Social Structure : Aaron Berman
|
Social Structure: Aaron Berman
1. Introduction Much contemporary political practice is self-consciously oriented toward the defense, modification, or abolition of social structures. To the extent that this statement holds true, it is unsurprising that structure ought to be counted as a live political concept. There are, however, a plurality of competing conceptions of social structure in circulation and my interest here is first in critically taxonomizing them in light of the divergent strategic orientations toward the practice of politics that they warrant and then in sketching a materialist model of emergent social structure. I am motivated by the worry that despite the incorporation of ‘structural’ analyses into both popular and academic critical discourse, the prevailing conceptions of social structure are inadequate to the task of fixing the subject matter of our most pressing contemporary concerns. In other words, a polyvocal critical discourse remains stalled in preliminary disagreements over how to characterize the objects of social critique, leaving us unable to advance toward more meaningful disagreement over how to address the interrelated structural crises of capitalism, care, global ecology, white supremacy, democratic legitimacy, and the like. This taxonomic inquiry is ontological in nature, not because I take politics to follow from ontology, but because political strategies—from winning hearts and minds to seizing the means of production or eliminating the opponent’s capacity to resist—are always, implicitly or explicitly, formulated, communicated, and enacted on the basis of decisions about which features of social reality are politically relevant.1 I am inspired in this endeavor by Douglas Porpora’s 1989 essay, “Four Concepts of Social Structure,” from which I borrow the taxonomizing approach and to which I am indebted in many ways, as will become clear to the reader.2 However, unlike Porpora, my primary focus here is on theories of structures as kinds of social entities, rather than on the roles that structures understood as theoretical entities may play in explanations of facts or events. Further, the taxonomy I present in what follows has a critical edge insofar as it is guided not only by ontological concerns, but also by pragmatic materialist political considerations. I will propose minimal criticisms of each model of structure that I present—excluding my own preferred alternative—which should be understood as tendential criticisms, in the Jamesian spirit of judging theoretical “trees” by their “fruits.”3 In other words, I am more concerned with the effects that various theories of structure have had on downstream scholarship and practice than I am with whatever philosophical resources a theory or family of theories has available to absorb or obviate critique. Finally, the model I offer is developed with a view to addressing practical deadlocks that I take earlier theories to have sustained or worsened. 2. Desiderata for a Realist Conception of Structure I am skeptical, for reasons that may become clear, of the value of etymologies and historical semantic investigations, but in the case of structure there are a few salient sedimented meanings that can orient the discussion. The first of these is that structure is an architectural term. Jean-Loup Bourget’s entry on structure in the Dictionary of Untranslatables gives the following definition: “[structure] designates the skeleton or armature, as opposed to the form or outer appearance.”4 He goes on to suggest that structure, by contrast with the related concepts of pattern and gestalt, carries with it an opposition between surface and depth and that this opposition has rendered the concept particularly appealing to scholars in the humanities who wish to “scientifize” their practice.5 It seems that “structure” carries this “scientifizing” sense because it suggests an explanatory gap between surface and depth that recalls Marx’s quip that “all science would be superfluous if the form of appearance of things directly coincided with their essence.”6 Structure, in this architectural sense, particularly appeals to those who hold rationalist and foundationalist epistemological commitments to respect the asymmetrical order of causes and effects and the parallel asymmetry of explanation. Knowledge of structures promises to found explanations of phenomena because structure itself “holds up” or grounds the reality of appearances. In the context of social theory, structuralist approaches tend to figure social objectivity in ways that brush against the grain of first-personal experience, suggesting that our lives in common are counter-intuitively determined by processes that go on “behind our backs,” to borrow another Marxian metaphor.7 In keeping with this, structure tends to guide those who use it toward some form or another of realism. In a memorable passage of Plato’s Phaedrus that has come to serve as a canonical reference for ontological realists, Socrates deploys the structural anatomical metaphor of the skeleton that “holds up” the organism when he evokes the figure of a philosophical butcher who knows how to “cut up each kind according to its species along its natural joints, and to try not to splinter any part as a bad butcher might do.”8 For contemporary realists, on Sally Haslanger’s gloss, “some properties are more important than others in structuring the world, and it’s not up to us, so to speak, which these are.”9 Whether ancient or modern, structural realism suggests that to classify adequately is to trace the articulated structure of reality that underlies and supports phenomena, rather than to haphazardly group together that which merely appears together. This basic semantic analysis suggests a first set of provisional desiderata for a theory of social structure. Social structures ought to be universals or real kinds whose concepts pick out a virtual multiplicity of realizations, rather than only concrete singulars. In other words, structures ought to stand in “one-to-many” relations with the social formations that realize them, such that we can draw inferences about societies that realize a given structure based on knowledge of that structure derived from the study of another realization of it. Social structural knowledge ought to be useful beyond concrete singular cases. Secondly, social structures ought to be understood as exhibiting distinct causal powers that can play a role in explaining social phenomena. Structures ought not be causally and explanatorily redundant—their invocation in explanations should be consequential and not merely an expression of a shorthand. A second semantic note will suggest two further desiderata and will help to guide the taxonomy I will present. Structure carries with it the idea of both elements and relations. A structure is not just one thing, but multiple things set into some kind of configurational relations. In this regard it is related to—yet not identical with—form. The difference is that a single object or event can have a form, a determination that renders it intelligible, without having a multiplicity of sub-components that are arranged into this form, while structure involves the unification, organization, or co-ordination of different elements or terms, in some way or another. From an erotetic perspective, the form of something typically responds to a “what is” question—as when “dog” responds to the question “what is that animal?”