by DAVID KAHN
Holden’s description makes explicit one requirement of successful operation of multiple anagramming (that the two messages be the same length) but presupposes the other (that their keys be the same). The technique rests on the fact that, if two messages of the same length are transposed in the same system with identical keys, their individual words will wind up in the same relative positions. To put it differently, if the first word of the plaintext becomes the 15th word of the cryptogram in the first message, the first word of the plaintext of the second message will equally wind up in the 15th position of the second cryptogram. This is transposition’s version of like causes producing like effects, and the principle holds for all transposition systems, letter as well as word, irrespective of their mixing process.
The principle may be illustrated with two five-letter cryptograms enciphered with the same key: GHINT and OWLCN. Suppose that the cryptanalyst begins trying to reconstruct the plaintext of the first message by assuming that it begins with th. This implies an encipherment key which moved the first plaintext letter (t, in this message) to the fifth position (in GHINT) and the second plaintext letter (h) to the second position (in GHINT). The cryptanalyst can determine that the same key would require the second message to begin with nw—hardly a promising beginning. If the cryptanalyst now tries to anagram the second message instead, he might try cl as a starter. The corresponding moves in the first cryptogram would bring n and i together at the head of the message. This gives good possibilities in both messages, which is, of course, more desirable. The cryptanalyst will continue juggling the two messages, checking one against the other, until he reconstructs them both as night and clown. The key he recovers will solve any other five-letter cryptograms enciphered by it. The process must be done individually for each key and each cryptogram of different length. Multiple anagramming cannot work with just a single message because without any control the single message could be anagrammed into too many equally likely texts. GHINT alone, for example, could be unscrambled to make thing as well as night, and there is no cross-check to tell which is right.
The word-transposition system carried the most explosive and the greatest number of messages sent by the Democratic politicians, but it was by no means the only one. Messages from Florida and South Carolina were evidently encoded by a dictionary, but the one used for the Oregon disclosures did not unlock them. The three tyro cryptanalysts had independently noticed that these dispatches included the word geodesy, which is a rather unusual term for the pocket dictionary that they reasoned would probably be used. Holden found the right one after an hour and a half’s search in the Library of Congress; he telegraphed the news to the Tribune just as a bleary-eyed staff member, who had examined 40 or 50 volumes without success, was about to go out and check the one that Hassard and Grosvenor rightly suspected—Webster’s Pocket Dictionary. It was used in the same way as the Oregon dictionary, though the number of pages turned to the front to select the codeword varied from one to five.
The Democrats also used pairs of numbers in a monalphabetic substitution. Hassard broke this system by guessing that the patterned ciphertext 84 66 33 87 66 27 27 mirrored canvass. He cracked a checkerboard substitution when he divined that ITYYITNS in a partially enciphered telegram from Florida stood for the name of the county of Dade. The coordinates of the checkerboard (which also served for the two-digit cipher) proved to be ten different letters that spelled a phrase of extraordinary suitability:
Of the 400 dispatches that were given to Hassard and Grosvenor, all but three (in a cipher not used elsewhere) were translated. The Democrats, unaware that their own machinations were being bared, raised the cry of fraud in the presidential election as the midterm campaign for Congress grew hot. On October 3, 1878, the Tribune reported that solution of the dispatches had been completed and published a few of them as a hint for the Democrats to confess. But they said nothing, and four days later the Tribune thundered out the story of Democratic intrigue in Florida and Louisiana. The first story detailed the operations of the ciphers; the second, next day, exposed the texts of the telegrams. On October 16, the South Carolina shenaningans came out. Their sum was that Colonel William T. Pelton, Tilden’s nephew and confidential secretary, had bargained through Marble and others for electoral votes.
