On 2 August 1964, North Vietnamese patrol torpedo boats attacked the USS Maddox (DD-731) while the destroyer was in international waters in the Gulf of Tonkin. There is no doubting that fact. But what happened in the Gulf during the late hours of 4 August—and the consequential actions taken by U.S. officials in Washington—has been seemingly cloaked in confusion and mystery ever since that night.
Nearly 200 documents the National Security Agency (NSA) declassified and released in 2005 and 2006, however, have helped shed light on what transpired in the Gulf of Tonkin on 4 August. The papers, more than 140 of them classified top secret, include phone transcripts, oral-history interviews, signals intelligence (SIGINT) messages, and chronologies of the Tonkin events developed by Department of Defense and NSA officials. Combined with recently declassified tapes of phone calls from White House officials involved with the events and previously uncovered facts about Tonkin, these documents provide compelling evidence about the subsequent decisions that led to the full commitment of U.S. armed forces to the Vietnam War.
Raids and Patrols in the Tonkin Gulf
In early 1964, South Vietnam began conducting a covert series of U.S.-backed commando attacks and intelligence-gathering missions along the North Vietnamese coast. Codenamed Operations Plan (OPLAN) 34A, the activities were conceived and overseen by the Department of Defense, with the support of the Central Intelligence Agency, and carried out by the South Vietnamese Navy. Initial successes, however, were limited; numerous South Vietnamese raiders were captured, and OPLAN 34A units suffered heavy casualties. In July 1964, Lieutenant General William C. Westmoreland, commander of the U.S. Military Assistance Command, Vietnam, shifted the operation’s tactics from commando attacks on land to shore bombardments using mortars, rockets, and recoilless rifles fired from South Vietnamese patrol boats.1
The U.S. Navy, meanwhile, had been conducting occasional reconnaissance and SIGINT-gathering missions farther offshore in the Tonkin Gulf. Destroyers carried out these so-called Desoto patrols. After missions in December 1962 and April of the next year, patrols were scheduled for 1964 in the vicinity of OPLAN 34A raids. In fact, one of the patrols’ main missions was to gather information that would be useful to the raiders.2 A top-secret document declassified in 2005 revealed the standing orders to the Desoto patrols: “[L]ocate and identify all coastal radar transmitters, note all navigation aids along the DVR’s [Democratic Republic of Vietnam’s] coastline, and monitor the Vietnamese junk fleet for a possible connection to DRV/Viet Cong maritime supply and infiltration routes.”3
The United States was playing a dangerous game. The South Vietnamese—conducted OPLAN 34A raids and the U.S. Navy’s Desoto patrols could be perceived as collaborative efforts against North Vietnamese targets. In reality, there was no coordination between the forces conducting the operations.
Daylight Attack on a Destroyer
On 28 July, the Maddox sortied from Taiwan en route to her Desoto patrol station. Specially equipped with a communications intercept van and 17 SIGINT specialists, she was to patrol in international waters off the North Vietnamese coast, from the demilitarized zone (DMZ) north to the Chinese border. On the night of 30-31 July, the destroyer was on station in the Gulf of Tonkin when a 34A raid was launched against Hon Me Island. From two boats, South Vietnamese commandos fired machine guns and small cannon at the island’s radar and military installations. At the same time, two other South Vietnamese commando boats carried out a similar attack against Hon Ngu Island, more than 25 miles to the south.4
After observing North Vietnamese patrol torpedo boats pursuing the vessels that had attacked Hon Me, the Maddox withdrew from the area. Nevertheless, when later queried by NSA headquarters, the destroyer indicated she had been unaware of the OPLAN raid on the island.5 That ignorance set the stage for a showdown between North Vietnamese forces and the U.S. Navy eavesdropping platform.
By 1 August, the destroyer had returned to the area and was back on patrol. In the early hours of the next day, Maddox communication technicians intercepted SIGINT reports of North Vietnamese vessels getting under way, possibly intent on attacking the destroyer. On board the ship, Commander, Destroyer Division 192, Captain John J. Herrick ordered the vessel out to sea, hoping to avoid a confrontation. But at 1045, he reversed orders, turning the Maddox back toward the coast, this time to the north of Hon Me Island.