—while the structure of something answers a “why” question about its determinations—as when the structure of the English language explains why “outside right we” is not a valid construction. By contrast with a “formal” analysis, a structural analysis is one in which a certain understanding of the causal and explanatory contribution of terms and their relations is at stake. While structure and form are not abstractly identical, considered singly, a given structure may nevertheless have a form that identifies it (e.g., English as opposed to Arabic). Social structures differ from other kinds of structures (such as mathematical, chemical, or anatomical structures) in that they are historical and artefactual entities whose persistence depends upon human activity.10 A structural theory is properly social only if it represents its elements, relations, or both as inherently involving temporally extended human participation. Most (although not all) models of structure involve explanatory reduction according to which phenomena that are structured are said to be explainable in terms of some aspects of their underlying structure such that non-structural properties of these phenomena can be disregarded. Above, I took this to be one of the desiderata for a theory of social structure—knowledge of a given structure ought to license non-trivial explanatory inferences about diverse social phenomena said to be structured in that way. There is a great deal of variation, however, in how theorists of structure mark explanatory priority according to their differing perspectives on the ontology of structure. Consequently, I will use the question of reduction to distinguish three distinct approaches to the ontology of structure: those that reduce structures to their elements or to one essential element, those that reduce structures to the relations that obtain between structured elements, and those that seek to theorize structure as a phenomenon that constitutively depends on both properties of elements and of relations. Explanatory reduction has political stakes for theories of social structure precisely because it shapes our perspective on the possibility and viability of various modes of strategic intervention. At the most extreme end of the elemental reductionist spectrum about social structure we find a still popular view: methodological individualism. Methodological individualism breaks down into two frequently conflated views. The first is explanatory individualism about social structure, the view that social phenomena must be exclusively explained in terms of the actions, beliefs, or properties of individuals; the second is ontological individualism about social structure, the view that social structures are composed of nothing but individuals (and their subcomponents, such as actions, beliefs, and properties).11 I am here concerned with the latter view as it is both more fundamental and more widely accepted. Ontological individualism follows Aristotle in denying the status of being to relations—relations are ways of speaking about properties of individuals, not types of things in the world.12 Structures cannot be unities of relations and terms for the ontological individualist. Rather, as Porpora presents the individualist view, structures are merely “patterns of aggregate behavior that are stable over time.”13 Consequently, even naming these patterns as patterns involves an intellectual accomplishment of abstraction from the basic social reality of individuals and their behaviors. As the individualist sociologist Randall Collins put it:
Microbehaviours can be treated as explanatory fundaments, but this commits social scientists to an extreme nominalist empiricism according to which behavior must be singularly catalogued before any explanation of structure, qua abstract pattern of aggregate microbehaviours, can get off the ground. If microbehaviours are instead taken to be explananda rather than fundamental explanans, as they are in the paradigm case of explanatory individualism in micro-economics, microbehaviours are said to be explained by collections of “preferences” held by individuals. In this case, the task of the social scientist is to produce abstract models—usually in the idiom of game theory—of behavior on the basis of given preferences. Individualism ultimately entails the dissolution of the concept of structure, as the latter becomes either a misleading shorthand for some additive empirical collection of repeated individual actions or a hypostatization of a game-theoretic scenario.15 It is well known that the individualist dissolution of structure tends to be mobilized in the service of political reformism if not outright pessimistic conservatism.16 Theoretically, ontological individualism denies the existence of social groups and thus the fundaments for group-based political strategies—which, if they are to be used at all, have to be redescribed as circumstantial convergences of individual interests. Because ontological individualism reduces the existence of social structures to patterns of individual behavior, to the extent that it will acknowledge phenomena like “structural” racism and gender oppression, it will suggest that these phenomena are ultimately grounded in individual beliefs and argue that strategies to reform or abolish these structures will have to work through either rational persuasion or (in the increasingly popular cognitivist mode) implicit bias training. Methodological individualism, while the most prominent, is not the only elemental reductionist view on social structure available. What is shared across reductionist views of this kind is the ontological priority of elements and their properties to relations. One could readily be, for example, a biological elemental reductionist about social structure without being an ontological individualist if one affirms that what determines the structured character of social phenomenon is the actualization of a human supra-individual species essence. On this hypothetical view, social life would be structured in virtue of some inherent properties or dispositions of humanity, such as a species instinct to build or pursue scientific knowledge that is held to be responsible for the phenomenal organization of collective life. Construing the problem of reduction according to the difference between elemental and relational reductionisms about social structure allows us to reconsider the model that Louis Althusser had polemically labelled “expressive totality,” which is typically seen as a form of holism absolutely opposed to methodological individualism, but that I take to be another variant of elemental reduction.17 Expressive totality, in the context of social structure, is the thought that structures have a central part that determines both their states as wholes and the states of all of their other parts.18 On this way of thinking structure, all aspects of a structured whole reflect or express some property, or—in the more nuanced version—some internal relation (like a contradiction) of a central element of the social. This is less strict a reductionism than ontological individualism in that, in this case, there is at least one configurational relation that is real, namely the unidirectional relation of expression that obtains between the central “essence” of the social and its multifarious appearances. All other apparent social relations between non-central elements