertility. Writers of farm manuals and horoscope charts bore no malice toward their readers’ children, but still they seem to have turned many parents’ hearts against infants born at an inauspicious time. Th
e authors of propaganda prints and
population policies—initially a mere handful of men—had every intention of changing the reproductive behavior of others. Th
eir partial success is yet another
illustration of how contingent demography is on ideas and on the individuals who formulate and spread them.
Human agency has little room in the sweeping theories of Modernization and Demographic Transition; at best, it hastens or delays the inevitable. In the account given here, the decisions of individual men and women, in palaces, studies, and earthen-fl oored huts, shaped Eastern Japan’s demography in ways that at diff erent historical moments pointed toward a range of starkly distinct possible futures. Th
e emergence of Japan as one of the most populous societies of the mid–twentieth century was only one of these, and from the perspective of the 1780s, one of the least likely.
O P E N Q U E S T I O N S
Many open questions remain. I have paid scant attention to the role of extramarital conceptions. I have passed over the culture and demography of child abandonment.24 Local elites populate these pages, but I have only touched on their concerns and motives.25 Samurai have only featured as administrators and commentators, but not as parents with diffi
cult reproductive choices of their own. Following the
example of many of my sources, I have tended to mention abortion in the same breath as infanticide. Yet abortion has its own history; to do it justice, a book needs to cross into disciplines for which there was little room in these pages, such as the history of medicine and the history of the body.26
While this book has documented the interplay of discourse and demography over three centuries, proof of causality oft en remains elusive. Th is is particularly
true for the fi rst and the fi nal decades of this study. While it is possible to suggest reasons for why infanticide became so common in Eastern Honshu around 1680, in many other regions of Japan the same package of commercial growth, Funerary Buddhism, and bans on partible inheritance does not seem to have added up to a culture of widespread infanticide. Perhaps the geography of religion and
Conclusion 241
urbanization made the diff erence, as Chapter 2 tentatively suggested; yet only a study that examines several regions in equal detail can hope to move beyond mere hypotheses about why starkly distinct demographic regimes took shape in seventeenth-century Japan.
Th
e fi nal decades of infanticide in Japan leave puzzles of their own. Th at its
incidence declined sharply in the Taishō period (1912–1925), a time not notable for determined eradication eff orts, may refl ect its incongruity with a modern Japan of Imperial Democracy and universal primary education. Perhaps the ideals of domesticity and national strength convinced many people to want more children, irrespective of what they thought of the possible means of limiting their number.
Th
is hypothesis, linking a worldview with reproductive outcomes, is appealing in that it would represent another iteration of the leitmotif of this book. Th is does not
mean that it is necessarily correct; the question cries out for careful examination.
Another riddle concerns the means of family planning. Prostitutes used moderately eff ective forms of contraception in the Edo period, and the role of semen in conception was well understood. Although other women, too, had a clear desire to limit the number of their children, pessaries, condoms, and coitus interruptus do not seem to have entered the conjugal relations of many people in Eastern Japan before the twentieth century. If confi rmed by future studies, this fi nding has large implications for how marital sex was viewed and practiced. For couples who resorted to infanticide, it suggests that sexual propriety was more important than the safety of mothers or the lives of newborn children.
M E A N S A N D E N D S
How we think of such choices hinges on whether we give each newborn the same moral weight as an older human being. If we do, the history of Japan assumes an unfamiliar shape. Th
e key achievement of the Tokugawa shogunate, more than
two centuries of uninterrupted peace, must then be balanced against the violent deaths of millions of infants in many corners of the realm. If the number of lives lost or saved confers importance on a historical topic, few events in Japan’s nineteenth century can compare to the fi ght against infanticide. If, however, we are more agnostic about the moral status of newborn children and willing to temporarily leave aside the perspective of those who died in their fi rst moments of light and air, we can ask how widespread infanticide aff ected the lives of parents and their chosen children.
