In an article in the Review of Metaphysics (1997), David Weberman argues for the “nonfixity of the historical past.” He starts from Arthur Danto’s argument that historical inquiry and writing cannot be a reconstruction available to an “ideal chronicler” who knows everything that happens the moment it happens and sets it down as it happens in just the way it happens. Historians can’t do this because they use “narrative sentences” like “The Thirty Years War began in 1618,” a sentence unavailable to the “ideal chronicler” because he didn’t know it was a 30-Years war when it began.
Weberman, though, wants to press the argument further. Danto’s argument is epistemological; our description of fixed events changes in the light of the outcome. Weberman’s argument is ontological: Not just our descriptions, but the events themselves change with the passage of time.
Weberman admits that what he calls the “skeletal past” is fixed. An event in the skeletal past could be described as “person s at place p raised his arm at time t.” Once that happened, it happened, and can’t be undone. But Weberman suggests it’s probably impossible to establish criteria to distinguish the skeletal past from what he calls the “thick past.” The fixed skeletal past is “a sort of artificial or unrealizable limit.”
By “thick past,” he means “any and all past events and past actions individuated and described in terms not restricted to basic physical states and movements.” “The Senate ratified the arms-limitation agreement” is the thick past (not, notice, just a thick description of a past event). He doesn’t claim that every thick event is unfixed, but only that “some thick events are - if not in perpetuity, for a finite duration of time.”
He analyzes the 1995 assassination of Yizchak Rabin by Yigal Amir as an example. At a certain time, Amir pointed a gun at Rabin and pulled the trigger so that the bullet hit Rabin. That’s close to a skeletal description. For the next several hours, the event was “Amir shot Rabin,” but when Rabin died a few hours later the original event changed to “Amir killed Rabin.” This is not simply a different description of the same event; had Rabin not died, the original event would have remained “Amir shot Rabin.” Nor does it seem right to say that Amir killed him between time A and time B, since during the intervening time he had been arrested. The killing did not last several hours, nor did the shooting. The shooting happened; after Rabin’s death the shooting became a killing.
Another example: “In 1713 the author of Rameau’s Nephew was born.” But for 56 years, Diderot did not write Rameau’s Nephew ; once he did, what happened on his birthday changed from “Diderot the author of other books was born” to “Diderot the author of Rameau’s Nephew was born.” It doesn’t make sense, Weberman points out, to say that the birth of the author of Rameau’s Nephew lasted 56 years. It’s much more sensible, despite its oddity, to say that what happened in 1713 changed as a result of a later event.
Aren’t these changes, though, simply “Cambridge changes,” not changes in the events themselves but only changes in the events in relation to other things? Weberman admits that these are “relational properties” of events, describing them as “delayed relational properties,” but insists that “such relational properties are indeed real properties and indispensable to any historical understanding of events.” I’ll summarize that argument, but in a different post. This one’s too long already.