For decades, one sentence from the House Select Committee on Assassinations has lived an unusually energetic afterlife: scientific acoustical evidence established a high probability that two gunmen fired at President Kennedy.
That sentence matters because it was not a witness impression or a novelist's theory. It was presented as science, and it helped support the committee's conclusion that Kennedy was probably assassinated as the result of a conspiracy.
Then the science was reviewed again.
The chain of inference
Dallas police communications were recorded during the assassination. One channel contained a period in which a microphone was apparently stuck open. Analysts working for the HSCA compared impulse patterns on that recording with acoustic test shots and concluded that several impulses were consistent with gunfire, including one associated with the grassy-knoll area.
The HSCA integrated those results with other evidence and reached its two-gunman finding. In other words, the conspiracy conclusion was not floating independently. It had an evidentiary dependency.
That dependency is crucial because good historical reasoning is not just a list of facts. It is a graph. If one high-weight node fails, every conclusion resting heavily on it must be recalculated.
The timing problem was already visible.
The committee's acoustic reconstruction identified impulse patterns separated by about 1.66 seconds. That created tension with earlier rifle testing suggesting that aimed shots using the telescopic sight required approximately 2.25 to 2.3 seconds between shots.
The HSCA explored whether open iron sights could allow faster fire. Its own testing suggested the rifle could be cycled faster under those conditions. That part of the debate matters because it demonstrates something often forgotten: the committee was not simply choosing between “one gunman” and “two gunmen.” It was trying to make multiple technical datasets coexist.
Then the recording itself came under attack.
The National Research Council's Committee on Ballistic Acoustics reviewed the prior acoustic studies and the Dallas police recordings. Its 1982 report did not merely say the earlier probability estimate was a little too confident. The committee unanimously concluded that the acoustic analyses did not demonstrate a grassy-knoll shot and that there was no acoustic basis for the claimed 95 percent probability.
A central issue involved timing and cross-talk between police channels. The NRC concluded that identifiable speech and timing relationships showed the supposed assassination impulses were not positioned in the recording as the HSCA analysis required.
Once that happens, pattern matching to simulated gunshots becomes much less persuasive. A waveform can resemble a gunshot and still not have occurred when the assassination occurred.
Similarity without chronology is a trap.
This is the general methodological point.
Pattern recognition is seductive. An acoustic impulse, radar track, blurry image, DNA partial match, or AI classification can look like the thing we are searching for. But resemblance is only one constraint. Time, location, chain of custody, sensor state, and competing causes also have to fit.
If the event candidate is in the wrong temporal window, the quality of the visual or acoustic resemblance becomes almost irrelevant.
Does the NRC report prove there was no second gunman?
No. This does not prove the absence of another shooter.
The NRC report attacks the acoustic basis for the second-gunman claim. It cannot establish the absence of another shooter by every possible line of evidence. A separate second-shooter theory would need separate evidence.
What it does do is remove one of the strongest scientific supports used by the HSCA. Therefore anyone citing the HSCA's “high probability” language today without also disclosing the subsequent National Research Council review is presenting a historical conclusion without its maintenance record.
Scientific conclusions need version history.
This case is a perfect argument for visible revision trails in journalism and public research.
Version 1: acoustic analysts report evidence consistent with multiple shots and a grassy-knoll origin.
Version 2: the HSCA incorporates that result into a probable-conspiracy finding.
Version 3: FBI review challenges the acoustic interpretation.
Version 4: the National Research Council reviews the technical record and unanimously rejects the acoustic basis for the grassy-knoll probability claim.
The correct historical representation includes all four versions. Otherwise we turn a changing technical assessment into a frozen quotation.
This is what “follow the science” is supposed to look like.
Science is strongest when it can embarrass its previous self.
The acoustic story does not show that expert analysis is worthless. It shows why independent review, alternative timing evidence, methodology criticism, and replication matter. The original analysts produced a testable claim. Other experts attacked it. The attack was specific enough to change how the evidence should be weighted.
That is not institutional weakness. That is the machinery working.
The HSCA's second-gunman conclusion cannot responsibly be repeated today as if its acoustic foundation remained unchallenged. The National Research Council's unanimous 1982 review substantially undercuts that foundation. This weakens one major route to a conspiracy conclusion; it does not logically disprove every possible conspiracy theory.
A rule for the whole site
Whenever Cyberdelia cites a famous technical conclusion, we should ask what happened to it afterward. Was it replicated? Reanalyzed? Withdrawn? Superseded? Did the underlying dataset change? Did later evidence break a hidden assumption?
The internet is excellent at preserving the first dramatic sentence. We should specialize in preserving the second report that made the first one uncomfortable.