Abstract
While the concept of functional differentiation is one of sociology’s oldest analytic tools, there is significant confusion about its meaning and purpose in the contemporary discipline. This article addresses one source of uncertainty: the conflicting array of ontological and methodological positions which are currently attached to the differentiation term. Drawing on Laudan’s philosophy of science, I argue that sociology does not house a unified program of differentiation theory, but is instead marked by at least two discrete traditions of differentiation thinking. Disentangling these traditions will, I contend, clarify our thinking, and help reduce the prevalence of internally contradictory conceptual models.
1. Introduction
“‘T is a very old strife between those who elect to see identity and those who elect to see discrepancies …” —Emerson, English Traits
From the very inception of sociology as a distinct field of inquiry, theorists have spoken of a process of change variously labelled as “social,” “structural,” or, more recently, “functional” differentiation (Luhmann 1995 [2000], 133).
1
As described by Eisenstadt, this process is one in which: [T]he main social functions or the major institutional spheres of society become disassociated from one another, attached to specialized collectivities and roles, and organized in relatively specific and autonomous symbolic and organizational frameworks within the confines of the same institutionalized system (Eisenstadt 1964, 367).
Despite the long history of inquiry on this topic, and the enthusiasm it has often aroused, many have questioned whether theories of differentiation show strong evidence of cumulative development. Surveying the situation in the 1970s, Rueschemeyer (1977, 1) noted that a genuine theory of structural differentiation was “still more a goal than an accomplished reality.” Rather than a well-integrated set of definitions, propositions, or causal hypotheses, so-called “differentiation theory” was, he argued, a ragtag assortment of metatheoretical assumptions, conceptual schemes, and descriptive generalizations about the directional tendencies of long-term social change (Rueschemeyer 1977). Today, the situation seems largely unchanged. While the notion of functional differentiation is regularly deployed in theoretical discussions, the concept continues to be marred by ambiguity, inconsistency, and general uncertainty over its specific explanatory purpose (Berger 2003).
In this article, I seek to address what I see as one major source of confusion: the problem of conceptual stretching (see Sartori 1970). Theorists with a synthetic impulse have, I believe, applied the differentiation term to widely divergent conceptions, leading to a loss of precision. Indeed, in collating all those authors who have recently been described as theorists of functional differentiation, we amass an extraordinarily diverse array of conflicting theoretical approaches. Kaldewey (2015, 89-90), for example, claims that the descriptive accounts of modern society offered by Dilthey, Weber, Parsons, Luhmann, and Bourdieu “have enough in common to be subsumed under the general label of functional differentiation” (cf. Ziemann [2007, 221], who adds Durkheim to this list). Differentiation theory is thus depicted as a broad tent, encompassing those concerned with the specialization of functional subsystems, those focused on the fragmentation of cultural values, those who comprise what is commonly referred to as “field theory” (see, for example, Lahire 2015), and those who focus, from a phenomenological perspective, on the proliferation of interpretive frames, unique realities, or distinct cognitive styles (with Thomas [2013, 30], for instance, claiming that “[f]unctional differentiation is the centrepiece of Berger and Luckmann’s (and almost everyone’s) description of the modern world”).
Integrative efforts of this kind certainly have their merits, and the theories thereby likened do seem to converge on a common image. Each speaks of a far-reaching process of change—broader and more inclusive than commonly described in accounts of the division of labor (Rueschemeyer 1977)—in which macrosocial entities of a particular type (whether systems, fields, spheres, or institutions) have gradually separated or diversified in accordance with specific teleological criteria (such as the function they perform for the wider society or the kinds of activity they encourage) (Knorr Cetina 1992). At the same time, by labeling these accounts as alternative theories of the same basic construct, we force the notion of differentiation to encompass vastly divergent metatheoretical positions. While the aforementioned scholars share certain impressionistic similarities in their view of modern society, they do not merely offer competing explanations of the same empirical phenomenon; rather they see this phenomenon in entirely distinct ways, and orient their descriptions to entirely distinct ends. Attempts to group their approaches under the same banner thus construct a category with dubious scholarly utility, giving an undue sense of a unified, coherent, and relatively homogeneous tradition of thought.
In contrast to such attempts, this article seeks to bring more clarity to theoretical discussions by sharpening conceptual distinctions. Following arguments that have largely remained confined to the German literature (see Schimank 1996; Tyrell 1998), and drawing on the work of Larry Laudan (1977), it argues that we should, rather than speaking of a single body of differentiation theory, instead consider how specific theories of differentiation are articulated within broader research traditions, marked by fundamental disagreements at the metatheoretical level. By tracing the use of the differentiation term in several major theorists, and highlighting points of divergence in their metaphysical and methodological commitments, the following piece argues that there are at least two distinct traditions of sociological theory within which accounts of differentiation have developed, and that these traditions supply us with notions of differentiation that are oriented to different aims, suited to different purposes, and work to solve different kinds of empirical and conceptual problems. As I will show (with special reference to the work of Niklas Luhmann), the conflation of these traditions has led to muddled and misleading thinking, not only in interpretive commentaries but from theorists of differentiation themselves, and must be avoided if we wish to define, discuss, and theorize differentiation in a logically coherent manner.
2. Research Traditions
In saying that notions of functional differentiation are articulated in distinct research traditions, we borrow a term from Laudan (1977), who introduced the concept in his general philosophy of science. 2 Science, for Laudan, is a problem-solving activity, and theoretical work is oriented to this aim insofar as theories offer potential solutions to cognitively significant problems (14). These problems can be empirical—having to do with substantive, unresolved questions about the behavior of objects in a domain of inquiry—or conceptual—having to do with either inconsistencies, ambiguities, or anomalies within a particular theory, or with incompatibilities between the elements of a theory and those of another (especially where the latter seems especially well-founded). Under this paradigm, theoretical progress involves increasing the ability of a theory to solve empirical problems while reducing any conceptual problems which may thereby be generated (68).
What is of most interest to us here is Laudan’s (1977, 71-72) claim that an analysis of scientific theory must distinguish between “two different sorts of propositional networks” to which the term “theory” is applied. On the one hand, theories are sometimes thought of as sets of related propositions which make “specific experimental predictions” or provide “detailed explanations” of a particular empirical phenomenon (71-72). On the other, the word theory is also used to describe general sets of presuppositions which inform these more specific claims (such as when we speak of “Darwinian theory” or “psychoanalytic theory”). In Laudan’s view, any sound understanding of the scientific process must recognize that these two kinds of theories are fundamentally distinct, not only in matters of generality and specificity, but also with regard to “the modes of appraisal and evaluation appropriate to each” (72). As such, he proposes a new term—research traditions—to describe theories of the more global kind.
Research traditions are, for Laudan (1977, 79), collections of interrelated metaphysical and methodological commitments: beliefs about the appropriate objects and methods of research. While traditions house more specific explanatory theories, and influence the kinds of entities and processes to which these theories are able to refer, they also extend beyond them, and can “go through a number of different, detailed (and often mutually contradictory) formulations” without dissolving (78-79). The unity of a research tradition is instead found in its metatheoretical commitments (79). One belongs to a tradition by sharing its commitments, and one repudiates it by referring to entities, methods, or explanatory mechanisms which the tradition does not recognize or consider as legitimate (78).
While Laudan’s notion of research traditions is explicitly indebted to Kuhn’s (1962) well-known concept of scientific paradigms, as well as Lakatos’s (1970) related idea of research programs, it is also pitched against them, and has some important advantages as an analytic tool. As Laudan (1977) notes, the concept of a paradigm is, in Kuhn’s work, “systematically ambiguous” (73). The exact nature of paradigms remains “obscure and opaque,” (74) and we are not given a clear set of criteria for identifying and articulating particular paradigms (see also Shapere 1964). In the case of Lakatos, the opposite problem applies: the criteria defining a research program are so strict that specific theories can only belong to the same program if one of the two entails and builds off the other (Laudan 1977, 77). By contrast, Laudan’s notion of research traditions presents an appealing, middle-ground method to identify and evaluate distinct theoretical paradigms. While it recognizes that the boundaries of such entities may be fuzzy and open to change, it also provides us with a well-specified method by which to locate points of cohesion, positing that classifications grounded in ontological and methodological similarities will have greater analytic utility than those which rely on arbitrary, ambiguous, or ad hoc criteria.
