Introduction
The Imjin War, beyond its significant political and economic impacts in East Asia, led to an immediate human consequence: the creation of a large population of Korean captives living in foreign lands.Footnote 1 Most research on these captives focuses on either the diplomatic aspects of repatriation or the cultural and technological contributions of certain groups in Japan (for example, Korean captive ceramists).Footnote 2 However, studies of the captive populations in Japan face structural limitations. When examined through a multidimensional lens, several issues emerge. First, single-factor models – labour demand, technological utility or strategic hostage-taking – cannot explain why some groups, such as the Naeshirogawa potters, were left to fend for themselves for years after their arrival. Furthermore, while existing scholarship tends to characterise the majority of captives as enslaved, with only exceptional individuals like scholars or artisans receiving better treatment, this narrative fails to explain the numerous accounts in historical sources of captives who moved freely, formed their own communities and maintained social networks.
This article therefore reconstructs how captives were allocated to Japan and how they lived once they arrived there. The analysis begins with a crucial observation: not all Korean captives were immediately destined for Japan. During the occupation, Japanese forces employed captives locally – maintaining order, constructing fortifications and harvesting crops. And for those selected for transportation to Japan, their fates were decisively shaped by which of the three transportation routes brought them to Japan: (1) direct sale to Japanese slave merchants in occupied ports, (2) private selection by individual samurai as war trophies and (3) mass transportation during army withdrawals. Each route determined not only their legal status but also their subsequent life chances in Japan. Those sold to merchants typically became transferable property at once. Those personally selected often received preferential treatment as valued acquisitions, while those transported en masse encountered the most variable outcomes.
The mass transportation also challenges prevailing scholarly assumptions.Footnote 3 Contrary to claims that mass-transported captives, save for Korean intellectuals and skilled artisans, uniformly became forced labourers or slaves, substantial evidence reveals a more complex reality. Many found themselves under what this article terms a ‘guarantor arrangement’: a de facto system born from administrative incapacity rather than policy design. Wartime conscription had stripped Japanese towns of clerks and guards, leaving officials unable to supervise the overwhelming numbers of foreign captives, particularly those arriving after 1597. In response, some local authorities either released captives or placed them under the nominal oversight of minor retainers, who rarely interfered beyond preventing escape. Under this arrangement, captives could engage in paid work such as copying texts, participate in seasonal festivities and build extensive social networks – yet they remained vulnerable to punishment if caught absconding. This arrangement explains the paradoxical coexistence of relative autonomy and acute poverty that pervades captive diaries and diplomatic reports.
Before proceeding, a conceptual clarification is needed: were these Korean captives ‘slaves’? Recent global‐slavery scholarship cautions that the term ‘slave’ embraces multiple forms of unfreedom – ranging from Atlantic chattel slavery to Ottoman ransom slavery – each defined by different degrees of commodification, legal dispossession and ‘social death’.Footnote 4 The term beiluren/p’iroin/hiryonin (被虜[擄]人) in the Sinosphere emphasised the fact of seizure rather than a settled legal status. The Great Ming Code, while maintaining a liang/jian dichotomy – a legal distinction between ‘good/common’ (liang 良) and ‘mean/debased’ (jian 賤) statuses – and ranking male nu and female bi among the debased, never stipulated that captives of war automatically became slaves.Footnote 5 Chosŏn law, like Kyŏngguk taejŏn, also contained no provision defining the status of foreign war captives. Yet Chosŏn practice consistently refrained from treating them as chattels. After the 1419 Tsushima expedition, Japanese captives were neither registered as slaves nor allocated as servile labour; they received protection and were subsequently returned as part of a diplomatic exchange.Footnote 6 During and after the Imjin War, the Chosŏn court likewise sought to ‘redeem’ its subjects captured by Japan, indicating that captives were viewed as recoverable persons rather than disposable property. Japanese statutes similarly imposed no rule of automatic enslavement.Footnote 7 Even Japan’s semi-servile statuses – genin and shojū – retained the capacity to marry, accumulate property and in some cases regain freedom.Footnote 8 Social reality in Japan thus operated along a spectrum of unfreedom rather than within the binary logic of Atlantic plantation slavery. For these reasons, this study therefore employs ‘captive’ for Koreans taken during the Imjin War. This distinction highlights the variability of their trajectories: some rebuilt kinship ties, secured marriages and negotiated return voyages; others sank into life-long dependency. Their experience was a staged, mutable process of constraint, not a seamless replica of Atlantic plantation slavery.
To achieve this analysis, this article employs a range of primary sources. It draws first on records left by the captives themselves, such as Kang Hang’s Kanyangnok and the Chŏng brothers’ diaries, which provide crucial insights into the conditions of their capture and subsequent lives. In addition, it utilises diplomatic documents produced by official institutions as well as Japanese domestic records. Korean officials’ memorials and interrogation records concerning the captives, along with reports from Korean communication envoys (t’ongsinsa) dispatched to Japan in the early seventeenth century, offer further perspectives on the captive experience.
