Interview with Stephen J. Collier on Foucault, Assemblages and Topology



For the original version of this interview, and other TCS material, go to the TCS Website

Simon Dawes interviews Stephen J. Collier on the turn to topology in Foucault’s later lectures, the relation between topologies and assemblages, and the use of concepts such as governmentality and neoliberalism.
Simon Dawes: In your article in the TCS Special Issue on Michel Foucault (26.6, Nov 2009), you diagnose in the use of terms such as ‘redeployment’, ‘recombination’, ‘reconfiguration’, ‘problematisation’, and ‘patterns of correlation’, a topological turn in Foucault’s approach. To what extent does this topology (this shift from searching for ‘conditions of possibility’ to identifying ‘patterns of correlation’) resemble or differ from the concept of ‘assemblages’, of which you have previously written?
Stephen Collier: I don’t think that it is a shift away from a concern with ‘conditions of possibility’ but a different way to investigate them. But let me focus on the second part of your question, concerning “topology” and “assemblage.” One way to answer would be to refer to Deleuze, Serres, and others who have written and theorized about these concepts. But since I am not a theorist, and since I took these terms up in an opportunistic way, I will just try to explain the problems I wanted to address in using them.
There is obviously a common attraction of assemblage and topology: they present alternatives to epochal or totalizing figurations of a certain political system, historical moment, form of power, and so on. But I think that I was after something quite different in referring to them. In Global Assemblages, Aihwa Ong and I were trying to highlight how, in the space of the “global”, heterogeneous things are thrust together in a fashion that is hard to pin down with the usual diagnostic terms (globalization, neoliberalism, marketization, etc.). So in using the term “assemblage” we wanted to emphasize a rather foreshortened time frame and very contingent relationships. I am not sure that is entirely consistent with Deleuze, but it is what we had in mind.
In my Foucault article, “Topologies of Power”, I was interested in something different – more enduring patterns of correlation among techniques, technologies, and forms of problem-making. This is not a return to a global or totalizing analysis. But it is, at the least, back to concepts of the “middle range,” as Clifford Geertz once put it. What I like about the vocabulary of topology is that it points to more general mutational parameters that allow us to make sense of particular governmental forms.  For example, in the 1978 lectures Foucault describes the interface between mechanisms of discipline and institutions of sovereignty in the Absolutist states of Europe. There are patterns of interrelationship among techniques, forms of knowledge-power, and institutions that you can discover in country after country, over many centuries. Of course there are particularities of any given case that are important, interesting, and worth studying. But it is crucial to have a general vocabulary for describing how particular formations become possible and intelligible. Topological analysis tries to capture these mutational parameters, this space of possibility. To borrow another Deleuzian distinction, in the way I have used these terms, perhaps topology refers to virtualities and conditions of possibility whereas assemblages refer to actualities.
On a related point: I wanted to add that the language of topology addresses something that has been criticized about Global Assemblages (for example, in a recent article by Neil Brenner, Jamie Peck, and Adam Tickell). Namely, that in place of totalizing analysis it seemed to offer only a set of particularistic cases. If you read the chapters carefully, I don’t think that this criticism is fair to the contributors. But it is a reasonable criticism of our introduction, which did not propose a vocabulary for describing how the singularity of an assemblage might be thought in relation to a broader pattern of correlation or topological space (without recreating a kind of structural analysis of the kind Brenner and his colleagues would prefer). I tried to make some progress in this direction in “Topologies.”
SD: How significant is this shift in Foucault’s vocabulary? To what extent could you be misdiagnosing as an ‘epistemic shift’ in the ‘single logic’ of his approach, what is little more than a new ‘pattern of correlation’ in his conceptual tools? (Are you not using a more traditional Foucauldian methodological approach in demonstrating this shift, and, if so, how would your interpretation change if you were to use a more ‘topological’ approach?)
SC: Of course we can say that at different moments Foucault was simply grappling with different problems so he selected different methodological tools. And of course I emphasized the shifts in Foucault, and perhaps I minimized obvious continuities in method, in key concepts, and in diagnostic style. That said, I do think that we can identify some quite important modifications in emphasis. And it seems to me that these arose because Foucault identified deficiencies in his prior analyses. He chose (or forged) new tools because they were better: more precise, more perceptive of distinctions that matter.