It is with this in mind that we can consider the case of differentiation theory in sociology. While the term itself suggests the existence of a unified research tradition, previous attempts to demarcate distinct strands of differentiation thinking suggest the possibility that there is substantial variance within this body of thought. For example, Kron, Berger, and Braun (2013) offer a five-fold categorization of “perspectives on differentiation,” speaking of evolutionary, functional, role-specific, cultural-theoretical, and form-theoretical conceptions. Likewise, Tyrell (1998) posits three major “lines” of differentiation discourse: a physiological line, developing from Spencer, in which differentiation is a process of functional specialization; a “social-scientific” line, seen in Simmel, in which differentiation is synonymous with individualization; and a specifically German line, traceable to Dilthey and Weber, in which differentiation refers to the growing autonomy of distinct cultural spheres (see also Kangas 2012). While both typologies are helpful, the precise manner by which they create these categories is largely unclear, and the evident differences between the two schemes—as well as the inclusion of theorists who, like Simmel, are describing a process of change quite different to that usually implied by notions of structural or functional differentiation (see note 1 above)—speak to the need for a more stringent approach.
If we are to follow Laudan’s model, such an approach should examine key differences in the metatheoretical commitments which inform theories of functional differentiation, some of which are hinted at in the above classifications. For example, Tyrell (1998), in distinguishing the accounts of Spencer and Weber, hits on a key metaphysical difference between the two: divergent attitudes to the notion of society. While the former adopts traits of organicism, holism, or “systems-theoretical” thinking, the latter has a more strictly individualist (or “action-theoretical”) habit of thought. Less noted is, however, a key methodological difference between these two theorists. 3 While Spencer treats differentiation as a universal comparative construct, applicable in principle to all social formations, Weber treats it as a singular historical event, wholly associated with the description and interpretation of Western modernity.
While these are not the only points of contrast between the two approaches, they are, I believe, the most prominent, and together they show the influence of two distinct—and diametrically opposed—traditions of thought. On the one hand, one sees a distinctly functionalist tradition, in which an organicist model of society is combined with a nomothetic vision of inquiry. On the other, we can perceive what I will describe as a Weberian tradition, which is fundamentally skeptical toward the concept of society, and which therefore adopts an idiographic approach. In the following sections, I show how these principles have filtered into two distinct strands of differentiation thinking, each carried across several prominent authors. While these two approaches have influenced one another, and have become deeply entangled in many contemporary accounts, they are almost entirely incompatible on the metatheoretical level, and ultimately remain embedded in analytically distinct traditions. As such, efforts which work across the two traditions, or freely draw elements from either approach, can easily fall into conceptual confusion, and often end up appealing to a contradictory set of metatheoretical commitments. If this is to be avoided, we must be clear on the points which separate traditions, and must ensure that we do not pitch our theories in an uncomfortable space between them.
3. The Functionalist Tradition
Let us begin by examining the functionalist tradition. The characteristic feature of functionalism is, as Turner and Maryanski (1988) argue, a mereological form of thinking in which the action of a part is linked with some quality, capacity, or condition of a whole. It is assumed that some aspect of a designated whole—a society, for instance—depends on the working of its parts, these workings being considered functions. 4 In this view, differentiation occurs when such parts become more specialized vis-à-vis the needs or capacities of their containing whole—i.e., when functionally homogeneous parts, each performing multiple functions, are replaced by functionally heterogeneous parts, with distinct and more limited functional orientations. This is most clearly stated by Parsons (1966, 22), who defines differentiation as the process by which a “unit, subsystem, or category of units having a single, relatively well-defined place in the society divides into units or systems (usually two) which differ in both structure and functional significance for the wider system.”
This conception of the differentiation process is not unique to sociology, and is best understood by way of comparison to the two major notions from which it was analogously derived: the concept of the division of labor in manufacture—as theorized by such figures as Smith, Mandeville, Hutcheson, and Ferguson—and the associated idea of a physiological division of labor—a nineteenth century doctrine evident in the work of several prominent naturalists, including Spencer (D’Hombres 2012). All three concepts rely on a mereological ontology, positing a systemic whole—such as a manufactory, an organism, or a society—and arguing that the needs of these wholes are met by the cooperative action of more or less differentiated parts. Thus workers are thought to fulfill the productive goal of their firm; cells, tissues, and organs are thought to maintain the life of an organism; and various social phenomena, such as roles, collectivities, or subsystems, are thought to preserve the existence of a society.
Commitment to a mereological model is a common feature of many theories of differentiation—most prominently those of Spencer, Durkheim, and Parsons—and theories based on this commitment can thus be located in a shared, functionalist tradition. This should not, however, lead us to conclude that these theories present a single, unified model of the differentiation process, nor that they agree on its various causes or consequences. As Laudan (1977, 81) notes, theories within traditions, especially evolving traditions, can be “mutually inconsistent rivals,” seeking to “improve and correct their predecessors.” In the case of differentiation, the functionalist tradition offers a range of distinct theories, and contains points of disagreement which—not broaching ontological or methodological questions—do not threaten the tradition’s integrity. For example, though functionalist theorists agree that societies are composed of parts, and that these parts perform functional labor within and for a societal whole, they disagree on what these parts are, what, precisely, it means for them to address functional ends, and what causes these various components to emerge or to change.
In Spencer, as well as in Parsons, the function term is tied to a systems-theoretical notion of equilibrium, inspired by analogies to organic phenomena. Spencer (1864, 74), for example, argues that an entity is in a state of equilibrium when there is a “continuous maintenance of such inner actions as will counterbalance outer actions” (as, for example, in the case of thermoregulation). If these inner actions are not present, the entity ceases to exist as something distinct from its environment: a state of death in organisms or dissolution in societies (Spencer 1873, 339). As such, Spencer (1864, 80) holds that living entities—whether organisms or societies—must contain “structural arrangements which enable [them] to adapt [their] actions to actions in the environment.” The forces produced by such arrangements are, in his view, functions; indeed he explicitly defines this term as “the statical and the dynamical distributions of force which an organism [or a society] opposes to the forces brought to bear on it” (Spencer 1864, 154; see also 1862 [1867], 82, 501). 5
In Parsons we find the same argument, updated with the more modern language of homoeostasis. Systems, note Parsons and Shils (1951, 107), “must have a tendency to self-maintenance, which is very generally expressed in the concept of equilibrium.” As in Spencer, systems are thought to contain functional components which collectively work to combat environing actions and maintain the “life” of the whole; in fact Parsons (1961, 36) explicitly argues that the function term designates processes which mediate between the exigencies imposed by the “relative constancy or ‘givenness’ of a structure,” on the one hand, and those imposed by the “givenness of the environing situation external to a system,” on the other.
Durkheim, too, is inspired by organic analogies; indeed he uses the function term precisely because it evokes the “relationship existing between [living] movements and certain needs of the organism” (see Durkheim 1893 [1984], 11). While these needs are, in the case of the “social organism” (1893 [1984], 155), sometimes described in terms of equilibrium (with Durkheim claiming, for instance, that “law and morality have the aim of assuring the equilibrium of society, of adapting it to surrounding conditions” (Durkheim 1886 [1974], 212), Durkheim’s search for a more specifically sociological vocabulary ultimately leads him to treat the notion of social solidarity—the need for individuals to be bound to the group, and for society to exist as a unified, individuated, and integrated whole (see 1893 [1984], 102, 23)—as his primary point of reference in functional analysis (Pope 1975). This move does not, however, challenge Durkheim’s place in the functionalist tradition, for he still retains a resolutely mereological ontology. Societies, he notes, are wholes “comprised of various parts in combination” (Durkheim 1895 [1964], 80; 1892 [1960], 54), and these parts perform functions when their actions contribute to the “general needs of the social organism” (1895 [1964], 95), or when they have as their objects the “defence, nurture, and the assurance of renewall [sic] and continuity of society” itself (1886 [1974], 212).
