01 The dossier and what was on the records
Spagna 364D is a register of the Secretariat of State’s outgoing correspondence to the nunciature at Madrid, described by the archive as “Lett. Orig. e cifre della Segretaria al nunzio, dal 4 gennaio 1718 al 6 gennaio 1720, ff. 1–403”. DECODE holds twenty-four of its ciphered items as records R154 to R177, one per item, ninety pages in all, classed homophonic and nomenclator, numerical, images behind a login. The catalogue entry here counted twenty-one, following DECODE’s own grouping, which leaves out the three items dated to 1717 (R172, R174, R176); they are in the same dossier and the same cipher, and all twenty-four are treated together below.
The twenty-four items are very unequal. Item 14 alone runs to 9,751 cipher tokens over twenty-one pages; six items are under three hundred tokens, and two are a single short paragraph. The full measured table:
| item | record | tokens | distinct | item | record | tokens | distinct |
|---|---|---|---|---|---|---|---|
| 364D/1 | R154 | 2,954 | 177 | 364D/13 | R166 | 484 | 97 |
| 364D/2 | R155 | 768 | 101 | 364D/14 | R167 | 9,751 | 244 |
| 364D/3 | R156 | 3,837 | 159 | 364D/15 | R168 | 519 | 94 |
| 364D/4 | R157 | 167 | 53 | 364D/16 | R169 | 715 | 100 |
| 364D/5 | R158 | 3,537 | 169 | 364D/17 | R170 | 2,234 | 187 |
| 364D/6 | R159 | 229 | 62 | 364D/18 | R171 | 1,175 | 114 |
| 364D/7 | R160 | 131 | 49 | 364D/19 | R172 | 899 | 119 |
| 364D/8 | R161 | 567 | 80 | 364D/20 | R173 | 770 | 122 |
| 364D/9 | R162 | 1,034 | 95 | 364D/21 | R174 | 80 | 46 |
| 364D/10 | R163 | 2,468 | 137 | 364D/22 | R175 | 79 | 45 |
| 364D/11 | R164 | 296 | 72 | 364D/23 | R176 | 926 | 108 |
| 364D/12 | R165 | 531 | 91 | 364D/24 | R177 | 92 | 36 |
Each record carries three attachments. The first is the raw transcription of that item, made from the ASV
images in 2016–17 by the DECODE transcribers EHum and KaJo: a bare stream of single digits, with the
contemporary marginal glosses flagged in <PLAINTEXT> lines. The second is the key,
headed #KEY: reconstructed, #TRANSCRIBER NAME: George Lasry, #DATE OF
TRANSCRIPTION: October 24, 2020, catalogued as “ASV - S364” and dated 1717–1720.
The third, identical on all twenty-four records, is the 800 KB corpus file with the segmented and deciphered
text of both 364C and 364D.
This is the third target in this project where DECODE’s status field understated what was on the record: “partially decrypted” describes the contemporary marginal gloss and leaves out the modern decipherment filed beside it. The rule, applied here before any analysis, is to open every attachment on the record before treating the cipher as open.
02 The contemporary decipherment is fragmentary
The catalogue entry took DECODE’s note, “(Partly) deciphered within the letter”,
to mean the despatches were deciphered on arrival and that the modern job was to copy out the decipherment. On the
leaves themselves the gloss is thin: on 364D/1 it gives four to eight words over a page of solid cipher, set above
the lines it belongs to. It is also unreliable. The transcribers recorded it faithfully and it contains
<PLAINTEXT ? CvYeono>, <PLAINTEXT ? Ee> and <PLAINTEXT IT
Narzio/Nunzio?>; where Lasry’s key decrypts the line, the gloss is sometimes wrong. It could
not have supported a decipherment of these letters, and none was based on it.
It does serve as a check. Where a gloss and the key agree over a run of words (“La St Sua. Hora deuo parteciparle” against the decrypted la N.S. … ora devo parteciparle), the key is confirmed on the page by a hand of 1718.
