Skip to main content

Japanese Subordinate Clauses: How Embedded Clauses Work (Relative, Complement, Quotation, Embedded Question)

Japanese subordinate clauses are full clauses that act like a single noun, adjective, or adverbial inside a larger sentence. They come in four types: relative, complement, quotation, and embedded question.1 Once you can embed one clause inside another, two rules carry across all four types: the predicate inside stays in plain form no matter how polite the outer sentence is, and the subject marker が can soften to の in the right environments.

Overview​

A simple sentence has exactly one clause, so the clause and the sentence are the same thing.1 A subordinate clause has been folded into a larger sentence. It serves the bigger structure rather than standing on its own.

This hub maps the four embedding types as one system and explains the two mechanics that govern all of them. Each type also has its own deep-dive article. Here, the goal is the shared skeleton.

What Counts as a Subordinate Clause in Japanese​

A subordinate clause (従属節, じゅうぞくせつ) depends on a main clause and cannot stand as a sentence by itself. The main clause it attaches to is the matrix clause (主節, しゅせつ). That matrix clause carries the sentence's final force, such as whether it is a statement or question, and its politeness.1

Matrix clause vs. subordinate clause​

The matrix clause is the outer, independent clause. The subordinate clause is the dependent one embedded inside it. It feeds the matrix clause as a noun-like modifier, and it cannot serve as the whole sentence alone.1

今日きょうは忙いそがしいから、ストレスが溜たまっている。1
"Today I'm busy, so I'm feeling stressed out."

In that sentence, the から clause is subordinate: it sets up a reason, but the force of the statement lands on the matrix clause that follows. A noun-modifying clause works the same way at the phrase level.

拾ひろった猫ねこ1
"the cat that I found in front of the supermarket"

A subordinate clause is not a standalone sentence

"the cat that I found" is a noun phrase, not a sentence. The embedded clause "I found in front of the supermarket" only describes 猫. It has handed its final force over to whatever matrix clause uses the phrase.1

Head-final ordering: the modifier always precedes its head​

Japanese is head-final: the modifying material always comes before the word it attaches to. A modifying clause sits directly in front of the noun it describes, just as a plain adjective would. The verb inside that clause stays in plain form.2 The structural template is subject, particle, plain-form verb, then the modified noun.3

いつも勉強べんきょうする人ひとだ。2
"It is a person who always studies."

Here 「いつも勉強する」 precedes and modifies 人. The same ordering scales to a clause with its own object.

赤あかいズボンを買かう友達ともだちはボブだ。2
"The friend who buys red pants is Bob."

This is the same ordering that governs SOV at the sentence level: modifying material consistently comes before the element it attaches to.2

The arrow follows the clause's order: the head noun sits at the right edge of its modifier, and the matrix predicate closes the sentence.

The Four Types of Subordinate Clause​

Four embedding patterns cover most subordinate clauses a learner meets. A clause can modify a noun (relative), become a noun-like argument (complement), be quoted (quotation), or be turned into an embedded question. They share the head-final frame and the plain-form-inside rule. What differs is how the clause connects to the matrix.

Relative (noun-modifying) clauses​

A relative clause modifies a noun directly, with no relative pronoun. The predicate inside is plain form.2 The clause attaches like an adjective: there is no Japanese word for who, which, or that.

先週せんしゅうに映画えいがを見みた人ひとは誰だれ？2
"Who is the person who watched a movie last week?"

A state-of-being clause modifies a noun the same way.

学生がくせいじゃない人ひとは、学校がっこうに行いかない。2
"A person who is not a student does not go to school."

This is the type where the が/の alternation lives most visibly, covered in its own section below.

Complement clauses: こと / の / と / よう nominalizers​

A complement clause is turned into a noun-like argument so a verb can act on it. The と member packages a clause as the complement of a verb of thinking or saying. Japanese requires と at the end of that clause to bundle it into a single unit.4 The clause inside stays in plain form (普通形), never the polite ます form.4

カレーを食たべようと思おもったけど、食たべる時間じかんがなかった。4
"I thought about eating curry, but I didn't have time to eat."

When the complement clause is a state-of-being for a noun, the declarative だ must appear before と.

彼かれは高校生こうこうせいだと聞きいたけど、信しんじられない。4
"I heard that he is a high school student, but I can't believe it."

A noun complement needs だ before と

For a noun or na-adjective complement, だ is obligatory: 高校生だと, not the bare 高校生と.4 This と is one member of the こと / の / よう nominalizer family. こと, の, and よう each get their own deep-dive treatment.