Though Durkheim speaks of political, administrative, juridical, military, and religious functions (1893 [1984], 2, xxxiii), these functions are not specifically enumerated or systematically deduced. By contrast, Spencer and Parsons both look to produce a universal catalogue of basic requisites, each consciously grounded in higher order principles (see Turner 1985, 58-59). Spencer, for example, bases his account of social functions on his general theory of physical force (see Spencer 1862 [1867]), which holds that one can distinguish between the accumulation, the expenditure, and the transfer of force. In his view, all societies must address these three general tasks, and they do so through more or less differentiated functional networks, respectively labeled as the “sustaining” system (addressing accumulation), the “regulative” system (addressing expenditure), and the “distributive” system (addressing transfer) (Spencer 1898a). 6 Similarly, Parsons (1977) presents the well-known “four-function” or “AGIL” paradigm, initially derived from the pattern variable scheme. Under this model, all living systems, including societies, are thought to be governed by four exigencies: adaptation, the need to provide facilities and resources, goal attainment, the need to minimize discrepancies between system needs and environmental conditions, integration, the need to coordinate differentiated parts, and pattern maintenance and latent tension management, the need to stabilize internal structure and provide sufficient motivation for units to follow established patterns (Parsons 1961).
Although these authors present different catalogues of functional exigencies, they nonetheless converge on a common idea: the view that there are an essential set of needs, requisites, or fundamental aims which must be addressed in every society. This leads to an important feature in functionalist notions of differentiation. As in the notion of the division of labor in manufacture, differentiation is treated as a process by which functions are concentrated. The functions being performed do not diversify or expand, but are simply redistributed, becoming the responsibility of a more specialized group of elements. 7 This account thus displays what has become known as a “decomposition paradigm,” in which differentiation is “thought of as the division of an antecedent unity into two new parts/systems deriving from this unity” (Stichweh 2013, 56).
Support for this model is readily found in the writings of Spencer, Durkheim, and Parsons. Spencer (1898a, 597), for instance, argues that through differentiation, “all institutions, at first confusingly intermingled, slowly separate.” In his view, differentiation divides what was originally united. It occurs when the functions performed by a multifunctional unit (entities such as classes, technologies, roles, and institutions) are redistributed to new units, each of which caters to a narrower subset of the original functions. Differentiation does not change the exigencies which must be met; instead, it simply creates more specialized structures to respond to them. For example, Spencer argues that every society has a “regulative” exigency: a need to control and direct individual behavior. While small, acephalous societies meet this exigency by the combined efforts of all individuals (who live, Spencer believes, without superiority or subordination), more “complex” societies are said to possess specialized organs of governance, the development of which involves a progressive differentiation of the regulative system (Spencer 1898a, 521-530).
Durkheim also demonstrates this way of thinking, though in a less consistent manner. Ostensibly, his major work on differentiation, The Division of Labor in Society, is focused on occupational diversification, or on what political economists have typically referred to as the social division of labor (specialization between firms, industries, or occupations), rather than the division of labor in manufacture (specialization within a firm, workshop, or manufactory). The difference between the two lies in the presence of the mereological model: in the former there is, as Marx (1867 [1976], 472) surmises, “exchange between spheres of production which are originally distinct from and independent of one another,” while in the latter there is a decoupling of functional tasks within an integrated, goal-directed whole. In Durkheim’s account, these two processes are conflated. Occupations are thought to perform social functions, and to cooperate to meet the needs of society in the same way that physiological components cooperate to preserve the life of an organism (see Durkheim 1893 [1984], 2, 212). Indeed, Durkheim (1893 [1984], 291-292) explicitly argues that the division of labor must involve “allocation of a common function,” such that the divided units “[concert] with others to maintain life more generally.”
Thus, while Durkheim does, in places, imply that the division of labor creates novel kinds of activity (e.g., 1893 [1984], lii, 216), the account of specialization presented in the Division generally follows the decomposition model. The division of labor is described as a process of disaggregation in which individuals come to “[concentrate] upon one part only of the total function that they fulfilled up to that time” (Durkheim 1893 [1984], 212), or in which a static set of social functions gradually disentwines. For example, Durkheim notes that: Originally, [religion] extended to everything; everything social was religious …Then gradually political, economic and scientific functions broke free from the religious function, becoming separate entities and taking on more and more a markedly temporal character. (1893 [1984], 119)
Likewise, we are told that the “functions of the scientist,” formerly “almost always exercised alongside another more lucrative one, such as that of doctor, priest, magistrate or soldier, are increasingly sufficient by themselves” (1893 [1984], 2); that “in the beginning it was the gathering of the whole people which fulfilled the functions of a court of law” (59), and that the functions performed by the family “at first undivided and overlapping, have gradually separated out and been constituted independently …” (78-79). In each of these cases, we see a process of decomposition: a “sharing out of functions that up till then were common to all” (218).
For Parsons, too, differentiation involves decomposition. Echoing Spencer, he describes differentiation as a process in which structural elements (roles, collectivities, norms, or values) come to focus on a single AGIL exigency, rather than contributing to all four functions in equal proportion (Blain 1971, 679). Once again, the functions are static. Every society will, in Parsons’s view, contain four analytically distinguishable “functional subsystems”: networks of structures and processes which address a particular functional exigency. The differentiation process simply reduces the extent to which these subsystems overlap on the empirical level, as units increasingly orient themselves to a single exigency (and thus participate only in a single subsystem) rather than having more diffuse functional roles (Parsons and Smelser 1956, 16). Structural differentiation cannot, in this model, alter the actual functions performed, or generate a fifth subsystem (though it can, Parsons believes, lead to further AGIL specializations within existing subsystems). Instead, differentiated units are, taken together, always “functionally equivalent” to the multifunctional units from which they disaggregate (Smelser 1959, 2). 8
The decomposition paradigm is not an arbitrary or incidental feature of the functionalist approach to differentiation theory, but is closely related to the methodological commitments of the functionalist tradition, and to the specific cognitive problems it has historically tried to solve. In particular, it is related to the desire to treat functional differentiation as a comparative metric, and thus to stabilize its meaning within comparative inquiry. One sees this in many of the biological accounts that initially inspired Spencer. As D’Hombres (2012) has argued, the notion of the physiological division of labor was developed to address a particular conceptual problem of nineteenth-century thought: the need to provide a set of rational criteria by which to rank certain organisms as more evolved than others. The concept of differentiation—thought of in terms of decomposition—became central to this aim. Looking to treat differentiation as the primary marker of evolutionary progress, many prominent naturalists, most notably Milne-Edwards, promoted the idea that every organism, regardless of its particular morphology, possessed the same set of functional aptitudes, differing only in the degree to which these aptitudes were the product of distinct anatomical parts (D’Hombres 2012, 9-10). As the aptitudes did not change, the “division of physiological labour” could, it was thought, act as a clear yardstick for comparison, and could, when combined with the idea (generalized from political economy) that specialization improves productive forces, establish a rational basis by which to position organisms on a hierarchy of being (D’Hombres 2012).
Each of the three figures examined above take up this way of thinking, and treat functional differentiation, to use Smelser’s (1959, 1) phrasing, as a marker of “‘growth,’ ‘advancement,’ and ‘civilization.’” Spencer (1904, 297), for instance, claims that an “increasing division of labour … characterises social evolution everywhere,” and generally treats evolution and differentiation as interchangeable terms (see Perrin 1976). Likewise, Durkheim, explicitly citing Milne-Edwards’s argument that an organism’s place in the “animal hierarchy” is determined by its specialization of functions (1893 [1984], 3), claims that mechanical solidarity becomes looser the more that we “mount the scale of social evolution” (105), and that the division of labor is most advanced among the “highest social types” (92). Even Parsons (1966, 2), writing well after the Darwinian revolution, upholds this view, claiming that evolution is a fundamentally directional process which proceeds from “simple to progressively more complex forms.” Indeed, Parsons explicitly argues that an “evolutionary perspective” requires a “criterion of evolutionary direction,” presenting “structural differentiation” (alongside “adaptive upgrading,” “inclusion,” and “value generalization”) as one such marker (26).
