Rendering assemblage dialectical

Introduction

In the twenty-first century, assemblage, as analytic, has achieved something akin to establishment status within anthropology. In the anglophone world, anthropologists encountered the term in the 1987 translation of Gilles Deleuze and Félix Guattari’s A Thousand Plateaus. It was, however, not until the early 2000s that assemblage theory truly gained disciplinary traction following its use and endorsement by prominent anthropologists like Tania Murray Li (2007), Aihwa Ong and Stephen Collier (2005), Paul Rabinow (2003) and Kathleen Stewart (2007). By then, the discipline had largely renounced overarching explanatory paradigms—varieties of functionalism and structuralism, most notably. Yet there remained no broadly accepted alternative for grasping the apparent order, however fleeting, encountered in complex social formations. It was in this role—as a ‘structure-like surrogate’ (Marcus and Saka, 2006)—that assemblage theory was made to serve.

For the most part, however, anthropologists employing assemblage as a conceptual frame for analysis have not fully engaged the philosophical premises underlying Deleuze and Guattari’s use of the term. Thus, in an early review of the literature, George Marcus and Erkan Saka (2006: 103) noted that anthropological ‘derivations of assemblage’ typically omitted ‘technical and formal analysis’ of how assemblage operates in Deleuze and Guattari’s writing. Of course, redeploying analytical concepts free of prior philosophical baggage may well enable more innovative analysis. But philosophical premises, even when unacknowledged, have a habit of clinging to the concepts they engender.

I therefore revisit in this article a particular premise in Deleuze and Guattari’s writing. Specifically, I consider their theorising of difference; I do this as a step towards rethinking assemblage theory more generally. In place of Deleuze and Guattari’s conception of discrete elements related only externally to each other within an assemblage, I propose a dialectical conception, whereby assemblage is taken to be a heterogeneous ensemble whose constituent elements also relate to each other internally—that is, co-constitutively. To arrive at this understanding, I consider the work of theorists who have pursued an ‘open’ approach to dialectical analysis. Although I give prominence to the arguments of Theodore Adorno, the approach I advocate aligns, as well, with that of Walter Benjamin, Ernst Bloch, Angela Davis, CLR James, and more recently George Ciccariello-Maher, among others. Reworked in this way, a dialectical notion of assemblage has implications for recent work on the anthropology of capitalism—a matter I consider in relation to several ethnographic examples drawn from my own and others’ research.

Assemblage as anthropological theory

Within anthropology, Aihwa Ong and Stephen Collier (2005) provided an early, and subsequently influential, framing of assemblage. This they did by theorising globalisation—itself an ensemble of diverse processes—as productive of heterogenous sociocultural phenomena stretched unevenly around the globe. As conceptual anchor for their broader analysis of globalisation, Ong and Collier (2005: 12) outline their use of assemblage as follows:

An assemblage is the product of multiple determinations that are not reducible to a single logic. The temporality of an assemblage is emergent. It does not always involve new forms but forms that are shifting, in formation, or at stake. As a composite concept, the term ‘global assemblage’ suggests inherent tensions: global implies broadly encompassing, seamless, and mobile; assemblage implies heterogeneous, contingent, unstable, and situated.

Such a dialectical tension is likewise evident, though again implicitly, in Tania Murray Li’s (2007) elaboration of the concept. For Li (2007: 264), assemblage as practice refers to ‘the hard work required to draw heterogeneous elements together, forge connections between them and sustain these connections in the face of tension’. The outcome is thus always a contingent, unstable arrangement. Turning her attention to community forest management in Sulawesi, Li (2007: 263) further disaggregates the labour of assemblage into six practices that, she argues, ‘are generic to any assemblage, whatever its specific contours’. These practices are, namely, forging alignments, rendering technical, authorising knowledge, managing failures, anti-politics, and reassembling. What this focus on practice illuminates is the negotiated character of every assemblage. Engaging more explicitly than do Ong and Collier with Deleuze and Guattari, Li (2007: 265) adds that a further strength of the latter duo’s framing of assemblage is their emphasis on a ‘diffusion of agency’—a feature that contributes to any assemblage’s inherent instability.

Overall, Ong and Collier, and Li have effectively raised the stature of assemblage theory within anthropology, thereby promoting ethnographic attention to the heterogeneity and instability of sociocultural formations, irrespective of any apparent order and durability. However, by not engaging the more problematic philosophical premises in Deleuze and Guattari’s writing—in particular, their repudiation of internal relations—these anthropological interventions made easier an uncritical disciplinary embrace of what Deleuze and Guattari had explicitly intended as an anti-dialectical analytic, the features of which I examine in the following section.

