Introduction
It happened in 1952 in a small village at the foot of Mount Fuji (case 1). For years, a village leader had gone from house to house asking residents whether they planned to use their election ticket, the form that enabled them to vote. If not, he asked if they could give it to him, lest it go to waste.
Teenager Satsuki Ishikawa was outraged. This was election fraud, and, still in middle school, she wrote an article for her school newspaper. School administrators collected every copy and destroyed them. Two years later, she decided to try again. She could not complain to city hall, since the man collecting the tickets worked there. She thought of complaining to the electoral commission, but she worried that they might be in on the scheme. She thought of the police, but she did not trust them either. Instead, she wrote to Asahi, a nationwide newspaper. The paper sent reporters to the village, and the election fraud hit the national news. The police arrested the guilty village leaders. The community responded by ostracising the Ishikawa family.
Readers of the Asahi wrote in from all over the country in support of Satsuki. Her teachers and classmates encouraged her too. But the Ishikawa family grew rice. In the pre-mechanised 1950s, transplanting rice required the assistance of the community. When it came time to transplant the Ishikawa fields, no one came to help.Footnote 1
Legal scholarship traditionally focuses on the courts. Scholars in both sociology and economics have increasingly treated informal social sanctions – norms – as a community's primary mechanism for controlling deviance. Formal legal sanctions, both civil and criminal, are more costly and secondary. Of the informal sanctions, ostracism is one of the most severe. Yet outside of the psychology laboratory, studies of ostracism barely exist.
We study the institution of ostracism within the tradition of socio-legal studies. To this end, we examine virtually every published, non-trivial legal opinion related to ostracism in Japan that we have been able to find. We focus on modern Japan because it has a long tradition of using ostracism to enforce social norms.
The cases involving ostracism include very few where a community actually used ostracism to try to control an anti-social member. Instead, most cases arose from disputes in which the community used ostracism opportunistically – to shield community-wide misconduct, to extract property from someone, or to harass a rival faction. Plaintiffs in civil suits did not primarily file suit for damages, and prosecutors in criminal cases did not seek imprisonment. Rather, we argue, they filed actions for informational purposes, aiming for the court to publicly certify their version of events. At root, they illustrate the informational role that courts so often play in modern society, a role aptly described by legal scholar Sadie Blanchard as that of ‘information intermediaries’.Footnote 2
We begin by reviewing the scholarly literature on formal and informal sanctions for deviance, with a particular focus on ostracism. We then relate this literature to the rapidly developing theory of the role that courts can play as informational intermediaries. Using Japan as our example, we describe the instances of ostracism that appear in modern court opinions. A formal model of the interaction between ostracism and its judicial review is included in the annex.
Ostracism, litigation, and information
Ostracism as norm enforcement
Ostracism is an informal social sanction against deviant behaviour. Informal sanctions are a community's primary sanction against deviance. Formal criminal and civil litigation serve only as secondary sanctions in the event that a deviant member fails to respond to informal sanctions. And among the many informal sanctions that a community can impose, ostracism is one of the most severe.
Much of the scholarship on deviance begins with Emile Durkheim, who classically presented deviance as a phenomenon that communities sought to constrain through their networks of informal ties. Communities did not so much ‘control’ an individual as, in Thomas J Bernard's words, provide a ‘structure of self-interest … such that people find it in their interest’ to follow community norms.Footnote 3 Edwin W Sutherland pushed sociologists to shift their attention from the individual to the society. He largely abandoned Durkheim's focus on individual-level factors. Instead (in the words of Laub and Sampson), ‘crime was viewed by Sutherland as a social phenomenon that could only be explained by social (i.e., non-individual) factors’.Footnote 4
Within economics, Gary S Becker brought a deliberately spare model of crime that turned exclusively on the individual. Scholars would do best, he argued, to posit a potential criminal who weighed his private benefit from a crime against his expected costs, and chose to commit crime when the net result was positive.Footnote 5 Although Becker focused on legal (particularly criminal) sanctions, modern scholars in the economic tradition examine not just legal rules but also informal sanctions. The literature is massive, but the classics include studies by economists Janet Landa and Avner Greif, and work by law professors Robert Ellickson and Lisa Bernstein.Footnote 6 Like sociologists, they observe how citizens help preserve public safety and order through informal social sanctions.
Curiously, perhaps, scholars have written almost nothing about actual cases of ostracism in modern societies. To be sure, experimental psychologists have been active in exploring the way people react to ostracism. Over the course of the past two decades, they have conducted a wide range of experiments. Most have been associated in one way or another with Kipling Williams at Purdue University.Footnote 7 A few scholars have studied the Amish practice of ‘shunning’.Footnote 8 Scattered ethnographies detail practices in hunter-gatherer societies.Footnote 9
Ironically, of the very few studies of ostracism in any modern democracy, the best may be the study of Japan by anthropologist Robert J Smith.Footnote 10 Smith focused on documentary evidence of eight cases: refusing to help maintain a footbridge, stealing millet, publicising election fraud (case 1), violating a village rest day to work on a building project, stealing potatoes, falsely claiming to police that village authorities had cut down a tree that the target owned, and refusing to attend the celebratory send-off of an army draftee in 1937. Only the last of these resulted in a lawsuit.
In this study, we focus on ostracism cases that do result in lawsuits. A Westlaw search in mid-2023 for murahachibu, the most common Japanese word for ostracism, returns 167 cases. The more formal kyodo zekko returns 22. Hamon, used especially by the organised crime syndicates, returns 225. From these cases, we extract the very few cases that actually concern ostracism by communities and include enough detail to discuss.
Ostracism as opportunism
Theory in the Durkheim and Becker tradition suggests that communities ostracise to enforce norms. Virtually none of the cases in our Japanese data set fit this template. Many of the cases do not involve attempts by a community to control anti-social deviance at all. More often they involve opportunistic tactics by the community itself. In several cases, the community tried to punish a member who attempted to stop the broader patterns of community misconduct (cases 1, 14, 15, 16, 20, 21, 22, 23). In some cases, the community used the ostracism to extort property from a minority of its members (cases 4, 17, 18, 19). In others, an opportunist manipulated the mechanism of ostracism for private advantage (case 13). And in still other cases, the community split, one faction using ostracism to penalise the other (cases 11, 12, 24, 25). We posit that in many of these cases the ostracised victims sue to demonstrate their innocence. They sue to obtain a judicial statement of what happened. They sue for that statement because they believe it will show that they have not violated any social norms at all.
Non-random selection
We have what we believe is close to a full set of those cases involving ostracism that resulted in a full, published court opinion. We believe these opinions reveal some of the ways in which community disputes arise, and some of the strategies by which participants in these disputes try to engineer the results they want. Necessarily, these strategies affect which disputes will end up in published court decisions. Crucially, we do not have a random sample of community disputes, of disputes that led to ostracism, or even of ostracism that led to litigation.
A first level of selection in the observed cases concerns the decision to resolve a dispute by ostracism or its threat. The effect of ostracism varies from place to place. It hurts more in an isolated village than in an anonymous city. It hurts workers in a tightly structured, regulated profession such as medicine more than those with widely useful skills such as manual labour. It hurts merchants who trade on credit within broad networks with high levels of social capital more than merchants who trade in cash on the spot market. Even within a given industry, the consequences of ostracism can vary according to location. For example, although both Japan and America were overwhelmingly agricultural in the nineteenth century, ostracism would have a much harsher impact on Japanese farmers than American farmers. Japanese farmers grew rice, and the technology of pre-mechanised wet-rice farming was harshly communal.Footnote 11 Without community help, a Kyushu farmer would have been forced to abandon his field and relocate to town. Kansas farmers, in contrast, grew grain, and could more easily manage without community help.
