The Mysteries of Tonkin Gulf

The Gulf of Tonkin incident in August 1964 proved to be America’s key entry point to war in Vietnam. The encounter sparked the first open fighting between the United States and North Vietnam, the first U.S. bombing of the North and an intensification of U.S. support for South Vietnam. It led to congressional passage of the Tonkin Gulf Resolution, which became the legal justification for America’s entry in the war. As with so much about Vietnam, events in the Gulf of Tonkin were not what they seemed at the time, and the consequences proved enormous. Even after five decades, we still struggle to understand what happened at the Gulf of Tonkin and why.

Things seemed clear-cut at the time. During the afternoon of Aug. 2, 1964, the U.S. destroyer Maddox was steaming in the Gulf of Tonkin – waters of the South China Sea between the coast of the Democratic Republic of Vietnam (i.e., North Vietnam) and the Chinese island of Hainan – when it came under attack from North Vietnamese torpedo boats. Maddox retreated down the gulf, where the destroyer C. Turner Joy joined it. Both ships headed back north in company.

On the night of Aug. 4, the warships, particularly C. Turner Joy, reported renewed attacks against them. President Lyndon B. Johnson, denouncing “hostile actions against United States ships on the high seas,” ordered retaliatory bombing against North Vietnamese naval bases. Sixty-four planes from the aircraft carriers Ticonderoga and Constellation attacked the North in what was called Operation Pierce Arrow. Two aircraft were shot down and one pilot – Lt. j.g. Everett Alvarez Jr. – was captured, becoming the first U.S. prisoner to be held by the North Vietnamese in Hanoi (the body of the other pilot lost, Lt. j.g. Richard C. Sather, was repatriated in 1985). Johnson asked Congress for a joint resolution approving his orders. Passed almost unanimously in the heat of the affair, this became the Tonkin Gulf Resolution, and its open-ended sanction of “all necessary measures to repel any armed attack ... and to prevent further aggression” became a crucial legal underpinning of the entire U.S. effort in the Vietnam War.

Doubts and Secrecy. In subsequent years – even while the fighting in Vietnam still raged – many aspects of the original account came into question. The first important element to come under scrutiny was the question of provocation for the North Vietnamese attack on Aug. 2. Senior administration officials, from Defense Secretary Robert S. McNamara on down, testified in hearings on the Tonkin Gulf Resolution that the U.S. warships had been in international waters and exercising free passage, their cruise unrelated to anything else. Certain “South Vietnamese” commando raids, “if there were any” (per McNamara’s testimony), were unknown aboard Maddox.

But suspicions to the contrary arose. By 1966, advocates were calling for a repeal of the resolution. Increasing doubts and political pressures led to new hearings in 1968, during which McNamara admitted that Maddox had been operating in close proximity to the commando raids, which he now represented as of South Vietnamese origin. The Tonkin Gulf Resolution was repealed in 1970; Johnson repeated the story of the “South Vietnamese” raids in his 1971 memoir.

Secrecy surrounding these events gradually unraveled. Maddox, it turned out, had been on a mission specifically aimed at collecting intelligence on North Vietnamese communications and coastal radars. Then it emerged that Maddox, though in international waters when it fought off the North Vietnamese torpedo boats, had been in territorial seas when the Vietnamese patrol craft left base to intercept it. The commando raids had not been South Vietnamese after all, but attacks using indigenous troops, unilaterally controlled by the U.S. special operations command in Vietnam. The strikes that took place during the Maddox cruise were partly intended to trigger the North Vietnamese to activate their nets so that the United States could record them. The 1971 leak of the Pentagon Papers revealed that the attacks themselves formed part of an extensive program of “graduated military pressures” against the North called OPLAN 34A. When audiotapes of Johnson’s telephone conversations on these days were declassified in the late 1990s, those with McNamara showed that both were aware of the connection between the Maddox mission and the coastal raids from the very beginning.

