first, latour's work is ground-breaking. for that reason alone it is difficult to understand since we are ill-equiped both to understand the precise ways in which certain words are meaningful, and in the more general patterns of how such words form discursive patterns. i've read a bit of him through howard becker (in the latter's telling about society [2007] for example), and have appreciated his thoughts more directly through what little i've read of science in action (1987).
i thought the audience, at least judging by the majority of responses and some conversations i had afterward, was a little cold (inhospitable even) to what he was proposing. then i thought, what a wonderful illustration of his premise: namely, that matters of fact have become matters of concern in the sciences. audience responses revealed thinkers more interested in (concerned with) semantics, it seemed, than the content of latour's talk. some, i'd argue, were more concerned with positioning themselves as various personae (mostly academically) in the room than they were with engaging latour's specific arguments, suggesting more a thing or two about institutional politics and ego-formation in hierarchical systems than about anything to do with problems about objectivity. (an exception: peter ryan's review, while brief, is at least even-handed and more giving to the territory latour was stridently marching through.)
perhaps i'm not the best person to ask (because i had far fewer qualms than did others about how he was framing things, and was more interested in what he was getting on about), but here are a few things i got from his talk.
1. first, that some objectivity can still be asserted in natural and social scientific research. however, the ways in which it's to be asserted have changed (i.e. objectivity cannot be divorced from the networked and interacting systems, local and global, of which everything is a part). i think this is what he was suggesting as a tension between the object and thing: he used the example of the challenger space shuttle (as object, symbol of scientific progress, and as thing, exploded scraps of forensic meaning) to suggest that there's a continuity of meaning between the two that often gets obscured (too often, and too quickly) as one assumes certain dominant interpretations of an object, rendering other equally viable interpretations invisible, even those which it is clear are no less important (like faulty o-rings known in advance to be faulty).
by way of comparison, and as an example entirely out of personal interest, i think of the photograph—any photograph. the last thing that's ever seen of a photograph is that it is a photograph: it's always, primarily, a photograph of something. yet to discount from any interpretation the enormous and elaborate systems of technological, economical, political, cultural and other forms of organization that have contributed to an author creating that image is to misread it entirely. (something we almost invariably do, of course, but that's another matter.) it's a representation, after all: an object holding some validity (contextual meaning), but not necessarily complete facticity (definitive or universal meaning).
thus what i gather latour means by taking an "object-oriented" approach is for researchers to pay attention less to the object as the source of meaning (which is ultimately partial and contingent), but to the object as a nexus of qualities, characteristics, and valuations. not as a fact, but as a vehicle (a rhetorical vehicle?) through which facts among other values (concerns) can be found, such as political orientations, struggles for real and symbolic power, and so on.
2. from this ultimately pragmatic position, how we choose our methodological and representational tools and why we do so is crucial to our practice. we'll do best, he suggests, to make this methodology explicit, rather than implicit. doing so will allow us to broaden our repertoire of representational techniques, sharing research findings in more ways than simply monographs or peer-reviewed journal publications, and work iteratively to refining representational forms. as our lack of knowledge in some of these areas becomes evident, greater collaboration becomes necessary as our practice expands outward to connect with others who possess better craft-skills in certain areas than do we ourselves. as our work needs to be shared amongst a broader-ranging audience of participants and collaborators, it has a greater likelihood of becoming more widely communicable and meaningful to a broader network of stakeholders. Both of these possibilities are politically progressive: first in making improvements to the ways institutional research gets conducted and shared; and secondly in engaging others in more meaningful, pragmatically grounded questions.
anyway, some rough thoughts. would love a ping back or two, especially if you feel i'm being unfair at all, or non-objective in my estimations ;-)
