W with Accent: Welsh Vowel Marks on ŵ ẁ ẃ ẅ, Waw in ẇ ẉ ẘ

Which w with accent is Welsh, and which stands for another alphabet?

Ask first where the w came from. A Welsh word, a Hebrew or Arabic name in Latin letters and a phonetic transcription each use a different w, and the mark means something different in each.

Each row condenses what Wiktionary's letter pages say about ŵ (opens in a new tab), ẃ (opens in a new tab), ẁ (opens in a new tab), ẅ (opens in a new tab), ẇ (opens in a new tab), ẉ (opens in a new tab) and ẘ (opens in a new tab). The Urdu row comes from the UN report quoted further down. In Unicode's own names, ŵ is the w with circumflex, ẃ the w with acute, ẁ the w with grave and ẘ the w with ring above.

Letter	Mark	Where it is written	What it tells the reader
ŵ Ŵ	circumflex, the Welsh to bach	Welsh; also Chichewa, Tumbuka and Yao	in Welsh, a long stressed w vowel
ẃ Ẃ	acute	Welsh; formerly Lower Sorbian	in Welsh, a short w vowel that takes the stress
ẁ Ẁ	grave	Welsh	a short w vowel in a word of one syllable
ẅ Ẅ	diaeresis	Welsh; phonetic alphabets	a w vowel said apart from the vowel beside it
ẇ Ẇ	dot above	ISO 259 for Hebrew; Bengali transliteration	Hebrew וּ, or Bengali য় said [o̯]
ẉ Ẉ	dot below	the UN system for Urdu; Americanist phonetics	a و that is not pronounced, or the sound [ɰ]
ẘ	ring above	ISO 233 for Arabic	وْ, the letter waw with a sukun

Unicode agrees on the first rows. Its names list (opens in a new tab) tags the small ŵ, ẁ, ẃ and ẅ with one word, Welsh, and gives the dotted and ringed w letters no language at all. That is why the languages card on this page shows Welsh and nothing else: it reads Unicode CLDR, which lists the alphabets of living languages, not transliteration schemes.

So there are two families here. The top half marks a vowel in Welsh spelling. The bottom half stands for a single Semitic letter, waw or vav, in romanization systems that needed a w different from the plain consonant.

Why does a Welsh w take an accent when an English w never does?

Because in Welsh, w is often a vowel. Wiktionary's entry for w (opens in a new tab) gives the Welsh letter the sound /uː/, the vowel of English moon. In cwm (opens in a new tab), valley, it is the short /ʊ/. English borrowed cwm with its Welsh spelling, which is why an English word can show a w with no other vowel near it.

A vowel can be long or short, stressed or not, so Welsh marks the w when the spelling alone would leave a reader guessing. One family of words shows the marks moving. The forms and sounds come from Wiktionary's entries for gŵr (opens in a new tab), gwŷr (opens in a new tab), gẃraidd (opens in a new tab) and cyẅres (opens in a new tab), northern speech first.

Word	Meaning	Said	What the w is doing
gŵr	man, husband	/ɡuːr/	a long vowel, so it carries the roof
gwŷr	men	/ɡwɨːr/	a consonant now; the roof moves to the y
gẃraidd	manly	/ˈɡʊrai̯ð/	a short vowel under the stress, so an acute
cyẅres	concubine, now obsolete	/kəˈʊrɛs/	a vowel of its own after the y, so a diaeresis

The plural is the telling row. In gwŷr the w is said as /w/, a consonant, and the long vowel is the y. The roof follows the vowel, not the letter. Gwŷr also shows that a w can be both kinds inside one language. In Wiktionary's Welsh entries, only the vowel kind ever carries a mark.

How do ẁ, ẃ and ŵ change the way a Welsh word is read?

A mark on a Welsh w warns the reader that the usual reading is wrong. Wiktionary's entry for w gives each mark its job: the grave flags a short vowel the spelling would not predict, the acute an unusually stressed short vowel, the circumflex a long or unusually stressed vowel, and the diaeresis a vowel that makes its own syllable.

Two pairs of words show the grave at work. Plain mwg (opens in a new tab), smoke, is said /muːɡ/ with a long vowel. The loanword mẁg (opens in a new tab), mug, is /mʊɡ/, short, and the grave says so. The same holds for mwd (opens in a new tab), vault, /muːd/, and mẁd (opens in a new tab), mud, /mʊd/. Unmarked, the mug would be spelled exactly like the smoke.

