は vs が in Japanese: A Beginner's First Pass

は vs が in Japanese is often the first comparison learners meet. The two particles can appear in what looks like the same slot in an English translation, but they do categorically different jobs in the Japanese sentence.¹²³ This article sorts the two roles, walks through the canonical question-and-answer pair every beginner needs, and sets clear limits on the most useful heuristic.

Overview

The one-sentence answer

は marks the topic (what the sentence is about). が marks the grammatical subject (the argument the predicate agrees with).¹² English packs both jobs into a single "subject" slot. That is why English speakers learning Japanese often look for one "right answer" between は and が, even though each marker fills a slot the other does not.³²

Both particles are introduced in lesson 1 of standard N5 textbooks, and the distinction between them is the most-discussed beginner question in the early curriculum.⁴⁵

私わたしは学生がくせいです。⁴
"I'm a student."

誰だれが来きましたか。¹
"Who came?"

Why this article is a first pass

At the upper-intermediate level, the full は/が problem breaks into at least four sub-problems: thematic versus contrastive は, neutral-description versus exhaustive-listing が, は in negation scope, and が in subordinate clauses.¹⁶ Each topic is its own teaching unit and has a dedicated article on the は particle or the が particle.

A beginner cannot resolve those sub-problems without first understanding the topic-versus-subject frame. The literature treats that frame as the prerequisite, not the conclusion.²³ The multiple-subject construction, especially the canonical 象は鼻が長い shape, also belongs in the dedicated topic-versus-subject orientation article because it shows that は and が are not competing for one slot.⁷¹

象ぞうは鼻はなが長ながい。⁷
"Elephants have long trunks."

What this article defers

Contrastive は, exhaustive-listing が, は inside negation, が inside subordinate clauses, and the multiple-subject 象は鼻が長い shape are named here only briefly. The dedicated は-particle, が-particle, and topic-versus-subject articles handle the full treatment.

JLPT level and prerequisites

Both particles are N5. The topic marker は appears in Genki I lesson 1 and Minna no Nihongo I lesson 1. The subject marker が appears by Genki I lesson 4 and Minna no Nihongo I lessons 9 and 10 in stative-predicate constructions with 好き, わかる, and ある/いる.⁴⁵ The prerequisite shape is the copula sentence [X は Y です], introduced in lesson 1 of both textbooks.⁴⁵

The article also assumes the label "topic-prominent language," a cross-linguistic term for languages, including Japanese, that make topics especially important.³ If you have not seen that concept yet, read the dedicated topic-versus-subject orientation article first.

今日きょうは月曜日げつようびです。⁴
"Today is Monday."

私わたしはコーヒーが好すきです。⁸
"I like coffee."

The two jobs each particle does

は marks the topic

Makino and Tsutsui define は as "a particle which marks a topic or a contrasted element," with the standard English gloss "as for X" or "speaking of X" before the natural translation.⁸ Kuno's analysis frames thematic は as marking a constituent that "stands for a theme." In this use, the theme must be generic or anaphoric, meaning already in the discourse or universally known.¹

The topic sets the frame. The comment that follows is the claim made about that frame.¹⁹

山田やまださんは大学生だいがくせいです。⁸
"Mr. Yamada is a university student."

犬いぬは動物どうぶつです。⁸
"Dogs are animals."

When は acts as the topic particle, it is pronounced wa. This is a kana-spelling exception: it preserves pre-1946 historical spelling and was codified in the 1986 Cabinet Notification on Modern Kana Usage. It is one of three particle-spelling exceptions, alongside を read o and へ read e.¹⁰ The Good to know section returns to this spelling exception in more detail.

は appears across all registers: plain (だ), polite (です/ます), and formal written prose. It is not limited to one register.⁸¹¹

が marks the grammatical subject

が is the nominative case particle. When it is overt, it marks the argument the predicate agrees with: the doer of the verb or the holder of the property.²¹ Shibatani classifies が at the syntactic case-marking level and は at the higher information-structural level. In other words, the two particles answer different questions rather than competing for the same slot.²

Kuno identifies two main uses of が: neutral-description が, which reports an event as news, and exhaustive-listing が, which asserts "X and no other."¹ This first pass treats only が as the subject case marker. The two sub-uses are deferred to the dedicated が-particle article.¹⁶

田中たなかさんが来きました。⁵
"Mr. Tanaka came."

