Although John Jay had been involved with various forms of secret writing since the Revolution began, notably distributing his brother Sir James Jay’s formulation for invisible ink,1 his departure for Europe in 1779 occasioned an intensified interest in cryptography, an enthusiasm he shared with many of his contemporaries. Frequent interception of mail meant that encrypted correspondence, and the discussion of codes and ciphers in correspondence, occupied a notable amount of Jay’s attention in Spain and France until just before the peace treaty in 1783.2 Jay’s correspondence with Robert R. Livingston, Silas Deane, William Carmichael, William Bingham, the various presidents of Congress, Charles Thomson (the secretary of Congress), the members of the Office of Finance, and key members of Congress’s Committee for Foreign Affairs, often had encrypted passages. Later, as secretary for foreign affairs, Jay shared codes with diplomats John Adams and Thomas Jefferson and, when negotiating with Britain in the 1790s, with Secretary of State Edmund Randolph. Jay also discussed the use of a complex code with Philip Schuyler.3 Private personal and family codes were employed in Jay’s correspondence with Catharine W. Livingston, Frederick Jay, Henry Brockholst Livingston, Lewis Littlepage, and others. Many of these coded passages went undecoded in early editions of diplomatic papers and remained so until the publication of Jay’s papers in modern editions.4

To understand these documents, some knowledge of the state of the art of cryptography at this time and of the relevant terminology is helpful. Jay and his contemporaries did not distinguish codes and ciphers, referring to both as “cyphers.” However, it is useful to understand the differences between these two types of encryption and the proper terminology involved, since annotation of the documents relies on this information to elucidate the state of the documents and the means of making them legible for this edition.5

First, a definition of basic cryptographic terms: When a passage is not encrypted, it is referred to as “plaintext.” When a passage is encrypted, generally either a cipher or code is used. A “cipher” features the substitution or transposition of individual letters according to a chart or algorithm. A “code” substitutes whole words and phrases, parts of words, and individual letters according to a particular key.6

Jay’s correspondence employed four general categories of ciphers and codes (in chronological order): simple substitution ciphers, book codes, polyalphabetic substitution ciphers, and nomenclator codes. It was not uncommon for hybrids of these to be produced—a cipher on occasion would be accompanied by a brief list of code symbols for frequently used words and names to speed up the job of encryption; likewise, a code might include a simple cipher to accommodate unlisted terms.

“Simple substitution ciphers” substitute a single symbol for each letter of the alphabet. These symbols can be numbers, letters, or other figures.

“Book codes” are based on commonly available books, usually dictionaries. A word or letter is substituted by the coordinates of the word or letter within the book: page and line numbers, and indications of columns within a page, if this applies to the book in question. Often the numerical values were offset to make code breaking more difficult. For example, Jay instructed Robert R. Livingston to count down the columns in Abel Boyer’s Royal Dictionary (1771) and then add seven to the figure in their code of February 1780 (WE078).7

“Polyalphabetic ciphers” used in this period substitute one of several characters for each individual plaintext letter according to a grid of substitutions and (often) a prearranged keyword that determines the position in the grid.

“Nomenclator codes” are named for the code sheets used for encryption and decryption. These “nomenclators” are long lists of numbers, usually with six hundred or more elements, each of which substitutes for a word or part of a word. The encryption component is arranged alphabetically, and the decryption section, numerically.

In his definitive work United States Diplomatic Codes and Ciphers, 1775–1938, noted cryptologist and historian Ralph E. Weber developed a numbering system for the military and diplomatic codes he identified. Whenever possible, these will be applied in discussion of the codes used in Jay’s documents in this editorial note and in the notes to the encoded documents appearing in this volume and, to a lesser extent, in other volumes of this series.