Hanoi’s choices obviously factored into the incident. Johnson based his decision for retaliatory bombing on an assumption that the North Vietnamese intended a confrontation in the gulf. The release in the late 1970s of the memoranda recording Johnson's meetings on this decision show that CIA Director John McCone assured the president that Hanoi had sought battle. But the CIA had no direct evidence at the time. A postwar Vietnamese official history attributes their response to the general staff and the naval command, not the North Vietnamese leadership. Similarly, senior Vietnamese officials told the U.S. delegation at a 1997 conference on “missed opportunities” in the Vietnam War that the dispatch of torpedo boats on Aug. 2 had been at the initiative of local commanders. This formula disguised the hand of the Vietnamese general staff but also contained the sense that no national decision to fight the Americans was made.

No Evidence. By far the deepest mystery of the Tonkin Gulf concerns the “second attack,” the notion that on the night of Aug. 4 the North Vietnamese came back to fight Maddox and C. Turner Joy together. It was this allegation of a repeated attack in the face of U.S. warnings that underpinned the retaliation. But unlike the sea battle of Aug. 2, there was no physical evidence for this engagement. The destroyers had maneuvered to avoid torpedoes, and Lt. Cmdr. Robert C. Barnhart Jr.’s C. Turner Joy pumped out more than 370 5-inch and 3-inch shells, yet there were no photos of attack boats, no shells hitting the ships, no prolonged observation of an enemy. Capt. John Herrick reported an initial radar contact. After that, Maddox dispatches recorded a mélange of sonar and radar contacts. When the ship trained its guns on the most solid of these, the vessel in the cross hairs was C. Turner Joy. Barnhart’s destroyer detected nothing on sonar, not even the torpedo some of his sailors said they saw.

As the on-scene naval commander, Herrick sent a dispatch warning against premature action – he and Maddox’s skipper, Cmdr. Herbert L. Ogier, doubted the authenticity of everything except the initial radar contact. Aircraft scrambled from Constellation and Ticonderoga also failed to spot anything, and when vectored to attack by the destroyers, found only the U.S. warships beneath them when they rolled in.

It turned out to be Washington on a hair trigger. The Johnson-McNamara telephone tapes show the president and defense secretary mulling over bombing targets before the alleged attack was even reported. This was possible due to National Security Agency (NSA) communications intercepts – and therein lies another tale. When the very first sighting report arrived, Washington and the Pacific Fleet presumed the expected attack and made their strike plans. McNamara and Johnson discussed specific targets before any meetings took place.

But the intelligence was wrong. NSA had pulled down a series of North Vietnamese messages related to the Aug. 2 attack, and a few between Aug. 2-4 pertained to making patrol, not torpedo, boats ready for sea. The volume of North Vietnamese communications decreased after Aug. 2, sighting and position reports disappeared from this traffic, and the patrol boat intercept could be interpreted as relating to sending someone to rescue survivors from the first engagement. A North Vietnamese naval officer captured in 1966 said under interrogation that he had been the senior deputy to the enemy torpedo boat commander. He had vivid recollections of the first fight, but insisted there was no battle on Aug. 4.

The NSA’s listening posts in the region included one at Phu Bai in Vietnam, another in the Philippines, and the special van installed aboard Maddox for the cruise. The Philippines base had decoded its intercepts and reported them in a timely fashion, but the Phu Bai base was delayed and reported late – in fact, during the time frame of the alleged second attack. Washington officials compiling communications intelligence summaries assumed that the late Phu Bai intercepts were up-to-the-minute reporting and included them as such. They did not check with Maddox, which itself recorded no messages indicating a renewed attack. In addition, one dispatch that analysts at Phu Bai decided was the attack order for Aug. 4 contained decoding and translation errors and was intercepted by no one except that station. Washington officials either did not notice or did not care that the messages corresponded to the actual events of Aug. 2, or that North Vietnamese Swatow-type patrol boats were not armed with torpedoes and lacked speed to maneuver as indicated in the destroyers’ spot updates on the wild night of Aug. 4. They also discounted Herrick’s warning dispatch. Johnson retaliated, and North Vietnam and the United States moved closer to war.

