The Pardon That Came Too Late: Life, Status and Killing in Japan, 1543–1700

Mercy on the Road

Why Was a Pardoned Boy Executed at Hirado in 1615?

On 23 December 1615 a boy of sixteen was put to death at Hirado for stealing a small boat and taking it to another island. He had been pardoned. The pardon was still on the road.

Richard Cocks, who ran the English factory in the port, had gone to the local ruler and asked for the boy’s life, and the ruler had granted the request. A messenger was sent out to hold up the execution. According to the diary, the executioner would not wait. The boy was dead before the messenger reached him, and his body was cut to test swords.

Four judgments went into that killing and four different people made them. Someone decided that a stolen boat was a capital matter. An English merchant decided the boy was worth saving, and had standing enough in the port to say so to a man who could act on it. The ruler decided to spare him. The executioner decided not to wait. Cocks records no consequence for the executioner, and he does not give the boy’s name.

The entry will not add up into an attitude. A pardon granted and an execution carried out were the acts of two authorities in the same town on the same morning, and a willingness to issue pardons did not make the system that produced the sentence a merciful one. Survival depended on obtaining an exception and then on having it arrive in time.

Out of material like this, European writers assembled a proposition. The Japanese, they reported, were quick to kill people and unusually reluctant to kill animals. It appears in missionary accounts and in merchants’ reports, and its force came from the inversion: human beings were receiving less protection than creatures Europeans ate without a second thought.

The violence behind those reports was not imaginary. Newborn children were killed, offenders were executed, and human bodies were used to test weapons, and all three require historical explanation rather than denial. What the phrase “life was cheap” does is put them in one sentence. They were done by different people, under different authorities, for different reasons, and the phrase delivers its conclusion in place of the relationships that made each of them possible.

Those relationships did not hold still either. Between the first Portuguese landfall in 1543 and the settled Tokugawa order of around 1700, civil war gave way to a political settlement, a foreign religion became grounds for prosecution, and the forms of labor dependence were redefined and renamed. A warrior, a newborn, a captive sold overseas and a condemned thief did not stand in the same relation to the power that could kill them. Inside a single household the authority to decide could belong to somebody other than the person whose life was at stake.

Two Lines for Two Civilizations

Do Multiple Missionary Reports of Violence in Japan Count as Independent Evidence?

In 1585 Luís Fróis set a European custom beside its Japanese counterpart, wrote the two as a matched pair, and then did it several hundred more times. His Tratado worked through clothing, food, religion and household conduct in short statements of that form. One pair opposed European reluctance to kill a person to Japanese reluctance to kill animals. Others compared European and Japanese practice on abortion and on the killing of newborn children. The form made a difference exceptionally easy to remember, which was the point, and exceptionally difficult to interrogate, which was the cost.

Fróis had lived in Japan for years and knew a great deal more than a traveler writing up a short landing. He also belonged to an enterprise whose declared purpose was to change the society he was describing. Both of those things are true of the same sentences. The modern critical edition sets the short statements inside a much larger apparatus of historical explanation, which is roughly the amount of scaffolding each one turns out to require. The catalog itself, and the century of Portuguese observers whose tradition it belongs to, are treated elsewhere on this site.

The pronoun did most of the work. “We” gathered societies divided by religious war, by jurisdiction and by rank into a single moral observer with consistent habits. European judicial killing and European warfare did not stop because a Jesuit had distinguished them from murder. They went into categories his readers never asked him to explain. Japanese actions were more likely to arrive without the corresponding distinctions attached.

The sentence had two sides and only one of them had to account for itself.

The same difficulty ran in the other direction. A reluctance to kill an ox could come from concern for a working animal, from a religious prohibition, from ideas about pollution, or from some combination of the three. Killing a condemned man belonged to an entirely different institutional setting. An observer could notice both accurately and still be nowhere near having shown that the people he watched preferred the existence of animals to the lives of human beings.