Take this relationship between discipline, sovereignty, and the new figure of power (regulation, security, biopower, it slips around) that Foucault is grappling with in the late 1970s. In Discipline and Punish Foucault suggests that societies of sovereignty give way to disciplinary societies. Discipline takes shape in the “circumscribed spaces” of classical power and spreads, leading to a generalization of discipline as the basic form of modern power. In 1976 he adds another term – biopower – and discipline is rethought as one axis of biopower (the other is the “biopolitics of population”). So in the initial introduction of the term, biopower is mapped on to this rather epochal story that we find in his prior work: from sovereignty to discipline/biopower.
But in 1978 things look quite different. The shift is partially on the level of substantive claims. But it is also on the level of method, and if you read the first few lectures of that year carefully I think you see that Foucault is conscious of this. He doesn’t talk about disciplinary society or a normalizing society or about discipline as an axis of biopower. Discipline, sovereignty, and security are analyzed as different logics. More importantly, there is no succession from one kind of power to the other, but rather different patterns of correlation among them. Foucault recognizes that discipline is both an “archaic” form that was present in classical monarchies and a form that can be found in modern formations of power. The vocabulary of redeployments, recombinations, critique and programming and so on, is used precisely to describe this more complex space of interfaces and correlations among different forms of power. I don’t think such concerns are totally absent in the prior work, but here they gain new centrality.
SD: In Society Must Be Defended, Foucault’s 1975-76 lectures, he’s already discussing how the norm is that which circulates between two mechanisms, and where the normalising society is one in which the norm of discipline and that of regulation intersect ‘along an orthogonal articulation’ (Foucault, 2003: 253). Could we see this idea of intersection as a precursor to the more fluid, topological approach of his 1978 and 1979 lectures? And how would you account for his concepts of normalisation and regulation in your topological understanding of his later approach, and its application to contemporary contexts?
SC: No, I don’t see it as a precursor to the approach of the later lectures; indeed, I think that passage from 1976 provides a window on how much Foucault’s understanding changes between Society Must Be Defended and Security, Territory, Population. The passage you cite is one part of Foucault’s first attempt – unsuccessful, I think – to grapple with a significant shift in his understanding of modern power. Foucault recognizes in this lecture another kind of power, very different from disciplinary power, that is operating in modern societies. Its object is not the individual body but the population; its aim is not discipline but “regulation” – actually “regularization” is closer to Foucault’s meaning here. This is certainly an important conceptual development. But in key respects his reading here is very much in line with those moments in Discipline and Punish that present the most totalizing view of modern power. Though Foucault says that this kind of power (he calls it biopower) is constituted through an orthogonal articulation of mechanisms of regulation/regularization and mechanisms of discipline, it is quite clear that he does not see this as the contingent joining of things that remain heterogeneous. There is, as he suggests, a kind of continuity, a fundamental isomorphism, between these mechanisms of regulation and discipline. These mechanisms, he says, succeed in “covering the whole surface that lies between the organic and the biological, between body and population.” That’s a rather totalizing vision of things!
If we turn our attention to the 1978 lectures we find at least two very important modifications in the analysis. First, Foucault no longer talks about modern power in general but examines diverse formations of power in which heterogeneous things are linked up in a contingent – although in some cases quite durable – fashion. The critical distinctions concern the Physiocrats versus the early British liberals, or the classical liberals versus the Ordo-liberals versus the American neoliberals. I see this as a move away from the epochal shift from the classical to the modern age that had preoccupied him since The Order of Things and The Birth of the Clinic and toward a finer set of distinctions among formations of power that are shaped in response to particular problems and situations. Second, Foucault seems to revise the prior image of power. Knowledge/power, almost by its very nature, was a continuous, enclosed system that covers everything. There is no outside of it, and Foucault said repeatedly that if one wanted to critique it, one had to do so from a position of immanence that accepts the terms of a certain epistemic opening. This is where it seems to me that the analysis of liberalism introduces something new into Foucault’s work. Liberalism is not a form of governmentality – although it may contribute to the creation of diverse governmentalities – or a form of knowledge/power. Instead, it is a kind of reflection on forms of knowledge/power that has a restless, critical aspect. So in contrast to the knowledge-power period, here Foucault is interested in a kind of relational play of critical reflection, existing techniques, new governmental forms, and, I suppose, contemporary problems. I think that this new interest in critique – and in a very different understanding of critique – is part and parcel of the topological approach I have tried to characterize.
