In the journal Futures and Foresight Science 2(3-4) there have been interesting discussions about the relationship between historiography and futures research.
The following passage from Martin Kunc’s “A modeler’s perspective: A commentary on Schoemaker 2020” was especially interesting:
“Schoemaker states, ‘The word forecasting captures this very notion, suggesting that the momentum of the past casts itself forward, with action and reaction producing repetitive cycles.’ We need to acknowledge that any forecast depends on the structural conditions of the past which repeat themselves in the future. This situation puts the observer in an interesting position because it has to recognize the existing past structure and be able to extrapolate it into the future through the forecasting model. Considering the biases that observers have, especially in terms of the choice of variables that they want to extrapolate, it is clear the reading of the past is subject to biases in the same way as the choice of variables in a forecasting model. Somehow, the belief that ‘the future is partly predictable because it is causally connected to the fabrics of the past and present’ may be an illusion from the observer in the present, as Keynes said ‘ideas shape the course of history’ or Twain suggested ‘the very ink with which history is written is merely fluid prejudice.’ Similarly, the accuracy of any forecasting model depends on the variables that are included in it, as well as the repeatability of the forecast since forecast models learn from their own mistakes to adjust their accuracy. However, I have not seen many historical books corrected based on lessons from similar events that happen later.” (2020, emphasis added).
There are some comments I would like to make on this.
First, some people (I certainly do not belong to this group) have argued that historiography explains evidence and all there is to historiography is to be able to account for evidence (see Virmajoki 2020, section 6). If taken literally, this means that a historiographical account is acceptable if it explains the evidence we have gathered thus far. Now, these accounts may turn out to be unacceptable because we later find further evidence that contradicts an account that was accepted before. In a sense, in this case the historical account is corrected on the basis of what happens later: It is corrected on the basis of a later event (the event being the finding of new evidence). Accordingly, historiography is a practice that accounts for evidence and attempts to predict what type of evidence will be found; and in this sense it would be essentially similar to forecast modelling.
The obvious objection is that the new evidence concerns the same – not merely similar – event in the past and therefore the case does not contradict Kunc’s observation. However, if one assumes that historiography only explains evidence and that evidence is the only object of study of historiography, the notion of “the same event in the past” does not make sense. Or at least it cannot be understood independently of the historical accounts that have been accepted.
As mentioned, I do not find this line of argumentation convincing because I think it is obvious that historical inquiry studies past events through sources, not sources themselves. However, we have to be aware of these types of philosophical acrobatics.
We can now turn to more natural interpretation where the relevant events are events in the past. The claim is that our historiographical interpretation of an event E1 rarely changes even if learn that a similar event E2 happened differently later.
I do not think this is true. First, it seems obvious that our substantial views concerning history have changed and therefore our interpretation of individual events. For example, we no longer believe in general historical progress, and the usual suspects for this are the horrors of the 20th century. It is no longer possible to view historical events as steps in some simple progressive schema. A similar example can be found in the historiography of science. The development of science has turned out to be somewhat messy. It is no longer possible to understand the development of science in terms of a simple schema like “the growth of knowledge” (see discussion here).
Secondly, other events may affect the historiographical account of an event indirectly. Every historical account and interpretation is based on some principles and ideas concerning the causal and structural aspects of the world. It might be difficult to test these principles directly (due to the problems discussed in a previous post) but it seems that the historical research as a whole can correct the principles and ideas. For example, Robert Merton argued, on the basis of multiple independent discoveries, that individual geniuses do not determine the development of science. In Merton’s analysis, many historical events are studied in order to show what did not, in fact, guide some individual event.
It seems obvious that historical interpretation of an event might change on the basis of what we learn about other events surrounding it. Of course, historical events can be understood only by contextualizing them.
However, the question of the significance of the later events is still open. Should we assume that the principles and ideas concerning the causal and structural aspects of the world that guide a historical interpretation remain invariant through time?
I do not think that we can guarantee the invariance a priori. It may be that history is a mess that cannot be structured under temporally invariant principles. However, I do think that a search for some invariances and generalization is necessary for healthy historiography. I have argued (here) that many important explanatory practices in historiography become impossible without generalizations and general considerations. Moreover, I do not think that it is possible to make the distinction between synchronic context and diachronic context. Synchronic context means that we can understand an event by comparing and contrasting it with other events that happened at the time. Diachronic context means that we understand the event by comparing and contrasting it to other events that happened at another time.
The argument: Assume that synchronic context is preferred. Take events E1 and E2 that happen in temporal proximity (i.e. share a synchronic context). Either E1 and E2 are causally connected or not. If they are not causally connected, understanding E1 through E2 requires that we assume that they are guided by similar principles. This assumption is based on either (i) intellectual economy, or (ii) the assumption (PA) that events that happen in close historical proximity are guided by similar principles. If the assumption is based on intellectual economy, then we can understand E1 through E2 even when they share a diachronic context. After all, the intellectual economy of historiographical principles can only be judged against the totality of our historical knowledge. On the other hand, it is difficult to accept PA without some substantial ontological claim about the history structuring itself into different eras that are guided by some principles that are expressed in that era. Such substantial claim is in tension with the whole point of historical contextualizing and cannot be accepted. If E1 and E2 are causally connected, understanding E1 through E2 becomes possible by understanding how E1 depended on E2. However, if causal understanding is all there is to the synchronic contextualization, then we can understand events equally well in their diachronic context, as long as the diachronic story is a causal one.
As we cannot prefer synchronic contexts to diachronic ones, later events must have the force to correct our historical accounts of some earlier events. If we drive for intellectual economy, knowledge of later events can change what principles we accept in historiography. These principles guide our interpretations of the historical events and therefore knowledge of later events can correct our interpretation. On the other hand, if our understanding is based on causal relationships then later events surely can shape our understanding about earlier events: only certain types of events can have certain types of consequences.
Given that healthy historiography must correct its account on the basis of what happens later, we can conclude that either (a) historiography is not healthy, (b) historiographical corrections happen unnoticed, or (c) Kunc’s claim is a bit of exaggeration.