Revisiting assemblage’s Deleuzian roots

When theory travels, observed Edward Said (1983: 227), its adoption elsewhere is never straightforward; ideas in circulation always transform ‘to some extent’, while retaining traces of their historical contexts of origin. Assemblage theory, too, must be understood within its historical moment—not 1980, when A Thousand Plateaus was published, but Paris 1968.

Engaged directly and enthusiastically with the tumultuous events of that year, Deleuze and Guattari played influential roles enunciating the spirit of the moment—‘developing intellectual, therapeutic, playful and meaningful alternatives to what was seen [at the time] as the stressful, disciplined, structure-bound Western individual’ (Rio and Bertelsen, 2018: 11). By way of contrast, the French Communist Party held firm to a calcified Stalinist orthodoxy, and intervened in May 1968 to arrest the revolutionary impetus of French workers. Years later, Louis Althusser (2006: 253) would recollect that his own theoretical interventions at the time had aimed to dislodge, from within, the party’s adherence to official Soviet dialectical materialism, with all the mechanistic and teleological baggage that this ideology entailed.

This, then, was 1968. And it was this very year that Deleuze published Difference and Repetition, the text that would lay the conceptual groundwork for A Thousand Plateaus. With an intent akin to that of Althusser, Deleuze (1994: xvii) states early on that Difference and Repetition is suffused with a ‘generalized anti-Hegelianism’. Thus motivated, the book sets out Deleuze’s critique of dialectical logic, which he sees as forcing a stifling subsumption of the particular to the universal. Against Hegel’s negative conception of categories grasped in relation to what they are not—the dialectical ‘identity of identity and non-identity’—Deleuze proposes an affirmative ontology of difference, whose elements are irreducible singularities. And whereas the Hegelian philosophy of identity posits a correspondence between concept and object—thus not respecting the individuality of the latter—Deleuze (1994: 13) offers instead a philosophy of difference unmediated by conceptualisation. Difference, ergo, must be understood ‘in itself’—a Kantian noumenon free of conceptual mediation (Deleuze, 1994: 222).

And so, when we encounter assemblage in A Thousand Plateaus, it is as ‘an intermingling of bodies reacting to one another’ (Deleuze and Guattari, 1987: 97–98). Or as the authors later define it: ‘We will call an assemblage every constellation of singularities and traits deducted from the flow—selected, organised, stratified—in such a way as to converge … artificially or naturally’ (448). Putting aside the observation that Deleuze pursued an ‘anti-dialectical program’ (Ellrich, 1996: 474) in which difference supplants dialectics (Smith and Protevi, 2018), how, specifically, is assemblage theory here un-dialectical? After all, at least some interpreters of the Deleuzoguattarian oeuvre believe the pair did, in fact, employ a dialectic of sorts, albeit one free of contradiction (Lambert, 2002: 74). As a processual ontology, for instance, assemblage theory, like dialectics, offers a relational account of internal dynamism, whereby transformation occurs through an agentive reconfiguration of elements (Colebrook, 2002: xx, and passim). And latter-day Deleuzians, like Ian Buchanan (2015), have developed assemblage theory in what is an arguably dialectical direction.

When I speak, therefore, of Deleuze and Guattari’s conception of assemblage being un-dialectical, I refer specifically to the following premises: elements of an assemblage are connected by external, to the exclusion of internal, relations, and thus retain an existence autonomous of the relations in which they are embedded. Deleuze, writing with Claire Parnet (2002: 55), was explicit on this point: ‘relations are external to their terms … [and] a relation may change without the terms changing’. The logic here follows from Deleuze’s (1994: 42, 333, 338) prior repudiation of internal relations. His concern, as always, is to avoid any analysis that would lead to a synthesis of elements, which any theory of internal relations, he argues, inescapably does. Manuel Delanda, one of assemblage theory’s most prominent interpreters, clarifies this premise of external relations, or relations of exteriority as he calls them, by defining that which they are not: internal relations, or relations of interiority. ‘A relation of interiority’, explains Delanda (2016: 62),

is one in which the terms constitute each other by the very fact that they are related; or, to put it differently, one in which the very identity of the terms is constituted by their relation, so that the terms have no autonomous existence.

