Is a Post-philosophical Sociology Possible? Insights from Norbert Elias’s Sociology of Knowledge

1. Introduction

The origins of the sociology of knowledge are most commonly associated with the work of Karl Mannheim ([1929] 1960). ¹ His canonical text, Ideology and Utopia, was published in German in 1929, to widespread admiration for its boldness, but equally widespread criticism of its conclusions and implications. To many, including some of the most illustrious of his contemporaries, the very idea of a sociology of knowledge was highly problematic. Hannah Arendt ([1930] 1994, 42) spoke not only for traditional defenders of philosophy when she pointed out, in a highly critical review in 1930, that, “[b]efore we can pose Mannheim’s question of the social and historical locus of sociological inquiry, we need to inquire first into the existential situation in which sociological analyses are historically legitimate.” In a rare show of solidarity between existentialism and critical theory, hostile (and separate) reviews by Theodor Adorno ([1955] 1967) and Max Horkheimer ([1930] 1993) raised similar questions. ² The common theme to these objections is that a sociology of knowledge runs up against insuperable problems when it attempts to overstep a (variously defined) dividing line with philosophy because at that point it encounters a crisis of legitimacy concerning its own position. This objection has, in contemporary reflections, been re-conceptualized as a problem of relativism. Indeed, historically speaking, the really decisive blow against Mannheim’s vision for a sociology of knowledge came not from more sociologically minded contemporaries but from the philosopher Karl Popper ([1945] 1963), who, in his tour de force against the enemies of the “open society,” declared it to be moribund on the basis of being self-defeating. For Popper ([1945] 1963, 216), sociologists of knowledge convert all questions of truth into questions of ideology and class interest but will always “produce an amendment to [their] theory in order to establish the objectivity of their own views.” Ever since, relativism has come to be seen as the central and insuperable problem with the sociology of knowledge, and the key objection to its various reincarnations in such forms as the strong programme of the Edinburgh School and feminist standpoint theories. ³

In this article, I examine one of the more notable attempts to put the sociology of knowledge on a “scientific” footing, namely, that provided by Norbert Elias, a former student and colleague of Karl Mannheim, who went on to achieve prominence as one of the foremost European sociologists of the twentieth century. ⁴ Elias was an indefatigable defender of the autonomy of sociology and hostile to what he perceived as the common Kantian element uniting acknowledged neo-Kantians such as Heinrich Rickert and Richard Hönigswald (important early influences against whom Elias rebelled), with their supposed opponents such as Arendt and Popper: a delusional conception of philosophy as both the underlaborer and the organ of legitimation for the sciences. He suggested that, instead of engaging philosophers on their own ground, sociologists would be well served to simply break off contact with them and pursue epistemological clarification of their discipline from within their own ranks, using their own methodologies. Elias’s contemporary defenders have mostly maintained his insistence on a full declaration of independence for the social sciences, and Richard Kilminster (2007) marshaled the full range of Elias’s arguments to this effect in his Norbert Elias: Post-philosophical Sociology.

Nevertheless, as I shall argue below, Elias’s averred declaration of independence from philosophy is not simply a diktat, but based on principled objections to the relationship between philosophy and sociology that was assumed to exist among orthodox philosophers of his own time. Elias believed that, from a pragmatic point of view, it was of no avail to continue a conversation with those who rejected, or indeed refused to recognize, his starting premises. But this does not imply that Elias’s premises would not be comprehensible and disputable on principled grounds under a different philosophical worldview. However one dates, disagrees with, or explains, the validity of Elias’s picture of the orthodox philosophical self-conception of his time, no doubt it has passed. ⁵ This opens up the possibility of re-examining the grounds for Elias’s post-philosophical move from a perspective that is not fatally tainted by the Kantianism which he reviled. Elias also has much to say concerning the themes of social reality, the unity of the sciences, and the role of language in human thought, all of which constitute long-running themes in the epistemology of the social sciences. Clarifying Elias’s positions on these issues can provide epistemology with a new perspective on the social sciences, as well as new resources to rethink the underlaborer model of the relation between philosophy and the disciplines.

The article proceeds as follows: First I discuss the connection between Elias’s better known work on the “civilizing process” and his later sociology of knowledge. I then discuss two philosophical objections to his approach that touch on questions of whether sociology can provide solutions to philosophical questions. I conclude with some reflections on the extent to which a post-philosophical sociology of the kind defended by Elias is feasible.

