THE CODEBREAKERS

by DAVID KAHN

  The earliest New World cryptogram extant: a cipher message of Hernán Cortés, June 25, 1532

  Spain administered its cryptography in the Despacho Universal, the nerve center of the government, from which couriers departed at all hours of the day and night for all parts of the world. When the capital was moved to Madrid in 1561, the Despacho was installed in the Alcazar with Foreign Secretary Gonzalo Perez in charge.

  Decipherment in the intrigue-filled court of Spain did not rest solely on a mere mechanical operation of cryptographic rules. If an ambassador requested the payment of his salary or solicited a bishopric, and the deciphering secretary was not his friend, the passage might remain undeciphered. Philip himself ordered the deciphering secretaries to suppress passages that he did not want his council to know about. On top of these sins of omission were piled those of commission, sometimes so serious that in at least one instance the codeword for king of England was confounded with that for king of France !

  The sweet smell of success in cryptanalysis never wafted through the Moorish chambers of the Alcázar. But Philip’s archfoes—Protestant England, France with its Huguenot king, and the rebellious Spanish provinces of the Netherlands—did not blind themselves to this providential source of information. Their cryptanalytic abilities had a pope and most of Europe snickering at Philip, played no small role in foiling his grandiose plans for the conquest and conversion of England, and helped ultimately to execute a sentence of death on that most romantic and captivating of royal ladies, Philip’s intended sister-in-law, Mary, Queen of Scots.

  In 1589, Henry of Navarre, who was destined to become the most popular king in the history of France (he coined the slogan “A chicken in every peasant’s pot every Sunday”), ascended to the throne as Henry IV and found himself embroiled still more fiercely in his bitter contest with the Holy League, a Catholic faction that refused to concede that a Protestant could wear the crown. The League, headed by the Duke of Mayenne, held Paris and all the other large cities of France, and was receiving large transfusions of men and money from Philip of Spain. Henry was tightly hemmed in, and it was at this juncture that some correspondence between Philip and two of his liaison officers, Commander Juan de Moreo and Ambassador Manosse, fell into Henry’s hands.

  It was in cipher, but he had in his government at the time one François Viète, the seigneur de la Bigotière, a 49-year-old lawyer from Poitou who had risen to become counselor of the parlement, or court of justice, of Tours and a privy counselor to Henry. Viète had for years amused himself with mathematics as a hobby—“Never was a man more born for mathematics,” said Tallement des Réaux. As the man who first used letters for quantities in algebra, giving that study its characteristic look, Viète is today remembered as the Father of Algebra. A year before, he had solved a Spanish dispatch addressed to Alessandro Farnese, the Duke of Parma, who headed the Spanish forces of the League. Henry turned the new intercepts over to him to see if Viète could repeat his success.

  He could and did. The plaintext of the long letter from Moreo, in particular, was filled with intimate details of the negotiations with Mayenne: “… Your Majesty having 66,000 men in those states [the Netherlands], it would be nothing to allot 6,000 to so pressing a need. Should your refusal become known, all will be lost…. I said nothing about that to the Duke of Parma…. The Duke of Mayenne stated to me that it was his wish to become king; I could not hold back my surprise….” The message was couched in a new nomenclator that Philip had specially given Moreo when he departed for France; it consisted of the usual alphabet with homophonic substitutions, plus a code list of 413 terms represented by groups of two or three letters (LO = Spain ;PUL = Navarre’, POM = King of Spain) or of two numbers, either underlined (64 = confederation) or dotted (94 = Your Majesty). A line above a two-digit group indicated a null.

  Moreo’s letter had been dated October 28, 1589, and despite Viète’s experience and the quantity of text, it was not until March 15 of the following year that Viète was able to send Henry the completed solution, though he had previously submitted bits and pieces. What Viète did not know was that, 110 miles from Tours, Henry had defeated Mayenne’s superior force at Ivry west of Paris the day before, making the solution somewhat academic.

  Any chagrin that Viète felt did not deter him from extending his cryptanalytic successes. As he wrote to Henry in the letter forwarding the Moreo solution: “And do not get anxious that this will be an occasion for your enemies to change their ciphers and to remain more covert. They have changed and rechanged them, and nevertheless have been and always will be discovered in their tricks.” It was an accurate prediction, for Viète continued to read the enciphered messages of Spain and of other principalities as well. But his pride led him straight into a trap in which a shrewd diplomat drew confidential information from him as deftly as he elicited the secret meaning from elegant and mysterious symbols. Giovanni Mocenigo, the Venetian ambassador to France, said that he was talking one day with Viète at Tours:

  He [Viète] had just told me that a great number of letters in cipher of the king of Spain as well as of the [Holy Roman] Emperor and of other princes had been intercepted, which he had deciphered and interpreted. And as I showed a great deal of astonishment, he said to me:

  “I will give your government effective proofs of it.”

  He immediately brought me a thick packet of letters from the said princes which he had deciphered, and added:

  “I want you also to know that I know and translate your cipher.”

  “I will not believe it,” I said, “unless I see it.”

