Grammatical Interlude: The Two Acceptable Uses of Quotation Marks

Wednesday, September 30, 2009
By dreeves

Grammatical prescriptivism annoys me to no end. (No ending sentences with prepositions? Ridiculous! Don’t split infinitives? Poppycock!) Well, except when I do it…

There are exactly two valid uses of quotation marks: when you mean something literally, and when you mean something nonliterally. It ought to be hard to mess that up.  But the following are not valid uses:

  1. I really, truly, extra mean this! I wasn’t going to mention this one because I presume that you (dear reader of Messy Matters) have never done this, but the blog of unnecessary quotes is just too funny. And I suppose it is a bit confusing: italics can sometimes serve the function of quotation marks (read on) but not the other way around.

On to more the more literate misuses (admit it, you’ve done these):

  1. I’m not going to tell you what I mean by this. E.g., “the result is ‘nearly’ optimal” by which the author means “the result is nearly optimal, for some definition of ‘nearly’, which I can’t be bothered to specify.”

  2. This sounds too colloquial/informal/idiomatic. E.g., “The subjects got ‘carried away’.” The author here is thinking, “this sounds too colloquial or idiomatic for formal writing but it’s OK because I’m not really writing it, I’m just quoting a more informal version of myself.” Please. Just write it or don’t write it. The only exception is when you expect the idiom to be unknown to the reader. In that case, the second acceptable usage below applies.

  3. I’m coining this term/phrase or introducing terminology. Use italics instead. (“We say that an allocation is satisficing if…”) Exception: it is acceptable if the usage is also literal. (“We use the term ‘satisficing’ to refer to…” Though italics are acceptable here as well.)

Which brings us to the two acceptable usages:

  1. I mean this literally. I’m referring to the word or phrase itself. E.g., “‘Promulgate’ is hard to pronounce.” Or, “Alice said ‘Look out!'” Another instance of this rule is referring to works by title, with the slew of exceptions that certain titles, like books and movies, should be italicized rather than quoted. There are also other times when italics may substitute for quotation marks, such as in the first example with the word promulgate. [1]

  2. I don’t mean this literally. You can use scare quotes when something is not the right word or phrase. Of course what I really mean by “not the right word” is that it simultaneously is and isn’t the right word. E.g., “the player then ‘wins’ with a loss of $100” (they technically won but really they lost). Or if the word is accurate only loosely by analogy, e.g., “the program bids ‘spitefully’” (there’s a tit-for-tat aspect to its bidding but, being a computer, it doesn’t actually have spite). Wrapping something in scare quotes is like adding a parenthetical “not really”. It should be done sparingly; typically only when the nonliteral usage might be confusing if not “called out”.

P.S. If the quotation marks in the last sentence were like fingernails on a blackboard to you then you didn’t need to read any of this. And, ironically, you’re likely the only one who didn’t find this all a painfully pedantic waste of time!

EDIT: Per Hartman’s Law of Prescriptivist Retaliation I had to fix a couple errors since this went to press (embarrassingly, “slough” instead of “slew”). Also, thank you to Eva Revesz, Kevin Lochner, and Reema Bennekaa for commenting on drafts of this.

Illustration by Kelly Savage.


[1] There’s little obvious rhyme or reason to these quotes-vs-italics rules but I believe the principle is just this: Both quotation marks and italics are serving the same functional purpose (marking a word or phrase as literal) but with different forms to suggest various dichotomies, such as book titles vs. chapters within a book. Although even I couldn’t bring myself to care if you violated these conventions, as long as you’re internally consistent.

Tags: ,

  • Chris Hibbert

    slough of exceptions –> slew of exceptions.

    A slew is a large quantity of something. a slough is a byway in a body of water.

  • dreeves

    Ha, thanks Chris! I actually caught that one before I saw your comment. But let this mark the opening of the flood gates…

  • sat

    The only prescriptivist essay I’ve ever liked is DFW’s Tense Present:

    [Don’t read that version; it was apparently hand-transcribed. Here’s the original: –ed. aka dreeves]

  • Jeff Doyle

    Well I certainly won’t argue with the suggestion that quotes be reserved for quoting. But this does suppose that there is a clear distinction between transparent and opaque uses of a word. (By transparent I mean using the word as a sign to indicate its ‘referent’. By opaque, the use of a word to refer to itself as a word, rather than to its meaning.)

