by DAVID KAHN
The number of well-known authors who have touched upon cryptology in their works is surprising. William Makepeace Thackeray, the six-foot four-inch author of Vanity Fair, employed a Cardano grille in 1852 in The History of Henry Esmond, the novel that has been called his most artistic and for which he did months of research in the British Museum. The cipher was used in the book’s account of the plot to put the would-be James III on England’s throne—the same plot in which Bishop Atterbury was convicted, largely on cryptanalyzed evidence. No cryptanalysis was involved here, though, simply the transmission of a message.
Jules Verne heightened the excitement of three of his novels with the mystery of secret writing. In general, he handled cryptology as well as he did other technical matters. But he marred his two solutions with a physical or a psychological improbability. These hurt the credibility of his cryptanalyses more than the technological improbabilities injured the credibility of his science-fiction fantasies, because they violated immutable laws of nature whereas the fantasies merely exceeded man’s then-current technical capabilities. On the other hand, just as he foresaw submarines, projectiles from the earth to the moon, and speedy trips around the world, so Verne anticipated a technique of cryptanalysis.
He began the book that was his second great success and that clinched his reputation, Voyage to the Center of the Earth, with a three-step cryptogram. On a slip of parchment, it flutters out of an old runic manuscript that Professor Otto Lidenbrock has just bought. It too is in runic—21 groups of six characters each. Lidenbrock converts it into roman letters; it makes no sense, so he rearranges the letters in the form of a columnar transposition, then transcribes them. Still no luck. Later, his young nephew, Axel, “mechanically” fanning himself with the paper bearing the transcription, sees the writing through its back and discovers that it is simply a Latin text written in reverse. Here the plausibility slips a little on several counts: Can one read through paper so easily? Would Axel have been able to grasp the significance of the letters, which would appear backward? Probably not. But all is forgotten in the excitement of reading the plaintext: “Descend the crater of the volcano of Sneffels when the shadow of [Mount] Scartaris comes to caress it before the calends of July, audacious voyager, and you will reach the center of the earth. I have done it. Arne Saknussemm.” And Lidenbrock and Axel, following the cryptanalyzed instructions, do it too.
In La Jangada, Judge Jarriquez, after failing to solve a cryptogram as an ordinary monalphabetic, concludes that it is a Gronsfeld (a Vigenère with numerical key) because it contains a group of three repeated letters. This induction is poorly founded—many ciphers produce three of the same letters in a row—and is the flaw in the solution. A Gronsfeld it is, however, and Jarriquez solves it by trying, on the basis of outside information, the name Ortega as the probable word that ends the message as a signature. He sees at once that the six final letters, SUVJHD, all stand farther back in the alphabet than the letters of Ortega, which supports his hypothesis. Jarriquez uses the probable word to recover the key 432513 and then tests it at the beginning of the cryptogram. He gets a lucky break when the one chance in six that the key begins with the 4 works out, and he reads the plaintext right off. Though this rather obvious technique may well have been used before, Verne’s exposition of it here in 1881 is the first to appear in print, and it may even be regarded as the forerunner of the identical method of using a probable word to recover a polyalphabetic key that Bazeries published twenty years later with the boastful cry that it was “a new procedure of decipherment which neither Kasiski, nor Kerckhoffs, nor Josse, nor Viaris, nor Valério have described!”
Verne’s final fictionalization of cryptology, in Mathias Sandorff, involved no solution, since Sarcany finds the grille that enciphered the message. In this book, published in 1885, Verne cited Eduard Fleissner von Wostrowitz’s Handbuch der Kryptographie, published four years earlier, which eulogizes grilles. Verne probably also read Kerckhoffs’ La Cryptographie militaire, published two years before Mathias Sandorff, since Verne’s language in discussing the requirements of good ciphers is strongly reminiscent of Kerckhoffs’. He did not absorb all Kerckhoffs’ information, however, since he proclaimed the grille and enciphered code as unbreakable.
