AP Updates Stylebook.

Merrill Perlman (who spent years as a copyeditor, first at The Southern Illinoisan, then The Des Moines Register, and finally The New York Times, and yet has retained sensible ideas about language) describes the new Associated Press Stylebook :

Many editors attending the AP presentation were elated that AP is now allowing the use of “%” with figures instead of requiring “percent.” Instead of “the stock rose 3 percent,” copy that follows AP can now read “the stock rose 3%.” […]

Until now, the AP also advised against transmitting any accents for the same reason. That wall now has a crack: a revised entry allows “accent marks or other diacritical marks with names of people who request them or are widely known to use them, or if quoting directly in a language that uses them.” There is still the caveat that some systems won’t accept them, and it is not blanket permission to use accents on words in English that have them or need them for pronunciation. As the “Ask the Editor” feature says: accents are for “people, not places, things, foods, weather systems or anything else. So, no accent mark in entree, cafe, decor or jalapeno. (Of course, as always, you can do it differently if you choose.)” Yes, Beyoncé, no “résumé.”

Two other small changes to the AP style guide that may have big impact:

Split infinitives are okay! (Er, OK!) The previous entry said: “In general, avoid awkward constructions that split infinitive forms of a verb (to leave, to help, etc.) or compound forms (had left, are found out, etc.” It continued: “Occasionally, however, a split is not awkward and is necessary to convey the meaning.” The new entry says this: “In many cases, splitting the infinitive or compound forms of a verb is necessary to convey meaning and make a sentence easy to read. Such constructions are acceptable.” But: “If splitting a verb results in an awkward sentence, don’t do it.” The myth that splitting an infinitive is somehow wrong is one of the hardest to kill. Perhaps it will now wither.

In another nod to sanity, the AP finally acknowledges that “data” can be singular. For years, the “data” entry said, “A plural noun, it normally takes plural verbs and pronouns,” but acknowledged that “data” could be a singular when regarded as a unit. The new entry says: “The word typically takes singular verbs and pronouns when writing for general audiences and in data journalism contexts: The data is sound. In scientific and academic writing, plural verbs and pronouns are preferred.”

Finally, the AP has also come around to the belief that using “sic” is rarely advised. The old entry called for it to “show that quoted material or person’s words include a misspelling, incorrect grammar or peculiar usage.” But as we’ve said, “sic” “can come off as snarky, giving a sense of “we know better,” at the expense of the original author.”

I’m meh on the “%” issue (6 of 1, half a dozen of the other), but I’m pleased about the accent marks, split infinitives, and singular “data,” and I have to agree that “sic” is often snarky and unnecessary.


  1. Sic is extremely snarky and extremely necessary.

  2. January First-of-May says

    Sic is extremely snarky and extremely necessary.

    This is essentially my opinion as well.

  3. Sometimes it’s necessary, yes, but not as often as snarky quoters think.

  4. David Eddyshaw says

    Sic is indissolubly connected in my mind with Private Eye, where the undoubted actual intention has ever been to be snarky. More power to them …

    That said, I included it four times in a shortish cited Kusaal text, on the grounds that (a) I thought it was important not to silently alter [sic] the original text, but (b) if I left the errors unmarked it would confuse the reader. So I can vouch for the fact that it doesn’t have to be snarky.

  5. January First-of-May says

    on the grounds that (a) I thought it was important not to silently alter [sic] the original text, but (b) if I left the errors unmarked it would confuse the reader

    OK, now that I think of it, I retract the “snarky” claim. This sort of thing is exactly what “sic” is for.

  6. Yes, in that case it’s irreproachable and irreplaceable.

  7. Don’t agree on “data”. I will continue to use it, as well as “media”, in the plural sense (“medium” for the singular). Old fogey.

  8. David Eddyshaw says

    My idiolect does the same for data and media, though in my case it may have more to do with the fact that I started Latin at school early enough (aged ten) to internalise them as plural. On reflection, that is probably just a more elaborate way of saying “old fogey.”