    Scare quotes are a case of translucent usage because they do both at once. With scare quotes I am using the word as some people use it, but also calling attention to it as a possibly defective word. In other words, I am reminding the reader that it is a word. A fact which, however obvious, is at once devilishly and fortunately easy to forget.

    So if I don’t believe in referents, and have grave doubts about the advisability of thinking about language in this way, and I use scare quotes (as in the first paragraph above) to indicate my distrust of the concept of referent and hence of the word ‘referent’, you may think I am improperly using quotes to introduce terminology but actually I am saying that I am a post-foundationalist and I think the very notion of referents is bogus.

    What I am getting at is that I think an argument can be made for using single quotes as markers of tainted language, in the way some programming languages tag text derived from user input as _tainted_. (No italics in comments!) In written language, we can use punctuation as a tagging mechanism, just as we use punctuation to indicate quoting.

    This is not ‘acceptable usage’ (I am at once quoting you and indicating my reservations about the conceptual framework in which the term is mobilized) but it is arguably something just as good: innovation. And since it appears to be possible to tell the uses apart by inspection, this overloading of the quote operator is unambiguous. (I am tempted to put that last word ‘unambiguous’ in scare quotes because there is no unambiguous use of language.)

    Language does change and so does punctuation. The danger in writing a normative essay is that one is on the side of the dinosaurs. Which may be better than being on the side of the Vandals.

    I enjoyed the essay.

  • Robert Felty

    I think that the whole italics vs. quotes thing is really an artefact of typewriters. Most typewriters did not have the ability to use italics, while traditional typesetting always has. Since the advent of computer typography, we now usually have the ability to use italics again. When I refer to a particular word, I use italics. For example: Find all occurences of foo in this article. (Note that I am also testing the ability to write italics in comments. Let’s see if it works)

  • dreeves

    This being a usage snobbery free-for-all, I was poised to pounce on Rob for “artefact” which looked totally wrong to me. But Merriam-Webster vindicates it as a legit variant. I actually got multiple complaints about “grizzly” instead of “grisly” in which I admit was borne of my own ignorance but that too turns out to be an established variant, even when not talking about bears.

    @Sat, that essay by David Foster Wallace is delightful!

    @Jeff, thanks! I really like your characterization of scare quotes as “translucent usage”.

  • Robert Felty

    I think that when I typed artefact I must have been influenced by the fact that I had just been reading about Wordnet, which uses artefact in its classification scheme. It is, as Merriam-Webster points out, a chiefly British variant of artifact.

  • John Cox

    There really needs to be an eponymous law for “every descriptivist, no matter how well-intentioned, has something they’re prescriptivist about”.

    I think there is (at least) one more acceptable case: when the style guide for the publication you’re in mandates it. I might think The New Yorker’s spelling choices and diaereses are unbearably twee, but if they’re publishing me I’m likely to overlook their odd focusses and coöperate.

    Why? Same reason you should adapt your coding style to what’s around you when you’re making a change to an existing file: consistency is sometimes more important than “correctness”.

  • John Cox

    A bit more:

    While I’m deeply skeptical of prescriptivism, I find it impossible to abandon completely. That we need some rules in order to communicate with each other is trivially true, otherwise orchid tesseractsly buffalo buffalo. A pure discriptivist position just doesn’t work.

    The mistake strong prescriptivists make, though, is that they want their rules of usage to apply always and everywhere. This is understandable: if grammar laws were universal and immutable, there would be one clear right answer in every situation. The laws of grammar would be clear and beautiful. As Newton arguing for a law of universal gravitation, so the strong prescriptivist arguing for a law of universal grammar.

    The problem, of course, is that language doesn’t work this way. For example, the injunction against split infinitives makes sense in French (or any romance language, or in Latin) where infinitives are one word. In English, where an infinitive is two words, things get more complicated and “if it’s good enough for Latin it’s good enough for English” is not a compelling rationale for prohibition. Then there more egregious examples like the rules about restrictive versus nonrestrictive pronouns, which as far as I can tell come from grammarians speaking ex cathedra.

    So what’s a well-intentioned writer to do? I use the following heuristic: when I’m writing with the goal of communicating something and I’m presented with multiple ways to do it, I pick the way that minimizes ambiguity.