Buddhist prohibitions on taking life did extend moral concern past the human boundary. They did not make Japan vegetarian, and they did not restrain rulers and warriors. What a doctrine supplies is a claim against which conduct can be defended, condemned or excused. Christianity ran the same combination from the other side, prohibiting killing while licensing war and execution. Neither religion’s stated prohibition, taken by itself, measured how securely any particular person was protected.

The comparison did register something real, which was what each side found unbearable in the other. Europeans accustomed to slaughtering animals for food met objections to the killing of useful ones. Japanese observers found the treatment of human beings aboard foreign ships intolerable. Each objection picked out a relationship between a living creature and whoever claimed authority over it, and the comparison becomes informative at precisely the point where those relationships are specified.

Then there is the question of how many witnesses there actually were. Missionary letters were copied, translated, excerpted and mined by later authors, so several similar descriptions are not automatically several independent observations. A vivid formulation travels further and faster than the event it came from, and the unanimity of the European record has to be examined passage by passage rather than counted by the nationalities in a bibliography.

None of which makes the testimony unusable. It changes the question put to it. An account of a particular killing, a report of what local people said, and an assertion about what all Japanese believed carry three different evidentiary weights, and a single author will place all three on the same page.

The Children a Household Would Raise

What Does Fabian Drixler's Research Show About Infanticide in Early Modern Japan?

The word was mabiki, thinning seedlings. A farmer pulls some of a row so that what stays in the ground has enough to grow on, and the metaphor did the same work for a household that limited the children it raised so that the ones it kept could be fed. It placed a newborn inside a decision about the family’s future rather than treating the birth as a question already settled. That is a claim about responsibility. It was made at the expense of the child excluded from it.

Missionaries reported the practice steadily: newborn children killed, pregnancies ended. Fróis connected some infant deaths to a household’s inability to feed another child and reported that girls were the more vulnerable. Valignano presented poverty as an excuse for an act he condemned. The difference in wording is not decorative. What one passage records as a material constraint becomes, in the other, a refusal to take up an obligation.

Children accordingly occupied a prominent place in the missionary indictment, and the reasons for that ran past compassion. Baptism and salvation made the death of an infant urgent to a priest in a way it was not urgent to anyone else keeping records, and the Christian prohibition supplied grounds for condemning it. A claim that conversion had ended infanticide in a district was simultaneously a claim that the mission was working. An assertion that Christian Nagasaki no longer killed infants is not the same document as a comparison of births and deaths in Christian and neighboring non-Christian households, and it was the first kind that got written.

Japanese records eventually supplied the second kind. Population registers let historians follow households across years, count the ages and the sexes of the surviving children, and reconstruct patterns of fertility. Fabian Drixler’s Mabiki sets that evidence beside the Japanese writing about reproduction in a study running from 1660 to 1950, and the long chronology is not padding. It is the argument. Practices changed, and so did what people were willing to say about them.

Drixler’s evidence begins in 1660.

Fróis was writing in the 1580s. Much of the detailed demographic work concerns later periods and particular regions, so a sex ratio from an eighteenth- or nineteenth-century village in eastern Japan can support an argument about selective child-rearing in that village and cannot establish a rate of infanticide in the Kyūshū communities Fróis knew two centuries earlier. Regional evidence stays regional even when its findings resemble an older description.

Poverty belongs in the history without finishing it. The upkeep of a household, the number and sequence of children a family wanted, and expectations about inheritance all bore on a decision, and so did the fact that a household is not a single mind. A report that “the family” chose to limit its births reveals nothing about how the power to choose was distributed among parents, older relatives, and the women attending the birth. European accounts that laid the act at the mother’s door are silent on that point.

The registers have limits of their own. A child who died before registration might never appear in one. A gap between two surviving siblings can be read as a decision or as ordinary infant mortality, and the document does not always distinguish. Demographic inference works on patterns across many households, and it does not recover the fate of any particular missing person.

Where the later evidence is strongest is as a corrective to the sweeping version. It shows that infanticide could be frequent, that local expectation helped sustain it, and that its prevalence varied by place and by decade. It also shows why a national population figure cannot explain what was decided in one room after one birth.

