Korean Particles You Only Learn by Speaking
은/는 versus 이/가 and the other particle choices that textbooks explain well and learners still get wrong in conversation — and how to drill them.
Particles are the most explained and least mastered part of Korean. Nearly every learner can recite that 은/는 is the topic marker and 이/가 is the subject marker. Far fewer can choose correctly at conversational speed, because the choice is not really grammatical — it is about what you are treating as known, contrasted, or newly introduced.
That is not a knowledge problem. It is a production problem, and production problems are solved by reps with correction.
The four that cause the most trouble
은/는 versus 이/가
The textbook rule is topic versus subject. The usable version is closer to this: 이/가 introduces or identifies something new or answers "which one?", while 은/는 takes something already in play and comments on it, often with a hint of contrast.
- 고양이가 있어요. — There is a cat. (introducing it)
- 고양이는 자요. — The cat is sleeping. (it is already the topic)
- 저는 커피요. — I'll have coffee. (implicitly contrasted with what others ordered)
The last one is the pattern learners under-use. 는 doing contrastive work is extremely common in speech and almost never drilled.
을/를 and when it disappears
Object marking is straightforward on paper and routinely dropped in casual speech. 밥 먹었어? is far more natural than 밥을 먹었어? in a relaxed register. Learners who mark every object sound stiff; learners who drop them in formal contexts sound careless. Both directions need exposure, not rules.
에 versus 에서
Both translate as "at" or "in", which is why they transfer badly. The distinction is roughly static location or destination (에) versus the place an action happens (에서).
- 학교에 가요. — I go to school. (destination)
- 학교에서 공부해요. — I study at school. (action location)
- 집에 있어요. — I'm at home. (existence, not action)
That third case is where the rule of thumb breaks for many learners: 있다 takes 에, not 에서, despite feeling like a location.
도, 만, 부터, 까지
These replace or stack with other particles and change meaning sharply. 저도 versus 저만 is the difference between "me too" and "only me". Learners often know them passively and never produce them, which flattens speech.
Why reading about them does not fix them
Particle choice happens in under a second, before the rest of the sentence. Any rule you have to consciously apply is too slow for conversation, so the goal is not understanding — it is automaticity.
Automaticity comes from producing the same structures repeatedly and being corrected each time. This is the specific thing that is awkward to ask a human tutor for and trivial in an app: in K-TutorCall, roleplay practice can be repeated without a cap, and after each attempt your sentences come back corrected one by one.
A drilling method that works
- Pick one particle pair per week. Do not drill all of them at once — you will not notice which correction is which.
- Choose a roleplay where that pair recurs. Directions and travel scenarios force 에/에서. Ordering and introductions force 은/는 versus 이/가.
- Run the same scenario daily, varying details. Different order, different destination, same structures.
- Read the corrections every time and specifically look for the particle you are drilling. Ignore the rest that week.
- Move on when you stop being corrected on it — not when you feel confident, which arrives earlier than accuracy does.
The daily challenge is useful here too: it is a short weak-spot drill, about three minutes, which is the right size for a single-particle focus.
What "getting it" feels like
You will not reach a point where you consciously know the rule better. You will notice that a sentence sounds wrong after you say it — and later, that the wrong particle stops arriving in the first place. That lag between noticing and preventing is normal and takes weeks of frequent practice, not days.
FAQ
Is there a simple rule for 은/는 versus 이/가?
No rule covers every case, which is why it stays hard. The most useful approximation is that 이/가 introduces or identifies, and 은/는 comments on something already in play or sets up a contrast. Beyond that it is exposure and correction.
Should I mark every particle when speaking?
In formal or careful speech, yes. In casual speech, object markers especially are often dropped and over-marking sounds unnatural. Practicing in both registers is why roleplay scenarios vary the listener.
Can I drill just particles?
Effectively, yes — by choosing roleplay scenarios that force the structures you are targeting and reviewing corrections with that one focus. Trying to fix everything at once is why particle errors persist for years.
Do particles matter if people understand me anyway?
Understanding is not the bar for most learners' goals. Particle errors rarely block meaning but consistently mark speech as non-native, and some — like 은/는 contrast — change the implication of what you said rather than just the polish.