Who Was Responsible for Japan's Dogs?

In 1881, Tokyo's police were instructed to feed certain dogs for a week. An unmarked animal found roaming was to be caught and kept in an enclosure. Its owner could reclaim it by paying twenty-five sen for each day's care. After a week without a claimant, the police could sell it, using the proceeds for feeding and repairs to the enclosure.^[1]

The provision turned a lost dog into several people's business. Somebody had to catch it, somebody had to keep it alive, and somebody might arrive with both a claim and a payment. The policeman holding the dog was not necessarily its owner. The person who owned it was not necessarily the person currently caring for it.

That separation runs through the history of dogs in Japan. The most revealing changes concern the duties attached to human relationships with animals: when a keeper had to provide care, when ownership justified intervention, and when decisions passed to an official. Following those changes from Tokugawa Edo through Meiji regulations and twentieth-century disease control gives a different history from either the rise of affection for pets or an uninterrupted campaign against strays.

A guide to Edo published in 1687 listed a specialist near Yushima Tenjin who trained animals to perform. The Tokyo Metropolitan Archives places the entry beside an order issued four years later prohibiting the training of animals for exhibition, expressly including dogs, cats, and mice. The order also addressed performers using snakes to attract customers for medicines.^[2]

Dogs belonged to a working urban world. Keeping one could be connected with earning a living, not simply with enjoying its company. A prohibition on performances consequently reached into an occupation: it restricted what a trainer could produce and what an exhibitor could show. The presence of such a rule is intelligible only if animals were already being handled, taught, and presented to an audience.

The measure belonged to policies associated with the fifth Tokugawa shogun, Tsunayoshi, who ruled from Edo in the late seventeenth and early eighteenth centuries. Their familiar collective name, shōrui awaremi, concerns compassion for living beings. As the Edo-Tokyo Museum emphasizes, this was a succession of measures rather than one comprehensive statute called the Edict of Compassion for Living Things.^[3]

That distinction matters. Different orders confronted different conduct. In one early measure of 1685, dogs and cats did not have to be tied up merely because the shogun was visiting. In the exhibition restriction, the issue was what people made animals do. A general injunction to show compassion became actionable through rules about particular encounters.

Those rules also make the usual portrait of a ruler indulging his favorite animal inadequate. The policies extended beyond dogs, and a keeper's closeness to an animal did not put that keeper beyond official scrutiny. Indeed, the people best placed to use an animal were often the people whose conduct a protective measure had to change.

The trainer, the exhibitor, and the customer had reasons to value the animal's acquired skills. The government's order challenged the use to which those skills were put. It reached beyond an isolated act of cruelty to interfere with a trade: protection could require people to surrender a source of income as well as refrain from harming a dog.

Other measures made care itself answerable to officials. In his historical account, the physician Takayama Naohide describes owners reporting a dog's coat, sex, and age for entry in a register. Illness, disappearance, and death could also require action or reporting. His account of the regulations includes obtaining treatment for a sick dog.^[4]

The details gave an individual animal an administrative identity. “A dog” was no longer enough: an entry had to remain connected with this animal, whose absence or condition could require an explanation. Recording age and appearance offered ways to maintain that connection without expecting an official to recognize every dog personally.

The register also extended responsibility through time. Feeding an animal on one occasion was an act that could end when the food was gone. Being identified as its keeper attached future events to the same person. If the dog became ill or disappeared, the relationship remained relevant even though the original act of registration was over.

Animals without an effective keeper presented another problem. In 1695, kennels were established at Yotsuya and Nakano. Nakano's municipal history describes the progressive addition of enclosures as accommodation expanded.^[5] The undertaking made protection occupy land. It required a continuing supply of food and the work of people who would care for animals gathered together.

The practical consequence is easy to miss behind the kennels' later notoriety. Forbidding harm did not itself supply a meal. Collecting dogs in one place transferred the immediate work from streets and households to a maintained establishment. The animals could be kept away from some encounters outside while becoming dependent on people bringing resources inside.

Registration and institutional care therefore addressed different weaknesses in the same undertaking. The first sought to keep a recognizable person attached to an animal. The second supplied accommodation when dispersed arrangements were insufficient. Neither allowed the government to dispense with human labor; each reorganized where that labor should occur and who should be answerable for it.

These were not modern municipal shelters simply operating under an older name. They belonged to Tsunayoshi's governing program, and their later history is more complicated than the familiar claim that every kennel instantly vanished when he died. Research by the historian Nesaki Mitsuo has challenged that blanket ending.^[6] For the comparison here, what matters is the demonstrated coexistence of duties imposed on keepers and care organized beyond any single household.

