by DAVID KAHN
Cryptanalytic techniques employed by the F.B.I. Laboratory to break these “bookie” ciphers are conventional in the sense that they rely heavily upon letter and digital frequency characteristics of the various types of bookmaking entries, and trial and error testing of probable betting data. F.B.I. experts have examined and broken literally thousands of encrypted betting entries during the past dozen years [1950 to 1961] and have developed a highly skilled acuity for penetrating this type of material. This skill is a combination of pure cryptanalytics and a very comprehensive knowledge of betting procedures and operations.
Indeed. To understand the plaintext, much less the cryptanalysis, would require an extended course of tuition at Churchill Downs, Pimlico, and Saratoga, with perhaps a week at Las Vegas for postgraduate work. The entries are highly abbreviated, so that a few code symbols can represent what would take a few dozen words to spell out. The bookmakers also rely heavily on their memory for many details, including especially the names of bettors, which are almost never entered. Nevertheless, the F.B.I. cryptanalysts nearly always master the systems. In one case, Hebrew script served to encipher the numbers, Yiddish phonetics the names of the horses, and the arrangements of numbers combined with symbols like parentheses the types of bets. After an F.B.I. cryptanalyst took the stand at the bookie’s trial in Lancaster, Pennsylvania, the bookie changed his plea from innocent to guilty. He was fined and sent to jail.
When Boston police raided a betting parlor, they confiscated racing forms, racing magazines, sports editions of newspapers, and a pocket notebook containing handwritten symbols like Greek letters. When questioned about this, the suspect told police that he was a student of the classics; he conceded that he was trying also to work out a system for beating the horses, but denied that he was a bookie. Privately, however, he boasted to friends that the cops would never break his code and that until they did he was going to continue to book horses. Boston police sent the notebook to the F.B.I. , who soon established, for example, that α = 1, ø = 2, σ = 9, β = 11, δ = 12, both H and O = 50,) = parlay, and so on. The entries proved so highly abbreviated that even after the plaintext meanings were established considerable interpretation was necessary. With the testimony of an F.B.I. agent, the Boston prosecutor was able to prove that the defendant had maintained “apparatus for the registering of bets on the result of a contest of speed or endurance of a certain beast, to wit, a horse,” and to win a conviction.
Another Boston case depended wholly upon the solution of 200 pages of coded information found in three notebooks in a raided clothing store suspected of being a bookmaking front. The F.B.I. solved it relatively quickly, and found that the basic encipherment involved transforming letters into numbers with the key
Other cipher symbols included BRD for $1 on a daily double, ERD for $2 on a daily double, NYD for 50¢ on a daily double, BLT for $1 across the board, NLT for $5 across the board, a check to indicate a payoff, and so on. Thus the cipher √IG BZBZ GB meant a bet on the No. 4 horse in the third race of $10 to win and $10 to place; the horse paid $6.20 on a $2 bet to place for a return of $31 on the $10. After a three-hour trial, during which an F.B.I. cryptanalyst testified to the solution, the clothing store operator was convicted of bookmaking.
Less common than bookie and policy systems are the occasional ciphers used to plan jail breaks, to note proposed robberies or burglaries, to record illicit activities. Nearly always these are of an extremely simple nature, usually a monoalphabetic substitution using symbols. When New York City police broke up a ring in 1959 that produced and sold pornographic films, they found enciphered records that listed the names and addresses of 300 actresses and what they would perform for the cameras. Police easily broke the cipher, a rather simple one consisting largely of abbreviations and symbol substitutions, and began making arrests wholesale. Prisoners use the ubiquitous knock cipher, based on the checkerboard, to transmit information from one cell to another. Thus one knock followed by two means b, two knocks followed by one indicates ƒ. The system serves in jails all over the world. Less specifically criminal but still outside the law was the “code” used by New York City employment agencies to designate the race or color preferences of employers. One agency used NO NFU’s—the latter originally standing for “Not for us”—to mean No Negroes. Another wrote RECOMMENDED BY REDBOOK to mean that Negroes are not wanted and MUST PLAY SAXOPHONE to mean that Jews should not bother to apply. Such designations are prohibited by the New York State law on fair employment practices, and in 1962 at least one employment agency signed a consent order in which it promised not to use them again. In 1960, the United States accused five big electrical concerns—including General Electric Company and Westinghouse Electric Corporation—of violating the antitrust laws. The firms’ executives replaced the names of the firms with numbers as a simple code to help conceal their agreeing to fix prices and to rig bids so that profits would be assured. Thus 1 stood for General Electric, 2 for Westinghouse, 3 for Allis-Chalmers Manufacturing Company, and so on. The system was transparent, however, and the executives got 30 days in jail.