The situation on the battlefield
A considerable proportion of the captives taken by the Japanese army on the Korean peninsula consisted of Korean soldiers. When the Christian daimyo Konishi Yukinaga, who commanded the Japanese vanguard, landed in Korea, his attendant monk Tenkei recorded that in Tongnae and Chungju, over 3,000 were killed and hundreds captured.Footnote 9 These accounts reveal both the brutality of war and the deliberate decision to spare some enemy troops. Korean officials were also prime targets.Footnote 10 The Veritable Records of the Chosŏn Dynasty recorded many instances of Korean officials and lower-ranking officials being captured. This suggests that the Japanese army may have considered these individuals to have higher value, whether as hostages, sources of intelligence or potential labour.
The Japanese army’s plundering was not limited to military and political targets. Fleeing Korean civilians also became targets for capture. Although the Japanese ruler Toyotomi Hideyoshi expressly prohibited this practice, his decree proved challenging to implement effectively. Furthermore, civilians often carried large amounts of luggage and valuables, making them attractive targets for plunder. For example, when Korean Confucian scholars and brothers Chŏng Hŭi-tŭk and Chŏng Kyŏng-tŭk were fleeing with their families by boat, they were unfortunately captured on the twenty-seventh day of the ninth month of 1597, by Mori Tadamura, a general under Hachisuga Iemasa, the daimyo of Awa domain.Footnote 11 From Chŏng Kyŏng-tŭk’s record of their friend Cha-p’yŏng’s silver cup being taken by Japanese soldiers, it is clear that their personal belongings were also plundered by Japanese soldiers.Footnote 12
These captives taken during the war were generally first gathered in Japanese army camps or occupied cities. In 1592, the high-ranking Korean scholar-official Ryu Sŏngnyong wrote in a letter to Ming generals that nearly two-thirds of the people in Hansŏng were Korean captives.Footnote 13 The Chŏng brothers also recorded in their captivity diary that after being captured, they experienced more than ten days of drifting at sea before being gathered with captives from other places at the camp of Japanese general Mori Tadamura.Footnote 14
Most captives taken during this period were not initially intended for transport to Japan but were used locally to support the Japanese army: maintaining order, building fortifications and harvesting crops. On the twenty-second day of the tenth month of 1597, Yi Tŏkhyŏng reported that Japanese soldiers near Kurye, Chŏlla-do, were forcing 200 Korean captives to harvest cotton and grain. Upon spotting Ming forces, the Japanese fled with the captives, though only sixty chose to escape. Many captives held ‘Freedom from Death’ certificates issued by Konishi Yukinaga’s troops, granting them protection and limited freedom of movement within Japanese camps.Footnote 15 This practice indicates that the Japanese army adopted a relatively flexible approach in managing captives, which differs from the traditional notion of captives completely losing their freedom. The example of the spy Sŏ Yun-bok, mentioned in Korean official Yi Ho-min’s 1593 report, further confirms this. Sŏ Yun-bok, disguised as a beggar, infiltrated the Japanese-occupied Imnan’po and communicated with captives, suggesting they retained a degree of mobility and intelligence value.Footnote 16
The battlefield was therefore the first site where Korean populations were assigned to divergent fates. As Rebekah Clements has shown, Japanese forces frequently took severed heads, and at times other body parts, as recorded proof of conquest. These trophy-taking practices were not merely lethal mechanisms but constituted a hierarchy of violence: while elites were often beheaded, commoners faced nose- and ear-severing – a practice that functioned as both a quantifiable tally of conquest and a ritualised form of punishment, often leaving victims alive but permanently stigmatised. This implies that captivity and mutilation were not mutually exclusive outcomes but parallel strands of a broader ‘hegemonic hunt’.Footnote 17 Nam-lin Hur’s demographic reassessment also demonstrates that at least 400,000 Koreans were systematically slaughtered during the invasions, amounting to roughly 3–5 per cent of the entire Chosŏn population within six years.Footnote 18 As villages were emptied through slaughter, flight and the collapse of local authority, abducting Koreans became routine: some were taken as labourers or porters near the front, others were retained as assets by individual commanders, and many more were assembled for transport to Japan. In this sense, captivity emerged directly out of the demographic and social disintegration produced by wartime violence, rather than as a separate or secondary phenomenon.
