Skip to content
Allin

Sentence Case and Title Case: The Rules Differ by Language

Published 8/10/2026 · 13 min read · Text & language tools

Daniel Okonkwo

Daniel OkonkwoFront-end developer and tech writer at Allin

Web performance · File formats

Checked against 4 sources

View profile
In short

Run the sentence-case tool and one regular expression fires: it uppercases any a-z that opens the text or follows a full stop, question mark or exclamation mark plus whitespace. It has no idea what language it is reading. Three consequences, all reproduced by running it. It never lowercases anything, so "The Quick Brown Fox Jumps Over the Lazy Dog" comes back unchanged — it cannot turn a title-cased headline into sentence case, which is what most people want from it. Its alphabet is ASCII, so "élan vital. étude de cas." comes back with no capitals at all, and Spanish "¿de verdad?" stays lowercase because the character after the space is the inverted question mark, not a letter. And it invents boundaries: "she works at acme inc. the office" becomes "...acme inc. The office". Then the rules themselves differ. English title case has three cutoffs, not one: since its 18th edition Chicago capitalises prepositions of more than four letters, where until then it lowercased them regardless of length; APA 7 lowercases minor words of three letters or fewer; AP capitalises words of four letters or more. A four-letter preposition splits them — Gone with the Wind in Chicago, Gone With the Wind in APA and AP. French, Spanish, Portuguese and Italian have no title case: first word and proper nouns, nothing else, which is why an English-style headline reads as a mistake in all four. German capitalises every noun wherever it stands, because § 55 of the official spelling rules says so — that is orthography, not house style — so Der kaukasische Kreidekreis, and no tool that only touches the first letter can produce it.

English title case has three different cutoffs depending on the style guide. French, Spanish, Portuguese and Italian have none at all. German capitalises every noun. The tool knows about none of this — here is exactly what it does.

What the tool actually does: one regex, no locale

The whole tool is a single substitution: find the start of the text, or a full stop, question mark or exclamation mark followed by whitespace, and uppercase the a-z that comes next. That is the entire behaviour. There is no language parameter, no dictionary, no sentence segmenter. Feed it the English sample and it does the obvious thing: "hello world. this is a test! really? yes indeed." becomes "Hello world. This is a test! Really? Yes indeed." Feed it anything the regex was not written for and it does nothing at all, silently.

The most important thing about it is what it does not do: it never lowercases. Paste "The Quick Brown Fox Jumps Over the Lazy Dog" and you get it back character for character. Paste "THE QUICK BROWN FOX. IT JUMPED." and you get that back too. Most people who search for a sentence-case converter are holding a title-cased headline or a line of shouting and want it flattened. This tool cannot do that, and it fails without saying so, which is the worst way to fail.

Two more observed cases. It invents a sentence boundary at any abbreviation: "she works at acme inc. the office is in boston." comes back as "She works at acme inc. The office is in boston." — a capital T where none belongs. And it requires whitespace after the punctuation, so "one.two.three" becomes "One.two.three". Both are consequences of the same missing thing: the tool has no model of a sentence, only of a full stop with a space after it.

English title case: three guides, three cutoffs

Everyone agrees on the skeleton: capitalise the first word, the last word, and all the nouns, verbs, adjectives, adverbs and pronouns in between; leave the articles a, an and the in lowercase, along with the coordinating conjunctions and, but, or, nor and for, and the to of an infinitive. The disagreement is entirely about prepositions, and it is not a small one.

The Chicago Manual of Style changed its own rule in the 18th edition: prepositions of more than four letters are now capitalised in titles. Until then, its instruction was to lowercase prepositions "regardless of length", a clarification that had been in the book since the 12th edition of 1969. Chicago's own worked example is The Gospel According to Matthew — According gets a capital, because it is longer than four letters and because it is a participle rather than a preposition, while the two-letter to stays down.

APA 7 draws the line differently: lowercase the minor words of three letters or fewer, capitalise everything else, prepositions included. AP draws it at four or more letters. So a four-letter preposition is the exact point where the three guides split. Chicago writes Gone with the Wind, because with is four letters and Chicago needs more than four. APA and AP write Gone With the Wind, because with reaches their four-letter floor. Neither is a mistake, and the same headline is correctly cased two different ways depending on which house you are writing for.

French, Spanish, Portuguese and Italian do not title-case at all

The rule in all four is the same shape, and it is short. A capital on the first word of the title, a capital on any proper noun inside it, and nothing else. Quebec's Banque de dépannage linguistique states it plainly: the general rule is to put a capital only on the first word of the title of a work, and to set the title in italics; if there is a subtitle, it too takes a capital when introduced by "or", but not when introduced by a colon. It calls this the simple and coherent approach and notes it is increasingly followed, precisely because it avoids the multiplication of capitals.