SD: You criticise the inflationary use of the ‘governmentality’ concept in Foucauldian literature. Can you give us some examples of this misuse, and advise on how you think the term should be used…if it should continue be used at all?
SC: A similar point about the inflationary use of governmentality in much literature has been made repeatedly – for example in the article that Nikolas Rose, Pat O’Malley, and Mariana Valverde wrote in the Annual Review of Law and Social Science. So I don’t think this claim about the literature is particularly original or, for that matter, controversial.
As for the use of the governmentality concept, this question is linked to the prior one – concerning the relationship among concepts in different moments of Foucault’s work, and whether one set of methodological tools supersedes the prior set. In “Topologies of Power” I argue that “governmentality” belongs with the analysis of knowledge/power. As such, governmentality is a very powerful concept. For example, Nikolas Rose has provided something invaluable in analysing the structure of advanced liberalism as a kind of governmental rationality or governmentality. There is no reason that such an analysis should not be complementary with a topological analysis. It certainly seems that Foucault thought they function together, as we can see in the 1978 and 1979 lectures.
So my point is not that we ought to discard governmentality as a concept. Instead, I wanted to highlight, and thereby try to avoid, a characteristic slippage, in which a scholar identifies a form of governmentality and then presumes that they have identified the formation of power in a particular society, country, and so on, at a particular moment; and in which they presume that, perhaps, in identifying a form of governmentality one has arrived at a kind of diagnosis of contemporary power relationships. The problem with this kind of analysis is that a governmental rationality is not a force that pervades all social relationships or political institutions. A form of governmental reasoning is invented by situated actors and deployed in particular situations to reform particular governmental institutions. You have to actually analyze all that to figure out what is going on, how expertise and expert reflection relate to structures of political authority, concrete policy, techniques of governing and so on. To do so, you have of course to analyze particular cases, and pay attention to what is singular about them. But it is of course much more powerful to find an analytical tool that provides leverage in describing forms of correlation that can be observed across many cases. This is the register of topology.
If we are more on guard against this kind of slippage then a concept like governmentality can be deployed in a more limited but also more precise way. An example might help here: It is very useful, in my view, to identify something like “government through calculative choice” as a key feature of an advanced liberal rationality of governing. But if you are interested in neoliberal reasoning about a particular sector or problem – health care, education, redistribution, tax policy, and so on – if you are interested in neoliberal reforms, or if you are interested in how institutional change relates to neoliberalism, you are going to need some other analytical tools. It isn’t possible to transform every relationship in every sector to function on principles of self-interested, individual calculation and choice. And in fact, this is not what any neoliberals propose! So how do things actually play out? What are the limitations of “government through calculative choice” and how are these recognized and addressed in neoliberal thought? How are choice mechanisms articulated with existing structures (welfare systems, regulatory apparatuses, material formations, etc.)? How does neoliberal reflection take up existing social norms and what does it do with them? The point of “Topologies,” again, was to identify some tools for addressing such questions, and to show that one need not proceed only through particularistic analyses that just multiply case study after case study. One can also try to identify broader patterns of correlation – between, for example, choice mechanisms and social values, or neoliberal reform and the social state – that are found across many cases.
SD: So continuing with the significance of pursuing this topological approach, and this move away from ‘governmentality’, for research into neoliberalism, how does it change our understanding of neoliberalism?
SC: This is the central question I address in my book on Russia, Post-Soviet Social, which is coming out this year. One crucial point is that a topological perspective forces us to think again about how neoliberalism or advanced liberalism (and I appreciate the difference between these terms in the relevant literature) relates to existing governmental structures, such as the institutions of social welfare and economic regulation associated with the social state. In a great deal of critical scholarship, the story has been that neoliberalism just displaces the existing norms and forms of the social state. Or, if elements of the social state are understood to persist, it is because they resist neoliberalism, but in any case they are opposed to it. But I have not found that such descriptions are very helpful for understanding neoliberalism and neoliberal reform in the sectors and the countries I have studied. It isn’t so much that neoliberalism displaces the social state. Instead, modifying Foucault, I have found that neoliberalism presents three things: first, a critique of the outcomes of the existing norms and institutions of social welfare, on the grounds of their inefficiencies and their inequities; second, a politico-philosophical critique of how norms such as social justice or public value are formulated and how the proper scope of governmental activity is conceived; third, a new programming that establishes a novel pattern of correlation between choice mechanisms and social welfare. If you actually analyze this form of critique and programming I think you get a much better understanding of the contours of a new topological space. And you also get a better grip on what is going on in particular countries and sectors.
