The Gulf of Tonkin incident that launched the Vietnam War was misrepresented

One real battle, and one that probably wasn't

On 2 August 1964, the destroyer USS Maddox was cruising the Gulf of Tonkin on a DESOTO patrol, a signals-intelligence mission that skirted North Vietnamese waters to map coastal radar. Three North Vietnamese P-4 torpedo boats came out to meet it. The Maddox fired first, in what it described as warning shots; the North Vietnamese boats pressed the attack with torpedoes and machine-gun fire, aircraft from the carrier USS Ticonderoga strafed the boats, and the exchange left the Maddox with a single bullet hole and the North Vietnamese with a damaged squadron and several dead sailors. Nobody seriously disputes that this happened. It is the least controversial fact in the entire episode.

The trouble starts two nights later. On 4 August, in heavy seas and poor visibility, the Maddox and a second destroyer, the USS Turner Joy, began picking up radar contacts, sonar returns and radio intercepts that their crews read as another torpedo-boat attack: one that, this time, would be cited by name in the resolution that sent America to war.

What the NSA's own historian found

For decades, all of this (Herrick's cables, Stockdale's account, the absence of any wreckage or bodies from a supposedly sunk North Vietnamese boat) remained circumstantial. Suggestive, but not proof that anyone had deliberately misrepresented anything; a bad night at sea is not, by itself, a scandal.

That changed with a document the National Security Agency fought for years to keep secret. In 2001, NSA historian Robert J. Hanyok completed an internal study titled Skunks, Bogies, Silent Hounds, and the Flying Fish: The Gulf of Tonkin Mystery, 2–4 August 1964, for the Agency's classified journal, Cryptologic Quarterly. Hanyok and other agency historians pushed from 2002 onward to make it public; senior NSA officials resisted, reportedly wary of comparisons to another war being justified by flawed intelligence, in Iraq. The impasse broke only after researcher Matthew M. Aid filed a Freedom of Information Act request in 2004 and reporter Scott Shane wrote about the standoff in The New York Times; the NSA then released Hanyok's study, alongside more than 140 other Gulf of Tonkin documents, on 1 December 2005.

Hanyok's conclusion, from inside the agency that generated the original intelligence, was unambiguous: the August 4 attack did not happen. He found that the signals intercepts cited at the time as evidence of a second attack in fact referred to North Vietnamese boats recovering vessels damaged in the real August 2 battle, not preparing a new one. Some of the confusion, his analysis showed, traced to translation: one intercepted Vietnamese phrase could be read as either a military operation or simply boats being towed a long distance for repair, and it was read the way that fit what everyone already expected. Analysts, he concluded, had made “an active effort... to make SIGINT fit the claim” of an attack. Of the intelligence actually available that night, he wrote, “the overwhelming body of reports, if used, would have told the story that no attack occurred.” Only the fragments that supported an attack were passed to policymakers.

“The overwhelming body of reports, if used, would have told the story that no attack occurred.”

None of this required North Vietnamese vessels to be present for the story to look convincing. It required only that the ambiguous evidence be filtered in one direction and the doubts filtered out, a bureaucratic distortion rather than an invented battle from whole cloth, but a distortion with the same consequence.

Two decades, two document releases

The full picture only assembled itself in stages, across two separate declassifications forty years apart. The first came in 1971, when military analyst Daniel Ellsberg leaked the Pentagon Papers (the Defense Department's own classified history of the war, commissioned by Secretary of Defense Robert McNamara) to the press. Among its revelations was that internal doubts about the August 4 incident had existed from the very first hours, well before the public was told otherwise.

McNamara himself spent decades defending the official account. Testifying before the Senate Foreign Relations Committee on 6 August 1964, he stated that “our Navy played no part in, was not associated with, was not aware of, any South Vietnamese actions, if there were any”, denying any connection to the covert raids conducted under the codename OPLAN 34A. Records that surfaced later show the claim was untrue: McNamara had already told Johnson privately that the 34A raids, with which the Maddox's patrol was in fact coordinated, “had some bearing on the attacks.” His public account shifted only decades later. In the 2003 documentary The Fog of War, he conceded there had been no attack on 4 August. In 1995, meeting his former adversary, North Vietnamese general Võ Nguyên Giáp, McNamara asked directly what had happened that night. Giáp's answer was two words: “Absolutely nothing.”

