If I gave you the word "안녕하세요" in written text and asked you to tell me how to say that, could you? I'm making a wide assumption here that those of you who are reading this do not speak Korean. What I just showed you is one of the many ways to say "hello" written in 한글, Korea's writing system.

I'm making a very deliberate choice not to romanize those words for you. Romanization is the act of converting text from one writing system to the Roman (latin) script. The act of romanization is messy, especially so for Korean. The word "한글" can be romanized in a few ways depending on which system you opt to use. You'll frequently see it as either "Hangul", "Hangeul", or even "Han'gŭl". To make things more difficult, the Korean language includes a number of morphophonemic constraints that cause one Hangeul syllable to be pronounced differently according to the phonological environment[1]. That is to say, many words are written differently to how they are pronounced given certain contexts.

One of the first attempts at romanization dates back to 1832 when Philipp Franz J B. van Siebold, a German doctor serving as an advisor to the Japanese government at the time, developed a system[2]. What followed was not widespread adoption but rather several other attempts at devising a romanization system. Only three of which made strong impacts; the Siebold system (1832), the Dallet system (1874) and the Ross system (1877). While these systems are no longer used in modern day Korea, both North and South, influences from them can be found in the systems used today.

To appreciate why romanization is a tricky feat, let's take some time to understand how Korea's writing system works. Hangeul was invented in 1443 by King Sejong, the great (세종대왕), the fourth king of the Joseon dynasty. Before its invention, Koreans used Chinese characters. In his book, "The Correct Sounds for the Instruction of the People"[3] (훈민정음) published in 1446, he described the principles behind his new alphabet:

The sounds of our country's language are different from those of the Middle Kingdom and are not confluent with the sounds of characters. Therefore, among the ignorant people, there have been too many who, having something they want to put into words, have in the end been unable to express their feelings. I have been distressed because of this, and have newly designed twenty-eight letters, which I wish to have everyone practice at their ease and make convenient for their daily use[4].

Hangeul is a phonographic script in which 24 (originally 28) graphs, also called Jamo (자모), depict the points of articulation of all the phonemes in the Korean language[5]. These are then arranged into syllable blocks containing a lead consonant (초성, choseong), a medial vowel (중성, jungseong), and an optional tail consonant (종성 / 받침, jongseong / batchim). For example, "한글" can be broken down into two syllables, and then further to two sets of three Jamo.

consonant + vowel + consonant = syllable -> word
ㅎ + ㅏ + ㄴ = 한 + ㄱ + ㅡ + ㄹ = 글 -> 한글

The optional tail may be constructed from two consonants and, while they are considered to be two separate letters, they are phonologically considered to be one[6].

consonant + vowel + consonant + consonant = syllable -> word
ㄷ + ㅏ + ㄺ = 닭 -> 닭

Syllables starting with an initial vowel, for example 안 (an), must be prefixed with ㅇ (eung), a special lead consonant that produces no sound when used at the start. This is not the case when placed as the tail consonant however, instead making the /ŋ/ (ng) sound.

As mentioned previously, there were three strong contendors in the early days of romanization - the Siebold system (1832), the Dallet system (1874) and the Ross system (1877). Most romanization systems can be traced back to these three systems as done by Do Hyong-su using the three voiced unaspirated consonants ㄱ, ㄷ, and ㅂ and the two vowels ㅓ and ㅡ[7].

The Siebold system used a "k" for ㄱ, a "t" for ㄷ, and a "p" for ㅂ. For the aspirated versions, an apostrophe was added - (ㅋ, k'), (ㅌ, t'), and (ㅍ, p'). The use of apostrophes to indicate aspiration can be seen as a direct influence to the latter produced McCune-Reischauer system.

The Dallet system used the same consonants as mentioned above but instead indicated aspiration via an "h" instead of an apostrophe - (ㅋ, kh), (ㅌ, th), and (ㅍ, ph). The use of the "h" can be seen in the official North Korean romanization system[8]. Another notable thing the Dallet system did was use "e" in addition to another vowel to indicate the two vowels ㅓ and ㅡ which became "eo" and "eu" respectively. This is used to this day in South Korea's official romanization system - The Revised Romanization of Korean. The Dallet system was also the first system to use diacritics via the two French accents, grave (è) and acute (é) over the "e" to indicate different vowels.

