PHẦN I · 01
One syllable, one tone, six words
You already control pitch with total precision. You do it every sentence of your life: you rise at the end of a question, you fall at the end of a statement, you swoop on "reeeally?", you drop flat on "fine." — and a listener reads all of it correctly without thinking.
The difference is only this: English spends pitch on the sentence. Vietnamese spends it on the syllable. Nothing new to build. Everything to re-aim.
Every Vietnamese syllable carries exactly one tone. No mark on the vowel means the first tone — which is a tone, not the absence of one.
This book is Northern (Hà Nội), which has all six tones. Southern has five — that section is near the end.
The operator's example, because it is the whole book in one word
English "no" and Vietnamese nó (he, she, it) look like the same word. They are three differences deep.
- The vowel. English "no" is a diphthong — it slides from "o" to "oo" and your lips close. The vowel in
nóis a single steadyo— the open one from Book 2, the "aw" of "saw". Your lips do not move at all. - The pitch. Say "No." as an English answer. Your pitch falls. The sắc tone on
nórises, sharply, from mid to high. - The length. English "no" is drawn out.
nóis short and clipped.
So an English speaker reading nó aloud produces something with the wrong vowel, moving in the wrong pitch direction. That is why "I said it exactly like it's written" and "nobody understood me" are both true at the same time.
The fix for the pitch is a re-aim, not a new skill: nó is said with the pitch of an English "what?" — a genuine question, going up. Say "what?" and then keep that rise while saying "naw". That is nó.
The six tones
| Mark · Name | Pitch shape | English feeling |
|---|---|---|
(none) ngang | mid, flat, held level | the steady middle of a calm statement — "it's fine, really" |
grave à huyền | starts low, sags lower, slightly breathy | the resigned sigh of "ohhh well" |
acute á sắc | mid, then rises sharply | a genuinely surprised "what?" |
hook ả hỏi | dips down, then curls back up | "so...?" said while waiting for the rest |
tilde ã ngã | rises, with a catch in the middle | "uh-oh" squeezed into one syllable |
dot ạ nặng | low, short, cut off hard | a grunted "uh!" that stops dead |
Tones are shapes, not notes. Each one is a path the pitch walks across a single syllable, and two of them are not even continuous. Isolate one below and the rest ghost out behind it.
Đường nét của sáu thanh — the shape of each tone. Two of the six hold a level, ngang and huyền; two change direction, sắc and hỏi; two are interrupted — ngã breaks mid-rise and nặng stops dead at about half its length. Pitch runs bottom to top, one syllable of time runs left to right. The contours are schematic, drawn from pronunciation-guide/tones.md: that file describes every shape in words and gives no Chao numerals, so the lines draw the description and not measured pitch.
The canonical set, one syllable, six words:
| Written | Tone | Means |
|---|---|---|
ma | ngang | ghost |
mà | huyền | but, which |
má | sắc | cheek (and in the South, mum) |
mả | hỏi | grave, tomb |
mã | ngã | code, horse (in compounds) |
mạ | nặng | rice seedling |
The two that need a bit more than an anchor
ngã is not just "high rising". It is a rise interrupted: the voice starts, gets choked off by a catch in the throat, and then restarts higher. That catch is a glottal stop, and it is the same thing you do in the middle of the English "uh-oh" — or, if you have a London accent, in the middle of "bu'er" for "butter". Say "uh-oh" and then compress it until it is one syllable. That is ngã.
The glottal break is the tell of a real Northern accent. Southern speakers do not make it, so if you are learning Northern this is one of the few places where getting it right marks you as having learned from Hanoians rather than from an app.
hỏi dips and then rises: down, then up. In connected Hanoi speech the rise is often small, and plenty of speakers make it with a bit of creak on the way down. Learners typically flatten it into ngang or drop it into huyền.
The English feeling that works: the way you say "so...?" when someone has stopped mid-story. The pitch falls on the vowel and then lifts as you hand the turn back.
A syllable ending in -p, -t, -c or -ch can carry only two tones: sắc or nặng.
