ABSTRACT
"To raise the level of cumulative knowledge in the scientific study of war," advised Stuart Bremer, we need "more comparative and competitive theory testing" (Bremer 1996: 1). While Bremer's contributions to the scientific study of war in the areas of testing national capabilities (Singer et al 1972; Bremer 1980), simulation (1977), probabilistic inference (1992), the dyad-year (ibid.), multivariate analysis (ibid.), democratic peace (1992, 1993a), the evolution of war (1993b, 1995), and data generation (Jones et al. 1996), are well known, what is perhaps less well known is his abiding interest in the philosophy of science. 1 One outcome of this interest was a growing concern over the utility of falsificationism: the prevailing view that we accede to the "truth" by successfully eliminating false theories. As Bremer was apt to point out, most philosophers of science no longer accept falsificationism as a valid means of accumulating knowledge (Laudan 1990). This is because causation cannot be observed; we observe instead the results of hypothesis tests. The construction of these tests requires the adoption of numerous auxiliary assumptions, particularly in the concretization of measures. This means that it is possible to "falsify" a theory when it is in fact true, simply because one assumption in the construction of the test is in error. Since we can never be certain that all of the auxiliary assumptions in the test itself are correct, we can never falsify a theory (or hypothesis) with certainty (see also Bremer et al. 2003: 8–9).
