My attachment to Michel Callon’s markets

Editors' note: This essay is the first in a three-part series devoted to reviewing the influence of Michel Callon's work, from his edited collection The Laws of the Markets in 1998 to the publication of Markets in the Making: Rethinking Competition, Goods and Innovation in 2021. The other two parts are collectively authored essays that bring together some of the most significant, and most provocative, established and emerging voices in the field to celebrate and review the impact of Callon's work.

No one will ever know the answer to the enigma of the market. (Michel Callon)

Things even people have a way of leaking into each other like flavours when you cook. (Salman Rushdie)

Let me be straight. Ever since The Laws of the Markets (1998) came out and threw a curve ball into the PhD I was almost finished writing, I have been trying to figure Michel Callon’s markets out. Back in 1998, I thought I knew what markets meant to me. Having come out of the interdisciplinary, polytechnical mangle of British academia’s cultural turn I had learned just about enough to disagree with things. I disagreed with Roland Barthes and Walter Benjamin, but I agreed with Michel Foucault. Often. I had reasons, some empirical, some contrarian, that left me impatient with arguments that made ‘the’ market a singular, monolithic force that could be modelled, measured, corrected, critiqued or damned, with scant reference to how human individuals encounter it (or, more pedantically, them). But neither Foucault, nor most of the post-structuralists in his wake, really wrote about economies or markets. Now there was a whole jumble of sociologists, anthropologists and accountants who did and none of them spoke my language. Anything can get lost in the sociology of translation – markets, economies, culture, anything.

At that time, I was in the thick of a meandering conversation about how carelessly anglophone cultural studies and adjacent sociological enquiries treated concrete economic practices that led eventually to the launch of this journal. To the extent that the group of people involved agreed on anything they agreed that culture was a tricky word and cultural not much better. ‘Cultural economy’ stuck around anyway. In a niche parallel to ‘actor-network theory’ none of the initial protagonists were attached to the label. The term actor-network theory, as Madeleine Akrich (2023, 171) remembers it, was seldom used by its best-known associates – Michel Callon, John Law and Bruno Latour – but still ‘met, for mysterious reasons, with a certain success.’ As Callon’s inquiry turned from scientific and technical innovation towards markets, this question about the mysterious conditions that allow certain innovations to become successful, to ‘find their publics’ (Mallard and Callon 2022, 154) has never gone away. In Markets in the Making (MITM) it is subjected to the serious, sustained, detailed, forensic, creative and entertaining analysis that it deserves. To understand the ‘felicitous conditions’ in which an innovation succeeds, and the infelicitous ones in which it fails, matters a great deal (Callon 2010). It matters for the same reasons that drive hostility to ANT – because questions of power, inequality and allocation matter. Without rekindling a debate about the difficult politics of ANT it is not, never has been, just about markets. If there is a political lesson to be learned from the decades of theoretical market modelling and concrete market studies that MITM so generously, so encyclopaedically, synthesises, for me it is about the purposeful orchestration of action. And that is as germane to the political necessity of organising things – risk, assets, public services and public infrastructures – differently, as it is to understanding markets per se. Callon’s basic question may be ‘what is a market’? but it is not only markets he is concerned about – ‘if one wishes to act upon markets, the target is not world visions. The target is the devices that frame markets’ (148).

That said, there is a LOT of market conceptualising in these 426, basically unillustrated pages (there is an exceptionally helpful Sketch of a Table on p. 144 but it takes until p. 388 to get to Figure 1). Much of that conceptualising is focused on correcting the market interface model. That is not what interests me though and since most of Callon’s many interlocuters – Neil Fligstein (2023) and Alvin Roth in this symposium, for starters – are better qualified than me to interrogate those arguments, I won’t. I have instead two other trails I want to follow: market agencement and affectio mercatus.

