The present article is mainly based on Bergenroth's groundbreaking work of 1862. Bergenroth succeeded in code breaking of various different ciphers, including very complex ones. Although he described the ciphers in some detail, the cipher per se was not his primary interest. Thus, it would be not without worth to collect and streamline descriptions throughout the volume about the ciphers used in the correspondence of the Catholic Monarchs.
Cipher was introduced to Spain by Miguel Perez Almazan. He started his career as an assistant secretary to Fernan Alvarez and soon rose to the place of Secretary, then to that of First Secretary of State (Bergenroth p.xvii-xviii). His taking over the conduct of business from the aged Alvarez brought about a great improvement in style and writing of documents (ibid. p.xii).
No cipher despatches earlier than 1480 were found in Bergenroth's research and even then the ciphers were rather primitive until about 1495. In modern parlance, they were a code rather than a cipher: Roman numerals represented words, mainly names of persons and countries. The vocabulary was rarely more than 500. (memorial p.79)
Some twenty different ciphers were identified by Bergenroth. It was only after this was achieved that one complete key and fragments of two others were discovered. The complete key was the one which had been the most used in the extensive correspondence between Doctor De Puebla, ambassador in England, and the Spanish government (Bergenroth. p.xiii). In the following, the latter cipher is referred to as "De Puebla's great cipher" for convenience's sake.
One early cipher consisted of Roman numerals representing a small number of names and words. It was used during De Puebla's first mission in England. De Puebla was sent to England late in the year 1487 or early in the year 1488 to negotiate a marriage between Catherine of Aragon and Arthur, Prince of Wales (Bergenroth p.xix). The cipher is used in a letter from De Puebla to Ferdinand and Isabella, dated 15 July 1488.
xi the King of England
xii the Queen of England
xix the King of Portugal
xvii the King of France
xxi the alliance
xxxv De Puebla
(Needless to say, the original letter was written in Spanish, though, for this particular case, a partially ciphered paragraph is a Latin translation of a clause of De Puebla's instructions that he prepared because the English could not understand or speak Spanish.)
Other instances below from the same year (memorial p.79) may or may not belong to the same cipher.
|iv the King of the Romans||xxiii Duke of York or Perkin Warbeck|
De Puebla was temporarily returned to Spain in 1489 but it appears he still represented Spain at the English court in 1491 (Bergenroth p.xx, cf. p.20; Burke p.158).
At the time of the last efforts towards the conquest of Granada to complete the Reconquista, Queen Isabella wrote to De Puebla "ambassador in England" on 26 May 1491 and told him to explain their situation to Henry VII.
A numerical cipher used in this letter has the following vocabulary (Bergenroth's conjecture based solely on the general meaning of the context).
3 their Highnesses|
(i.e., Ferdinand and Isabella)
4 the King
8 the King of France
9 Madame de Bourbon?
10 the King of England
42 at present
78 the Duchess of Brittany
81 the treaty
82 in Brittany
188 the King of the Romans
In November 1494, De Puebla returned to his embassy in London for a renewed marriage negotiation.
The cipher used in a letter from De Puebla to Ferdinand and Isabella, dated 19 July 1495, includes the following vocabulary.
xvi el Rey|
lvi.ci vuestras altezas si
cccxi el Cardinal
cccxxiiii dispojo de la Yglesia
cccxxv violencia del Papa
cccxxvi restituir al Papa lo suyo
High frequency of numbers above 100 is impressive, though this is a small number compared to the volume the Spanish cipher would soon attain.
The contemporary deciphering of this letter is confusing at places and Bergenroth observes that there seems to be errors in the ciphertext in more than one respect.
De Puebla was urged by his masters to report in common writing, not cipher. Considering the continued use of cipher in these despatches from the Catholic monarchs, the problem was not in the cipher itself.
Bergenroth says letters entirely in cipher first occur in the year 1495 and are composed of Roman numerals (p.cxxxvii-cxxxviii). He may be referring to these letters.
There is a partially ciphered letter of the same date, 28 December 1495 (Bergenroth p.75-77). Though written in the form of a commercial letter, it was written in the hand of Alvarez and was actually an instruction to De Puebla. Apart from the codes listed below, the letter refers to the King and Queen of Spain as "the Directors of the Company" and their ambassadors as "factors." The cipher used was one used in a great many other documents and was deciphered by Bergenroth. Apparently, the vocabulary exceeds 2000 elements. In the following, this cipher is called "Alvarez's Great Cipher" for convenience's sake.