What the w vowel does	Mark	Example	Said
long and stressed, in a one-syllable word or the last syllable	ŵ	dŵr, water	/duːr/
short in a word of one syllable	ẁ	mẁg, mug	/mʊɡ/
short, carrying the stress at the end of a longer word	ẃ	cibẃts, kibbutz	/kɪˈbʊts/
a syllable of its own beside another vowel	ẅ, in practice often ŵ	cyẅres; dŵad	/kəˈʊrɛs/; /duː.ad/

The examples are Wiktionary's, from dŵr (opens in a new tab), cibẃts (opens in a new tab) and dŵad (opens in a new tab). Notice how many marked words are borrowings. Mug, mud and kibbutz came into Welsh from English, kibbutz from Hebrew before that, with sounds that Welsh spelling would read otherwise.

The last row has a quirk of its own. Wiktionary's note on ŵ says that in practice the roof also does the diaeresis's job, so sŵau (opens in a new tab), zoos, and dŵad are written with ŵ where you would expect sẅau and dẅad. It adds that no other Welsh vowel with a circumflex works this way. So a ŵ before another vowel can mean "say me apart", not only "say me long".

Is ŵ a Welsh letter only?

No. Wiktionary also lists ŵ in Chichewa, Tumbuka and Yao, and there it is a consonant, not a vowel. The sounds are from its entry for ŵ (opens in a new tab).

Language	ŵ is said	Note in the entry
Welsh	/uː/	a long stressed vowel
Chichewa	/β̞/	rare
Tumbuka	/β/	no note
Yao	/ʋ/	labelled Malawi

IPA writes the African sounds with β and ʋ, consonants close to a v. Welsh ŵ is a vowel like the oo of moon. The same character, U+0175, carries both, so a name spelled with ŵ needs its language before anyone can say it. None of these African alphabets is in the CLDR data our languages card reads, which is why the card names Welsh alone.

Which romanizations spell waw as ẇ, ẉ or ẘ?

Hebrew, Arabic and Urdu script share one letter, vav in Hebrew and waw in Arabic and Urdu, that writes a consonant and also long vowels. The UN Urdu table shows both jobs: و alone is v, and after a consonant it helps spell ū, o or au. Romanization systems that must tell such uses apart add a mark to the w. Wiktionary names the schemes, and the Urdu rule comes from the UNGEGN report on Urdu (opens in a new tab).

Latin letter	Stands for	Script	Scheme	When it is used
ẇ	וּ	Hebrew	ISO 259	vav used as a vowel letter, read u
ẇ	য়	Bengali	transliteration	when the letter is said [o̯]
ẉ	و	Urdu	UN system	when و is not pronounced, after خ in a few words of Persian origin
ẘ	وْ	Arabic	ISO 233	waw carrying a sukun, the sign for no vowel

In every row the mark tells the reader not to say a plain w. In the UN Urdu table, و is v unless it is silent, and the dot below flags the silent case. The same report says the UN system is not used in Pakistan, which follows the Hunterian system instead.

Phonetic alphabets reuse two of these letters for sounds. Wiktionary says Americanist phonetics writes ẅ for the sound IPA writes [ɥ], a consonant form of ü, and ẉ for IPA [ɰ]. The General Alphabet of Cameroon Languages writes ẅ for [β̞].

Only ẘ lacks a capital, a gap the FAQ below explains. A transliterated name in capitals therefore spells its ring as a separate combining mark on the W.

Which character sets were built to hold ŵ and the Welsh w letters?

Two Celtic ones were, and the common Windows one was not. Windows-1252, the Western code page, has no w with accent at all; Unicode's CP1252 mapping (opens in a new tab) lists none.

Welsh text needed ISO 8859-14 (opens in a new tab) or Apple's Mac OS Celtic (opens in a new tab). Apple's file credits Mac OS Celtic to Michael Everson, and his type foundry's page on Celtic encoding (opens in a new tab) says Latin 8 adds the Irish Gaelic consonants and the Welsh vowels to Latin 1, and was approved as ISO/IEC 8859-14 in 1998. Apple's file says Mac OS Celtic served the Welsh localization of Mac OS 7.1 and adds w with acute, grave, circumflex and diaeresis to Mac OS Roman.

Letters	ISO 8859-14	Mac OS Celtic	Windows-1252
ŵ and Ŵ	F0, D0	F9, F8	none
ẃ and Ẃ	BA, AA	FF, FE	none
ẁ and Ẁ	B8, A8	FD, FC	none
ẅ and Ẅ	BE, BD	FB, FA	none

Neither file has a byte for ẇ, ẉ or ẘ, and Windows-1252 has none either. For the dotted and ringed w, save the file as UTF-8.

These bytes are behind the odd symbols in old Welsh files. A Python 3.14.6 run on 2026-10-10 showed the effect: saved as ISO 8859-14 and opened as Windows-1252, dŵr came out as dðr, mẁg as m¸g and cibẃts as cibºts.

A Mac OS Celtic file read as Mac OS Roman fares worse: byte F9 is a lone breve in the Mac OS Roman table (opens in a new tab), so dŵr turns into d˘r. The text is not lost. Open it again with the encoding it was written in, and every roof and accent is where it was.