    “Data” and “media” are not really so much singular in the usual sense in the Tongue of the Avocado Eaters as mass, like “water.” I don’t hear anyone saying “a data” or “a media.”

  9. January First-of-May says

    I think “media” is still plural for me, weirdly enough. Not very sure on “data”; I think it could go either way.

  10. David Eddyshaw says

    It’s still more hopeless (from a fogeyish standpoint) with phenomenon versus phenomena. But then I’m one of the last people left in the UK who did (proper) Greek at school.

    And when I talk about a noumenon the Young People of Today just look blank. I blame Facebook. Or possibly Tomb Raider. In any case, it’s a Jolly Poor Show. By any criteria.

  11. @David Eddyshaw: The singular phenomenon is in a much stronger position than general singular datum and medium (although each of those has specific senses in which those singular forms still predominate) or singular agendum (which has been virtually replaced by agenda item).

    When it comes to Tomb Raider, I think your ire is misdirected. The original Tomb Raider video game (and the first two sequels, made by the same team) were a huge phenomenon in the video gaming community. In the early days of the Sony Playstation, it was one of three games (the others being Final Fantasy VII and Oddword: Abe’s Oddysee) that really seemed to be getting the most out of the console’s capabilities. Moreover, the heroine, Lara Croft, was a classically educated Englishwoman with a unique character. The media outside the video gaming world eventually took notice of the game series’s popularity and made a Tomb Raider film starring Angelina Jolie. This casting choice (and other aspects of the film) was a terrible fit to the existing video game series. I remember reading a preview article about the film (obviously written by somebody who had never actually played the games) that described Lara Croft as an “uber-bimbo,” which was about as wrong as could be. While the character was widely noted for her very large breasts, she was also intentionally made rather ugly. (Twenty years later, I confess that I am a still a bit sore about the way this bit of nerd culture was coopted and ruined by the wider corporate media.)

  12. “This time they had two competing media at the carnival, one of them doing a tarot shtick and the other a crystal ball one…”

  13. Lars (the original one) says

    I once looked up the ancient Greek dual of phenomenon — I think it is φαινομένω — in case I might outpeeve somebody with it. Does the alternative-only-applies-to-one-of-exactly-two-possibilities peeve exist in English, or is it unique to Denmark?

  14. I really can’t understand the (admittedly qualified) prescription against the use of accents and other diacritical marks. Writers should know how to produce them. They are needed for both pronunciation and disambiguation in résumé and pâté, and for pronunciation in every word which has ç or ñ. (As regards ñ, I thought Americans cared about Spanish.)

  15. Lars (not the original one) says

    > alternative-only-applies-to-one-of-exactly-two-possibilities peeve

    I’ve never heard of this particular peeve in reference to Danish, only to English. Odd.

  16. He Who Lars Last says

    alternative-only-applies-to-one-of-exactly-two-possibilities peeve

    My great aunt’s peeve was against saying “There are two alternatives” when you mean there are two choices. Because in that case there is only one alternative to the original. But the same criticism works for more than two (eg 3 alternatives for 4 choices).

  17. AJP Crown says

    I like sic. “The bar is open on Saturdays but never on weekends [sic] …” is pretty useful because it shows weekends wasn’t a typo substitution for weekdays, made at the newspaper. It may also imply that the speaker was drunk, which is also good to know if you’re planning a visit.

  18. David Marjanović says

    For data, the analogy to information is so strong that scientists treat it as a mass noun when they speak, and most make a conscious effort to treat it as plural when they write.

  19. AJP Crown says

    And it REALLY annoys me when English language publications don’t use diacritics. It’s just a jolly good excuse for English speakers to pronounce words wrong: Byawk for Bjørk, for example. I feel more strongly about retaining diacritics in languages I’m familiar with, I can’t think why.

  20. David Marjanović says

    And then there are the hypercorrectivisms like habañero

  21. I really can’t understand the (admittedly qualified) prescription against the use of accents and other diacritical marks. Writers should know how to produce them. They are needed for both pronunciation and disambiguation in résumé and pâté, and for pronunciation in every word which has ç or ñ. (As regards ñ, I thought Americans cared about Spanish.)