are epiphenomenal echoes of the determinate expression relation. Althusser takes a certain common, albeit problematic, interpretation of Marx’s presentation of historical materialism in the widely read Preface to the 1859 Contribution to the Critique of Political Economy as a paradigm case of expressive totality.19 Here, a society’s superstructure is determined by its base, the base, in turn, is the unity of the forces and relations of production, and this contradictory unity is itself an expression of the internal dynamism of the developing forces of production. On this technological determinist interpretation of historical materialism, social structures organizing practices from artistic production to the family can ultimately be explained with reference to the state of technological development of the forces of production. This reduction of multiple relations to one unidirectional relation should also be considered a form of elemental reduction. Although many expressive totality views take the central element in a structure to be a dynamic internal relation rather than a static object, I do not take this to be taxonomically relevant here.20 What matters from the point of view of the concept of social structure is that this central internal relation counts as a central element whose state expressively and unidirectionally determines the states of all others such that relations between non-central elements are disregarded. Criticisms of the political use of expressive totality are well known. Indeed, they are the bread and butter of dismissive post-war treatments of Marxian thought. This view of structure lends itself to a hermeneutically myopic politics, dismissive of the ersatz and the accidental that merely adorn the development of the fundamental element. While it can lend itself to a determinist quietism warranted by prognostication about the development of the central contradiction, it is equally available to the most voluntaristic of knowing militants. 4. Reduction to Relations Another general framework for conceptualizing social structure sets up the explanatory reduction in the opposite direction, rendering relations primary and elements either secondary or unreal. I will discuss two divergent cases. A first paradigm case can be identified with the positivist social scientific search for lawlike regularities that obtain between social facts. Here, social structure can only be explained in terms of social facts, not in terms of individual facts. Structure names a complex of social laws—on the model of natural laws—that are robustly indifferent to individual psychological factors. Porpora, again, neatly glosses the paradigmatic Durkheimian example:
Reduction to lawlike relations is difficult to square with conceptions of human freedom, which is why it was so stridently challenged in anglophone sociology during the “structure vs agency” debates. In a similar vein, critical theorists have typically taken the positivist conception of structure to reify the social.22 This conception of structure can be politically invoked as support for the ideological naturalization of oppressive social relations, but it need not necessarily be. The clear strategic partner of positivistic holism is technocratic governance, which trades in the increasingly unpopular and inflated currency of trust in the counter-intuitive explanatory and instrumental successes of the natural sciences. A second case takes relations to be ontologically prior to their relata without thereby taking relations to be laws. Post-Saussurean structuralism and contemporary anglophone ontic structural realism both take structure to refer to complexes of relations such that structures can be individuated according only to their characteristic relations. A slogan of Karen Barad’s sums up the common core of relational reductionism well: “relata do not precede relations.”23 On this view, relations are figured as ontologically prior to, and constitutive of, their terms. In its strongest version, Steven French’s fundamental metaphysical structural realism, relational reductionism holds, contra the substantialist ontological mainstream, that the basic metaphysical building blocks of reality are not objects and properties, but relations.24 It is helpful in thinking about these views to distinguish between two aspects of the ‘terms’ or ‘elements’ of a structure as I have been discussing them: we can talk about a term qua object and a term qua placeholder. For example, we might consider members of a family as individuals or as occupiers of relational places—fathers, daughters, aunts, and so on. The relational reductionist view takes places, qua aspects of relations, to be ontologically (rather than merely analytically) distinguishable from the objects that occupy them as placeholders, and further, takes the structurally relevant features of objects to be constituted by their occupation of a relational place. The French structuralist tradition that derives from Lévi-Strauss’ adoption of Saussure’s linguistics is the most well-known exemplar of the relational reductionist conception of structure in social theory.25 Recall that for Saussure the phonemic terms of a language are constituted by the system of oppositional relations that they enter into with one another. A given language is just a virtual system of differences that couples oppositions between terms and oppositions between ideas. In principle, if we were to replace every phoneme in English with a different one the language would remain the same so long as the relations between phonemes remains the same. The intrinsic sonic properties of a phoneme are only relevant to the extent that they serve as bearers of negative differential relations. Hence the phoneme /b/ can be said in many voices while retaining its identity, but it loses this identity when it passes into the sonic territory covered by the opposed phoneme /p/. Here we have to acknowledge the conceptual movement of “post”-structuralism, which departs from the recognition that a system made up only of negative oppositional relations is entirely indeterminate. The various streams of post-structuralism represent alternative resolutions to this problem. For example, Deleuze posits a metaphysically fundamental pure difference, Derrida concludes that all determination takes the form of deferral, Lacan thematizes the interplay between the “Real” body of the drives and the differential symbolic order, and so on. For present purposes, these solutions are less relevant than the initial conception of structure as nothing but relations of co-constitutive negative difference that motivate them. What happens when the philosophical reflection on Saussurean structuralism returns to social theory? Given the significant challenge of establishing criteria for the non-metaphorical transposition of the Saussurean model beyond its linguistic context of origin, the most straightforward path is toward a metaphysically general linguistic idealism for which social structure is a set of symbolic oppositions organized around a paradoxical “empty square” signifier that is self-referential and consequently indefinitely contestable.26 Hence, the social appears in Laclau and Mouffe, for example, as a series of chains of equivalence that meet at intrinsically paradoxical or nonsensical nodal points (e.g., the populist signifier of the ‘people’).27 The causal and explanatory significance of the material properties of