Thanks for your comments Blair and Pieter.
Blair, I think your frustration with Feyerabend is really interesting because it speaks to a larger debate about the relationship between science and philosophy of science. Many (if not most) scientists practice science without much thought to philosophy of science, or without much meta-examination of what it is they do and how they see the world as a distinct social group. They simply work with the theories, strategies and problems of their day (and within the institutional constraints of the day too). Richard Feynman takes the extreme (if facetious) view:
“Philosophy of science is about as useful to scientists as ornithology is to birds.”
On the one hand I agree with this sentiment in the sense that, though philosophy of science can be useful to scientists, the two disciplines actually have quite different goals. The goal of science is the pursuit of scientific truths (e.g., ‘laws of nature’), while the goal of philosophy of science is to make compossible or consistent the truths that science produces with other regimes of discourse (e.g., history, logic, politics). On the other hand, I think that when it comes to the big conflicts/advances in scientific understanding, philosophical questions are usually there bubbling up from under the surface in the discourse.
Theory choice is a good example to illustrate how these questions come into play (what counts as evidence would be another). While scientists may abstractly agree that there is no ultimate theory of the scientific method (yet), when two competing theories appear to be equally supported by evidence, those debates about method suddenly become more important. You mention usefulness, for instance, as a solid criteria for theory choice. But should we really determine the truth of a theory based only on usefulness? And what exactly is the relationship between usefulness and truth? Most scientists would likely say well, usefulness is just one among a few or several key criteria. But this mix of criteria is hardly objective. Feyerabend’s goal is, in my view, to attack the view of philosophers of science like/especially Popper, who insisted that there is an objective method or set of practices which makes science distinct from everything else (pseudoscience, superstition, ideology, tradition, etc.). I’m not sure if Feyerabend has convinced me yet that “anything goes,” but I agree that there are also serious problems (for philosophy of science, not necessarily for science) with the notion that the practice of science follows a consistently objective logic.
One further comment: The way I’m trying to understand this phrase “anything goes” is that scientific truths, by definition, come into the world as an entirely new thing – they can’t help but break with the body of knowledge that came before it. Thus, the logic of the pursuit of scientific truth must necessarily be one whose rules can only come to be ‘known’ (i.e., philosophized) as that truth is integrated into (i.e., transforms) the body of scientific knowledge to which it is relevant. In other words, the possibility of the production of scientific truth is founded on the literal void in our knowledge of the world, and our knowledge of any consistent method of science is thus irreducibly historical (backward-looking). I think recognizing this kind of radical ignorance that makes doing science possible leads to, or at least implies the kind of theoretical anarchism that Feyerabend is articulating. This perspective admittedly comes more from my reading in Alain Badiou than it does my understanding of Feyerabend, so as I get further into the book my understanding of his meaning of the phrase may change.