Contents
Introduction
Anyone writing about words, phrases, and especially names that come from languages that use different writing systems faces the question of how to represent these in the language being written. Sometimes, one may elect to leave these in their source orthography even if the writing system is different—texts in Japanese or Chinese might be interspersed with foreign terms and names given in the Roman alphabet. But frequently it is preferable to represent such terms in the same writing system as the rest of the text. In such cases, one needs to convert text between different writing systems.
For those writing in English, French, German, or any of the countless languages that use the Roman alphabet, such a solution is called romanization. Because of the ubiquity of the Roman alphabet and the fact that it is used for many of the most spoken and most influential languages in the world (starting with English, the closest thing there is to a global language), romanization is the most common form of interscriptal transcription, and romanization schemes of some sort have been proposed for virtually all languages of note written natively in a different writing system. But those writing in a language using the Cyrillic, Arabic, or Chinese scripts can also find themselves needing to transcribe other languages into their own writing systems.
Korean, the official language of the Republic of Korea (South Korea) and the Democratic People’s Republic of Korea (North Korea) and spoken natively by around 8 million people in total, is written in Hangul—the Korean alphabet invented in 1444 by King Sejong the Great—a writing system it shares with no other language. Whereas those writing in English, Finnish, Maltese, Swahili, Vietnamese, or any other language that uses the Roman alphabet can draw upon a vast reservoir of existing work on romanization of different languages, the Korean language communities have been left to their own devices in devising transcription schemes for different languages into Hangul.
Because of the phonetic nature of the Korean alphabet, transcription into Hangul is closely linked with the issue of loanword adaptation into Korean. Like any other language, Korean adapts terms and names from other languages into its own phonology, using sounds that are available in Korean in combinations that are permitted in Korean. Because Korean pronunciation is largely predictable from the spelling, the way a foreign term or name is written in Hangul shows quite precisely how it is adapted to Korean phonology. Hangul transcription and loanword adaptation are therefore mostly equivalent problems.
The topic of loanword adaptation into Korean has been an active are of research. However, the existing scholarly literature often examines loanwords in Korean in comparison to their source language as if they arose naturally and intuitively among Korean speakers. The fact that the loanword adaptation process has been a target of standardization for the better part of a century seems to elude most such studies.
History of transcription into Korean
English has never had a formal language authority to codify and prescribe the standard language, but many languages around the world do have such regulators, such as the Académie française for French and the Asociación de Academias de la Lengua Española for Spanish. Korean is regulated by the National Institute of the Korean Language (국립국어원/國立國語院 Gungnip Gugeo-won) in the South and by the Language Research Institute, Academy of Social Science (사회과학원 어학연구소/社會科學院 語學研究所 Sahoe Gwahak-won Eohak Yeon’gu-so) in the North. These bodies by and large continue the foundations laid by the Korean Language Society (조선어학회/朝鮮語學會 Joseon-eo Hakhoe) in the first half of the 20th century, when Korea was under Japanese occupation.
In 1933, the Korean Language Society published the “Proposition for the Unification of Hangul Orthography” (한글 마춤법 통일안/한글 마춤法 統一案, or 한글 맞춤법 통일안 Han’geul Matchum-beop Tong’il-an in modern orthography). This still forms the basis for the orthography of Standard Korean in both North and South Korea, even though subsequent revisions have led to the divergence of the two standards. The “Proposition” of 1933 included a very brief section on the transcription of loanwords, quoted in its entirety here:
제6장 외래어 표기
제60항 외래어를 표기할적에는 다음의 조건을 원칙으로 한다.
- 새 문자나 부호를 쓰지 아니한다.
- 표음주의를 취한다.
VI. Transcription of loanwords
60. When transcribing loanwords, adopt the following conditions as principles.
- Do not use new letters or symbols.
- Follow the phonetic principle.
This was followed in 1941 by the “Proposition for the Unification of Loanword Transcription Rules” (외래어 표기법 통일안/外來語 表記法 統一案 Oerae-eo Pyogi-beop Tongil-an), which clarified these principles by stating that the transcription into Hangul should be based not on the spelling but solely on the pronunciation of the original terms as transcribed in the International Phonetic Alphabet (IPA), and provided conversion tables between IPA symbols and Hangul letters.
After the end of Japanese occupation and the division of the peninsula in 1945, language regulation went separate ways in the North and the South. In South Korea, Hangul transcription regimes have gone through several iterations—the radically phoneticist “Rules for Writing Borrowed Words” (들온말 적는 법/들온말 적는 法 Deuron-mal Jeongneun Beop) of 1948 which overturned the 1941 “Proposition” to introduce new letters to represent foreign sounds, the “Hangul Transcription Rules for the Roman Alphabet” (로마자의 한글화 표기법/로마字의 한글化 表記法 Roma-ja-ui Han’geul-hwa Pyogi-beop) of 1958 which went back to the previous principle of using only the letters used for Korean to represent foreign sounds, and the current regime, called the “Loanword Transcription Rules” (외래어 표기법/外來語 表記法 Oerae-eo Pyogi-beop), introduced in 1986.
The current South Korean language norms governing standard speech and orthography were also introduced around the same time, as 1988 saw the introduction of both the “Standard Speech Regulations” (표준어 규정/標準語 規定 Pyojun-eo Gyujeong) and the “Korean Orthography” (한글 맞춤법/한글 맞춤法 Han’geul Matchum-beop). The latter represented the most recent significant reform of Korean orthography (notably in changing the form of the formal polite declarative suffix for sentence-final verbs previously written as ~읍니다 eumnida to ~습니다 seumnida), though minor revisions have continued to this day.
The basic principles of transcriptions of the current “Loanword Transcription Rules” are as follows:
- Write loanwords only with the 24 letters currently used for Korean.
- In principle, one phoneme in a loanword is written with one sign.
- For syllable codas, use only ㄱ k, ㄴ n, ㄹ l, ㅁ m, ㅂ p, ㅅ t, and ㅇ ng.
- Adopt the principle of not using tense consonants to write plosives.
- Respect common usage for loanwords that are already established, but the scope and examples are to be decided separately.
