Update: I cannot agree with an angelic view of science that places it on a pedestal above philosophy. When Feyerabend talks about the “way of the scientist” as opposed to the “way of the philosopher”, he is talking about a small number of philosophically-cultivated scientists freely exploring bold new speculative conjectures. The vast majority of scientists he considered to be “human ants”: “As opposed to its immediate predecessor, late 20th-century science has given up all philosophical pretensions and has become a powerful business that shapes the mentality of its practitioners. Good payment, good standing with the boss and the colleagues in their ‘unit’ are the chief aims of these human ants who excel in the solution of tiny problems but who can not make any sense of anything transcending their domain of competence. Humanitarian considerations are at a minimum and so is any form of progressiveness that goes beyond local improvements” (from “These on Anarchism”).
For Feyerabend there is no such thing as the “scientific image”, and common sense (or so-called “folklore”) is for him often much richer than the pale abstractions one tries to replace it with on the basis of a handful of one-sided cherry-picked ill-analysed results that one has taken from one or other of the scientific paradigms enjoying scholarly approval at the moment.
For Sellars too, there is no one Scientific Image. What he called the “scientific image” is the idealized convergence constructed out of the diverse particular scientific images proposed by the various science:
“Thus the conception of the scientific or postulational image is an idealization in the sense that it is a conception of an integration of a manifold of images, each of which is the application to man of a framework of concepts which have a certain autonomy. For each scientific theory is, from the standpoint of methodology, a structure which is built at a different ‘place’ and by different procedures within the intersubjectively accessible world of perceptible things. Thus ‘the’ scientific image is a construct from a number of images, each of which is supported by the manifest world”.
Thus for Sellars the scientific image is an idealized “construct”, out of the convergence or “integration” of diverse scientific images, taking place on the substantive level, making use of a unified “framework of concepts”, rather than on a meta-level methodology.That is, Sellars still presupposes a monistic principle of convergence. That the scientific image is substantive means that it is not a Feyerabend-style methodological notion prioritising pluralist heuristics and epistemological anarchism, i.e. it is not an image of thought in the Deleuzian sense. Rather, it is a substantive image in Heidegger’s sense of a world view, or a specific understanding of Being.
Feyerabend was a process-philosopher, although this is often not noticed. He drew his process thought from a combination of Hegel and Kierkegaard, whom he did not see as opposed on this point. He pluralised Hegel by means of Mach and Mill, and kept to this sense of processual pluralism that he found in the work of David Bohm, Niels Bohr, and Wolfgang Pauli. His idea was that there is not just one scientific image but many, and that this plurality is not just a sign of the uncompleted state of our research, but indicative of the nature of reality itself. Such a view makes the epistemology/ontology distinction largely irrelevant to understanding the heuristic nature of science, representing a superficial scholasticism that would shackle research in poorly analysed binary oppositions.
Ontological plurality goes deeper and wider than such naive logic chopping can take us. It goes deeper because there is no way that one can simply dogmatically posit that “Reality is one”. That too is an assertion that is to be answered by research. Research implies necessarily the exploration of alternative views, so the hypothesis that reality is multiple must not be ruled out of court in advance. It’s no use loudly proclaiming that you are in favour of heuristics and then proceeding to dogmatize, and to assert that Science is on your side.
Every ontology involves value judgements in terms of (1) the very types of concepts it mobilises, (2) the image of research it projects and conforms to (3) the communities that pursue and promote that ontology. A good example is precisely the difference between diachronic and synchronic ontologies. This is made very clear by Deleuze in his explication of why Spinoza’s fundamental ontological treatise was not called “Being” but “Ethics”. Feyerabend agrees and tries to protect us not only from arrogant conformist philosophers but from their equivalents in the sciences.