At a glanceKey from adjacent plaintext · partial (72% of the cipher)
- The letter
- Nicolò Guarna, writing from Milan as Count Francesco Sforza’s envoy at the ducal court, to Sforza, on Sunday 21 May 1447. It is signed “Servitor N. Guarna”. The opening, the connectives and the close are in clear; most of the body, about 3,960 cipher signs over three pages, is in cipher.
- Where
- Paris, Bibliothèque nationale de France, italien 1584 (the Archivio Sforzesco, 1447), ff. 153r–154r, on Gallica (views 146–147). No DECODE record and no edition found. The letter is not in the catalogue; the lead for Guarna’s other letters in this volume (ff. 163–164 and others) does not list ff. 153–154.
- The cipher
- Homophonic substitution with nulls (79 distinct signs), clear decoy words written inside the cipher (quia, adviso, mo, primum …) and a small code (cer = Signore, mea = non, tuy = gente d’arme, tri = Re de Ragona, dn = dinari). Several signs carry more than one value in our table, which may partly reflect our own confusion of look-alike signs.
- How it was read
- Brice Dudout rebuilt the key from Guarna’s letter of 27 May 1447 (ff. 163r–163v), aligned with its 18th-century clear copy (ff. 161r–162v, 165r), and applied it to this letter, which has no copy. The tests do not validate it: both held-out measurements are partial, and there is no fresh held-out block left.
- What it says
- An audience with Duke Filippo Maria Visconti about Sforza’s coming to Lombardy: enemies in the field, troops poorly provided, the plague, nine thousand ducats and then “not another florin for a long while”. The Duke answered “more openly than the other times” and shows a fervent wish for Sforza to come quickly.
- Still open
- About 236 words are unresolved (an estimate: each unread stretch counted as its length divided by five), mostly on f. 153r and the first half of f. 153v. 72.3% is below the 95% read bar. The key is partial, and the end of the letter (f. 153v from about l. 46 to f. 154r) was already in clear on f. 156, so it is not new.
- Credit
- Brice Dudout (Europe chiffrée), working with Claude Opus 5.5 agents: transcription, key, tests and reading, contributed by pull request #280. The clear text of f. 156 is the contemporary reading of the end. The training pair came from Daniel Bourdeau’s survey of the late copies in this volume.
01 The letter
Guarna writes to the Count from the court of Duke Filippo Maria Visconti, the last year of the Duke’s life, when Sforza, his son-in-law and condottiere, was far off in the Marche and the Duchy was short of money and men. The letter has about fifty lines on f. 153r, fifty-two on f. 153v and three on f. 154r. The salutation, many connectives and the close are in clear; the substance is in cipher, with clear decoy words scattered through it.
The letter belongs to the Sforza archive volume that also holds the Duke’s own letters to Sforza and those of Barbavara and Pusterla, each in a different cipher; the Duke’s second key and the Barbavara and Pusterla letters have their own pages, and nothing here is read with them. f. 156, a 15th-century clear text, gives the end of a letter of the same day, which matches the end of this cipher.
02 The cipher
The signs are letter-like and geometric, written without word division. Most stand for a single letter, and several letters have more than one sign (homophones). Some signs are nulls, and a few clear words (quia, adviso, tanto, tame, mo, cio, oto, si, ot, primum, tua, spes) are written inside the runs and dropped when reading. A short code gives whole words: A = Papa, cer = Signore, mea = non, con = qua, tus = que, tuy = gente d’arme, ba = per, ra = perché, qt = quantunque, tri = Re de Ragona, SF = Signore de Faenza, dn = dinari.
In the table as we have it, many signs list two or three values, most frequent first. That may be the cipher’s own polyphony or the contributor’s confusion of look-alike signs on a microfilm (a four-like sign and hk, a sigma and an o with a hook, t, T and z, hs and h). A blind shape sheet on those pairs was prepared and is not yet answered, so it was not applied. The key is published as targets/guarna1447/key.tsv (CC BY 4.0).
03 The key and its tests
Where the key comes from. No Guarna key was known: DECODE lists none among the keys examined, and nothing in print was found. Brice Dudout rebuilt it from the cipher of Guarna’s letter of 27 May 1447 (ff. 163r–163v), aligned with its late clear copy (ff. 161r–162v and 165r), the pair listed in the survey of late copies of this volume. The key was frozen in three stages (K, K2, K3, each fixed by SHA-256 before its test) and applied to this letter.
The tests. The thresholds were fixed in advance: validated at 0.60 (words of five letters or more in Italian) or 0.85 (characters against a clear), and at least 0.30 above the best of 200 shuffled keys; partial when the gap holds without the absolute figure.
| test | held-out block | key | metric | score | best control | result |
|---|---|---|---|---|---|---|
| 1 | f. 153r ll. 1–20 (this letter) | K2 | Italian words of 5+ letters | 0.441 | 0.118 | partial (gap 0.323) |
| 2 | f. 164r (the training letter) against its copy f. 165r | K3 | characters | 0.759 | 0.315 | partial (gap 0.445) |
From targets/guarna1447/validation.txt. These are the contributor’s own pre-registered tests, not the project’s _validate.py.
What that means. The key is partial by our protocol, twice, and not validated. Both held-out blocks are now used, so no fresh block is left to test it on. Declared slips: line 20 of the first block (clear only) was seen before the block was frozen, and the copy f. 165r was seen at low resolution while the second block was chosen, though it was not used to build the key. The end of the cipher agrees word for word with the clear text of f. 156, which the key never used, but that is not a pre-registered test.
04 The reading
The reading covers the cipher segments only (the clear frame of the letter is not counted): 852 words, of which 583 are certain and 33 probable, none restored and 236 unresolved. Certain and probable make 616 words, 72.3%. The unresolved count is an estimate, because a stretch that does not read cannot be cut into words: each is counted as its decoded length divided by five. By part: the held-out block 71%, the rest of f. 153r 71%, f. 153v 73%, f. 154r 91%. That is below the 95% bar, so the letter is read in part. The reader is one AI agent working on a microfilm, and no palaeographer has checked the signs.
