Don’t get your dashes mixed up!
U+002D
; colloquially known as a hyphen), used to join words;U+2013
; also known as nut dash), used to represent numeric or temporal domains; also used to denote interaction or connection in conceptual domains—e.g., “gene–environment interaction;” “the ε–δ definition of continuity;” “the Dunning–Kruger effect;” &c. Traditionally, the en dash was half the width of an em dash, but it is now specified to equal the width of the capital letter N (though sometimes it is the width of the lowercase letter n);U+2014
), used to split up sentences, though its uses vary across cultures—e.g., in Polish it may be used as an opening quotation mark for each turn in dialogue; or to represent copula in Russian, unpronounced in the spoken language. The em dash is specified to equal the width of the capital letter M (though sometimes it is the width of the lowercase letter m). Traditionally, the em dash should be written with thin spaces (U+2009
) or hair spaces (U+200A
) around them, however, nowadays, set closed (no spaces) is typically fine, given the inflexibility of mainstream word processors. In Spanish, em dashes can be used parenthetically, but are spaced like parentheses. You should not substitute an em dash with a spaced en dash, as is unfortunately commonplace (this is the predominant style in German and French typography, but is becoming more popular due to popular word processors that “correct” hyphen-minuses in some contexts). Typographer Robert Bringhurst (who wrote The Elements of Typographic Style) asserts that the length and visual magnitude of an em dash “belongs to the padded and corseted aesthetic of Victorian typography,” but Bringhurst is wrong. If you must use ASCII, you can use two or three hyphens to approximately represent an em dash (this is a practice that was common in the typewriting era, and LaTeX recognises an em dash as three consecutive hyphens);U+2212
; not to be confused with monus).Caveat: characters can differ in appearance vastly with different fonts. Using the semantically correct character might not always look the best.
See also: dashes (ext.), monus, tildes, crosses