social elements is sidelined—despite protestations to the contrary—as the social is assimilated to language, which is more readily accounted for by the model. This point is best illustrated pragmatically by the fruits of the post-Saussurean tradition: what features of social reality are actually thematized and what features are neglected in works that follow from this paradigm?28 Underlying both the internal debates within post-structuralism and Marxian critiques of the tradition is a veiled reappearance of the medieval scholastic quarrel over the status of universals. If one is a relational reductionist, it is natural to also embrace a Platonist ante rem view of universals. Structure, like Platonic forms, possesses a mode of existence that is prior to and independent of its instantiations.29 When it comes to social theory, however, we are typically trying to understand structures that historically emerge and are much harder to make sense of outside of their instantiations. For these purposes, a background Aristotelianism about universals according to which universals exist only as instantiated in particulars and metaphysically depend on the existence of their particular instantiators is more attractive.30 Another route, neither Platonic nor Aristotelean, is followed by Althusser’s posthumously published writings, which sketch a “theory of the encounter” as a view that couples a nominalist rejection of universals with relational reductionism. Althusser’s concept of contingently formed immanent structure departs from a heterodox reading of Lucretius according to which encounters between elements are retroactively constitutive of the elements themselves.31 In the context of social structure, this yields a view of structures as consequences of contingent encounters between heterogeneously developed historical elements, where encounters generate novel formal relations between elements that in turn reconstitute these elements. Developing on a brilliant observation made by Étienne Balibar in his contribution to Reading Capital, Althusser recognizes that three key elements of the structure of capital—accumulated money capital, “doubly free” commodified labor power, and industrial mechanisms—are “not contemporary products of one and the same situation”; rather, “each of these elements has its own ‘history,’ or its own genealogy.”32 Althusser offers an interpretation of Marx’s treatment of primitive accumulation (or original expropriation) in these terms, as the encounter between the relatively independent genealogies of “doubly free” proletarians—free from bondage to any particular firm and free from the ownership of means of production—and amassed capital in the money form is a contingent one, but once it occurs it “takes hold” in the form of a durable and self-reproducing structure. Although here elements have a causally relevant pre-structured existence, the structurally relevant features of elements are nevertheless their relationally constituted properties, not their intrinsic properties. Thus, within a structured situation: “an effect B (considered as an element) is not the effect of a cause A (another element), but is instead the effect of element A insofar as this element A is inserted into relations that constitute the structure in which A is situated [and caught up].”33 In his later work, Althusser increasingly prefers to speak of structural relations as singular, a move that undermines the explanatory ambitions of structural social theory insofar as knowledge of singularities is inferentially null. On this view there is, strictly speaking, no such thing as, for example, “the wage relation”—there are only concrete singular wage relations. A nominalist theory of singular relations cannot furnish an account of the characteristic unity of a social structure like capitalism across the different actual overdetermined social relations that contingently take hold and reproduce themselves over time and space. Instead of retaining a realist conception of the basic structure of capital, Althusser turns to a Spinozist conception of the “singular essences” of compositional entities in order to figure the always exceptional “conjuncture” as the proper object of historical materialism.34 This theoretical gesture proliferates an infinite continuum of conjunctural social formations and modes of production in the place of the structure of capitalism—a consequence that I fear produces far more confusion than it is worth.35 While the late Althusser’s “aleatory materialist” concept of social structure does non-reductively register the causal contributions of both elements and relations in the constitution of structures, it ultimately gives way to a fully immanent concept of social structure that licenses no distinctions between relevant and irrelevant relations and is consequently a descriptive rather than an explanatory concept of structure. While the exposition of a dialectical tension between singular description (or “conjunctural determination”) and structural explanation is taken by Panagiotis Sotiris to be the ultimate lesson of Althusser’s career in social theory and a horizon for us, I remain optimistic that learning from Althusser does not condemn one to accepting a dualistic oscillation between realism and nominalism in the face of the real complexity of political life.36 5. Toward a Non-Reductive View: Emergent Practical Materialism The risks of elemental and relational reductionism in theorizing social structure have not gone unnoticed. Both practice theorists and critical realists in sociology have sought to develop co-constitutive models of social structure that recognize the causal contributions of both elements (typically individuals) and relations while retaining the core insights that the social world is structured and that inferentially valuable explanations of social life can be given in terms of structures and their effects. The aim of such models is to reduce social complexity to its structural determinants while nevertheless maintaining a non-reductive view of structures themselves. One of the most well-known attempts at co-constitutive concept social structure is Anthony Giddens’ structuration theory. For Giddens, the elements of structures are individuals endowed with agency and resources that individuals organize their behavior around. Structure as such refers to complexes of rules that govern the way that individuals interact with resources. Giddens gives the following definition of structure: “Rules and resources, recursively implicated in the reproduction of social systems. Structure exists only as memory traces, the organic basis of human knowledgeability, and as instantiated in action.”37 A benefit of Giddens’ view is that it generates a dynamic concept of structure that is neither transcendent to social practice nor dissolved into a mere shorthand for an empirical patterning. Structures are taken to be “dual” in that they are “both the medium and the outcome of the practices which constitute social systems.”38 Structure, in this view, is changeable: it is a site of contestation in a way that affirms political aspirations to target, reform, or abolish social structures as such, rather than to simply change the individuals who partially constitute them. Further, it is a concept of structure that can be used to identify constraints as well as enabling affordances on practice and thus can play a straightforward role in forming political