Something about NPR has been driving me crazy for a long time now. At first, I couldn't identify it. But then my training in linguistics kicked in, and it all made sense. Many NPR commentators and announcers consistently exhibit affected speech patterns, and that's what I want to tell you about. But before I do that—and in order to make my rant enlightening—I need to go into some basic facts about phonology. It's my hope that the following lesson will give my rant more weight, and explain clearly why and how NPR made me nuts.
Flapping
American English has, as a part of its phonological system (its system of speech sounds), a rule concerning the sound called a flap. What is spelled "tt" in the word butter is pronounced as a flap. What is spelled "t" in the word metal is pronounced as a flap, also. We spell it with t's, but we don't pronounce it with t's. No American English speaker in his or her right mind pronounces these words as though there's a "hard" t in them. When I say "hard," I mean what linguists would call an aspirated t. That is, a t produced with a tiny puff of breath following the release of the sound. In a word like top, the t is aspirated. But in stop, the t is unaspirated. Say them. You'll see what I mean. (Introductory phonetics textbooks usually tell people to hold their hands in front of their mouths to feel the difference between the t in top and the t in stop. It’s more noticeable doing this with spot and pot to see the difference between the unaspirated and the aspirated p.) When I say, "in this case, t is aspirated, but in this other case, it's not," I don't mean there's some rule in a book that says American English speakers have to do this in order to speak "correctly." No, this is just one of many, many things Americans know about their language without even knowing that they know it. No one taught it to them any more than they were explicitly instructed to put adjectives before nouns, or to pronounce the plural of dog with a z sound at the end and the plural of cat with an s sound.
Okay, so far, so good. American English has flaps. Not only that, we have a rule that governs when to flap. In other words, flapping is systematic. Predictable. You could even say "grammatical." Look at the word markets, as in "The markets were crowded today." That t in market is a regular old t. But what happens when market gets a different ending attached to it, like -ing? What about the word marketing? What happens to the t in that case? It gets turned into a flap. I'll write that flap sound with a D. In other words, marketing is actually pronounced [markeDing], with a flap where the spelling of "marketing" might lead you to expect the t sound. Sure, in jolly old England, they might pronounce it [marketing], but this is America. Let's be clear here. In jolly old, they'll actually pronounce the t in marketing in its aspirated form. (This is one of the main things you should do if you're trying to sound like the British upper crust: aspirate your t's.) I'll write that aspirated t like so: T. (Hey, this is even present in the word British. We Americans say [briDish], but over there, they call themselves [briTish].) To sum up, in America, it's [markeDing] (and [buDer] for butter), while in England it's [markeTing] (and [buTer]—or more like [buTa], but that's a different issue). By the way, these bracketed forms aren't "official" phonetic representations of the words in question. I'm greatly simplifying things, so I can make my case about flapping.
What is a flap, anyway? Phonologically, is it just a kind of t or d? Well, no, it's not. Might be simpler if it was, but it's not, really. Even though flaps get spelled with t's and d's in English—and I'm using [D] to represent them here—they're actually a different sort of sound. If you were to look at a sound spectrogram of a flap—a depiction of the acoustic energy involved in the transmission of a flap—you'd see that flaps have more in common with sounds like r and l than they do with t and d. In some languages, what is typically spelled with an r is actually a flap. (But let's not get hung up on the spelling. What's important here is sounds, not letters. We have to use symbols to write down those sounds, but let's not lose sight of the fact that it's the sounds we're interested in.)
I said before that flapping—or To Flap or Not to Flap—was a matter of rule. Not a matter of whim, or laziness, or what have you. The reason we Americans flap isn't because we don't have the energy to pronounce everything written with a "t" in that very British aspirated form. We flap because that's just how American English works. (Hey, we also say [rather] instead of [rawther]—it's just part of what makes American English American English.) So what does the rule look like? It's simple:
When a t or d comes after a vowel and before an unstressed vowel, it becomes a flap.