The initial version published in 1986 included rules for transcribing English, German, French, Spanish, Italian, Japanese, and Chinese. Successive iterations added rules for more languages—Polish, Czech, Serbo-Croatian, Romanian, and Hungarian in 1992; Swedish, Norwegian, and Danish in 1995; Malay/Indonesian, Thai, and Vietnamese in 2004; and Portuguese, Dutch, and Russian in 2005. This makes a total of at least twenty-one languages served by the “Loanword Transcription Rules”, and more if you count e.g. Croatian, Bosnian, Serbian, and Montenegrin separately or Malay and Indonesian separately. In 2006, rules for Greek and Turkish were considered, followed by rules for Arabic in 2007, but these last three were never formalized.
Of course, South Koreans at large are not familiar with the detailed rules governing the transcription of different languages, if they are aware of the existence of such rules at all. But the standard forms of loanwords as well as foreign terms and names that they encounter in schools and in the media are the result of the application of these normative rules, so South Koreans are very familiar with at least some of the results of these normative transcription guidelines, especially if they have been educated after the introduction of the “Loanword Transcription Rules” in 1986. By no means does this mean that all loanwords and foreign terms or names are written according to the rules, as older South Koreans especially may be more familiar with earlier transcriptions that are no longer standard, or as those not familiar with the norms introduce nonstandard transcriptions according to their own judgements. It cannot be ignored however that the “Loanword Transcription Rules” have been hugely influential in shaping loanword adaptation and the transcription of foreign terms and names in South Korea.
Following the announcement of the “Loanword Transcription Rules” in 1986, the Korean Language Research Institute (국어연구소/國語硏究所 Gugeo Yeon’gu-so), as the National Institute of the Korean Language was known at the time, published two volumes of standard transcriptions based on the new rules—Loanword Transcription Examples: General Terms (외래어 표기 용례집: 일반 용어 外來語 表記 用例集 Oerae-eo Pyogi Yongnye-jip: Ilban Yong’eo) and Loanword Transcription Examples: Names of Places, Names of People (외래어 표기 용례집: 지명·인명 Oerae-eo Pyogi Yongnye-jip: Jimyeong·Inmyeong). The latter included some additional principles used for transcribing English and German as well as Latin, Greek, and Russian (the last of which would be superseded in 2005 by the introduction of rules for Russian) along with other languages not covered by the existing rules. The standardized transcriptions (soon followed by the new spellings due to the 1988 reform of Korean orthography) were incorporated into textbooks, which in South Korea are prepared or approved by the Ministry of Education.
The news media also helped to promulgate the new standard transcriptions. This was a time of greatly expanded contact with the outside world for South Korea, as Seoul hosted the Summer Olympic Games in 1988 and all restrictions on travelling to foreign countries were lifted for South Korean citizens in 1989. There was no shortage of new foreign terms and names to transcribe into Korean, and the news media were at the forefront of introducing these to the public. The introduction of the transcription rules alone was not enough on its own to guarantee that all such new terms and names would be transcribed the same by everyone, so a need was recognized for periodic review of these new transcriptions.
In 1991, the Joint Committee of the Government and Press on Loanword Review (정부·언론 외래어 심의 공동 위원회/政府·言論 外來語 審議 共同 委員會 Jeongbu·Eollon Oerae-eo Simui Gongdong Wiwon-hoe) was formed by the language regulator, by then renamed the National Research Institute of the Korean Language (국립 국어 연구원/國立 國語 硏究院 Gungnip Gugeo Yeon’gu-won), and the Korea News Editors’ Association (한국 신문 편집인 협회 Han’guk Sinmun Pyeonjip-in Hyeophoe) to decide on the standard transcriptions of new terms and names that were appearing in the news. The first meeting on 10 September 1991 produced the standard transcriptions for Mongolia 몽골 Monggol (in parallel with the existing Sino-Korean form 몽고/蒙古 Monggo, which was to be phased out), Myanmar 미얀마 Miyanma, Saint Petersburg 상트페테르부르크 Sangteupetereubureukeu, and Yangon 양곤 Yanggon (the last three having just recently been renamed from Burma, Leningrad, and Rangoon respectively).
After convening a few more times, notably in 28 January 1993 to rule on the transcriptions of the names of the independent states following the break-up of the Soviet Union, the Joint Committee decided to start holding regular meetings. Since 1995, they have met every two months to produce standard transcriptions. Even this was felt to be inadequate for keeping up with the news cycle, so in 2009 a working group of the committee started to produce a weekly list of standard transcriptions. The full committee continued to meet every two months to make decisions on the transcriptions that were not unanimously agreed upon in the weekly deliberations of the working group.
Over the years, the volumes of transcription examples (at least eight, including ones that were published for the addition of the transcription rules for new languages) and the decisions of the Joint Committee and its working group have produced a fairly robust body of standard transcriptions, over 65,000 of which are searchable on the website of the National Institute of the Korean language (http://kornorms.korean.go.kr/example/exampleList.do).
Korean phonology and alphabet
Before we take a deeper look at the rules themselves, let us examine the phonology of Korean and how the sounds correspond to the letters of the Korean alphabet. The current standard orthography reflects a conservative pronunciation of Standard Korean.
There is little debate about the inventory of consonant phonemes in Standard Korean. They correspond to the consonant letters of Hangul:
Table 1. The consonant phonemes of Korean as represented by Hangul consonant letters
| Bilabial | Alveolar | Palatal | Velar | Glottal | ||
|---|---|---|---|---|---|---|
| Nasal | ㅁ m | ㄴ n | ㅇ ng | |||
| Stop and Affricate | lenis | ㅂ b | ㄷ d | ㅈ j | ㄱ g | |
| tense | ㅃ pp | ㄸ tt | ㅉ jj | ㄲ kk | ||
| aspirated | ㅍ p | ㅌ t | ㅊ ch | ㅋ k | ||
| Fricative | non-tense | ㅅ s | ㅎ h | |||
| tense | ㅆ ss | |||||
| Liquid | ㄹ l | |||||
Korean is notable for the three-way phonation contrast of its stops and affricates, which can be lenis (also called plain or lax), tense (also called fortis or glottalized), or aspirated. Hangul treats the lenis series as the basic, unmarked consonants, writing them as ㅂ, ㄷ, ㄱ respectively for the bilabial, alveolar, and velar stops, and as ㅈ for the affricate (analysed further, ㅂ and ㄷ are based on the corresponding nasals ㅁ and ㄴ, and ㅈ is based on the corresponding fricative ㅅ). These letters are doubled for the tense series, giving ㅃ, ㄸ, ㄲ, and ㅉ respectively, or are altered in form for the aspirated series, giving ㅍ, ㅌ, ㅋ, and ㅊ respectively.