    The prescription has nothing to do with not caring about the issues you mention, it is purely a matter of what the newspapers who pay for AP’s existence can print.

  22. Byawk for Bjørk, for example.

    Björk. There’s no “ø” in Icelandic.

  23. Yeah, you can still get AP feeds in 7-bit ASCII, which does not include any characters with diacritics. It may seem obsolete now, but there are usually reasons to maintain compatibility with older standards.

    In the course of double-checking a fact or two about 7-bit ASCII, I finally found out why computer keyboards have a “control” key. The first 26 (or maybe 31, depending on your definition) ASCII characters are called “control characters” and are produced from the keyboard by holding down control key while you hit a letter. The control characters do things like issue line feeds (ASCII 10, ctrl-J or “^J” in old WordStar notation) or carriage returns (ASCII 13, ctrl-M). I knew all these facts that already, but I had never before put together that the reason for the “control” key was that it was used to produce the characters that control hardware devices.

  24. January First-of-May says

    in old WordStar notation

    Probably still elsewhere to this day; I had problems recently when I had to insert something into… a command line, I think… and my naive attempts at just using Ctrl-V like usual just put a “^V” in the line.

    I did find the right option in the menus, but I forgot what it was.

  25. Lars (the original one) says

    To several: The form of my tag here originated as a desperate attempt to get past some dislike that Akismet had formed against my plain name (“Lars”). And when it worked, I dared not change it again. At the time I was in fact the only regular Lars to post here, but it was not intended as a slight against other Larses (of which there were, as I said, none in evidence).

    I was taught the rule about alternatives by my grandmother some time around 1970, for Danish, but it may since have vanished from Danish peeverdom at large. And yes, the form was more like “You either have one alternative, or multiple other choices” (relative to what is being talked about). But if there are indeed only two possibilities and none of them is previously mentioned, you can introduce them as “the alternatives”, and saying “the two alternatives” would be a pleonastic and redundant superfluosity according to this rule.

    But now-a-days the semantics have changed, and I would not want to assume that any concept of duality is engendered in the minds of my recipients if I say alternativ.

  26. Lars (the original one) says

    @January First-of-May: Maybe you were trying to use the CMD widget in Windows, its keyboard shortcuts were never updated from MS-DOS until Windows 7. The one in Windows 10 did get an overhaul and can now accept many of the editor shortcuts, but you may need to disable ‘Legacy Mode’ and then enable them manually.

  27. “Habañero” always annoys me, companero.

  28. To Original Lars, that was just me. I wasn’t having a go at you, it just amused me (and for a sec. confused me) to see you and Other Lars follow each other and so since I was talking about the same thing I used a Lars tag myself.

    Re alternatives, my great aunt didn’t care what other people said. She wasn’t trying to change anything; she had merely thought about it and found it odd.

  29. Athel Cornish-Bowden says

    Pedantic, necessary, or wrong? I was writing a reference the other day to a paper that has an author with a ğ and an ı in his name when he writes in Turkish. Should I write them like that in citing it, or just use g and i? It’s no big deal, as he’s the fourth of about 20 authors, so his name won’t appear in the text of the paper, just in the reference list.

    An author I know writes his name as Gutierrez. I asked him, shouldn’t that be Gutiérrez, and he said yes, but he omits the accent in English. Especially with é, I don’t think that’s a right decision. With ğ and ı it’s perhaps more arguable, as even people who know how to use their keyboards may not know how to type them.

  30. Athel Cornish-Bowden says

    In the early days of email, when many systems couldn’t cope with accents (or users thought they couldn’t), I noticed that in general Spanish speakers were quite happy to write á as a, etc. However, they systematically refused to write ñ as n, though they differed as to how to write it: instead of año, for example, one saw agno, anyo and anio, but never the one I thought the most logical and historically accurate, anno. I found agno and anio very weird-looking, inviting wrong pronunciations, but anyo OK, as it would be in Hungarian or Catalan.