That just points to the topic. But you get a sense that the picture is quite different from the usual account of neoliberalism.
SD: Would you say that there has been a recent ‘topological turn’ in the social sciences, and, if so, how much do you think it is influenced by the recent translations of Foucault’s later work?
SC: In the critical social sciences it seems to me that we see, on the one hand, a lot of abstract theorizing about the condition of the present with thin invocation of “examples” that serve to illustrate what a theorist has already claimed to be broadly true. On the other hand, there is a lot of fine-grained empirical work (much of it influenced by ANT and sundry offshoots) whose conceptual contribution can be a little obscure. Topological analysis probably belongs, again, to some middle range of theorizing or concept-work. Paul Rabinow would call this middle range “inquiry.” Of course there are many terrific examples of this kind of work, but I don’t know if there has been a “topological turn.”
I am also not sure that Foucault’s work of the late 1970s has, for the most part, been taken in this direction. The literature on governmentality took up some available fragments from this period. But my view is that these fragments were read too much in light of the Foucault of Discipline and Punish and the framework of knowledge/power, and that they didn’t capture important aspects of his analytical style during this period, which you only get by reading the accounts of Physiocracy, the British liberals, the ordo-liberals, and so on. We also have a massive range of theorizing that has derived from the 1976 lectures, particularly the last lecture of that year. But something quite different is going on in 1978 and 1979. I don’t know that it radically changes our understanding of Foucault, but it might change our understanding of the kind of inquiry that can be conducted with Foucaultian tools.
SD: How have you continued to work with these concepts since writing the TCS article in 2009? Does your forthcoming book on Russia, for example, deal explicitly with ‘assemblages’ within a framework of ‘topological space’?
SC: I would avoid talking about assemblages ‘within’ a topological space. An assemblage is not contained within or constrained in its forms by a particular topology. Perhaps my language has been too loose in that respect, or perhaps here the metaphor simply breaks down. I would rather say that there are two analytical dimensions: that of topologies and that of assemblages. I am interested in inquiry that moves between these two levels, and that employs each as one tool (but not the only tool) for rendering the other intelligible.
In the work on Russia, I tried to examine the singularity of the post-Soviet case – the interface between the Soviet project of social welfare and post-Soviet reforms – as a site for examining more general forms of correlation between mechanisms of calculative choice and the institutions of the social state. The point is not that the post-Soviet offers a window on the essence of neoliberalism; indeed, for a variety of reasons I discuss in the book, contemporary Russia is quite peculiar and untypical when viewed in comparative perspective. Nor is it that post-Soviet Russia can be understood as a “case” of neoliberalism, an example of a global logic working itself out, however much attention one plays to local specificities and circumstances. Instead, two dimensions of the inquiry proceeded in conjunction and communication – a linked investigation of conceptual tools and particular sites or multi-sites.
My current project on the government of catastrophe in the post-World War II United States with Andrew Lakoff has a similar focus. Part of our interest is to trace out the specificity and contingency of developments in one case – albeit a centrally important case – to understand how the government of catastrophe became a central concern of government in the United States. But we are also investigating a more general topological space. Our argument is that in the post-World War II period we see the emergence of a new technology of government that is quite different from the security of “population” that Foucault described in 1978. The point is not that this new technology of government replaces the prior one. Instead, there are new configurations of techniques, technologies, institutions, and forms of critical reflection that are popping up in many different contexts, not just in other countries but on a global level as well. So in that sense these methodological concerns are very relevant to my current work.
Stephen J. Collier is Assistant Professor in the Graduate Program in International Affairs at the New School, New York. His research interests include neoliberalism, socialist and post-socialist urban planning, contemporary security, infrastructure, and welfare. In all these areas he has examined forms of governmental rationality, their recent past, and their present transformations. He has worked in Russia, in post-Soviet Georgia, and in the United States.
Simon Dawes is the Editor of the TCS Website, and Editorial Assistant of Theory, Culture & Society and Body & Society
To read Stephen J. Collier’s articles in TCS, click on the titles below:

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

This site uses Akismet to reduce spam. Learn how your comment data is processed.