When the Revolutionary War broke out in Lexington and Concord in April 1775, Benjamin Franklin, almost 70, had already achieved an international fame as a scientist. When the Committee of Secret Correspondence was set up in November to seek foreign connections, Franklin was its most prominent member. To begin with, Franklin wrote to his correspondent Charles William Frederick Dumas [Charles Guillaumes Frédéric Dumas], a scholar residing in the Netherlands and an ardent supporter of the American cause, and asked him to serve as an American agent.
Dumas accepted the offer and sent a cipher to Franklin in his letter dated 30 April 1776. Dumas' cipher was a passage cipher in which a 682-letter passage taken from Dumas' own writing in French served as the key. Each letter of the passage was assigned numbers 1-682. Thus, a letter can be enciphered in as many ways as the letter occurs in the passage. For details, see here.
Dumas' early use of this cipher includes:
Soon after the declaration of independence, Franklin was named as one of the three commissioners for obtaining French aid. Leaving Philadelphia on 26 October, he reached France on 4 December. Franklin had told Dumas to continue to report to the Committee of Secret Correspondence (24 October). The Committee, now consisting of Robert Morris and three others, told Dumas that they took over a cipher from Franklin.
Thus, Dumas continued to write to America, while he corresponded with Franklin in Paris. In the mean time, the Committee of Secret Correspondence was renamed the Committee of Foreign Affairs as of 17 April 1777. Dumas used the cipher in the following letters.
This cipher was also used to explain Lovell's cipher and Franklin would repeatedly express his preference of this cipher, as discussed below. It is noted that Franklin used this cipher as late as in 1781:
Franklin's fame was such that a collection of his works went into a fourth edition in 1769, which was translated into French by Jacques Barbeu-Dubourg in 1773.
Barbeu-Dubourg, in his despatch written during 10 June to 2 July 1776, sent Franklin a cipher, which was probably a passage cipher similar to Dumas'.
The above quotation is followed by an exemplary deciphering as seen in the image below.
Inspection of this example reveals the following correspondence:
The first six letters suggest that Barbeu-Dubourg's cipher was also a passage cipher like Dumas'.
When Franklin reached the French shore on 4 December 1776, he addressed his first letter to Barbeu-Dubourg: "My dear friend will be quite surprised to receive a letter from me dated in France ...." (Franklin to Barbeu-Dubourg, 4 December 1776, back-translated from French by the author)
Dumas, Franklin, and John Paul Jones, the naval commander, used a one-part code WE003, which was designed in Europe, possibly by Dumas (Weber, p. 79). Judging from the timing, this seems to have replaced Dumas' passage cipher WE041, which Dumas apparently ceased to use after June 1779. Jones may have referred to this when he wrote to Franklin on 25 July, "Before I depart I will send you a Cypher for a private Correspondence".
This is a code list assigning numerical codes of 1(Aberdeen)-928(Your) to words. Basically, the words are arranged alphabetically ("one-part code"), though some blocks are swapped.
Since WE003 did not have codes assigned for single letters or syllables, a substitution alphabet was attached for use in enciphering words or names not in the list.
The news of the victory of Saratoga allowed Franklin to bring France into open alliance with the Americans in 1778. In 1779, Jones was given command of the Bonhomme Richard, a converted merchantman renamed in honor of Franklin after his old publication. Jones set sail on 14 August 1779 from Lorient and won a victory over the British Serapis on 23 September. Jones lost the Bonhomme Richard and transferred his flag to the Alliance and arrived at a neutral port in Holland on 3 October.
A letter written by Dumas probably shortly after this used WE003 (see here).
Apparently, a new cipher was introduced between Dumas and Franklin in early 1780. Dumas' letter to Franklin on 8 February 1780 includes a passage partly in cipher, as shown below.
This scheme (WE004) was almost identical with WE003 but codes were offset by 33. Thus, instead of 1(Aberdeen)-928(Your), it ranged from 34(Aberdeen) to 961(Your).
Apparently, the codes 1-33 were assigned to some names and grammatical endings. The following may be extracted from the existing letters:
In subsequent correspondences between Franklin and Dumas, some key words were occasionally enciphered with this scheme. Names and words not listed in the table were enciphered with a substitution table (e.g., "Brunswick", "Monsieur", "between", "state").
