Monday, April 17, 2006

On History, Causation and E J Tapp

In my last post on history I made a distinction between topic selection and the approach that should be adopted in dealing with the topic. Here I argued in part that the purpose of historical analysis is to test, not prove. When proving or justifying becomes the central point, the discipline is lost.

Upon reflection, while I still think that this is a valid point, it's also not the end of the story.

When writing, I do try to tell a story, to make things comprehensible. I also think that the role of the historian is an part that of custodian of the past of the country, tribe or group. How then, does this fit with the idea that the purpose of historical analysis is to test, not prove or justify?

Perhaps the most influential course that I have ever done measured by personal long term impact is Ted (EJ) Tapp's philosophy of history course at the University of New England.

Ted was a reflective man. His personal view was that without a concept of causation there could be no history, no way of knowing anything. He was therefore opposed to the concept of history as simply description. However, his view of causation was deeply rooted in philosophy, and he therefore introduced his students to a range of philosophical texts concerned with the concept of causation.

Central to his thinking was the difference between correlation (a and b) as compared to causation (if a then b). Drawing from the philosophy of science, he argued (at least as I saw it) that all historical theses were essentially refutable. You tried to establish your case from the evidence, to develop the case that best fitted. But you did so in the knowledge that later work might invalidate your position.

Now, and this drives to the heart of my point about method, whatever one's view about the role of the historian, all historians must write in such a way that the reader can understand both the evidence and the logic chain. That is, we must set up our arguments for later test by others.

No comments: