Read an Excerpt

THE READER OF GENTLEMEN'S MAIL

Yale University Press

Chapter One

Codes are a way of making messages secret. But not all of what are called codes are codes. Some are ciphers. There's a difference. Codes are books. They are like foreign-language dictionaries, except instead of translating English into, say, French, they translate English into code. A small part of a codebook may look like this portion of one from World War I:

stop 3514 stopped 3329..4017 storm 4211 strength 1740..2329 strength of enemy unknown 3961 strengthen 1679 stretcher bearers 3166

This means that, in turning the original English message into code, the word stop will be replaced by 3514 and the word stopped by either 3329 or 4017. These numbers are what will be transmitted to the receiving station.

The receiver has a similar book, only with the numbers in numerical order:

SHORT COURSE IN CODES AND CIPHERS

1674 favorably 1675 make ready 1676 no patrols 1679 strengthen 1681 -nt 1684 49 1685 question mark

The gaps in the numbering make it harder for the enemy cryptanalyst to reconstruct the code. This is called a two-part code because it has one part for encoding, one for decoding. The code elements can be letters instead of numbers. It is harder to solve than a one-part code, in which the codenumbers or codewords run parallel to the plaintext elements.

While codes work mainly by words, ciphers work by letters. They use not books but tables or mechanisms to convert their messages, letter by letter, into secret form. The simplest cipher-one found in the Bible-replaces the individual letters of the plaintext message with other letters:

plain a b c d e f g h i j k l m n o p q r s t u v w x y z cipher L B Q A C S R D T O F V M H W I J X G K Y U N Z E P

The message attack at dawn would be enciphered into LKKLQFLKALNH.

The weaknesses of this monoalphabetic system are evident. The ciphertext reflects the letter pattern of the underlying plaintext. And since each plaintext letter is represented by a single ciphertext letter, the system is vulnerable to frequency analysis. This technique is based on the fact that-given several different pieces of writing of about two hundred letters or more in a single language-the percentage of each letter in each piece stays about the same. Thus, in English, the number of e's-the most common letter-will average around 12.5 percent. The next most common, t, will stand around 9 percent, and on down to z, at around 0.5 percent. These frequencies hold whether the text is a military telegram, the Gettysburg Address, or "To be or not to be." So a codebreaker, faced with a cryptogram enciphered in this system, can count the letters in the cryptogram and can assume that the most frequent cryptogram letter stands for the most frequent letter of the language of the plaintext. He can then insert this guess into the cryptogram and try to fill in the missing letters. Sometimes the guess is wrong, so the cryptanalyst must try another possible plaintext letter. Cryptanalysis involves much trial and error; it is not as straightforward as proving a theorem in plane geometry.

Because frequency cryptanalysis has been known in the West since the Renaissance, cryptographers have proposed systems that would defeat it. One uses several alphabets, so that a single plaintext letter would have diverent ciphertext representations, according to the alphabet used. Such systems are called polyalphabetic; the best known, published by a French diplomat in 1586 and named for him, is called the Vigenere. A mnemonic, usually called the keyword, species which alphabets are to be used and in which order. Both the encipherer and the decipherer must hold the keyword. For example, if the keyword is COMET, the alphabet beginning with the letter C will encipher the first plaintext letter, the alphabet beginning with O the second, and so on, repeating the keyword until the entire message has been enciphered. To avoid the weakness caused by the regular repetition of the keyword, cryptographers use a long text, such as a poem or novel, as a running key. Some cipher machines generate extremely long incoherent keys for polyalphabetic cryptosystems. Other machines create many different cipher alphabets. Solving repeating-key polyalphabetics requires the cryptanalyst to first determine the number of alphabets used, to separate out the letters enciphered in each alphabet, and then to apply frequency analysis to each such set. Other polyalphabetics demand more complicated solutions.

