According to the philosophical position of epistemological holism, the statements of the empirical sciences do not relate to observations singly, but collectively. This is because these statements belong to logically complex theoretical structures, which are to a large extent indivisible, and also because further theoretical assumptions (an “observational theory”) underlie the observations made to check them empirically. As a consequence of this basic claim, all brands of epistemological holism include an underdetermination thesis, to the effect that the scientists’ decisions about hypotheses are underdetermined by the evidence available to them, and in particular, by the results of the tests they perform. Pragmatic reasons must eventually prevail in the choice of attributing the evidence to this or that part of the theoretical whole, and when elaborating on these reasons, philosophers of science will never offer more than partial and context-dependent guidelines.