Possible Duplicate:
Should I include punctuation marks in \emph, or should I place those marks after the command?

When emphasizing the last part of a sentence by making it italic, is it recommended to also typeset the terminal punctuation in italics, or leave it as it is?


  • What is this? This is nothing!
  • What is this? This is nothing!

Quoting Bringhurst's “Elements of Typographic Style” section 3.5.2 Don't clutter the foreground:

When boldface is used to emphasize words, it is usually best to leave the punctuation in the background, which is to say, in the basic text font. It is the words, not the punctuation, that merit emphasis in a sequence such as the following:

... on the islands of Lombok, Bali, Flores, Timor and Sulawesi, the same textiles ...

But if the same names are emphasized by setting them in italic rather than in bold, there is no advantage in leaving the punctuation in roman. With italic text, italic punctuation normally gives better letterfit and thus looks less obtrusive:

... on the islands of Lombok, Bali, Flores, Timor and Sulawesi, the same textiles ...

If spaced small caps are used for emphasis - changing the stature ond form of the letters, instead of their weight or slope, and thereby minimizing the surface disturbance on the page - the question of punctuation does not arise. The punctuation used with small caps is (except for question and exclamation marks) usually the same as roman punctuation; it is only necessary to check it for accurate spacing.


For some reason the following points seem not to be quite as obvious as I thought:

  • Other typographers may have a different opinion than Bringhurst, probably for equally good reasons.
  • There is no need to follow these rules just because Bringhurst says so. In fact that would be the wrong reason to follow them.
  • Bringhurst's opinion can be a basis for a decision but shouldn't be the only basis.
  • The actual decision one makes should depend on various points: is a single word emphasized or a whole sentence? Does the emphasis have a special meaning? Does one have to follow a specific style? Which fonts were chosen (what looks okay with the one might look ugly with another)? Probably there are a lot more I didn't think of... Some of them are mentioned by user14996 in his answer.
  • I have neither said “follow these rules” nor “don't do so” but have merely quoted from Bringhurst's “Elements of Typographic Style”
  • 13
    A quote from the TeXbook (p. 321): "Some manuals of style say that punctuation should inherit the font of the preceding character, ... The author heartily disagrees."
    – egreg
    Sep 12 '12 at 19:40
  • 1
    @egreg Well, typography isn't science... :) I have two books about typography and they contradict each other on more than one matter...
    – cgnieder
    Sep 12 '12 at 19:46
  • 1
    Only in parts... Bringurst says one should use italic punctuation but not the bold one.
    – cgnieder
    Sep 12 '12 at 20:22
  • 1
    That's really a matter of opinion and may depend on the fonts. Apply your taste and don't use boldface for emphasis. :)
    – egreg
    Sep 12 '12 at 20:35
  • 4
    @egreg whether boldface is a good choice or not also depends on the language and context and on tradition. (Eg, Mandarin Chinese sometimes uses dots below their characters.) For a very useful scenario, a textbook author might sensibly boldface all new technical expressions in their first occurrence or wherever they're first defined in the main text, alongside other italicization. Sparse usage is usually best, but that too is a practice one might want to question: if lots of different semantic emphasis weren't sensible per se, code editors wouldn't do syntactic highlighting (for example). Sep 13 '12 at 1:26

Update: The Chicago Manual of Style (16e, §6.2) explicitly recommends against italicization "with" the adjacent text:

All punctuation marks should appear in the same font—roman or italic—as the main or surrounding text, except for punctuation that belongs to a title in a different font (usually italics).

It also offers the following example sentence "For light amusement he turns to the Principia Mathematica!" with the final exclamation mark being not italicized.

In brief: Publishers wouldn't emphasize the punctuation in your two examples. In other cases the answer is usually semantically "no" (if you're looking for a linguistically backed-up answer), but some people make other recommendations based on consistency of visually proximate elements. Note also that many areas of typesetting have no widely agreed-upon conventions.

I was also going to quote the same passage from Bringhurst as user @cgnieder just did, but with a strong caveat:

Styleguides are linguistically prescriptive, not descriptive. Prescription is full of opinion and very often based on personal history and tradition than on either descriptive reality or what makes the most sense.

Bringhurst is a typographer with an established track record, good taste, and generally a critical mind, unafraid to question the practice of others.

