Lessons from the Great Kantō Earthquake

 

Abstract: This article shows how security forces in Japan in the early 1960s used studies of the violence and unrest that followed the Great Kantō Earthquake as templates for speculation about the challenges they would face in the aftermath of Tokyo’s next disastrous earthquake. Both studies reiterated the ambiguities associated with earlier state-sanctioned descriptions of the circumstances surrounding the massacres of Koreans and others in 1923, while maintaining that the Imperial Japanese Army and the police had done all they could to prevent that violence. The Self-Defense Agency and police analysts responsible for the two new studies concluded that if the capital district were to suffer another earthquake disaster like the one in 1923, then it was quite likely that the spread of misinformation – among other factors – would once again lead to outbreaks of vigilante violence and political instability, leaving the police and the SDF with no choice but to respond as their counterparts had forty years earlier.

 

Keywords: police, Self-Defense Forces, disaster preparedness, rumors, vigilantes, earthquake, disaster planning

 

 

Introduction

This article offers a close reading of two government studies of the unrest that followed the 1923 Great Kantō Earthquake. The reports were written in the early 1960s by analysts from Japan’s self-defense forces and Tokyo’s Metropolitan Police Department, respectively. Lessons from the Great Kantō Earthquake (Kantō Daishinsai kara eta kyōkun) was released in pamphlet form by a unit of the Defense Agency’s Ground Staff Office (Rikujō bakuryō kanbu dai 3-bu, in Tokyo) in March 1960. Earthquake Disaster Countermeasures Research Materials (Daishinsai taisaku kenkyū shiryō) published two years later, was the product of a collaboration between the Security Bureau of the Metropolitan Police Department (Keishichō keibibu) and staff in the Headquarters of the Ground Self Defense Forces (GSDF) Eastern Army (Rikujō jieitai tōbu hōmen sōkanbu), also based in Tokyo. Lessons is short, only 40 A5 pages or so; at 250 pages, Research Materials is much longer.

The two studies were products of institutions with histories as first responders in the aftermaths of disasters. Their analytical interest in the 1923 earthquake in service of that work, however, was new and significant on several registers. Both Lessons and Research Materials combined retrospective accounts of how security forces had responded to that disaster with speculative assessments of the challenges they would face if another, equally powerful earthquake were to strike the capital in the present day. The retrospective elements of the studies reiterated long-standing state-sanctioned narratives about the actions of the security forces that had taken root shortly after the disaster, first in media accounts, then in various official histories that followed.¹ Like those responsible for those earlier accounts, the reports’ authors portrayed the actions of the Imperial Japanese Army and the Tokyo Metropolitan Police as they related to the restoration of public order after the Great Kantō Earthquake as necessary responses to threats of or acts of violence that they themselves had done nothing to provoke, and as having produced positive outcomes in the end. The security forces were said to have helped rescue thousands of Koreans and others from potential vigilante attacks by taking them into protective custody, for example, even as the capital’s Japanese residents – supposedly convinced they were about to be overrun by marauding colonial others – were reportedly overjoyed at the arrival of the army, the declaration of martial law, and the subsequent restoration of order (Keishichō keibibu, Rikujō jieitai tōbu hōmen sōkanbu, 1962, 237). When the authors of Research Materials acknowledged that the army had been responsible for the murder of leftist activists at Kameido, and that military police captain Amakasu Masahiko had killed Itō Noe, her nephew and Ōsugi Sakae, they framed those well-known incidents as isolated, extraordinary events, and as not representative of the otherwise lawful actions of the army or the police (Keishichō keibibu, Rikujō jieitai tōbu hōmen sōkanbu, 1962, 70-71).