We are not aware of any studies that try to measure how frequently various communities use formal ostracism. We know of no studies of its incidence in the US, and although we have a little more information about Japan, it does not support the claim that ostracism is widespread. We searched the electronic databases of two of the leading national newspapers, Asahi and Mainichi, in 2020 and found almost no cases of ostracism beyond those discussed in this article.
Japanese government records contain evidence of about 20 instances of ostracism per year, as shown in Table 1. Even if he chooses not to file suit, someone who believes that others have infringed his human rights may report the offence to the local office of the Ministry of Justice. Ministry officials will then investigate. If they believe that the complainant's rights have been infringed, they may try to help him obtain relief. They explain; they negotiate; they mediate; and they introduce private or public organisations that might help.
Table 1. Cases of murahachibu reported to human rights offices of the Ministry of Justice (by year and region)Footnote 13
In 2018, Japanese residents reported 19,600 putative human rights violations to the Ministry.Footnote 12 Few of these concerned ostracism. Table 1 gives the number specifically of ostracism cases that people reported to the Ministry of Justice human rights offices over the past five years. Of course, we have no reason to believe that people reported all cases. Subject to this qualification, however, they did report around 20 cases each year from 2015 to 2019. The Ministry does not report details beyond location. Cases disproportionately come from northeastern Japan (the Tohoku region), central Japan (the Chubu region), and the areas adjoining the Seto Inland Sea (the Chugoku and Shikoku regions). While these areas do exclude the highly urbanised areas of Tokyo and Osaka, they have little else in common.
A second kind of selection observed in the cases is that, even though ostracism is available as a punishment in a community, it, like prison, will not be used unless someone makes a mistake. Criminals do not expect to end up in prison, and the ostracised do not expect to end up shunned. Two people interact only when doing so benefits them both. Ostracism forces them to cease these voluntary interactions and turn towards autarky. If they stop interaction to enforce a norm, they necessarily incur a cost. Ostracism destroys gains from trade. Thus, both the dissenter and his community gain if they can avoid ostracism and settle the dispute peacefully. Provided they both anticipate the same outcome from pushing a conflict into ostracism, they both benefit by avoiding it via compromise or surrender. They ordinarily settle their dispute according to the expected losses to each side from ostracism actually being inflicted.
Crucially, however, a dissenter and his community can only reach this mutually beneficial negotiated settlement if they can agree on what will happen if they do push the dispute into ostracism. In a community that has remained unchanged for a century, with stable families, economy, roles, and power relations, such consequences are easier to predict. The parties know how much each community member would lose from ostracism. They know each other's feasible alternatives. They know whether anyone in the community would refuse to go along with the ostracism or would secretly violate it. With shared estimates of the consequences of ostracism, they would settle without having to carry out the threat.
When a community is in transition, however, the parties are less likely to agree on what might happen. A dissenter may believe he can find profitable employment in a nearby city; the rest of the community may know better. The community may believe it can cheaply replace the dissenter's services; the dissenter may know how much they will miss his talents. With change comes uncertainty.
Most of the disputes in our set involve communities in flux. Many involve agricultural villages located near rapidly expanding municipal centres. Necessarily, in these cases both sides to a dispute will need to estimate the alternatives available to each other in the greater municipal area. They will be relying on information that is less certain than in a remote and stable agricultural village.
A third level of selection comes when ostracism has occurred, but the person ostracised must decide whether to go to court. This point is true of litigation generally, and involves the same logic as selection into ostracism. Litigation is expensive for both parties, so people avoid it when they can. Provided they agree on the probable outcome at trial, they settle their disputes ‘in the shadow’ of that outcome.Footnote 14 If the person ostracised knows he would lose in court, he will not bother to try.
Some of the cases in our set involve criminal prosecutions, where an analogous logic applies. Prosecution is costly for both the state and the defendant. Provided both agree on the probable outcome upon trial, both have an incentive to avoid it. In the US, they avoid it by plea bargaining. Japan has not allowed plea bargains, but prosecutors do generally recommend lighter sentences for defendants who plead guilty. This process does not quite replicate plea bargains, but the dynamic is obviously similar (and Japanese law does now allow formal plea bargains in limited cases). The result is that prosecutors bring to trial only a small, non-random minority of people arrested by the police.
The informational logic of litigation
The plaintiffs in our set seem not to have sued for large amounts of money. As these are mostly appellate decisions, few cases give the amounts actually recovered by the plaintiffs. Those that do give the numbers, however, do not report large amounts.
The prosecutors in our set seem not to have demanded heavy penalties either. Again, as these are appellate decisions, most do not report the penalties imposed. Those that do, however, report only suspended sentences.
Instead, the plaintiffs and prosecutors seem to have filed the suits for the informational role that courts play. As in Sadie Blanchard's work tying the courts to the economics of information, we believe they sued to obtain public certification and dissemination of their stories. They sued to capitalise on what Blanchard called the court's role as an ‘information intermediary’.Footnote 15
In the course of litigation, courts produce information. When they ultimately decide a case, they certify that information, announce it to the public, and disseminate it. All told, courts produce, certify, and publish information.
Crucially, this litigated outcome can change the character of the public understanding of a dispute. If a victim sues and wins, the dispute becomes the judge's word against that of the village leaders. And if the dispute has any news value, the local press will convey the judge's word broadly. Through litigation, a victim who wins both can increase the credibility of his account, and convey that information broadly.
This certification and dissemination matter because of the impact that information about a dispute can have on the relative reputations of the leaders and the dissenters. Those reputations, in turn, determine the capacity of both groups for advantageous trade. The more public the information, the greater the impact on future economic transactions.
Courts, explains Blanchard, increase a ‘[r]eputation's effectiveness’ because they spread ‘information about past behavior … more widely among potential counterparties.’Footnote 16 One of us has written about the ‘stigmatization’ function of punishment.Footnote 17 What the courts illustrate in this context is their capacity to ‘destigmatize’ when communities punish the wrong party.
Note that what Blanchard calls the informational logic to a dispute interacts with the selection of disputes for litigation. Suppose a plaintiff sues to obtain public certification of his version of the dispute. If the plaintiff and the village leaders agree that a court will ultimately decide in the plaintiff's favour, both gain by negotiating an out-of-court settlement in which the village leaders publicly acknowledge the plaintiff's version of the dispute. On the other hand, if both agree that the court will side with the village, it would be foolish for the ostracised person to go to court and have the accusation of deviant behaviour officially certified.
Consistent with the cases in our set, plaintiffs and prosecutors in informationally driven cases will tend to select cases in which the ostracism appears improper. To the extent that plaintiffs sue to have the court publicly endorse their claims of innocence, they will not sue if the court would instead shame them. Prosecutors, too, will select cases for the message they want to convey to the rest of the community. In no country do prosecutors have the resources to pursue all (or even most) cases referred to them by the police. Instead, they focus on the cases that most forcefully reinforce the norms they want people to follow. In the context of ostracism disputes, they will focus on the most egregious cases. If, on the other hand, the village gets ostracism right, the prosecutor will leave it alone.