2. Process-Sociology and Knowledge

Elias is one of the few sociologists to have advanced the discipline in three quite distinct areas. First, he produced a number of empirically rich studies tracing social changes in many areas, including, for example, the long-term development of state structures, shifts in the balance of inequality between the sexes, and the changing functional role of sport in culture. Second, he developed an elaborate sociological theory to guide these studies, in which his distinctive concept of a social process was a central element. Third, he intervened in important debates within what has been seen as the philosophy of social sciences, but which he sought to address “from within,” as sociological questions themselves. His sociology of knowledge therefore has a circular quality, insofar as he sought to explain the social conditions of its own development. It is this explanatory feature that constitutes the core of Elias’s attempt to respond to the problem of the legitimacy of the sociology of knowledge raised by both its early critics and contemporary conventional wisdom within philosophy, which has tended to cast the problem in terms of relativism.

The idea of justifying sociology with reference to its own historical conditions arguably goes back to Comte, who in turn was drawing on Hegel’s speculative philosophy of history, which is supposedly self-grounding. Marx radicalized this feature by according world-historical significance to his own understanding of world history. Nevertheless, it was Mannheim who provided the most developed formulation of this thesis (cf. Kettler, Meja, and Stehr 1984, 57-58). The idea of a self-grounding social science involves, fundamentally, the rejection of the underlaborer model of the relationship between philosophy and sociology. On the underlaborer model, the validity of the social (and natural) sciences presupposes a philosophical account of the structure of the reality within which they work, and an account of how their theories and concepts hook onto this reality. Whether one conceptualizes reality and its relation to thinking in terms of the vocabulary of fundamental ontology, as language, or as a reality and its “mirror,” the general model of philosophy as underlaborer remains the same. But for Mannheim ([1929] 1936, 404), this image is inherently problematic because it assumes the absence of “existential boundedness” in thinking, that is, philosophy cannot evade the same attempt to exercise power and influence in the world that mark all intersubjective human relations, and which in the field of knowledge becomes a struggle over “the public interpretation of reality.” The underlaborer model fails then, on Mannheim’s account, because it refuses to recognize that knowledge does not stand outside the reality that it seeks to know, but is itself a product of struggles within that reality. This position, of course, leads Mannheim to his relativistic difficulties. However, it is worth noting that the charge of relativism, pace Popper, or of illegitimacy, pace Arendt, both take aim at the conclusions Mannheim draws from these reflections, namely, that sociology is capable of discerning the structures of power that “existentially bind” knowledge (including philosophy), but they do not explicitly confront his critique of the underlaboring model.

Elias was close to Mannheim and, though the junior partner in the relationship, influenced the development of Mannheim’s own sociology of knowledge project (Kilminster 2007, 40-71). But it was only in 1956—nine years after Mannheim’s death—that Elias began to publish material devoted explicitly to his own perspective on the sociology of knowledge. The most important of these texts is the two-part book-length article, titled “Problems of Involvement and Detachment,” which first appeared in the British Journal of Sociology. By this time, Elias had already published what became known as his magnum opus, The Civilizing Process, although the book had not attracted the widespread attention it subsequently achieved. In this book, and his other empirical writings, Elias presents many events in European and world history as related sequences that demonstrated an overall change in a particular direction, that is, as part-processes that together form an overall process. In his sociology of knowledge, he extends the scope of this project by conceptualizing the development of knowledge as a long-term historical process, intimately connected with those analyzed in The Civilizing Process. He includes the rise of both the natural and the social sciences in this analysis, seeking to explain, not the meaning of concepts such as causation and scientific explanation as (arguably) had been the principal aim of the underlaboring model of philosophy, but what long-term sociohistorical forces had allowed these concepts to appear as meaningful to those who used them in their language and actions. In so doing, he may be said to have abandoned Mannheim’s relativism in favor of perspectivism (Kilminster 2007, 50). However, Elias’s perspectivism does not seek to “relativize” the natural or social sciences, but to demonstrate how they are themselves part-processes that, together with others, comprise together the “civilizing process.” Given its centrality to both Elias’s general perspective, and to his specific claims about knowledge—including how we should understand the differences between the social and natural sciences and their relations to philosophy—it is worth looking more closely at his concept of a social process in general.

