Conjuration ciphers in Trithemius’s “Steganographia”
“Steganographia” by Johannes Trithemius was written in 1500 ca. and first printed in 1606. In a 1998 paper, Jim Reeds summarizes the enigmatic nature of Trithemius’s work by this question: “Is it primarily an exposition of cryptographic techniques disguised as angel magic, or is it primarily a magic work disguised as cryptography?”
Reeds’ paper discusses the numeric ciphers in the third part of Steganographia. Here I intend to focus on the much simpler ciphers which appear in the first pages of the book, disguised as magical conjurations.
While the ciphers in the third part remained a mystery for several centuries, other ciphers in Steganographia were soon disclosed in printed books, apparently based on explanations by Trithemius himself. The system I will discuss is well described in Johannis Trithemii Steganographia Vindicata (Wolfgang Ernst Heidel, 1721) p.132–133. The cipher text (Conjuratio I, first conjuration) is reproduced before the explanation (Clavis / key). The conjuration originally appeared at p.2 of the 1606 edition.
My translation of the “key”: The reader will be instructed about the single key for the conjurations. 1. The first and last words are not relevant. 2. After removing them, only the first, the third (and so on alternating) of the remaining words contribute to the message. 3. All the valid words should be concatenated. [In the resulting character sequence], the first character is irrelevant, the second one is valid, the third character is again irrelevant, the fourth valid and so on. 4. The Greek ‘Y’ stands for the Latin ‘I’. 5. From time to time, a simple ‘e’ is used instead of the diphthong Æ written in the style of the ancients. The reader can then arrange this and all the following conjurations in the described way. He will easily read the message.
After the key, the concatenation of the valid words is printed, with the meaningful characters appearing in upper case. If one reads these letters, the following message is revealed:
Sum tali cautela ut primae literae cuiuslibet dictionis secretam intentionem tuam reddant legenti.
Meaning: I am so careful that [only] the first character of each word conveys your secret message to the reader. — This describes a simple “acrostic” cipher method: each letter of the original message is represented by a whole word in the cipher message, with only the first character of each “cipher” word being meaningful.
The 1606 includes a “key” to the work as an appendix. The following fragment (p.7 of the appendix) is an example of the simple acrostic method:
… nutritor egentium, salusque tribulatorum, miserorum adiutor, nobilium triumphator, administrator gratiarum …
The passage is grammatical Latin and can be translated as: feeder of the needy, saviour of the troubled, helper of the miserable, conqueror of the proud, donor of favours…
Considering the first character of each word, the text can be deciphered as “nest mantag”, corresponding to the German Nächst Montag (Next Monday). Apparently, in order to fit the message into correct Latin, it was necessary to take some liberties with German spelling.
The conjuration cipher is in some respects more sophisticated than the “acrostic” cipher it describes, tough it has the similar problem of a low rate of meaningful vs meaningless characters (only 25% convey meaning). It also has a major flow that is not shared with the acrostic cipher: brittleness due to words and characters being selected according to position. If two words are not accurately spaced, or if a typo like skipping a single character occurs, the whole text gets very hard to recover. Apparently, this happened in the 1606 printed edition of Steganographia. If you compare the text in the 1606 edition with the text in the 1721 Steganographia Vindicata, you will see that the sixth word is “charustea” instead of the correct “charustrea”: the missing letter causes a misalignment of all the following letters, so that the irrelevant letters are selected instead of the valid ones. ‘Atloor’ for ‘atleor’ is another typo, but since the number of characters is unchanged, a single character is affected. A third typo, ‘thalmon’ instead of ‘thalmo’, accidentally corrects the shift due to ‘charustea’, restoring the correct alignment.
Ignoring other typos in the 1606 edition, the two variants produce the following clear text messages:
1721 Steganographia Vindicata:
pamersiel oShUrMi delmuson ThAfLoYn peano ChArUsTrEa melany LyAmUnTo colchan PaRoYs madyn MoErLaY bulre+ aTlEoR don mElCoUe peoloin IbUtSyL meon mYsBrEaTh alini DrIaCo person CrIsOlNaY lemon aSoSlE mydar iCoRiEl pean ThAlMo asophiel IlNoTrEoN baniel oCrImOs esteuor NaElMo besrona ThUlAoMoR fronian bElDoDrAyN bon oTaLmEsGo merofas ElNaThYn bosramoth
SUM TALY CAUTELA UT PRYME LYTERE CUIUSLYBET DICCIONYS SECRETAM INTENCIONEM TUAM REDDANT LEGENTY (sum tali cautela ut primae litterae cuiuslibet dictionis secretam intentionem tuam reddant legenti)
1606 Steganographia:
pamersiel oShUrMi delmuson ThAfLoYn peano CHARUSTEA melany lYaMuNtO colchan pArOyS madyn mOeRlAy bulre+ ATLOOR don MeLcOuE peoloin iBuTsYl meon MySbReAtH alini dRiAcO person cRiSoLnAy lemon AsOsLe mydar IcOrIeL pean THALMON asophiel IlNoTrEoN baniel oCrImOs esteuor NaElMo besrona ThUlAoMoR fronian bElDoDrAyN bon oTaLmEsGo merofas ElNaThYn bosramoth
SUM TALY CAUTAYM NO AOSORA ALOMLO EBTYMS RAHRAOR SLAAO LIOILHLO INTENCIONEM TUAM REDDANT LEGENTY (the central part of the hidden message is lost)
Not only word spacing in the cipher message is vital to the correct selection of meaningful characters, spaces must also be inserted manually in the deciphered message on the basis of the reader’s knowledge of the clear-text language. A last interesting detail about this cipher text is that the supposed conjuration includes the cross symbol, mimicking the medieval convention used to mark points in incantations where the officiant was supposed to cross himself.
