code talkers

[William J Poser]

Aidan Wilson wrote:
>One wonders if any Japanese intelligence organisations recorded any
>intercepted transmissions. Might be some good historical data...

Unfortunately, I doubt it, for two reasons. First, as far as I can tell,
in WWII nobody was making audio recordings of inteligence intercepts.
Audio recording technoogy was still rather primitive and not very
portable, and most military communication was in any case in Morse code,
so the raw data for intercept intelligence normally consisted of a
radio operator's transcription of a bunch of code groups. Making
audio recordings or finding radio operators who could transcribe
unfamiliar languages was something they never got to.

Second, even if audio recordings were available, they probably would not
be of much value as linguistic data because, it appears, the code talkers
did not communicate in Navajo. Although the literature on the code talkers
is now rather large, it foccusses heavily on the lives of the code talkers
and of how clever the military people who thought of this were. It is
infuriatingly vague as to what the code talkers actually did. As best I
can tell, except for casual chitchat, they did NOT speak Navajo in the usual
sense of the term. Rather, they communicated in English, using what you
might call a Navajo-based system for representing English.

This representational system had two components. One was a system for
representing English letters, where each English letter was represented
by a Navajo word. Instead of the usual a = alpha, b = bravo, c = charlie,
etc., they had a = bilasaana ("apple"), b = nahashchid ("badger"), etc.
The other was a code for a set of a few hundred words. If you wanted
to say "tank", for example, instead of spelling it out as t-a-n-k,
you said chaydagahi "turtle". The syntax was that of telegraphic English.
If this is correct, audio recordings would not contain any Navajo
sentences or phrases. All you would get would be the isolated Navajo
words used for the letters of the alphabet and the small set of words
represented by codes rather than spelled out.

If anyone knows of evidence to the contrary, I'd be thrilled to see it,
but this is what I've been able to reconstruct from the rather limited
descriptions of what the code talkers actually said.
  
Curiously, it seems to me that this system was not as secure as having
the code talkers communicate in actual Navajo would have been. Had the
Japanese devoted more effort to this and perhaps had more or better
linguists involved, it shouldn't have been all that hard to break this
system. Given a decent phonetic transcription of a reasonable amount of
material, it would soon become obvious that the language consisted of
Navajo words but was not Navajo and the skewing of the word frequencies
would reveal that it was basically a Navajo-word-for-English-letter
substitution cipher. Such simple substitution ciphers were trivial to
break even in WWII. In contrast, if the code talkers had used
real Navajo, even if the Japanese caught on to what it was, they would
have been hard put to find people who could understand it.

I also wonder what the practice was with the even more poorly documented
code talkers who used other languages. For those, one generally doesn't
see descriptions of alphabets and codeword lists like those for Navajo.
I suspect that in at least some cases those less formally organized
systems involved the use of the actual native language.

Bill