Finally, the Ross system used voiced English consonants "g", "d", and "b" for the voiceless unaspirated consonants ㄱ, ㄷ, and ㅂ. For the aspirated consonants ㅋ, ㅌ, and ㅍ, the Ross system uses the voiceless English consonants "k", "t", and "p", thereby avoiding the need for additional information for aspiration.

Following the colonization of Korea by the Empire of Japan in 1910, swift actions were taken by the Japanese in an attempt to abolish and erase Korea's history[9]. This was done under the pretext of the racial theory known as Nissen dōsoron - a theory such that Japanese and Korean people had once shared a common ancestory where the Japanese people had descended from the Japanese deity Amaterasu and the Korean people from Susanoo, her younger brother. This lead to the eventual ban of Korean names under the Sōshi-kaimei policy which forced Koreans under Japanese rule to adopt Japanese names and identity as such[10]. What followed in 1938 was a complete ban on the use of Korean which retained until the defeat of the Japanese in 1945. During this period, the Korean Language Society (한글 학회) - a society with the goal of preserving Korean language during the Japanese occupation[11] - maintained strong activity despite many of their members being put into prison. One of the outputs from this group is considered to be the first romanization system developed by Koreans. Although it was never widely used, the work done by the Korean Language Society was appended to the Unified Hangeul Orthography system[12].

Systems continued to be developed and by 1934 Japanese linguist Shinpei Ogura (小倉進平) counted at least twenty-seven different systems[13]. Due to the Korean languages complex phonetic structure, romanization is extremely difficult. One major difficulty faced is the challenge of mapping Korea's ten vowels to the five present in the Latin script. Sacrifices needed to be made such as vowels being rendered with multiple letters or the usage of diacritics as seen in the previously mentioned systems. Another difficulty presented itself in the form of linguist assimilation - the change of sound in certain phonemes based on the context of nearby sounds[14]. The Korean dynasty 신라 (Silla) for example is written as "Sin-la" but pronounced "Sil-la"[15].

The liberation from the Japanese was short-lived as soon after came the Korean War bringing an end to a unified Korea and dividing it into North and South. Although romanization was the least of concerns for these two states, efforts had to be made due to each of their close ties with one of the cold-war superpowers. North Korea was the first to invent a new romanization system in 1956[16]. It took influences from both the Dallet and Unified Hangeul Orthography systems, notably adding "h" to signify aspiration.

It has since been updated a handful of times with the latest change happening in 2012[17]. Notable changes include differentiation between initial and final consonant sounds and a handful of vowel reworkings. The system was also given an official name - "Rules of the Latin Alphabetical Transcriptions of the Korean Language"[17].

Meanwhile, South Korea was still attempting to come to an agreement on how romanization should be done. In 1959, the Ministry of Education developed a system to represent each Hangeul graph with one Roman letter. This was done in an attempt to make romanization for Koreans easier. The devised system would allow those who "think in Hangeul" to produce accurate romanized Korean without the need to think about the many phonological changes that cause deviation from the orthography[18]. This system became the center of controversy due to its use of voiced consonants. While this makes sense for English speakers as their voiced consonants are much closer to the unaspirated voiced consonants in Korean, it unfortunately led to many romanized words resembling combinations of english words. 독립문, for example, would be romanized as "Dogribmun" leading to many jokes about "Dog ribs"[19]. Further romanizations would follow the same unfortunate pattern. Despite the controversy, the use of "eo" for ㅓ and "eu" for ㅡ is still popular to this day as the use of a breve is inconvenient for most people.

Displeased with the systems available, George M. McCune and Edwin O. Reischauer worked together with help from Korean linguists Choe Hyeon-bae (최현배), Jeong In-seop (정인섭), Kim Seon-gi (김선기), and others to devise a new system which they aptly named the McCune–Reischauer system. As stated in their 1993 published work "The Romanization of the Korean Language Based on Its Phonetic Structure", the main goal of their romanization system was not meant for use in phonetic or technical philological research but rather for "general scholarly and non-scholarly use where phonetic symbols would be cumbersome and annoying and where strict phonetic exactness is not demanded"[18]. Their method differed greatly from what existed at the time. Instead of employing a letter-by-letter transcription they instead opted for a system that is able to represent the phonetic pronunciation of words. Although attempts were made to keep the romanization system simple, diacritical marks were still required. Due to differences in pronunciation caused by varying Korean dialects, the McCune-Reischauer system defined the standard Korean pronunciation as that spoken by the "educated middle class of the capital, Sŏul (Keijō)"[20] and designed its romanization around such pronunciation.

What follows is a simplified table showcasing the McCune-Reischauer system. There are cases and irregularities not covered by the table below. The full system can be viewed in the following wikipedia entry on the McCune–Reischauer system or sections "II. THE VOWELS" and "III. THE CONSONANTS" in "The Romanization of the Korean Language Based on Its Phonetic Structure".

Due to the rising criticms, the Ministry of Education decided to reexamine the issues presented by their romanization system. In 1983, the Korean National Academy of Sciences invited three phonologists to help aid an investigation into a way forward. Of the three invited, only Lee Sang-oak (이상억) was in favour of changing the system to the McCune-Reischauer system. The system put into place in 1984 following these discussions was, for the most part, the same as the McCune-Reischauer system. There were minor changes to the original McCune-Reischauer system, notably 시 romanized as "shi" instead of "si", ㄹㅎ romanized as "lh" instead of "rh", and divisions in words indicated by a hyphen instead of an apostrophe. A full breakdown of the changes can be viewed in the following wikipedia entry on the McCune–Reischauer system. This was released not without pushback, notably from Koreans who found the system confusing as many were not used to representing phonological changes with different graphs[21]. Although the system gave a rough approximation to the sounds of the Korean language, it came at the cost of convenience.

Although the McCune-Reischauer and Ministry of Educations systems were the most commonly used, it did not stop others attempting further iterations on existing systems. In 1968, Professor Samuel Martin of Yale University developed the Yale system. It was widely adopted by linguists as it allowed readers who know Hangeul to come up with accurate romanized representations. Beyond its use in scholarship, it never saw much reach. This could be traced to its unsightly presentation (독립문: toklipmwun) and failure to provide approximate pronunciations (전라: cenla, pronounced closer to "jeolla").

The spread of computer and internet usage in the 1990s, lead to growing complaints with the McCune-Reischauer system. Although fonts for breves and other diacritics were available, they were not widespread. This lead to many internet users dropping the breve when romanizing text or replacing it with the circumflex. The use of diacritics is especially confusing to Koreans whose main exposure to a Roman-alphabet language is English. Participants in a 1980 workshop on romanization conducted by the University of Hawaii argued to replace the usage of a breve with a more common diacritic mark such as the accent mark from Spanish. The group also suggested replacing the apostrophe with "h" to indicate aspiration[22]. Another criticm from Koreans is that the system requires too much overhead when reading and writing. Since it requires native-speakers to think about phonological changes, it often times leads to them having to sound out the words as they write. Many also argue it doesn't sit right to them that the system was developed by foreigners during Japanese occupation. Linguistic nationalists feel the system violates the genius of Hangeul by representing allophones and phonological changes with different Roman letters[23].

In 1986, the International Organization for Standardization requested that both North and South Korea work together to devise a unified North-South romanization system. Discussions continued through to 1995 with a clear consensus failing to be reached. In 1991, The South Korean National Academy for the Korean Language (국립국어연구원) announced a new system it had been working on. Around the same time, Kim Bokmoon (김복문) developed his own system and launched an accompanying media campaign in an attempt to get his system adopted. In response to the burst of new systems, many jumped to the current systems (The Ministry of Education System, 1984) defense.

It wasn't until 1997 that the South Korean government decided to consider changing the system driven by the decision to repaint road signs with the inclusion of romanization. This was not without its controversy and heated debates sparked online. By the end of June 1998, the Ministry of Culture and Sports announced that all further debates on the changing of the official romanization system be stopped due to a lack of funds and nationwide inability to come to a consensus.

It wouldn't be until July 2000 that the Ministry of Culture and Sports would announce that South Korea would be adopting a new system titled the Revised Romanization of Korea (국어의 로마자 표기법). It had been being developed behind the scenes since 1995 by the National Academy of the Korean Language. The new system was designed with the explicit goal of only using the 26 letters of the Roman-alphabet[24]. This meant ridding of all breves and diacritics; one of the main criticms of the McCune-Reischauer system. Along with the announcement of the new system, it was mandated that all Korean textbooks, maps and signs to do with cultural heritage were to comply with the new system by 28 February 2002. This was at an estimated cost of roughly 500 billion to 600 billion won[25].