The most famous of fictional detectives, Sherlock Holmes, encountered ciphers not once but three times in his distinguished career (excluding a simple signaling system of light flashes and a word puzzle). In “The ‘Gloria Scott,’” the great detective soon discovers that a secret message is hidden within an open-code text as every third word. In “The Valley of Fear,” given a numerical code message from an accomplice of his arch rival, Professor Moriarty, he reasons his way brilliantly not only to the conclusion that it is a book code, but to the very volume used. The book must be both readily available and standardized as to format. This excludes the Bible, which meets the requirement of availability to perfection but that of standardization of pages not at all—and also because “I could hardly name any volume which would be less likely to lie at the elbow of one of Moriarty’s associates.” The only volume which fits both requirements is Whitaker’s Almanac. The current edition yields the senseless Mahratta government pig’s bristles, but last year’s gives perfect sense. Thus Holmes solves the cryptogram purely by use of his famed deductive powers and without really needing to know cryptanalysis.
But his thorough knowledge of that subject, as of all others needed in his chosen profession, becomes manifest in his “Adventure of the Dancing Men.” The dancing men—little stick figures with their arms and legs in various positions—constitute the cipher symbols. An American gangster, Abe Slaney, “the most dangerous crook in Chicago,” writes threatening notes in them to a former childhood sweetheart, Elsie, who has married an English squire. The squire copies the messages, which are chalked on window sills and tool houses, and brings them to Holmes. Holmes solves them, but the squire is killed by Slaney in an exchange of shots before Holmes can prevent the tragedy. Slaney escapes. Holmes, who knows where he is from the solved cryptograms, carefully composes a message out of cipher symbols that he has recovered and sends him a note urging him to Come here at once. (Holmes perhaps borrowed this scheme from Thomas Phelippes, who, Holmes knew, had in 1587 forged a cipher postscript to a letter of Mary, Queen of Scots, to learn the names of the intended murderers in the Babington plot against Elizabeth.) Slaney, naively believing that only Elsie and others of his Chicago gang at the Joint could read the cipher and that the note must therefore have come from her, returns to the squire’s home. He is at once arrested and, naturally, confesses.
A message in the Dancing Men cipher, solved by Sherlock Holmes
Holmes is, as he himself says, “fairly familiar with all forms of secret writings, and am myself the author of a trifling monograph upon the subject, in which I analyse one hundred and sixty separate ciphers, but I confess that this is entirely new to me.” He referred, of course, to the use of the dancers “to give the idea that they are the mere random sketches of children,” and not to their nature as a monalphabetic substitution. That he promptly recognized that they belonged to this class of ciphers is proved by his embarking at once upon a solution without any false starts. His task was considerably more difficult than that of any other fictional cryptanalyst, because his text was exceedingly short, disconnected, and elliptical and loaded with proper names. It eventually consisted of five messages in telegraphic English: (1) Am here Abe Slaney, (2) At Elriges, (3) Come Elsie, (4) Never, (5) Elsie prepare to meet thy God. But to begin with Holmes had only the first message, on which he made his start, and he broke the cipher only with that message plus the next three. They total only 38 letters, eight of them occurring but once; out of the nine words, four are proper names, and of the other five none is among the ten most frequent words in English, which normally make up a quarter of English text.
The difficulty of such a solution demonstrates the power and flexibility of the great detective’s mind. Holmes would quite evidently
have preferred to solve the cryptogram with his usual rigorous deductions, which means by frequency analysis. He began that way. The first message contained 15 dancing men, of which four are in an ecstatic spread-eagle position and three have their left leg bent. Holmes at once marked down the four spread-eagle dancers as e. Now, neither letter frequencies nor any other statistical phenomena are reliable in small samples; it was quite possible that the three bent-left-leg dancers represented e, or that one of the single dancers did, or even that no e at all occurred in the first message. It is inconceivable that Holmes did not know this. Nevertheless, he fixed the symbol for e “with some confidence.” He was right, of course, but why? No doubt Holmes, having recognized that the figures holding flags marked the ends of words, noticed that two of the four spread-eagle dancers carried flags, and instantly connected this with the well-known fact that e is the most frequent terminal letter in English. His swift mind may also have observed the variety of the e dancers’ contacts. But all this flashed through his great brain just below the threshold of his consciousness—this perhaps helps explain the characteristic rapidity of his deductions—and consequently he did not articulate it in his explanation to Watson. Or perhaps he did not want to burden Watson with all those details.
