Excerpt from Ellsberg's Memoir, Secrets

The Tonkin Gulf: August 1964

Chapter 1. The Tonkin Gulf: August 1964

The Most Dangerous Man in America - Secrets, Ellsberg Memoir Book Cover

Ellsberg, Daniel. Secrets. Penguin, 2003.

On Tuesday morning, August 4, 1964, my first full day on my new job in the Pentagon, a courier came into the outer office with an urgent cable for my boss. He'd been running. The secretaries told him Assistant Secretary John McNaughton was out of the office; he was down the hall with Secretary of Defense Robert McNamara. They pointed him to me, his new special assistant. The courier handed me the cable and left. It was easy to see, as I read it, why he had been running.

It was from Captain John J. Herrick, the commodore of a two-destroyer flotilla in the Tonkin Gulf, off North Vietnam in the South China Sea. He said he was under attack by North Vietnamese patrol boats and had opened fire on them. He was in international waters, over sixty miles off the coast of North Vietnam. One torpedo had been heard by the sonarman on his command ship, the USS Maddox, and another had just passed by the other destroyer, the Turner Joy.

As soon as he gave me the cable, the courier returned to the message center of our department in the Pentagon, International Security Affairs (ISA), part of the Office of the Secretary of Defense, the civilian part of the Department of Defense. Within ten minutes he was back to me with another one in the same series: "Am under continuous torpedo attack."

A few minutes later Herrick reported another torpedo had run by him, and two more were in the water. His ships were firing at the attackers and might already have destroyed one of them. They were firing by radar, without visual contact. The encounter was taking place in total darkness, on an overcast night without moon or stars, in the hours close to midnight.

Gulf of Tonkin incident. Source: U.S. Navy Historical Center, Washington Navy Yard / WikiCommons

Illustration, Gulf of Tonkin Incident. Source: U.S. Navy Historical Center, Washington Navy Yard

This was no ordinary event. It was exactly the second attack on a U.S. Navy vessel since World War II. But the first had been less than three days earlier. That was on Sunday, August 2, also on Herrick's ship, the USS Maddox, on patrol in the Tonkin Gulf. In broad daylight in the middle of the afternoon, twenty-eight miles out to sea, three North Vietnamese PT boats had attacked and launched torpedoes at the Maddox. All the torpedoes had missed, and there was no damage to the destroyer, except for a single 14.5-mm bullet that lodged in one of its stacks. The boats were driven off, all damaged, by fire from the Maddox and from navy planes from the carrier Ticonderoga nearby.

Since there had been no American casualties or significant damage, President Johnson had decided to take no further action, except to add another destroyer, the Turner Joy, to the mission. The two destroyers were directed to continue what was described publicly as a routine patrol in order to assert U.S. rights to navigate freely in international waters. But the president also announced on Monday his orders that in case of any further attacks, the attacking boats were to be not only repulsed but destroyed. He had sent a formal protest to Hanoi, warning that "any further unprovoked offensive military action against United States forces" would "inevitably" result in "grave consequences." All this, except for the latest announcement, I'd read in the Monday morning newspapers. That afternoon, reading classified accounts of the episode, I'd learned a good deal more.

Now, as each new message came in, I looked at the date-time group, the six-digit number (followed by a letter indicating the time zone, then the month) at the upper-left-hand corner of the cables. The first two digits indicated the day of the month; the next four, in military time (2400 for midnight), the exact time the message had been transmitted. The first cable had been transmitted from Herrick's command ship at 10:42 a.m. Washington time (9:42 p.m. in the Tonkin Gulf). I compared the time of transmission with the clock on the wall of my office in the Pentagon, which showed, as I recall, that it was about half an hour later, an extremely short time in this precomputer age for this message to reach me. The same was true for the second, sent at 10:52 a.m. Washington time and handed to me about 11:20, and for the others that kept arriving every few minutes. Herrick was giving them "Flash" priority, the highest priority for message handling, so they were taking precedence at every terminal for handling, retransmission, and distribution.

