Status-based social structure

To delineate the features of marital age differences before and after the Imjin War, it is imperative to consider the social structure that influenced marriage during that period. Since the Chosŏn society was structured along status-based lines, a person's socio-familial background was pivotal for spousal selection. In this context, variations in marital age disparities would likely correlate with social standing.

The Chosŏn government officially stratified its populace into two distinct legal categories. The first group comprised the yangin 良人, who were the king's subjects with various obligations, including taxation. Nevertheless, they were eligible to partake in civil service exams (munkwa) 文科 and hold government positions (Hwang Reference Hwang2004; Palais Reference Palais1996). The second group consisted of the lowborn (ch’ŏnmin) 賤民 (Deuchler Reference Deuchler1992, p. 13), most of whom were unfree laborers known as nobi 奴婢, who were either owned by the government or owned privately (Hwang Reference Hwang2004; Kim Reference Kim and Campbell2004).Footnote ⁶

In contrast to Chosŏn society's two-tier stratification system, this study adopts a three-class system (Ch'oe Reference Ch'oe1974, Reference Ch'oe1987, pp. 103–4; Deuchler Reference Deuchler1992, pp. 302–3; Kim Reference Kim and Campbell2004, p. 153) to account for the significant disparity between the ruling and ruled classes within the yangin sector of the population. Yangin thus require further subdivision into yangban elites and commoners, traditionally referred to as sangmin 常民.Footnote ⁷ The yangban constituted a group of people who held government posts or were bureaucratic candidates preparing for civil service exams. The Chosŏn government exempted them from military service. Hence, individuals with non-military responsibilities, such as bureaucrats and young students (yuhak) 幼學 (Ch'oe Reference Ch'oe1987, pp. 135–37; Duncan Reference Duncan2000, p. 195; Han Reference Han2021, note 15), should be classified as yangban. The commoner status largely consisted of peasants with diverse obligations to the monarch, including military service. Individuals below this status were categorized as lowborn, including nobi.

The three statuses – the yangban, commoners, and lowborn nobi – shaped the patterns and opportunities of marriage and remarriage in Chosŏn society. Gender also wielded considerable influence; for instance, societal tendencies regarding remarriage under the sway of the Neo-Confucian norm of chaste widowhood varied not only across social statuses but also along gender lines. In this context, examining spousal age gaps across different statuses can reveal the extent to which Neo-Confucian marital norms permeated society. This is particularly pertinent when discussing the period after the Imjin War when the deaths of spouses reached unusually high levels.

Spousal age differences and marriage norms

Several studies have investigated age differences between spouses in traditional East Asian societies (Lee and Campbell Reference Lee and Campbell1997; Lundh and Kurosu Reference Lundh and Kurosu2014; Tsuya et al. Reference Tsuya, Wang, Alter and Lee2010, pp. 82–87; Wolf and Huang Reference Wolf and Huang1980). However, due to a lack of data, the spousal age gaps specific to Chosŏn society around the time of the Imjin War have not been methodically explored. While some notable studies, such as the work of Kuen-tae [Kŏnt'ae] Kim (Reference Kim2004, Reference Kim2005, Reference Kim2006), have shed light on marriage age, customs, and marital age gaps, they focus on the eighteenth and nineteenth centuries, rendering them unsuitable for understanding the Imjin War's impact on nuptiality. Furthermore, Yi (Reference Yi2001), who analyzed age differences between couples in the sixteenth and seventeenth centuries, considered a limited number of cases, making it infeasible to ascertain the overall trend with confidence.Footnote ⁸

As shown in Table 1, an indirect method of exploring age difference patterns within marriages before and after the Imjin War is to observe the disparity in age at first marriage during the mid-Chosŏn period. Studies reconstructing age at first marriage in the sixteenth and seventeenth centuries – drawing on data such as biographies recorded in personal anthologies and letters that mention marriage – reveal that couples from yangban families had an age difference of approximately 0.5 to 2.5 years (Pak Reference Pak2006; Park Reference Park2008). This trend of age-similar marriage persisted into later periods; for instance, the average age at first marriage in the Tansŏng 丹城 region from the late seventeenth to the late eighteenth centuries was 18 years for men and 17.5 years for women (Kim Reference Kim2005, pp. 196–98).Footnote ⁹ Taken together, these findings suggest that first marriages were typically with peers, with a maximum age difference of less than three years. Given that the age at first marriage for women belonging to yangban families was influenced by the onset of menarche and displayed no distinct fluctuations in the longue durée compared to men (Pak Reference Pak2006), it is reasonable to infer that the cohort of women affected by the Imjin War at the time of their first marriage was born in the mid-1570s and later.

