Colons, semicolons, and dashes can create capitalization boundaries in titles, while quotation marks and parentheses may introduce embedded structures that require additional analysis. This guide separates punctuation-boundary detection from word-class decisions so it can explain when a new segment begins and when it must leave the result unresolved.
ReviewedEvidence4 sourcesSectionWriting & Style
Quick answer
Treat clear title/subtitle segment boundaries separately from ordinary internal punctuation. Quoted or embedded titles may require their own title structure and should not be flattened by an automatic converter.
Key details
Rule Familypunctuation-boundaries
Tracked Boundariescolon, semicolon, em dash, en dash, parentheses, quotation marks
Further guidance
Quotes And Parentheses
Balanced parentheses can be tracked without discarding the surrounding title. Quotation marks are tracked structurally; embedded-title semantics remain left unresolved when this guide cannot identify the quoted material’s role.
Segment Boundaries
The source tokenizer marks colon, semicolon, em-dash, and en-dash boundaries so the explanation trace can show why a following word is treated as a segment start.
Traceability
Every experimental token decision carries a stable reason code plus boundary/phrase context so QA can distinguish a correct rule application from a lucky-looking output.
Sources and evidence
Sources are shown with the role they play in this guide. Historical or style-sensitive claims are kept within the evidence boundary described above.
The first word of a subtitle is treated as a new capitalization boundary in major title-case systems. APA explicitly capitalizes the first word of a subtitle, Chicago capitalizes the word after a title/subtitle colon, and MLA has published the same treatment in its Style Center guidance.
Multiword prepositions such as “according to,” “because of,” “instead of,” “out of,” and “up to” should not be handled by a single crude small-word list. Different style systems analyze components differently, and some phrases contain words that can change grammatical role.
APA title case capitalizes the first word, the first word of a subtitle, major words, and words of four letters or more while keeping short minor words lowercase in ordinary mid-title positions.
Chicago 18 uses title case for titles of works and changed a long-standing rule: prepositions of five letters or more are now capitalized. Shorter prepositions generally remain lowercase internally, while first and last words are capitalized and “to” remains lowercase as an infinitive marker in ordinary mid-title use.
MLA 9 describes title style as capitalizing the first word, last word, and all principal words. MLA’s public Style Center also documents the traditional lowercase treatment of internal prepositions, but its detailed preposition article is edition-labeled for MLA 8, so this engine does not yet treat the full MLA 9 converter profile as complete.
Apostrophes can mark possession, contractions, or omitted initial letters, and those uses should not be confused with quotation marks. Chicago also has specific title-case examples for initial-apostrophe contractions and for the compound form “O’Clock.”
A title appearing inside another title has its own identity and may need quotation marks, italics, or language-specific treatment. Flattening the inner title into the outer title’s token stream can destroy both formatting and capitalization.
Hyphens and dashes are not interchangeable in title-case logic. Chicago 18 changed the treatment of later elements in prefix-based hyphenated compounds, while an em dash can act like a parenthetical or subtitle-like boundary depending on structure.
