Weberman returns to Danto to advance his argument. He notes that Danto shies away from saying that past events change, and suggests that he did this in order to remain a realist about past events; Danto held that “past events are what they are in a mind-independent way.” He insists that his views on the nonfixity of the past are not in conflict with this realism.

Danto, he says, works with a dualism between the past-in-itself and the past as “the changing narrative organization of the past.” To protect the past-in-itself from the interests of historical description, Danto “cordons off narrative organization and all relational properties of past events from the past itself.” Weberman answers this by distinguishing between the interest-relative description and the relational properties that adhere to the events:

“while it is true that which relational properties we attend to and thus how we narratively organize the past may be interest-relative, whether a past event has a given relational property is not interest-relative. In other words, it is true that whether we describe a past event in light of this later event or that later event is up to us. Historians tell different stories about past events according to which later events they find to be most significantly tied to the earlier events. This does not mean that whether a given narrative sentence is true is something subjective. The fact that we are the ones who describe the earlier shooting in light of the later death by calling it a killing does not make the killing subjective or relative to the knower. Once we see this, it should be clear that the properties of past events that emerge only as a result of later events are ontological, not epistemic. Realism about past events is not incompatible with recognition that the past is not fixed in virtue of its delayed relational properties.”

Weberman then addresses the question whether the changes in the thick past he describes are Cambridge changes or “real” changes. Why would someone doubt that changes of relational properties are “bogus” (Peter Geach’s word). Weberman notes that there is a widespread intuition that changes in intrinsic properties are real while relational changes are not, allied to the notion that “only changes that are causal effects are real.” Finally, if relational changes are real, then every change in anything is a change in everything, since everything is related to everything else; this seems absurd.

Yet, there are examples of relational change that appear to be real. Becoming a husband or wife is relational, but we regard that as a real change. A dollar that lies in a drawer until the Treasury withdraws the bill from circulation has not changed at all in its intrinsic properties, but the change in the dollar’s use and value is much greater than a tear in the corner (an intrinsic change). Weberman argues that ” Social phenomena in general exemplify how relational changes and relational properties are constitutive of the core properties of a thing or event since social phenomena fundamentally depend for their very existence on people’s beliefs about and behavior toward those phenomena.” Weberman concludes not that all relational properties and changes have ontological weight, but that some do. Whether we think the dollar bill in the drawer has changed depends on what properties we are attending to.

History is another area in which relational changes cannot be siphoned out, and this because events are “historically what they are only in virtue of their being part of narrative wholes that encompass other events.” Which later historical events change earlier ones then? What relationships among historical events are relevant? Weberman argues that there must be a chain of causes and effects connecting the earlier event to the later event. The birth of Diderot is one of the causes of the writing of Rameau’s Nephew , and Amir’s shooting of the gun is a cause of the death of Rabin.

Causation is a necessary, but not sufficient, criterion. The further requirement is “the relevance or the significance of the later event to the earlier event to which it is causally connected.” Someone reading the account of Rabin’s assassination in the paper is a later event, causally connected with the shooting, but it doesn’t “bear on what kind of event the assassination was.”

Significance depends, he admits, on “contingent human customs, conceptual schemes and interests.” This does partly turn Weberman’s thesis back into an epistemological one, but not entirely: ” Whether a relational property is genuine rather than bogus or negligible does depend on our concepts and our interests. But once our conceptual schemes and interests are there and a given relational property is counted as genuine, then whether or not that property has changed is not a matter of description but of the properties of the event itself. What we attend to and thus what we take to be genuine is a matter of our ways of thinking. But once attended to, properties and changes therein are not solely epistemic. ” It is social custom that makes “I do” in the setting of a wedding an event that changes relational properties of a man and woman; in another culture, those words in that kind of setting may mean nothing. But, given our customs, the relational property established by the I do is real.

The consequences of treating all relational properties as bogus or less than real, is disastrous, especially to historical investigation:” In the end, it is crucial to see that denying, à la Geach, the reality of all relational properties and changes would be tantamount to dismantling the ontological infrastructure of various aspects and regions of reality, including the entire infrastructure of historical reality. This is because history is essentially about events characterized in light of later events and thus as parts of greater wholes. To rule out relational properties and changes as unreal would leave us with a type of history that is no more than an atomistic or pointillistic chronicle of events devoid of all reference to eras, movements, revolutions, wars, and even killings where there is a temporal interval between the shooting and the death of the victim. To limit ourselves to such a chronicle would be to acquiesce in the unnecessary exclusion of structures and aspects of the way the world is and was.”