So Le rouge et le noir, Cien años de soledad, Os Maias, Il nome della rosa. Not Le Rouge Et Le Noir. Chicago itself acknowledges the difference when it handles French titles: the nouns get initial capitals in an English-language context, but they "would normally be lowercase in a French title when not the first word". If you translate an English headline and carry its capitals across, a native reader sees not a design choice but an error — the same way an English reader sees one in "the new rules of remote work" set as a headline.

The Accademia della Crusca gives the same guidance for Italian from the other direction: capitals are certain at the start of a sentence and on proper nouns, and there is no reason to go beyond that. Modern Italian is deliberately more sparing with capitals than English. Portuguese adds a small trap of its own, because so many titles are proper nouns: Os Maias needs its second capital, and no rule about first words will find it.

German capitalises every noun, and that is spelling, not style

German is the one language in the set where getting this wrong is a spelling error rather than a style slip. The official rulebook of the Council for German Orthography, the 2024 Amtliches Regelwerk, splits the job over two paragraphs. § 53 says the first word of a heading, a work title, an address and the like is capitalised. § 55 says, in four words, that nouns are capitalised — everywhere, always, whatever position they occupy.

The rulebook's own examples show what the combination produces: Der kaukasische Kreidekreis, Hundert Jahre Einsamkeit, Unter den Dächern von Paris, Ein Fall für zwei. The adjective kaukasische stays lowercase because German title case does not exist; the nouns Kreidekreis, Jahre, Einsamkeit, Dächern and Fall are capitalised because they are nouns, not because they are in a title. An English-style title-caser would wrongly raise kaukasische. A first-letter-only sentence-caser leaves the nouns down. Both produce misspelled German, and running the sentence-case tool on "der hund läuft über die straße" demonstrates it: you get "Der hund läuft über die straße" where the correct form is "Der Hund läuft über die Straße".

The other capitaliser on this site is worse, and here is exactly how

The case converter offers a Title Case button, and it is a single line: lowercase the whole string, then uppercase every character that follows a word boundary. Run it and four separate problems appear at once. It capitalises articles and prepositions, which no style guide does: gone with the wind becomes Gone With The Wind, with a capital T on the. It treats an apostrophe as a word boundary, so it's a wonderful life becomes It'S A Wonderful Life. Its word-boundary class is ASCII, so a letter after an accented character starts a new word: cien años de soledad becomes Cien AñOs De Soledad, and état de siège becomes éTat De SièGe — lowercase é, capital T, inside the same word.

And because it lowercases the whole string first, it destroys any casing that was deliberate: NASA launches iPhone app comes back as Nasa Launches Iphone App. That one is not a locale problem at all — it is the reason no automatic title-caser can be trusted with a headline that contains a brand, an acronym or a surname. Between the two tools on this site, one capitalises too little and one capitalises too much, and neither implements any of the three English style guides.

What to actually do with a headline

Decide the target language first, because that decides whether there is anything to do. In French, Spanish, Portuguese and Italian, sentence case is the answer, and the sentence-case tool gets it right as long as the first letter is unaccented and there is no inverted question mark — otherwise you type the capital yourself. In German, write the nouns in capitals as you type; no tool can find them for you. In English, pick a guide and stick to it, because the reader notices inconsistency far more than they notice which cutoff you chose.

One practical habit covers most of it: type the headline in ordinary sentence case, then raise only the words your target language and style guide require, by hand. That is three or four keystrokes on a headline, and it is the only method that survives a brand name, an accented first letter and a German compound noun in the same line.

The same lower-case headline in six languages: what the rule requires, and what the sentence-case tool returns
LanguageWhat gets a capital in a titleCorrectly casedWhat the tool returns
EnglishChicago 18: everything but articles, coordinating conjunctions, infinitive to, and prepositions of four letters or fewer. APA 7 and AP: every word of four letters or more, prepositions includedGone with the Wind (Chicago) / Gone With the Wind (APA, AP)Gone with the wind — wrong for every guide
FrenchFirst word and proper nouns onlyLe rouge et le noirLe rouge et le noir — right
SpanishFirst word and proper nouns onlyCien años de soledadCien años de soledad — right, but an inverted question mark stops it dead
PortugueseFirst word and proper nouns onlyOs Maias — the second capital is a proper nounOs maias — wrong, it cannot see proper nouns
GermanFirst word (§ 53) plus every noun wherever it stands (§ 55) — orthography, not styleDer kaukasische KreidekreisDer kaukasische kreidekreis — a spelling error, not a style choice
ItalianFirst word and proper nouns onlyIl nome della rosaIl nome della rosa — right
Sentence case converterCapitalize the first letter of every sentence in your text.Try the tool