The Cases
Conventional cases
Let us start with a Supreme Court case from 1921. The case concerned a rural hamlet that had received subsidies from the larger village (mura) and county (gun) governments to build a road (case 2). Sadaji Kodama owned part of the land over which the road would pass. He refused to convey it to the community. Whether his objection was that the community wanted more land from him than it took from others; whether it offered them higher compensation than it offered him; whether the road benefited others more than him, we cannot tell from the court opinion. Three times, however, the county head visited Kodama to plead with him, but to no avail. After seven or eight years passed and the hamlet had still to finish the road, the county withdrew its subsidy. Furious, the hamlet members assembled and voted to cut all ties with Kodama and with anyone – ‘whether or not related by blood’ – who might continue to have contact with him.Footnote 18 Kodama sued in response, and (as discussed in more detail below) the court declared the ostracism a tort.
The Supreme Court faced a similar case in 1939 (case 3). Here, too, a hamlet planned to expand a road, and here, too, a landowner refused to cooperate. The hamlet needed to remove a hedge on the edge of his property, but the owner refused permission. After long and complicated negotiations involving not just the owner but his adult nephew, community workers started to clear the hedge. The owner called the police, and the community responded by imposing ostracism.Footnote 19 In turn, prosecutors brought charges, and (as discussed below) the court declared the ostracism a crime.
A 1952 Tokyo High Court case involved a hamlet's liability to the national government (case 4). Under the stringent economic controls of the early post-war years, the national government requisitioned rice from farming hamlets (effectively, but not formally, a tax). Community leaders then allocated the requisitioned amount among the hamlet members.Footnote 20 One resident of the 45-household hamlet thought his allocation unfair and refused to provide the full amount demanded. The community responded with ostracism.Footnote 21 The farmer sued, and again (see below) the court held the ostracism to be a tort.
In the reported cases, Japanese courts almost always declared the ostracism illegal. Sometimes the police arrested the hamlet leaders, and sometimes the victims themselves sued the leaders. When prosecutors pursued criminal charges, the courts generally convicted the leaders. Section 222 of the Criminal Code made intimidation – conduct that would ‘threaten the life, body, freedom, reputation, or property of another’ – a crime, and judges have called ostracism criminal intimidation. When victims sued hamlet leaders, the courts generally called the ostracism a private wrong. Section 709 of the Civil Code made intentional harm – the ‘intentional or negligent invasion of another person's rights or legally protected interests’ – a tort, and judges have called ostracism an intentional tort.
Criminal cases have also been brought against ostracisers. One of the earliest reached the Supreme Court in 1911 (case 5). It involved a man who had failed in business. He had largely brought failure upon himself, and had caused his neighbours considerable harm in the process. The community imposed ostracism. Lest his friends decide to ignore the sanction, some members of the community contacted his likely sympathisers. Should the sympathisers ignore the decree, they warned, they would meet the same fate.
The court declared this threat to the sympathisers a crime.Footnote 22 No one has a right to social interchange, it reasoned. If anyone finds that a neighbour no longer speaks to him, he has not necessarily suffered a legal wrong. But should his neighbours cut off contact collectively, they do commit a crime. ‘When the residents in an area decide collectively to punish a member, and then declare that they will cease all contact with him, they have excluded the member from their society. They have degraded his personhood, and harmed his good name.’ They have, in violation of section 222 of the Criminal Code, committed criminal intimidation.Footnote 23
Government scepticism toward ostracism did not start in 1911. Even during the Tokugawa shogunate, the government was sceptical. In 1822, twenty-six villagers in Komono village (in present-day Mie prefecture) sued in the local (domainal) court to expel their neighbour Kishichi. He was not, they complained, ‘conforming to the customs of the village’ (case 6).Footnote 24 Kishichi had moved to the hamlet from a nearby village. He was farming land that his family had already owned, but the villagers wanted him evicted anyway. The court considered the attempted expulsion an overreaction, and punished the village leaders.Footnote 25
The Tokugawa government did accept ostracism as a general tool of village control, even though its dangers were recognised. In 1827, one Kyujiro, along with two other villagers from what is now Saitama prefecture, claimed that another villager, Chojiro, owed them five ryo from loans. Chojiro denied that he owed them money. When self-help attempts to collect the debt escalated to the point of grabbing bales of rice, a fight ensued. The crowd started breaking farm implements, and eventually a dozen other villagers, led by the headman, intervened to break up the fight. In the end, Kyujiro paid three of the five ryo.
Two days later, the village formally ostracised Kyujiro. He immediately sued nineteen of the villagers, including the village headman. Unless they rescinded the ostracism, he claimed, he would not be able to pay his taxes. The defendants denied that they had ever ostracised him, and the case was settled: the debt was declared paid, and the defendants admitted the ostracism, apologised, and reinstated the plaintiff.Footnote 26
Although the Supreme Court announced a flat ban on ostracism in the 1911 case (case 5), courts generally took a more measured approach. In the 1939 road expansion case (case 3), the Supreme Court did conclude that the ostracism was criminal. But it held it criminal only because the offenders had imposed it “without a reason deemed appropriate by social convention.” As ‘judged by social convention, their ostracism had lacked a recognisably proper reason’. Given this lack of a ‘proper reason’, it violated ‘public order and good morals’.
Yet, if in 1939 the Court declared only unreasonable ostracism illegal (case 3), the courts usually found the ostracism in the reported cases to be unreasonable. One would have to search long to find any village ostracism that the courts permitted.
More common than criminal cases, however, were tort cases. Return to the 1921 Supreme Court case where Kodama refused to provide land for a road (case 2). The case did not stem from a criminal prosecution. Instead, Kodama had brought it as a tort case against the hamlet members who had orchestrated the ostracism against him. Through the case, the Court made clear the tort equivalent of Criminal Code Section 222: the collective ostracism of a member of a community is an intentional tort. ‘Leave aside doctors and innkeepers, for whom special rules exist’, the court explained. Its reasoning paralleled the principles it would later apply to criminal prosecutions in 1939 (case 3):Footnote 27 while no one has ‘a right to demand’ participation in social interchange, the fact that each person can individually refuse to interact with another does not mean that a group can collectively refuse to do so.
The defendants argued that Kodama had ‘damaged the collective interest’ of their community, and that they were merely trying to ‘preserve its good customs and order’. The court would have none of it: if community members collectively decide to terminate contact with an offending member, they are committing a tort under section 709 of the Civil Code.Footnote 28
Recall the 1952 Tokyo High Court decision on the government's rice requisition programme (case 4). The offending farmer refused to deliver the share of the collective rice burden assigned to him by the hamlet, and the community responded with ostracism. The court declared the retaliation a tort.
If a tort is threatened, it can become the crime of extortion. The potential criminal sanctions do not end with section 222, which we discussed earlier. If neighbours vote to ostracise someone in their community, they do indeed commit criminal intimidation under section 222 of the Criminal Code. But if they demand money in return for cancelling that sanction, they are committing the more serious crime of extortion under section 249. In 1923, a man named Kurosawa in a small community in Akita prefecture made charcoal from material he had stolen from the hamlet and from a local contractor (case 7). Upon discovering his theft, the hamlet leaders called a general meeting and voted to terminate all contact with him.