The need to construct a universal scale of evolutionary progress has—in light of the ascendance of the Darwinian paradigm—generally become regarded as a pseudo-problem: today, one no longer speaks of organisms as more or less evolved in a general, context-independent sense. The link between differentiation and evolution established by the functionalist tradition does, however, speak to a more contemporary ambition. In establishing a conceptual relationship between these two terms, functionalist theorists relied on, and sought to extend, a host of empirical theories about the causes and consequences of functional specialization. For example, Spencer links differentiation with evolution because he believes that differentiated systems can fulfill their functional exigencies in a more efficient manner, and can thus maintain their equilibrium in a wider variety of external conditions (Spencer 1864, §36). This exact argument is also present in Parsons, who argues that differentiation leads to an increase in what he terms adaptive capacity: the “capacity of an action system … to maintain and generate resources which can serve to improve the system’s level of adaptation to the environments in which it is situated” (1977, 297, 231). Stripped of any normative gloss, these claims are simply causal propositions about macrosocial dynamics, and it is to the development and articulation of such propositions—i.e., to a nomothetic model of sociological inquiry—that the functionalist tradition is ultimately oriented.
Evidence of this commitment is clearly present in each of our three theorists. Spencer, for instance, distinguishes his own “general sociology”—which aims to uncover “facts of structure and function displayed by societies in general, dissociated … from special facts due to special circumstances” (1898a, 36-37)—from what he calls “special sociology”—his term for ethnographic or historical case studies. Durkheim, likewise, claims that he is only interested in the “constant” causes of the division of labor—those that explain its regular historical development—rather than the various “special causes” which might explain its particular form in particular societies (1893 [1984], 79). Finally, Parsons explicitly argues that his conceptual framework is designed to produce general, transferable insights about regularities in social phenomena (see Parsons 1951, 20), contrasting this with a more “historistic” position in which the “meaningfulness … of a structural unit or process [such as differentiation] could be defined only within one very specific system” (1977, 283).
It is in light of such claims that we can best understand why each of these theorists conceive of functional differentiation as a process of decomposition, rather than one in which novel functions may emerge. As Smelser (1967, 35) indicates, nomothetic inquiry demands a conceptual framework which can maximize the possibility of systematic comparison, for to compare is to control, and to control is to augment one’s ability to make reliable inferences about causal regularities (see also Sartori 1991; Zelditch 1971). The nomothetically minded theorist must therefore operate at a high level of abstraction: the meaning of their concepts cannot be limited to specific societies because propositions including these concepts would only then apply to these societies. The decomposition paradigm creates a concept of differentiation which meets this particular demand. As it posits that all societies have the same, invariant set of functions, defined with reference to universal requisites of social existence rather than contingent problems of specific societies, it allows one to treat functional differentiation (or the degree of “functional specificity vs. functional diffuseness” in a particular society’s structure [Parsons and Smelser 1956, 34]) as a generic, cross-cultural variable, identical in meaning across all societies and logically comparable across them (on this method see Goldschmidt 1966). 9
From this, we can see that the metaphysical and methodological commitments of the functionalist tradition are logically related, not in the sense that a commitment to nomothetic inquiry will inevitably produce a concept of functional differentiation, but in that this commitment will, where it informs attempts to conceptualize such a process, encourage theorists to do so in ways that facilitate comparative inquiry. In the case of the above theorists, this is not only evident in their adoption of a decomposition model, but can also be seen in one of their most basic ontological commitments: the mereological notion that societies are systems. The logic of comparative control demands not only that units of comparison share some common variable, but also that they are statistically independent: events in one not influencing events in another (Zelditch 1971, 267). The system paradigm supplies this, for it allows one to treat societies as discrete, boundary maintaining entities, with common traits of organization or behavior (Ragin and Zaret 1983, 737). As such, it postulates that the comparison of societies—and the development of nomothetic knowledge on this basis—is logically acceptable. In fact, by appealing to an abstract system concept, it even allows one to compare societies to other kinds of entities, and to speak of concepts, methods, and causal generalizations which hold across these various “systemic” objects.
Comparisons of this kind were essential to the development of sociological functionalism, and to its particular notion of functional differentiation. Appealing to an interdisciplinary register of general systems science, functionalist theorists developed their ideas on the differentiation process by drawing on theories in more established disciplines. For example, Spencer’s thought that differentiated systems can fulfil their exigencies in a more efficient manner is clearly generalized (via Milne-Edwards) from economic discussions of the division of labor, as is Parsons’s related idea that differentiation acts to enhance adaptive capacity. In a similar vein, Durkheim’s (1893 [1984], 212) claim that the division of labor is causally preceded by a rise in “dynamic density” (the frequency and intensity of social intercourse in a given area, which Durkheim believes creates competitive pressure for specialization [205]) is an explicit sociologization of Darwin’s famous “principle of divergence,” an ecological notion itself inspired by work in economic theory (Richards 2012; Schweber 1980).
4. A Weberian Alternative
The above discussion demonstrates how the key claims, commitments, and cognitive problems of the functionalist tradition gave rise to a particular way of theorizing functional differentiation. While the authors examined disagree on what structural units exist, and what, precisely, it means for these units to perform social functions, each posit the existence of a general process of change in which these functional performances become increasingly specialized, and each argue that one can compare the degree to which this process has advanced across distinct societies. Assumptions of this kind led more empirically-minded thinkers to produce quantitative measures of functional differentiation, and to deploy them in comparative research. For example, Marsh’s 1967 work Comparative Sociology presents a universally applicable measure of structural differentiation—relying on indicators for which there are “at least convertible data in all societies” (Marsh 1967, 31)—and empirically correlates this variable with a range of other macrosocial traits. This approach clearly demonstrates the methodological commitments of the functionalist tradition, and is the logical outcome of the way its key theorists conceptualize the differentiation process.
Things are entirely different in what I have thus far referred to as a largely Weberian strand of differentiation theory, whose foundational text is Weber’s (1915 [1946]) famous Zwischenbetrachtung (“Intermediate Reflection”). In this text, Weber introduces a typology of distinct “value spheres” or “life orders,” each characterized by a distinct intrinsic logic or set of “inherent laws” (Eigengesetzlichkeit). A development on Dilthey’s (1910 [2002], 187) concept of “cultural systems”—“uniform and homogenous productive system[s] … formed by the activities of specific individuals” and directed toward the fulfilment of a “cultural function”—Weber’s account of the value spheres has had a range of interpretations, with theorists disagreeing over their precise character (Terpe 2020). In general, however, they are thought of as differentiated sets of rules or norms that evoke specific patterns of action in particular contexts (Bruun 2008, 102). Weber explicitly outlines five such spheres—the economic, political, aesthetic, erotic, and intellectual—with most commentators agreeing that the religious constitutes a sixth (Terpe 2020).
In Weber’s account, the six conflicting life orders are ideal types, the cataloguing of which is “merely intended to show that at certain points such and such internal conflicts are possible and ‘adequate’” (1915 [1946], 323). At the same time, Weber argues that the value spheres, though “prepared with a rational consistency which is rarely found in reality … can appear thus in reality and in historically important ways …” (323-324). As in Parsons’s account of subsystem differentiation, Weber thus speaks of circumstances in which empirical phenomena begin to resemble analytic distinctions. As he notes: … the rationalization and the conscious sublimation of man’s relations to the various spheres of values … have then pressed towards making conscious the internal and lawful autonomy of the individual spheres; thereby letting them drift into those tensions which remain hidden to the originally naive relation with the external world. (Weber 1915 [1946], 328)
Thus, while Weber (1915 [1946]) believes that the distinct ends of different spheres are always analytically distinguishable, certain processes of change can increase the degree to which conflicts between them are evident to the social actor. In such cases, the spheres become more empirically conspicuous (“more and more consciously grasped independent values which exist in their own right” [342]), leaving us with a sharper awareness of incompatibilities between available life ends. With, for instance, the differentiation of the religious, scientific, and aesthetic spheres, we come to feel a growing disjuncture between what is beautiful, true, and good, recognizing that these attributes may not necessarily overlap (Weber 1922 [1946], 148). On the level of social structure, this can be felt in a greater differentiation of roles and values. As individuals are increasingly compelled to internalize the ethic of a specific value sphere as a personal calling, roles previously oriented to multiple spheres are replaced by those directed to the operative norms of a single sphere alone (a trend which Weber moralizes in the two “Vocation” lectures). Today, for instance, the “political man” internalizes the values of the political sphere, and acts only in accordance with the “rational rules of the state order”: “‘without regard to the person,’ sine ira et studio …” (Weber 1915 [1946], 334).
