Assemblage theory

Jump to navigation Jump to search

in a multiplicity, what counts are not the terms or the elements, but what is ‘between’ them, the in-between, a set of relations that are inseparable from each other

— Gilles Deleuze & Claire Parnet, Dialogues, Columbia University Press, 1987.

From a pre-history in ideas around "desiring machines" in the work of Gilles Deleuze and Félix Guattari, agencement (commonly translated as assemblage) became a "provisional analytical tool" in A Thousand Plateaus (one of the more celebrated philosophical texts of the 1980s) before being appropriated diverse fields and being assimilated into disparate traditions in ways which have not always been seen as reflecting Deleuze and Guattari's distinctive contribution.


In 2015, Martin Müller suggested that "If language, representation and discourse were the pet concepts of the 1990s, assemblage, actor networks and materiality might well be those of the 2000s"[1]. Müller's focus was on connections with Bruno Latour's Actor Network Theory: both prominent in academic efforts to "decentre reified totalities" and he picks up on the way assemblage as a concept in Gilles Deleuze and Félix Guattari's A Thousand Plateaus (1980, 1987) was "foreshadowed in Anti-Oedipus and its focus on desiring machines."

As many commentators have noted, English language rendering of agencement as assemblage is deeply problematic. One issue, widely identified, is that agencement (the term being translated) is quite distinct, in French, from assemblage. The former stresses the agency involved in arranging a multiplicity of parts into a whole: the event-like characteristic of bringing-together. If we consider assemblage as a verb, rather than as a noun, or Deleuze and Guattari's preference for becoming over being, we might better appreciate this critical distinction (cf. Andrew Wilson on Verbing Nouns within Ecological Psychology[2] and Rita Felski on "Thinking about the humanities as a series of actions"[3]).

PS. See the Ian Buchanan section (below) for arguments that "the consensus understanding of the concept has been shaped as much (if not more than) by a plain language understanding of the English word ‘assemblage’ as it has by any deep understanding of the work of Deleuze and Guattari."[4]

Illustrative Examples of Assemblages

In A Thousand Plateaus (1980, 1987), Deleuze and Guattari discussed assorted assemblages. These included crystals, languages, sedimentary rock, anvils, and societies. Over the last four decades, many more illustrative examples have been produced.

Books as Assemblages

In A Thousand Plateaus (Deleuze and Guattari, 1987) we find discussion of how "a book is an assemblage of this kind, and as such is unattributable" - but Thomas Nail makes the example more accessible:

We cannot extract the being of a book from the vast historical conditions of the invention of an alphabetic language, distribution of paper, the printing press, literacy, and all the social contingencies that made possible the specific book of our inquiry, with all of its singular features (color, lighting, time of day, and so on) and all the conditions under which someone is inquiring into the book. A vast network of processes continues to shape the book and thus there is no final product. We do not know what the book might possibly become or what relations it may enter into, so we do not yet know its universal or essential features. We know only its collection of contingent features at a certain point in its incomplete process. As Deleuze says, “If one insists, the word ‘essence’ might be preserved, but only on condition of saying that the essence is precisely accident, the event” (Difference and Repetition 191).

— Thomas Nail, "What is an Assemblage?", SubStance, Volume 46, Number 1, 2017 (Issue 142), pp. 21-37 (Article)[5]

This ties in with Rita Felski's argument, explicitly linked back to Assemblage Theory through the work of Bruno Latour), in The Limits of Critique (2015) that "actors only become actors via their relations with other phenomena, as mediators and translators linked in extended constellations of cause and effect."[6]

Constellations of Stars as Assemblages

Without stars in the sky there are no relations between stars, but without relations between stars there is only radical heterogeneity. In this example, the abstract machine is the relational lines that connect the stars and the concrete assemblage is the stars that are connected. However, since new stars are born and old stars die and all of them move around in relation to our point of view, there is no eternal essence of the constellation. The constellation is a singular event: a set of relations that change as the elements change in a kind of reciprocal feedback loop.