In time, Kurosawa sought reconciliation. He asked his older brother to act as intermediary. The hamlet convened a second meeting. Kurosawa apologised, and most of the members seemed inclined to end the sanction. The defendant (unnamed) however, intervened. Rather than forthrightly forgiving Kurosawa, he urged the others to require that Kurosawa first pay a penalty. He demanded 200 to 300 yen. Kurosawa eventually paid 100 yen, still a significant sum for a poor farmer. When the court announced its decision in 1927, it called this extortion under section 249.Footnote 29
Gangs present a special case. Organised crime syndicates in Japan (the yakuza) routinely ostracise insubordinate members. Lest rival syndicates attribute any misconduct by a deviant member to them, they send a notice (typically a printed post-card) to their local rivals announcing his ostracism. One case in 2011, for example, involved fratricidal battles within the massive Yamaguchi gumi crime syndicate (case 9). The leader of one faction shot the boss of the Yamaguchi gumi in a hotel lounge, the syndicate expelled (hamon) the faction, and war ensued.Footnote 30
A curious variation on this practice occurred in 2018 in Shizuoka City (case 10). The unnamed plaintiff was a long-time member of the local mob. Now in his 60s and suffering from liver cancer, he was no longer of much use to the organisation. Anticipating heavy medical expenses, he applied for public welfare. The welfare office turned him down. He was still in the syndicate, and the office did not pay welfare benefits to members of the mob.
The ageing gangster called a police officer he knew. He explained that he needed cancer surgery and planned to leave the mob. How, he asked, could he prove to the welfare office that he was no longer a member in good standing? The officer suggested that he produce the usual expulsion postcard (hamonjo). Unfortunately, the gangster replied, although his boss would sign a certificate saying he had left the organisation, he was too scrupulous to circulate an ostracism notice: ‘You haven't done anything wrong. How can I circulate an expulsion notice?’ The plaintiff pleaded with the welfare office to approve him anyway. The office refused; he sued, and the district judge ordered the welfare office to proceed with the application.Footnote 31
Troubling cases
In many of the reported ostracism cases, the community did not ostracise a member in order to enforce welfare-enhancing norms; instead, it ostracised a member to enforce seriously anti-social norms. There are exceptions, to be sure. In one case, it punished a man who imposed costs on his neighbours by repeatedly making bad bets in business (case 5); in two others, it punished a man who refused to contribute toward infrastructure improvements (cases 2 and 3); in yet another, it punished a man who reneged on his share of the community tax burden (case 4).
Yet these plausibly benign cases are not the rule: most of the published opinions involve more troubling disputes. Some seem to involve reasonable disagreements about community policy. Some involve hamlets dominated by criminals. Some involve hamlets that punished members for reporting criminal activities. Some involve hamlets that took property from other members. And several involve hamlets engaging in electoral fraud. We will discuss these categories below.
The first set of cases involves a community that punishes a member simply for disagreeing about village policy. In a 1935 Supreme Court case, a firm had planned to build a synthetic textile factory near the mouth of the Yagyu river in Toyohashi city (case 11). Most residents opposed the factory on the grounds that the effluents would threaten the amount and quality of the fish, shellfish, and seaweed harvested. When three villagers announced their support for the factory, the rest of the community retaliated by ostracising the three. Absent more detail, one cannot tell what was at stake. Perhaps the three dissenters had invested in the factory. Perhaps the factory had bribed them. The court does not say. Instead, it treats the dispute as an honest disagreement about village policy, and held the ostracism to be criminal intimidation.Footnote 32
A second 1935 Supreme Court case involved a small island off the southern coast of Kyushu (case 12). Part of the Amami Oshima chain, it lay a seventeen to eighteen hour ferry ride from the city of Kagoshima. In 1935, the island became the scene of what historians would call the great ‘Lily Bulb War’. The residents were primarily growing lily bulbs for export. In 1932, a Yokohama nursery owner formed the Japan Lily Export Association and obtained exclusive control over the government-mandated export inspections. Now able to block rival exporters, he planned to dominate the market. At about the same time, however, Mitsubishi Trading decided to challenge his control. Mitsubishi offered the farmers an exclusive trading contract. The local farming association held a meeting. The farmers debated the two options. About 2,000 members voted in favour of the Mitsubishi contract and 138 voted against. The majority argued that the 138 opponents were jeopardising the deal with Mitsubishi for private gain and hit them with ostracism. The court held the ostracism to be criminal intimidation.Footnote 33
Cases 11 and 12 at least involved ostracism for the benefit of most of the community. Contrast this with a 2007 Niigata District Court case (case 13) involving a village bully, Taro Kono (a pseudonym), who dominated his village through wild and unpredictable violence: he picked fights, he beat people. His neighbours had called the police on him multiple times: when he started to strangle someone; when he swung a metal bar at someone; when he attacked a man with a sake bottle. Notably, Kono also ran the annual village festival. According to the other residents, he ran it autocratically and stole community funds. Several members tried to distance themselves from the event, only to face Kono's retaliation. He used intimidation and coercion, forcing the other village members to ostracise them. The Niigata District Court ultimately declared the ostracism a tort.Footnote 34
Perhaps even worse are the cases where ostracism was used to punish community members who cooperated with the police. Six decades ago, the anthropologist Robert Smith observed that Japanese who reported community misdeeds to the police could suffer ostracism.Footnote 35 So they did. So they still do. Akimitsu Fujii ran a general store in Kumamoto with his wife and three daughters (case 14). One January afternoon, he watched the local fire brigade training. After practice, the firemen shared drinks. Several of them started a fight with a firefighter who had missed practice. When the police interviewed Fujii several days later, he detailed what he had seen. The firefighters retaliated by organising a boycott of Fujii's store, and drove him and his family out of town. He sued, and the court held the firefighters liable to Fujii.Footnote 36
Another ostracism victim worried that the local residents’ association was cheating the community (case 15). The association was constructing a new building, and he suspected that the contractor was shaving costs. He began to circulate a complaint. Steadily, he ramped up the tension. The association leaders were (in the court's words) ‘crazy in the head’, he asserted. They were evil. They were ‘liars’, perpetrating a fraud. The community sued him for slander and won. They also expelled him from the association. When the victim sued in response, the Tokyo District Court reasoned that expulsion from the neighbourhood association would have a major impact on his life, and vacated the sanction (slander or no slander).Footnote 37
In 1954, the Fukuoka High Court was faced with a case of ostracism by an 18-household hamlet against four members (case 16). The opinion does not describe the full scope of the offending conduct (opinions rarely do), but the precipitating event seems to have been something one of the victims told the village government. The national government was still requisitioning rice from farming villages. Apparently, one of the four victims told the government how much rice it could safely demand from the hamlet. The other members were outraged and expelled all four. The court convicted the hamlet leaders of criminal intimidation.Footnote 38
Ostracism may also be used to enrich the ostracisers. Tomoyuki Arakawa was a nationally prominent potter in the town of Yagusa (within Toyota city, Aichi prefecture) (case 17).Footnote 39 His family had lived in the village since the Tokugawa period. Apart from seven years in the nearby Nagoya city, he himself had spent his entire life in Yagusa.
Arakawa made pots from clay he dug from the communal mountain. He built his kiln on the mountain. He fired his pots with wood he collected on the mountain. Sometimes he would leave his home for days on end to work at the kiln. His neighbours considered him an odd fellow, but no one really minded how he made his pots.
The mountain covered about forty per cent of the ‘town’. Gardens and paddies occupied most of the rest. The national government had conveyed the mountain to the village in 1913. Ever since, title had lain with the descendants of the seventy-five families who lived there in 1913, including the Arakawa family. But over time, the humble mountain became extraordinarily valuable. A mining company discovered that it contained valuable deposits of silica. Near as it was to the Nagoya metropolis, it had development potential. Close to the Toyota factory network, it could provide land for access roads. By 2008, the constituent families had so shrewdly exploited its potential that they had amassed two billion yen (about US$20 million).
