THE CODEBREAKERS

by DAVID KAHN

  Then, too, there is the matter of the ship’s captain, that dignitary whom we are accustomed to think of as a strong, silent man—alert, commanding, and always on the job. The code-book picture of him is different and more than a little disconcerting. By the time we have finished reading messages like Captain lost overboard, Captain not to be found, Captain drunk, Captain refuses to leave vessel, and Captain insane, it is with considerable relief that we light upon the entry Arrest the Captain. It would seem to have come not a moment too soon.

  But even if the captain avoids these pitfalls, and the ship itself escapes the ravages of the storm, the code has still other hazards in store for the unhappy voyager. At any moment the ship may be captured by pirates (ENIMP) or plundered by natives (YBDIG). There is always the chance that the captain will receive cabled instructions to arrest all passengers (ZEIBI). Even less consoling to the prospective traveller are such glimpses of world hygiene as IDDOG (Ship in port, all hands down with scurvy) and OAVUG (An epidemic of foot-and-mouth disease has broken out here). The only ray of light is provided by EWIXI (Very few cases of cholera are now reported), and even that statement is not without its depressing implications.

  There is no denying, however, that the code book is full of helpful information. Should you ever, for example, feel the need of lard, in bladders, the word is CHOOG. Flannel shirts are GOLPO. Cod-liver oil is called GAHGU—and a very good word for it, too, as is FOOLP for ship’s biscuit. A niblick is, of course, a GAZEB, but the word for foot-warmer is FREIZ. No matter what commodity it is you desire, you will find it covered in the code, which includes a list of nearly a thousand necessities of life, ranging all the way from arsenic to ostrich feathers, from blasting charges to porch umbrellas. Even the commodities, however, are blighted by the same spirit of melancholy that pervades the entire code, and the result is such decadent listings as ZOKIX (unhealthy trees) and GNUEK (rubber, slightly moldy).

  But it is in its cross-references that the code book reaches the logical limits of pessimism. For gratuitous gloom, it would be difficult to equal such groupings as “Ankles: see Accidents” or “Chief topic on the Stock Exchange: see Failure.” In other cases, however, the effect is merely rococo, as in “Marriage: see Hotel Accommodations” or “Noses: see also Fittings, Machinery, and Spare Parts.”

  Valuable as it is from both the literary and the practical standpoint, it is plain that the code book was intended for people who get around more than I do. Such well-meant suggestions as DEOBI (A great battle is now raging here), PUMZI (Can you combine horses and grain?), and EZUCZ (Calling at Elephant Point for orders) can hold for me at best only an academic interest. But any of these messages is a monument of utility beside the picturesque YBTUA, which deals with the transportation to Mecca of pilgrims—at the prevailing price per head! And however much I may regret my inability to send a message like WUMND (Have every reason to believe oil will be struck), at least I feel certain that I shall never rise, Phoenix-like, from my own ashes to cable that most fantastic of all code words, AHXNO: Met with a fatal accident.

  * Reprinted by permission, copyright © 1934, 1962 The New Yorker Magazine, Inc.

  23

  CIPHERS IN THE PAST TENSE

  ALL CRYPTANALYSTS have not borne arms for Mars. Some of the most prolific have served Clio, the muse of history. Many of these unsung heroes—the only cryptologists whose contributions enlightened all mankind—worked in the 19th century. The immense influence of Leopold von Ranke’s objective school of history, which demanded a study of the original documents, sent droves of historians to mine state papers and diplomatic correspondence in the archives, whose doors had been unlocked for the first time by the nationalism and democracy of the 1800s.

