Palatalization (구개음화) is the rule that turns the everyday word 같이 ("together") into [가치] rather than the "gat-i" the spelling promises. In one line: when a batchim ㄷ or ㅌ is followed by the suffix vowel 이 (or the suffix 히), the alveolar stop slides forward in the mouth and becomes the palatal affricate ㅈ or ㅊ. The tongue is already heading up toward the [i] position, and dragging the ㄷ/ㅌ along with it is easier than making the stop and then jumping to the vowel. Physiologically it is the same instinct behind English did you → "didja" and got you → "gotcha."
- ㄷ + 이 → [지]
- ㅌ + 이 → [치]
The core cases
주말에 같이 영화 볼래요?
jumare gachi yeonghwa bollaeyo
Want to watch a movie together this weekend? (같이 → [가치])
굳이 사과할 필요는 없어요.
guji sagwahal piryoneun eopseoyo
You really don't need to apologize. (굳이 → [구지])
새해 첫날 해돋이를 보러 갔어요.
saehae cheonnal haedojireul boreo gasseoyo
We went to watch the sunrise on New Year's Day. (해돋이 → [해도지])
이 채소는 밭이 아니라 화분에서 키웠어요.
i chaesoneun bachi anira hwabuneseo kiwosseoyo
I grew this vegetable in a pot, not in a field. (밭이 → [바치])
저는 삼 남매 중 맏이예요.
jeoneun sam nammae jung majiyeyo
I'm the eldest of three siblings. (맏이 → [마지])
In 같이 the ㅌ batchim of 같 palatalizes before the adverb-forming 이 → [가치]. In 굳이 the ㄷ of 굳 palatalizes → [구지]. 해돋이 (해 "sun" + 돋- "rise" + 이) gives [해도지], and the subject-marked 밭이 ("the field [subject]") gives [바치]. Revised Romanization reflects the change directly: gachi, guji, haedoji, bachi.
It also fires on derived verbs, where the 이 is a transitivizing/passivizing suffix rather than a particle:
여기에 우표를 붙이면 돼요.
yeogie upyoreul buchimyeon dwaeyo
You just stick a stamp on here. (붙이다 → [부치다])
굳힌 초콜릿을 냉장고에서 꺼냈어요.
guchin chokolliseul naengjanggoeseo kkeonaesseoyo
I took the set chocolate out of the fridge. (굳히다 → [구치다])
The ㅎ twist: ㄷ + 히 palatalizes too (via aspiration)
When the suffix is 히, watch two rules chain. First the ㄷ and the ㅎ fuse by ㅎ-aspiration into ㅌ; then that fresh ㅌ meets the 이 and palatalizes to [치]. The output looks like the stop simply became [치]:
문이 바람에 세게 닫혔어요.
muni barame sege dacheosseoyo
The door slammed shut in the wind. (닫히다 → [다치다])
Trace 닫히다: 닫 (ㄷ) + 히 → the ㄷ and ㅎ aspirate to ㅌ ([다티]) → the ㅌ palatalizes before 이 → [다치다]. Two rules, one smooth result. 굳히다 works identically: 굳 + 히 → [구티] → [구치다]. This is why RR writes dachida and guchida — the reading aid already bakes both steps in.
The boundary condition everyone else skips
Here is the detail that separates a good explanation from a great one, and it is exactly where most textbooks stop too soon. Palatalization is not a free-floating "ㄷ/ㅌ before 이 always changes" rule. It fires only across a morpheme boundary — that is, only when the 이/히 is a separate grammatical piece (a particle, an adverb-former, a verb suffix) attached to a stem that ends in ㄷ/ㅌ. When ㄷ/ㅌ and 이 live inside the same root morpheme, absolutely nothing happens:
잔디밭에 들어가지 마세요.
jandibate deureogaji maseyo
Please keep off the grass. (잔디 stays [잔디], never [잔지])
이 다큐멘터리는 끝까지 견디기 힘들었어요.
i dakyumenteorineun kkeutkkaji gyeondigi himdeureosseoyo
This documentary was hard to sit through to the end. (견디다 stays [견디다])
잔디 ("lawn/grass") is one indivisible morpheme, so its internal ㄷ + 이 stays [잔디] — never ×[잔지]. 견디다 ("to endure") likewise keeps its ㄷ: [견디다]. Compare that with 굳이 [구지], where 굳- is a stem and 이 is a separate suffix — the seam is what licenses the shift. The contrast is the whole rule:
| Word | Structure | Boundary? | Pronunciation |
|---|---|---|---|
| 같이 | 같 (stem) + 이 (suffix) | Yes | [가치] |
| 굳이 | 굳 (stem) + 이 (suffix) | Yes | [구지] |
| 밭이 | 밭 (noun) + 이 (particle) | Yes | [바치] |
| 잔디 | 잔디 (one morpheme) | No | [잔디] |
| 마디 | 마디 (one morpheme) | No | [마디] |
| 티끌 | 티끌 (one morpheme) | No | [티끌] |
손가락 마디가 다 아파요.