DCCCLV De Puebla
DCCCXCIIII between England
DCCCCXXVII the Constable of Navarra
DCCCCLII the marriage of Prince Arthur
DCCCCLIII marriage of Princess Katharine
As used herein, De Puebla's Great Cipher refers to the only one cipher for which a complete key is extant and it was the most used one in the extensive correspondence between Doctor De Puebla and the Spanish government (ibid. p.xiii). Although Alvarez's Great Cipher was also used in a great many documents, its key is not extant.
Although Bergenroth's descriptions are tantalizingly varying, the boldfaced references below may all refer to De Puebla's Great Cipher. (Someone who has access to the Archives is kindly asked to check the manuscripts.)
As seen in the above table, from 1496, combined use of two ciphers became common. In addition to the Roman numerals, a cipher alphabet composed of symbols that represent letters of the alphabet was used. In the first instance, each vowel was represented by five different symbols and each consonant by four. (Bergenroth p.cxxxviii)
The number of signs for each letter very soon increased to thirteen and fourteen. One instance of 1498 had more than seventy signs for twenty letters (memorial sketch p.78).
In addition to these, monosyllabic codes came into use. For example, "bax" represented "ciertamente", "dem" represented "gente de armes," "ham" represented "Yo, el Rey Catolico," and so on (Bergenroth p.cxxxviii). Other examples are "hep" for "Dios" and "hip" for "Diablo" (Bergenroth p.xlvi). Further examples are "la" for "Venetin," "dam" for "war", "hop" for "King of England", "feb" for "England", "fib" for "Englishmen," "cov" for "Scotchmen" (memorial sketch p.78)
Further, nulls, that is, signs without meaning, were interspersed in the ciphertext. Nulls may be some symbols similar to those representing letters or monosyllables. In some cases, nulls were words in plain writing such as "Semper ille Cesar" or "Je vous prie" or any other words or phrases of any other language but generally different from the language in which the letter itself was written. (Bergenroth p.cxxxviii)
Ordinary words of some language were not only used for nulls. In a cipher of Don Pedro de Ayala, "etiam" represented "ll" and "malus" represented "rr". (Bergenroth p.cxl)
De Puebla also used code words for representing names and keywords. A letter addressed to him in June 1489 included a passage in code:
Different systems may be used in combination to represent even one word. For example, "DCCCCLXVIIII le N o γ malus ζ" represents a word "enviando" (sending): DCCCCLXVIIII(en) le(vi) N(a) o(n) γ(d) malus(-) ζ(o) (Bergenroth p.cxxxviii). In this example, a Roman numeral, a monosyllabic code, symbols (in this case approximated by Latin and Greek characters for typographical reasons) and a null are all used to represent one word.
The following example Bergenroth used in explaining his feat of deciphering is probably an actual instance: "Cox Ω MDCIX Δ" where "cox" represents "Po", Ω represents "d", MDCIX represents "rey" (king), and Δ represents "s", all amounting to one word "podreys" (you will be able). (Bergenroth p.cxliv)
The "great key of Latin numbers used by De Puebla", probably referring to De Puebla's Great Cipher or Alvarez's Great Cipher, uses the following Roman numerals to represent numbers (Bergenroth p.cxlii):
MMCCCLXXIII -- 1|
MMCCCLXXIV -- 2
MMCCCLXXV -- 3
MMCCCLXXVI -- 4
MMCCCLXXVII -- 5
MMCCCLXXVIII -- 6
MMCCCLXXIX -- 7
MMCCCLXXX -- 8
MMCCCLXXXI -- 9
MMCCCLXXXII -- 10|
MMCCCLXXXIII -- 20
MMCCCLXXXIV -- 30
MMCCCLXXXV -- 40
MMCCCLXXXVI -- 50
MMCCCLXXXVII -- 60
MMCCCLXXXVIII -- 70
MMCCCLXXXIX -- 80
MMCCCXC -- 90
MMCCCXCI -- 100|
MMMCCCC -- 1000
Of course, symbols representing letters of the alphabet may also be used to spell any number.
In addition to these, letters of the Latin alphabet may represent numbers. For example, "i" represents "1"; "y", "u", "n" represents "2"; and "m" represents "3". It is only the number of vertical strokes that counts. Thus, "y m" represents "5" (i.e., 2+3). Ordinary Roman numerals such as "x" (10), "L" (50), and "C" (100) were also used. (Bergenroth p.cxxxix)
Parisi (2004) deals with cipher correspondence in 1495-1948 between King Ferdinand and Escrivà de Romaní i Ram, Spanish ambassador in Naples. This work in Spanish seems to be an achievement from the author's cryptanalysis based on ten cryptograms. It lists two sets of ciphers, both consisting of monosyllabic codes (nomenclator) and a symbol alphabet. Symbols used are similar to those illustrated in the next section, albeit with different assignment.
