The comte d’Affry to the Marquis de Stainville, August 1757

This letter was filed with d’Affry’s letters to the foreign secretaries Rouillé and Bernis, which share one code whose key was rebuilt from the Dutch codebreaker Lyonet’s decipherments. It is in a different cipher; those letters have their own page: the comte d’Affry’s intercepted letters, 1757–58.

01 The document

D’Affry’s despatches went by the Dutch post, and the stadholder’s cabinet had them opened, copied and passed to Pierre Lyonet, the Republic’s cipher secretary. Inv. 192 of the Prins Willem V archive holds the copies of nineteen of them, catalogued on DECODE as R1052–R1076 and R2067. All but this one are to Rouillé and then Bernis at Versailles, January 1757 to January 1758. This one is to Stainville at Vienna, written days after the French victory at Hastenbeck (26 July), when France and Austria were fighting Prussia and Hanover together. The copy carries the number 485 and the date 4 August; DECODE gives 7 August.

The three pages are photographed sideways and hold only figures, separated by dots. The groups used here are DECODE’s transcription, not re-read from the images. Comparing the first page with it for the figure above showed two differences: the image has 240 where DECODE has 24 in the first line, and 1116 and 1158 at the end of the seventh line are run together in the transcription.

02 How it was tried

• A comparison of group frequencies across all the inv. 192 letters set this one apart. The letters to Paris, first split into two codes the same way, turned out to share their common values: one code.
• The key of that code, 651 groups recovered by aligning Lyonet’s decipherments with their cipher, gives nonsense on R1071.
• On 27 October 2026 every machine-readable key of the period in this project was applied to R1071 and scored against shuffled copies of itself: d’Affry’s code, the Van Swieten–Cobenzl table, the Hanoverian army code of 1743, the Prussian code of the Michell letters (1751) and Croiset’s Dutch codebook of 1765. The same test picks out d’Affry’s key on one of his own letters at once; on R1071 none fits (keytest_R1071.py).
• Eight of the sibling records carry Lyonet’s clear copy among their images. R1071 does not.
• The DECODE catalogue was searched for other records of d’Affry, Stainville or Choiseul: only the inv. 192 letters and a Choiseul record of 1763–64 in the British Library, and no key.
• Karl de Leeuw’s history of the Dutch black chamber (2000) says Lyonet broke d’Affry’s first code by June 1756 and “two more French ones”, and d’Affry’s code to Choiseul from December 1758 later. Bussemaker’s 1906 extracts of Lyonet’s decipherments do not give this letter’s text.

03 What would settle it

Lyonet’s decipherments of d’Affry’s letters for July 1757 to December 1758 are in the Nationaal Archief, Stadhouderlijke Secretarie 1.01.50 inv. 223, “grouped by the cipher used”, and not digitised. They, or the French embassy code for Vienna in 1757, would give the key. Without them, 761 groups of a large code with no crib are too few for a reconstruction from the ciphertext alone. With either, the letter should first be transcribed again from the images.

04 Sources

• Koninklijk Huisarchief, The Hague, Prins Willem V, inv. 192, via DECODE R1071 (images 5592–5594), with DECODE’s transcription.
• K. de Leeuw, Cryptology and statecraft in the Dutch Republic (thesis, Amsterdam, 2000).
• Th. Bussemaker (ed.), ‘Uittreksels uit de brieven van D’Affry aan de Fransche regeering (December 1755–Mei 1762)’, Bijdragen en Mededeelingen van het Historisch Genootschap 27 (1906).

Transcription and notes: targets/affry1757c/ in the repository.