But twenty or thirty minutes was a long duration for an action like this. The whole exchange on Sunday, surface and air, had lasted thirty-seven minutes. It could have been all over, on the other side of the world, by the time I read the first message, or the latest one. Or a destroyer might have been hit, might already be sinking, while we were reading about its evasive maneuvers or its success at destroying an attacker. But there was no way for anyone in Washington to know that as he read these.

There was then no CNN on which to watch live action half a world away. There was not even any direct voice contact between Washington and destroyers in the western Pacific. The closest to it was radio and telephone contact with Admiral Ulysses S. G. Sharp, commander in chief Pacific (CINCPAC), at his command post in Hawaii, as far away from the Tonkin Gulf as Washington was from Hawaii. CINCPAC cables, and many others, were now adding to the pile on my desk, but they weren't arriving as frequently or as fast as the flash cables from the destroyers. Following Captain Herrick's stream of messages, we weren't really watching the action in real time, but they were coming in such quick sequence that it felt as if we were.

Captain John J. Herrick, USN, Commander Destroyer Division 192 (at left) and Commander Herbert L. Ogier, USN, Commanding Officer of USS Maddox (DD-731), on board Maddox on 13 August 1964. They were in charge of the ship during her engagement with three No

Captain John J. Herrick, USN, Commander Destroyer Division 192 (at left) and Commander Herbert L. Ogier, USN, Commanding Officer of USS Maddox (DD-731), on board Maddox on 13 August 1964. Photographed by PH3 White. Official U.S. Navy Photograph, from the collections of the Naval Historical Center.

The messages were vivid. Herrick must have been dictating them from the bridge in between giving orders, as his two ships swerved to avoid torpedoes picked up on the sonar of the Maddox and fired in the darkness at targets shown on the radar of the Turner Joy: "Torpedoes missed. Another fired at us. Four torpedoes in water. And five torpedoes in water....Have...successfully avoided at least six torpedoes."

Nine torpedoes had been fired at his ships, fourteen, twenty-six. More attacking boats had been hit; at least one sunk. This action wasn't ending after forty minutes or an hour. It was going on, ships dodging and firing in choppy seas, planes overhead firing rockets at locations given them by the Turner Joy's radar, for an incredible two hours before the stream of continuous combat updates finally ended. Then, suddenly, an hour later, full stop. A message arrived that took back not quite all of it, but enough to put everything earlier in question.

The courier came in with another single cable, running again, after an hour of relative quiet in which he had walked in intermittently at a normal pace with batches of cables from CINCPAC and the Seventh Fleet and analyses from the State Department and the CIA and other parts of the Pentagon. I was sitting at my desk -- I remember the moment -- trying to put this patchwork of information in some order for McNaughton on his return, when the courier handed me the following flash cable from Herrick: "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."

It was a little after 2:00 p.m. The message had been sent at 1:27 p.m. Washington time. Half an hour later another message from Herrick, summarizing positive and negative evidence for an attack, concluded: "Entire action leaves many doubts except for apparent attempted ambush at beginning. Suggest thorough reconnaissance in daylight by aircraft." The reconnaissance in daylight, still three or four hours away in the gulf, would search for oil slicks and wreckage from the boats supposedly hit, indications that an attack, not just a fight with radar ghosts, had actually taken place.

