CtrlK
Discord Join our cozy Discord community to chat with like-minded people and stay up to date with MangaBaka (and get the BAKA server tag)

MangaBaka Korean Romanisation Style

> [!NOTE]
> This romanisation guide is not entirely finished and might change at any moment, be it partially or entirely.
> Last update: 2025-01-18

## Core Romanisation Style

This style guide defines how to consistently romanise and style Korean-language titles (Manhwa/Webtoons) for metadata purposes.
The goal is to provide a standard that is easy to read, type, and apply across various databases, while respecting the linguistic norms of **Revised Romanisation (RR)**

### The Basis

We follow the **[Revised Romanisation of Korean](https://en.wikipedia.org/wiki/Revised_Romanization_of_Korean)** with the following specific enhancements and deviations to align with our styling standards:

- **Loanword Restoration**: Translate clear **English/foreign loanwords** into their standard spelling (e.g., 택시 → "Taxi", 헌터 → "Hunter") instead of strict transliteration (*Taeksi*, *Heonteo*) ― **major deviation**.
- **Particle Separation**: All grammatical particles are **separated by a space** and written in **lowercase**, regardless of standard Hangul agglutination (e.g., 나를 → *Na reul*)
- **Pronunciation-based Transcription**: Romanisation reflects **pronunciation** (sound changes/assimilation) rather than strict Hangul ortography, strictly adhering to standard RR rules (e.g., 독립 → *Dongnip*, not *Dokrip*).
- **Names**: Family names are capitalised, and given names are hyphenated (e.g., 김민수 → *Kim Min-su*).

### Capitalisation

As we are romanising **titles**, we use the **Chicago Headline Style**:

- Capitalise all **nouns**, **verbs**, **adjectives**, **adverbs** and **auxiliary verbs** (when separated).
- Keep all **particles** (e.g., eun, neun, i, ga, ui), **suffixes** when hyphenated, and **verb endings** (when fused) in lowercase.
- **Copulas** (ida, anida) follow the specific fusion rules (see [Copulas](#copulas-ida--anida))

### Word Spacing, Fusions, and Hyphenation

Korean is an agglutinate language, but for title legibility, we standardise distinct word boundaries.

- **Nouns and Compounds**: Keep compound nouns separate unless they are established single dictionary words.
- **Verbs and Adjectives**: The verb stem and its grammatical endings (suffixes) are always **fused** (e.g., 가다 → *Gada*, 갔습니다 → *Gasseumnida*).
- **Auxiliary Verbs:** When an auxiliary verb follows a main verb (usually connected by a particle or ending), it is **separated** and **capitalised** (e.g., 하고 있다 → *Hago itda*).
- **Particles**: Always **separate** particles from the preceding noun/pronoun.
- **Hyphens**: Used for **personal names**, **honorifics**, **titles**, **specific prefixes**, and **Arabic numerals with counters**.

!!! warning Sound Change (Assimilation)
Rule Revised Romanisation dictates that spelling changes to reflect natural pronunciation changes between syllables.
- **Sinmun** (신문) not *Sin-mun*
- **Jongno** (종로) not *Jong-ro*
- **Halla-san** (한라산) not *Han-ra-san*
- **Gungmul** (국물) not *Guk-mul*
**Exception**: To preserve the meaning of words in a title, semantic boundaries in **separated words** do *not* undergo assimilation across the space (e.g., *Guk Mul* if written as two separate nouns, though rare).

!!!

## Actual Title Showing the Style

나 혼자만 레벨업: 헌터스 길드의 초대 ～이세계에서 귀환한 랭커가 다시 튜토리얼을 클리어하는 방법～
Na Honjaman Level Up: Hunters Guild ui Chodae ～Isekai eseo Gwihwanhan Ranker ga Dasi Tutorial eul Clear Haneun Bangbeop～

- **Loanwords**: 레벨업 → Level Up, 헌터스 → Hunters, 길드 → Guild, 랭커 → Ranker, 튜토리얼 → Tutorial, 클리어 → Clear.
- **Particles**: *ui*, *eseo*, *eul* are all separate and lowercase.
- **Assimilation**: *Bangbeop* (Method) follows RR rules.
- **Verbs**: *Gwihwanhan* (Returned) is fused; *Haneun* (Doing) is fused.
- **Noun + Verb spacing**: *Clear Haneun* is separated because "Clear" is a restored English loanword acting as the object noun phrase concept.
- **Punctuation**: Tilde `～` preserved for subtitles.