机つくえの上うえに本ほんがあります。⁵
"There is a book on the desk."

Stative predicates, which describe states rather than actions, include 好き, 嫌い, 上手, 下手, わかる, できる, ある, and いる. In their textbook frames, they canonically take a が-marked object or holdee. Genki I introduces this pattern in lesson 5 with 好き and 嫌い, and Minna no Nihongo I in lessons 9 and 10.⁴⁵

コーヒーが好すきです。⁴
"I like coffee."

が is the standard case marker across registers. It appears identically in polite (来ました) and plain (来た) sentences.⁸¹¹

Why English crams them into one slot

Li and Thompson group languages by two independent axes: how prominent the subject is, and how prominent the topic is. English is usually subject-prominent. It requires every clause to have a subject, even when that means inserting a dummy "it." Japanese is usually topic-prominent, with a dedicated topic slot and no requirement for dummy subjects.³

A topic-prominent language can express the same proposition without an English-style subject. Japanese 雨が降っている presents a thetic "rain-falling" event, meaning the whole event is introduced as news. It has no equivalent of the English dummy "it."³²

雨あめが降ふっている。¹
"It's raining."

Because English forces every sentence into a [SUBJECT VERB OBJECT] mold, both "what we are talking about" and "who does the verb" surface in the single English subject slot. A learner who reads only the English translation cannot see the Japanese contrast. That is why the topic-versus-subject orientation article is the prerequisite to this one.³²

私わたしは頭あたまが痛いたい。¹
"I have a headache."

日本にほんは山やまが多おおい。²
"Japan has a lot of mountains."

The canonical contrast pair

誰が来ましたか → ジョンが来ました

A question-word subject forces が because a question word is neither generic nor anaphoric, so it fails the thematic-は constraint.¹⁸ The answer to a question-word-が question copies the case-marking pattern. The answering noun is the new piece of information, the value that fills the wh-slot, so it also takes が.¹⁶

Q: 誰だれが来きましたか。 A: ジョンが来きました。¹
Q: "Who came?" A: "John came."

This pattern is the most reliable single rule at N5: a wh-subject question is answered with が every time. A beginner will not encounter main-clause exceptions.¹⁸

Q: 何なにがありますか。 A: 本ほんがあります。⁵
Q: "What is there?" A: "There's a book."

ジョンは何をしましたか → ジョンは本を読みました

Once a referent has been introduced, here by the previous question-and-answer turn, it is anaphoric and can satisfy the thematic-は constraint. That is why it switches to は in the next turn.¹⁹ The new information has shifted from "who" to "what action." The comment now carries the news, while the established referent sits in the topic slot.¹⁶

Q: ジョンは何なにをしましたか。 A: ジョンは本ほんを読よみました。⁸
Q: "What did John do?" A: "John read a book."

This new-to-old shift across consecutive turns is also the textbook pattern for the canonical Japanese folktale opener. Kuno uses it to illustrate the same generalization at the discourse level.¹ Once the topic is set, maintaining it with は is the default behavior in spoken Japanese discourse. Restating the referent with が a second time would re-introduce it as new and reads as marked.⁹¹

昔々むかしむかし、おじいさんがいました。おじいさんは山やまへ行いきました。¹
"Long ago, there was an old man. The old man went to the mountain."

What the pair teaches in one line

が introduces a referent into the discourse. は continues it across later turns.¹ Kuno states this as the asymmetry between neutral-description が, which reports new information, and thematic は, which operates on already-established or generic referents.¹ Heycock restates the same generalization in information-structure terms: が-marked subjects can sit inside the focus of a sentence, while は-marked nominals sit in the background.⁶

The flow across the two turns of the canonical pair, and across the folktale opener, is the same shape:

友達ともだちが来きました。友達ともだちは本ほんを持もっていました。⁵
"A friend came. The friend was holding a book."