In my mind, these messages erased the impact of the two-hour-long "live" drama that we'd been following. This new information was a cold bath. Around three o'clock, in response to frantic requests for confirmation, Herrick cabled, "Details of action present a confusing picture although certain that original ambush was bona fide." But how could he be "certain" of that, or why should anyone else be, when he had seemed equally confident, an hour earlier, of all the succeeding reports up till now? Herrick continued to assert at 6:00 p.m. Washington time (5:00 a.m. in the gulf) that "the first boat to close the Maddox probably fired a torpedo at the Maddox which was heard but not seen. All subsequent Maddox torpedo reports are doubtful in that it is suspected that sonarman was hearing ship's own propeller beat." But his acknowledgment that all the other vivid reports he had been sending were unreliable undercut his assertion of continued confidence in his initial messages and the first torpedo. As negative evidence accumulated, within a few days it came to seem less likely that any attack had occurred on August 4; by 1967 it seemed almost certain there had been no second attack, and by 1971 I was convinced of that beyond reasonable doubt. (In 1966 credible testimony from captured North Vietnamese officers who had participated in the August 2 attack refuted any attack on August 4. In late 1970 journalist Anthony Austin discovered and gave me evidence that intercepted North Vietnamese cables supposedly confirming an August 4 attack actually referred to the attack on August 2. Finally, in 1981 journalist Robert Scheer convinced Herrick -- with new evidence from his ship's log -- that his long-held belief in the first torpedo report was unfounded.) However, on August 4, given Herrick's repeated assurances and those of a number of seamen over the next few hours, I concluded that afternoon, along with everyone else I spoke to, that there probably had been an attack of some sort. At the same time, there was clearly a good chance that there had been none. In that light, Herrick's recommendation to pause and investigate before reacting seemed prudent, to say the very least: Reverse engines, stop the presses! But that was not how things were moving in Washington that Tuesday afternoon.

Herrick's new cables didn't slow for a moment the preparations in Washington and in the Pacific for a retaliatory air strike as quickly as possible, preferably at first light in the Tonkin Gulf. What they did stimulate was a flurry of probes for evidence and witness testimony that would support his earlier descriptions of the attack or at least confirm the fact that some attack had occurred.

As these were arriving in Washington, the president was meeting with the National Security Council (NSC) basically to inform it of the planned actions. Next he briefed congressional leaders. Carriers were moving into position to launch their planes at first light or as early in the morning as possible. In Washington time that could be anywhere from six o'clock in the evening to nearly midnight. But the president was determined to tell the American people of the U.S. attacks more or less as they were happening. He didn't want them to hear about the strikes in the morning news the next day, hours after they had taken place and after the rest of the world, in earlier time zones, had already heard.

The navy was concerned, on the other hand, not to have the president's public announcement warn Vietnamese antiaircraft gunners that an attack was coming before the planes had entered North Vietnamese radar. The president undertook not to do that. He asked for airtime for 7:00 p.m., which shifted to 8:00, then to 9:00, because the carrier Constellation had still not reached its launching station or finished briefing its pilots. The president was determined to speak no later than 11:30 p.m. After that his entire audience on the eastern seaboard would be in bed. Through McNamara to CINCPAC (Admiral Sharp, in Hawaii), he was pressing to see if he could make his announcement before the planes were over their targets, perhaps when the first ones started to launch. Would they be picked up immediately on radar, he asked, so that it wouldn't be his announcement then that broke the news to Hanoi? The answer was yes, but Hanoi wouldn't know where the planes were heading, so he should take numbers and types of targets off the TelePrompTer.

At this point in the evening I was sitting with John McNaughton in his office along with his director of Far Eastern affairs and other members of his staff, reading cables from the carriers and CINCPAC on progress toward the launch and trying to help answer questions from McNamara or the White House. The large TV in McNaughton's office was on continuously, with the sound turned down, in case the president decided to break in on the programming.

Word came in that planes had taken off, then word that they had not; requests arrived that the announcement be delayed till the planes were on enemy radar, but it was too late for that. Admiral Sharp (CINCPAC) told McNamara at 11:20 p.m. that the Ticonderoga had launched its planes, and the president went on TV at 11:37. He announced that "air action is now in execution," though in fact the Constellation had not yet launched its planes and no other planes had as yet reached the coast of North Vietnam or entered its radar. So the announcement did give Hanoi warning, which it passed down quickly. Our navy concluded from the results that surprise had been sacrificed.

McNamara gave a press conference at the Pentagon after midnight. We were up all night in the office following the raids, to prepare for another McNamara press conference the next day. My first full day in the Pentagon had been over twenty-four hours long.

The president's announcement and McNamara's press conference late in the evening of August 4 informed the American public that the North Vietnamese, for the second time in two days, had attacked U.S. warships on "routine patrol in international waters"; that this was clearly a "deliberate" pattern of "naked aggression"; that the evidence for the second attack, like the first, was "unequivocal"; that the attack had been "unprovoked"; and that the United States, by responding in order to deter any repetition, intended no wider war.