Card sharks mark their cards, which constitutes a kind of cryptography, and they communicate clandestinely with one another by various body or voice signals. Particular ways of holding a cigarette or scratching an itch can indicate various suits or cards. In one common hand-signal, a gambler quickly, almost casually, places his hand against his chest, thumb spread, to mean I’m going to take this game. Anybody want to partner with me? A right hand, palm down, on the table, means yes; a fist on the chest or table means, No, I’m working single, and I discovered these guys first, so scram. In one of the English whist clubs of Victorian days, a player would tell his partner which suit to lead by a casual comment, the first letter of which was the same as the first letter of the suit. Thus “HAVE YOU SEEN OLD JONES IN THE PAST FORTNIGHT?” would mean that hearts was to be led. Such systems are limited only to the ingenuity of the gamblers. The position of cards on the table can represent ace, king, queen, or jack. Knees can be touched under the table. Accusations of cheating by such means at the world bridge championships in Buenos Aires in May, 1965, were hurled at the British team of Boris Shapiro and Terence Reese. But the subtlety of the signals—a way of puffing a cigar, of scratching an ear, of breathing, of anything that people might ordinarily do—made it extremely difficult to prove the accusation true or false.
Jargon, because of its allusive nature and easy comprehension, seems to be widely used in illegal activities that involve considerable contact with the public, such as vice. In 1961, police in Graz, Austria, noticed that curio dealer Alexander Kotzbeck got more customers than a nearby milk store. Checking, they discovered that he was telephoning his customers to report that the BAROQUE ANGEL had arrived, or that the ROCOCO STATUE could be picked up immediately. The “angels” and the “statues” turned out to be living dolls—call girls aged between 18 and 24. The police, having broken the code, broke up the vice ring.
Spoken jargon of this kind borders upon argot, the language of thieves. Argot is just one of many specialized vocabularies used in various social groups, from children to sailors to printers. Because such social factors as exclusiveness, mutual experiences, the need to discuss common technical operations, and delight in word games engender such specialized “languages,” they arise all over the globe. Tinkers in Ireland use one called Shelta. Cockneys speak rhyming slang (STORM AND STRIFE for wife). Medical students at London University occasionally used Medical Greek, a transformation of pairs of English words, so that smoke a pipe would become POKE A SMIPE. Primitive tribes like the Langos and the Todas have secret languages. The Chinese use them. Children are prolific in inventing them; each natural language seems to have at least one particular children’s cipher. Well known in English is Pig Latin: the speaker beheads the initial consonant sound of the word, attaches this sound to the end of the word, and adds the syllable AY. Thus third becomes IRDTHAY. For words beginning with a vowel, the speaker simply adds WAY to th
e end—and becomes ANDWAY. Tut Latin, or the King Tut language, interpolates a TUT between syllables.
Argot differs from these languages in being, perhaps, more developed, more extended; in addition, while, like the others, it includes many necessary technical terms and serves as a sign that the speaker belongs to an in-group, it incorporates a much darker strain of secrecy. Argot is not an international thieves’ language, though some terms may be international. It varies from country to country because it is essentially a modification of the national language. When speakers of argot desire secrecy, they can transmute either the meaning of a word or its form. Metaphor plays a large role in the former. A lawyer is a MOUTHPIECE; money is BREAD or DOUGH (this metaphor occurs in French as well); the electric chair is the HOT SEAT; solitary confinement is the COOLER. Substitutions of form include abbreviation (ALKY for alcohol, CON MAN for confidence man), and systems like Pig Latin. These are not widely used in English argot but are relatively common in French. One French system is largonji, so-called from its deforming the word jargon according to its own system: moving the initial consonant to the back of the word, adding an I, and putting an L on the front of the word. French argot devised many ingenious systems, most of them short-lived; several replaced a word’s normal ending with a peculiar-sounding one. Thieves in Peshawar, India, insert a syllable with a z in it into words, so that piu (“father”) becomes PIZEO and usko bula (“call him”) becomes UZUSKUZO BUZULEZA.