strategy. However for Giddens, to the extent that structures exist, they exist only in the form of the ideal rules that are instantiated in practices or, recursively, in systems of diachronically connected practices.39 While Giddens’ proposition of pairing rules and resources suggests a non-reductive conception of structure, he ends up locating what is “structural” —qua organizing —about structures in the ideal (rules governing practical consciousness) and is thus vulnerable to collapse back into a form of mentalist ontological individualism about structure.40 William Sewell picks up the pieces of Giddens’ theory of structure and bolsters it with an account of the materiality of structure that acknowledges the role played by intrinsic properties of actual resources in constituting, constraining, and enabling structuration. On his view, structures are mutually sustaining unities of ideal, virtual, transposable schemas and actual, material or immaterial, resources.41 The causal powers of social structures are in principle dependent on both relations and elements, both human and non-human. Sally Haslanger, finally, takes up Sewell’s modification of Giddens’s views to produce not a theory of structure as such, but a theory of social practices.42 For her, social practices are unities of mutually active and co-constituting resources and schemas. Resources are constituted and unified qua resources (and negatively or positively valued) within the process of a given practice by one or more embodied transposable social schemas, which are themselves learned objective complexes of public social meanings.43 To use a Marxian example, “[o]nly when a tribe hunts, does a particular region of the earth become a hunting ground.”44 The practice of hunting constitutes the resource “hunting ground.” Reciprocally, schemas are themselves constituted, individuated, and in some cases modified over time in response to both material features of the resources that they organize and the properties of other social practices that overlap with or otherwise bear upon them—for example, the practice of hunting transforms in response to regulations on hunting seasons, to the introduction of new technologies like night-vision, or to transformations in animal migratory or feeding patterns. Following Sewell, this view recognizes the role that resources play in both causally and constitutively constraining schemas, in both reflectively available and subdoxastic ways, according to their mind-independent material properties. Mutually sustaining and conflicting practices can be organized into structures with a degree of causal and ontological autonomy (as, for example, when practices of commodity exchange and commodity production come to both constrain and sustain each other in the transition to capitalism, generating unexpected effects that in turn transform both production and exchange in line with novel structural imperatives to competitively economize). Structures are causally potent insofar as they may constrain or enable the development of their constitutive component practices such that the practice-structure relation, like the resource-schema relation, is dynamic. On this view, social structures involve multiple practices, in turn composed of multiple resources and schemas that are mutually constraining and enabling, often in ways that exceed the consciousness of actors engaged in them. There is no mereological constraint on ‘leveled’ causality, such that, for example, a structure may be causally responsive to transformations amongst the practices that compose it or to transformations involving a resource that is coordinated by one of its composite practices.45 Correspondingly, practices, resources, and schemas, can play roles in multiple structures simultaneously. On this view, what is “structuring” about structure is no longer derivative only of either properties of elements pertaining to structures or the set of relations that obtain between those elements. While structures are the abstract configurational forms of related social practices, they are not entirely indifferent to the practices that instantiate them, such that the “object” aspect of structural elements cannot be reduced to the “place” aspect. While they have ideal and virtual components—schemas—these are constitutively constrained by their relations to both resources and other schemas, and that constraint is actual. In other words, the specific indifference of a structure to its material instantiations is a matter of degree and determining the degrees of freedom relative to a structural-relational ‘place’ or ‘node’ is an in-principle empirical task. It cannot be assumed a priori that just any instantiation of an abstract relational form can be actual. Correlatively, changing a structure may require changing not only meanings, but also material or immaterial resources. Haslanger’s concept of social structure consequently suggests an “infrastructural” practice of politics, attuned to the semiotic and material interdependencies responsible for the reproduction of “order” in social life.46 Might Haslangerian structures not, nevertheless, be reduced to being mere names for aggregate properties of their concrete particular instantiations? Does her view, like that of the late Althusser, establish a descriptively compelling model of social structure at the cost of the concept’s explanatory power? Otherwise put, what secures structures’ individuation and explanatory transposability across concrete spatio-temporal situations if structures are constituted by practically united virtual schemas and actual resources? To what extent is a structure more than its singular parts? I propose that the contemporary practical materialist concept of structure can be strengthened in such a way as to circumvent the kind of worries I have raised in relation to Haslanger’s otherwise powerful view, by appealing to Jessica Wilson’s contributions to the metaphysics of emergence. Wilson’s work is complex and developed against a background thicket of distinctions that I will risk ignoring for the sake of brevity, but its core contribution to the matter at hand can be developed intuitively. The aim of the concept of metaphysical emergence in general is to provide a non-dualist account of the apparently novel causal powers exhibited by composed entities in excess of the powers of their components. The class of emergent phenomena is characterized by “cotemporal material dependence” on a micro-level “dependence base” coupled with “ontological and causal autonomy” from that base.47 In our case, social structures can be figured as emergent from the dependence base of social practices. Earlier emergentists from the British nineteenth-century school around J.S. Mill onwards had taken the causal and ontological distinctiveness of emergent entities to be fundamental such that, for example, when the brain is in a functioning state the complex pattern of its physical parts gives rise to a new fundamental and irreducible ontological unit of mind with distinctive novel causal powers not possessed by any of its material constituents individually or in aggregate. However, over the course of the 1990s, Jaegwon Kim developed a persuasive and influential criticism of emergence, primarily as it was being deployed in the philosophy of mind.48 Kim had argued that emergence, as a “halfway house” between dualism and reductive physicalism, is incompatible with any form of physicalism—and thus that the very project of an emergence-based “non-reductive physicalism” is incoherent—because a key component of physicalism is the “causal closure” of the physical, that is, the view that all causality is exhaustively physical causality. It would seem that any macro-level causal autonomy will either have to be non-physical and thus in violation of physical causal closure, or it will be problematically overdetermined causality insofar as it will have to be taken as an unnecessary repetition of fundamental physical causality that should be singly sufficient to cause the effect at issue. Against Kim’s skeptical criticisms, Wilson argues that there is a coherent and ontologically parsimonious form of emergence—what she terms “weak emergence”—that avoids the problem of causal overdetermination.