While these partake essentially of mathematical and statistical analyses, solutions of code messages more resemble the reconstruction of lost languages. It is not surprising that the ace French cryptanalyst of the 1890s, Commandant Etienne Bazeries, was called the Champollion of codebreakers. Code solutions require many more messages than cipher solutions, and their cryptanalysts often begin by determining, from their frequent appearance and their positions in the cryptogram, the codewords that represent period or full stop. This outlines the structure of the plaintext. Then, events in the world, such as the sailing of a warship from various ports during a cruise, are related to intercepted messages, guesses are made and confirmed and other guesses added to them. Gradually the code is built up.

Codes and ciphers form one of the kingdoms of cryptography, the substitution kingdom. Letters or words are replaced by other letters or numbers. The other kingdom consists of transposition ciphers. In these, the letters of the plaintext message are scrambled: enemy might become NEEYM. The cryptograms of transposition ciphers retain the original letters and thus may be seen as weaker than substitution systems, where the plaintext has to be recovered. Transposition cryptograms are more difficult to decipher correctly than substitution systems if a letter is dropped in transmission. So they are used less often than substitution systems. Substitution systems may operate in a continuous stream; transposition systems must run in batches.

Cryptographers apply substitution or transposition systems to codes to double the work of those trying to break a code. These systems conceal the codewords or codenumbers. For example, the letters of the codeword PEDED may be replaced, either individually or in pairs, as agreed on, with other letters or numbers. Thus PEDED might become, in one system, RBIBI. In a different system, the PE might become RL, the DE might become MI, and the D, X, so that PEDED becomes RLMIX. This is enciphered code, or superencipherment.

Cryptanalysis is a practical, not an abstract, technology. Many cryptosystems fail not because of their own flaws, but because they are poorly used. Commanders repeat themselves. Cipher clerks err. Generals begin messages with "To the colonel of the 14th Regiment," diplomats with "I have the honor to ..." Cipher clerks send messages in the wrong key that they have to redo in the right key, creating two cryptograms with the same plaintexts that cryptanalysts can work like a crossruff in bridge. Codes may list many different code equivalents for stop, but the clerks quickly remember one or two and use them instead of looking in the codebook for the others. Even when codes or ciphers change, the habits of the communicators do not. And often messages are not solved at all.

The terminology of cryptology has become more precise since Yardley's time. In those days, "cryptography" ambiguously meant both making and breaking codes and ciphers. Today cryptography means only making. Breaking is cryptanalysis, a term that came into general use only in the 1930s. Cryptography plus cryptanalysis now combine into cryptology. Likewise, under Yardley, decoding and deciphering stood both for the authorized turning of a cryptogram into plaintext by the legitimate receiver and for the unauthorized solution of a cryptogram by the enemy. Popular writing often prolongs this confusion. But in the modern taxonomy, "deciphering" means only the legitimate reconversion of a cipher message from its secret form into its plaintext; "decoding," of a code message. The term for solving a message is "cryptanalysis." I use the clearer modern terminology except in quotations and sometimes, when the meaning is plain, I use "codes" to stand for all cryptography, as in the first sentence of this glossary, and "codebreaking" for all cryptanalyses. The modern term encrypt usefully means either "encipher" or "encode"; likewise decrypt for the restoration to plaintext.

How Yardley Wrote His Best-Seller

In 1929, when Secretary of State Henry L. Stimson withdrew his department's funds from codebreaking on the ground that it was immoral, Yardley's agency, quartered in New York, was closed down. Unable to find a job in the Depression, Yardley decided to use his sole asset-his knowledge of his secret activities-to support his wife, their boy, and himself by telling that story. In a memorandum for literary agent George T. Bye he set down how he wrote his book, The American Black Chamber, which became a best-seller and one of the most famous books in the literature of intelligence. The memorandum gives a good insight into Yardley, including his misspellings.

I was a cryptographer, not a writer. Friends suggested that I return to New York and consult you. They told me that you could make anyone write, no matter what his training. So I came to New York with my last few dollars, took a room at the Commodore Club Hotel, and called you up. Your suave secretary told me that you were not in; so I left my telephone number and sat around the rest of the day waiting for a call from you. None came. The same thing happened the next day, and then the next. I was a bit discouraged, not only because my funds were low, but because no one seemed to recognize that I had a story to tell.