(Okay, I lied. It's more complicated than this, but let's just go with this streamlined version.) The t in markets doesn't fit that pattern, so it doesn't get transformed into a flap. But the t in marketing does fit that pattern—it comes after a vowel and before an unstressed vowel—so it's transformed into a flap. Nothing could be simpler, once you've learned your language's system. Notice that a word like atomic is pronounced like [aTomic]. That's no flap in there, because the vowel that follows the t is stressed, so the flapping rule doesn't apply. The flapping rule does apply to atom, though, to give us [aDom]. (The same holds for the flap and nonflap in thirty and thirteen, respectively.) Potato has one aspirated t and one flap: [poTaDo]. Again, and I have to emphasize this, it's not a matter of laziness. In fact, in a sense, Americans are actually more concerned about such things than the British are. We go out of our way to include a rule about this in our language. We "care," linguistically speaking, about something some British English-speakers don't even think about. They don't flap. We do—but only sometimes, in these certain special cases. See? We need to be careful and make distinctions, where they just have a blanket policy.
What does it mean to say that flapping is a rule? It means that When To Flap (and What To Flap) can be found in the heads of every native American English-speaker. Oh sure, there might be some nonflappers hanging on for dear life, like maybe some so-called Boston Brahmins, but for the most part, flapping is everyday, garden-variety American English. There's nothing uneducated about it. It's not a Southern thing, or a Western thing, or a youth thing, or something engaged in by those lower socio-economic classes over there. And there are many, many such rules at work in our (and everyone else’s) language. Remember what I said about aspirated t vs. unaspirated t? That's a matter of a rule also. And the rule doesn't govern only t's, but also p's and k's. The rule, like most such rules, is pretty abstract. And it's something even very young English-speakers handle like pros. (The rule says something like: If a voiceless stop—the kind of sound that p, t, and k are—comes at the beginning of a word, or right before a stressed vowel, make it aspirated. Otherwise, don't bother.) It's because of this rule that we say pie with an aspirated p, but spy without one. Likewise, we pronounce akimbo with an aspirated k, and skit without one.
It might seem like this isn't actually a rule, but instead just a description of something we have to do. As though linguists just wrote down something natural and inevitable and said, "Behold the Rule of Aspiration!" After all, it's hard not to aspirate in those cases. It's hard to pronounce these words any differently. Try saying top without an aspirated t at the beginning. (It can even be hard for us to hear the difference!) Maybe this is because p's and t's and k's just automatically come out aspirated or unaspirated depending on what word they're in. Nope. Sounds reasonable, but that's wrong. Spanish-speakers and Korean-speakers, to name just two groups of many, don't feel compelled to aspirate all voiceless stops that come at the beginning of words. It's not that their mouths work differently; it's that they have learned a different set of rules. So you see, the idea that we carry around weird-looking rules in our heads isn't really all that strange, once you examine some of these rules and see what they might do.
Postlexical Rules
Back to flapping. What do linguists say about the flapping rule? How do they explain it, define it, describe it, and so on? Well, at least when I got my Master's in linguistics in the 90s, they called this rule a postlexical rule. It's not important right now why they called it that. What is important is what it means for a rule to be a postlexical rule. There are all sorts of phonological rules—rules that describe how we manipulate the symbols that stand for sounds. And I'm not talking about letters. Linguists are almost never interested in how we spell things. They're interested in how people represent language in their minds, whether they've ever learned to read and write or not. Like that rule above about aspiration. That rule involves all kinds of abstract, "technical" concepts like "voiceless stop"—which implies that speakers organize sounds into different categories—and "stressed vowel"—which suggests that we know the difference, on an abstract level, between stressed and unstressed vowels, and that the difference matters. And notice that the rule isn't a rule about t, or an unstressed a. It's a rule about general, abstract categories of things. This is what linguists believe our internal grammars—our symbolic representations of our language and the rules that govern it—actually look like. No, not with little t's written somewhere, not with anything written down, but rules do involve abstract representations of the features of language.
So what is a postlexical rule? A postlexical rule is a kind of phonological rule that operates across the board, with no exceptions, whenever and wherever it can. Any place the conditions for the rule are met, the rule applies. Between words, between roots and suffixes, within simple words... Postlexical rules don't care where they apply. In the case of flapping, this means that wherever the conditions of the rule are found (when a t or d comes after a vowel and before an unstressed vowel) the rule automatically applies and the t or d becomes a flap. There's no override switch, no “yes, but.” Inside a simple word like meteor—by simple, I mean that meteor is one single root, with no suffixes stuck onto it—the condition is found. (There's a t or d after a vowel and before an unstressed vowel.) And so what we write with a t in meteor is spoken as a flap: [meDeor]. The only place you'd expect to hear [meTeor]—with an aspirated t—is in England, or some other nonflapping region. Between a root like rocket and a suffix like the "-ed" that indicates past tense, the condition for flapping is found. That's why we say "the spaceship [rockeDed] across the galaxy," instead of [rockeTed]. And postlexical rules can even apply between words. Look at the sentence "Halley's comet is burning brightly." The conditions that trigger the flapping rule are here, too, and the rule doesn't care that part of the flap-trigger (the vowel and the t or d) is found in one word and another part of the flap-trigger (the following unstressed vowel) is found in another word. And so that sentence comes out "Halley's [comeD is] burning brightly" and not "Halley's [comeT is] burning brightly." Not all rules in English—or in other languages—work this way. Some rules never apply between words, and some rules never apply within simple words. Postlexical rules like the flap rule, though, will apply all over the place.
