In an interview (2012), discussing his own methodology, Latour affirms:
“this a thing that I learned from the “scientists” I studied, i.e. that: “Anything goes as long as it leads to what you want to find”. Just as in a laboratory you have instruments of all sorts, including the most archaic and the most contemporary, because that is what is necessary in production, I myself have learnt a lot from “true” scientists, hard scientists, i.e. total indifference to questions of method” (page 123, my translation).
This is very exactly what Feyerabend proposes in AGAINST METHOD (already in the essay version published in 1971). Latour continues in the same vein for two and a half pages, applying this notion of epistemological anarchism to his own work, and concludes: “So, how do you produce objects that resist what is said of them?, well, anything goes” (126).
As in the case of Feyerabend’s philosophy, epistemological anarchism is explicitly tied to a thesis of realism: the goal is to produce objects that “resist what is said of them”. This anarchism is “epistemological” precisely because Feyerabend wants to get away from the need to posit a dogmatic “anarchist” method that has to be applied in every case. He explains that he did propose such a dogmatic anarchism for science in the early 60s, but then the encounter with the needs of the practicing scientist, and later the encounter with the needs of the more diverse population of students that were enrolled after more democratic education policies were adopted in the US (at the end of the 60s), led him to reject even the most open set of rules as long as they were supposed to be applied dogmatically as fixed, universal, and binding principles, instead of heuristically as variable, local, suggestions or rules of thumb. This is similar to Latour’s intellectual evolution from the methodological anarchism that he espouses above to his later concerns with composing democractic assemblages.
This evolution from methodological anarchism to democratic pluralism parallels the progression in Feyerabend’s work from his Popperian methodological anarchism applied inside the pre-constituted and pre-demarcated sciences to his epistemological anarchism (circa 1966) which puts that demarcation totally up for grabs and argues for all sorts of transversal composites as necessary for what we commonly think of as scientific progress, and on to his democratic relativism as diplomatic guard rail for a pluralist free society.
Latour tells us that this epistemological anarchist phase of his work, that can be seen in his actor-network phase were in danger of entrenching a hyper-reductionist approach. He argues that the actor-network analysis always reduces everything to the same sort of explanation in terms of networks, and needs to be supplemented and pluralised by his new theory of modes of existence. Feyerabend too fell into that sort of undifferentiated theorising in the early 60s, expounding a sort of radicalised Popperian universal pluralist methodology covering art, science, religion, myth etc. But he broke with that at the end of the 60s, thanks to his Machian, Hegelian, and Wittgensteinian inheritance.
Feyerabend spent much time analysing the typological distinctions between different sorts of traditions and cosmologies. For example, in his last (unfinished) book CONQUEST OF ABUNDANCE he distinguishes between Homeric, Judaic, and Rationalist traditions, and inside science itself between Einsteinian and Bohrian traditions. His typology is different from Latour’s but he is emphatic that we do need a typology to distinguish between different modes of existence, and between different forms of knowledge.