tcmay@netcom.com (Timothy C. May) said:

They were the Navajo (or Navaho) code talkers. Kahn's "The Codebreakers" has a discussion of this, as I recall. Probably the "Encyclopedia Britannica" will also a mention of it.

It's mentioned in a bunch of crypto sources, but coincidentally, PBS just ran an entire show about this precise subject a couple weeks back. Not just passing mention...the whole show was about this. Therefore talking to PBS people would likely yield some good info... and possibly videos.

...

2) If one used a natural language for encryption, and the would-be code crackers did not know it was natural language (say, Hittite), could they crack it? I seem to remember that hieroglyphs were undecipherable until the Rosetta Stone was discovered. But maybe current techniques would do a better job....?

These are codes, not ciphers, and are of course not very secure.

Err...."of course"??? Codes are (all else being equal) quite a bit more secure than ciphers. (or do I have "code" and "cypher" reversed...whatever. :-) There's a classic SF story, whose title I forget, about anthropologists trying to figure out writings of a dead species on Mars. The table of elements finally proved to be the Rosetta Stone equivalent. The interesting thing about codes, which in a sense includes all natural languages, versus ciphers, is that code systems represent semantics. If the underlying semantics is radically different than what the code-breaker knows...too bad.

Germans and Japanese in WW II obviously did not have enough time to find native Navajo speakers, and I suspect few books on that language were available at that time, hence the scheme was temporarily secure.

Otherwise, forget it.

It was more complicated than that. They used a hybrid system that native Navajo speakers could not decrypt, because the system used not only Navajo, but on top of that, arbitrary (and newly invented) metaphors for concepts, and (newly invented) puns to represent ciphers, too. It is certainly true that part of the security was "through obscurity", but (A) that part was effective..."security through obscurity" can be effective over short periods of time...and (B) they layered ciphers on top of codes. The obscure linguistic aspects of Navajo vs. other modern languages is also said to have played a part, but I haven't researched this yet, so I won't comment. Arthur Chandler <arthurc@crl.com> said:

I seem to remember that hieroglyphs were undecipherable until the Rosetta Stone was discovered. But maybe current techniques would do a better job....?

In the absence of a Rosetta-Stone-sort-of-thing, we're still lost. For instance, the Easter Island hieroglyphs are still completely mystifying. Some of Nyquist's mathematical results are still classified, so one never knows, but... Arbitrary semantic systems encoded in writings are not decipherable, period, barring some breakthrough in mathematical semantics...don't hold your breath. :-) Doug