This single rule covers a large majority of N5 textbook dialogues. The exceptions are contrastive は in new-information contexts and exhaustive が with old referents. They are upper-N4 and N3 material and live in the dedicated articles.¹⁶

The "new vs. already in the conversation" heuristic

The heuristic stated plainly

If the noun is new to this conversation, mark it with が. If it is already in the conversation, universally known, or about to be talked about as a frame, mark it with は.⁸⁴ This heuristic is a teaching simplification of Kuno's two-condition rule for thematic は: the marked nominal must be either anaphoric or generic.¹

A: あ、猫ねこがいる。 B: その猫ねこは誰だれのですか。⁵
A: "Oh, there's a cat." B: "Whose cat is that?"

犬いぬは動物どうぶつです。⁸
"Dogs are animals."

The heuristic correctly predicts N5-level main-clause sentences in roughly the cases the textbooks treat. It begins to fail at the edges, especially contrastive は and exhaustive が. That is why this article sets its limits explicitly below.¹⁶

Question words always take が

Interrogative subjects, such as 誰, 何, どれ, どちら, and いつ, always take が in main clauses when they are in subject position with a stative predicate.¹⁸ The rule follows from Kuno's generalization that themes must be anaphoric or generic. A question word can satisfy neither condition.¹ The answer to a question-word-が question also takes が, because the answer noun is the new value that fills the wh-slot.¹⁶

Question-word subjects never take は

In a main clause, a question-word subject takes が, full stop. The form *誰は来ましたか is ungrammatical because 誰 is neither anaphoric nor generic, so it cannot satisfy the thematic-は condition.¹⁸ Genki I and Minna no Nihongo I introduce the pattern with explicit "who/what + が" templates for exactly this reason.⁴⁵

誰だれが日本語にほんごを話はなしますか。⁴
"Who speaks Japanese?"

どれがあなたの本ほんですか。⁴
"Which one is your book?"

どちらが好すきですか。⁴
"Which (of the two) do you like?"

Where the heuristic stops working

The new-versus-old heuristic fails in at least four well-documented constructions. Each one is named here briefly and routed to its dedicated article.

1. Contrastive は can attach to a new referent when the contrast frame is what matters; treated in the dedicated は-particle article.¹
2. Exhaustive-listing が can attach to an old referent when the speaker is asserting "X and no other"; treated in the dedicated が-particle article.¹⁶
3. は inside subordinate clauses disappears, so a topical noun reverts to が; treated in the dedicated は-particle article.¹
4. Multiple-subject sentences like 象は鼻が長い split topic and subject between two nouns; treated in the topic-versus-subject orientation article.⁷¹

Heycock summarizes the cumulative effect: a focus-marker analysis of が works for exhaustive listing but fails for neutral description. An information-structure analysis is needed to cover both uses uniformly.⁶ This framing matters at N4 and above. At N5, the right move is to recognize that the simple heuristic has known edges, not to import the deeper machinery yet.

ジョンが一番いちばん速はやいです。¹
"John is the fastest."

私わたしが買かった本ほんは高たかかった。²
"The book I bought was expensive."

Quick-reference table

Side-by-side: when each particle wins at N5

The eight rows below cover the N5-stable cases. Contrastive は and exhaustive が are deliberately excluded because they belong in the downstream dedicated articles.¹⁶

Usage condition	Particle (at N5)	Example	Why
The noun is already in the conversation or generic	は	犬は動物です。8	Anaphoric or generic referent satisfies the thematic condition.1
The noun is being introduced for the first time	が	昔々、おじいさんがいました。1	Neutral-description が reports the existence of a new referent.1
The answer to a question word in subject position	が	誰が来ましたか → ジョンが来ました。1	The answer noun is the new value that fills the wh-slot.6
A question word is the subject	が	何がありますか。5	Question words are neither generic nor anaphoric.1
Universal or generic statement (definition, kind-level)	は	犬は動物です。8	Generic referent satisfies the thematic condition.1
Sudden-noticing or event-report ("oh look, X is happening")	が	雨が降っている。1	Thetic, neutral-description が; the whole event is news.16
Existential or location ("there is X at Y")	が	机の上に本があります。5	The existent is new information; が marks the located noun.15
Stative-predicate object (好き, わかる, ある, できる)	が, with topic-は on the experiencer	私はコーヒーが好きです。84	Stative predicates take a が-marked holdee; the experiencer is the topic.84

私わたしはコーヒーが好すきです。⁸⁴
"I like coffee."