Some examples of the monosyllabic codes are as follows (see Parisi(2004) for more).
xud en mucho servicio
xer [vuestras altezas]|
The paper also prints many images of the original cipher letters: two pages of 11 May 1495 (Cipher No.1); a page of 26 June 1495 (Cipher No.1); a page of 4 July 1497 (Cipher No.1); and six pages of 31 August 1498 (Cipher No.2). Of these, the first three letters use both symbols and monosyllabic codes. On the other hand, the last letter is written almost exclusively in symbols and the ten lines at the end of the letter are written in a different system, though consisting of similar symbols, and are left undeciphered by Parisi,
A facsimile of a two-page letter from Ferdinand to Gonzalo Fernández de Córdoba [Gonzalve de Cordoue], known as the Great Captain (Gran Capitan) (Wikipedia), dated 16 August 1502, is printed in Speziali (1955). In this letter, a symbol alphabet, monosyllabic codes, and nulls (one symbol and Latin words "nemo" and "animabus") are used.
The symbol alphabet is as follows. (Some variant styles are included for reference's sake.)
Some examples of the monosyllabic codes are as follows (see Speziali (1955) for more).
xap da, d-
xeq de lo
xiq de la
It appears the King and the general used a newer cipher (see the next section) as early as May 1502, which is before the above cipher.
A ciphertext (1496) sent by the Great Captain is reproduced in Arturo Quirantes' blog. Some symbols (Q, U, E) are similar to the above.
In 2018, another cipher used in letters from Ferdinand to the Great Captain (27 May 1502, 14 April 1506) was identified by CNI (the National Intelligence Center of Spain) (ABC; El Mundo; Jesus Garcia Calero and Juan Fernandez-Miranda, "El CNI descifra uno de los grandes misterios de la historia de España" (1, 2)). (I thank Sergio Miguel for letting me know of this news.) The decipherment of a few sentences on the manuscript led to the discovery of the 88 symbols and 237 letter-combination codes. In one letter, King reproached the general for his contacting the King of the Romans (Wikipedia) without his sanction (see ABC2 for more).
Arturo Quirantes noticed the cipher is the same as one discovered by Bergenroth in the 19th century (his blog). Although there are symbols that do not match, it is understandable that both versions, being reconstructions, are incomplete. According to his appraisal of the cipher in historical context, this was the first Spanish cipher to employ a cancellation sign. In addition, the nomenclature introduced some irregularity in the arrangement, though it is not a full two-part code.
Galende Díaz (2006), an article in Spanish, which seems to deal with classic code breaking techniques for encrypted Spanish text, prints a manuscript cipher between Don Juan Manuel and the Marquis of Villena. It consists of a symbol alphabet (including double letters ll and rr), several monosyllabic codes such as "daz", "da" and some null words such as "magno."
Letters of Almazan often told an ambassador that he had changed the cipher and the old one was no longer to be used (Bergenroth p.cxxxix), though few of such communications are calendared. In July 1498, the Knight Commander Sanchez de Londoño visited England on his way to his embassy in Flanders (Bergenroth p.xx-xxi). It appears he brought a fresh cipher for De Puebla.
Mattingly says De Puebla held two ciphers of his own, and one in common with the Netherlands embassy between 1496 and 1507. The latter may be the one brought by Londoño. (Some ciphered letters of Londoño and his colleague, Sub-prior of Santa Cruz, are calendared. Bergenroth p.182-183, p.199-202, and Supplement p.54.))
This marks the end of the use of De Puebla's Great Cipher (with a possible exception of the letter of 5 December 1504, which is marked "in two keys of cipher; only one of which is extant" (Bergenroth p.345)).
De Puebla's letters to Ferdinand and Isabella, dated 27 August 1498, 25 September 1498, 11 January 1500, 16 June 1500, and as late as 15 April 1507 and 7 September 1507 used the following codes (Bergenroth p.194, 197, 213, 226, 228, 410, 429). The codes 881, 888, 889, and 890 collate well with Alvarez's Great Cipher. But since the latter cipher was already used by De Puebla in 1495, this may not be the new cipher brought by Londoño. The code "CCCCLXXXVIII with" is different from the code "CCCLXXXVIII with" in Alvarez's. "The King of England" is represented by DCCCLXXXVII and DCCCLXXXVIII. It may be such slight discrepancies that left Bergenroth in doubt as to whether two keys are really the same key (Bergenroth p.xiii).