What was new about these narratives in the context of the early 1960s was not their content, which was familiar and orthodox, but the uses to which they were being put. The studies’ authors used their accounts of the 1923 earthquake’s aftermaths to argue first that Tokyo’s next disaster would also be followed by violent civil unrest, and second that Japan’s security forces would once again have no choice but to step in and restore order. The experts believed that the security forces would be called upon once again to contain post-disaster threats to society posed by vigilantes, radical activists, and foreigners. Some of the activities the report’s authors imagined the military performing in that context would not have been part of the GSDF’s normal operational repertoire in the 1960s but were at least adjacent to it; these included dispatching troops to safeguard government buildings and the Imperial Palace, for example. A significant subset of the GSDF’s post-disaster plans described in Lessons, however, was both unprecedented (by postwar standards, anyway) and almost certainly unlawful. In a section on “Preserving the Public Order” (Chian iji ni tsuite), for example, Lessons’ authors observed that:

 

Although taking into custody or placing under surveillance Koreans or other resident foreigners, as well as other individuals of interest (those suspected of thought-crimes (shisōhan) or of other violations of the law) are also means of guarding [the capital], there is concern that doing so could develop into an international problem. Special care will be necessary (Rikujō bakuryobu dai 3-bu, 1960, 32).

 

The prospect of taking people into custody was entirely aspirational on the GSDF’s part at that point, but it was also just one of several forms of direct intervention in civil society that Lessons’ analysts hoped would be available to security forces in the event of another earthquake. The two reports’ authors avoided going into specifics, but both studies referenced the need for new legislation or other policy initiatives that would give the police and the SDF the legal authority they said they would need to restore order after a major disaster in the capital district.

This essay explores how the studies’ authors drew parallels between the unrest in Tokyo after the earthquake in 1923 and conditions in the capital in the early 1960s. Their argument, essentially, was that the city was still just a disaster away from disorder and mayhem on a grand scale. My argument is that understanding how the security forces’ experts came to that conclusion is valuable for what it tells us about how the violence that followed the Great Kantō Earthquake has been remembered, for the questions it raises about how “lessons” from one disaster shape preparations for others, and for what it reveals about how vulnerable the police and the military believed Japan’s postwar stability really was.

 

Planning for Disasters in the Early 1960s

The security forces’ novel focus on the Great Kantō Earthquake was in part a product of a growing awareness of Japan’s vulnerability to disasters in general in the early 1960s. The Isewan Typhoon in September 1959, which left more than 5,000 people dead, set the stage for a series of high-profile debates over how best to make sure that the country was better prepared for other catastrophes to come. One of the outcomes of those debates was the government’s designation of September 1, 1960 as Japan’s first-ever “Disaster Preparedness Day” (Bōsai no hi), thereafter an annual occasion for exhibitions of fire-fighting equipment and techniques, simulated emergencies, and evacuation drills in which the public was encouraged to participate (Mizuide, 2019). The 1961 Disaster Countermeasures Basic Law, another legislative response to the typhoon’s effects, made comprehensive disaster preparedness planning mandatory at all levels of government (Kazama, 2002). Speculation about the effects of disasters that hadn’t happened yet was set to become something that the media, policy makers, and state agencies did as a matter of course.

There was also a more particular precedent of sorts for the security forces’ interest in the relationship between what had happened to Tokyo in 1923 and the capital’s vulnerabilities in the present. Starting in the mid-1950s, the Tokyo Metropolitan Fire Department had begun using data gathered in the wake of the Great Kantō Earthquake to produce detailed estimates of how quickly and how far fires would spread if an earthquake as powerful as the one in 1923 were to strike the capital again. The department’s experts had no reason at that point to think that another such earthquake was imminent, but mapping the data from 1923 onto the city Tokyo had become was an excellent way to arrive at a sort of “worse-case” scenario for planning purposes. The police and the SDF were briefed on the fire department’s data and concerns in 1961 (if not earlier), and many of its main findings were included in Research Materials when it was completed in 1962 (Tokyo Shōbōchō, 1961; Ōya, 1977, 205).

