Alex Steer

973 words | ~5 min

On Wednesday, the wonderful elves at Qikipedia (the Twitter presence of QI) announced that:

The word 'twitter' was first used by Geoffrey Chaucer in 1374.

Now, this is obviously the sort of thing to make lexicographers (even washed-up former lexicographers) sit up and take notice. So let's get the big, obvious and pedantic problem out of the way first.

If you look at the OED's entry for twitter, v.1 (originally published in 1926; included in the 1989 Second Edition), the first quotation in the first sense ('intr. Of a bird: To utter a succession of light tremulous notes; to chirp continuously with a tremulous effect.') is:

1374 CHAUCER Boeth. III. met. ii. 54 (Camb. MS.) The Iangelynge bryd..enclosed in a streyht cage..twiterith desyrynge the wode with her swete voys.

There are no other quotations that seem to antedate 1374, so Chaucer appears to claim the prize for earliest use.

But that doesn't, of course, mean that he invented the word. Just that his translation of Boethius's Consolatio Philosophiae (normally known as Boece) is recorded in the OED as containing the earliest example of the word that had been found by 1926. This is one of those things that lexicographers will tell you until they die: earliest citation does not necessarily mean invention.

There's another problem, though. A lot of work has been done on medieval books and writing since 1926, and a lot more has been discovered. Language changes, but so does our knowledge of language in use. That's one of the reasons why the OED is being continuously updated, with new and revised entries now being published online every quarter. (See the website for information.) It's common, with the tools now at lexicographers' disposal, to find earlier examples (known as 'antedatings') for words.

In the case of twitter, no new example has been found. Instead, something more complicated has happened to knock Chaucer off his perch. Welcome to the surprising world of historical bibliography...

The second quotation given for the first sense of twitter is:

1387 TREVISA Higden (Rolls) I. 237 Þe nytyngale in his note Twytereþ wel fawnyng Wiþ full swete song.

For those not familiar with Middle English, the weird 'þ' character is called 'thorn', and pronounced 'th'. For those baffled by the OED's slightly cryptic citation, this is the translation by the fourteenth-century Cornish vicar John Trevisa of the Polychronicon, a history of the world by Ranulf Higden, a Benedictine monk from Chester. We know exactly when Trevisa finished his translation, because he noted it at the end:

God be {th}onked of al his nedes {th}is translacioun is I ended in a {th}orsday {th}e ey{gh}te{th}e day of Aueryl {th}e {y}ere of oure lord a {th}owsand {th}re hondred foure score and seuene {th}e ten{th}e {y}ere of kyng Richard.

This might seem irrelevant if the Chaucer quotation comes from 1374. The reason it matters is that we need to know not when we think Geoffrey Chaucer might have finished writing Boece or when John Trevisa might have finished writing Polychronicon, but when the earliest surviving manuscript containing the word twitter dates from.

You see, books and manuscripts were copied and recopied - it was a huge industry in fourteenth and fifteenth-century England, before the invention of print (and after, for some time) - and the copyists would introduce changes. Sometimes this would be to replace words from one local dialect to make the work comprehensible to readers elsewhere in the country; sometimes it just seems to have been personal preference or error. It's quite rare to have an author's original copy of a work in his or her own hand. (These are known as holographs, which sounds quite exciting but isn't. There's a possible holograph of Chaucer's Equatorie of the Planets in the manuscript collection of Peterhouse, Cambridge.) All of this means that you can't guarantee that a word in a manuscript was put there by the author. (You can get more and more sure by comparing different manuscripts, but that's about all.)

Here comes the science. The manuscript of Chaucer's Boece in which the word twitter appears is called Cambridge, University Library MS Ii.1.38, and has been dated to the first quarter of the 15th century, somewhere around 1425. (There are lots of ways of dating manuscripts, which thankfully I'm not going into here.) The manuscript of Trevisa's Polychronicon is Cambridge, St John's College, MS H.1, and dates to the late 14th century, somewhere between 1387 and 1400. It is the earliest manuscript copy of the Polychronicon that survives.

That means that our earliest example of twitter appears in the Polychronicon, not in Boece. Until someone finds an earlier one, anyway.

So next time you use Twitter, spare a thought for John Trevisa. He may or may not have invented the word, but he was one of English's great organisers and sharers of content. Not content with a massive chronicle, he also translated the Middle Ages' most popular encyclopedia (Bartholomeus Anglicus's De Proprietatibus Rerum, 'On the Properties of Things'), and may also have been involved in the effort by John Wycliffe's friends and followers to translate the Bible into English.

Though not, admittedly, in 140 characters or fewer.