Johnson's own doubts surfaced in private almost as quickly. According to journalist Stanley Karnow, in his 1983 history Vietnam: A History, Johnson told an aide within days of the alleged attack, “Hell, those dumb, stupid sailors were just shooting at flying fish.” None of that private skepticism reached the public account at the time. Decades later, historian Edwin Moise's exhaustive 1996 study of the declassified record and interviews with participants on both sides reached, independently, the same conclusion the NSA would only confirm in 2005: the August 4 attack was a mistake, not a real engagement.

The second release was the NSA's own, in 2005, not a leak this time, but a forced disclosure under freedom-of-information law, of the agency's internal admission that its own signals intelligence had been mishandled. Between them, the two releases turned a persistent rumor into a documented finding: not from outside critics, but from the Department of Defense's leaked internal history and the NSA's own declassified one.

What the doubtful night actually bought

The speed of the congressional response was not entirely improvised. Assistant Secretary of State William Bundy had a draft war-powers resolution finished by early June 1964 (two months before any Gulf of Tonkin incident) after his brother, national security adviser McGeorge Bundy, told the National Security Council that Johnson wanted a North Vietnamese attack on US forces, rather than Viet Cong action inside South Vietnam, as the trigger before sending it to Congress. Johnson held the draft back through the summer, reportedly so it would not compete with the Civil Rights Act for congressional attention. The events of 2–4 August gave him the occasion he had been waiting for.

None of this remained an academic dispute about naval sensors, because of what followed within seventy-two hours. On 7 August 1964, citing “renewed hostile actions against United States ships in the Gulf of Tonkin,” Congress passed the Gulf of Tonkin Resolution, authorizing President Johnson “to take all necessary steps, including the use of armed force,” to defend US forces and allies in Southeast Asia. The Senate passed it 88 to 2, with only Senators Wayne Morse and Ernest Gruening opposed; the House recorded no votes against it at all, passing it 416 to 0, though Kentucky Republican Eugene Siler, who opposed the resolution but was absent that day, arranged to be formally “paired” against it, calling it an attempt to “seal the lips of Congress against future criticism.”

That resolution became the legal basis for the escalation of American combat operations in Vietnam for the next nine years, in the absence of a formal declaration of war. It rested, in significant part, on an incident that (on the evidence of the task force commander who reported it, the pilot who overflew it, and the intelligence agency that generated the signals used to justify it) almost certainly did not occur as described.

How the record reads now, in the Navy's own histories

The most important shift since 2005 is that the finding is no longer confined to one classified NSA monograph fought loose by a Freedom of Information Act request. It has migrated into the official, public record of the institutions that were closest to the events. The Navy's own Naval History and Heritage Command now states flatly, on its public history pages, that “later investigation revealed that the 4 August attack did not happen” and that no North Vietnamese vessels had been present. This is the service whose destroyers reported the attack, writing off its own alarm.

The U.S. Naval Institute, whose Naval History magazine speaks to and for the Navy's officer corps, went further. In a February 2008 analysis, The Truth About Tonkin, Lieutenant Commander Pat Paterson worked through the roughly 200 documents the NSA declassified across 2005 and 2006 and concluded that senior officials had distorted the intelligence record before the American public. Read beside Hanyok's study, it is a second, independent institutional pass over the same primary sources reaching the same place.

And the man who sold the incident to Congress ended up agreeing. In the 2003 documentary The Fog of War, former Defense Secretary Robert McNamara said plainly that the August 4 attack had never happened. When the intelligence agency that produced the intercepts, the naval service whose ships were involved, the institute that represents its officers, and the defense secretary who briefed the resolution all arrive at the same conclusion, the burden has shifted: it is now the denial of the finding, not the finding itself, that stands against the weight of the record.

“Later investigation revealed that the 4 August attack did not happen.” That is not a critic's line. It is the U.S. Navy's own history of its own ships.