Enter content here


Published January 27, 2012

Beale Treasure Story
How Cipher No. 2 Was Decoded


The Ward pamphlet does not explain how the author went about his business of decoding cipher No. 2. Neither does the pamphlet explain the counting errors made by Beale in preparing his key. Nor does it discuss the differences in Beale's and the author's versions of the Declaration. Nor does it explain why the author chose to include an obviously misnumbered copy of the Declaration in the pamphlet.
The pamphlet leaves these important aspects of the treasure story entirely to our imagination and conjecture. Little wonder that some readers of the pamphlet have become frustrated and confused. Some have gotten things totally wrong, as evidenced by the following three statements:  
  • How is it possible for Beale to have used the same misnumbered Declaration as the author's misnumbered Declaration to decode cipher #2?
  • In order for the author to decode cipher #2, he would have had to start with a Declaration misnumbered in the same way that Beale's Declaration was misnumbered. 
  • As there is no way for the author to have duplicated Beale's misnumbered Declaration, it means that Beale and the author had to be the same person. 
Some are just confused:    
  • How was it possible for the author to figure out the counting errors made by Beale in preparing his key to No. 2?
  • Why did the author fail to mention anything about the misnumbered Declaration printed in the pamphlet?  
People go astray because they look at the problem in the wrong way. Instead of using computer printouts and tables of accumulated data, one should put oneself in the place of the author and imagine how he would have solved the problem, using only pencil and paper and a copy of the Declaration.

The Author's Method Explained

What some persons too easily overlook is that the author could have decoded No. 2 without first having a copy of the Declaration misnumbered in the same way that Beale misnumbered his Declaration. The pamphlet's author did not have to start with a misnumbered Declaration. Instead, he was able to decode No. 2 using a correctly numbered Declaration, and once that was accomplished he was able to figure out the counting errors that Beale made, and hence was able to prepare the purposely misnumbered Declaration printed in the pamphlet. Beale and the author were two different persons, not the same person.  

The counting errors made by Beale in preparing his key and the Declarations used by Beale and the pamphlet's author are important aspects of the treasure story. Thus, their relevance to the matter must be explained. In the discussion that follows, I shall refer to the creator of cipher No. 2 as Beale. There are two kinds of clerical errors made by Beale, clerical errors made in constructing his key (called counting errors) and clerical errors made in referencing the key while enciphering Paper No. 2 and of no importance to the present discussion.  

A counting error (e.g., counting 9 words as 10, 11 words as 10, or 20 words as 10) is a serious error, as it causes all words in the key to be misnumbered beyond the point where the error occurs. Thus, during decipherment the affected cipher numbers are decoded as incorrect letters. An equally serious problem could potentially arise if the author's copy of the Declaration happened to contain an extra word not found in Beale's Declaration, or vice versa. The consequences would be the same as with a counting error.  

Beale's Declaration and the Declaration selected by the author (reprinted in Ward's pamphlet) are different—they have slightly different wordings. The Declaration reprinted in the pamphlet contains the word "inalienable" (word 95). Yet, cipher number 95 in Beale cipher #2 is decoded as letter "U" not letter "I." It is also known that word 95 in the Declaration printed by John Dunlap on the night of July 4, 1776 (referred to as the Dunlap broadside), is the word "unalienable" not the word "inalienable." Many in fact, most Declarations printed before 1823 contain the word "unalienable." Thus, it may be surmised that Beale's Declaration contained the word "unalienable," not "inalienable," and therefore that the two Declarations are different and taken from two different source works (probably books). There are other differences in the two Declarations, as well.

The source of Beale's Declaration is unknown. However, Beale's Declaration can be partially reconstructed using the decipherment of cipher No. 2 and a copy of the Declaration. For those who wish to learn more about the Declaration of Independence, please visit my website at www.USDeclarationOfIndependence.com. You will find a copy of my book entitled DECLARATION OF INDEPENDENCE: A Checklist of Books, Pamphlets, and Periodicals, Printing the U.S. Declaration of Independence, 1776–1825, printed in 2009. A free PDF copy of the book can be downloaded and viewed on your own personal computer. I recommend that you read the preface. There are 358 checklist entries in the book. Of the 358 entries, 321 are works printed prior to 1823 and thus cover the period of the Beale treasure story. In preparing the book, I obtained a copy of each of the different Declarations, and I examined the wording in each of them. So I can speak with some authority about these 321 Declarations. The findings are detailed in the preface to the book. Beale's reconstructed Declaration is consistent with 26 of the 321 candidate Declarations, all taken from books. The Declaration printed in Ward's pamphlet is not consistent with any of the 321 Declarations, showing that it came from a source work printed after 1822.

By the late 1970s, members of the Beale Cypher Association expressed some concern that no one had yet explained how the pamphlet's author went about his work of deciphering cipher No. 2. To address this, a colleague (Frank Aaron) and I put together a talk entitled "How the Message in Paper No. 2 was Recovered." The talk was presented at the Second Beale Cipher Symposium, 1979, and published in its Proceedings.