The concept of a process is ubiquitous within Elias’s work. His empirical studies are conceived of explicitly as studies of what he calls part-processes and he sometimes characterizes his own distinctive approach as “process sociology” (see Mennell 1989, 252). Nowhere, however, does Elias explicitly defend his usage of the term, although he consistently suggests that the success of the natural sciences depended on their becoming both sciences of processes and processes themselves (Elias 1987, 123-4; cf. Dunning and Hughes 2013, 50; Mennell 1989, 177-179). The (fewer, by comparison) successes of the social sciences are dependent on their viewing themselves the same way. To my knowledge, no one—including Elias—has traced the actual history of the transference of the concept of a process from the natural to the social sciences. That such a transference did take place, however, is undeniable, and occurred following the recognition of the enormous time spans involved in the explanations offered by the physical, geological, and biological sciences. The explanations developed within these sciences—of the evolution of life-forms over millions of years, and of the changes affecting the surface of the Earth and the universe as a whole over billions—became exemplars of processes. The early social sciences borrowed the language of processes explicitly to extend the historical horizons of their disciplines. ⁶

Elias demonstrates the use of this concept most famously in his classic, The Civilizing Process ([1937] 1978, [1937] 1982). Here, he is concerned with the examination of apparently trivial alterations in features of everyday life in European societies that helped to bring about larger scale changes that are among the most momentous in world history. In the first part of the book, Elias discusses the histories of table manners, patterns of ceremony within Aristocratic courts, and attitudes toward bodily functions. Over the period from about 1300 to 1900, all exhibited long-term alterations in a discernible direction. Manners became more elaborate, signifying a gradual and greater orientation of people to the consciousness of others’ perceptions of them. Forms of interpersonal address between social unequals became less formal, signifying a gradual overall increase in interdependence between groups. Attitudes toward bodily functions that people share with animals went in the direction of an “advancing threshold of shame and disgust,” indicating an increase in the degree of self-restraint and control of both bodily processes and social affect.

These developments, he argues, combined to form, eventually, a social habitus within a social figuration. A habitus refers to a socially expected set of preconscious attitudes toward things and others that become embodied in people’s unexamined attitudes, and eventually normalized for individuals within an entire figuration. A figuration refers to the overall set of relationships and interdependencies within which people find themselves embedded ⁷ :

[Gradually], a social apparatus is established in which the constraints between people are lastingly transformed into self-constraints. These self-constraints, a function of the perpetual hindsight and foresight instilled in the individual from childhood in accordance with his integration in extensive chains of action, have partly the form of conscious self-control and partly that of automatic habit. They tend towards a more even moderation, a more continuous restraint, a more exact control of drives and affects in accordance with the more differentiated pattern of social interweaving. (Elias [1937] 1982, 374)

In other words, a combination of part-processes produced eventually a more even, steady, time-conscious and self-aware social habitus existing within a distinct figuration, to which new generations oriented themselves and then reproduced. This shift aligns with more general “macro-level” changes that Elias discusses in the later part of the book. These changes revolve around the retreat of overt violence from everyday life, and the overall tendency toward functional democratization.

In “Problems of Involvement and Detachment,” Elias explicitly takes up the problem of how to connect changes in knowledge practices with the changes associated with the civilizing process. He makes use of a longer time-scale, including in his analysis the earliest tool-using human societies, but focuses mostly on the “spurt” of (relatively) rapid changes in people’s attitudes toward nature that followed the discoveries of Galileo, and which were subsequently disseminated within the natural sciences, aesthetics (especially painting), and philosophy. That such attitudinal changes were integral to both the processes of “civilization” and to the growth of scientific knowledge is implied in the meanings of the terms “involvement” and “detachment.” The decline of involvement means a lessening of the tendency to think emotively about objects in the natural world as personified, purposive, incalculable forces that are first and foremost personally threatening or fearful. The decline is complemented by the growth of detachment, that is, the ability to see natural objects or forces as autonomous from humans, morally neutral, and, ultimately calculable and manipulable by human action. The process of the changing balance between these two attitudes is dependent on the development of the social habitus described above—one that tends toward a “more even, continuous constraint.” This attitude then prescribes people’s affectual relations to nature and natural objects, and leads to the development of the natural and later the social sciences. But the involved attitude is never entirely left behind, and indeed inhabits many of our modes of speech and basic categories in complex and often misleading ways.