— Thomas Nail, "What is an Assemblage?", SubStance, Volume 46, Number 1, 2017 (Issue 142), pp. 21-37 (Article)[7]

Lifeskills as Assemblages

Popular manifestations of athlete leadership in (modernist) sport include but are not limited to calming agitated teammates, acting as a liaison between athletes and coaches, mediating a conflict, and taking initiative to solve a problem […] actualisations always necessarily occur in unique ways and in fleeting moments […] the athlete may desire to mediate the conflict to improve team dynamics, foster a psychologically safe team environment, and allow all parties involved to express their concerns respectfully. The resonance between form of content and form of expression is the assemblage, coming together in a living purposeful arrangement through the deliberate realisation of a leadership plan that has performative and legitimated functions

— Martin Camiré, A move to rethink life skills as assemblages, [8]

In 2021, Martin Camiré produced a paper "to propose a move to rethink life skills as assemblages through a postqualitative inquiry, with the work of Deleuze and Guattari, Delanda, and Buchanan informing how the concept of assemblage is advanced." Echoing Nikolai Bernstein's line about repetition without repetition[9], he argues that "every enactment of leadership is an event that is relationally woven into reality in unique ways; there are no sequels, only the new."[10]

Source and Related (accessible) discussion:

Martin Müller on Assemblage Theory and Actor Network Theory (Bruno Latour)

Martin Müller is one of many who take Deleuze and Guattari as having used agencement as "a provisional analytical tool rather than a system of ideas geared towards an explanation that would make it a theory" - but outlines what might tentatively be taken as constituent features. Leadership is taken as a paradigmatic instance:

1 Assemblages are relational. They are arrangements of different entities linked together to form a new whole. The crucial thing to note here is that for Deleuze, assemblages consist of relations of exteriority. This means two things. First, it implies a certain autonomy of the terms (people, objects, etc.) from the relations between them. Second, ‘the properties of the component parts can never explain the relations which constitute a whole’ (DeLanda 2006, 10).

2 Assemblages are productive. They produce new territorial organisations, new behaviours, new expressions, new actors and new realities. This also means that they are not primarily mimetic; they are not a representation of the world.

3 Assemblages are heterogeneous. There are no assumptions as to what can be related – humans, animal, things and ideas – nor what is the dominant entity in an assemblage. As such, one can also say they are socio-material, eschewing the nature–culture divide (Bennett 2010).

4 Assemblages are caught up in a dynamic of deterritorialisation and reterritorialisation. Deterritorialisation/reterritorialisation is a central axis of an assemblage, where ‘reterritorialized sides, … stabilize it, and cutting edges of deterritorialization, … carry it away’ (Deleuze and Guattari 1987, 88).2 Assemblages establish territories as they emerge and hold together but also constantly mutate, transform and break up.

5 Assemblages are desired. ‘Desire constantly couples continuous flows and partial objects that are by nature fragmentary and fragmented’ (Deleuze and Guattari 1983, 6).3 Assemblages thus have a corporeal component.

— Martin Müller, Assemblages and Actor-networks: Rethinking Socio-material Power, Politics and Space, Geography Compass 9/1 (2015): 27–41, 10.1111/gec3.12192

Müller notes that for Deleuze and Guattari, the state "becomes an effect rather than the origin of power" He references Geographical work in which uses Assemblage Theory notions of deterritorialisation / reterritorialisation to focus upon "reconstructing the socio-material basis of its functioning" and which draws on Deleuze's framing of power in mapping the "establishment and exercise of power through the making of alliances that join entities together." In relation to materials, objects and technologies, the two theories have enabled researchers to "articulate a sensitivity to the material interventions of matter and the animal world in how agency and politics are constituted." His fourth framing is around topological space, where he draws on McFarlane to suggest an assemblage "is a more adequate concept to describe social movements than the dominant network metaphor, for assemblage retains a focus on emergence and allows the component parts to exceed the network."

Thomas Nail on Assemblage Theory

In 2017, Thomas Nail argued that although Deleuze and Guattari never formalised it, their work did reflect a fully fledged "general logic of assemblages." He builds from a keen appreciation that the French word agencement "does not simply entail heterogenous composition, but entails a constructive process that lays out a specific kind of arrangement" to suggest that whilst all assemblages may be singular and heterogenous, "they also share three features that define their arrangement: their conditions, their elements, and their agents" These are his take on Deleuze and Guattari's abstract machine, concrete assemblage and personae.

