Quotation marks are punctuation symbols used to denote direct speech, citations, titles of works, and words used in a special sense. English uses double quotation marks ("") and single quotation marks (''), while Japanese uses corner brackets and double corner brackets. French uses guillemets (angle quotation marks). Because each language has its own conventions for quotation mark style and usage, multilingual text processing requires careful attention.

The Japanese quotation mark system is straightforward. Corner brackets are used for dialogue and direct quotations, and double corner brackets are used for quotations within quotations or for book and work titles. In horizontal writing, Western-style double quotes are sometimes used, but corner brackets remain the standard in vertical writing. Japanese corner brackets are full-width characters, each counted as a single character.

English distinguishes between "smart quotes" (curly quotes) and "straight quotes." Smart quotes have distinct opening and closing forms and are used in published text and word processors. Straight quotes are the typewriter-era characters used in programming and plain text. Unicode defines these as separate characters, so they are counted differently. The opening double quote (U+201C) and the closing double quote (U+201D) are distinct from the straight double quote (U+0022).

In programming, quotation marks serve as string literal delimiters and carry syntactic meaning. JavaScript supports three kinds: "hello", 'hello', and `hello` (template literals). Python similarly allows both " and ' interchangeably, but when a string itself contains a quotation mark, you must either escape it (\") or wrap the string in the other type of quote.

The "smart quote" auto-conversion feature is a character-counting pitfall. Microsoft Word and macOS automatically convert straight quotes to smart quotes. A straight quote " (U+0022) is 1 byte in ASCII, but a smart quote such as the left double quotation mark (U+201C) takes 3 bytes in UTF-8. When text is copied and pasted, smart quotes can silently infiltrate the content, causing unexpected truncation in systems that enforce byte-based character limits. Typography books on Amazon discuss these nuances in detail.

Multilingual support introduces even more quotation mark varieties. French uses guillemets, German uses low-high quotation marks, Chinese uses the same corner bracket shapes as Japanese but with different usage conventions, and Russian uses a combination of guillemets and low-high marks. Failing to convert quotation marks during localization (translation) degrades the quality of the translated text.