Weber’s portrait of the value spheres is often described as a theory, or proto-theory, of functional differentiation, for like Dilthey he does at times suggest that the spheres have different functions (Weber 1915 [1946], 342, 331; see Berger 2003). The notion of functionality being deployed in such cases is, however, entirely different from the mereological approach that defines the functionalist tradition. As is well known, Weber’s commitment to methodological individualism made him particularly wary of holistic or systemic models of society, as well as any related notions of societal level needs (Roth 1976; Stone 2010). As such, the value spheres are not, in his view, oriented to exigencies or requirements of an overarching system, but are—to take Habermas’s (1981 [1984], 187-188) formulation—ultimately concerned with the “material and ideal needs” of “sociated individuals.” The spheres thus “function” in much the same manner as competing products in the marketplace. For example, Weber (1915 [1946], 342) argues that art and religion compete over the individual need for “this-worldly salvation.” In that sense, these two spheres are not actually functionally differentiated at all; rather they offer competing means for fulfilling the same end and are in fact functionally equivalent.
Weber’s metaphysical stance thus leads to an entirely different understanding of differentiation than one finds in the functionalist tradition, for he does not posit a societal entity for which functions are performed (Schwinn 2003; Tyrell 1998). His view is thus, to return to aforementioned analogies, closer in form to the notion of the division of labor in society rather than the more mereological notion of the division of labor in manufacture. One sees this in the language used: while the functionalist tradition, operating via the latter model, primarily describes relationships between differentiated entities in terms of cooperation, integration, or interpenetration, the primary nouns in Weber’s account of the value spheres are conflict, competition, and tension (see Kangas 2012; Tyrell 1998). Indeed the very aim of the “Intermediate Reflection” is to demonstrate how the value spheres, by encouraging distinct and irreconcilable ends, increasingly clash with religious ethics, such as when the demands of an increasingly rationalized economic sphere come into conflict with the religious ethic of brotherly concern (Weber 1915 [1946], 331). Thus while Weber speaks of relative compromises between these ends, and occasional alliances between them in “empirical historical reality” (337, 343), his presentation ultimately stresses the totalizing character of individual spheres, each of which postulates an ultimate end whose full realization relies on one’s capacity to repudiate and reject alternatives. One cannot, for example, speak of a fundamental value beneath which these spheres could be unified; in Weber’s view any hierarchical arrangement or normative integration of the value spheres is utterly ruled out (Oakes 2003, 29-30).
While Weber’s account of the value spheres is relatively underdeveloped, it has nonetheless had a significant influence on many contemporary theories, such that we can, noting that many of these theories also adopt Weber’s major metatheoretical commitments, speak of a distinctly Weberian strand of differentiation thinking, proponents of which participate in a shared tradition of thought. For example, Habermas (1981 [1984], 176-177) refigures Weber’s account of rationalization into an explicitly neo-Kantian theory of modernization, based on the differentiation of cognitive, expressive, and evaluative rationalities, and the corresponding formation of three autonomous “cultural systems of action”: science, morality, and art. One can, however, also see the Weberian heritage in the work of Zetterberg (1991)—whose account of modernity focuses on the differentiation of six discrete institutional realms, each oriented to a distinct “cardinal value” (see Hallonsten 2021)—Schwinn (2001)—whose action-theoretical account of differentiation is explicitly posed against the mereological ontology of functionalist analysis—and many theorists connected with the broad enterprise of field theory—at least one branch of which stems, as Martin (2003, 20) has noted, from attempts to translate “Weber’s sociologization of the idea of ‘spheres of value’” into a more empirically grounded analytic tool (see also Schimank 2021, 487). 10
While such theories offer distinct analytical possibilities, and rest on a diverse array of background assumptions, what is common to each is a strong rejection of functionalist metaphysics: in particular its mereological ontology. Nowhere do such theories argue that the differentiated entities of which they speak make a specialized, collaborative contribution to the needs of a systemic whole, and if they deploy notions of systems or subsystems they do so only in a loosely metaphorical manner, rather than to establish interdisciplinary analogies or generalized laws. Such theorists thus participate in an entirely distinct ontological tradition than Spencer, Durkheim, or Parsons; rejecting holism as a metaphysical principle, and jettisoning any organicist beliefs, they instead operate with a weak, or in some cases explicitly empty, concept of society (Tyrell 1998). This is perhaps most clearly conveyed by Schwinn (2001), whose sophisticated account of differentiation, situated squarely in the tradition of Weber, is presented under the telling title Differenzierung ohne Gesellschaft (Differentiation without Society).
A similar divergence is present on the methodological level. Although Weber’s account of differentiation follows a decomposition model, in that he depicts a process in which analytic distinctions between the value spheres increasingly take on empirical form, rather than one which creates additional spheres, his use of this model is merely incidental, and is by no means linked to the methodological demands of systematic-statistical comparison. Indeed, Weber’s fundamental hostility to “strong” concepts of society (see Schwinn 2011) renders any such comparisons—based as they are on the assumption of statistical independence—invalid from the outset. Weber’s approach is, as such, not oriented toward universal generalizations, but is instead designed to produce what Nagel has termed genetic explanations: attempts to demonstrate, through “singular statements about past events … why it is that a given subject of study has certain characteristics …” (Nagel 1961, 25). As Roth (1981, xxi) explains, such attempts require both “configurational” and “developmental” modes of analysis. In the former, one identifies unique types or historical constellations (or what Weber (1904-5 [2001], 13), following Rickert, describes as historical individuals: “complex[es] of elements associated in historical reality which we unite into a conceptual whole from the standpoint of their cultural significance”), while in the latter, one seeks to provide contextual explanations of their particular course of development.
While such approaches involve an analysis of causality, they generally reject the “necessary and lawlike” causes targeted in more nomothetic accounts (Roth 1981, xxv). Likewise, while they rely on comparative inquiry, it is a qualitative-historical model of comparison, designed to evince the heterogeneous character of cultural constellations, rather than a quantitative-statistical one, designed to minimize context in favor of abstract, “permanent causes” (Ragin and Zaret 1983; see Weber 1904 [2012], 164). The Weberian tradition of theory is thus, in comparison to the functionalist, oriented around entirely different cognitive problems. In Weber’s case, for instance, one sees an enduring interest in the “specific and peculiar rationalism of Western culture,” and a lifelong attempt to diagnose and explain its uniquely Occidental path of development (Weber 1904-5 [2001], xxxix, xxxviii; see Schluchter 1979 [1981], 14). The descriptive configurations mobilized to this end—such as the portrait of the differentiated value spheres—and the genetic causes thought to explain them—such as the value patterns of ascetic Protestantism—are thus pitched as specific features of this culture, rather than generic variables. While the logic of such analyses may appeal to underlying regularities in psychology or social behavior, nomothetic knowledge is, in Weber’s (1904 [2012], 118-119) view, merely the means, and never the end, of cultural inquiry, which must ultimately deal with the “concrete conditionality” and general “cultural significance” of individual historical phenomena.
Weber’s work thus exemplifies the key contours of idiographic or historicist inquiry—if by such terms we can merely designate a commitment to limited generalization, grounded in “specific historical causes and contexts” (Beiser 2011, 19). In this approach, emphasis is placed on particularism, subjectivism, and contingency. Social phenomena are interpreted with reference to their particular cultural contexts, and the formulation of generic, cross-cultural variables is thought to be a gross violation of ethnographic fact (see, for example, Leach 1964). Thus while one might, in a historicist analysis, argue that modernity is marked by a process in which “legal” elements become differentiated from “religious” elements, one would not, as Durkheim did (see note 9 above), present a universal definition of these two clusters, nor aim to determine their level of differentiation (and its correlates) across different societies. Instead, one would hold that the meaning of terms like law and religion (and, indeed, of differentiation itself) is always embedded in a historico-cultural milieu, abstraction from which is intrinsically damaging to the goal of understanding (Zelditch 1971, 277; for a discussion of this position with regard to religion see Schwinn 2024). This principle is clearly evident in Weber, and one sees it carried forward in a tradition for which “functional differentiation”—or what is perhaps better labeled as the differentiation of institutions (Zetterberg 1991), systems of action (Habermas 1981 [1984]), or “fields of organized striving” (Martin 2003)—is not an abstract variable, but a concrete historical transformation, with a contextually unique set of causes and consequences.
