Language

Why English spelling doesn't match how words sound

30seconds
The 30-second version

Around 1400, English spelling roughly matched speech. Bite sounded close to "beet", meet roughly like "mate", and house like "hoose".

Then, over about three centuries, the long vowels of English moved up in the mouth one after another. This was the Great Vowel Shift. At the same time, printing arrived with William Caxton in 1476, and spellings began to settle. The letters stayed where they were while the sounds kept moving.

Add French scribal habits, a Flemish h in ghost, and Renaissance scholars who put a silent b into debt, and you get modern English spelling.

Deep dive · 5 min
Deep dive · 5 min

Letters that stopped moving

Say bite, meet and name out loud. The letter i makes an "eye" sound, the double e makes an "ee" sound, and the a makes an "ay" sound. In most European languages those letters stand for "ee", "ay" and "ah". English looks as if it borrowed the Latin alphabet and then scrambled the vowels.

It didn't. Around 1400 the letters meant roughly what they mean in Italian or Spanish. Bite sounded close to modern "beet". Meet sounded roughly like "mate". Name had a broad "ah", nearer "nahm". The spelling was a fair record of speech. Then two things happened at the same time: the spelling began to set, and the sounds kept moving.

The Great Vowel Shift

Between about 1400 and 1700, the long vowels of southern English slid upwards in the mouth, one after another. Linguists call it the Great Vowel Shift. Vowels made with the tongue low moved higher. Vowels that were already as high as they could go, "ee" and "oo", had nowhere left to climb, so they broke into two-part sounds (diphthongs).

  • bite: "ee" became "eye".
  • house: "oo" became "ow". It was hūs in Old English. German had a similar change and respelled the word as Haus. English kept its spelling.
  • meet: roughly "ay" rose to "ee".
  • boot: roughly "oh" rose to "oo".
  • name: "ah" rose, step by step, towards "ay".

Nobody decided this. It spread over several lifetimes, unevenly across regions and social classes. No speaker would have heard their own grandparents as strange. After three centuries, though, most long vowels in the language had changed, and the letters written for them had not.

Try it: the spelling stays, the sound moves

Pick a word and drag through time. The spelling above stays the same. The vowel under it changes. The chart shows where in the mouth the vowel sits: front to back across, tongue high to low down. The faint trail is the route that word's vowel took.

Things to try:

  • Compare meet and meat. In 1400 they had different vowels, which is why they are spelled differently. Meat caught up later. Great and break never did, so they still have the "ay" sound.
  • Watch bite and house: they are mirror images. One starts at the top front of the mouth, the other at the top back, and both end up low with a glide.

Printing froze the snapshot

In 1476 William Caxton set up England's first printing press at Westminster. He had learned the trade on the Continent and printed his first English book in Bruges. Printing needed consistent spellings, and it multiplied them in a way hand-copying never had. Printers mostly followed London usage and the habits of the scribes before them. As spellings settled during the 1500s and 1600s, they kept Middle English vowel letters, just while those vowels were moving.

Printing also added some oddities of its own. Caxton's workmen included compositors from the Low Countries. Ghost was gost in Middle English. The h probably came from Flemish or Dutch habit (compare Middle Dutch gheest) and it stuck. Several other gh- spellings from that period later dropped the h.

Older layers: French scribes and Latin show-offs

The mismatch didn't start with printing. After the Norman Conquest of 1066, many scribes were trained in French conventions. They replaced Old English cw with qu (cwēn became queen), wrote sh for sc (scip became ship), and used ou for the long "oo" sound (hūs became hous). Handwriting played a part too. The letters u, n and m were written as rows of identical short strokes, so a u next to them was hard to read. Scribes often wrote o instead. That is why come, love and son have an o where you hear "uh".

Then came the Renaissance. Scholars who admired Latin added letters to show where words came from, even though nobody pronounced them. Middle English dette and doute became debt and doubt, after Latin debitum and dubitare. The letter b had never been pronounced in either English word. Sometimes they got the history wrong. Island comes from Old English īegland and has nothing to do with Latin insula. It gained an s in the 1500s through confusion with the unrelated word isle.

So what is English spelling?

It is partly a record of speech from around 1400, with French scribal habits, some Flemish typesetting and some Renaissance guesswork mixed in. That makes it bad at telling you how to pronounce a word. It is often better at showing where the word came from and which words are related. Sign keeps its silent g, and you hear it in signature. The spelling of bite works less like a sound recording and more like a fossil: it keeps the shape of a word as English speakers said it over 500 years ago.

More in Language

View all →
Why a language gets grammar when nobody designs it: the iterated-learning experiment How does a language get grammar when nobody designs it? In a 2008 experiment, volunteers learned an alien language of 27 random words for pictures, saw only 14… Language4 min read How babies find words in a stream of sound: transitional probabilities Babies hear speech as one unbroken smear, with no gaps between words, yet somehow they start picking out words long before they can say one. One clue sits in… Language4 min read Why Russians have two blues: how a colour word can speed up your eyes Russian has two basic words for blue, goluboy (light) and siniy (dark), and in a 2007 experiment Russian speakers were about 124 milliseconds faster at telling… Language4 min read Why a sentence can nest forever on paper but not in your head: centre-embedding The rat the cat the dog chased killed ate the malt. It is perfectly grammatical, and almost nobody can read it. English lets you tuck one clause inside another… Language4 min read Why 'The old man the boat' stops you cold: garden-path sentences and a brain that guesses Read "The old man the boat." and your brain probably crashes at the fifth word. It guessed "the old man" was one person and waited for a verb that never came… Language4 min read Why Latin p became English f: Grimm's law and the sound-shift chain Latin says pes, English says foot. Latin tres, English three. Latin cornu, English horn. The swap is not random: when the Germanic languages split off, their… Language5 min read