Abstract
An important insight of scholarship on epistemic iteration is that scientists have to start from somewhere, but without sufficient justification to determine which ‘somewhere’. The engine of epistemic iteration is supposed to help us see how the epistemology of science gets along anyway. By adopting some working assumptions to get the whole process up and going, scientists afford themselves material to work with and to improve upon. The applicability of the adopted view can be tested until it breaks down and invites refinement or replacement. Through many iterations, the content of scientific theory thus earns its epistemic credentials. Indeed, thinking about epistemic iteration reveals the important roles of anomalies for epistemic progress. It is by probing the places where a theory seems to crack that scientists can often make headway. I will compare two cases in which scientists adopt a framework to work within, and then probe the failures of that framework in the hope of making epistemic progress. The first case involves astrophysical modeling of core collapse supernovae, and serves as a relatively clean example of how probing failures can lead to epistemic advances. Astrophysicists initially adopted one-dimensional (that is, spherically symmetric) computational models of supernovae, largely due to the limitations of computing power available at the time. These model supernovae would not evolve explosions—they were consistent duds. Thanks to some key empirical evidence demonstrating the asymmetry of debris (isotope titanium-44) ejected from a supernova (1987A), modelers realized that asymmetry (and thus, models in more than one dimension) is needed for successful explosions. This insight led to better and more accurate models of core-collapse supernovae. The second case concerns the abundance of primordial lithium in our universe—the so-called ‘the lithium problem’—and is less clean. Less clean in part because it is ongoing, but for that very reason a fertile testing ground for philosophy of science. The ‘lithium problem’ is the discrepancy between the abundance of lithium predicted by modeling big bang nucleosynthesis and that implied by empirical evidence. It arises from adopting the standard model of cosmology together with nuclear physics, and has been noted for decades. The discrepancy is not subtle—it amounts to something like a 4-5 sigma difference. Interestingly, the approach that many scientists have taken in attempt to resolve this discrepancy has been to connect it up to other outstanding problems in cosmology and physics including the Hubble parameter discrepancy, the nature of dark matter, and open questions about the lifetime of the neutron. Drawing on these two cases, I discuss the connections between epistemic iteration and scientific research strategies such as gathering new empirical evidence and attempting to unify anomalies.