  And as I had three kinds of cipher—an ordinary which I used, a different one which I did not use, and a third, called dalle Caselle—he showed me that he knew the first. Then, to better probe so grave an affair, I said to him,

  “You undoubtedly know our dalle Caselle cipher?”

  “For that, you have to skip a lot,” he replied, meaning that he only knew portions of it. I asked him to let me see some of our deciphered letters, and he promised to let me, but since then he has not spoken further about it to me, and, having left, I have not seen him any more.

  Mocenigo was reporting to the Council of Ten, and it was after hearing his remarks that they so promptly replaced their existing keys.

  Meanwhile, Philip had learned, from his own interceptions of French letters, that Viète had broken a cipher that the Spanish—who apparently knew little about cryptanalysis—had thought unbreakable. It irritated him, and, thinking that he would cause trouble for the French at no cost to himself, told the pope that Henry could have read his ciphers only by black magic. But the tactic boomeranged. The pope, cognizant of the ability of his own cryptologist, Giovanni Batista Argenti, and perhaps even aware that papal cryptanalysts had themselves solved one of Philip’s ciphers 30 years before, did nothing about the Spaniard’s complaint; all Philip got for his effort was the ridicule and derision of everyone who heard about it.

  One of those who must have been laughing the hardest was probably a 50-year-old Flemish nobleman who had himself just completed solving a cipher of Spain. This was Philip van Marnix, Baron de Sainte-Aldegonde, right-hand man of William of Orange, who led the united Dutch and Flemish revolt against Spain. Marnix, an intimate of John Calvin and composer of what is today the Dutch national anthem, was also a brilliant cryptanalyst. An adversary described him as “noble, wise, gracious, sagacious, eloquent, experienced and with a very acute understanding, knowing the finer points of dealing with people. He is learned in Greek, Hebrew, Latin; he understands and writes the Spanish, Italian, German, French, Flemish, English, Scottish, and other languages very easily—better than any other man of this country. He is about 40 years old, of medium height, of dark complexion, but ugly of face. He is the greatest and most constant anti-Catholic in the world, more than even Calvin himself.”

  The Spanish cipher message that Marnix solved had been intercepted by Henry IV during his siege of Paris. The writer was, as withViète’s solut
ion, the luckless Juan de Moreo; the addressee was, as before, King Philip. Marnix had joined Henry at the siege; his reputation had apparently preceded him, for the French king himself turned over the three-and-a-half page cryptogram to his Protestant ally. Marnix’s solution revealed some of the jealous Moreo’s vituperations against the Duke of Parma (who served also as the Spanish governor of the Low Countries), accusing him in venomous terms of subverting Philip’s programs there. Henry had Marnix send both the decrypt-ment and the substantiating ciphertext to the duke in August of 1590, in the faint hope of stirring up some discord; the duke, who knew of Moreo’s calumnies and considered them beneath his contempt, found Marnix’s solution of sufficient interest to preserve it, but took none of the hoped-for actions.

  It was not the first time that Marnix had solved Spanish ciphers. Thirteen years before he had done it, and his demonstration of the value of cryptanalysis set in motion a train of events that culminated on a headsman’s block.

  In 1577, Philip was ruling the Netherlands through his half brother, Don Juan of Austria, its governor. The ambitions of Don Juan, then the first warrior of Christendom by virtue of his crushing defeat of the Turks at Lepanto, were not to be circumscribed by those constricted borders. He dreamed of crossing the Channel into England with a body of troops, dethroning Elizabeth, then marrying the seductive Mary, Queen of Scots, and sharing a Catholic crown of England with her. Philip consented to the invasion and the marriage, both to be begun as soon as Juan had restored peace in the Netherlands.

  But England did not sleep. Sir Francis Walsingham, Elizabeth’s satanic-looking minister, had built up an efficient organization for secret intelligence that reportedly had 53 agents in its pay on the Continent at one time. Walsingham first got wind that something was afoot when he heard of the marriage proposal. But his suspicions remained unconfirmed until the Huguenot general François de la Noue intercepted some of Don Juan’s enciphered letters in Gascony near the end of June, 1577. Since they presumably dealt with affairs in the Low Countries, they were sent to authorities there.

  Somehow they reached Marnix. Within a month he had broken the cipher. It was a typical Spanish nomenclator of the period, with a total vocabulary of about 200 words, the usual syllabary, and an alphabet. A peculiarity was that each plaintext vowel was given, in addition to its one literal and two numerical substitutes, a swash symbol as a substitute. Then, if a consonant preceded a vowel, this flourish was joined to the consonant’s ciphertext number to form a single combined character representing the two letters.

