THE CODEBREAKERS

by DAVID KAHN

  Cryptography seems to have been not at all uncommon in the Roman state. Suetonius’s phraseology implies that the two Caesars employed it habitually and not on a single isolated occasion. Cicero used SAMPSICERAMUS and ARABARCHES and HIEROSOLYMARIUS as mocking codenames for Pompey; they all allude to persons and places of importance in Pompey’s career. Many Latin writers mention rudimentary forms of secret communication. A grammarian named Probus, probably Valerius Probus, even wrote a treatise on the ciphers of Julius Caesar; this has not survived, and is the first of several Lost Books of Cryptology.

  It must be that as soon as a culture has reached a certain level, probably measured largely by its literacy, cryptography appears spontaneously—as its parents, language and writing, probably also did. The multiple human needs and desires that demand privacy among two or more people in the midst of social life must inevitably lead to cryptology wherever men thrive and wherever they write. Cultural diffusion seems a less likely explanation for its occurrence in so many areas, many of them distant and isolated.

  The Yezidis, an obscure sect of about 25,000 people in northern Iraq, use a cryptic script in their holy books because they fear persecution by their Moslem neighbors. Tibetans use a kind of cipher called “rin-spuns” for official correspondence; it is named for its inventor Rin-c’(hhen-)spuns(-pa), who lived in the 1300s. The Nsibidi secret society of Nigeria keeps its pictographic script from Europeans as much as possible because it is used chiefly to express love in rather direct imagery, and samples appear to be at least as pornographic as they are cryptographic. The cryptography of Thailand developed under Indian influence. An embryonic study of the subject even appears in a grammatical work entitled Poranavakya by Hluang Prasot Aksaraniti (Phe). One system, called “the erring Siamese,” substitutes one delicate Siamese letter for another. In another system, consonants are divided into seven groups of five letters; a letter is indicated by writing the Siamese number of its group and placing vertical dots under it equal in number to the letter’s place in its group. A system called “the hermit metamorphosing letters” writes the text backwards.

  “The erring Siamese”—a form of Thai cryptography with plaintext in upper lines, cipher in lower

  As isolated an area as the Maldive Islands in the Indian Ocean uses two kinds of secret writing. “Harha tana” involves reciprocal substitution between consecutive letters of their alphabet, the “gabuli tana,” so that h =RH and rh = H, and so forth, the first equivalent perhaps giving rise to the name of the system. “De-fa tana” effects substitutions between the halves of the gabuli tana. In Malaya, natives call their cryptographic alphabet the “gangga malayu”; it consists of the slightly altered or inverted characters of the Malayan Arabic alphabet, with some Javanese marks. In Armenia in the 16th century, two scribes employed a Polybius-like checkerboard to inject an air of special hidden knowledge into religious texts; a third composed his ciphertext by writing two letters whose numerical value equaled that of the plaintext letter—z, with value 6, became GG, each G having a value 3.

  Persia, in the first half-millennium after Christ, apparently made use of cryptography for political purposes. A chronicler mentioned a “script called ‘shāh-dabīrīya,’ and the kings of the Persians used to speak it among themselves to the exclusion of commoners and prevent the rest of the people of the kingdom from [learning] it for fear that one who was not a king should discover the secrets of the kings.” He also referred to “another script called ‘rāzsahrīya,’ in which the kings used to write secrets [in correspondence] with those of other nations that they wished, and the number of its consonants and vowels is forty, and each of the consonants and vowels has a known form, and there is no trace in it of the Nabataean language.” Though the historian gave no examples, a 10th-century compiler of a handbook for secretaries, in setting down two monalphabetic substitutions, said that they were of Persian origin. One substituted the names of birds for the letters of the alphabet. The other equated the letters of the alphabet with the names of the 28 astronomical lunar mansions: the two horns of the ram, the ram’s belly, the Pleiades, and so on.