The arguments of “Problems of Involvement and Detachment” were refined in Elias’s later writings, especially What is Sociology? (1971) and The Symbol Theory (1991), into a fully fledged “post-philosophical sociology.” In these later works, Elias developed a series of concepts that were designed to occupy the spaces of adjudication and legitimation of both validity and explanatory adequacy that had been developed by traditional philosophy. Such concepts as “sociogenesis” instead of causation (Elias 1971, 162-63) and “reality-congruence” instead of truth (Elias 1987, 37) were added to the conceptual armory intended to replace philosophical concepts. ⁸

3. Objections and Problems

Elias’s conclusions about the civilizing process and its effects have been subject to a great deal of misinterpretation. His sociology of knowledge has also been tainted—though more by association than by substance—by these, so it is well to dismiss some of the cruder ones initially. Elias does not claim, pace Steven Pinker (2011), that the changes associated with the civilizing process were the outcome of a “coherent philosophy” of the Enlightenment. Rather, human beings have become subject to greater restraints on their impulsive actions, Elias argues, through their interdependence with others. The growth of detachment, which allowed the development of the natural sciences, cannot therefore be identified with some Enlightenment conception of reason. Indeed, Elias explicitly avoids this term, the meaning of which he thinks has been established largely to justify the ideological imperatives of the present rather than the facts of the past. It is also misleading to conceptualize the relationship between the civilizing process and the growth of knowledge as in any way causal. As Elias rejects the usefulness of the concept of causation more generally within the social sciences, this is part of a larger dispute that Elias has with conventional social science itself.

Neither is the civilizing process a teleological theory, in the Hegelian or Marxist mode. ⁹ Elias explicitly points out that the changes that led in the direction of self-constraint, democratization, and detachment were completely unplanned and their later developments unknowable at an earlier stage. Moreover, de-civilizing processes also occur and are in no way less likely to be dominant over a long-term course of events.

Finally, Elias’s theory has been cast as irretrievably Orientalist, given the associations of the concept of civilization with the European history of colonialism (see Goody 2006, 154-79). Elias was fully aware that the concept of civilization had been used as a means to legitimate the domination of colonized peoples by their colonizers. He chose the term for three primary reasons: first, because it already had a strictly technical meaning in the German sociological tradition in which Elias was schooled, as a term contrasting with Kultur, referring to changes derived from primarily economic factors, which Elias was concerned to preserve (though seeking to detach it from its primarily economic meaning). Second, Elias chose the term precisely to counteract the myths that European peoples had been telling themselves about their role as “civilizers” to “primitive” peoples. By recovering the history of civilization, he sought to de-naturalize the elements that Europeans took for granted about themselves (cf. Kilminster 2007, 74-75). Third, Elias was deeply influenced by Freud, and particularly his perspective in Civilization and its Discontents. Like Freud, Elias maintained a neutral detached attitude about the so-called benefits of civilization. He made no judgments about its inherent worth relative to other historical or cultural forms of social life.

Having cleared away some of the simpler objections to Elias’s general approach, I now wish to consider two more serious ones. Both of these derive from the meaning of the concept of a process and have implications for Elias’s vaunted shift toward a post-philosophical sociology. A first objection relates to the excessive objectivity suggested by Elias’s use of the concept of a process. ¹⁰ Michael Oakeshott (1975, 24) provides a classic version of this objection, insisting that the concept of a process in the social sciences is fundamentally misleading. He suggests that the term applies only to natural or, in humans, to physiological sequences of events, and therefore its importation into the social sciences is a fundamental “category error.” For Oakeshott, human beings in their social lives do not engage in processes, but in practices; to imply otherwise is to strip human practices of their essential character, and to deny human agency to a greater or lesser degree. Collingwood (1946, 215) presents another version of this objection in terms of the contrast between the inside and the outside of an event:

The processes of nature can therefore be properly described as sequences of mere events, but those of history cannot. They are not processes of mere events but processes of actions, which have an inner side, consisting of processes of thought.

Oakeshott and Collingwood are concerned, in other words, about the problem of human agency, which in contemporary terms, has come to be seen, I argue below, as bound up with the problem of personal reflexivity.

Second, there are problems about how one establishes the conceptual or descriptive range of an Eliasian process. In other words, there are difficulties in justifying one particular frame of reference in social explanations rather than another. This is not simply a problem of establishing precise enough terms to capture what is being described, but, raised to a sufficiently high level of abstraction, becomes a problem of explanation per se, because frames of reference at a certain level compete with each other. For example, as I shall argue below, there are disagreements about the scope of such processes as secularization that appear to be irresolvable, taken in their own terms. This raises problems with how we should understand the metaphorical content of descriptions of processes. Alfred North Whitehead (1926, 11) termed this the problem of “misplaced concreteness,” and it long traveled under the moniker of reification, although the conceptual over-flexibility of that term led to problems of its own. I shall therefore hew more closely to Whitehead’s term and treat it as a problem of explanation that raises relativistic difficulties for Elias.

