Umlaut and Diaeresis (¨): Two Jobs for the Same Two Dots

What does umlaut mean, and where do the names diaeresis and trema come from?

Umlaut is a German word: um, around or re-, plus Laut, sound. Wiktionary's entry (opens in a new tab) says English borrowed it in the 19th century and gives the pronunciation /ˈʊm.laʊt/. In German it first named a change of sound, not a mark. Duden (opens in a new tab) still lists that sense first: a, o, u and au turning into ä, ö, ü and äu. Its second sense is the vowel that results, and the letter that spells it.

So the umlaut is a job. The two dots are only how German writes it. Other languages put the same dots to a different job and give them older names. Each name in the table says what its word meant before it named the mark. The sources are Oxford for diaeresis (opens in a new tab), Duden for Trema (opens in a new tab), and Wiktionary for trema (opens in a new tab), diéresis (opens in a new tab) and dialytika (opens in a new tab).

Name	Who uses it	What the word meant first	The job it names
umlaut	German, and English for German	um, around or re-, plus Laut, sound	a changed vowel: a to ä
diaeresis, US dieresis	English	Greek diairesis, separation, from diairein, take apart	say this vowel on its own
trema, tréma	Dutch, French, German	Greek trēma, a hole, used for the dots on dice	the same: a separate vowel
diéresis	Spanish	borrowed from Latin diaeresis, the same Greek word	the u you pronounce in güe and güi
dialytika	Greek	the Greek name of the mark	a diaeresis in Greek

Oxford adds a detail the others leave out. In the late 16th century English used diaeresis for the split itself, one syllable divided into two. The name of the mark came from what the mark does.

Umlaut or diaeresis: how do you tell which job the two dots over a letter do?

Look at what would change if you removed the dots. An umlaut changes the vowel itself, so the word without the dots is a different word, or no word. A diaeresis keeps the vowel's own sound. It only tells you that the vowel starts a new syllable, so the word without the dots is the same word, read badly. FundéuRAE (opens in a new tab) gives the Spanish rules, and Oxford's entry for naive (opens in a new tab) shows the English one.

Test	Umlaut	Diaeresis or trema
Typical letters	ä, ö, ü, and äu in German	ë and ï after another vowel; Spanish ü after g
Example	für, Ströme	naïve, pingüino
What the dots do	turn a, o, u into a new vowel	say this vowel apart from the one before it
Remove the dots	a different vowel, often a different word	the same word, read with one sound or syllable too few
Write it without the key	add an e: fuer	usually drop the dots: naive
Spanish poetry	no such use	dots on a diphthong split it: rüido, ru-i-do

The last row is the rarest job. FundéuRAE notes that Spanish poets may put the diéresis on the first vowel of a diphthong. It tells the reader to say the two vowels in two syllables.

Some letters change job by language. ä is an umlaut in German but, in Finnish or Swedish, a letter of its own with its own place in the alphabet. The languages table above shows who writes each letter. The a with umlaut, o with umlaut, u with umlaut and e with umlaut pages cover one letter each.

Why does an umlaut turn back into ae, oe and ue, and a diaeresis does not?

Because the umlaut dots were an e to begin with. Wiktionary's entry for Ä (opens in a new tab) traces the German letter to Middle High German aͤ: an a with a small e written over it. The e first stood beside the a, then moved on top. The same entry notes that Finnish writings of the 16th century, set in blackletter, still show a clear lowercase e there. Over time the e shrank to two short strokes, and then to two dots.

Unicode still has that small e. It is U+0364 COMBINING LATIN SMALL LETTER E, one of the small letters written above another. Chapter 7 of the Unicode Standard (opens in a new tab) ties this group to the Latin historic manuscript tradition. Seen that way, ue for ü is less a workaround than the old e written out again at full size.

A diaeresis has no e behind it, so the same trick breaks a word. These are the fallbacks the sources above give, one per job:

Word	Job of the dots	Written without them	Why
für	German umlaut	fuer	the dots stand for a small e
naïve	English diaeresis	naive	Oxford lists naive first, with naïve as the other spelling
pingüino	Spanish diéresis	no safe fallback	without the dots, gui reads with a silent u
ä in a Finnish word	a letter of its own	a	Wiktionary: Finnish writes a, not ae as in German

So "add an e" is a German rule about the umlaut, not a rule about the mark. Where a name must lose its dots, as on a passport, look up the rule for that language. The a with umlaut page gives the passport rule for ä.

Why does Unicode use one mark, U+0308, for the umlaut and the diaeresis?

Because Unicode encodes these marks by shape, not by meaning. Chapter 7 (opens in a new tab) uses those words to explain why Greek can share the combining marks, so U+0308 is also the Greek dialytika. Its section on combining marks names U+0308 as diaeresis, trema, umlaut and double derivative. It adds that such multiple functions for one mark are not encoded separately.

The result is that ä in German and ä in Finnish are the same character, U+00E4. In decomposed form, ä is a followed by U+0308, whatever the dots mean. A search, a spell checker or a sort sees one character and cannot know the job.

Unicode does document one way to mark the difference when it matters. Section 23.2.4 of chapter 23 (opens in a new tab), on the combining grapheme joiner U+034F, gives German as its example. A German u-umlaut is often sorted like ue; a u-diaeresis, which is rare, should sort like u. An invisible joiner between the u and U+0308 marks the second kind.

We tested that sentence on 2026-10-13 with Intl.Collator in Node.js 26.0.0 (ICU 78.3, CLDR 48.0). We compared German phonebook order, which files ü as ue, with the default German order:

Text stored	German phonebook order	Default German order
Müller, one character ü	equal to Mueller	sorted with Muller, after it
Mu, U+0308, ller, decomposed	equal to Mueller	sorted with Muller, after it
Mu, U+034F, U+0308, ller	sorted with Muller, after it	sorted with Muller, after it

The joiner did what the standard says. The phonebook rule treated the marked ü as a u with dots, not as ue. The joiner has no visible glyph, the names list says, so a reader sees the same word. Without it, the decomposed and the single-character forms behave alike.

Which characters look like the umlaut but are not the same mark?

Several characters show two dots, or something close to them, and they behave differently in text. Some are other marks; Greek ϊ and Cyrillic ё carry the same U+0308. Unicode's names list (opens in a new tab) keeps them apart:

Character	Code point	What it is
◌̈	U+0308 COMBINING DIAERESIS	the mark itself; also called double dot above, umlaut, Greek dialytika and double derivative
¨	U+00A8 DIAERESIS	a spacing character: it stands alone and never attaches to a letter
◌ͤ	U+0364 COMBINING LATIN SMALL LETTER E	the small e that the umlaut dots grew from
◌̋	U+030B COMBINING DOUBLE ACUTE ACCENT	Hungarian and Chuvash; it makes ő and ű, not ö and ü
◌̤	U+0324 COMBINING DIAERESIS BELOW	two dots under a letter; in IPA, breathy voice
ϊ	U+03CA GREEK SMALL LETTER IOTA WITH DIALYTIKA	ι plus U+0308: the same mark in Greek
ё	U+0451 CYRILLIC SMALL LETTER IO	е plus U+0308: the same mark in Russian Cyrillic

The spacing ¨ is the usual trap. It looks like the mark in a character list, but typed after a letter it stays a separate character, so the text holds u and ¨, not ü. Use the letter itself, or U+0308 straight after the base letter.

The double acute is the other one. Hungarian ő is not a stylish ö: the names list gives U+030B as a separate mark, used in Hungarian and Chuvash.