The exact pronunciation of the stops and affricates depend on the phonetic environment. Utterance-initially, they are all voiceless while maintaining the three-way contrast. The typological rarity (even uniqueness) of contrasting three kinds of voiceless stops and affricates has been frequently noted, and the exact nature of this contrast has been debated extensively. The initial allophones of the lenis stops and affricate are often transcribed as unmarked voiceless sounds in IPA, e.g. [p], [t], and [k], but these sounds are much weaker than the tense or aspirated counterparts—the duration between the closure and the release is shorter and is accompanied by a much smaller build-up of air pressure. In addition, these are also usually weakly aspirated, although not as much as the aspirated series (at least in older speakers). However, between voiced sounds (i.e. vowels, glides, nasals, and liquids), the lenis series becomes voiced, e.g. [b], [d], and [ɡ].
The exact nature of the tense series is not agreed upon, but it has been found that at least in word-initial position, the tense series is not perceptibly different from the voiceless unaspirated consonants of languages such as Chinese and Spanish (Charles B. Chang). Moreover, between a voiceless consonant and a vowel, the lenis series neutralizes to the tense series, suggesting that it is simply the voiceless counterpart of the lenis series in this position. For example, 입다 |ib-da| ipda “to wear” is pronounced as if it were spelled 입따 |ib-tta| iptta.
A three-way contrast featuring voiced, voiceless aspirated, and some kind of voiceless unaspirated stops or affricates is typologically quite common, and can be found in such languages as Thai, Lao, Eastern Armenian, Kurdish, Punjabi, Classical Tibetan, and Classical Greek. Kim and Duanmu (2004) have argued that the Korean stops and affricates should be analysed as having an underlying contrast of voiced, voiceless aspirated, and voiceless unaspirated, with devoicing occurring initially.
The stops at each point of articulation are neutralized in coda position as unreleased stops [p̚], [t̚], or [k̚]. Moreover, all sibilants (affricates and fricates) also become [t̚] in coda position. The underlying contrasts resurface when followed by a vowel-initial particle or verb ending. So 낟 |nad| nat “grain”, 낱 |nat| nat “each”, 낮 |naj| nat “daytime”, 낯 |nach| nat “face”, and 낫 |nas| nat “sickle” are all pronounced [nat̚], though followed by the topic-marking particle 은 eun, they are all distinguished in pronunciation: 낟은 nadeun, 낱은 nateun, 낮은 najeun, 낯은 nacheun, 낫은 naseun. The morphophonemic orthography of Korean extends to writing consonant clusters at coda position which only surface in such combinations. The verb stem 읽- |ilg| “to read” is pronounced like 익 ik in the basic form 읽다 |ilg-da| ikda (pronounced as if spelled 익따 iktta), but the written cluster is pronounced in the connective form 읽어 |ilg-eo| ilgeo.
In the literature, the Korean affricates have been described as postalveolar ([dʒ, tʃ, tʃʰ]), more recently as alveolo-palatal ([dʑ, tɕ, tɕʰ]), and even more recently as alveolar ([dz, ts, tsʰ]) (Kim 2001). There seem to be regional and generational variations. It may be that the alveolar pronunciation is more likely before back vowels.
The sibilant fricatives ㅅ s and ㅆ ss show a two-way phonation contrast. The latter is uncontroversially tense, but the former has been observed to pattern with the lenis series in some ways (it neutralizes with the tense series after voiceless consonants just like lenis stops and affricates) and with the aspirated series in others (it does not undergo voicing in voiced environments, at least canonically, and retains aspiration). The basic place of articulation is alveolar ([z̥ʰ, s]), but it becomes alveolo-palatal ([ʑ̥ʰ, ɕ]) before /i/ and /j/, and postalveolar ([ʒ̊ʰ, ʃ]) before /y~wi/.
The glottal fricative ㅎ h is normally /h/, but word-initially it can be velar ([x]) before /ɯ/ and palatal ([ç]) before /i/ and /j/, and bilabial ([ɸ]) before /u/ and /o/.
Table 2. The vowel phonemes of Conservative Standard Korean
| Front | Central | Back | |||
| unrounded | rounded | unrounded | unrounded | rounded | |
| Close | ㅣ i /i/ | ㅟ wi /y/ | ㅡ eu /ɯ/ | ㅜ u /u/ | |
| Close-mid | ㅔ e /e/ | ㅚ oe /ø/ | ㅗ o /o/ | ||
| Open-mid | ㅐ ae /ɛ/ | ㅓ eo/ʌ/ | |||
| Open | ㅏ a /a/ | ||||
Different dialects of Korean exhibit different vowel systems, but conservative Standard Korean has a 10-vowel system as represented above. Hangul originally represented the vowels of 15th-century Middle Korean using the seven basic vowel letters ㅏ a, ㅓ eo, ㅗ o, ㅜ u, ㅡ eu, ㅣ i, and ㆍå, and the combination of the first four with a /j/ on-glide as ㅑ ya, ㅕ yeo, ㅛ yo, and ㅠ yu respectively (the phonemeㆍå disappeared in Standard Korean, though it survives in conservative versions of the Jeju dialect). Combinations of these eleven letters are supposed to have represented diphthongs, though in Modern Korean, ㅐ ae from ㅏ a and ㅣ i as well as ㅔ e from ㅓ eo and ㅣ i have become monophthongs (i.e. pure vowels). Others, such as ㅘ wa from ㅗ o and ㅏ a as well as ㅝ wo from ㅜ u and ㅓ eo, have developed into vowels with a /w/ on-glide (if they did not represent such sounds in the first place).