4. Agency and Process

Elias believes that philosophy tends to get stuck in dualisms that are more often than not unhelpful in working out problems for sociological theory. The problem of agency would appear to be a cardinal example of this. In the terms of the philosophy of his day, it appeared to him as an outmoded Kantian either/or of freedom and determinacy. Neither Oakeshott’s nor Collingwood’s more nuanced elaborations alter this picture, as both couch their claims in terms of binaries between events/processes and actions/practices. Their claims imply only that people, as individuals, reflexively make choices at particular junctures that, ceteris paribus, could have been, counterfactually, different.

In contrast, Elias argues that the widespread experience of oneself as a willing agent, of being able to think about the actions one carries out as “choice”, is restricted in time and place. People do not always have such self-experiences, and the fact that the capacity to have them is very widely disseminated in contemporary societies is a function of the social habitus. Put another way, the capacity to engage in the abstracting activity that allows oneself to appear to oneself as an agent capable of acting either in this or that way is made possible by a particular set of social arrangements. Elias points out that historical spurts in the direction of more people experiencing themselves and acting in this way can be detected with reference to certain historical indices, including, for example, the rise of novel reading in the seventeenth century, and the popularity and resonance of existentialist philosophy in the mid-twentieth century. These point to the novelty of this mode of self-experience for relatively large numbers of people at this time. In other words, the individual “will” which has come to be seen both as an individual possession and a human universal, on the Kantian picture, is neither. It becomes available to people only within certain social figurations that are closely linked with the increasing capacity for detachment and self-distancing. Elias’s sociology of knowledge therefore contains the problem of agency, as a particular question about the effects of the spread of a particular kind of self-experience.

Nevertheless, Elias’s perspective is vulnerable to a further objection that vitiates the Kantian framework which he opposes, but which does not justify a post-philosophical sociology. This objection has been advanced in the work of Margaret Archer (2003, 6), who points to the way talk of processes blurs the distinction between “the causal powers of structures from the causal powers of persons.” Indeed, the causal powers of persons and the causal powers of structures are not simply “divided,” they are fundamentally different. Beliefs, intentions, and deliberations exist on an entirely distinct ontological plane; personal reflexivity resides within the same plane, both as the self-relation necessary for consciousness as such, and as the thought-act by which the contents of one’s consciousness can be actively examined, understood, and acted upon. Archer (2010a, 7-8) has extended this criticism successfully to a broad range of sociological theorizing and empirical work that, according to her, denies the “reality” of thinking in one way or another. Her general point is that the absence of a significant ontological distinction between agency and structure directly implies a failure to take account of, let alone theorize adequately, the role of individual reflexivity in action. Although Archer does not explicitly mention Elias, her critiques of Bourdieu and Giddens (a former colleague of Elias’s) also apply to him. Process-theories, such as structuration and Bourdieuian practices, lack an adequate ontology, and therefore underplay the significance of agency and reflexivity. ¹¹

Recently, Archer (2010b, 279) has presented this point in terms of the contrast between reflexive and habitual action, which augments Weber’s typology of social action. Habit, or “relatively unmotivated” action tends to be the predominant mode of thinking and acting where morphostatic social conditions prevail, that is, relatively undifferentiated hierarchical societies with “a high and lasting degree of everyday contextual continuity for the populations in question: repetitive situations, stable expectations and durable relations” (Archer 2010b, 279). Under the opposite conditions—of “morphogenesis”—reflexivity is enhanced and social repertoires are exposed to creativity. Individual agency, in other words, for Archer and for Elias is made possible by particular social arrangements.

However, in contrast to Elias, Archer’s framework for conceptualizing the relationship between structure and agency entails a set of ontological claims, which are very much in the tradition of the philosophical underlaborer model of the relations between philosophy and social science. Although certain social conditions foster reflexive relations “within” agents, explaining any given case of reflexive (non-habitual) activity must also make reference to the causal powers of the agent, and these in turn must comprise a distinct set of properties, such as volition and the capacity to will, that can only be understood in their own terms. Denying or reducing the causal powers of people to the causal powers of structures (or figurations) commits what Archer has called the fallacy of “central conflation” of identifying the two. This is in contrast to downward conflation (voluntarism), and upward conflation (reification), but is still a fallacy.