Tokyo's dog regulations of 1873 approached the street through a visible distinction. Owners were to attach a collar and a wooden tag bearing their name and address. The rules threatened unmarked dogs with destruction. They also assigned owners responsibility for restraining dangerous dogs and compensating injury caused to other people's animals.^[7]

The tag made an absent human available to inquiry. A dog encountered away from home could carry information about where an official or an injured party should turn. Ownership was being presented to strangers through an object attached to the animal, rather than remaining knowledge shared only by the people who usually fed it.

That arrangement could work in two directions. A named keeper could distinguish a dog from animals targeted as unclaimed. The same identification could direct costs and accusations back toward the keeper. It was therefore more than a badge of belonging. It made belonging consequential when the dog crossed a household boundary.

The earlier Tokugawa register and the Meiji collar should not be collapsed into one unchanged institution. A register held information in an administrative setting; a tag placed selected information on a moving animal. Their purposes also differed. The 1873 provisions emphasized injuries and dangerous dogs, joining identification to the protection of people and property.

Yet the comparison reveals a continuing difficulty. The authorities wanted to attach an animal's conduct to a person who could be required to act. That person might be absent when the dog was encountered. The administrative solution had to survive the separation between animal and keeper.

Rabies made the limits of private control especially stark. Under the 1873 provisions, other people could kill a rabid dog found in the road, with expenses charged to its owner. Ownership did not reserve every decision about the animal's life to the owner. A relationship that supported a claim to possession could also support a demand for payment after someone else intervened.

The collar thus moved part of the administrative work into ordinary keeping. Before letting the animal enter a space shared with strangers, a keeper was expected to give those strangers a way to identify the human answerable for it.

The revised Tokyo rules of 1881 gave the missing relationship a procedure. An owner looking for a lost dog was to provide the police with identifying details, including size, coat, and kind. A roaming dog without a mark could be detained, fed, and held for reclamation.^[1]

The arrangement connected two streams of information. One began with a person missing an animal; the other began with an animal missing an identifiable person. Descriptions and custody gave the police a means of bringing them together. A collar was useful precisely because it could reduce this work, but the revised rules allowed for the possibility that it was absent.

The one-week holding period gave reclamation a deadline. During that interval, the police temporarily occupied the keeper's position without simply becoming an ordinary owner. Their authority included the possibility of sale after the deadline, while a claimant's route to recovery included payment for the care already supplied.

The daily charge put a price on continuing custody. It converted the period in the enclosure into a bill payable by the person seeking return. The longer the separation lasted, the more days of care stood between loss and recovery. Reclamation required the owner to settle the expense of work somebody else had performed.

Selling an unclaimed animal connected its future to the maintenance of the institution that had held it. The proceeds were assigned to feeding and enclosure repairs. A buyer could acquire the animal while helping fund the premises and care needed for other dogs awaiting collection. The regulation thus tied disposal of one case to the continuation of the service.

The same regulations remained capable of severe intervention. They required restraint and notification where contagious disease was suspected, and allowed the police to kill a dog once infection was confirmed.^[8] The lost animal and the infected animal did not necessarily follow the same route through the rules.

That coexistence prevents an easy chronology from cruelty to kindness. Feeding could form part of policing. Sale could end one household's claim while beginning another's. Disease control could authorize killing within regulations that also required temporary care. The historical question is which classification applied and what it permitted people to do next.

The week in the enclosure also exposes the cost of declaring an animal publicly manageable. Custody required food while officials waited. The fee and sale provisions were attempts to keep that waiting from becoming an undefined obligation with no assigned source of support.

Ownership did not necessarily keep a dog from spreading danger. A study of Osaka's rabies epidemics between 1914 and 1933 found that more than half of the recorded canine cases involved dogs identified as owned. The researchers reconstructed the outbreaks from surviving official and other historical records.^[9]

That finding makes the next regulatory step easier to understand. In January 1915, Osaka revised its dog rules to require registration and owner identification, together with restraint for dogs likely to bite. During an epidemic, dogs had to be tethered when ordered. A dog left roaming contrary to that order could be treated as an unowned stray even if somebody claimed it.

The decisive question had shifted from “does this animal have a person?” to “is that person exercising the required control?” A claimed relationship was insufficient when the animal's movements defeated the purpose for which the relationship had been made administratively useful. The category applied in the street could depend on the owner's conduct.

This also made responsibility immediate. Finding a name afterward might help recover expenses, but it could not prevent the encounter that had already occurred. Requiring restraint sought to change what happened before injury. The collar directed attention toward a person; the tether was supposed to alter the dog's range of possible contacts.