A crucial implication of the aforementioned premises is the displacement of contradiction by difference. In all of their individuality, elements of an assemblage are unambiguous—ontologically univocal, in other words, not equivocal (Deleuze and Guattari, 2004: 254). A curious implication here is that, while Deleuze rejects the identity of concept and object (a position consonant with the dialectical approach I present in the following section), he nonetheless affirms the identity of each element with itself. There is thus ‘identity of elements or components but no identity of compound substances’, which is how Deleuze and Guattari (2004: 51) present the matter in their favourable account of French naturalist Geoffroy Saint-Hillaire. Yet there is an irony here, whose reasoning Lutz Ellrich (1996: 484) traces as follows: ‘Differential play generates sameness-with-itself and thus reverts to exactly that from which it attempted to distance itself: identity’. Being univocal, elements are without internal contradiction; any ambivalence in their being is disavowed. It is this evacuation of contradiction that motivates Benjamin Noys’ (2010: 62) critique of Deleuzian ‘affirmationism’—a ‘positivisation of difference’ that sees things as they are as ‘a multiplicity that has no structure of antagonism’. This is a premise the politics of which Alberto Bonnet (2009: 49) lays bare: ‘liberalism is a ghost that persistently haunts Deleuze’s pages’.

Deleuze and Adorno in conversation

From the start, post-structuralism in France was marked by the absence of any serious engagement with the Frankfurt School. It was a shortcoming that Michel Foucault, for one, would later acknowledge—stating as much in a 1978 interview:

When I recognize all these merits of the Frankfurt School, I do so with the bad conscience of one who should have known them and studied them much earlier than was the case. Perhaps if I had read those works earlier on, I would have saved useful time, surely: I wouldn’t have needed to write some things and I would have avoided certain errors (Foucault, 1991: 119).

Following from these premises, Adorno opposed Hegel’s identity thesis not with hypostatised difference, as would Deleuze, but with non-identity. This is because philosophy has no direct access to noumenal elements unmediated by conceptualisation. Instead, for Adorno, difference as concept is always already abstracted—conceptually severed from its constitutive relations. Dialectical criticism must therefore proceed by exposing the contradictions of conceptual non-identity—revealing, in other words, the ‘constitutive character of the non-conceptual in the concept’ (Adorno, 1973: 12). This constitutive character is what Bertell Ollman (2003) recasts as ‘internal relations’, and what for Adorno was a ‘reciprocal self-production’ (2017: 215). We thus arrive here at the very relations of interiority that Deleuze rejected. It is, however, precisely by attending to such constitutive relations that Adorno militates against the fetishism of concepts, which are inescapably abstractions, irrespective of claims to non-conceptual immediacy.

What we have, therefore, are two quite different understandings of dialectics. The first—a ‘closed’ dialectic—operates as an overarching, expressive totality cited to deductively explain the particular. This is what Stalin enshrined as official dialectical materialism. And this is exactly what Adorno, Althusser, Delueze and so many besides them have rejected. Of course, there has long been another approach to dialectics that refuses overarching synthesis. This is an approach Ernst Bloch (1977: 38) called ‘polyphonous dialectics’, and Adorno labelled, first, ‘an open and fractured dialectic’ (2017: 95), and subsequently, ‘negative dialectics’ (1973). In line with this approach is also what George Ciccariello-Maher (2017: 1) calls decolonial dialectics—a dialectics that refuses ‘harmonious closure’ and attends instead to the ongoing dynamism of heterogeneous conflicting relations. Such a dialectics is no ‘airtight deductive system’ (Adorno, 1984: 163), but a logic that proceeds from the interdependence of elements, such that ‘one moment sheds light on the other’ (Adorno, 2008: xvii). Or as Borges (1998: 252), in this article’s epigraph, more poetically put it, ‘to say “the jaguar” is to say the jaguars that engendered it’.

That the idea of constitutive relations has been central to much anthropological theory is an observation I would not be alone in making. Here are Harri Englund and James Leach (2000: 229) commenting on a prominent trend in late twentieth-century ethnography, after having cited Ollman on internal relations: ‘Rather than being external to the person, social relationships are here imagined as internal to its being. The person is a composite, an embodiment of others’ contributions’. Notwithstanding the irreconcilability of theories of constitutive internal relations, and anti-dialectical theories of difference that repudiate such relations, anthropologists have at times brought both approaches to bear on a single analysis. Such is the case in a recent body of literature on the anthropology of capitalism, to which I now turn.

Capitalism as assemblage

First published in 1996, The End of Capitalism (As We Knew It), by geographers Julie Graham and Katherine Gibson, went on to influence an impressive line of scholarship on the anthropology of capitalism (see Bear et al., 2015). Writing under the singular pseudonym, JK Gibson-Graham, the pair pursued a poststructuralist deconstruction of what they saw as ‘totalizing’, ‘capitalocentric’ representations of the economy. Such representations, they argued, discursively subsumed heterogeneous economic relations and practices to an imagined capitalist system, blinding us thereby to the world’s many non-capitalist economic forms. Illustrative of such economic forms, the authors proposed, were self-employment, workers’ cooperatives, indentured servitude, and slavery (Gibson-Graham, 2006a: xiii). These forms of labour were non-capitalist by virtue of the authors having defined capitalism as a social relation ‘in which nonproducers appropriate surplus labour in value form from free wage laborers’ (Gibson-Graham, 2006a: xxiv). Going further, Gibson-Graham (2006a: 35) posited such ‘non-capitalist’ economic forms as ‘outside’ and ‘autonomous’ of capitalism. Capitalism, on the other hand, limited to the formal exploitation of free wage labour, was seen as constituted by its outsides—that is, the manifold relations in which it is embedded.

