[hist-analytic] Analytic Philosophy of History

Danny Frederick danny.frederick at btinternet.com
Fri Sep 17 15:19:19 EDT 2010

Hi Steve,

I don't pretend to understand what you are up to, but I will offer a few
comments on Popper and singular causation.

For Popper, there is no real connection between causation as ordinarily
understood and historical explanation (or, indeed, social scientific
explanation in general). As you know, Popper advocated the
hypothetico-deductive method in natural science, where we have:

a law of nature,

the initial conditions (= the cause), and

the predicted event (= the effect). 

But in the historical and social sciences we have human agents. And human
agents have free will (on Popper's view). But Popper also advocates the
unity of science, so he wants to apply the hypothetico-deductive method of
explanation in the social and historical sciences too. He achieves this by
proposing that, in place of a natural law, we have 'the rationality
principle,' which says that the agent acts appropriately to his situation
(as he sees it). The rationality principle is not a law of nature; in fact
it is false, because people sometimes act irrationally. But, Popper says, we
must assume the rationality principle in social scientific and historical
explanation, otherwise our hypotheses about the initial conditions will not
be testable/falsifiable. For Popper's most detailed exposition of this stuff
see his 'Models, Instruments and Truth' in his book 'The Myth of the

Methodological individualism is Popper's stand against collectivist
conceptions of society as an agent. Historical and social explanation should
be in terms of the actions of individual human agents in social situations,
these situations including the agents' aims and their understanding of the
situation (and also including social institutions of which the agent takes
account). Popper does not make much use of the TERM 'methodological
individualism' (excuse me not checking, but it might be that he never uses
it). I think the term comes from Weber, and it was made much of by Ludwig
von Mises and the Austrian economists, who had a big influence on Popper. It
probably got strongly associated with Popper because one of his closest
students and colleagues, John Watkins, was a vigorous advocate of it (though
he later relented somewhat in a paper on Hobbes).

I am not aware that Popper ever discussed the possibility of singular
causation. I think he just assumed that causation was bound up with natural
laws. When he developed his propensity metaphysic, he generalised the notion
of causation to include one event making another more probable (other
philosophers have done this too, notably, Reichenbach). But the generalised
notion of causation was now linked to probabilistic laws. So still no
singular causation.

I do not think that Popper's neglect of singular causation was simply an
oversight. I think he would have objected that such a notion was
scientifically useless because we could not test statements of singular
causation (since tests require repeatable observations). I guess he would
have been happy to allow singular causation as a metaphysical possibility
(he was not a positivist, after all); but he would not have taken such an,
in principle, untestable notion at all seriously. I write this paragraph
off-the-top-of-my-head: if anyone knows better, please say so.

Best wishes,


More information about the hist-analytic mailing list