Forty years ago, a short passage of film - a none-too-smooth panning shot - was exposed. Never mind the great sequences in Godard or Angelopoulos, no piece of moving film has ever been subjected to the scrutiny that fell upon the 8mm frames shot in Dealey Plaza, Dallas, on 22 November, 1963, by Abraham Zapruder.
Zapruder was the ordinary guy with a home movie camera who thought to record the presidential motorcade. It says something about the relative innocence of 1963 that the Zapruder film was "the only continuous chronological visual record of the assassination [and] the best available photographic evidence of the number and timing of the shots that struck the occupants of the presidential limousine." That comes from the 1979 report of the Select Committee on Assassinations (known as the Stokes Report, after its chairman, Louis Stokes). It also echoes the assumptions made by the earlier Warren Report.
Both the Warren and the Stokes reports took the Zapruder film as the only available clock on the split seconds in which firing occurred. Read those detailed reports, and you'll find references to Zapruder 312 (or whatever), which means the frame number from Zapruder's footage. Eye-witness evidence from the Plaza was confused on how many shots there had been, or where they had come from. The Zapruder footage became the template that supplied the Warren Report's description of what had happened - this included the "single bullet theory", the notion that one bullet had struck the president, emerged from his body, struck Governor John Connally (sitting in the front of the limousine), pierced his body and then entered his wrist. That theory was vital to Warren's claim that only one shooter - Lee Harvey Oswald - had fired on the car. This reconstruction of time was all based on what the ordinary eye could see from the Zapruder film - the instant at which a body was registering shock and impact from a bullet.
I am not going over this again to raise old doubts about what really happened. What's more interesting, I think, are the lessons in how we relate to what you might call random film observation or surveillance. You might think that, minus the Zapruder film, the Warren Report would have been at a loss to explain what happened. On the contrary, in that case the conclusions would have been obvious: for we can place Oswald and the gun on the sixth floor of the Texas School Book Depository, and we know that Kennedy was killed and Connally wounded. There would have been no room for doubt.
The film's record - trusted implicitly after precautionary examinations of Mr Zapruder and his camera (some paranoids argued that the footage might have been "filmed" in advance with actors!) - encouraged Warren into a meticulous recreation of time. But it also, helplessly, introduced doubt. The shots came quickly: according to Warren, there were three in under six seconds. Could the gun in question perform that well? Could Oswald have handled it?
But these doubts were minor compared with the observation made by everyone who has ever seen the film that the decisive shot, the killing shot, the one that takes away a piece of the president's head, seems to come from the side and ahead of him - from the general location of what is still known as the grassy knoll. Why? Because the president's head jerks backwards, as if thrown by the force of the shot.
Common sense says there was a second shooter, on the knoll. But then the experts, the anatomists, said, no, a body struck from behind can easily jerk the "wrong" way, towards the direction of the shot.
I should add that the Stokes report found a sound recording which, when analysed by acoustical experts, decided that there had been two shooters. The case is closed by history, yet open in legend. At great cost to America's stability, I suggest. But much of the doubt comes from the ways in which film records what it sees without necessarily revealing the truth. All that film does is alert us to the great test in looking, and seeing.Reuse content