Weather conditions were clear, and seas were calm. At 1440, the destroyer detected three North Vietnamese patrol boats approaching her position from the west. Aware of North Vietnamese intent from the earlier SIGINT message, Captain Herrick ordered gun crews to open fire if the fast-approaching trio closed to within 10,000 yards of the destroyer, and at about 1505 three 5-inch shots were fired across the bow of the closest boat. In return, the lead vessel launched a torpedo and veered away. A second boat then launched two “fish” but was hit by gunfire from the destroyer. Re-engaging, the first PT boat launched a second torpedo and opened fire with her 14.5-mm guns, but Maddox shell fire heavily damaged the vessel.6
Overhead, meanwhile, four F8 Crusaders that the Maddox had called in earlier from the USS Ticonderoga (CVA-14) were rapidly approaching. One of the pilots, Navy Commander James Stockdale, commanding officer of VF-51, recalled that they passed over the unscathed Maddox at 1530, minutes after the 22-minute surface engagement had ended. All of the enemy boats were heading northwest at about 40 knots, two in front of the third by about a mile. The destroyer was retiring to the south.
Stockdale and the other pilots, with orders to “attack and destroy the PT boats,” made multiple firing runs on the enemy vessels. The two lead boats maneuvered evasively but were nevertheless heavily damaged. The third was left dead in the water and burning.7
Fighting Phantoms on 4 August
The next day, the Maddox resumed her Desoto patrol, and, to demonstrate American resolve and the right to navigate in international waters, President Lyndon B. Johnson ordered the USS Turner Joy (DD-951) to join the first destroyer on patrol off the North Vietnamese coast. That night, the South Vietnamese staged more OPLAN 34A raids. Three patrol craft attacked a security garrison at Cua Ron (the mouth of the Ron River) and a radar site at Vinh Son, firing 770 rounds of high-explosive munitions at the targets.8 North Vietnamese installations had been attacked four separate times in five days.
On the morning of 4 August, U.S. intelligence intercepted a report indicating that the communists intended to conduct offensive maritime operations in the Gulf of Tonkin. In contrast to the clear conditions two days earlier, thunderstorms and rain squalls reduced visibility and increased wave heights to six feet. In addition to the difficult detection conditions, the Maddox‘s SPS-40 long-range air-search radar and the Turner Joy‘s SPG-53 fire-control radar were both inoperative.9 That night, Herrick had the two ships move out to sea to give themselves maneuver space in case of attack.
The Maddox nevertheless reported at 2040 that she was tracking unidentified vessels. Although the U.S. destroyers were operating more than 100 miles from the North Vietnamese coastline, the approaching vessels seemed to come at the ships from multiple directions, some from the northeast, others from the southwest. Still other targets appeared from the east, mimicking attacking profiles of torpedo boats. Targets would disappear, and then new targets would appear from the opposite compass direction.
Over the next three hours, the two ships repeatedly maneuvered at high speeds to evade perceived enemy boat attacks. The destroyers reported automatic-weapons fire; more than 20 torpedo attacks; sightings of torpedo wakes, enemy cockpit lights, and searchlight illumination; and numerous radar and surface contacts. By the time the destroyers broke off their “counterattack,” they had fired 249 5-inch shells, 123 3-inch shells, and four or five depth charges.10
Commander Stockdale was again in the action, this time alone. When his wingman’s aircraft developed trouble, Stockdale got permission to launch solo from the Ticonderoga. He arrived overhead at 2135. For more than 90 minutes, he made runs parallel to the ships’ course and at low altitude (below 2,000 feet) looking for the enemy vessels. He reported later, “I had the best seat in the house to watch that event and our destroyers were just shooting at phantom targets—there were no PT boats there . . . there was nothing there but black water and American firepower.”11
Captain Herrick also began to have doubts about the attack. As the battle continued, he realized the “attacks” were actually the results of “overeager sonar operators” and poor equipment performance. The Turner Joy had not detected any torpedoes during the entire encounter, and Herrick determined that the Maddox‘s operators were probably hearing the ship’s propellers reflecting off her rudder during sharp turns.12 The destroyer’s main gun director was never able to lock onto any targets because, as the operator surmised, the radar was detecting the stormy sea’s wave tops.