Table 3. Glide-vowel combinations in Korean
| /j/ | /w/ | /ɰ/ | |
| ㅣ i /i/ | ㅟ wi /wi/ | ㅢ ui /ɰi/ | |
| ㅔ e /e/ | ㅖ ye /je/ | ㅞ we /we/ | |
| ㅐ ae /ɛ/ | ㅒ yae /jɛ/ | ㅙ wae /wɛ/ | |
| ㅏ a /a/ | ㅑ ya /ja/ | ㅘ wa /wa/ | |
| ㅓ eo /ʌ/ | ㅕ yeo /jʌ/ | ㅝ wo /wʌ/ | |
| ㅗ o /o/ | ㅛ yo /jo/ | ||
| ㅜ u /u/ | ㅠ yu /ju/ |
Conservative Standard Korean also has a phonemic distinction of long and short vowels which was only preserved in the first syllable of words. The long vowels have virtually identical qualities as their short counterparts for the most part, but the long version of ㅓ eo is a more central [əː] or [ɘː] as opposed to the short ㅓ eo [ʌ]. However, vowel length contrast has largely disappeared in the contemporary Standard Korean of younger speakers, though it continues to be taught in school as part of the “Standard Pronunciation Rules” (표준 발음법/標準 發音法 Pyojun Bareum-beop), itself part of the “Standard Speech Regulations” introduced in 1988.
The front rounded vowels ㅟ wi and ㅚ oe also became monophthongs /y/ and /ø/ in conservative Standard Korean, but further shifted to vowels with a /w/ on-glide in most contemporary varieties. Speakers of contemporary Standard Korean overwhelmingly pronounce these as /wi/ [ɥi] and /we/, the latter identical to 웨 we. Note that following the conservative pronunciation, we could reserve ㅚ oe to represent /ø/ to distinguish it from ㅞ we /we/, but Hangul does not distinguish between the pronunciations /y/ and /wi/ for ㅟ wi. The “Standard Pronunciation Rules” still prescribe the monophthongal pronunciations /y/ and /ø/ for ㅟ wi and ㅚ oe as primary, but also permit the pronunciations /wi/ and /we/.
A notable aspect of contemporary Standard Korean is the merger of ㅔ e /e/ and ㅐ ae /ɛ/ to a single front mid vowel /e/. This also means that ㅖ ye /je/ merges with ㅒ yae /jɛ/ and ㅞ we /we/ with ㅙ wae /wɛ/ (along with ㅚ oe /ø/ as already discussed). However, the “Standard Pronunciation Rules” do not recognize this merger. Nevertheless, together with the /w/ on-glide pronunciations of ㅟ wi and ㅚ oe, the merger of ㅔ e /e/ and ㅐ ae /ɛ/ means that contemporary Standard Korean has a 7-vowel system, with /y/, /ø/, and /ɛ/ removed from the 10-vowel system of conservative Standard Korean.
A Korean syllable is maximally CGVC, with one or no consonant in each of the onset and coda positions and a nucleus consisting of one or no glide (/j/, /w/, /ɰ/) and a vowel. Only the nasals and stops can occur in coda position. The velar nasal ㅇ ng /ŋ/ only occurs as a coda, never as an onset.
How different sounds are transcribed
The “Loanword Transcription Rules” provide various tables to indicate how the sounds or letters in the source language should be transcribed in Hangul. The table below is used for the transcriptions of English, German, and French.
Table 4. Comparison table of the International Phonetic Alphabet and Hangul letters
| IPA symbol | Hangul: before vowel | Hangul: before consonant or word-final |
|---|---|---|
| p | ㅍ p | ㅂ p, 프 peu |
| b | ㅂ b | 브 beu |
| t | ㅌ t | ㅅ t, 트 teu |
| d | ㄷ d | 드 deu |
| k | ㅋ k | ㄱ k, 크 keu |
| ɡ | ㄱ g | 그 geu |
| f | ㅍ p | 프 peu |
| v | ㅂ b | 브 beu |
| θ | ㅅ s | 스 seu |
| ð | ㄷ d | 드 deu |
| s | ㅅ s | 스 seu |
| z | ㅈ j | 즈 jeu |
| ʃ | 시 si | 슈 syu, 시 si |
| ʒ | ㅈ j | 지 ji |
| ts | ㅊ ch | 츠 cheu |
| dz | ㅈ j | 즈 jeu |
| tʃ | ㅊ ch | 치 chi |
| dʒ | ㅈ j | 지 ji |
| m | ㅁ m | ㅁ m |
| n | ㄴ n | ㄴ n |
| ɲ | 니* ni/ny | 뉴 nyu |
| ŋ | ㅇ ng | ㅇ ng |
| l | ㄹ r, ㄹㄹ ll | ㄹ l |
| r | ㄹ r | 르 reu |
| h | ㅎ h | 흐 heu |
| ç | ㅎ h | 히 hi |
| x | ㅎ h | 흐 heu |
| IPA symbol | Hangul |
|---|---|
| j | 이* i/y |
| ɥ | 위 wi |
| w | 오, 우* o, u/w |
| IPA symbol | Hangul |
|---|---|
| i | 이 i |
| y | 위 wi |
| e | 에 e |
| ø | 외 oe |
| ɛ | 에 e |
| ɛ̃ | 앵 aeng |
| œ | 외 oe |
| œ̃ | 욍 oeng |
| æ | 애 ae |
| a | 아 a |
| ɑ | 아 a |
| ɑ̃ | 앙 ang |
| ʌ | 어 eo |
| ɔ | 오 o |
| ɔ̃ | 옹 ong |
| o | 오 o |
| u | 우 u |
| ə** | 어 eo |
| ɚ | 어 eo |
You can see the full transcription rules for English, German, and French here.
Comparison tables are provided for other languages, but instead of using the symbols of the International Phonetic Alphabet as for English, German, and French, these use the letters in the original orthography or in the standard romanizations. The tables for the transcription of Italian and Chinese are provided here as examples.
Table 5. Comparison table of the Italian alphabet and Hangul letters
| Letter | Hangul: before vowel | Hangul: before consonant or word-final | Examples |
|---|---|---|---|
| b | ㅂ b | 브 beu | Bologna 볼로냐 Bollonya, bravo 브라보 beurabo. |
| c | ㅋ k, ㅊ ch | 크 keu | Como 코모 Komo, Sicilia 시칠리아 Sichillia, credo 크레도 keuredo. |
| ch | ㅋ k | — | Pinocchio 피노키오 Pinokio, cherubino 케루비노 kerubino. |
| d | ㄷ d | 드 deu | Dante 단테 Dante, drizza 드리차 deuricha. |
| f | ㅍ p | 프 peu | Firenze 피렌체 Pirenche, freddo 프레도 peuredo |
| g | ㄱ g, ㅈ j | 그 geu | Gallileo 갈릴레오 Gallille’o, Genova 제노바 Jenoba, gloria 글로리아 geulloria. |
| h | — | — | hanno 안노 anno, oh 오 o. |
| l | ㄹ r, ㄹㄹ ll | ㄹ l | Milano 밀라노 Millano, largo 라르고 rareugo, palco 팔코 palko. |
| m | ㅁ m | ㅁ m | Macchiavelli 마키아벨리 Makiabelli, mamma 맘마 mamma, Campanella 캄파넬라 Kampanella. |
| n | ㄴ n | ㄴ n | Nero 네로 Nero, Anna 안나 Anna, divertimento 디베르티멘토 dibereutimento. |
| p | ㅍ p | 프 peu | Pisa 피사 Pisa, prima 프리마 peurima. |
| q | ㅋ k | — | quando 콴도 kwando, queto 퀘토 kweto. |
| r | ㄹ r | 르 reu | Roma 로마 Roma, Marconi 마르코니 Mareukoni. |
| s | ㅅ s | 스 seu | Sorrento 소렌토 Sorento, asma 아스마 aseuma, sasso 사소 saso. |
| t | ㅌ t | 트 teu | Torino 토리노 Torino, tranne 트란네 teuranne. |
| v | ㅂ b | 브 beu | Vivace 비바체 Bibache, manovra 마노브라 manobeura. |
| z | ㅊ ch | — | nozze 노체 noche, mancanza 만칸차 mankancha. |
| Letter | Hangul | Examples |
|---|---|---|
| a | 아 a | abituro 아비투로 abituro, capra 카프라 kapeura. |
| e | 에 e | erta 에르타 ereuta, padrone 파드로네 padeurone. |
| i | 이 i | infamia 인파미아 inpamia, manica 마니카 manika. |
| o | 오 o | oblio 오블리오 obeullio, poetica 포에티카 po’etika. |
| u | 우 u | uva 우바 uba, spuma 스푸마 seupuma. |
You can see the full transcription rules for Italian here.
Table 6. Comparison table of Chinese phonetic symbols and Hangul letters
| Pinyin | Zhuyin | Hangul |
|---|---|---|
| Bilabial | ||
| b | ㄅ | ㅂ b |
| p | ㄆ | ㅍ p |
| m | ㄇ | ㅁ m |
| Labiodental | ||
| f | ㄈ | ㅍ p |
| Denti-alveolar | ||
| d | ㄉ | ㄷ d |
| t | ㄊ | ㅌ t |
| n | ㄋ | ㄴ n |
| l | ㄌ | ㄹ r |
| Velar | ||
| g | ㄍ | ㄱ g |
| k | ㄎ | ㅋ k |
| h | ㄏ | ㅎ h |
| Alveolo-palatal | ||
| j | ㄐ | ㅈ j |
| q | ㄑ | ㅊ ch |
| x | ㄒ | ㅅ s |
| Retroflex | ||
| zh [zhi] | ㄓ | ㅈ j [즈 jeu] |
| ch [chi] | ㄔ | ㅊ ch [츠 cheu] |
| sh [shi] | ㄕ | ㅅ s [스 seu] |
| r [ri] | ㄖ | ㄹ r [르 reu] |
| Dental sibilant | ||
| z [zi] | ㄗ | ㅉ jj [쯔 jjeu] |
| c [ci] | ㄘ | ㅊ ch [츠 cheu] |
| s [si] | ㄙ | ㅆ ss [쓰 sseu] |
| Pinyin | Zhuyin | Hangul |
|---|---|---|
| Monophthong | ||
| a | ㄚ | 아 a |
| o | ㄛ | 오 o |
| e | ㄜ | 어 eo |
| ê | ㄝ | 에 e |
| yi (i) | ㄧ | 이 i |
| wu (u) | ㄨ | 우 u |
| yu (u) | ㄩ | 위 wi |
| Diphthong | ||
| ai | ㄞ | 아이 ai |
| ei | ㄟ | 에이 ei |
| ao | ㄠ | 아오 ao |
| ou | ㄡ | 어우 eou |
| Nasal coda | ||
| an | ㄢ | 안 an |
| en | ㄣ | 언 eon |
| ang | ㄤ | 앙 ang |
| eng | ㄥ | 엉 eong |
| Rhotic coda | ||
| er (r) | ㄦ | 얼 eol |
| Composite rimes | ||
| [j]-glide | ||
| ya (ia) | ㄧㄚ | 야 ya |
| yo | ㄧㄛ | 요 yo |
| ye (ie) | ㄧㄝ | 예 ye |
| yai | ㄧㄞ | 야이 yai |
| yao (iao) | ㄧㄠ | 야오 yao |
| you (iou, iu) | ㄧㄡ | 유 yu |
| yan (ian) | ㄧㄢ | 옌 yen |
| yin (in) | ㄧㄣ | 인 in |
| yang (iang) | ㄧㄤ | 양 yang |
| ying (ing) | ㄧㄥ | 잉 ing |
| [w]-glide | ||
| wa (ua) | ㄨㄚ | 와 wa |
| wo (uo) | ㄨㄛ | 워 wo |
| wai (uai) | ㄨㄞ | 와이 wai |
| wei (ui) | ㄨㄟ | 웨이 wei (우이 u’i) |
| wan (uan) | ㄨㄢ | 완 wan |
| wen (un) | ㄨㄣ | 원 won (운 un) |
| wang (uang) | ㄨㄤ | 왕 wang |
| weng (ong) | ㄨㄥ | 웡 wong (웅 ung) |
| [ɥ]-glide | ||
| yue (ue) | ㄩㄝ | 웨 we |
| yuan (uan) | ㄩㄢ | 위안 wian |
| yun (un) | ㄩㄣ | 윈 win |
| yong (iong) | ㄩㄥ | 융 yung |
You can see the full transcription rules for Chinese here.
