He now had to prove it correct. In cipher practice this step consists in substituting the letters obtained from guessing the probable word wherever else the same letters occur in the message being solved; the step Tychsen had taken with the name Arsaces. If the substituted letters make sense, when occurring in different combinations in other places in the message, then the probable word is right. But this is precisely the step Grotefend could not take; for ancient Persian was a lost language, there was no way of telling whether the letters made sense or not in other combinations.
Therefore, he was compelled to erect still taller towers of hypothesis in the air, hoping that he would someday be able to shove foundations underneath. In medieval Persian inscriptions the first word is always the name of a king, followed by the title “king of kings.” The “king of kings” phrase in two long inscriptions with which Grotefend was working occurred in the right places (as shown by the word-division marks) to justify the idea that the ancient Persians had begun their inscriptions in the same way as the medieval. Grotefend guessed, therefore, that the first words in these two long inscriptions were the names of kings. If he could discover what kings were represented, he would be a long way toward solving the whole business, for the names of the Persian kings were known.
The first word, which he had taken for the name of a king, in inscription A also occurred in inscription B (of the pair he was studying) not first this time, but lower down and with a small word dependent from it, the termination of which was the same as the termination of the long form of the word Grotefend had accepted as king. By his hypothesis this long form was the genitive of kings; therefore this new word must be a genitive also—of something. Most likely son of, thought the professor, given the connection with the names.
In short Grotefend had now arrived at the provisional identification of the beginning of the two inscriptions as:
A) X——, king of kings, son of Y——, king of kings —
B) Z——, king of kings, son of X——, king of kings —
This meant that the inscriptions had been set up by a father and son, both Persian emperors, and that the father’s father had also been a Persian emperor. The succession order of the emperors was well known from Greek histories. There was only one instance of father, son and grandson following one another to the throne, and the three in question were Hystaspes, Darius and Xerxes.
Assuming always that his identifications were accurate, Grotefend now had something that would give him a clue to alphabetical identifications and pronunciations of the letters in these names. “Hystaspes,” “Darius” and “Xerxes” were Graecized forms; he had to throw the three names back into old Persian, taking his cue as to what the old Persian form would be from medieval Persian and Hebrew texts. “Darius” thus appeared as “Darheush” and “Xerxes” as “Khshayarshtra.” When these forms were applied to the texts in question the number of letters came out exactly right. It was the first crumb of proof he had obtained after a long Barmecide banquet on pure theory; and with this crumb of proof Grotefend’s work came to an end for, like Niebuhr and Tychsen before him, he had now spent a lifetime on the Persian inscriptions.
He had, however, taken the essential step. The three names he had thus provisionally read gave to later workers ten or fifteen letters of the old Persian alphabet with their pronunciations. Those later workers applied the letters thus obtained to the remaining inscriptions of type I and discovered other recognizable names, which in turn yielded still further dividends of letters. Within twenty-five years of Grotefend’s passing the archaeologists were able to read old Persian phonetically; within another five years they were able to translate it with the aid of medieval Persian. Today anyone who wishes to take the trouble can decipher a Persian inscription as easily as a German and sometimes more so.
II
Scientific history is filled with the strangest repetitions, as though new ideas float into the world on some invisible medium and are caught through senses attuned by study in many places at once. The planet Uranus was discovered twice within a month; the periodic law which forms the basis of modern chemistry was propounded separately by two men who had never heard of each other and were working along different lines. Similarly, at about the time that Georg Friedrich Grotefend was painfully spelling out the names of forgotten kings of kings, another archaeological cryptographer was using the same methods to work out the other great puzzle of antiquity—the Egyptian hieroglyphics.
He was Jean François Champollion, an infant prodigy, whose father had been an archaeologist before him and had talked shop over the dinner table so entertainingly that at the age of fifteen the boy was already publishing a learned essay on “The Giants of the Bible” which won the applause of the bewigged professors at the French Institute.
Champollion’s problem in dealing with hieroglyphic was radically different from the one Grotefend of Göttingen had faced. The latter had before him various combinations of markings which were altogether meaningless except as the letters of an unknown language. Champollion was trying to read verbal sense into long strings of pictures which were considered by many very good scientists to have no more than a mystic religious sense, like the work of certain savage races which draw a picture of a deer when they feel hungry, expecting the gods to send them the real article in exchange for the pictured image.
Again, Niebuhr had identified forty-two different alphabetic signs, or letters, in ancient Persian; but the scientists who had already held hieroglyphic under investigation for centuries had discovered over a hundred and sixty signs—far too many to constitute any alphabet, beside which they were unmistakably conventionalized pictures. Moreover Grotefend had plunged into a new field, where all thought was independent thought; Champollion entered a domain already strewn with the wreckage of hypotheses, where it would be fatally easy to accept the errors along with the logic of some previous failure.
Particularly since the discovery of the famous Rosetta Stone. That celebrated chunk of crockery had been found by the scientists who accompanied Napoleon’s expedition to Egypt, and was surrendered to the English with the remains of that expedition. It bore an inscription in Greek, together with two other inscriptions, one in hieroglyphic and one in a third form known as Egyptian Demotic, then as unreadable as hieroglyphic. No great intelligence was required to make the supposition that all three inscriptions said essentially the same thing; but some of the best brains in Europe had spent years trying to resolve the hieroglyphic into an intelligible language, and even with the aid of the Greek texts it had proved impossible. The general conclusion was that the problem was insoluble.
For everything seemed to indicate that if the hieroglyphic were a language at all (and not a series of mystical pictures) it was that extremely rare thing, a purely syllabic tongue. For example, in the place where the word king appeared in the Greek text, the hieroglyphic had a picture of an extraordinarily tall man with a sword in his hand. This was a logical symbol for king; a whole word in one picture-letter. And if this were true, many of the other symbols stood for entire words or syllables; there would be no clue from the interrelation of letters as to how the language had been pronounced, and it would be forever unreadable.
There was also another difficulty. The British scientists who first handled the Rosetta Stone had taken the obvious step of making parallel lists of Greek words and the hieroglyphics that supposedly represented them. To their dismay they discovered that Greek words which appeared more than once in the inscription were represented on these different appearances by wholly unrelated sets of hieroglyphics, and that the same hieroglyphics were sometimes used to represent different words of the Greek text. Even the names, through which Grotefend was even then breaking ancient Persian, were of no help in this case. The only personal name in the Greek text was that of King Ptolemy V; in the hieroglyphic it was represented by four symbols—too few to spell it out with letters, too many to spell it in syllables. There seemed no conclusion but that the hieroglyphics were