In the Occident, argot apparently began in France in the 1100s, in the warren of thieves’ dens and culs-de-sac on the site of the present Place Maubert in Paris that was the most famous Court of Miracles, inhabited by highly stratified and closed bands. In England it originated in the 16th century, in America in the 18th. Argot was for hundreds of years a secret language, as much owing to the isolation of its speakers as to any inherent cryptography. Some poems written in argot by the 15th-century poet-thief François Villon remain partly incomprehensible even today. Argot lost most of its secret character when, in the early 19th century, the demolition of the old quarters, the creation of municipal police, the breakdown of social barriers, destroyed the old criminal bands. They melted into the population at large and their language filtered into the common speech. Most terms disappeared; some of the more colorful became slang. Though today criminals still use argot, its secret character has faded—the police understand nearly all of it; scholars write about it—and it consists mainly of technical and professional terms. Crime, like everything else, has become a business.
And business, too, uses secret languages. The need to conceal financial information from the unscrupulous led the early church to encipher some of its financial transactions by using Greek letters to represent numbers. The Knights Templars, the semimilitary, semireligious order of the Middle Ages, enciphered letters of credit that Templars carried instead of cash from one of their 9,000 commanderies in Europe to another. The cipher alphabet, like their chapels, was based on the Maltese cross. In the 19th century, the Ottoman Turks in Egypt used qirmeh, an abbreviated sort of writing, to record their tax and financial transactions.
Free enterprise entails secrecy almost inevitably. Entrepreneurs must keep secrets not only from competitors, but often from consumers. Thus, cloth merchants in Peshawar insert MIRI as a kind of null into words to create their secret languages, much as Pig Latin adds -AY at the end of words. Goldsmiths in Kashmir drastically alter the form of words. Where the normal forms of “one,” “two,” “three” and “Will you sell?” in their Zergari dialect are akara, sanni, trewai, and choande the goldsmiths convert them into BIN, HANDISH, YANDIR, and PHETZU WAHNO. Other trades in India have their own dialects—probably reinforced by the caste barriers. Butchers in Hanoi move the initial consonant of a word to the end, replace it with a CH sound, and add an -IM onto the end of the word. Fish dealers at the Fulton Fish Market in New York use impromptu codes to conceal their prices. Professional magicians use codes with their assistants in “mind-reading acts”; the assistant, passing among the spectators, keeps up a patter that tells the magician what objects the spectators hold, or the serial number of a dollar bill, or their names. Among written secret languages within various trades, perhaps the oldest, most widespread, and best known is that of the hoboes, who chalked various marks on the walls or doorposts of houses to inform one another that the owner is a soft touch, or has a vicious dog, or calls the police upon being approached, or makes one work for the handout. But Social Security is gradually eroding this particular cryptography.
Often, retail merchants will encipher wholesale prices on price tags so that they will know how much discount they can afford to give while keeping the customer ignorant of the original price. Generally they will replace the numbers with letters according to a keyword, rather like the bookmakers. Dealers in antiques and other flexibly priced items probably use these systems more than anyone else, though in the 1890s the impresario George Broadhurst, then manager of a theater in Minneapolis, encoded the night’s receipts by means of the word REPUBLICAN before telegraphing them to the Baltimore headquarters of the theater chain. Such systems have been solved—sometimes with valuable commercial results. In one case Macy’s, the giant New York department store, was bound to observe manufacturers’ minimum prices until it could prove that others did not. It cracked the price-tag code of Masters, Inc., a large New York discount house, and used the information for business tactics that improved its own competitive position.
From these simple beginnings commercial secrecy evolves into levels of much greater complexity. The Swedish match king, Ivar Kreuger, employed cryptography to mask the hollowness of the enormous financial empire he built up. For his own use, he carried about with him a tableau of 26 mixed alphabets graven on a small ivory plaque. In his office he used cipher machines, which he employed often enough to have a sign printed “Ciphering in progress here” to deny access to that office. Rumor said that the giant J. P. Morgan Company banking house had tried to break his messages. Perhaps because of this, Kreuger hired Yves Gyldén to teach two of his employees cryptography. The course of instruction ended suddenly when Kreuger committed suicide and his business honeycomb collapsed.