Despite differences in their specific formulations of the differentiation process, none of the authors who have taken up this approach would thus give license to the kind of systematic-statistical comparative study envisaged by Marsh. While they typically see differentiation processes as being of general, world-historical significance, they also claim, to mention some prominent examples, that the “extensive differentiation” of autonomous institutions is a unique trait by which “European civilization differ[s] from others” (Zetterberg 1991, 318, 309); that “cultural spheres of value” were, from the sixteenth to the eighteenth century, “differentiated out of the traditional residues of religious-metaphysical worldviews transmitted by the Greek and above all the Judaeo-Christian traditions” (Habermas 1981 [1984], 165); and that the breakthrough of a modern, “occidental pattern of differentiation” is not, as in Parsons, causally dependent on “evolutionary universals,” nor, as in Durkheim, on generic traits like “dynamic density,” but rather stems from “cultural characteristics of the Jewish-Christian religion” (Schwinn 2017, 360). Such statements give clear examples of genetic, historicist forms of explanation, in that they emphasize a pattern of social differentiation unique to Western culture, and speak to the specific concatenation of historical conditions responsible for its development.
5. Theory at the Crossroads
The contrast between the approach of the above authors and those situated in the functionalist tradition gives ample evidence that theorists associated with the differentiation term are often committed to vastly different metatheoretical positions, and often work within entirely distinct research traditions. As Ragin and Zaret have pointed out in their attempt to distinguish the “Durkheimian” approach to sociological analysis—based on quantitative-statistical comparison—from its “Weberian” counterpart—characterized by a qualitative-historical method—these two strategies (here reformulated as alternative research traditions) are “neither congruent nor convergent with regard to units of analysis, conceptions of causality and adequate explanation, or logic of analysis” (1983, 749). As such, efforts which seek to combine elements across these two traditions, or which appeal to a notion of functional differentiation which is supposedly shared between them, are likely to generate a number of conceptual problems, and will—without careful consideration—struggle to maintain a coherent set of metatheoretical commitments.
Problems of this kind have long hampered theories of functional differentiation. Indeed, while the specific theories outlined above can be broadly grouped into two distinct traditions (particularly when we focus on their more explicit statements about matters of ontology or methodology), many of them make arguments which undermine the consistency of their metatheoretical positions. Durkheim, for instance, famously argues that the function of the division of labor is to produce organic solidarity by integrating social units in an interdependent totality, an argument explicitly based on physiological analogies. As, however, many critics have noted (e.g., Limoges 1994, 330), Durkheim’s explanation of what causes the division of labor to emerge drops this model entirely, turning instead to an ecological analogy in which specialized occupational roles—now treated as equivalents to biological species—are thought to emerge as a response to competitive pressure: as populations multiply, less competitive individuals will, Durkheim (1893 [1984], 213) believes, pursue more specific niches “in which [they] can render a service.” Similarly, while Parsons consistently defines differentiation with reference to AGIL functions, and generally sees societies as equivalent to organisms, he also argues that one can, with the aid of the differentiation concept, see cultural history, “like the system of organic species, [as] an immensely ramified ‘inverted branching tree’ of forms …” (1977, 284; emphasis added). In both of these examples we find muddled analogies, and an inconsistent use of the mereological model. Differentiation is likened to speciation, but speciation is not a process of functional differentiation as these theorists define it, for species do not perform complementary functions for a broader totality in the way that physiological elements do in an organism. 11
While these discrepancies are largely confined to ontological matters, one can also find examples where functionalist theorists have, when offering empirical case studies of the differentiation process, shown little fidelity to their more explicit methodological commitments. Homans (1964), for example, famously argued that functionalist analyses of social change often forgo reference to abstract propositions about system dynamics when they come to the task of concrete causal explanation. For example, Smelser’s (1959, 80) analysis of the Industrial Revolution argues that there was, in early 1760s Lancashire, “excited speculation about instantaneous fortunes for the man lucky enough to stumble on the right invention,” and that said excitement spurned innovation, and eventually structural differentiation, in the cotton industry. As Homans (1964, 816-817) notes, this explanation “refer[s] to the behaviour of men and not the conditions of equilibrium of societies or other social groups as such.” It is, as such, distinctly genetic, in that it examines the specific actions of specific individuals in specific historical circumstances, rather than appealing to generalized laws about the behavior of social systems.
If theorists of differentiation have themselves constructed inconsistent positions, then it is little wonder that interpretive commentaries often conflate the two traditions outlined above. Schluchter (1979 [1981], 28), for instance, folds Weber’s account of the value spheres back into a functionalist register, arguing that there are, for Weber, “basic functions that must be fulfilled through differentiation,” and that one finds, in his account, an “elementary model of functions” which has certain “superficial similarities to Parsons’ AGIL scheme.” Likewise, Alvehus and Hallonsten (2022, 11), discussing a purported tradition of “functionalist differentiation theory,” note that “twentieth-century functionalism was developed into an elaborate theoretical framework by followers of Max Weber, and Weber provided it with its theoretical fundament.” As our above discussion demonstrates, such claims ignore the distinctive commitments which separate Weber from the functionalist tradition (for further discussion see Schwinn 2003).This not only gives a misleading impression of his position, but may also deter engagement with contemporary Weberian approaches, assimilating them to a functionalist tradition which—having had its core assumptions successfully problematized by decades of trenchant critique—has few contemporary adherents and little disciplinary purchase.
This is not to say that theories which cross these traditions should be ruled out entirely, nor that theorists working in one tradition cannot successfully import elements developed in another. As Laudan (1977, 94) readily admits, specific theories may break with the traditions that “initially inspired or justified them,” seeking legitimation in a new or alternative tradition. Parsons, for example, was indeed influenced by Weber, and explicitly took on elements of both his historical sociology (see Parsons, 1947) and his broader theory of action (see Parsons 1990). At the same time, the above comments miss the fact that Parsons was, for the most part, careful to assimilate Weber’s insights within his own particular set of metaphysical and methodological commitments, and remained committed to an organicist model of society which is entirely absent from Weberian thinking (Roth 1976). 12
Efforts of this kind remain common today. For example, Abrutyn, one of the more prominent contemporary advocates of the differentiation concept (see Abrutyn 2021), has recently proposed a sophisticated theory of institutional differentiation which explicitly aims to “rescue aspects of functionalism and its close cousins,” while also “revisiting Weber’s social psychological work on worldviews, ideas, and interests surrounding social orders” (Abrutyn 2016, 208-209). Abrutyn is, however, cautious to integrate these theories within a cogent set of metatheoretical commitments. For example, while Abrutyn is able to leverage aspects of Weber’s historical analyses, he clearly sees differentiation as a generic social process. As such, he not only considers how differentiation might ultimately be measured (see Abrutyn 2021, 152), but also seeks to develop a series of nomothetic propositions about its causes and consequences (some explicitly inspired by Spencer) (Abrutyn 2009; Abrutyn 2021). In addition, while he is careful to avoid the pitfalls of functionalist explanation (see note 4, above) he is clear that he continues to see institutions in functional terms, arguing that institutions “are adaptive responses to exigencies faced by all societies” (Abrutyn 2009, 450), that the universality of certain institutions reflects the fact that “all groups, regardless of size, face the same basic problems even if the contents of those problems in addition to their size and scale vary” (Abrutyn 2021, 151; see also Abrutyn 2016, 213-214), and that “differentiation is intimately tied to selection, or the pressure to resolve real, imagined, or manufactured internal or external exigencies through structural and/or cultural solutions” (Abrutyn 2021, 157; see also Abrutyn and Turner 2022).
In my view, such efforts demonstrate the right way to integrate elements from different traditions. While theorists can, as Laudan (1977, 103) notes, engage with more than one research tradition, they cannot accept traditions which are “inconsistent in their fundamentals” without raising “serious doubts about [their] capacity for clear thinking.” In other words, rather than working across traditions, successful synthetic efforts will generally incorporate insights from one tradition into another. For this, two major requirements must be met. First, the theorist must ensure that their tradition—regardless of whether it is an established approach or a novel construction—is “sufficiently rich conceptually, and its partisans sufficiently imaginative, to allow it to justify and rationalize theories which prima facie are more naturally related to very different … traditions” (Laudan 1977, 95). Second, they must recognize that while research traditions may, at times, show certain compatibilities, such that amalgamating them is a fairly straightforward manner, synthetic efforts will typically require a “repudiation of some of the fundamental elements of each of the traditions combined” (103-104). In the case of our two traditions, for example, it seems obvious that one cannot be both a positivist and a historicist (See Turner 2006); cannot believe that society is both a living system and that it lacks any “objectively delimited structure” (Roth 1976, 308); cannot believe that functional differentiation is both a universal comparative construct and a “historical-individual.” As Turner (2006, 262) has noted, there is “no easy middle ground” between these two explanatory strategies, and while there can be “cross fertilization” between them, our theories would be stronger if they were more clearly committed to one or the other.