03 The system
From Lasry’s key sheet, whose editorial notes state the design outright:
- The despatch is an undivided stream of digits, with no word division and no group separation.
- Ordinary elements are two digits; nomenclator elements are four digits beginning with 4. Division into groups is therefore deterministic: a 4 opens a four-digit group, anything else is a pair.
- The single digit 3 is a null, except when it falls among the three digits after a 4 prefix, where it belongs to the nomenclator group and is not a null. This is the one place the scheme is not context-free, and Lasry flags it.
- Homophony is modest for letters and heavy for punctuation:
50alone for a,17|71for i,12|21for c, but six variants for the full stop (303|323|343|363|373|393). - A few two-digit groups are syllables or short words rather than letters:
91|94= che,55= con,78|88= del,01= al.
The nomenclator is a series of runs, each ordered alphabetically within itself, and this is what makes the missing values recoverable:
| range | contents |
|---|---|
| 4001–4078 | particles, pronouns and word endings, loosely alphabetical (ancora, ben, come, essere, fare, molto, quando, quanto, questo, tutto, -zione) |
| 4079–4296 | nouns, titles and names, A–R (Arcivescovo, Armi, Auditore, Camera, Capitolo, Chiesa, Concistoro, Corte, Ecclesiastica, Guerra, Imperatore, Italia, Napoli, Pace, Regni, Religione) |
| 4401–4479 | the same alphabet continued, S–V (risoluzione, Roma, S. Sede, scrittura, sempre, Sig.re, Spagna, spedizione, subito, tempo, troppo, Turco, Ungheria, Vienna) |
| 4485–4531 | calendar words and the day numbers, 4501–4531 = 01–31 in plain order |
| 4532–4543, 4552–4563 | two parallel month runs, Gennaio to Dicembre (see below) |
| 4564–4619 | a second alphabetical run, A–T (aggiustamento, alcun, altro, attenzione, Breve, conforme, costì, dispaccio, giustizia, parola, possibile, pubblico, riflessione, stesso, tempi) |
| 4681, 4706, 4870+ | scattered later blocks (Cristian-, contro, espulsione, present-) |
04 How much Lasry’s decipherment covers, measured
The corpus file aligns each cipher group with its value in a column underneath. Parsing those columns gives a token-by-token count rather than an impression. Over the twenty-four items of 364D:
| cipher tokens | valued | unvalued | read | |
|---|---|---|---|---|
| Lasry’s filed reading | 34,243 | 33,408 | 835 | 97.56 % |
| with the groups recovered here | 34,243 | 33,521 | 722 | 97.89 % |
Most of that 2.4 per cent residue comes from the transcription: a long tail of two-digit
groups they could not read off the page, recorded as 70?, 00?, 6?, and
single stray digits at line ends. A hundred and thirty-five of the 151 distinct residual groups occur once or
twice. Closing them needs the images; the key is not at fault.
The nomenclator is a different matter, and it is the part that was unfinished:
| distinct 4-digit groups in use | without a value | tokens affected | |
|---|---|---|---|
| 364D alone, Lasry | 279 | 97 | 198 (8.0 %) |
| 364D alone, with this session | 279 | 64 | 85 (3.4 %) |
| whole S364 corpus, Lasry | 399 | 189 | 715 (11.0 %) |
| whole S364 corpus, with this session | 399 | 135 | 221 (3.4 %) |
05 Method: brackets and contexts
No cryptanalysis was needed and none was done. The method has two halves, and a value was only accepted when both agreed, or when one of them was overwhelming on its own.
The bracket. Because each run is alphabetical, a group without a value lies between its nearest
neighbours that have one. 4467 sits between 4466 troppo and 4468 Turco, so it begins
tr… to tu…. 4198 sits between 4197 maggior and 4200
necessari. 4233 sits between 4231 occasione and the opportun recovered at 4236,
so it is of….
