Japanese Intonation: Questions, Emotion, and Emphasis

Accent belongs to the word, intonation to the sentence

Pitch accent is a property of the word. In Tokyo Japanese, ame meaning rain drops after the first mora, so it sounds like A-me, high then low; ame meaning candy has no drop and sounds like a-ME. That difference is stored with the word and does not move when you make a question. The pitch accent guide on this blog covers the four patterns; here it matters only because intonation is a separate layer laid on top.

Intonation is the tune of the whole sentence: whether the last mora rises, how far the line drifts down, and where the range widens. In Ame desu ka, the rain word still drops on its second mora, then desu sits low, and only ka rises. Learners who raise the whole word, or who flatten the accent because the sentence is a question, are mixing the two layers. Keep the word's shape and add the rise at the very end.

The rise on questions with ka

A yes-no question ends with ka and the pitch rises on that single mora. Everything before it keeps the shape of a statement. This is the opposite of an English yes-no question, where the melody often starts climbing several words before the end. If you begin rising on ikimasu, the Japanese listener hears an odd emphasis on the verb and then a question tacked on. Hold the line flat or gently falling, then lift ka cleanly.

Question-word questions, with doko, nani, itsu, or dare, also end in ka and usually rise, but the rise tends to be smaller because the question word has already announced a question. A strong rise on doko desu ka is not wrong, only slightly more insistent than a native speaker would be when simply asking the way.

A natural reply to gakusei desu ka is hai, sō desu with a clear fall, or iie, chigaimasu. Listen for the fall in the answer as well as the rise in the question; once you can hear the answer land, your own questions stop trailing upward for an extra beat.

Questions without ka

In casual speech ka is usually dropped, and in relaxed polite speech even ikimasu can stand as a question with no ka at all. The rise on the final mora is then the only thing that separates a question from a statement, which is why it has to be unmistakable. Iku with a rise is are you going; Iku with a fall is I am going. Learners who rise weakly get an answer to the wrong sentence.

The pattern extends to nouns and adjectives. Oishii with a rise asks whether the food is good, oishii with a fall says it is. Because the rise carries the whole question, the final mora is often lengthened very slightly to give the pitch room to move; that small stretch is normal and does not sound like the lengthening used for emphasis.

The statement falls, and the whole line drifts down

A plain statement ends low. The final mora of desu or masu is not stressed and not raised; it simply completes the fall. Across the sentence, Japanese also shows downdrift: each new phrase starts a little lower than the previous one, so the pitch range narrows as the sentence goes on. A long sentence sounds like a gentle staircase going down, with a small reset when a clause boundary such as kara or kedo opens a new section.

English speakers often keep their pitch high throughout and drop only on the last word, which sounds emphatic and slightly breathless in Japanese. Practise a long sentence chunk by chunk, letting each chunk begin lower, and let the final verb settle near the bottom of your range. If the last word feels too quiet, you are probably doing it right.

Emphasis without English stress

English emphasises by stress: the important word gets louder, longer, and higher on one syllable. Japanese cannot do that, because pitch on each mora is already spoken for by the word's accent. If you shout the second mora of ame meaning rain to stress it, you have raised a mora that should be low, and the listener hears candy. That is why stressing a word for emphasis sounds foreign even when the listener cannot say what went wrong.

Instead, Japanese emphasises with grammar and timing. Add an intensifier such as totemo, sugoku, hontō ni, or zenzen. Switch the particle: wa after an object makes a contrast, ga on a subject picks that person out. Move the important chunk earlier in the sentence. Lengthen a final particle such as ne or yo. You may also widen the pitch range of the whole emphasised word, high moras a little higher and low moras a little lower, as long as the shape stays the same.

Notice that none of these lines contains a loud syllable; the emphasis lives in the added word or the changed particle, and the voice stays even. Say totemo samui desu once with English stress on sa and once with a slightly wider range across the whole of samui. The second version is what a Japanese listener hears as strong feeling rather than as an accent mistake.

Ne: the rising check and the falling nod

The particle ne has two tunes and they mean different things. With a rise, ne is a check: you think something is true and want the listener to confirm it. Tanaka-san desu ne with a rise means you are Tanaka, aren't you, and it expects hai. With a fall, ne asserts something you believe you both share and invites a nod rather than an answer. Ii tenki desu ne with a fall is not a question about the weather; it is an offer to agree.

Learners often use the rising ne everywhere, which makes shared feelings sound uncertain, or the falling ne when they genuinely do not know, which sounds as if they are not listening. The reply gives you feedback: a falling ne usually gets sō desu ne with its own fall, and a rising ne gets hai, ee, or a correction.

Surprise, doubt, and politeness

Sō desu ka is the clearest case of intonation carrying feeling. With a fall it means I see and closes the topic; it is an acknowledgement, not a question. With a rise it means is that so, signals doubt or surprise, and invites the speaker to say more. The wrong tune makes you sound uninterested when you were curious, or sceptical when you were simply listening. Hontō desu ka works the same way, and a high, wide rise on it reads as genuine surprise.

Politeness has its own shape. A softened refusal or request is often left unfinished, with ga or kedo trailing off low and quiet, and the missing clause is understood. Chotto with a falling, fading tone is a complete polite no. Short reactions also carry tone: a rising e with a short vowel is surprise, and a longer ē with a fall is a warm yes. The sound is the same; the tune decides which one you said.

Sentence types and their tunes at a glance

Use this table as a checklist rather than a rulebook. Regional speech varies: Kansai speakers often use a wider range and different particle tunes, and the details differ by city and generation. Standard Tokyo shapes are the safest default and what most audio material and workplaces will expect.

How to check your own tune

Intonation is hard to hear in your own voice while you speak, so checking happens afterwards. Record a pair, such as sō desu ka falling and rising, and listen with your eyes closed to decide which is which. If you cannot tell, a listener cannot either. Then compare against a native recording and hum both versions without words; humming strips away the consonants and leaves only the melody.

• Record each pair in one take, falling then rising, and listen back before you look at the label.
• Hum the native line, then hum yours, then say both; melody first, words second.
• Say a long statement and mark on paper where each chunk started; the marks should step downward.
• Ask your AI Sensei in the 3D classroom to reply to sō desu ka as it heard it, and see whether it closes the topic or keeps explaining.
• Read the pronunciation coaching in the session report for the lines where your rise or fall was flagged as unclear, and re-record only those.

A pair drill for the week

Choose five pairs from this guide, one each from the question, ne, sō desu ka, emphasis, and trailing sections. Say each pair aloud five times, changing only the tune, and finish every pair with the reply you would expect to hear. Ten minutes a day for a week is enough for the shapes to stop feeling like a performance; after that, try to predict the reply from the tune alone whenever you listen to Japanese.

• Day one and two: iku falling and rising, then hai and iie as replies.
• Day three: Tanaka-san desu ne rising and falling, replying hai, sō desu to the rise and sō desu ne to the fall.
• Day four: sō desu ka falling and rising, then continue the story after the rising one only.
• Day five: samui desu, totemo samui desu, samui desu nē, with an even voice and no loud syllable.
• Day six and seven: chotto and sumimasen ga trailing off, then the full long statement stepping downward.