Before a Name Was a Place in the Family

What Does the Saying 'Until Seven, a Child Belongs to the Kami' Actually Mean?

Until the age of seven, the saying went, a child belonged among the kami, the divine beings, and not yet altogether to this world.

The saying is quoted often enough to have become an explanation, and it will not carry the weight. It is not a nationwide legal permission to kill children under seven. No ceremony marking a child’s entry into a household meant the same thing in every family. A proverb about the vulnerability and the religious standing of small children is not an account of what parents did.

What the underlying ideas do explain is how such a decision could be made intelligible to the people taking it. Language about returning a child to another realm distinguished the killing of a newborn from the killing of somebody already established in a community. Naming, ritual recognition and incorporation into a household gave social existence a set of stages that did not line up with biological birth. A person could be alive without yet being, in the full sense, a member.

That language could reduce the apparent finality of a death without reducing anything else about it. It coexisted with grief, with fear and with disagreement. Calling an act customary does not establish that the people performing it were untroubled, and the later remembrance of a dead child does not establish that everyone present had understood its status identically beforehand. Action and commemoration answer different needs, at different times.

The modern evidence is where this goes wrong most often. Mizuko kuyō, the memorial rites associated with pregnancy loss and abortion, are regularly produced as proof of a timeless Japanese accommodation between ending a life and acknowledging it spiritually. Helen Hardacre’s research places the popularization of those rites in the 1970s and examines the religious marketing that came with them.

The rites are from the 1970s.

Older beliefs and devotional practices did exist, and Jizō’s associations with children and with the dead supply a genuine religious setting. But a roadside Jizō image, a memorial service in a modern temple, and a sixteenth-century Jesuit report of an infanticide cannot be assembled into one continuous explanation without somebody tracing the connections between them. The statue gives no access to the thoughts of a parent Fróis wrote about.

The same caution applies from the other side, to any account in which Japanese religion had quietly settled the question for everybody. Drixler documents a long later struggle over infanticide, conducted from the late eighteenth century onward with moral argument, with images and with government intervention, and institutions that supported children, monitored pregnancies and set out to make an accepted practice unacceptable. A practice that requires justification and attracts organized opposition belongs to a disputed social order, not to a national character. Those developments lie well beyond the Nanban period, and their use here is negative: they show what goes wrong when later evidence is read as the unfolding of something that never changed.

For the earlier period the defensible conclusion is narrower and much less quotable. The missionaries described real instances and real customs. Their explanations of them were frequently incomplete. A different understanding of infancy helps to explain how the act could be permitted, and it does not establish that Japanese people valued every life less, or that the killing left nothing behind in the people who survived it.

A Sword and the Authority Behind It

Did Samurai Really Have an Unrestricted Right to Cut Down Commoners (Kirisute Gomen)?

A blade’s quality was established by cutting something with it, and the European accounts of how that was done in Japan turned the human body into material.

The cutting was performed on a corpse, or, in some accounts, on a condemned prisoner who was still alive. The distance between those two circumstances is the whole of the question, and a visitor could put them on the same page as evidence of the same cruelty.

Tameshigiri, test cutting, covers an activity whose targets and institutional settings varied. Tsujigiri, usually rendered as crossroads cutting, means attacks on passers-by. The occurrence of the second establishes no lawful entitlement to it, and the authorization of testing on the bodies of the condemned does not convert every reported killing into an official test. Stories about lords going out at night need evidence beyond their usefulness as illustrations of a violent reputation.

Cocks’s entry joined the punishment to the testing. A condemned body became available for another use, and the pardon, which had been granted, did not prevent it. The diary does not disclose whether the testing was intended when the sentence was passed. What it does show is why a foreign observer would connect the authority to judge with the treatment of a person as an object. Both had happened to the same boy inside a day.

Rank governed exposure. Masters and lords held powers that dependents did not, and foreigners recorded killings inside households that they believed would go unpunished. A diarist’s expectation of impunity, however, is an expectation. Without the proceedings that followed, it establishes neither what a court decided nor which rule was applied.

