New Evidence on the
Gulf of Tonkin
BY JOHN PRADOS
Just before midnight on the final
day of November 2005, more than 41 years after the event, the
National Security Agency declassified a large swath of material
related to the August 1964 Gulf of Tonkin Incident. That famous
episode in the history of the Vietnam War is now revisited in the
release of significant documentation on perhaps the final piece of
the puzzle: the communications intercepts that bore on the battle.
The new evidence conveys several significant insights.
On August 2, 1964, the U.S.
destroyer Maddox fought a battle against several North
Vietnamese torpedo boats not long after a coastal raid on the
North carried out by American and South Vietnamese special forces.
The Johnson administration promptly sent the Maddox, which
had withdrawn after the action, back into the Gulf, reinforced by
destroyer C. Turner Joy. On the night of August 3, the
ships radioed that they were again under attack. The Johnson
administration took this repeat attack as a token of aggression
and used it to justify the first American bombing of North Vietnam
as well as the passage of the Gulf of Tonkin Resolution, which was
the legal basis for waging the war in Vietnam.
Events in the Gulf of Tonkin have
remained controversial. In the middle of the supposed second
attack, the naval commander on the scene sent a dispatch that
called into question his own action reports, without any impact on
Washington’s decisions. Johnson officials staunchly insisted that
its accounts of both incidents were true. The general verdict of
history has been that the August 2 battle indeed took place, but
that the second incident was a fiction. Even the author of the
Navy’s official history of the Vietnam War covering this period,
who followed the Johnson script, now concedes the inaccuracy of
claims for the night of August 3. The last holdouts among those
who sided with the Johnson administration’s version of events
relied upon the NSA intercept evidence to buttress their position.
The significance of the new
declassifications is that they make the full range of intercept
evidence available for the first time. It turns out that the
claims originally based on the communications intelligence used
only a handful of the NSA intercepts, and they were selectively
culled from a much wider array of material. This fact is
illuminated in considerable detail in a lengthy paper by NSA
historian Robert J. Hanyok, which also is among the newly released
Intercepts used by then-Secretary
of Defense Robert S. McNamara to defend the official version of
events in 1968 and 1969 investigations have long been contested,
and actually do reflect the events of August 2, rather than the
alleged second incident. That was the view of NSA deputy director
Louis Tordella, CIA senior official Ray Cline, the Senate Foreign
Relations Committee, and many historians.
The new material also reveals these
- Only one of four relevant NSA
intercept stations took in the message that became central to
the debate, while one of the other stations reported a version
of the same dispatch that did not sustain the inference that
an attack was taking place.
- Washington officials relied
upon the times at which NSA transmitted the intercepted and
translated messages to place them within the timeframe of the
alleged attack, rather than the times at which North
Vietnamese radio operators had originated the messages, which
were before any of the alleged events were occurring.
- The Vietnamese-language
original of the key message in which a phrase was subject to
interpretation (hurt “sailors” versus damaged “boats”) had
disappeared from the files, even though NSA practice typically
included plaintext at the end of a message as a technical
assistance feature. Two of the five points of the
administration argument for the veracity of events relied upon
this message. All the others were based on alleged radar,
sonar, or visual observations that were increasingly disputed.
- Confusion over whether a
particular radio call sign referred to an individual boat or a
unit permitted further distortion to occur.
- NSA headquarters reports
combined elements of different messages without the
acknowledgment that was standard procedure to issue a summary
containing the starkest possible rendering of available
- The vast majority of NSA
intercepts for August 3 conveyed a picture of a quiet North
Vietnamese naval communications net, with the bulk of those
messages sent pertaining to an effort to recover and tow the
two torpedo boats damaged in the August 2 battle, and none of
the sighting reports or command traffic that would have been
involved in a battle with the Maddox-Turner Joy force.
- The after-action chronology
prepared by an NSA branch chief clearly relied upon the NSA
summary reports and the few culled dispatches rather than the
large mass of traffic. The agency never updated or modified
that chronology and repeatedly issued that document as its
sole element of evidence when Tonkin Gulf was investigated by
White House boards and congressional panels.
When the Tonkin Gulf incident was
investigated by the Senate Foreign Relations Committee and Robert
McNamara had to testify, he used only the NSA chronology to
prepare for his appearance. Therefore the original distortions in
the NSA intercept evidence were perpetuated and observers dug in
their heels based on the flimsiest material. See also John
Prados’s “Six mysteries of the Tonkin Gulf,” in the August 1989
issue of the Veteran.