Resale of Korean captives
Many captives on the Korean peninsula were employed to assist the Japanese forces in maintaining local governance and agricultural production in the occupied areas. As a result, they enjoyed a certain degree of mobility and were not always held under strict confinement. However, this relative liberty did not prevent the practice of reselling these captives by the Japanese soldiers. Reports from Admiral Yi Sun-sin in 1594 frequently highlight that Korean captives were often sold in the port of Pusan subsequent to their capture.Footnote 19 Hŏ Nŭng-nyŏn, a regular soldier from Kŏje County, was captured by Japanese forces on the twenty-fourth day of the seventh month of 1593, together with his wife, their two daughters (aged twelve and seven) and four others, and was held at the Changmun’po Ochŭngkam camp. Their twelve-year-old daughter was subsequently separated from the family and sold on elsewhere (her whereabouts are not recorded). Hŏ, his wife and their seven-year-old daughter were later resold to Japanese in Pusan. Taking advantage of the Japanese going out to buy alcohol, they stole a boat and escaped to Ungp’o.Footnote 20 Another example is Cho Yun-sin, who was a special guard from Chise-p’o in Kŏje. In the seventh month of 1593, while hiding in a mountain valley, he was suddenly captured by ten Japanese soldiers and was eventually sold to Pusan.Footnote 21 The latter example provides important clues for us to consider under what circumstances captives might encounter being resold. Compared to captives taken in large-scale military operations, those captured in smaller actions were more likely to be privately sold by Japanese soldiers.
The primary purchasers at ports such as Pusan were Japanese slave traders, the hitoakibito. Naturally, they also capitalised on opportunities to acquire Korean captives during the Imjin War period. Thomas Nelson and Lúcio de Sousa both point out that the early seventeenth-century Vocabulário da língoa de Japam included entries for Fitoaqibito (hitoakibito) and fitocaibune (hitokaibune, a boat that carried slaves) and note the close relationship between these slave merchants and the Portuguese.Footnote 22
These hitoakibito may have voyaged to Korea on merchant ships and other civilian vessels, purchasing captives from Japanese soldiers at ports like Pusan. As Keinen, a monk doctor accompanying Ōta Kazuyoshi, noted in his diary, Chōsen nichinichiki or ‘Daily Diary of Chosŏn’, he witnessed Japanese merchants in Pusan forcibly enslaving Koreans. In Kyŏngsang-do, he also observed Japanese troops plundering the countryside, enslaving captives and restraining them with bamboo chains; even children were not spared.Footnote 23
The captives purchased by the hitoakibito were subsequently transported back to Japan for resale. Francesco Carletti, an Italian merchant who visited Kyushu in 1597, provided a first-hand account of this phenomenon. In his chronicles, Carletti described witnessing large numbers of Korean men and women being sold as slaves at remarkably low prices. To illustrate the extent of this price depreciation, Carletti himself purchased five Korean slaves for only 12 scudi, a relatively small sum at the time.Footnote 24 On the twenty-ninth day of the sixth month of 1604, a Japanese trading ship bound for Luzon that had drifted to the coast of Kyŏngsang Left Province due to strong winds was captured by the Korean navy. On board was a Korean captive named Pak Ch’ung, who had been captured and taken to Japan in 1597. After being captured in Yŏng’am County, Korea, he was taken by ship to Nagasaki and subsequently resold to someone in Hirado.Footnote 25 He was probably sold by Japanese soldiers to hitoakibito at the time and taken to the slave market in Nagasaki.
The proliferation in Japan of Korean slave purchases by European traders, particularly Portuguese merchants, was not solely attributable to lower prices. A significant contributing factor was the contemporaneous prohibition on enslaving Japanese natives. This created a vacuum in the labour market, which the influx of Korean captives inadvertently filled. The extensive acquisition and overseas transportation of Japanese slaves by Portuguese merchants ultimately incurred the wrath of Japanese authorities. On the thirteenth day of the seventh month of 1587, Toyotomi Hideyoshi issued an edict via the Jesuits banning the purchase of Japanese slaves and ordered their repatriation from overseas.Footnote 26 The order had a certain deterrent effect. However, the Portuguese demand for slaves did not cease. As a result, Korean captives taken to Japan provided a substitute for Japanese slaves. As de Sousa has noted, just as European merchants such as Carletti purchased Korean slaves in Nagasaki, large numbers of Korean captives were bought by Europeans, led primarily by the Portuguese, and taken out of Japan. The scale of this phenomenon was such that it drew the attention of leading members of the Society of Jesus in Japan, who, at a meeting in September 1598, explicitly discussed the problem of Portuguese traders transporting Koreans abroad, many of whom were very likely being sold as slaves or bound as indentured labourers.Footnote 27
Large-scale transport of captives
From the beginning of the Imjin War until its end, the Japanese army systematically transported Korean captives back to Japan, most notably during the Ming–Japanese peace negotiations in the sixth month of 1593 and again at the end of the war in 1598. During these phases, large numbers of captives were evacuated alongside the retreating forces. Korean scholars Kang Hang and Chŏng Hŭi-tŭk (themselves captured and taken to Japan) describe how hundreds of captives were forcibly loaded onto ships as collective spoils of war for various daimyo.Footnote 28