“Cipher Mumm(er)y. Exhumed by the New York Tribune.” Cartoon of Samuel Tilden by Thomas Nast in Harper’s Weekly
The results were sensational. The public marvelled at the ingenuity of the cipher-solvers. Thousands of readers tested the keys and satisfied themselves as to the accuracy of the solutions. The Democrats argued that the telegrams were strictly for Pelton’s information, but it seemed clear that only Pelton’s hesitation at the price and the subsequent bungling and delay subverted his intentions. The Tribune had prepared its exposé thoroughly and presented it skillfully; even its Democratic rival, the Sun, was forced to a grudging tribute. The timing, too, was perfect: election was only a few weeks off. In that election, the G.O.P. made emphatic gains in Congress. New York, Pennsylvania, Massachusetts, and Connecticut voted, as the Tribune inferred with pardonable pride, to rebuke the cipher fraud.
But the effects did not stop there. The telegrams had been addressed to Pelton at 15 Gramercy Square, New York, Tilden’s home, and though Tilden, haggard and with his perpetual cold, swore before the Congressional investigating committee that he had no personal knowledge of what his nephew was doing in his house, and that anything that was done was done without his permission, his reputation was sullied. The disclosures ended his presidential aspirations. As his old supporter, the Sun, sadly conceded, “Mr. Tilden will not again be the Presidential candidate of any party.”
In fact he was not, and in the election of 1880, James A. Garfield, a personal friend of Reid’s, defeated Winfield S. Hancock, the Democratic candidate, by only 7,000 votes out of 9,000,000 in the popular tally but by an unchallengeable 214 to 155 in the electoral ballot. Even a sympathetic biographer of Tilden acknowledged that “As a result of the cipher telegrams the Republicans won an advantage which probably gave them the national election of 1880. Much of the public became convinced that the millionaire candidate for the Presidency had permitted his party directors to dip into his purse to win a decision for the party that was willing to pay the highest price.” Cryptanalysis had helped elect a President. The Tribune’s triumph stood forth as one of the first great journalistic exposés of governmental corruption, which helped elevate American newspapers to their role of public watchdog. It also carried the Tribune into the citadel of Republican power. Reid later banqueted at its tables when he was named ambassador to the Court of St. James’s. But perhaps the most lasting value of the Hassard-Grosvenor cryptanalysis and its dramatic disclosure by the Tribune was noted by Reid’s biographer: “It had pilloried once and for all the single manifestation in our annals of the idea that the Presidency was a purchasable honor.”
* These do not include the ciphers—mostly simple word transpositions—that the several military departments employed within their own territory. The Department of the Missouri used these more extensively than any other. A number of other cipher systems were proposed by members of the infant Signal Corps. They generally consisted of various types of poly-alphabetic systems, and one of them—a fanlike set of 26 wooden tablets, each with a different ciphertext alphabet on it, designed for use with a keyword by Sergeant Edwin H. Hawley—matured into the first United States patent granted for a cipher device (No. 48,681, July 11, 1865).
8
THE PROFESSOR, THE SOLDIER, AND THE MAN ON DEVIL’S ISLAND
ONLY A FEW BOOKS in the history of any science may be called great. Some of these report a technical innovation that radically alters the content of the science. Through the 19th century, Alberti’s and Kasiski’s were the two great books of this kind in cryptology. Such books look inward.
Other great books look outward. They bring the science up to date—make it consonant with its time—and so renew its utility to men. This th
ey do by assimilating developments in relevant fields (for example, improvements in instrumentation), by summing up the lessons of recent experience and deducing their meaning for the current age, and by reorganizing the concepts of the science according to this new knowledge. This does not mean simple popularization, though such a work usually does have an organic persuasiveness. Rather, it amounts to a reorientation, a new perspective.
For 300 years, the only great book of this kind in cryptology was Porta’s. He was the first to delineate a coherent image of cryptology. His ideas remained viable so long because cryptology underwent no essential change; communication was by messenger, and consequently the nomenclator reigned. But his views no longer sufficed after the invention of the telegraph. New conditions demanded new theses, new insights. And in 1883 cryptology got them in the form of its second great book of the outward-looking kind, La Cryptographie militaire.