Frequently asked questions

Why does the tool do nothing to my title-cased headline?
Because it only ever uppercases. Its substitution finds a lowercase a-z at the start of the text or after sentence-ending punctuation and raises it; there is no branch that lowers anything. "The Quick Brown Fox Jumps Over the Lazy Dog" already has a capital in every position the regex looks at, so nothing matches and the string is returned unchanged. To flatten a title-cased headline you need two steps: lowercase the whole line with the case converter, then run the sentence-case tool on the result. Check the output afterwards, because step one will have destroyed any acronym or brand name in the line.
Why is my Spanish sentence after an inverted question mark not capitalised?
Because the character sitting where the regex expects a letter is the inverted question mark itself. The pattern looks for sentence-ending punctuation, then whitespace, then an a-z. In "esto es una prueba. ¿de verdad? sí, claro." the position after ". " holds ¿, which is not in a-z, so the match fails and the whole clause stays lowercase. Running it confirms the output: "Hola mundo. Esto es una prueba. ¿de verdad? Sí, claro." — every sentence raised except the interrogative one. The same shape of failure hits any sentence opening with an accented letter in French, Spanish, Portuguese, German or Italian: "élan vital." gets no capital at all.
Which English style guide should I follow?
Whichever your publisher, journal or newsroom already uses — that decision is not yours to make and looking it up takes a minute. If nobody has decided, the practical difference is narrow: all three capitalise the first and last word and every noun, verb, adjective, adverb and pronoun, and all three lowercase a, an, the, and, but, or, nor, for and infinitive to. The whole divergence is prepositions of exactly four letters, such as with, from, into, over and past. Chicago's 18th edition keeps them down, APA 7 and AP raise them. Pick one, write it in your style sheet in one sentence, and stop re-deciding per headline.
Can any tool produce correct German capitalisation automatically?
Not with pattern matching. § 55 says nouns are capitalised, and deciding whether a word is a noun in a given sentence needs grammar, not characters: laufen is a verb but das Laufen is a noun, and the rulebook devotes several paragraphs to nominalisations recognised by a preceding article, an attribute or a case-determined role. A part-of-speech tagger can do it with real accuracy; a regular expression cannot do it at all. Neither tool on this site attempts it, which is the honest outcome — the sentence-case tool leaves the nouns lowercase and the case converter's Title Case would raise the adjectives too, and both give you misspelled German.
Is title case ever right in French, Spanish, Portuguese or Italian?
Only when the title is itself an English name that you are quoting, or a proper noun that happens to contain capitals — a company name, a product, an album released under an English title. Those keep the casing their owner gave them. Everything you write yourself follows the local rule: first word, proper nouns, nothing else. There are older traditional conventions in French for titles beginning with a definite article, which is why you will see both La Chartreuse de Parme and La chartreuse de Parme in print; the modern simplified rule covers both and is the one to standardise on if you have no house style telling you otherwise.

Articles you may find interesting

All guides
ExplainerDetecting the Language of a Text, and Why Short Texts FailMeasured, not asserted: 90 short real phrases across six languages, none declined and 68 right — 76%, falling to 64% under sixteen letters. Four of the wrong answers came back at 100% confidence.ExplainerRemoving Accents Breaks Search — Until You Do It on Both SidesFolding diacritics is a normalisation step, and normalisation only works when the same function runs on the index and on the query. NFC against NFD with the code points shown, and the letters — ø, ł, ß, œ, ı — that survive the strip untouched.How-toCleaning Messy Text: The Order of Operations That Actually MattersStrip tags before decoding entities, trim before deduplicating, collapse whitespace last. Three orderings run in Node, a nine-step pipeline in the right sequence, and the invisible characters — U+00A0, U+200B, U+FEFF — that survive every naive cleanup.ExplainerCounting Characters Against a Limit Someone Else SetOne emoji is 1 character, or 7, or 11, or 25, depending on who is counting. Which one your form, your database and your SMS gateway mean — and a one-paste test that tells you which you are facing.ExplainerWord Count: Reading Time and LimitsA word count is words separated by spaces. Here's how it's counted, why limits exist, how it maps to pages and reading time, and when characters count instead.GuideStripping HTML Safely: What a Tag Remover Can and Cannot DoRemoving tags and sanitising HTML are two different jobs. One real fragment run through a naive regex and through a formatting-aware stripper, with script and style contents, block breaks, comments, CDATA and entity order all shown as output.

Related tools

This describes what these text tools do today, checked by running their own functions against the exact inputs quoted here. Where a tool gets a case wrong, that is written down rather than smoothed over, because a tool you cannot predict is worse than one you know the limits of. Nothing here is a rule any tool is obliged to follow: capitalisation, character counting and line numbering are conventions, and the conventions differ by language, by style guide and by whichever program is on the other end. Before running any of these over text you cannot retype, run it over a copy and compare both ends.

Sources

Spotted a mistake in this article?