A final note about Lessons and Research Materials is that both would have been written and read with the unrest associated with the Anpo protests in mind. The first demonstrations against Prime Minister Kishi’s attempts to ratify a new security agreement with the U.S. began in the spring of 1959, and reached their peak the following June, when hundreds of thousands of protestors marched on the Diet (Kapur, 2018). Lessons was finished a few months before those final, massive demonstrations in Tokyo, and Research Materials a year or so after Prime Minister Kishi’s resignation brought the crisis to a close. Neither study’s authors cited the recent protests as a factor in their analyses of the capital’s vulnerabilities, and yet the intensity and scale of the disruptions in 1959 and 1960 almost certainly made the possibility of unrest in a post-disaster Tokyo imaginable in ways it wouldn’t have been before. The Anpo protests in addition raised questions about how far the government was willing to go to restore order, ones that were likely to come up again in the aftermath of a major earthquake. The police’s actions left little doubt that it was willing and able to use force against civilians, for example, but the SDF’s utility as domestic peacekeepers was not yet clear. Prime Minister Kishi’s June 1960 attempt to deploy GSDF troops against the protestors faltered because the Defense Agency Director, a civilian, personally opposed it, not because there were absolute institutional prohibitions against their use (Skabelund, 2022, 163-164).

With these contexts in mind, we can turn to the studies’ main arguments about the “lessons” of the Great Kantō Earthquake. 

 

Reflecting on Post-Earthquake Violence

Most of Lesson’s content consisted of descriptions of the earthquake’s physical effects on Tokyo and summary accounts of the activities of the army and the police in the days and weeks after September 1. Research Materials covered the same topics in more detail, while also cataloging the changes to the capital’s built environment and patterns of habitation since 1923, and relating experts’ estimates of how much harm another M7.9 event would do were it to strike the city as it was in 1962. Both studies conclude with attempts to connect their analyses of Tokyo’s past and present vulnerabilities to speculation about how to protect the capital in the future. It was clear, for example, that in the event of another powerful earthquake the military and the police would again be responsible for making sure that survivors had access to food, water, shelter, and medical care, that damaged communication and transportation infrastructures would have to be repaired as quickly as possible, and so on. The reports argued that the security forces had to be prepared to perform the same roles after the next big earthquake that their 1920’s counterparts had after the last one. Which was, incidentally, more or less the same conclusion that the Tokyo Metropolitan Fire Department had reached at about what its role would need to be under those same circumstances.

Unlike the fire department’s experts, however, the authors of Lessons and Research Materials also took it upon themselves to reflect on the 1923 disaster’s effects on “public safety” (chian). More specifically, they described the post-earthquake spread of misinformation targeting Koreans in conjunction with the emergence of the jikiedan (vigilante groups), and acknowledged (to a degree) that acts of violence had followed in the wake of those two developments. (Lessons refers to Koreans as senjin throughout; the other study uses chōsenjin.) Lessons implicates the jikeidan in unnamed criminal acts without identifying who was harmed by them, or on what scale. Research Materials is more specific, and supported its descriptive summaries of post-earthquake disquiet with data about when and where the jikeidan were formed (there were 562 in the capital as of September 16, it noted), how many people participated in their activities, and so on (Keishichō keibibu, Rikujō jieitai tōbu hōmen sōkanbu, 1962, 60). The report goes on to state that the jikeidan in and around the capital had targeted Koreans, and that they had harassed, assaulted, robbed and even murdered some of those they encountered. How many may have been harmed is mentioned only in passing. According to Research Materials, the vigilantes in Tokyo had been directly responsible for the deaths of only twenty or so Koreans after the earthquake and injuries to another hundred (Keishichō keibibu, Rikujō jieitai tōbu hōmen sōkanbu, 1962, 73). Those numbers are of course much lower than the ones provided by witnesses at the time and by historians later, but they are in line with some of the tallies that officials had settled on after the earthquake and stuck with since.

