In September I write of dipping into Ancient Records of Assyria and Babylonia, a fine old volume of translations from Meso-potamian cuneiform inscriptions, from which it is easy to see that the part of the world now known as Iraq was, even in antiquity, a bloody battleground ruled by ferocious tyrants. I quoted from the boastful inscriptions of such ancient Assyrian kings as Sennacherib and Assurbanipal, of which this is a typical sample: “For two days, from before sunrise, I thundered against them like Adad, the god of the storm, and I rained down flame upon them. . . . A pillar of living men and of heads I built in front of their city gate, seven hundred men I impaled on stakes in front of their city gate. The city I destroyed, I devastated, I turned it into mounds and ruins; their young men I burned in the flames.” And I noted that what I found most interesting about these horrifying testaments of atrocity wasn’t their ghastliness but the mere fact that we are capable of reading them at all, written as they were on tablets of clay in what is now a lost language and a strange wedge-shaped script. So let’s look now at how we came to understand the Assyrian and Babylonian inscriptions in the first place.
European scholars had been puzzling over this ancient writing—”cuneiform,” it was called, from the Latin word meaning “wedge”—since the seventeenth century. Three different kinds of cuneiform inscriptions had been discovered at the ancient Persian capital of Persepolis. By 1778 one of them had been shown to be a forty-two-character alphabetic script. The other two were vastly more complex. And no one knew which language these characters represented, although it was reasonable to think that one or perhaps all three scripts were in the ancient Persian tongue.
To decipher an unknown script, there has to be some point of contact with the known. If both the characters and the language they represent are enigmas, it becomes impossible to get very far with a decipherment. François Champollion was able to decipher hieroglyphics because he had the use of the Rosetta Stone, which provided a long Egyptian text in two kinds of Egyptian script, plus a translation into Greek. But no Rosetta Stone for the cuneiform script was available.
One big break came toward the end of the eighteenth century as philologists began to study existing texts of an archaic form of the Persian language that had been written using a decipherable alphabet called Pehlevi. One of them guessed that the simplest of the Persepolis inscriptions, the so-called Class I ones, were Old Persian texts written in the wedge-shaped cuneiform letters. A German high-school teacher named Georg Friedrich Grotefend, whose hobby was solving puzzles, went looking for some repetitive phrase in the Persepolis inscriptions that might give him a clue to the meaning of a few of the wedge-shaped symbols.
It was already known that Persian official inscriptions almost always began with, and constantly reiterated, a formula that went, “So-and-So, Great King, King of Kings, King of This and That, Son of So-and-So, Great King, King of Kings.” Grotefend first identified the eight most frequent of the forty-two Class I characters and decided that these probably stood for vowels. Then he went looking for the words of his royal formula, and the names of the kings.
Quickly he found clusters of repeated words, the most frequent of which was a seven-letter group that he suspected meant “king.” In Old Persian that word was “khsheihioh,” and by lining up the characters he arrived at guesses for seven letters. Then, finding what seemed to be a royal name at the proper place in the formula, he matched the letters he had already identified against the name of a known king—Xerxes, “Khshershe” in Persian—and then tested his growing list of letters against the name of Xerxes’ father, Darius—“Darheush.” Bit by bit, by trial and error, he was able to claim identification of twenty-nine of the forty-two Class I characters by 1803. Although it turned out that he was wrong about some of these, he had provided entry into the mysteries of Class I cuneiform.
But Class I was still a long way from a complete decipherment and the Class II and III inscriptions were still total mysteries when, in 1835, a swashbuckling English scholar-adventurer named Henry Creswicke Rawlinson entered the picture. Rawlinson, a lieutenant in the service of the British East India Company, had been stationed in various Asian posts—first in India, then in Persia—since the age of seventeen. He had a natural knack for languages and quickly mastered several Indian tongues, Arabic, and Persian. And, like Grotefend, he was inclined toward puzzle-solving as an amusement. Knowing little of Grotefend’s work, he attempted a decipherment of Class I using the same method, and worked out thirteen letters on his own. When the East India Company transferred him to its Persian base at Kermanshah, he swiftly learned that a lengthy cuneiform inscription in all three scripts was to be found carved on the Behistun Rock, a seventeen-hundred-foot-high cliff twenty miles from town, and rode out to take a look.
But the inscription was all but inaccessible. The ancient Persian rock-carvers had removed the steps leading up to it after they were finished, so later vandals could not deface the words. Rawlinson, a considerable athlete, scrambled up the bare, slippery face of the rock without the aid of ropes or ladders until he reached a ledge, two feet in breadth at its widest point, where he could stand and copy part of the inscription.
The Behistun text was studded with names out of Persian history—King Darius and his whole ancestral line—and Rawlinson, using inspired guesswork, his linguistic skills, and his knowledge of history, was able to match names to cunei-form characters and work out a nearly complete translation by January 1838, correcting many of Grotefend’s errors and demonstrating certain knowledge of eighteen of the forty-two letters. In the months that followed he was able to decipher and translate some two hundred lines of the Class I Behistun inscription.