  31. David Marjanović says

    With ğ and ı it’s perhaps more arguable, as even people who know how to use their keyboards may not know how to type them.

    I simply don’t type them, I copy everything that’s not on the German keyboard layout from the character map.

    In my latest paper I cited a few works that are in Russian. I transcribed the authors in the main text, sorted the references alphabetically by these transcriptions, put the transcribed names in brackets there and then let the whole citation follow in untranscribed Russian, with translations in brackets. The journal only publishes online, though, so there’s no worry about what can be printed.

  32. The other Spanish food name that I find tends to get an annoying erroneous “ñ” is “empanada.”

  33. “Habañero” always annoys me, companero.

    I ranted about that back in 2006.

  34. David Marjanović says

    “Oh hear bloody hear.”

  35. Giacomo Ponzetto says

    In Spanish, omitting accents from common surnames poses no problem because everyone knows where they belong anyway. I suspect, but don’t have data to prove, that people may be a bit more reluctant to drop accents from uncommon surnames that are liable to be mispronounced even by fellow Spanish speakers.

    The distinction between accented vowels and ñ is also a normative one. Only the latter counts as a separate letter in the alphabet. Incidentally, as a substitute digraph I don’t see why gn (as in French and Italian) is weirder than ny (as in Catalan) or nh (as in Portuguese), but I acknowledge my bias.

  36. J.W. Brewer says

    Only now have I belatedly noticed the studied choice of eyebrow-raised conjunction in “yet has retained sensible …” in the original post. A belated bravo!

  37. In The American Way of Death, Jessica Mitford uses “[sic]” to signal a misspelling in a direct quote (compliment for complement) and then adds this footnote:
    “I do not like the repeated use of sic. It seems to impart a pedantic, censorious quality to the writing. I have throughout made every effort to quote the funeral trade publications accurately; the reader who is fastidious about usage will hereafter have to supply his own sics.”
    More here, for anyone curious: https://stancarey.wordpress.com/2014/04/29/the-pedantic-censorious-quality-of-sic/

  38. Excellent, thanks for the very apposite quote!

  39. But the reader cannot supply his own sics; the whole point (apart from smugness) is to make sure that the reader knows at which stage the apparent mistake was introduced.

  40. habañero

    We’ve come down with a slightly similar case in Finnish: “jalopeno” for jalape⁽ñ⁾o. Might be some sort of analogy from jalo ‘noble’.

  41. How does [sic] interact with a quote within a quote? Really there should be “[sic1]” or “[sic2]”. There must be instances of quoting an incorrect sic and adding a sic of one’s own


    Smith writes

    Jones screamed out in an access of fury.

    Johnson writes

    Smith says Jones screamed “in an access [sic] of furry”.

    Andrews writes

    According to Johnson, ‘Smith says Jones screamed “in an access [sic (Smith)] of furry [sic (Andrews)]”.’

  42. @mollymooly: The one time I can remember a [sic.*] in a quote of a quote, where there was thus ambiguity about who was responsible for inserting the “[sic.]” it was done clarified with something like:

    As Jones reminds us, “Allen’s idea of ‘being highly affective [sic.]’ is really nothing more than…” [“sic.” in Jones].

    * I would be tempted to quote something with an abbreviation “sic” missing the period with “[sic [sic.]].”

  43. Stu Clayton says

    On the model of cis/trans, one could have sic/snart to distinguish a dumb sic from a snart one.

  44. The problem of quoting something that already contains “[sic]” is a special case of the problem of quoting anything that contains square brackets. Not that I have any good solution.

  45. @Kieth Ivey: One thing I like about writing for physicists is that it is accepted that you can nest the delimiters { [ (…) ] } in running text, just like in mathematical equations. If a square bracket has been used, you just make the next one out a French bracket. That does not solve all the problems, but it eliminates some of them.

    On the other hand, trying to use that convention when preparing a departmental newsletter (for physics graduates!) led to a telephone conversation the involved me yelling, “Do not understand what ‘STET’ means?!” in to the receiver.

