Keepapitchinin, the Mormon History blog » “The Qmlbwpnygax Eujugec Have Not the Power to Ktgjie the Wzznlhmpygtg”: Codes and Ciphers in Mormon History (part 3)
 


“The Qmlbwpnygax Eujugec Have Not the Power to Ktgjie the Wzznlhmpygtg”: Codes and Ciphers in Mormon History (part 3)

By: Ardis E. Parshall - December 11, 2008

Cable/Telegraph Codes

The substitution ciphers we have looked at can encrypt any message, even words or names unknown at the time a cipher was devised. True codes, on the other hand, are a closed system – when people are assigned new labels (Gazelam for “Joseph Smith,” for instance), you cannot introduce a new name without notifying correspondents of the additional code word.

Nineteenth century Mormon leaders used true codes following the completion of the telegraph, and especially following the completion of the transatlantic cable. Cable codes could be used to maintain secrecy, but generally they were intended to lower the expense of cabling between, say, Liverpool and Salt Lake City, because a single code word could replace the multi-word name of a person or city, or signal the completion of an assignment.

Actual 1868 message telegraphed by Brigham Young (Salt Lake City) to Hiram B. Clawson (New York City):

Yankee cherub highness frog tablet darby dungeon mint granby budget. Horace walrus magnet it ramble Ireland time.

signifying:

???

(I have not found the key, although I’m told the First Presidency telegraph
code books do still exist somewhere in the archives.)

A more dramatic example is the telegraph code used among polygamists “on the underground” in southern Utah, whereby men could be warned of subpoenas issued to their wives, the approach of federal marshals, and the retreats of their associates. Complained one marshal to a judge, “[P]eople along the [rail]road are advised of the presence of a Deputy Marshal on board any train … [and men] are employed to watch the movement of officers and to … aid the criminal … to escape.” Telegraphed codes played a major role in such escapes.

Simulated message in a code actually employed by Mormon men “on the underground”:

RING LOST JAMES WHITE POCKETS BLACK

signifying:

The U.S. marshals are hunting for Erastus Snow in St. George. Move your quarters to Beaver.

Cannon’s codes

Because of often intense political antagonism between Mormondom and the broader American culture, Utah Territory’s delegates to Congress understandably might have taken special care to protect their correspondence. Surprisingly, though, I have found little evidence of coded communications in the first 20 years of Mormon presence in Congress. The letters of John M. Bernhisel reflect his personality – courtly and dignified, reporting political affairs without gossip or intrigue.

The letters of William H. Hooper are likewise matter-of-fact and indicate little awareness that security was an issue – although he occasionally discussed sensitive political matters that could easily have justified encryption.

George Q. Cannon, on the other hand, was a far more sophisticated man of affairs than either of his predecessors. He had waged fierce battle in the newspaper wars, and had engaged in politics in Hawaii, California, New York, and Washington as well as in Salt Lake City, before his 1872 election to Congress. Cannon was keenly aware of the lengths to which an opponent might resort to gain advantage, and he took steps to protect his communications. The first such step was to avoid writing directly to Brigham Young, whose name on an envelope would have attracted attention all along the mail route. Instead, he often wrote to trusted middlemen, such as Daniel H. Wells, who would deliver letters to Brigham Young.

George Q. Cannon -to- Daniel H. Wells -to- Brigham Young

One way in which Cannon disguised his messages was to write in a roundabout way that presumably would be understandable only to his brethren in Salt Lake City. We have seen how he avoided mentioning Kane by name. A far more elaborate illustration occurred in 1874:

In January that year, Cannon wrote to relay critical information about steps that Congress would soon take. In a separate letter, Cannon stressed the reliability of his information. Rather than naming his source, however, Cannon directed Young to a letter written a year earlier, to note whom Cannon had met at a certain dinner party.

From that letter, Young discovered that Cannon’s reliable source was William Tecumseh Sherman, a powerful figure and one not generally known to be friendly to Mormon interests.

Another method of preserving secrecy in his letters was one which arose as a consequence of the peculiar history of Mormondom: When Cannon wished to say more than could be said by the roundabout method of allusion, and when there wasn’t time to employ a more formal method of encryption, Cannon simply dropped into Hawaiian, knowing that fellow missionary Joseph F. Smith would be able to translate at the Salt Lake end. In this example, the desirability of confidentiality is obvious: This letter proposes to bribe members of Congress to kill a disastrously anti-Utah bill before it reached the House floor. “We think its defeat is important, and we think we can defeat said Bill by giving a few thousand dollars. … Shall we try?”