  At the Coptic monastery of St. Jeremias in Saqqara, Egypt, perhaps just before it was abandoned late in the sixth century A.D., a man enciphered a message in monalphabetic substitution and scratched it on the wall inside the door to a courtyard in a curious bid for immortality. “In the name of God before all things,” the inscription calls out beseechingly across the centuries, “I, Victor, the humble poor man—remember me.” Victor’s encipherment of his plea gave him his wish. At the site of another Coptic monastery, the seventh-century one of Epiphanius at Sheikh-abd-el-Gourna in southern Egypt, there was found an unusual object in the cell of a priest named Elias. It was a dried-out piece of wood about a foot long and four inches high, bearing two lines of writing in black ink. The top line is a slightly garbled verse in Greek, notable not for its beauty but because it includes all the letters of the Coptic Greek alphabet. It spills over for five letters into the bottom line, which contains 21 letters of that alphabet, divided into four unequal sections that are reversed and shuffled. How Priest Elias used it is not known, but it does seem fairly certain that this wooden tablet, now in the Metropolitan Museum of Art in New York, is the oldest surviving cipher key (as distinguished from the codelike cuneiform tablets) in the world.

  The hardy plant of cryptography sprouted not only in these sunblasted climes but also in the damp, chill lands to the north. Two non-Latin scripts of Europe, Teutonic runes and Celtic oghams, were occasionally enciphered.

  Runes flourished in Scandinavia and in Anglo-Saxon Britain during the seventh, eighth, and ninth centuries. They were nearly always used for religious purposes. A stark, angular script, its alphabet was divided into three groups of eight runic letters each. The letter thorn, for example, which looked somewhat like a modern p and represented the initial sounds of “thin” and “then,” was the third letter of the first group. All systems of runic cryptography replaced runic letters by groups of marks indicating the number of a letter’s group and the number of its place in that group. Isruna used the short i rune, a short vertical stroke named “is,” to give the number of the group, and the long i rune to give the place number. Thus thorn—group 1, letter 3—would be replaced by a single short vertical mark and three longer vertical marks. Another system of runic cryptography, hahalruna, attached diagonal strokes representing these numbers to a vertical shaft, putting the group marks on the left, the place marks on the right. Sometimes shafts were crossed. Other variations on this theme were lagoruna, stopfruna, and clopfruna. Cryptographic runes occur in many places, most profusely on the Rök stone, a 13-foot-high slab of granite standing at the western end of the Rök churchyard in Sweden. It includes among its more than 770 runic letters a veritable catalog of runic cryptography.

  The 13-foot-high Rök stone of Sweden, covered with enciphered runes

  Three forms of enciphered ogham: head of quarreling, interwoven, and well-footed ogham as shown in the “Book of Ballymote”

  Ogham survives chiefly in inscriptions on tombstones. Its alphabet consists of five groups of five letters, represented by one to five lines extending away from a horizontal line. In the first group, the lines extend above the horizontal line; in the second group, below it; in the third, perpendicularly above and below; in the fourth, diagonally above and below; the fifth group is heterogeneous. Methods for enciphering them are catalogued in the “Book of Ballymote,” a 15th-century compilation of historical, genealogical, and other facts of importance.

  The most delightful thing about these systems is their names, the most charming in cryptology, which have been bestowed with all the Irish flair for poetry, blarney, and wit. There is, for an example, a system called “the ogham that bewildered Bres,” in which the name of the letter stands for the letter, as if one were to encipher who as DOUBLE-YOU AITCH OH. The name comes from a story that a message thus concealed was given to the ancient hero Bres as he was going into battle
, and so confused him by its complications that he lost the battle while trying to figure it out. “Sanctuary ogham” puts a stroke between every pair of letters. “Serpent through the heather” runs a wavy line above and below the successive letters. “Great speckle” has a single mark of appropriate slant and length for the letter, followed by as many dots, less one, as there are strokes in the letter. In “twinned ogham” each letter is doubled; in “host ogham,” tripled. “Vexation of a poet’s heart” reduces the lines to short marks extending beyond an empty rectangle. In “point against eye,” the alphabet is reversed. In “fraudulent ogham” the letters are replaced by symbols one step further on. And a system in which the chaotic order of the substitutes seems to have resulted from an infuriated Irishman’s knocking them about with a shillelagh is called “outburst of rage ogham.” Probably none of these ever actually enciphered ogham. They seem to have been just dreamed up for fun. But the bottom of one of the pages of the “Book of Ballymote” is written in another system called “Bricriu’s ogham.” With some emendation, it can be interpreted as a fragment of an ancient Druidic liturgy —probably the only one known to the modern world, and, fittingly, the only place in which enciphered oghams were ever used.

