が: The Subject Marker

が is the particle that marks the grammatical subject — the one who does the action or the thing that exists or has a quality. If は tells you what a sentence is about, が tells you who or what is actually doing or being something. But the label "subject marker" undersells it. The real work が does in a conversation is presenting information as new: it spotlights something the listener could not have predicted, and says "here — this is the one."

You blurt this out the moment the bus appears. The bus is new to the conversation — it just entered the scene — so it takes が, not は. This "just appeared / just noticed" flavour is the heart of が, and everything else on this page grows out of it. (There is no pronunciation trap here, by the way: が is simply read ga, unlike its cousin は.)

New information and neutral description

The most basic use of が is to report something as a fresh, whole event — what grammarians call neutral description. Nothing is assumed; you are describing a situation from scratch, all of it new.

In every case, the が-marked noun is the newsworthy part. Compare English: we would say "the phone is ringing," but that definite "the" is misleading — the phone is being introduced, not referred back to. English marks new-vs-old mostly through articles (a/the); Japanese has no articles and does the same job with は vs が. That parallel is the fastest way in for an English speaker, and it is developed fully on the は vs が page.

が answers "who?" and "what?"

Here is the most reliable rule on this page, and one you can lean on with total confidence: a question word like 誰 (who), 何 (what), or どれ (which) always takes が, and so does the answer to it. A question word is by definition brand-new, un-guessable information — the exact thing が exists to mark. は is impossible here.

Notice how the answer 田中さんが来ました keeps が. This is because 田中さん is precisely the piece of information the question was hunting for — the new, identifying part. If you answered ×田中さんは来ました, it would sound as if you were changing the subject to talk about Tanaka, not answering "who." This question–answer が is one of the clearest, most testable signals of when to reach for が instead of は.

Exhaustive listing: "it is X (and only X)"

が has a sharper cousin of the "new information" use called exhaustive listing. When you put が on a noun in answer to an implicit "which one?", you are saying this one, to the exclusion of all others.

私がやりました is not the neutral "I did it" — it singles you out: of everyone here, I am the one. It is the natural sentence for owning up to a mistake, or for volunteering. Swap in は — 私はやりました — and the meaning flips to "I did it (whatever the others did)," a contrastive statement about you. This split between neutral-description が and exhaustive-listing が is subtle, and it is one of the edge cases the choosing guide drills.

が with existence: いる and ある

The verbs of existence — いる (for animate things: people, animals) and ある (for inanimate things and abstractions) — take their subject with が. This is where beginners meet が constantly, because "there is a ~" sentences are everywhere.

The logic is identical to neutral description: you are introducing the dog, the question, the car into the conversation for the first time. Existence is new information by nature — you cannot presuppose something you are announcing exists.

The が that surprises English speakers: stative predicates

Now the pattern that trips up every English speaker, because it defies English intuitions about "objects." A whole class of Japanese predicates — describing liking, ability, understanding, wanting, skill, and perception — mark the thing they are about with が, not を, even though English makes that thing a direct object.

Every English translation there has a direct object — apples, Japanese, the piano, water — so the reflex is to reach for the object particle を. Resist it. These predicates (好き "likeable," 分かる "understandable," できる "possible," 〜たい "want to," 上手 "skilled at," 見える "visible," 聞こえる "audible") are grammatically stative: they describe a state in which the person stands relative to the thing. The trick is to read りんごが好きです not as "I like apples" but as "apples are pleasing [to me]." Once you see the thing as the subject of a state — apples-are-pleasing, Japanese-is-understandable, water-is-want-worthy — the が falls into place naturally. This pattern is important and systematic enough to get its own page: が with 好き, ほしい, できる, 分かる.

が inside modifying clauses

Because は is banned from noun-modifying (relative) clauses, が becomes the only ordinary subject marker inside them. Whenever you build a clause that describes a noun, its internal subject takes が.

Inside 母が作った ("[that] my mother made"), the subject 母 takes が. The whole clause then modifies ケーキ, and ケーキ is the topic of the outer sentence, so it takes は. This layering — が for the inner subject, は for the outer topic — is the normal shape of a complex Japanese sentence. In these modifying clauses が can even be swapped for の, a quirk covered on the の replacing が page.

Common mistakes

The recurring theme: reach for が whenever the noun is new, identifying, or the answer to a wh-question, whenever it is the subject of an existence verb or a stative predicate (好き, 分かる, できる, 〜たい, 上手), and whenever it is the subject inside a modifying clause. Everywhere else, weigh it against は — which is exactly what the next page does.

Key takeaways

• が marks the grammatical subject, and its discourse job is presenting information as new or identifying.
• A question word (誰, 何, どれ) and the answer to it always take が — the single most reliable が rule.
• Existence verbs いる／ある take が.
• Stative predicates (好き, 分かる, できる, 〜たい, 上手, 見える, 聞こえる) mark their "object" with が because the thing is really the subject of a state.
• Inside a modifying clause, the subject takes が (or の) — は is forbidden there.