Contemporary estimates place the number of Koreans transported to Japan at between 20,000 and 100,000, yet Chinese and Korean reports reveal a far larger captive pool that never left the peninsula.Footnote 29 In the ninth lunar month of 1593, Chinese commissioner Song Yingchang recorded between 10,000 and 20,000 captives clustered on islets off Pusan who had sought passage with the retreating Japanese to avoid Korean retribution, but were refused.Footnote 30 As this figure concerns only one occupied sector in a seven-year conflict, the total held in Japanese service must have greatly exceeded those actually conveyed overseas, foregrounding the question of how Japanese commanders selected who would depart. Age appears to have been a decisive consideration. The fate of the Chŏng family is instructive. Captured at sea and brought to Tagŭng-p’o with Mori Tadamura’s fleet, the brothers witnessed an explicit age-based triage: their father and three small boys (aged two, four and seven) were released as ‘too weak’, whereas their eight- or nine-year-old niece was retained for transport.Footnote 31 This aligns with the earlier case of Hŏ Nŭng-nyŏn, whose seven-year-old daughter was transferred along with her to another Japanese soldier. These examples suggest an operational threshold at around seven years of age, below which children were considered burdensome, and above which they were seen as viable for relocation to Japan.
The Japanese army in some cases maintained the integrity of captive families. Kang Hang records that after capture at Sunchŏn Left Naval headquarters, Japanese general Sado-no-kami gathered Kang, his brothers Kang Chun and Kang Hwan, his father-in-law Kim Pong and other family members on the same ship before escorting them to Japan.Footnote 32 The Chŏng brothers record similar arrangements: on the thirteenth day of the tenth month of 1597, Japanese commanders moved Ch’oe Tŏk-yang from another ship to reunite him with his wife and children.Footnote 33 After arriving at Awa Castle (now Tokushima city), Chŏng Hŭi-tŭk wrote that since the Japanese did not separate families when allocating captives, he pretended to be his friend Chŏng Cha-p’yŏng’s brother to avoid separation.Footnote 34
The practice of maintaining family integrity may have been based on multiple considerations. First, the Japanese army may have hoped to reduce anxiety and unease among captives, thereby reducing potential resistance or escape attempts. The mutual support and comfort between family members might have made it easier for captives to accept their situation, at least in the short term. Second, this practice also facilitated management. By keeping relatives together, Japanese commanders reduced the risk of flight: any would-be runaway knew that a parent, spouse or child could be punished in reprisal. The emotional dependence within a family thus served as a built-in restraint while also simplifying supervision. For instance, in 1617, Yi Kyŏng-chik, the document officer of Korean embassy to Japan, reported that Korean captive Yang Mong-in served Wakizaka Yasuharu, a lord of the Ōtsu domain. As Yang Mong-in was a favourite of Wakizaka, Wakizaka took his mother hostage, fearing that he might escape.Footnote 35 This form of leverage was not unique to the treatment of foreign captives; it was paralleled domestically by the Tokugawa shogunate’s practice of holding daimyo families in Edo as hostages to ensure the lords’ loyalty.Footnote 36
Captives with special talents or political value received special treatment. For example, Kim Tŏk-hoe, a Japanese language interpreter who surrendered to Tsushima daimyo Sō Yoshitoshi in 1592, was appointed magistrate of Yŏngyu County for providing information and assisting the Japanese army. Konishi Yukinaga planned to take him to Japan in the twelfth month of 1592 along with his eighty-year-old mother, wife and children.Footnote 37 This decision might have been to ensure Kim Tŏk-hoe could stay in Japan with some peace of mind. Another example is Hong Ho-yŏn, captured at the age of twelve, who later became secretary to Nabeshima Naoshige and Katsushige of Saga Domain. Nabeshima noticed him carrying a huge writing brush when captured and sent him to study at a Gozan temple.Footnote 38 Japanese scholar Tsuruzono Yutaka notes that many young captives during the Imjin War demonstrated proficiency in writing and poetry, and as a consequence attracted daimyo attention for cultivation of their literary talents once they arrived in Japan.Footnote 39
While Song Yingchang’s account indicates that some captives desired transportation to Japan but were unable to secure it, other captives employed diverse strategies to evade such relocation. Captives exhibited agency throughout their captivity experience. Taking the previously mentioned case of the Japanese-language interpreter Kim Tŏk-hoe as an example, when he learned that Yukinaga planned to take his entire family to Japan, he used the argument of family responsibility, telling Japanese guards: ‘We must stay with our aged mother, wives and children, and cannot travel far, so we humbly request to be released.’ The guards then sent his family out of camp first, allowing Kim and others to escape at midnight.Footnote 40 While Kim Tŏk-hoe’s account lacks detail on his dialogue with the guards or their motives, the episode suggests that decisions about captive transport could be negotiated and were not always mechanically enforced. The case of Kim Ŭng-ji, a Kyŏmsabok cavalryman, sheds further light on the complexities of captive management. Captured by the Japanese on the nineteenth day of the seventh month of 1593, he was transferred several times before reaching the Japanese camp at Pusan port. There, he witnessed multiple shipments of captives to Japan but avoided being selected and eventually escaped.Footnote 41 These cases show that the transfer process was less tightly controlled than assumed, with room for negotiation. Whether through rapport with guards or by leveraging informal channels within the camps, captives actively sought to influence their fate.
