Where Our Whacky English Spelling Comes From, Pt. 1: Vowels

English words mostly fall into one of three groups: native English words, loanwords from French, and loanwords directly from Latin. The loanwords from Latin are the most recent newcomers to the lexicon and have the most regular spelling; if you know Latin, a lot of the time you can pretty much tell how the English loanword will look just from the spelling. But the most common, and the most irregularly spelled, words are the native English ones.

The thing about the spelling of native English words is that it doesn’t reflect modern pronunciation. Instead, it reflects the pronunciation of Middle English, which was about different enough from Modern English to be a different language. And to understand some of the idiosyncrasies of Middle English you have to know a thing or two about Old English. So to really understand our spelling you need to be pretty well acquainted with the history of English.

But never fear! Your friendly neighborhood nerd is here to save the day!

Vowel Spelling

Old English had 7 vowels.

i, y    u
e       o
æ      a

The rows are arranged by the height of the vowels, or how closed or open they are. The first row of vowels, i, y, and u, were the ones where the tongue went closest to the roof of the mouth (closed vowels). E and o were in the middle (mid vowels), and æ and a had the tongue lowest in the mouth (open vowels). The columns are arranged by how far forward or backward in the mouth each vowel was. The vowels in the left column were in the front of the mouth, and those on the right were in the back.

Each vowel came in short and long flavors. But the thing is, in Old English, the long and short vowels were pronounced more or less the same. So how were they distinguished?

Get this. The long vowels were literally held longer than the short vowels. Who’da thunk it??

As for how they were pronounced, I guess most of them were pretty much the way you would think based on a stereotypical European language. The only really unusual ones were y and æ (called “ash,” or rather “æsc”). Y was pronounced just like i, but with the lips rounded, so that it sounded like a French u. Æ was pronounced just like the modern short a.

In Middle English, the vowel system had a bit of a makeover. The Middle English short vowels were the following:

i       u
e      o

The short æ’s and a’s merged into one vowel, and both short and long y’s became i’s. The other short vowels aren’t particularly interesting.

The long vowels are where all the fun happened.

i       ou
ee   oo
ea  oa

The two long a’s, æ and a, each closed up a bit and turned into long open mid vowels. So now we have two long e’s, a close mid one from Old English e and an open mid one from Old English æ, and two long o’s, a close mid one from Old English o and an open mid one from Old English a. So if all the a’s were morphed, why is there still a long a in the chart? That comes from a delightful little thing called “open syllable lengthening” (hereafter OSL). OSL was a lengthening of vowels that occurred in syllables that were not “closed” by a consonant. So for instance, the Old English word for “name” was “nama”; since the m in the middle of the word goes with the start of the second syllable rather than the end of the first, the first syllable is open, which means that it’s subject to OSL, hence the long a in the modern English descendant. Whereas in a word like “lick,” which was “liccian” in Old English, since there are two c’s in the middle of the word, one c went to the second syllable and the other went to the first syllable, closing it off (Old English double consonants were literally pronounced twice, just like the s’s in the Italian suffix -issimo). Thus the first syllable was not subject to OSL, hence the short vowel in the modern word. As you can see from the “name” example, the vowels that allowed OSL to take place often dropped out; this trend goes all the way back to Proto-Germanic, where the Germanic languages first acquired their distinctive stress accent on the first syllable of every word. Since then, unstressed vowels at the ends of words have been gradually becoming indistinct and disappearing. The silent e’s that we use today were actually pronounced as indistinct vowels, like the unstressed e in accelerate, in Middle English. This is why we use silent e’s to indicate long vowels and doubled consonants to indicate short vowels; the doubled consonant blocks OSL, showing that the preceding vowel must be short, while the silent e opens up the preceding syllable, indicating that it has a vowel that’s long by OSL. Of course, as time went on people got confused about which words actually used to end with vowels and which didn’t, so a silent e doesn’t always indicate that there used to be a vowel there; for instance, the Old English word for stone was “stan,” not “stona” or anything like that. By the same token, sometimes words that aren’t spelled with a silent e actually did have a vowel at the end, like steal, from Old English stelan.