Class II and Class III remained unknown, though, and Rawlinson had only fragmentary copies of those texts, which he believed were the Class I inscription written in the scripts of two other languages, one of them very likely Babylonian or Assyrian. In 1844 he returned to Behistun, erected a wooden folding ladder on a narrow ledge three hundred feet above the ground, stood on its topmost rung, and, bracing himself against the rock with his left arm and holding his notebook in his left hand, copied the inscription with his right hand. “The interest of the occupation,” he wrote, “entirely did away with any sense of danger.”
Now he had all of Class I and most of the Class II inscription. In 1847 he returned equipped with ladders, planks, ropes, nails, hammers, and pegs, and hired a Kurdish boy to scramble out over the abyss on a scaffold to make paper casts, “squeezes,” of the almost inaccessible Class III. Equipped with the full texts, Rawlinson set about to match his Persian Class I text against the far more intricate Class III, which had hundreds of characters instead of only forty-two, in the hope of solving the riddle of Babylonian cuneiform. In England, meanwhile, a clergyman named Edward Hincks started work on the same difficult task.
Hincks showed that it was wrong to talk of a Babylonian “alphabet.” More than five hundred different Class III characters were known, and no language could have that many basic sounds. Hincks guessed that some of the symbols stood for individual syllables, and others for entire words. Comparing Class I’s royal formulas with their likely Class III counterparts, he showed that the seven signs of Xerxes’ name—KH-SH-Y-A-R-SH-A in Hincks’ reading—lined up with Babylonian signs that could have been pronounced KHI-SHI-I-AR-SHI-I. By 1847 he had deciphered twenty-one syllables and had identified the ideographic symbols that stood for such words as “and,” “son,” “great,” “house,” and “god.”
Rawlinson, at the same time, was finding syllables—ka, ki, ku, ak, ik, uk, etc.—and deciphering certain words, helped by his familiarity with Hebrew and Arabic, languages closely related to the one that the Babylonians and Assyrians had spoken. The word for “dog,” keleb in Hebrew and kalbu in Arabic, turned out to be kalbu in Babylonian-Assyrian. “To burn” was saraf in Hebrew and sarapu in Babylonian-Assyrian. So it went, until by 1850 he could claim to know the meaning of 150 Class III characters and two hundred Babylonian-Assyrian words. Each solution, though, brought with it a host of new complications. The Meso-potamian civilizations had lasted thousands of years, and over that time Mesopotamian scribes had invented all manner of new ways of transforming words into wedge-shaped symbols, so that by the time the Behistun inscription was carved the system in use was full of bewildering overlaps of meaning and linguistic shortcuts. Rawlinson devoted years to untangling these puzzles, while Hincks carried on parallel research of equal value. In 1851 Rawlinson made public a translation of a text carved on a clay cylinder that confirmed an event described in the Bible, the defeat of King Hezekiah of Judah by the armies of the Assyrian king Sennacherib, which I quoted in last month’s column. (“As for Hezekiah, the Jew, who had not submitted to my yoke, forty-six of his strong, walled cities and the cities of their environs, which were numberless, I besieged, I captured, as booty I counted them.”) Its announcement caused a mighty stir in England.
Excavation of the ruined Assyrian palaces at Nineveh in what is now northern Iraq produced whole libraries of clay tablets bearing cuneiform inscriptions, among them, of all useful things, what Rawlinson called “a perfect encyclopedia of Assyrian science,” a cuneiform “book” describing “the system of Assyrian writing, the distinction between phonetic and ideographic signs, the grammar of the language, explanation of technical terms. . . .” These royal libraries also contained dictionaries of the Assyrian language. What he already knew of the script allowed him to penetrate these tablets and greatly extend his knowledge. His rival Hincks kept at it also, and a Franco-German scholar named Jules Oppert entered the field as well. These three, working independently, steadily refined the understanding of the cuneiform inscriptions by mastering, step by step, the bizarre tangle of complexities that governed the Mesopotamian system of writing. The climax came in 1857 when the Royal Asiatic Society of England sponsored an event in which Rawlinson, Hincks, Oppert, and an English scholar named William Fox Talbot were each given copies of an untranslated Class III text and told to translate it independently. When the results came in, Rawlinson’s and Hincks’ versions were virtually identical, Oppert’s quite similar, and Fox Talbot’s, though sometimes vague and incorrect, fairly close in a general way. And that is how we came to unlock the secrets of Mesopotamian cuneiform. (The Class II, which turned out to be in a language called Elamitic, was not deciphered until 1879. Sumerian, yet another cuneiform script and the ancestor of all the others, was decoded around the same time through the use of Assyrian-Sumerian dictionaries.)
It will not, of course, be so easy to decipher the writings we find on alien worlds, when and if we discover such things. We won’t be provided with convenient multilingual inscriptions that can partly be understood, as were Champollion in Egypt and Rawlinson in Persia. On the other hand, perhaps future advances in computer science will sweep away such obstacles. Until then, though, the nineteenth-century decipherment of Mesopotamian cuneiform must stand as one of the most miraculous of linguistic achievements, opening, as it does, a doorway into civilizations thought forever lost.
