His first step was to count the total number of symbols in the Greek and hieroglyphic texts, a method which is now a commonplace of decipherment, but which Champollion seems to have been first to take in this science. The count revealed that there was something radically wrong with all previous efforts to solve hieroglyphic; for there were three times as many Egyptian as Greek letters. If the hieroglyphics were, then, either symbols for syllables or for ideas expressed as directly as the cave man’s deer, the Egyptian inscription must be more than three times as long as the Greek. But the very basis of any deduction must be that the inscriptions say the same thing; and the nature of the Greek text (a hymn of praise to Ptolemy V by a corporation of priests) made it seem unreasonable that there could be any great difference. If the inscriptions were identical, then the hieroglyphics must, after all, be letter-symbols. There were too many of them for any other theory.
On the other hand an alphabet of 160 letters remained inadmissible. But since other scientists had allowed themselves to be hung up on the horns of this dilemma, Champollion neglected it and plunged ahead on the alphabetic theory, attacking the names as Grotefend had in Persian. The name of Ptolemy was neatly enclosed in an outline, preceded by the symbol the English investigators had taken to represent the word for king. Now “Ptolemy” is a Greek word; Champollion made the reasonable deduction that in Egyptian it would have to be spelled phonetically. If the four symbols that stood for the name on the Rosetta Stone were letters, some letters in the name must have been omitted—which? The vowels, Champollion answered himself, remembering that Hebrew, which had a considerable Egyptian heritage, also omitted the vowels. The four symbols of the name were the letters pronounced P, T, L and M.
At this point the investigator turned to some older hieroglyphic inscriptions to check his conclusions. He had at hand a couple whose origin in the reigns of Kings Rameses and Thutmoes were proved by portraits and other evidence. The symbol he had adopted as M appeared in both names, and the T twice, in the proper places, in the second name. Thus it checked and, checking, gave him values for R and S; and with six letters to work on the scientist-cryptographer began to work through all the Egyptian inscriptions containing known names, obtaining new letter values at every step.
Very rapidly as scientific processes go—that is, in a matter of a few years—he accumulated enough data from names to provide the correct symbols for every possible consonant sound. There remained many letters of the impossibly extended alphabet for which he had no values; letters which never appeared as part of a name. Of these Champollion formed a separate list.
Returning to the Rosetta Stone inscriptions, he noted that one of these unidentified symbols appeared before every noun in the hieroglyphic text, and a few of them appeared before verbs. Now one such symbol was the picture of a tall man that had preceded King Ptolemy’s name. Later, where a temple was mentioned the word was preceded by a conventionalized picture of a building, and when the sun-god Ra’s name appeared there was a conventionalized solar disc. Champollion therefore reasoned that such characters were “determinatives”—special signs placed in the text by the Egyptian writers to indicate the character of the object they were talking about.
He died at the age of thirty-four without having worked out all the alphabet, and without having accounted for the remainder of the enormous surplus of letters, for even with the determinatives taken out, most of the words were far too long. It remained for later investigators to show that the Egyptians, in writing words, were never satisfied by expressing a sound in a single letter, but must repeat the same sound in three of four other ways to make certain the reader got the idea. It is as though one were to write the word “seen” as S-C-SC-EE-IE-EA-N. In a cryptological sense hieroglyphic was thus a substitution cipher with suppression of frequencies and the introduction of a prodigious number of nulls; and Champollion’s great merit as a decipherer was that he held to the main issue without allowing these things to throw him off the track.
CHAPTER II
THE ELEMENT OF DOUBT
I
THE cryptographer, however, must balance more delicately than an aerialist between not being distracted by collateral issues and failing to recognize root objections. He may hold so tightly to a main line of theory that he fails to see the facts have been fatally strained; and apparently this is what happened in the case of another famous decipherment.
In 1912 a New York bibliophile and dealer in old documents named Wilfrid Voynich bought in Italy a chest full of ancient manuscripts. Most of them were the material of his ordinary business and as such were catalogued for sale, but one had special interest for him. It was a volume of about eight by six inches, written on vellum in a fine clerkly hand, with its pages ornamented by some extremely curious drawings, the whole being bound in with a later but still very old dedication sheet which declared the volume to be the work of Roger Bacon, the famous medieval philosopher.
Dr. Voynich had handled Bacon manuscripts before. If the handwriting were any indication this was from the same pen as the others, and the materials used established it as being indubitably from the thirteenth century, when Bacon had lived. A long course of investigation which stretched across Europe from Italy to England, via Prague and Vienna, established with practical certainty that the manuscript was the work of the famous friar.
The particularly interesting thing about the manuscript was the language. It was not in clerk’s Latin, in which Bacon had written all his other known works, nor in any of the other six languages with which he was known to be familiar. It was not in any language whatever; it was in cipher. The characters in which it was written were not those of any alphabet ever seen, but a set of arbitrary signs, apparently belonging to the alphabet of a substitution system.
As soon as he had established the work as being genuinely from Roger Bacon’s pen, therefore, Dr. Voynich submitted it to several cryptographers. The discovery of a substitution cipher from the thirteenth century, employing an arbitrary alphabet, excited them greatly for, although ancient historians had mentioned the alphabet of Julius Caesar, no ciphered text of a date earlier than 1500 had ever before been found, and all the known early cipher texts were simple substitutions using either figures or the normal Latin alphabet. When the signs in the Bacon manuscript were copied off and classified, their appearance and number were found to be consistent throughout, both internally and with the idea that the manuscript had been composed in a simple single-alphabet substitution cipher.
But to the dismay of the cryptographers, their utmost resources proved unequal to the task of extracting a sensible message in any language from the text. This was astonishing in view of the fact that some of the same cryptographers had already performed such feats as reading messages which had first been translated into Chinese and then thrown into a complex cipher—and this without any previous knowledge of Chinese. Nevertheless there the fact was. The cipher experts reported they could go no farther without a great deal of time and expense and that, even then, they would not answer for the results.
Dr. Voynich therefore turned to scientists in other fields. The drawings that accompanied the text were partly of plants, roots, seeds and the process of germination; partly of astrological symbols; partly of stars, among which Aldebaran and the group Hyades were readily recognizable. It seemed logical to believe that the captions which appeared under these drawings were descriptions of the objects, and the manuscript was accordingly submitted to several botanists and astronomers, as well as to experts in ancient languages for attack along the lines used by Grotefend and Champollion. None of them was able to make the slightest impression on it; and that simple statement may be qualified by mentioning that they put several years of effort into the task.
At this point it occurred to Dr. Voynich that Roger Bacon had also been an expert in the interpretation of the mystic-symbolical Jewish Kabbala, and that the manuscript might require interpretation in the light of cabalistic lore. He therefore turned it over to Dr. William E. Newbold, of the University of Pennsylvania, one of the greatest students of medieval philosophy and science. Over the other investigators who had looked into the work Newbold possessed this advantage—that he was familiar with medieval methods