Trithemius.com provides an English translation of the Latin text of Steganographia. The site can also show the decipherment of the conjurations, but since the text includes the typos that appear in early printed editions, the output is not always fully meaningful. The following are a few more conjurations with the corresponding clear-text derived from Steganographia Vindicata.
“Lamarton Anoyr” Conjuration p.2
lamarton aNoYr bulon MaDrInEl traschon EbRaSoThEa panthenon NaBrUlGeS camery iTrAsBiEl rubanthy NaDrEs calmosi OrMeNu lan YtUlEsRaBiOn hamorphin
Hidden message: NYM DIE ERSTEN BUGSTABEN DE OMNY UERBO (i.e. nim die ersten buchstaben de omni verbo)
The clear text is a mixture of German and Latin meaning “take the first character of each word”. The filler word “panthenon” exemplifies how the text was intentionally made to look like Greek.
“Padiel Aporsy” Conjuration p.8
The second chapter of the book is devoted to the powerful spirit-prince Padiel. Since the first word of each conjuration plays no role in the cipher, Trithemius was free to use it by explicitly naming the object of the spell.
padiel aPoRsY mesarpon oMeUaS peludyn mAlPrEaXo condusen VlEaRo thersephi BaYl merphon PaRoYs gebuly Ma ilthomion IlThEaR tamarson aCrImY lon pEaTa casmy ChErTiEl medony ReAbDo lasonti OaCiEl mal ArTi bulomeon AbRy pathulmon ThEoMa pathormyn
Hidden message: PRYMUS APEX VERBY PRYMI TERCIY ET CETERA DOCEAT ARTEM (i.e. primus apex verbi primi, tercii et cetera doceat artem)
Translation: “the first character of the first word, of the third etc. This will teach you the art” — The hidden message describes a variant of the acrostic cipher where the first character is only meaningful in half the words, so that the quantity of meaningless text is even lower than with the other techniques.
This conjuration is discussed in Jim Reeds’ 1998 paper Solved: The Ciphers in Book III of Trithemius’s Steganographia. Also in this case, the 1606 edition contains a typo: the two words ma ilthomion were printed as the single word “mailthomyon”. This causes the decipherer to consider irrelevant words as containing cipher text and to ignore valid words. Presumably because of this typo, Reeds could only decipher the text up to where the error occurred. Reeds’ paper was quoted in a wide number of following publications (e.g. “Investigator’s Guide to Steganography”, Kipper, 2003; “Hildegard of Bingen’s Unknown Language: An Edition, Translation, and Discussion”, Higley, 2007) which repeat the typo and only present the decipherment of the first few words of the conjuration.
“Padiel Ariel” Conjuration p.8
padiel aRiEl vanerhon ChIo tarson PhYmArTo merphon AmPrIsCo ledabarym ElSePhRoY mesarpon aMeOrSy paneryn AtLe pachumgel ThEaRaN beslonty lAs gomadyn TrIaMy mefarnothy
Hidden message: RECIPYAT APICES PRYMOS ALTERNATIM (i.e. recipiat apices primos alternatim)
Translation: “Take the first characters alternating” — The same “alternating acrostic” cipher discussed in “Padiel Aporsy”.
The conjuration itself exemplifies the great freedom provided by the low rate of meaningful characters. Six of the 21 words end with -n and five by -y, so that the whole text is also suggestive of Greek.
Trithemius could also make so that 3 of the 21 words end with the suffix -el, as appropriate for spirit names in the tradition of Christian kabbala. The word “Ariel” is the only one of the three to contribute to the cipher message, providing the first two characters of “recipiat”. Ariel also is the name of a spirit in other parts of the book (p.172).
The spirit name “Ariel” occurs in Agrippa’s “De occulta philosophia”: it could derive from Trithemius, who was one of Agrippa’s teachers at Würzburg. But it is also possible that both authors shared a common earlier source, in particular the word “ariel” appears in the Hebrew bible. In his 1951 paper The Genesis of Ariel, Stacy Johnson discusses a scholar (Abel Lefranc) who believed that Trithemius was the source of Shakespeare’s Ariel; Stacy Johnson also points out that The Geneva Bible (London, 1594) mentions the Hebrew word and appears to be a more likely source.