It was through their reading of Althusser, Ernesto Laclau, and Chantal Mouffe that Gibson-Graham came to this idea of capitalism’s constitutive outsides. The argument originates, however, with Rosa Luxemburg (2003), who employed it to conceptualise the dialectic propelling capitalist Europe’s imperialist expansion into non-capitalist colonies. That the argument is dialectical is by no means lost on Gibson-Graham (2006a: 13, 26–27, 161), who recognise dialectical non-identity as key to challenging the fetishism of capitalism as an abstract category—much as Karl Marx employed dialectical logic to critique the fetishism of commodities. In The End of Capitalism, however, ‘dialectical’ has the specific meaning of overdetermination, in the sense employed by Althusser, whom Gibson-Graham (2006a: 26–27) read as philosophically aligned with anti-dialectical Deleuze and Guattari. Indeed, the latter are foundational to The End of Capitalism. And in subsequent work, Gibson-Graham (2014) present their argument in Deleuze and Guattari’s preferred language—rephrasing ‘economic difference’ as an assemblage of economic forms. What we encounter in Gibson-Graham is therefore a curious pairing of constitutive relations and anti-dialectical difference.

To be clear, most scholars, even those otherwise sympathetic to Gibson-Graham’s project, do not subscribe the latter’s narrow definition of capitalism. Kalyan Sanyal (2007: 7–8), who takes Gibson-Graham’s critique as a point of departure, notes that capitalism is not coterminous with formal wage labour; he thus speaks, instead, of capital’s outside. Anna Tsing (2015: 65), despite aligning herself intellectually with Gibson-Graham, argues that what the latter called non-capitalist forms ‘are never fully shielded from capitalism’. It is, however, Sandro Mezzadra and Veronica Gago (2018: 584–586) who most effectively parse the distinction between outside capital and outside capitalism. To do this, they stress capital’s parasitic appropriation—ongoing in processes of primitive accumulation—of its non-commodified outsides, which cannot at present be conceived as spatially exterior to, or autonomous of, capitalist relations.

Capitalism, then, is more widely understood as an ensemble of social relations that are not reducible to free wage labour—the labour, that is, of individuals ‘free’ to enter into contract with the employer of their choice. The latter is but one form of exploitation, which in fact pre-dates capitalism (Banaji, 2010). Insofar as wage labour has capitalist content, this derives not from its form, but from the constitutive relations in which it is embedded. Gibson-Graham are therefore correct to critique the fetishism of capitalist labour as an abstract category. But their critique does not go far enough, for it leaves untouched the fetishism of ostensibly ‘non-capitalist’ economic forms. What defines capitalism is not labouring for a wage, or engaging in market transactions, but rather, as Li (2014) emphasises, the market compulsions to do so. Amid such compulsions, owners of capital have in diverse contexts sought profit through varied relations of exploitation—at times unfree, and at times unwaged. We see such arrangements in Lisa Lowe’s (2015) study of European liberalism’s illiberal foundations, wherein she documents the interpenetration of metropolitan free wage labour with colonial slavery and indentured servitude. What this heterogeneity challenges is thus not capitalism per se, but the axiomatic centrality once granted to industrial wage labour (Kasmir and Carbonella, 2014).

Refusal/recuperation as dialectic of capitalist assemblage

How, then, might we conceive of capitalism as a dialectical assemblage—bringing, in other words, multiple logics together with constitutive relations in a single analytic? A useful point of departure, I suggest, is Li’s (2007: 264) framing of assemblage as practice: ‘the hard work required to draw heterogeneous elements together, forge connections between them and sustain these connections in the face of tension’. This focus on practice allows for analysis of class struggle as a catalyst of capitalist development—displacing, thereby, analyses that would read capitalist development solely in terms of capital’s own logic.

From this perspective, capitalist development proceeds not from entrepreneurial innovation or corporate initiative, but from proletarian refusal. This is a line of argument I come to by way of Mario Tronti. Seeking to make sense of the emergent Fordism–Keynesianism of 1960s Italy, Tronti (1964) argued that workers’ struggles had pushed employers and the Italian state to institutionalise trade unionism as a means to facilitate industrial peace and thereby stabilise capital accumulation. In this way, institutionalisation aimed to transform what were at first subversive workers’ struggles into that which Max Gluckman (1954) called rituals of rebellion—ritualised hostility to centres of power, which, counterintuitively, serves to stabilise and reproduce the status quo. And so, like this, capitalist development proceeds by a dialectic of proletarian refusal and capitalist recuperation.