Kirisute gomen is the formula that has suffered most from retelling, routinely expanded into an unrestricted samurai right to cut down any commoner who gave offense. The expansion erases the difference between a defense recognized in specified circumstances and a general license, and it imports later legal formulations into earlier decades. An eighteenth-century code can illuminate the history of Tokugawa law, but only for as long as its date and its relationship to earlier practice stay visible.

Kenka ryōseibai, the rule punishing both parties to a fight, has been compressed the same way. Its existence does not prove that every private quarrel in Japan ended in two death sentences, or that restraint was always pointless. Keiichi Kōno’s study of the Mōri examined the measures that house took against fighting among its retainers and how those measures changed, found the circumstances of a fight being attended to, and warned against assuming that one formula had hardened into a universal rule.

Regulation of that kind exposes a tension inside warrior government. A lord needed armed retainers, and the private disputes of armed retainers disrupted the order he was trying to command. Assertions of honor did not always coincide with the interests of the house above them. The same act of violence could demonstrate standing in one relationship and challenge authority in another.

So the history here is larger than a contrast between European compassion and Japanese brutality. It is about the distribution of armed power and the restraints placed on it, a thing that changed. The protections available to a merchant with access to a ruler were not the protections available to an unnamed boy accused of stealing a boat.

Same port. Same day.

The People Carried Out of Japan

Why Did Hideyoshi's 1587 Order to Free Enslaved Japanese Include a Refund for the Buyers?

In 1587 Hideyoshi demanded that the Portuguese give back the Japanese people they had bought and shipped abroad, and, where that could not be managed, that those still in Portuguese hands be freed and the purchase price refunded to the men who had bought them.

The second half of that demand is the interesting one. An objection to the destruction of families arrived attached to an acknowledgment that money had changed hands and would have to change hands again. The same persons appear in one document as subjects to be recovered and as purchases whose release would require a settlement of accounts. Neither the ruler’s concern nor the financial mechanism cancels the other, and both of them are in the text, which survives in Fróis’s history.

The trade itself, and the Jesuit reckoning that eventually shut down the permits, are told in full elsewhere on this site. What belongs here is the narrower question of who was protected by what.

Portuguese demand arrived into existing supplies. Warfare, kidnapping and sale already connected Japanese brokers and Japanese households to foreign merchants, and the resulting traffic crossed several jurisdictions at once, so that a person’s departure from Japan could put recovery beyond the reach of the authority that later objected to the sale. Export changed the practical possibilities of protection, and not merely the location of the confinement.

Missionaries stood inside that commercial world rather than beside it. Their relationship to the trafficking ran through accommodation, through attempts at regulation and through opposition, and it was never one position held consistently by every Jesuit. Rômulo Ehalt’s research follows how Japanese categories of bondage were translated into European legal and theological ones, and how the translation helped connect a local practice to an expanding trade.

That work makes two convenient descriptions unusable. Calling every dependent servant a chattel slave obscures real differences in duration, in transferability and in obligation. Calling Japanese bondage a form of service conceals the coercion, and the processes by which a person became saleable. The comparison has to follow the conditions of dependence rather than assume that a familiar word has settled them.

The theological verdicts had practical effects, though not always on the person they were about. A judgment that somebody could lawfully be held in bondage protected the conscience of the buyer and left the captive exactly where he was. Arguments about the validity of an enslavement were part of how the trade operated, which puts them beside the missionaries’ declarations about the sanctity of life rather than in a compartment marked commerce.

Hideyoshi was not abolishing captivity. The demand concerned the removal of his own subjects and the severing of their relationships at home, and it proclaimed no equal right to freedom for every captive in every jurisdiction. Protection followed political membership. People outside that membership had no comparable claim on a ruler’s intervention, which is why the mirror between the European and the Japanese accusation is uneven: each side could identify the suffering produced by the other’s arrangements while accepting the coercion inside its own. That unevenness marks the edge of what each side was claiming to protect.