  The researchers found many of the documents in cipher, or partly so. Invariably, it seemed, the crux of a dispatch was enciphered. In the mid-1500s, a Venetian ambassador wrote home about his talk with Henry II of France concerning English affairs. “His Majesty suddenly turned to me, taking a troubled aspect and shrugging his shoulders, added to me these very words….” and the rest is in cipher! Historians realized that the most important parts were the most likely to be put into cipher. Some, unfamiliar with crypt-analysis, apparently regarded the resultant cryptogram as an act of God, an insuperable obstacle which they would have to live with as with a hole in the document. “Were we able to decipher the letters written on congressional politics by Richard Henry Lee and his correspondents … no doubt much of the cloud which hangs over the congressional intrigues of that critical period would be removed,” mourned Francis Wharton in 1889 in The Revolutionary Diplomatic Correspondence of the United States.

  But other scholars looked upon the cryptograms as a challenge. One of the first of these was a transplanted German whose services to English historiography were of high importance.

  Gustave Adolph Bergenroth was born February 26, 1813, at Marggra-bowa, which his biographer called “an insignificant town in the remotest and dreariest corner of East Prussia.” He attended the University of Konigsberg, where he was very popular with his fellow students and where he sustained a severe injury to his right wrist in duelling. After working in Cologne and Berlin as an assessor, with time out for a trip to Italy necessitated by his liberal views, he quit his job and sailed in 1850 for California as a pioneer. The racy style of his first composition in English, “The First Vigilance Committee,” drew favorable attention, and he determined to write. After some literary work, he began a history of Tudor England. Finding the available materials insufficient, he set out, late in his forties, for that great repository of documents for those years when Spain bestrode the world, the Archivio General at Simancas in northwest Spain. His letters home soon won him a stipend from England’s Master of the Rolls to find, list, and summarize the state papers at Simancas that related to English history and to prepare a volume for the Spanish series of the endless Calendars of State Papers. He forgot his Tudor history.

  He arrived at Simancas in September, 1860, and established himself in a kind of hotel, the Parador della Luna, where he would do much of his crypt-analysis. An Englishman who visited him painted the scene: “Simancas is a collection of wretched hovels, half buried in dust and sand. There is not a good house in the place. The one in which Mr. Bergenroth lives belongs to a farm bailiff, consists of two storeys, all the rooms of plaster, and the floors of brick. No fireplace in any of the rooms, and, as the winter is very intense here from November to February, and the walls full of holes, nothing but the strongest desire to do service to history could reconcile any man to so much hardship.” Bergenroth had, moreover, to overcome some of the oddest phenomena ever to interfere with cryptanalysis. The plaza beneath his room was crowded with shouting donkey-drivers and visited frequently by a dulciana, whose “shrill notes, continually playing an air from Traviata and one Spanish melody, and nothing else, drive me almost mad.” His landlady liked to strum on her guitar, and “none but drivers of bullock-carts could, for a single night, stand the music of the Lady della Luna.” The kitchen girl “hangs my linen and that of the whole family over my balcony for drying, and then, with laudable resolution, sets to ironing it on my writing-table.”

  More troubles faced him at the Archivio General. It consists of an old castle, with crenellated walls pierced by loopholes, surrounded by deep moats and drawbridges. Its 46 rooms contain more than 100,000 bundles, or legajos, in each of which are filed from 10 to 100 documents, making a total of several million. From this staggering accumulation Bergenroth had to select the pertinent items. It was hard for him even to get at them. When Spain’s archives administration finally granted him entrée, the crabbed Renaissance semiuncials made long and dogged practice necessary before he could read the handwriting. Indeed, the archivist himself had often been defeated by it, and in his jealousy at Bergenroth’s success he deliberately hampered the historian’s work by refusing access to such cipher keys as were in his possession. Bergenroth had to recover them by himself, as well as those
keys that had been lost.

  The story of his cryptanalytic endeavors can be pieced together from several of his writings.

  I did not go to Spain quite unprepared for my work. I had carefully studied the Paléographie of Christoval Rodriguez; I had also spent much time in deciphering such old Spanish documents as were to be found in the libraries of London and Paris….

  [At Simancas.] The first thing I considered it necessary to do was to study most carefully, not only the Spanish orthography of the period, but that of each statesman in particular who could be supposed to have written any of these letters. Even this was not sufficient. I had to study the turns of thought, and the favourite words and expressions of each statesman. Long and curious lists, covering many sheets of paper, lay during many months on my writing-table, and were stuck up against the wall of my room.