But even Hawaiian, as mystifying as it would have been to an interceptor, was not the ultimate weapon in Cannon’s code arsenal. Cannon also made use of an ingenious cipher – one that was easy to remember, yet so complex that it would be almost impossible to break without the key. There is no reason to suppose that Cannon or any of his associates invented this cipher, because codes of this type were being produced commercially by this time. One such cipher – Larrabee’s – was evidently used by the Mormons, because Kane owned a copy sent to him by Brigham Young. Larrabee’s was not the precise code used by Cannon, but Cannon’s code worked in precisely the same way.

Such ciphers were built around tables such as this one, with a different cipher associated with each letter of the alphabet.

(This page shows only the first 6 of 26 keyword charts;
there is a separate chart for each letter of the alphabet.)

The first step in encoding a message is to write out the plain text:

BRIGHAM LOVES AMELIA

The next step is to write a keyword (in this example, the word MILK) above the plain text, repeating to correspond to the length of the message. The keyword could be changed as often as desired, as long as both the sender and receiver were informed.

MILKMIL KMILK MILKMI
BRIGHAM LOVES AMELIA

(I have not learned how to change fonts in WordPress to allow me to use a monospace font. If you have trouble understanding the following explanation with the garbled illustrations, write out the message yourself with a pencil and paper, being careful to put “M” immediately over “B”; “I” immediately over “R”; “L” immediately over “I”; and so on.)

Then a chart is consulted. In our example, we locate the chart for the keyword’s letter M, and find that “B” is represented by N, and we write N immediately under the M/B.

MILKMIL KMILK MILKMI
BRIGHAM LOVES AMELIA
N____________________

Next, we locate the chart for the keyword’s letter I, and find that “R” is represented by Z, and we write Z immediately under the I/R.

MILKMIL KMILK MILKMI
BRIGHAM LOVES AMELIA
NZ___________________

And so on, letter by letter, until the full message has been encrypted.

MILKMIL KMILK MILKMI
BRIGHAM LOVES AMELIA
NZT__________________

MILKMIL KMILK MILKMI
BRIGHAM LOVES AMELIA
NZTQ_________________

MILKMIL KMILK MILKMI
BRIGHAM LOVES AMELIA
NZTQT________________

MILKMIL KMILK MILKMI
BRIGHAM LOVES AMELIA
NZTQTI_______________

MILKMIL KMILK MILKMI
BRIGHAM LOVES AMELIA
NZTQTIX______________

MILKMIL KMILK MILKMI
BRIGHAM LOVES AMELIA
NZTQTIX VADPC MUPVUI

Only the encoded message NZTQTIX VADPC MUPVUI is transmitted.

When the message is received by someone knowing the keyword, the process is reversed to decipher the message.

The reason, of course, why such a code is difficult to break is because each letter may be represented by a different coded letter almost every time it appears: In our sample, the letter “A” is represented by I, M, and again by I. Someone else could send exactly the same message, using the same tables, but as long as they used a different keyword, the encrypted message would look entirely different. Such a code was secure enough even for use over the public telegraph wire.

The title of my paper is taken from a Cannon letter. Although Cannon wrote primarily in plain English words, he encoded enough to make the message unintelligible.

“Suppose it were to be decided that the qmlbwpnygax Eujugec did not have the power under the selisbgepy to ktgjie upon uztuexr hwzkxw the wzznlhmpygtg they have been exercising …”

Cannon wrote a related letter to David McKenzie, a clerk in Brigham Young’s office, providing him with a list of the encoded words but not their context. McKenzie decoded the words – we can see here that the keyword used at this date was “fifteen” – and provided Young with their translation, all without McKenzie’s understanding Cannon’s message.

Combining the text written in plain English with the encoded words deciphered by McKenzie, Brigham Young was able to read a legal analysis concerning the future of Utah’s probate courts:

Suppose it were to be decided that the qmlbwpnygax Eujugec [legislative Assembly] did not have the power under the selisbgepy [Organic act] to ktgjie [confer] upon uztuexr hwzkxw [probate courts] the wzznlhmpygtg [jurisdiction] they have been exercising … “

Incidently, the cipher used by Kane in 1858 probably worked in much the same way as Larrabee’s or the variation used by Cannon, because it was based on Robert Patterson’s code. The 26 separate tables of Larrabee‘s were combined into a single grid in Patterson’s code:

The Kanes’ version had a diabolically difficult twist. Rather than using a predetermined keyword such as fifteen or milk, both the sender and the receiver had to construct a keyword based on the first letter of each line in a certain portion of Genesis as it appeared in a particular edition of the Bible; the scripture used to generate the keyword changed according to the date of the encrypted letter. Printing variations in the Bibles used by Kane and his wife, and perhaps carelessness in constructing the keyword, would account for the tremendous difficulty Elizabeth Kane had in deciphering the coded messages sent by Kane.

