ㄹ-Irregular Predicates (ㄹ 탈락): Full Table

Every Korean verb or adjective whose stem ends in ㄹ — 살다 "live," 알다 "know," 만들다 "make," 팔다 "sell," 열다 "open," 놀다 "play," 울다 "cry," 길다 "be long" — belongs to a single pattern that textbooks file under "irregular" but which is, in truth, the most regular class in the language. This page is the lookup sheet: one table that shows, for the eight most common ㄹ-stems, exactly where the ㄹ vanishes and where it stays. Learn the trigger set once and you never conjugate a ㄹ-stem by guesswork again.

The one rule, and its silent companion

Two facts govern the whole class, and they always run together.

1. The stem ㄹ drops before four sounds: ㄴ · ㅂ · ㅅ · 시. That is the classic Korean mnemonic — the drop fires before the formal -ㅂ니다, the honorific -(으)시-/-세요, the causal -(으)니까, and the present modifier -는 and past modifier -(으)ㄴ (all ㄴ-initial). Before anything else — a vowel, ㄱ, ㅁ, ㄷ, ㅈ — the ㄹ holds.
2. A ㄹ-stem never inserts the buffer 으. Where an ordinary consonant stem like 먹다 takes the "heavy" allomorphs (먹으세요, 먹으니까), a ㄹ-stem takes the light, vowel-stem ones (사세요, 사니까) — as if the stem had no batchim at all.

The full table

Columns run left to right through the four drop environments, then close with the -아/어요 form where the ㄹ survives. Cells give the Hangul over its Revised Romanization.

A few cells reward a second look. The formal column is where the ㄹ-drop is most invisible and most often botched: 살 + ㅂ니다 loses the ㄹ and the ㅂ becomes the batchim → 삽니다 (pronounced [삼니다], because the ㅂ then nasalizes to [m] before the following ㄴ). And 길다 is a descriptive verb (adjective): it has a formal 깁니다 and a modifier 긴, but no plain honorific command 세요 and no propositive — you cannot order a state to be long. That is why its 세요 cell is empty.

Where the ㄹ survives: vowels, ㄱ, ㅁ, ㄷ, ㅈ

The other half of the sheet is the forms where nothing drops. Before the polite -아/어요, the connective -고, the conditional -(으)면 (whose surface consonant is ㅁ, not a trigger), and the retrospective -던, the ㄹ stays put.

The future keeps exactly one ㄹ

The prospective ending -(으)ㄹ begins with ㄹ itself, so a ㄹ-stem meeting it does not double the letter: the stem ㄹ and the ending ㄹ merge into a single 살, 만들. The everyday future -(으)ㄹ 거예요 therefore shows one ㄹ, spelled just like the plain stem.

Why English speakers should stop calling this "irregular"

Here is the reframe that fixes the class for good. A genuine irregular — the ㅂ, ㄷ, ㅅ, 르, ㅎ classes — is lexical: some verbs follow the pattern and other identically-spelled ones do not, so you memorize membership word by word (돕다 → 도와요 but 잡다 → 잡아요). The ㄹ-drop is nothing like that. There are effectively no "regular" ㄹ-final stems to contrast against — every ㄹ-stem drops in exactly the same four environments, without a single exception. It is a phonological reflex, closer to English "a → an before a vowel" than to "sing → sang."

So why does it still trip people? Because the change is invisible in the dictionary form. You look up 살다 and see the ㄹ sitting there, so your instinct is to keep it — and the ㄹ then wrongly survives in precisely the high-frequency forms where it should vanish: the formal 삽니다 and the honorific 사세요. The fossilized errors ×살습니다 and ×알으세요 come from trusting the spelling instead of the rule.

Common Mistakes

1. Inserting the 으 buffer. ㄹ-stems never take 으; the ㄹ drops instead.

2. Keeping the ㄹ (and using -습니다) in the formal form. The ㄹ drops and the ending is -ㅂ니다.

3. Keeping the ㄹ before the causal -(으)니까. ㄴ is a trigger, so 만드니까, not ×만들으니까.

4. Keeping the ㄹ before the modifier -는. ㄴ is a trigger, so 파는, not ×팔는.

Key Takeaways

• The stem ㄹ drops before ㄴ · ㅂ · ㅅ · 시: 삽니다, 사세요, 사니까, 사는, 산 — and stays before vowels, ㄱ, ㅁ, ㄷ, ㅈ (살아요, 살고, 살면, 살던).
• A ㄹ-stem never inserts 으: 사세요 / 사니까, never ×살으세요 / ×살으니까.
• The future keeps one ㄹ: 살 거예요, 만들 거예요 (stem ㄹ + ending ㄹ merge).
• Watch -면 vs -니까: 살면 keeps the ㄹ, 사니까 drops it.
• This is a fully rule-governed elision with no membership list — the most predictable class of all, not a true irregular.