  I did not discover any of the keys to the ciphers in a methodical manner. Whilst engaged in copying I was constantly on the watch for a weak point, convinced that no man can for any length of time succeed so completely disguising his thoughts but that he will occasionally betray himself to a close observer. Wherever I thought that that was the case, I tried to guess the meaning of the signs. A hundred times I may have done so in vain, but at last I triumphed….

  When copying an instruction to the Duke [de Estrada], I discovered little dots, like full stops, behind two signs of cipher. As interpunction is never used in cipher of this kind, the dots could only be signs of abbreviation. But even abbreviations (a skilful writer would never have made use of them) offer so many difficulties that they can be employed only on the most common occasions, as, for instance, V. A. for Vuestra Alteza, or n.f. for nuestra fija, or nuestro fijo. From obvious reasons [in this case], I decided in favor of “nuestra fija,” and inferred further that the preceding signs must correspond to “princesa de Gales.” The breach was opened, and before three o’clock in the next morning I was in possession of eighty-three signs, representing the letters of the alphabet, and of thirty-three monosyllables, signifying words. The key is far from being complete, but there remain no longer unconquerable difficulties…. [This cipher] of the Duke de Estrada is the most difficult, and at the same time the most important of all, as a greater number of undeciphered despatches are written in it than in any other kind of cipher….

  The question may be asked, whether my decipherings are trustworthy? I answer with full confidence in the affirmative. I have more reason than one for doing so. After I had deciphered the despatches, I found, in some instances, that they were only ciphered copies of drafts in plain writing. Thus I had an opportunity of comparing my interpretations with the originals, and found that in all essential points they were identical. The key of De Puebla and the fragments of the two other keys, which were given to me after my return from Madrid, provided me with an additional test. The keys which I had already formed before seeing them coincided perfectly with them…. But the general and most decisive proof consists in the meaning of the despatches, concealed behind the cipher.

  In ten months, Bergenroth surpassed the feats of many professional crypt-analysts by reconstructing 19 nomenclators—an average of about one every two weeks, some with 2,000 or 3,000 elements. This was in addition to his own copying, his supervising of a copyist, his searching for documents, his battles with the bureaucracy, and his frequent letters home. He did not like the cryptanalysis: “Nothing but sheer necessity would have forced me to attempt such a task, which, I think, is one of the most laborious that any man could undertake.” Yet by July 23, 1861, ten months after his arrival, he could report, “The despatches in cipher are all copied and deciphered, with the exception of two small letters (the one of them from John Stile to Henry VII.), which I intend to decipher in Barcelona or in London. I am now too fatigued for a work which requires so much concentration of thought as the discovery of keys to unknown cipher does.” He did solve the Stile letter, but not the other, a short one from King Ferdinand and Queen Isabella dated at Segovia on August 20, 1503, the only one in that key. This key was the only one of those used by Spain during the reign of England’s Henry VII (1485-1509) that he failed to read.

  Gustave Bergenroth’s reconstruction of a Spanish cipher

  One long dispatch, whose solution took a week, typifies the treasures he unearthed. It is a letter of July 25, 1498, from Don Pedro de Ayala in London to Ferdinand and Isabella, reporting on England’s fitting out of an expedition to some islands in the New World which, Ayala thinks, had already been discovered by Columbus and were owned by Spain. He apparently referred to the second voyage of John Cabot, on whose discoveries the English claims to North America rested. Though some of the nomenclators that Bergenroth recovered were later found in the archives, many others never were, and only his cryptanalyses brought the documents to light. Bergenroth died in 1869 of a fever contracted at Simancas, but the results of his labors shine today in the close-printed pages of his Calendars of Letters, Despatches, and State Papers Relating to the Negotiations Between England and Spain. To their résumés of hundreds of documents, the historians return time and again, with gratitude.