Sometimes, a code was followed by an ending enciphered with the substitution table (e.g., "598<north>PBX<ern>", "657<pention>LBI<ary>" in 2 March 1780).
While word endings might be indicated with specific codes (e.g., "26", "27", "28" in 2 March 1780) or enciphered with the substitution table, they were often just implied. At times, they were simply spelled in plaintext (e.g., 3 August 1780). Further, adjectival forms were often represented by the code for the noun. For example, the word "French" was not in the list and code 337 for "France" was substituted. Similarly, 783<Russia> was used to represent "Russian".
Dumas usually wrote in French but the vocabulary of WE004 was in English. While codes for names as well as codes such as 196<Congress> and 78<arrive> could also be used in a French sentence (e.g., 29 May 1780), some words had to be enciphered letter by letter (e.g., 11 April 1780).
Dumas often switched to English when he used cipher in the overall French letter (e.g., 18 February 1780). At times, Dumas used English for only part of a sentence (e.g., 25 February 1780).
See here for examples of use of this cipher.
The letters between Franklin and Dumas at this period often referred to Henry Laurens, who was to sail from his hometown Charleston on a mission to Holland. However, in 1780, Charleston was attacked by the British and fell in May. As late as 26 July, Franklin thought Laurens sailed as planned. Actually, however, Laurens could not leave Philadelphia till 13 August 1780 . Moreover, his ship was captured by the British on 3 September off New Foundland and Laurens was taken prisoner.
Franklin's letter of 9 October still shows that he did not know of the fact. Dumas told it to Franklin on 10 October. "You will have already learned of the misfortune that befell Mr. Laurens of being taken with his secretary and taken to England. I have written to him, by a sure hand, a letter that offers him the services which he might consider I am capable of doing him." (the author's translation from French)
From this letter on, Dumas stopped using WE004. (According to David R. Chesnutt and C. James Taylor, The Papers of Henry Laurens, pp. 526-528, Henry Laurens' use of cipher is only extant in letters to James Bourdieu. This is different from WE004, assigning "45" to "peace commission".) Apparently, there was no definite agreement for the disuse and Franklin used WE004 at least on two subsequent occasions (3 December 1780 and 18-20 January 1781).
Back in 1777, James Lovell, who had joined the Committee of Secret Correspondence after Franklin left for France, sent Franklin in Paris a polyalphabetic cipher he gained "by accident". (The text of the undated letter is here. Franklin acknowledged receipt of it on 17 October 1777.)
Lovell described a 27x27 alphabet square as follows. (Lovell included the ampersand (&) in his regular alphabet.)
He gave an example of enciphering with this table.
Some elaboration would be desired to supplement his typically terse explanation: "each letter of the Alphabet will be referable to every figure of the 27".
The text indicates a nonce keyword "chardon". Each of the letters in this keyword is used in turn to encipher respective letters of the plaintext "Powder & Ball" "Cannon", and "May".
First, to encipher the letter "P" with the keyletter "C", one would look at the column headed by the letter "C", i.e., the third column from the left. Then, looking down the column, the plaintext letter "P" is located. The figure on the left side of the table, 14, is used for "P" in the cipher.
Next, to encipher the letter "O", another column of the alphabet is consulted (thus, "polyalphabetic"). Now, in the column headed by "H" (i.e., the second letter of the keyword CHARDON), the plaintext letter is found on row 8. Thus, the "O" is enciphered as "8".
The remaining letters would be similarly enciphered. It is noted there is one error in Lovell's enciphering (or at least in the text the author used). While BALL is enciphered as 27.21.12.2., the second "L", being enciphered with the keyletter "R", should be "22" instead of "2".
Lovell advised Franklin to use a new keyword if the letter arrived "under dubious circumstances". The full 27x27 alphabetic square allows enciphering with any keyword.
The table for this particular CHARDON cipher would be as follows.
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27
C D E F G H I J K L M N O P Q R S T U V W X Y Z & A B
H I J K L M N O P Q R S T U V W X Y Z & A B C D E F G
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 &
R S T U V W X Y Z & A B C D E F G H I J K L M N O P Q
D E F G H I J K L M N O P Q R S T U V W X Y Z & A B C
O P Q R S T U V W X Y Z & A B C D E F G H I J K L M N
N O P Q R S T U V W X Y Z & A B C D E F G H I J K L M
The correspondence between the cipher and the plaintext is as follows.