At the very start of the book (12), Callon says that the ‘conceptual disintrication’ (a phrase I have already stolen in my head) of markets and capitalism is the most important result of getting into the ‘empirical weeds.’ The tool he has been sharpening for decades for the purpose is ‘market agencement.’ In English, MITM is the most sustained and explicit account of what this magnificently confounding term means. Agencement is part of a whole glossary of closely scrutinised words that belong primarily to the market branch of ANT’s sociology of translation (I know; there is some label redundancy here but so be it). Cochoy’s (2008) useful coining ‘qualculation,’ of course, together with a few new ones – passiv(a)ction, pass(act)ivation – sit alongside the technical re-specification of old words; device, calculation, attachment, detachment, affectation, with Latour’s (2005) reassembled ‘social’ working in the background to accommodate the active market participation of non-humans (74). Agencement sits in the centre of this, a word clear enough in French, but notoriously impenetrable and unpronounceable in English. (Anecdotally, there is a tendency among first-language English speakers to pronounce it in a French-sounding fashion when its meaning would be more effectively conveyed by a received English pronunciation). Agence-ment connotes ‘agency’ and that is the crucial distinction from the more general synonyms, arrangement, apparatus, assembly and assemblage. As Callon explains, agencement is ‘necessary to understand distributed and strategically oriented collective action and the role of the material environment’ (367). This has been made clear enough in the past through repeated and compelling elaborations of the necessity of a word that signals that action is produced by the manner in which human, material and non-material things are arranged (Calıskan and Callon 2010; Callon 2007). So far, so good.

My concern is that I still can’t, after all this attention, arrive at a clear distinction between market agencement and market device. Martha Poon’s meticulous editorial translation flags the difficulty, well, of translation (427–428). Her notes explain that the terms agencement and dispositif (usually rendered as device in English) have both a quasi-technical nature and methodological implications that distinguish them from their synonyms. In the section ‘Why the term market ‘agencements’ and not ‘dispositifs’? Callon tackles this head on with a conceptual disintrication [there it is] of the trajectory of dispositif such that it eventually ‘implies that the human and non-human components can be separated’ (358). At length,

[t]he dispositif’s main function is to dispose of man. Seen from this perspective, it shapes man’s dispositions, which implies that the human and non-human components can be separated with the dispositif without challenging its internal economy. The dissymmetry introduced between different types of scientific knowledge is striking: nonhuman technosciences are limited to being resources that can be mobilized without interfering with human subjectivities whereas humanities play a unique role in shaping and enacting the latter. (358)

This is fair. Despite the prominence of material architectures in Foucauldian post-structuralism, there is always a lean to human practice first, evident in exemplary analyses of how institutions, knowledge and power operate in statistical, accounting and risk practices and in the life and psy sciences for example.¹ Paradoxically though, the anglophone translation of dispositif as device simultaneously privileges ‘the material and mechanical world of technical artefacts’ (359) in a sensibility that I have described before as ‘a bit too materially, too technologically, observant’ (McFall 2014, 17). This applies even to literature explicitly addressing market devices as agencements.² These tidal drifts in emphasis on how material, non-human and human elements connect, and with what consequences, as Callon is painstakingly aware, are exacerbated in translation. For all the care Callon takes in MITM to clarify why agencement should be preferred over device, it does not quite resolve the issue if for no other reason than that the words device and dispositif have sometimes been rendered, in English as well as French, as encompassing human and non-human elements generatively, if not purposefully, assembled (Deleuze 1991; Muniesa, Millo, and Callon 2007).

Does it matter whether market device and market agencement do the same conceptual job? Probably not very much. Both have their adherents and their critics and both have been subject to reframing designed to accommodate more dynamic, more empirically versatile, accounts out of a static sounding concept – hence ‘agencing’ (Cochoy, Trompette, and Araujo 2016; Guzman, Diedrich, and Cochoy 2023). I have also tried this by using ‘devising’ (2014) instead of device or agencement but with a slightly different target that traces back to the much older meanings of the word device in English. In its ‘darker English vernacular meaning’ (Erturk et al. 2013, 337) device can refer to a contrivance, a jack-in-the-box, which involves disguise, deception and opportunism. The phrase ‘devices and desires’ in the Anglican Book of Common Prayer – and in P.D. James’ eponymous ‘whydunnit’ novel – refers to a motive force, even a commercial mechanism like advertising, that drives conduct unawares (McFall 2004). It is this residual inexplicability, that the word ‘device’ conveys more clearly than ‘agencement.’ Not everything that happens in market settings can be explained and it is these darker, cloudier elements that tended to fade from view in The Laws and much of the literature that followed in its wake.