Good to know

Why は is written は but pronounced wa

When は serves as the topic particle, it is pronounced wa. This is one of three particle-spelling exceptions formalized in 「現代仮名遣い」 (Modern Kana Usage, Cabinet Notification No. 1, 1986), alongside を read o and へ read e.¹⁰ The exception preserves pre-1946 historical kana orthography for grammatical particles even though general kana spelling was reformed in 1946.¹⁰

Inside ordinary words, such as the は of はな ("flower") or the は of はる ("spring"), は is pronounced ha. The wa reading applies only when は is acting as a particle attached to a noun phrase.¹⁰

Reading every は as the topic particle

A frequent beginner trap is reading the は inside an ordinary word as if it were the grammatical particle. The kana は inside 花 ("flower") is pronounced ha and has no grammatical role. Only the particle は that follows a noun phrase, pronounced wa, marks topics.¹⁰

花はなは赤あかいです。⁸
"The flower is red."

In the example, the first は (inside 花) is ha and part of the word; the second は (between 花 and 赤いです) is wa and is the topic particle.

Question-word subjects mistakenly marked with は

After learning は first in lesson 1, beginners often over-apply it to question-word subjects. The form *誰は来ましたか is ungrammatical because the thematic-は constraint requires the marked nominal to be anaphoric or generic. A question word is neither.¹⁸ The correct form is:

誰だれが来きましたか。¹
"Who came?"

Sudden-noticing sentences mistakenly marked with は

A second over-application happens when a referent is being announced for the first time. Beginners reach for は because は is the first particle they learned. The form *あ、犬は来ました for an unintroduced dog is wrong. A sudden announcement of an unintroduced referent uses neutral-description が, because the referent has not yet been established as a theme.¹ The correct form is:

あ、犬いぬが来きました。¹
"Oh, a dog came."

The "spotlight" mnemonic and where it misleads

A widely used teaching metaphor among learners says は puts the spotlight on what follows it, and が on what precedes it. The intuition tracks new-versus-old information at a surface level, but it collapses two distinct grammatical jobs, topic and subject, into one visual.⁸ The mnemonic also fails on multiple-subject sentences like 象は鼻が長い. In that sentence, は and が both appear and each marks a different role; a single spotlight cannot explain why both particles are needed.¹⁷

A frame-versus-doer mnemonic is more durable: は is the frame, が is the doer. This pair maps cleanly onto Kuno's topic-versus-subject distinction. It also survives the multiple-subject case, where the frame is "elephants" and the doer is "noses."¹⁷

私は in every sentence reads as foreign

私は is the textbook self-introduction frame. But once the topic is established, Japanese drops the topic and does not re-state 私 in every following sentence. Repeating 私は reads as foreign-learner overspecification or, in some contexts, as contrastive "I, as opposed to others."¹⁹

The は/わ split is older than the 1946 reform

Modern Japanese spelling generally collapses historical /ha/ to /wa/ in word-medial position. 川 kawa comes from Old Japanese kafa. The grammatical particle は kept its old kana form even after the 1946 reform, on the explicit grounds of preserving grammatical recognizability. This is codified in the 1986 Cabinet Notification on Modern Kana Usage.¹⁰

See also

• Topic vs. Subject in Japanese: The Hidden Slot
• Dropped Subjects in Japanese: Pro-Drop Explained
• The Three Hiragana Spelling Exceptions: は, へ, and を as Particles
• Japanese Focus Prosody: Pitch Widening, Contrastive は, and Information Structure
• Japanese Particles (助詞): The Eight Categories Explained
• The Japanese Copula: です, だ, である Explained

References

Footnotes

1. Kuno, Susumu. The Structure of the Japanese Language. Cambridge, MA: MIT Press, 1973. Chapters 2 (thematic and contrastive wa) and 3 (neutral-description and exhaustive-listing ga). ↩ ↩² ↩³ ↩⁴ ↩⁵ ↩⁶ ↩⁷ ↩⁸ ↩⁹ ↩¹⁰ ↩¹¹ ↩¹² ↩¹³ ↩¹⁴ ↩¹⁵ ↩¹⁶ ↩¹⁷ ↩¹⁸ ↩¹⁹ ↩²⁰ ↩²¹ ↩²² ↩²³ ↩²⁴ ↩²⁵ ↩²⁶ ↩²⁷ ↩²⁸ ↩²⁹ ↩³⁰ ↩³¹ ↩³² ↩³³ ↩³⁴ ↩³⁵ ↩³⁶ ↩³⁷ ↩³⁸ ↩³⁹ ↩⁴⁰ ↩⁴¹ ↩⁴² ↩⁴³ ↩⁴⁴ ↩⁴⁵ ↩⁴⁶ ↩⁴⁷ ↩⁴⁸ ↩⁴⁹ ↩⁵⁰ ↩⁵¹ ↩⁵²

