Abstract
The recovery of the underlying code book forming part of an additive system depends upon the successful decryption of many intercepted messages, thus exposing the bare code equivalents. The skills required for successful decoding of a bare code message are more akin to the linguistic challenge of determining the nature and meaning of an unknown written language than to those needed for elucidating the operation of a cipher machine or the form of a superencipherment. This is particularly so when the underlying plain text is in a foreign language or in the jargon peculiar to a restricted class of users. Two historical examples are given.
This is a preview of subscription content, log in via an institution to check access.
Access this chapter
Tax calculation will be finalised at checkout
Purchases are for personal use only
Notes
- 1.
The well-known book Gods, Graves and Scholars by C. W. Ceram (Kurt Marek) gives much more information on the Rosetta Stone. A copy of it is displayed prominently in the National Cryptologic Museum in Fort George Meade, Maryland. Michael Coe’s Breaking the Maya Code, Thames and Hudson, 1992, is also of interest here.
- 2.
Another well-known example of the reading of ancient texts is the post-war work on Linear B using certain written material found in Crete. The work of Michael Ventris (1922–1956) and John Chadwick (1920–1998) initially attracted some controversy. Chadwick had worked on Japanese Naval Attaché material (not covered in this book) at Bletchley Park in 1944–1945.
A. P. Treweek was a classicist and mathematician who worked on Japanese material at Frumel. He published a paper Chain Reaction or House of Cards: An Examination of the Ventris Decipherment in the Bulletin of the Institute of Classical Studies 4 (1) December 1957, pages 10–26. He concluded that the Ventris work was basically sound and set an interesting related puzzle for the reader. This gave the names of various geographical locations in the Pacific in kana syllables. The challenge was to work out the meanings of the kana syllables used.
You begin by noting that \(\spadesuit\)-\(\heartsuit\)-\(\diamond \)-\(\diamond \) probably refers to HO-NO-LU-LU and then use the identifications \(\spadesuit =\mathrm{ HO}\), \(\heartsuit =\mathrm{ NO}\), \(\diamond =\mathrm{ LU}\) to work on the rest of the names in the list. The work on Linear B had made similar use of reference to geographical locations.
- 3.
The word ‘crib’ has one meaning in the context of code books and a quite different one in the saga of attacking the Enigma machine.
- 4.
The basic source of information on Scovell’s code breaking work is Appendix XV The Scovell Ciphers in Vol. V of Charles Oman’s History of the Peninsular War, published from 1902 onwards. Oman had access to papers of Scovell, now held by the British National Archives (TNA) as file WO 37. Mark Urban’s stimulating biography of Scovell—The Man Who Broke Napoleon’s Codes—appeared in 2001.
The French language would have been widely understood among the better educated officers of the British Army at the time of the Napoleonic wars. The Japanese language was understood by only a few members of the Allied armed forces in December 1941. Despite the introduction of special courses in that language, the decryption process was necessarily more difficult. The capture of the documentation for JN-25C in August 1943 taken together with the thousands of archived JN-25C intercepts must have provided invaluable insight into IJN telegraphic language.
Likewise Scovell would have been able to access a limited supply of intercepts without any assistance. In the case of JN-25, with thousands of book groups and thousands of intercepts, some mechanisation of the task of finding all occurrences of a given book group would have been most useful. This was one place where the IBM punch cards and associated machinery (see Sect. 1.15) came in very handy.
The CBTR include a somewhat jocular Part L—Bibliography put together by various students of Latin and ancient Greek in 1945. Thus one Aeneas the Tactician (360 bc) is mentioned but not Sun Tzu (Chap. 20) or Scovell. The authors of the CBTR also missed quoting Virgil’s Aeneid (ii 390): Dolus an virtus, quis in hoste requirat? This translates to ‘Trickery or courage: which is needed in conflict?’.
A fuller account of early codes could mention the French optical telegraph (semaphore line) dating from 1794. It signalled information by varying the position of movable wooden arms on towers. This could be read from the next tower in the chain and the signal re-transmitted. At one stage in the Napoleonic Wars the RN benefited from reading the signals transmitted from a tower near the coast.