Tronti, however, never used the term recuperation. I get the word instead from Guy Debord (1967), who in expressly Hegelian language argued that capitalist society continuously neutralises radical ideas and practices by way of co-option; that which starts subversive becomes commodified and redeployed as mortar to the status quo. It was an idea Delueze (1971) was quick to incorporate into his own analysis, minus the Hegelian language: ‘when we say recuperate we mean: each time something seems to escape capitalism, seems to pass beneath its simili-codes; it reabsorbs all this, it adds one more axiom and the machine starts up again’.

I suggest, however, that returning to recuperation’s Hegelian roots could address a concern that Gibson-Graham (2006a: xxxi) raised regarding claims of capitalist co-option. What concerned the authors, in particular, was the politically demobilising implications of presuming that political alternatives were always already co-opted, or that co-option is ever complete and final. Hegel, I believe, provides a way to conceive of recuperation without presuming unambiguous continuity in the status quo, or complete extinction of that which is recuperated. For we can read recuperation as a moment of sublation (aufheben), the dialectical character of which Hegel (2010 [1812]: 81-82) elucidated by pointing to the contradictory meanings of the term in its original German:

What is sublated does not thereby turn into nothing … it still has in itself, therefore, the determinateness from which it derives. The German ‘aufheben’ (‘to sublate’ in English) has a twofold meaning in the language: it equally means ‘to keep’, ‘to preserve’, and ‘to cause to cease’, ‘to put an end to’. … Something is sublated only in so far as it has entered into unity with its opposite.

In order to elaborate this dialectic of capitalist transformation, I consider in the remainder of this article three examples (workers’ cooperatives, self-employment, and gift exchange) as found within the existing ethnographic literature. My aim is to show how seemingly non-capitalist social forms can never be adequately grasped ‘in themselves’, outside of the relations in which they are embedded. This is because these otherwise alternative social forms may very well, through a process of recuperation, become their opposite—a means of reproducing, rather than contesting, the status quo.

Re-embedding the workers’ co-op

Consider the workers’ cooperative—an economic form that is, argue Gibson-Graham (2006a: xiii), non-capitalist by definition. Given this non-capitalist form, the authors posit the establishment of cooperatives as an unambiguously post-capitalist project, exemplary of which they celebrate the ‘community economy’ of Spain’s Mondragón Cooperative Corporation (Gibson-Graham, 2006a: xxvi). In Gibson-Graham’s telling, Mondragón is an ethical project outside and autonomous of capitalism. From this perspective, what has determined Mondragón’s development since its founding in 1956 are its core tenets of solidarity and democracy, which have enabled the cooperative to pursue an ethical experiment challenging ‘present-day economic orthodoxy’ (Gibson-Graham, 2006b: 101–103).

But something happened to Mondragón by the late 1980s. Mirroring patterns of corporate restructuring globally, the cooperative corporation initiated ‘aggressive strategies to defend and increase its market position’, while ‘adapting to the international market’ (Gibson-Graham, 2006b: 121). This restructuring entailed mass employment of temporary contract workers excluded from cooperative membership, and an international expansion of production aimed at taking advantage of cheap labour in the global South (121). Consequently, by 2006, when Gibson-Graham published their study, half of Mondragón’s 68,260 employees laboured as a low-waged, precarious underclass, sweating to secure the corporation’s financial viability in a highly competitive global market (123). The impetus for Mondragón’s restructuring the authors trace to an isolated ethical failing, wherein the ‘apparent dictates of the market’ came to be erroneously seen as ‘economic imperatives’ (123). Proceeding in this way from a premise of Deleuzian difference, Gibson-Graham remain committed to grasping Mondragón ‘in itself’—an ethical project autonomous of any constitutive relations.

An initial anthropological critique of Gibson-Graham’s study would point to the authors’ reliance on a rather superficial literature review. Anthropologists wishing to interrogate Gibson-Graham’s portrayal of Mondragón would thus want to confer with more detailed ethnographic research grounded in long-term fieldwork. Fortunately, such research exists in the form of Sharryn Kasmir’s 1996 ethnography, The Myth of Mondragón (see also Kasmir, 2018).