Lovell used the polyalphabetic system with his various correspondents and the scheme was often called Lovell's cipher. Lovell took care to use different keywords with different correspondents. Franklin was given keyword COR. Other keywords (WE042 to WE052) include CR for John Adams (see here (in Japanese)), YO for Henry Laurens (as noted above, actual use of this is not known), BY for John Jay (see here), and UNT for William Palfrey.
The table for the COR cipher would be as follows.
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27
C D E F G H I J K L M N O P Q R S T U V W X Y Z & A B
O P Q R S T U V W X Y Z & A B C D E F G H I J K L M N
R S T U V W X Y Z & A B C D E F G H I J K L M N O P Q
PCC, Roll 72, Page 142 explains the Lovell scheme and permutation of Franklin's keyword "COR" in Dumas' passage cipher (WE041).
About the same time Dumas started to use WE004 to Franklin, James Lovell urged Franklin to use cipher in his correspondence. Lovell, in his part, was prompted by Luzerne, French minister at Philadelphia.
From the first, Franklin just ignored Lovell's cipher. But Lovell sent him a letter in his cipher.
Franklin tried to decipher this on the reverse side of the address sheet.
The two phrases should read "may not boast" and "the merchant & farmer" but Franklin could not make it out. (In the above, highlighted letters are correct plaintext letters.)
In the first phrase, Lovell made an error when enciphering "o" in "boast". Following "b" enciphered with the C-row (first line) in the COR cipher table above, he should have used the O-row (second line) but he used the C-row again. With hindsight, Franklin correctly deciphered all but one letter but Franklin was at a loss and wrote other candidates below.
For the second phrase, Franklin got right the first three letters 3(t) 6(h) 18(e) but somehow misapplied his table and put "y" for 23. After this, he switched his way of deciphering and just wrote down the candidate letters below each cipher letter.
After all his effort, Franklin could not decipher this cipher.
Here, Franklin expressed his preference of the cipher he left with Robert Morris when he left the Committee of Secret Correspondence (see above), that is, Dumas' passage cipher WE041.
While the above letter did not reach Lovell, Franklin communicated his trouble to Francis Dana, to whom he sent an excerpt of Lovell's letter and the instructions to the COR cipher in Lovell's hand. Francis Dana had been sent to Europe in 1779 as a secretary to John Adams and stayed in Paris during December 1780 to April 1781, before his new mission to the court of St Petersburg.
Actually, Dana had been troubled with Lovell's cipher himself. Dana had just received a letter dated 6 January 1781 from Lovell but could not read it because he could not recall until some days afterwards the name used for a key (BRA) Lovell hinted in his letter. He may have talked about Lovell's enigmatic cipher over dinner with Franklin and offered to make a try.
Dana reported Franklin's letter to John Adams and promised to send a copy of the ciphers by a private opportunity. Dana wondered if he could read the ciphers and only said "I will make the attempt."
The copy in Adams Papers has plaintext letters written over cipher numbers but it does not mean deciphering was successful, for every plaintext letters for the second phrase is placed on a wrong number! (Weber p.32, 43; Helen Jones p.460)
(In addition to Lovell's own mistake, Franklin's excerpts erroneously put "36." for "3.6.") The correct reading would be as follows.
In June 1781, Congress resolved to appoint Franklin and others as peace commissioners in addition to John Adams. Peace instructions were sent to Adams, enciphered with keyword CR, but Adams could not decipher it. On the other hand, Franklin had no trouble in deciphering his.
This does not mean, however, that Franklin, being a scientist, was more adept in cipher than Adams. While one authority considers the cipher referred to was Lovell's cipher based on the keyword CR (WE042) (Weber, p.111, n.25), there is reason to believe otherwise. Franklin later confessed his preference of the cipher "in which those instructions were written that relate to the future peace", as discussed below.
In July 1781, Robert Morris told Franklin that he took office of Superintendent of Finances and sent Franklin new code tables.
Actually, Morris failed to enclose the cipher and sent it with another letter the next day.
