The lost letters of the English alphabet
And why they didn't survive
Westminster, AD 1476.
The scent of ink and slowly cooling lead hangs heavy in William Caxton’s London workshop. The hustle and bustle of the street outside can hardly be heard against the rhythmic, almost hypnotic, thud of the greatest technological advancement of the age: the printing press. The Londoners going about their daily lives have no idea that a revolution is happening behind Caxton’s wooden doors.
A young German apprentice named Wynkyn de Worde sweeps up as his master examines a freshly printed page of Chaucer’s Canterbury Tales, the ink still glistening on the rag paper.
As Caxton examines the page, a frown crosses his face. It does not go unnoticed by his apprentice. Masters’ frowns rarely do.
“I still don’t like it,” growls Caxton.
“Master?” asks de Worde.
“Come here and see,” says Caxton, handing him the page
De Worde reads out loud, “For joy he thought he clawed him on… ye back.”1
“On the bak!” says Caxton.
De Worde begins to see Caxton’s point: English used one very strange letter called “Thorn.” It was written like this: ‘þ,’ and it represented the equally strange sounds the English said in words like think, three, and the… or, as they usually wrote it, þe.
But when Caxton had bought his printing press — and not cheaply — in Flanders, there was no thorn in the type case. The closest thing they had was the letter ‘y.’2
Caxton sighs. “Ye. Even after all this expense, I’m not able to produce anything more than a crude imitation of the simplest English word — and the most common.”
“Could we not make a new sort,3 Master?” De Worde asks, his voice struggling to lift the gloom that had settled over the workshop. “A proper Thorn?”
Caxton runs a hand over the ink stains on his apron and begins to pace about the room. De Worde can guess what is going through his master’s mind: the cost, the time… a wrestling match between purist and pragmatist playing across the man’s face.
De Worde knows the pragmatist will win in the end. He imagines what it would take to make a proper Thorn: the carving of the punch4 would require the labour of a goldsmith for days. More expense.
At last, Caxton returns to the page and speaks: “It doesn’t look so bad, does it?”
“Not at all, Master.”
De Worde understands it as well as Caxton: the press, the first to make its way to these shores, must turn a profit. And if that means that the English language has to adapt to fit into a continental mould, so be it.
The English alphabet — the 26 characters that schoolchildren learn across the world — might seem fixed and eternal. But, like the English language itself, the English alphabet is also a living thing. It has changed over the centuries, innovating new letters and discarding the old along the way.
The Thorn (þ) that so troubled our dear Mr. Caxton is a good example of what English left behind. But it’s not the only one. Let’s take a look at all the letters English has lost over the years — and what has replaced them.
You're reading The Dead Language Society. I'm Colin Gorrie, linguist, ancient language teacher, and your guide through the history of the English language and its relatives.
Subscribe for a free issue every Wednesday, or upgrade to support my mission of bringing historical linguistics out of the ivory tower and receive two extra Saturday deep-dives per month.
This Saturday, we’ll be exploring the long history of schemes to reform English spelling — and why they have (almost) all failed. I’ll explain the linguistic and historical reasons that English spelling is so difficult, as well as the factors which have kept it that way despite 900 years of attempts to fix it.
Adapted alphabets
Before you meet the lost letters of English, however, I’d like to reacquaint you with the 26 that we still have. This alphabet comes to us in English from the Romans — hence its full name, the Roman or Latin alphabet. And they themselves got it from the Etruscans, who got it from the Greeks, who got it from those clever Phoenicians, the inventors of the alphabet.
And with every new language that employed the alphabet, it changed. The Phoenicians didn’t write vowels — their alphabet wrote only consonants. But the Greeks were very keen on writing all of their sounds, vowels included. So, to represent Greek vowels, the Greeks repurposed certain Phoenician letters that the Greeks had no need for.
For example, the Phoenician letter heth ‘𐤇’, which represented a consonant made far back in the throat, was repurposed for the Greek letter eta: ‘Η.’ Since the Greek language didn’t have the heth sound, it was free to reuse it for a vowel sound, which is a long ‘eh’ sound (the same vowel as in English bed).5
The Etruscans didn’t distinguish ‘k’ sounds from ‘g’ sounds, so they used one letter ‘𐌂’ for both. Then the Romans, who did make this distinction, added a slash to invent the letter ‘G.’
These adaptations were necessary because each language has its own sound system, which means an alphabet designed for one language won’t always work for another.
For instance, Latin distinguished five vowels — perfectly represented by the Latin alphabet’s five vowel letters: ‘a e i o u.’ Modern English distinguishes between 14–25 vowel sounds in pronunciation, depending on the dialect and how you count them; hence the need for ‘ee,’ ‘oo,’ ‘ea,’ ‘ai,’ etc.