By 0127 on 5 August, hours after the “attacks” had occurred, Herrick had queried his crew and reviewed the preceding hours’ events. He sent a flash (highest priority) message to Honolulu, which was received in Washington at 1327 on 4 August, declaring his doubts: “Review of action makes many reported contacts and torpedoes fired appear doubtful. Freak weather effects on radar and overeager sonarmen may have accounted for many reports. No actual visual sightings by MADDOX. Suggest complete evaluation before any further action taken.”13
Confusion in Washington
Messages declassified in 2005 and recently released tapes from the Lyndon Baines Johnson Library reveal confusion among the leadership in Washington. Calls between the Joint Chiefs of Staff; the National Military Command Center; headquarters of the Commander in Chief, Pacific; and Secretary of Defense Robert McNamara were frequently exchanged during the phantom battle. Vietnam was 12 hours ahead of Washington time, so the “attacks” in the evening of 4 August in the Gulf of Tonkin were being monitored in Washington late that morning.
In Hawaii, Pacific Fleet Commander-in-Chief Admiral U. S. Grant Sharp was receiving Captain Herrick’s reports by flash message traffic, not voice reports. At 0248 in the Gulf, Herrick sent another report in which he changed his previous story:
Certain that original ambush was bonafide. Details of action following present a confusing picture. Have interviewed witnesses who made positive visual sightings of cockpit lights or similar passing near MADDOX. Several reported torpedoes were probably boats themselves which were observed to make several close passes on MADDOX. Own ship screw noises on rudders may have accounted for some. At present cannot even estimate number of boats involved. TURNER JOY reports two torpedoes passed near her.14
McNamara phoned Sharp at 1608 Washington time to talk it over and asked, “Was there a possibility that there had been no attack?” Sharp admitted that there was a “slight possibility” because of freak radar echoes, inexperienced sonarmen, and no visual sightings of torpedo wakes. The admiral added that he was trying to get information and recommended holding any order for a retaliatory strike against North Vietnam until “we have a definite indication of what happened.”15
Other intelligence supported the belief that an attack had occurred. An intercepted SIGINT message, apparently from one of the patrol boats, reported: “Shot down two planes in the battle area. We sacrificed two comrades but all the rest are okay. The enemy ship could also have been damaged.”16 Amid all the other confusion and growing doubt about the attack, this battle report was a compelling piece of evidence. At 1723 in Washington, Air Force Lieutenant General David Burchinal, the director of the Joint Staff, was watching the events unfold from the National Military Command Center when he received a phone call from Sharp. He admitted that the new SIGINT intercept “pins it down better than anything so far.”17
McNamara considered the report, coupled with Admiral Sharp’s belief the attack was authentic, as conclusive proof. At 2336, President Johnson appeared on national television and announced his intent to retaliate against North Vietnamese targets: “Repeated acts of violence against the armed forces of the United States must be met not only with alert defense, but with positive reply. The reply is being given as I speak to you tonight.”18
Back on board the Ticonderoga, Commander Stockdale had been ordered to prepare to launch an air strike against the North Vietnamese targets for their “attacks” of the previous evening. Unlike Captain Herrick, Stockdale had no doubt about what had happened: “We were about to launch a war under false pretenses, in the face of the on-scene military commander’s advice to the contrary.”19 Despite his reservations, Stockdale led a strike of 18 aircraft against an oil storage facility at Vinh, located just inland of where the alleged attacks on the Maddox and Turner Joy had occurred. Although the raid was successful (the oil depot was completely destroyed and 33 of 35 vessels were hit), two American aircraft were shot down; one pilot was killed and the second captured.20
On 7 August, Congress, with near unanimity, approved the Gulf of Tonkin Resolution, which President Johnson signed into law three days later. Requested by Johnson, the resolution authorized the chief executive to “take all necessary measures to repel any armed attack against the forces of the United States and to prevent further aggression.” No approval or oversight of military force was required by Congress, essentially eliminating the system of checks and balances so fundamental to the U.S. Constitution. On hearing of the authorization’s passage by both houses of Congress, the delighted President remarked that the resolution “was like Grandma’s nightshirt. It covers everything.”21
Analysis of the Evidence
Historians have long suspected that the second attack in the Gulf of Tonkin never occurred and that the resolution was based on faulty evidence. But no declassified information had suggested that McNamara, Johnson, or anyone else in the decision-making process had intentionally misinterpreted the intelligence concerning the 4 August incident. More than 40 years after the events, that all changed with the release of the nearly 200 documents related to the Gulf of Tonkin incident and transcripts from the Johnson Library.
