Korean Family Name Distribution in the Past

Hoang Anh Tuan Kiet, Seung Ki Baek, Hawoong Jeong, Beom Jun Kim

Abstract

We empirically study the genealogical trees of ten families for about five centuries in Korea. Although each family tree contains only the paternal part, the family names of women married to the family have been recorded, which allows us to estimate roughly the family name distributions for the past five hundred years. Revealed is the fact that the unique Korean family name distribution, characterized by a logarithmic form of the cumulative distribution and an exponentially decaying rank-size plot often called the Zipf plot, has remained unchanged for a long time. We discuss the implications of our results within a recently suggested theoretical model and compare them with observations in other countries in which power-law forms are abundantly found.

Introduction

In the study of sociophysics and statistical demography, certain patterns are often considered universal. One of the most prominent is Zipf’s Law, which typically describes a power-law relationship in family name distributions: a few names are extremely common, while a massive "long tail" of rare names exists. This pattern is observed in diverse nations, from the United States to Japan.

However, Korea presents a striking anomaly. With fewer than 300 family names for a population of millions, the distribution does not follow a power law. Instead, it exhibits a unique exponential decay. This study investigates whether this "Korean exception" is a product of modern urbanization or a deeply ingrained historical characteristic by analyzing five centuries of genealogical records.

Methodology: A Proxy for the Past

To reconstruct historical name distributions, researchers analyzed ten digitized jokbo (genealogical books), representing ten different paternal lineages. While these records primarily track male descendants, they meticulously document the original family names and birth years of in-married women. Because Korean women traditionally retain their family names after marriage, these records serve as a randomized "snapshot" of the broader population's name distribution at different points in history.

The study analyzed 221,609 women, with records dating back as far as 1254, though the bulk of the data covers 1510 to 1990. Statistical consistency was verified using the Kendall tau rank correlation, which showed a high degree of stability ($\bar{\tau} \approx 0.73$) in name rankings over centuries, indicating that the relative popularity of names has remained remarkably constant.

The Persistence of Exponential Decay

The primary finding is the remarkable stability of the Korean name distribution. Regardless of the specific family book or the historical era, the rank-size plots consistently show an exponential decay rather than the power-law behavior seen in other countries.

Rank-size plots of family names across ten datasets (a) and four historical time periods (b), showing consistent exponential decay.

As shown in the figure above, the functional form of the distribution remains nearly identical across five centuries. This suggests that the social mechanisms governing the Korean naming system have been static for a very long time, predating modern industrialization.

Logarithmic Growth and Innovation Rates

A key mathematical distinction between power-law and exponential distributions lies in how the number of unique names ($N_f$) scales with the total population ($N$). In power-law systems, $N_f$ typically grows algebraically ($N_f \sim N^\gamma$). In Korea, however, the number of names grows logarithmically ($N_f \sim \ln N$).

The number of distinct family names ($N_f$) grows logarithmically with total population size ($N$) across different datasets (a) and time periods (b).

This logarithmic growth is a direct consequence of an extremely low "innovation rate." In many cultures, new surnames are frequently created based on occupations, locations, or nicknames. In Korea, the set of family names has remained largely closed for nearly two millennia. Once the system was established, it became a closed loop where the introduction of new names was negligible, reinforcing the dominance of a few major clans even as the total population grew.

Discussion: Regional Origins and Social Class

The researchers also explored the role of bon-gwan (regional origins). When names are subdivided by their ancestral seat (e.g., "Kim of Gimhae" vs. "Kim of Gyeongju"), the distribution broadens and begins to resemble the power-law patterns found elsewhere. This implies that the "uniqueness" of the Korean distribution is partly a result of how names are categorized at the highest level.

While the data primarily reflects the yangban (aristocratic) class who maintained these records, the consistency of the pattern as naming practices spread to the general population suggests that the aristocratic distribution eventually became the national standard.

Conclusion

By leveraging historical genealogical records, this study confirms that the unique statistical profile of Korean family names is a stable cultural feature spanning at least 500 years. The exponential decay and logarithmic growth of names point to a social system with a near-zero rate of name innovation, setting Korea apart from the power-law distributions commonly observed in global demography.

Summary updated

Abstract

We empirically study the genealogical trees of ten families for about five centuries in Korea. Although each family tree contains only the paternal part, the family names of women married to the family have been recorded, which allows us to estimate roughly the family name distributions for the past five hundred years. Revealed is the fact that the unique Korean family name distribution, characterized by a logarithmic form of the cumulative distribution and an exponentially decaying rank-size plot often called the Zipf plot, has remained unchanged for a long time. We discuss the implications of our results within a recently suggested theoretical model and compare them with observations in other countries in which power-law forms are abundantly found.
BibTeX
@article{kiet2007korean,
  author = {Hoang Anh Tuan Kiet and Seung Ki Baek and Hawoong Jeong and Beom Jun Kim},
  title = {{Korean Family Name Distribution in the Past}},
  journal = {Journal of the Korean Physical Society},
  year = {2007},
  doi = {10.3938/jkps.51.1812},
}
Figure