By midnight on the fourth, or within a day or two, I knew that each one of these assurances was false.

"Unequivocal"? In the president's initial public announcement and in every official statement afterward, it was implicit that the August 4 attack on our ships, which had triggered our retaliatory strikes, was a simple fact. There was no official hint, either to Congress or to the public, that in the minds of various experienced navy operators and intelligence analysts at the time of our retaliation, as well as earlier and later, doubt adhered to every single piece of evidence that an attack had occurred at all on August 4.

A "routine patrol in international waters"? The two destroyers were on a secret intelligence mission, code-named DeSoto patrols, penetrating well within what the North Vietnamese regarded as their territorial waters. We assumed, correctly, that the North Vietnamese claimed the same limits as other Communist nations, twelve miles from their coastline and from their islands. The United States did not officially "recognize" this extended limit; nevertheless U.S. Navy ships were prudently directed to keep at least fifteen miles out from the Chinese islands or mainland. But before the August 2 incident the Maddox had been frequently eight miles from the North Vietnamese mainland and four miles from their islands. The purpose of this was not merely to demonstrate that we rejected their claims of limits on our "freedom of the seas" but to provoke them into turning on coast defense radar so that our destroyers could plot their defenses, in preparation for possible air or sea attacks. Thus it was true that the August 2 attack had been twenty-eight miles out to sea, but that was because a warning of attack when the Maddox was just ten miles from the coast had led the skipper to change course and to head out to sea, with torpedo boats in pursuit.

"Unprovoked"? Hanoi had claimed that "puppet" forces of the Americans had shelled two of its coastal islands, Hon Me and Hon Nieu, on the night of July 30-31. In public releases, the State Department denied any knowledge of any such attacks, as did McNamara in his press conferences on August 4 and 5. In top secret testimony to congressional committees in closed hearings over the next two days, Secretary of State Dean Rusk and McNamara acknowledged such attacks but insisted that they could not realistically be considered U.S. provocations that justified or were intended to evoke North Vietnamese counterattacks because they were entirely "South Vietnamese" operations, run by the South Vietnamese navy, aimed at stopping infiltration from the North. The United States supported them and knew about them in general terms but, Rusk claimed, not in detail; there was little knowledge of them in Washington. They had no relationship at all with our destroyer patrols, they were in no way coordinated, and in fact the commander on the destroyers knew nothing of them at all. It was implicit in this testimony, and not challenged, that in any case no such raids were taking place in the context of the second attack or since July 31. The resolution that Congress was being asked to pass quickly and as nearly unanimously as possible was nothing other than a gesture of support for the president's action, to demonstrate solidarity to Hanoi and to deter future attacks on our forces. Each of these assertions was false.

In my new job I was reading the daily transcripts of this secret testimony, and at the same time I was learning from cables, reports, and discussion in the Pentagon the background that gave the lie to virtually everything told both to the public and, more elaborately, to Congress in secret session. Within days I knew that the commander of the destroyers not only knew of the covert raids but had requested that his patrol be curtailed or terminated after the first attack on August 2 because he expected retaliatory attacks on his vessels as a result of the raids. His request was denied. Moreover, I learned, these weren't South Vietnamese operations at all, not even joint operations. They were entirely U.S. operations, code-named 34A ops. The anti-infiltration operations by South Vietnamese junks that McNamara described in some detail to Congress were entirely separate and different, as he knew. For the raids against North Vietnam, of which Hanoi had publicly complained, the United States owned the fast patrol boats known as Nastys (which the CIA had purchased from Norway), hired the crews, and controlled every aspect of the operations. The CIA ran the training, with help from the U.S. Navy, and recruited the crews; some of them were recruited, as individuals, from the South Vietnamese navy, but others were CIA "assets" from Taiwan and elsewhere in Asia, along with mercenaries from around the world. The operations had been run originally by the CIA but now were jointly controlled by the CIA and Military Assistance Command, Vietnam (MACV), in coordination with the navy. Despite the use of foreign personnel, to provide "plausible deniability" if captured, the 34A operations were exactly as much American operations as were the U.S. Navy DeSoto patrols of the destroyers. Moreover, the North Vietnamese were not mistaken to believe that the two types of American operations were coordinated at various levels. For one thing, the DeSoto missions in that particular area were timed to take advantage, in their plotting of coastal radars and interception of communications, of the heightened activity that was triggered in North Vietnamese coastal defenses by the 34A raids.

