When Is One Explanation Better than Another in Historiography?

It is often the case that, in historiography, we have two explanations for some outcome, and we wish to know when one is better than the other. We assume here that both explanations are true in the sense that they explain.

Consider the following explanations:

E1: Eddington observed gravitational deflection in 1919 because Einstein formulated the theory of relativity. Had there not been the theory, the observation would not have been made.

E2: Eddington observed gravitational deflection in 1919 because Einstein survived stomach problems in WW1. Had he not survived, the observation would not have been made.

These both explanations are true, but it seems that E1 is more interesting and deeper. How should this notion of explanatory betterness be understood?

First, there is a straightforward reading of the betterness: One explanation is better when it answers the kind of question we are interested in. An explanation is better when it satisfies an explanatory goal. The explanatory goal is what the explanation seeks to establish in addition to providing information about conditions on which the explanandum depends. The notion of explanatory goal is based on the idea that historians want to highlight certain aspects of the process that led to the explanandum.  An explanation may seek to pinpoint the role of some types of factors (F) in history and this attempt is centered around the (meta)explanatory goal of understanding the role of F in historical processes. In this sense, the choice of the explanatory goal aims at second-order understanding (whereas an answer to a particular explanandum gives us first-order understanding).

One can say that E1 is most often deemed more interesting because we are interested in understanding the role of scientific theories in scientific practice and empirical testing. E1 enables us to satisfy an explanatory goal. However, in some contexts. E2 could be more interesting. If one wishes to understand how embodied scientific ideas are in particular persons, E2 is a nice explanation to be given.

Secondly, the betterness of explanation can refer to explanatory depth. Whereas the ability to answer a type of question that we are interested in is an external property of an explanation, the explanatory depth is an internal property of an explanation. (Although the separation is not all that clear, as I point out below). One common way to define explanatory depth is to define it in terms of the ability to answer what-if-things-were-different questions (Woodward and Hitchcock 2003). One explanation is better (deeper) when it answers more what-if questions. For example, we can say that

(I) E is a deeper explanation than E* if E gives information about counterfactual alternatives to the explanans Z that E* is silent about.

For example, if E says

Had Einstein not formulated the theory OR had he provided a different formulation, the observation of the gravitational deflection would not have been made.

and E* says

Had Einstein not formulated the theory, the observation of the gravitational deflection would not have been made.

then E is a deeper explanation than E* since it answers a what-if question about Einstein providing a different formulation. We can notice that an early version of Einstein’s theory, in fact, included a mistake that led to a wrong prediction about the deflection. If this mistake had not been corrected, the observation would perhaps not have been made in 1919.

Notice that the number of what-if questions is not the only dimension that is relevant in such cases. Some what-if questions are more important than others. Here the notion of explanatory depth becomes dependent on issues external to the structure of the explanation. For example, in some context, it might be more important to know what would have happened if Einstein provided a different formulation of the theory than to know what would have happened if no theory was provided. If we wanted to know the influence of personal prestige, this question would be interesting. Would Eddington still have concluded that the observation was in accordance with Einstein’s theory?

Another dimension could be the following:

(II) E is deeper that E* if E would be true in background conditions B while E* would not.

Here E answers questions about what would happen in B while E does not. For example, take the following:

E1: Eddington observed gravitational deflection in 1919 because Einstein formulated the theory of relativity. Had there not been the theory, the observation would not have been made.

E2: Eddington observed gravitational deflection in 1919 because Einstein survived stomach problems in WW1. Had he not survived, the observation would not have been made.

E2 is true because Eddington was motivated by uniting the international scientific community. If Einstein, a pacifist from the Central Powers had died during the war, the observation would have been less likely to have the impact Eddington thought it would have. The war created a situation where Einstein’s survival was crucial for the testing. Had there not been a war, there would have been fewer obstacles and less reluctance to give attention to Einstein’s theory. This means that, in the absence of war, the theory would have likely been tested even if Einstein died because not so much would have depended on Eddington’s plan. Changing the background condition so that there is no war makes E2 untrue. However, E1 would still be true if there was no war.

We can see that this is one of the dimensions that explain why E1 seems better than E2.

Just to point out one more example of how many dimensions there can be, another dimension is the following:

(III) E is deeper than E* if E provides information about a factor F that is not provided by E*. In other words, E is a more complete explanation than E* as it makes explicit a factor that is only a background condition from the perspective of E*.

For example:

(A) The observation was made in 1919 because Einstein formulated the theory and because Eddington executed the expedition. Had there not been such theory or expedition, the observation would not have been made in 1919.

(B) The observation was made in 1919 because Einstein formulated the theory and because Eddington executed the expedition. Had someone else formulated the theory or had someone else executed the expedition, the observation would not have been made.

(A) and (B) together

Had there been no theoretical work or no expedition OR had the work or the expedition been conducted by someone else than Einstein and Eddington, the observation would not have been made

provide a deeper explanation than either of them alone.

According to Hitchcock and Woodward “[this] is, perhaps, the most fundamental way in which one [collection of counterfactuals] can provide a deeper explanation than another.” (Hitchcock and Woodward 2003, 188).

We can see that explanatory betterness in historiography is interesting notion and progress can be made in explicating the notion.

Leave a Reply

Your email address will not be published. Required fields are marked *