Between Kasmir’s study and that of Gibson-Graham, several significant discrepancies are evident. First, whereas Gibson-Graham (2006a: xxvi) celebrate Mondragón as a progressive alternative to capitalism built in the face of ‘fascist persecution’, Kasmir (1996: 75) details the ways the cooperative system was ‘appropriated by and accommodated within a fascist discourse’. In 1965, for instance, Franco’s government awarded Mondragón founder Father Arizmendiarrieta the nation’s ‘Gold Medal for Merit in Work’. The move was a right-wing recuperation of cooperatives that saw worker–ownership as a means to divide and de-radicalise the workers’ movement (75). And while the Mondragón cooperatives were not simply tools of the fascist state, Arizmendiarrieta was from the start a Catholic reformer intent on promoting a socially conservative ideology wherein class struggle would be ‘overcome’ through labour-management collaboration (65–68).

Despite this ideology of classless production, Kasmir documents widespread class antagonism between workers and cooperative managers (Kasmir, 1996). Managers, for instance, pushed to widen the pay differential between themselves and workers, and to expand technocratic management of production in pursuit of ‘efficiency’ and market competitiveness (Kasmir, 1996: 168). As a result, ‘workers’ self-management’ devolved into rubber-stamping managers’ decisions (Kasmir, 1996: 137). Workers, in turn, complained of alienation, exploitation, and a sense of manipulation under the cooperative’s ideology of classlessness. Tellingly, one cooperative worker told Kasmir she felt exploited, ‘just like any worker in any firm’ (Kasmir, 1996: 122). More recent anthropological research on cooperatives documents similar sentiments. In Kathleen Millar’s 2018 (166) ethnography of informal waste collectors in Brazil, an informant stated that working for a waste collectors’ cooperative ‘was like any other wage labor job’. Illustrating such grievances, Kasmir (1996: 112) offers the case of a strike involving 700-plus workers at two Mondragón cooperatives in 1974, to which cooperative managers responded by firing 24 strike organisers, using to do so the cooperative’s anti-strike by-laws. Ironically, by glossing over the contradictions of class antagonism and presenting cooperatives as post-capitalist, Gibson-Graham’s position mirrors the historical claim of the Soviet government that class exploitation could not exist in the Soviet Union because Soviet workers were employed by a ‘workers’ state’. This was a claim the fallacy of which CLR James, Raya Dunayevskaya and Grace Lee Boggs (1950) exposed in their classic critique of the Soviet Union as an instance of state capitalism.

In contrast to Gibson-Graham’s ethical determinism, Kasmir (1996: 9) argues that class dynamics and internal restructuring within Mondragón cannot be effectively understood when dis-embedded from the cooperative’s ‘social and political milieu’. Significantly, this critique was also voiced locally regarding Mondragón’s flexibilisation beginning in the mid-1970s. According to one Basque political group, ‘the increasing integration of the cooperatives into the world market obliged co-ops to increase the number of contract laborers [and] produce according to profit motives rather than principles of social necessity and to reduce wages while work pace increased’ (Kasmir, 1996: 116). Given how closely post-1973 labour flexibilisation in Mondragón mirrors that of private firms globally, it seems disingenuous to analytically reduce these dynamics to an endogenous ethical failing. Ultimately, Mondragón cannot be understood as an instance of economic difference ‘in itself’—autonomous, that is, of its constitutive context. The alternative is certainly not to replace ethical determinism with a mechanistic political economy—to put ‘the object on the orphaned royal throne once occupied by the subject’, as Adorno (1973: 181) quipped—but rather to employ dialectical analysis ‘to abolish the hierarchy’.

How to be your own tyrannical boss

By now, the post-1973 flexibilisation of labour that Kasmir documented in Mondragón has become a staple topic of research across the social sciences, a critical policy concern for governments, and a central activist focus globally. A prominent line of analysis sees these developments—variously involving casualisation, informalisation, sub-contracting, and increased employment of migrants and workers hired through labour agencies—as the outcome of post-Fordist or post-socialist market liberalisation. Offering a Deleuzian reading of such restructuring, Brian Massumi (1992: 202–204) argues that, as a result, ‘class no longer exists’ as a relation of value extraction, and thus, within the capital-labour relation, ‘contradiction has been abolished’.

Bracketing for a moment the effects of late twentieth-century capitalist restructuring, David Harvey (1990: 147) sees the process as flowing from a capitalist logic, such that flexibilisation was a capitalist response to the ‘rigidities’ of unionised and legally protected employment under Fordist labour arrangements. By contrast, analyses informed by Mario Tronti and CLR James have highlighted how workers themselves militated against the stifling rigidities of Fordist production—a refusal of work expressed in absenteeism, sabotage, and wildcat strikes, which exposed unresolved contradictions in the Fordist social compact (e.g. Glaberman, 2002). My suggestion is that neither perspectives is adequate alone, as each points to one side of a dialectical relation; flexibilisation here was an unstable capitalist recuperation of proletarian refusal, as well as an aggressive assault on prior working-class gains.