That is not a tendency, it is how the sound system works. Those endings cut the syllable short, and a short syllable has no room for a contour that dips or breaks.
So the moment you see một, học, biết, đẹp, sách, thuốc, you know before you look at the mark that it is one of two possibilities: high-and-short (sắc) or low-and-stopped (nặng). About one syllable in ten across FluentViet's own scene and story text ends this way, and for every one of them the tone problem is a coin flip rather than a six-way choice.
The other four tones — ngang, huyền, hỏi, ngã — only ever appear on syllables that end in a vowel, a glide, or -m -n -ng -nh.
Where the tone mark sits, and why it looks like it moves
Two marks can stack on one letter: the shape mark from Book 2 and the tone mark from this one. ầ is â with a grave. ữ is ư with a tilde. ế is ê with an acute.
The tone mark always lands on the main vowel. When a syllable has two vowel letters in a row, the rule is:
- Nothing after the vowels — the mark goes on the first:
mía(sugarcane),của(of),lửa(fire),mùa(season). - Something after the vowels — the mark goes on the second:
muốn(to want),tiếng(language),người(person),nguyễn(the surname).
Once you see that, nguyễn stops looking like a spelling puzzle: it is ng + the sliding uyê + n, with the ngã tilde on the ê because something follows.
Style note: for the w-glide spellings, older typesetting writes hòa, thúy and newer writes hoà, thuý. Both are current in Vietnam, both are correct, and it changes nothing about how the word sounds. If you type with Telex the choice is usually made for you by the software.
PHẦN II · 02
Same letters, one different mark
Not every tone pair matters. These do, because both words are common and the situations overlap.
The phở family — same three letters, three different words:
| Written | Tone | Means |
|---|---|---|
phở | hỏi | the noodle soup |
phò | huyền | to escort, to prop up — also crude slang, so avoid it in the wild |
phó | sắc | deputy, vice- |
Order breakfast with the wrong mark here and the misunderstanding is memorable. And note separately that phô (to show off) is not in this set at all — that is ô, a different vowel from Book 2, not a different tone.
sắc · huyền
New, just now — versus to invite. One tone apart and both are everywhere. Mời anh means "after you". Mới đến means "just arrived". This is the danger pair for daily life.
sắc · huyền · nặng
To sell, a table, a friend. All three appear in the same café sentence, which is exactly why the set is worth drilling as a set rather than as three separate words.
huyền · sắc
But — versus cheek, and mum in the South. One is grammar and one is a person, and nothing separates them except the direction the pitch moves.
hỏi · hỏi — the vowel, not the tone
Door — versus of, belonging to. Both turn up in every other sentence. Both are also hỏi: what separates them is the vowel, ưa against ua, which is exactly why this pair sits on a tone sheet as a warning. Not every confusable pair is a tone problem, and drilling the pitch here will not fix it.
huyền · sắc · ngang
Coconut, pineapple, melon. One fruit stall, three answers, and nothing but the tone to separate them.
sắc · nặng
To urinate — versus big, and the đại of đại học, university. Mind this one when you are telling someone where you study.
The hall of fame, from the same drill sheet: bún chả (grilled-pork noodles) versus bún chưa (noodles yet?); buồn (sad) versus buồi (vulgar) if you do not finish the syllable. Both are ordinary sentences right up until they are not.
More sets, laid out for drilling, are in [[../cheatsheets/tones-minimal-pairs]]. This book is about how the six tones work; that sheet is the practice material.
PHẦN III · 03
The English habit that fights you hardest
Do not let questions rise. In English, pitch at the end of a sentence carries the question. In Vietnamese, the question is carried by a word — không, à, chưa, hả — and the pitch of every syllable stays exactly where its own tone puts it.
Anh ăn cơm chưa? (have you eaten?) ends on chưa, which is ngang: flat. If you lift it into an English question rise, you have not asked more clearly, you have said a syllable that does not exist.
