Unsolved ciphers / write-up and files
What is enciphered
Not whole letters. Vande Perre wrote in Dutch and put only the sensitive phrases in cipher, a few words at a time. Thurloe’s office intercepted the letters and translated the clear parts into English, leaving the ciphered phrases as numbers; Birch printed the translations that way. The verified outbound inventory contains twenty ciphered runs across five letters, 322 printed groups in all, including the continuation on p. 523. This is a checked inventory of these passages, not proof that no further cipher occurs elsewhere in the correspondence. The longest is 70 symbols, one Dutch sentence (P500.2); most are one to five words. The alphabet reads most of these fragments; two short strings remain unresolved.
Below, each run as it stands on Birch’s page, then the decoder’s output, the Dutch and an English translation. In the decoded line [..] marks a proposed repair or inferred value; in the Dutch and English, square brackets hold editorial supplements. English labels in braces identify code meanings without claiming to recover their exact Dutch wording. The small English words over some runs are glosses transmitted by Birch from the contemporary translations; these passages had already been read in 1653. Each image links to the full page on the Internet Archive.
The ciphered passages
P431.1 · p. 431 · Westminster, 29 August 1653
- Decoded
ickseggheeenc[ij]ffer- Dutch
- ick segghe een cijffer
- English
- I mean a cipher
P500.1 · p. 500 · Westminster, 3 October 1653
- Decoded
in[ee]nviandichland- Dutch
- in een viandich land
- English
- in a hostile country
P500.2 · p. 500 · Westminster, 3 October 1653
- Decoded
eenredelijckedisposi[t]ietothandelingewertgespe[u]rtdieweldiendewaergenomen- Dutch
- een redelijcke dispositie tot handelinge wert gespeurt, die wel diende waergenomen
- English
- a reasonable disposition to negotiate is noticed, which ought to be taken up
P500.3 · p. 500 · Westminster, 3 October 1653
- Decoded
regeringemetimpatientieopnaderordrevano[n]sverwac- Dutch
- regeringe met impatientie op nader ordre van ons verwac[hten]
- English
- the government await further orders from us with impatience
P522.1 · p. 522 · Westminster, 10 October 1653 NS
- Decoded
degoededisposi[t]ie- Dutch
- de goede dispositie
- English
- the good disposition
P522.2 · p. 522 · Westminster, 10 October 1653 NS
- Decoded
diemnicvndament- Dutch
- unresolved
- English
- Birch: “who upon good grounds”
P522.3 · p. 522 · Westminster, 10 October 1653 NS
- Decoded
wijgeq[u]alifi[c]eerteenigenaderepropositi[e]doen- Dutch
- wij gequalificeert eenige nadere propositie doen
- English
- we [were] authorised to make some further proposal
P522.4 · p. 522 · Westminster, 10 October 1653 NS
- Decoded
tenprincipaln- Dutch
- ten principal [e]n
- English
- chiefly, and
P522.5 · p. 522 · Westminster, 10 October 1653 NS
- Decoded
[14]daghen- Dutch
- 14 daghen
- English
- fourteen days
P523.1 · p. 523 · Westminster, 10 October 1653 NS
- Decoded
{Major-General Harrison}- Dutch
- wording not recovered
- English
- Major-General Harrison (printed gloss)
P523.2 · p. 523 · Westminster, 10 October 1653 NS
- Decoded
ana[ba]ptisten- Dutch
- ana[ba]ptisten
- English
- Anabaptists
P523.3 · p. 523 · Westminster, 10 October 1653 NS
- Decoded
{Denmark}- Dutch
- wording not recovered
- English
- Denmark (printed gloss)
P523.4 · p. 523 · Westminster, 10 October 1653 NS
- Decoded
{queen of Sweden}- Dutch
- wording not recovered
- English
- queen of Sweden (printed gloss)
P576.1 · p. 576 · London, 4/14 November 1653
- Decoded
onstotvoordeelgerekent- Dutch
- ons tot voordeel gerekent
- English
- reckoned to our advantage
P582.1 · p. 582 · Westminster, 11/21 November 1653
- Decoded
thien[s]chepen- Dutch
- thien schepen
- English
- ten ships
P582.2 · p. 582 · Westminster, 11/21 November 1653
- Decoded
Jarmuyengeloope[n]- Dutch
- [Jarmuyen] geloopen
- English
- [at Yarmouth] run [away]
P582.3 · p. 582 · Westminster, 11/21 November 1653
- Decoded
bewaren- Dutch
- bewaren
- English
- to guard
P582.4 · p. 582 · Westminster, 11/21 November 1653
- Decoded
ote- Dutch
- unresolved
- English
P582.5–6 · p. 582 · Westminster, 11/21 November 1653
- Decoded
gebreck | masten- Dutch
- gebreck … masten
- English
- want [of] … masts [found]
Birch printed short English glosses over some runs: “the good dispositions do”, “who upon good grounds”, “we were qualified with some farther propositions that would do”, “chiefly”, “fourteen days”, “ten ships at Yarmouth”, “were run away”, “left”. They show that Thurloe’s office read at least these passages in 1653. The alphabet was found without them. With the explicitly listed repairs, it reads wij, gequalificeert, nadere propositie, principal, thien schepen and geloopen under the matching glosses. The London letter of 4/14 November has no gloss: its run reads “The said change [in the council] is ons tot voordeel gerekent”, reckoned to our advantage.
