What makes this all the stranger is that while a general line of development can be traced through the Middle Ages with the suggestion that the Roman method of ciphering was first lost, then rediscovered, bit by bit, the line has such enormous gaps that it may not really exist at all. In one of the earliest manuals on cryptography in existence (di Lavinde’s from the fifteenth century) the use of an enciphered code is recommended: an extremely modern and complex development, which has not been surpassed today. It argues decades and perhaps centuries of effort to defeat decipherers who have become extremely acute at their business. We have not the slightest clue as to how and where that skill was acquired. It may have been at Rome; it may have been at Constantinople; or it may have been at Venice and Genoa. If it was a Venetian-Genoese discovery there is a possibility, rather faint, that ciphers were originally developed in that Near East from which the two great commercial cities drew so much else.
In short, when we begin to investigate the history of ciphers, we are digging in a graveyard whose limits we do not know and where there are headstones only for the failures. Decipherments that have changed the course of history (and they are not a few) are often recorded. The triumphs of encipherment, of messages that got through without being read by interceptors are never mentioned—if for no other reason, because people who have used a cipher successfully wish to keep it secret and use it again.
It is probably for this reason as much as because cryptographers are naturally proud of their own performance that the statement is often made that there is no such thing as an insoluble cipher. Strictly speaking it is not true. Roger Bacon in the early Middle Ages wrote a whole manuscript in a cipher that has thus far defied analysis. It is extremely probable that an insoluble cipher could be produced by mathematical means today.
This is true, however, only if the production of an insoluble cipher and the recording of some relatively short message in it were the only end in view. All ciphers in actual use break down on repetitions, not merely repetitions which can be avoided by careful phrasing in a single message but the necessity of repeating the same words or sentences in several messages. The redundancies of action defeat the best efforts of those who would send secret communications.
Nor is the phrasing always careful, even in a single message. Wherever ciphers are most frequently used, they must be written in a hurry, usually by men without much special training, and always without special apparatus. The effort to break them down, to read the messages they contain, will always be made by experts with ample training, a wealth of time at their disposal and whatever special apparatus they need. No systematic method of defeating this analysis has ever been found or is likely to be.
Moreover, another factor enters here. The only method of delaying expert analysis is by complicating the enciphering process; and complication is fatal. An officer of the British Black Chamber estimates that one-third of all the cipher messages which passed through that department during the World War were garbled; that is, mistakes had been made in the enciphering process. The more complex and safer the cipher, the greater the likelihood of these errors; and in some of the better ciphers they are progressive, so that a single error renders all the rest of the message gibberish, even to the man with the key.
The utmost any modern cipher can hope to accomplish is to force the decipherer to employ his last resources, particularly his resources of time; to delay decipherment until the information obtained by the process is no longer of value. That the information obtained by decipherments will always be of some value is the reason why navies use codes.
But this is already trespassing on matters that should be reserved for the text of this volume.
II
A few definitions of special terms will help to make that text clearer.
A cipher is a method of writing a message so that it cannot be read by anyone ignorant of the method.
A cryptogram is a message written in cipher.
The clear is the communication which it is desired to make.
The message is that communication after it has been written in cipher.
A substitution cipher is one in which letters of the clear are replaced by letters, figures or symbols.
A simple-substitution cipher is one in which one letter of the clear is represented by one, and always the same, letter, figure or symbol of the cipher. Example: if the clear be “Come here” and each letter be represented by the one following it in the alphabet, the resulting message will read DPNF IFSF.
A simple-substitution cipher with suppression of frequencies is one in which each of the very common letters (such as E) is represented by several figures or symbols. Example: the same clear as above, “Come here,” with each letter still represented by the one following it in the alphabet. But it has been concerted that in addition to F, the figures 2 and 3 will also represent E. The message now reads DPNF I2S3.
A double-substitution cipher is one in which letters of the clear are represented in the cipher by letters which vary according to a system, the basis of which is a key-word. To be more fully explained and illustrated in the text. See Chapter Six.
A two-step cipher, not at all the same thing as the last, is one in which the message, usually obtained by enciphering by simple or double substitution, is now enciphered for a second time. The second substitution is usually made according to a table, of which both sender and receiver have copies. This table may (for instance) make the second substitution on the basis of two or more letters at a time. Example: the same “Come here,” enciphered as DPNF IFSF. The cipherer refers to his table; suppose it indicates as the value for DP, 416; 317 as that for NF; 96 for IF and 138 for SF. The message would then be 416 317 96 138. In two-step ciphers one of the steps is usually substitution, the other transposition.
A transposition cipher is one in which the letters remain the same as in the clear, but are shuffled according to a prearranged pattern. Example: the same clear as above, written in two lines:
which is read off line by line, the message sent being CMHR OEEE.
A combination cipher is a two-step cipher in which the steps are transposition and substitution. Example: “Come here,” enciphered by substitution to DPNF IFSF, which is now written:
and the message taken off as DNIS PFFF.
A grill cipher is one in which a grill or mask with holes in it is placed over the paper on which the message is to be written. The message, in clear, is then written through the holes; the grill removed and the spaces between the words or letters filled up with others to give the whole the appearance of an innocent letter containing no cipher message.
A syllable cipher is one in which substitution or transposition is made on the basis of syllable or pairs or triplets of letters instead of single letters. This is very rare.
Nulls are letters or words having no connection with the clear, introduced to confuse a decipherer.
Stops are punctuation marks, usually sentence endings, for which special characters are provided, sometimes placed after each word.
To encipher a message is to put it in cipher. To decipher it is to reduce it to clear. To break a cipher or code is to discover the system by which it was composed. It is perfectly possible for a cryptographer to read the content of a ciphered message without being able to discover the system on which it was written. Another message in the same cipher forces him to repeat the work of decipherment.
Frequency