  The solutions seem to have all but disclosed Juan’s plan of landing his Spanish veterans in England under the guise of seeking refuge from storms that had blown him off course. William of Orange revealed the contents to Daniel Rogers, one of Walsingham’s agents, at a dinner at Alkmaar on July 11, in an attempt to persuade Elizabeth to come to his aid. Wrote Rogers in his report to Walsingham:

  The Prince [William] told me this that her Majesty might perceive how this negotiation of Don John and the Pope’s Nuncio agreed with the letters written by Don John and Escovedo [Juan’s secretary] in April last, and now intercepted. With that he called for M. de Sainte-“Allagunde,” whom he would have to bring the letters with him…. Sainte-Aldegonde brought nine letters, written all in Spanish, the most part of every one in cipher, excepting one. Three of these were written by Don John, two of them to the King [of Spain], the third to the King’s Secretary, Antonio Perez. The rest were all written by Escovedo to the King; it appeared by the seals and signatures they were no forged letters. The Prince also showed me the letter of La Noue, in which were enclosed all the said letters, as he had intercepted them in France. I thought good to pick out of them notes of the chief things contained in them.

  This report gave England tangible evidence of Philip’s aggressive intentions and perhaps provoked an increased watchfulness that served England in good stead when Philip finally did mount his invasion attempt in the Grand Armada eleven years later. Juan’s plot came to naught because he failed to negotiate peace with the rebels, which he needed before he could begin with England. But Walsingham, having learned of Marnix’s rare talent, induced the nobleman to solve messages for him. On March 20, 1578, he wrote to William Davison, an English agent in Flanders: “It is very important to her Majesty’s service to have this letter of the ambassador of Portugal deciphered with speed. Please therefore deal earnestly and speedily with St. Alagondye in that behalf. The cipher is so easy that it requires no great trouble.”

  Philip van Marnix’s solution of a nomenclator used by Don Juan de Austria, in 1577

  For once the sanguine expectations of a superior concerning the lack of difficulty of an assigned task proved correct. On April 5, Davison replied: “Sainte-Aldegonde is this day gone toward Worms…. His leisure before going did not suffice to decipher the letters you sent me with your last, but he procured me another to perform it. I send it herewith….” The lengthy letter revealed the ambassador complaining to his king about how Elizabeth feigned illness to avoid an audience he was seeking.

  Walsingham must have been dazzled by possibilities he never suspected when he first received Marnix’s solutions, for he took steps to assure himself of the rich flow of information provided by cryptanalysis without having to depend on foreign experts. Later that very year he had a bright young man in Paris devouring enciphered messages. This was Thomas Phelippes, England’s first great cryptanalyst.

  Phelippes was the son of London’s collector of customs, a not inconsequential post to which he himself later succeeded. He traveled widely in France, probably as a roving representative for Walsingham. On his return, he became one of the minister’s most confidential assistants. He was an indefatigable worker, corresponding tirelessly in his calligraphic hand with Walsingham’s numerous agents. His letters show a fair acquaintance with literary allusions and classical quotations, and he appears to have been able to solve ciphers in Latin, French, and Italian and, less proficiently, in Spanish. The only known physical description of him comes from the pen of Mary of Scots herself, who describes Phelippes, whose hair and beard were blond, as “of low stature, slender every way, eated in the face with small pocks, of short sight, thirty years of age by appearance.”

  Mary’s unflattering comments betrayed her suspicions about Phelippes—suspicions that were well founded. For Phelippes and his master, Walsingham, were casting a jaundiced eye on Mary for reasons that, in their turn, were equally well founded. Mary was the heir apparent to the throne of England. She was also nominally queen of Scotland, though she had been ejected in a tangled series of events and had been prevented from returning by the opposition of the strong Protestant party there to her indiscretions. She was a remarkable woman: beautiful, possessed of great personal charm, commanding the loyalty of her subordinates, courageous, unshakably devoted to her religion, but also unwise, stubborn, and capricious. Various Catholic factions had schemed more than once to seat her on the throne of England and so restore the realm to the Church. The chief result had been to confine Mary to various castles in England and to alert Walsingham to seek an opportunity to extirpate once and for all this cancer that threatened to destroy his own queen, Elizabeth.

  The opportunity arose in 1586. A former page of Mary’s, Anthony Babington, began organizing a plot to have courtiers assassinate Elizabeth, incite a general Catholic uprising in England, and crown Mary. A conspiracy that involved the overthrow of the government naturally had ramifications all over the country, and Babington also gained the support of Philip II, who promised to send an expedition to help, once Elizabeth was safely dead. But the plan depended ultimately on the acquiescence of Mary, and to obtain this Babington had to communicate with her.

  This was no easy task. Mary was then being held incommunicado under house arrest at the country estate of Chartley. But a handsome former seminarian named Gilbert Gifford, recruited by Babington as a messenger, discovered a way of smuggling Mary’s letters into Chartley in a beer keg.
It worked so well that the French ambassador gave Gifford all the correspondence that had been accumulating for Mary for the past two years.

  Much of it was enciphered. But this was only part of the care that Mary took to ensure the security of her communications. She insisted that important letters be written within her suite and read to her before they were enciphered. Dispatches had to be sealed in her presence. The actual encipherment was usually performed by Gilbert Curll, her trusted secretary, less often by Jacques Nau, another secretary. Mary not infrequently ordered changes in her nomenclators, which were much smaller and flimsier than the diplomatic ones.