Her philosophical intervention is to be found in condition (ii), which is the way that she secures the causal autonomy of emergent entities. By contrast with the “strong” conception of emergence endorsed by the British emergentists and criticized by Kim, Wilson’s view takes the causal autonomy or “distinctive efficacy” of emergent entities to derive from their realization of a constrained subset of the causal powers of their lower-level constituents. This constrained subset has a “causal power profile” that is distinct from the causal powers of 1) the aggregate of micro-level entities’ causal powers and 2) the causal powers of any individual micro-level entity or subgroup of individual entities. While the “strong emergentist” contends that emergent entities have more causal powers than their dependence base, a Wilsonian weak emergentist will argue that emergent entities are characterized by the fact that they have fewer causal powers than their dependence base. They are nevertheless ontologically distinct from their base by virtue of their characteristic subset of causal powers and causally distinct by virtue of the distinct contours of their “power profile” vis a vis the “power profile” of their aggregated dependence base. To illustrate this, imagine a case where we have three related practices: buying, selling, and growing wheat. Each practice has three causal powers (represented by letters): Buying (a,b,c), Selling (d,e,f), and Growing (x,y,z). These practices are configurationally related in some way such that a fourth emergent entity, call it “the wheat economy,” with the constrained causal power profile (a,e,x) exists. The emergent profile (a,e,x) is “proportional” to some effects (such as supply-side inflationary price trends) that are not proportional to any of the individual base-level practices (buying, selling, and growing wheat), nor to the aggregate of their causal powers (a,b,c,d,e,f,x,y,z). Here, causal proportionality refers to the distinctive efficacy of some set of powers to produce, without under or overdetermination, some set of effects. A second case that draws out how the subset of powers view can render “downward” structural causality intelligible may bring the point home: “in a recession, your boss caused you to be fired” can be correctly expressed as “in a recession, capital caused you to be fired.” This is a case of downward causality insofar as your boss’s management practice is simultaneously 1) a part of the dependence base of capital and 2) determined by the expression of capital’s distinctive causal powers. In this case, your boss’s management practice is structurally constrained by its embedding within the structure of capital. Capital, in turn, under conditions of recession, expresses its distinctive causal power in such a way as to determine your boss’s practice in a way that is “proportional” or causally relevant to producing the effect of you being fired as an austerity response. The subset of powers model allows us to articulate classical structuralist claims about the asymmetrical causal relevance of positions over occupiers of those positions—a kinder boss would have done the same—while simultaneously accounting for the material dependence of structures on practices and allowing for an excess of causal powers borne by occupiers of structural roles beyond their structurally relevant powers. Conceiving of practically emergent structures according to Wilson’s model admits of multiple realizability without vitiating the causal contributions of structural elements. A social structure may be an ontologically and causally distinct emergent entity that is constitutively dependent on a set of practices and can be individuated by the proper distinct subset of those practices’ causal powers, which the structure possesses in virtue of its configurational form. Nevertheless, a given social structure may be indifferently realized by a multiplicity of concrete practices (themselves unities of schemas and resources), provided that they are configurationally related so as to jointly exhibit that structure’s distinctive causal power profile. To return to our first example, it may be that there is an identical emergent structure that depends on and affects the buying, selling, and growing of some other agricultural commodity which shares some relevant intrinsic causal powers with wheat. Thus, contra the nominalism of the Althusserian theory of the encounter, the composite emergent practical materialist view of structure I have sketched here makes it plausible to assert simultaneously that social structures: 1) are universals or real kinds whose concepts pick out a virtual multiplicity of realizations, rather than only concrete singulars; 2) exhibit distinct causal powers that can play a role in explaining social phenomena; 3) are historical and artefactual entities that depend on human activity; and 4) may be the products of contingent compositions of distinct social practices and be neither teleological nor functional. 6. Conclusion I want to conclude with an appeal to the political importance of conceiving of structure both non-reductively and non-idealistically. The various forms of reduction I have sketched carry very different political consequences. Methodological individualism suggests that political action ought to be directed toward changing the minds of individuals, who are the exclusive authors of the social; expressive totality directs attention toward a single privileged aspect of the social, which is taken to be the key to the transformation of the whole; lawlike holism suggests a form of quietism, as the social obeys laws of its own, indifferent to human intervention, and structuralist relational reductionism, depending on the form it takes, either collapses into a transcendent version of expressive totality or an absolute nominalism, which avoids transcendence by entirely abandoning the difference between surface and depth that gives semantic coherence to the concept of structure in the first place. A fully immanent concept of structure that licenses no action-guiding distinctions between relevant and irrelevant relations and element is, in other words, no longer a concept of structure that is relevant to the strategic practice of politics. The limitations of reductive conceptions notwithstanding, there are surprising challenges that complicate attempts to generate non-reductive concepts of structure. It does not seem prima facie very difficult to conceive of structure as mutually constituted by elements and relations. However, attempting to do so coherently and systematically requires overcoming a persistent Western