Other languages have solved this same problem by adding accents, or, to use the technical word, diacritics: think of French ‘è’ or German ‘ü,’ each of which represents a vowel not in the original Latin set of five.
This pattern of adaptation would play out once more when Latin letters encountered English speech. This happened beginning in the 7th century AD, when the Latin alphabet first came to be used for English as a result of the Christianization of England. This religious conversion was undertaken (partly, at least — we’ll return to this) by the Roman Church. So the alphabet came along as part of a package that changed England, and the English language, forever.
The runic legacy
In the centuries before Christianity, English had been written — when it was written, which was not very often — in a runic alphabet called the Anglo-Saxon futhorc, named after the sounds of its first few letters. This runic alphabet was a pretty good match for the sounds needed to write Old English.
The Latin alphabet was… less good. There was, for example, no letter for the ‘th’ sound in the ancestors of words like think or three. Our solution for this sound in Modern English is to write it with a combination of two letters: ‘t’ + ‘h.’ But, really, the sound ‘th’ isn’t a ‘t’ or an ‘h.’ It’s entirely its own thing.
The writers of Old English chose a different solution: they just brought the rune for the ‘th’ sound, which was called thorn, into the Roman alphabet. The thorn rune looked like this: ᚦ — but with a Roman makeover, it looked more like this: þ.
Another sound the English had trouble writing was the ‘w’ sound used in the ancestors of the words wood, well, or weather. This wasn’t a sound that Latin had,6 so it didn’t appear in their alphabet, even though it was hugely important in English. But never fear: there’s a rune for that, too!
This ‘w’ rune was called wynn, which meant ‘joy.’ It looked like ᚹ in the Runic alphabet, but its romanized version looked like this: ƿ. If you read Old English manuscripts, you’ll see a wynn on every page.
Anglo-Saxon innovations
There was another Old English sound that the Latin alphabet had no way to handle — it’s a sound that we have in most dialects of Modern English today: the short ‘a’ in cat or hat. In Modern English, we use ‘a’ for this, but that wasn’t an option in Old English: the ‘a’ was already taken for another ‘a’ sound, one that sounded more like the long ‘a’ we use in words like father.7
Anglo-Saxon scribes could have taken a runic letter for this too, since there was one: ᚫ, called æsc, meaning ‘ash (tree)’ and pronounced the same way. But they didn’t do that.
Instead, their solution was to take the two letters which represented sounds close to this very English ‘a’ sound — that is, ‘a’ and ‘e’ — and mash them together to produce the letter ‘æ,’ which they called the same name as that runic letter: æsc, or, in Modern English, ash.
The Old English alphabet added one more letter to the Roman alphabet, but this one is different from the others, since it was neither a runic import nor strictly necessary for writing the language. This was the letter ‘ð,’ called today eth, with the ‘th’ sound of either rather than the ‘th’ of ether. The Anglo-Saxons just called it ðæt ‘that.’
As you might have guessed from its name, the letter ‘ð’ also makes the ‘th’ sound in Old English — which is why it’s not, strictly speaking, necessary. Any word you could spell with an eth, you could also spell with a thorn.
In its origin, ‘ð’ was a modified version of the Roman letter ‘d.’ It wasn’t actually part of the original adaptation of the Roman alphabet to writing Old English, but it became popular in the 8th century before fading in use over time. Nevertheless, both ‘þ’ and ‘ð’ appear in most Old English manuscripts, with no difference in sound.8
So if ‘ð’ is technically unnecessary, why does it exist? Well, the Roman Church wasn’t the only way Christianity came to England. While the Roman Church sent missionaries who operated out of Kent, the Irish Church was also evangelizing in the north of England.
So literacy in England and, therefore, English spelling conventions, have dual origins: Roman and Irish. Admittedly, the history here is murky, but it seems as though the ‘ð’ character may reflect the Irish side of that influence.
These four Old English letters were each used for a sound we still have in Modern English (although ‘þ’ and ‘ð’ are slightly redundant). And yet we use none of them today. What happened to them?
Letters lost and found
Ultimately, thorn, wynn, ash, and eth were lost from the English language largely because they were so particular to the English language,9 and only English scribes were trained to write them. This becomes important because of two major events in the history of England, each of which forced the English language to conform to continental European writing practices.
The first of these was the Norman Conquest of AD 1066. In brief, this occurred when William of Normandy (aka William the Conqueror) took the English throne by force. At a stroke, the language of power in England became French and stayed that way for three centuries. Scribes trained in France were brought over to England and replaced the scribes who wrote according to English customs.
