Arrant Pedantry


Right, Wrong, and Relative

A while ago at work, I ran into a common problem: trying to decide whether to stop editing out a usage I don’t like. In this case, it was a particular use of “as such” that was bothering me. To me, “as such” is a prepositional phrase, and “such” is a pronoun that must refer to some sort of noun or noun phrase, as in “I’m a copy editor; as such, I fix bad writing.” In this sentence, “such” refers to the noun phrase “a copy editor”; in other words, it means, “I’m a copy editor; as a copy editor, I fix bad writing.”

But most of the time when I encounter it nowadays, it’s simply used to mean “therefore” or “consequently” (for more on that, see this post I wrote several years ago for Visual Thesaurus). And when I encountered it on that day, I changed it, as I always had before. But this time, I kept thinking about what makes a usage right or wrong and how we as editors decide which rules to enforce and which ones to let slide.

“As such” may be a simple transitional adverb for most people, but I still reflexively look for a noun phrase for that “such” to refer to. And I do this even though I know I’m in the minority. I can look at the evidence and see that the shift has happened, but it hasn’t happened in my own mental grammar.

And I think this tells us a lot about why it’s so hard for us to change our minds about usage. Knowing that I’m in the minority hasn’t magically changed how the phrase works in my head. Some things are so habitual that it’s hard to root them out. And of course there’s more than a bit of snobbery at work too—the adverbial use of “as such” sounds less educated to me, so I don’t have much incentive to give up my meaning for the new one.

Sometimes editors insist that it’s our job to preserve older meanings and slow language change, but I don’t believe it is. Nobody hired us to preserve the language. We’ve simply been hired to fix errors and make text clear and readable. And anyway, changing “as such” to “therefore” might make me feel slightly less annoyed, but it’s not going to have any measurable impact on Standard English. Even if all the copy editors in the English-speaking world were to edit it out, it will likely continue to thrive in speech and unedited text. The rest of the language will keep marching on without us.

Some editors might say that even though usage is changing, the new meaning isn’t correct or accepted yet, as if there will come some point at which it becomes correct or accepted and then everything will magically change. But the question of what’s correct or accepted is much less clear than most people realize.

What makes a particular usage correct? Is it official sanction by usage commentators? Inclusion in a reputable dictionary or style guide? Usage by well-regarded writers or some other elites? A critical mass of popular usage? Some combination of the above? And even those questions raise other questions. What if one usage commentator accepts it and another doesn’t? How do you know if a dictionary or style guide is reputable? How many well-regarded writers need to use it, and for how long? How big a mass of popular usage do you need before you decide it’s a critical one? Is it a simple majority, or maybe 75 percent or 90 percent? Does it matter if the rule in question has some sort of history behind it or if it’s a pure invention? Does it matter if the allegedly incorrect usage arose from ignorance or by some other means? Does it matter how vociferously people object to the allegedly incorrect usage?

The questions go on and on. And my answer is that, honestly, I don’t believe it’s possible to come up with any reliable test for deciding which rules to enforce and which to abandon. Even if you can answer all of the questions above, there is no formula that you plug them into that will tell you what’s correct. And even though it’s sometimes said that language is the ultimate democracy, with every user casting a vote, the truth is that there isn’t really a vote either. Nobody every tallies up the numbers and declares a winner.

That is, the answer is that there is no answer.

This doesn’t stop people from trying to come up with answers, of course. The American Heritage Dictionary had its usage panel, but that was just an opinion poll of mostly older, mostly male, and mostly white scholars and writers. Some usage dictionaries have relied on corpus data to find out what actual usage is, though finding out what usage is doesn’t tell us which usage is right. Bryan Garner gives some first principles in his usage dictionary, but they’re not true first principles—they’re inconsistently applied and occasionally contradict each other, so it often feels like they’re applied after the fact to justify the desired judgment.

This is one reason why I love Merriam-Webster’s Dictionary of English Usage so much. It mostly doesn’t attempt to declare what’s right and wrong. It basically says, “Here’s how this word has been used, and here’s what people have said about it; now make up your own mind.” It embraces the relativity.

A lot of editors find that approach frustrating because they just want to know if they should leave the word or phrase in question or change it, but I find it refreshing. It doesn’t try to pretend that there are objective answers to questions of opinion. That is, when you’re asking if you should accept a usage, you’re not asking a question that can be answered with facts.