The same goes the other way: English statements fall at the end. A one-word Vietnamese answer like Rồi. (yes, already) is huyền — already low — and a final English fall pushes it down into something unrecognisable. Có. (yes, there is) is sắc and must rise, even though your English instinct for a one-word answer is to drop.
This is the interference the operator's nó example names, generalised: your accent is not the problem, your sentence melody is.
How to practise this so it actually transfers
Isolated syllable drills build the wrong thing. You can nail ma mà má mả mã mạ in a mirror and still lose all six tones the moment you form a sentence, because real speech puts tones next to each other and each one starts from wherever the last one left off.
Ngâm nga — hum before you speak
Say the pitch shape with no consonants and no vowels — just a hum. Get the shape right with nothing else to think about, then add the word.
Nói theo cụm — shadow phrases, not syllables
Play three to five words of native Hanoi audio, speak along with it in near-real time, do not translate. Chunks first, then sentences.
Chậm 0.7x — slow the first pass
Slow the audio to 0.7x for the first pass, then take it back to full speed. Do not stay slow — the point is to arrive at native speed.
Ghi âm — record and compare
You cannot hear your own tone errors reliably from the inside. In the app, this is what the Nói exercises do: when a transcript comes back as a different word, the tones are what changed it.
Chỉ và nói — point-and-say with a native
They shuffle a minimal-pair set and say one; you point. This trains the ear, which is what the mouth is copying from.
Steps 1 to 3 are five minutes a day. That is enough. The two that need another person — record-and-compare and point-and-say — are what you bring to a session, not what you do alone at midnight.
You may also hear — Southern and Central
Southern has five tones, not six. ngã and hỏi merge — both come out as one mid dipping tone. So mã and mả are the same word in Sài Gòn, and Southern speakers writing quickly sometimes have to stop and think about which mark a word takes. Southern sắc and nặng also lack the sharp glottal cut that Northern gives them, so the whole system sounds smoother and less clipped.
Central is different again. The accent the Central journey teaches you to speak is South-Central — Đà Nẵng and Quảng Nam — which merges hỏi and ngã the way the South does. Huế, an hour north, moves the tone shapes themselves and not just the count, and ngã and nặng behave there in ways that surprise Northern and Southern speakers alike; FluentViet teaches Huế as a listening skill inside the Central ride rather than as a target accent. See [[../cheatsheets/central-dialect]].
The pattern one more time: the North keeps all six tones and merges consonants; the South merges tones and keeps consonants apart.
Chọn dấu — pick the mark
Eight words with the mark taken off, each with the meaning you are aiming for. Pick the tone that makes that word. Getting one wrong is the cheapest possible way to meet a contour — the shape comes back with the answer.
Dấu nào cho nghĩa này?
Items are the "Classic tone sets" rows of [[../cheatsheets/tones-minimal-pairs]].
If you remember nothing else
Không dấu cũng là một thanh
ngang is a deliberate flat mid pitch, not the absence of a tone. Treating unmarked syllables as "neutral" is how a sentence loses its melody.
Ngâm nga trước khi nói
Hum the shape with no consonants and no vowels, then add the word. Five minutes a day on steps 1 to 3 is the whole dose.
Câu hỏi không đi lên
The particle asks the question — không, à, chưa, hả. Every syllable stays exactly where its own tone puts it.
-p -t -c -ch chỉ có hai thanh
Those endings cut the syllable short, so only sắc or nặng can sit on them. About one syllable in ten — and for each one the six-way choice collapses to a coin flip.
Prove it to yourself
Say each pair. If they sound the same, you are producing one tone, not two.
| Pair | The move |
|---|---|
ma and mà | flat, then sagging low |
má and mả | rising, then dipping-then-rising |
mã and mạ | rise with a catch, then low and stopped dead |
nó and English "no" | rise, not fall — and no lip movement |
rồi and rời | same tone, different vowel — check you did not also move the pitch |
Có. as an answer | it must go up, even though English wants it to go down |
Anh ăn cơm chưa? | flat on the last syllable — no English question-rise |
Back to the start: [[0-INDEX]].