Embedded wh-Questions: 疑問詞 + か

Everything in this group so far has treated a question as a whole speech act — something you ask. But a question can also become an ingredient: "I decided what to buy," "I don't know who did it," "Tell me when it starts." In each, a smaller question is packed inside a bigger sentence and fed to a verb. Japanese has a strikingly clean tool for this: take the content question exactly as you would say it out loud, cap it with , and set the whole thing into a slot. This page is about where that fits in your questioning toolkit — and, crucially, how it pairs off against its yes/no twin, 〜かどうか.

The recipe: 疑問詞 + plain clause + か + verb

Form the question in plain form, keep its か, and hang a main verb off the end — typically a verb of deciding, knowing, telling, or asking: 決める, 分かる, 知る, 教える, 聞く, 調べる, はっきりする.

何が問題かはっきりしない。

nani ga mondai ka hakkiri shinai

It isn't clear what the problem is.

だれが犯人か分かった。

dare ga hannin ka wakatta

I figured out who the culprit is.

いつ出発するのか教えて。

itsu shuppatsu suru no ka oshiete

Tell me when we're leaving.

Notice the shape: a complete little question (何が問題か, だれが犯人か, いつ出発するのか) sits in the object slot of the outer verb, filling it exactly as a noun would. The embedded clause is in plain form — politeness lives only on the outer verb — and the question word stays put. That "stays put" is the mechanical heart of the construction, and because it is where English speakers' instincts fight hardest, it gets its own detailed treatment on the syntax page embedded wh-questions: 〜か + 疑問詞. Here we care about the inventory view: what this frame is, and what it stands next to.

何を注文するかまだ決めていない。

nani o chūmon suru ka mada kimete inai

I still haven't decided what to order.

どこで乗り換えるか駅員に聞いた。

doko de norikaeru ka ekiin ni kiita

I asked the station attendant where to transfer.

The one decision: wh or yes/no?

Here is the insight that organizes the whole area. Japanese embeds exactly two kinds of question, and they are distinguished by a single choice:

Inner question is…CloserExample
open-ended (what/who/where/when/why/how)疑問詞 + どこに行く分からない
yes-or-no〜かどうか行くかどうか分からない

The logic is beautifully economical. In an open-ended question the question word itself supplies the openness — 何 already means "which of many things," so か alone can close the clause and nothing more is needed. In a yes/no question there is no question word to do that work, so Japanese spells the openness out with どうか ("…or how / …or not"). Put the two side by side:

どこに行くか決めていない。

doko ni iku ka kimete inai

I haven't decided where to go.

行くかどうか決めていない。

iku ka dō ka kimete inai

I haven't decided whether to go (or not).

So the choice between them is never a grammar puzzle — it is just a content question: is the inner question open-ended, or a yes/no? If it has a 疑問詞, use bare か. If it is a yes/no, reach for 〜かどうか (fully treated on 〜かどうか: whether or not). This is also why sticking どうか onto a wh-question is redundant — the question word already carried the load.

💡
The 疑問詞 is the tell. See a question word in the embedded clause → close with plain か. No question word (a yes/no inside) → close with かどうか. You never need both, because 何・誰・どこ… are the "which-of-several" that どうか otherwise provides.

The embedded question can be marked for its own role: 〜かが / 〜かを

Because the whole 疑問詞〜か clause behaves like a noun, it can itself take the case particles が and を to show how it slots into the outer sentence — as that verb's subject or object. This is a step beyond the bare frame and worth seeing, because it is very common in careful and written Japanese.

誰が正しいかが問題だ。

dare ga tadashii ka ga mondai da

Who is right is the real question.

何が必要かを考えよう。

nani ga hitsuyō ka o kangaeyō

Let's think about what we need.

In the first, 誰が正しいか is the subject of だ, so it takes が. In the second, 何が必要か is the object of 考える, so it takes を. In everyday casual speech these particles are often dropped (何が必要か考えよう), but in writing and formal speech the 〜かが/〜かを marking is standard and makes the structure unambiguous.