When I first learned about the Beale treasure story, I spent some time trying to break the ciphers, by selecting different books and documents, numbering them by word and by letter, and making trial decipherments. I examined the U.S. Constitution, the Articles of Confederation, passages from Shakespeare, and many others, all to no avail. It wasn't long before I realized that such work was extremely tedious. Also, it didn't take long to learn that the process could be greatly speeded up by selecting only a few short sequences of low-valued numbers that would serve as test cases. And this would mean that only a few hundred words would need to be numbered in each candidate key text, greatly reducing the amount of time necessary to examine each passage of text.

I recalled that the author of the pamphlet said "With this idea, a test was made of every book I could procure." It sure sounded like the author had examined many books and documents before selecting and using the Declaration to decode No. 2. If true, I felt that the author would likely have adopted the same short cut that I hit upon after attempting to decode No. 1 and No. 3 using just a few different candidate key texts. Moreover, the author would likely have recognized that mistakes made in numbering the words in a candidate key text would cause some letters in the decoded plain text to be incorrect. Thus, the author was probably careful to avoid making mistakes in numbering the words or letters in each text.

It so happens that a rather long string of mostly low-valued numbers occurs right at the beginning of cipher No. 2. The twenty-one cipher numbers are these:

115, 73, 24, 807, 37, 52, 49, 17, 31, 62, 647, 22, 7, 15, 140, 47, 20, 107, 70, 85, 56

If the pamphlet's author had used this string of cipher numbers, he would have needed to number only the first 140 words in each candidate key text. The large numbers 807 and 647 could be skipped, as the remaining nineteen numbers would be enough to recognize gibberish and rule out most texts. And, it would most likely be enough to recognize syllables or words produced by a correct key text.

If words one through 140 are numbered in the pamphlet's Declaration and the string of 21 cipher numbers taken from the beginning of No. 2 are deciphered (omitting 807 and 647), the recovered text looks like this:

I H A _ E D E P O S _ T E D I N T H E C O

The author could have used other strings of low-valued numbers in No. 2 to confirm that he was on the right track. Moreover, the fact that cipher numbers 807 and 647 are not decoded, but instead represented as spaces, actually makes it easier to see probable words in the decoded plain text. A bit of luck never hurts.

In fact there was a bit of luck involved in the selection of Beale's Declaration and the author's Declaration, which is worthwhile to point out. As I say, a little bit of luck never hurts. Word 155 ("A") in the phrase "institute a new government" found in Beale's Declaration, though not present in the original Dunlap broadside printed by John Dunlap the night of July 4, 1776, is also found (by luck) in the pamphlet's Declaration. It so happens that the phrase "institute a new government" is a major variant found in 112 of the 321 different Declarations printed prior to 1823, and it is found in many Declarations printed afterwards. Also, it is fortunate for us that Beale did not select one of the 64 of 321 Declarations in which words 148 and 149 ("or to") are replaced by a single word ("or" or "and"). Hence, the first perceived discrepancy between the two Declarations occurs at word 510, spelled as two words ("mean" and "time") in Beale's Declaration and one word ("meantime") in the pamphlet's Declaration. The two words "mean time" or "mean-time" occur in 305 of the 321 Declarations printed before 1823; the single word "meantime" is a more common variant found in Declarations printed in the latter part of the nineteenth century. The first counting error in Beale's Declaration occurs somewhere between word 241 ("invariably") and word 247 ("design"). Thus, no discrepancy occurs between the two Declarations until after word 241.

This means that if the author had correctly numbered the first 140 words in his Declaration (as conjectured), and deciphered the first twenty-one cipher numbers in No. 2 skipping over numbers 807 and 647 (as conjectured), he would have obtained the decipherment given above, namely, I HA_E DEPOS_TED IN THE CO.

Even before numbering additional words in his Declaration and decoding additional cipher numbers in No. 2, the author may have recognized that number 807 was likely letter "V" and number 647 was likely letter "I."

It seems likely that the next step taken by the author would have been to number the remainder of his Declaration, say to word one thousand. Then, he would have discovered that cipher numbers 807 and 647 were not decoded as letters "V" and "I." He probably rechecked the numbering of the words in his Declaration, but alas numbers 807 and 647 could not be decoded correctly. From this, the author must have realized that Beale had likely made an error in preparing his key—perhaps a counting error.

Before continuing, it is worthwhile to note that with only 140 words numbered in the author's Declaration, 535 of the 763 cipher numbers in No. 2 (70 percent) could be correctly decoded. Moreover, if the pamphlet's Declaration was numbered to word 241, or beyond, then 606 of the 763 cipher numbers in No. 2 (79 percent) could be correctly decoded.