The sentences are built from stretches that construe as 15th-century Lombard Italian (de li quali, per respecto del spaccio, senza comparatione) with many holes, most of them on f. 153r and the early lines of f. 153v. The part where the reading is steadiest is the end, f. 153v from about line 46 to f. 154r, and that part is also the part that was already in clear on f. 156. The passage below is the end of the letter in the normalised text (a bare word is certain, {x} probable, [?n] about n unresolved words).
ne {recevessimo} [?1] per le genti de Cremona restano [?1] milia li quali spero havere fra otto o diece di o non prima … ma per {dirve} quello che io credo pareme esser certo voi non haveriti piu uno fiorino per uno grande spacio se la S.V. aspecta de moverse. … vole bene voria che facesti ultimum de potentia de venire presto et Dio volese che potesti farla
of which we received […] for the troops of Cremona remain […] thousand, which I hope to have within eight or ten days, not before … but to tell you what I believe, I seem certain you will not have another florin for a long while, if Your Lordship waits to move … [one who] wishes you well would want you to do your utmost to come quickly, and God grant you may do it.
- The danger of the state. Guarna reports that the enemies are in the countryside and that the state stands in great danger; the plague adds to it. What he wants from Sforza is not clear from the opening lines, which are mostly holes: they speak of remedying “these cursed troubles”, of how Guarna stands in the world on account of the dispatch, and of what was hoped from Sforza.
- Money. The main thing to remedy is to find money. Besides nine thousand ducats there is what is expected from the gabelle, and money for the troops at Cremona, to be had in eight or ten days; after that, not another florin for a long while.
- The Duke. He answered Guarna “more openly than he had done the other times” and shows a fervent wish for Sforza’s coming, though in effect he has no money to give. Sforza should come quickly, with all or part of his company.
The gist (an audience with the Duke, money short, the Duke wanting Sforza to come) holds across the parts that read; the details of the first half are not secure.
05 What remains uncertain
- About 236 unresolved words (estimated), mostly on f. 153r and f. 153v ll. 4–19. Blocker: illegible; the look-alike signs are not separated on the microfilm. The next step is the pending blind shape sheet and a finer re-transcription.
- The key is not validated (section 03), and there is no fresh held-out block. Another Guarna leaf with a complete late copy (the lead for Guarna’s letters in this volume) would serve as a new test.
- Not new at the end. f. 153v from about line 46 to f. 154r agrees with the 15th-century clear of f. 156, so the contribution of the reading is the rest, where no decipherment is known.
- Multi-valued signs. Most signs list more than one value in the key. In the reveal above the orange signs are the ones where the reading takes a value other than the first, or corrects one from the sense.
- One reader. An AI agent transcribed and read; no palaeographer checked the signs.
- Not seen. Cerioni, La diplomazia sforzesca (1970), and the printed calendars of BnF italien 1584. Four web searches (the last on 5 November 2026) found no edition of this letter; if one prints it, we will credit it.
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.
ff. 153r-154r, cipher segments of the letter of 21 May 1447 read in part
Guarna reports an audience with Duke Filippo Maria Visconti about Sforza's coming: the state in danger (enemies in the field, troops poorly provided, plague), 9,000 ducats, the gabelle and the troops at Cremona, then no more money for a long while; the Duke answered more openly than before and wants Sforza to come quickly.
Builds coherent sentences, but under the read bar of 95% of words read as sense (72% here); the holes are marked in the reading.
Transcription