Vowels
The vowel transcriptions in the “Loanword Transcription Rules” are mostly straightforward, but some remarks are in order about the use of 애 ae to transcribe [æ] instead of [ɛ] which is the normal pronunciation of 애 ae in Korean. The motivation seems to be practical. For French and German as well as for many other languages that use the Roman alphabet (which only has the letter ‘e’ for a mid front unrounded vowel), making a systematic distinction between [e] and [ɛ] is fraught with difficulty as one cannot always rely on the spelling to do so. Sometimes, both pronunciations are possible according to dialect or even in free variation—French les can be [le] or [lɛ], and German Daniel can be [ˈdaːni̯eːl] or [ˈdaːni̯ɛl]. On the other hand, there is no Korean vowel normally pronounced [æ] as in English (although according to dialect English /æ/ may approach [ɛ] in quality), but 애 ae is the closest approximation. By reserving 애 ae for the transcription of [æ], the “Loanword Transcription Rules” leaves 에 e to cover both [e] and [ɛ], which are often written with the same letter ‘e’ in the source language.
While front rounded vowels have largely disappeared from contemporary Standard Korean, the “Loanword Transcription Rules” still make use of the letters that represent these sounds in conservative Standard Korean. Therefore, the close front rounded vowel [y] is mapped to 위 wi while the close-mid front rounded vowel [ø] and the open-mid front rounded vowel [œ] are both mapped to 외 oe. The “Loanword Transcription Rules” inherit the inherent ambiguity of the Hangul letter, as 위 wi is also used to represent the glide-vowel sequence [wi].
As Korean does not have nasalized vowels as are used in French, they are imitated with sequences of a vowel and ㅇ ng [ŋ]. So the French nasalized vowels [ɑ̃], [ɔ̃], and [œ̃] are transcribed as 앙 ang, 옹 ong, and 욍 oeng respectively. Interestingly, [ɛ̃] is transcribed as 앵 aeng despite the fact that [ɛ] is transcribed as 에 e, not 애 ae. However, this seems to be a concession to the fact that in the contemporary French pronunciation of Paris, the nasalized vowel traditionally transcribed as [ɛ̃] has shifted towards the vowel space of [ɑ̃], often ending up as [æ̃].
Consonants
Voiceless stops and affricates
English, German, and French all have a two-way distinction in stops and affricates based on the presence or absence of voicing, as opposed to the three-way phonation contrast in Korean. The question then becomes how to map the two-way contrast in the source language to the three-way contrast in Korean.
The solution adopted by the “Loanword Transcription Rules” can be gleamed from the fourth of its five basic principles: “Adopt the principle of not using tense consonants to write plosives.” Voiced stops and affricates are mapped to the lenis series, while unvoiced ones are mapped to the aspirated series. The tense series goes unused. So [b, d, ɡ, dz, dʒ] get mapped to lenis ㅂ b, ㄷ d, ㄱ g, ㅈ j, and ㅈ j respectively before vowels while [p, t, k, ts, tʃ] get mapped to aspirated ㅍ p, ㅌ t, ㅋ k, ㅊ ch, and ㅊ ch. Tense ㅃ pp, ㄸ tt, ㄲ kk, and ㅉ jj are not used.
The choice of the lenis series to represent voiced stops and affricates is obvious as these are indeed realized as voiced in Korean in many environments. But voiceless stops and affricates in different languages can be perceived as closer to either the aspirated series or the tense series in Korean. Voiceless stops at the onset of stressed syllables are generally aspirated in English and in German, while they are not aspirated in French. So the English and German stops sound more like the aspirated series, while the French, Spanish, and Italian stops sound more like the tense series. But since aspiration is not contrastive in languages such as English, German, and French, the actual realization of the stops and affricates can be quite complex with regard to aspiration. English stops are unaspirated in clusters such as [sp], [st], and [sk]. Stops are not aspirated in certain dialects of English (e.g. Traditional Scottish English, Broad White South African English) and German (e.g. southern varieties for [p] and [t]), Affricates are generally aspirated in English (at least at the onset of stressed syllables), but not in German. While French [p], [t], and [k] are perceived to be closer to the tense series, in clusters such as [pʁ], [tʁ], and [kʁ] they are usually perceived to be closer to the aspirated series since the [ʁ] is devoiced and surfaces as [χ], which sounds like aspiration.
The “Loanword Transcription Rules” go around this difficulty by mapping all voiceless stops and affricates to the aspirated series, whether they are aspirated or not in the source language. This simplifies the task of transcription, especially when extended to languages where it is difficult for the average Korean speaker to find information about aspiration, say Albanian or Malagasy. However, this stricture on using tense consonants for stops and the blanket conversion of voiceless stops and affricates in the source language to the aspirated series have their fair share of detractors, led by the influential publishing house Changbi Publishers who use their own transcriptions in their publications. A political dimension is added by the fact that the transcription method used in North Korea takes a similar approach.
For example, the standard Korean name for the capital of France is 파리 Pari based on the French pronunciation of Paris [paʁi] (or [pari] in the simplified transcription scheme used by the “Loanword Transcription Rules”), but Changbi Publishers use 빠리 Ppari instead which sounds closer to how Korean speakers perceive the French pronunciation. 빠리 Ppari is also the standard spelling in North Korea.
Chinese also has a two-way distinction in stops and affricates, but both series are voiceless. Instead, the distinction is based on the presence or absence of aspiration. The aspirated stops and affricates, written p, t, k, q, ch, and c in pinyin, are naturally mapped to the aspirated series in Korean—ㅍ p, ㅌ t, ㅋ k, ㅊ ch, ㅊ ch, ㅊ ch. The unaspirated stops and affricates are usually perceived to sound like either the tense series or the lenis series in Korean, leaning towards the former when accompanied by a high tone or when pronounced more forcefully and towards the latter when accompanied by a low tone or when pronounced less forcefully, especially in voiced environments. The “Loanword Transcription Rules” maps all the unaspirated stops and two of the unaspirated aspirates of Chinese to the lenis series in Korean, so b, d, g, j, and zh are transcribed as ㅂ b, ㄷ d, ㄱ g, ㅈ j, and ㅈ j. However, the alveolar affricate [ts] written z in pinyin is transcribed as ㅉ jj, the tense counterpart of ㅈ j. This allows one to distinguish z from the retroflex initial zh [ʈʂ] in Korean transcription when they take the same rimes (the alveolo-palatal j [tɕ] does not occur with the same rimes as z), though no such distinction is possible for the initials ch and c in Korean transcription. The same idea is extended to transcribing the alveolar fricative [s] written s in pinyin as tense ㅆ ss while the retroflex sh [ʂ] is written as ㅅ s, as is the alveolo-palatal x [ɕ]. The use of ㅉ jj to write z and of ㅆ ss to write s in Chinese does not violate the basic principle of not using tense consonants to write plosives since these are an affricate and a fricative respectively, not stops (plosives).
