up cOqqmenD 4m deD pO mpxeG qCaO …
il beccaria ha fra le altre cose …

Regensburg → Milan · homophonic substitution of Italian · 2 September 1576 · key recovered from the ciphertext

Alfonso II del Carretto to his agents, 2 September 1576

Go to the full text ↓

The marquis of Finale writes from the imperial Diet to his agents in Milan about a man sent to the King as trustworthy, in a cipher whose key was recovered from the ciphertext alone.

Daniel Bourdeau · posted

Method: key recovered from ciphertext-only · Extent: complete · On the map

Contributed by Brice Dudout (solution) · Also credited · Daniel Bourdeau (pointer to a source)

At a glanceCiphertext-only break · complete

The letter
Alfonso II del Carretto, marquis of Finale, to his agents Lodovico Allaria and Paolo Fenocchio in Milan, Regensburg, 2 September 1576, signed “il principe marchese del Final”. The first eleven lines are in cipher, the rest in clear.
Where
Biblioteca Estense Universitaria, Modena, Autografoteca Campori, Lettera D, Del Carretto, Alfonso II, c. 10 (Lodovico Media Library). Not on DECODE.
The cipher
A substitution of Italian with homophones for the vowels and Latin, Greek-like and digit shapes for the consonants; word division kept, an isolated o as a null, no nomenclator seen.
How it was deciphered
The key was broken on lines 1–8 alone by annealing with a model of 16th-century Italian, frozen, and checked on lines 9–11 against an independent machine break of those lines: 0.961 of characters against 0.263 for the best of 200 shuffled keys.
What it says
Beccaria has said that no one could carry out His Majesty’s orders better than Count Spinola; the marquis will make the truth known, though the man was sent to His Majesty as trustworthy, and “neither he nor the one behind him will reach his aim”.
Still open
Six words of the torn margin note, and 25 probable words that depart from the key by one sign. The persons named are not identified.
Credit
Brice Dudout (Europe chiffrée) transcribed the cipher, recovered and validated the key and read the letter, working with Claude Opus 5.5 agents. The letter came from Daniel Bourdeau’s catalogue (no. 2164).

01 The letter

In the late summer of 1576 the Diet of the Empire sat at Regensburg, where Maximilian II would die in October. Alfonso II del Carretto, marquis of Finale on the Ligurian coast, was there and wrote to his two agents in Milan. The upper part of c. 10r is eleven cipher lines with a torn note in the left margin and one interlinear insertion; the rest of the letter is in clear. The neighbouring c. 9, a letter of 1574 to the same agents, is in clear and mentions letters “in ziffra”. No decipherment of the time is on the leaf, and no edition was found.

02 The key and its test

The vowels have several signs each (a = D, m; e = G, O, t, l; o = C, i and two groups written oPo and oP; i = u, n; u = 7 and a dotted rho), and the consonants single shapes (h = 4, d = ezh, f = d, r = e, s = a, t = x, n = k, m = h). The key was found by simulated annealing on lines 1–8 with a four-letter model of 16th-century Italian learnt from other texts; six seeds gave the same key. One value was corrected by sense on those lines (d = f, in fra, far, difficultà), and the key was frozen.

Lines 9–11 (54 words) were fixed as the held-out part before any calculation. They were broken a second time, alone and automatically, and that independent break was frozen before the frozen key was applied to them. The two agree on 0.961 of characters, against 0.263 for the best of 200 shuffled keys; the thresholds set in advance were 0.85 and a margin of 0.30. The reference is a machine break, not a clear text of the time, and the transcription was made in one pass.

03 The reading

182 cipher words, 174 in the block and 8 in the margin: 149 certain, 25 probable, 2 restored and 6 unresolved, all six in the torn margin. That is 95.6% without the restored words. The probable words depart from the key by one sign each; several are probably slips of the single-pass transcription, and harbbian for habbian, twice, looks like a habit of the encipherer. The Italian and an English translation are under Full texts.