The Norman scribes, for example, used a double ‘u’ for the ‘w’ sound, which eventually got written together, and turned into the single letter that we call “double ‘u’” — ‘w.’
As for the ‘th’ sound that the Anglo-Saxons had spelled ‘þ’ and ‘ð,’ it was respelled ‘th.’ This is how Norman scribes were used to spelling the sound, which also existed in Old French, in words like feith ‘faith.’
The letter ‘æ’ became unnecessary without any French help, as the English language lost the sound it represented. The two kinds of ‘a’ sounds in Old English (like hat vs father) collapsed into one, which was just spelled ‘a.’
The letter ‘þ’ did, however, stick around for longer than the others, although it was used less and less as the years went on — that is, until a second event happened that would bring English spelling in line with continental customs: the arrival of the first printing press in England in 1476.
As I alluded to in the opening vignette, the printing press was brought to England from Flanders. As a result, it only had the characters — the individual letters — needed to print Latin and other continental European languages such as German and French. The physical limitations of the printing press, and what was probably the significant expense of customisation, had the effect once again of making English spelling more like that of continental European languages.
But don’t worry: English spelling would get very weird again very quickly, in no small part because of the printing press.
At any rate, there would be no thorn for Caxton’s press — although he may not have been as upset about it as I portrayed him to be.
And, because of this, there would be no thorn for English anymore either: because it couldn’t be printed, it dropped out of the language. Although this was just the death stroke: by the late 15th century, ‘þ’ had already dropped out of use in most words. It was mostly just used in a few common words, such as þe — that is, the. So it wasn’t the end of the world to respell þe as ye.
This is the origin of the Ye in the faux-medieval Ye Olde… found on the signs of pubs and souvenir stores the world over (which, yes, would have actually been pronounced “the”).
But it’s not all loss: English has gained some letters too over the years. The replacement of wynn — our ‘w’ — is a perfect example. The letters ‘j’ and ‘v’ are fairly recent additions too: earlier forms of English spelling made no distinction between the consonant ‘j’ and the vowel ‘i.’ Similarly, the consonant ‘v’ and the vowel ‘u’ were both spelled the same. Up to the 15th and 16th centuries, you could read spellings such as ioie (pronounced ‘joy’) and loue (pronounced ‘love’).
The history of these lost letters shows us the forces at work shaping our writing systems. Most languages receive their writing systems from the outside — war, religion, or technology all play a role. Rarely is an alphabet chosen because it’s the best way to represent all the sounds in a language.
And yet, writing finds a way.
The history of writing is the history of human ingenuity in the face of traditions which aren’t always perfectly adapted to the moment. From ancient Greece to medieval England (and beyond!), scribes, and later, printers, have invented, borrowed, or made creative reuse of old material. And, gradually, they have devised ways of writing that work with existing technology and the expectations of their cultures… even if the result sometimes frustrates.
Ye end.
for ioye he toughte he clawid hym on ye bak
When written by hand, ‘þ’ looked a bit like a ‘y,’ as it was often written without connecting the loop to the line.
A sort is a block with a character written on it. When a page was printed, all of the characters used on the page had to be removed and placed back in the right place in the type case. In other words, they needed to be sorted: hence the name sort for one of these characters.
A punch was a piece of metal with a character on the end of it. This was the first stem in making type — the punch would stamp the shape of the letter into a softer metal matrix, which acted as a mould. The sorts were then cast from this matrix.
‘Η’ did, however, represent the consonant ‘h’ in western versions of the Greek alphabet, including the one that the Latin alphabet is based on. But that’s another story.
Latin actually did have this ‘w’ sound hundreds of years before, during the classical period. It’s usually spelled ‘v’ in modern editions of Latin texts, but it was pronounced ‘w.’ So, yes: Caesar would have pronounced it weni, widi, wici. But, by the 6th century, the Latin pronunciation of this sound had changed to something more like a ‘v.’
Obviously, we don’t have any problem with using the same letter for both of these sounds today, but, as I’ve written about elsewhere, that kind of chaos is typical for Modern English spelling.
There is a difference in where ‘þ’ and ‘ð’ show up within a word, though, with ‘ð’ preferred after vowels and ‘þ’ elsewhere. So you’d likely write þancian ‘thank’ with ‘þ’ but baðian ‘bathe’ with ‘ð.’
Particular but not totally unique: modern Icelandic has ‘þ’; Icelandic and Faroese have ‘ð.’ The letter ‘æ’ is also found in modern Icelandic, Faroese, Danish, and Norwegian.
The sound of “th” in either and ether are two different sounds — one voiced and one unvoiced. We only think of them the same because we write them the same.
This is a nice summary, but is there a reason for not mentioning yogh?