Vietnamese Alphabet, Tone Marks and How to Type Each Letter

Why is ấ not a letter of the Vietnamese alphabet?

Because Vietnamese has two kinds of mark, and only one kind makes a letter. A breve, a circumflex, a horn or a stroke makes ă, â, ê, ô, ơ, ư and đ. A tone mark only sets the pitch. Ấ is the letter â carrying dấu sắc.

Lê Vinh Quốc's article (opens in a new tab) in the HCMUE Journal of Science calls Ă Â Đ Ê Ô Ơ Ư variant letters, made by adding a hat or a horn to A, D, E, O and U. Tone marks are not on his list.

Unicode CLDR draws the same line. Its alphabetic information chart (opens in a new tab) gives Vietnamese an index heading for each letter. A toned vowel gets none. So the grid above holds the letters, and the table under it holds the tones.

The tone still changes the word, as Basic Vietnamese (opens in a new tab), Michigan State University's open textbook, shows with ma:

Word	Tone	Meaning
ma	thanh ngang, no mark	ghost
mà	dấu huyền	but
má	dấu sắc	mother, cheek
mả	dấu hỏi	grave
mã	dấu ngã	horse
mạ	dấu nặng	young rice stalk

A dropped tone mark is a misspelling, even with no letter missing.

Which goes first on a Vietnamese vowel, the letter mark or the tone?

The letter mark goes first, then the tone, whenever a toned vowel is built from loose combining marks. Picked whole, it is one character: the Latin Extended Additional chart (opens in a new tab) lists ấ as U+1EA5.

We tested both orders on 2026-10-10 in Node.js v26.0.0 (ICU 78.3, Unicode 17.0). Each string went through NFC, the composed form defined in UAX #15 (opens in a new tab).

Typed as	After NFC	One character?
â, then a combining acute	ấ, U+1EA5	yes
a, combining acute, combining circumflex	á plus a loose circumflex	no
a, combining breve, combining dot below	ặ, U+1EB7	yes
a, combining dot below, combining breve	ặ, U+1EB7	yes

The second row is the trap. Both marks sit above the letter, so Unicode keeps the order typed. Tone first leaves á with a circumflex beside it. It can look right and still miss ấ in a search.

The last two rows are the control: a dot below and a breve sit on opposite sides, so either order gives ặ. Every key in the table above is already one character.

Where does the tone mark go in a Vietnamese word like Nguyễn?

On a vowel, never on a consonant, and which vowel depends on how the syllable ends. Basic Vietnamese gives the rule in three cases. Nguyễn falls in the third, so its tone sits on the ê:

The syllable has	The tone goes on	Examples
one vowel	that vowel	mình, bạn, cùng
two or three vowels, no final consonant	the second-to-last vowel	hòa, kiểu, mỗi
two or three vowels and a final consonant	the last vowel	tiếng, việt, nguyễn

Nguyễn has the vowels u, y and ê, closed by n. In kiểu no consonant follows, so dấu hỏi lands on ê, the second-to-last vowel. Both tones stack on a letter mark.

How did Vietnamese telegrams spell tones with F, J, S, R and X?

With letters the alphabet does not use. The HCMUE article describes a telegraph language devised by the scholar Nguyễn Văn Vĩnh early in the twentieth century. It sends Vietnamese with no marks at all. The article says it is still in use. A tone becomes a letter after the syllable:

In normal spelling	In a telegram
dấu huyền, sắc, hỏi, ngã, nặng	f, s, r, x, j
ư, and the marks on ă and ơ	w
â, ô, ê	aa, oo, ee
d	z

The article's sample telegram writes Haf Nooij for Hà Nội. F, J, W and Z sit outside the Vietnamese alphabet, so they were free for the job.

The same letters still type Vietnamese on a computer. Microsoft's Telex input (opens in a new tab) takes s, f, r, x and j for the tones, w for ă, ơ and ư, and aa, oo, ee for â, ô, ê; đ is dd. The Vietnamese keyboard page types Telex in the browser.

Why does Vietnamese need both D and Đ?

Because D was given another sound. The HCMUE article says Alexandre de Rhodes, who shaped the script, used D for a sound that belongs to Z. He then needed Đ for the sound of English d. That is the sound at the start of Đà Nẵng.

Đ is not the Icelandic capital eth. They only look the same. The Latin Extended-A chart (opens in a new tab) lists Đ as U+0110, D with stroke. The Latin-1 chart (opens in a new tab) lists Ð as U+00D0, the eth. In the same Node.js test, Đ lowercased to đ and Ð to ð. Paste the wrong one and a Vietnamese name stops matching. Our eth letter page covers the other side of the pair.

Đ is also the first letter to get lost. In the article's example, Trương Văn Đường gets the email address truongvanduong. In Vietnamese text, keep the đ.