An Idea Born From a Reservation Childhood
The code talker program traces back to Philip Johnston, a civil engineer and World War I veteran who was not Navajo himself but had grown up on the Navajo reservation as the son of a missionary and was one of the few outsiders fluent in the language. In early 1942, after reading a news article about Native American soldiers relaying messages in their own languages during Army training exercises, Johnston approached the Marine Corps with a specific pitch: Navajo was essentially unknown outside a small, geographically contained population, had no written form or alphabet an enemy could study, and was grammatically complex enough that even the rare outside linguist who had studied it could not easily improvise translations in real time. The idea itself was not entirely new; American commanders had experimented with Choctaw speakers to secure field-telephone traffic during the trench stalemate of World War I, but no one had tried building a full, purpose-designed military code around a Native American language before.
The Marine Corps tested the idea and was convinced. On May 4, 1942, 29 Navajo recruits were sworn in at Fort Wingate, New Mexico, becoming what the program's records still call the "original 29." They were sent to Camp Pendleton in Oceanside, California, with an assignment that had no existing template: build a military code, from scratch, in a language most of the Marine Corps' communications officers had never heard spoken.
Two Layers, 411 Words, and No Paper
The code the original 29 built worked on two levels. The first was a spelling alphabet: each letter of the English alphabet was assigned one or more Navajo words (the letter A, for instance, could be sent as either Wol-la-chee, meaning ant, or Tse-nill, meaning axe), letting code talkers spell out names, places, or terms with no direct Navajo equivalent. The second, faster layer was direct word substitution for roughly 211 recurring military terms, later expanded to about 411 as the war went on. A fighter plane became Da-he-tih-hi, hummingbird, because of how it darted in and out like the bird among flowers. A submarine became Besh-lo, iron fish. A tank, for which Navajo had no native word at all, became the word for turtle, chosen for the same hard shell moving low across the ground. A squad became Debeh-li-zine, black street.
None of it was ever written down in the field. Code talkers memorized the entire dictionary and transmitted and received it purely by voice over radio and field telephone, decoding each line in their heads in real time under combat conditions, a discipline the program's designers built in deliberately: a captured codebook would have been useless against an enemy who did not speak the underlying language, but a captured code talker's memory was the only copy that existed, and losing it meant losing nothing that could be reconstructed from paper.
The Language Their Own Schools Had Once Forbidden
The program carried a specific irony that its own recruits recognized immediately. Many of the original 29 and the roughly 400 Navajo Marines who eventually served as code talkers had attended government or mission boarding schools built around an explicit policy of cultural assimilation, where speaking Navajo was punished, sometimes with corporal discipline or having a student's mouth washed out with soap. Code talker Chester Nez, one of the original 29, later described the disorienting experience of being asked by the same government that had once punished him for his language to now build a weapon out of it.
The irony ran deeper than symbolism. Navajo worked as an unbreakable code because it had survived, spoken and largely unwritten, outside the reach of the assimilation policy meant to erase it. The isolation the boarding school system had treated as a problem to be corrected turned out, one generation later, to be the war's most reliable communications security advantage.
Iwo Jima: The Numbers That Proved the Idea
The clearest single data point for how well the code performed comes from Iwo Jima in February 1945, fought in the same island-hopping Pacific campaign that had already cost the Navy the battleship entombed at Pearl Harbor three years earlier. Major Howard Connor, signal officer for the 5th Marine Division, had six Navajo code talkers working in shifts around the clock through the battle's first 48 hours. In that window alone, they sent and received more than 800 messages with zero errors, a level of speed and accuracy that mattered enormously in a battle where artillery coordinates and troop positions changed by the hour. Connor's own verdict, recorded afterward and preserved in the CIA's official account of the program, was blunt: "Were it not for the Navajos, the Marines would never have taken Iwo Jima."
Code talkers served in every Marine assault in the Pacific from 1942 to 1945, across all six Marine divisions, and the discipline that made Iwo Jima possible, memorized transmission with nothing committed to paper, held throughout. It is also why so few written artifacts of the wartime code survive outside the training dictionaries the Marine Corps itself retained: the system was designed so that capturing a code talker or his equipment yielded no code to reverse-engineer, only a language almost no one outside the Navajo Nation could speak in the first place.
Silence Until 1968
The code's success came at the cost of decades of public silence. The military treated the underlying technique as a reusable communications security method rather than a one-time wartime measure, so the program stayed classified after 1945, and the men who built and ran it were barred from discussing what they had done, even with their own families, for more than twenty years. Declassification did not come until 1968, a full generation after Iwo Jima. Formal recognition took longer still: the original 29 code talkers received the Congressional Gold Medal, and the roughly 300 additional Navajo Marines who later qualified as code talkers received the Congressional Silver Medal, at a Capitol ceremony in 2001, nearly 60 years after the first 29 were sworn in at Fort Wingate.
Most closely guarded wartime secrets surface within a few years of the fighting ending; this one took two decades longer, because the underlying method, not any single message, was what the military wanted kept out of view. Britain's 1917 dilemma over how to expose a decoded German cable without revealing that its own codebreakers had cracked it is a useful contrast: that secrecy problem resolved itself in weeks, the moment the telegram was published. The Navajo program's classification outlasted the war that produced it by a margin measured in decades, not weeks.