In July 2023, Sotheby's released a notice of an auction of papers including the instructions of the Continental Congress to the American peace commissioners (Sotheby's (Lot 1006)). They shed a new light on use of ciphers by the founding fathers.
Independence of the United States from Great Britain was finalized by the Treaty of Paris in 1783. The American peace commissioners who worked through the negotiations were John Adams, Benjamin Franklin, John Jay, and Henry Laurens.
It is well-known that the original commissioner, John Adams, could not decipher the instructions from the Continental Congress dated 15 June 1581. His instructions were enciphered in James Lovell's cipher (equivalent to the Vigenere cipher) with a keyword "CR". Lovell knew Adams had had troubles with his cipher, and wrote to Adams "If any Thing prevents your coming at the Purport Doctr. Franklin can certainly decypher his." As late as March 1782, Adams (still in Holland) had to ask Franklin in Paris to send him a copy of the deciphered instructions. (See another article.)
Indeed Franklin appears to have had no trouble in reading his instructions. But it was not that Franklin, as a scientist, could better handle Lovell's polyalphabetic cipher. There has been reason to believe instructions to Franklin were in a different cipher from Lovell's. (See another article.)
Now, the papers presented for auction show that the instructions for Franklin were encoded in John Jay's book code.
The papers put on auction include:
(1) a deciphered copy of the instructions, with attestation by Franklin "A true Copy B. Franklin" and endorsements "Instructions relative to the Peace Treaty" (annotated "Benj. Franklin's handwriting" in later hand) and "Ultimately to govern ourselves by their advice & opinion" ("John Adams' handwriting"; extracted from the instructions, presumably considered problematic by Adams).
(2) the instructions in code, dated 15 June 1785 and signed "Sam[ue]l Huntington President", attested and countersigned by Charles Thomson, Secretary to Congress. There is a note: "Mr Jay has the key."
(3) partial decipherment in John Jay's hand.
The auction title "The Treaty of Paris | "A True Copy of the congressional instructions to the American peace commissioners, attested by Benjamin Franklin and John Adams" refers to (1), but the item (2) is of much more interest in terms of historical cryptography.
The instructions are clearly in a different code from that used for Adams.
The ciphertext "13372 1650 11371 11510 11537 15488 15555 ..." immediately reminds one of a book code, with each codegroup consisting of a line number and a page number (there is slight spacing between the line and page numbers). Even without looking at the note for Franklin "Mr Jay has the key", it is remembered that John Jay used various book codes. In particular, the bars over/under digits (omitted in my transcription) point to a variant of one specific book code used by Jay and described as WE079 by Ralph E. Weber (see another article).
Jay proposed to Thomson (29 February 1780) to use "The second part of Boyers Dictionary where the English is placed before the French", specifically, "the thirteen Edition & was printed in London in the year 1771." The code group consists of (line number counting from the bottom + 10) and (page number + 5), with a bar under the first [second, third] digit indicating the first [second, third] column.
In reply, Thomson proposed (7 June 1780) an improvement whereby the line number may also be counted from the top, in which case the bar should be placed above the digit rather than below it.
With such information, my informant found a copy of this edition available on the web:
Abel Boyer, The Royal Dictionary Abridged, The Second Part, containing the English before the French (1771) (Google)
(The book is not paginated. Although Jay instructed to manually number the pages, the pdf file saves the work. Moreover, the first page of the dictionary is p.6 of the pdf. So, by coincidence, the pdf page number accounts for Jay's page offset of +5!)
The following is the attached plaintext, annotated with corresponding code groups in brackets. (I omitted the bars indicating columns.)
The association of code groups with words is mine. There are a few mismatches, but I believe it is generally correct.
I confirmed that the page number in the book code matches the pdf page number (i.e., the book page + 5) for all the words starting with A and words SUCH, THE, TO, WITH, YOU (the last page). Thus, we can safely tell that the page offset is +5 (as specified by Jay).
As to the line offset, sometimes it's 10 (as specified by Jay), but there are too many cases where it is different from 10. I have not been able to figure out this seeming discrepancy.
Some observations on usage of the book code:
• It seems the word "plural" is used as a plural ending "-s". (The word "plural" occurs on the right page, but on the first line (matching the line offset 10) in the second column, indicating the bar should be over the second digit, whereas the code group 11401 has a bar over the first digit in two of the three instances. (The third instance omits the bar.)
• Verbs are given as to-infinitives in the dictionary. So, the code group "26 10" encodes "to accede", rather than "accede."
The substance of the instructions is, of course, the same as that of the known text. But there are slight textual differences.
• The text in code reads "on behalf of", while the plaintext reads "in behalf of."
• The word "belligerent" in the plaintext is coded as "power-s at war", because "belligerent" is not in the dictionary. (Words "power", "-s (plural)", and "at" also occur elsewhere, and the word "war" (11547) at least appears on the correct page, the correct left column, and the bottom line (matching "11" with a line offset of 10).) (By the way, "belligerent" in the plaintext is an adjective, modifying "powers" after the insertion clause between double dashes.) This means the plaintext attached is not the result of decoding, but it came from some other source (record in the Continental Congress?).
• The text in code reads "you are to make the most candid and confidential communications we all subject to the ministers of our generous ally", which in plaintext corresponds to "you are to make the most candid and confidential communications upon all subjects to the ministers of our generous ally". This seems to be an encoding error.
• The text in code reads "make a formal acknowledgement of our independence", while the plaintext simply reads "acknowledge our independence."
• The decoded copy omits the date line.
The instructions are accompanied by partial decipherment in Jay's hand. It is described by the auction website as "John Jay's deciphering of the instructions: John Jay. Autograph document, being a worksheet transcribing the enciphered passages of Huntington's instructions to the peace commissioners."
One section reads:
The blank before "to concur" corresponds to "authorized and instructed", not in code. Similarly, the second and the third lines omits "his most Christian Majesty" and "the Empress of Russia and the Emperor of Germany." So, this section corresponds to the first sentence in the body of the instructions. But it is not very clear whether Jay decoded the instructions on these sheets.
The instructions have a note "Mr Jay has the key." This suggests Franklin may not have had the key himself. At least, I don't know any instance whereby Franklin used this book code prescribed by Jay.
When the instructions were sent in June 1781, however, Jay was in Spain and would not arrive in Paris until June 1782. Did the Continental Congress expect Jay to arrive in Paris soon, or consider the instructions need not be decoded until Jay's arrival? I need to check history books on these points.
However, there is evidence indicating that Franklin already had a deciphered copy as of April 1782: he sent a copy by courier to Adams, who acknowledged its safe receipt on 2 May (see another article).
In the first place, this means the deciphered copy was sent from Paris to Holland in clear. Probably, the land communication between Paris and Holland was relatively safe, compared to the trans-Atlantic conveyance, which was subject to interception by the British navy.
Then, Franklin may have forwarded the instructions in code to Jay in Spain, and received the decipherment by land. But I have not found supporting correspondence in franklinpapers.org. For example, the cover letter of Huntington dated 19 June 1781 says, "I have also herewith enclosed, Instructions (in Cyphers) for your Government, in Addition to those formerly given Mr Adams for negotiating Peace." and also "You will please to communicate immediately to Mr Adams & Mr Jay the Receipt of these important Dispatches." but nothing about decoding.