In keeping with their search for “lessons” from the disaster, the authors of both studies pivoted from describing what the vigilantes had done to speculating about why they had acted as they did. The police and the military’s experts came to similar conclusions about that, with both settling on a list of factors that they said helped explain the aberrant behavior of their fellow citizens in 1923. Perhaps not surprisingly, both blamed the spread of misinformation about potential sabotage or insurrection by Koreans for the creation of an environment that tacitly justified violence. The sheer volume of misinformation and the stakes of the threats they conveyed had convinced frightened earthquake survivors to take matters into their own hands, the reports implied. Lessons and Research Materials repeated the gist of many of those “rumors,” and advanced several theories about why so many people had been willing to believe that they were true. One factor was of course the intense personal anxiety that survivors of such a traumatic disaster could hardly avoid; another was the absence of more reliable sources of information than whatever news was being spread by word-of-mouth. To those factors the analysts added others that were unique outgrowths of Japan’s colonial policies in Korea and the experiences of Koreans as colonial subjects. Lessons, for example, suggested that (presumably negative) ideas about Koreans harbored in the subconscious of most Japanese were to blame for the violence (Rikujō bakuryobu dai 3-bu, 1960, 12). The 1962 study took a somewhat more nuanced approach, which was to point out that many Japanese assumed that Koreans would of course be resentful over how they had been treated under colonial rule, and therefore found it easy to believe that they would want to take revenge when the opportunity arose. (Keishichō keibibu, Rikujō jieitai tōbu hōmen sōkanbu, 1962, 69).

 

Anticipating Post-Earthquake Unrest in the 1960s

One conclusion that could be drawn from the reports’ analyses of circumstances that were so clearly unique to Imperial Japan in 1923 and their links to the horrors that began on September 1 is that there was no reason to think that another disaster, even one on the same scale as the Great Kantō Earthquake, would once again be followed by such systemic violence and mayhem. If the post-earthquake killings in 1923 had been primarily a phenomenon of empire, in other words, one might conclude that Japan in the 1960s need not fear a repeat of mob violence, vigilantism, or widespread unrest in the wake of the next major seismic event to strike the capital, whenever that might be.

That was not, however, the conclusion at which the authors of the two studies arrived. The police and the military instead interpreted the violence in 1923 as directly relevant to their preparations for Tokyo’s next earthquake disaster. Officials acknowledged the importance of the colonial context and long-standing prejudices against Koreans in their analyses of the violence in 1923, but in the end they assigned much of the blame for that unrest to factors that were neither products of empire nor unique to the 1920s. Both studies, for example, blamed the rapid spread of misinformation (including but not only in the form of rumors) for amplifying the fear and anxiety the authorities claimed had precipitated widespread acts of violence after the earthquake. Lessons and Research Materials took it as a given that the next disaster would generate conditions similar to the last, and that if official channels of communication fell silent (as they well might), false and inflammatory information would once again spread quickly. The security forces also anticipated that the capital’s residents in the early 1960s would react to misinformation and rumors in more or less the same ways as their counterparts in the 1920s had, namely by forming “self-defense organizations” (jikeisoshiki) in misguided and dangerous attempts to maintain “public order” (Rikujō bakuryobu dai 3-bu, 1960, 33). To be clear, the reports’ authors were not suggesting that these groups would necessarily once again target Koreans; rather, their assertion was that frightened and desperate civilians would be highly susceptible to rumors inviting them to focus their fears on some external threat, with unpredictable but potentially lethal consequences.

