This past Thursday I was faced with the delightful task of trying to defend the “performativity of economics” literature from two of the best informed skeptics I have ever encountered. I can’t quite remember how the conversation shifted to the topic, but I soon found myself trying to re-justify the entire program from the ground up in order to defend my take on it, as well as the work I would like to do using the concept. So, here is a brief gesture in that direction. [Note: If you haven’t heard about performativity before, you should jump to the big block paragraph near the end of the post and read that then come back here for the more historical argument.]
First, I want to note that I am only going to talk about Michel Callon’s notion of performativity, and not any of the myriad derivative works (many of which I rather like, but each work’s use of the term deviates in small or large ways from Callon’s, as such works always do). In particular, I am not going to speak to the whole MacKenzie set of performativity, counterperformativity, effective performativity, generic performativity, Barnesian performativity, etc. I think in trying to multiply the concept to make it more analytically rigorous, MacKenzie may have (likely unintentionally) moved the term away from its original concern (see below) and towards a different set of questions (about the truth or falsity of economic models in general). For Callon, the success of a model and the truth of its statements are not distinguishable – a principle derived from his pioneering work in science studies and in particular the Actor-Network Theory. So the idea of a model ‘making itself true’ is not so radical – any model that becomes successful in some sense makes itself true (again, truth within a particular kind of discourse, not capital T Truth which science studies scholars usefully avoid). From Callon’s 2007 update of his programmatic statement in Do Economists Make Markets?:
As MacKenzie shows… at a certain point in time, in certain places, the world of the formula is actualized, in such a way that it can be said that the formula describes and represents the world correctly. We are no longer in the register of truth as a reference but – to stick to the same word – in that of truth as success or failure, in truth as fulfilled conditions of felicity.
However, Callon’s original interest in economics, I will argue, was not in its success or failure per se (a classic science studies topic), but rather in its model of the agent. Actor-Network Theory has often been criticized for its reduction of agency to something seemingly silly: if scallops and mosquitoes can have agency, what does agency mean? (See this earlier post for some thoughts on that.) In 1999, a year after his first programmatic statement about the performativity of economics in the introduction to The Laws of the Markets, Callon published a chapter in an edited volume titled “Actor-Network Theory – the market test”. In classic French theory style, Callon argues that ANT’s lack of a theory of the actor is actually one of its strengths, as it allows ANT to explain how markets work. Economics (broadly construed to include marketing, accounting, etc.) frames the world in order to make it calculable, and thus make being the rational economic actor (homo economicus) possible. Thus, ANT’s ability to explain how markets work justifies its lack of an overarching theory of the actor – the ‘way of acting’ (agency) is endogenous to the system, in a sense. Economics has been very successful (and thus its models have been “true” in many times and places), which is not utterly surprising or particularly interesting from a science studies perspective, but the way in which it has been successful is interesting: it’s a clear example of the creation of a kind of agency.
OK, so I’ve probably rambled too long so far without actually explaining Callon’s performativity. So here’s my paragraph long summary of what Callon is doing and how, taken from an edited version of my ASA submission:
In his programmatic statement, Michel Callon (1998) argued that sociologists interested in the economy and economics ought to focus their attention on what economics does rather than emphasizing the ways in which economic models fail to capture the complexities of social relationships. In particular, Callon argued that economics plays a constitutive role in modern markets, that markets are “embedded” in economics. Economics, which for Callon is broadly construed to include accounting, marketing and related disciplines, plays two essential roles in the creation of markets. First, economics frames the world in such a way as to make it calculable. Second, economics creates calculative agencies that make use of the newly-framed world in order to act in rational ways. In Foucault’s (1982) terms, economics forms certain kinds of objects (through framing) and subject positions (calculative agencies). Thus, homo economicus is not an inaccurate description of human behavior, but a product of economic theory. Callon labeled these processes of object and agency-formation “the performativity of economics.”
Before finishing this post, and my defense of Callon, I want to note the connection I try to draw between Callon and Foucault’s Archaeology of Knowledge. In AoK, Foucault lays out the methodology underlying his early works, a methodology of uncovering discourses and how they have changed. In examining the changes in medical, psychological, and social scientific discourse, Foucault realized he had trouble even describing what unified those discourses. Was it a single object – disease, insanity, etc.? No, it could not be, as discourses constitute their objects, and the way they constitute that object can and has changed over time. For example, economics before the 1930s examined markets, but not “the economy” (the argument of my ASA paper, and some work by Mitchell and Emmison). Afterwards, macroeconomic discourse focused on the economy (which it had helped to construct).