The work of Niklas Luhmann, another prominent advocate of the differentiation construct, provides a clear example of this issue. At first blush, Luhmann seems quite clearly situated in the functionalist tradition. He is, for example, explicitly committed to a central element of functionalist analysis: the thought that certain entities—in his case functionally differentiated subsystems—are specialized in the performance of social functions (Luhmann 1997 [2013], 88). In addition, he remains committed to the interdisciplinary paradigm of general systems science, and, much like earlier functionalists, consciously borrows elements from theoretical biology. For example, his claim that social systems are autopoietic systems (systems which create their own elements and maintain their own boundaries) relies on a model initially developed to describe the self-maintaining chemistry of living cells (see Maturana and Varela 1972). Likewise, his claim that social change follows an evolutionary scheme (one of variation, selection, and restabilization) is an explicit generalization of Darwinian thought, indebted to the work of Donald T. Campbell (1960).
At the same time, Luhmann jettisons many of the functionalist tradition’s key ideas. Curiously, for example, he explicitly rejects the mereological model, repeatedly claiming that the theory of autopoietic systems replaces the distinction between part and whole with the distinction between system and environment (Luhmann 1984 [1995], 5-6). As such, Luhmann does not see functional differentiation as the specialization of distinct parts in a pre-existing, multifunctional totality, but as the process by which functionally distinct, autopoietic subsystems, each oriented to a specific “problem of society,” are created within existing societal systems (Luhmann 1984 [1995], 6-7, 53). With this comes a clear move away from the decomposition approach, toward what has been termed an “emergence” perspective (Schwinn 2011; Stichweh 2013). Unlike his functionalist predecessors, Luhmann does not define functional differentiation with reference to a static set of societal prerequisites, but rather considers the “historically variable” structures of societal systems, and the contextual nature of their problems (1997 [2013], 89; 1984 [1995], 54-57). In his view, functional subsystems (entities such as the economy, politics, art, law, science, or education) need not form by detaching from an existing complex, or specializing in some preexisting societal need, but can emerge via the synthetic combination of elements into a new, functionally distinct entity, oriented to a contingent, historically-specific problem (1997 [2013], 58, 89; 1984 [1995], 54-57). 13
In making this move, Luhmann must repudiate the major methodological commitments of the functionalist tradition. By historicizing the problem states to which the function term refers, rather than defining functions with reference to a theoretically deduced catalogue of social needs (see Luhmann 1997 [2013], 89; 1970, 10), he rules out the possibility that one could treat functional differentiation as a generic comparative variable. This repudiation is quite prominent in Luhmann’s work, which is notably dismissive of variable analysis (Luhmann 1997 [2012], 14); largely indifferent toward empirically-minded standards of scientific explanation (Luhmann 1986, 129; see also Herting and Stein 2007); and explicitly opposed to the idea that differentiation can be treated as a simple quantitative trait (1997 [2013], 14). Instead, his own analysis is distinctly genetic, in that it posits a catalogue of structurally distinct “forms” of differentiation (these being segmentary, center-periphery, stratificatory, and, finally, functional differentiation: the dominant form in modern societies), and offers a contextually-based, developmental account of historical transitions between them (Luhmann 1997 [2013]).
Luhmann thus presents a curious and genuinely novel combination of elements which crosscut the two traditions discussed above, in that he shows a clear commitment to systems-theoretical thinking, but entirely rejects the methodological ambitions which originally inspired this framework. Even on the metaphysical level, Luhmann’s depiction of functional differentiation resonates with elements of both traditions. In accordance with the functionalist approach, it retains the idea that modern societies are characterized by functionally differentiated subsystems, each addressing referential problems of a wider, container society. This is, however, combined with the thought—drawn from autopoietic theory, but highly reflective of Weberian thinking (see Kangas 2012; Tyrell 1998)—that these systems are self-producing, operationally closed, and incommensurate domains of meaning, which do not receive inputs or outputs, and cannot be directed by a surrounding environment, but autonomously produce their own internal elements (functionally distinct forms of communication) with reference to system-specific binary codes (such as truth/falsity in science or legal/illegal in law) (Bruun 2008). Thus, while Luhmann keeps one foot firmly in the functionalist camp, continuing to retain the holist assumption that functional subsystems are oriented to problems of society as a whole (Schwinn 2011), 14 his portrait of modernity is notably Weberian, in that he depicts a polycontextual world of autonomous, incompatible, and inwardly directed system rationalities, each offering an irreducibly unique way of perceiving and evaluating the world (Müller 1994; Rasch 2000; Tyrell 1998).
While this combination of metatheoretical elements is an importance source of novelty for Luhmann’s theory, it also generates a range of tensions, for his Parsonian-inspired analysis of functional subsystems is in constant competition with a more Weberian argument about the proliferation of unique life-worlds or domains of meaning (see Berger 2003; Kangas 2012; Schwinn 1995). This is, to a large extent, because of fundamental contradictions between the concept of autopoiesis—radically non-functionalist in its original formulation (see Mingers 1995, 38)—and the mereological ontology of functionalist thought. As a number of German commentators have noted, the autopoiesis paradigm, which Luhmann gradually adopted in the 1980s, was never fully synthesized with the functionalist foundations of his general theory, initially developed in the 1960s (see Borch 2012). Instead, the former was essentially overlaid onto the latter, with little consideration of key contradictions between them (see especially Berger 2003; Schimank 2005). As Kangas (2012) notes, subsystems cannot be both autopoietic and functional. Autopoietic systems are self-referential and self-producing: by definition, they can have no other purpose than their own self-regeneration (Kangas 2012, 68). A functional system, by contrast, would need to be allopoietic, bound to some particular outside purpose or overarching goal. There is thus a fundamental tension at the heart of Luhmann’s theory, which he continually struggles to address or resolve.
Luhmann’s basic problem is, in essence, that while he attempts to maintain a functionalist position, he is constantly bringing in elements from alternative traditions (or developing his own conceptual novelties) which undermine the logic of a functionalist approach. For example, Luhmann states that he wishes to retain the “‘holistic’ aspect of earlier systems theories,” and he continues to speak of system-level problems (arguing, for example, that differentiation means the “establishment of functionally specific differences within [a] system, to whose problems the functional mechanisms relate”) (Luhmann 1984 [1995], 53; emphasis added). As Schwinn (2003) has noted, any such method, i.e., any convincing attempt at functionalist explanation, requires an analytically consistent, “strong” notion of society. This is, however, precisely what Luhmann’s attempts at innovation tend to undercut. In earlier functionalist theorists, the concept of society was given shape and coherence by the decomposition paradigm, and by the delineation of an objective set of societal needs. Under this approach, there is a clear “whole” which differentiates: a preexisting unity that gradually splits into functionally distinct parts. By contrast, Luhmann’s historicization of functionalist analysis, based in an “emergence” perspective, explicitly orients functionalist thinking away from the question of societal prerequisites (see Luhmann 1997 [2013], 89), arguing, in a more Weberian mode, that problems are constructed within the communicative context of individual functional subsystems, each of which offer a unique way of understanding and accessing the world (Schwinn 2011). In this more constructivist frame, “society” is no longer seen as an objectively delimited entity, with a clear and static set of needs, but a subjective construct, perceived and defined within a polyperspectival framework of competing system rationalities (Schimank 2005; Schwinn 2001).
