What we have here is a failure to communicate ;-) which is because your reading of Latour -- which I generally agree with -- links his scientific ontology of the social to a dualist/representational philosophical ontology that I think is neither what Latour is up to nor self-evidently what we social scientists/analysts ought to be engaged in. "If we live in an ontologically hybrid world we need modes of analysis suited to that world. Social constructivism or discourse analysis in the vein of Foucault, Laclau, Wittgenstein, etc. are well suited to particular areas of our world but they miss out entirely on vast swathes of it because of their blinkered and entirely unnecessary dualist predispositions..." I would disagree pretty fundamentally here, because like most philosophical mind-body dualists you're trying to "put ontology first" and make categorical claims about how we should study the world based on the character of the world -- a gesture that just side-steps the basic conundrum of how we might get to that character of the world without the conceptual equipment that you are suggesting follows from that character...leaving us with a profession of faith in the character of the world (in this case, that it is made up of "hybrid networks"). Instead of this, why not spend time demonstrating how regarding the world as made up of "hybrid networks" allows us to explain it better? Once one does that -- once one shifts one's philosophical ontology from dualism to monism, come to the dark side, we have cookies -- a lot of the supposed "controversy" dries up and withers away.I appreciate that our opinions are maybe not as different as I have suggested (oh, academic-types and their nitpicking!) but riddle me this: if discourse analysts by and large do nor operate on the basis of a 'bifurcation of nature,' as Whitehead put it, then why are the non-human things upon which ANT analyses focus almost completely absent from DA texts?
I'll say again: no discourse analyst who actually has read Foucault can consistently claim that discourse is somehow opposed to practice or "the material." The scientific-ontological duality you keep bringing up is a misreading of the social constructivist and discourse-theoretical claim, because it persistently misunderstands such explanations as ideational determinism. Which they are not. I am not putting a line between sociality and materiality, and you are not doing so in your Latour-inspired account of the social world. You are doing so only in the straw man you persistently attack ;-) Let it go, no one is actually arguing that here, and let's get down to business: whether Latour is better or worse than, say, Bourdieu or Foucault or Luhmann (or, what the heck, Gramsci or Braudel) as a way of making sense of the social world. In that vein, for my money what Latour is doing is relational discourse analysis, with a considerably broader sense of the extent of the relevant networks of discursive practice than most Foucauldians have and perhaps than Foucault himself had. Latour's networks make meaning possible, so we can sensibly refer to cities and the time and the results of an experiment; in that sense they're meaning-making practices just like Foucault's epistemic grammars are. So I personally don't see the fundamental difference.
Let's back up: the ANT concept of 'actant' comes from semiotics (mostly from Propp and Greimas) where it can refer to any character or thing within a narrative that does something -- it can refer to mountains and unicorns, equally, if either of those things does something in the course of the narrative. (This is where the 'principle of symmetry' between human and non-human things in ANT comes from.)
Looking at an ANT analyses in these terms the actants in these texts are many and varied: Latour's aforementioned 'Paris: Invisible City' invokes bollards, security cameras and atoms as well as people, texts and so on. If you look at any given discourse analysis text in the same way you will find a far less heterogeneous assemblage of actants: you'll have signifiers, meanings, subjects, texts -- at a push you might have human bodies and printing presses and so on, but rarely (I don't think that Foucault is especially representative of this creed, as it happens).
DA texts just don't draw on the same range of resources as ANT ones do. Doesn't this suggest some more profound philosophical differences beyond simply substituting 'discourse' for 'network'?
To take another example, in his book on Pasteur, Latour describes an event that Pasteur hosted at Pouilly-le-fort in 1881. What Pasteur did was stage essentially an act of scientific theater. He gathered together scientists, journalists, politicians and others and made a grand wager: He took two groups of sheep and infected them with anthrax. One group were inoculated with his vaccine, others were not. The inoculated group would survive, Pasteur claimed, and the rest would be dead in a few days. If he was right he would have proved his theory of microbial infection (or at least he would have submitted an extremely convincing proof of it to all in attendance and, consequently, also to their colleagues, readers, constituents, etc.).
Latour considers the whole event as theater, quite literally. He shows how it was carefully stage managed down to the last detail in terms of how the farm-cum-lab was set up, how the guests were treated, how the vaccines were administered and so on. But, and this is the point, who (or what) is the star that comes onto the stage at the end -- the fat lady, if you will --, at the triumphant moment? It is the microbes themselves. They enter the fray as fully fledged actors when they kill the unvaccinated sheep and spare the vaccinated ones. Not the microbes as they are represented -- not microbes as linguistic terms or microbes reduced to their meaning (and he is very careful to make this point). Sure, the whole socio-linguistic apparatus was necessary to make their appearance thinkable but it is still the things themselves that enter into the performance, as actors or actants.
Could or would a discourse analyst ever write such a thing? Could a non-human actant ever occupy such a role in a DA text? Would a discourse analyst ever even be interested in such an event?
Of course this is an impossible argument in the abstract -- there is no such thing as 'Discourse Analysis' only discourse analyses. If we include, for instance, 'Discipline and Punish' within the corpus of 'discourse analysis' (and I'm not totally on board with that) then perhaps we can say that some DA texts do allow for non-human actants to enter the narrative on an equal basis with signifiers and so on. But even then such instances are extremely rare.
Which brings me back to my previous claim: DA texts are extremely actantially homogeneous compared to ANT texts. Why is this so if DA does not, by and large, bifurcate nature and bracket off a sociolinguistic realm from everything else? I believe that it does, although I appreciate that my own 'theater of proof' may be less than convincing and is somewhat under-evidenced.
p.s. I wasn't trying to be unduly polemical in previous comments -- although sometimes that mode of discourse can be fun ;)
This discussion now bears little resemblance to the original post but hopefully it has been interesting to someone, somewhere!
Related to the OP, I've just started reading Robert Kaplan's article 'Revenge of Geography' and I can see that geographical determinism of the most blunt and ignorant sort is very much alive and kicking. This is food for critique, yet I do think that ANT and the like stand a better chance of taking apart this sort of paleo-materialist block-headedness than social constructivism as it has been traditionally practiced -- and for the reasons I've mentioned above.