How we learned to read hieroglyphs

Walk into almost any large museum and you'll find a case of Egyptian things covered in small, beautifully cut birds, hands and coils of rope. For roughly fourteen hundred years, nobody alive could read a word of it.

That's the part the story usually skips. Egypt didn't mislay a few obscure texts. It lost the ability to read anything at all — its own tax records, its own love poetry, its own instructions for the dead. Temples stood in full sun, covered floor to ceiling in writing, saying nothing to anyone.

Getting it back took about a thousand years of confident wrong answers, followed by twenty-odd years of hard work by people who disliked each other.

How the script actually died

It wasn't a catastrophe. It was a slow professional extinction.

Hieroglyphs were always specialist work. Very few Egyptians could read at any point in the three thousand years the script was in use, and the people who could were mostly attached to temples. When Egypt became Christian and the old temples closed, the training pipeline closed with them. Nobody was teaching it because nobody needed it any more.

The last dated hieroglyphic inscription anyone has found was cut at Philae, near the southern border, in AD 394 — a graffito by a temple scribe called Esmet-Akhom. Demotic, the everyday cursive form, limps on at the same site until about AD 452. After that the sacred script is finished, and everything Egyptian is written in Coptic instead.

Which is the detail that matters later. Coptic is the Egyptian language, still spoken, still evolving — just written in Greek letters, plus half a dozen extra signs kept from demotic for sounds Greek didn't have. The language never died. It changed clothes, moved into the church, and stayed there.

The assumption that cost a thousand years

The obstacle wasn't a shortage of effort or intelligence. Serious people worked on this for centuries. They were all starting from the same false premise, and it seemed like common sense.

The signs are pictures. So each sign must stand for an idea. An owl means an owl, or wisdom, or night; a snake means danger or eternity or renewal. Under that reading, hieroglyphic writing isn't writing at all — it's a symbolic code that expresses concepts straight, without passing through anybody's mouth. Mystical, universal, independent of language.

A late antique text called the Hieroglyphica, attributed to one Horapollo, hardened this into doctrine. It gives solemn symbolic explanations for individual signs, and after a manuscript of it surfaced in Europe in 1419 it was treated as the key to the whole thing. Some of its entries aren't bad. Most are inventions. All of it pointed away from the answer.

Athanasius Kircher, working in seventeenth-century Rome, is the best example of how badly this went. He produced enormous, learned, gorgeously illustrated volumes of hieroglyphic translations that are, essentially, complete fiction — a cartouche spelling a pharaoh's name gets rendered as a paragraph of theology. And yet Kircher got one thing exactly right that nobody else did: he argued that Coptic descended from the language of the pharaohs. He had the key in his hand and used it to open the wrong door.

What was actually dug up at Rosetta

In July 1799, French soldiers strengthening a fort near the town of Rashid — Rosetta, to Europeans — pulled a broken slab of dark stone out of the works. An officer named Pierre-François Bouchard realised it was worth reporting.

It's granodiorite, not basalt, as it's often described. It's about 1.12 metres tall and weighs something like 760 kilograms. And it carries the same decree three times: fourteen surviving lines of hieroglyphs at the top, badly damaged; thirty-two lines of demotic in the middle; fifty-four lines of ancient Greek at the bottom.

The Greek was readable immediately, and it turned out to be extremely boring. A decree issued by a council of priests at Memphis in 196 BC, in the ninth year of Ptolemy V, confirming his honours, remitting some taxes and arranging for statues. Nobody would look at it twice for its contents.

The dullness didn't matter. Two unknown scripts were sitting next to a known language, and the Greek text ends by instructing that the decree be carved in sacred characters, native characters and Greek — which is to say the stone contains its own explanation of why it exists in three versions.

One warning about the popular version. The Rosetta Stone wasn't a key that unlocked hieroglyphs. It was one input. The hieroglyphic section is the most damaged of the three, it's short, and the decree is a translation of a translation from a period when Egyptian writing was at its most baroque. Plenty of the decisive work was done on other monuments entirely.