These new documents and tapes reveal what historians could not prove: There was not a second attack on U.S. Navy ships in the Tonkin Gulf in early August 1964. Furthermore, the evidence suggests a disturbing and deliberate attempt by Secretary of Defense McNamara to distort the evidence and mislead Congress.
Among the most revealing documents is a study of the Gulf of Tonkin incidents by NSA historian Robert J. Hanyok. Titled “Skunks, Bogies, Silent Hounds, and the Flying Fish: The Gulf of Tonkin Mystery, 2-4 August 1964,” it had been published in the classified Cryptological Quarterly in early 2001. Hanyok conducted a comprehensive analysis of SIGINT records from the nights of the attacks and concluded that there was indeed an attack on 2 August but the attack on the 4th did not occur, despite claims to the contrary by President Johnson and Secretary McNamara. According to John Prados of the independent National Security Archive, Hanyok asserted that faulty signals intelligence became “vital evidence of a second attack and [Johnson and McNamara] used this claim to support retaliatory air strikes and to buttress the administration’s request for a Congressional resolution that would give the White House freedom of action in Vietnam.”22
Almost 90 percent of the SIGINT intercepts that would have provided a conflicting account were kept out of the reports sent to the Pentagon and White House. Additionally, messages that were forwarded contained “severe analytic errors, unexplained translation changes, and the conjunction of two messages into one translation.” Other vital intercepts mysteriously disappeared. Hanyok claimed that “The overwhelming body of reports, if used, would have told the story that no attack occurred.”23
The historian also concluded that some of the signals intercepted during the nights of 2 and 4 August were falsified to support the retaliatory attacks. Moreover, some intercepts were altered to show different receipt times, and other evidence was cherry picked to deliberately distort the truth. According to Hanyok, “SIGINT information was presented in such a manner as to preclude responsible decision makers in the Johnson Administration from having the complete and objective narrative of events of 04 August 1964.”24
And what about the North Vietnamese battle report that seemed to provide irrefutable confirmation of the attack? On further examination, it was found to be referring to the 2 August attacks against the Maddox but had been routinely transmitted in a follow-up report during the second “attack.” The North Vietnamese were oblivious to the confusion it would generate.
What should have stood out to the U.S. leadership collecting all the data of these attacks was that, with the exception of the battle report, no other SIGINT “chatter” was detected during the attacks on 4 August. In contrast, during the 2 August attack NSA listening posts monitored VHF communications between North Vietnamese vessels, HF communications between higher headquarters in Hanoi and the boats, and communication relays to the regional naval station. None of these communications occurred on the night of 4 August.