Table 1. Age of the first marriage in Chosŏn in the sixteenth and seventeenth centuries (unit: years)

In addition to the normative customs regarding marriageable age, it is important to examine the socio-legal contexts of marriage during the Chosŏn dynasty to identify the factors affecting marital age differences (Kim Reference Kim2004, Reference Kim2006; Pak Reference Pak2020; Park Reference Park2008). While Chosŏn society legally upheld strict monogamy, it held unfavorable views of extramarital affairs, but only toward women. Men were permitted to have secondary or minor wives (ch’ŏp) 妾 and faced relatively lenient consequences for extramarital affairs. Conversely, women encountered harsh judgment even in cases of forced relationships, let alone remarriage.

Although both late imperial China and the Chosŏn dynasty – where the Neo-Confucian social order was strict – emphasized the significance of chaste widowhood, the actual situations in each society differed. In China, women could remarry after a three-year mourning period following the death of their husbands, so remarriage was more frequent than lifelong abstinence (Brook Reference Brook2010, pp. 140–41; Sommer Reference Sommer2000). However, the Chosŏn government maintained a stipulated statute prohibiting widow remarriage from the fifteenth century until the Kabo Reform in 1894 (Deuchler Reference Deuchler1992, pp. 276–80). The rationale behind the enactment of legal prohibitions against widow remarriage can be analyzed from various angles, including demographic imbalances such as a surplus of women, political exigencies, and the consolidation of Neo-Confucian moral values. A recent study (Kim Reference Kim2022) posited that the negative portrayal of remarriage found in Elementary Learning (Xiaoxue) 小學, a Confucian textbook authored by Zhu Xi and esteemed among Chosŏn scholars, may have influenced the proscription against women remarrying. Further, the consequences of breaching the law regarding chaste widowhood extended to the descendants of remarried women. They were barred from participating in the civil service exams, thereby preventing them from attaining high-ranking positions (Ch'oe Reference Ch'oe1987; Deuchler Reference Deuchler2003).Footnote ¹⁰ Under such circumstances, widows tended to eschew remarriage due to this discriminatory regulation.

In Chosŏn society, the loss of chastity was referred to as sirhaeng (失行) or shilchŏl (失節).Footnote ¹¹ The Imjin War triggered a surge in female victims who experienced sirhaeng involuntarily as well as cases where husbands sought divorce because their wives had suffered sirhaeng. For instance, King Sŏnjo (r. 1567–1608) received a request from a man in the royal family to grant a divorce to his wife who had been captured by the Japanese army during the Imjin War and subsequently returned.Footnote ¹² Despite King Sŏnjo's opposition to abandoning returned wives, he was unable to alter the prevailing societal atmosphere. Following the Second Manchu Invasion in the twelfth lunar month of 1636, the Chosŏn government allowed the divorce of such returned wives to placate public opinion, particularly among social leaders (Pak Reference Pak2020). Accordingly, women who experienced sirhaeng during the Imjin War or returned after being captured by the Japanese army faced significant disadvantages in their marriages. This situation was especially precarious for women in the yangban, who were expected to adhere strictly to Neo-Confucian norms regarding female chastity.

It is therefore worthwhile to investigate the extent to which the Imjin War impacted marriage and remarriage patterns in Chosŏn society. Since the marital age difference in Chosŏn society typically fell below three years, the present study examines whether the war and post-war periods showed a different trend by considering key factors influencing mate selection, such as status, wealth, and region. Identifying patterns in marital age difference in this manner aids the analysis of marked changes in spousal bereavement and remarriage resulting from the prolonged war. This study pays special attention to a specific group – women born in the last quarter of the sixteenth century – whose age gap with their husbands may have diverged from that of previous generations due to their first marriages being directly affected by the Imjin War. As will be discussed later in the analysis of the Chosŏn household registers, these women were more likely to marry considerably older husbands, partially because of the war's impact on the sudden increase in the number of widowers.