philosophical metaphor that identifies form with mind. It is puzzling to think about human affairs as formal and structured without for that matter being ideal. I take this to be the paradox at the heart of Marx’s “Theses on Feuerbach” and thus at the modern headwaters of “materialism” as a current in social theory.50 The gesture of Sewell and Haslanger to register the active structuring power of mind-independent properties of the object (in Marxian language) or resource (in the language of contemporary practice theory) pole of social practices constitutes genuine progress along the road to a materialist conception of structure that conceives of “human activity itself as objective activity” without thereby determining in advance the historical limits of social objectivity.51 Further, to the extent that post-Sewellian practice-theoretic conceptions of social structure are attentive to the structuring powers of resources, they are uniquely suited to addressing the anchoring of our most pressing social crises in socio-natural and material artefactual components of social structure (like the infrastructures of segregation, technologies and geographies of mineral extraction, workplace and urban surveillance technologies, and so on). A non-reductive and materialist conception of social structure is, consequently, a bare minimum requirement for the work of critique in the twenty-first century. Conceiving of social structure as an emergent phenomenon, dependent on interrelated social practices themselves constitutively constrained and potentiated by both their co-constitutive semiotic and material components not only affords ecological and infrastructural critique, but also accords with an experimental practice of politics without guarantee.52 Structures, as I propose that we conceive of them, need not be rigid and can be multiply realized to differing (in principle empirically determinable) degrees, but they do not float transcendently above the fray of social conflict and natural history. Rather, they are historical artefacts that objectively endure so long as the practices that jointly sustain them continue. We may, in principle, obtain knowledge of a structure such that we are able to better discern which material and semiotic factors in the jumbled world of our social practices are “skeletal” with respect to that structure, a knowledge essential to carrying out the pressing work of transformation or demolition. Social structural knowledge cannot, however, tell us much of use in advance about what grows in ruins. Published on August 15, 2025 * Aaron Berman * 1. I use “ontology” in the weak sense, indicating an inquiry into existence and its modes. My concern is with what analytic philosophers tend to refer to as “ontological commitments” as I am asking after the claims that various theories of structure, implicitly or explicitly, make about what kinds of things exist. This should not be confused with understandings of ontology that take the subdiscipline to be concerned exclusively with the substantial or the fundamental, nor with the Heideggerian “ontological difference” between “being” and the so-called “ontic.”↩ 2. Douglas Porpora, “Four Concepts of Social Structure,” Journal for the Theory of Social Behavior 19 (1989):195–211.↩ 3. William James, Varieties of Religious Experience (Oxford: Oxford University Press, 2012), 24.↩ 4. Jean-Loup Bourget, “Structure,” in Dictionary of Untranslatables: A Philosophical Lexicon, ed. Barbarra Cassin (Princeton: Princeton University Press, 2014), 1066.↩ 5. Bourget, “Structure,” 1066.↩ 6. Karl Marx, Capital, vol. 3 (London: Penguin Classics, 1991), 956.↩ 7. Karl Marx, Capital, vol. 1 (New York: Penguin, 1976), 135. On the theme of structuralism’s tension with phenomenology in the twentieth-century French context, see Knox Peden’s Spinoza contra Phenomenology (Standford: Standford University Press, 2014).↩ 8. On the allegory of the philosophical butcher, see Plato, Phaedrus, in Complete Works, ed. John Cooper (Indianapolis: Hackett, 1997), 265e.↩ 9. Sally Haslanger, Resisting Reality (Oxford: Oxford University Press, 2012), 156. The link between structure and metaphysical realism is deep and becomes increasingly pronounced in response to the challenge of idealism. For an early realist response to Berkeley that locates the realist/anti-realist différend in the mind-independence of the relational structure of reality, see Mary Shepherd’s Essays on the Perception of an External Universe (Oxford: Oxford University Press, 2020).↩ 10. Bracketing the sociality of some non-human animals, which I do not dispute.↩ 11. For an overview of the distinction and the decline in plausibility and popularity of explanatory individualism, see Julie Zahle and Finn Collin, introduction to Rethinking the Individualism-Holism Debate (London: Springer, 2014). For a strong critique of ontological individualism, see Brian Epstein, “Ontological Individualism Reconsidered,” Synthese 166:1 (2009): 187–213.↩ 12. In Categories 7, Aristotle engages in an ambivalent and unsettled discussion of the criterion according to which relative statements can be properly made. At issue is not the reality of relations themselves, but whether or not there are any substances that are categorially relative (like parts of an organic whole), or whose mode of being is, in a categorical sense, being-in-relation-to. For Aristotle, it is clear that talk of relations as ontologically independent of, or even prior to, their terms is absurd; what is debatable is whether or not terms have an absolute ontological priority vis-à-vis their contrary relata—can a head exist as a head without a body? A relation cannot precede its terms, but it may either succeed them or co-originate with them (see Aristotle, Collected Works, vol. 1, ed. Jonathan Barnes [Princeton: Princeton University Press, 1984], 6b1–8b24).↩ 13. Douglas Porpora, “Four Concepts of Social Structure,” 195.↩ 14. Randall Collins, “The Microfoundations of Macrosociology,” American Journal of Sociology 86 (1981): 994, 989.↩ 15. Douglas Porpora, “Four Concepts of Social Structure,” 197.↩ 16. Adam Przeworski’s impressive work in the tradition may well still represent its leftmost edge (see Adam Przeworski, “Material Interests, Class Compromise, and the Transition to Socialism,” Politics and Society 10 [1990]: 125–153).↩ 17. Louis Althusser, “The Errors of Classical Economics,” in Louis Althusser, Étienne Balibar, Pierre Macherey, Roger Establet, Jacques Rancière, Reading Capital (New York: Verso 2015), 237–267.↩ 18. Strictly speaking, Althusser identifies three theses of expressive totality: a mereological thesis, a causal thesis, and a temporal thesis. For the sake of brevity, I am conflating these three moments under mereological terminology.↩ 19. Karl Marx, Marx and Engels Collected Works, vol. 30 (London: Lawrence and Wishart, 2010), 263. See G.A. Cohen’s Karl Marx’s Theory of History (Princeton: Princeton University Press, 2000) for a reconstructive defense of the “orthodox” interpretation of this passage as the seed of a theory of historical determination according to the progress of technological development.↩ 20. Althusser illustrates the distinction by way of a contrast between Montesquieu’s The Spirit of Laws (Cambridge: Cambridge University Press, 1989) and Hegel’s Philosophy of Right (Cambridge: Cambridge University Press, 1991), which he takes to be structurally analogous in spite of Montesquieu’s preference for ‘principles’ and Hegel’s emphasis on ‘contradictions.’ See Chapter 3 of Panagiotis Sotiris, A Philosophy for Communism: Rethinking Althusser (Leiden: Brill, 2020), chapter 3, on Althusser’s interpretation of Montesquieu. Perhaps the clearest case of a static structural expressivism can be found in Leibniz’s Monadology (Edinburgh: Edinburgh University Press, 2014).↩ 21. Douglas Porpora, “Four Concepts of Social Structure,” 198.↩ 22. See, e.g., Theodor W. Adorno, “On the Logic of the Social Sciences,” in The Positivist Dispute in German Sociology, trans. Glyn Adey and David Frisby (London: Heinemann, 1976), 105–122.↩ 23. Karen Barad, Meeting the Universe Halfway (Durham: Duke University Press, 2007), 334.↩ 24. Steven French, The Structure of the World: Metaphysics and Representation (Oxford: Oxford University Press, 2014).↩ 25. Claude Lévi-Strauss, Structural Anthropology, trans. Claire Jacobson and Brooke Grundfest Schoepf (New York: Basic Book, 1963); Ferdinand de Saussure, Course in General Linguistics, trans. Wade Baskin, ed. Perry Meisel and Haun Saussy (New York: Columbia University Press, 2011).↩ 26. Phillip Pettit had already clearly identified the problem of extending structuralism beyond the linguistic in his The Concept of Structuralism (Berkeley: University of California Press, 1975). See Gilles Deleuze “How do we Recognize Structuralism,” 184–189 on the ‘empty square’ that is “its own metaphor, its own metonymy.”↩ 27. Ernesto Laulau and Chantal Mouffe, Hegemony and Socialist Strategy (London: Verso, 2001). “The necessity of the social is the necessity proper to purely relational identities—as in the linguistic principle of value . . . necessity, in this sense, is simply equivalent to a ‘system of differential positions in a sutured space” (ibid., 100). Laclau and Mouffe claim to reject the distinction between “discursive and non-discursive practices” insofar as “no object is given outside every discursive condition of emergence” (ibid., 93). This rejection, however, is accompanied by a collapse of the asemiotic causal dimension of practices into their semiotic relational differentiation.↩ 28. To further illustrate, consider what one would take to be relevant to the practice of politics if their intellectual diet consisted of nothing but such writings. Why given the flow of academic attention and research funding is there no ‘climate’ or ‘eco’ post-structuralism other than by means of ideology critique or hermeneutics?↩ 29. Deleuze, clear eyed on this issue, takes the modality of Saussurean structure to be virtual. Lacan and J.A. Miller prefer talk of absence, void, and the non-ontological, but what philosophically differentiates this talk from talk of the virtual is ambiguous. See Gilles Deleuze, “How do we recognize structuralism?,” in Desert Islands and Other Texts (Cambridge: Semiotext(e), 2002), 170–196; Jacques Lacan, Seminar XI: The Four Fundamental Concepts of Psychoanalysis (New York: W.W. Norton, 1998); Jacques-Alain Miller, “Matrix,” trans. Daniel G. Collins, available at https://www.lacan.com/symptom13/matrix.html (last accessed June 18, 2025).↩ 30. Karl Marx, Holy Family, in Marx and Engels Collected Works, vol. 4 (New York: International Publishers, 1975), 5–211. Here, Marx criticizes those post-Hegelians who develop in a more Platonic direction by means of resources drawn from both Aristotelean post-Hegelians and nominalists. See also Jonathan Pike’s From Aristotle to Marx (New York: Routledge, 1999) for an interpretation of Marx’s Aristotelianism about social structure.↩ 31. Louis Althusser, “The Underground Current of the Materialism of the Encounter,” in Philosophy of the Encounter (London: Verso, 2006), 163–207.↩ 32. Étienne Balibar, “On the Basic Concepts of Historical Materialism,” in Louis Althusser, Étienne Balibar, Roger Establet, Pierre Macherey, and Jacques Rancière, Reading Capital: The Complete Edition (London: Verso, 2015), 448; Louis Althusser, “On Genesis,” Décalages 1:2 (2012): 1.↩ 33. Louis Althusser, “On Genesis,” 2.↩ 34. For an account of how Althusser interprets and applies the concept of “singular essences” see Fourtounis, Giorgos. “An Immense Aspiration to Being: the causality and temporality of the aleatory,” in Encountering Althusser: Politics and Materialism in Contemporary Radical Thought, edited by Katja Diefenbach et al, Bloomsbury, 2013.↩ 35. For a defense of the political consequences of Althusser’s singularist aleatory materialism, see Ronjon Paul Datta, “From Foucault’s Genealogy to Aleatory Materialism” in Critical Realism and the Social Sciences: Heterodox Elaborations, (eds.) Jon Frauley and Frank Pearce (Toronto: University of Toronto Press, 2007), 273-295.↩ 36. See Panagiotis Sotiris, “Structure and/as Conjuncture,” in A Philosophy for Communism (Chicago: Haymarket, 2020), 200–214.↩ 37. Anthony Giddens, The Constitution of Society (Berkeley: University of California Press, 1984), 377.↩ 38. Anthony Giddens, A Contemporary Critique of Historical Materialism (Berkeley: University of California Press, 1981), 27.↩ 39. Giddens, A Contemporary Critique of Historical Materialism, 26.↩ 40. For a detailed substantiation of this charge, see Alex Callinicos, “Anthony Giddens: A Contemporary Critique,” Theory and Society 14:2 (1985): 133–166.↩ 41. William H. Sewell Jr., “A Theory of Structure,” American Journal of Sociology 98 (1992): 1–29.↩ 42. Sally Haslanger, “What is a Social Practice?,” Royal Institute of Philosophy Supplements 82 (2018): 231–247.↩ 43. Haslanger sometimes replaces “resources” with “sources” in order to emphasize that the practical valuation that partially constitutes a resource need not be a positive valuation—toxic waste, for example, is no less a resource than is lithium.↩ 44. Karl Marx, Marx and Engels Collected Works, vol. 5 (London: Lawrence and Wishart, 2010), 417.↩ 45. Dave Elder-Vass neatly makes the point that social entities, unlike many other kinds of entities, are not constrained to be “a part of [only] one entity at the next highest level of organization at a time,” which allows for the “interleaving” or overlapping of social entities across scales (Dave Elder-Vass, “Top-down causation and social structures,” Interface Focus 2:1 (2012): 88.↩ 46. Sally Haslanger, “Failures of Methodological Individualism: The Materiality of Social Systems,” Journal of Social Philosophy 53:4 (2022): 512–534.↩ 47. Jessica Wilson, Metaphysical Emergence (Oxford: Oxford University Press, 2021), 1.↩ 48. See Jaegwon Kim, “Making Sense of Emergence,” Philosophical Studies 95 (1999): 3–36; Jaegwon Kim, “Emergence: Core Ideas and Issues,” Synthese 151:3 (2006): 547–559.↩ 49. Jessica Wilson, Metaphysical Emergence, 72.↩ 50. In his Political Concepts entry, Patrice Maniglier characterizes materialism through the ‘secondness’ of thought relative to practice (Patrice Maniglier, “Materialism,” Political Concepts 3:1 [2016]: 1–25).↩ 51. Karl Marx, Marx and Engels Collected Works, vol. 5, 6.↩ 52. The reference is to Stuart Hall: “Understanding ‘determinacy’ in terms of setting of limits, the establishment of parameters, the defining of the space of operations, the concrete conditions of existence, the ‘givenness’ of social practices, rather than in terms of the absolute predictability of particular outcomes, is the only basis of a ‘Marxism without final guarantees’” (Stuart Hall, “The Problem of Ideology,” in Selected Writings on Marxism, ed. Gregor McLennan [Durham: Duke University Press, 2021], 156).↩ |