  Bergenroth had been assisted in his cryptologic work at Simancas by Paul Friedmann, who appears to have been some kind of itinerant archivist. For France’s Bibliothèque Nationale he compiled a collection of ciphers employed by various French political correspondents of the 16th century. In 1868 he became interested in the cipher messages of Giovanni Michiel, Venetian ambassador to the court of England’s Queen Mary, sister and predecessor of Elizabeth J. No one in the archives of Venice could read the dispatches, and, though photographs of them were sent to England for attempts at solution there, they continued to baffle everyone. In Venice, Friedmann examined the Michiel correspondence and “soon arrived at the conviction that the cipher was not one of extraordinary difficulty, that it was not always used with sufficient care, and that with a little labor the sense might be discovered.” It used about 200 signs, and within a few months he had solved it. Thus d11 was bo, d12 was g, t25 Sua Maesta, and so forth.

  “Michiel’s correspondence is of a considerable value,” Friedmann wrote. “It will redress many errors, and fill many a gap in the narratives….” For example, historians had generally considered that the transfer of the future Queen Elizabeth from Woodstock to Hampton Court took place in June, 1554, and was a release from prison after all hope of Mary’s having a child had faded and she no longer needed to keep the presumptive Protestant heiress under her control. Michiel’s letters make clear that what happened was just the reverse. The removal took place not in June but in April, at the very moment when Mary was expected to give birth, and it was not a release but a tightening of security in the face of the grumblings of the populace against the thought of a Catholic offspring of Spain’s Philip and Mary upon the English throne. No child was born, of course, and Elizabeth later ascended the throne, to the delight of her people. Thus did cryptanalysis help rectify knowledge of a tense episode in the life story of one of England’s greatest sovereigns.

  Friedmann bitterly complained that an Italian archivist, Luigi Pasini, attempted to claim credit for the Michiel solution when, in fact, he merely augmented it. This was true, but Pasini did achieve some notable cryptanalyses. He had begun work in the Archivio di Stato of Venice in 1855, when he was 20. Ten years later, he became interested in the Venetian ciphers and began collecting keys and documents concerning them.

  His enlargement of the Michiel nomenclator won him the commendation of the Master of the Rolls, and when a French scholar, Armand Baschet, heard of it, he encouraged Pasini to attack the enciphered dispatches of the Venetian ambassadors to the court of France for the last four years of the reign of Henry II, the three-year reign of Francis II, and the first five years of Charles IX, for all of which no decipherments or keys could be found. Pasini, an intelligent and likable young man, succeeded in solving about 5,000 lines, and Baschet, who had at first considered not publishing the dispatches for those years because the most v
aluable information was concealed under the cipher, could declare: “Thanks to his [Pasini’s] extraordinary aptitude, the dispatches of six Venetian ambassadors for a period of twelve years, written in the impress of great events such as the last struggles of the King of France with the [Holy Roman] Emperor and the Spaniards, and the first religious wars, have recovered their extreme interest.” Pasini also assisted Baschet cryptologically in a study of the letters of Aldus Manutius, the great printer. He continued to collect material on the history of Venetian ciphers until his death in 1885.

  Another Italian archivist, Domenico Pietro Gabbrielli, an abbot, was positively awesome in his cryptanalytic industry. He was appointed an apprentice archivist in the diplomatic section of the Archivio di Stato of Florence early in 1854, when he was 30. Ten years later he began a nine-year marathon of historical cryptanalysis, becoming probably the most expert solver of nomenclators who ever lived. He solved 400 in his first seven years, a rate of better than one nomenclator solution per week. He owed this apparently incredible facility to his progression from the simple ciphers of the 1400s to the full nomenclators of the 1700s, which familiarized him with the quirks and trends and phraseology of Florentine cryptography, to the probable similarity of many of the nomenclators, to the great volume of material at hand, and to his own ability. In addition to his solutions, he reconstructed twice as many nomenclators from existing plain and cipher versions of dispatches.