The contexts. The corpus is 84,961 tokens with everything else already decrypted, so each group without a value comes with its sentences. 4467 appears in “far retrocedere le numerose — che si dicono già incamminate” and “presi nelle proprie case con — di fanteria”. Both halves give truppe.
The agreement of the two is the check. 4469 lies between Turco and Vescovo, so it is va–ve; its contexts are “li — di Spagna hanno preso camino”, “il soccorso de’ — da mandarsi in Levante” and a list of “due palandre et undici barche”. Vascelli satisfies both and nothing else does.
The calendar block, and the control it supplies. The month codes were recovered together,
as a structure. Lasry’s sheet has only three of them: 4535 “Apte”,
4539 Agosto, 4562 Novembre. The corpus uses 4533 to 4539 contiguously and 4560 to 4562,
always after a day number and before passato or corrente. Two parallel Gennaio–Dicembre
runs at 4532–4543 and 4552–4563 place Agosto at 4539 and Novembre at 4562 exactly, and make
Lasry’s “Apte” Aprile, the fourth slot.
That scheme puts Luglio at 4538, and one sentence of the corpus confirms 4538 = Luglio. The corpus has “in data delli 4499 4538 passato, qual appunto fu il giorno della promozione del Sig. Card.” Giulio Alberoni was created cardinal by Clement XI on 12 July 1717. The month run, derived from an offset, lands on Luglio; the day code beside it is 12; and the sentence names the day of the promotion. Three independent things agree, and 4499 = 12 follows.
06 The sixty-one groups
Each value below is carried by at least two independent contexts, and in most cases by the alphabetical bracket
as well. They are in nunzio1718/key_extra.txt in the repository, in the format of Lasry’s own
key file.
| group | value | evidence |
|---|---|---|
| 4007 | stesso | “in questo stesso dispaccio”; “lo stesso Sig. Card. Alberoni” |
| 4008 | ciò | “è vero che ciò seguirebbe”; “spero che ciò sia per succedere” (26 tokens) |
| 4013 | già | “già è abbastanza risposto”; “non sarebbe già che un deplorabile effetto” |
| 4030 | quel | “diverso da quel prescrittomi negli indulti” |
| 4033 | ogni | “superare ogni renitenza”; “per togliere ogni equivoco” |
| 4082 | articoli | “i due consaputi articoli”; bracket Armi (4081) – Auditore (4084) |
| 4086 | avvis- | “divulgato ne’ pubblici avvisi”; “si contenti di avvisarmi” |
| 4101 | cifra | “con la sua cifra de 12”; “l’altra cifra unita a questo dispaccio”; bracket Christianità – Commission |
| 4110 | Congregazion- | “la Congregazione Concistoriale”; “i padri della Congregazione di San Filippo Neri” |
| 4113 | Card.le | “il Sig.e Card. Alberoni”, 25 tokens |
| 4114 | corriere | “col ritorno del corriere”; “la spedizione fatta costì del corriere sotto li 31” |
| 4172 | general- | “il Capitolo generale”; “il vicario generale di quella Chiesa” |
| 4174 | giurisdizion- | “nel libero esercizio della sua giurisdizione”; “l’autorità e la giurisdizione Ecclesiastica” |
| 4176 | governo | “per sua regola et governo”; “al buon governo di coteste Chiese” |
| 4191 | istruzion- | “le precise istruzioni che N.S. le diede”; bracket Imperatore – Italia |
| 4193 | leg- | “il Sig. Card. Legato di Ferrara”; “una lega da Valenza” |
| 4198 | maniera | “nella maniera già detta”; bracket maggior (4197) – necessari (4200) |
| 4210 | Monsignor | stands where the parallel plain copy reads “Mons.e Nunzio” |