He did realize, however, that neither frequency analysis nor anything else could go further in the first message, and so he awaited more text. Upon the arrival of the next three messages, he saw that frequency analysis would not serve with so short a text. Unable to progress with his beloved deductions, he deftly switched to induction. He performed brilliantly, guessing first that a five-letter word with e as the second and fourth letters and comprising a message in itself must be never, and then conjecturing that the name Elsie might occur in the messages and finding it. With these values he was fairly on his way, and with further arduous labor completed the solution.
Some cryptologists have affected to sneer at Holmes’s taking two hours to solve these cryptograms, covering “sheet after sheet of paper with figures and letters” as he did so. With so short and difficult a text, however, the time is not only understandable, but admirable. Moreover, the dancers caper in no recognizable pattern when placed in alphabetical order, and when they pose in a graduated order of choreography, no regularity appears in the letters. In other words, the cipher of the dancing man is purely arbitrary. Some members of the Sherlock Holmes fan club, the Baker Street Irregulars, which included Alexander Woollcott, Christopher Morley, and Franklin D. Roosevelt, have kept their gaslights burning late in attempts to discover a regular basis of construction. It is wasted energy. The fact that Holmes limited himself to already recovered letters in his “Come here at once” message to Slaney suggests that he did not discover any regularity which would have permitted him slightly more latitude in composing that message. And surely had there been such a key pattern, Holmes would have discovered it. The inventor of the cipher, Elsie’s father, Patrick, “the boss of the Joint,” may have gotten the idea for the dancing men from a cipher based on human figures in the semiofficial Manual of Signals by Albert Myers, the founder of the U.S. Army Signal Corps, or from the same unknown place as the inventor of a slightly later United States patent that uses manikins for cipher symbols, or from the ubiquitous Carbonari, whose call-sign is made by extending the arms horizontally in the form of a cross and the reply by pressing two fists one above the other on the breast. Holmes may well have known of these possible sources. But even if Patrick did borrow the idea from one of them, he has altered the arrangement so thoroughly that cryptanalysis is left as the only way of resolving the problem.
A final point remains to be cleared up in the case of the dancing men: the source of the cryptographic errors that appear in all printed accounts. In the very first publication of “The Adventure of the Dancing Men,” the cryptograms use the same dancer for the v in Never and the p’s in prepare, and use an identical dancer for the b in Abe and for the r in Never. The Baker Street Irregulars have expended a great deal of energy on this problem. It is in their attempts to find the “correct” version that they have falsely assumed a regularity in the cipher alphabet, constructing tables of arm and leg positions and extrapolating the ciphertext symbols for the eight letters (f, j, k, q, u, w, x, z) that do not occur in the messages. They have also sought to determine the cause of the errors. Their efforts, however, have served only to show why they are the disciples and Holmes the master. All of them engage in armchair thinking without investigating the facts. There has been a suggestion that the errors “are in the messages of the villain of the story and may be laid, if one so wishes, to the poor devil’s confusion and despair,” but no one has raised the equally likely possibility that the squire may have made the mistakes while copying the messages to bring them to Holmes. In fact, however, the errors are neither Slaney’s nor the squire’s, for the errors were not present when Holmes solved the cryptograms. If the same symbol had been used for v and p in the originals, Holmes would have produced the partial plaintext vrevare in the fifth message after guessing Never instead of the ?re?are that he shows, with the two p’s as unknowns. Similarly, if the r and b had been confounded in the original, he would have shown a partial solution ?re (for the correct Abe) after guessing Never, but in fact he shows a partial solution??e with the b still unknown. Holmes’ own account thus proves that the errors did not exist in the original messages—and it is fortunate that they did not, for they occur at junctures crucial to the analysis and, coupled with the other difficulties, might have rendered the cryptograms almost impossible to read, even for Holmes. The errors must therefore have been made by Dr. Watson in transmitting the canon to the world. Later publications have compounded Watson’s original errors, but these have passed through the hands of literary and journalistic types, notoriously frivolous and unreliable as to facts, and need not be considered.