Its author was born Jean-Guillaume-Hubert-Victor-François-Alexandre-Auguste Kerckhoffs von Nieuwenhof on January 19, 1835, at Nuth, Holland, son of a well-to-do landlord and a member of one of the oldest and most honorable families of the Flemish duchy of Limburg. He went to school at a little seminary near Aachen. Afterward, to improve his knowledge of English, he lived in Britain for a year and a half, then returned to the University of Liège, where he received two degrees, one in letters, one in science. After teaching modern languages for four years at two schools in Holland and joining a number of literary societies there, he accompanied a young American, Clarence Prentice, son of the founder of the Louisville Journal, through England, Germany, and France as traveling secretary, then went to Meaux, near Paris, where he again taught modern languages.
In 1863, he obtained the chair of modern languages at the high school at Melun, a large town 25 miles southeast of Paris. The next year he married a girl from the area and in 1865, when he was 30, they had their only child, a daughter, Pauline. He stayed at Melun for 10 years, teaching English and German. He supplemented his salary of about 1,600 francs by taking students in to lodge with him—a practice that was officially prohibited but winked at.
During these years he participated in a variety of activities that show the great diversity of his interests. He gave lectures on the formation of languages and on literature, founded a society for the encouragement of education in Melun, gave free courses in English and Italian, served as delegate of the local branch of the French Society of Archaeology to the international congress at Bonn in 1868, and got embroiled in some minor political difficulties after the French defeat of 1870. His learning was broad enough for him to fill in at different times for teachers of Latin, Greek, history, and mathematics.
By this time he had shortened his name to Auguste Kerckhoffs. Bearded, dignified, slow of speech, Kerckhoffs, despite an inability to maintain discipline in his classes and some eccentricities of character, was a “learned, zealous, capable” teacher who awoke his students’ interest in their work; his superiors said “his students like him and work with success.” Thus when a hostile official wanted to turn down Kerckhoffs’ request for a leave for further studies, he discovered that the teacher had “ardent protectors,” and the leave was granted.
Kerckhoffs went from 1873 to 1876 to the universities of Bonn and Tübingen, getting his Ph.D. He earned his living by teaching the young Count de Sao Mamede, who later became secretary to the king of Portugal; Kerckhoffs was made a commander of the Order of Christ for this. He then returned to Paris, where he worked as a private instructor, teaching two younger sons of the Sao Mamede family. He demonstrated an interest in things military by applying for the chair of German at the Ecole Militaire Superieure in 1878, losing it because a clerk failed to note that he had become naturalized as a French citizen in 1873. In 1881, Kerckhoffs became professor of German at the Ecole des Hautes Etudes Commerciales and at the Ecole Arago, both in Paris. It was during this time that, aged 47, he wrote La Cryptographie militaire. It was not his first book: he had already written a Flemish grammar, an English grammar, a German verb manual, a study (in German) on the origins of German drama, and a work examining the relation of art to religion.
His busiest years followed the publication of La Cryptographie militaire. A new international language called Volapük (“World-Speak”) had been invented by a German priest, Johann Martin Schleyer. About 1885, it caught on in France, and flashed with express-train speed all over the country, not only among intellectuals but among all classes: it was even heard in the streets. From France it radiated throughout the world. The most active propagandist of Volapük was Auguste Kerckhoffs, who, at the second Volapük congress in Munich in 1887, was acclaimed director (“Dilekel,” in Volapük) of the International Academy of Volapük. To this body were submitted questions of the grammar, vocabulary, and orthography of the expanding tongue.
As secretary of the French Association for the Propagation of Volapük, Kerckhoffs proselytized the artificial language with ability and vigor. In 1888, 182 textbooks on Volapük appeared—a publication rate of one every other day—and the Macy’s of Paris, the Grands Magazins des Printemps, sponsored courses in it. By 1889, 25 periodicals in or about the language were being published and 283 Volapük clubs were meeting all over the globe. When the third Volapük congress was held at Paris in May of 1889, with Kerckhoffs presiding, even the waiters and porters conversed in World-Speak. A new Golden Age of brotherhood, unencumbered by the chains of Babel, seemed to shimmer just ahead.