# guarna1447 - Paris, BnF, italien 1584 (Archivio Sforzesco), ff. 153r-154r (Gallica btv1b100373864, views 146-147): # Nicolò Guarna to Francesco Sforza, Milan, Sunday 21 May 1447. Cipher segments only, one line per segment, by shape names; # [word] = a clear word written inside the cipher (decoy or real, see key.tsv). Transcribed by one AI agent, 5 Nov 2026. # H02-H19 = held-out block (f. 153r ll. 1-20; l. 1 and l. 20 clear only), hashed before decoding (SHA-256 95e803c0...). # L22b-L51 = f. 153r ll. 21-51 (the token file was offset by one: L22b = l. 21); V1-V52 = f. 153v; P1-P3 = f. 154r. # No images here. H02 hk 7 ym oe h 4 + ff bp gp hd t 2# pi gp pc ot pi 6 oe qn 7 a bp oe ^ tc x pi nd 2 8 hz oe H03 qn pi oe nd 2 4 h 7 ls tc [mo] 2# 8 + bp fi pi bp H04 ym x q hs 2 hk + nd hd a b h 7 pr + ff qq q hk 4 ff 4 t pi hz [nec] pc oe hk pi 4 bp 2 q ff gp x a H05 oe to t che x bp [nec] to x ym q pi b hs oe ot hk q x H06 [primum] [tua] c q ls 7 2 8 2 oe nd ot bp x H07 2# pi ls qn bp 2 b hd ym 4 2 hk x hz q ph oe k + pi x 4 nd 2# pi oe pc ot b ym x qn h qo x pr 7 hk ff oe hs T gp ^ b [cio] h 7 fi pc pr + pp pi q H08 a gp ls 2 8 2 7 pi c u mea 8 gp h 7 8 + pc oe 2 hz tc hk oe 2# pi 7 W 4 gp 4 7 ^ 4 che pi b pc pr 7 ff + pi t bp h q fi t hs hd pi x H09 pr + ls bp 2 8 ot gp H10 pi x hd ym 4 2 hk x pi + hz t 4 pc 7 nd k q 2 8 pr + ff t 2 4 x nd oe bp T 7 z q 2 8 2 7 qn ot b [mo] che mea gp hk t hs 2 + ot H11 pi x k ls 4 h pi pr + ph q bp 2 tc bp pi ls+ hk t pc q nd oe [tame] 7 hk q 2# pi oe pc 2 + [ac] t fi pi a tc pr q ph hd q 2 4 bp tc x oe ff t H12 pc oe q ls z 2 8 ot gp nd 2 b hs q ls x T 7 hk ls oe ff + pc b nd 4 h tc ^ gp hd x ym + q ls T ff 7 pc oe nd ot gp hs 7 nd 2 gp qn a b [mo] fi 4 qn tc ym pi c H13 fi tc hd x ph gp hk oe # hk x ph tc x u 8 t pi + ls c q hk oe pr gp ff x fi + c t ym T + # qn q bp pc 7 nd k t hd x 2 pc 2 + c x ls q [cio] gp x bp H14 bp a ot 6 4 2 + hk 7 [manifestamente] che ls x pc 2 + ot q h 7 to + pc 2 t hz + 4 nd hz ff t nd a 7 pr oe hk 4 c b ls q H15a c b ls ot bp x pr 7 hk 4 c q fi b pc + hd x + H15b ls x q H16 7 pi 4 a q nd ot gp to + hz 4 c bp q fi x qo h gp fi pc 2 + ot b a gp to + ls x t 4 q c t b ym 4 x hd 7 hk + bp tc nd 2 b W 7 ff q 6 4 ls gp bp H17 pr 7 hk x che 2# pi t nd a b pi oe qn 7 ls x 4 + c 2 b h 7 ls+ hs hd x pi bp q nd h x ph 7 [nec] + qn a t 8 bp q x pi ls x pc 2 + ot b h gp to + H18 pc pr hs # fi 8 t ph 7 pc gp ym T hk 7 to + [cio] T b qq 4 + ff 7 ym oe h 4 + hk gp t ls pi 4 q c 2 4 pi q pc ot ff pi pi 7 ff + bp c hz ff ot q H19 che 2# pi t nd h q ls b 2 7 q T x # fi tc [primum] hs 7 ff [cio] pi gp ph oe to + o + qq 4 t pi 7 hk hs b fi q pc bp 8 qo L22b ym + pc 4 ls x L23 hs 2 + ot b hk gp to + mea t qn a q ph 7 6 gp qo hd pc oe hk hk + ym + ym q ls 7 4 qq gp c 8 pr oe ff + hz 2 4 ls hs 7 bp pc 4 pr 7 hk to + L24a pi hs tc + ot ff pi hz 4 [cio] L25b h + ls q hd pc hd pi bp L26 + ls 2 8 bp 2 gp pr hk x c 7 a oe mea hs 7 nd k t 3 ff + qn a 4 ph pi ym x bp 7 qq x q hk 7 hz 8 pc 4 a gp hk + 2 b che pr 7 hk ls + pi 4 t h 7 fi L27 2 8 ot oe ls + hs x [mo] hz x [cio] pc 4 pr pi x hs + fi hz + hk 7 u mea [solamente] a gp ph 7 nd h oe ff pi oe ym + pi nd q 2 8 ls q fi 2 hk 7 # oe L28 qn ot 4 a oe ls q pc qo 8 x bp to q 3 7 hk oe fi q T 4 8 # fi x hk 7 x hk pi q pc + hz 4 c 2 b ff tc + che pc 4 q hs 2 + ot t 4 nd tc 2 tc fi 4 t L29 # hk + nd ot gp ym pr x L + u T 7 hk hd h t ff gp bp q pi a 4 7 nd 2 tc q [mo] + a c to tc to + pi x fi hz 2 b h pi ff oe e q ls gp hd L30a ls + hd hs [primum] hz [mo] pr + 2 gp h 7 ym oe ^ gp ff 2 bp fi x pc 2 + ot b hs hz q 4 nd [ena] T gp hk 4 hd c q fi x hd pi mea hs bp x L31 z hz + ls 7 pr hk b hz t hs oe T q 2 7 ph gp bp 4 q ffo 4 qn 2# pi bp 7 hs ot x hd 4 q nd h b [mo] che 4 q ph 7 ym t 4 q hk 7 h gp 2# L32 pi 7 W t che to + 4 q ^ t ls x 2 8 ot gp bp hd 2 8 gp gx pr hk 7 ph t oe gx T oe hk 4 gp qn 2 pi t [tame] h 7 fi q 4 7 h oe pc x t L33 pi 4 nd c fi tc nd + 2 pi x qn gp pi gp hk pc b ls + pc hd [cio] hz x bp L34 + ss [Adiungendo adesso pare che] [nca] 6 x nd + hd pi x ls b qn 2 + h 7 2# hz t nd 2 q ls q hs x 8 hd to q pi 4 3 q 8 gp hk pc x 7 ls L35 ^ 8 hz pi [tua] 7 gx pr oe ff 4 oe q 2 pi t bp che 4 + 4 q ot ss pr ot b 7 ls 2 8 ot gp hz 7 fi + pc x q 4 7 h gp u a 7 pi b 2 4 q nd gp L36 3 7 hk pc b ls + hs qo pi bp mea hs 4 pr x ph b nd q hd a 7 4 gp nd h oe hk 7 h q to b ym 4 nd tc T 7 hk pi oe ff pc pi hd u h 4 q qo x L37 fi tc