We encounter such a dialectic in recent anthropological writing on the informal economy. Exemplary in this regard is Kathleen Millar’s study of informal waste collectors in Rio de Janeiro (Millar, 2018). Millar’s argument is that the individuals engaged in this labour—precarious though it is—have not been simply excluded from the formal economy. Rather, their decision to take up informal waste collection expresses a refusal of the workplace discipline found in ‘normative forms of capitalist labour’ (Millar, 2018: 92). In my own research with informal waste collectors in Thailand, individuals I spoke with expressed similar sentiments, while also valuing this work as an ‘honest living’ in contrast to illicit activities like theft (Campbell, 2018b: 276). Critically, however, in the context I researched, the Thai government promoted independent waste recycling amid broader economic restructuring as a means to privatise and subsidise the country’s waste disposal sector. Here, then, in the capitalist transformation of waste disposal, everyday forms of proletarian refusal were recuperated in the government’s own marketisation agenda, resulting in a contradictory unity of opposites. Informal waste collectors pursued a form of self-employment that was subsumed to large-scale, private recycling firms. Millar (2018: 8) makes a related point: if informal waste collectors are imagined as ‘superfluous to capitalist accumulation, then it becomes impossible to ask how the materials they collect are tied into a 200-billion-dollar global recycling industry’.

Cases such as these of informal self-employment subsumed to formal capitalist enterprises challenge Gibson-Graham’s claim that self-employment is non-capitalist by definition. The reason is that self-employment has become a widely adopted strategy of labour flexibilisation, whereby capitalist enterprises free themselves from the constraints of labour protection legislation, while offloading the risks of market fluctuations onto self-employed contractors, freelancers, gig workers and the like. For this reason, self-employment cannot be adequately understood independent of the constitutive relations in which it is embedded.

On the one hand, outsourcing production to self-employed workers puts the labour process outside of direct capitalist management, limiting the extent to which managers are able to regiment and mechanise the production process. Self-employed workers can therefore obtain some measure of control over the time and pace of work—a level of autonomy that attracted Millar’s (2014) informants to informal waste collection. At the same time, however, endeavouring to get by in a situation of market dependence, self-employed workers confront enduring pressures to produce commodities whose value can only be realised on the market. Consequently, and despite the contemporary celebration of entrepreneurial self-employment, there have been documented trends of heightened psychological distress, as flexibilised workers put increased pressure on themselves to be more productive amid intensified competition (Molé, 2013). The result here is a contradictory dynamic akin to Chayanovian self-exploitation, wherein the precariously self-employed worker becomes her own tyrannical boss, enforcing labour discipline on herself in the absence of any direct capitalist manager (see Day, 2018).

Gift exchange in the gears of capitalism

In the disciplinary revival of moral economy analysis, anthropologists have understood the concept as synonymous with Karl Polanyi’s notion of embeddedness, and Marcel Mauss’ understanding of gift-giving. That is to say, anthropologists have posited moral economy, the embedded economy, and gift economies as positive alternatives to—as outside, autonomous, or interstitial to—the depersonalised market rationale that presumably characterises capitalism (e.g. Gregory, 1997). Yet, as recent critical interventions make clear, such a distinction creates a false dichotomy, for all economies, capitalist or otherwise, are moral economies—inevitably shaped by specific value judgements and ideological claims (Carrier, 2018; Palomera and Vetta, 2016: 419). The irony is that, by spatially distinguishing moral economy from political economy, such analyses end up reifying the economistic political economy they otherwise aim to critique.

Consider workplace cooperation and gift exchange. Pursing a Maussian analysis informed by Gibson-Graham, David Graeber (2014) highlights the significant role of workplace cooperation in the day-to-day operations of capitalist firms. Identifying such cooperative relations as ‘everyday communism’, Graeber (2014: 68) believes he has uncovered a ‘scandal’, whereby capitalist firms have come to depend—despite themselves—on a non-capitalist, even anti-capitalist, mode of operation. But scandal, I would argue, it is not. As recent studies in the anthropology of labour have made clear, workplace cooperation has become a key strategy of flexible capitalist production. Such is the case, to take one example, with the cooperative production model pursued by the US-based Saturn automotive company, a subsidiary of General Motors (GM). Rather than distinct from a capitalist logic, Kasmir (2014) documents the ways Saturn’s managers strategically promoted ‘cooperative’ labour relations as a material/ideological project of flexible production—a project aimed at dividing and undermining the unionised workforce at GM’s other plants.

