[Cryptography] How crypto killed Admiral Yamamoto

[Arnold Reinhold]

> On Apr 19, 2021, at 6:55 PM, Peter Gutmann <pgut001 at cs.auckland.ac.nz> wrote:
> 
> Arnold Reinhold via cryptography <cryptography at metzdowd.com> writes:
> 
>> The big intelligence coup from these breaks came from Japanese Vice Admiral
>> Abe, based in Berlin, who radioed back to Tokyo detailed reports on German
>> thinking and plans using the CORAL machine.
> 
> Minor nit, it was Hiroshi Ōshima, an army general, who was the US' man in
> Berlin, not Abe.
> 
> Peter.
> 
> 


You are correct that General Ōshima, who became the Japanese ambassador to Germany, was the biggest source of info on Nazi plans, but Admiral Abe also provided valuable intelligence. I refreshed my memory using this NSA publication: https://www.nsa.gov/about/cryptologic-heritage/center-cryptologic-history/pearl-harbor-review/early-japanese/ which mentions Abe but not Ōshima for some reason. 

And Ōshima did use the Purple cipher, so my implication that the Purple break did not provide ongoing intelligence during WW II was wrong. I should have known better. The NSA’s National Cryptologic Museum has on display the original machine the US built to decode Purple, using similar stepping switches. But there is later industrial strength version on display as well. It has three wiring matrices with a lot of empty positions, suggesting it was built to handle other codes as well, likely Jade and Coral.

On Sun, 18 Apr 2021 10:38:12 +0000 Ray Dillinger wrote:

> JN-25 was notoriously difficult to break; because it operated at the
> level of words and phrases rather than individual letters, it required
> deep knowledge of the Japanese language and how it was used specifically
> in naval communications to even build the frequency tables needed to
> attack it.
> 
> The Japanese are believed to have underestimated the availability of
> American cryptographers familiar with the Japanese language. But we
> were then as we still are today, 'A Nation Of Immigrants'. However
> badly we may have treated Japanese-Americans during the war, recruiting
> and training cryptographers familiar with that language was not a
> show-stopping problem.



The NSA publication linked above mentions that in the late 1920s the "the Japanese embassy was entered surreptitiously and code books were photographed” including a Japanese Navy code. While that code was soon superseded, the code book gave US cryptanalysis’s a good idea of the semantic content of future codes, aiding future attacks. It’s my understanding that the US did not use Japanese-Americans as translators for decrypted messages due to security concerns. The NSA pub mentions that the stolen code book was translated by a missionary couple. I knew one older gentleman, since passed on, who would not talk about what he did in the war but he was taught Japanese. Much military traffic was formulaic and the translated code books might have been enough to make sense of many messages. 

The Japanese Army also extensively used code book with additive systems. These too were broken, with devastating effect. It’s worth noting that the US and UK decision to employ thousands of women in their massive code breaking effort was likely crucial to the level of success and gave Allies a strategic advantage over the Axis. Both Japan and Germany refused to use women in most war work for ideological reasons.

Arnold Reinhold