- 5.
Gannon Room 40, page 169.
- 6.
A letter from Lieutenant Fabian to Commander Safford dated 30 August 1941 (in RG38 of NARA but also available on pages 143–144 of Timothy Wilford’s thesis via ProQuest Dissertations) makes it clear that the long USN experience in using TA and DF on IJN signals was instrumental in getting started on the JN-25B code book.
- 7.
(This note repeats much of Sect. 7.1.) A bare code may be insecure without it being capable of being fully solved. Thus in NARA RG457 SRH-349 there is the following general comment: ‘Reconstruction of a code book is a long laborious process. Each code group needs to be identified singly. The larger the code, the longer is the time needed.’ Thus a decrypted (but incompletely decoded) message may look like ‘xxxxx xxxxx submarine xxxxx attack xxxxx xxxxx merchant ship xxxxx xxxxx’. This state of knowledge must be what Alastair Denniston meant by the phrase skeleton code books in Sect. 3.2.
- 8.
The code book was saved from a fire by a Marine early in the Guadalcanal fighting. Possession of a complete code book in fact assists in finding entries in its replacement code book, provided data is available on the frequency of occurrence of commonly used book groups in both systems. One may reasonably assume that the meanings of, say, the 40 most commonly occurring book groups are much the same and use this to reduce considerably the work needed to identify individual book groups in the new system. Comparison of signals from an active base that regularly uses the same commonly used book groups could help accelerate the process.
- 9.
Joe Rochefort did give an oral history interview which is preserved in the United States Naval Academy. Key extracts may be found on the Naval Cryptological Veterans’ Association website by asking Google to search for NCVA Cryptolog. In view of the insecure encryption of JN-25B messages from December 1941 to May 1942, his Combat Intelligence Unit was in effect working on an unenciphered code. (Even if this last comment is an exaggeration, the decryption had become a simple, standard chore.) He stated there that he needed and had a very good memory of previous messages to work out the likely meaning of key intercepts in which some but not all the book groups used were ‘evaluated’. He described the use of ‘two or three million’ cards in a month. He states that the practice in his CIU was to use twice as many cards as there were groups in the message. This seems obscure.
In June 1943, Rudolph Fabian wrote ‘A careful survey of the requirements of this unit, set against the possible space available, indicates that the total IBM machinery allowance should be two (2) each of the following: Type 405 Tabulators; Type 513 Summary (?) Producers; Type 075 Sorters; Type 077 Collators; Type NC4 machines; plus six (6) Type 035 Alpha punches.’ See NARA RG457, SRH-275.
- 10.
This description was made possible by the memories of George Aspden, who was at CBB in 1945, and of Ian Watson, who worked with IBM cards in the early 1960s. It is confirmed on page 11 of Part G of the CBTR. Laurie Robertson’s Arlington Hall Station: The US Army’s World War II Cryptanalytic Center, in the IEEE Annals of the History of Computing, 2004, has also been of use.
The NAA Canberra file A6923 SI/5, being a technical report on CBB procedures as they were in April 1945, refers to automated decoding in the words ‘The IBM stage decodes the message and produces a new work copy for the translators’.
- 11.
NAA Canberra item A5799 5/1940 is entitled Japanese Language in the Services and contains a draft report abandoned on 20 January 1940 due to the war in Europe. It contains the rather dismal sentence ‘At the present time the services have no first class interpreters of the Japanese language.’ See also Note 12 of Chap. 7.
Author information
Authors and Affiliations
Rights and permissions
Copyright information
© 2014 Springer International Publishing Switzerland
About this chapter
Cite this chapter
Donovan, P., Mack, J. (2014). Recovery of a Code Book. In: Code Breaking in the Pacific. Springer, Cham. https://doi.org/10.1007/978-3-319-08278-3_11
Download citation
DOI: https://doi.org/10.1007/978-3-319-08278-3_11
Published:
Publisher Name: Springer, Cham
Print ISBN: 978-3-319-08277-6
Online ISBN: 978-3-319-08278-3
eBook Packages: Computer ScienceComputer Science (R0)