Franklin acknowledged receipt of the duplicates on 12 September and Major Franks, who bore the originals, arrived thereafter. Franklin, however, did not like the cipher sent by Morris.
The "old one" Franklin mentioned is probably Dumas' passage cipher (WE041), which Franklin left with the Committee of Secret Correspondence, of which Morris was a member, when he left America.
Robert R. Livingston took oath as the first Secretary for Foreign Affairs for the United States on 20 October and reported it to Franklin the same day. He would take over the task performed by Lovell of the Committee of Foreign Affairs.
The Americans had just gained an important victory at Yorktown over the British in October 1781. General Lafayette obtained leave to visit his family in France in November and took Livingston's letter of 26 November to Franklin.
On 16 December, Livingston informed Franklin that, of the two ciphers No. 3 and No. 4 sent to Franklin, he would use No. 4.
WE008 was a code list of about 660 elements. Though the list is smaller than Dumas' WE004, it is more difficult to break because of the random mapping of code numbers (i.e., two-part code), in contrast to the alphabetical ordering of WE004 (i.e., one-part code). To the original list, Livingston added 16 codes in a postscript to a letter to Franklin dated 13 February 1782.
Franklin, however, repeated his preference of Dumas' cipher to Livingston.
After receiving Livingston's letter of 13 February, Franklin said again:
Thus, Franklin repeatedly expressed his preference of "the old one ..., which I [Franklin] left with you [Robert Morris]", "that in which those Instructions were written, that relate to the future Peace", and "the old one of Dumas", that is, Dumas' passage cipher (WE041).
In June 1782, John Jay, who had been working at the court of Spain, joined Franklin in Paris for the last stages of peace negotiations. In a letter of 8 August to Jay, Livingston used "Doctor Franklin's cipher" (WE008). Livingston enclosed the resolution of Congress of 7 August encoded in both WE007 and WE008. (Originally, WE007 had been sent to Jay but Jay reported that it did not arrive safe. See here for details.)
On 9 August, Livingston used WE008 in his letter to Franklin and expressly disapproved of Franklin's preferred cipher.
However, Livingston knew that Dumas' cipher was easier to handle. When he knew John Adams was troubled by Lovell's cipher, he had advised to use Dumas'.
Anyway, to Franklin, Livingston continued to use WE008 (see here).
In June 1783, Livingston resigned and Elias Boudinot temporarily took over correspondence with foreign ministers.
Franklin replied three months later.
Probably, Boudinot did not know the code WE008 used by Livingston was Morris'.
About the time Livingston was writing his disapproval of Franklin's preferred cipher, Franklin on his part used a code in his letter of 12 August to Livingston (see PCC, Roll 127, Page 286 (clean transcript), PCC, Roll 108, Page 207). Somehow, this is different from WE008 of Livingston's recommendation and from WE004 between Franklin and Dumas, still less Dumas' passage cipher WE041.
The translation of the coded portion is known to be as follows (shown in braces { }).
Using repeated letters and syllables as clues, the plaintext may be speculatively mapped to the codes as follows.
Assignment in parentheses ( ) are considered probable. Five occurrences of 167(th), four of 109(y), four of 481(i), three of 268(si), etc. comfortably fit the plaintext. Codes sandwiched between such highly probable guesses may also be deemed safe.
Assignment in brackets [ ] are tentative. Separation "navy-ga-tion" may seem artificial but there are examples for such separation (e.g. Livingston's letter in WE007 and Adams' letter in WE013).
Assignment in angle brackets < > are dubious. 440<coop> is classified as dubious because "coop" seems hardly worth listing in a code list of less than 700 elements. 187<en> and 64<en> are dubious because of collision.
It is remembered that Morris sent Franklin two codes: No. 3 and No. 4 and Franklin apparently received both. Though Livingston chose to use No. 4, Franklin may have taken up No. 3 for encoding this passage.
Morris' Code No. 3 is considered to have been prepared on the same printed template as No. 4 with different mapping (see here). Use of the same template is supported by (i) presence of code numbers for all the words and syllables of the assignment including "hope", "fre", "navy", "tion" (but excepting 440<coop>), (ii) lack of a code number for "my" (cf. 599(m) 109(y)), and (iii) lack of a separate code number for "V" from "U" (cf. 170[ri] 399(u) 250[er]).