With this in mind, do consider Bringhurst's recommendations. I however perceive language in strongly structural terms. Punctuation is intended to aid in parsing; that is in fact the only purpose of punctuation, unless you use it artistically (eg, in poems) or for special effect (eg, in advertisements). Whenever I read language, I parse it, and whenever punctuation is discordant with the semantic structure or underlying linguistic constituency, I have to stop and I am bothered. Semantically, using boldface is an emphasis of an expression, but a comma is almost never part of the emphasized expression (unless the expression is an entire sentence or the comma is somewhere in the middle). So semantically, it "shouldn't" be boldfaced or italicized, though Bringhurst's argument is clearly a visual one (based on what's visually proximate). I very much respect his point of view, but for someone like me who pays a lot of attention to semantic structure, I must state that my preferences differ, for a good reason.

(A note on the side: Semantic structure is also the reason why so many people disagree with the American prescription to always place a comma and period before a closing quotation mark.)

As for the examples in the original question: italicization here emphasizes only the "this" and the "nothing", and the following punctuation mark is not part of the emphasized material. The sentence structure is something like [[This [is EMPHASIS(nothing)]].]!, with the period being eaten up by the exclamation mark. I guarantee you that there is nary a publisher who will italicize the exclamation mark (or the question mark) in such a context. The string "nothing!" does not form a linguistic constituent, so if you emphasize both the "nothing" (deepest in the sentence) and the "!" (most outside in the sentence), either (a) you need to emphasize the entire sentence to cover both, or (b) you emphasize them separately. But your question was: "If I emphasize [only] 'nothing', does emphasis of the punctuation mark automatically go with that?"

So my argument is really about semantic constituency: when looking at the tree structure, "function application" (by for example the "quote me" function or the "emphasize me" function) usually applies only to constituents (that is, subtrees), with exceptions being rare.

Bringhurst's italicization example breaks that rule (or you could italicize the "islands of" and the "and" as well, treating the emphasized material as one big constituent and not as an enumeration of separately emphasized ones). But his breaking of the rule I'm stating is not the end of the world, esp if you're more of a visual person and don't do structural parsing as much.

I stated this as a "rule", but is is really one? First, I think that as a descriptive generalization it is correct. (Note that the rule applies to the underlying semantic structure: certain expressions with discontinuous surface constituents, such as German verbs whose separable parts are somewhere else in the sentence, still form an underlying constituent.) Second, it follows from linguistic/syntactic considerations: function application only ever applies to things that you can structurally group together. So it makes sense to be wanting to apply this rule to the scenario here as well.

Back to your question: There are for sure special contexts (fonts, visual layouts, etc) that can make italicization an okay choice there. Or you could decide to independently emphasize sentence-final punctuation in a sentence (for some special reasons).

Btw, note that Bringhurst's examples do not actually cover the cases presented in your question.

Generally speaking: Styleguides and typesetting manuals differ widely in their prescriptions. Some of this is unregulated area. Beware especially of dogmatic adherents to one philosophy (eg, beware of people rabidly defending American practice for punctuation mark placement without ever having seen other linguistic areas do things differently, beware of people treating Chicago as dogma even though they've never read even a fraction of it, and beware of people who cite Knuth's decisions as ultimate truth – I'm not finger-pointing and I don't have anyone in particular in mind, but just be aware that his practice and terminology are not universal dogma and thus please be suspicious of guides and manuals that treat it as such). Especially in areas where there is no obvious consensus and where sources are scarce, you are free to make your own decisions.

PS: I wish SE'd add an italic correction ... this makes the question's option #2 look ridiculous, even though it shouldn't. If all books were printed like SE posts, italicizing the "?" and "!" here would become a necessity to prevent horizontal glyph overlap.

  • Semantics is certainly the first thing that came to my mind here as well, so I can totally understand your reasoning here. It's not the only thing, and I think you made a crucial observation at the very end: The overlap of glyphs can make the semantically correct emphasis look visually discomforting. This is font dependent, specifically how much italics are inclined. f is a good candidate for failure: Gnarf! The same goes for kerning with a full stop: \emph{STOP.} will allow for proper kerning of P., \emph{STOP}. won't. This is what motivated the question in the first place. Sep 13 '12 at 8:44
  • @Nico Great to hear your perspective on this. From my point of view and in an ideal world, kerning tables would take care of that (and any residual visual discrepancy would need to be subordinate to semantics), but your view is definitely reasonable and defensible. I know that Adobe InDesign lets you very easily adjust the kerning between any two glyphs (part of the reason is simply its WYSIWYG nature), just as it lets you adjust tracking or paragraph-level parameters straightforwardly. I wish LaTeX were easier to use in those regards. Sep 13 '12 at 10:24

Not the answer you're looking for? Browse other questions tagged or ask your own question.