DCCCLXXV Ferdinand and Isabella
DCCCLXXVIII the Queen of Castile
DCCCLXXXI the King of France
DCCCLXXXVII the King of England
DCCCLXXXVIII the King of England
DCCCLXXXIX the Queen of England
DCCCXC the Prince of Wales
DCCCCXXI the King of Scots
DCCCCLXXVIIII an embassy
MCCXVIII Archduchess Margaret of Austria|
MDLXXXIIII Muy altos y muy poderosos Senors Rey y Reyna, &c. (i.e., Ferdinand and Isabella)
MMCCCXXXI De Puebla
In this period, letters between De Puebla and the Catholic monarchs of 25 February 1500, 25 July 1500, 13 August 1500, 3 October 1500, 29 July 1501, 10 August 1502, 29 June or July 1506 used two keys of ciphers, neither of which is extant. That is, neither of these is De Puebla's Great Cipher.
There was a rival ambassador at the English court, Don Pedro de Ayala. His several cipher letters of 1499 and 1500 (Bergenroth p.205, 216, 217, 218) and one in 1501 (Supplement, p.1) are calendared.
Catherine of Aragon married Prince Arthur in 1501 but Arthur suddenly died in 1502. Ferdinand and Isabella sent the Duke of Estrada as an ambassador, who was empowered to arrange a marriage between Catherine and the new Prince of Wales, later Henry VIII.
Estrada's letters of 14 June 1502, 24 January 1504 and many others use two ciphers, for which only a fragment of the key for one is extant. That is, Estrada had a cipher different from De Puebla's. Further, the cipher used in a short letter from Isabella to Estrada, of 20 August 1503, was not used in other letters extant and is the only cipher Bergenroth could not decipher (p.309, xii). Thus, Estrada used at least three ciphers.
Spanish cipher had developed into so complicated a system that some of the deciphered despatches have marginal notes such as "Nonsense", "Impossible", "Cannot be understood," or "Order the ambassador to send another despatch." (Apparently, these remarks are not calendared.) After the death of Queen Isabella in 1504, it was found necessary to return to a simpler cipher (Bergenroth p.xii).
Catherine of Aragon, who had been only a girl when she came to England as a bride for Prince Arthur, had to fend for herself for years and was accredited as an official ambassador to negotiate with the English court. She learned ciphering in 1507. After she married Henry VIII, Ferdinand relied on cipher correspondence with Catherine in communicating sensitive matters with his son-in-law. (See another article.)
Ferdinand the Catholic died in 1516. Shortly before his death, he updated the cipher for Bernardino De Mesa, ambassador in England (1515-1518, 1519-1523).
Galende Díaz (2006) prints a piece of a cipher letter of Charles I (Emperor Charles V) of 1523 to his ambassador, from which it appears that the basic form of cipher consisting of symbols and monosyllabic codes was inherited into the next reign.
Bergenroth, G.A. (ed.) (1862), Calendar of State Papers, Spain (Google: Vol. 1, Vol. 2, Supplement, Further Supplement; British History Online), in particular, "Remarks on the Ciphered Despatches in the Archives at Simancas", Vol. 1, p.cxxxvii *The copy at Google lacks pages after p.464 (7 August 1508). (Vol. 1 is simply referred to as "Bergenroth" hereinabove.)
Cartwright, William Cornwallis (1870), Gustave Bergenroth: a memorial sketch (Internet Archive, Google) This reprints "Remarks on the Ciphered Despatches in the Archives at Simancas" in the appendix. (Simply "memorial" hereinabove.)
Ivan Parisi (2004), 'La Correspondencia Cifrada entre el Rey Fernando el Católico y el Embajador Joan Escriv è de Romaní i Ram,' Pedralbes, 24, pp.55-116 (PDF) (in Spanish)
Pierre Speziali (1955), 'Aspects de la Cryptographie au XVIe Siecle', Bibliotheque d'Humanisme et Renaissance, Tome XVII
Juan Carlos Galende Díaz (2006), 'Basic Concepts of the Cryptology: The Manuscript of the Biblioteca Nacional,' Documenta & Instrumenta, 4, pp.47-59 (PDF) (in Spanish)
Garrett Mattingly (1955), Renaissance Diplomacy (Internet Archive) p.274 n.137
Ulick Ralph Burke, History of Spain from the Earliest Times to the Death of Ferdinand the Catholic 2nd ed. Vol 2 (1900) (Internet Archive)