This is not something that MITM can be accused of. In numerous places throughout the book, Callon explicitly addresses the enigmatic core of markets. In Chapter 5, affectio mercatus is introduced as a new element in the market scene that lies;

at the heart of the market. It is the source that animates us as agents; it communicates energy and inclines us; it shapes our preferences, nourishes our expectations, stimulates our infatuations, and leads us to the point where some of us are ready to pay in order to attach ourselves to the goods we covet. (246)

The section that follows elaborates: the forces leading to consumer attachment are not general resources ‘pent up in some huge reservoir, waiting to be tapped.’ Instead need and desire exist only in relation to objects. Affectio mercatus is the mode of framing that shapes, interrupts, and channels market activities such that attachment between people and objects can result. This is the closest I have read to the formulation advanced in Hennion, Meadel and Bowker’s Artisans of Desire, viz. desire happens when we;

have in front of us not a strange object, but an object that already contains us since we have been incorporated in it by a thousand techniques from the moment of its production; and it is to be ourselves but the simple addition of the objects through which we are defined. The product traces out the consumers, the consumers the product: the familiarity of the couple has replaced the otherness of the confrontation between the reality of things and the illusions of desire. (1989, 208)

This passage was, for me anyway, the singlemost important inspiration for the arguments presented in Markets and the Arts of Attachment (Cochoy, Deville, and McFall 2017). Callon’s (2017) incisive contribution to the collection distinguished between three devices of attachment; listening, co-production, and addiction. This programmatic distinction is typical of Callon’s analytical interventions and it is offered with the qualification that in practice ‘actual attachment devices combine all three’ (Callon 2017, 183). That movement between an organised accounting of the various modes of framing that make markets and an accommodation of an enigmatic, messier in practice, residue is typical of MITM and the book is the better for it. And yet … 

Reading the book took me a long time. I knew I would review it so i was reading both to understand it and to find my position. Callon does not make that easy. There is so much there and such an elegant dance between descriptive engineering and enigmatic disclaimer that every time I thought I had something it was addressed a few pages later. This manoeuvring itself eventually became my frustration. Quite simply, I want more room for doubt and less room for explanation. Maybe after all that’s a culture thing, or even a cultural economy thing.

When Hennion (2017) writes that attachment cannot be registered through structures, determinations, causes or intentions and ‘does not belong to the vocabulary of action’ he tilts the concept just a little away from Callon’s original formulation in the early 1990s. Hennion’s notion of attachment explored through the (more cultural) matters of taste and addiction seems less amenable to accounting. It is true, he remarks, ‘that attachment has lacked the body, not in principle but due to the terrain and the objects of inquiry privileged by Callon, namely science, techniques and markets’ (2017, 115). I am not sure I can buy this. Did the socio-technical study of markets that Callon almost singlehandedly triggered have to disregard bodies? Did that enhance it? ANT had so many corrections to make that perhaps taking seriously the matter of calculation necessitated scrupulously clearing away the soft bedding of economic sociology. Certainly, through most of the 2000s and 2010s, despite the occasional supermarket trolley, it seemed that the most important markets were the abstracted ones where calculation and the performativity of formulae could be tested. To be clear, this is not a mistake that MITM makes, every kind of market is here, and every kind of emotion is channelled into chains of actions and reactions. To make affects observable;

requires specific equipment and protocols. Emotions, and this is essential for our purposes, are linked to devices in which they participate and outside of which they lose all meaning. Affects and their effects cannot be separated from the material and discursive contexts in which and through which they are expressed. (253)

This, a legacy perhaps of the ‘generalized symmetry’ between social actors and discourses and between human and nonhuman actors, (Blok and Jensen 2023, 17) that Callon with Latour so decisively introduced, seems to cover everything. In Blok and Jensen’s recent reassessment of Richard Rorty, they argue that his ‘postepistemological skepticism’ can help resist the ‘creeping objectivism’ of Latour’s later writing. While STS’s ‘philosophical wing’ is not my local there is something in their analysis that mirrors my mild frustration with the kind of over-engineering ‘markets in the making’ always returns to. It’s not possible to see all the joints that link emotions to devices to market attachment no matter how specialised the equipment. Callon, I suspect, both knows this and doesn't know it. It has not helped that the attention lavished on certain kinds of ‘big white men,’ markets, notably in ‘high’ finance, has not been matched by attention to markets driven by ‘small brown women’ (or dogs, or cats, or babies³). In a podcasted conversation between Richard Rorty, Isabelle Stengers and Donna Haraway at Stanford back in 2006 intended to nudge Rorty to take non-human conversation and action more seriously, Haraway remarked that she would do this by taking ‘the debased relationship of human beings and a 35 billion dollar a year global industry with their pet dogs’ seriously. She did not actually have much to say about this 35 billion dollar market either in the podcast or the chapter (2008) that followed, but her remark stands out because such markets are so seldom taken empirically seriously in the Callonian market studies field. If they had been the field just might have come quicker to researching affectio mercatus and the discursive, linguistic, aesthetic equipment that makes markets. While the role of ‘digital-affective’ technologies (Bassett et al. 2024), programmatic advertising, and style analytics might be making it clearer that the discursive too is accessible only through techno-material equipment there is nothing new in this. If acknowledging the fact of the consumer-product couple replaces the supposed confrontation between the objective reality of things and the illusion of desire – to paraphrase Hennion, Meadel, and Bowker (1989) – these 1000s of techniques are just as efficiently deployed in C19th insurance selling techniques as in behavioural design (Doyuran 2024), neuromarketing or sentiment analysis. For markets to even function as markets they must already have their people (McFall and Deville 2017).

