How did the Navajo code talkers' code actually work?

The Navajo code worked on two levels. Operators could spell out English words using Navajo words that stood for letters, and they could use Navajo words that stood for whole military terms. Because the underlying language was complex and unwritten, the Japanese never broke it. [1] [2]

The first 29 Navajo recruits trained in May 1942 and then created the code at Camp Pendleton in California. They first chose English words for each letter, using familiar things such as animals, and then translated those words into Navajo. A speaker recalled in the Smithsonian's account that A was a red ant, B a bear, C a cat, D a dog and E an elk. [3] [1]

A receiving operator heard what seemed to be a string of unrelated Navajo words. He translated each one into English and kept only the first letter. The Navy's account gives the example of the Navajo words for needle, ant, victor and yucca, which spell N-A-V-Y. Most letters had several Navajo words, so the letter A could be wol-la-chee (ant), be-la-sana (apple) or tse-nill (axe). That variety made the code harder to decipher. [1] [4]

Spelling every word would have been slow, so the developers assigned Navajo words to about 450 frequently used military terms that had no Navajo equivalent. Examples include besh-lo (iron fish) for submarine, dah-he-tih-hi (hummingbird) for fighter plane and debeh-li-zine (black sheep) for squad. America became Ne-he-mah, meaning our mother. One source says the initial project covered 211 keywords, so the size of the vocabulary depended on the stage of development. [1] [4] [2]

The Smithsonian notes that Code Talkers from several nations were given picture charts and invented words that seemed to fit the images. A Comanche speaker described calling a tank a turtle because it has a hard shell and moves. [3]

The dictionary and all code words had to be memorized during training. Code talkers usually worked in pairs. One translated and the other operated the radio when sending, and the roles were similar when receiving. Messages arrived in English, were translated without being written down, and were logged in English after receipt. Messages were coded only when security was essential. [1] [2] [3]

In a test staged by Philip Johnston, Navajos encoded, transmitted and decoded a three-line English message in 20 seconds, whereas machines of the time needed 30 minutes. Johnston argued that Navajo suited the task because it is unwritten and has a complex syntax and tonal qualities, and he estimated that fewer than 30 non-Navajos understood it. Japanese intelligence chief Seizo Arisue said Japan decoded Army and Army Air Corps codes but never the Marines' code. [1]

The Marine Corps Museum describes two types of code. Type one codes were formally developed by Comanche, Hopi, Meskwaki and Navajo speakers, with native words for each letter. Type two was informal, a direct translation from English into the native language, with descriptive words for terms that had no equivalent. [5] [3]