Note one consequence of flapping being classified as a postlexical rule: there won't be any words in American English that could have flaps in them, but don't. You won't find any words with aspirated t's where flaps would otherwise be expected. If it can flap, it will. Case closed. You won't find any place where the rule could apply, but where, for some reason, it doesn't.
(Of course, to be picky about it, it's not the classification of the rule that means there won't be exceptions; it's the lack of exceptions that leads to the classification of the rule. It's easy to start talking like the rules come first, but they don't. It's like the "law" of gravity. The law came after the facts, as an attempt to describe them, not the other way around. Even if Newton had never formulated the law of gravity in the first place, apples would still fall out of trees.)
It's very important that this point be understood: flapping or not flapping isn't a matter of the "correct" vs. "incorrect" use of language. Flapping is perfectly standard behavior for Americans. Everyone does it. Educated people, people in power, working people, people from all regions of the country and all walks of life. Professors flap, doctors flap, lawyers flap, CEOs flap, farmers flap. And not only is flapping utterly commonplace, it's also grammatical. That it is to say, it's systematic and rule-based. It's grammatical, in the sense that linguists use that term. It's a part of the language system we all learn and use, not something to be avoided or "hypercorrected" out of our mouths.
The Shame of It All
And now we return to NPR. (At last!) NPR's on-air talent appears to suffer from the misconception that there's something wrong with flapping. As though it's unseemly, or careless, or, well, fine for the Masses but not for the Ivory Tower crowd. Oh, poor misled NPR. Of course, once you start listening, it's easy to hear fellow Americans who harbor a deep-seated insecurity about their linguistic heritage. This insecurity goes back hundreds of years! Historically, Americans have always been worried that their speech embodied a sort of corruption of the beauty and perfection that was British English. (Who knows? Maybe it's the name of the language that confuses people. "It's called English, so the English must know more about it than we do!")
Here's an example of this kind of hyper-Briticism that Americans have been susceptible to for centuries: every year, if you listen for it, you'll hear some American sports commentator pronounce the name of a prestigious British tennis tournament wrong. You'll hear Wimbledon pronounced [wimbleTon], with an aspirated t where you'd expect a flap (or maybe a d). Where does this mispronunciation come from? We all make mistakes. We all have weird, idiosyncratic pronunciations in our heads, ready to leap out and brand us as mavericks. I don't think that's what's going on with "Wimbleton," though. What I think is going on is a reanalysis of the word on the part of a linguistically insecure announcer. Allow me to read his mind. He's thinking, "I know people usually pronounce that word [wimbleDon]. It looks like that flapping rule must have operated on the word. We Americans are so darn lazy and informal! We've corrupted another perfectly good British pronunciation and made ourselves look uneducated in the bargain. Well, I'll prove I'm an educated, sophisticated person: I'll pronounce the word as though that dirty old flapping rule had never applied to it. I'll say [wimbleTon], which is how I imagine it's really meant to be pronounced by the people who invented it!" So the poor fool bungles it and sounds worse than he would have if he had just been content with his humble linguistic heritage. (It was never [wimbleTon]—not even to the British!) No, I can't prove this is what's really going through people's minds when they come up with Wimbleton—and surely they're not aware of the rules they use—but it does fit nicely with our facts about phonological rules, and with what we've all seen of American English speakers' inferiority complex.
I remember a lowbrow example of this, too. Chris Elliott, in his brief but weird Fox sit-com Get a Life!, once cursed someone with the imprecation "Damn you to [haTes]" instead of [haDes], in a misguided attempt to appear sophisticated. (And, have you ever heard someone say, with a very formal tone, "[laTies] and gentlemen" instead of [laDies]? Again, same thing.)