The abstract machine is abstract in the sense that it is not a thing, but it is absolutely real in the sense that the relations that arrange concrete elements are real. It is a machine in the sense that it is defined only by extrinsic relations and not intrinsic relations of organic unity […] In every case, the abstract machine functions as a kind of local condition of possibility—a set of relations in which elements appear to be meaningfully related […] designated by a proper name through which concrete objects and agencies speak and attribute their similarities and differences from each other.

concrete elements of an assemblage are the existing embodiment of the assemblage […] the relationship between the abstract relations and the concrete elements is not pre-constructed but has to be constructed piece by piece […] The concrete elements are like the archipelago or skeletal frame, whereas the conditioning relations are the breath that suffuses the separate parts.

Personae are not autonomous rational subjects, nor are they simply decentered or fragmented subjects incapable of action. Rather, the personae of an assemblage are the mobile operators that connect the concrete elements together according to their abstract relations. In other words, personae do not transcend the assemblage but are immanent to it […] The persona marks out the conditions under which each machine finds itself filled with concrete elements.

— Thomas Nail, "What is an Assemblage?", SubStance, Volume 46, Number 1, 2017 (Issue 142), pp. 21-37 (Article)[11]

Those exploring Substrate-Independence Theory might find resonance with Nail's summary: "There is no essence of the event; there are only concrete elements that are defined by their external relations, i.e., what they are concretely capable of at any given point. If we want to understand how an assemblage works, we do not ask what its essence is, but rather what it can do."

Nail goes to explain how, "For Deleuze and Guattari, there are four major kinds of assemblages: territorial, state, capitalist, and nomadic" and stresses that these are four major types or ways of arrangement "in which the conditions, elements, and agencies of different assemblages are laid out." This stresses the inherently (small-p) political nature of assemblages: these assemblages structure perceptual horizons -

Territorial assemblages divide the world into coded segments […] They express the pregiven, essential, and proper limits and usage of persons and objects in a given assemblage by explaining how the world is related to the past, to an inscription of memory.

State assemblages are arranged in such a way that the conditioning relations attempt to unify or totalize all the concrete elements and agencies […] the abstract machine attempts to cut itself off from and rise hierarchically above the concrete relations and personae of the assemblage.

In the capitalist assemblage, it is no longer the concrete elements that drive the process of progressive itinerant change (as in the territorial assemblage), nor the abstract machine that centralizes the control over the concrete elements (as in the statist assemblage), but the agent or persona that becomes disengaged from the assemblage and tries to force unqualified concrete elements into strictly quantitative relations.

the nomadic assemblage makes possible a truly unlimited qualitative transformation and expansion of the assemblage. Without the abstraction and dominance of any part of the assemblage, a truly reciprocal change occurs. Thus the nomadic assemblage does not simply affirm the chaos of heterogeneity or qualitative difference, it constructs a participatory arrangement in which all the elements of the assemblage enter into an open feedback loop in which the condition, elements, and agents all participate equally in the process of transformation.

— Thomas Nail, "What is an Assemblage?", SubStance, Volume 46, Number 1, 2017 (Issue 142), pp. 21-37 (Article)[12]

Nail argues that for Deleuze and Guattari, four "political types of change" might be at work in any given assemblage and highlights deterritorialization as their term for "the way in which assemblages continually transform and/or reproduce themselves" - and his paper ends with discussion of four potentially distinguishable deterritorialization processes: relative negative / positive and absolute negative / positive.

Manuel Delanda's "Neo-Assemblage” Theory

In a review of Manuel DeLanda’s A New Philosophy of Society: Assemblage Theory and Social Complexity (2006), Jason Read highlights how Deleuze and Guattari "constructed their own vocabulary and ontology to address social relations, rethinking society as made up of abstract machines, assemblages, strata, full bodies, and desiring machines: a vocabulary that has as its specific goal an attempt to avoid the pitfalls of holism and individualism." He argues that "rather than return to the sources of Deleuze (and Guattari’s) social ontology, DeLanda focuses on the central concept of assemblage, expanding it to encompass a general redefinition of society that encompasses the work of Fernand Braudel and Max Weber" to offer "not so much an interpretation of Deleuze as an application to the central problems of social ontology."