As for Washington knowledge of them, top officials read and signed off personally on schedules for them in advance, based on incredibly detailed descriptions of the planned actions. I soon knew this because I came later that month to be the courier who carried these highly secret plans around Washington from one to another of these officials for their signatures. These included Deputy Secretary of Defense Cyrus Vance, Deputy Secretary of State Llewellyn Thompson, and finally, National Security Adviser McGeorge Bundy in the White House. They were among the members of the 303 Committee, which oversaw and approved all covert operations for the president. While they read the documents, I sat in their offices, along with a colonel from the covert operations branch of the Joint Chiefs of Staff (JCS) who had initially brought the file to me.

The contrast between what the senators had been told by the secretaries in a secret joint session of the Senate Foreign Relations and Armed Services committees, as I read the testimony, and what I soon knew as a first-week staffer in the Pentagon was striking. Pressed by Senator Frank Church to acknowledge that "our government which supplied these boats" (supposedly, as he had just been told, to the South Vietnamese) did know that they would be used for attacks on North Vietnam, Secretary Rusk replied, "In the larger sense, that is so, but as far as any particular detail is concerned we don't from Washington follow that in great detail."

In contrast with this disclaimer, as I knew very well, it would have been more accurate to say that every particular detail of these operations was known and approved by the highest authorities in Washington, both military and civilian. The monthly plan for September 1964, the month following the August raids, which I carried over to the State Department to be read and initialed by Mr. Rusk's deputy and then to McGeorge Bundy in the White House, included the following scheduled actions:

Two junk capture missions; remove captives for 36-48 hours interrogation; booby trap junk with antidisturbance devices and release; captives returned after interrogation; timing depends upon sea conditions and current intelligence;...Demolition of Route 1 bridge by infiltrated team accompanied by fire support teams, place short-delay charges against spans and caissons, place antipersonnel mines on road approaches;...Bombard Cape Mui Dao observation post with 81 MM mortars and 40 MM guns from two PTFs;...Destruction of section of Hanoi-Vinh railroad by infiltrated demolition team supported by two VN [Vietnam] marine squads, by rubber boats from PTFs, place short-delay charges and anti-personnel mines around area....

Some of these operational details, such as the placement of antipersonnel weapons and 81-mm mortar rounds, might have seemed rather petty to be occupying the attention of these officials, but this was the only war we had. Of course it was precisely the "sensitive" nature of the operations-their illegality, the danger both of exposure and of escalation, and their covertness, defined as "plausible deniability"-that required such high-level officials to lie to the Senate if questions were raised and therefore to need such detailed prior awareness and control of what it was they would have to lie about.

This wasn't the end of the coordination in Washington. After a monthly program like this was approved, General William Westmoreland, U.S. military commander in Vietnam, requested approval for execution of each individual maritime mission, and I again carried these around for approval. When an attack that had earlier been approved in Washington for the following month actually took place-the exact timing would depend on weather and sea conditions-that fact and its results were reported back to Washington before another attack was approved by Washington. On August 2, during the Sunday morning meeting in which President Johnson was told of the daylight attack on the Maddox, there was discussion of the results of the July 31 covert attacks on the islands, and the president personally approved the next proposed covert raids, for the nights of August 3 and August 5.

On the evening of the fourth, at an NSC meeting when the president asked, "Do they want war by attacking our ships in the middle of the Gulf of Tonkin?" Director of Central Intelligence John McCone answered: "No. The North Vietnamese are reacting defensively to our [sic] attack on their off-shore islands. They are responding out of pride and on the basis of defense considerations." He was referring to the July 31 raids, but his answer covered the supposed attack that morning, since there had been another raid, this time on the North Vietnamese mainland, the night before. This estimate did not prevent the president from saying, in his message as he urged Congress to pass the resolution days later: "We have answered their unprovoked aggression...."