Danish is another language where voiceless unaspirated stops written b, d, and g are transcribed as lenis consonants in Korean, though these come from historically voiced stops.
The stricture against the use of the tense series for stops was loosened when rules were added for transcribing Thai and Vietnamese in 2004. For the bilabial and alveolar stops, Thai maintains a three-way distinction of voiced, voiceless unaspirated, and voiceless aspirated sounds, [b], [p], [pʰ] and [d], [t], [tʰ] respectively, that map well onto the Korean lenis, tense, and aspirated series. This was followed by the “Loanword Transcription Rules”, leading to transcriptions such as ‘Phuket’ 푸껫 Pukket and ‘Pattani’ 빠따니 Ppattani.
The Vietnamese dental/alveolar stops have a similar three-way distinction of [ɗ], [t], and [tʰ], where the voiced stop is also glottalized. The other places of articulation lack the full distinction, but treating some of the fricatives as if they were stops allows a similar mapping into Korean. In fact, as reflected in the Vietnamese alphabet, the Middle Vietnamese values ca. the 17th century of ph [f] and kh [x] are thought to have been [pʰ] and [kʰ] respectively. Using these mappings, the “Loanword Transcription Rules” give us such transcriptions as ‘Cần Thơ’ 껀터 Kkeonteo and ‘Duy Tân’ 주이떤 Ju’itteon.
Coda consonants
The distinction between the plain, aspirated, and tense series is neutralized in coda position in Korean. Only the bilabial, alveolar, and velar stops ([p̚, t̚, k̚]) as well as the corresponding nasals ([m, n, ŋ] plus the lateral approximant ([l]). This accounts for the third of the five basic principles in the “Loanword Transcription Rules”: “For syllable codas, use only ㄱ k, ㄴ n, ㄹ l, ㅁ m, ㅂ p, ㅅ t, and ㅇ ng. In the morphophonemic spelling used for Korean, the coda [k̚] for example may be spelled ㄱ, ㄲ, ㄳ, ㄺ, ㅋ to indicate underlying sound distinctions which may surface when a vowel-initial particle or morphological ending is attached. But such distinctions were judged to be unnecessary for transcribing sounds from other languages.
The choice of ㄱ k and ㅂ p respectively for coda [k̚] and [p̚] make sense, as these are the same letters that are used for the plain unmarked consonant in onset position—ㄱ g and ㅂ b. The choice of ㅅ t over ㄷ t seems more puzzling on the surface, as in onset position ㅅ s stands for a sibilant fricative. However, coda ㅅ t is treated as the default spelling for coda [t̚] even in “Korean Orthography” (III 3.7): “Write ㅅ t for coda ㄷ t sounds that have no special reason to be written as ㄷ t.” This tendency to treat the neutralized [t̚] as underlying /z/ can also be observed in frequent nonstandard pronunciations. For example, 꽃이 kkot-i (“flower” as subject) is supposed to be pronounced as if it is written 꼬치 kkochi in Standard Korean, but many speakers will pronounce it as if it is written 꼬시 kkosi instead.
However, the use of coda ㄱ k, ㅂ p, and ㅅ t is rather limited in scope in the “Loanword Transcription Rules”. They are used most often for voiceless stops that are followed by obstruents, as in transcribing English October 옥토버 oktobeo and September 셉템버 septembeo, German Oktober 옥토버 oktobeo and September 젭템버 jeptembeo, and French octobre 옥토브르 oktobeureu and septembre 셉탕브르 septangbeureu. The exact scope where coda ㄱ k, ㅂ p, and ㅅ t can be used differs slightly according to the language—in English, it needs to be after a short vowel, and in French, it needs to be after an oral vowel and before a voiceless consonant (the scope is not defined explicitly for German).
However, word-final [p], [t], or [k] in German or French are never written as ㅂ p, ㅅ t, or ㄱ k but rather as 프 peu, 트 teu, or 크 keu respectively, with the epenthetic vowel 으 eu appended. The high back unrounded vowel 으 eu [ɯ] can be argued to function as the default epenthetic vowel in Korean. For example, the connecting ending -니 –ni becomes -으니 –euni following a stem with a coda consonant other than ㄹ l. The epenthesis of 으 eu is in fact the usual way of writing both voiced and voiceless stops that are not followed by a vowel.
However, word-final [p], [t], and [k] in English are written as ㅂ p, ㅅ t, and ㄱ k respectively when they are preceded by a short vowel. Therefore, ‘peak’ [piːk] is written as 피크 pikeu because the [k] is preceded by a long vowel [iː], but ‘pick’ [pɪk] (or [pik] in the simplified transcription used by the “Loanword Transcription Rules”) is written as 픽 pik because the [k] is precede by a short vowel [ɪ]. This seems to be because English syllable-final stops are often unreleased in fluent speech. Limiting the use of Korean coda consonants to the cases where the preceding vowel is short makes this the only case that vowel length can be distinguished between pairs such as [iː] and [ɪ] in English, which are both mapped to 이 i.
Dutch is the only other language besides English in the “Loanword Transcription Rules” where word-final [p], [t], and [k] are written as Korean coda consonants after short vowels and with 으 eu appended in other cases. All other languages always append 으 eu for all word-final stops except for Malay/Indonesian, Thai, and Vietnamese, where word-final stops are written as coda consonants. This is because in these Southeast Asian languages, the stops are indeed unreleased just as in Korean. For Spanish, the rules seem to include writing word-final ‘c’ as ㄱ k, with the example given being writing ‘bistec’ as 비스텍 biseutek, but this is a minor and perhaps accidental exception given that word-final ‘c’ is confined to loanwords in Spanish.