  In the Europe of the Latin alphabet—from which modern cryptology would spring—cryptography flickered weakly. With the collapse of the Roman empire, Europe had plunged into the obscurity of the Dark Ages. Literacy had all but disappeared. Arts and sciences were forgotten, and cryptography was not excepted. Only during the Middle Ages occasional manuscripts, with an infrequent signature or gloss or “deo gratias” that a bored monk put into cipher to amuse himself, fitfully illuminate the cryptologic darkness, and, like a single candle guttering in a great medieval hall, their feeble flarings only emphasize the gloom.

  The systems used were simple in the extreme. Phrases were written vertically or backwards; dots were substituted for vowels; foreign alphabets, as Greek, Hebrew, and Armenian, were used; each letter of the plaintext was replaced by the one that follows it; in the most advanced system, special signs substituted for letters. For almost a thousand years, from before 500 to 1400, the cryptology of Western civilization stagnated. An “advanced” system is as likely to appear in the 600s as in the 1400s—though the really simple systems do fade away by the end of the period.

  A few names glimmer through the mists. Tradition attributes to St. Boniface, the Anglo-Saxon missionary who founded monasteries in Germany in the eighth century, the importation to the continent of cryptographic puzzles based on a dots-for-vowels system. The brilliant monk Gerbert, who reigned as Pope Sylvester II from 999 to 1003 and whose learning became legendary, kept notes in a syllabic system called “tyronian notes,” a shorthand reputedly developed by Tullius Tyro, a freed slave of Cicero’s. He even wrote his name in it on two of his bulls. Hildegard von Bingen, an 11th-century nun who saw apocalyptic visions and was later canonized, had a cipher alphabet which she claimed came to her in a flash of inspiration. In the early 800s, an Irishman named Dubthach concocted a cryptogram while at the castle of the king of Wales as a kind of malicious IQ test for visiting compatriots. He apparently wanted to embarrass them in revenge for some humiliation he had suffered at home, and was confident that “no Irish scholar, much less British,” would be able to read it. But four clever sons of Eire—Cauncho-brach, Fergus, Domminnach, and Suadbar—turned the tables on him by solving the cryptogram, which consisted of a short Latin plaintext written in Greek letters. Then they prudently sent the answer back to their teacher, urging him to “give this information to such of our simple and unsophisticated Irish brethren as may think of sailing across the British sea, lest perchance otherwise they might be made to blush in the presence of Mermin, the glorious king of the Britons, not being able to understand that inscription.”

  The only writer of the Middle Ages to describe cryptography instead of just using it was Roger Bacon, the English monk of startlingly modern speculations. In his Epistle on the Secret Works of Art and the Nullity of Magic, written about the middle of the 1200s, Bacon stated: “A man is crazy who writes a secret in any other way than one which will conceal it from the vulgar,” and then listed seven deliberately vague methods of doing so. Among them are the use of consonants only, figurate expressions, letters from exotic alphabets, invented characters, shorthand, and “magic figures and spells.”

  A cryptogram composed and written by Geoffrey Chaucer

  Far and away the most famous of all those who had an acquaintance with cryptology in the Middle Ages was an English customs official, amateur astronomer, and literary genius named Geoffrey Chaucer. In a work called The Equatorie of the Planetis, which describes the workings of an astronomical instrument and which appears to be a companion piece to his Treatise on the Astrolabe, Chaucer included six short passages in cipher. He enciphered them with a symbol alphabet in which, for example, a is represented by a sign resembling a capital V and b by one looking like a script alpha. One passage reads: “This table servith for to entre in to the table of equacion of the mone on either side.” The encipherments give simplified directions for using the equatorie—never mind about the complicated technical explanation, just do this and that and the answer comes out right. The cryptograms are in Chaucer’s own hand, making them some of the most illustrious encipherments in history.