In addition to their worries about revenant jikeidan, security officials expressed concern about other potential provocateurs. Research Materials claimed that “hooligans” (gurentai) and juvenile delinquents were all but certain to take advantage of the chaos after an earthquake, raiding warehouses and stores, looting abandoned homes, and so on (Keishichō keibibu, Rikujō jieitai tōbu hōmen sōkanbu, 1962, 239-240). Its authors warned too that one or more groups of political activists might launch crime sprees of their own, or exploit survivors’ fears and uncertainties in service of campaigns to force the Diet, the ministries, and/or Tokyo’s government to acquiesce to their demands, whatever those might be. Lessons provides fewer details about who might be behind the threats it expected the GSDF would have to confront, but it is clear from the document’s description of the many locations in Tokyo it anticipated having to guard – the Imperial Palace, government offices, and so on – that it took those threats seriously. The cryptic reference in Lessons to plans to detain “Koreans or other resident foreigners, as well as other individuals of interest (those suspected of thought-crimes (shisōhan)),” cited earlier, is the only time the document hinted at whom the military thought it would be guarding those locations against.

As Kenji Hasegawa has pointed out, official accounts of the post-disaster violence in 1923 were often deliberately ambiguous “about whether or not the reported attacks by ‘Korean malcontents’ after the earthquake were real or not”(Hasegawa, 2020, 116). That rhetorical strategy is also present in the studies under discussion here; both anticipated that misinformation would spread in the wake of Tokyo’s next disaster and that it would again lead to vigilante violence, but both also concluded that gangs of “hooligans,” Koreans and other bad actors actually existed and that they would pose real threats to lives and property. The ambiguities embedded in Lessons and Research Materials are also reflected in the studies’ discussions of the security forces’ plans for preserving the peace, in which their authors describe taking control of the media, putting some “people of interest” under surveillance and taking others into custody, among other interventions, while admitting that they lacked the legal authority to do any of those things (Keishichō keibibu, Rikujō jieitai tōbu hōmen sōkanbu, 1962, 241). Lessons’ call for changes to “the legal and legislative structures related to the maintenance of order” stopped well short of explaining what those revisions would actually entail, or who would pursue them (Rikujō bakuryobu dai 3-bu, 1960, 20).

 

Conclusion

As Aaron Skabelund and others have noted, the SDF’s very visible association with the successes of the 1964 Olympics was one of the key factors in the public’s embrace of increasingly positive views of the military in that era. Its routine participation in relief and rescue operations and in disaster preparedness exercises was another (Skabelund, 2022, 181; Murakami, 2013). Even Tokyo, which for many years had refused to allow the SDF to be part of its disaster preparedness programming, normalized the armed forces’ participation in those events over the course of the 1980s. 

Part of what made those developments possible was that the police and the SDF appeared to abandon the idea of turning to the Great Kantō Earthquake for lessons not long after they first embraced it. Neither the Self-Defense Agency nor the police publicly endorsed either of the two studies, and neither seem to have subsequently lobbied policy makers to grant the security forces the new powers that Lessons and Research Materials implied they would need going forward. The SDF’s spokesmen have not, so far as I can tell, suggested at any point since the early 1960s that Japan’s armed forces were making preparations for dealing with civil unrest or “hooligans” as part of its post-disaster planning, much less that it anticipated taking people suspected of thought crimes into custody.² The security forces’ public-facing plans for dealing with Tokyo’s next earthquake disaster focused instead on activities that had few obvious corollaries with their more problematic interventions in 1923 – providing personnel, helicopters, and heavy equipment in support of large-scale search-and-rescue exercises, for example.

The publication in 1963 of the first in a series of new, analytically rigorous works documenting the actual scale of the post-earthquake violence against Koreans and the state’s complicity in it was an important first step toward a clearer understanding of that history, and thus of any lessons it might yield (Kang Tŏk-sang and Kŭm Pyŏng-dong, 1963). Scholars since have shed yet more light on the killings, on those responsible for them, and on the lingering implications of their crimes (Smith 2023). These developments would certainly have made it difficult for the security forces to stand by the claims that Lessons and Research Materials relied on.