  46. Brett—”@Kieth [sic] Ivey”?

  47. AJP Crown says

    If it’s not really needed currently, most of the time, the best use of [sic] might be as a reference to Emily Dickinson’s poem Sic Transit Gloria Mundi.

    So you could write: mortality is fatal, insolvency, sublime [sic]. Even though it isn’t sic.

  48. The “sic” that I find most annoying and patronising is where the writer feels the need to insert it for minor spelling errors. It seems to be saying, “Don’t blame me, it wasn’t MY mistake”.

  49. January First-of-May says

    How does [sic] interact with a quote within a quote?

    Highly ambiguously.

    Language Log commented on it back in 2003.

  50. Good find!

  51. Brett, that might work, though it’s not quite the same as the mathematical use, since in most cases the interpolations are not nested. And when they are nested, it’s not really necessary to use separate symbols. “When in the [curse [sic]] of human events” would theoretically be OK.

    Also, what makes you think “sic” is an abbreviation? It’s a perfectly cromulent Latin word.

  52. AJP Crown says

    I would be tempted to quote something with an abbreviation “sic” missing the period with “[sic [sic.]].” [Sic]

    I had to. God made me do it.

  53. @Keith Ivey: The “sic.” used to mark errors is an abbreviation of “sic erat scriptum.” It just happens to be a full word that abbreviates a whole phrase.

  54. I saw a fairly recent paper with an elaborate scheme for representing an interpreted manuscript. There were separate notations for filled-in missing material, for removing superfluous material, and for corrections to scribal errors. That involved several different kinds of parentheses and brackets. Does this ring a bell with anyone?

  55. John Cowan says

    Oh yes, all that is the standard notation of an apparatus criticus from the 19C onward. The linked paper is mostly a glossary of Latin abbreviations, but explains the various brackets on the penultimate page. I learned damn. ‘thought to be corrupt’ and several other doozies from the glossary, and found that the abbreviations s.v. ‘see the entry for’, recte ‘correctly’, and vel sim., ‘or something similar’ that I’ve been using for years come from the app. critt. tradition. (Vel sim. is the dual of etc., speaking mathematico-logically.)

    The verb obelize ‘mark with an obelus (†) as a sign of corruption without any obvious amendment’ is pretty neat too.

  56. January First-of-May says

    I’m reminded of Nick Nicholas’ description of someone’s edition of Philodemus…

    TL/DR: the thing with Philodemus is that our texts of his works were mostly recovered from Herculaneum, in fairly burnt condition, and the 18th century researchers working on them weren’t very careful about preserving the original for the sake of future science.
    So a lot of the text is only known from the early (18th century) copies, and those early copies had problems in places, and occasionally the early copyists weren’t quite sure what the original was either.
    And in some cases – less often than not – there’s actually still enough surviving of the original to tell that the 18th century guys were wrong; but usually that copy is the best we have.

    As such, the author of that edition had to make separate notation for where the original probably had mistakes and the 18th century text had faithfully copied them, and where the 18th century copy had mistakes that probably weren’t actually there in the original, and where the text as given is clearly corrupted but since the original isn’t around any more we’re not sure if the mistakes were original or 18th century, and where the 18th century version made emendations that we don’t have the evidence for any more because the original isn’t around… there was a lot of options.
    It was fairly complicated, and involved inventing some symbols to work (which is why it came up in the Nick Nicholas article I found it in – an extensive description of Unicode and non-Unicode Greek).

  57. Ah, thanks! This is not the particular article I saw (which didn’t use any of the abbreviations), but the various brackets look familiar.

  58. Walter Scott’s edition of the Corpus Hermeticum is full of such brackets. I made myself a card, which I keep in the book, to tell myself how to reverse all his editorial changes, since he had very tendentious ideas about what the authors Must Have Meant.

  59. Brett, I’d say the “erat scriptum” is understood. That doesn’t make “sic” an abbreviation. Are there other cases where a whole word is used as a short form for a phrase and is therefore followed by a period? It seems bizarre.