Quotation clauses with と​

A direct quotation encloses the statement and adds と plus a verb such as 言う or 聞く.4 An interpreted, or indirect, quotation drops the quotation marks but keeps と. The quoted clause stays in plain form.4

アリスが、「寒さむい」と言いった。4
'Alice said, "Cold."'

Remove the marks and the same content becomes an interpreted quote, still anchored by と.

先生せんせいから今日きょうは授業じゅぎょうがないと聞きいたんだけど。4
"I heard from the teacher that there is no class today."

Embedded questions: か / かどうか​

A question can be embedded inside a larger sentence by treating it as a phrase ending in the question marker か.5 A question-word question embeds with bare か. A yes/no question appends どうか to stand in for the other choice, giving かどうか.5

田中たなかさんはいつ来くるか、分わかりますか。5
"Do you know when Tanaka-san is coming?"

For a yes/no question, かどうか marks the embedded "whether or not."

明日あした晴はれるかどうか知しらない。6
"I don't know whether or not it will be clear tomorrow."

それが正ただしいかどうかわからない。6
"I'm not sure whether or not that is correct."

The Plain-Form Rule Inside Subordinate Clauses​

Inside a subordinate clause, the predicate takes the plain (non-polite) form, no matter how polite the outer sentence is.3 A noun-modifying clause ends in a plain-form verb before the modified noun. An embedded or quoted clause must be in the normal form (普通形), keeping its own tense and polarity but never the polite ます form.4

Why the matrix politeness does not leak inward​

Politeness is marked once, at the sentence's outer edge, on the matrix predicate. The embedded predicate defaults to plain because politeness is a property of the whole utterance, not of each clause inside it.3

友達ともだちが私わたしに貸かしてくれた本ほん3
"the book that my friend lent me"

The clearest demonstration is a polite matrix wrapped around a plain embedded predicate.

それはかなりむずかしいと思おもいます。7
"I think that is quite difficult."

Politeness lives once, at the outer edge

In the と思います example, the matrix verb 思います is polite, but the embedded predicate むずかしい stays plain.7 You do not say むずかしいですと思います. The です belongs only to the sentence's outer edge.

The が / の Subject Alternation​

In a noun-modifying clause, the subject can be marked with の where が would normally appear.3 This が/の alternation (sometimes called ga-no conversion) is specific to subordinate environments. It shows up most often in relative clauses.

When の can replace が​

The substitution is smoothest in short, single-subject modifying clauses. The same relative clause works with either particle. The の variant reads as slightly more literary or written than the が variant.3

マギーが／の作つくったおにぎり3
"the rice ball that Maggie made"

A stative single-subject clause is the environment where の slots in most comfortably.

彼かれが住すんでいる街まち3
"the town where he lives"

Here, 彼が住んでいる街 and 彼の住んでいる街 both work, with の leaning more written.3

When の is blocked​

The の substitution resists longer or internally complex clauses, especially when other material sits between the subject and the verb. In that environment, が is the more natural choice.3

友達ともだちが私わたしに貸かしてくれた本ほん3
"the book that my friend lent me"

The intervening 私に makes this clause complex enough that が is preferred and の substitution is resisted.3 There is also a stylistic block. When another の already appears later in the phrase, が is chosen to avoid stacking two の sounds together.3

は can never mark a subordinate subject

は is not an option inside a relative clause at all. The subordinate subject must be が or の.3 This is separate from the が/の choice and admits no exception in a plain relative clause.

Nuance and Usage Contexts​

Embedding scales up: clauses can stack, each embedded clause keeps its own tense, and the topic marker は normally stays out of subordinate subjects.

Tense inside the clause is independent of the matrix​

The embedded predicate carries its own tense, anchored to the clause's own time rather than the matrix verb's.2 A past-tense clause can sit inside a past matrix. Each clause is tensed on its own.

子供こどもだったアリスが立派りっぱな大人おとなになった。2
"The Alice who was a child became a fine adult."

The same independence holds for progressive and other tense-aspect forms inside the clause.

彼女かのじょが昨日きのう、着きていたドレス3
"the dress that she was wearing yesterday"

Why は rarely marks a subordinate subject​

The topic marker は belongs to the matrix clause. When a phrase becomes subordinate, は is replaced by が with no special nuance attached to the swap. A subordinate clause normally cannot carry a marked topic.8

私わたしが読よんだ本ほん
"the book that I read"

The subordinate subject here takes が, and 私は読んだ本 is rejected as a relative-clause subject.3 When は does surface inside a clause, it reads as contrastive rather than as a neutral topic. That is why が is the unmarked default for a subordinate subject.89

Good to know​

The "read it backward" reflex for long modifiers​

For a long modifier, find the head noun at the end first. Then read the modifying clause that precedes it. In 「赤いズボンを買う友達」, the head is 友達 and 「赤いズボンを買う」 describes it. Locating the noun first lets you parse a stacked modifier without getting lost.2 The reflex works because the head-final principle guarantees that the modified noun sits at the right edge of its modifier. The head is always the anchor to read from.2

English relative pronouns have no Japanese equivalent​

There is no Japanese word for who, which, or that; the relative clause attaches directly and the linker gap is silent.2

A common beginner error is inserting a stray particle where English would put a relative pronoun. The same error often appears as marking the subordinate subject with は. The wrong form is 私は読んだ本. The correct form is the one below.