Selective transfers
Although large-scale transportation was the primary method used by the Japanese army, small-scale, individualised transfers also occurred. For example, on the eighteenth day of the fifth month of 1592, a vessel arrived at Kushikino port in Satsuma carrying only six or seven Korean captives.Footnote 42 Even in the final phase of the war, when Kang and Chŏng were captured, captives were not always moved in large groups. A sailing certificate issued by Shimazu Tadatsune on the seventeenth day of the seventh month of 1598 for captain Ozaemon’i recorded just two Korean captives among fourteen passengers.Footnote 43 This distinction between mass and sporadic transport was driven less by vessel capacity than by the captives’ ownership status. Individually transferred captives had often been preassigned to specific owners, whereas group transports typically involved unassigned captives representing domain spoils.
Contemporary sources show that junior samurai and commanders often selected captives from battlefields or markets, sending them home as enslaved war trophies. Ōshima Hisazaemon Tadayasu jyū Kōrai no bun utsushi (The Letters of Ōshima Hisazaemon Tadayasu from Korea), a rare epistolary record from the Imjin War by Shimazu vassal Ōshima Tadayasu, offers key insights here. In one letter to his wife, he described sending a male and female Korean captive home via his servant Kakuemon. He also mentioned plans to send a Korean woman as a servant for his daughter, and hinted at another captive intended for Zaemon’i, probably a friend.Footnote 44 Similarly, Luís Fróis reported on 20 October 1595 that the Tsushima daimyo Sō Yoshitoshi had recently gifted his wife (Konishi Yukinaga’s daughter) two Korean boys taken from Korea.Footnote 45
However, the personal selection and repatriation of captives was largely limited to daimyo and high-ranking samurai. For lower-ranking samurai and common soldiers, it was difficult and costly. In 1597, Shirao Magokurō, a Satsuma vassal, wrote to his relative Shirao Goei’i requesting supplies from home, noting: ‘I cannot send back the teruma (Korean male) and kakusei (Korean female) this time because I cannot afford the ship fare.’Footnote 46 He probably lacked access to military vessels or funds to hire escorts. In a letter the following year, he reported looting silver from Korean corpses in battle, though their general had warned that the priority was killing the enemy, not taking money. With the silver gained, he then hired a monk to escort his two captives home. He also mentioned asking Tanejima Rokubei to return a Korean woman requested by someone named Fujisuke and asked to be informed once she arrived.Footnote 47
The cases of Ōshima and Shirao illuminate a crucial distinction in captive allocation outcomes. Those pre-selected by daimyo or samurai for individual transport arrived in Japan with predetermined destinations and masters: a selection process that inherently favoured captives deemed valuable enough for personal acquisition. This elite patronage typically translated into superior treatment, as high-ranking warriors possessed both the resources to maintain captives properly and the social incentive to display their magnanimity. By contrast, captives transported en masse during military withdrawals faced an entirely different fate. Distributed impersonally as domain property, primarily to common soldiers as war rewards, occasionally even to lower-class Japanese as servants, these captives entered households that often lacked both the means and motivation to treat them well. A 1655 memo by Sago Shikiemon, a Tsushima vassal involved in Chosŏn affairs, reflects this system. Writing at the age of 72, he recalled that Hideyoshi’s campaign aimed to bring Korean men, women and children to Japan, and that the policy allowed even those of low status to become temporary masters. There was anticipation that future Korean wars might yield more slaves, underscoring how wartime population control involved not only mass transfers but also a deliberate and hierarchical distribution of captives.Footnote 48
The aftermath of mass transport and the ‘guarantor system’
Yet the thousands of Korean captives shipped to Japan were not simply parcelled out to local households or lower-class soldiers as slaves. When these large consignments arrived, especially in the final years of the war, Japan’s wartime bureaucracy was too depleted to allocate them systematically. Their post-disembarkation trajectory therefore cannot be reduced to automatic enslavement. Substantial historical records contradict this narrative, revealing the existence of many captives who possessed a degree of mobility and appeared to be without masters. For example, in 1600, Sin An-nam, a scholar from Yangsan, escaped from Japan and returned to Chosŏn Korea. He informed the Korean authorities that the captives were generally idle and not forced into labour.Footnote 49 While Sin An-nam’s description may be somewhat exaggerated, his testimony, alongside similar accounts, suggests that the scale of the influx, combined with post-war administrative chaos, impeded systematic allocation and supervision, creating spaces in which some captives were left unassigned, comparatively idle and able to move about without a fixed master.