For the captive, the edge determined who could intervene. A lord seeking the return of a subject, a missionary disputing a sale and a buyer defending his title could all recognize the same person and disagree entirely about what the recognition obliged them to do.

Recognition was not release.

The Lives Claimed by Religion

Was Seppuku Always a Freely Chosen Death, or Could It Be a Punishment?

Baptism changed nothing about the person. It changed what could be required of him.

Japanese converts entered communities in which birth, marriage and the deaths of children were interpreted through Catholic teaching, and taking that teaching on could set them against a ruler or a household demanding a different allegiance. Once Christianity was prohibited, the conflict became prosecutable. A believer who went on practicing could be punished for doing what the Church understood as an obligation, so that a missionary account of steadfastness and an official account of disobedience describe the same refusal, on the same day, by the same person. Salvation and loyalty did not license the same act.

Martyrdom also complicates the proposition that Europeans simply valued continued biological life above everything else. Catholic teaching could commend accepting death rather than abandoning the faith while condemning suicide, and the distinction ran on intention, agency and religious obligation. A martyr did not hold his life to be worthless. He held something else to be worth more.

Warrior death needs the same care. Seppuku, the ritual suicide associated with warrior status, could be a punishment, and it could be performed under social and political pressure that left the verb “chose” doing very little work. Its presentation as honorable does not establish that it was freely undertaken in every instance. That a condemned man retained dignity through the form of his death describes one feature of a procedure. It does not describe the balance of coercion and choice inside it.

The Hagakure is not a transcript of what sixteenth-century warriors believed. It was compiled in the early eighteenth century, and a text urging readiness for death is an argument about service, made from a particular position at a particular moment. Exhortation implies a gap between what the author is demanding and what his readers might otherwise do.

Nobody writes that book for men who are already doing it.

Its most quoted sentences have since made the demand look considerably more universal than its setting warrants. The comparison with martyrdom has limits of its own: the doctrines classified the two acts differently, and the institutions that rewarded them were not the same institutions. What both require is attention to the reasons a death was accepted or imposed, and neither is explained by ranking whole populations according to their attachment to survival.

Missionary polemic against the Buddhist clergy raises a different problem. Reports of sexual misconduct and of terminated pregnancies could carry an observation about reproductive practice and an allegation against religious competitors in the same sentence. A claim introduced as local hearsay does not become direct evidence because several missionaries repeated it. That abortion existed does not by itself prove an organized clerical trade in the medicines, and the mobility of Japanese religious women does not demonstrate sexual misconduct. Condemning a practice could win converts; defending an institution could preserve its authority. Both interests are real without making every accusation false or every denial dishonest.

The dead were subject to the same arithmetic. Ritual attention to a corpse could coexist with unequal treatment of bodies, and with the delegation of the difficult work to groups the society stigmatized for doing it. Fear of pollution is not necessarily indifference. Invoking pollution also does not explain the circumstances of every corpse a traveler walked past, which requires evidence about that place, that death, and the people whose business it was to deal with it.

The Record That Leaves a Person Out

Why Were Japan's Population Registers Originally Created to Find Christians?

The paperwork that lets a modern historian count children was, in part, created to find Christians.

By the later seventeenth century Japanese authorities had developed forms of registration that made religious affiliation and household membership continuously visible, and some of the documentation later used by historical demographers came out of exactly that scrutiny. Those documents were instruments of government before they were evidence about fertility, and they recorded what the institutions compiling them wanted recorded.

A register can give a person’s age, position in a household and temple affiliation, and say nothing whatever about affection, fear, or a birth that nobody entered. A Jesuit letter, written for a different institution with different requirements, might preserve a death precisely because the death demonstrated the need for baptism or the success of a conversion. Each archive made a different set of lives visible, and each made them visible for a reason.

Which is why the question of what a human life was worth in Japan cannot be answered by collecting the most severe incidents out of one archive and the most generous doctrines out of another. European prohibitions have to be compared with European conduct. Japanese justifications have to be compared with the people they were used on, and with the Japanese who disagreed with them. The unit that does the explaining is usually a relationship of power, not a civilization.