Key
304 of 322 printed groups map through the cryptanalytic key (grade S), 5 are read from contemporary glosses (C), and 13 are proposed repairs or inferred values (I). These count token mappings, not securely understood meaning: the letters of both unresolved fragments are included in S. No surviving key has been located.
| 3 | 4 | 5 | 6 | 7 | 8 | 9 | 10 | frac | 12 | 13 | 14 | 15 | 17 | 21 | 22 | 23 | 24 | 26 | 27 | 28 | 29 | 77 |
|---|
| a | b | c | d | e | f | g | h | ij | i | k | l | m | n | o | p | q | r | s | t | v | w | ee |
The five C groups are 617, the article de or die; 115, Yarmouth (the Dutch spelling Jarmuyen is supplied); 109, Major-General Harrison; 76, Denmark; and 52, the queen of Sweden. The p. 523 letter says Harrison and the Anabaptists are losing influence, and that Colonel Wurts, claiming to act for Denmark, receives a Swedish pension. These meanings were already supplied by the historical translators.
The printed 77 reads ee, but an intentional digraph and joined 7.7 are both possible. The singleton 57 is read as ee by inference. In P523.2, 43 is left unknown in the key and provisionally read ba in ana[ba]ptisten; splitting it into 4.3 gives those letters, while a syllable code is another possibility. frac denotes the fraction-like printed sort read as ij.
A conditional key-shuffle test on the 322 printed groups gives z = 18.0, p = 0.001 (1,000 shuffles, repository Dutch corpus). Code and unknown boundaries, ij and ee are held fixed; only the single-letter S values are shuffled, and no repairs are applied. This compares the chosen key with random relabellings, not with all keys obtainable by a search, and gives no probability that an individual repair is correct. A separate ciphertext-only annealing experiment on the archived 308-group discovery corpus recovers 17–18 of 22 letter values in its best seeds. See the method notes.
Proposed repairs
The printed digits are kept as they are in transcription.txt; each change below is applied only in the reading.
| Run.position | Printed | Read as | Reason |
|---|
| P500.2.20 | 14 | t | dispositie: 14 (l) stands for t here and again in P522.1 at the same place in the same word; 27 is t everywhere else. A slip in the writer's or the decipherer's key for this word, or the printer's; the text needs t |
| P500.2.45 | 17 | u | gespeurt: printed 17 (n). u is also wanted in gequalificeert (P522.3); either u was written with the n-symbol or the print confuses 16 and 17 (16 is never printed) |
| P500.3.40 | 27 | n | 'van ots verwac[ht]': ons is required; printed 27 (t) for 17 (n) |
| P522.1.14 | 14 | t | dispositie, as P500.2 position 20 |
| P522.3.6 | 17 | u | gequalificeert, as P500.2 position 45 |
| P522.3.12 | 15 | c | gequalificeert: printed 15 (m) for 5 (c); the 77 that follows is the ee symbol |
| P522.3.37 | 71 | e | propositie: '71' for '7.' with a stray 1; the next word, doen, is clean |
| P522.5.1 | 14 | 14 | the numeral fourteen left in clear inside the run: Birch glosses 'within [14 daghen]' as 'fourteen days'; as the l-symbol it gives no Dutch word |
| P582.1.6 | 27 | s | thien schepen: printed 27 (t) for 26 (s); Birch's gloss 'ten ships' |
| P582.2.9 | 1 | n | geloopen: '1' at the line end is 17 (n) with the 7 dropped; Birch's gloss 'were run away' |
| P431.1.13 | 3 | ij | een cijffer (a cipher): printed 3 (a) where the ij symbol is needed; the context ('a character to open a lockt chest, I mean a cipher') is not in doubt |
| P523.2.4 | 43 | ba | Splitting printed 43 into 4.3 gives ba under the existing alphabet and completes anabaptisten, agreeing with the English gloss. A dedicated ba group is an alternative; neither mechanism is established from this singleton. Do not add b or ba as a secure cryptanalytic key value. |
Open points
- P522.2
diemnicvndament is glossed “who upon good grounds”. The conjecture die met fvndament requires three consecutive substitutions: position 5, 17 → 7 (n → e); position 6, 12 → 27 (i → t); and position 7, 5 → 8 (c → f). Normalizing v to u is separate. The original wording is unresolved.