Advantages and disadvantages of household registers

This study uses extant early seventeenth-century Chosŏn household registers to assess the effect of the Imjin War on changes in nuptial patterns. Since its inception, the Chosŏn government compiled national household registers triennially.Footnote ¹³ The primary objective of crafting household registers was to identify individuals and households subject to taxation and corvée service.

Household registers contained vital details concerning subjects' obligations to the state, which were contingent upon their occupation or social status. While the occupational classifications appeared in household registers may not always align perfectly with individuals' actual professions, they nonetheless represent their official status and provide valuable insights. Household registers from the Chosŏn dynasty meticulously documented various social statuses, ranging from royal family members and high-ranking government officials to socially marginalized subjects like monks, butchers, and nobi at the lowermost tier.

One notable advantage of household registers lies in their ability to unveil ancestral and individual identities. Given the status system in Chosŏn society, ancestral records were indispensable for determining the status of male household heads and their wives. Hence, the names and occupations of the “four fathers” (sajo) 四祖 – the father, grandfather, great-grandfather, and maternal grandfather – were documented. Furthermore, household registers comprehensively recorded various personal particulars, such as a person's surname, ancestral origin, age, place of residence, and relationship to the household head. They also contained migration details; for example, instances of absence owing to escape or registration in a different household from three years prior were noted with a brief explanation (Dong et al. Reference Dong, Campbell, Kurosu and Lee2015a).

However, the Chosŏn household registers have certain drawbacks. First, their triennial compilation hampers the ability to ascertain the precise timing of vital events such as births, deaths, and marriages that took place between two consecutive surveys (Park and Lee Reference Park and Lee2008). Particularly pertinent to the present study is the inability to determine the spousal age difference in first and subsequent marriages as these were not recorded separately. Second, since the primary purpose of Chosŏn household registers was to identify individuals or households liable for taxation, only specific households were documented based on the requisite tax threshold.Footnote ¹⁴

Despite these limitations, household registers remain the sole source for understanding the overall population and family structure of premodern Korea. For elite individuals who could compile their family records into genealogies, their personal lives, and family details can be corroborated via these biographical records. However, genealogical records (chokpo) 族譜 of the Chosŏn dynasty were prone to inaccuracies because of retrospective interpretations rather than real-time recording.Footnote ¹⁵ Moreover, given that the vast majority of the population did not compile genealogies until the nineteenth century, information about these segments is limited. As discussed, research on marriageable age and marital age differences is confined to the yangban who left behind chronological records and marital correspondences (Pak Reference Pak2006; Park Reference Park2008); thus, generalizations about broader societal trends are restricted. In this context, a dataset reconstructed from household register information serves as the most robust source for portraying family formations of all social strata (Park and Kim Reference Park and Kim2010, p. 110).

Five household registers in post-Imjin Korea

All the extant Chosŏn household registers were compiled following the Imjin War.Footnote ¹⁶ Because of the catastrophic war, the Chosŏn government suspended the compilation of household registers until before 1606 when post-war restoration efforts began to show progress. Surviving household registers from the first half of the seventeenth century are scarce, with only six types from five administrative districts known to exist. The present study utilizes all but one of these registers, covering five household registers assembled in one prefecture and three counties.Footnote ¹⁷

This study focuses on the five oldest Chosŏn household registers – the 1606 Sanŭm 山陰, Tansŏng, and Chinhae 鎭海 registers, the 1609 Ulsan 蔚山 register, and the 1630 Sanŭm register. These four administrative districts were located in the southeastern part of the Korean peninsula, specifically in Kyŏngsang Province 慶尙道 (Fig. 1). Kyŏngsang Province faces the Japanese archipelago across the Korea Strait; its southern coastal areas were major battlegrounds and suffered cataclysmic damage during the Imjin War due to their geographic proximity to Japan.

Figure 1. Locations of the four studied administrative districts in Korea.