songarak madiga da apayo
Every one of my knuckles aches. (마디 stays [마디])
책상 위에 먼지랑 티끌이 가득해요.
chaeksang wie meonjirang tikkeuri gadeukhaeyo
The desk is covered in dust and specks. (티끌 stays [티끌])
마디 ("joint/knuckle") and 티끌 ("speck") are single roots, so their ㄷ/ㅌ never palatalizes. This is the crucial asymmetry: the seam, not the [i], is what triggers the change.
Why English speakers get it wrong in both directions
English spelling makes learners want to read each consonant letter literally, so the first error is under-applying: reading 같이 as "gat-i." But once the rule clicks, a second, opposite error appears — over-applying it to every ㄷ/ㅌ + 이, mangling 잔디 into ×[잔지] or 티끌 into ×[치끌]. Both mistakes come from missing the boundary condition. The fix is a single habit: palatalization needs a morpheme seam. With a seam, trust the sound over the spelling (같이 = [가치]); without one, trust the spelling (잔디 = [잔디]).
Common Mistakes
1. Reading 같이 literally as "gat-i." The single most common palatalization error in the language. It is always [가치].
- ✗ 같이 said [가티] → ✓ [가치]
2. Over-applying inside a root. Because 같이 → [가치], learners "fix" every ㄷ/ㅌ + 이.
- ✗ 잔디 said [잔지] → ✓ [잔디]
- ✗ 견디다 said [견지다] → ✓ [견디다]
- ✗ 마디 said [마지] → ✓ [마디]
3. Missing the ㅎ-suffix chain. With 히, learners stop after aspiration ([다티]) and forget the second step.
- ✗ 닫히다 said [다티다] → ✓ [다치다]
- ✗ 굳히다 said [구티다] → ✓ [구치다]
4. Forgetting palatalization with a subject particle. 밭이 is not exempt just because 이 is a case marker.
- ✗ 밭이 said [바티] → ✓ [바치]
Key Takeaways
- 구개음화 / palatalization: a stem-final ㄷ/ㅌ becomes [ㅈ]/[ㅊ] before a suffix 이 or 히 — 같이 [가치], 굳이 [구지], 해돋이 [해도지], 밭이 [바치], 붙이다 [부치다].
- With 히, aspiration fires first, then palatalization: 닫히다 [다치다], 굳히다 [구치다].
- The rule needs a morpheme seam. Inside a single root there is no change: 잔디 [잔디], 마디 [마디], 티끌 [티끌], 견디다 [견디다].
- The trigger is the seam, not the vowel — so avoid both errors: don't read 같이 as [가티], and don't turn 잔디 into [잔지].
- Revised Romanization reflects palatalization, so the reading is on your side:
gachi,guji,dachida.
Now practice Korean
Reading grammar gets you part of the way. The exercises are where it sticks — free, no signup needed.
Start learning Korean→Related Topics
- Aspiration 격음화: ㅎ + Plain Stop → Aspirated (좋다 → 조타)TOPIK 1 — 격음화: ㅎ and an adjacent plain stop or affricate fuse into a single aspirated consonant, in either direction — 좋다 [조타], 축하 [추카], 입학 [이팍] — a change that Revised Romanization actually shows, unlike tensification.
- Liaison 연음: Batchim Moves to the Next SyllableTOPIK 1 — The highest-frequency Korean sound rule: when a syllable ends in a batchim and the next begins with a vowel (the silent ㅇ), the final consonant slides forward to become that syllable's onset. Spelling keeps morpheme boundaries visible, but speech relinks right across them — so you glide, never pause, and a neutralized final is restored to its true value when it links.
- The Sound Inventory & the Seven Final ConsonantsTOPIK 1 — The map for the whole Pronunciation group: Korean's 19 consonants built on a three-way plain/tense/aspirated contrast that is NOT English voicing, its vowel system, and the master fact behind every sound-change page — in final (받침) position only seven sounds survive, so spelling and pronunciation systematically diverge.
- ㄴ-Insertion at Compound Boundaries (한여름 → 한녀름)TOPIK 2 — Why a [ㄴ] appears out of nowhere at a compound seam — 한여름 [한녀름], 담요 [담뇨], 꽃잎 [꼰닙], 십육 [심뉵] — whenever the first part ends in a consonant and the second begins with 이/야/여/요/유/예. It targets the seam, so 십육 is [심뉵] but 육 alone is [육], and it is partly lexical (담요 [담뇨] but 금요일 [그묘일]).