Conclusion

So what does this aspect of our history say about us as a people?

First, that we were a complex society with secrets to keep – just as any organized people might guard financial interests against rivals or political interests against opponents. The earthly Kingdom of God was and is as complex as any worldly kingdom.

But the use of codes and ciphers in 19th century Mormondom illustrates how very much “the Other” we were, outsiders in the very society where we had been born. Based on hard experience from our earliest days, reinforced throughout the century, Mormon leaders anticipated bloody, devastating opposition, and guarded secrets to preserve every possible advantage. The use of codes and ciphers is a manifestation of the seriousness with which our forebears took the scriptural caution: “I send you forth as sheep in the midst of wolves: be ye therefore wise as serpents, and harmless as doves.” The use of codes was an expression of both vigilance and ingenuity.

Part 1, Part 2



22 Comments »

  1. Great stuff, Ardis. I was not very aware of the use of codes in 19th century church history. This is fascinating.

    The use of a short key like the word “fifteen” that you show in the example here would be sufficiently complex to hinder anyone working manually to try and decrypt the messages. However, with the advent of computers, a 7 letter key would be ridiculously easy to decrypt. Typical encryption algorithms these days run from 128 to 256 bit keys (the equivalent of keys of up to thousands of alphanumeric characters), and employ multiple passes of encryption, breaking up the messages and rotating specific parts multiple times so that someone trying to break a code by trying successive random keys would have to take days, months, or even years to break. Most recently, such encryption schemes as DES (using a 56 bit key) or WEP for your wireless routers at home (similar to DES) can be broken in minutes to hours using typical laptop computers and available software. The more advanced encryption technologies used today are currently restricted for export under US law to many countries.

    It would be interesting to know for sure, but one can only assume that the church probably uses encrypted email communications for sensitive information, just as banks and businesses do. And paranoid individuals.

    Comment by kevinf — December 11, 2008 @ 9:42 am

  2. Ardis, you really should pull this together and publish this in one of the Journals. It is simply to splendid not to.

    Kevin, I have heard that the temple instruction encryption is mighty and strong.

    Comment by J. Stapley — December 11, 2008 @ 10:41 am

  3. Paranoia has left its stamp on Mormon psyche. It shows in our how we share church financial records and how we guard the CHI. It is even at the root of the Word of Wisdom.

    Behold, verily, thus saith the Lord unto you: In consequence of evils and designs which do and will exist in the hearts of conspiring men in the last days, I have warned you, and forewarn you, by giving unto you this word of wisdom by revelation. D&C 89:4

    Just because we are paranoid, however, doesn’t mean there aren’t people out to get us.

    Comment by BruceC — December 11, 2008 @ 10:44 am

  4. Even working manually, such a cipher can be broken fairly easily without knowing the keyword initially, given a surprisingly short amount of encrypted text. Short messages such as the examples Ardis gave above were probably reasonably secure, but if the same keyword is used for enough text, they can be broken quickly.

    The key length can be found by comparing the encrypted message with a copy of itself shifted by various numbers of letters (1, 2, 3, etc.) and checking how many times the same letter appears in the original message and the shifted message.

    For example, one would compare

    NZTQTIXVADPCMUPVUI
    _NZTQTIXVADPCMUPVUI
    and get 0,
    NZTQTIXVADPCMUPVUI
    __NZTQTIXVADPCMUPVUI
    and get 1 (the two Ts match), etc. The shift with the most coincidences is the most probable length for the keyword. (This works because when the shift is the same length as the keyword, then letters like E that appear more frequently in English are more likely to be matched with other occurrences of E, while if the shift is a different length than the keyword, the distribution of letters will be more random.)

    Once the length of the keyword is known–say it’s 7, as in the word “fifteen”–it’s easy to take every 7th letter of the original message and perform a standard frequency analysis. Since in this case each letter corresponds to a simple shift of the alphabet, these are not hard to decrypt. (For instance, if you guess that G stands for A, but the message has lots of F’s and D’s, your initial guess is probably wrong unless the original message has lots of X’s and Z’s in it, which isn’t too likely. Such arguments can eliminate most possibilities.)

    This is a standard topic in mathematical cryptography courses; more information can be found here, and the Java applet at this site is a fun way to play with decrypting such messages. Variations of the code can make encryption more difficult, but this type of cipher certainly doesn’t live up to its historical billing as le chiffre indéchiffrable.