Long u was written “ou,” a spelling borrowed from French, where it’s used to indicate the oo sound as opposed to the French u. When the ou would have occurred at the end of a word or before a vowel, it would be written “ow.” So technically, the ou/ow sound is long u. The sound that we moderns like to call “long u” would really more fittingly be called “French u,” since it’s an approximation of that sound. I guess the y vowel from Old English must have already merged with i by the time we got all the French loanwords, because otherwise you’d think people would just have used the y sound for the French u and we would have ended up with i’s instead of “long u’s.”

I guess it must have been because of a preference to end words with consonants rather than vowels that, like the ow instead of ou, y came to be written instead of i at the ends of words. I’m not sure when or how y came to be associated with the consonant that we think of it as today. Originally, it was borrowed by the Romans to represent upsilon (Υ/υ) in Greek loanwords. At that time, upsilon had the same sound as the French u, which is why it makes sense that it was used to represent that sound in Old English.

Oh yeah, and there’s one funny thing about OSL. Any vowel that was lengthened by OSL was also opened up. With a, that doesn’t make a difference because it’s already the most open vowel, but the other vowels would actually turn into different vowels when they were lengthened. So if an e was lengthened in an open syllable, the resulting long vowel was not ee, but ea; if an o was lengthened, the result was not oo, but oa; if an i was lengthened the result was not i, but ee; and if a u was lengthened the result was not ou, but oo. Notice that since there’s no vowel higher than i or u, it’s impossible for a long i or long u to result from OSL. Which is why it drives me crazy that long i’s and “long u’s” are always written as if they came about from OSL. Take life, for example. If there had actually been a vowel at the end of the word life, there is no way the word would have an f, because, as we’ll see later when I get to consonants, an f between vowels in Old English turned into a v. Even if there were a vowel at the end, when an i is lengthened by OSL, it turns into ee; week, for example, was wicu in Old English, with a short i becoming ee by OSL. So it is impossible for there to have been a vowel at the end of life, and even if there had been it’s impossible that i should have been lengthened to a long i. So the spelling life is really just silly on multiple levels. But such is life.

At first, the long vowels were still literally long vowels, that is, they were held longer than the corresponding short vowels. But after a while, Middle English underwent the Great Vowel Shift. This was a sweeping overhaul of the vowel system. The results were:

-The high vowels, long i and ou, underwent breaking, which is when a simple vowel is broken up into two vowels. (English has a long history of breaking, going all the way back to Old English, where vowels were regularly broken before certain consonants. Breaking is still around to this day, as in the dialects where words like cat are pronounced almost like kyat.)
-The closed mid vowels, ee and oo, closed up from closed mid vowels to bona fide closed vowels.
-After the closed mid vowels had already closed up, ea closed up to ee’s former position as a closed mid vowel, then closed further to become a closed vowel just like the new ee. Oa closed once but then stayed put, and also underwent breaking. (Incidentally, apart from the breaking and OSL, the Greek vowel situation is uncannily similar to the English one up to this point.)
-After the open mid vowels had already closed up, the long a resulting from OSL closed up to become the same open mid sound that ea had formerly represented, then underwent breaking to become the modern long a sound. As it happened, a few ea’s had escaped the change mentioned above, instead persisting as open mid vowels; since this was now the same sound as the long a, they underwent the same breaking. Break and great are two cases of this.

When all this was done with, length ceased to be the distinguishing factor between long and short vowels; now, the quality of the vowels did that instead. And that just about brings us to the modern situation.

Oh, and one other little thing. In Middle English, some u’s were written as o’s. The reason for this is that u, m, and n look really similar, especially in old-fashioned fonts, so a u next to an m, n, or w was written o to make it easier to tell the letters apart. That’s why comesome, son, wolf, and so on have a short u sound spelled with an o. This was just a spelling convention, not a reflection of pronunciation.

See how, in the lower case letters, the vertical lines are thicker than the diagonal and curvy ones? If you had a u side by side with an n, an m, or a w, it would just look like a whole bunch of vertical lines next to each other.