Ultimately, this position corrodes Luhmann’s commitment to a systems-oriented functionalism (see especially Schwinn 1995, 2011). Society becomes an ambiguous reference point, devoid of substantive content, and as a result, Luhmann’s various statements on the “functions” of functional subsystems are rife with contradictions and points of ambiguity. For example, in Die Gesellschaft der Gesellschaft, Luhmann’s most mature statement of his theoretical position, we are told that “functions should not be defined,” as was the case in “older sociological theory,” as “conditions for the continued existence of the societal system,” and that “it would not make an essential difference if we were to replace ‘continued existence’ by ‘autopoiesis’” (1997 [2013], 89). Yet just two lines later Luhmann seems to do exactly that, arguing that his approach to functionalism is based around “a sort of cognitive experiment” in which one tests “how the societal system would have to change its structures to maintain its autopoiesis if certain functions were no longer to be fulfilled” ([1997] 2013, 89; emphasis added). Likewise, while Luhmann consistently maintains that functional differentiation takes place within a system, with respect to problems of that system (e.g., 1984 [1995], 53), the various subsystems he describes do not all seem to operate for the same entity, with some (such as art, religion, or science) addressing problems of a postulated “world society,” and others (such as politics and law) remaining oriented to more distinctly national concerns (Münch 2013, 79; Schwinn 1995, 205).
These ambiguities cause persistent problems in Luhmann’s work. Because, for example, his functionalism is not disciplined by any underlying principle or deductive starting point (such as Abrutyn’s [2009; 2016] notion of basic human needs or ubiquitous human concerns), he is free to make a number of dubious claims about the “functions” of entities which—while easily describable in Weberian terms—strain the limits of the functionalist register. Art, for instance, can certainly be seen as a differentiated domain of activity—or, in Weberian language, as an autonomous value sphere—but to argue that it is also a functional subsystem, oriented toward a referential problem of society as a whole—is a far more demanding endeavor, not only because we are, under the emergence model, given little sense of what this whole actually is (and thus what problems it could have), but also because any attempt to specify the function of this domain sits uncomfortably with statements about its inner autonomy—autopoietic or otherwise. Luhmann claims, for instance, that the function of art is “to [demonstrate] the compelling forces of order in the realm of the possible” (1995 [2000], 148). It is, however, not only difficult to see what societal-level reference problem is thereby addressed (with Luhmann [2000, 207] himself asserting that the “collapse of art would not be very harmful for the total society”), but also hard to understand why art—given its supposedly autopoietic character—must be concerned with this particular task. In this case, as in many others, Luhmann’s functionalism simply gets in the way, and it seems clear that his analysis would be far more cogent if it was more thoroughly detached from this particular tradition of thought.
6. Conclusion
It is not possible here to give a full account of Luhmann’s theory, whose idiosyncratic range of influences, and general ambitiousness of scope, strain the possibilities of a brief summary. What is relevant for the present piece is simply the way that Luhmann’s theory can, by demonstrating the fundamental tensions that are likely to arise when theorists do not make a sufficient distinction between the two traditions of theory I have delineated, provide us with a cautionary tale. Although elements from one tradition can, as I have explained earlier, be grafted onto another, what makes Luhmann’s work struggle—particularly in comparison to some of the other contemporary options discussed above (such as those presented by Abrutyn or Schwinn)—is that it attempts to combine elements which are fundamentally incompatible on the metatheoretical level. While Luhmann remains committed to sociological functionalism, and to the interdisciplinary paradigm of systems science, he not only rejects the methodological positions that have historically justified these commitments, but also adopts a conceptual model of differentiated systems which, while not directly drawn from the Weberian tradition, is highly evocative of its alternative metaphysical position. The result is an uneasy synthesis of metatheoretical elements, and a conceptual framework in which the precise nature of differentiated entities, and the specific meaning of “functional” differentiation, is unstable and difficult to grasp.
From this comes a clear recommendation for the development of future theories. If we are to develop a cogent account of the differentiation process, we must ground our theories in a coherent research tradition, and be clear about our own metaphysical and methodological foundations. Likewise, we must remind ourselves that the articulation of a consistent position will, in many cases, require us to abandon ideas which are simply more suited to alternative lines of thought (Laudan 1977, 104). In Luhmann’s case, for example, it seems clear that the continued presence of functionalist elements is a constant source of problems. Uncoupled from a commitment to systematic comparison, and connected to novelties that push one away from holistic notions of society, the functionalist elements in Luhmann’s account create unnecessary friction, and make the fundamental weaknesses of their associated tradition more obvious to the critical observer. The theory would, then, be better served by discarding them entirely, and one sees this quite clearly in the fact that it is, in the contemporary German literature, the more action-theoretical approaches to differentiation, arising either as novel products of the Weberian tradition (e.g., Schwinn 2001), or as conscious attempts to—by jettisoning all reference to functions, and all but rhetorical reference to systems—fold Luhmannian theory into that tradition (e.g., Schimank 2011), which show greater coherence and theoretical promise.
Choices, then, must be made, and it is here that the research tradition framework can demonstrate its ultimate utility. To speak of different research traditions is not simply to offer an organizing principle by which to clarify our positions; rather it also opens up new avenues to evaluate bodies of thought. For the theorist, the construction of a typology or categorical scheme should never be an end in itself; rather one should, as Van den Braembussche (1989, 5) puts it, aim to construct a “relevant framework within which the question of the relative advantages and disadvantages of the different types can be formulated and discussed.” To speak of distinct research traditions is, as such, always to invite evaluative comparisons. If we are to follow Laudan’s account of the scientific process, such evaluations should ultimately consider the problem-solving nature of scientific inquiry: a fruitful research tradition is, under this framework, one which leads to the “adequate solution of an increasing range of empirical and conceptual problems” (Laudan 1977, 82). To abandon a research tradition is not, then, to judge it false, nor to “[relegate] it to permanent oblivion,” but merely to make a “tentative decision not to utilize it for the moment because there is an alternative to it which has proven to be a more successful problem solver” (83).
A detailed evaluation of the two traditions outlined above awaits future efforts, but it seems clear from the general direction of the field that the functionalist tradition has generated a host of conceptual and methodological problems, and that these problems have—despite the efforts of many innovative thinkers—proven largely insoluble. While it is, then, likely that further attempts to develop theories of functional differentiation will gravitate toward the Weberian tradition, with all the metaphysical and methodological loyalties this entails, we are also seeing the emergence of new constructions which, while not explicitly functionalist, can be read as attempts to refigure many of this tradition’s key ideals into a more workable and cogent form (I take Abrutyn’s [2009] work as an example of this approach). Though elements of the functionalist tradition will undoubtedly play a role in the development of these theories, and have shown promise when incorporated into alternative accounts (e.g., Abrutyn 2021; Schimank 2011), theorists must be conscious of the risks involved, and must locate and neutralize potential points of friction. 15
Ultimately, my attempt to distinguish different traditions of differentiation thinking is intended as an aid to this process: a tool to correct the ad-hoc, “eclectic pragmatism” by which theories are often constructed (Schwinn 2011, 41). While one can certainly take issue with many specific features of Laudan’s philosophy of science, and may disagree with the contours of the traditions I have constructed in this piece, it seems clear that commentaries on functional differentiation have created a significant stumbling block for theoretical development by improperly entangling distinct lines of thought, and that future efforts would benefit from greater attention to the kinds of metatheoretical distinctions which the research tradition concept compels us to consider. Although theorists may not wish to adopt Laudan’s exact formulation of this concept, I hope they will at least agree, with Schwinn (1998, 75), that “fragments from theories originating in different sociological traditions cannot be combined arbitrarily,” it being all too easy to create a construction with endemic incongruities, or to develop specific theories which, unmoored from a consistent set of metatheoretical commitments, come to lack purpose or a clear rationale.
If this is the case, then the delineation of traditions offered in this piece has clear value. By distinguishing the traditions in which theories of differentiation have thus far been developed, and drawing attention to potential conflicts in their metaphysical and methodological positions, we flag points of danger for future constructions, which will inevitably draw from earlier efforts, and which must be cautious not to combine elements that are fundamentally incompatible on the metatheoretical level. Knowing the tradition which we operate within, and ensuring that our claims do not appeal to a contradictory set of metatheoretical commitments, is essential to the development of internally coherent theory, and it is only in a theory of this kind that the concept of differentiation can, I believe, maintain its longevity, justify its utility, and live up to its postulated promises.
Footnotes
Declaration of Conflicting Interests
The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Funding
The author(s) received no financial support for the research, authorship, and/or publication of this article.