## Further Clarification

### Loanword Restoration Rule

Unlike Japanese, where loanwords are often heavily modified, Korean loanwords often map directly to English. To improve searchability and readability:

**Standard English Terms**: If a Hangul word corresponds to a common English word, use the English spelling.
- 버스 → **Bus** (not *Beoseu*)
- 던전 → **Dungeon** (not *Deonjeon*)
- 아카데미 → **Academy** (not *Akademi*)
- S급 → **S-Class** or **S-Geup** (prefer English *S-Class* if broadly understood contextually, otherwise romanise the counter).

**Unclear/Faux-English**: If the term is a "Konglish" abbreviation or unique concept without a direct English equivalent, romanise the Korean sound.
- 스킨십 → **Skinship** (Accepted Konglish)
- 알바 → **Alba** (German/Korean slang for part-time)

!!! tip Loanword Hybrid Components
If a compound consists of a Loanword + Native Particle/Noun:
- **Skill**-eul (Skill + particle) → **Skill eul**
- **Bus**-tago (Bus + riding) → **Bus Tago**
!!!

### Verbs and Adjectives

Korean verbs consist of a **Stem** + **Endings**. In titles, these are **FUSED**.

- **Dictionary Form**: 가다 (*Ga-da*) → **Gada**
- **Honorific/Formal**: 갑니다 (*Ga-pni-da*) → **Gamnida**
- **Past Tense**: 갔다 (*Gat-da*) → **Gatda**
- **Modifier FOrm**: 예쁜 (*Yeppeun*) → **Yeppeun** (Pretty), 먹는 (*Meokneun*) → **Meongneun** (Eating - *note the sound change k → ng).

!!! warning Auxiliary Verbs
When a sentence structure involves `Main Verb + Space + Auxiliary Verb`, romanise them as two separate capitalised words.
- 먹고 싶다 (want to eat) → **Meokgo Sipda**
- 해야 한다 (must do) → **Haeya Handa**
- 되어 버렸다 (ended up becoming) → **Doseo Beoryeotda**
!!!

### Hyphenation

#### Personal Names
Follow the pattern: **Surname** + **Given Name**. The Given Name is **hyphenated** with the second syllable in **lowercase**.
- 김민수 → **Kim Min-su**
- 박지민 → **Park Ji-min**
- 이재욱 → **Lee Jae-wook** (Common conventional spellings for surnames like Lee/Park/Choi are acceptable if the author is well-known, otherwise follow RR: *I*, *Bak*, *Choe*).

#### Honorifics
Honorifics and titles attached to names are **hyphenated** and **lowercase**.
- 민수씨 → **Minsu-ssi**
- 선생님 → **Seonsaeng-nim** (Teacher)
- 형님 → **Hyung-nim** / **Hyeong-nim** (Older brother - honorific)
- 누나 → **Nuna** (Older sister)
- 아저씨 → **Ajeossi** (Mister/Uncle - often lexicalised without hyphen if standalone, hyphenated if used as address).

#### Prefixes
Specific prefixes regarding numbering or size are hyphenated
- 제1화 → **Je-1-hwa** (Chapter 1)
- 고(故) → **Go-** (The late...)
- 대(大) → **Dae-** (Great/Big - only when acting as a distinct prefix, otherwise fused in compounds like Daehak (University)).

#### Arabic Numerals and Counters
Arabic numerals are hyphenated with their counters.
- 10년 → **10-nyeon** (10 Years)
- 3명 → **3-myeong** (3 People)
- 100번째 → **100-beonjjae** (100th)
- 24시간 → **24-sigan** (24 Hours)

#### Copulas (Ida / Anida)
The copula **Ida** (to be) and **Anida** (not to be) are unique. In standard RR, they are fused to the preceding noun. However, for readability in titles:
**Ida (이다)**: Fuse to the noun/particle **if it is the verb ending**.
- 학생입니다 → **Haksaengimnida** (It is a student)
- 남자이다 → **Namjaida**
**Anida (아니다)**: Usually behaves as a separate adjective requiring a subject particle.
- 학생이 아니다 → **Haksaeng i Anida**
**Future/Guessing (-gess-)**:
- 알겠습니다 → **Algesseumnida**