Who took the stone from whom

This part gets skated over, so here it is plainly.

The French took it from Egypt. Then the British took it from the French. When the French army in Egypt surrendered at Alexandria in 1801, the terms of the capitulation handed over the antiquities the expedition had collected, and the stone was shipped to England as war booty. It reached the British Museum in 1802, a gift from George III, and it's been on display there ever since — the most visited object in the building.

You can still read the arrangement on the object. Painted lettering on the sides records that it was captured in Egypt by the British Army in 1801 and presented by the king. At no stage was anybody in Egypt asked.

Egypt has requested its return, publicly and repeatedly, since at least 2003, and Egyptian archaeologists have petitioned for it again more recently. The British Museum has kept it. Whatever you think the right answer is, the tidy schoolroom story — French find, British scholars decipher, everyone benefits — leaves out the only party with an original claim on the thing.

Thomas Young got much further than he's given credit for

The usual telling has Champollion as the lone genius and Thomas Young as a footnote. That isn't fair, and it wasn't how it looked at the time.

Young was an English polymath — physician, physicist, the man behind the double-slit experiment — who took up the problem around 1814 and made several real advances. He established that demotic wasn't a separate symbolic system but a cursive relative of the hieroglyphic script. He worked on the cartouches, those oval loops drawn around certain sign groups, and reasoned correctly that they enclosed royal names. He matched the Ptolemy cartouche to the Greek and assigned phonetic values to its signs, several of which were right. He published a long account in a supplement to the Encyclopaedia Britannica in 1819, anonymously.

He had help from an unlikely direction. An obelisk brought back from Philae by William John Bankes carried a Greek inscription on its base and cartouches on the shaft, and the name Cleopatra could be picked out by comparing the two. More names, more sound values, no Rosetta Stone required.

So why doesn't Young get the credit? Because of what he concluded. He decided phonetic spelling was a special measure for foreign names — Greek names, Roman names, sounds the picture system had no way to handle — bolted onto a script that was otherwise symbolic. He'd found the mechanism and then walled it off as an exception. That single reservation is the whole distance between him and the answer.

Champollion had one advantage nobody else did

Jean-François Champollion had been obsessed with Egypt since childhood and had done something no rival codebreaker bothered with: he'd learned Coptic properly, well enough to write notes to himself in it.

That gave him a hypothesis about what the sounds ought to add up to. A codebreaker can extract phonetic values and end up with gibberish and no way to tell. Champollion could extract values and check whether the result was a word in a language he knew.

In 1822 the architect Jean-Nicolas Huyot sent him copies of reliefs from Abu Simbel — a site with no Greek or Roman rulers anywhere near it. In one cartouche Champollion recognised a sun disc, which he knew was ra in Coptic, followed by signs he read as mes and s. Ramesses. Another gave him the ibis of Thoth plus the same mes: Thutmose. Both are native Egyptian pharaohs from a thousand years before any Greek ruled in Egypt.

That's the moment. Phonetic writing wasn't a workaround for foreigners. It was how the script worked, and had always worked.

He set the argument out in a letter to the secretary of the French academy on 27 September 1822, expanded it into a full system in 1824, and finally got to Egypt himself in 1828, where he spent a year reading inscriptions off walls that nobody had understood for fourteen centuries. He died in 1832, aged forty-one, with his Egyptian grammar unpublished. His brother saw it into print.

The priority argument between Young's supporters and Champollion's started while both men were alive and has never entirely stopped.

What the writing system turned out to be

Both camps were partly right, which is exactly why the thing held out so long.

Hieroglyphic writing is mixed. Some signs are phonograms, standing for sounds — one consonant, two, or three at a time. Roughly two dozen of them carry a single consonant each, which is as close as Egyptian gets to an alphabet, and it's sitting there in plain view for anyone who's willing to believe pictures can spell. Some signs are logograms, standing for whole words. And some are determinatives, which aren't pronounced at all: silent category markers stuck on the end of a word to tell you whether you're reading about a man, a journey, an abstract idea or something unpleasant.