Jay prearranged codes and ciphers with his correspondents while still in Philadelphia in 1779, as evinced by the complaint of a lost “cypher” in his letter to Robert R. Livingston dated 23 May 1780 (below). The first cipher to appear in the Jay Papers was devised while he was on board the Confederacy about to depart for Europe from Reedy Island in October 1779. As a last-minute thought, he devised a simple substitution cipher for private communication with Robert R. Livingston. The numbers 1 through 26 randomly substitute for the letters of the Roman alphabet. After this point it was not uncommon for Jay to use such simple ciphers in private correspondence. Jay’s code memorandum (EJ: 7590) records a similar cipher used with Frederick Jay. Although Robert R. Livingston almost immediately noted the inherent insecurity of a simple cipher for international communication, the cipher did find some use as a means of communicating the keys of more complex systems, especially during times of confusion. Livingston used it in his 10 February 1780 letter for lack of a better cipher, which he proposed to send. Jay would, as a consequence of this letter, propose a series of book codes.8

Although simple substitution ciphers were known to be insecure and liable to cracking by the analysis of word and letter frequencies, Jay would continue to use them on occasion, often as supplements to other schemes. In the nomenclator code he shared with Silas Deane, he included as a supplement a simple alphabet substitution for spelling out words not included on the list.9 This practice would later be replaced by improved nomenclator codes in which single letters were included in the nomenclators themselves. Jay also gave William Bingham a simple cipher (WE076) to supplement a book code (WE083), and ciphers continued in family correspondence.10

Because of the inadequacies of simple ciphers, Jay began proposing book codes in 1780. Entick’s New Spelling Dictionary (1777) had been in use in the 1770s, notably by the Lee family, and usage continued through the early 1780s. Entick’s was perhaps the most popular dictionary for cryptography during the war. Congressman James Lovell, who was one of the major figures in cryptography of the period (see below), even uncovered a code based on it used by the British command in 1781. John Adams and Francis Dana used another edition.11 Gouverneur Morris used Entick’s (1777) in early 1780, adding 23 to the page number and 96 to the line number, counted from the bottom right. Since no extant previous correspondence about this code has been found, Jay and Morris probably had agreed upon it before Jay’s departure.12

For Livingston, Jay proposed a code based on Abel Boyer’s Royal Dictionary (1771). This code proved problematic because Livingston did not have the correct edition. Jay also proposed a Boyer-based code with Charles Thomson. The scheme of this code (WE079), along with an accompanying repertory of frequent names and places (substituted by Roman numerals), can be found in Jay’s code book (EJ: 7590). Jay also encoded part of his 23 September 1780 notes on a conference with Floridablanca in a Boyer code (WE081).13 A code similar to this one was used with President of Congress Samuel Huntington, with some orthographic differences.14

Jay used codes based on Entick’s (1777) extensively during the early stages of his Spanish mission, as did William Carmichael and Charles Thomson (who would also decrypt letters for Samuel Huntington). Weber designated this group of Entick’s codes as WE080. Initially, Jay wrote to Thomson with a code using Entick’s in which the first figure was the line number plus 11; the third figure was the page number plus 181; and the middle figure signified the column and direction (c = right from the top, d = left from the top). This code is identified as WE080a in the notations. Carmichael used a simpler code with no displacement of line or page numbers, with a signifying the left column (from the top) and b signifying the right (henceforth, WE080b). When Jay wrote to Carmichael, the displacements and letters were the same as in WE080a, but he counted from the bottom of the columns (WE080c). Jay also used this scheme with Huntington but complicated matters further with the addition of a monoalphabetic cipher (WE080d). Jay expressed dislike of this system and would soon move on to other means of encryption.15

Jay also used Entick’s (1777), supplemented by a cipher, with William Bingham.16 In late 1780, Jay suggested an Entick’s code to Robert Morris (WE082). Morris did not adopt it. The Finance Office was soon to take up its own encryption standard (see below). Jay did write to Gouverneur Morris in a new Entick’s code in November 1780.17

In response to the various problems with book codes and ciphers, Robert R. Livingston devised a polyalphabetic cipher with the keyword “XZA” (WE034). When encrypting using a keyword cipher, the encoder writes the keyword over the plaintext. Each keyword letter represents a column in a grid of numerical substitutions. In this code a can be 1 (in column X), 4 (Z), or 6 (A). Jay noted that the “XZA” was entirely arbitrary. Jay suspected his mail had been examined, so he told Livingston to abandon XZA in favor of a five-column grid of his own device, designated “YESCA” (WE033; see figure 8). This keyword is also arbitrary.18

In Lovell’s keyword ciphers, the keyword itself would reveal the structure of each column. He would suggest a name or a word to a person according to some mutual knowledge unknown to the enemy, such as the name of a favorite servant. The first two or three letters of the name would be the keyword. In each column of the grid, the alphabet would proceed from the letter in the keyword. For instance, for the keyword “CR,” a would be represented as c or r, b by d or s, and so on.