私わたしが読よんだ本ほん
"the book that I read"

The subordinate subject must be が or の, never は.23

の-marked subordinate subject reads more written than が​

「マギーの作ったおにぎり」 and 「マギーが作ったおにぎり」 both mean "the rice ball Maggie made." The の variant skews more literary or written, while が is the neutral spoken default.3 Choosing between them is a register decision, not a meaning change.

Dropping だ before と in a noun or na-adjective quotation​

When the quoted or complement clause is a state-of-being for a noun or na-adjective, the declarative だ must be explicit before と. The wrong form is 彼は高校生と聞いた. The correct form is below.

彼かれは高校生こうこうせいだと聞きいた。4
"I heard that he is a high school student."

Without the だ, the noun cannot serve as a quoted state-of-being.4

従属節 vs. 主節: the terms in a Japanese grammar book​

主節 (しゅせつ) is the main or matrix clause. 従属節 (じゅうぞくせつ) is the subordinate or dependent clause. The 従 carries the sense of "follow" or "be subordinate," so 従属 means "to be subordinate to." Knowing the pair lets you navigate Japanese-language grammar references, where these are the standard terms.

See also​

References​

Footnotes​

  1. Tofugu (Mami Suzuki et al.). "Japanese Sentence and Clause Structure." Tofugu. https://www.tofugu.com/japanese-grammar/sentences-and-clauses/ (limitation: learning-publisher tier) ↩ ↩2 ↩3 ↩4 ↩5 ↩6 ↩7

  2. Tae Kim. "Using descriptive subordinate clauses to describe nouns." Tae Kim's Guide to Learning Japanese. https://www.guidetojapanese.org/subclause.html ↩ ↩2 ↩3 ↩4 ↩5 ↩6 ↩7 ↩8 ↩9 ↩10 ↩11 ↩12 ↩13

  3. Maggie Sensei. "How to modify a noun in Japanese." Maggie Sensei, 4 Feb 2017. https://maggiesensei.com/2017/02/04/how-to-modify-a-noun-in-japanese/ (limitation: learning-blog tier; used for が/の alternation examples) ↩ ↩2 ↩3 ↩4 ↩5 ↩6 ↩7 ↩8 ↩9 ↩10 ↩11 ↩12 ↩13 ↩14 ↩15 ↩16 ↩17 ↩18

  4. Tae Kim. "Making and performing actions on quotations." Tae Kim's Guide to Learning Japanese. https://www.guidetojapanese.org/quotation.html ↩ ↩2 ↩3 ↩4 ↩5 ↩6 ↩7 ↩8 ↩9 ↩10 ↩11 ↩12

  5. Tae Kim. "The Many Different Ways to Embed Questions" (Phrasing questions / embedded questions lesson). Tae Kim's Guide to Learning Japanese (Grammar Guide, "Complete" edition). https://guidetojapanese.org/learn/complete/embedded_questions ↩ ↩2 ↩3

  6. JLPT Sensei. "JLPT N4 Grammar: かどうか (ka dou ka) Meaning." JLPTsensei.com. https://jlptsensei.com/learn-japanese-grammar/%E3%81%8B%E3%81%A9%E3%81%86%E3%81%8B-ka-dou-ka-meaning/ ↩ ↩2

  7. JLPT Sensei. "JLPT N4 Grammar: と思う (to omou) Meaning." JLPTsensei.com. https://jlptsensei.com/learn-japanese-grammar/%E3%81%A8%E6%80%9D%E3%81%86-to-omou-meaning/ ↩ ↩2

  8. Seraphim Editorial / "Japanese with Anime" reference (は particle). Japanese with Anime. https://www.japanesewithanime.com/2019/06/wa-particle.html (limitation: hobby-reference tier; used only as a corroborating statement on は in subordinate clauses) ↩ ↩2

  9. Massachusetts Institute of Technology, OpenCourseWare. Course 24.946 Linguistic Theory and the Japanese Language (Fall 2004), handout "The Uses of Wa and Ga." https://ocw.mit.edu/courses/24-946-linguistic-theory-and-the-japanese-language-fall-2004/ ↩