- Document
- London, British Library, Add MS 72438, f. 10 (DECODE record R8624). [Sir Edward Nicholas, Secretary of State,] to King Charles I, Oxford, 13 May 1646. The letter was written in Oxford during the siege and taken by Fairfax's army. The name of the writer is our inference.
- New
- The first connected reading of the letter: 89 % of the tokens read. The key is rebuilt from the printed cipher of the King with Captain Titus of 1648: the numbers of 1646 are the Titus numbers with a small shift.
- Not ours
- The cipher text is the transcription of A. Aymeloglu, made in one pass from the DECODE image (aaymeloglu/unsolved-ciphers, folder
royalist-1646). He also had about 45 code values from the glosses of Nicholas. We did not see the manuscript, and we have no image.
The key
The cipher has 261 codes in the rebuilt table: numbers for single letters, for common words and syllables, for names, and nulls. The word list is the list of the Titus cipher, which George Hillier printed in 1852 (Narrative of the Attempted Escapes of Charles the First, pp. 153-161). The letter alphabets of the two keys differ; the one of 1646 was solved from context.
| Words | Shift | Examples: word, number of 1648 → number of 1646 |
|---|---|---|
| a to h (an, and, be, for ...) | +4 | an 102 → 106, as 104 → 108, at 105 → 109 |
| i to n (I, if, in, it, me, my, not ...) | 0 | it 210 → 210, is 211 → 211, in 212 → 212 |
| o to t (of, or, send, so, the, to ...) | +1 | our 284 → 285, so 335 → 336, some? 337 → 338 |
| w (was, where, which, with ...) | +8 | was 397 → 405, will 399 → 407 |
| write, you, numbers, days and months | +9 | one 634 → 643, two 635 → 644, four 637 → 646, five 638 → 647 |
| second list of names and nouns | +10 | desire 479 → 489, letter 532 → 542 |
The codes of the three worked lines, in order of number. Third row: the basis of the value. E = gloss by Nicholas in the print of Evelyn (iv. 178-179); E2 = our reading of the part of the King's letter of 16 August 1646 that Nicholas left without gloss; T = Titus cipher with the shift, and it fits the context; C = context only; ? = doubtful; U = not read.
Full table: key1646.tsv (261 codes, each with its basis). The Titus key: titus_key.tsv.
Worked example
We have no image of the manuscript. The numbers below are three lines of A. Aymeloglu's transcription, file royalist-1646/f10_ct.txt of his repository. The rest of his transcription is not copied here. Each box holds one token: the number, its value, and the basis of the value. Words in italics are clear words of the manuscript. The line under a group of boxes gives the word.
Line 3: the summons of Fairfax
to render Oxford to him for the use {of ye} Parliament , and {express}ing that he may have honorable terms {for} himselve and all with{in ye} garrison if he [...] {reasonably} accept . t r(?)
And he [sent] to have Oxford rendered to him for the use of the Parliament, and said that the Governor may have honourable terms for himself and all within the garrison if he accepts in good time.
The key gives hav and al. The reading file writes "have" and "all".
Line 4, first part: the safe conduct
directions was {That ye} y e (they) desired {a} safe conduct for Sir Io. Munson and {Mr} <629: Warwick> to come to <384> with Fairfax
The direction was that they asked a safe conduct for Sir John Monson and Mr [Warwick] to come to [confer] with Fairfax
The reading file adds the gloss "(they)" and writes "come" for com. The codes 629 and 384 are not read. The reading file gives "Warwick" for 629 as a guess; the printed answer of the governor names "Master Philip Warwick".
Line 6, first part: no word from the King
{what} condition , for since your {depart}ure , we have not received {a} word from your Majesty
for since your departure we have received not a word from your Majesty
The key gives hav. The reading file writes "have".
Marks of the reading file: {..} = clear words of the manuscript; (?) = doubtful; <n> = code not read; [...] = illegible in the transcription.
How sure
| Measure | Value |
|---|---|
| Tokens in the transcription | 735, of which 20 are illegible there |
| Tokens read with no doubt mark | 652 (89 %) |
| Tokens read with a doubt mark | 43 |
| Tokens not read | 20 (17 codes) |
| Words that Nicholas glossed in 1646 and that fit the Titus list with the shifts | 20 of 20; with the Titus values shuffled 20,000 times: 5 at most |
Outside check. The printed papers of 11 May 1646. Fairfax's summons says: "honourable termes for your selfe, and all within the Garrison, if you seasonably accept thereof". The governor's answer asks "a safe conduct for Sir Iohn Mounson, & Master Philip Warwick". Lines 3 and 4 above give both. On 2 June 1646 the King wrote to Nicholas that he had received "but one letter from you, wch was of the 5th of May"; this letter begins "Since mine of ye fift of May".
What is not resolved
- 17 codes are not read (20 tokens), and 20 tokens are illegible in the transcription.
- The transcription had one pass, and it is not ours. The decipherment had one pass. A second, blind pass needs the image.
- 130 of the 261 code values rest on context only. The key table marks them.
Reading in words, with an English version: reading_f10.txt. Folder: oxford-1646.