“My dog,” Jim Taylor writes, “a Chesapeake Bay Retriever, says ‘woof’ or perhaps ‘wuff’ – I’m not completely sure of the central vowel. But I’m quite sure there is an opening diphthong ‘woo’ and a closing fricative ‘ff’.
“But physically, anatomically, I don’t understand how a dog can make those sounds. The fricative requires an upper lip against the lower teeth; a dog doesn’t have an upper lip. The ‘woo’ should require pursed lips; a dog doesn’t have lips it can purse.
“So how does it make those sounds?”
Ah, acoustic phonetics. It’s a fascinating area, and one that can make many people nervous and confused pretty quickly (see my note on cepstrum for a wee taste). But the language we speak and hear is a complex fabric woven of many threads. And actually it’s amazing that we can understand what we hear, in fast speech and slow, casual and careful, spoken and sung, in different voices and different accents; a lot of it has to do with what we expect to hear, and what sounds are reasonable combinations in a given language and context.
But first: It’s obvious that a big dog says woof, right? You can hear them say it. People have been writing it down as woof since… well, at least since the mid-1800s. Hmm, and what did dogs of that sort say before then? Well, there are some slightly older whooghs attested. And, uh… hmm… well, some other things before that.
The thing is, in any given language, there are phonemes – sounds that are recognized as being distinct sounds. And these aren’t crisply defined things; they’re more like regions, buckets, circles, subdivisions of the possible sound from the articulatory space of the mouth. Any two languages will divide the articulatory space up at least slightly differently. Sounds that are heard as different in one language may be heard as the same in another, and vice versa.
One of the first things we do when we start to learn language in our infancy is to learn what the buckets are to sort sounds into. We come to understand that one pair of sounds are treated as different sounds, while another pair no less different are treated as the same sound. And we come to actually hear them as the same, especially if we’re not paying close attention; we also tend to expect certain sounds in certain places, as some words will be more likely than others in any given place. This is why speakers of some languages have so much trouble distinguishing between English pairs such as bit and beat.
And it’s why Shakespeare represented the Welsh name Llywelyn as Fluellen. And partly why when the voiceless velar fricative (that we hear in German ach) disappeared from English, in some words it was just dropped but in others it became a [f] sound (think of the words that end in gh). And why the voiceless bilabial fricative in Maori, which is spelled as wh, sounds like [f] to us. And why some Anglophones have so much trouble with the vowel in German Tür and French tu. And so on.
And, of course, why a sound that seems so clearly [f] to your ears and mine might sound like something else to someone else. We sort it into the buckets available, and if it doesn’t fit neatly into one or another there may be differences of opinion on what bucket it best fits into. So the spelling of animals’ sounds varies from time to time and from place to place.
And then there’s the question of how a dog, which can’t round its lips, can make a pretty clear [w] sound and a certainly distinguishable mid-high back rounded vowel [ʊ] sound.
But, really, what about a sound conveys roundness of lips? In the shape of the sound wave, what do you suppose it might be? When we think about it, we must acknowledge that speakers (as in the ones on your stereo) make all these sounds without lips to round, and you can make a saw sound like a human singing, and then there’s the wah-wah pedal you can use for an electric guitar…
The number one thing to understand about the sounds we make is that they are not simple even sound waves. A violin, a piano, and a voice all sound different because of the shapes of the sound waves and the different resonances they have. Those resonances involve structures of harmonics. If you hear an A at 880 Hertz, unless it’s produced by some electronic sound generator and heard in a non-resonant environment, you will also be hearing a structure of resonances at various multiples of 880 Hertz. Any resonating space that can fit one sound wave of a given length can fit two of half the length, three of a third the length, and so on.
The shape of the resonating space has an important effect on what resonances come through. Now, what affects the shape of the space in your mouth? The movement of your tongue and your lips. There are two main resonating areas, determined by where your tongue constricts your mouth: one is between the larynx (voice box) and the point of constriction, and the other is between the point of constriction and the lips.
If you take a speech sound and analyze it acoustically, you can get a thing called a spectrogram. It looks a bit like an unevenly made fabric, a rather blurry one; the x axis is time, and the y axis is frequency, and what you see is areas of certain frequencies that are very dark, meaning strong, and others that are very light, meaning weak. They look like bands of fuzzy threads going across at certain points. There will be two main ones you will see, and more above them. The lower one is called formant 1, or F1, and is mainly the resonance from behind the tongue. The higher the tongue is, the larger that space is and so the lower the F1 is. The one above it is of course formant 2 (F2), and is mainly the resonance in front of the tongue. The farther forward the point of constriction is, the smaller the space, and the higher the F2 is.
So, in brief, low F1 and F2 means the sound is like [u]. High F1 and F2 means the sound is like [æ]. High F2 and low F1? That’s [i]. And so on. Yup, we follow the thread of speech sounds by following the dark threads woven across the tapestry of harmonics.
Oh, and the effect of rounding the lips? Well, that constricts the sound wave at the opening, where it would normally be fullest, and so it lowers all of the formants, including the ones above F1 and F2. The higher formants are much fainter, but F3 does have something of a role to play too – otherwise lip rounding would be entirely equivalent to shifting the tongue up and back.
The point being, anything that produces that harmonic profile will seem to have that sound. How does a wah-wah pedal work? Basically by varying between emphasizing lower harmonics and higher ones. It’s really just an adjustment of the equalization (you know, like fiddling with the sliders on a higher-end stereo or sound board). It’s really the contrast between the sounds (as it slides in a “wah”) that leads you to hear it as a contrastive vowel sound.
And how does the dog’s mouth produce the “woof”? Well, I can’t say exactly what, in the shape of a dog’s mouth, would produce that sound. Alexander Graham Bell probably knew. He used to demonstrate speech sound articulation and production by manipulating a dog’s mouth with his hands.
Woof isn’t just the sound a big dog makes, by the way. Nor may we limit ourselves to adding the sound a saw makes (when sawing wood, not when singing), or the noise a pile of gas-soaked rags makes when ignited, or the sound of a strong gust of wind through a window. We may move quite away from onomatopoeia, to weaving.
On a loom, you see, there are two sets of threads. The ones that are perpendicular to the weaver, attached to the frame, are the warp, a word that used to mean “throw” (the word that throw comes from originally meant “twist”, so the two words have pretty much changed places semantically – how they did so is a tale for a whole other note). The ones that run cross-wise, like the formants on a spectrogram, are called the woof, a word coming from the same old Germanic root as weave.
And the weaving is done with the aid of a shuttle, which is thrown back and forth between the threads. You might or might not see something in common between it and the gesture your mouth makes when saying “woof”. Say it a few times and you’ll see how it starts with the tongue tense and the lips forward, and then the lips pull back and spread and the tongue at the same time lowers. As you repeat the word, the whole assembly of your mouth moves back and forth like a shuttle or a saw.
Whatever a dog’s mouth is doing, though, I guarantee it isn’t that. But it doesn’t really need to be, either.