Well, this went on for about two weeks. Finally I found FPA [columnist Franklin P. Adams], whom I had known during the war, and asked him to recommend a literary agent. He asked me if I could write, and I told him no, but that perhaps someone else would oblige me. He suggested that I see you, and I said that was what I had been trying to do for two weeks without success.

I said, "Bye must be drunk; he's never in."

He replied, "No, he isn't drunk all the time."

"What's the matter with him?" I asked.

"He's busy-out selling stuff."

"Perhaps he is," I said, "but can't you get me another agent? Bye won't see me."

"There is no other agent," FPA retorted. "I'll fone him and tell him to see you."

Well, George, by this time, the seat of my pants was quite thin, and I went to bed and sent them out to be mended. I foned and foned and foned. Finally I caught you unaware and you said to come up.

I had had some tough spots in my life, but talking to you for the first time was the toughest of all. I was conscious of my patched trousers, and though I had rehearsed the glowing terms in which I would recite my tale of romance, adventure, and intrigue, I was confused by the sudden interruption of telephones and your conversations with the Great.

But I think you took pity on me. Anyway, you told me to come back the next day to interview Mr. Costain of the Saturday Evening Post [Thomas B. Costain, later the best-selling author of The Silver Chalice and other books].

I was a good half hour early and sat in your anti-room while your secretary wrote checks and introduced me to the successful as they streamed into your office. For two hours I waited, most of the time listening through closed doors to the exchange of conversation between Mr. Costain, you, and Mr. Franklin, the American bull-fighter, who was negotiating for a series of articles.

Finally, in despair, I started to pick up my hat and ragged coat, when suddenly Mr. Franklin ended his conversation, and you asked me to come in and tell my story to Mr. Costain.

Mr. Costain has an overpowering personality, and I felt very small in my rags and could scarcely open my mouth. Poverty had done strange things to me, though only a few months before I had stood at the top of my profession. Now I suddenly found myself with no voice, no matter, no confidence. Mr. Costain, however, was polite enough to listen for a few moments, then rushed from your office to catch a train. I followed, discouraged, and to tell the truth, in a hopeless mood.

I presumed this ended my association with you, but much to my surprise you telephoned me the next day to catch a train to Philadelphia, as the Saturday Evening Post wished to talk to me there.

I saw Mr. Costain again, whose distinguished head and restless eyes intrigued me. He introduced me to their most distinguished writer, Mr. Stout. To the latter I told my story. He seemed not particularly impressed, and I wondered at his chubby face and round eyes. He asked me if I could write the articles, and I told him no, that I was, I believed, the only person in America who couldn't write. This statement warmed him to me a bit.

Mr. Costain finally came back, and after discussing the matter with Mr. Stout, told me that he would take three articles, but that Mr. Stout was engaged on another series and could not do mine for several weeks.

This was a bit discouraging, for I wanted quick action. I came back to New York and we discussed the matter. You told me to write the articles myself, and I said I couldn't. You then promised, in order to save time, to get someone else to write them at once.

I waited several days, but did not hear from you. Finally I came back and had the nerve to show you a chapter that I had scribbled. You took it home with you and read it, or at least you said you did (I doubt if you even looked at it), and told me that all I needed was a typewriter and some paper.

These I had within a few hours, and moved from the hotel to a dark cheap room. Before me sat a typewriter, and [by] my side laid 500 sheets of paper. But I could do no more than stare into space. For days I pecked out a few lines and threw them into the fire.

Continues...

---

Excerpted from THE READER OF GENTLEMEN'S MAIL by DAVID KAHN Copyright © 2004 by David Kahn. Excerpted by permission.
All rights reserved. No part of this excerpt may be reproduced or reprinted without permission in writing from the publisher.
Excerpts are provided by Dial-A-Book Inc. solely for the personal use of visitors to this web site.

---