On August 7 Congress approved the Tonkin Gulf Resolution, which reads: "Congress approves and supports the determination of the President, as Commander in Chief, to take all necessary measures to repel any armed attack against the forces of the United States and to prevent further aggression....The United States is...prepared, as the President determines, to take all necessary steps, including the use of armed force, to assist any member or protocol state of the Southeast Asia Collective Defense Treaty requesting assistance in defense of its freedom" [emphasis added].

There was some unease expressed regarding the unusually vague and open-ended scope of the resolution drafted by the administration. Senator Wayne Morse called it a predated declaration of war. Senator Gaylord Nelson offered an amendment expressing a sense in Congress that "[o]ur continuing policy is to limit our role to the provision of aid, training assistance, and military advice," and "we should continue to attempt to avoid a direct military involvement in the Southeast Asian conflict." Senator Fulbright, who managed passage of the resolution in the Senate, said he believed this amendment was "unobjectionable" as "an accurate reflection of what I believe is the President's policy." He rejected it only because (as Johnson had stressed to him in private) the delay in passage to resolve differences in language between the House and Senate versions would weaken the image of unified national support for the president's recent actions. At this moment it was announced that the House had passed the resolution 416 to 0 after forty minutes of debate. Fulbright hoped the Senate would approach that unanimity. Soon after this the Senate voted 88 to 2, with only Senators Morse and Ernest Gruening voting against it.

Several senators, including George McGovern, Frank Church, Albert Gore, and the Republican John Sherman Cooper, had expressed the same concern as Nelson. Fulbright acknowledged that the language was broad enough to permit the president to launch direct combat involvement, including U.S. infantry divisions, which was what worried them. But they accepted Fulbright's assurances-reflecting his talks with officials including the president-that there was no consideration in the administration of using the resolution as an authorization for changing the American role in the war. He had "no doubt that the president will consult with Congress in case a major change in present policy becomes necessary."Most of the Democrats saw the resolution mainly as a way to get a strong expression of bipartisan support for the president's forceful action, undercutting Goldwater's campaign claim that Johnson was uncertain in foreign affairs and indecisive in Vietnam. By thus helping to defeat Goldwater, they saw their support for the resolution as a way of avoiding escalation in Vietnam, which only Goldwater was promising.

But Fulbright's assurances, all of them, were as unfounded as those of Johnson, Rusk, and McNamara. The difference was that he didn't know it. He had been deceived, and in turn, unwittingly, he misled the Senate. Of all the week's deceptions, these were by far the most significant.

We seek no wider war? But the president that summer was secretly and explicitly threatening the Hanoi regime with a wider war against North Vietnam itself, unless its leaders took steps to end the conflict that no one in the administration thought they were likely to take. Johnson's messages to Ho Chi Minh, through a Canadian intermediary, amounted to a secret promise by the president of the United States to the leaders in Hanoi to widen the war unless they called it off.

The warnings were being delivered to North Vietnam by Blair Seaborn, the Canadian member of the International Control Commission (ICC), set up to monitor observance with the 1954 and 1962 Geneva Accords. In his first meeting in Hanoi on June 18, he had met privately with Prime Minister Pham Van Dong. Seaborn had relayed the warning, drafted by U.S. officials and coordinated with the Canadians, that "U.S. public and official patience with North Vietnamese aggression is growing extremely thin," and that if the conflict should escalate, "the greatest devastation would of course result for the DRV [Democratic Republic of Vietnam, or North Vietnam] itself."

Among those who had advocated these threats-virtually all of the president's civilian and military advisers-no one regarded them as bluffs. The Joint Chiefs of Staff had been directed to make detailed plans for air attacks on North Vietnam. By the end of May it had completed studies and preparations, down to target folders for a recommended list of ninety-four targets. The targets for retaliation selected so quickly on August 5 had simply been drawn from this ninety-four-target list. Both this planning and the warning by a Canadian intermediary figured in detailed scenarios coordinated within the government since March and April-most recently on May 23-leading up to a "D-Day" air assault on North Vietnam, to continue until "terrorism, armed attacks, and armed resistance to pacification efforts in the South stop." Another key element, scheduled for D-20 (twenty days before the attacks began), was:"Obtain joint resolution [from Congress] approving past actions and authorizing whatever is necessary with respect to Vietnam."