When not followed by a vowel, the fricatives [f], [v], [θ], [ð], [s], [z], and [x] as well as the affricates [ts] and [dz] are also written with 으 eu appended, but the fricatives [ʃ], [ʒ], and [ç] as well as the affricates [tʃ] and [dʒ] are treated differently. The voiceless palatal fricative [ç] is written 히 hi, which accords with the fact that [ç] can be an allophone of word-initial ㅎ h before [i] or [j]. This contrasts with the voiceless velar fricative [x] being written as 흐 heu, [x] being a possible allophone of word-initial ㅎ h before [ɯ].
An epenthetic vowel other than [ɯ] usually indicates that the consonant has a palatal nature, a rounded nature, or both. For the postalveolar affricates [tʃ] and [dʒ], which can be considered palatal in a broad sense, the epenthetic vowel used is 이 i, allowing them to be contrasted with the dental or alveolar affricates [ts] and [dz]. For the postalveolar fricatives [ʃ] and [ʒ], the picture is rather complicated. In English, [ʃ] is written 슈 syu before a consonant and 시 si word-finally, and [ʒ] is written 지 ji before a consonant or word-finally. Meanwhile, in French and German, [ʃ] is always written 슈 syu before a consonant or word-finally, while in French, [ʒ] is written 주 ju. The latter is used because 쥬 jyu is pronounced identically to 주 ju in Standard Korean. The choice of ㅠ yu as the epenthetic vowel represents both the palatal and rounded nature of the consonants.
Liquids
Korean has only one liquid phoneme ㄹ r/l which stands for both [r] and [l] word-initially, so [ra] and [la] are both written 라 ra. Between vowels, [r] is written as a single ㄹ r whereas [l] is written with this letter doubled as ㄹㄹ ll so that [ara] is written 아라 ara but [ala] is written 알라 alla. The same applies after a consonant that takes on an epenthetic vowel, so [abra] is written 아브라 abeura but [abla] is written 아블라 abeulla. Word-finally or before a consonant, [l] is simply written as a coda ㄹ l but [r] takes on an epenthetic 으 eu so that [al] is written 알 al but [ar] is written 아르 areu. This way, the “Loanword Transcription Rules” take advantage of the allophonic pronunciation of coda ㄹ l as [l] and of doubled ㄹㄹ ll as [ll] and of intervocalic ㄹ r as [ɾ].
For the transcription of English, however, the non-rhotic pronunciation is taken as the basis. So ‘star’ [stɑː(ɹ)] and ‘computer’ [kəmˈpjuːtə(ɹ)~kəmˈpjuːtɚ] are written 스타 seuta and 컴퓨터 keompyuteo respectively. The Chinese rhotic coda written in pinyin as er and pronounced as [ɤɻ~ɚ] or [aɻ~aɚ] is written as 얼 eol. This could be because the rhotacized vowels of English and Chinese are not perceived to have a prominent enough consonant segment at the end that sounds like a tapped [ɾ] or rolled [r] to warrant an epenthetic vowel.
Geminates
English, German, and French share with one another the aversion for geminate consonants within a morpheme. French allows for a stylistically marked pronunciation of words such as ‘addition’ [a(d)disjɔ̃], ‘grammaire’ [ɡʁa(m)mɛːʁ], ‘annexe’ [a(n)nɛks], and ‘horreur’ [ɔ(ʁ)ʁœːʁ] with optional gemination, but the usual pronunciation is one without such gemination, even in cases where the doubled consonants are due to transparent prefixes such as ‘illegal’ [i(l)leɡal] and ‘innomé’ [i(n)nɔme].
However, there are languages where gemination is distinctive, such as Italian where all consonants other than /z/ can be geminated. The “Loanword Transcription Rules” ignore this gemination for most consonants, writing ‘Pinocchio’ [piˈnɔkkjo] as 피노키오 Pinokio, ‘Puccini’ [putˈtʃini] as 푸치니, and ‘Sorrento’ [sorˈrɛnto] as 소렌토 Sorento. However, -mm- and -nn- are written ㅁㅁ mm and ㄴㄴ nn respectively: ‘Emma’ [ˈɛmma] as 엠마 emma, ‘Gianni’ [ˈdʒanni] as 잔니 janni.
Other than [mm] and [nn], the other Italian geminates are difficult to capture in Korean. Voiced consonants become voiceless when geminated in Korean—compare 아가 aga [aɡa] and 악가 akga [akka]. Even 아카 aka [akʰa] and 악카 akka [akkʰa] cannot reliably be distinguished in pronunciation unless carefully enunciated. So [mm] and [nn] are the only geminates that are treated as such in Korean transcription, along with [jj] and [ww] in transcribing languages such as Arabic (eg. ريان Rayyān as 라이얀 Raiyan, عواد ʿAwwād as 아우와드 Auwad) though the transcription rules for Arabic have not yet been officially codified.
Conclusion
The “Loanword Transcription Rules” govern the standard transcription of terms and names from different languages into Korean. First introduced in 1986, they continued to add rules for transcribing more languages until 2005 and the rules are still being refined (with the latest revision dated 28 March 2017). Many of the most salient practices in loanword adaptation in Korean such as the avoidance of the tense series and the choice of specific epenthetic vowels are due to these rules, which exist in order to standardize the transcription of foreign sounds in Korean. Awareness of these rules is helpful not just for learners of the Korean language, who would find these useful to learn how to write their own names and talk about their own cultures and places of origin in Korean, but also essential to any academic study of Korean loanword adaptation as background in how the official norms have shaped what is still often assumed to be a spontaneous unguided phenomenon.
References
외래어 표기법. 문화체육관광부 고시 제2017-14호. 2017.3.28. at http://kornorms.korean.go.kr/regltn/regltnView.do
Kim, Hyunsoon. “The place of articulation of the Korean plain affricate in intervocalic position: an articulatory and acoustic study.” In Journal of the International Phonetic Association Volume 31, Issue 2, December 2001, 229-257.
Loanword Transcription Rules | Examination |