Perhaps more importantly, it was also the case that legislators (primarily on the left) reacted quite strongly to both documents once their contents were known. Throughout the 1960s and 1970s, opposition politicians took to quoting the documents’ more provocative passages during deliberations in the Diet – including the excerpt from Lessons, above – when the government introduced proposals that would have significantly expanded its powers or those of the security forces in times of crisis.³ They argued that the studies’ semi-nostalgic sketches of the authority granted to the military and the police in the 1920s and their complaints about the unfortunate limits on their powers in the present day were nothing more than thinly veiled appeals for a return to pre-war legal and political norms. Self-Defense Agency spokesmen were put in the difficult position of having to admit that Lessons had been written in-house while denying that the document necessarily reflected the military’s thinking or goals. For legislators and members of the public in the 1960s and 1970s already dubious about the purposes to which the SDF might be put, documents like Lessons and Research Materials only heightened their anxieties and encouraged them to push back against a number of government initiatives that might otherwise have faced less opposition. The studies were ultimately far more effective in that cohort’s hands as leverageable evidence that the state was not to be trusted than they ever were in their authors’ as blueprints for enhancing the security forces’ authority.

 

References

Hasegawa, K. (2020) The Massacre of Koreans in Yokohama in the Aftermath of the Great Kanto Earthquake of 1923. Monumenta Nipponica, 75, 91-122.

Kang Tŏk-sang and Kŭm Pyŏng-dong (Eds.) (1963) Kantō daishinsai to Chōsenjin Misuzu Shobō, Tokyo.

Kapur, N. (2018) Japan at the Crossroads: Conflict and Compromise after Anpo. Harvard University Press, Cambridge, Massachusetts.

Kazama, N. (2002) Saigai taisaku kihonhō no seitei: bōsai seisaku netowaku no keisei. Kinki daigaku hōgaku, 50, 1-82.

Keishichō keibibu, Rikujō jieitai tōbu hōmen sōkanbu. (1962) Daishinsai taisaku kenkyū shiryō. Keishichō keibibu, Tokyo.

Magnier,M. (2000) Tokyo Governor Assails Critics, Says Remarks Were Misunderstood. Los Angeles Times, April 13, https://www.proquest.com/docview/421511257?accountid=9758.

Mizuide, K. (2019) ‘Saigo’ no kiokushi: Medeiya ni miru Kantō daishinsai – Isewan taifū. Jinbunshoin, Tokyo.

Murakami, T. (2013) Jieitai no saigai haken no shiteki tenkai. Kokusai anzen hoshō, 41, 15-30.

Ōya, S. (1977) Shinsai ni tsuyoi toshi zukuri no mondaiten to kadai. Hōritsu jihō, 49, 204-211.

Rikujō bakuryobu dai 3-bu. (1960) Kantō Daishinsai kara eta kyōkun: Kantō Daishinsai ni okeru gun, kan, min no kōdō to kore ga kansatsu. Rikujō Bakuryō Kanbu Dai 3-bu, Tokyo.

Sims, C. (2000) Tokyo Chief Starts New Furor, on Immigrants.” New York Times, April 11, https://www.proquest.com/historical-newspapers/tokyo-chief-starts-new-furor-on-immigrants/docview/91695644/se-2.

Skabelund, A. (2022) Inglorious, Illegal Bastards: Japan’s Self-Defense Force during the Cold War. Cornell University Press.

Smith, K. (2023) Introduction, Japanfocus special issue on the 100th anniversary of the Great Kantō Earthquake.

Tokyo Shōbōchō. (1961) Tokyo-to no daishinkasai higai no kentō: taisaku ni taisuru shiryō (dai ippō). Tokyo Shōbōchō, Tokyo.

Tolbert, K. (2000) Old Words Can Still Wound in Japan; Governor’s Use of WWII Reference to Foreigners Stirs Furor.” The Washington Post, April 13, https://www.proquest.com/docview/408610587?accountid=9758.

Yoshikawa Mitsusada. (1949) Kantō Daishinsai no chian kaiko. Hōmufu Tokubetsu Shinsakyoku, Tokyo.