Thoughts in a Haystack

A generation's worth of philosophers sought and failed to find a clean demarcation between science and non-science. For our purposes, it's better to think of this as the line between epistemically worthy and unworthy pursuits. No plausible necessary and sufficient conditions were ever found for being epistemically worthy. That doesn't mean there isn't a difference, however. There is a large difference between the modern synthesis in biology and creationism, between chemistry and homeopathy, and so on. The failure to articulate a sharp division means only that what we count as epistemically worthy is quite diverse and assessed along so many dimensions that it's hard to narrow the criteria down to something simply state-able. The marks are clear enough: empirical adequacy, simplicity, novel predictions, novel explanations, unification, consilience, and more. The metric by which we tolerate trade-offs among these virtues is less clear.

Now, as it happens, we have these concepts, 'metaphysics' and 'science'. There is no sharp difference between the two. To a rough approximation, we can think of metaphysical claims as more abstract and distantly related to experiment than scientific claims. Bear in mind that there is of course a lot of theory and metatheory in science, but at some point we start classifying the theory and metatheory 'metaphysics'.

-Craig Callender, "Philosophy of Science and Metaphysics"