The group decided to distribute the two billion yen to the constituent owners, but refused to pay Arakawa his share. Arakawa sued for the money, but he had also sued to stop the development. At root, he seems to have cared less about the money than about stopping the mining, the construction, and the roads. The other villagers invented one reason after another for not paying him his share, but mostly they wanted him gone. Ostracism came naturally. ‘Just leave Yagusa’, one village official begged. As of 2020, the litigation was apparently still in progress.
In Arakawa's case, it was a poor man whose share was expropriated, but ostracism is even more attractive for the poorer majority to use to extort the wealthy few. In 1946, the Miyamoto family on the island of Shikoku decided to cancel their leases with several families who had been renting their land (case 18). Both the Miyamoto family and the lessees had been members of the local Japan Farmers’ Union (Nihon nomin kumiai), a hard-left group with alliances with both the Socialist and Communist Parties. Sixty of the eighty lessee households were part of this Union. Once the Miyamoto family announced their plan to terminate the tenancies, the local Union expelled and ostracised them. The Miyamotos could find no one from the hamlet willing to work on their land. The local court declared the ostracism a tort, and the parties settled out of court.Footnote 40
The year 1946 was also the year of the US-imposed ‘land reform’ programme.Footnote 41 The Miyamotos may have cancelled the leases in the hope of getting better terms for land they tilled themselves. Under the programme as eventually imposed, the government took land from farmers owning more than three hectares (with nominal compensation) and gave it to their former renters (at a nominal price). Subject to modest variation, the redistribution applied to all farm land.
The programme famously did not apply to mountain land. Although worth less than farm land, the mountains had real value. Obviously, they provided lumber. They also supplied the firewood and grasses needed by farm households. Close to metropolitan centres, many had development potential (as the Yagusa families discovered, see case 17). And many mountains also contained food – the ‘mountain vegetables’ (sansai) used in some dishes, and the extraordinarily expensive (sometimes US$1,000 per kg) mushrooms known as matsutake.
A town in Hyogo prefecture managed its local mountain collectively through a voluntary association (case 19). The group included 103 households, a majority of the local residents. In 1950, the association decided to require all villagers owning more than two hectares of mountain land to transfer all rights to the sansai and matsutake on their land to the association without compensation. The national government had not redistributed the mountain land, so the locals decided to do it on their own.
Five families refused to cooperate with the expropriation. When the association then withheld their share of the communal profits, they sued. In retaliation for their lawsuit, the association declared all members of the five families ostracised. In the ensuing criminal case, the District Court acquitted the association members on the grounds that the sanctions were not binding, but the High Court reversed the decision. In 1958, the Supreme Court affirmed.Footnote 42
The most common of the troubling cases, however, involve elections. In 1913, the Supreme Court used an electoral dispute to decide perhaps the oddest of all its ostracism cases (case 20). The villagers of a hamlet had agreed to vote for a particular candidate, and had further agreed to punish anyone who deviated from the agreement. Two villagers reneged. The others imposed ostracism, and the prosecutors initiated criminal cases against several of the ostracising villagers. The Supreme Court reversed the convictions. Ostracism is not always criminal, it explained. Villagers can ostracise members for a wide variety of reasons, some of which are morally justified and some not. In this case, the two offenders had reneged on their promise to vote for the community's chosen candidate. When a community punishes someone to force him to do something he has no obligation otherwise to do, its members are committing criminal intimidation. So too when they punish someone to stop him from doing something he has every right to do. Here, however, they were simply punishing the two members for breach of contract. They had agreed to vote a certain way, and they had done otherwise. The court ignored the obvious electoral context, and reversed the convictions.Footnote 43
In 1920, the Supreme Court took a more typical approach to these election-related ostracism disputes (case 21). In the national Diet election in May 1920, most of the voters in a town in Mie Prefecture favoured one candidate. Katsunosuke Oku favoured another. Outraged by his independence, the other villagers decided to sever all ties with Oku and his family. The prosecutor brought charges, the judge convicted, and the Supreme Court affirmed. The Court followed what would become a standard formula: no one has a right to social intercourse; no one breaks the law by refusing it; but when members of a community refuse that intercourse collectively, they commit criminal intimidation.Footnote 44
A 1924 Supreme Court case followed the same pattern (case 22). In September 1923, four people had been arrested for violating electoral law in the Miyagi prefectural elections. A certain Mr Honda, living in the same hamlet as they did, had turned them in. The arrested villagers complained to their neighbours, and the hamlet's mutual aid society voted to expel and ostracise Honda and his father. The trial court convicted the villagers of criminal intimidation, and the Supreme Court affirmed.Footnote 45
A second 1924 Supreme Court case involved not an actual hamlet sanction, but a threat by an influential leader in Nara unilaterally to oust an uncooperative villager (case 23). The leader had told the villager to vote for a particular candidate. If he tried anything else, he warned, he would expel him from the village. The prosecutor brought criminal charges against the leader. Expelling someone from a village is, of course, not a technical term, and the defendant's lawyer professed not to know what it meant. The Supreme Court declared that it was easy to see that the defendant meant murahachibu. The lawyer also protested that the defendant had no authority to expel anyone anyway. The Supreme Court noted that the defendant was an influential man, and that a resident could reasonably worry about the threat. It affirmed the conviction.Footnote 46
Zones of judicial neutrality
Introduction
These cases suggest that the courts generally decide whether to intervene in ostracism disputes on the basis of the conduct involved: they intervene when they believe that a community ostracised for improper reasons. In a small number of disputes, however, the courts purport not to intervene at all. In cases involving political parties and religious groups, for example, the courts instead announce that they will let the losses lie where they fall.
However, this summary is potentially misleading. First, many of the troubling cases do not involve the typical community ostracism at stake in the conventional cases we first discussed. Instead, they involve the arguably distinct question of who controls the membership rosters of voluntary associations. Moreover, they do not involve a village deciding to shun a non-conformist resident; rather, they involve associations that people joined precisely to express their political and religious preferences.
The disputes arise when the leaders of an association decide to expel someone over questions involving those political or religious beliefs. Japanese courts – taking a stance much like that of American courts – try not to intervene. As Douglas Linder puts it in the US context, courts are hesitant to compel the ‘benefits of membership in a voluntary association’. Instead, they tend to declare that ‘the associational freedom at stake, the right of an association to define its own membership, is fundamental to a conception of a pluralistic free society’.Footnote 47
Second, although Japanese courts claim not to intervene in these disputes, effectively they do intervene – in fact, they have no choice but to intervene. In the two cases below, for example, an association decided to evict a dissident member from association housing. Although in each case the court declared that it would not intervene, non-intervention meant that it did not stop the association from evicting the dissident.
Political parties
The best known of the political cases involved a struggle for power at the centre of the Japan Communist Party (JCP) (case 24). Kenji Miyamoto, Satomi Hakamada, and Sanzo Nozaka had helped lead the party during the stormy pre-war years. Miyamoto had studied economics at Tokyo Imperial University; Hakamada had studied in the Soviet Union. Together, in 1933 they tortured to death Tatsuo Obata, a colleague they suspected of spying for the police. Sanzo Nozaka had found himself in Moscow during Stalin's purges, and had survived by inventing charges against another JCP member in Moscow – a man whom Stalin promptly had shot. After the war, the American-run occupation welcomed all three into the public realm.Footnote 48
The three men promptly took over the JCP. Nozaka won election to the national legislature in April 1946. When Stalin ordered the party in 1950 to abandon peaceful tactics, Miyamoto went underground and masterminded the party's bombing and sabotage campaign, while Hakamada stayed with the party's legal faction. After Stalin's death, Miyamoto returned to electoral politics, eventually rising to the post of Central Committee chairman in 1970, with Hakamada serving as vice chairman. Decades later, Miyamoto continued to insist that Obata had died a natural death; Hakamada wrote that they had strangled him to death. Hakamada also nursed a long-standing suspicion that Nozaka remained a Soviet spy – an accusation that the party leadership declared treasonous (but which Soviet archives would later prove to be true).
