The Gulf of Tonkin incident that justified US escalation in Vietnam was fabricated

Gulf of Tonkin: When Signals Were Signals of Deception

On August 2, 1964, American destroyers reported being attacked by North Vietnamese torpedo boats in the Gulf of Tonkin. Two days later, another attack was allegedly reported. These incidents gave President Lyndon B. Johnson the political cover he needed to push the Gulf of Tonkin Resolution through Congress, which authorized an unprecedented expansion of military operations in Vietnam. Within months, Operation Rolling Thunder began—a sustained bombing campaign that would last three years and kill an estimated 52,000 American servicemen and over a million Vietnamese civilians.

For decades, the second attack existed in a murky space between fact and assumption. Military officials insisted it happened. Congressional leaders voted to approve the resolution based on briefings claiming it had occurred. Textbooks taught it as history. But serious questions persisted among researchers and former military personnel who had been there that night. The waters were rough. The radar signals were unclear. Some officers at the time expressed doubts about whether North Vietnamese boats had actually fired on the American destroyer USS Turner Joy.

Those skeptics were dismissed as second-guessers or Vietnam War opponents with an axe to grind. The official position from the Department of Defense remained firm: both attacks happened, and the military response was justified. Critics were told they lacked access to classified information. The government had signals intelligence—intercepted North Vietnamese communications—that proved the attack occurred. That was enough to end most public debate.

Then, in 2005, forty years after the incident, the National Security Agency its internal report on the affair. 's own analysis revealed what many suspected but could never prove: the second attack on August 4 never happened. More damaging still, the showed that signals intelligence had been manipulated and presented selectively to support a predetermined conclusion.