  60. On “alternative”: that peevery seems to crop up everywhere where a language has that specific Latinism and people have a smattering of a classical education. I remember it from my German school days forty years ago, when teachers of both German and Latin told us that “Alternative” can only properly be used when the choice is between two options. I assume that peeve still exists; it was especially topical because of the new Alternative Listen (they didn’t want to be parties back then) that made their entries into German parliaments at that time.

  61. David Marjanović says

    I suppose you could argue that the Greens saw a choice between just two options – themselves and The Establishment…

  62. I have never seen this peeve in English for alternative (although a similar peeve does exist for dilemma), and according to the OED, there is no evidence of English usage ever restricting alternative to only two options (unlike dilemma, which does have such a history). The oldest relevant cite in the OED is from Swinburne: The alternatiue or disiunctiue speech of the testator….

  63. John Cowan says

    Well, the AHD definitely has:

    A traditional view holds that alternative should be used only when the number of choices involved is exactly two. This reasoning is based on the word’s historical relation to Latin alter, “the other of two.” Even in the 1960s, some 58 percent of the Usage Panel did not favor this edict, and now that majority is overwhelming. In 2009, fully 87 percent of the Panel accepted the sentence There are plenty of alternatives to straightforward advertising, and 85 percent accepted There are many new antibiotic alternatives to penicillin. Constructions like a number of alternatives must now be considered standard. · As an adjective, alternative can mean ‘allowing or requiring a choice between two or more things’, as in We wrote an alternative statement in case the first was rejected by the board. It may also refer to a variant or substitute in cases where no choice is involved, as in We will do our best to secure alternative employment for employees displaced by the closing of the factory. In our 2009 survey, 87 percent of the Usage Panel accepted this sentence. Interestingly, only 52 percent accepted alternate when used in the same sentence.

  64. @John Cowan: Understanding of this issue has apparently evolved significantly in recent decades. I suspect that this “traditional view” may be a comparatively recent invention, although it was given credence in the previous version of the OED (1989). The 1989 entry traced the usage with more than two alternatives as far back as John Stuart Mill in 1848: The alternative seemed to be either death, or to be permanently supported by other people, or a radical change in the economical arrangements; and it gives, “Extended to, A choice between more than two things; or one of several courses which may be chosen,” as a separate definition.

    However, the updated (2010) entry has many citations older than that which are unambiguous about there being more than just two possibilities, and the earlier relevant senses have all been updated to encompass “two or more” things. For example, Wee sometime finde Syrup of Violets, sometimes a gentle infusion of Roses, sometime sennes, sometime Rhewbarbs, sometimes Aloes, sometimes Zallop, sometimes an alternative potion or decoction only of some pearles produce such sad symptomes, from ca. 1659, and Alternative promise is where two or more are engaged to do a thing..though if either of them discharge it, both are acquitted, from 1753. The 1998 edition of The New Fowler’s Modern English Usage explicitly bases its entry for alternative (including the claim that the two-options usage is “traditional,” although it “can no longer be maintained”) on the erroneous 1998 OED entry. So I will take other usage authorities similar claims with a hearty grain of salt.

  65. either of them — also prescriptively limited to two people, but clearly not in 1753.

  66. David Marjanović says

    It is followed by both, though.

  67. I missed that. So, two or more people are engaged […] both are discharged — somebody forgot where they started, I guess.

  68. John Cowan says

    Brett: Though the OED2 is dated 1989 as a whole, the great bulk of its contents is unmodified OED1, and I’m pretty sure that’s true of alternative: in which case the entry, with its strictly/loosely-marked senses, is from 1884 or earlier, not that long after the Mill quotation. The evidence of the OED3 clearly establishes the age of alternative-beyond-two, but Johnson (1755) clearly establishes the prescription as being at least that old, and the AHD evidence shows that it was still in force in 1969.

    Lars: Or (and I think more likely) went back and inserted “or more” without fixing the rest of the sentence.

Speak Your Mind