It was a mirage. For the congresses, which seemed to be the harbingers of that great day, were actually symptoms of critical tensions within the movement. Schleyer’s goal of creating the richest and most perfect literary language, in which he was supported by the German Volapükists, clashed with the desire of Kerckhoffs and the other active Volapükists to have the simplest and most practical language for commerce and science. From the beginning, Kerckhoffs had eliminated from his grammatical manuals some of the forms that Schleyer had carried over in Volapük from his native German, such as the endings for the jussive and optative moods of verbs. But the priest insisted that, as the father of Volapük, he should have the final decision on any changes. The tensions mounted, and when the Academy refused to grant Schleyer the full veto he wanted, the movement broke in two.
It splintered into bickering factions entirely unable to agree when Kerckhoffs submitted to the Academy, not individual questions, but a complete grammar, and other members of the Academy retorted with projects of their own. The movement crumpled with unbelievable swiftness: in 1889, it seemed as though it would conquer the world; in 1890, it was moribund. Kerckhoffs resigned as Dilekel in 1891, and, by 1902, of the estimated 210,000 enthusiasts the language had once had, only 159 remained on its List of Correspondents, and only four little clubs clung weakly to life. Kerckhoffs’ Cours complet de Volapük, his Dictionnaire Volapük-Français et Français-Volapük, his Vollstän-diger Lehrgang des Volapük remain only as forgotten monuments to a splendid dream.
Crushed and perhaps embittered by the collapse of what had seemed so needful and so certain, Kerckhoffs one day exploded with some intemperate criticisms of the handling of the state’s school examinations so that his contract at the Ecole des Hautes Etudes Commerciales was not renewed in 1891. It was only through the intervention of influential friends that he managed to get a post teaching German at the high school at Mont-de-Marsan, near Bordeaux. Here, his superiors reported on him: “Very diverse and extended knowledge taught with more method, exactness and precision than I would have expected in a spirit that embraces so many things. Highly regarded and highly appreciated.” The following year, trying to get closer to Paris, Kerckhoffs moved to the Brittany seaport of Lorient, where he again taught German. In the middle of that school year, his daughter died. He stuck it out for another year, but by 1895, then 60, his health failing, his spirit broken, but living in Paris not far from the Sorbonne, he applied for a year’s leave. He renewed it annually until his death in Switzerland, apparently while on vacation, on Augus
t 9, 1903.
But if his works on Volapük are defunct, his cryptologic ideas still flourish. La Cryptographie militaire first appeared as two installments in the Journal des Sciences militaires in January and February of 1883, being reissued later that year as a paperback book by the journal’s publisher. It is the most concise book on cryptology ever written. Kerckhoffs had the instinct for the cryptographic jugular, and he compressed into 64 pages virtually the entire known field of cryptology, including polyalphabetics with mixed alphabets, enciphered code, and cipher devices. The book is also one of the most scholarly on cryptology. Its footnotes cite most classical and many modern sources; comments such as “This is not the only historical or bibliographic error for which the Austrian writer must be reproached” show how carefully the author has studied those sources. And the book throbs with life. Kerckhoffs selected an enciphered news-service dispatch as the specimen for a demonstration solution. He discussed current German practice and contrasted it with what was then going on in France. He scrutinized the most recent ciphers, such as the Wheatstone device. He concentrated upon it all his extraordinary range of knowledge, and it is perhaps significant that at least three of the great books of cryptology—Kerckhoffs’, Alberti’s, and Porta’s—were written not by narrow specialists but by well-rounded men who had one foot in each of what C. P. Snow would later call “the two cultures” of science and humanities.