    For completeness, it’s important to mention that not every piece of writing is about communicating one particular meaning clearly. When writing poetry, for example, ambiguity is just another artistic technique. A really good example of this the word “Meister” in Felstiner translation of Celan’s “Todesfuge”: its multiple connotations (slave owner, orchestra director, teacher, expert artisan, etc.) reinforce each other. This makes the poem more affective for me.

    Most pieces of writing, though, are about communicating one particular meaning clearly. In this case, eliminating ambiguity is essential if you want you want to communicate with your readers without loss or distortion.

    In that case, this heuristic becomes a really good way to decide between approaches. Serial commas? Use them, because “I’d like to thank my parents, Ayn Rand and God” can accidentally leave some readers thinking you have a very interesting family tree while “I’d like to thank my parents, Ayn Rand, and God” doesn’t run that risk.

    The downside, though, is that not everything will be equivalently clear to every audience. I’m pretty sure I could use the abbreviations “i.e.” and “e.g.” here without confusion. With a general audience, though, there’s no way for me to be sure, so I’ll tend to say “that is” and “for example” instead.

    Prescriptivists would prefer one rule in all situations. This doesn’t work, because not all audiences are the same. So grammar — like punctuation, level of diction, and orthography — needs to be audience-specific for a particular piece of writing to be maximally effective.

  • John Cox

    And, yeah, Hartman’s law. s/mistakes/fixed/gc, thanks.

  • David Pennock

    “when you mean something literally, and when you mean something nonliterally”

    At first “I” thought you defined a “tautology” freeing me to scarequote “anything”, “but”…

    …now I see: to mean something nonliterally does not mean to not mean something literally.

    Wonderful essay.

    See also Chapter 41 of “Bugs in Writing”

  • Shirley

    Aren’t periods and commas “always” supposed to be inside quotation marks? (I know, I’m off on another subject)

  • dreeves

    @John Cox, great thoughts! I agree with almost everything. But I don’t think I would condone scare quotes for mocking a house style that you’re nominally conforming to.

    @David Pennock, I was relieved to see that “Bugs in Writing” vindicates my take on quotation marks. But I had mixed feelings about it as I noticed that it also includes laughable proscriptions of terminal prepositions and split infinitives (with especially laughable rationalizations thereof).

    @Shirley, excellent use of scare quotes around “always”. In fact, US and British conventions differ on this point. I prefer the British style as it’s a little more logical and less ambiguous (and even the US style allows for exceptions in the case of real ambiguity, like “Did he say ‘I rule’?”). Or maybe it’s because I’m a computer scientist. In a computer program, sticking punctuation inside a quoted string is out of the question. Like splitting infinitives in Latin! (Look what I’m turning into. (Look at that into which I’m turning?)) But seriously, this is too much nitpickery even for me. Pick a convention that makes sense to you and stick with it.

    Speaking of terminal prepositions, there was once a little boy who called down to his mother to bring up a book for a bedtime story. She arrived with a book about Australia, something of little interest to the child. So he asked, scornfully,
    “What did you bring me that book I didn’t want to be read
    to out of about Down Under up for?”

    More here:

  • Pingback: Abusing quotation marks | Scripturient: Blog & Commentary

  • Pingback: sites

  • Pingback: Decubre más

  • Pingback:

  • Pingback: speaking of

  • Pingback: haga clic aquí ahora

  • Pingback: electricity saving device

  • Pingback: Cyber Monday: OK, Fine | Beeminder Blog

  • Pingback: horoscope

  • Pingback: roswell bicycles

  • David Pennock

    A great example of misuse #0 (or refreshing honesty), spotted at mile 254.6 of westbound PA turnpike.

  • Amy Leskowski

    I’m wandering around the internet trying to figure out if it’s acceptable to use scare quotes in this situation:

    He has what his father calls an old soul.

    Should it be:

    He has what his father calls “an old soul.”

  • Daniel Reeves

    Not necessary there! “Old soul” isn’t the father’s coinage. Kind of like how “my roommate called me a deadbeat” doesn’t want quotes around “deadbeat” even if my roommate literally called me that.

  • Amy Leskowski

    Thank you, Daniel! Wow, I didn’t know if I’d get a response as this seemed like an old post! Thank you :)