It follows from this that the capacity to will may be a human universal, but this capacity is not universally exercised, and may be “unknown” under certain social conditions. Although Archer does not consider this limiting case, it is worth considering one piece of evidence in its favor. This is the acknowledged fact that philosophical discussions of the will are largely confined to modern philosophy. This suggests that the modes of self-experience which we characterize in terms of “will” were rarer or different in pre-modern societies from our own. This was Hannah Arendt’s conclusion, in The Life of the Mind, when she notes that one finds no straightforward correlate for the word “will” in Ancient languages. The Aristotelian term proaeresis conveys deliberation and confrontation of alternatives, but not some internal self-experience that could correspond to our notion of individual will. Therefore, she writes:

Proairesis, the faculty of choice, one is tempted to conclude, is the precursor of the Will. It opens up a first, small restricted space for the human mind, which without it was delivered to two opposed compelling forces: the force of self-evident truth, with which we are not free to agree or disagree, on one side; on the other, the force of passions and appetites, in which it is as though nature overwhelms us unless reason “forces” us away. But the space left to freedom is very small. We deliberate about means to an end that we take for granted, that we cannot choose. Nobody deliberates and chooses health or happiness as his aim, though we may think about them. (Arendt 1971, 62)

Arendt goes on to argue that the capacities to think, to will, and to judge are each themselves reflexive powers that are historically variable. She seeks, in the first two volumes of The Life of the Mind, to provide a phenomenology of their various forms and properties. The “I-will” becomes fully developed only with the advent of Christianity, and finds its historical expression in the philosophy of St. Augustine.

Arendt’s (1971, 5) claim that the “I-will” names a human capacity that “did not always exist,” but which came to be developed in the context of particular social/spiritual relations, is no doubt open to contestation. Nevertheless, understood as a limit case, it suggests that the concept of agency is not timeless, as both the Kantianism of Elias’s time and the perspectives of Collingwood and Oakeshott would suggest, but is nevertheless ontologically distinct. Because the causal powers of people and of structures are distinct in terms of both their particular qualities, and in their modes of existence (cf. Searle 1998, 44), they presuppose an ontological level of inquiry that must be discursively distinct from the empirical investigation into how these powers intertwine in any given, historically embedded event. One need not be committed to a “transcendental” realm of inquiry to insist that differentiating the distinct reality levels that underlie social action cannot be “post-philosophical” in the way Elias supposes.

5. Description, Process, and Language

I turn now to the problem of the conceptual range of process terms in the social sciences. One can address the problem of how a concept fits an object at different levels. Elias treats it primarily as a problem of language, but deals with it at both a fundamental and a concrete level. At both levels, the question of the extent to which metaphor plays a role in language—both technical and ordinary—is in play. Again, as a discipline, philosophy perennially finds it difficult to deal with this question because of a professional polarization between the logical approach that has dominated within the Anglo-American “linguistic turn,” which may be said to be oriented to attempting to eliminate metaphorical elements in language, and “deconstruction” or “genealogy”, which suggest that all meaning is metaphorical, and aim at historicizing and relativizing all meanings within the sciences.

Elias cites communicability as the “meaning of meaning,” and that languages should be understood as dynamically changing, long-term historical processes. Our use of particular words and sentences, both in everyday and technical languages, is therefore dependent on the background web of meaning that constitutes the language as a whole, together with its history. Elias insists that one cannot do without metaphors in either the social or the natural sciences and that the choices of metaphorical vocabulary shape the development of the field. The generation of new terms and concepts within a science is therefore a fraught and difficult activity. When a new object of investigation comes to light, it does not come with a ready-made label tied to it, which the investigator can simply read off and communicate to the scientific community. For example, as the historian of science, Evelyn Fox Keller (2000) has argued, the concept of the gene has never had a stable meaning, and in its current usage is being stretched beyond its capacity to convey clearly what it is supposed to refer to. But concepts may also enter a discipline in other ways, most commonly perhaps by being borrowed from other disciplines.

