Engma Enigma

You may remember that a few weeks ago, I was sorting through Adam’s spelling worksheets from the past school year, looking at how they handled teaching [ə], and the cot/caught vowel merger. I was also interested to see what they did with another question about vowels:

What vowel is in the word sing, and what vowel in the word sang?

I’d never really given it much thought until I got to college and took my first linguistics class, but up until then if you had asked me, I would have said the vowel in sing was long E, and the vowel in sang was long A. Transcribing them phonetically, I’d have written them as [siŋ] and [seŋ]. (There’s a whole introduction to engma, aka the ng sound, the one written [ŋ] in IPA, in this post from the Linguistic Mystic.) In class, though, I was surprised to learn that the expected transcriptions were [sIŋ] and [sæŋ] — in other words, with short I and short A, respectively.

I refused to believe it at first. Sing definitely sounded more like seen than sin. Sang definitely sounded more like sane than San. Didn’t they? I tried to trick myself by starting to say seen, stretching out the ee and then suddenly finishing off with [ŋ] instead of [n]: seeeng. I had to admit, it didn’t sound like a proper sing, even after I was able to shorten the vowel down to a reasonable duration. It sounded like I had a really strong Mexican accent. But when I started with sin and finished with [ŋ] instead of [n], it sounded much more natural. I tried the same experiment with sang. I started to say sane and then slipped in the [ŋ] instead of the [n] to hear if it sounded like sang. It made for a passable sang, but didn’t really sound like what I was used to. Then I started saying San, and again plugged in the [ŋ] for the [n] at the end. Now that was more like it: sang.

My wife thought I must have been crazy to ever have been confused about these vowels. It was a simple case of sounding out the letters — or digraph in the case of in the ng for the engma ([ŋ]) sound. Furthermore, although none of Adam’s worksheets answered the question, I did find this online phonics resource that explicitly gives the vowel in ing as short I, and the vowel in ang as short A. And looking at it from a phonological point of view, it would be strange to have [i] and [e] before [ŋ], since no other tense vowels can appear before it: There are no English words with long O or long U before [ŋ]. So how on earth did I ever get the idea that these vowels were long E and long A?

First of all, I don’t think my phonics lessons ever talked about it. I think, like the worksheets of Adam’s that I saw, that they just taught the vowel+ng sequences as a chunk. Second, it’s easier to confuse [i] with [I], and [e] with [æ], in an environment where you only get one. With seen and sin, it makes a difference whether you utter [i] or [I]. With sane and San (or bane and ban, if you don’t like the borrowed word San), it makes a difference whether you say [e] or [æ]. But there are no minimal pairs like that that end in [ŋ]. You could say one, the other, or something in between, and still be understood, and in fact, that’s what people do. Third…

…as it turns out, I’m not the only one to have heard ing and ang with long E and long A. Doug identified them as such when I asked him. And in a Phonoloblog post from 2007, Eric Bakovic says:

One thing I’ve always found hard about teaching English phonetics is convincing my students that the high front vowel preceding [ŋ] is (lax) [ɪ], not (tense) [i:]. It’s not hard to convince them that there’s no contrast between the two in this context, but no matter how many spectrograms I show them, they’re convinced that the vowel is more like [i:] than it is like [ɪ] — i.e., that bing sounds more like bean than like been. The biggest problem is that I can’t say that I disagree, no matter how many spectrograms have “convinced” me to the contrary.

During the past month on the American Dialect Society mailing list, they’ve been going around and around on the issue, with one contributor doggedly maintaining that he pronounces the word Chinglish as “Cheenglish”. Other contributors have been sure that he’s wrong, but Herb Stahlke did concede that:

It’s pretty well known in phonetics that a syllable-final [engma] raises a high front vowel a bit, so the lax high front vowel of “sin” and the vowel of “sing” are not phonetically identical. (link)

He still doesn’t agree that it’s a long E, noting as I did above that no other tense vowels can come before [ŋ]. Matthew Gordon, however, raised this point:

Anyway, it seems to me that the real question here isn’t whether the vowel of wing, sing, king, etc. is PHONETICALLY identical to that of win, sin, kin or to that of ween, scene, keen. It obviously isn’t phonetically identical to either. The only interesting question (for me, and even I’m rapidly losing interest) is whether the vowel is PHONEMICALLY part of the /I/ class or the /i/ class. … [I]f we take seriously native-speaker intuitions, doesn’t the fact that many people identify the pre-velar nasal vowel as the same as other clear cases of /i/ argue that the allophone has been reallocated to /i/? (link)

Finally, David Peterson has an interesting paper on this issue, showing that in Southern California (where Bakovic teaches, incidentally), both [æ] and [I] are tensing up before [ŋ] to become [e] and [i]. He notes that:

When classifying the vowels of their own language, … it is unclear how native speakers would classify the vowels before velar nasals. (p. 49)

So what’s your judgment? What vowels do you have in sing and sang? Did you learn them in school, or arrive at your own judgment based on the spelling, or arrive at your own judgment based on the sound?

: : : : : : : : : : : :