A more specific example comes from Chŏng Hŭng-pang. He was repatriated to Korea on the sixteenth day of the third month of 1606 by Tachibana Tomomasa, a vassal of Tsushima domain. In his debriefing, he stated that he had been captured on the twenty-fifth day of the ninth month of 1597 and had led the life of a beggar in Hizen for eight years until his relocation to Tsushima Island by the Tsushima daimyo at the end of the previous year.Footnote 50 Together, Sin An-nam’s report and Chŏng Hŭng-pang’s ordeal show that late-Imjin captivity ranged along a spectrum that ran from tightly controlled household slavery to quasi-autonomous, legally unanchored survival on society’s margins.
Such accounts force us to reconsider fundamental questions about the captive experience: Why were some captives able to obtain relative freedom while others were reduced to slavery? How did Japanese society manage and integrate these involuntary immigrants?
Part of the answer lies in wartime habit. As Fujiki Hisashi has shown, the private seizure and sale of persons was endemic to Sengoku-period warfare; Japanese soldiers appear to have projected these domestic habits onto the Imjin front.Footnote 51 Senior commanders might select a few captives as trophies or gifts, but those seized by rank-and-file troops were shipped home en bloc as domain property. Decisions about the fate of these captives were postponed until after disembarkation, and urban authorities, overstretched by wartime mobilisation, often lacked the capacity to impose immediate control. This administrative vacuum explains why some captives roamed freely and temporarily evaded forced labour.
Indeed, after many Japanese male labourers were conscripted into the army, managing the sudden influx of captives into ports in certain areas (particularly in Kyushu and the domains of Western Japan) became practically unfeasible. This was exemplified by the considerable proportion of Korean captives in the streets of Awa Castle Town, as described by Chŏng Hŭi-tŭk. Furthermore, amidst the chaos of war, providing sustenance and shelter for captives and their families posed a significant burden, despite the potential utilisation of captives as labourers. Many Koreans therefore enjoyed what might be called ‘relative freedom’, in that they were spared conscript labour and were able to manage their daily routines largely on their own, circulate locally and mingle with fellow captives, although they still lacked the resources and ability to leave Japan. It seems plausible to conclude that some captives existed outside official management or control.
The case of Che Man-ch’un offers an interesting perspective on this issue. Che Man-ch’un was from Kyŏngsang Province and was a low-rank probationary military officer (Hullyŏn pongsa) in Kosŏng. Captured at the onset of the Imjin War, he was taken to Hizen Nagoya in Japan. According to his report, Hideyoshi initially intended to burn him, but upon discovering Che’s literacy, he was instead sent to the house of a Japanese man named Hansuke, who acted as his guarantor. For ‘guarantor’, Che uses the word posu (to take charge of a bailed person). Under Hansuke’s guarantee, Che was exempted from the arduous labour that other Korean captives were forced to undertake. Instead, he devoted significant time to locating other captives and strategising his escape. He mentioned that in Nagoya there were over twenty Korean captives in the larger houses, eight or nine in the medium-sized houses, and three or four in the smaller ones. It is plausible that these captives were also beneficiaries of the guarantor system, and thus they were able to meet with each other from time to time, even at night.Footnote 52
Che’s account provides a possible explanation for this apparent ‘relative freedom’ among some Korean captives, namely, the placement of multiple captives under a Japanese guarantor rather than under individual household masters. In some cases, as described by Che, multiple captives were placed under the management of a Japanese guarantor. Furthermore, under this arrangement, they were not required to serve the masters, but were allowed freedom of movement and communication, to the extent that they were even able to plan escape strategies at night. This suggests that the guarantors functioned more as nominal overseers than actual masters, exercising minimal interference in the captives’ daily activities, largely because they lacked the capacity to control effectively and manage such substantial numbers of Korean captives. The absence of official documentation suggests this was not formal policy but rather a localised response to overwhelming numbers. This arrangement may also explain why the Jesuits stationed in Nagasaki were able to interact with and educate a substantial number of Korean captives during the war. Jesuit records from 1594 document the baptism of over 2,000 Koreans in Arima, Ōmura and Nagasaki,Footnote 53 with Luis Frois reporting an additional 1,300 Korean baptisms to Claudio Aquaviva, the Jesuit General, in 1596.Footnote 54 Some of these captives even found time to assist the Jesuits in translating Christian prayer books (devocionarios) into the Korean language.Footnote 55
Figures such as Kang Hang and Chŏng Hŭi-tŭk – Korean literati captured during the war – appear to have lived under an informal guarantor arrangement that exempted them from manual labour while keeping them under discreet supervision as well. Chŏng Kyŏng-tŭk records that he and Chŏng Hŭi-tŭk were treated as ‘esteemed guests’, yet a diary entry on the first day of the fourth month of 1598 notes that Japanese escorts prevented them from speaking with a friend, implying the quiet presence of a guarantor who was also conveying them to meet Tōshūza, abbot of the Hachisuka clan temple.Footnote 56 A comparable pattern emerges in Kang Hang’s case: seized by Tōdō Takatora in 1597 and brought to Ōzu, he was spared hard labour, provided with a male and a female servant, and placed in the custody of the captor’s retainer Nobushichirō.Footnote 57 When Kang’s attempted escape failed in 1598, it was Nobushichirō, acting in his guarantor role, who recaptured him, saved him from execution (others were beheaded) and returned him to Ōtsu, illustrating that monitoring and retrieving runaways formed part of a guarantor’s tacit duties.