- P582.4
ote, after English “fleet”, is unresolved. A retained Dutch word ending, [vl]ote or [vlo]ote, is a possibility: the paragraph also juxtaposes English “to look to them” with Dutch bewaren, and “want of masts” with gebreck van masten bevonden. No prefix is encoded by the surviving three groups.
- P522.4 literally reads
tenprincipaln. The proposed ten principal [e]n supplies an e; word division alone does not produce it. The boundary and exact wording remain uncertain.
- The letter values have no u: 28 is v, and gespeurt and gequalificeert print 17 (n) where u is wanted. Either u was written with the n symbol or the print confuses 16 and 17 (16 never occurs).
- The 29 August letter announces a cipher sent “by a third hand”: “a character to open a lockt chest, ick segghe een cijffer”. The cipher was discussed again on 5 September and its safe receipt acknowledged on 3 October. These references do not establish whether the enclosure contained an alphabet, a code list, or both.
- The Beverning and Vande Perre letter to Boreel of 1 September 1653 (p. 435) uses a different cipher and remains unread.
Further manuscript evidence
The strongest next source is Bodleian MS. Rawl. A.7, old p. 47: Van de Perre to Bruyne, 17 October 1653, explicitly catalogued as a copy partly in Dutch and partly in cipher. Macray’s 1862 catalogue, col. 12, item 4, lists it among material omitted from Birch. The manuscript has not been seen, its calendar style is unspecified, and use of this key is unverified.
Two Zeeland bundles are verified catalogue leads: access 2, inventory 3188, Johan de Brune’s incoming letters, 1652–1658; and inventory 3101.1, England correspondence, 1652–1656, explicitly including copies returned by the other envoys after Van de Perre’s death. Both public image manifests were empty on 19 September 2026. Neither bundle’s contents have been inspected.
For the damaged October passage, Birch p. 522 prints the manuscript reference “Vol. X p. 471”, although A.10 mainly covers December 1653–February 1654. A.6 contains other 10 October letters at pp. 463 and 474, making A.6 p. 471 a plausible alternative to check, not a verified correction. The November passage points to A.8 old p. 102. Old pagination needs confirmation against modern foliation.
The reverse letter, De Bruyne to Van de Perre, 31 October 1653 NS, Birch p. 552, contains 118 without a resolving gloss. It is a likely code group with unknown meaning and unverified key, recorded separately from the 322 outbound groups.
Searches of Birch’s OCR, Colenbrander’s first-war documents, the First Dutch War indexes and relevant text, and the Bodleian and Zeeland catalogues found no parallel wording that resolves either damaged fragment. These were targeted searches, not an inspection of every manuscript or printed page. The 1725 embassy Verbael remains incompletely inspected. The evidence audit records coverage and retained sources; the retrieval specification identifies the next documents. No key or target manuscript image has been obtained.
Sources and files
Thomas Birch, ed., A Collection of the State Papers of John Thurloe, vol. 1 (London, 1742), Internet Archive collectionofstat01thur:
- p. 431, Westminster, 29 August 1653
- p. 500, Westminster, 3 October 1653 NS
- p. 522, Westminster, 10 October 1653 NS
- p. 523, Continuation, 10 October 1653 NS
- p. 576, London, 4/14 November 1653
- p. 582, Westminster, 11/21 November 1653
The envoy is identified as Paulus in Birch’s index and the NNBW biography.
Satoshi Tomokiyo, Unsolved Historical Ciphers, “Dutch ciphers (1653)”. The page is generated from transcription.txt, key.json and repairs.json in the repository, where verify.py reproduces the reading and solve.py the ciphertext-only search.