Helge Kragh (Copenhagen) gives a talk at the Workshop on "Why trust a Theory?" (7-9 December, 2015) titled "Fundamental Theories and Epistemic Shifts: Can History of Science serve as a Guide?". Abstract: Epistemic standards and methodologies of science inevitably reflect the successes and failures of the past. In this sense, they are in part of a historical nature. Moreover, the commonly accepted methodological criteria have to some extent changed over time. Faced with the problem of theories that cannot be tested empirically, perhaps not even in principle, it may be useful to look back in time to situations of a somewhat similar kind. Roughly speaking, previous suggestions of non-empirical testing have not fared well through the long history of science. Ambitious and fundamental theories of this kind have generally been failures, some of them grander than others. So, is there any reason to believe that they will not remain so in the future? Can we infer from history that empirical testability is a sine qua non for what we know as science? Not quite, for it is far from obvious that older scientific theories can be meaningfully compared to modern string theory or multiverse physics. History of science is at best an ambiguous guide to present and future problems, yet it does provide reasons for scepticism with regard to current suggestions of drastic epistemic shifts which essentially amounts to a new “definition” of science.