Perhaps more than any other of the physical sciences, cosmology exemplifies the inevitable contact between science and philosophy, including the problem of the demarcation criteria that distinguish science from non-science. Although modern physical cosmology is undoubtedly scientific, it is not obvious why it has this status, and nor is it obvious that all branches of theoretical cosmology satisfy ordinarily assumed criteria for science. While testability is generally admitted as an indispensable criterion for a theory being scientific, there is no agreement among cosmologists what testability means, more precisely. For example, should testability be taken to imply falsifiability in the sense of Popper? I discuss this and related questions by referring to two episodes of controversy in the history of modern cosmology, the debate over the steady state theory in the 1950s and the recent debate concerned with the anthropic multiverse. In addition, I draw attention to the use of historical analogies in cosmological and other scientific arguments, suggesting that such use is often misuse or otherwise based on distortions of the history of science.
Read more: arxiv.org/pdf