Although the detailed thirty-day scenario approach was shelved by the president's top advisers in late May, they recommended to him as separate items that month nearly all of its pre-D-Day elements, including those above. They also recommended an initial strike against North Vietnam to underline the secret warning. This followed a proposal by Ambassador Henry Cabot Lodge in Saigon, a strong advocate of attacks on the North who had earlier in the spring introduced the notion of the warning through Canada. On May 15, in a message to the president, he suggested:

If prior to the Canadian's trip to Hanoi there has been a terroristic act of the proper magnitude, then I suggest that a specific target in North Vietnam be considered as a prelude to his arrival....

This had not occurred prior to Seaborn's first visit to Hanoi in June. But his second visit was scheduled for August 10. The events of August 2-7 allowed the United States to point out, in case of any doubt in Hanoi, just what that warning meant in concrete terms. Moreover, the second discussion would allow the administration to make clear what it felt entitled to do with the authority granted by the Tonkin Gulf Resolution, lest Hanoi had been misled by the interpretation Senator Fulbright had given to his fellow Democrats.

To these ends my new boss, John McNaughton, was asked to draft instructions for Seaborn's August 10 session. That was why McNaughton chose to tell me about and to show me a file on the threat process, describing it as one of the most closely held secrets in the administration. He told me that Imust not hint of the existence of this process to anyone, including any of his own deputies. One reason for the extreme secrecy of the information McNaughton gave me was that it was a very dubious role for an ICC commissioner to be conveying U.S. threats to Hanoi. (An intermediary was needed because the United States had no formal representation or contact with the Hanoi regime.) That role could not be known to the other members of the ICC, Poland and India, or to the Canadian Parliament or public, which would not be as quick to accept it as Canadian Prime Minister Lester Pearson. But what was most "sensitive" about this information was that this official warning by the president to the heads of an adversary state came very close to committing him to the course of action that his Republican opponent, Senator Goldwater, was advocating and that President Johnson was opposing and describing in his campaign as dangerously reckless. Moreover, it put the administration's intentions with respect to the Tonkin Gulf Resolution in an entirely different light from what Congress was being told. Indeed, on August 7, as Congress was voting on the Tonkin Gulf Resolution, John McNaughton was drafting instructions on the message Seaborn should (and later did) deliver that precisely reversed the emphasis on the two key clauses in the resolution that Senator Fulbright had been encouraged by the administration to convey to his fellow senators. His draft, which was adopted by the administration and followed by the Canadians, told Seaborn to conclude his comments with the points:

a. That the events of the past few days should add credibility to the statement made last time, that "U.S. public and official patience with North Vietnamese aggression is growing extremely thin."

b. That the congressional resolution was passed with near unanimity, strongly reaffirming the unity and determination of the U.S. government and people not only with respect to any further attacks on U.S. military forces but more broadly to continue to oppose firmly, by all necessary means, DRV efforts to subvert and conquer South Vietnam and Laos.

c. That the U.S. has come to the view that the DRV role in South Vietnam and Laos is critical. If the DRV persists in its present course, it can expect...to suffer the consequences.

Pham Van Dong's reaction on August 13, as a State Department report described it, was "extremely angry" and cold. And unyielding, as on the first visit (when the exchange had been friendlier, despite the threat). Then he had said that the prospect for the United States and its friends in South Vietnam was "sans issue": no way out, a dead end. Now, in the aftermath of the American raids, he said that the United States had found "it is necessary to carry the war to the North in order to find a way out of the impasse...in the South."

He had gotten the message. (It remained a secret from the American electorate, and from Congress, for the next eight months.) A wider war was on the way.

From Secrets: A Memoir of Vietnam and the Pentagon Papers by Daniel Ellsberg, Copyright © October 2002, Viking Press, a member of Penguin Putnam, Inc., used by permission