Escalation. NSA post-mortems in late 1964 cast some doubt on the chronology but hewed to the official line. Deputy Director Louis Tordella satisfied himself by the early 1970s that a mistake had been made, as did CIA intelligence chief Ray Cline. The Senate's Foreign Relations Committee staff director, Carl Marcy, also discovered the discrepancies. Yet they remained concealed for a very long time. In 2005 and 2006, the NSA finally declassified full texts of the most important intercepts, revealing the chronological transpositions in its messages. The agency’s official history on Vietnam, declassified in 2007, examines the evidence and concludes that there was no incident in the Gulf of Tonkin on Aug. 4, 1964. The NSA also released the full record of its Tonkin Gulf messages, demonstrating how reporting from Phu Bai differed from other intercept stations. In the meantime, at the “missed opportunities” conference in 1997, North Vietnamese Gen. Vo Nguyen Giap openly told McNamara that Vietnam had never made any second attack.

The most important effects of this episode were in Hanoi and Washington. Shortly after the incident, the North Vietnamese sent an infiltration group to South Vietnam that, for the first time, stayed together as a unit instead of dispersing among the National Liberation Front as cadres. While it was true that Hanoi decided in December 1963 to increase its support for the insurgency in the South, it had not previously committed regular army formations. In the fall of 1964, Hanoi decided to send a full regular infantry division to the South as a constituted force. The 325th Division began preparatory training for this move about September 1964. For the Democratic Republic of Vietnam, Tonkin Gulf provided evidence that the United States was determined to wage a full-scale war. Hanoi moved to meet that challenge.

In Washington, as the Pentagon Papers first documented, top officials had struggled for months to contrive a scenario making it politically feasible to broaden the U.S. effort in Vietnam. Part of that was to obtain congressional approval. An interagency working group had written several drafts for a proposed resolution. The Tonkin incident became the perfect opportunity. The White House sent its draft resolution to Capitol Hill the night of the Pierce Arrow attacks; it passed Aug. 7.
Johnson’s concern for “the scenario” and his careful search for political cover reflect his determination to press ahead in Vietnam but not the misgivings he also expressed, or his ambitious plans for social legislation. The president had acquired a stake in Vietnam policy. Like his predecessor, John F. Kennedy, the seeming inability of the South Vietnamese allies to progress against the National Liberation Front frustrated him. U.S. military capability and expertise appeared to be the solution. But Johnson still thought that Vietnam could be fought on the side without impacting his domestic goals. That turned out to be a major error.

During August 1964, shortly after the incident, the United States sent B-57 jet bombers to Da Nang and Bien Hoa, and also deployed F-100 and F-102 jet fighters to Vietnam. Air Force personnel in South Vietnam increased by nearly 20 percent, while the overall U.S. force level – which had held steady for half a year – swelled from 16,500 to 23,300 between June and December. The B-57s at Bien Hoa offered the target for a spectacular commando raid that November, helping persuade Johnson to escalate still further.

Today, the Gulf of Tonkin reminds us that small events can have enormous consequences. Only one American died – Sather during Pierce Arrow – as well as a small number of North Vietnamese sailors. But the bombing gave the war new intensity. It also challenged the North to strike directly at Americans, as Hanoi did starting with Bien Hoa. And the Tonkin Gulf Resolution would be stretched beyond its context to cover a commitment that no one except the war managers foresaw. Events can be crucial not just for their intrinsic importance, but also as a hinge for the plans and purposes of others, and for what an adversary might conclude from them. Americans in Vietnam, and since, would have to relearn that lesson.

John Prados is a historian based in Washington. He is a senior fellow and director of the Vietnam Documentation Project of the National Security Archive. His current book is “Islands of Destiny: The Solomons Campaign and the Eclipse of the Rising Sun” (NAL Caliber).