| 4216 | negoziat- | “tutti i suoi negoziati conclusi in questa Corte”; bracket necessaria – notizie |
| 4230 | obblig- | “l’obbligazione che gliene avrà a N.S.”; “divotissimo et obbligatissimo servitore” |
| 4233 | offici- | “il tribunale et officiali della Nunziatura”; bracket occasione (4231) – opportun (4236) |
| 4236 | opportun- | “gli opportuni provvedimenti”; “valersene opportunamente” |
| 4238 | ordin- | “in ordine alla reintegrazione”; “disordini et inconvenienti” (36 tokens) |
| 4239 | pace | “turbare la pace della Christianità”; “pace o tregua” |
| 4262 | particolar- | “una giunta particolare deputata dal Re”; bracket N.S. (4261) – pendenz (4264) |
| 4270 | port- | “portarsi alla Corte”; “seco portati da Roma”; “importantissimo affare” |
| 4279 | promozione | “in proposito della promozione del Sig. Card. Alberoni”, 45 tokens — the commonest unvalued group in the corpus |
| 4290 | Re Cattolico | “specialmente dell’Imp.re et del —”; bracket Re (4289) – Spagna (4293). Less certain than the rest |
| 4403 | risoluzion- | “prenderà quelle risoluzioni che saranno più proprie” |
| 4408 | scriv- | “mi ha fatto l’onore di scrivermi”; “scrivercene subito”; bracket scrittura (4407) – S.M. Ces.a (4410) |
| 4437 | subito | “subito che si fosse concluso”; “risolversi subito, altrimenti” (a second homophone beside 4447) |
| 4462 | tratt- | “avesse potuto ivi trattare gli affari secondo le sue istruzioni” |
| 4463 | Trattato | “la conclusione del Trattato”; “il Trattato della neutralità d’Italia” |
| 4465 | tribunal- | “il tribunale della Nunziatura di Spagna”; “registrata in quel tribunale” |
| 4467 | truppe | “le numerose truppe già incamminate”; bracket troppo (4466) – Turco (4468) |
| 4469 | vascelli | “li vascelli di Spagna hanno preso camino”; bracket Turco (4468) – Vescovo (4470) |
| 4474 | Vicerè | “un espresso spedito dal Sig. Vicerè di Sardegna al Sig. Vicerè di Napoli” |
| 4482 | corrente | “li 31 del corrente” and the other day-of-month formulae. Probable rather than certain |
| 4491 | quattro | “li quattro motivi addotti”; sits beside cinque (4492) |
| 4499 | 12 | “in data delli 12 Luglio passato, qual appunto fu il giorno della promozione” |
| 4564 | aggiustamento | “il bramato aggiustamento delle note pendenze”; “il consaputo aggiustamento” |
| 4565 | accennat- | “la seconda parte sopra accennata”; “gli accennati due indulti” |
| 4569 | bisogn- | “di non averne bisogno, et quando ve ne fosse bisogno”; bracket attenzione (4568) – Breve (4570) |
| 4573 | contrari | “furono sì contrari alla risoluzione presa dall’Imp.re”; bracket conforme – costì |
| 4583 | giust- | “niun Re potrebbe giustamente pretendere”; “senza darsi alcuna giusta occasione di querela” |
And the calendar block, two parallel runs. Lasry’s own three values, italicised, fix both offsets:
| Gen | Feb | Mar | Apr | Mag | Giu | Lug | Ago | Set | Ott | Nov | Dic | |
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| run A | 4532 | 4533 | 4534 | 4535 | 4536 | 4537 | 4538 | 4539 | 4540 | 4541 | 4542 | 4543 |
| run B | 4552 | 4553 | 4554 | 4555 | 4556 | 4557 | 4558 | 4559 | 4560 | 4561 | 4562 | 4563 |
07 Two corrections to the filed key
Working through the runs turned up two errors in Lasry’s sheet, both of the kind that comes of reconstructing a key from short contexts rather than a whole corpus.