Note: This map was adapted from resources provided by the Chosun Culture Electronic Atlas, a project of the Research Institute of Korean Studies at Korea University.

First, Ulsan, situated along the seashore, served as a major military base, hosting the headquarters of the provincial military commander of Left Kyŏngsang Province 慶尙左兵營. Its military significance was further emphasized after the Imjin War. In 1599, immediately after the war, it was elevated to a strategic prefecture (Tohobu) 都護府.

This study concentrates on the earliest surviving records of the Ulsan household registers, specifically the 1609 register, which provides a vivid depiction of the post-war landscape.Footnote ¹⁸ Ulsan underwent profound demographic shifts during the Imjin War, emerging as one of the most affected regions. This was attributed to the presence of Japanese commander Katō Kiyomasa, who established a Japanese castle (waesŏng) in Ulsan and held out until the end of the Imjin War, retreating to Japan in late 1598 (Swope Reference Swope2006). Hence, the 1609 Ulsan household register, compiled a mere eleven years after the ceasefire, offers insight into local conditions at that time, marked by the devastation wrought by the Siege of Ulsan in 1598 and the Japanese military stationing. It shows, for instance, state-led immigration initiatives aimed at repopulating the Ulsan area as part of post-war revitalization efforts. In Nongsŏ-ri 農所里, Ulsan, 166 out of a total of 237 households consisted of relocated soldiers (somogun) recruited from other regions. The 1609 Ulsan household register also portrays the migration of various individuals from diverse backgrounds to the area due to the Imjin War. This included individuals abducted by the Japanese army and taken to Japan, some of whom managed to escape and return, as well as those who arrived on the Korean peninsula with the Japanese army but defected during the war (Lee and Han Reference Lee and Han2020).Footnote ¹⁹

Second, Sanŭm and Tansŏng, situated adjacent to each other (see Fig. 1), have extant household registers dating back to the seventeenth century. Unlike Ulsan, which was valued for its military importance and mainly inhabited by non-elite populations, Tansŏng was strongly influenced by the yangban. The yangban in Tansŏng, in contrast to neighboring regions, produced more officials and civil service exam candidates, and exerted strong control over the local community until the late nineteenth century (Hojŏktaejang yŏn'gut’im 2003).

While Sanŭm and Tansŏng did not experience a prolonged Japanese military presence, their proximity to Chinju 晋州, one of the major battlefields, exposed them to several attacks and substantial damage. Tansŏng lost its function as a county due to depopulation resulting from the war, leading to its annexation by neighboring Sanŭm in 1599. Only in 1613 did Tansŏng regain its status as a separate county. Owing to these circumstances, the 1606 Tansŏng household register was compiled as part of the Sanŭm register, confirming the lasting impact of the Imjin War in this region. The 1606 household register of Sanŭm and Tansŏng reveal records of individuals currently absent who were taken captive during the Imjin War (Han Reference Han2020a). Furthermore, the surviving household registers of Tansŏng have garnered scholarly attention. This is because, unlike the Sanŭm registers, the Tansŏng registers were kept in a state of preservation from 1678 until the late nineteenth century.Footnote ²⁰

This study analyzes two surviving Sanŭm household registers from 1606 and 1630. Despite being among the oldest surviving Chosŏn household registers, these Sanŭm household registers have received limited attention in related research, with only a few exceptions (Han Reference Han2020a; Ro Reference Ro1983). This is because, unlike the Ulsan and Tansŏng registers, the Sanŭm registers did not survive beyond the mid-seventeenth century, making it impossible to conduct long-term investigations. Therefore, this study holds significance as it conducts a comprehensive re-examination of two early seventeenth-century Sanŭm household registers.

Third, albeit in small quantities, the records of the 1606 Chinhae household register are available. Chinhae, a small county on the southern coast where major naval battles occurred during the Imjin War, suffered significant damage from the conflict. While Chinhae compiled its household register in 1606, most of the records have disappeared, with only a few remaining in private possession. Hence, no research has been conducted on the 1606 Chinhae household register to date. Despite its limited volume, it holds importance as one of the earliest surviving household registers. This study therefore includes the 1606 Chinhae household register to identify the impact of the Imjin War on the trend of marital age gaps.

Information on social stratification in the five household registers

The occupational records and genealogical information in household registers are important as they represent an individual's state-recognized social status. During the Chosŏn dynasty, individuals inherited the socio-legal status of their parents. As mentioned, the household register established a person's social standing by documenting the names and occupations of four male ancestors across three generations: the father, grandfather, great-grandfather, and maternal grandfather.Footnote ²¹

This study improves the precision of social status information by cross-verifying individuals' occupational records with genealogical data recorded in household registers. When it was difficult to determine social status, the family details of the household head (chuho) 主戶 were consulted to establish the person's social status. For household members other than the head, except for nobi, their occupation records were employed to ascertain their social status. In cases where occupational information was lacking, the same social status as the household head was assumed.

The Chosŏn government did not mandate labor service obligations for women of yangin status; hence, their occupational records are absent in the household registers. Women with occupational records belonged to the ch’ŏnmin group, except for yangban women, whose honorific titles corresponded to their bureaucrat husbands. In this study, the primary criterion for differentiating women's social strata consisted of official appellations indicating their positions, adopting established research methods such as those developed by Yi (Reference Yi1997, pp. 233–41). Both yangban and commoner women typically identified themselves in official documents using their surnames and appellations, reflecting their social status, rather than their given names. In household registers, yangban women designated themselves as “lady” (ssi) 氏, while commoner women used either “wife” (sŏng) 姓 or “woman” (choi) 召史 (Han Reference Han2021, note 76, 77; Lee and Han Reference Lee and Han2020, note 26; see also Wagner Reference Wagner, Kendall and Peterson1983). Given that women listed in the household registers as either the household head or the wife of the household head had their ancestors' information recorded, they could not arbitrarily use appellations indicating a social position higher than their own.

In addition to social status, household assets wielded significant influence in the spousal selection process. Since family estates in Chosŏn society primarily comprised land and nobi (Kim Reference Kim and Campbell2004, p. 153), the information on nobi in household registers serves as a valuable gauge of a household's prosperity. Despite many nobi fleeing from their owners in pursuit of freedom amid the chaos of the Imjin War, owners persisted in registering these runaways in household registers (this is particularly evident in the 1606 register of Sanŭm and Tansŏng) to maintain ownership claims (Han Reference Han2020a, p. 181). In light of such complexities, the present study considers whether the household head had ever owned nobi as an indication of property wealth, rather than simply tallying the number of nobi in the household at the time of registration.

Social statuses and spatial compositions

The total sample size included in this study was 11,749 individuals, with the breakdown by household register and social status presented in Table 2. The yangban accounted for only 7.54 percent, while the lowborn constituted over half (54.77 percent) of the sample. This suggests that no large-scale emancipation of nobi took place, despite the turmoil caused by the Imjin War – such as mass refugees and individuals abducted by the Japanese – and the practice of expunging them from records of lowborn (myŏnch’ŏn) 免賤 by presenting a Japanese head.Footnote ²²

Table 2. Composition of social status by household register

The distribution of social statuses varied by region. In Tansŏng, the yangban and lowborn were prominent, while commoners accounted for only around 10 percent. This implies the dominance of influential local families who owned nobi and the overwhelming yangban culture in this area, confirming the abovementioned discussion. In contrast, Sanŭm had a relatively smaller population of yangban and lowborn individuals but a significant proportion of commoners at around 40 percent. Ulsan had a minimal share of yangban individuals at 3 percent, with half of the population being commoners (50.84 percent). In sum, Tansŏng and Chinhae exhibited greater social inequality while Ulsan demonstrated less inequality, with Sanŭm in the middle. These results indicate the possibility of regional variations in marriage-related norms and practices, which will be empirically validated by analyzing spousal age differences.

To accurately calculate marital age gaps, this study minimizes errors by excluding individuals who were deceased or secondary wives at the time of household registration, in addition to instances where information on social status was missing. Since the household registers were updated every three years, events occurring within that period were recorded in the subsequent census. For example, if someone died during the three-year interval, the household register would indicate “deceased” (ko) 故 before the name. Based on this notation, the present study excludes spouses who were deceased at the time of household registration. Furthermore, while it was common in Chosŏn society for yangban men to have secondary wives, these relationships were not considered legitimate and thus were not suitable for the analysis of marital age differences.