A newborn could be excluded from the children a household chose to raise. A condemned offender’s body could become available for testing a blade. A captive could be recognized as a person and still carry a price that somebody would have to refund. A Christian could be ordered by one authority to keep the faith and by another to abandon it, with a penalty attached at both ends. The categories did practical work. They determined who could act, who could object, and which objection counted for anything.

None of which establishes that the people involved felt nothing. Attachment and unequal protection are entirely compatible. Parents invested heavily in some children and refused another. Rulers granted individual pardons and maintained severe punishments, in the same jurisdiction, on the same morning. An institution does not require universal indifference in order to produce deaths. It requires accepted powers, workable classifications, established procedures, and people able or willing to carry them out.

The European witnesses are most useful where they preserved the particulars, including the particulars that resisted their own larger claims, and weakest where one reported event becomes proof of what everyone in Japan thought. A diary will contain both kinds within a few lines of each other.

Cocks left an unusually narrow record of the boy condemned at Hirado. It preserves his age, the theft he was accused of, and the attempt to save him. It gives no name, and nothing at all from his family.

The next paragraph records wine, bread and oranges, sent out to the master of a junk that was preparing to sail for Siam.

Sources & Further Reading

Cocks, Richard. Diary of Richard Cocks, Cape-Merchant in the English Factory in Japan, 1615–1622. Edited by Edward Maunde Thompson. Vol. I. Hakluyt Society, 1883. The entry for 23 December 1615 supplies the granted pardon that did not arrive in time, the use made of the body afterward, and the wine and oranges in the paragraph that follows.

Drixler, Fabian. Mabiki: Infanticide and Population Growth in Eastern Japan, 1660–1950. University of California Press, 2013. The demographic evidence behind everything said here about what the population registers can and cannot show, and the source for the later Japanese campaign against infanticide; the introduction was consulted and the full study is further reading.

Ehalt, Rômulo. “Geninka and Slavery: Jesuit Casuistry and Tokugawa Legislation on Japanese Bondage (1590s–1620s).” Itinerario 47, no. 3, 2023, pp. 342–356. Traces how Japanese categories of bondage were translated into European legal and theological ones, and why neither “slave” nor “servant” settles the question by itself.

Fróis, Luís. The First European Description of Japan, 1585: A Critical English-Language Edition of Striking Contrasts in the Customs of Europe and Japan. Translated, edited and annotated by Richard K. Danford, Robin D. Gill and Daniel T. Reff. Routledge, 2014. The critical edition of the Tratado, carrying the historical apparatus that the paired sentences require; the contrasts are paraphrased here rather than quoted from its English wording.

Hardacre, Helen. Marketing the Menacing Fetus in Japan. University of California Press, 1997. Dates the popularization of mizuko kuyō to the 1970s and describes the religious marketing that accompanied it, which is what keeps the rites from being read backward into the sixteenth century.

Kōno, Keiichi. “A Study of the Mōri’s Method of Dealing with Fighting between Vassals.” Legal History Review, volume 2000, no. 50 (published 2001), pp. 87–116, English abstract p. en6. Examines how one house’s measures against fighting among its retainers changed over time and warns against treating kenka ryōseibai as a rule that applied everywhere; the English abstract was consulted, not the full Japanese text.

Sousa, Lúcio de, translator. “Source: Portugal and the Japanese Slave Trade.” Teaching Medieval Slavery and Captivity, undated. Supplies Hideyoshi’s 1587 demand as preserved in Fróis’s História de Iapam, including the proposal to compensate the purchasers when captives were released.

Willman, Olof Eriksson. The Journal of Olof Eriksson Willman: From His Voyage to the Dutch East Indies and Japan, 1648–1654. Translated and annotated, with an introduction, by Catharina Blomberg. Brill, 2014. A later European account of Japanese justice and punishment, valuable for Blomberg’s apparatus; its dependence on Caron and its more sensational anecdotes are reasons to read it beside other witnesses rather than as independent corroboration.