A single ordinary word can use all three kinds at once.

Then the layout. Vowels mostly aren't written. Text runs in rows or columns, left to right or right to left, and you work out which by looking at the animals and people — they face the start of the line. Signs get shuffled inside their little blocks for visual balance rather than strict order, and sometimes for respect, so a god's name jumps to the front of a phrase regardless of how it's spoken.

The classical script uses several hundred signs. In the temples of the Greek and Roman period the priests expanded that into the thousands, with elaborate new spellings and puns, at least partly because a script that outsiders couldn't read was the point. The Rosetta Stone comes from exactly that phase, which is another reason it was a harder starting text than it looks.

So anyone studying the signs could find plenty behaving like pictures, because plenty do. The error was assuming that was the whole system rather than a third of it.

What decipherment still can't give you

Egyptian can be read. It can't really be pronounced.

The script recorded consonants and left the vowels out, which was fine for people who already spoke the language and is a permanent problem for everybody since. Egyptologists get round it by dropping an e between consonants by convention, which is why "Nefertiti" and "Tutankhamun" sound the way they do in English. Those aren't reconstructions. They're a professional habit that hardened into names.

A lot of the familiar royal names reach us through Greek anyway. Khufu built the Great Pyramid; Cheops is what Herodotus called him. Coptic and Greek transcriptions let scholars work backwards towards the vowels in some cases, but large stretches of the sound of spoken Egyptian are simply gone.

And 1822 wasn't the finish line, whatever the anniversaries suggest. Champollion could read names and a good deal of text; a working grammar took decades more, largely through Adolf Erman and the Berlin school later in the century, and the great Berlin dictionary that grew out of it was begun in the 1890s and took most of the following century to complete.

The scripts nobody has read

Hieroglyphs are the famous win, and famous wins make the job look more tractable than it is.

Linear B is the other clean success — Michael Ventris cracked it in 1952 and, to general astonishment, the language underneath turned out to be an early form of Greek. Linear A, the older Cretan script the signs were borrowed from, is a different matter. Because many signs are shared, a lot of it can be pronounced. The sounds just don't make words in any language anyone recognises.

The Indus Valley script has the opposite trouble: thousands of inscriptions, nearly all of them a handful of signs long, no bilingual, no agreed language behind it, and a live argument about whether it encodes a language at all. Rongorongo from Easter Island and proto-Elamite from Iran are stuck for related reasons.

The pattern is unglamorous. To read an unknown script you generally need a bilingual text, a known relative of the language, or such a mountain of material that statistics can do the work. Egyptian had a bilingual and a surviving descendant. Most undeciphered scripts have none of the three, and no amount of cleverness substitutes.

What Egypt said once it could speak

Before 1822, ancient Egypt was known mostly through Greek and Roman visitors writing centuries after the fact, with all the confidence of tourists. Afterwards it could answer for itself, and what came out was far less mystical and much more recognisable than a thousand years of symbolic theory had promised.

Contracts. Court records. Medical papyri with prescriptions and prognoses. Schoolboys' exercises with the mistakes still in them. Letters complaining about relatives.

The village of Deir el-Medina, where the men who cut the royal tombs lived with their families, is the richest seam of it, because they wrote on limestone flakes and threw them away and the desert kept everything. There are attendance registers listing why each man missed work — ill, brewing beer, building his own tomb, a family member's funeral. There's a record of what looks like the earliest documented strike, in the twenty-ninth year of Ramesses III, when the rations didn't arrive and the workmen downed tools and sat outside the mortuary temple until somebody dealt with it.

None of that was retrievable while everyone believed the owl meant wisdom. The scholars who spent a millennium on symbolic readings weren't stupid or lazy. They'd just never seriously tested whether the thing in front of them was a language.