のか and the dropped だ

Two small refinements complete the inventory. First, the embedded question very often appears as 〜のか rather than bare か. The の nominalizes the clause and adds a faint "the matter of…" flavor — a sense that there is a genuine puzzle to account for. It is especially natural with なぜ/どうして and in reflective contexts:

なぜ失敗したのか、よく考えてみた。

naze shippai shita no ka, yoku kangaete mita

I thought hard about why it had failed.

Second, when the embedded question ends in a noun or na-adjective, the copula だ does not survive in front of か — and if you use the のか variant, だ surfaces as な. Compare:

原因が何か調べている。

gen'in ga nani ka shirabete iru

I'm looking into what the cause is.

これが何なのか分からない。

kore ga nan na no ka wakaranai

I don't understand what this is.

So 何か (bare, だ gone) and 何なのか (with のか, だ→な) are both correct; ×何だか in this meaning is not. The だ→な shift before の is the same one you meet across the explanatory 〜のだ / んです.

Common mistakes

Piling どうか onto a wh-question. The commonest structural slip: treating every embedded question as if it needed どうか. The 疑問詞 already supplies the openness.

❌ 何を買うかどうか決めた。

Wrong — どうか is for yes/no only. With 何, plain か is enough.

✅ 何を買うか決めた。

nani o kau ka kimeta

I've decided what to buy.

Stacking a redundant の and か wrongly, or forgetting だ→な. のか is fine; the error is keeping だ in front of it.

❌ これが何だのか分からない。

Wrong — before のか a noun predicate takes な, not だ: 何なのか.

✅ これが何なのか分からない。

kore ga nan na no ka wakaranai

I don't understand what this is.

Dropping the clause-closing か. Without か the inner question isn't nominalized and the two clauses simply collide.

❌ だれが来る分からない。

Wrong — the embedding か is missing; nothing hands the question to 分からない.

✅ だれが来るか分からない。

dare ga kuru ka wakaranai

I don't know who's coming.

Leaving polite ます/です inside the embedded clause. The inner clause normalizes to plain form; keep the politeness on the outer verb only.

❌ いつ出発しますのか教えてください。

Wrong — the embedded clause must be plain (出発するのか); ます belongs to the outer 教えてください.

✅ いつ出発するのか教えてください。

itsu shuppatsu suru no ka oshiete kudasai

Please tell me when we're leaving.

Key takeaways

  • Embed a content question as 疑問詞 + plain clause + か, then hang a verb of deciding/knowing/telling off the end (何を買うか決めた).
  • The whole area is one choice: wh → bare か; yes/no → 〜かどうか. The question word supplies the openness, so a wh-clause never needs どうか.
  • The embedded clause acts like a noun and can take が/を for its own role in the outer sentence (誰が正しいかが問題だ).
  • 〜のか adds a reflective "the matter of…" nuance; a noun/na-adjective drops だ before か (何か) and shows な before のか (何なのか).
  • The internal "the words don't move" mechanics — Japanese never fronts the wh-word — live on the syntax page; this page is the inventory view of where the frame sits.

Now practice Japanese

Reading grammar gets you part of the way. The exercises are where it sticks — free, no signup needed.

Start learning Japanese

Related Topics

  • 〜かどうか: Whether or NotN3The same question particle か you put at the end of a yes/no question can hold that question open inside a bigger sentence: 来るかどうか分からない — か is the interrogative, どうか spells out the 'or not.'
  • だれ / なに・なん / どこ / いつN5How to build who/what/where/when questions: drop the question word into the exact slot the answer would fill, give it the particle that slot demands, and mark the sentence — the word never moves to the front.
  • Embedded Wh-Questions: 〜か + 疑問詞N3To embed a content question — 'I decided what to buy' — Japanese leaves the question word exactly where it stands and closes the clause with か; there is no wh-fronting, so English speakers must suppress the reflex to yank the question word to the front.
  • Embedded Yes/No Questions: 〜かどうかN3To fold a yes/no question inside a bigger sentence — 'I don't know whether he'll come' — Japanese uses plain form + かどうか, dropping だ before か exactly as a direct question does.