    Comment by XiGauss — December 11, 2008 @ 11:06 am

  5. XiGauss, I agree: my crypto courses had stuff like the Cannon code as do-by-hand homework assignments. 19th-century cryptanalysts were very good at exactly the analytic process you describe. Even a message with only several hundred characters in such a code can generally be decrypted by hand, although it may take a few hours and some false starts. As these things go, such codes are only a bit trickier than straight substitution ciphers and were recognized as such by 19th-century cryptographers and analysts. A longer key would help, and Mormons had natural resources in this department — the Book of Mormon would have been a good key, or the Journal of Discourses. Use one page per message and the resulting code will be substantially hard to break by hand.

    My guess here, once again, is that any secrecy gained by using these codes results primarily from a lack of serious effort to break the code. Which may have sufficed for the time, of course.

    Comment by J. Nelson-Seawright — December 11, 2008 @ 11:29 am

  6. XiGauss,

    Way easier with a computer than manually. The technique you describe becomes more plausible as the length of encrypted text increases. Encrypting half a dozen words would be difficult to crack. Several pages would become much easier, relatively speaking, given sufficient time, reams of paper, and lots of motivation. However, it would discourage all but the most monomaniacal of enemies.

    Note to the paranoid: There are several companies offering encrypted email service that even eliminates copies of emails that are normally left on the server. The message is written on one end, encrypted, transmitted, and displayed at the other end, but cannot be copied, printed, or saved. Once you close the email, it’s gone forever.

    It’s a fascinating concept, but apart from drug dealers or members of secret combinations, you really have to be paranoid to use it.

    Comment by kevinf — December 11, 2008 @ 11:30 am

  7. This was great; thanks Ardis.

    I second J’s suggestion.

    Comment by Ben — December 11, 2008 @ 1:22 pm

  8. Outstanding.

    Comment by BHodges — December 11, 2008 @ 2:15 pm

  9. XiGauss and you others with cryptographic experience or interest — do you know when the techniques for breaking these codes were developed? A code like Patterson’s at the time of the Lewis and Clark expedition would probably have been entirely safe because virtually no one who hadn’t been initiated into the secret would have had a clue what was going on. By the 1870s, though, when published manuals like Larrabee’s were for sale in stationery stores, then lots of businessmen would have had an understanding of how to encode. I’m still thinking there would have been a lag between having the idea to encode this way, and figuring out techniques for decoding without the key.

    Comment by Ardis E. Parshall — December 11, 2008 @ 2:41 pm

  10. Wonderful, Ardis. Truly fascinating.

    Comment by Ray — December 11, 2008 @ 2:46 pm

  11. Great post Ardis. Amazing.

    Kevin & J., I can’t speak for all temple related correspondence, but I know from a good friend who was a Temple President that certain sacred documents are only distributed in person, and that he has gone to the airport to receive said documents and watched the deliverer turn around and get back on a plane.

    Comment by Bud Fox — December 11, 2008 @ 3:54 pm

  12. As to the history, ciphers go back to Imperial Roman times, and were in common usage for military communications in Europe during the 1700’s using codes like these. They were fairly common during the English Civil Wars between the puritans and the royalists in the 17th Century. John Wallis, Cambridge mathematician, had the distinction of doing codebreaking for both Parliament under Cromwell, and later for the royal court. He is credited for introducing the concept of “keys”, making coded messages more secure than previous substitution ciphers and simple algorithms.

    They were almost certainly used by merchants during those time, but only likely if they had access to the folks with mad hot crypto skills, generally meaning that they had to seek the good graces of the royal courts. And I suspect that being the royal cryptographer may also have come with some liabilities; you knew all the secrets, and so you were probably always viewed with suspicion, not trust.

    Comment by kevinf — December 11, 2008 @ 4:00 pm

  13. David Kahn’s book The Codebreakers: The Comprehensive History of Secret Communication from Ancient Times to the Internet is probably the best book on the history of cryptography. Simon Singh’s The Code Book is a more readable history. A lot of what I know about the history of cryptography comes from those sources, with a little coming from cryptography textbooks such as Trappe/Washington or Stallings.

    I don’t have Kahn’s book in front of me right now, but Singh’s chapter on the Vigenère cipher (of which Larrabee’s seems to be an example) states that Charles Babbage probably broke the cipher in 1854 but didn’t publish his findings, and that Friedrich Wilhelm Kasiski independently published a solution in 1863. However, Kasiski’s work (in German) does not seem to have been widely known, and as late as 1917, Scientific American still characterized this cipher as “impossible of translation” (Supplement LXXXIII, 1/27/1917, page 61, according to Trappe and Washington). Thus, these solutions don’t seem to have spread as quickly as the cipher did.