Is it good to know what people’s opinions on usage are? Absolutely. But opinions can’t tell me what I should do. They can’t tell me whether I should accept “as such” to mean “therefore” or whether I should keep editing it out at work. Ultimately, I have to decide for myself what to do.

So the next time a new “as such” came across my desk, I made a decision: I let it go.


Umlauts, Diaereses, and the New Yorker

Several weeks ago, the satirical viral content site Clickhole posted this article: “Going Rogue: ‘The New Yorker’ Has Announced That They’re Going To Start Putting An Umlaut Over Every Letter ‘O’ And No One Can Stop Them”. I’ve long enjoyed poking at the New Yorker for its distractingly idiosyncratic style,* but I had a couple of quibbles with the article, so I took to Twitter to explain the history of those little dots.

First off, those little dots that appear in words like coöperation aren’t umlauts: they’re diaereses. But a few paragraphs into the article, they actually correct the headline with this fake quote from New Yorker editor David Remnick: “We already know some of you don’t like the dots. You probably call them umlauts. Well, you’re wrong: They’re actually called diaeresis, so try thinking twice before trying to correct us on how we use them.” But this just introduced another problem: diaeresis is the singular form. The plural form is diaereses. That is, the second o in coöperate has a diaeresis over it, but you’d say that the New Yorker uses diaereses in words with doubled vowels.

A diaeresis is a pair of dots that appear over a vowel to indicate that the vowel is pronounced separately from an adjacent vowel. For example, in English oo is generally pronounced as a single vowel sound, usually either the /u/ sound in boot or the /ʊ/ in book. The New Yorker puts a diaeresis over the repeated vowel in words like cooperate to show that those two o’s are pronounced as two distinct vowels. This also applies to other words with repeated vowels like reelect.

English doesn’t use very many diacritical marks, and the ones that it does use are almost entirely from foreign borrowings. But the diaeresis is uncommon in English even compared to other diacriticals. It mostly appears in French borrowings like naïveté (though naïve is often simplified to naive), where it serves the same purpose: showing that the two adjacent vowels are pronounced separately and not as a diphthong or a single long vowel. (In French, for example, ai is pronounced with the /ɛ/ sound in bet, so without the diaeresis, naive would be pronounced like Neve Campbell’s first name.)

The diaeresis goes all the way back to Ancient Greek, where it was also used the same way. Its first use, though, was to separate a vowel at the start of a new word from a vowel at the end of a preceding word, because Greek was originally written without any spaces between words. The word diaeresis comes from the Ancient Greek word for ‘division’, from diairein ‘to divide, separate’, from dia– ‘apart’ + hairein ‘take’. That is, it was simply a mark that divided two words or two adjacent vowels. Some later European languages saw the utility of a mark that indicated that two vowels were meant to be pronounced individually, and they adopted it.

But it has never been common in English outside of the pages of the New Yorker. In Confessions of a Comma Queen, former New Yorker copy editor Mary Norris briefly recounts the rationale behind the magazine’s style choice (excerpted here on Merriam-Webster’s website):

Basically, we have three options for these kinds of words: “cooperate,” “co-operate,” and “coöperate.” Back when the magazine was just developing its style, someone decided that the first could be misread and the second was ridiculous, and so adopted the third as the most elegant solution with the broadest application.

Norris also says that the style editor was on the verge of changing his mind on the diaereses back in 1978, but then he died, and “no one has had the nerve to raise the subject since.” Norris herself admits that “most people would not trip over the ‘coop’ in ‘cooperate’ or the ‘reel’ in ‘reelect'” and that diaereses are number one complaint from readers, but apparently they’re not going anywhere anytime soon. (I think that if you’re afraid to talk about changing your style guide—especially when readers find your style distracting and annoying—then you have either a bad style guide or a bad culture surrounding your style guide or both.)

But on to umlauts. Umlauts look just like diaereses—you could call diaereses and umlauts homoglyphic—but they’re used in a very different way and have a distinct origin. The umlaut symbol originated in German but has been borrowed into other languages, including Swedish, Hungarian, Turkish, and Finnish. But to understand what an umlaut does, you need to understand a little bit about where vowels are produced in the mouth.

