indicated, the following result is obtained:
It is apparent how nearly this finishes the task. Obviously nothing will do at the end of group 11 but the letters L and D, to complete the word should, which gives the equations:
F = L; M = D
Similarly replacing the G of the message in group 6 with M yields a satisfactory result, and the U’s in groups 4 and 14 work out nicely as V’s. LON-blank in groups 13–14 now becomes clear as LONG, and H = B is required in group 17. The remainder can now be filled in:
V = W; X = Y; L = P; I = C.
The message is solved and the cryptographer now draws up his table of equivalents:
The key-word was evidently “chimpanzee” with the final E dropped off (repetitions are not permissible) and it is now possible to fill in the whole table and wait for the appearance of the next message written with the same key.
V
This is the basic method in all decipherments of substitution ciphers. Admittedly the example shows the process at its shortest and simplest. In normal English text the alphabetic frequency tables are unreliable until two hundred or more letters have been reached in one message or two or three written with the same key. Still with the backing of the bigram and trigram tables any cryptographer can dismember any simple substitution cipher in a few minutes and with a minimum of trials. The fact was evidently widely known when Sicco Simonetta wrote that first book on ciphers, for he included alphabetic frequency tables in it.
At the same time the earliest codes were also being found wanting. In the sense that they are mentioned earlier, they antedate ciphers; and, like ciphers, appear to have grown out of the then new custom of keeping at foreign courts resident ambassadors who found it necessary to send reports home and ask for instructions. Venice and the Papal Curia were the first powers to use resident ambassadors; and, though the latter may well have used means of secret communication before the republic of the lagoons, the oldest reference to any code system is in an instruction to the Venetian ambassador at the Court of Austria. In his dispatches it is ordered to refer to the Doge as “V,” the King of Hungary as “P” and the Pope as “Q.”
The context alone would apparently betray the secrets of such messages if they were intercepted, an observation with which the Venetians evidently agreed; for only twelve years after this first crude code, another Venetian instruction orders the city’s diplomats to refer to important personages by periphrasis—that is, to speak of Austria as the “Sun,” since the sun rises in the east; the east = Ost = Österreich; and to replace all the verbs in their dispatches with meaningless words according to a regular system outlined in a little code dictionary they were given when setting out on a mission.
Then comes another of those gaps in cipher history, followed by the appearance of the Trattati in Cifra of G. di Lavinde in 1480. It is quite a remarkable book, showing cryptography already in a state of considerable development, for he recommends a method of decipherment by attacking the vowels, which is still classic for the Latin languages, and a method of defeating this decipherment, by first throwing messages into a kind of jargon-code, and then enciphering them by simple substitution with suppression of frequencies. Perhaps more remarkable still, di Lavinde’s official position was that of special secretary for secret communications before the Pope. Ciphers had already become so important a business at the Curia as to require the attention of a full-time expert.
Within the next hundred years every major court and minor principality of Italy, Spain and France was using ciphers, and all the great systems of encipherment but one had been invented. Decipherment does not seem quite to have kept pace; indeed it was nearly four hundred years before the classic method of deciphering double substitutions was discovered. Now in cryptography as in other fields of human activity necessity is so much the mother of invention that it is difficult to believe that if the complex systems of cipher had been widely used, means of breaking them down would not have become as widely known as the ciphers themselves. As a matter of fact, practically all the ciphers of which samples have been preserved to us from periods before the Napoleonic era belong to relatively simple types. It is difficult to avoid the conclusion that the more elaborate systems, though they were developed early, remained buried for some time in the theoretical manuals of the men who invented them.
This theory is considerably strengthened by the fact that we know northern Europe, where the greatest development in deciphering processes subsequently took place, did not take readily to ciphers in the early days of resident ambassadors. There were two or three striking historical incidents which painted the unreliability of the existing systems of cipher across the continent (they will be described later); and the North had long since developed its own device for private communication. Under the pressure of the rising art of diplomacy this device began to develop into a regular system. Its basis was a code which is not a code—jargon, cant, euphuism, the erection into a regular method of expression of that system which had formed part of the means of communication used by the Venetian agents in the fourteenth century.
VI
Partly this method is inherent in the Teutonic languages themselves, which drop the precision of the Latin tongues for a tendency toward double meanings and involved images. One of the earliest written works in any northern language, the Prose Edda, is a handy manual of involved periphrastic metaphors for the use of poets.
“What shall we call the air?” says one passage. “The air may be called the ravens’ causeway; or the bearer of storms; or the woof of the winds.”
If figures of speech are constantly used, if they be sufficiently farfetched and involved, or if they have reference to something known only to a few persons, this sort of thing can be developed into a code which can be either written or spoken. It is difficult to state positive facts when dealing with so indefinite a subject, but it would seem that the inspiration for the use of this type of code for diplomatic purposes comes ultimately from the very root of the language growing through medieval thieves’ slang, which was already highly developed in the fourteenth century.
“These Babes of Grace,” says an early Elizabethan instruction for the training of young thieves, “should be taught by a master well verst in the cant language or slang patter, in which they should by all means excel.”
A good deal could be and has been written about these peculiar crooks’ codes, about which it is enough to remark here that they seem to have reached their greatest development in the early part of the nineteenth century. This was the period when a pair of initiates could carry on a whole conversation without an eavesdropper catching any more of it than the prepositions. Both Victor Hugo (Les Misérables) and Eugène Sue (Les Mystères de Paris) give some interesting examples, most of them already out of date when they were distilled into the books; and the famous Vidocq, who was on both sides of the law at different times, quotes a long song by means of which young members of the profession were taught its special language, not very good as a song but intended more as a school exercise:
J’ai fait par comblance
Gironde languepé,
Soiffant picton sans lance,
Pivois non maquillé,
Tirants, passe à la