The Defense Secretary’s Role
Subsequently, Secretary McNamara intentionally misled Congress and the public about his knowledge of and the nature of the 34A operations, which surely would have been perceived as the actual cause for the 2 August attack on the Maddox and the apparent attack on the 4th. On 6 August, when called before a joint session of the Senate Foreign Relations and Armed Services committees to testify about the incident, McNamara eluded the questioning of Senator Wayne Morse (D-OR) when he asked specifically whether the 34A operations may have provoked the North Vietnamese response. McNamara instead declared that “our Navy played absolutely no part in, was not associated with, was not aware of, any South Vietnamese actions, if there were any.”25
Later that day, Secretary McNamara lied when he denied knowledge of the provocative 34A patrols at a Pentagon news conference. When asked by a reporter if he knew of any confrontations between the South and North Vietnamese navies, he responded: “No, none that I know of. . . . [T]hey operate on their own. They are part of the South Vietnamese Navy . . . operating in the coastal waters, inspecting suspicious incoming junks, seeking to deter and prevent the infiltration of both men and material.” Another reporter pressed the issue, “Do these [patrol boats] go north, into North Vietnamese waters?” McNamara again eluded the question, “They have advanced closer and closer to the 17th parallel, and in some cases, I think they have moved beyond that in an effort to stop the infiltration closer to the point of origin.”26
In reality, McNamara knew full well that the 34A attacks had probably provoked the 2 August attacks on the Maddox. On an audio tape from the Johnson Library declassified in December 2005, he admitted to the President the morning after the attacks that the two events were almost certainly connected:
And I think I should also, or we should also at that time, Mr. President, explain this OPLAN 34-A, these covert operations. There’s no question but what that had bearing on it. On Friday night, as you probably know, we had four TP [sic] boats from [South] Vietnam, manned by [South] Vietnamese or other nationals, attack two islands, and we expended, oh, 1,000 rounds of ammunition of one kind or another against them. We probably shot up a radar station and a few other miscellaneous buildings. And following 24 hours after that with this destroyer in the same area undoubtedly led them to connect the two events. . . .”27
Intelligence officials realized the obvious. When President Johnson asked during a 4 August meeting of the National Security Council, “Do they want a war by attacking our ships in the middle of the Gulf of Tonkin?” CIA Director John McCone answered matter-of-factly, “No, the North Vietnamese are reacting defensively to our attacks on their offshore islands . . . the attack is a signal to us that the North Vietnamese have the will and determination to continue the war.”28
Johnson himself apparently had his own doubts about what happened in the Gulf on 4 August. A few days after the Tonkin Gulf Resolution was passed, he commented, “Hell, those damn, stupid sailors were just shooting at flying fish.”29
Can the omission of evidence by McNamara be forgiven? Within time, the conflict in Vietnam would likely have occurred anyway, given the political and military events already in motion. However, the retaliatory attack of 5 August marked the United States’ first overt military action against the North Vietnamese and the most serious escalation up to that date. The Tonkin Gulf Resolution, essentially unchallenged by a Congress that believed it was an appropriate response to unprovoked, aggressive, and deliberate attacks on U.S. vessels on the high seas, would open the floodgates for direct American military involvement in Vietnam. McNamara’s intentional distortion of events prevented Congress from providing the civilian oversight of military matters so fundamental to the congressional charter.
Some historians do not let the Johnson administration off so easily. Army Colonel H. R. McMaster, author of the highly acclaimed 1997 book Dereliction of Duty, accused Johnson and McNamara of outright deception:
To enhance his chances for election, [Johnson] and McNamara deceived the American people and Congress about events and the nature of the American commitment in Vietnam. They used a questionable report of a North Vietnamese attack on American naval vessels to justify the president’s policy to the electorate and to defuse Republican senator and presidential candidate Barry Goldwater’s charges that Lyndon Johnson was irresolute and “soft” in the foreign policy arena.30
For his part, McNamara never admitted his mistakes. In his award-winning 2003 video memoirs Fog of War, he remained unapologetic and even bragged of his ability to deceive: “I learned early on never answer the question that is asked of you. Answer the question that you wish had been asked of you. And quite frankly, I follow that rule. It’s a very good rule.”31
We may never know the whole truth behind the Tonkin events and the motivations of those involved. However, it is important to put what we do know into context. The administration’s zeal for aggressive action, motivated by President Johnson’s election worries, created an atmosphere of recklessness and overenthusiasm in which it became easy to draw conclusions based on scanty evidence and to overlook normally prudent precautionary measures. Without the full picture, Congress could not offer the checks and balances it was designed to provide. Subsequently, the White House carried the nation into the longest and one of the most costly conflicts in our nation’s history.