Late in the 1970s, Hakamada began writing about the way he and Miyamoto had murdered Obata. Miyamoto immediately moved to push him out of the party. In 1977, Miyamoto successfully dropped Hakamada from the party Central Committee. When Hakamada retaliated by publishing yet more information about the murder, Miyamoto led the party to expel him.
Since 1963, Hakamada had lived in party housing. For a house with market rental pegged by the court at 132,000 yen per month, Hakamada paid just 22,000. Now that he had been expelled, the party administration ordered him to leave. Hakamada refused, and the party sued. Hakamada explained that he was growing old and – having worked for the party all his life at low pay – had very little savings.
The District Court ordered Hakamada out, and the High and Supreme Courts affirmed. The courts declared internal party disputes beyond their jurisdiction. Said the High Court:
Political parties are indispensable for supporting representative democracy, and effective bodies for helping citizens structure their political thoughts.
Never mind that Miyamoto had fought for decades to end representative democracy. The court continued:
The expulsion and other punishment of party members are matters internal to the parties themselves. These matters follow from the right of party self-governance, and the courts should treat them with ample respect.
The courts would not intervene in party affairs. One might have thought this meant that the state would not help the party to evict Hakamada from party housing. Whatever their logic (they did not explain it), the judges decided it meant the opposite.Footnote 49
Religious organisations
Courts show the same reticence toward disputes within religious organisations. In the United States, writes Eric Posner, the ‘Free Exercise Clause and common law principles of free association … prevent people from suing a religious group for expelling them’.Footnote 50 The same is true in Japan, for the most part.
During the last decades of the 20th century, Japanese courts faced several cases involving the highest profile religious revitalisation movement in modern Japan: the Soka Gakkai. The Nichiren Shoshu Buddhist denomination traces its roots to its namesake, the 13th-century priest Nichiren.Footnote 51 In 1930, Nichiren Shoshu adherents established the Soka Gakkai as their lay organisation. After the Second World War, the Gakkai grew explosively. It remains enormously popular and has steadfastly maintained its roots in the blue-collar working-class community. In 1960, Gakkai leadership passed to Daisaku Ikeda. Ikeda proved to be a polarising figure, attracting both international publicity and domestic hostility.
In time, clerical leaders within the Nichiren Shoshu denomination grew suspicious of Ikeda. By 1991, tensions had reached the point where the denominational leaders demanded that their priests attack the Gakkai leadership and pledge loyalty to the denomination. Soon, they would expel the Gakkai itself. One of the Nichiren Shoshu priests, however, refused to attack the Gakkai (case 25). He lived with his wife in temple housing, but despite enormous pressure, he refused to sign the proffered statement. In response, the denomination expelled him from the priesthood, slashed his pay, and evicted him and his wife from temple housing.
The priest sued for tort damages, but the Shizuoka District Court refused. For the court, religious denominations were like political parties. The Constitution of Japan protected their self-governance, and unless they violated ‘public order and good morals’ or threated ‘basic human rights’, the courts would not intervene. Here, that meant that the court would not order the denomination to compensate a dissenting priest whose career it had ruined.Footnote 52
Discussion
The basic observation
Having described the cases, it is now clear that most court disputes over ostracism in Japan share one basic characteristic: they have nothing to do with the theory of informal social sanctions against deviance. Many of the cases do not involve attempts by a community to control anti-social deviance at all. Instead, they involve opportunistic tactics by the community itself. In some cases, the community tried to punish a member who attempted to stop the broader patterns of community misconduct (cases 1, 14, 15, 16, 20, 21, 22, 23). In some cases, the community used ostracism to extort property from a minority of its members (cases 4, 17, 18, 19). In some, an opportunist manipulated the mechanism of ostracism to his private advantage (case 13). And in some, the community split, and one faction used ostracism to penalise the other (cases 11, 12, 24, 25).
This raises three questions:
a) How did the courts adjudicate the disputes presented to them?
b) In which disputes did people decide to use ostracism, and which cases of ostracism did they litigate in court?
c) What might the plaintiffs or prosecutors have hoped to gain by filing the cases they did?
How did the courts adjudicate these disputes?
In cases of ostracism in Japan, the government may intervene. Victims may sue the ostracisers, and the courts may award them relief. Prosecutors may file charges against the ostracisers, and the judges may convict.
In the US, by contrast, the courts hesitate. They do recognise the tort of intentional infliction of emotional distress. Yet usually they intervene in ostracism cases only when they see a group threatening someone in a ‘protected’ category. Under labour law, they may intervene if they think a company is using ostracism to isolate a union organiser. Under Title VII of the Civil Rights Act of 1964, they may intervene if they attribute the ostracism to sex, race, religion, or national origin. But absent a protected category, they usually leave socially isolated victims to their own devices.Footnote 53
Japan, however, is not the outlier among wealthy democracies: America is. As James Q Whitman notes, German and French courts are much more ready than US courts to protect people who find themselves socially harassed.Footnote 54 American courts worry about the government intruding into the private sphere; European courts ‘aim to protect people from shame and humiliation, from loss of public dignity’.Footnote 55
Whitman plausibly attributes the continental phenomenon in part to its aristocratic past. Germany and France both maintained an ‘honor culture’ among their elites, he writes. When they democratised, they extended to everyone the honour and dignity formerly confined to those elites. Whatever the case in Europe, however, the democratisation of aristocratic ‘honor culture’ probably does not explain the cross-national contrasts. Japan, too, has an aristocratic past. But Japanese courts did not wait until after democratisation in the late 1940s to start policing ostracism. They had already held ostracism illegal during the distinctly non-democratic decades of the early 20th century.
The informational logic redux
Although the plaintiffs in the ostracism cases filed monetary claims, they probably did not collect substantial compensation. Given that the discussed cases are mostly appellate decisions, few give the amounts recovered by the plaintiffs. Those that do give figures do not report large amounts.
Nor could the prosecutors have expected to obtain heavy penalties. Again, as these are appellate decisions, most do not report the penalties imposed. Those that do give them, however, report only suspended sentences.
Instead, the plaintiffs (and prosecutors) seem to have filed the suits for the informational role that courts can play. Consistent with the recently developed theory tying the courts to the economics of information, they sued to obtain public certification and dissemination of their story. They sued in a way that reflected the role that the courts themselves can play in producing, certifying, and publishing information. They sued to capitalise on what Sadie Blanchard called the court's role as an ‘information intermediary’:Footnote 56 In the course of litigation, courts produce information. When they ultimately decide a case, they certify that information. They then announce it to the public and disseminate it.
Crucially, litigation can change the character of the public understanding of a dispute. If a dissenter sues and wins, the dispute now becomes the judge's word against that of the village leaders. And if the dispute has any news value, the local press will convey the judge's word broadly. Through litigation, the victim who wins both increases the credibility of his account and conveys that information more broadly than he could do otherwise.
This certification and dissemination of information matters because of the impact that information about the dispute can have on the relative reputations of the leaders and the dissenters. Those reputations, in turn, determine the capacity of both groups for advantageous trade. The more public the information, the greater the impact on future economic and social transactions.