Similarly, subject positions could not unify discourses – i.e. medical discourse is what doctors do. The problem is that discourses can constitute and reconstitute those subject positions – “enunciative modalities” – and thus they too fail to serve as good unifying principles of discourses. In the end, Foucault argues that discourses are unified in a given moment by the rules of formation of statements as a whole, but not by the particular objects or subjects involved. In realizing that discourses constitute objects and subjects, Foucault comes face to face with the interaction of discursive and non-discursive practices, and soon abandons purely archaeological work for genealogy and then the analysis of power/knowledge, which focuses much more on those interactions and constitutions.
Callon follows Foucault, and although ANT looks at a much lower level than Foucault’s work, many of the principles are the same. I think, in many ways, (Latour and) Callon’s work continues Foucault’s, challenging ontologies and looking at the way that constellations of elements define situations and their outcomes. So, much as Foucault showed how disciplinary techniques reshaped schools and prisons, and constitute new ways of acting on individuals as objects of knowledge (and thus a new form of agency, in ANT terms), Callon shows how economics (and related practices, like marketing and accounting) shape and reshape markets, commodities and the ways in which people act in those markets and on those objects. “The performativity of economics” thus follows in the tradition of studying power/knowledge. Like much of Foucault’s work, Callon focuses on what economics succeeds at doing (“framing”, creating “calculative agencies”) and less on what it fails to do (the “overflowing” of externalities which sociologists have already spent years cataloging and will continue to do). It’s not that the framing part of the story is the only part, but rather that sociology had ignored it while focusing exclusively on the overflowing – the ways economics fails, misleads, or ignores other ways of being.
Ok, I think that’s enough for now. What do you think? Does it make sense? It’s worth noting that this particular ‘defense’ of Callon was aimed at someone skeptcal of his work, but less so of Foucault’s (his theoretical positions anyway, if not the details of the empirical arguments).
A (Budding) Sociologist's Commonplace Book 
Peter
/ February 21, 2009I can’t quite remember how the conversation shifted to the topic, but I soon found myself having to re-justify the entire program from the ground up in order to defend my take on it, as well as the work I would like to do using the concept.
It’s a fake out, Dan. There is no need to feel like you have to re-justify the entire program of STS or SS of finance. Don’t let intellectual bullies play that game.
Dan Hirschman
/ February 21, 2009It was friendly, and they were buying me a nice dinner. I was happy to oblige… and I seemed to make some headway. They didn’t actually require me to re-justify the program, I should note, but that seemed the best way to make my argument. I think they were reading Callon (and SSF) outside of the context of STS, and that was the problem, not that they were rejecting the premise of STS.
I changed having to “trying” to better reflect the events of the evening.
But thanks for the support!
neville
/ February 22, 2009I really enjoyed the post. Thanks!
It is, indeed, about time somebody started linking F to ANT in a rigorous way (there is a little of that going on in org studies but the rest, to my knowledge, is rubbish).
Still, I am little cautious of this link between AoK and ANT/Callon. For example, when you indicate that Callon is interested in “truth within a particular kind of discourse” this does not seem to correspond to any of the Callon that I am familiar with. Discourse, in Foucault at least, always seems to me to rely on some kind of multi-scale ontology (despite the disclaimer he will sometimes give, I can’t help but see his earlier work as a glorified Plato or Durkheim. Just replace Forms or society with discourse!). On the other hand, Callon (and the F of DandP onwards) seems to only want to formulate a flat ontology. And, here, by flat ontology I mean something like a desmology, a focus on circulating traces, relations and processes – what Deleuze calls rhizomes or assemblages, what F calls an apparatus (although they are conceived as more closed and rigid), what Callon calls networks, or what Latour calls (when he is not using the term network) modes of attachment).
Similarly, when you write that “Foucault argues that discourses are unified in a given moment by the rules of formation of statements as a whole, but not by the particular objects or subjects involved” I wonder how much *further* from Callon we can get. Callon, to my mind, wants to avoid talking about anything that even resembles rules of formation (and this seems to be precisely why he ‘rejects’ so much sociology, because of its emphasis on bubbling underground practices – I wrote an essay last semester on a possible rationale for this move, in the context of his work on economics, if you are interested). Additionally, subjects and objects *do* mutually constitute each other for Callon. All that exists are actors… One could even imagine Callon exclaiming that discourse displaces the action from where it really lies…
I do, however, most certainly agree with you(as if you even need my agreement!) if you are only suggesting that ANT seems to follow Foucault out of that type of problematic that comes from his pre-DaP work. Latour, incidentally, retorts in interview (he is asked on F) that slashing power and knowledge evades the real question of how truth gets done. But you are right that they become/were always interested in the same things. It makes sense, too, when you consider Foucault’s and ANT mutual accomplices : F was close to Serres in the 60’s (they effectively write Words and Things together), while from the 70’s onwards Serres became close with Latour and influences ANT a lot; similarly both Foucault and ANT draw on Nietzsche, Deleuze, and pragmatism a lot.