Many captives who appear to have been managed under this guarantor arrangement, although they enjoyed a degree of limited autonomy, largely depended on their own efforts for survival. The Chŏng brothers often struggled with procuring food. An entry in Chŏng Kyŏng-tŭk’s diary, dated the twentieth day of the ninth month of 1598, mentions his nephew Chŏng Ho-rye being bedridden due to lack of food.Footnote 58 To generate income, they had to resort to transcribing books and inscribing letters for the Japanese.Footnote 59 The Chŏng Hŭi-tŭk brothers’ diary frequently references foraging trips to the mountains to collect mugwort for soup,Footnote 60 and when food was scarce, they would mix mugwort with rice to stave off hunger. They even resorted to begging at Buddhist temples and neighbouring villages during food shortages.Footnote 61
Literate captives, such as the Chŏng brothers, occasionally received relief from Japanese intellectuals and daimyo, owing to the respect they had garnered. However, other common captives – those managed under the guarantor arrangement or left without proper oversight due to manpower shortages, especially those transported to Japan after 1597 – led lives of abject misery as a result of the Japanese army’s retreat and subsequent political upheaval. Some captives were often abandoned in one location and left to their own devices. The Korean potters from Naeshirogawa, credited with pioneering Satsuma ceramics, were dispatched to Kagoshima in three separate groups in 1598. They were housed by the Satsuma domain in Shimadaira (presently known as Kijino).Footnote 62 The aftermath of the Imjin War and the subsequent onset of the battle of Sekigahara, however, resulted in these Koreans receiving scant attention from the Satsuma clan. Until 1606, they were compelled to grow their own crops, with those possessing pottery skills earning a living through their craft. This implies that some captives, like Chŏng Hŭngpang, who were transported to Japan towards the end of the war, were largely neglected upon their arrival. There were also unskilled Korean captives who had to beg in the towns.
Although this loosely supervised guarantor arrangement contributed, in some cases, to the impoverishment of captives, who had to make a living in an unfamiliar country while lacking stable resources and support, it also meant that they could enjoy a greater degree of freedom of movement and opportunities for social interaction at the local level. This dynamic is vividly illustrated in the narrative of Chŏng Hŭi-tŭk and his companions’ arrival in Awa in 1597, as news of their presence quickly spread throughout the area. Just four days after their arrival, on the third day of the lunar new year in 1598, fellow Korean captive Ryu Chungwŏn rushed to visit them upon learning of their presence.Footnote 63 Korean captive Kang Hang likewise chronicled in his captivity narrative encounters with other Korean captives in Ōtsu, with whom he engaged in poetic exchanges.Footnote 64 Che Manch’un, a Korean assigned to labour for Hansuke, articulated in his captivity narrative a vivid account of his struggle with severe rheumatism. Upon recuperation, he likened his existence to that of ‘a bird trapped in a cage’, longing for his homeland and yearning to conspire with fellow Korean captives towards a future escape. Therefore, he regularly searched for Korean captives in the city and successfully contacted Korean officials and some private slaves from Kimhae, Changwon, Milyang and Ulsan.Footnote 65 Chŏng Hŭi-tŭk also recorded in his diary, Wŏlbong haesangnok, that:
there is a river below Awa Castle and a rainbow bridge across it. Every time there are ten people on the bridge, eight or nine of them are Koreans … On moonlit nights, Koreans gather on the bridge. They sing, talk, express their feelings and sometimes cry. They stay there until late at night. The bridge can hold more than a hundred people.Footnote 66
Chŏng Hŭi-tŭk himself often met with other captive friends on the street, confiding in each other and even crying out of deep emotion.Footnote 67
For the captives dwelling in an alien environment, the cultivation of communal activities in their free hours became a means of sustaining a semblance of cultural identity and a sense of belonging. These interactions transcended mere social gatherings, serving as attempts to replicate and preserve the customs and rituals intrinsic to their homeland. A striking illustration of this is found in Chŏng’s narrative, in which he details the events of the fifth day of the fifth month of 1598. On that day, at the Korean Tano Festival, Korean captives assembled in the streets to engage in traditional folk entertainment. Adhering to their cultural heritage, they constructed a Korean-style swing and indulged in playful activities, providing amusement to an observing Japanese crowd.Footnote 68 The incident offers a vivid tableau of cultural resilience, as the captives sought to recreate a cherished celebration from their homeland amidst their captivity. On New Year’s Day, Korean captives would embark on visits to each other’s dwellings, engaging in festivities and exchanging salutations in alignment with their indigenous social hierarchy. This practice underscores not merely a yearning for communal connection but an unyielding adherence to cultural norms and etiquette, even in the most unpropitious circumstances.