c pi che 4 + c b pr q pc 2 q 2 hz ^ q ls b 4 nd ot 7 tuy gp ^ b fi x hk b [tanto] + h c t c pi fi q hk oe ls q ym gp nd 2 oe h 7 fi q pc L38 [adviso] pi bp x 8 c h 7 hs 4 # qn b hk 7 mea hz b 3 fi 4 + 2 7 bp pr q ph ff h pi q T 7 hk pc tc hz t nd pi pr 7 qn hs hk 7 hk tc ym 7 t gp L39a hk oe ls q pc 2 + ot b h 7 ls t hs bp 8 [tanto] hk nd [tua] pr 7 hk tc c b fi q L39b pi q x ls 7 hk pc pi h 7 ls L40 2 8 q oe A T x tc che to t pi 7 hk pi 2 4 4 + c ot tc ls pi oe q 8 2 pi [mo] pi b pc ot hk 4 2 ff + c ot t hk gp ls x 2 8 2 oe hd x 3 oe ff b L41a pc 7 2 8 h q ls tc ym gp ff pi 2 4 hs hz b tc b ss gp 4 x L41b gp fi 2 8 ot 7 pc 2 t hk + L42 mea pc b fi x pr + c 4 7 nd ot oe ym q 2 8 2 oe qn ot tc ph pi c b bp t a b hz nd pi ls q pc 2 4 q hd T 4 + c 7 ff gp T h 7 2 gp ff ym tc nd t L43a ot hk b nd 7 L44b bp x qo L45 ls x hd pc pi hs pr 7 ^ b hk q L45b pi 2 8 4 pi hk t ym oe nd 2 tc hd che + c c 4 mea 2 8 4 pc b # nd + pi q h oe L46 to t pi bp 7 ffo oe hd hs pi ym tc fi 2 pr + ph tc x qo bp u che hs gp 4 b ph 7 + fi c pi nd b a gp 2 + nd ot t [mo] q pi h + 2 pi t ot oe L47 ff 7 pc hz ym 2 4 b qn oe che hz b ls 7 ph gp ym 7 t gp ff oe hd fi + c oe 7 c oe qn 2 7 h gp fi + pc hd hs b + 4 nd pc hz hs pr 7 ^ q h 7 ls t hs bp L48a ss pr 7 ff 2 gp ym T b hd x x T 7 hk ot + hk a x hs oe nd gp T 7 nd ot 4 ff pi + L50b x ls x nd b pc ot ff q pr 7 hk 4 4 c b fi x pi oe h 4 c gp ym x [tanto] bp che h h oe L51 tc 2 pi fi x hz oe hk x u 8 b # nd q pc c 7 4 bp q e b ls 2 x 2 gp qn 7 ls q 4 + c 2 q nd b pc ot ff q pc 2 + ff gp hk qn # hk t nd V1 pr 7 hk 4 c x fi q [mo] pc tc T oe ff ls + T x ph t nd k q h oe ls 4 nd pi ym 4 c tc pi pr 7 hk fi + T x 8 t T hk x bp pi tc pc 4 b qn 7 che pc pi hd L + bp V2 + a fi 7 # oe nd ot tc nd q pc 2 ff gp u pc tc pr 7 hk pc pi hs pr 7 ^ b [tame] [adviso] che to t hz 7 ym x a gp x t fi c hz nd b a 7 ls 4 qn x pc ot ff tc V3a ym t V3b nd 3 # to + pi 7 ym bp x 2 8 c q nd h + 2 b pi 2 8 ym t nd h q ym x hd b # nd tc h pi pi pc tc V4 q ym x c 7 hk 2 pi che hs 7 8 pc gp pi pc q ph 7 V4b hs 4 4 + 4 4 t hd q ph t tc ss pr 4 pi che mea qo oe + pi nd pi 4 V5 pc oe ym b 4 qn to t ot pi fi q hd h 7 ls + pr gp ss ot x nd + x che mea T b 4 4 t ym q x pr q 4 ot q ff bp gp pc 4 ym tc ls 4 4 + 2 4 che u c q ph V6 tc ls + c q pc + pi ym q ym t ls 7 [cio] x V7 pi 7 ph gp ff oe hd fi x pr hk 4 nd c tc bp T + ls ym bp che to + pi 7 ph 4 ym b + a hk 7 c + ^ t ff gp pi + h ot ff b hz q hk 7 [spes] pi che to + V8 mea 4 + c 7 hz + h pi hz tc b che 2 8 pr + hk 7 nd h q tc b pi qn hd nd x ym 7 h gp ls q pc hd x pi bp hs 7 2 ff x hz q hk 4 + nd q a 7 ls gp 8 4 7 V9 + ph q tc bp h hz q gp c + ^ t ffo [spes] V9b pr 7 hk h pi 4 ff tc pc T 7 ^ V10 pi pr ff tc ym b [cio] mea hs x hd + ^ b nd oe h 4 pc T x hs 2 q bp t a hs 4 ym 4 4 ym tc fi 7 c q hs oe [primum] ls b pc 7 2 8 h b ff tc hz T oe ^ q pi V11 T ot hk bp mea t 7 3 4 pc 2 + hk pi tc c q fi ot pi hd 4 nd tc c pi c 4 hd 4 nd 2# pi oe pc ot + T q 2 ff pi q [cio] pr 7 hk c + hz t ff # ls tc [spes] V12 h 7 ls + 6 b hk pc q hd 2# pi + nd h q pi b to 7 ff 4 qn 7 # + 2 b V13 h 7 6 q nd oe hd pr + hk b ls gp ym + hk tc # pi h q pi hs pi hs 7 c + fi 4 4 + ^ pi V15 ym tc nd x c 4 nd b c 7 ph gp q b hs c gp pi + pi q bp ot x q V16 4 x # hz tc a b 2 q x q bp x x 2 8 h 4 4 oe che V16b pc b nd x hd tc q c ls 4 c nd t ot 4 t fi q pc hd 8 [primum] 4 + hk 4 t nd x V17 pi gp gx oe 2 8 tc ff 4 q nd b bp pr 4 pi hd hz b ls q nd 2 7 ff pi fi q oe + c ot b pi ot b pi q pc 2 ff q [tanto] V19 ls + 6 7 bp 4 tc nd ot q hd V20 pi 7 nd 4 [cio] hk 7 bp pi b pc ot ff q bp hd qo x [quia] gx cer V21 pc 4 pr x ph t nd b x to + pi gp ff oe q fi 2 ff + 2# hz gp pc ot pi nd q pi gp c 4 W tc q [tame] h pi c + hd V22 2 4 bp che + pc pr gp ^ t ym x 4 q pi 7 ff oe bp h 7 ls q # t 6 7 W q hd a gp fi tc 2 pi + fi h pi [adviso] 6 7 nd oe h 4 c gp che hz hz + nd V23 h q bp ls + pc [quia] pi [si] 4 q ph 7 h gp 2# pi + hs 7 ot 4 x hz 4 tc bp q nd x a 7 fi pi [spes] t ph hd q tc bp V24 ls+ x hs [mo] 8 [quia] hz mea + hs pr 7 ^ t ph gp + fi 2 ff q 2# pi + nd 2 4 ot + bp a 7 [spes] hd pi che V25 4 + c 7 ph gp pr pi b u che 7 8 7 pc ot b hk gp pc 2 + mea T gp hk x mea pi b ls 7 # oe ym + T oe ff mea pr x ot oe ff gp pi che 2# pi + nd V26 h b mea pr b 2 7 pc ot 4 ym gp nd q 4 oe ot 8 ^ + fi t 2 8 T q # nd 4 t ym gp qn q hs 2 tc 2# pi 7 W + pr t ff ot oe pi 4 bp 4 b ph 7 T x ph V27 4 6 tc ls oe T pi ff 7 che mea 8 oe qn hk pc hd 2 tc 2 8 c + nd c b che 2 8 ls + c 4 ot + [mo] h 7 fi t 2 8 pr q # nd 4 t c u 2 hz + nd h b # ls tc pi 7 V28 nd 4 pc 2 pi 2 8 c gp q x hs 7 nd t pi nd b [tanto] pr 7 ff a gp ffo oe a 7 bp 4 gp T hz 2 + ot 4 q nd oe T 7 ff to + pi T 7 ff ls + pi b pc 2 ff hd t V29 ls tc hz x pc ot 4 pi [mo] ym + nd c t c oe nd 2 tc pi hz 4 hs b # qn tc pi che pr 4 pi bp pi x ls q nd ot 7 ff 4 8 7 nd 7 4 tc pc V30 ot 4 2 8 T b c to 4 [spes] b 2 8 nd 4 7 nd 2 gp [quia] che 2 8 t ph q bp x 4 pr 7 hk 4 + ff oe c q pc + # hk t 2 q bp + to + ym + ls q hs bp V31 u x pi hk 7 pc 2 q pc b fi x pr 7 ff 4 q pr b 2 gp c 2 tc + V32 mea ym 4 ym 4 bp gp h 7 pc pr 4 + c tc pi ot x hd pr 4 4 c b c 4 gp T t V33 hk pi ot x bp che cer c 7 hd 4 nd 2 7 pc gp pi x ls pi nd 2 7 ff 4 pi hk tc hs T x pc 7 ym 4 [mo] T 4 pi + pr 7 hk ot t ym 7 nd 2 oe che hd x mea V34 to + 4 t ^ x ls oe q ls ot ffo pi b fi 2 gp a 2 pr t 4 x fi oe + ph q tc pr 4 8 T tc t c 7 fi x ls pi bp u ls q 4 hz oe h 7 ls pi pc + 2 b V35 4 nd ym q qn hs 2 ff + 4 oe [cio] hz b ls q nd 2 + h oe ls + hz gp qn 2 t 8 b hs 2 ff q qt 4 nd ls q 7 4 gp ^ b a oe fi pi gp qn tc 4 7 [mo] k q gp V36 h 7 [spes] mea # fi tc gp + ls ot ff x che 2# pi oe ym b che 4 nd 2 oe qn h tc 2 pi [primum] pr pi 4 7 T gp ff 2 hz gp hs ot + pi b fi 2 t V37 ls q pr + hk ff bp fi q hk 7 a gp to + c 7 hd gp x T + hk hs x + ph t 4 pr tc pi h b c oe pc ot 4 c x pi V39 x ph t 4 + hk oe h t c 6 pi che o x hk pc oe 2 8 hs 4 ym tc fi pi ym x h tc pi 4 nd a pi hs 2 ff tc gp 2# pi 7 pc ot + tc nd 4 4 hk ym + ot t h 7 pc 4 V40 pr x ot 7 ph gp hd c pi 4 q ffo oe bp h 7 ls che mea to b pi qn t # ff + qn hd pc pr 7 4 q nd k t nd 7 V40b h 7 bp pc T V41 7 4 t ff T ym oe nd gp V42 pi b [mo] hs ot 4 bp + [cio] hz oe nd 8 x hd 2 t bp x V45 pi x # nd 4 bp t tc V46 ls q hd c b pc + x gp pr hk x ls q nd hz + 2 t V46b h pi c + hd ot 4 bp V47 h 7 c gp ym 4 ls tc q h 7 fi pi 2# hz t ls tc V47b nd 7 hk gp c oe ph bp tc ym x c pi W oe T 7 hk ls oe # gp nd ot 4 h 7 ? hk 7 pc 2 + nd x [mo] V48 q x pi oe ym pi ls tc + hd ls tc 2# hz + ls 4 bp pc pr 7 hk x to + pi oe ff 7 bp 4 hk + b 2# b bp b h 7 c tc a pi hd pi mea pr hk 4 ym + k x gp x V49 2 pi ^ 4 ls tc nd b hz oe ym pi fi 4 + V50b ym + pr 7 hk [mo] h 4 ff pi ym gp 2# 3 7 W V51 x che pi x c ff oe h b hd u pr + hk 7 c pi gp ph 7 ff oe [cio] h 7 hk ot b pi 8 tc mea to + 8 7 ff 4 2 tc T pi pi hd 8 nd x 4 4 x ff tc nd V52 q pr 7 hk pc qn b # hk q qn a 7 [primum] pc pr + c 4 x hd x pi pc gp ls + pc hd [quia] 8 [tame] + hs T 7 ^ q h gp ym x 8 gp hk hs tc hd P1 2 hz + nd h x 4 t 8 oe hk q h gp ls 4 [adviso] q fi ot ff pi hd h tc x hs + 2# 8 q qn a q bp pc oe q [quia] x # nd 4 pi nd q che pi tc P2 hz b ls 7 6 gp qo bp 8 q qq 4 + che 4 + c 7 pc 2 pi hz fi ot 4 ym pi e h 7 pr b 2 ym gp nd ot tc + h oe [mo] pi oe nd pi 4 gp pr 4 oe pc P3 2 q pi h tc x pi b ls oe ph 7 che T b ot gp [quia] pc ot tc 4 + hk ls hd x q
Decryption
# guarna1447 - mechanical decryption of transcription.txt with key K3 (key.tsv, first value of each sign), no correction. # Clear words inside the cipher are dropped; ' | ' marks a cut at a clear word. One line per segment, same order as transcription.txt. H02 remediareaquestibenedetiincove H03 nventideliqoali H04 mostrandodeparnorfrhavu | servicared H05 ehache | hmnvosetra H06 malecotent H07 qvlncomitrovoseaiinquestomondoperrespetodelspaccio H08 delccceetcunonvedevasetuirequelieietichevosperavadalasv H09 paltote H10 vmhcriavahsenocoparationepetatotentochenonerastat H11 ilidipasatietlarasoneeraquesta | alidipasochiera H12 sealttotentosoloperlerasoniditemaolpresentesentendolinimetc H13 liseregrsiouoaialcareperlamampagnosenatstamlae H14 dtbitare | chelostatodehastavahngrandeperimola H15a moltperimalosaa H15b la H16 euhdantehavicolodelstatodehalainmaomieraintolerabile H17 perchequandovenelfactodelasvandose | andavailostatodeha H18 spsglcasesemprehaporiaremediarealifactivostriveramvrta H19 chequandoloteapgli | seruesehaariaversolasv L22b masil L23 statorehanonandosebeserramamaleinecoperagthlsesiperha