I’ve always been fascinated by the way languages other than English use letters other than our familiar 26—not so much completely different alphabets, like Greek, Arabic or Cyrillic, but those little tweaks to the Latin alphabet, ranging from unusual diacritical marks to additional letters, that other languages use to communicate particular sounds to their readers. So this is the first of what may or may not grow into a series of posts entitled “Letters From Abroad”—and it deals with the letter ŋ, which is called (among other things) an eng.

Eng symbolizes the nasal sound most commonly indicated by the letter pair “ng” in English—technically called a velar nasal, it’s the sound we make at the end of words like sing and thing. Less obviously, it’s also the sound we usually give the letter “n” when it comes before a “k” or a “g”—as in pink or bingo. If you try to say pink or bingo with a pure “n” sound, you’ll be aware that your tongue has to do a little leap backwards between the “n” and the next sound. The “n” is an alveolar nasal sound, with the tongue tip pressed forward behind the teeth while air flows through the nose, whereas “g” and “k” are velar stops, with the tongue moving back to block the airflow against the soft palate—hence the sense of rapid tongue movement if we try to sound an “n” followed by a “k” or “g”. In normal speech we tend to let one sound slide into the next, and so when pronouncing pink and bingo normally we just keep our tongues against the soft palate and cycle smoothly from the velar nasal to the velar stop. Since ŋ is the phonetic symbol for that velar nasal sound, we can say that pink is pronounced /pɪŋk/, not /pɪnk/.

The reason we ended up with the odd letter combination “ng” to symbolize the velar nasal sound in English is because we used to pronounce the “g” at the end of words like sing. So exactly the same thing happened to the “n” when we tried to say “sin-g” as happens to the “n” in bingo. We said /sɪŋɡ/, not /sɪnɡ/. And then, after a while, when English dropped the terminal hard “g” sound, we hung on to the velar nasal pronunciation and the original spelling. *

Long ago, the Ancient Greeks noticed the velar nasal sound that occurred before their velar stops, symbolized by their letters gamma (γ) and kappa (κ). They marked the eng sound in these locations with a letter gamma, so that γγ was pronounced /ŋɡ/ and γκ, /ŋk/. Latin writers tell us that the Greeks had a special name for a letter gamma when it was performing this duty—with a simple sound transposition, they called it an agma. That word was adopted in English as a name for the velar nasal sound, but by association with the sound it symbolized, it has mutated into angma.

Eng is an old word, dating back to at least the twelfth century, when it appears in a treatise on the sounds of Old Norse, Fyrsta Málfræðiritgerðin (First Grammatical Treatise)—in that text it’s associated with the velar nasal sound, but assigned a completely different symbol, like a letter “g” with a bar through it. The etymology of eng is probably simply imitative, like other letters of the alphabet. And it seems to have had an influence on angma, which sometimes appears in a hybrid form, engma. So we have actually four words for the same sound, but eng is the word that’s nowadays most often associated with the symbol ŋ, so that’s the name I’m using here.

So much for the sound and its name. Where did the symbol ŋ come from? William Holder seems nearly to have got it into print in 1669, but his printers let him down. In the errata to his book Elements Of Speech, he wrote that:

… there was intended a Character for Ng, viz. n with a tail like that of g …

But the printers (not having such a character in their type-box) had replaced his carefully crafted manuscript symbol with a random selection of  n‘s and y‘s.

Benjamin Franklin seems to have been the first person to both usher the eng character into print and propose it as a new letter of the alphabet (rather than just a symbol for a sound), in his “Scheme for a new Alphabet and a Reformed Mode of Spelling” which appeared in his essay collection Political, Miscellaneous and Philosophical Pieces (1779).

And it has been hanging around as a potential letter of the alphabet ever since. It featured (in a slightly curlier form, emphasizing its relationship to both “n” and “g”) in James Pitman’s Initial Teaching Alphabet, which enjoyed a brief vogue in UK “progressive” schools during the 1960s, ensuring that some of my contemporaries had to learn to read twice (and that some have been unable to spell since).

So the eng never made it into the English alphabet. But it turns up elsewhere:

This sign is on Easter Island, and it’s a new one. An older sign for the same placename looks like this:The older Rapa Nui alphabet used “ng” for the velar nasal sound, but that alphabet was a colonial inheritance, and the digraph makes little sense, particularly since the Rapa Nui language has no other use for the letter “g”. But at the turn of the millennium there was a resurgence of interest in written Rapa Nui, with textbooks and a newspaper being produced for what had previously been largely a spoken language, and the opportunity was taken to rationalize the spelling, producing a new alphabet that ditched the colonial “ng” and replaced it with an eng. In the example of Oroŋo, above, it dissuades English speakers from reading the name of the place as O-ron-go, with a hard “g”, and reminds us to say O-rong-o instead. (Although, in a hair-tearing moment on Easter Island, I overheard an English couple solemnly deciding that “ŋ” was a combination of “n” and “j”, after which they carefully repeated O-ron-jo to each other a few times, nodding sagely.)

Elsewhere, too, old orthographies are being discarded, and native speakers are replacing them with character sets that better suit their purposes. Alphabetical restrictions that are inherited, in many cases, from the keyboard of a missionary’s typewriter, are being replaced using the rich resources available from modern electronic typesetting. So the eng now appears in a slew of African languages, extending in a line from Senegal to Sudan (with a couple of outliers in Zambia). The other big eng consumers are the Sámi languages of the indigenous people of northern Scandinavia. And there are a scatter of others, from Australian Aborigines to Alaskan Inuit. The Native American Lakota alphabet also uses the eng, but in a different way from the other languages. It only ever appears after a vowel, and it indicates that the preceding vowel is nasalized—so the eng is performing the same function in Lakota as the tilde accent has in Portuguese.

Finally, there’s the matter of the capital eng. It comes in two styles, which appear at the head of this post. You can either preserve the style of the lower case letter, but make it bigger, or you can add a curved tail to a conventional capital “N”. These are called the “n-form” and “N-form” respectively. Which form you get depends on the font you’re writing in—at the head of the post, Times New Roman on the left is n-form, but the same text converted to Cambria on the right is N-form.

A factoid that echoes around the Internet claims that the African languages take the n-form capital, whereas the Sámi take the N-form. But in practice, there’s not much evidence of this. Electronic publishing often means that the writer has little or no control over the typeface seen by the reader, and it’s possible to find African websites sporting the N-form, just as one can find Sámi texts using the n-form. The distinction, if it ever existed, is certainly fading.

Every now and then, though, problems occur. Sometimes the typesetter doesn’t have either the character set or the patience to hunt down an eng. So occasionally another similar character is pressed into service. Here’s another example from Easter Island, in which some despairing Rapa Nui sign-maker has pressed a Greek lower-case eta (complete with diaeresis) into use, enlarging it to match the capital letters on either side:

* In fact, we didn’t all drop the “g”—in parts of England, notably around Liverpool, Manchester and Birmingham, people still say /sɪŋɡ/ rather than /sɪŋ/, a pronunciation that’s been called “velar nasal plus”.