Courts, explains Blanchard, increase a ‘[r]eputation's effectiveness’ because they spread ‘information about past behavior … more widely among potential counterparties.’Footnote 57 One of us has written about the ‘stigmatization’ function of punishment.Footnote 58 What the courts illustrate in this context is their capacity to ‘destigmatize’ when communities punish the wrong party.
Conclusions
Scholars have long treated informal social sanctions as a community's primary means of controlling deviance. They have treated formal legal sanctions – both civil and criminal – as a more costly secondary mechanism. Among informal sanctions, ostracism is one of the most severe. Yet very few scholars have studied actual instances of ostracism.
We have examined legal cases brought over ostracism in modern Japan – a wealthy democracy in which (according to most Western scholarly accounts) communities traditionally enforced local norms through informal sanctions. Some of these cases are civil, others criminal. As with any other kind of dispute, very few cases of ostracism actually reach the courts. Of those cases that do reach the courts, however, very few involve a community that has used ostracism to restrain deviance. Instead, most cases involve disputes where the community has used ostracism opportunistically – to extract property from a member, for example, to hide community-wide malfeasance, or to harass a rival faction.
The plaintiffs who bring these ostracism cases do not bring them primarily for damages (nor are prosecutors aiming for criminal sanctions). Instead, they appear to bring them for informational purposes: to have their version of events publicly certified by the court. They bring these cases because these were the cases where ostracism was unjust, even by village standards. At stake is the informational role that, as scholars increasingly recognise, courts can play in modern society. It is the role suggested in Blanchard's pioneering work: that plaintiffs bring cases to obtain the court's imprimatur on their claims of innocence.
Annex: A Formal Model in the Law & Economics Tradition
Ostracism has been modelled using game theory in various highly abstract contexts, as the idea that if one member of a group offends, the other members will refuse to engage in mutually profitable interactions with him.Footnote 59 Usually the focus is on how the group incentivises individual members to ostracise the target, for example by ostracising the non-ostraciser, and the problems created by the resulting infinite chain of penalties. Here, we will set aside the problem of how the villages enforces ostracism and focus on mistaken ostracism and the role of courts.
The Basic Model
Consider a simple model to address several of the situations that can arise:
(a) a villager whose actions hurt everybody but him;
(b) a villager who hurts the village but helps society;
(c) a villager who helps the village but hurts society;
(d) a villager who is mistakenly believed to have hurt the village; and
(e) other variations on the basic situation of ostracism.
We will model a village that uses ostracism to deter deviant conduct, and a broader society that establishes a court system which may or may not wish to restrict ostracism. Let the payoffs for each player be normalised to zero if the target villager does not deviate, the villagers do not ostracise, and the court does not get involved. The target villager chooses to comply with village custom (x = 0) or to offend (x = 1). The village sees evidence that he has complied (y = 0) or offended (y = 1). If the target offends, he is always detected: Prob(y = 1|x = 1) = 1. If the target complies, the evidence sometimes mistakenly indicates that he offended: Prob(y = 1|x = 0) = m, where 0 < m < 1.
If the target offends, he obtains a personal benefit B > 0 from this act, but imposes a cost C on the village and a cost D on the rest of society. The costs and benefits B, C, and D are unobserved until later (otherwise the villagers could look at them to determine whether the target has offended). The village can either continue to associate with the target, or ostracise him at a cost Z > 0 to itself and impose a cost P > 0 on him.
The costs C and D need not be positive. If they are, the target's offending is harmful; if they are negative, his ‘offending’ is beneficial. It could be, for example, that C > 0 and D < 0, which would mean that offending hurts the village but helps outsiders, as with reporting village corruption (cases 1, 14, 15, 16). If B < C + D, offending is wealth-diminishing for society as a whole – the sum of target, villagers, court, and outsiders. If B > C + D, offending is wealth-increasing for society as a whole, but still bad for the villagers if C > 0.
At cost L to himself, the target can take his case to court. At cost J to the outside world, the court can agree to hear it, to decide whether or not the target truly offended, andto announce its decision publicly (as discussed in the section ‘The informational logic of litigation’ above).
Whether or not the target has gone to court, the model then moves to a second period: the long-term. The village decides whether it wants to continue ostracising the target, in which case the costs are incurred a second time: Z for the village and P for the target.
In interpreting the model, note that it is unimportant that we have assumed that a target who has truly offended is detected by the village with a probability of one, and that the court never makes mistakes. While descriptively unrealistic, adding parameters to incorporate these sources of error would make no significant difference to the model or our conclusions.
The Outcome
Consider three regimes: (1) No-Penalty, (2) Unconstrained Ostracism, and (3) Constrained Ostracism. We will compute the payoffs under these regimes and compare them to see which regime would be chosen by the village and which by the court. We assume that the target maximises his own utility, the other villagers maximise the sum of their utilities, and the court maximises the sum of the utilities of everyone in society – the utilities of the target (amounts B, P, and J), the villagers’ utilities (amounts C and Z), the utilities of people outside the village (amount D), and the public court costs (amount J). Each type of norm violation by the target will have its own values for each of these parameters, and it is quite possible to have different regimes for different offences. Reporting corruption, refusing to give up land for a road, and murder will differ in their parameter values, and this will be related to how the village and the courts react to them differently.
(1) The no-penalty regime: villagers do not ostracise.
In this regime, the villagers never ostracise anyone.
The no-penalty regime is the base case. The target will offend, for a payoff of B, since he will incur no penalty. The villagers will have an aggregate payoff of –C. The court has no role. Society's overall welfare is B – C – D.
(2) The unconstrained ostracism regime: villagers ostracise; the court refuses to hear ostracism cases.
(a) The villagers ostracise if they see evidence of deviant behaviour.
(b) The court refuses to hear any ostracism case brought before it, ruling that ostracism is never illegal.
Under unconstrained ostracism, the target's expected payoff is –2mP if he complies, since with probability m he will be ostracised by mistake in both periods. If he offends, it will be B – 2P, since he will definitely be detected (under the model's assumptions) and ostracised. Thus, he will comply if and only if –2mP > B – 2P, which is true if B < 2P(1–m); that is, if the reward from offending is small compared to the penalty of being ostracised, and if the probability of mistaken ostracism is small enough. Note that if the probability of mistaken ostracism is high enough, the target will offend even if his benefit is small, because he can expect to be ostracised whether or not he actually offends.
The village's expected payoff depends on what the target does, so it depends on B, P, and m. If B < 2P(1–m), then the target complies and the village's payoff is –2mZ, the cost of mistaken ostracism. If B > 2P(1–m), then the target offends and the village's payoff is –C – 2P.
Compare the village's payoff in the no-ostracism regime with the unconstrained ostracism regime. If B < 2P(1–m), then the village benefits from having the ostracism regime if –2mZ > –C, which is true if 2mZ < C. Hence the first proposition:
Proposition 1.
The village prefers the ostracism regime if it does effectively deter and if the cost of mistaken ostracism is small relative to the cost from the offence.
On the other hand, if B > 2P(1–m), ostracism fails to deter and the target will offend anyway. In that case, the ostracism regime is clearly worse for the village, because it simply adds the cost of inflicting ostracism to the cost of the offence. Punishing offenders is worse than useless if it fails to deter.
We add detail to the simple statement in Proposition 1a.
Proposition 1a.
In the absence of courts, villages will adopt a custom of ostracising people who commit offences that impose a relatively high cost on the village (high C) but have a relatively low benefit to the target (low B), but only if the evidence for that kind of offence is reliable enough (low m) and the cost to other villagers of ostracising someone is not too high (low Z), while the cost to the target is high enough to deter him (high P).
