I've been reading a lot of academic history recently. For a quantitatively oriented social scientist, it's an interesting experience.
- Historians tend to see a counterexample as a disproof. So, e.g., if Christopher Hill says that Puritanism was the ideology of the middling sort of people, that is disproved if there are Puritans among the upper classes.
I would assume that social scientific hypotheses are about probabilities, not absolute laws – because life and society are complicated. But there is some merit in not just treating the "error term", so to speak, as a nuisance, instead looking for its structure and trying to accommodate that.
- Sometimes they seem to think that the history is what people thought at the time. In fact, some books give the impression of literally not caring whether contemporaries' ideas were accurate or not.
When Robert Boyle says that reading "accustom'd his thoughts to such a Habitude of Raving, that he has scarce ever been their quiet Master since", my first thought is "Wow, is that true? I wonder what effects persistent reading did have on the mind, when the printing press was new on the scene ?" and to reach for Marshall McLuhan, Jack Goody or Walter Ong. But Adrian Johns just says that such experiences "were widely credited". Dammit, tell me if they were real! Is this some kind of hangover of postmodernism?
- We do hypotheses. They do narratives.
That's not exactly true. It's more that a hypothesis plays a different role. It is a peg that a narrative hangs off, and it's not there to be tested so much as to structure the narrative and be qualified by it.
This is probably in conformance with the nature of the historical process, which cannot really be understood by estimating conditional means. I'm still trying to understand the epistemological role narrative can play. Difficult stuff.