For all the cosmologies and metaphors that justifi ed infanticide, it is likely that anguish and guilt attended it in some measure. Yet if we read the historical record without the preconceptions of our own time and place, we must conclude that the evidence for parental trauma is at best equivocal: stone statues of an ambiguous purpose, euphemisms that may have refl ected a need to palliate guilty consciences,
242 Conclusion
and the resort to prediction methods that allowed parents to convince themselves that the rejected child would have failed to thrive anyway.27 Th e barrage of words
and images that opponents of infanticide directed at the population was in part designed to induce remorse and guilt, and as we saw in Chapter 9, these eff orts seem to have succeeded in some measure.28 In gauging how much guilt parents really felt, however, it is worth noting that in most accounts, parents carried out the killing of their children themselves, rather than delegating it to a midwife.
At the same time, there are grounds for believing that infanticide actually less-ened the pain parents experienced in their lifetimes. Th e death of an older child
oft en left them devastated.29 As we have seen, even the opponents of infanticide noted that the “merciless” parents whose ways they sought to change loved their chosen children dearly. Th
e best extant fi rst-person account of infanticide,
Tsunoda Tōzaemon’s diary, records not only the killing of three newborns but also its author’s distress whenever his older children sickened or died. Infanticides may have lowered the mortality risks of the chosen off spring by permitting more comfortable intervals between children and perhaps also by reducing the strain on household budgets. Even if infanticides did not improve the survival chances of the child’s siblings, they did reduce the number of children to whom parents could grow attached. Before the twentieth century, roughly every third chosen child in Eastern Japan died before its fi ft h year. An infanticidal parent would suff er such bereavements more rarely than one who raised every son and daughter.30 One possible reading of the lengths to which parents went to determine whether a particular newborn was promising enough to raise is that they were trying to spare themselves the pain of losing a child with whom they had formed a bond.
Whether infanticidal parents actually succeeded in diminishing the pain they experienced is impossible to know. We have no metric to measure one form of distress against another. Even if such a metric could be found, how should we weigh the pain avoided against the pleasures forgone? Would the joy of rearing another child have balanced out the economic strain, the worries about how to provide for its future, the fear and all-too-frequent reality of losing the child?
Would it have off set the guilt of unmet obligations to parents, ancestors, and older children? We cannot know. Th
e fact that infanticide was so widespread, however,
suggests that many parents judged that in certain circumstances killing their newborn child was the lesser evil.
One of the most poignant aspects of the culture of infanticide is that most of its aims could have been achieved by using contraceptive technologies that by the seventeenth century were already known at least in certain sections of Japanese society. Nonetheless, the sacrifi ces the people of eighteenth-century Eastern Japan made to keep their reproductive potential in balance with their means were not necessarily greater than those of many societies that inhabited the same historical moment and did not practice infanticide on a demographically important scale. In
Conclusion 243
the parts of Japan that fell into this category, many younger sons and daughters delayed marriage, accepted dependent employment in distant provinces, or looked for work in cities. In Japan as elsewhere in the world, the lives of urban migrants may have been more exciting than those of people who stayed in the villages, but they were also on average far shorter. Most societies of early modern and modern Europe excluded many young people from marriage, condemned some proportion to lifelong celibacy, and stigmatized women who conceived out of wedlock.31
Th
e bourgeois of Paris and other French cities farmed out their children to wet nurses in the countryside, in whose care they were at least 40 percent more likely to die.32 When parents in parts of southern Germany felt that they had too many children, they treated infants in such a way that 40 percent “went to Heaven” before their fi rst birthday.33 Th
roughout Catholic Europe, it was not uncommon for
a fi ft h of infants to be abandoned to foundling homes, which were typically inept at keeping them alive.34 Each of these systems had its mental and physical costs—
sexual frustration and social exclusion, the foreshortened lives of adults and children, and many shades of guilt.
In the most fortunate parts of the world, two achievements of our civilization, reliable contraception and safe abortions, have potentially liberated us from such trade-off s. For benefi ciaries of these achievements, it is an eff ort to imagine a world in which a newborn baby did not necessarily elicit the fi erce aff ection of its parents, and could be killed without much guilt or fear of social censure. Th is
mental exertion has great rewards in that it helps us take a measure of the distance we have traveled in terms of our means. It also reminds us that in our ends, we are not always so diff erent from people who thought they had a choice whether to welcome or to return their newborn children.