Case 2 illustrates this: ostracism of the target for refusing to join fellow villagers in giving up land to construct a road that would benefit them all. Proposition 1 implies that ostracism will be used for relatively minor offences, not major ones. It is not suitable for dealing with a villager who steals his neighbour's stash of coins. That offence is profitable to the target (high B), unimportant to everyone except the victim (low C), and false accusations can easily be made since the deed is secret (high m). For such offences, villages need government courts and official penalties such as fines or imprisonment.
In the unconstrained ostracism regime, society's payoff will depend on whether ostracism deters or not. On the one hand, if B > 2P(1–m), the target offends, so the sum of everyone's payoffs is his B – 2P, plus the village's –C – 2Z, plus outsiders’ –D, a total of B – C – D – 2(P + Z). Under the no-ostracism regime, total welfare is B – C – D. Thus, social welfare is lower by amount 2(P + Z) with unconstrained ostracism – since it fails to deter, all it does is impose costs on society.
On the other hand, if B < 2P(1–m), the target complies, so the sum of everyone's payoffs is his –2mP plus the village's –2mZ, a total of –2m(P + Z). Under the no-ostracism regime, total welfare is B – C – D. Thus, social welfare is higher with unconstrained ostracism if –2m(P + Z) > B – C – D; that is, if the cost of mistaken ostracism to target and village is less than the offence's benefit to the target minus its cost to the village minus its cost to outsiders.
In sum: if ostracism fails to deter offending, it merely imposes costs and it hurts the village and society. As a result, we would not expect it to persist as a social custom unless we introduced something not in the model – for example, manipulation of the problem of group action for private gain (see, eg, case 13). On the other hand, if ostracism does deter, it can increase village and societal welfare, depending on how accurate and costly it is, and whether the ‘offence’ is really harmful to the village and to outsiders.
(3) The constrained ostracism regime: villagers ostracise; the court hears cases; villagers listen to the court.
(a) The villagers ostracise if they see evidence of deviant behaviour.
(b) The court hears any case brought before it.
(c) The villagers end ostracism if the court declares that the target did not deviate.
Regime (3) introduces court intervention. Under unconstrained ostracism, the court does not review ostracism cases. In the constrained ostracism regime, it reviews the cases, declares whether the target offended, and the villagers cease to ostracise if the target did not offend. In this regime, the target may or may not choose to comply, depending on the parameter values, as we will discuss shortly. The villagers will sometimes observe apparent offending even if the target is complying, and if they observe it, they will ostracise the target. The target may or may not go to court if he complies and is unjustly ostracised, depending on his legal cost L, but if he does, he is always vindicated.
Consider the target's payoff. If he offends, his maximised payoff is his personal benefit B minus his cost of being ostracised for two periods, 2P, for a total of B – 2P. If he were to go to court, he would lose and only subtract L from his payoff. If he complies instead of offending and goes to court if he is ostracised, his payoff is made up of the expected cost of one period of ostracism, –mP, minus the cost of going to court, L. Because he will be vindicated, however, he avoids the second period of ostracism and his overall payoff is –mP – L. If he complies but does not bother going to court, his ostracism will continue, so his payoff is –2mP. Thus, he will choose to go to court if L < mP.
Consequently, the target will compare his offending payoff of B – 2P with his complying payoff of Max(–mP – L, –2mP). On the one hand, if the legal costs are high (L > mP), he will offend if B – 2P > 2mP. In this case, the existence of the court is irrelevant, since it is too expensive to use, and we are back to the same outcome as in the unconstrained ostracism regime; the court exists and is now willing to hear ostracism cases, but access to justice is too expensive, so the result is the same as if it refused to hear cases. Thus, we can immediately conclude that for offences complicated enough to require costly legal proceedings, the equilibrium payoffs end up being the same as in the unconstrained ostracism regime, and we can refer back to those results to explain what the village custom will be and whether it is a good outcome for society.
On the other hand, if legal costs are low (L < mP), the target will offend if B – 2P > –mP – L. We will continue our analysis assuming that legal costs are low, so the target will go to court if he is unjustly ostracised. First, consider the village's payoff. If the target complies, the village sometimes ostracises him unjustly, but only for one period, so its payoff is –mZ. If the target offends, this hurts the village directly, plus it ostracises for two periods, so the village payoff is –C – 2Z.
Second, consider society's payoff, which is what the court cares about. If the target complies, we must subtract the public's cost of the court, J, from the sum of the target's payoff, –mP – L, and the village's payoff, –mZ, for an overall social welfare of –J – L – m(P + Z).
Thus, if the target offends, the implications for aggregate social welfare are composed of the target's payoff, B – 2P; the villager's payoff, –C – 2Z; and the harm to outsiders, –D; for social welfare of B – D – 2(P + Z).
The Social Preference
Will the village and court prefer unconstrained ostracism or constrained ostracism? Unconstrained is closer to the US regime; constrained is closer to the Japanese (and to Whitman's European regime discussed in the text).
If the target's personal benefit B is high enough, the target will offend no matter what. Ostracism fails to deter, and the courts are unimportant. In that case, the no-ostracism rule is best for the village and for society. This sounds bad, but if B is large, and C and D (the costs to the village and to outsiders) are also large, society can simply turn to criminal law, which lies outside our model. Hence Proposition 2:
Proposition 2.
When a villager has violated custom and obtained enormous benefits for himself at high cost to his village and the outside world, the offence is criminalised.
If the offence benefits the target enough, ostracism is insufficient to deter. Instead, the court will put him in jail and prosecutes him. Failing that, the village engages in self-help: it does not just ostracize, but lynches him—though that is outside of our model. For seriously harmful offenses, a society will prefer a regime in which the courts review ostracism:
Proposition 3.
If C > 0 (the offence harms the village),
and D > 0 or D = 0 (the offence harms outsiders or leaves them unaffected),
and if the cost to the public of hearing cases, J, is not too high,
then both village and society prefer constrained ostracism to unconstrained: court intervention is valuable.
The villagers know that sometimes they wrongly conclude that an innocent target has misbehaved. They know that judges are skilled at weighing evidence and have the advantage of hearing both sides of the story. The villagers are happy to cease ostracising (and save the cost Z) once they learn the truth. And since the target knows that any ostracism will only be temporary if he complies, he will comply rather than give up and decide to be truly as bad as they would think anyway by mistake. Increased accuracy helps everybody – as long as it is not too expensive.
For many offences, constrained ostracism also dominates a fourth regime, which we might call the ‘no-ostracism criminal law regime’. In this regime, villages would not ostracise, but could choose to take an offender to court (at some cost). At that point, the court would not only determine what is true, but could also impose a penalty on the target (something that is not possible in the above model). The potential superiority of the constrained or unconstrained ostracism regimes over this fourth possibility is that they are cheap. Village gossip may not be as accurate as court proceedings, but it is quick and low-cost. For modest offences, a cheap and quick process will often dominate more accurate but costly regimes.
One unintuitive implication of the model is that if the village is somewhat inaccurate in its assessment of deviance, but not too inaccurate, then the possibility of court review actually increases the usefulness of ostracism. If courts did not exist, inaccurate village ostracism could lead to so many mistakes that over time we would expect villages to abandon it as a tool of social control. If, however, there is the possibility of the target going to court, the village does not need to worry so much about unjust, village-harming ostracism. If the ostracism is unjust, the target will go to court, the court will inform the village of that fact, and the village will relent.
Open access