  During all these years, cryptology was acquiring a taint that lingers even today—the conviction in the minds of many people that cryptology is a black art, a form of occultism whose practitioner must, in William F. Friedman’s apt phrase, “perforce commune daily with dark spirits to accomplish his feats of mental jiu-jitsu.”

  In part it is a kind of guilt by association. From the early days of its existence, cryptology had served to obscure critical portions of writings dealing with the potent subject of magic—divinations, spells, curses, whatever conferred supernatural powers on its sorcerers. The first faint traces of this appeared in Egyptian cryptography. Plutarch reported that “sundry very ancient oracles were kept in secret writings by the priests” at Delphi. And before the fall of the Roman empire, secret writing was serving as a powerful ally of the necromancers in guarding their art from the profane.

  One of the most famous magic manuscripts, the so-called Leiden papyrus, discovered at Thebes and written in the third century A.D. in both Greek and a very late form of demotic, a highly simplified version of hieroglyphics, employs cipher to conceal the crucial portions of important recipes. For example, in a section telling how to give a man an incurable skin disease, the papyrus uses secret signs to encipher the words for “skin disease” and the names of the lizards: “You wish to produce a skin-disease on a man and that it shall not be healed, a hantous-lizard and a hafleele-lizard you cook them with oil, you wash the man with them.” The plaintext in most of the cipher sections (including one telling how to make a woman desire a man, which doesn’t work) is in Greek, and the cipher alphabet consists basically of Greek letter signs. Cryptology served magical purposes frequently throughout the Middle Ages, and even in the Renaissance was still disguising important parts of alchemical formulas. A manuscript compiled at Naples between 1473 and 1490 by Arnaldus de Bruxella uses five lines of cipher to conceal the crucial part of the operation of making a philosopher’s stone.

  The association of magic and cryptology was reinforced by other factors. Mysterious symbols were used in such esoteric fields as astrology and alchemy—where each planet and chemical had a special sign, like the circle and arrow for Mars—just as they were in cryptology. Like words in cipher, spells and incantations, such as “abracadabra,” looked like nonsense but in reality were potent with hidden meanings.

  A very important factor was the confusion of cryptology with the Jewish kabbalah, a mystical philosophy that also interested many Christians of an occultistic turn of mind. One of its basic tenets was that language, which comes from God, reflects the fundamental spiritual nature of the world, and so expresses creation itself. Kabbalists thus produc
ed new revelations about existence by wringing hidden meanings from every word, every letter, even every vowel point and accent mark in the Torah. “Truth,” they would say, “stands more firmly than falsehood”—an assertion based on the fact that the letters of the Hebrew word for “falsehood” all balance precariously on one leg, somewhat like an English r, whereas those of the word for “truth” all rest solidly on two feet, like h. Among their devices was gematria, which gave the letters of Hebrew words their numerical values, added them up, and then interpreted the result, often by comparing it with other words having the same total. For example, Genesis 14:14 says that Abraham came to the aid of his nephew Lot with 318 servants. But 318 is the numerical value of the name of Abraham’s servant, Eliezer. Hence the 318 were really only one—Eliezer. Less important than gematria were notarikon, which regarded the letters of words as abbreviations for whole sentences, and temurah, an interchange of letters according to various rules, including atbash. These practices work upon the same raw material as cryptology, but unlike cryptology they are flexible and speculative. Some of their laxness seemed to infect cryptology, while their mystical pronouncements seemed to add further magical elements.

  Later writers boasted of their ability to solve ciphers in the same breath that they bragged of their prowess in recording human voices, in telepathy, and in communicating with people far underground or miles away. One influential writer, an abbot who believed in magic, then under condemnation by the church, wrote about it under the guise of the more innocuous cryptology—and thus intensified the association between them. Later writers discussed the two together either because they believed they went together or to impress their readers with their own dread powers. Much of this supernatural claptrap besmirched cryptology.