    Of course, some of those who knew the solution would have had obvious motives to keep it secret, if they were using it to read others’ mail. Tracking the history of cryptography can be difficult, since there are so many secrets involved.

    Comment by XiGauss — December 11, 2008 @ 4:01 pm

  14. Bud,

    I also recall hearing that when the Mexico City temple was opened, the films were delivered in person, but had considerable footage of a boring organ recital put on the front end, so if anyone lost/stole the films, they would likely give up viewing after a few minutes, and not get to the good stuff.

    Comment by kevinf — December 11, 2008 @ 4:03 pm

  15. Sorry for all these comments, but the use of Hawaiian for a simple code obviously brings to mind the use of Navajo radio operators in the Pacific during WWII.

    However, a more interesting use of foreign languages relates to BYU basketball days when Danny Ainge and Timo Saaralinen both played. Timo was a native of Finland, and I believe one or two other players or coaches had been on missions to Finland, so the team used to call out inbounds plays in Finnish, which pretty much made it indecipherable by the opposing scholar/athletes.

    Things like that don’t always work, especially in Utah. There is the true story of two Finnish tourists who got pulled over in Sandy after an illegal u-turn, and feigned not knowing English by speaking only in Finnish. The cop who pulled them over happened to have served in Finland, and replied in fluent Finnish that their turn was illegal in any language.

    Comment by kevinf — December 11, 2008 @ 4:14 pm

  16. Sorry for all these comments

    Don’t be! Comment early and often (Chicago political news must be on my mind for some reason).

    The Navajo code talkers went even farther than using Navajo to communicate — they created a code-within-a-code whereby a “turtle” was a specific kind of watercraft and a “hawk” was a specific kind of aircraft, or whatever. Even if somehow the Japanese had had a Navajo on their staff, he couldn’t have known the code within the language. What a terrific skill!

    XiGauss, I’m guessing from your handle and your seemingly intimate knowledge of these sources that you use mathematics in real life. You’re adding a lot to this conversation. It’s good to have you here.

    I’ll be away from the computer much of the afternoon and evening. A lot of comments have been getting trapped in the filter for some reason. If something doesn’t show up when it should, please accept my apologies and know that I’ll free the comments as soon as I can. I appreciate all the interest in this topic and the encouragement to prep it for publication.

    Comment by Ardis E. Parshall — December 11, 2008 @ 4:38 pm

  17. XiGauss has the same sources on the Vigenere cipher that I do, and he’s right that the decryption methods were only published in the last half of the 19th century.

    The cipher, one might note, was already very old by the 19th century, having been formulated at least as early as 1467 by Renaissance man Leon Battista Alberti.

    Comment by J. Nelson-Seawright — December 11, 2008 @ 5:12 pm

  18. Ardis,

    This is positively fascinating. How many days / weeks / months / years of research does this represent?

    I affirm the suggested publication, for what it’s worth!

    Comment by John Turner — December 11, 2008 @ 9:44 pm

  19. Ardis,

    I’ve been putting off reading this series until after I finished all of my papers for the semester. It was well worth the wait and anticipation. This was very cool stuff. Thanks.

    Comment by Christopher — December 11, 2008 @ 10:15 pm

  20. Aw, shucks, guys [scuffing toe modestly]. As I’ve said on one of these parts, I’ve always been interested in codes. At one point I even kept my personal files alphabetized not by ABC… but by ETAONRISHDLFCMUGYPWBVKXJQZ, the letters in order of their English frequency. I notice whenever I run across code-events in the Mormon materials I use for normal work, and this paper really only took a week or two to pull together, after having made those mental notes for several years.

    Matt Grow advised on the Kanes’ 1858 experiences, and the staff at LDS Archives were kind enough to let me have images of the snippets of letters they don’t normally allow to be copied, because transcriptions really wouldn’t do the job in some cases.

    Comment by Ardis E. Parshall — December 11, 2008 @ 11:01 pm

  21. I struggle to keep my materials organized using the plain old alphabet.

    On the other hand, next time I’m accused of turgid writing, I’ll claim it’s mid-nineteenth-century Latter-day Saint code.

    Comment by John Turner — December 12, 2008 @ 7:50 am

  22. Love it. I ran across some coded passages in the diaries of Nathaniel Baldwin at the UofU. Baldwin was famous for inventing the headphones used on military radios. He was also a key figure in the founding of Mormon fundamentalist groups in the early 1900s. From context, I believe the coded passages dealt with prospective plural wives and meetings with other polygamists. I feel motivated now to go back and see if I can decode them.

    Comment by Morgan Adair — December 12, 2008 @ 10:21 am

Leave a comment

RSS feed for comments on this post.
TrackBack URI