Reading Delandra's contributions against a backdrop of contemporary discourse haunted by what Derrida termed Spectres of Marx, Read presents the seminal 2006 text as "an attempt to either dispense with such totalities as ‘the state’ or ‘capitalism’ altogether, or to at least radically reconfigure our understanding of them such that we grasp them as nothing other than the more or less transient intersections of specific and local processes." He concludes that " DeLanda’s work, from A Thousand Years of Nonlinear History to A New Philosophy of Society, should be credited for shifting the focus away from the task of simply defining the dense network of neologisms that make up Deleuze and Guattari’s writing and turning attention towards the crucial philosophical and political positions that underlie them" but warns that Delanda was, in his eyes, "predisposed to overlook" Deleuze and Guattari's insights into how "totalities become real through the very way in which people subordinate their actions to them."[13]

For an accessible, up to date introduction to DeLanda's thinking in this area, see Assemblage Theory (Delanda 2016). Edinburgh University Press supply a free PDF of the Series Editor’s Preface by Graham Harman and Delanda's own Introduction.


DeLanda, Manuel. A new philosophy of society: Assemblage theory and social complexity. A&C Black, 2006.
DeLanda, Manuel. Assemblage theory. Edinburgh University Press, 2016.

Ian Buchanan Critique

One cannot help but wonder how different the uptake of Deleuze and Guattari’s concept of ‘agencement’ would be if it had not been translated as ‘assemblage’ and an alternate translation such as ‘arrangement’ (which is my preferred translation [Buchanan 2015: 383]) had become the standard? It may be that assemblage theory as we know it today would never have taken off

— Ian Buchanan, Assemblage Theory, or, the Future of an Illusion, [14]

Buchanan has been brutal in his assessment of much of what passes as "Assemblage Theory" in the social sciences. In Assemblage Theory, or, the Future of an Illusion, he argued that "detachment from Deleuze and Guattari’s work seems to compel a plain language approach which, to borrow a term from translation studies, puts them at the mercy of several ‘false friends’, that is, words that look like they should mean one thing but in fact mean something else."[15]

Among much else, Buchanan warns that "the assemblage itself is, by definition, self-sustaining: it requires labour to actualise it, to be sure, but its existence does not depend on that labour" and that "we need to be wary of assigning every local variety of an assemblage an independent identity distinct from the abstract assemblage."

1. The assemblage is not a thing in the world – it is assemblages that explain the existence of things in the world, not the other way round;

2. Assemblages are structured and structuring (not purely processual), that is one of their principal processes; 3. Assemblages have a logic, an operational sense if you will, that can be mapped – one always knows what is possible and what is not possible within a given assemblage;

4. Assemblages always strive to persist in their being […] they are subject to forces of change, but ultimately they would always prefer not to change

— Ian Buchanan, Assemblage Theory, or, the Future of an Illusion, [16]

These argument builds into consideration of policy as an assemblage:

If policy is to be understood as an assemblage, as I want to suggest it should be, then we have to first of all grasp that the assemblage is not a thing and it does not consist of things. I would even go so far as to say the assemblage does not have any content, it is a purely formal arrangement or ordering that functions as a mechanism of inclusion and exclusion (Deleuze and Guattari 1987: 45). It does not consist of relations; rather, it is a relation, but of a very particular type

— Ian Buchanan, Assemblage Theory, or, the Future of an Illusion, [17]

Buchanan is far more generous in relation to Latour, and discusses how "Latour’s insight that agency can and should be thought in narratological terms is helpful in deepening our understanding of Deleuze and Guattari’s concept of the assemblage." This reworking is marked as "helpful in deepening our understanding of Deleuze and Guattari’s concept of the assemblage" because it "cuts through the illusory veil of scientificity that has been wrapped around the concept of the assemblage by many of its erstwhile admirers."

A passing reference to "something else the new materialists and the realists neglect in their appropriation of Deleuze and Guattari" perhaps highlights Buchanan's wider target: the way the concept holds (retains) an irreducible tension between ‘nondiscursive multiplicities’ and ‘discursive multiplicities’ such that "one dimension does not map onto the other without remainder." This is the assemblage as the yoke that points to both non-discursive and discursive dimensions that "should bring about a new way of seeing something and not simply fix a label to something we think we already know about."