I am rambling too much. I will stop:
One question: do you think that ANT looks at a much “lower level” than Foucault’s work for any theoretical reason, or do you think this focus is incidental to the ANT/Callon approach?
And a more pedantic question: do you know roughly how much of an impact is Callon actually making in economics (generally) and economic sociology (in particular)?
Thanks again for the post. And sorry if my response a little negative (unfortunately, it does reading over it; not my intention, at all).
Dan Hirschman
/ February 23, 2009Neville: Thanks for the comment and the historical background! While I hang out with some hard core Foucauldians (and some hard core ANTers), I have thus far only dabbled in each. So it was very helpful for me. I think Foucault’s ontology is messy, and does change over time (or at least his emphasis). I was writing a term paper at the end of last semester, all set to argue how Callon was like Foucault in certain ways (emphasis on the creation of subject positions and the framing of objects) but with a seriously different and more micro (or, in your better term, ‘flat’) ontology. But I couldn’t make it stick, esp. for post D&P Foucault, for whom practices are an important part of the story. I do think the Foucault of AoK is trying to stick to this big D discourse as an autonomous realm, which is about as far from Callon as you can go, but it falls apart. So I think we are the same page with the idea of ANT and later Foucault both moving past that vision and into one of a flatter world.
As to the first question, I’d need to read and re-read the later Foucault more to see how different they are at that moment – it’s hard to compare, say, the Order of Things to Science in Action. I don’t think there are really incompatible differences in scale. But ANT has often focused on controversies, and moments when nature/society were obviously unsettled, whereas Foucault (even the Foucault of D&P) has focused more on the big shifts in discursive and non-discursive practices rather than the micro-level pathways through which they are (and always have been) fought out. So, I think compatible, but aiming at slightly different sorts of thing most of the time. Does that make sense?
As to the second question… I doubt Callon is having much of an impact on economics at all. Economic Sociology, bless its heart, has had a relatively modest impact if any in its 25+ years as a re-emergent subfield. Psych and evolutionary biology seem to be making a splash over there, but that’s about it. Within econ soc, Callon is certainly being cited more, and the social studies of finance crew (esp. MacKenzie) have been very prominent. His article with Millo won the best article award in.. 04? And his book, “An Engine, Not a Camera”, won best book award last year. I am not 100% sure how closely aligned MacKenzie and Callon are at this point (something I argued, albeit somewhat ramblingly, in the post), but they are certainly not too far apart! So, broadly speaking, the performativity literature is getting some play in Econ Soc, but not (to my knowledge) in Econ proper, which is not surprising.
One of my favorite things about this literature is how it helps us recognize that we had been previously trying to have a dialogue with someone utterly uninterested in talking with us. By trying to point out what economics was missing (‘overflowing’), we were, often explicitly, trying to get economics to be better or to be less powerful. That tactic was failing, perhaps in part because we were failing to examine what economics did (‘framing’) and thus the way it was embedded in markets and economies. Economics isn’t just another social science, sitting back in its offices speculating about human nature, it’s also out there in the trenches getting its hands dirty (or, more often, making the world ‘cleaner’, or ‘safe for economics’).
Does that answer your questions? And thanks again for the thoughts/comments! And definitely send your essay my way (email address available on the about page).
neville
/ February 24, 2009Thanks for the swift answers. Much appreciated.
As someone who is most positioned well outside of economic sociology, its interesting to get a closer perspective on what is going on over there.
And even though I would dearly love it to be, I must admit that the phrase ‘flat ontology’ is not mine. If my sense is not failing me, it comes from Deleuze. And like all Deleuze it is a little bit misleading taken at face value. But that is another story entirely.
Will send essay presently – I think I should write a disclaimer first.
diogo
/ February 25, 2009Just thought you’d be interested in this:
http://www.wired.com/techbiz/it/magazine/17-03/wp_quant?currentPage=1
So economists might not want to hear it, but the moment is right for a sort of pop breakthrough.