It seems probable that after numbering additional words in the Declaration, No. 2 was decoded by continuing with the 22nd cipher number, and decoding each number one at a time until reaching the end of cipher No. 2. In other words, the cipher would be decoded in the usual manner, like any other cipher whose key was available, with one exception. Letters that didn't seem to decode correctly were examined further to see if they could be decoded by replacing the decoded letter with some other letter on the basis of the surrounding context. All decoded cipher numbers up to and including number 241 would have been decoded correctly by consulting the key. Only numbers greater than 241 would have decoded incorrectly.

For sake of argument, suppose the process is moved forward in time in order to see what the decoded letters would have looked like after forty cipher numbers had been decoded using only the key:

115, 73, 24, 807, 37, 52, 49, 17, 31, 62, 647, 22, 7, 15, 140, 47, 20, 107, 70, 85, 56,

I       H    A   R      E    D    E    P    O   S    C     T    E   D   I      N   T    H     E    C    O 


239, 10, 26, 811, 5, 196, 308, 85, 52, 160, 136, 59, 211, 26, 9, 46, 316, 554, 122 

U      N   T    F      O   F     G     E    D    F      O     R    D     A   B   O   A     T      F

Notice that number 554 in the author's correctly numbered key happens by accident to reference word "to" beginning with the letter "T." The word "to" is the wrong word but it begins with the right letter. As I say, a little bit of luck never hurts. From a perusal of the cipher numbers and the matching decoded letters, it seems clear that the author could have deduced that numbers 807, 647, and 811 stood for letters "V," "I," and "Y." Likewise, he may have noticed that number 308 should decode as the letter "B" to form the word "Bedford" and that number 316 should decode as the letter "U" to form the word "about." The process most likely consisted of parsing the decoded string of letters into words while at the same time correcting certain of the decoded letters. Such an approach is possible only if enough letters in the decoded text are correct to start with. In our case, it works.

The author may have noticed that word 309 ("Britain") begins with letter "B" and word 317 ("usurpations") begins with letter "U." Again, this would be further evidence that Beale very likely made a mistake in preparing his key, and that number 308 was intended to refer to word 309 and that number 316 was intended to refer to word 317. Even if the author was not so astute and did not recognize the emerging pattern immediately, he would most likely have noticed it eventually. And, once that happened, he would have been on his guard to decode cipher numbers by first looking at the referenced word and if that didn't seem to give the right letter, then attempt to find a correct decoding by consulting an adjacent word in the key. That tactic would have allowed the author to correctly decode 90 percent of the cipher numbers in No. 2. Moreover, as numbers like 807, 647, 811, 308, and 316 were successfully decoded, on the basis of their context in the partially decoded No. 2, a tally of these cipher numbers and their corresponding decoded letters could have been maintained, perhaps in a list. This would permit other occurrences of these same cipher numbers to be conveniently decoded as well, which would also serve as an additional confirmation that the entries in the accumulated list were indeed correct. If an unusually long string of large numbers was encountered, without being able at once to be decoded, such strings could be passed over and decoded later, after additional portions of the cipher text had been decoded. As more and more cipher numbers in No. 2 were decoded, and as more and more correct decoded letters were obtained, the author's task would have become easier and easier.

By the time the author completed his decipherment of No. 2, he would have recognized that Beale made one or more clerical errors in preparing his key. Much later, when the author decided to publish the Beale Papers in pamphlet form, it is supposed that he recognized two important things. Attempts to explain how No. 2 had been decoded or attempts to explain the clerical errors that Beale made in preparing his key were issues much too complex to be dealt with in a small pamphlet.

Moreover, it was the author's intention to say in the pamphlet that ciphers No. 1 and No. 3 could be solved by finding the right key texts. But, if Lynchburg residents had known the amount of work involved in decoding No. 2, many of these people would have been discouraged before ever beginning the task. This could negatively impact the sale of pamphlets.

However, the author also recognized that readers would want to see how No. 2 could be decoded using the Declaration. His compromise, or way out, was to provide a copy of the Declaration whose words were  purposely misnumbered, thus compensating for the mistakes made by Beale in numbering his copy of the Declaration, and thereby permit No. 2 to be decoded by readers of the pamphlet. The author prepared the pamphlet's misnumbered Declaration using the decoded No. 2 and his copy of the Declaration. The method involves adding or subtracting a trial displacement to one or more cipher numbers in sequence and consulting the key to see if the referenced words decode correctly. The procedure is repeated until a correct displacement is found. Additional cipher numbers are tested to provide additional confirmation and to identify the full set of cipher numbers corrected on the basis of the identified trial displacement. Once the full set of cipher numbers has been identified, the affected words in the Declaration are renumbered so that the affected cipher numbers will reference the correct words in the Declaration. The same procedure can be repeated to renumber additional affected words in the Declaration.

The author also made several editing changes to the actual decipherment of No. 2, apparently to make the text read "better."

I will be the first to admit that the treasure story would be more appealing and convincing, if the author had explained how he went about his work of decoding No. 2, rather than leave this entirely to conjecture. But, I can also see, from a marketing perspective, why he elected to omit this information in the pamphlet.