The Telltale
So how do I know that NPR isn't just fancier than the rest of us? Heck, maybe all those commentators were brought up and educated in England, so that's why they talk the way they do. A reasonable hypothesis (sort of), but wrong. NPR is a hotbed of inconsistent nonflapping. It doesn't appear that there's a rule at work (such as "never flap" or "flap like normal"). There doesn't appear to be anything systematic about it at all. It's arbitrary. And this is what we'd expect from people who are winging it. Or putting on airs the best they can. Some flaps squeeze through, undetected. Let's face it—it's hard to notice all these things, and it can be exhausting trying to pay attention to them. This inconsistency is the telltale sign of affectation. The on-air personalities don't want to flap, as though there were something wrong with it, but they just can't help it sometimes.
My second hypothesis was that flapping, at least in the heads of NPR people, is no longer a postlexical rule. Perhaps the rule is becoming lexical—in other words, moving toward the level of individual words. (Lexical rules have exceptions, unlike postlexical rules.) Therefore, some words might have internal flapping, whereas others wouldn't. One thing derailing this hypothesis (besides the fact that I don't really understand it) is that NPR flapping often occurs between words (not the environment for lexical rules). It's almost as though NPR-speak favors the between-word environment for the application of the flapping rule. Again, this just sounds bizarre to me. It's like NPR invented a new category of phonological rule: the postpostlexical rule, or the cryptolexical rule, which applies ONLY between, and never within, words.
Of course, this isn't really a rule; it's just a tendency. (Perhaps it's easier to monitor our own within-word behavior, so it's easier to eradicate within-word flaps?) And on NPR some individual words have flaps sometimes, and aspirated t's other times. Even when spoken by the same person! Again, I conclude that this is a matter of affectation. Putting on airs. Carelessness in the attempt to sound more careful.
There also appears to be a sort of hyperaspiration at work here. It's not enough to fight unruly flaps wherever they appear—or wherever they're noticed. One must aspirate all t's. This leads to pronunciations like [waiT waiT, don'T tell me]. And it might, somehow, be responsible for one of KUOW's most annoying habits, referring to the station as "your news an dinformation station." Dinformation? It's as though every t or d must find its full expression! (I've even heard the word modern come out as [moTern] on NPR! Now you're refusing to flap d's and insisting on turning them into aspirated t's!)
Are there other examples of this Britishish speech at NPR? Well, yes, I can think of one. British English—at least the form of British English we Yankees think of as proper and high class—never has the [u] sound after the set of consonants known as coronals. So after n and t, for instance, you hear [yu] and not [u]. The word news gets pronounced [nyuz], instead of [nuz], and tune comes out [tyun]. NPR-speak employs this British-esque feature as well. And, again, it employs it inconsistently. Which is what drives me nuts.
Now, if you can bear a little more of my obsessive nonsense, I'd like to present some actual NPR data to you, to illustrate the preceding sermon.
The (old) Data
KUOW
12/18/2000
[waDer] as expected
[forDy] as expected
But: [concreT and waDer] (for concrete and water, where [concreD and waDer] would be expected)
KUOW
1/10/01
[forDy-eight] as expected
But: [forTy-five] where [forDy-five] expected
[uTiliTies] where [uTiliDies] is expected
[universiDy] as expected
NPR
1/22/01
[senaTor] alongside [democraDic]
NPR
2/20/01
[ciTy] alongside [ciDies]
KUOW
3/13/01
[staDe of the economy] and [universiDy of washington]
But: [staTe of washington] where [staDe of washington] is expected
KUOW
5/23/01
[universiDy of washington]
But: [social securiTy numbers]
NPR: "Wait, Wait! Don't Tell Me!"
7/14/01
[mathemaDical] as expected, alongside [activiTy]
NPR: "Marketplace"
8/17/01
[varieTy] alongside [varieDy]
NPR: "Wait, Wait! Don't Tell Me!"
9/8/01
[biTTen] (where you wouldn't expect aspiration or flapping!)
[manhatn] (as expected), but briTain (linguistic insecurity in action!)
[creaTed] and [relaTed] alongside [exciDed]
[noT officially over]
[marTin amis]
[starTed]
[auTomaTic] (two flaps that didn’t flap!)
KUOW
10/30/01
[vaniTy] alongside [mortaliDy]
KUOW
11/03/01
[compuTer] alongside [compuDer]
[thirTy] alongside [forDy]
Conclusion
Embrace the flap, NPR! Flap on, you lovers of freedom!