L24a usiatrivi L25b dalasv L26 alcccepromedenonsenagrandisimenorevosideratocheperlaviadel L27 totelasvsipusalvareunon | desendervemainatolaltrege L28 ntidelaschaverelapiogloreriasavictoriachesiostatainicilia L29 grantempauperdarenidientioadchihavolutodiremale L30a las | gpacedemeterclostatosuoin | perimolunons L31 zualeprovasepoteseinminquestiondocheiosemaiaredeq L32 uelachehaintaltotecoexpresaexperientiadelafedesa L33 iinclinationeversolasv L34 ac | | bnavolontadequantolaschavignoersel L35 tvui | experientiachefaintcptoeltotevelasoiedeudevocione L36 versolasvnonsipsonadefenderedahominiperetersetudin L37 licetchefacopostotutalointegentedarmeetoloroadmacilorelamentedelas L38 uccdesignorenonvogliaceposrdetapersivanetpensrerimeae L39a relastatodelascrn | perimola L39b ualersidel L40 conepapapichehaveritifactilienocivostritractarelococevero L41a secodalimeritisuoiosei L41b eltotestara L42 nonsolopachentematotentisimoadognilastiapiacerepdetermina L43a trone L44b L45 lsispetora L45b utofiramenticheamminoncvisognavade L46 havemesimiltpasioucheseiosealcunodecantaaetdatiate L47 resumcionechevolesemeaerelaceementedelassoainsuspetodelas L48a spertempopertardosenepentiria L50b lnostroperfimoletedicemochedde L51 itulveruoognoscehamolttenelofactanostrastarerngran V1 perimolasiperlaposanodelinimimietperlapocaprovisionechesia V2 adlegentinostreusipersispetochehavemodealcunodelinstri V3a ma V3b ngghavemocomandatoetcomandomogniduvsi V4 nmcercichesevseisose V4b sfhafhadasaisphvchenoneainii V5 semohnhatilodelapestonachenonpoiiamopoitoresimilifacicheumas V6 ilacosaimamale V7 ueserelprinmipalmchehavesimoadrematarevadtrovare | vcheha V8 nonfacevadvgiochecoparendoioetnnomedelasvsetrovarianodeleche V9 asaidunecatam V9b perdvhrispet V10 uprimononsatonedispostoadsimiimilecose | losecodorivpetav V11 ptrnonaevistarvicaltiinicumiinquestapatriapercagargli V12 delaborsoquandoioherinegato V13 deboneparolemarigvdovsvsemaliiati V15 minminomesenoscevainto V16 hguidocantodifeche V16b soninclhcnatialasv | farian V17 iexetcirianopiuvolanterilneactoutoiostra V19 labefinto V20 venirevostroxsignore V21 sipsanohaverealtraquestinavemiliadica V22 cicheaspetamohaveredelagabelodelitialdubenedicecheguan V23 dalasvfosedequasetfvhiondeli | asoi V24 lascunonaspetasealtroquantitade | vche V25 facesepioucheevestorestanonpernonvolegemapernonpotereetchequan V26 dononpotestimenoietotalacopagniamenostiquelapartevhiosepos V27 hbilepurechenonvenrsticomancochecolacitadelacopagniamucvandoglive V28 nisticomenosenaunoperdemedefeputationeperhaetperlavostra V29 livostiimanmamentivuisognivchepiuilantericenefis V30 titopochi | oconhentechetcasoiperiarecosagratoahamalas V31 uirestasoloperiapotectia V32 nonmhmhedespfaciutopiimochepa V33 rutochesignoreceintesevolunterfirisposemipiuapertamentechenon V34 hafatolealtevoltedtpaileasoiphopiamelluulnivedelisato V35 inmonstrafevolantadelaventavostroquantunqueinloefetodelienihene V36 de | nongliealtrochequemocheintenditi | pvhepercuestavolta V37 loparrloredehamedeparsasaipiudomestici V39 saiaredambvchersecosimilimodietfndistriequestainfirmatadesi V40 ptesemiiamedelchenonhoinagranspeionane V40b desp V41 efarpmene V42 vostiavenoca V45 igniai V46 lacosaeprlanuaca V46b dicati V47 dememiliadeliquali V47b nerecesimocileperlegentiderestan V48 avemilialiqualisperhaverefraoqoodemidiinonprimae V49 titilinovemilia V50b maperdhrimeqvel V51 cheiocredouparemieseredertouoinonhaveritipiucnfirin V52 opersnogrande | spachiselasvaspetodemoersi P1 tvandohaveradelialtretdisaqcondosenognfinochevi P2 volebevariachefacestivlthmvmdepotmentiadevenuhepies P3 toetdivolesechepotestifarla
Reading
# guarna1447 - reading of Nicolò Guarna to Francesco Sforza, Milan, 21 May 1447 (BnF ital. 1584, ff. 153r-154r)
Read in part (5 Nov 2026). Counted by script on the lines below (cipher segments only; the clear frame of the letter is not counted): **852 words: 583 certain, 33 probable, 0 restored, 236 unresolved** (the unresolved count is an estimate, see below), so 616 = **72.3%** certain or probable. Under 95%: read in part. Our working notes gave 862 words and 72.6%; the recount by script on this file, counting each word split at a line end once, gives the figures above.
Conventions (spelling of the cipher kept; 15th-century Lombard Italian):
- a bare word is **certain** (given by key K3, clear sense);
- `{word}` is **probable** (one doubtful sign, or one of the sign's other values);
- `<word>` would be **restored** (none here);
- `[?n]` is a stretch of **about n unresolved words**: n is estimated as the length of the decoded stretch divided by 5, because a stretch that does not read cannot be cut into words. The total of 236 unresolved words is therefore an estimate.
The end of the letter (f. 153v from about V47b, and f. 154r) agrees word for word with the 15th-century clear text on f. 156 of the same day ("le gente de Cremona restano … fra otto o diece di, o no prima … piu uno fiorino per uno gran spacio … ultimum de potentia de venire presto"). That part is therefore **not new**: it was in clear in the volume. The rest (f. 153r and f. 153v to about l. 46) has no known decipherment.
## f. 153r — H (l. 2–19)
H02 remediare a questi benedetti inconve-
H03 -nienti de li quali
H04 mostrando de {parere} [?3] servi{re} [?1]
H05 [?2] che ne [?2] {vostra}
H06 [?1] {contento}
H07 [?2] {come} mi trovo [?1] in questo mondo per respecto del spaccio
H08 [?2] et {cum} non vedeva [?3] {che} {vostra} sperava da la S.V.
H09 [?1]
H10 [?2] senza comparatione [?1] tanto che non era stato
H11 li di passati et la rasone era questa [?1] li di passati {che} era
H12 [?2] tanto solo per le rasoni dite ma al presente sentendo li nimici
H13 [?3] per la campagna [?3]
H14 dubitare manifestamente che lo stato de [?1] stava {in} grande {pericolo}
H15 molto {pericoloso}
H16 [?2] del stato de [?2] era intolerabile
H17 perche quando venne el facto de la S.V. [?2] lo stato de [?1]
H18 [?2] sempre [?1] poria remediare a li facti vostri [?2]
H19 che quando [?2] {servesse} [?1] verso la S.V.
## f. 153r — outside H (l. 21–50)
L22 [?1]
L23 stato [?1] non andase bene {serra} [?1] male [?3]
L26 [?2] non senza grandissime [?1] desiderato che per la via del
L27 [?1] la S.V. si pò salvare non solamente defenderve ma [?1] laltre gen-
L28 -ti [?2] haver la piu gloria [?1] victoria che sia stata in {Italia}
L29 gran tempo [?2] ad chi ha voluto dire male
L30 [?2] {pace} de metere lo stato suo in [?1] {pericolo} [?1]
L31 [?2] prova se potese [?1] in questo mondo che io se mai [?2]
L32 quella che ha in tal [?1] expressa experientia de la fede [?1]
L33 [?1] inclinatione verso la S.V.
L34 [?2] volontade quanto [?2]
L35 [?2] experientia che fa [?2] devocione
L36 verso la S.V. non si possono defendere da homini [?2]
L37 licet che faza posto tuta [?1] in gente darme et loro [?2] de la S.
L38 [?2] de Signore non voglia [?3] {pensare} [?1]
L39 [?1] lo stato de la S. [?1] {pericolo} [?1] valersi del
L40 [?1] Papa [?1] che haveriti facti li [?1] vostri tractare [?1] vero
L41 secondo li meriti suoi [?1] el [?1] stara
L42 non solo [?2] tanto [?1] ad ogni [?1] piacere [?1] determina
L46 havemo [?3] che se io se alcuno de [?2]
L47 resolucione che volese [?2] de la S. in suspecto de la S.
L48 per tempo per tardo se ne pentiria
L50 el nostro [?1] dicemo che [?1]
L51 [?1] cognosce [?1] molto [?1] lo facto nostra stare in gran
## f. 153v (l. 1–46, cipher)
V1 {pericolo} si per la possanza de li inimici et per la poca provisione che sia
V2 ad le genti nostre [?1] per respecto che havemo de alcuno de li nostri
V3 [?1] havemo comandato et comandamo [?1]
V4 [?4]
V5 [?2] de la peste [?1] non possiamo [?1] simili [?2]
V6 la cosa [?1] male
V7 [?1] el principale che haveressimo ad remediare [?1] trovare spese et che ha
V8 non faceva [?1] che comparendo io et nome de la S.V. se trovariano de le [?1]
V9 [?3]
V10 [?1] primo non sa [?1] disposto ad simile cose [?3]
V11 [?3] in questa patria per {cavare} [?1]
V12 de la borsa quando io ho {negato}
V13 de bone parole [?3]
V15 [?3]
V16 [?5]
V17 [?2] piu {volunteri} [?3]
V19 [?2]
V20 venire vostro Signore
V21 si possono havere altra questi nove milia {ducati}
V22 [?1] che aspetamo havere de la gabella [?3] che quan-
V23 -do la S.V. fose de qua [?2] de li spesa soi
V24 [?1] non aspeta se altro quantita de spese et che
V25 facese [?1] che [?2] non per non volere ma per non potere et che quan-
V26 -do non potesti meno [?1] tota la compagnia [?1] la parte [?1] se pos-
V27 -sibile pure che non venesti come [?1] con la {cita} de la compagnia [?1] {quando} gli ve-
V28 -nisti [?1] per [?1] reputatione per [?1] et per la vostra
V29 li vostri {comandamenti} [?1] che piu [?2]
V30 [?2] poco spes{o} [?2] cosa grato [?1] la S.