## Detailed Tables

### Particles
Particles must be **separated** from the word they modify and written in **lowercase**.

| Expression | Romanisation | Function |
| ---------- | ------------ | -------- |
| 이 / 가 | i / ga | Subject markers |
| 은 / 는 | eun / neun | Topic markers |
| 을 / 를 | eul / reul | Object markers |
| 의 | ui | Possessive particle (pronounced *e* often, but written **ui**) |
| 에 | e | Location / Time (at, in, to) |
| 에서 | eseo | From (location), At (action) |
| 로 / 으로 | ro / euro | Direction / Instrument (towards, by means of) |
| 와 / 과 | wa / gwa | And / With |
| 하고 | hago | And / With (colloquial) |
| 도 | do | Also / Even |
| 만 | man | Only |
| 부터 | buteo | From (time/place) |
| 까지 | kkaji | Until / To |
| 에게 / 한테 | ege / hante | To (a person) |

### Common Suffixes and Honorifics
| Hangul | Romanisation | Type | Usage Rule |
| ------ | ------------ | ---- | ---------- |
| 님 | -nim | Honorific | Hyphenate after name/title (e.g., *Sajang-nim*) |
| 씨 | -ssi | Honorific | Hyphenate after name (e.g., *Minsu-ssi*) |
| 군 | -gun | Title (Boy) | Hyphenate (e.g., *Kim-gun*) |
| 양 | -yang | Title (Girl) | Hyphenate (e.g., *Lee-yang*) |
| 들 | -deul | Plural | Hyphenate (e.g., *Saram-deul*) |
| 층 | -cheung | Floor | Hyphenate after Number (e.g., *2-cheung*) |
| 년 | -nyeon | Year | Hyphenate after Number (e.g., *2026-nyeon*) |
| 월 | -wol | Month | Hyphenate after Number (e.g., *5-wol*) |
| 일 | -il | Day | Hyphenate after Number (e.g., *15-il*) |

### Special Characters
- **Symbols (♥, ★, etc.)**: Keep as-is, with spaces around them unless stylisticaly fused on the cover.
- **Tildes (~)**: Keep as `～`.
- **Ellipsis**: Keep as is.
- **Brackets**: Convert `『 』`, `「 」`, `【 】` to standard double quotes `"` or single quotes `'` or square brackets `[ ]` depending on if they denote dialogue or technical stats.
> - System Prompts / Skills in brackets usually use `[ ]`.
> - Dialogue uses `"`.
___

## LLM Ruleset and Prompts

### YAML ruleset
```yaml
romanization_style:
  name: "MangaBaka Korean Style"
  version: "1.0"
  base_system: "Revised Romanization of Korean (RR)"
  loanword_rule: "restore_english"
  
  capitalization: >
    Capitalize all nouns, verbs, adjectives, adverbs, and auxiliary verbs.
    Lowercase all particles, suffixes (when hyphenated), and bound verb endings.

  loanwords: >
    If a Hangul word is a transliteration of a common English word (e.g., 택시, 버스, 헌터),
    romanize it using the English spelling (Taxi, Bus, Hunter).
    If it is obscure slang or a hybrid, adhere to RR or established Konglish (e.g., Skinship).

  particles: >
    Always separate grammatical particles from the preceding word with a space.
    Particles are always lowercase.
    List: eun, neun, i, ga, eul, reul, ui, e, eseo, ro, euro, wa, gwa, hago, do, man, buteo, kkaji.

  assimilation: >
    Apply strict RR sound assimilation rules for pronunciation changes within a word.
    (e.g., Shinmun instead of Shin-mun for 신문).
    Do not apply assimilation across spaces between separate words.

  verbs_and_adjectives: >
    Fuse verb stems with their endings (e.g., Gada, Gatda, Haneun).
    Separate auxiliary verbs from the main verb (e.g., Hago Sipda).
    Adjectives follow the same fusion rules as verbs.

  names: >
    Surname First (Capitalized) + Space + Given Name (Capitalized-Hyphen-Lowercase).
    Example: Kim Min-su.
    Exceptions permitted for famous authors with established alternative spellings (e.g., Park, Lee).

  hyphens: >
    Use hyphens for:
    - Honorifics (e.g., -nim, -ssi)
    - Plural marker -deul (optional, but preferred for clarity in titles: Saram-deul)
    - Arabic numerals + Counters (e.g., 10-nyeon, 1-level)
    - Prefixes like Je- (ordinal) or Go- (late).

  copula: >
    Fuse the copula 'ida' and its inflections to the preceding noun (e.g., Haksaengimnida).
    'Anida' is treated as a separate adjective (e.g., i Anida).

  example_validation:
    input: "나 혼자만 레벨업: 헌터스 길드의 초대"
    output: "Na Honjaman Level Up: Hunters Guild ui Chodae"
    explanation: "Particles 'ui' separated. Loanwords 'Level Up', 'Hunters', 'Guild' restored."
```