Business espionage, which in some respects is almost as elaborate as governmental espionage, with undercover agents, long-range cameras, hidden microphones, and bribes to get the contents of firms’ wastebaskets, has rarely gone as far as cryptanalysis, however. In addition to the Macy’s and alleged Morgan cases, the only other case known is that of a firm in Hong Kong that obtained the messages of a rival firm from an employee of a cable office. These had been encoded with a commercial code that was sold publicly, and the intercepting firm had no trouble reading the messages—and then submitting bids of its own that were half a cent lower than those of its rival, thus stealing considerable business. When the other firm learned about this, it began enciphering its code messages. Evidently this proved too much for the intercepting firm, for it no longer won bid after bid.
Nevertheless, to prevent this sort of thing from even getting started, many firms encipher their messages or—more frequently—encode them with their own private code. Usually they are not superenciphered, secrecy being obtained by small printings, restricted distribution, and careful supervision of the codebooks. Though these would not guarantee protection against the attacks of government cryptanalysis, they adequately guard the firm’s message from the casual inspection of cable clerks or from cryptanalysis by their rivals. Firms in such highly competitive businesses as oil and mining, where information about a possible rich field can be worth hundreds of millions of dollars, are the greatest users of such private codes.
In the 1920s, with the sudden expansion of international trade following World War I, a number of individuals saw opportunities in what they thought was the need for secrecy in the booming capitalistic economy. Inventors directed promotional efforts for the first time more at business than at armies or diplomats. The understandable desire to get rich motivated the
efforts of such men as Damm, Hebern, and Scherbius, the inventors of the rotor; the A. T. & T. Company promoted Vernam’s machine in the hope of profit. The inventions of these men enriched cryptology if not themselves, but the efforts of many others contributed nothing either to cryptology or to their own pockets. The best known was Alexander von Kryha, a handsome engineer of Ukrainian extraction who energetically promoted his machine. It consisted of a simple cipher disk, attractively housed, with gears controlling the number of spaces it turned and a spring mechanism driving it. The encipherer found the plaintext letter in an outer alphabet and took the letter inside it as its ciphertext; he then depressed a button, allowing the disk to turn an irregular distance and stop, presenting a new set of cipher equivalents. Von Kryha got a German professor of mathematics, Georg Hamel, to calculate the number of different permutations of alphabets that were possible with the movable letters of the cipher alphabet, multiplying this immense figure by the number of possible gear combinations, and then by all the other variables to “prove” that only immortals could break the cipher. Unfortunately, the mechanism came down to a simple polyalphabetic cipher with a single cipher alphabet and a period of a few hundred letters—solvable within hours, not millennia. But the company probably failed less because of its product’s technical weaknesses than because of the same lack of interest that firms like Damm’s and Scherbius’ encountered in the business world at large.
The effort to sell secrecy to the business world runs like an irregular thread through the fabric of cryptologic history. A number of firms manufacture cipher machines today and offer them to the commercial world. Ottica Meccanica Italiana, a Roman company, produces a rotor machine. Standard Elektrik Lorenz of Stuttgart makes the Mi-544, a heavy, solid, one-time tape machine. Hagelin sells his machines. Gretag, Limited, of Regensdorf-Zurich, Switzerland, produces two machines. The KFF-58 is an electromechanical device using sprocket wheels as the keying mechanism. The TC-534 is a solid-state digital device that generates a pseudo-random key for use in a Vernam-like device. And other companies manufacture cipher machines upon occasion. But the sales effort never really succeeds because the commercial market is too small. Mining and oil companies may buy a few, but almost no one else. At the time of the coronation of England’s Queen Elizabeth II, when the National Broadcasting Company and the Columbia Broadcasting System were competing hotly to get their films on television first, N.B.C. encoded its transatlantic messages to keep its plans away from C.B.S. But even here, with millions of dollars at stake, N.B.C. probably did not use machines. The commercial cipher machine market is minuscule.