After applying exclusions, the analytic sample consisted of 2,336 married couples. Table 3 displays the results of classifying them according to the household register dataset and the husband's social status.Footnote ²³ The proportions of yangban husbands follow a similar pattern to Table 2, with Tansŏng exhibiting the highest ratio. A distinctive feature of Ulsan is a significant number of lowborn husbands, along with the lowest share of yangban husbands. This phenomenon correlates with the demographic makeup of the region, characterized by a higher influx of migrants compared to other areas. Most of the lowborn husbands in Ulsan were immigrants who were privately-owned nobi (sano) 私奴 fleeing from their residences to become household heads. As examples of inter-status marriages, some of them were married to commoner women.Footnote ²⁴

Table 3. Composition of husband's social status by household register

The pattern of age gaps between spouses

The age difference between spouses varied markedly across regions (Table 4).Footnote ²⁵ Sanŭm and Tansŏng show distinct patterns despite their proximity. Conversely, Ulsan exhibits the lowest average age difference at 5.02 years. As explained earlier, Ulsan experienced a colossal outflow of its population to Japan due to the Japanese military presence and subsequent withdrawal during the Imjin War, leading to a significant influx afterward. Such demographic changes likely affected the chances of finding spouses of similar ages in Ulsan. Moreover, the proportion of yangban, constrained by norms prohibiting widows from remarrying, was the smallest in Ulsan compared to other regions (Table 3). This characteristic was conducive to the narrower marital age difference in this region.

Table 4. Basic statistics: Spousal age difference by household register and husband's social status (based on wife's age)

Notes: SD (Standard Deviation), Q2 (Median), IQR (interquartile range: Q3 – Q1).

Furthermore, the spousal age difference differed remarkably according to the husband's social status. On average, yangban husbands were over 10 years older than their wives, whereas lowborn husbands were only 3.51 years older, with commoner husbands falling in between at 7.08 years older. The reason for the age difference by social status can be interpreted through the prism of marriage customs and laws discussed earlier. Especially in the context of chaste widowhood, remarriages of yangban widowers typically involved younger first-time married women. Such remarriages became more pronounced as the men rose in social status, leading to an increase in the marital age gap. In contrast, for lowborn men, for whom the Neo-Confucian norm of chaste widowhood was not strictly implemented, the median age difference with their wives was only two years, with the bottom 25 percent (Q1) value at −1, indicating that over a quarter of the cases featured husbands younger than their wives.

The overall trend of age differences within marriages evident in the early seventeenth-century Chosŏn household registers yielded a mean of 6.61 years and a median of 5 years. The top 25 percent (Q3) value of eleven years suggests that in one-quarter of the couples in this analytic sample, the age difference was eleven years or more. Such a large age difference between spouses may be closely linked to the unique environment of a long-term war (to be discussed later).

Figure 2 visualizes the detailed marital age differences by the husband's social status. Among the yangban, only 7.62 percent of husbands were younger than their wives. Meanwhile, marital age gaps of 11–15, 16–20, and 21 years or more constituted 16.23 percent, 11.59 percent, and 12.91 percent, respectively. Hence, for couples where the husband was eleven years or more older than the wife, the rate exceeded 40 percent. These findings imply that couples featuring a remarried husband and first-married wife, in which the husband was notably older, were prevalent among the yangban.

Figure 2. Distribution of spousal age difference by husband's social status.

In contrast, about half of commoner husbands were 1–10 years older than their wives, with 15.68 percent of wives being older. For couples where the husband belonged to the lowborn, the share of older wives reached 29.10 percent, and when couples of the same age were added, it approached nearly 40 percent. However, less than 20 percent of husbands were more than eleven years older than their wives.