The Revised Romanization of Korean system also specified a process to follow when romanizing words[24]. The first step in the process is to write the word to follow the standard pronunciation spellings. Each syllable is then broken down into individual graphs/jamo. Referring to the romanization table above, each graph is then transcribed depending on whether it is a vowel or a consonant in the initial or final position. Finally, spaces are inserted between family and given names and proper nouns are capitalized. Hyphens are also optionally added to either disambiguating pronunciation or between syllables of a given name. The National Institute of Korean Language has stated however that the use of optional hyphens should be discouraged[26].

안녕하세요 → ㅇㅏㄴㄴㅕㅇㅎㅏㅅㅔㅇㅛ → annyeonghaseyo
종로구 → 종노구 → ㅈㅗㅇㄴㅗㄱㅜ → jongnogu → Jongno-gu

The Revised Romanization of Korean romanization process

When you see "Hangul", "Hangeul", or "Han'gŭl", you're seeing historical layers, political choices, and cultural values. Phonetic accuracy for typing convenience, native speakers for foreign learners, technical precision for practical accessibility. Every system has to make sacrifices. The "best" system is the one that serves the people who need to use it. This is precisely the principle King Sejong declared in his original works.

Footnotes

[1] [2] [12] [16] [18] [21] [23] [24] Fouser, Robert J. (1999). “Nationalism and Globalism in Transliteration Systems: Romanization Debates in Korea” (PDF). Language Research 어학연구. 35 (1/4). Seoul National University Language Education Institute 서울대학교 언어교육원.
[3] Ahn, Pyong-hi (2018) [2007]. Studies in Hunminjeongeum. Translated by Jung, Ha-yun (English ed.). National Hangeul Museum. ISBN 979-11-89438-00-5.
[4] Ledyard, Gari K. (1966) The Korean language reform of 1446: The origin, background, and early history of the Korean alphabet, Unpublished Ph. D. dissertation, University of California at Berkeley.
[5] Pae, Hye K.; Winskel, Heather; Kim, Say Young (2025a). “An Introduction to the Korean Language and the Special Status of the Korean Alphabet”.
[6] [7] [14] [15] [20] McCune, George M.; Reischauer, Edwin O. (1939). “The Romanization of the Korean Language Based on Its Phonetic Structure”. Transactions of the Royal Asiatic Society Korea Branch. 29. Royal Asiatic Society Korea Branch: 1–55.
[8] “Working Paper No. 46” (PDF). UNGEGN.
[9] Oguma, Eiji; 小熊英二 (1995). Tan’itsu minzoku shinwa no kigen: “Nihonjin” no jigazō no keifu = The myth of the homogeneous nation (Shohan ed.). Tōkyō. ISBN 4-7885-0528-2. OCLC 34028233.
[10] “Korea under Japanese rule”. Library of Congress Country Studies. Washington, D.C.: United States Library of Congress. 1982.
[11] Lee, Peter H.; Bary, William Theodore De (1997). Sources of Korean Tradition: From the sixteenth to the twentieth centuries. Columbia University Press. p. 321. ISBN 9780231120302.
[13] Ogura Shimpei 小倉進平: Ommun no Romaji Hyokiho 諺文の ローマ 字表記法 (A System of Romanizing Ŏnrnun) in Oda Sensei Shoju Kinen Chosen Ronshu 小田先生頌籌記念朝鮮論集 Keijo, 1934.
[17] “E/CONF.101/CRP15 - The Rules of Latin Alphabetic Transcription of Korean Language” (PDF). UNGEGN.
[19] 강만수 (1 September 2000). 신판 斥和碑 새 로마자표기법이 불러올 국가적 위기. Monthly Chosun (in Korean). “2005년까지 연차적으로 도로표지판을 바꾸는 데 5000억~6000억원이 들고”.
[22] Austerlitz, Robert, et al. (1980) ‘Report on workshop conference on Korean romanization,’ Korean studies.
[25] “Romanization of Korean”. National Institute of Korean Language.
[26] “국어의 로마자 표기법 해설 – 제1장 표기의 기본 원칙”. hangeul.or.kr. National Institute of Korean Language. 8 October 2001.