Blau, A., 2011. Uncertainty and the History of Ideas. History and Theory, 50(3), pp.358-372.
Adrian’s interest in this article is to improve the way in which intellectual historians manage the uncertainties associated with their research:
Adrian notes two broad types of empirical phenomena faced by intellectual historians: actions and belief. Actions refer to ‘what happened’ (did this guy do X?). Beliefs refer to why it happened (why would he XYZ?).
Why did Plato write The Republic? When did Machiavelli write the Discourses? What did Jefferson mean by “virtue”?
It’s dubious whether we can know what was going on in the head of some crazy genius hundreds of years ago. [I added the word ‘crazy’, of course. If today’s intellectual are a bit awkward already, just try and imagine those daring to think when that could have gotten you killed.]
As such, Adrian believes that the idea of ‘proof’ (or ‘disproof’) in intellectual history is a bit of a misnomer. Any explanation is inherently uncertain.
What’s the best next step to take after you accept uncertainty? Adrian’s answer: marry it! (I swear – more or less. OK, not in those words).
To offer a framework Adrian grounds himself on King, Keohane, and Verba’s (p. 9 & 185):1
A researcher who does not report uncertainty “is either asserting that he or she knows everything perfectly or that he or she has no idea how certain or uncertain the results are.” We must thus give “honest statements of the degree of uncertainty entailed in each conclusion.”
That said, he thinks reporting uncertainty is easier said than done due to challenges at 3 levels.
Level ‘numero UNO’ / Objective versus subjective uncertainty
There are ‘objective’ sources of uncertainty for which estimation is possible, as it the example of an error coefficient in a quantitative analysis. There are, however, subjective sources of uncertainty that are difficult to weigh, as are qualitative judgments (why did the chicken cross the road?).
Regardless, Adrian believes there’s room for improvement if authors focus on “not saying what happened, but how strong they think their evidence is”.
Level ‘numero DOS’ / Reporting uncertainty
There is no universal language to communicate uncertainty (e.g. what does ‘I’m quite certain that…’ mean?) That noted, four type of considerations are possible.
- Order/Magnitude: Some statements indicate more certainty than others. He gives the example of ‘almost certain’ being stronger than ‘very likely’.
- Completeness: Instead of choosing an explanation (even if acknowledging others), intellectual historians could discuss “how likely each interpretation is”.
- Don’t just dismiss: Always propose an alternative when rejecting an explanation [although to be frank, I did not really understand Adrian’s rationale here].
- Forget quantifications: “Subjective estimates of uncertainty are themselves so uncertain that there is no point seeking the precision of statistic[s]”.
Level ‘numero TRES’ / Actually weighing uncertainty
There’s still the problem of knowing how much uncertainty there is in a given situation. What if you think something is fairly certain? Maybe I think it’s totally not likely! Adrian’s solution to this dilemma is to move focus away from the exact level of uncertainty an author thinks there is to the reasons why said author thinks so.
After this, Adrian enters justifications and examples of best practice that are very specific to intellectual history. That’s the point of the article after all. Since this is not a summary but a commented reading of what I thought/learned whilst reading the article, I’m going to take the liberty of not commenting these examples (this post is getting too long anyway).
This article was written with a very specific audience in mind. However, it could speak to most empirical approaches.
In a way, it is true that what Adrian calls ‘subjective’ uncertainty is perhaps most challenging for intellectual historians. I mean, their cases of analysis are, almost by definition, at least a couple meters below ground. However, this seems to also be the case of any qualitative judgment in any empirical situation.
Some of us hardly understand why we do half the stuff we do. We struggle to understand ourselves, we change our minds, we say things that come out exactly in the opposite way of how we want it. What Adrian calls ‘subjective’ uncertainty seems inherent to the very fabric of human existence. I cannot think of a way a [meaningful] empirical study could properly control for that.
Should we then live by ‘Adrian’s rules’ in all fields of empirical research?