In addition, nuptial patterns in premodern societies, whether in the East or the West, were closely tied to family wealth (Lundh and Kurosu Reference Lundh and Kurosu2014; Wolf and Huang Reference Wolf and Huang1980). As discussed above, this study considers nobi ownership to assess the economic status of the husband's family. In the household register records, 501 married couples were identified whose household ever owned at least one nobi (Table 5 and Fig. 3).Footnote ²⁶ A comparison of Tables 4 and 5 indicates that the age difference for commoner husbands who had ever owned a nobi was greater than for those who had not. Consequently, the remarkable age difference observed between yangban and commoner couples in Table 4 is no longer relevant.Footnote ²⁷

Table 5. Basic statistics: Spousal age difference by husband's social status in a sample of households that ever owned nobi

Figure 3. Distribution of spousal age difference by husband's social status: A sample of households that ever owned nobi.

According to Figure 3, the percentage of commoner husbands whose wives were older (9.43 percent) is substantially lower than those in Figure 2 (15.68 percent). In contrast, the proportion of commoner husbands whose households had ever owned a nobi and who were at least 16 years older than their wives (20.76 percent) is 5.15 percentage points higher than for all commoner husbands (15.61 percent). These results demonstrate that the mechanism for selecting a spouse's age was largely influenced by family property within the commoner status.

The Imjin War and the widening of marital age gaps

This analysis involved 2,336 couples, predominantly born in the late sixteenth century, who wedded their current spouses around the time of the Imjin War. It is inferred that the notable prevalence of age-discrepant marriages was closely correlated with shifts in the local marriage market, prompted by demographic changes in mortality and migration during the war. This study therefore investigates the existence of a temporal trend in age disparity, with a focus on a potential period effect of the Imjin War.

To accurately measure the period effect, having information about the timing of marriage to the current spouse is crucial; however, the datasets lack such details. As an alternative, birth cohort information was utilized. The decision of whether to set the birth cohort of wives or husbands as the reference point was pivotal for the analysis. As such, it was necessary to examine the trends in the marriage ages of men and women during the Chosŏn dynasty. As previously noted, men's age at marriage varied owing to social status and mortality, while women's age at marriage remained relatively stable (Pak Reference Pak2006). Accordingly, the wife's birth cohort was chosen as the reference category, aligning with the research of Wolf and Huang (Reference Wolf and Huang1980, pp. 138–39) on marital age differences in China.

Table 6 presents the age differences between wives and husbands, delineated by the wife's birth cohort, as observed in the datasets of the five household registers. Except for the 1630 Sanŭm household register, the remaining four registers exhibit a notable trend toward marriages with significant age disparities over time. During the recorded period in either 1606 or 1609, those born in the 1560s (581 individuals) and the 1570s (507 individuals) constituted over half of the data. The former cohort displayed an average age difference of about five years, while the latter exhibited a substantial increase to 8.22 years. These findings are attributed to the influence of the Imjin War on women born in the 1570s as they were of marriageable age during the war. The prolonged conflict affected societal norms and practices regarding the age of the groom, leading families with marriageable-age daughters to accept older husbands.

Table 6. Age difference with husband by wife's birth cohort

Notes: The couple's birth cohort was constructed based on age and animal birth year information recorded in household registers.

The 1630 Sanŭm household register reveals a trend spanning three decades following the end of the Imjin War in 1598. Age disparities with husbands continued to rise for women born up to the 1580s, aligning with previous findings from other registers. Beginning with the 1590s birth cohort, the age gap diminished somewhat. Nonetheless, wives born in the 1600s still had an average age difference of around nine years, which is meaningfully higher than for couples married before the war, especially wives born prior to 1570.

These findings underscore the significant impact of the Imjin War on nuptial trends and social customs. Considering that the marriages of women born in the 1570s and 1580s closely overlapped with the Imjin War, it notably influenced norms surrounding marriage ages. This impact persisted in subsequent generations born after the war.

This study further examines whether the observed period effect remained consistent across all social statuses or exhibited variations. Figure 4 illustrates the variation in age differences between spouses while controlling for the husband's social status. The results indicate that women who faced the hardships of the Imjin War during marriageable age had the largest age differences with their husbands. This trend was pronounced for women who were children during the war and reached marriageable age afterward (born around 1580 or later). On average, they had an age difference of nearly ten years with their husbands. Their remarkably high age difference can be attributed to changes in the local marriage market resulting from the war, where an increase in young men's mortality and a rise in widowers led to a surge in remarriages between widowed men and first-married women. Furthermore, the post-Imjin rise in age gaps varied depending on the husband's social status, with the lowborn showing a smaller increase vis-à-vis other groups because they had more flexibility regarding strict Neo-Confucian marriage norms, including chaste widowhood.

Figure 4. The temporal trends in spousal age difference by wife's birth cohort and husband's social status.

Notes: Lowess (local weighted regression analysis), one of the smoothing techniques, has been applied to the trend line.

Marriage, remarriage, and widows

When investigating spousal age disparities in post-Imjin Korea, it is vital to acknowledge an inherent limitation of the household registers: the challenge of distinguishing whether a recorded marriage represented a first-time marriage or a remarriage. These registers only contained personal details about the current spouse, with occasional exceptions (as mentioned earlier) where information about the deceased spouse was included.

To tackle this issue, some prior studies, like those by Kim (Reference Kim2004, Reference Kim2006), categorized couples under the age of thirty as first-time marriages and restricted the analysis to this age bracket.Footnote ²⁸ However, this approach fails to resolve the fundamental limitations of the Chosŏn household register data. In traditional Korean society, the average age for men's first marriage was around eighteen years (Pak Reference Pak2006; Park Reference Park2008), indicating that instances of remarriage in their twenties due to spousal death cannot be disregarded. This may be particularly relevant during the Imjin War when occurrences of deaths, disappearances, or abductions of young men and women were more frequent compared to other periods. Hence, considering individuals under thirty as first-married when calculating the age gap between spouses may cause misleading conclusions.

The current study addresses these limitations by examining genealogical records from yangban families to verify cases where a significant marital age difference was recorded in the household register, indicating male remarriage. One such example is found in the 1606 Sanŭm household register, where Min Hyochong (閔孝宗, born in 1534) married Lady Yun (尹氏, born in 1579) of the P'ap’yŏng 坡平 Yun descent group; Lady Yun was a staggering forty-five years younger. According to the Genealogy of the Yŏhŭng 驪興 Min descent group compiled in 1923, Lady Yun was the second wife (huch’ŏ) 後妻 of Min Hyochong, who had previously been married to Lady Yi 李氏 of the Kangnŭng 江陵 Yi descent group. Another example in the 1606 Sanŭm household register is the case of O Yi (吳儞, born in 1534), who was recorded as being married to Lady Yi (李氏, born in 1569) of the Kŭmgu 金溝 Yi descent group; she was thirty-five years younger. In the Genealogy of the Hamyang 咸陽 O descent group compiled in 1909, however, O Yi's wife was listed as Lady Min 閔氏, instead of Lady Yi. This difference is likely because O Yi married Lady Yi after the death of his first wife, Lady Min, who was not recorded in the 1606 Sanŭm register.

These intriguing cases imply that the majority of couples in which the husband was twenty years older than the wife comprised those of remarried men and first-married women.Footnote ²⁹ The age difference at the time of remarriage was particularly pronounced among the yangban. The effective imposition of the socio-legal norm of chaste widowhood can explain this societal trend. Since the yangban placed a high value on prohibiting women from remarrying, widowers belonging to the yangban had to seek out virgins instead of widows. In contrast, this tendency was less conspicuous among commoners and lowborn individuals, where women's remarriage was not a significant social or normative concern.

Finally, this study tests the validity of its argument by examining cases where widows maintained chaste widowhood throughout their lives. Table 7 presents the number of female-headed households recorded in the household registers. Most of these household heads were widows, except for a few lowborn women (government- or school-owned nobi) who were unmarried and lived separately from their owners. Some of these households headed by widows had adult sons capable of assuming the role of household head after their fathers passed away (Kim Reference Kim2018). Examining the ratio of female household heads to the general female population across different social statuses indicates that a significant number of yangban women did not remarry and remained widows. This was particularly pronounced in Tansŏng where the yangban culture was strong. More than one-fifth of yangban women endured hard times following the Imjin War as widows.Footnote ³⁰ Hence, it is possible to deduce that the social culture of Chosŏn – where women were strongly discouraged from remarrying, combined with men's desire to remarry and the demographic aftermath of the Imjin War – served as the main driver of the unusual marital age differences.

Table 7. Share of female household heads of by household register and woman's social status