Osaka's mass vaccination campaign began in 1923, alongside continued restrictions on movement. Vaccination added another task for owners and officials rather than making restraint disappear. The same study records repeated orders and a combination of measures, not a single intervention whose independent effect can be extracted from these historical records.

The comparison with Tokyo's earlier rules is therefore about the demands attached to recognized ownership. An address, a reclamation payment, and compliance with an outbreak order could each matter at a different stage. None could substitute for all the others.

For keepers, the growing requirements changed what it took to maintain an accepted place for an animal. Obtaining identification was an initial act; restraint demanded continuing supervision; vaccination required another encounter with the prevention system. Officials could require these acts regardless of the attachment between a dog and the person who kept it.

The epidemiological finding supplies the most important corrective. Danger was not confined to animals without human connections. Any system organized around owners had to make those connections do practical work.

The Rabies Prevention Act enacted in 1950 specified a wider division of responsibility. It required prefectural governors to appoint prevention officers from veterinary-qualified employees. These officers, and authorized dog catchers, had to carry identification. The person acting upon a dog now also had an identity that could be checked.^[10]

The act retained registration and vaccination requirements, but it did not pretend that the legal owner always performed the daily work. For the vaccination obligation and subsequent provisions, it expressly included the person managing the dog when that person differed from its owner. The wording directed duties toward the human actually positioned to carry them out.

Article 8 gave suspected illness a route through government. A diagnosing veterinarian—or the owner where there had been no veterinary examination—had to notify the municipality. The information then went to the prefecture, the responsible ministry, and neighboring prefectures. A case first encountered locally could become relevant beyond the place where the dog was kept.

The reporting chain addressed a problem that no enclosure could solve. Keeping a dog apart from others changed its immediate contacts. Informing neighboring authorities changed what people elsewhere could know and prepare to do. The same event required both physical action and communication.

Such a chain depended on decisions at its beginning. Someone had to recognize a reason to report and act while the case was still uncertain. The statute therefore included suspected infection and dogs bitten by infected or suspect animals. It did not wait for every uncertainty to disappear before assigning an obligation.

This was a different administrative world from Tsunayoshi's registers, although both made information about an animal consequential. The older duties concerned maintaining an accountable keeper under a protective program. The 1950 act joined the household to specialized veterinary authority and territorial reporting in a law expressly directed toward public health.

The extension of government did not make the household irrelevant. It made household action one component of a system that also needed trained officers, records, facilities, and neighboring administrations. A failure at the first encounter could weaken work assigned to people much farther along the chain. Responsibility had become more distributed, even while the law continued to place substantial demands on the person keeping the dog.

One of the 1950 act's most revealing provisions prohibited killing. A dog isolated under its disease-control requirements could not be killed without permission from a prevention officer. Other provisions required the surrender of a dead suspect animal for examination, unless the officer waived it, and authorized killing for diagnostic purposes under specified official approval.^[11]

Read together, these clauses placed decisions about death inside an evidentiary procedure. The point was not that every isolated dog would be preserved indefinitely. It was that an owner or bystander should not independently destroy an animal whose condition had become a matter for official investigation. The act retained an emergency exception where human life was in immediate danger.

Here, keeping a dog alive could serve disease control. Under the earlier compassion policies, restraining human harm had belonged to a different governing purpose. Similar-looking prohibitions therefore need their institutional setting: who was being restrained, which decision was being reserved, and what investigation or care should occur in the meantime?

The law also made that meantime somebody's expense. Prefectures had to establish detention facilities. Owners bore specified costs, including care during detention and return, while other expenditure fell to public authorities. Registration revenue was earmarked for the act's purposes.^[12] The statute distributed powers without making the food free.

This arrangement returns us to the 1881 police enclosure, but with a changed structure around it. Holding an animal still generated work while its status was resolved. The later system specified professional oversight and reporting beyond the locality; it continued to charge particular costs to keepers rather than having the national government undertake every obligation itself.

Across these histories, responsibility for dogs was repeatedly divided. The person claiming an animal, the person feeding it, and the person authorized to decide its fate could be different people. Ownership became useful to government when it identified someone who could be required to restrain, report, pay, or arrange treatment. It ceased to be sufficient when those actions were missing or when disease created claims beyond the household.

The unmarked dog waiting for a week in Tokyo's police enclosure was therefore not merely between homes. It stood between duties: private care had failed to keep it identifiable, public custody had begun, and a deadline would determine what claim could come next. Understanding those duties explains more of its historical position than deciding, from a distance, whether the society around it loved dogs.