What the tapes actually say

Nixon’s recording system began in February 1971 and expanded to several locations and telephones. Most room recording was sound-activated; the Cabinet Room was an important exception with manual activation. The Secret Service maintained the equipment. The result was an unusually extensive record, not a perfect recording of every presidential conversation everywhere. [1]

On June 23, 1972, Nixon and Haldeman discussed the FBI’s tracing of burglary money and a plan to use the CIA to limit that inquiry. Listen to the surviving exchange, not just the “smoking gun” nickname. Its force lies in the president’s participation in an obstruction-related discussion only six days after the arrests. It does not itself prove that Nixon had ordered the original burglary. [2]

On March 21, 1973, Dean described the cover-up as a danger to the presidency and discussed blackmail, money demands and perjury. This recording documents a detailed conversation about an ongoing problem. It is not the same tape as June 23, and it should not be used to imply that Nixon first learned anything about obstruction that March. [3]

Audio also has limits. These recordings can be difficult to hear; a transcript is an aid to listening, not a guarantee that every uncertain word is settled. Read the participants, date and surrounding discussion before turning a memorable sentence into a claim about an entire conspiracy. The strongest conclusions come from the recorded conduct in context. [1][2][3]

An approximately eighteen-and-a-half-minute stretch of a June 20, 1972 Nixon–Haldeman recording was replaced by a hum. Prosecutor Jill Wine-Banks’s firsthand account describes questioning Nixon’s secretary Rose Mary Woods about her explanation that she accidentally recorded over material while reaching for a telephone. That explanation was evidence to test, not a judicial finding identifying who caused the entire gap. [1]

Court-appointed audio experts investigated the physical tape and machines. Their technical work supported multiple erasures through manual machine operation, rather than treating the gap as a simple untouched blank. The record is materially more troubling than an ordinary bad microphone. But a forensic finding about how erasure occurred cannot, without more evidence, name the operator or establish that person’s motive. [2]

Later technology did not produce a recovered conversation. In 2003 the National Archives reported controlled tests using comparable recording and erasure equipment. Participants recovered no intelligible speech from the relevant simulation tapes. NARA retained the original in hope of future methods; it did not announce a secret successful reconstruction. [3]

Attempts to examine Haldeman’s handwritten notes were a separate avenue of investigation. Notes can supply context; they are not restored audio. This leaves a legitimate historical unknown. Claims that confidently quote the erased words—or use their absence to erase the surviving June 23 and March 21 conversations—go beyond what these sources establish. [3][4]