NSA’s Gulf of Tonkin Files: The Declassified Record of an Attack That Didn’t Happen (1964)

OFFICIAL FILES · Events of August 1964. The NSA declassified these records in two releases and later re-posted them on media.defense.gov (path dated 14 Jul 2021).

The centerpiece is NSA historian Robert J. Hanyok’s article “Skunks, Bogies, Silent Hounds, and the Flying Fish: The Gulf of Tonkin Mystery, 2-4 August 1964.” It examines the signals intelligence behind the reported second attack on U.S. destroyers on 4 August 1964, the event that led to the Gulf of Tonkin Resolution. The surrounding files include messages, chronologies and interviews (REL1/REL2 series, listed below).

See our historical case file gulf-of-tonkin-1964 for the full story and grades.

CASE FILE · NSA-TONKIN
Publisher: National Security Agency

Timeline: from secret history to public file

• 2–4 August 1964: The destroyer USS Maddox is attacked by North Vietnamese torpedo boats on 2 August. That attack is not disputed. On the night of 4 August, Maddox and USS Turner Joy report a second attack. Congress passes the Gulf of Tonkin Resolution on 7 August.
• Winter 2000/Spring 2001: Hanyok’s article appears in NSA’s internal, classified journal Cryptologic Quarterly.
• October 2005: The New York Times reports his conclusions.
• 30 November 2005: NSA puts out the first installment of Tonkin material, including a “moderately sanitized” version of the article. Our REL1_ files come from that release, and the REL2_ files are from the follow-up.
• January 2008: NSA releases Hanyok’s wider history of SIGINT in the Indochina wars, which contains the Tonkin study as one chapter.

What the article actually says (quoted from our local PDF)

Hanyok describes “two startling findings”:

1. No attack happened. “It is not simply that there is a different story as to what happened; it is that no attack happened that night.” He writes that on 4 August “Hanoi’s navy was engaged in nothing that night but the salvage of two of the boats damaged on 2 August.”
2. The intelligence was presented selectively. Through October 1964, the SIGINT “was presented in such a manner as to preclude responsible decisionmakers in the Johnson administration from having the complete and objective narrative.” The intercepts left out of the summary reports “constituted nearly 90 percent of all available SIGINT.”

Was it a conspiracy? The NSA’s own historian says no

The mainstream account and the conspiracy account split here, and Hanyok takes a position on it. He writes that the mishandling “was not done in a manner that can be construed as conspiratorial, that is, with manufactured evidence and collusion at all levels.” In his account, the people involved wanted to back up the Navy’s claim that the patrol had been deliberately attacked, and couldn’t do that while also passing along all of the evidence. Hanyok traces the error back to analysts at Phu Bai, who wrongly believed an attack was coming, and to the deference NSA headquarters gave to people closer to the action.

The 2005 New York Times report summarized his work as finding that staff “deliberately skewed” the evidence. Readers should keep both framings in mind: the evidence was knowingly presented in a one-sided way, but there was no fabricated master plan.

Context the documents add

• The patrol wasn’t routine. The Maddox was on a DESOTO intelligence patrol while covert South Vietnamese raids under OPLAN 34-A were hitting the North Vietnamese coast. On 30 July 1964, commandos attacked a radar site on Hòn Mê island. President Johnson told Congress the attack was “unprovoked.”
• The task group commander, Capt. John J. Herrick, had doubts in real time. According to Hanyok, he was “not entirely certain what had transpired.” The release includes a file on him (REL1_HERRICK.PDF).
• McNamara’s later admission. In 1995, former Defense Secretary Robert McNamara asked Gen. Võ Nguyên Giáp about 4 August. In the 2003 documentary The Fog of War, he acknowledged that the second attack did not happen.

How to read the local files

The REL1 and REL2 PDFs are a mix of chronologies, message traffic, interviews and later reviews. Many pages are heavily redacted and the scans are poor. Our quotes come from the text layer of rel1_skunks_bogies.pdf.

Grades