It is also no coincidence that the journal Hau—foundationally Maussian in name—pursued a collaborative production model (Graeber and Da Col, 2011), which facilitated the exploitation of precarious junior academics while obscuring the exploitative character of these relations (Kalb, 2018; Neveling, 2018). That ideological claims of cooperation are commonly employed to legitimate exploitative relations is, of course, no novel insight. Feminists in the 1970s, for instance, employed this critique in their demand of wages for housework. Whereas bourgeois ideology had, they argued, framed the home as a non-capitalist sphere organised according to love and familial obligation, this ideology had in fact served to mobilise women’s unwaged domestic labour while masking its exploitative character. To this, Sylvia Federici (2012: 28) added that the exploitation of unwaged ‘housewives’ had been all the more effective than waged exploitation ‘because the lack of a wage hid it … where women are concerned, their labour appears to be a personal service outside of capital’.

But a capitalism parasitical on non-commodified relations is not free of contradiction. This is clear from ethnographic research on, for example, gift exchange among factory workers. In his study of a diamond-cutting factory in Tamil Nadu, Jamie Cross (2012) found that gift exchange among workers engendered the cooperative relations on which factory production—and the workers’ exploitation—depended. And yet, workers may nonetheless engage in gift exchange as a refusal of the alienation experienced in factory production. Such was the case in my own research with Myanmar migrants employed at garment factories in Thailand (Campbell, 2016: 262). Moreover, in the context that I researched, cooperation and gift-giving among factory workers established the social relations and solidary necessary for labour strikes and other forms of collective action. In such cases, cooperative relations among workers remain, in their implications, contradictory, and thus indeterminate—equivocal, that is to say, not univocal. They are neither simply manipulated, nor necessarily subversive. What we therefore need in order to analytically grasp this ambiguity is a dialectical critique that refuses ‘the tendency to think in simple alternatives’, as Adorno (2017: 185), for one, long advocated.

Conclusion

Against the backdrop of Paris 1968, Gilles Delueze—with Félix Guattari, as well as alone—made important intellectual interventions against the mechanistic and teleological dialectical materialism prevalent in France at the time, particularly within the (Stalinist) French Communist Party. To do this, the authorial pair repudiated the totalising closed-system character of such a dialectic, which left no conceptual space for heterogeneity or agency. And to do that, the pair reframed complex formations, social or otherwise, as assemblages—arrangements of autonomous elements interacting externally, but free of constitutive internal relations. Their reasoning was that any theory that accepted relations of interiority would result in a synthesis of elements, and from there fall into conceptual homogenisation and totalisation. However, by construing differences as autonomous singularities related only externally, the result was a reified notion of difference free of contradiction and ambiguity. That is to say, identity—the correspondence of form and content, concept and object—was not removed, but instead retained under the name of difference.

Anthropologists, such as Tania Li and Aihwa Ong, have, to their credit, productively employed assemblage theory to attend ethnographically to the heterogeneity of logics and diffusion of agency evident in social formations, such that sociocultural phenomena cannot be deduced from a singular systems analysis in the manner of classical structural functionalism. However, the premise of Deleuzian difference underpinning assemblage theory has, among its advocates in the discipline, remained uncontested. For the anthropology of capitalism, the result has been a romanticised notion of economic difference outside and autonomous of capitalist relations. Such a framing has inhibited sober analysis of the ambiguity inherent in what are seemingly progressive economic alternatives, such as cooperatives, self-employment, and gift exchange, which hold a particular fascination for anthropologists within a context of deindustrialisation and Fordism’s decline. By contrast, an open dialectical analysis of such economic forms enables critique of their recuperation into capitalist relations, without such a critique reducing these economic forms to the functional imperatives of an all-encompassing capitalist system.

I have therefore sought in this article to push assemblage theory towards a fuller critique of identity, which persists in Deleuze and Guattari’s logic under the guise of difference. What an engagement with Adorno shows is that a critique of totalising identity theory does not necessitate a hypostatisation of difference; nor does recognition of co-constitutive relations require a synthesis of elements. An open dialectics escapes such false dichotomies, allowing multiple logics to be read as internally related without this becoming a unified system-logic. Difference, it needs restating, pace Deleuze, cannot be adequately grasped ‘in itself’, conceptually severed from its constitutive relations.

A properly relational—indeed, holistic—anthropology can thus not proceed from a logic that treats sociocultural phenomena as unambiguous singularities (Kalb, 2015). This, I have argued, is evident in the anthropological study of contemporary capitalism. When situated within their social and political milieu, the heterogeneous labour arrangements that have proliferated since the end of last century show themselves to be more than autonomous ethical experiments. Yet they are also not straightforward products of a functionalist capitalist logic, for they have variously developed out of the conflicting relations of historically particular actors. If, however, assemblage theory cannot be rendered sufficiently dialectical so as to account for such relational phenomena, then the theory’s continued relevance for our discipline is suspect; anthropologists would have little reason to retain the analytic as a theoretical tool in our disciplinary repertoire.