However, as noted above, the guarantor arrangement emerged organically from wartime labour shortages, and once the conflict neared its end and Japanese troops came home, the captives who had enjoyed this limited autonomy faced reallocation. The Chŏng brothers witnessed this shift. They documented an encounter with a Korean captive, Chu Hoe-paek, in Awa. Despite Chu’s limited writing skills, he and the Chŏng brothers bonded over their shared aspiration to return to their homeland. They met each other daily and discussed their return plan. However, their daily connection faced a turning point in the fifth month of 1598. This shift coincided with the daimyo Iemasa’s return to Awa, leading his troops in retreat from the Korean peninsula. The captives in Awa city underwent a reassignment, probably due to the military presence increasing the manpower capable of overseeing other captives who were not under a master. Upon learning on the fifteenth day of the fifth month that the reallocation would separate the brothers, Chŏng Kyŏng-tŭk penned a detailed letter to Iemasa, imploring him to reconsider their separation.Footnote 69 Ultimately, Iemasa acceded to their request. Yet their friend Chu Hoe-paek was delegated to serve under Iemasa’s vassal, Inada Shirō, and sent to the mountains for forestry work.Footnote 70
Conclusion
To understand the captives during the Imjin War better, we must avoid a simplistic view that once taken to Japan, they inevitably became subservient slaves. In fact, the complex dynamics of agency and autonomy among these individuals offer vital insights into their historical circumstances. This complexity stems largely from their personal abilities and limitations, as well as the deeper factors that shaped their particular predicaments, all of which can be traced back to the captive allocation system employed during the conflict, a system that ultimately produced widely divergent fates.
As the war progressed, captives were gradually sent to Japan. Some were given as gifts by lower-ranking samurai or generals to their families, while the majority were gathered in Japanese camps and transported by ship alongside military forces. Unlike the domestic war captives of earlier Japanese conflicts, these foreign captives were continuously dispatched even while the fighting continued. Japan’s castle towns, already strained by conscription shortages, could not cope with the sudden influx of Korean captives – a point illustrated in Chŏng Hŭi-tŭk’s writings, which frequently describe encountering Korean captives in the streets during the day. Confronted with numbers far exceeding their capacity, local officials and bureaucrats devised an unwritten system to manage the situation, relinquishing direct control over some Korean captives and, in certain cases, merely assigning them nominal overseers. Under this arrangement, while the captives were spared the burden of hard labour, they nonetheless faced the challenge of lacking basic sustenance. For instance, the Jesuit Carlo Spinola recorded an intriguing case: a Korean captive in Nagasaki, though granted his freedom, chose to continue working for the Japanese army in order to earn a living, rather than attempting to escape back to the Korean peninsula.Footnote 71 These sources show that captives navigated a wide range of constraints and opportunities, but always within structures of dependency shaped by the circumstances of their seizure and the institutional context into which they were absorbed.
This study has therefore argued that the fates of Imjin War captives cannot be reduced to a single model. Their experiences were produced at the intersection of multiple forces: the demographic devastation of the war, the classificatory violence that began on the battlefield, the pressures created by the mass transport of civilians, and Japan’s improvised systems for incorporating foreign captives. Situating these trajectories within broader discussions of coercion and early modern mobility highlights the need to understand captivity as a spectrum of conditions rather than a fixed status. Such an approach not only clarifies the internal diversity of the Imjin War captives but also positions their stories within wider global patterns of early modern coerced movement.
Open access