04 What remains uncertain

Full texts

The transcription and the decryption of every document on this page, so that the reading can be used and checked. Short files are shown here; all are in the repository.

c. 10r: cipher block L01-L11 and torn margin M1-M5

Beccaria has said that no one could carry out His Majesty's orders better than Count Espinula (Spinola); the marquis has not failed and will not fail to make the truth known, for his agents' honour and His Majesty's service, although it is awkward that the man was approved and sent to His Majesty as trustworthy; let the agents do their duty on their side, for the man is better known to His Majesty and his ministers than is believed, and neither he nor the one behind him will reach his aim.

Transcription

transcription.txt

# delcarretto1576 - Biblioteca Estense Universitaria, Modena, Autografoteca Campori, Lettera D, Del Carretto, Alfonso II, c. 10r
# (Regensburg, 2 Sep 1576). Cipher block, lines L01-L11, and the torn left margin note M1-M5.
# One pass by an AI agent, 4 Nov 2026, reading sign shapes before any value was known; frozen (SHA-256 d307d72f...) before any key work.
# Not blind-checked by a second reader. No images here (Lodovico IIIF, no licence stated: study only).
# Codes (one character per sign shape): a b c d e h k m n p q r t u x z 4 7 as written; C = c with a dot (ç); D = theta-like sign
# (looped 8 with a stem); E = e with a dot; G = looped 8 with an upper ring and a foot; l = tall l; O = reversed c (ɔ);
# P = rho (o with a stem); Q = q with a barred stem; z = ezh (ʒ).
# [..] = struck through or lost; ^..^ = interlinear insertion; space = gap between words; . = dot in the line; ? = doubtful sign.
# Tokenisation used by the key: 'oPo.' and 'oP.' -> 3 ; 'oPo' -> 1 ; 'oP' -> 2 ; remaining '.' dropped ; then sign by sign.
L01: up cOqqmenD 4m deD pO mpxeG qCaO [struck, long] ^zOxxC^ q4G koPok O bGeaCkm q4G [struck: aut bn7 DxxC Ox]
L02: [struck: ardluqnxkxG] bCaam hDrpuoPo kO qCau cGkO qoPohO upqCkxO labuk7pm GaaOQoPo.ueG pn oPoezukn zu a7D hmuzaxD
L03: Gx qqO [struck] Daanq7em koPk hDkq4Gem zu OaaGQ7ueO hGkkxeO aum aCPoPo . zD Q7OaxC aoPoPC boPC dme ru7zn?
L04: up anrkie buGxeoP zOp eGaxC , Ox zGPPm Q7DPuxm zGP 4oP.ChoPo , hD aum b7e qOexC q4G koPok 4Deccumk hDkqm
L05: kO aumhC bGe hDkqme boPokxC zu ame qoPokCaqPeG up 7GeoPo qCau nk 4oPokCeO zu aoPom anrkoPeum qChO uk
L06: aGe7uxnC zu aoP.om hDuOaxD Gx koPoaOeC DkqoPem . cGkq4O kG DDbCexu Q7mPq4G [struck] Ouddnq7PxD p' GaaOe
L07: qCax7u [struck] DDbeoPooP.mxC zGPPD hmkuOet zm a7D aurkoPenD [struck] 4m7GkzoPPC qoPohO bGeaCkm zu dG?
L08: zGaxukmxC D a7m hDuOaxm , Ox qChO q4G au [struck] anD bu7 xCaxoP zubCexmxoPo cGkO q4O hmpG , bOeGp?
L09: axum b7e aoP.D aurkCenD , Ox mxxGkzD m dDe up zGeuxC zD QoP.OPPm bDexO au qoPohO [struck] 4Deccumk
L10: dmxxC Ox dDeGhoPo kCu zm Q7GaxD . q4O aGkzC qoPax7u qCkoPaqu7xC zD aoP.om hDuOaxm Gx zDPPua ^en^ . a7En hukna
L11: bu7 q4G poPn dCeaO koPk au qeOzG , an beoPohOxxumhC q4G kO P7u kG Q7OPPC bDe qoPu OrPn beCqreD [struck] 4DoPemkkC p'ukxOkxC
M1: [lost]Oxq4G mp
M2: [lost]OixxC poPoC
M3: [lost] qCaO
M4: [lost]DaaGemkC
M5: [lost]C OkG .
Decryption

decryption.txt

# delcarretto1576 - mechanical decryption of transcription.txt with key.tsv (frozen key K_T), no correction.
# Struck-through and lost passages dropped; interlinear insertions read as words in place; one decoded word per cipher word.
L01	il beccaria ha fra le altre cose detto che non e persona che
L02	possa maglio ne cosi bene come ilconte espinula essequire li ordini di sua maidsta
L03	et cce assicura non manchera di essequire menntre sia solo da questo solo poo far giudi
L04	il signor pietro del resto et della qualita del huomo ma sia pur certo che non harbbian manca
L05	ne siamo per mancar ponto di sar conosclre il uero cosi in honore di soa signoria come in
L06	seruitio di suia maiesta et nosero ancora benche ne aaporti qualche eifficulta l esser
L07	costui aaprouato della maniere da sua signoria hauendolo come persona di fe
L08	destinato a sua maiesta et come che si sia piu tosto diportato bene che male perel
L09	stia pur sua signoria et attenda a far il derito da quella parte si come harbbian
L10	fatto et faremo noi da questa che sendo costui conosciuto da suia maiesta et dallis ri su?i minis
L11	piu che loi forse non si crede si promettiamo che ne lui ne quello par coi egli procgra haoranno lintento
M1	etche al
M2	eotto loo
M3	cose
M4	asserano
M5	o ene
Reading

reading.md

# delcarretto1576 - reading of the cipher (Italian, spelling of the letter)

Marks: plain = certain (every letter given by the frozen key); {x} = probable (one sign departs from the key, settled by the
sense); [x] = restored (the word is given by the sense, against or beyond the cipher); ‹…› = unresolved (margin torn).
Spelling of the letter kept (ponto, soa, essequire, huomo, seruitio); accents and apostrophes added in the reading only.

## Reading, line by line

- L01: il Beccaria ha fra le altre cose detto che non è persona che
- L02: possa {meglio} né così bene come il conte Espinula essequire li ordini di sua {maiestà}
- L03: et {ci} assicura non mancherà di essequire {mentre} sia solo da questo solo {poi} far {giudi[ce]}
- L04: il signor Pietro del resto et della qualità del huomo ma sia pur certo che non {habbian} {manca[to]}
- L05: né siamo per mancar ponto di {far} {conoscere} il vero così in honore di soa signoria come in
- L06: servitio di {sua} maiestà et {nostro} ancora benché ne {apporti} qualche {difficultà} l' esser
- L07: costui {approuato} della maniere da sua signoria havendolo come persona di fe
- L08: destinato a sua maiestà et come che si sia più tosto diportato bene che male {però}
- L09: stia pur sua signoria et attenda a far il {debito} da quella parte si come {habbian}
- L10: fatto et faremo noi da questa che sendo costui conosciuto da {sua} maiestà et {dalli} [suoi] {ministri}
- L11: più che {lui} forse non si crede si promettiamo che né lui né quello {per} {cui} egli {procura} {hauranno} l'intento
- M1: ‹…› al
- M2: ‹…› ‹…›
- M3: cose
- M4: ‹…›
- M5: ‹…› ‹…›

## Count (Daniel's rule: certain + probable over all cipher words; restored words not counted)

182 cipher words (block L01-L11: 174; margin M1-M5: 8): 149 certain, 25 probable, 2 restored, 6 unresolved.
**174/182 = 95.6% read by sense** (176/182 = 96.7% with the two restored words): **read**.
The cipher block alone: 172 certain or probable + 2 restored of 174 words; the 6 unresolved words are all in the torn margin.

Most probable words differ from the key by one sign (D/G, ʒ/ɔ, e/c, ρo/u); several are probably slips of our single-pass
transcription rather than of the encipherer, and a blind second reading of those signs has not been done. "harbbian" (twice,
for habbian) looks like a habit of the encipherer.

## Word by word (cipher token | key decryption | reading | status S/P/R/I | note)

- L01 `up` | il | il | S
- L01 `cOqqmenD` | beccaria | Beccaria | S
- L01 `4m` | ha | ha | S
- L01 `deD` | fra | fra | S
- L01 `pO` | le | le | S
- L01 `mpxeG` | altre | altre | S
- L01 `qCaO` | cose | cose | S
- L01 `zOxxC` | detto | detto | S | interlinear word replacing the struck-through one
- L01 `q4G` | che | che | S
- L01 `k1k` | non | non | S
- L01 `O` | e | è | S
- L01 `bGeaCkm` | persona | persona | S
- L01 `q4G` | che | che | S
- L02 `bCaam` | possa | possa | S
- L02 `hDrpu1` | maglio | {meglio} | P | 2nd sign D (= a) where the sense wants e; to check on the image (D/G)
- L02 `kO` | ne | né | S
- L02 `qCau` | cosi | così | S
- L02 `cGkO` | bene | bene | S
- L02 `q1hO` | come | come | S
- L02 `upqCkxO` | ilconte | il conte | S | written without a gap
- L02 `labuk7pm` | espinula | Espinula | S | = Spinola (Spanish form Espínola); l = e only here
- L02 `GaaOQ3ueG` | essequire | essequire | S
- L02 `pn` | li | li | S
- L02 `1ezukn` | ordini | ordini | S
- L02 `zu` | di | di | S
- L02 `a7D` | sua | sua | S
- L02 `hmuzaxD` | maidsta | {maiestà} | P | 4th sign read ʒ (= d); the sense wants ɔ (= e); ʒ and ɔ are close in this hand
- L03 `Gx` | et | et | S
- L03 `qqO` | cce | {ci} | P | qqɔ = cce; the sense wants ci/ce ('et ci assicura'); spelling to check
- L03 `Daanq7em` | assicura | assicura | S
- L03 `k2k` | non | non | S
- L03 `hDkq4Gem` | manchera | mancherà | S
- L03 `zu` | di | di | S
- L03 `OaaGQ7ueO` | essequire | essequire | S
- L03 `hGkkxeO` | menntre | {mentre} | P | n doubled (kk)
- L03 `aum` | sia | sia | S
- L03 `aCP1` | solo | solo | S
- L03 `zD` | da | da | S
- L03 `Q7OaxC` | questo | questo | S
- L03 `a1PC` | solo | solo | S
- L03 `b2C` | poo | {poi} | P | ρo-C = o-o; the sense wants poi; last sign perhaps another (a dot)
- L03 `dme` | far | far | S
- L03 `ru7zn` | giudi | {giudi[ce]} | P | word at the edge of the leaf, end cut or abbreviated
- L04 `up` | il | il | S
- L04 `anrkie` | signor | signor | S
- L04 `buGxe2` | pietro | Pietro | S
- L04 `zOp` | del | del | S
- L04 `eGaxC` | resto | resto | S
- L04 `Ox` | et | et | S
- L04 `zGPPm` | della | della | S
- L04 `Q7DPuxm` | qualita | qualità | S
- L04 `zGP` | del | del | S
- L04 `43Ch1` | huomo | huomo | S
- L04 `hD` | ma | ma | S
- L04 `aum` | sia | sia | S
- L04 `b7e` | pur | pur | S
- L04 `qOexC` | certo | certo | S
- L04 `q4G` | che | che | S
- L04 `k1k` | non | non | S
- L04 `4Deccumk` | harbbian | {habbian} | P | one sign e (= r) too many between a and cc; same spelling in L09 (habit of the encipherer)
- L04 `hDkqm` | manca | {manca[to]} | P | end of line at the edge; 'habbian mancato'
- L05 `kO` | ne | né | S
- L05 `aumhC` | siamo | siamo | S
- L05 `bGe` | per | per | S
- L05 `hDkqme` | mancar | mancar | S
- L05 `b1kxC` | ponto | ponto | S | = punto, spelling of the time
- L05 `zu` | di | di | S
- L05 `ame` | sar | {far} | P | 1st sign a (= s) where the sense wants f (d); s and f are close in clear writing too
- L05 `q1kCaqPeG` | conosclre | {conoscere} | P | ρ (= l) where the sense wants e
- L05 `up` | il | il | S
- L05 `7Ge1` | uero | vero | S
- L05 `qCau` | cosi | così | S
- L05 `nk` | in | in | S
- L05 `41kCeO` | honore | honore | S
- L05 `zu` | di | di | S
- L05 `a1m` | soa | soa | S | = sua, spelling of the time
- L05 `anrk2eum` | signoria | signoria | S
- L05 `qChO` | come | come | S
- L05 `uk` | in | in | S
- L06 `aGe7uxnC` | seruitio | servitio | S
- L06 `zu` | di | di | S
- L06 `a3om` | suia | {sua} | P | isolated o = null (the only isolated o of the text, with L10)
- L06 `hDuOaxD` | maiesta | maiestà | S
- L06 `Gx` | et | et | S
- L06 `k1aOeC` | nosero | {nostro} | P | ɔ (= e) where the sense wants t (x)
- L06 `Dkq2em` | ancora | ancora | S
- L06 `cGkq4O` | benche | benché | S
- L06 `kG` | ne | ne | S
- L06 `DDbCexu` | aaporti | {apporti} | P | 2nd θ where the sense wants p; same in L07 (θθb = app?)
- L06 `Q7mPq4G` | qualche | qualche | S
- L06 `Ouddnq7PxD` | eifficulta | {difficultà} | P | 1st sign transcribed ɔ, a different shape (g/9); = d by the sense and by dd = ff
- L06 `p` | l | l' | S
- L06 `GaaOe` | esser | esser | S
- L07 `qCax7u` | costui | costui | S
- L07 `DDbe13mxC` | aaprouato | {approuato} | P | as in L06
- L07 `zGPPD` | della | della | S
- L07 `hmkuOet` | maniere | maniere | S
- L07 `zm` | da | da | S
- L07 `a7D` | sua | sua | S
- L07 `aurk2enD` | signoria | signoria | S
- L07 `4m7Gkz2PC` | hauendolo | havendolo | S
- L07 `q1hO` | come | come | S
- L07 `bGeaCkm` | persona | persona | S
- L07 `zu` | di | di | S
- L07 `dG` | fe | fe | S | fé (= fede); edge of the leaf
- L08 `zGaxukmxC` | destinato | destinato | S
- L08 `D` | a | a | S
- L08 `a7m` | sua | sua | S
- L08 `hDuOaxm` | maiesta | maiestà | S
- L08 `Ox` | et | et | S
- L08 `qChO` | come | come | S
- L08 `q4G` | che | che | S
- L08 `au` | si | si | S
- L08 `anD` | sia | sia | S
- L08 `bu7` | piu | più | S
- L08 `xCax2` | tosto | tosto | S
- L08 `zubCexmx1` | diportato | diportato | S
- L08 `cGkO` | bene | bene | S
- L08 `q4O` | che | che | S
- L08 `hmpG` | male | male | S
- L08 `bOeGp` | perel | {però} | R | word at the torn edge, end lost; 'però stia pur…'
- L09 `axum` | stia | stia | S
- L09 `b7e` | pur | pur | S
- L09 `a3D` | sua | sua | S
- L09 `aurkCenD` | signoria | signoria | S
- L09 `Ox` | et | et | S
- L09 `mxxGkzD` | attenda | attenda | S
- L09 `m` | a | a | S
- L09 `dDe` | far | far | S
- L09 `up` | il | il | S
- L09 `zGeuxC` | derito | {debito} | P | e (= r) for c (= b)
- L09 `zD` | da | da | S
- L09 `Q3OPPm` | quella | quella | S
- L09 `bDexO` | parte | parte | S
- L09 `au` | si | si | S
- L09 `q1hO` | come | come | S
- L09 `4Deccumk` | harbbian | {habbian} | P | as in L04
- L10 `dmxxC` | fatto | fatto | S
- L10 `Ox` | et | et | S
- L10 `dDeGh1` | faremo | faremo | S
- L10 `kCu` | noi | noi | S
- L10 `zm` | da | da | S
- L10 `Q7GaxD` | questa | questa | S
- L10 `q4O` | che | che | S
- L10 `aGkzC` | sendo | sendo | S
- L10 `q2ax7u` | costui | costui | S
- L10 `qCk2aqu7xC` | conosciuto | conosciuto | S
- L10 `zD` | da | da | S
- L10 `a3om` | suia | {sua} | P | as in L06
- L10 `hDuOaxm` | maiesta | maiestà | S
- L10 `Gx` | et | et | S
- L10 `zDPPua` | dallis | {dalli} | P | final a wrongly joined to the word? (word break to check)
- L10 `a7En` | suEi | [suoi] | R | é not found elsewhere; o by the sense
- L10 `hukna+^en^` | minis+ri | {ministri} | P | interlinear 'en' = ri; t left out or in the interlinear
- L11 `bu7` | piu | più | S
- L11 `q4G` | che | che | S
- L11 `p2n` | loi | {lui} | P | ρo (= o) for u; the dot of ρ• perhaps lost
- L11 `dCeaO` | forse | forse | S
- L11 `k2k` | non | non | S
- L11 `au` | si | si | S
- L11 `qeOzG` | crede | crede | S
- L11 `an` | si | si | S
- L11 `be1hOxxumhC` | promettiamo | promettiamo | S
- L11 `q4G` | che | che | S
- L11 `kO` | ne | né | S
- L11 `P7u` | lui | lui | S
- L11 `kG` | ne | né | S
- L11 `Q7OPPC` | quello | quello | S
- L11 `bDe` | par | {per} | P | D (= a) for e
- L11 `q2u` | coi | {cui} | P | as in 'loi'
- L11 `OrPn` | egli | egli | S
- L11 `beCqreD` | procgra | {procura} | P | r (= g) for u
- L11 `4D2emkkC` | haoranno | {hauranno} | P | ρo for u (cf. loi, coi)
- L11 `pukxOkxC` | lintento | l'intento | S
- M1 `[lost]Oxq4G` | …etche | ‹…› | I | start lost (tear)
- M1 `mp` | al | al | S
- M2 `[lost]OixxC` | …eotto | ‹…› | I | start lost
- M2 `p1C` | loo | ‹…› | I
- M3 `qCaO` | cose | cose | S
- M4 `[lost]DaaGemkC` | …asserano | ‹…› | I | start lost
- M5 `[lost]C` | …o | ‹…› | I
- M5 `OkG` | ene | ‹…› | I | = bene? first sign lost
Translation

translation.md

# delcarretto1576 - English translation of the cipher

Translation of reading.md (Brice Dudout, with a Claude Opus 5.5 agent, 6 Nov 2026). Probable words are not marked here
(see reading.md); the torn margin is left out except for its legible words.

> Beccaria has said, among other things, that there is no one who could carry out His Majesty's orders better or as
> well as Count Espinula (Spinola), and he assures us that he will not fail to carry them out, as long as it is left to
> him alone; then let Signor Pietro judge the rest and the quality of the man. But be sure that we have not failed, nor
> shall we fail in any point, to make the truth known, as much for the honour of Your Lordship as for the service of His
> Majesty, and ours too, although it brings some difficulty that this man was approved in the way he was by Your
> Lordship, who sent him to His Majesty as a trustworthy person, and that he has behaved rather well than badly.
> So let Your Lordship rest easy and attend to doing your duty on that side, as we have done and shall do on this one;
> for since this man is known to His Majesty and to his ministers better perhaps than is believed, we trust that neither
> he nor the one for whom he acts will achieve their aim.

Margin (torn): "… al … … things … …".

Context (to be checked by a historian): Alfonso II del Carretto, marquis of Finale, at the Diet of Regensburg (1576),
writes to his agents in Milan, Lodovico Allaria and Paolo Fenocchio, about a man recommended to Philip II and about
Count Spinola; Finale was then contested between its marquis, its subjects, Genoa, Spain and the Empire. "Signor
Pietro" and "Beccaria" are not identified.

Key: key.tsv

05 Sources

Notes, transcription, decryption, reading, translation and the key are in targets/delcarretto1576.

Cite this page

Brice Dudout and Daniel Bourdeau, “Alfonso II del Carretto to his agents in Milan, Regensburg, 2 September 1576”, Cyphersolver, posted 6 November 2026, https://cyphersolver.com/delcarretto1576.html.