With all the problems caused by the complexities of book codes and polyalphabetic ciphers, it became expedient to improve nomenclator codes. With their rationally arranged lists of word substitutions, nomenclators were easy and quick to use and fairly unambiguous.

As stated above, nomenclators were not new to Jay—he and Silas Deane used one as early as 1780. However, longer nomenclators (six hundred or more words) and more inclusive nomenclators (with letters and word particles included to account for words and names not on the list) made the added complexity of hybrid systems unnecessary. Robert Morris and Gouverneur Morris adopted several nomenclator codes, one of which was used extensively with Jay, the 660-element “Office of Finance Cipher Number 1” (WE006).19 Charles Thomson used a six-hundred-element nomenclator with Jay and Robert R. Livingston (WE007).20 Jay also employed a nomenclator with Benjamin Franklin (WE008).21

The only problems with nomenclators were making copies of the code and distributing them safely to correspondents, especially those in Europe. Early nomenclators were written out longhand, such as the Deane code found in Jay’s code memorandum. By the time of Thomson’s code and the Finance Office’s codes, preprinted forms, apparently devised by Gouverneur Morris, were being used to ease the copyist’s job. The copyist was often Gouverneur Morris himself.22 As for distribution, Jay ultimately had to depend on couriers, such as Col. David S. Franks, to hand-deliver the codes. Those that went through the mail were usually intercepted and therefore had to be considered compromised.

Although the nomenclator codes turned out to be the favored type of encryption in diplomatic circles, ciphers like Lovell’s, or Jay’s YESCA code, continued to find use, particularly when the preferred code was not available or had become suspect. Correspondence between Jay and Robert R. Livingston continued to be confused at times, with Livingston occasionally using or sending the wrong codes or modifying compromised codes (such as the Thomson code WE007, which appeared to have been examined in the mail) rather than replacing them, thus making it necessary to fall back on previous methods.23

The most confused period followed Robert R. Livingston’s appointment as secretary for foreign affairs in October 1781. Charles Thomson gave Livingston at least two nomenclator codes: the above-mentioned Thomson code, a copy of which reached Jay via Franks, and a different one given to consul William Palfrey for delivery to Jay. Both Palfrey and the code he carried were lost at sea and so never reached Jay. Livingston encrypted two letters using this lost code before discovering his error, only the second of which he resent. Therefore, his first letter to Jay as secretary remained undecrypted.24 This sort of muddle resulted in Jay and Livingston’s subsequently relying on WE008 and YESCA.

On returning to America, Jay continued to use nomenclator codes for international communication. In his capacity as secretary for foreign affairs, he shared codes with John Adams, minister to England (WE013), and Thomas Jefferson, then minister to France (Papers of Thomas Jefferson code 10).25 Adams’s reports to Congress would pass through Jay’s office, where they would be decrypted before being submitted to Congress.26 Similar practices were adopted when Jay returned to Europe as a diplomat to negotiate what became known as Jay’s Treaty.

Jay recommended that codes be deciphered on separate pages, rather than above the relevant passages, adding the instruction, in the case of Robert R. Livingston, to then “destroy the paper immediately.”27 That recommendation is one of the reasons many extant texts remained undecrypted in the public document files and were published in early editions without what were often their key passages.

A note on orthography: encrypted words and passages are indicated by small capitals. Instances of decoding that is by the editors rather than in the original source, as well as interpolations for sense, are enclosed in square brackets. The bracketed sections therefore may not have been legible to the original recipients.