Towards a Popper-ized History

I’m reading (a book about) Karl Popper right now. It has lots of long quotes by him, though, and is as far as I can tell a pretty good summation of his philosophy (he seems to have been involved in its publication), so I’m saying I’m reading him. It’s the little lies I tell myself that hurt the most.

Anyway, for those who don’t know Popper, his main idea—or at least the main one that I am grappling with—is a reversal of the scientific method. The old Baconian model was to generalize from observed data and try to reproduce your results to prove your hypothesis correct through verification. Popper saw that no number of positive observations could ever prove a theory, though, because of the problems with induction that have been evident since at least Hume. If no number of positive results can verify a theory, however, one contradictory observation will show that the theory was wrong.* This allows us to refine our theories and, by trying (and succeeding) in refuting them, come up with more and more informative theories.

(One thing I’m not totally clear about is how you properly refine your theories. One of the critiques Popper developed about Marxism was that the Marxists will either change the theory or the evidence to make sure that the theory is always correct; this makes the theory unfalsifiable and therefore just dogma. While I understand that changing the evidence is bad, changing the theory seems to be a pretty good response to evidence that contradicts it. Isn’t that what falsification is all about?) 

Anyway, that aside aside, I’m wondering whether there is a way to incorporate Popper’s theories into a historical, rather than merely scientific, method.** Now, this requires us to think a bit about what history is, and whether we can consider it a science.

History can be seen as one of the humanities—in an extreme case, no more than the study of literature from times gone past. This is valuable, and in my essay about why we should study history I alluded to a model of historical research that sought to expand our understanding of ancient people in this way. This can branch into cultural and intellectual questions.

Generally, however, I prefer to see history as something of a (social) science, one in which we can put forward hypotheses about what happened in the past and why, which can then be compared with the evidence we have. Examples: trade in bulk goods presupposes a luxury market and low risk of predation (maybe true) or nomadic horse-riding tribes are unable to administer the territories they conquer (probably false). One might even hope that people who make things happen in our world today take some heed of what we, as historians, discover.***

While it may seem that this latter model of asking testable historical questions primarily works for what I’ll calls old school political or military history, it can also work for cultural and intellectual history. The problem is then how to pose questions so that they are falsifiable. “Late antique people refused to incorporate magical traditions outside of their personal religious tradition” seem to be an eminently falsifiable proposition, as a quick look at a few amulets and spells will show.

The broader question, and one for which I do not have an answer, is whether these are helpful questions to ask. There are arguments on either side.  The argument against these being useful questions—or more precisely, whether Popper’s method is a useful way to formulate questions—is that Popper’s methodology is designed to help us discover a system that can be profitably described with laws. History is not in this sense a science. It deals with people, who are infinitely variable, and life, which is messy and doesn’t hew to the broad causalities that, e.g., the Marxists Popper broke with would have endorsed.

The argument for a Popperized history, which I think I support (but I’m still working this out) is that it helps us be clear about how we use evidence. It requires us to frame questions in such a way that they could be falsifiable. If you make a broad claim that cannot be falsified, there isn’t much information we gain about the past from it. Take the claim that “late antique people did X.” That is pretty much unfalsifiable: short of a time machine and a panopticon, how could we ever determine that no one ever did X? And, moreover, do we really care?

However, if we change our approach so that we put forward falsifiable theses about the way people acted in the past, the impetus then becomes to find the counterexample that will falsify the claim. If we posit that “late antique people did not do X” or “ties of dependence in the pre-modern world were based exclusively on Y,” we are put in a position where one counterexample can lead us to the conclusion that “late antique practices in regard to X were varied” and “ties of dependence in the pre-modern world were based on a number of factors.”

Now, neither of the conclusions reached above is really that groundbreaking, and that’s why I’m not totally sold on applying this method to history. But it seems to me that by putting forward many theses like these and refining them, we can actually increase our knowledge of the world, rather than merely asserting claims that may or may not be true. And that seems like a way to fight back against the sense that we can’t ever really learn anything about the past.



*The classic example is swans: a million white swans don’t mean that all swans are white, but one black swan means that not all swans are white.

**I should note that this train of thought has nothing to do with Popper’s views about historicism, or the open society, or any of his more political projects. This is about applying the methodology he brought to the sciences to history.

***Free lesson to any big machers reading this blog: stay out of Afghanistan. Not worth it, even if you think you’ll get to marry Roxane.


© Henry Gruber 2013