Ghosts in the Parchment · Chapter 2

The Cliff of Three Tongues

Henry Rawlinson and the Cuneiform Trilingual of Behistun

Three hundred feet up a limestone precipice at Mount Behistun in western Persia, British cavalry officer Henry Creswicke Rawlinson dangled from ropes above a vertical drop in the summer of 1835. Before him, chiseled into a smoothed rock face twenty-five feet high and fifty feet wide, stood a massive bas-relief carved in 520 BCE by order of the Achaemenid emperor Darius the Great. The sculpture showed Darius placing his foot on the chest of the defeated usurper Gaumata while nine bound rebel kings stood roped together. Surrounding the figures were over a thousand lines of wedge-shaped cuneiform script arranged in three parallel columns. A coat of siliceous varnish had preserved every chisel stroke for twenty-three centuries.

Unlike Egyptian hieroglyphs—which Jean-François Champollion decoded in 1822 using the Greek text on the Rosetta Stone—Mesopotamian cuneiform had died out by the first century CE without leaving a single bilingual Greek or Latin inscription. For two millennia, travelers walking the mounds of Nineveh, Babylon, and Persepolis stared at clay tablets covered in bird-track wedges. Nobody knew whether the marks were letters, ideograms, or floral ornaments. In the seventeenth and eighteenth centuries, European visitors Pietro della Valle and Carsten Niebuhr copied short inscriptions from Persepolis. Their transcripts showed that the carvings always grouped three wedge systems together. The first was a simpler script of forty-two signs (Old Persian). The second was a middle script of over a hundred signs (Elamite). The third was a script of over five hundred characters (Babylonian Akkadian).

In 1802, German teacher Georg Friedrich Grotefend in Göttingen made the first crack in the simplest script. Inspecting Niebuhr's copies of two Persepolis doorway inscriptions where diagonal wedges divided words, Grotefend noticed a recurring royal formula: [Name A], [Title], [Title] of [Title]s, son of [Name B]. Comparing the Achaemenid kings recorded by Herodotus, Grotefend looked for father-and-son rulers whose names were roughly the same length. Moreover, the father of the first king did not carry the royal title himself. Cyrus and Cambyses failed because both names began with the same letter, whereas the cuneiform names began with different signs. That left Hystaspes (a non-king), his son Darius, and his grandson Xerxes. Grotefend slotted in a dozen phonetic values, but he lacked long texts and Iranian philology, so his work stalled.

Stationed at Kermanshah in 1835, Rawlinson realized that the Behistun cliff was the Rosetta Stone of Mesopotamia. Returning between 1835 and 1837 with ladders bridged across eighteen-inch ledges, Rawlinson copied the Old Persian columns into his notebooks. Because Darius's proclamation listed dozens of rebellious provinces and generals across the empire—Babylon, Assyria, Media, Armenia, and Bactria—Rawlinson possessed nearly a hundred proper names. He cross-checked their phonetic spellings against Greek historians, the Zoroastrian Avesta studied by Eugène Burnouf, and classical Sanskrit:

In May 1857, the Royal Asiatic Society in London sealed Rawlinson, Hincks, Julius Oppert, and William Henry Fox Talbot in separate rooms with an unpublished cuneiform prism of the Assyrian king Tiglath-Pileser I. When their four independent translations matched line for line, Mesopotamian philology passed its decisive blind test.