In 1799, French soldiers digging near the Nile town of Rashid (Rosetta) pulled out a dark stone slab covered in writing. The same decree was carved on it three times: once in Egyptian hieroglyphs, once in a cursive Egyptian script called Demotic, and once in ordinary Greek. Anyone could read the Greek. The other two had been unreadable for well over a thousand years. The last dated hieroglyphic inscription, at Philae, is from AD 394.
You might expect that a translation key would solve it in a week. It took 23 years. This page lets you repeat the crucial step yourself.
Among the hieroglyphs, a few groups of signs sat inside an oval ring, called a cartouche. Scholars guessed these were royal names. The Greek text on the stone mentions King Ptolemy again and again, so the ringed group that kept recurring was probably Ptolemy. The English scholar Thomas Young made that match around 1814 and got several sign values right. But he believed sound-signs were used only to spell foreign names such as Ptolemy, and that ordinary Egyptian words worked some other way.
A second cartouche came from an obelisk from Philae, copied by the collector William Bankes. Its Greek base mentioned Ptolemy and Cleopatra, so one ring should say Ptolemy and another Cleopatra. Two names that share letters are a gift, because a sign that appears in both must stand for the same sound.
Once you have the Ptolemy signs, Cleopatra's ring is half solved. The sign for P (a block) appears in both names, and so do the signs for L (a lion) and O (a loop of rope). They sit in different places in each name, which is exactly what you expect if they are sounds and not whole-word pictures.
The leftovers were then a puzzle with a built-in check. Cleopatra has two A's, and the ring contains the same bird sign twice, in sixth place and again at the very end. Fill in one and you have filled in both. That is the kind of consistency that separates a real reading from a lucky guess.
Jean-François Champollion was in his early thirties and had spent his life on languages, especially Coptic, the last stage of the Egyptian language written in Greek letters. He used it as a bridge. In a ring from the temple at Abu Simbel he saw a sun disc followed by two signs he could read as M and S. Coptic gave him the word for sun, something like re. Together: Ra-m-s-s, Ramesses, a native Egyptian king with no Greek in sight.
That was the breakthrough Young missed. The sound-signs were not a special trick for foreign names. They were part of how Egyptians wrote their own language, mixed with signs that stand for whole words and signs that hint at meaning. Champollion announced his system in September 1822, in a letter to the scholar Joseph Dacier.
Notice what carried the day: not a dictionary, but overlap. Two different names that share some sounds create constraints. Fix a few values and the rest of the puzzle starts to fall into line. Codebreakers use the same idea whenever they can find a known word, a "crib", inside a message.
Hieroglyphic writing was not finished off by Champollion. Scholars are still refining how signs combine and how the language changed over three thousand years. But before 1822 nobody could read the temples. After it, they could.
Sketches above are simplified for the activity. The real shapes and some sound values (such as the two kinds of T) are more varied than shown, and real royal names carry extra signs for titles and endings that we left out.