___
## Old Guidelines (Kept temporarily for internal reference. DO NOT USE)

## Overview

This public guide explains how Korean-language titles are romanised and styled for display on MangaBaka. The goal is consistent, readable, and publisher-independent romanisations that make titles clear and searchable for users.

**Canonical system:** **Revised Romanization (RR)** with a practical exception for loanwords (see Loanword Rule).

---

## Title field priorities

* **Native Title:** the original Hangul title as shown on the cover or source.
* **Romanization:** the romanised title shown to users (RR with the Loanword Rule applied).

---

## Subtitle & punctuation rules

* Follow MangaBaka's General Title Style Guide for English titles: use **colon (:)** and **em-dash (—)** to separate subtitle and sub-subtitle. Swap colon/em-dash if the subtitle contains a colon or for visual clarity.
* Replace straight quotes/apostrophes with typographic quotes only in English/translated fields. The display romanisation uses plain ASCII punctuation unless a decorative symbol is necessary.
* Native Hangul punctuation and spacing are kept exactly as presented in the native title.

---

## Romanization rules (Revised Romanization)

* Use **Revised Romanization (RR)** as the standard for Korean romanisation.

> Examples: 학교 → **Hakgyo**; 서울 → **Seoul**; 김민수 → **Kim Min-su**.

* Preserve Hangul word segmentation when producing romanisation (i.e., keep spaces that reflect Korean word breaks), except for particles that are separated regardless of the spacing (see next bullet for examples).

> Example: 우리들의 평화로운 일상 → **Uri-deul ui Pyeonghwaroun Ilsang**

* Particles (이/가, 은/는, 을/를, 에서, 로/으로, 와/과, 도, 의) should be shown as separate tokens and written in lowercase for readability in the romanised title.

> Example: 나는 학교에 간다 → **Na neun Hakgyo e Ganda**

* Double consonants and aspiration follow RR mapping (kk, tt, pp, ss, jj). e.g., 학교 → **Hakgyo**, 빵 → **Ppang**.

---

## Loanword Rule

* When a Hangul token clearly represents a foreign loanword and there is a common real-world/English form, prefer the real-world term in the **romanisation**.

> Examples: 택시 → **Taxi** (not "Taeksi"); 버스 → **Bus** (not "Beoseu"); 스킬 → **Skill** (not "Seukil"); 바스켓볼 → **Basketball** (not "Baseukesbol").

* The native Hangul title remains unchanged in the native field; the romanised display uses the real-world term so users immediately recognise the concept.

---

## Names & honorifics

* Use family name first (capitalized), then given name: 김민수 → **Kim Min-su**.
* Honorifics and titles attach with a hyphen: **-ssi**, **-nim**, **-seonsaeng**.

> Example: 박지민씨 → **Park Ji-min-ssi**

---

## Hyphenation and counters

* Use hyphens for name+honorific, number+counter, and ordinal/part markers when it helps clarity and sorting:

> 10년 → **10-nyeon**
> 제3부 → **Je 3-bu**

* For lexicalized compounds, prefer splitting unless fusion improves readability (editorial judgment applies).

---

## Capitalization

* When an official English/romanised capitalization exists and is canonical, use it.
* For fan-romanised titles, apply **Chicago-style title capitalization**: capitalize nouns, verbs, adjectives, adverbs, and main auxiliaries; keep particles and grammatical endings lowercase.

---

## Examples

* Native: 택시 타는 날

> Romanization: **Taxi Taneun Nal**

* Native: 학교에 간 고양이

> Romanization: **Hakgyo e Gan Goyang-i**

* Native: 김선생님과의 하루

> Romanization: **Kim Seonsaeng-nim gwa ui Haru**

---

## Notes for editors

* Prefer clarity and recognizability for readers. Use the Loanword Rule when loanwords are obvious.
* Do not rely on publisher romanisation; publisher styles are often inconsistent.
* Keep the native Hangul title intact and visible in the metadata.