V31 [?1] resta solo per [?1] potentia
V32 non [?3] che pa-
V33 -ruto che Signore [?1] volunteri [?1] risposemi piu apertamente che non
V34 ha fato le altre volte [?4]
V35 [?1] monstra {fervente} volontade de la {venuta} vostra quantunque in lo efecto [?2]
V36 de spese non glie altro che [?2] che intenditi [?1] per questa volta
V37 lo {parlare} de [?2] piu domestici
V39 [?2] cosi simili modi et [?2] questa infirmita de si
V40 [?3] che non ho in gran [?2]
V41 [?2]
V42 vostri [?1]
V45 [?1]
V46 la cosa [?2]
V47 [?1] de [?1] milia de li quali
V47b ne {recevessimo} [?1] per le genti de Cremona restano
V48 [?1] milia li quali spero havere fra otto o diece di o non prima
V49 [?1] tutti li [?1] nove milia
V50 ma per {dirve} quello
V51 che io credo pareme esser certo voi non haveriti piu uno fiorino
V52 per uno grande spacio se la S.V. aspecta de moverse
## f. 154r (cipher, l. 1–3)
P1 quando havera de li altri [?1] sa quando sera [?1] fino che vi [?1]
P2 vole bene voria che facesti ultimum de potentia de venire pre-
P3 -sto et Dio volese che potesti farlaTranslation
# guarna1447 - English translation (working translation of the passages read) Read in part: 72.3% of the cipher words certain or probable. "…" = a gap (unresolved stretch). S.V. / Your Lordship = Francesco Sforza; "the Lord" / Signore = Duke Filippo Maria Visconti. [153r, H] … to remedy these cursed troubles, of which … showing that he seemed … to serve … discontented … how I stand in this world on account of the dispatch [of your troops] … and since he did not see … what you hoped from Your Lordship … beyond comparison … so much that it had not been … in past days, and the reason was this: … in past days, that he was … only for the said reasons; but now, hearing that the enemies … through the countryside … to fear plainly that the state of … stood in great danger, very dangerous … of the state of … was intolerable, because when the matter of Your Lordship came … the state of … always … could remedy your affairs … that he should serve … towards Your Lordship. [153r, ll. 21–50] … the state … was not going well … badly … not without very great … wished that by way of the … Your Lordship can save yourself, not only defend yourself but … the other troops … have the greatest glory … the victory there has been in Italy for a long time … to whoever wished to speak ill … to put his state in danger … in this world that I if ever … she who has in such … express experience of the loyalty … inclination towards Your Lordship … will … devotion towards Your Lordship; they cannot defend themselves from men … although he has put all … into men-at-arms … that the Lord should not wish … think … the state of His Lordship … danger … avail himself of the … Pope … that you would have done … treat … according to his merits … not only … at every pleasure … we have … that if I, if any of … a resolution he might want … of His Lordship in suspicion of His Lordship; in time or too late he would repent of it … our … we say that … knows … our affair in great [153v] danger, both through the power of the enemies and the small provision made for our troops … on account of some of ours … we have ordered and order … the plague … we cannot … the thing … badly … the main thing we should remedy is to find money, and that … if I appeared in the name of Your Lordship, there would be found … first he is not disposed to such things … in this country to draw money from the purse, when I refused … good words … more willingly … your coming, Lord … one can have, besides these nine thousand ducats, … which we expect from the gabelle … when Your Lordship were here … at his own expense … not for want of will but for want of means, and that if you could not at least … the whole company … the part … if possible, provided that you come … with … of the company … when you come … reputation … and for your … your orders … little money … a pleasing thing … there remains only … power … the Lord … willingly … answered me more openly than he had done the other times … he shows a fervent wish for your coming, although in effect … of money there is nothing else but … that you mean … for this time … speak … more intimate … in similar ways, and … this illness … that I do not have in great … thousand, of which … we received … for the troops of Cremona remain … thousand, which I hope to have within eight or ten days, not before; … all the … nine thousand. But to tell you what I believe, I seem certain you will not have another florin for a long while, if Your Lordship waits to move [154r] until you have the others … knows when that will be … until … wishes you well would want you to do your utmost to come quickly, and God grant you may do it.
Key: key.tsv
06 Sources
- Paris, Bibliothèque nationale de France, italien 1584, ff. 153r–154r, Gallica btv1b100373864, views 146–147; f. 156 (clear text of the end); ff. 161r–165r (Guarna’s letter of 27 May 1447 and its late copy, the training pair).
- Brice Dudout, Europe chiffrée, with Claude Opus 5.5 agents: transcription, key (CC BY 4.0), tests and reading; pull request #280.
- Daniel Bourdeau, the survey of late copies in italien 1584, which gave the training pair.
Notes, transcription, decryption, key and tests: targets/guarna1447/.
Cite this page
Brice Dudout and Daniel Bourdeau, “Nicolò Guarna to Francesco Sforza, Milan, 21 May 1447”, Cyphersolver, posted 5 November 2026, https://cyphersolver.com/guarna1447.html.
