A bit of background might help elucidate how and why I arrived at my last post on things and agency. I was suddenly compelled to write about it by Jeremy's post but I'd been thinking about it for a while.
Until quite recently I too was quite happy to use 'agency' in the more-or-less Latourian fashion (as a catchall term for 'causal efficacy' of any sort) and considered any objection to it to be simply reactionary. It was attending the Millennium journal's conference on 'Materialism in World Politics' a few months ago that changed my mind. The conference itself seemed to be split down the middle to some degree in terms of the theoretical persuasion of the attendees: half were dyed-in-the-wool historical materialists of one sort or another; the other half were acolytes of materialism after the fashion of Latour, Connolly, Bennett, Haraway, etc. I found it interesting that the to-and-fro of the argumentation between the two camps eventually ended up orbiting around one issue in particular: that of the 'agency' of things.
Some, Colin Wight and David Chandler for two, argued strongly against the attribution of agency to things. Another person even called it 'fascist' since it 'eliminated the difference between people and things.' This objection is, of course, totally reactionary. Granting 'agency' to things doesn't destroy the difference between humans and other kinds of things; it makes the difference one within a plurality rather than across a duality and ontologically secondary rather than primary. However, some objections were more nuanced. The argument that really persuaded me was that the concept of political agency is a hard-won spoil of ideological conflict and that it mustn't be taken for granted because many political ideologies deny the notion of human beings in general having the capacity for reflexivity, self-consciousness and intentional, purposive, intelligent action. We see this today with the creeping technocratisation of governance in Europe, disguised and rationalised by economics and the cult of There Is No Alternative, for example. Such ideologies are real today and they were even more so in the past.
This made me realise that agency in science and in politics are quite different things. In science human agency, conventionally defined, has been granted far too much but in politics this is not the case; in politics the very capacities that were so overstated in scientists need to be cultivated, nurtured and promoted. What in one instance was so hyper-abundant as to be a pest in the other is, if not endangered, not in overwhelmingly good health.
We certainly need to understand the impurity and complexity of politics in terms of the massive proliferation of things qua 'issues' (as in 'dingpolitik') and we need to be rid of the myth of naked humans talking amongst themselves but we also need to recognise that there is an enduring truth in the various theoretico-political myths of 'humans amongst themselves' -- that there are specificities to human being that require special theoretical terms. Human being may not be unique in type but it is singular in importance and so we need to be careful when we are taking apart the concepts with which its specificity has hitherto been captured, even if we are justified in performing such disassembly work in general.
So, I don't pretend to have any answers here. I'm not saying that the old humanist version of agency must necessarily be preserved. I'm largely won over by Levi's articulation of agency as gradational and to be attributed in its varying forms on the basis of empirical evidence rather than a priori. Jeremy's splitting of the term into three works well too.
I suppose what I'm trying to say programmatically is that a lot of work has been done un-bifurcating nature, flattening ontology, pluralising duality, post-ing humanism and so on and a large part of this has involved dethroning the naked human Subject. However, that, for any social or human science, can only be an initial, critical move; an earth-clearing exercise. The attribution of 'agency' to things worked well at that stage but it has become dehistoricised and ANT has become known as 'the theory that things have agency' -- as ANT quietly steamrollers its way through the human sciences this claim becomes gradually indissociable from everything else anyone has ever said about it. A far greater degree of nuance than this is now required
It seems that Levi and Jeremy are both well ahead of me on this but I met numerous people at the Millennium conference who were still very much stuck with the impression that the claim that 'things have agency' is very much the insight of Latour, ANT, etc. I now think that this is not the important insight so much as it was one way of articulating a broader point in a way that worked analytically, polemically and philosophically in that particular context (science studies in the 1980s) but that is less than adequate now.
I suppose this is ultimately just a 'baby and bathwater' argument: I think that 'agency' as traditionally defined has some conceptual and political importance, despite its flaws, so we should be careful when redefining it, though redefine it we must.