Tim Ingold Critique

Tim Ingold's reflections on his own flirtations with assemblage as a concept articulate a powerful critique. Commentators might question whether that critique is of Gilles Deleuze and Félix Guattari's "provisional analytical tool" or of what Assemblage Theory became in the hands of Latour, Delanda and others.

As a concept, the assemblage seems to provide a convenient escape from the classical alternatives of having to think of the group either as nothing more than an aggregate of discrete individuals or as a totality whose individual components are fully specified by the parts they play within the context of the whole. Yet just as much as the alternatives it displaces, assemblage-thinking rests on the principle of the blob. In place of five little blobs or one big blob, it gives us five blobs that have partially run into one another while yet retaining something of their individuality.3 But whether the parts add up to the whole or not, what is missing from the additive logic is the tension and friction that make it possible for persons and things to cling. There is no movement. In the assemblage, it is as though the dancers had turned to stone.

— Tim Ingold, "Line and Blob", Life of Lines[18]

In his On human correspondence paper, Ingold elaborates a little:

I am reluctant to refer to the gatherings of social life as assemblies, or as incidents of ‘assemblage’, as the word is commonly rendered through awkward translation from the French agencement […] some of the plethora of senses that have clustered around ‘assemblage’, as something like a bundling of life-lines reminiscent of sheaves of corn at harvest (Ingold 1993: 168-9), do approximate to what I have in mind with the principle of interstitial differentiation […] But others definitely do not. An example is Manuel DeLanda’s (2006) appropriation of the term to denote a transitory and contingent coming together of heterogeneous components that cohere only through an exterior contact or adhesion that leaves their inner natures more or less unaffected, and that can therefore be detached and reconfigured in other arrangements without loss.

Bruno Latour (2010) has advanced much the same idea in his manifesto for what he calls ‘compositionism’. The idea of composition, for Latour, ‘underlines that things have to be put together (Latin componere) while retaining their heterogeneity’. Bits and pieces that are ‘utterly heterogeneous’, as Latour admits, ‘will never make a whole, but at best a fragile, revisable and diverse composite material’ (Latour 2010: 473-4). For this reason the composition, in his terms, may indeed be as readily decomposed as composed. This cannot be said, however, of materials joined in sympathy, or of the lives they animate. Precisely because they go along together and because their continual regeneration is nourished and impelled by the memory of their affection, they cannot be parted without a sense of loss, if not grief. Something has to give from the inside; or in other words, it is necessary to forget.

An articulated structure, since it remembers nothing, has nothing to forget. But the knot remembers everything. Untying the knot, therefore, is not a disarticulation or decomposition. It does not break things into pieces. It is rather a casting off, whence lines once bound together go their separate ways. Consider, for example, how children in the human family outgrow their parents. In anthropological parlance the relation between parent and child is known as ‘filiation’, and it is typically drawn on genealogical charts as a straight line connecting two points, representing parent and child, respectively.

Doubtless influenced by this genealogical depiction, Deleuze and Guattari associate filiation with the figure of the tree, which they go on to contrast with their favoured figure of the rhizome. Whereas the tree stands for filiation, the rhizome, they say, is nothing but alliance. ‘The tree imposes the verb “to be”, but the fabric of the rhizome is the conjunction, “and . . . and . . . and . . . ”’ (Deleuze & Guattari 2004: 27). With respect, this is grossly unfair to living trees, which, unlike their diagrammatic counterparts, grow, branch, and swerve from within the midst of things every bit as much as do the tangling roots of the rhizome.

In real life, likewise, filiation is a process of becoming in the course of which, through ‘growing older together’ (Schutz 1962: 17), the child carries on the life of its parent while progressively differentiating its own life from that which engendered it. Indeed, filiation is not the connection of parent and child so much as the life of parent with child.

— Tim Ingold, "On human correspondence", Journal of the Royal Anthropological Institute (N.S.) 00, 1-19

See also Ingold's Essays on livelihood, dwelling and skill for argument that "if every organism is not so much a discrete entity as a node in a field of relationships, then we have to think in a new way not only about the interdependence of organisms and their environments but also about their evolution"[19]