The term process itself, as discussed above, depends for its meaning on certain associations. These associations are, I think, actually at odds with the dominant vocabulary of the contemporary social sciences, which eschew organic for architectural elements. This is apparent in the profusion of terms like building, development, construction, maintenance, and, most ubiquitous of all, structures, that pervade contemporary mainstream sociology. The term process is often used in studies of social change, but is often paired with the concept of a mechanism. The emphasis is almost exclusively on static rather than dynamic terms. This suggests the persistence in the social sciences of process reduction, a critical term that Elias levels both at sociologists, such as Talcott Parsons, and also more broadly at the philosophical tradition of thinking in terms of binaries such as substance and accident and subject and predicate. The general point is that such thinking tends to conceptualize objects as things rather than as developments, and thereby to mystify their origins and histories. Sociologists have generally addressed this as the problem of reification, although the term suffers from being overdetermined. ¹² Nevertheless, most of those theorists who have taken the problem seriously have also sought to solve it from the standpoint of the philosophical underlaborer. In other words, they have accepted the view that we need an epistemological account of the meanings of such terms as society or race or value to avoid attributing a misplaced thingness, or even an agency, to these concepts. Elias’s sociology of knowledge provides an alternative strategy: seeking the origins and histories of the metaphors of the sciences reveals how they have come to be used, rather than their “essential” meanings.

Of course, there is a venerable tradition of such histories, from Nietzsche’s genealogy to Foucault’s “archaeological” studies. The distinctive element of Elias’s approach is the length of his time-scale and his focus on how grammatical rules shape meaning. In his late work, The Symbol Theory, Elias argues that disputes over meaning can be explained in part by the historically dynamic character of the grammatical rules governing the way we conceptualize the world. He was impressed by the work of the anthropologist Benjamin Lee Whorf in this respect. Whorf ([1937] 1956) pioneered the idea that the relationships between words laid down by grammar rules affect meaning at a very fundamental level, and that European languages are particularly prone to misrepresenting everything in terms of an idiom of subject and action. As examples of Whorf’s thesis, Elias cites the sentences “The wind is blowing” and “The river is flowing.” He points out that the use of the words “river” and “wind” are referred to as if they are entities that act. Their connection with a verb such as flowing and blowing then appears—once we reflect on it—as redundant. As Elias (1971, 112) asks, “Can there ever be a wind that does not blow or a river that does not flow?”

In The Symbol Theory, Elias refines these ideas. He does not accept the so-called Sapir–Whorf thesis that differences in grammatical rules between languages produce incommensurable forms of experience specific to the groups using those languages. Such rules do not determine meaning, but they constrain or limit it in ways that may be more or less resistant to reflection and correction. But the fact that our languages are constructed in a particular way reflects their earlier formation. The subject–action conjunction reflects the fact that in the earliest human societies, the involved attitude was dominant, and natural objects and processes were perceived as personified. All events, as Elias describes it, were understood by the groups that experienced them, in terms of their significance for “us.” “Purposiveness” was not yet perceived as something that applied to humans and not to natural processes. Indeed the distinctions that we take for granted, and apply as second-order categories of experience, between inert, living and human were not yet developed. ¹³

Thinking in terms of processes, then, is intended to combat static ways of thinking in the social sciences. Once an object is named and enters into the canonical vocabulary of a discipline, it is difficult to discern its origins, and to see it as a coined concept that is only more or less distinct from other things, rather than as a “thing” that the concept picks out sharply and uniquely. In Elias’s vision, the sociology of knowledge includes the study of meaning associations deriving from the usage of concepts, as an antidote to process reduction. The detachment of the social scientist is therefore both an outcome of the long-run development of meaning associations, and the condition that allows her or him to discern this development. This is, in other words, a re-worked vision of the sociology of knowledge that, like Mannheim, is based on a suitably de-mystified Hegelian self-completing account of history.

Elias’s distinctive, long-run approach to the problem of description is both original and fruitful in many ways, including its potential as an antidote to misplaced concreteness. Nevertheless, it has repercussions for his strategy of treating his own descriptive techniques as unproblematic. This points to a particular problem of relativism, originally identified by Max Weber that, I argue, Elias cannot evade. To show this, I discuss first how Weber set up the problem, and then use the example of the term secularization to show how it constrains Weber’s own historical analysis and clarifies the problem with Elias’s averred “post-philosophical” move.