4474 is Vicerè. His entries 4470–4477 read Veso, Vesconi,
Vescoua, Viscon, Vesconi|Vienna, Viscoui: garbled
placeholders around a Vescovo / Vescovi group. 4474 sits inside that run, and its contexts name a viceroy: “un espresso
spedito dal Sig. Vicerè di Sardegna al Sig. Vicerè di Napoli”, and again
“ordine da legato al Vicerè di Valenza di esiliare il detto vicario generale”. It is
Vicerè, and it belongs alphabetically where it sits.
4535 “Apte” is Aprile. The abbreviation came from a heading; the calendar run shows it is the fourth month slot, plainly written.
08 What the despatches are about
These are Clement XI’s Secretariat writing to the nuncio at Madrid in the two years when relations with Philip V collapsed: Alberoni’s Sardinian and Sicilian expeditions, the Quadruple Alliance, and Spain’s retaliation against the Roman jurisdiction. Two words run through everything and both of them were among the groups without a value, which shows what the gap in the key cost. The first is promozione, the 45-token group: whether Rome would make the promotion Alberoni wanted, and whether it should be held back as leverage. The second is aggiustamento, the settlement of the “note pendenze”: the giurisdizione, the tribunale of the nunciature in Spain, the abolition of the Nunziatura at Naples, the espulsione of the nuncio. Around them run the vascelli of Spain taking their course for the Levante, the Emperor’s truppe in the Ferrarese, the Turco, and the Trattato of the neutrality of Italy.
A specimen of the decrypted text from 364D/3, where the transcription is clean and every group in the span has a value. The Secretariat writes on the Spanish ambassador at Rome, Cardinal Acquaviva:
… Card.e Acquaviva, a cui non solo per mezzo mio ma da sé stesso ha procurato N.S. di fare conoscere l’indispensabile necessità di tal sospensione. Non mostrando d’appagarsene, pretese di rimuoverne N.S. con dedurre le sue ragioni compilate nella scrittura, di cui, non con altra mira, se le manda copia che per fare vedere costì la loro insussistenza et la stranezza delle [ragioni] con le quali si sostengono…
09 What remains open
- 135 nomenclator groups still without a value, in 221 tokens, 3.4 per cent of nomenclator usage across the corpus. Almost all occur once to four times, which is why they remain open: a single context, often in a badly transcribed line, does not settle a value against a bracket several letters wide. The frequent ones among them are 4434 (13 tokens, somewhere in the S run), 4130 (10, a person named “il Sig.e —” who carries news to N.S., bracketed D–E), 4795, 4340, 4213 and 4063, all in passages the transcribers left half-legible.
- 722 tokens unvalued in 364D overall (2.1 per cent), the great majority of them two-digit groups the 2016–17 transcribers could not make out. Closing them needs the images, which are behind the DECODE login.
- Volume 364C (items 1–32) is covered by the same key and the same corpus file and was used here only as context for the key extension. Its decipherment has not been checked item by item here.
- Three of the twenty-four items (R172, R174, R176) are dated 1717 and fall outside the dossier’s stated span; they are in the same cipher and were measured with the rest.
10 Sources
- Archivio Segreto Vaticano, i. 1025, Segretario di Stato, Spagna, dossier 364D, ff. 1–403: “Lett. Orig. e cifre della Segretaria al nunzio, dal 4 gennaio 1718 al 6 gennaio 1720”.
- DECODE records R154 to R177 (24 records, 90 pages). Transcriptions by EHum and KaJo, 2016–17. Images and attachments require a login.
- George Lasry, reconstructed key “ASV - S364”, 24 October 2020, and the segmented decipherment of Spagna 364C/1–32 and 364D/1–24, both filed as attachments on every one of those records. The decipherment of these letters is his; this page measures it and extends his key.
- Copies of the key, the corpus decipherment and the per-item transcriptions, with the parsing and measurement scripts, are in the repository folder.
Working files, the recovered key values and the measurement scripts: nunzio1718/.