This, in the end, is what inclines me to keep a bit of space for Michel Foucault and a little more for Walter Benjamin.⁴ Trying to write a history of advertising practices in the 1990s in a discipline that already knew exactly what it thought of this manipulative ‘sin industry’ full of tricks that had just got better and better, I leant heavily on Foucault’s non-progressive account of history. Insisting that ‘chance, discontinuity and materiality’ need to be thought through, Foucault advocated for a genealogy of events ‘outside of any monotonous finality; it must seek them in the most unpromising places, in what we tend to feel is without history – in sentiments, love, conscience, instincts’ (Foucault 1984, 76). For Rorty (1986, 47) these moves leave Foucault without a theory of anything because there is nothing that can ‘be accurately represented in thought.’ With no method, as Blok and Jensen (2023, 45) put it, ‘for accessing and “mirroring” the objective properties of the material world discourses and language games are on a par, epistemologically speaking.’

In many ways MITM is just not guilty of the ‘creeping objectivism’ Blok and Jensen (2023) find in Latour’s later writings. If anything, it enlarges the enquiry from that set out in The Laws and admits more stuff, more sentiment, more chance, more uncertainty into the making of markets. But the paradox is that there is an answer, a process, there always to explain how it all comes together when it does. The fragments of market talk that Walter Benjamin left behind leave no sign of such impulses. The virtuoso rebuttal of The Work of Art in the Age of Mechanical Reproduction in Hennion and Latour’s (2003) essay depicts Benjamin as a thinker with a comprehensive, if terribly mistaken, account of things that somehow misses the mechanical and technical character of art itself. But Benjamin does not always write that way. While I was ready to duck Benjamin for his production of culture sympathies in the 1990s, I have repeatedly slipped back to the fragments collected in One Way Street and the Arcades Project (1985; 1999). When Benjamin writes, however impressionistically, that sentimentality is restored in ‘the huge images across the walls of houses, where toothpaste and cosmetics lie handy for giants’ (1985, 89) he is outlining an orchestration, of technique and of sentiment, that is at the heart of how markets are. This is easier to see in toothpaste or Spotify marketing than in abstracted financial markets, but it is of course a feature of both. In my, quite possibly contrary, reading, Benjamin’s reasoning prefigures the relationship between description and action laid out in the concept of market agencement. If the Work of Art essay consistently mistakes the substance of artistic technique, Benjamin’s writings about markets and things, particularly in the unfinished Arcades Project, are full of inconsistencies and contradictory fragments. What emerges is a sense of markets as a material atmosphere, a magical mirror world, that presses on action and can never be satisfactorily expressed. Despite the necessity, the scope and the scale of the movement Callon provoked in how non-economists treat markets (though it remains fair to say economists themselves remain largely unscathed) I miss this in MITM. I originally planned to call this essay Hey, it’s me, I’m dynamite and I don’t know why after the Van Morrison (or the Waterboys) lyric. I’m not entirely sure why but that is the point. As much as possible we should try to know markets in their making but we won’t, not quite.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Notes

1 This is a vast literature eloquently represented in the early work Hacking (1990), Miller (2001); Ewald (1991) and Rose (1990).

2 See for example many of the contributions to Callon, Millo, and Muniesa (2007) and the considerable literature that followed it.

3 I’ve misappropriated Akrich’s (2023) remarks about CSI in the 1980s here but the point stands.

4 If not Roland Barthes who I keep, respectfully, on my window sill.