After Holmes’ feat, all other solutions look pale. Westrell Keen, the Tracer of Lost Persons, an elderly, distinguished-looking, unfailingly courteous man, conceived by Robert W. Chambers, is given a cipher consisting of rectangles with crossing diagonals, some of the lines of which are crossed by ticks. Like every fictional cryptanalyst, Keen must set forth superexcellent credentials and then marvel at the surprising uniqueness of the present system. Keen’s statement is classic:
“It’s the strangest cipher I ever encountered,” said Mr. Keen—“the strangest I ever heard of. I have seen hundreds of ciphers—hundreds—secret codes of the State Department, secret military codes, elaborate Oriental ciphers, symbols used in commercial transactions, symbols used by criminals and every species of malefactor. And every one of them can be solved with time and patience and a little knowledge of the subject. But this”—he sat looking at it with eyes half closed—“this is too simple.”
Notwithstanding, he solves it in short order, finding that the ticks mark the lines that in each rectangle form a crude drawing of a number. These numbers serve as the cipher message in a system where 1 = a, 2 = b, and so on. The message, directed at handsome Captain Kenneth Harren of the Philippine Scouts, reads, I never saw you but once. I love you. Edith Inwood. Keen discovers from his voluminous files that Miss Inwood, 24, a 1902 graduate of Barnard, is an assistant to Professor Boggs of the American Museum of Inscriptions and an authority on Arabian cryptograms. Keen unites the lovers.
H. Rider Haggard, author of She and King Solomon’s Mines, had the eponymous hero of his Colonel Quaritch, Q. V., poor, plain, and middle-aged, solve a cipher that enabled him to find a buried treasure and marry the young woman who loved him. This saved her “from a fate sore as death,” namely, marrying a young, rich, and handsome man whom—alas!—she did not love. The solution of the cipher, a null system, must rank among the strangest in fiction: “Now, as the match burnt up, connected probably with the darkness and the sudden striking of light upon his eyeballs, it came to pass that Harold [Quaritch], happening to glance thereon, was only able to read four letters of this first line of writing of the cryptogram. A
ll the rest seemed to him but as a blur connecting those four letters. They were: D. … e …. a …. d, being respectively the initial letters of the first, the sixth, the eleventh, and the sixteenth words of the line given above.” It is a cryptanalysis by optical illusion. Interestingly, a niece and a nephew of Haggard later served in England’s Room 40. Perhaps if Haggard had written the book after their experience there, his solution might have been a little less farfetched.
Lloyd C. Douglas inserted a journal kept in a railfence cipher in Magnificent Obsession. It began:
O. Henry penned a sardonically amusing story about “Calloway’s Code.” Calloway, a newspaper correspondent, gets a scoop past the censor’s eye by using the first half of a newspaper cliché as the codeword for the second half, which forms the plaintext. Thus, FOREGONE meant conclusion; DARK, horse; BRUTE, force; BEGGARS, description. And—sad to say—the journalists in New York understand it. Robert Graves gave the ancient Romans two ciphers, an ordinary (the Caesar) and a special (a polyalphabetic), in I, Claudius. India’s great Rabindranath Tagore recorded the directions to a secret treasure in gūdhalekhya, a classical Indian system of cryptography, in his story “Gupta-dhana” (“The Hidden Treasure”).
But most cryptograms appear in mystery stories. Bram Stoker, the author of Dracula, based The Mystery of the Sea upon an enormously complicated concealment system; the book, like Byrne’s Silent Years, may have been written partly to show off the method. The solution leads to a fabulous treasure. Agatha Christie, one of the most artistic of mystery writers, employed an open code using flower names in “The Four Suspects,” solved, of course, by her prim Aunt Jane Marple. E. C. Bentley, author of Trent’s Last Case, has his Philip Trent likewise tangle with an open-code flower system in “The Ministering Angel.” Lillian de la Torre has her Dr. Samuel Johnson solve a skytale cipher that used a peg leg as the skytale in “The Stolen Christmas Box.” In Montague Rhodes James’ ghost story “The Treasure of Abbot Thomas,” a solution leads to a treasure. And there have been scores of others—sometimes in short stories, sometimes as decoration for the major plot in books. Even Fu Manchu thrust The Hand of Fu Manchu into crypto-logy. The cryptograms have been of all kinds, based on everything from the Dewey Decimal classification system used in libraries to the roulette wheel.