2. Shibatani, Masayoshi. The Languages of Japan. Cambridge: Cambridge University Press, 1990. Chapter on Japanese grammar, sections on wa and ga. ↩ ↩² ↩³ ↩⁴ ↩⁵ ↩⁶ ↩⁷ ↩⁸ ↩⁹ ↩¹⁰

3. Li, Charles N., and Sandra A. Thompson. "Subject and Topic: A New Typology of Language." In Subject and Topic, ed. Charles N. Li, pp. 457–489. New York: Academic Press, 1976. ↩ ↩² ↩³ ↩⁴ ↩⁵ ↩⁶ ↩⁷

4. Banno, Eri, Yoko Ikeda, Yutaka Ohno, Chikako Shinagawa, and Kyoko Tokashiki. Genki I: An Integrated Course in Elementary Japanese, 3rd ed. Tokyo: The Japan Times, 2020. Lessons 1–4. ↩ ↩² ↩³ ↩⁴ ↩⁵ ↩⁶ ↩⁷ ↩⁸ ↩⁹ ↩¹⁰ ↩¹¹ ↩¹² ↩¹³ ↩¹⁴ ↩¹⁵

5. 3A Corporation. Minna no Nihongo Shokyū I, 2nd ed. Tokyo: 3A Corporation, 2012. Lessons 1, 9, 10. ↩ ↩² ↩³ ↩⁴ ↩⁵ ↩⁶ ↩⁷ ↩⁸ ↩⁹ ↩¹⁰ ↩¹¹ ↩¹² ↩¹³

6. Heycock, Caroline. "Japanese -wa, -ga, and Information Structure." In The Oxford Handbook of Japanese Linguistics, ed. Shigeru Miyagawa and Mamoru Saito. Oxford: Oxford University Press, 2008. ↩ ↩² ↩³ ↩⁴ ↩⁵ ↩⁶ ↩⁷ ↩⁸ ↩⁹ ↩¹⁰ ↩¹¹ ↩¹² ↩¹³

7. Mikami, Akira. Zō wa hana ga nagai: Nihon-bunpō nyūmon (『象は鼻が長い：日本文法入門』). Tokyo: Kurosio Publishers, 1960. ↩ ↩² ↩³ ↩⁴ ↩⁵

8. Makino, Seiichi, and Michio Tsutsui. A Dictionary of Basic Japanese Grammar. Tokyo: The Japan Times, 1986. Entries: は (wa), が (ga). ↩ ↩² ↩³ ↩⁴ ↩⁵ ↩⁶ ↩⁷ ↩⁸ ↩⁹ ↩¹⁰ ↩¹¹ ↩¹² ↩¹³ ↩¹⁴ ↩¹⁵ ↩¹⁶ ↩¹⁷ ↩¹⁸ ↩¹⁹ ↩²⁰ ↩²¹

9. Hinds, John, Senko K. Maynard, and Shoichi Iwasaki, eds. Perspectives on Topicalization: The Case of Japanese wa. Typological Studies in Language 14. Amsterdam: John Benjamins, 1987. ↩ ↩² ↩³ ↩⁴

10. 文化庁 (Agency for Cultural Affairs). 「現代仮名遣い」内閣告示第一号 (Modern Kana Usage, Cabinet Notification No. 1), 1986. https://www.bunka.go.jp/kokugo_nihongo/sisaku/joho/joho/kijun/naikaku/gendaikana/index.html ↩ ↩² ↩³ ↩⁴ ↩⁵ ↩⁶

11. Iwasaki, Shoichi. Japanese (Revised Edition). London Oriental and African Language Library 17. Amsterdam: John Benjamins, 2013. Chapter on particles. ↩ ↩²