Weber distinguishes between value-freedom (Wertfreiheit) and value-relevance (Wertbeziehung). The researcher, he says, can attain value-freedom in the sense of freedom from bias—or detachment, in Elias’s terms—in how he or she treats a particular problem, and this is a core component of an objective social science. However, values enter into social scientific explanation at two other, more fundamental, levels. These comprise (1) the selection of objects that the social scientist considers relevant or important to a problem, and (2) the concepts he or she deploys to capture them. At both these junctures, value-relevance guides the explanation, and in both cases. such value-relevance is, ultimately, relativistic. In clarifying the implications of value-relevance, Weber focuses explicitly on (1), most famously in his late essay, “Science as a Vocation.” Here he insists that the scientist must simply presuppose that “what is yielded by scientific work is important in the sense of ‘worth being known.’ In this, obviously, are contained all our problems. For this presupposition cannot be proved by scientific means” (Weber [1919] 1946, 150).

It is possible to re-interpret and/or contain Weber’s explanatory relativism, by speaking of, for example, “pre-theoretical commitments” or the “theory-ladenness of observation”. Alternatively, one can revert to some trans-historical/cultural account of rational explanation (Hollis 1982; cf. Runciman 1969, 157-60). Nevertheless, the idea that the “existential boundedness” of explanation can be simply dispensed with was rightly denied by Weber and, as we saw, later exploited by Mannheim’s sociology of knowledge. Elias also came to accept it as bound up with his perspectivism. Nevertheless, given the choices Elias made about what he thought were the important things to study—manners, rather than statecraft, chains of interdependence rather than patterns of religious belief, forms of grammar rather than logic—he seems to be particularly vulnerable to the problem of the value-relevance of his choice of objects.

But explanatory relativism also infects description. This problem is best captured by drawing attention to the existence of what W. B. Gallie (1955) dubs “essentially contested concepts”. To illustrate this, it is worth referring back to the first explicit conceptualization of a series of events as a “social process” within classical sociology, namely, what Comte sought to capture under his “law of the three stages,” and which subsequently came to be characterized under the moniker of “secularization.” Conventional wisdom suggests that this means that religious belief declined in importance in people’s mental lives over a period roughly between the seventeenth century and the present, being replaced by other beliefs of a more “scientific” kind. But closer examination of the explicit meaning of this “process” suggests that its meaning is much more complex and essentially contested. Comte’s ideas were refined by Durkheim who proposed a vision of a transition from a society where the notion of sacredness was strongly associated with “society” (in the amplified version of this concept that Durkheim advanced), to one where it attaches to individuals. He aligns secularization with the advance of individual human rights, the eclipse of traditional churches, and an emerging cosmopolitanism (see Durkheim [1893] 1984, [1957] 2003). In contrast, Weber suggests instead that modern belief is increasingly characterized by “disenchantment.” Weber ([1919] 1946, 129) understands this as a process, but one that encompasses secularization, and is characterized by the increasing withdrawal of inexplicable magical forces from people’s experience and explanations of the everyday world. These differences reveal not only irresolvable divergences over the meaning of the concept of secularization but also the provenance of Elias’s own conceptions of the advance of detachment as a part-process. The salient feature of an essentially contested concept is that there is no court of appeal to which reliable appeal for adjudication may be made. In other words, everyone agrees that “secularization” captures something important about a process of social change, but no one is able to define the concept with sufficient precision to fix its meaning.

As concepts, the existence of essentially contested concepts is entirely consistent with Elias’s long-run anthropological account of the “meaning of meaning.” But—if I may express it this way—their essential contestedness is not consistent with the conception of sociology as post-philosophical. We can put this point another way by saying that post-philosophical sociology disallows conceptual or theoretical pluralism because it rules out tout court, the possibility of philosophical contestation of the adequacy of description, which has rightly been regarded as a mainstay of philosophical reflection. In this respect, sociology is no different from other disciplines, which have drawn on the underlaboring power of philosophy to clarify or (often perennially) contest the meaning of their concepts.

6. Conclusion

Elias sought to free sociology from the influence of philosophy altogether, insisting that its epistemological and ontological problems could be successfully countered from within the discipline of the sociology of knowledge. In this article, I have shown how Elias’s ambitions for the sociology of knowledge present credible answers—or at least first steps toward answers—to some questions that have been thought to be soluble only within the philosophy of social science, but that this does not obviate the need for a philosophy of sociology that operates along the lines of the underlaboring model. I have pointed to two areas in particular where Elias’s sociology of knowledge runs into difficulties: in its failure to establish an effective ontology that takes account of the distinction between structure and agency, and its intolerance for the theoretical pluralism implied by the existence of essentially contested concepts. This does not, contra its various philosophical critics, imply a problem with the very idea of a sociology of knowledge, but suggests limits to its ambitions.