Dan Hirschman
/ February 25, 2009Thanks for the link, digging in now. My first thought is that statements like these are shockingly prominent in the world: “His method was adopted by everybody from bond investors and Wall Street banks to ratings agencies and regulators. And it became so deeply entrenched—and was making people so much money—that warnings about its limitations were largely ignored.”
I wonder if anyone in econ soc or science studies has already written about this (perhaps Beunza et al.’s new work on reflexive modeling is close enough), but there seems to be a sort of story like this for a lot formulae and other statistics. A mathematical/statistical/economic wizard figures out a new way of looking at something, with all sorts of technical problems, that something travels out into the world and gets embedded into markets (Callon-style) and then the world overflows (in often the ways the wizard foresaw) and everybody ends up blaming the model and its unreflective users while the wizard ends up decrying how everyone missed all the caveats he’d been writing about since day 1. I wonder how inevitable this cycle is. If statistics (and similar models) make things that “hold together” (cf. Desrosieres, although he was focusing on governmental data collection and publication and not modern statistical modeling, I think similar dynamics may be at play) then those statistics are going to travel out and be used in ways the original creator expressly forbid – the goldilocks amount of “holding together”, tightly enough to be used, not so tight that it gets misused, may be impossible. I wonder if we have a short name for that sort of story already and I just don’t know it.
Of course, sometimes the world never overflows enough to break the formula. In 1948, Simon Kuznets wrote a scathing critique of the way national income accounting had progressed during WWII, and especially the adoption and institutionalization of a particular set of decisions for how to count certain things (modeled on the British version). His criticisms seemed to go nowhere, although they reappear in critiques made by development scholars. But it’s harder to overflow a description than a model of this sort, since the description appears to be just that, a neutral observation, and not an imposition on the world every bit as precarious as a complicated financial model.
diogo
/ February 26, 2009Hopefully Millo and Mackenzie will get to that as they move along their analysis of the Black-Scholes model. After all, it is not often that one wins the Nobel prize in economics in one year and loses 5 billion dollars the next.
On a personal note, that last point is what my dissertation is partially about. Country risk ratings, their spectacular failure in predicting most crisis, and the differing paths Brazil and Argentina took this decade. Why on one country the discourse of economic stability and risk reduction remained relevant and rhetorically important as a left wing president played nice with “the market,” while the other experience a resurgence of developmental economics (even though it was the poster boy for the benefits of risk management a few months earlier).
In other words, why did reality overflowed enough in Argentina to make them discard the formula, but not enough in Brazil.
diogo
/ February 26, 2009Oh, forgot to add: even Brad Delong liked that article, which is somewhat surprising given how strongly he defends neoclassical economics.
In fact, I think it is incredibly interesting that so many prominent neoclassical economists have blogs nowadays. Such a neat contrast between their informal, quasi-neoinstitutional discourse about politics and the economy on their blogs and their strict micro-foundations approach in their professional writing.
Dan Hirschman
/ February 26, 2009@Diogo – Good luck with your dissertation! It sounds very interesting. Out of curiosity, do you look at what happens with country risk ratings after their failure? That is, are they still used the same way? What if anything changes? How are the failures justified? Among other things, it’d be a really good place to see how discourses about the truth of the market are constructed and defended.
And yes, I am addicted to economics blogs. Big names like DeLong, Krugman, Mankiw and Rodrik, profs who have become big names in part through their blogs (e.g. the marginal revolution folks), and some more specialized things as well. I’m starting to have to cut back due to overload, and double-count reading those blogs as both getting my news and doing research!
diogo
/ February 26, 2009Re: failures justified:
On the part of the big institutions, it is seen as sort of a minor “normal science” failure. Or, to use Peter Hall’s terminology applied to policy evaluation, you have a sort of second order change in paradigm. In the case of Argentina, it was pretty much a “oops, should have paid more attention to the deficits at the provincial level,” and then they change how they evaluate certain instruments and policies (i.e., currency pegs go from being the silver bullet to inflation to a disaster, and free floating currencies are more valued). Of course, given their perspective and their role, these institutions never go through the sort of paradigmatic change. But alas, I digress.
Keep up the good work, as I have no idea how you have the energy to be a grad student, do this blog and also that sociology life one.
diogo
/ February 26, 2009ps: forgive the typos and other minor mistakes. It is way to late to be coherent. Is there an “Edit” function I am missing?