The vowel /u/ (the sound in “boot”) is a high back vowel: it’s pronounced with the tongue pulled back and the mouth only slightly open so that the tongue is close to the roof of the mouth. The vowel /i/ (the sound in “beet”), on the other hand, is a high front vowel: it’s similarly pronounced with the mouth only slightly open so that the tongue is close to the roof of the mouth, but the tongue is pushed forward instead. If you alternate between saying “oo” and “ee”, you should be able to feel the difference. The vowel /i/ is pronounced a little behind your top front teeth, while /u/ is pronounced towards the soft palate, also known as the velum. (The vowel /u/ is also pronounced with the lips rounded, which has the effect of enhancing the distinction between it and /i/.) And the vowel /a/ (roughly like “ah”, though not every dialect of English has that exact vowel sound) is pronounced with the tongue in the middle or towards the front of the mouth and with the mouth wide open. The International Phonetic Alphabet considers it a low front vowel, but it’s also sometimes treated as a low central vowel.

What an umlaut symbol does, then, is indicate that a vowel is produced further forward in the mouth (and sometimes also higher in the mouth) than normal. For example, ü is pronounced in the same place as /i/, but it retains the lip rounding of /u/. Try saying /i/ or /ɪ/ (“ee” or “ih”) with your lips rounded, and voilà: you just made the sound of the German ü. An ö, by contrast, is like an /e/ or an /ɛ/ (roughly an “ay” or an “eh”) with lips rounded, while an ä is raised to an /e/ or an /ɛ/. (The vowel /a/ doesn’t have any lip rounding, so neither does the umlauted version.)

The term umlaut, which comes from a German word roughly meaning ‘sound change’, is also used in Germanic linguistics to refer to certain kinds of vowel changes, especially when a vowel moves closer to /i/. Sometimes, when a back or central vowel is followed by a front vowel, we start moving our tongue forward a little early in anticipation of that front vowel. In other words, the frontness of one vowel can spread backwards through the word to the preceding vowel.

English doesn’t use the umlaut mark, but it’s full of words that were produced by the phonological process of umlaut. Plurals like men, geese, feet, and mice were formed by umlaut. In Proto-Germanic, an ancestor of English that was spoken between about 500 BC and the first few centuries AD, the singular form of the word for ‘man’ was mann, and the plural was manniz. That /i/ vowel in the suffix eventually pulled the /a/ up and forward to /ɛ/, yielding the word men in English. At some point the suffix dropped away entirely, leaving only the changed vowel in the stem as evidence that it was there. In the case of geese, feet, and mice, the umlauted vowels also lost their rounding after they moved forward.

Umlaut also shows up in English in some less expected places. Have you ever wondered why words like busy and bury aren’t spelled like they’re pronounced? It’s because those words evolved in different ways in different Old English dialects. In some dialects, the first vowel umlauted and then lost its rounding, ultimately yielding an /ɪ/ or an /ɛ/. But in other dialects, they didn’t undergo umlaut, retaining the original /u/. At some point the two forms mashed up, and we got the spelling of one dialect and the pronunciation of another. The weird alternations in words like bring/brought and teach/taught are also the product of umlaut, with a couple other phonological changes thrown in for good measure.

So if the phonological process of umlaut is common to English, German, and other Germanic languages, why does German use the umlaut character but not English? It’s simply because the writing systems of each language developed separately after many of those sound changes had happened. For example, the modern German word schön was written schoene in Middle High German (around 1050 to 1350 AD). That final -e on the end, which has since dropped off, caused the o to become umlauted. But then, to make it clear that the o was pronounced with an umlaut, people started writing another e after the o too. Then they started writing that e above the o rather than after it to show that it was affecting the vowel but wasn’t really pronounced, and eventually this superscript e simplified to two short vertical strokes or two dots.† And thus the confusion between diaereses and umlauts was born.

So there you have it: The diaeresis is originally a Greek thing that indicates that two adjacent vowels are pronounced separately. In English, you’ll mostly see it in a few French borrowings or in the pages of the New Yorker. And the umlaut is originally a German thing, though it also represents a phonological process found in English and other languages. There aren’t a lot of German borrowings in English that use umlauts, so you mostly see it in the names of bands that are trying to look a little more metal.

Nöw yöu knöw.

* The New Yorker’s style inspired one of my favorite style-related tweets, from the inimitable Benjamin Dreyer:

† The tilde and cedilla were formed in similar ways. A tilde was originally just a superscript n, while a cedilla was a subscript z. The history of the latter is even right there in its name: a cedilla is a little ceda, an Old Spanish form of zeta.

%d bloggers like this: