The Nexus of Practices: Connections, Constellations and Practitioners

Thomas Alkemeyer is professor of sociology and sociology of sport, and director of the research centre 'Genealogy of the Present', at the University of Oldenburg, Germany. His research foci are theories of practice, sociology of the body and practices of subjectivation.

Stanley Blue is a lecturer in sociology at Lancaster University, UK. His work traces the reproduction of everyday practices that matter for sustainability and health. His current research examines the temporal organisation of working practices in large institutions.

Nikolaus Buschmann was post-doc at the DFG (German Research Foundation) Research Training Group 'Self-Making. Practices of Subjectivation in Historical and Interdisciplinary Perspective' and is now member of the research centre 'Genealogy of the Present' at the University of Oldenburg, Germany. His main areas of research are Modern German and European history, theories of practice and subjectivation, and the cultural history of politics.

Silvia Gherardi is senior professor of sociology of work and organisation at the Faculty of Sociology of the University of Trento, Italy, where she founded the Research Unit on Communication, Organizational Learning and Aesthetics (www.unitin.it/rucola).

Allison Hui is an academic fellow in sociology and a researcher within the DEMAND Centre at Lancaster University, UK. Her research examines transformations in everyday life in the context of changing global mobilities, focusing particularly on theorising social practices, consumption and travel.

Cecily Maller is a vice-chancellor's senior research fellow at RMIT University's Centre for Urban Research in Melbourne, Australia. As co-leader of the Beyond Behaviour Change Research Program, she researches human–environment interactions, health and sustainability in cities.

Janine Morley is a senior research associate in the DEMAND Centre at Lancaster University, UK. She is a sociologist who studies the relationships between everyday practices, infrastructures and resource-use, with particular interests in digital technologies, energy demand and social change.

Davide Nicolini is professor of organisation studies at Warwick Business School, UK. He co-directs the IKON Research Centre at Warwick Business School and the Warwick Institute of Health. His research interests include practice-based approaches to the study of knowing, learning and change in organisations and innovation process in healthcare and other complex environments.

Andreas Reckwitz is a professor of cultural sociology at European University Viadrina in Frankfurt (Oder), Germany. His research focuses on social theory and sociology of (late-)modern culture. His most recent publication is The Invention of Creativity (2016).

Theodore Schatzki is a dean and professor of philosophy and geography in the College of Arts and Sciences at the University of Kentucky, USA. Among his books are Social Practices (1996), The Site of the Social (2001) and The Timespace of Human Activity (2010).

Robert Schmidt is professor of process-oriented sociology at Catholic University Eichstätt-Ingolstadt, Germany. His research focuses on praxeology, praxeography and process-oriented methodology. His publications include Sociology of Practices (2012) and Siting Praxeology: The Methodological Significance of 'Public' in Theories of Social Practices (with Jörg Volbers, 2011).

Elizabeth Shove is professor of sociology at Lancaster University, UK and is co-director of the DEMAND Research Centre (Dynamics of Energy, Mobility and Demand). She has written about social practices, everyday life, and the changing demand for energy and mobility.

Nicola Spurling is lecturer in the Institute for Social Futures and sociology at Lancaster University, UK. Her research explores everyday futures and their implications, focussing on change in mobility, working and daily lives since 1950, theories of practice and futures methods.

Alan Warde is professor of sociology in the School of Social Sciences and professorial fellow of the Sustainable Consumption Institute at the University of

As the book's subtitle suggests, its chapters offer new approaches to describing and analysing connections within the nexus of practices, to questioning the dynamics and implications of the complexes and constellations that practices form and to understanding the entities – the practitioners – that carry on or enact practices. Since individual chapters speak to more than one of these agendas, there is no one way of grouping them and thus no one way of reading the book. We have therefore taken the unusual step of presenting the chapters in alphabetical order, by author's surname. The resulting table of contents is complemented by a table which identifies themes that each chapter addresses (see p.vii).

Five significant themes are woven through the book, namely: suffusing, threading through, largeness, changing connections and practitioners. These themes did not arise by chance. Some authors, in tackling questions of largeness and in examining the roles of practitioners, respond to the prominent criticisms of practice theory mentioned above. Other authors concentrate on the problem of better understanding how practices hang together. The value of distinguishing between qualitatively different processes of interconnection (such as suffusing and threading through) emerged as the book took shape, as did an interest in characterising the dynamism, not of isolated practices, but of complexes and constellations of them. When read together, the chapters combine to provide new understandings of the nexus of practices and of its connections, constellations and practitioners.

The remainder of this introduction is designed to help readers navigate the collection as a whole. It describes the five cross-cutting themes, explains which chapters engage with each, and briefly comments on what the contributions have to offer.

Thomas Alkemeyer and Nikolaus Buschmann
(Translated by Robert Mitchell and Kristina Brümmer)

Practice theorists have a vested interest in studying how social order emerges via the interplay of things, artefacts and bodies. However, they rarely reflect upon how the people involved in order-making come to be or, indeed, put themselves in the position to be able to participate as competent 'players'.¹ One reason for this oversight, from our perspective, is that practice theoretical discussions still mostly operate within the framework of familiar alternatives, namely, whether social structures constitute social action and its actors, or whether the structures are constituted by activities within the bounds of preformed agency. Accordingly, two – oversimplified – perspectives can be differentiated by the relationship they propose between praxis² and participants. If praxis is seen as pre-structured, i.e. as practice, then participants are mere dependent variables keeping routinised action 'going'. If, on the other hand, praxis is conceived as a contingent accomplishment, then participants become autonomous actors with stores of practical knowledge that enable them to deal with the contingency of practice in a skilful and creative manner. These differences notwithstanding, the perspectives both pay very little attention to how play-ability³ arises. Furthermore, within the framework of their respective conceptions of praxis, the participants are conceived of in one way or another as pretty much 'play-able' out of the box, so to speak. In contrast, we understand play-ability not to be an always given property of individuals, but rather to be formed within the framework of 'distributed agency' (Rammert and Schulz-Schaeffer, 2002): things, artefacts and bodies mutually enable one another in their play-ability and, by doing so, bring forth a specific reflexivity of praxis.

In this chapter, we reconstruct how play-ability arises qua learning through participation in shared practices. We do this in connection with our thesis that practice theories need to be complemented by theories of subjectivation if they are to resolve the issues of both how practices are transmitted and reproduced and how they are accomplished in situated praxis. Against this background, we conceptualise

subjectivation as a process inherently embedded in praxis, in which the ability intelligently to orient one's own action towards practice-specific requirements is continually being formed and in which the process of doing can also entail the critique and transcendence of these requirements. Learning cannot be understood as the passive acquisition of operational skills. Instead, people (trans)form themselves via their engagement as recognisable subjects and cultivate their play-ability by learning to comply with the normative standards unfolding in praxis.

In this context, we utilise the processual and relational category of en-ablement.⁴ This term indicates that people only become carriers of specific abilities through participation in practices. It also suggests that the status of a participant is dependent upon mutual recognition. We assume that 'our participation in ... practices enables us to become the agents we are through our mutual accountability to the possibilities those practices make available and to what is thereby at stake for us in how we respond to those possibilities' (Rouse, 2007b: 53). Accordingly, we introduce the idea of en-ablement as subjectivation in order to incorporate activity alongside passivity, adaptation and defiance, routine and reflexivity, which are all involved in the formation of play-ability.

By conceiving of the unfolding of practices and the formation of play-ability as co-constitutive and co-extensive, we attempt to do justice to the ultimate, not yet fulfilled aim of abandoning the traditional contrast between methodological holism and methodological individualism. In pursuit of this aim, we try to grasp the emergence of play-ability within a practice theoretical framework without succumbing to old habits of either attributing deterministic power to practices over the activities of participants or presupposing fully formed actors who are ready for action.

In taking this approach, we enter a research field that is usually dominated by such terms as learning or education. Hence, the first section discusses how these terms are conceptualised within the framework of practice theories. Subsequently, the second section argues that a practice theoretical perspective must seek to reconstruct the processes in which play-ability emerges in praxis. The following section portrays the development of competent participation as a process in which specific knowledge, identity and social membership are formed, thus embedding learning within power relations and conflict. The section thereafter shows that the process of becoming a 'player' cannot be reduced to a purely mental interaction with the world, but rather emerges as part of one's bodily being and acting in the world. The chapter concludes by shedding light on the relationships between normativity, enablement and subjectivation.

In many ways, 'hospital life' has changed dramatically in the last 150 years. In other ways it has remained remarkably the same. These changes (and lack of changes) can be seen in the design of the New Alder Hey Children's Hospital in Liverpool. While architects and hospital planners have done away with the traditional, department – connecting central corridor, the New Alder Hey maintains many of the now essential features of hospital design that were advocated by Florence Nightingale in the 1850s, including the spacing between patient beds and the use of verandas for patient access to daylight and fresh air (Prior, 1988). Moreover, it has been designed in such a way as to facilitate a whole host of practices that would have

been completely out of place in the old pavilion style hospital, but that are now essential to hospital design, including socialising and shopping. 'Resembling a more salubrious version of an airport concourse, the internal street [of the New Alder Hey] is the entry point and social condenser, colonised with shops, café, specially designed furniture and a giant, conical structure containing a multi-faith space' (Slessor, 2015: 6). Hospital life is made up of different combinations of activities at different points in history.

But how do these combinations of activities change? And why do they stay the same? These are our starting questions in exploring how hospital life has changed over time. In order to tackle these questions, we begin with complexes of practices and not 'a practice'. This starting point reflects other contributions in this volume and echoes recent developments in theories of practice that are moving away from a focus on the constitution and trajectories of specific practices to an emphasis on the interdependencies, connections and configurations that are central to the constitution, reproduction and transformation of social life.

Conceptual headway regarding how practices hang together has been made on several fronts. Steps that have been taken include recognising that practices become organised in time in different ways (Southerton, 2006); practices gather around particular places (Shove et al., 2012: 84); and that practices compete and collaborate for time (Schatzki, 2010b). Multi-practice configurations have been described as bundles, complexes, constellations and systems in order to capture issues of scale, fixity, flexibility and structuration in connection. While these ways of conceptualising complexes of practices are useful for understanding how one practice is connected to another, they are of less value in helping us understand relationships between the connections that hold practices together. To date, much less has been said about how different types of connection impact on and matter for each other. For example, how are we to understand the changing material-spatial organisation of practices in hospitals and how those changes are affected by and shape the temporal sequencing of activities that take place within and beyond the hospital?

In what follows, we propose that in order to understand how complexes of practices change, we need to consider the multiple ways in which practices hang together and to show how these different types of connection matter for one another and ultimately for the reproduction of the complex itself. Our ambition is to build the foundations of a theory of practice (and of change) that is concerned from the outset with relationships between connections (interconnections). In our schema, new elements of practice are not the source of change. Rather trajectories of change are an expression of the ways in which practice complexes interconnect. In our description of these interconnections we put forward an understanding of complexes of practices as held together by a connective tissue that is itself an essential feature of practices.

We develop and illustrate this idea through a discussion of the transformations of complexes of practices that make up hospital life. A hospital is an intriguing site in which multiple activities necessarily combine and coordinate in routine and observable ways. What seem to be outwardly static structures have changed radically over the past century, meaning that any one hospital is constituted by historically

specific complexes of practices. So how has that happened? The multiple activities that go on in a hospital and their orchestration to make an institution that functions 24 hours a day, 365 days a year, means that hospitals are sites in which temporal aspects of connection are especially visible. Hospitals have multiple socio-temporal peaks, sites and cycles and they have been studied as such (Zerubavel, 1979). But they also vary in terms of the kinds of infrastructure, built environment and technology which constitute them and the past design and implementation of these material arrangements is obviously important for contemporary hospital life. As such, hospitals provide a revealing setting in which to pursue our goal of understanding change as an outcome of the constitution of practice complexes.

Drawing from more recent developments in theories of practice (e.g. Shove et al., 2012; Schatzki, 2010a; Schatzki, 2010b; Shove, Watson and Spurling, 2015), as well as from established ideas in social theory (e.g. Zerubavel, 1979; Abbott, 1988; Prior, 1988), we develop the beginnings of a conceptual scheme that can better account for changes in the complexes of practices that make up hospital life from the 1850s to 2015, as described in the two architectural accounts of hospital design introduced at the start of the chapter. Analysing the organisation of activity in architectural designs has been proposed by others (Prior, 1988; Shove et al., 2012). For example, Shove, Pantzar and Watson write: 'Since buildings represent sites in which practices are contained, separated and combined, the history of domestic architecture provides a telling record of how daily life is organised and how this changes' (2012: 84).

We recognise that activity cannot be explained in full from building design alone, and indeed that is one of our central arguments. We claim that the architecture and layout of hospital buildings, such as the newly built Alder Hey Children's Hospital in Liverpool, is not simply the outcome of design trends, but represents one facet of multiple registers of transformation in institutional life. Others include: vast advances in medical science and in theories of infection and disease; a reconfigured and extended system of health professions, with their altered, emergent and redundant jurisdictions and areas of expertise; cultural shifts in social categories like children and childhood, social class, gender and age, all of which have implications for ideas of good hospital care; along with changed schedules and rotas of staff and patients associated with new forms of training, departmental opening hours, and different kinds of treatment, therapy and surveillance.

Given these multiple registers of transformation, it seems obvious that changes in hospital life cannot be revealed only through a story of developments in architectural design and professional planning, nor through a narrative of changing medical knowledge and professional organisation, nor through an understanding of the temporal organisation of the hospital. Instead, what is required is an account that reveals the dynamic interplay between these registers and their historical development. Such an account would get at changes in configurations of complexes of practices. What is needed is a method of understanding how different types of connection between practices matter for each other and how these connections matter for the reproduction of complexes of practices over time. Our

original contribution is to focus on the different qualities of what we call the connective tissue (Shove et al., 2012)¹ of complexes of practices, and the interconnections within the connective tissue, in order to understand how the combinations of activities that make up routine and everyday hospital life have changed and how they have stayed the same.

We begin by briefly mapping some of the ways in which theorists of practice have dealt with connections between practices in order to situate our notion of connective tissue. In the subsequent sections, we develop three qualities of connective tissue in more detail, namely jurisdictional qualities, temporal qualities and material-spatial qualities. In the final section we focus on interconnections among these three qualities to show how such a focus helps to understand better how complexes of practices change and stay the same.

Having argued that complexes of practices are held together by forms of connection that are multiply interlinked (by connective tissue) and that have various qualities, we now consider how complexes of practices change. Our starting question was this: how can we account for the changes (and lack of changes) over the last 150 years in the complexes of practices that make up hospital life? To answer this, we need to account for changing complexes over time and for changes in how these complexes are constituted. One response is to suggest that the connective tissue of a given complex of practices, its qualities of connection and relationships between the different types of connection, has a history. What we mean is that past interconnections shape qualities of contemporary complexes of practices that matter for the kinds of connections they make and enable in the present and the future. Understanding, in these terms, the connective tissue that holds complexes of practices together allows us to consider the significance of past and present interconnections. These lines of enquiry offer an alternative, sociological way of thinking about how past activity matters for present and future social action.

Prior (1988) provides a more classical historical account of the development of the design of children's wards. She demonstrates that the design of children's wards is both shaped by and shapes changing medical knowledge. She describes that in 1852 (when the first children's ward in an English hospital was opened) there were no special design requirements in that both adult and child medicine were underpinned by a miasmic theory of disease (the idea that prominent diseases of the time were spread by 'bad' or stagnant air). The pavilion plan of the hospital, advocated by Florence Nightingale (1859), was designed to facilitate the dissipation of miasma and allow the flow of fresh air. 'Space in the pavilion hospital is necessarily, then, full of light and air. The use of verandas, to which patients can be expelled during the hours of daylight, facilitates the circulation of air...' (Prior, 1988: 95). The architectural features of openness that allow the flow of air and light are repeated in contemporary hospital design despite medical knowledge having moved on from this theory of disease. The design director for the New Alder Hey Children's Hospital writes: 'The sense of openness extends to the clinical areas... to optimise observation and daylight. In the Critical Care Unit this approach has produced an innovative layout with patient bays curved around a central staff base and a rooflight that floods the eight-bed cluster with daylight' (Zucchi, 2015: 13).

While medical knowledge, and with it the organisation of medical, nursing and administrative practices, has clearly progressed significantly in the last 150 years, many design features remain. According to Prior '... it is essential to underline the fact that these features of physical environment were woven into entirely different discourses on disease and medical practice' (Prior, 1988: 101). To understand why some features of hospital life remain the same, while others have changed, we need to turn not just to aspects of knowledge and design, but to a more subtle account of the organisation and hanging together of hospital life.

The shift in medical knowledge from the miasmic theory of disease to 'germ theory' is important, but knowledge is only one aspect of the connective tissue that holds complexes of practices together. Instead of viewing architectural features (the isolation cubicle or glass doors to restrict the flow of air and germs) as a material expression of changes in medical knowledge, we would see these developments as part and parcel of changing jurisdictions that connect nursing with practices of sanitation and isolation and that disconnects it from practices involving socialisation and interaction. We would see that these changing jurisdictions matter for the temporal organisation of activity on the ward (so that nurses spend less time interacting with patients and more time cleaning up, etc.) and beyond. These changing jurisdictional and temporal qualities of how practices on the ward hang together matter for and are shaped by the material-spatial qualities imposed by isolated beds and separate wards.

Similarly, jurisdictional qualities change as practices of education, socialisation and recreation fall under the medical and nursing remit and, as they do so, temporal qualities change as play and interaction, and not just observation, become fixed practices in nursing schedules. Visiting hours become more flexible as children are no longer isolated from their parents. Instead, parents are encouraged to be with their children to lessen the 'emotional shock of a hospital visit'. The inclusion of 'mother's divans' and 'playrooms' in wards connects medical and nursing practices to practices of parenting and socialising. And as all these examples show, material-spatial qualities do more than just reflect changes in other domains.

The contemporary design of the large glass-sliding doors to patient's rooms in the newly built Alder Hey Children's Hospital embody and reproduce this multiple layering of historical interconnections. The doors can be closed completely and blinds pulled down to isolate and facilitate patient privacy. They can be closed and transparent to enable observation and they can be fully drawn back to open the patient's room into the ward to assist socialisation and interaction with staff and other patients.

Our point is that past interconnections shape features of contemporary practices that matter for the connective tissue, its qualities and interconnections in the present. The design of the hospital does not simply reflect changes in the complex of practices that make up hospital life, rather, the spatial organisation of the hospital is in a recursive relationship with the organisation of activity, but so too is the temporal organisation of the hospital and the organisation of jurisdictions within it. In connecting through the built environment, through patterns of time and through systems of professional responsibility, those multiple interconnections become features or characteristics of the complex of practices, the connective tissue that holds practices together.

Finally, while this historical layering of interconnections, and the development of connective tissue, appears structuring and directive, we stress that human activity remains indeterminate and that past connections and interconnections do not directly determine present activity or what will happen in the future. That current activities are indeterminate does not mean that they have no relation to the past. We

follow Schatzki when he writes: '... past phenomena circumscribe, induce-orient, and underwrite the public manifestation of – but do not cause or antecedently pin down – present activity' (Schatzki, 2010b: x).

In any specific instance, the relation of the past to the present and the influence of past interconnections (of the connective tissue) in current complexes of practices is an empirical question. Our aim has been to lay some foundations from which to begin investigations of these relations and thus of how practices hang together.

We are calling for a version of practice theory which begins with complexes of practices and not 'a practice' and that focuses on relationships between connections and on how different types of connection matter for each other. We have developed an approach which is concerned from the outset with how complexes of practices hang together. Central to our framework is the idea of a connective tissue that both holds complexes of practices together and that is itself an essential feature of the practices involved. With this idea in place, it is possible to focus on the different qualities of connective tissue, and the interconnections between those qualities, to understand changing complexes of practices over time.

Taking the interconnections of connective tissue as a starting point is significantly different to focussing on the constitution and trajectories of specific practices. It also contrasts with work that has focussed on singular dimensions of connectivity. Instead, we have sought to account for multiple registers of change at the same time. We consequently argue against the conceptualisation of connective qualities as either background or as part of a practice, instead we contend that they are both. Finally, we suggest that no particular interconnection should have ontological privilege, but rather, that understanding the character of interconnection is a question for empirical research.

These ideas point to new ways of thinking about the relationship of the past to the present. In viewing the present as an outcome of intersecting registers of reproduction and change, we challenge the view that the past somehow causes the present and the view that future trajectories can be anticipated by extrapolating from the past. Providing an exhaustive catalogue of processes by which past interconnections influence present shapes or forms of connective tissue is beyond the scope of this chapter. However, we hope to have demonstrated the potential of such an approach and given some clues as to how it might be developed.

The literature that has produced and consolidated the turn to practice and put in motion the 'bandwagon of practice-based studies' (Corradi, Gherardi and Verzelloni, 2010) is now in need of better understanding of the differences among approaches to practice. Since the beginning of what has been called a re-turn to practice (Miettinen, Samra-Fredericks and Yanow, 2009), it has been stressed that a unified theory of practice does not exist and 'disagreements reign about the nature of embodiment, the pertinence of thematizing it when analyzing practices, the sorts of entities that mediate activity, and whether these entities are relevant to practices as more than mere intermediaries among humans' (Schatzki, 2001a: 11).

At the beginning of the practice turn, much work was devoted to identifying the elements of a practice and some models were named 'element-based practice theory' (Morley, this volume). The elements of a practice were defined by Reckwitz (2002b) as bodily and mental activities, objects or materials and shared competences, knowledge and skills; Shove et al. (2012) identified them in competencies, meanings and materials; Warde (2005) pointed to understandings, procedures and engagements, moreover, for Warde (2005: 134), 'practical activity and its representation' are components of practice. According to Schatzki (2006), the elements that comprise a practice are linked to each other through five main mechanisms: practical understanding, rules, teleoaffective structure, general understanding and social memories. It is evident that the identification of the elements within a practice varies from author to author; nevertheless – according to Guzman (2013) – a momentary agreement regarding the identification of elements is apparently in place. Nevertheless, the issue is not whether or not materiality matters within practice theory; rather, it is whether materiality merely mediates human activities – as in human-centred

theories – or is constitutive of practice, as in posthuman practice theories. The term ‘sociomateriality’ enters this debate without resolving the tension between a substantialist ontology that assumes that the social and the material, human beings and things, exist as separate entities that interact and impact on each other and a relational ontology that assumes the constitutive entanglement of the social and the material.

At present, the debate on the sociomateriality of practices is still in a precarious equilibrium between the social and the material, on the edge of interpreting practice as an object of change (privileging humans) or an agent of change (privileging its agency vis-à-vis humans). I would point out that one of the reasons for a re-turn to practice theory was to move beyond problematic dualisms (Schatzki, 2001a: 10) and for this reason, I want to explore how the dualism between the social and the material may be overcome with a concept of sociomateriality in which it is the ‘glue’ that connects all the elements of a practice. Jones (2014), who conducted a literature review of the 140 journal articles published since 2007 in which sociomateriality appeared, distinguished a strong and a weak sociomateriality. Strong sociomateriality addresses all the concepts that Orlikowski (2010) suggests are entailed in sociomateriality – materiality, inseparability, relationality, performativity and practices – while the weak version employs only some of these concepts selectively. I shall take a strong sociomaterial approach to practice.

I shall use the term ‘posthumanist’ better to qualify an approach to practice (Gherardi, 2015) in which relational materialism is the assumed epistemology that differentiates it from human-centred practice theories (for an articulation of the difference, see Monteiro and Nicolini, 2015). A posthumanist practice theory has its roots in the sociology of translation (Latour, 1992; 2005), in the principle of symmetry between humans and non-humans and in a relational epistemology (Law, 1994; Law and Hassard, 1999). In order to stress how within a relational epistemology both the idea of performative accomplishment and becoming are central, I propose to leave aside for the moment current definitions of practices as arrays of activities (human, non-human or intertwined) and to focus on the image of a practice ‘as a mode, relatively stable in time and socially recognised, of ordering heterogeneous items into a coherent set’ (Gherardi, 2006: 36). Practice is thus seen as a mode of ordering, rather than an ordered product, an epistemology rather than an empirical phenomenon. In fact, considering practice as epistemology ‘enables scholars to theorize the dynamic constitution of dualities and, thus, to avoid the twin fallacies of “objectivist reification” on the one hand and “subjectivist reduction” on the other’ (Feldman and Orlikowski, 2011: 1242).

In this chapter I shall use three examples to ground the concept of sociomateriality in bodily practices such as fathering, in organisational and working practices such as caring for the elderly, and in the aftermath of Hurricane Katrina that made visible the interdependencies of the practices that had constituted New Orleans over time. The rationale for choosing examples ranging from male bodies in fathering

practices, through caring discourses situated in changing technological practices, to the co-dependencies of the practices producing a territory, is to show how sociomateriality overcomes also another classical dualism, that between 'micro' and 'macro' analysis. In fact, as Welch and Warde (2015: 84) note, this dualism creates an analytic problem whose clarification and resolution would enhance the power of practice theories, since it concerns the relationship between the minutiae of everyday performances of practices and the macro-institutional context.

The chapter will open by positioning the concept of sociomateriality within ongoing conversations in organisation studies and their connections with new feminist materialisms. I shall then offer three examples to illustrate the 'viscous porosity' in between entangled practice elements.

Sociomateriality, with or without a hyphen, is a key topic of research in organisational and working practices, following the tradition of sociotechnical literature in which the social and the technological are inextricably linked. Similarly, the concept of sociomateriality has a long history in the study of working practices connected to information systems (Cecez-Kecmanovic, Galliers, Henfridsson and Vidgen, 2014). Within these fields of research, 'the material' in sociomateriality is mainly technological but not solely so, since the technological effects in social life and in epistemic practices have been considered.

The term 'sociomateriality', without a hyphen and in reference to the feminist onto-epistemology of Barad (2003), has been introduced into organisation studies by Wanda Orlikowski (2006; 2007; 2010) together with Susan Scott (Orlikowski and Scott, 2008; Scott and Orlikowski, 2014). The term 'entanglement' or 'generative entanglement' is present in their work together with a relational ontology and an acknowledgement of relational materialism discussed by Law (1994) and performativity from actor network theory. These terms refer to the fact that within a practice, meaning and matter, the social and the technological, are inseparable and they do not have inherently determinate boundaries and properties; rather, they are constituted as relational effects performed in a texture¹ of situated practices.

Donna Haraway (1991; 1997), who has also deleted the hyphen (in sociomateriality), proposes to talk of natureculture as a unity, a universe rich in material-semiotic actors and metaphors, like the cyborg, the OncoMouse, the multispecies crowd and otherworldly conversation. Her questions revolve around what 'nature' means in the complex practices of our contemporary society. The image of the cyborg is based on a science fiction imagination, but a woman taking contraceptive pills is already a cyborg, a hybrid of nature (body) and technology. The OncoMouse, manufactured in scientific laboratories and not born, is another example of artificially produced nature. An oncogene, the gene that produces breast cancer, has been transplanted into it, so that the OncoMouse is the techno-body par excellence and at the same time the mammal rescuing other mammals.

Haraway's work dislocates the centrality of the human in favour of the in/non/posthuman and of bio-centred egalitarianism. In fact, what we call a 'body' is a multispecies crowd if we consider that human genomes can be found in only about 10 per cent of cells, while the other 90 per cent of the cells are filled with the genomes of bacteria, fungi and other tiny messmates. And she concludes by saying that 'to be one is always to become with many' (Haraway, 2008b: 3). The image of the companion species (Haraway, 2008b) links to the idea of otherworldly conversation (Haraway, 2008a), in which various non-human entities participate as subjects. Her work 'calls for a renewed kinship system, radicalized by concretely affectionate ties to nonhuman "others"' (Braidotti, 2006: 199). And in fact, Braidotti (2013) proposes using the term 'more-than-human' instead of 'non-human' to overcome the human/non-human dichotomy.

Karen Barad again removes the hyphen between the social and the material and with her work, the term 'sociomateriality' has entered the debate on technology as a social practice (Orlikowski, 2010; Suchman, 2007a; 2007b; 2011). Barad's (2007; 2013) work defines her epistemological position as agential realism. In order to understand what she means with this term, we must bear in mind that her book begins with a conversation between Heisenberg and Bohr on the so-called particle-wave duality paradox. For the former, quantum theory represents an epistemological concern and the particle-wave duality demonstrates that we can only make probabilistic predictions (uncertainty principle); for the latter it is an ontological issue, since particles do not have determinate values of position and momentum simultaneously (indeterminacy principle). Barad uses the term 'intra-action' to denote the Bohrian ontological inseparability of all words (culture) and all things (nature). Rather, culture and nature are entangled, since Barad seeks to avoid the issue of representationalism, so that 'realism' is not 'about representations of an independent reality but about the real consequences, interventions, creative possibilities and responsibilities of intra-acting within and as part of our world' (2007: 37). Reality is defined as things-in-phenomena and not as things-in-themselves. In fact, 'phenomena' are considered as the primary ontological units, recalling Bohr's definition of them as observations under specific circumstances, including an account of the whole experimental arrangement. Both Haraway's and Barad's work are reflections on epistemic practices and both talk of a way of knowing in which the knower is not external to, or pre-existing the world. Rather, the knower and the 'things' do not pre-exist their interactions, but emerge through and as part of their entangled intra-relating. This assumption allows us to reformulate the notion of agency and to transcend the duality of social versus material agency, human versus more-than-human agency, material versus discursive. From these points of view, the idea of engagement with sociomateriality requires an onto-epistemology that does not separate nature from culture.

Building on these ideas, the challenge is to develop appropriate methodologies for the empirical study of sociomaterial practices. My first step is to consider methods of analysing the sociomateriality of the body.

The concept of sociomaterial practices implies not only that the social and the material are co-constituted, but also that nature and culture are entangled. It has a methodological corollary that entails studying how, within a practice, bodies, matter and discourses are expressions of the same sociomaterial world. The term 'embodiment' expresses how the nature/culture divide is blurred in the materiality of bodies encountering a material-semiotic environment. When we study working practices empirically, we focus on how practical knowledge is embodied and how practitioners rely on sensible knowledge in order to bring a practice forward (Gherardi, 2012a). The centrality of bodies in approaching practices is self-evident, yet it has been overlooked even when humans are considered the carriers of practices. One reason for this may be the Cartesian and idealist tradition in Western culture that undervalues the sociomateriality of human bodies and the other-than-human bodies with which human bodies meet.

But what is a body? Do we 'have' bodies or 'are' we bodies? From Merleau-Ponty's (1962) widely cited concept of 'body subjects', ideas of the indivisibility of mind and body and of human beings as embodied social agents spread into the sociology of the body (Crossley, 1995), feminist theorising on embodiment (Howson and Inglis, 2001; Fotaki, Metcalfe and Harding, 2014), and corporeal ethics within organisation studies (Pullen and Rhodes, 2015b).

Human perception is intrinsically embodied for Merleau-Ponty: 'we are in the world through our body, and... we perceive that world within our body' (1962: 206). He argues that to be a body is to be tied to a certain world; the perception of the world is always embodied and the perceiving mind is an incarnate mind. Moreover, embodiment is bidirectional in that the body is sentient and sensible: it sees and is seen, hears and is heard, touches and is touched. Thus, embodiment is neither idea nor matter, neither subject nor object, but both at the same time.

Merleau-Ponty's concept of embodiment enables us to see how sensible knowing is enacted in situated practices and how aesthetic judgements sustain working practices (Strati, 1999; Gherardi and Strati, 2012). Moreover, Dale (2005) sees in this conceptualisation of the embodied subject a tool for understanding the negotiation of the material and the social (that she keeps in interaction as mutually enacted), the organisational and the subjective, the everyday relationship of the individual with organisational control systems.

Embodiment is a concept present in practice theories. Reich and Hager (2014) consider it one of the six threads of the literature on practice (the others being: knowing-in-practice, sociomateriality of practices, relationality, historical and social shaping of practices, emergent nature of practices). In fact, it is now widely accepted within the social sciences that selfhood is not only social, but also embodied. Nevertheless, the collapse of the mind/body and psyche/soma dichotomies originated from many different sources. Here, I shall mention only a few that may be relevant to a discussion of sociomaterial bodies.

A telling example of the intra-connection of space as both symbolic and material, is the liminal space of in-betweenness aptly described by Ettinger (2006) as 'matrixial', in the sense of uterine, with reference to the pregnant body and to the permeable membrane between the body of the mother and that of the foetus. The concept of matrixial space as a space of non-separation and non-distinction has been taken up in organisation studies by Kenny and Fotaki (2015), because the concept of the co-emergence of partial subject (the mother and the baby) with the matrixial borderspace provides the image of an emerging subjectivity in a sociomaterial encounter. The authors argue that Ettinger's work provides a fruitful new direction for the study of corporeal ethics within organisation theory. The idea of corporeal ethics draws on Diprose's (2002) notion of an ethics grounded in embodied experience.

When we consider subjects as embodied subjects and the body as 'neither brute nor passive' (Grosz, 2011: 18) but as 'agential intra-activity in its becoming' (Barad, 2003: 818), we can consider practices as embodied and emplaced (Ingold, 2000), since the issue is not the body per se, but the shifting sociomaterial intra-actions of bodies and matter across time and space.

By emphasising the notion of 'emplacement', Elisabeth Hsu (2007) adds another aspect to the body-mind paradigm, since emplacement implies a body-mind-environment interrelationship. Drawing examples from Chinese medicine, she illustrates how body concepts derive from ecological experiences and she illustrates the relationality of the body and its composition in diverse networks of desire, practices and habits. Our understanding of a 'body' is changed into a 'body-assemblage', a series of affective and relational becomings. At the same time, Hsu also suggests that space ought to be understood as a dimension of social relationships, a site where capacities can be enacted.

A good exemplification of what I consider the entanglement of a body-mind-environment is provided by Doucet's (2013) research on the practice of fathering. She considers parent-child relations as processes of 'intra-active becoming' (Barad, 2007: 146) or 'generative becoming' (Bennett, 2010: 3). In her research on bread-winning mothers and care-giving fathers, she stresses the recurring invisibility of the body in studies of parental care-giving and especially the invisibility of male embodiment. The embodied relations of parents and children are issues of 'matter':

[P]eople – mothers, fathers, babies, children, others – are in a perpetual state of becoming, and this 'becoming' posits the fundamental units of analysis, not as things or words, or subjects or objects, but as dynamic phenomena that are constituted by and through entangled and shifting forms of agency.

(Doucet, 2013: 295)

When we take 'becoming' as the unit of analysis – as Doucet does – we may see not only how embodiment is the entanglement of a body-mind-environment, but also

how its meaning changes in time and within 'a choreography of becomings'. Doucet gives a full description of the agencements² of entangled elements in fathering:

The meanings of fathers and infant care have also been gradually changing, partly through the constitutive intertwining of sociocultural practices, ideologies, and embodied subjects (via workplace policies and cultures, state policies, popular culture, media, and social media, including the burgeoning of daddy blogs), but also through women's rising employment, men's slow take-up of parental leave, and the social acceptance of that leave.

(Doucet, 2013: 295)

We can see how fathering as a sociomaterial practice is embedded in a wider texture of practices connecting the situated and everyday encounters of embodiments to the macro-institutional context. Therefore, gender becomes not a fixed biological identity, but a social practice (Poggio, 2006), a performance playfully and cannily enacted, that changes from context to context and from day to day, as men and women adjust their ideas of themselves to fit changing sociomaterial circumstances.

The title of Doucet's article – A choreography of becoming – is inspired by Coole and Frost's (2010: 10) expression, where the issue is 'that matter becomes', rather than that 'matter is'. The same expression may be extended to Nick Hopwood's (2014; 2015) work on parental embodiment in care as the enactment of a texture of practices through the four dimensions of times, spaces, bodies and things. Similarly, Maller (2015) in health settings research assumes a posthumanist approach, illustrating how health and wellbeing are outcomes of participation in a set of social practices mutually constructed by the materiality of everyday life.

The physical 'matter' of the body, its material-discursive production, its sensible knowing, its choreography of becoming, are all instances of embodiment as irreducibly material, social and emplaced in practices. By implication, a posthuman practice theory can inform a research programme that takes serious account of the concepts of both entanglement to undermine oppositional dualism, and intra-action to explore how practice components intra-act.

I offer another example of sociomateriality, drawn from my own research, to discuss working practices in relation to discourses on ethics.

In this chapter, I have stressed how the concept of sociomateriality is at the very heart of a posthumanist practice theory. The diffusion of Karen Barad's work and vocabulary outside feminist theory can be considered the beginning of a conversation among separate fields of research. In writing about the shift from 'inter' to

intra-thinking, John Shotter (2013: 41) notes that 'small changes in words can provide big changes in our orientations'. In fact, sociomateriality has already become the symbol for the disappearance of the hyphen that connects but separates oppositional dualisms.

The chapter has offered a selective immersion in the concept of sociomateriality and argued that it enriches the study of practices with special consideration of the materiality of bodies, technologies, discourses, meanings and material-institutional contexts of interconnected practices. This argument has been illustrated with references to empirical research projects, since the problem of how to craft a methodology for the empirical study of the entanglement of sociomateriality within the practices under scrutiny is still open and crucial.

While a humanist approach to practice assumes the centrality of humans as sites of embodied understandings and then proceeds to analysis of humans and their practices, a posthumanist approach instead interrogates how all the elements within a practice hold together and acquire agency in being entangled. In the latter case, 'sociomateriality' may be considered an attribute of any practice, and 'sociomaterial' an adjective that stresses how a practice is constituted by matter and culture. Moreover, on considering sociomaterial practices, the situatedness of a practice being practised in a contingent space and time is linked to other sociomaterial practices that sustain and allow the situated performance of that practice.

In this chapter, on considering the male body in the social practices of fathering, we could consider at the same time how fathering practices are linked to institutional contexts, to women's work, to social regulations on parental leaves and many other practices in which the social and the material are entangled. Similarly, in looking at a working practice like artificial nutrition in a nursing home, we could trace how the sociomateriality of bodies no longer able to take food, their nursing, medical practices and family care become entangled with ethical discourses on life and death as sociomaterial issues. In the last example of the sociomateriality of a territory – New Orleans – disrupted by a so-called natural event, we could see the social practices deriving from the past and shaping the territory in the way that a hurricane impacted on it with such violence. In all the three cases, the materialities of bodies, technologies, discourses could not be separated from the society that formed them and vice versa, the social was not external or separate from materialities. My argument is therefore constructed around the evidence that, in the three cases, both ethics and politics are sociomaterial and are part of the agencement of the heterogeneous elements in the becoming of a practice.

A final consideration concerns how the practices of representation of the topic of study (practice in our context) are linked to the concept of epistemic responsibility (Code, 1987). Not only are we – as researchers – internal to the object of study, but also our epistemic practices cannot be separated from ethical analysis. Hence, as researchers, we should be responsive to how the distinctions that we embrace construct our experiences and how these distinctions conceal as well as reveal what we research in the practices that we study.

I would like to express my great appreciation to all the other colleagues who took part in the discussion of the first drafts of the chapters of the book and especially to the three editors who generously took care of the whole process. Moreover, I wish to thank Marie Manidis for the long email conversations over the chapter and for her warm friendship.

As noted above, theories of practice have already addressed several forms of variation, even if these have not always been articulated as characteristics of variation per se. In order to summarise briefly what this work has revealed, this section identifies three different forms: variation as a basic feature, variation as meaningfully constructed and variation in the constituent elements of a performance.

As this chapter has made clear, dimensions of variation are themselves varied. Identifying variation depends upon a point of reference in terms of which judgements of similarity and difference are made. There are endless possibilities for specifying terms and instances of variation. My aim in this chapter was not to characterise all possible variations within a world of practices – such a task is impossible. Instead, I have articulated different analytic strategies that can be used to name and discuss select types of variation and I have considered some of the implications these distinctions raise for understanding a world of overlapping and interconnected practices.

Variation is a feature of all practices and of relations between them. Analysing variations therefore generates further questions about how practices interconnect and about change within a field or nexus of practices. Discussions of variation prompt further thought about how varied temporalities and spatialities are constitutive of not only practices but also nexuses of practices. Articulating dimensions of variation is also crucial for addressing questions about power and norms within a field of practices: how do variations between practices inform understandings of accessibility, power and interdependence? On what basis might normative assessments of acceptable or intolerable variations in practices be made? Extending analyses of variation along these lines could also inform more empirically focused questions such as: on what basis should practices be selected for empirical study? While this question must be answered in relation to

the concerns and questions at hand, being able to articulate dimensions of variation within a nexus of practices is important for justifying why some practices might be more suitable for investigating particular issues than others. While these questions have not been taken up in detail here, this chapter has helped develop a language and set of analytic tools for articulating variations within and between practices that will support their exploration in the future.

Many thanks are given to the participants and discussants at the Windermere (2014) and Lancaster (2015) workshops, who provided very helpful comments and suggestions on previous versions of this chapter.

In the spirit of investigating new frontiers and territories, this chapter is situated between two different types of theory: social practices and epigenetics, a sub-discipline of genetics. Its purpose is to examine the effects of practices on practitioners' bodies¹ and in so doing, advance theories of social practice towards new frontiers in health. So-called 'lifestyle diseases' are a persistent and growing global issue and the leading cause of death and disability despite decades of intervention (Daar et al., 2007). In regard to the many varieties of practice theories and practice composition, this chapter draws on the work of Shove et al. (2012) as it incorporates central tenets of theoretical and applied work to date. Explorations of the possible benefits and tensions of joining ideas from social and natural theories, or the 'soft' and 'hard' sciences, are not new. In fact, they are consistent with the materialist tradition of which theories of social practice are a part (Nicolini, 2012; Schatzki, 2001a).

Aside from sub-disciplines such as evolutionary psychology and human socio-biology, other more metaphorical explorations venture between the social and natural sciences (e.g. Shove and Pantzar, 2005b). Although these explorations have potential, they also have their challenges, some of which are highlighted by Schatzki (2001b).

For example, recent work on theories of social practice has benefited from numerous concepts sourced from the natural sciences of population biology and ecology, including terms like 'recruitment', 'reproduction', 'convergence' (Shove et al., 2012), 'competition', 'cooperation' and 'predation' (Pantzar and Shove, 2010a). Similarly, ideas about muscle memory from sports medicine and exercise science are useful in understanding how practices are remembered in bodies through 'practice memory' (Maller and Strengers, 2014). Practice memory, like muscle memory (Shusterman, 2011; Staron et al., 1991), relies on the notion that through performance, imprints

of practice elements are codified and can remain linked in the mental and bodily patterns of the performer' (Maller and Strengers, 2014: 150). These theoretical moves, although not free from the problem of incommensurability (Kuhn, 1970), have the potential to break new ground by offering transformative ideas about social change and how it might be (re)directed in more sustainable, equitable or healthier directions.

There is a need for positive change in regard to health and wellbeing and the rise of chronic non-communicable diseases (NCDs), characterised as typically not passing from person to person and being largely preventable (Daar et al., 2007). NCDs are attributed to 'unhealthy lifestyles' associated with post-industrialisation (WHO, 2015a). However, evidence suggests that the distribution of NCDs is uneven, since aside from socioeconomic differences, 'social habits, routines and conventions provide a source of general resistance' (Warde et al., 2007: 381). NCDs include cardiovascular diseases, respiratory illnesses, some cancers and type 2 diabetes – responsible for killing 38 million people each year (WHO, 2015a), or 60 per cent of all deaths globally (Daar et al., 2007). Obesity and being overweight – 'a highly predominant nutritional problem and one of the world's greatest health issues' (Patel, Choksi and Chattopadhyay, 2015: 430) – are associated with NCDs, in particular cardiovascular disease and type 2 diabetes. According to the WHO (2015b), in 2014, an estimated 1.9 billion adults were considered overweight and 600 million were obese. Public health programmes have had only marginal success so far and despite decades of apparently productive research into methods of intervention, 'no country to date has reversed its obesity epidemic' (Roberto et al., 2015: 2400). NCDs and obesity are particularly interesting health challenges as they are defined by the complex interplay of people, their bodies and environments (Roberto et al., 2015).

The chapter has two aims. First, to work through possible points of intersection between theories of social practice and epigenetics to explore the effects of practices on performers' bodies. Second, to see if ideas from epigenetics can extend the reach of theories of social practice to encompass intergenerational and environmentally related health problems attributed to 'lifestyle'. As there are a number of possible diseases attributed to lifestyle, the chapter focuses on obesity. The challenges that arise when aligning two very different theories and when reconciling the objects of study, the sites where action occurs and conceptualisations of bodies are presented.

The chapter begins by providing an overview of the main ideas in epigenetics relevant to a discussion about social practices and practitioners' bodies. The next section explores possible points of intersection between theories of social practice and epigenetics, highlighting tensions and potential incompatibilities. Using obesity as an example, the chapter takes familiar ideas about performance and recruitment and explores them at a detailed level. This exercise raises a series of questions about the inner workings of practice, the intricacies of action and doing and effects on practitioners' bodies. The subsequent section uses ideas outlined previously to consider how a social practice-epigenetics perspective might further encourage the application of theories of social practice to the problems of understanding and intervening in lifestyle disease.

Epigenetics is a relatively new and somewhat ‘controversial’ (Bird, 2007) field of study in genetics that seeks to understand and explain the effects of gene–environment interactions on bodies, in other words, ‘the interface between genetics and the environment’ (Milagro, Mansego, De Miguel and Martínez, 2013: 784). Its focus is at the cellular and molecular level: how environmental exposures change the way genes are expressed, thereby changing the physical appearance – or phenotypes – of bodies without changing the underlying genetic makeup, or genotype (Skinner, 2014; Tammen, Friso and Choi, 2013); in other words, epigenetics studies changes to the physical form of individuals that occur in response to environmental instead of genetic factors (Handel and Ramagopalan, 2010).

Various environmental exposures change the way genes are expressed (switched on or off). They include weather, drought, chemicals and environmental toxins, high or low calorie diets, stress, exercise, drugs, including tobacco and alcohol, as well as pathogens (Skinner, 2014; Tammen et al., 2013). Exposures can result in changes to the phenotype, affecting conditions such as physical shape, disease susceptibility, stress response, behaviour and longevity (Tammen et al., 2013). These conditions are associated with a range of pathologies, from obesity, diabetes, cardiovascular disease and cancer (Milagro et al., 2013), to changes in social behaviour (Keverne and Curley, 2008; Skinner, 2014) and mental states (Karlic and Baurek, 2011), although the evidence for social and mental change is inconclusive (Januar, Saffery and Ryan, 2015). This means:

Epigenetic marks are therefore a reflection of an individual’s environmental exposures and as such change during the lifetime of a cell/tissue. Thus, we are ‘acquiring’ changes to our epigenome all the time.

(Handel and Ramagopalan, 2010: 2)

Epigenetics is proposed as an explanation for why monozygotic twins with identical genes often diverge in their physical appearance and disease susceptibility over time (Fraga et al., 2005).

One of the most exciting prospects of epigenetics is that it represents a form of ‘soft’ inheritance (Handel and Ramagopalan, 2010: 2) or ‘epigenetic memory’ (Karlic and Baurek, 2011: 279). Accordingly, past environmental exposures and their effects on bodies pass to future generations, despite the fact that the genes themselves are unchanged and that future generations do not necessarily experience the same exposure as parents and grandparents (Skinner, 2014; Tracey, Manikkam, Guerrero–Bosagna and Skinner, 2013). Epigenetic inheritance arises from a number of mechanisms. In multigenerational inheritance, more than one generation is exposed to the same stressor at the same time (e.g. via pregnant females). Transgenerational inheritance occurs when transmission of epigenetic changes passes through the eggs and sperm to future generations not directly exposed to the same environment (Skinner, 2014).

Theories of epigenetic inheritance have been developed in both the social and natural sciences, variously focusing on the prospect of inheriting artistic talent or a propensity for genius (Karlic and Baurek, 2011; Shenk, 2011) or for disease and obesity (Patel et al., 2015; Rhee, Phelan and McCaffery, 2012). Not surprisingly, epigenetics is considered to represent 'some of the most exciting contemporary biology' (Bird, 2007: 396).

How could these ideas link to theories of social practice? To begin, environmental exposures may occur during the performance of practices – where genes and environment interact. Following this, the logic of epigenetic changes and how human–environment interactions leave imprints on bodies parallels theories of practice where 'a social practice is the product of training bodies in a certain way' and 'can be understood as the regular, skilful "performance" of (human) bodies' (Reckwitz, 2002b: 251). This is a principal point of intersection between these theories.

Theories of social practice have long recognised the importance of doing and performance (e.g. Bourdieu, 1990; Giddens, 1984), more recently distinguishing moments of enactment from the 'pattern' or 'block' of the recognisable practice entity that persists over time (Reckwitz, 2002b; Schatzki, 1996; Shove et al., 2012). In the moment of doing or performing a practice, all the elements are integrated. In this state of action, two things occur: (1) bodies interact materially with the world and its artefacts; and (2) in carrying out action, bodies experience certain material conditions, receiving a range of environmental exposures. Both affects can result in material changes in and to bodies. Performance is where genes, bodies and material environments interact – and consequently where epigenetics and theories of social practice interface.

Theories of social practice are concerned with the 'enduring' practice entity (Shove et al., 2012: 8), whereas epigenetic theories are concerned with the gene, or more precisely, its expression (Patel et al., 2015; Tammen et al., 2013). These differences in the objects of study are where the theories potentially diverge. However, despite having different foci, both involve human bodies and change over time. This possible point of tension serves to highlight the point that action and its effects are taking place at multiple levels or bodily scales and have varying impacts over time and space. This multiplicity generates further questions of timing and sequence. In epigenetics, there are likely to be multiple instances of exposure that trigger a series of cellular events or molecular processes culminating in a change of gene expression and phenotype at a later date – or even in another generation (Milagro et al., 2013). Similarly, a practice does not arise from a single performance; instead, it requires multiple performances by many performers across time and space (Reckwitz, 2002b; Shove et al., 2012). Recognition of these temporal patterns leads to further speculation about how these theories describe different types of action that fit together,

or link, in a causal or dependent fashion such that epigenetic changes to bodies are reliant on, and the result of, the performance of practices. As performances are repeated over time, bodies are continually shaped by and through practice in different ways in order to generate 'skilful performances' (Reckwitz, 2002b: 251). These include mental patterns and corresponding bodily development of muscles, tissue and bones consistent with theories of social practice on one hand, and biochemical or metabolic processes leading to epigenetic changes on the other.

Given that material changes to performers' bodies are central to both theories, this observation raises questions about how bodies are conceptualised in each and whether correspondence, commensurability or compatibility should be expected (Kuhn, 1982).

In contrast to epigenetics, in theories of social practice, bodies have a 'present-absent' status. They are clearly present in that bodies are the carriers of practice, essential for the continuing performance and persistence of practice entities and in that practices are 'embodied' by their performers. Bodies are also 'absent' in the way that their physical and sensory qualities are largely unrecognised or dematerialised. It could be argued that in theories of social practice, human bodies are recognised as carriers: as contributing to the elements of meanings and skills or competence (Reckwitz, 2002b; Shove et al., 2012) and as 'sites of embodied understanding' (Reckwitz, 2002a: 212). However, it is not clear how they might be conceptualised beyond this. In particular, there is no specificity or delineation as to what parts of bodies might be implicated in certain practice performances. This is despite the trend to '[put] the body back into sociology' (Shilling, 2003: 17). Reckwitz briefly acknowledges the materialities of bodies in this way:

One can say that both the human bodies/minds and the artefacts provide 'requirements' or components necessary to a practice. Certain things act, so to speak, as 'resources' which enable and constrain the specificity of a practice. (2002a: 212)

However, in this instance, Reckwitz again refers to bodies in general. There is clearly a bodily presence in Bourdieu's concept of 'habitus' (1984) as 'a system of dispositions' (Bourdieu, 2005: 43) where habitus is considered both 'a medium and outcome of social practice' (Wainwright and Turner, 2006: 240). More recent work pays scant attention to the materiality of bodies, perhaps because there has been greater interest in the material roles of non-humans in the conduct of practice, or because there has been less practice theoretically inspired research in the area of health. For some of these reasons, the materiality of bodies has been underplayed; exceptions include Pink (2012; Pink and Leder Mackley, 2014) and Wallenborn (2013; Wallenborn and Wilhite, 2014).

Concerning the health outcomes of practices, the central role of bodies is paramount. Most practices have health implications, whether for good or ill (Blue, Shove, Carmona and Kelly, 2016; Maller, 2015). Furthermore, as NCDs and obesity

arise from people–environment interactions (Patel et al., 2015; Roberto et al., 2015), there is an opportunity for theories of social practice to contribute to understanding and intervention in these issues (Andrews, Chen and Myers, 2014; Blue et al., 2016; Maller, 2015).

Before proceeding, it will help to provide a brief review of some of the ways bodies could be further conceptualised in theories of social practice to foster links with epigenetics. As an obvious first step, bodies and their constituent parts could be considered material elements, depending on the practices at hand, and; second, bodies could be considered as an ‘assemblage’. In what follows, I discuss these conceptualisations and in particular, consider bodies as material elements.

That bodies rarely count in a physical sense as contributing to the material elements of practices is somewhat surprising as bodies – literally – can provide a number of possible material aspects of practice/s. These include energy in the form of physical strength, cellular and molecular processes and substances of various kinds necessary for basic and higher order functioning. Furthermore, they are a substrate or repository for practices (Bourdieu, 1984; Maller and Strengers, 2014; Shove and Pantzar, 2005b; Wallenborn and Wilhite, 2014). Bodies also offer a range of sensory capacities that enable or participate in practices (Pink, 2009; Pink and Leder Mackley, 2014).

Some practices require more active bodily involvement, movement and skill than others, for example, dancing or playing a sport compared to working at a desk or commuting to work. To do professional ballet, or indeed any sport or artistic performance, performers’ bodies must be ‘fast, strong, supple and have impressive stamina’ (Koutedakis and Sharp, 1999, as cited in Wainwright and Turner, 2006: 242). In these performances, bodies move in deliberate and precise ways, often formed through years of training in which movements are tightly coupled with the other practice elements (competences and meanings) (Shove et al., 2012; Wallenborn and Wilhite, 2014). Another way bodies materially contribute to physically demanding practices is aesthetically, meaning that they must be shaped, sculpted, presented and tightly maintained through rigid routines of training and body care to meet visual requirements and expectations (Wainwright and Turner, 2006). The bodily requirements of these practices contrast with others that are arguably less demanding in terms of corporeal precision, aesthetics and discipline. In other practices, bodies have a more biological role. For example, in relation to pregnancy and childbirth, bodies contribute a large range of materials to the developing foetus, including genetic information, proteins and blood. Whether bodily components of genes, fat or muscles will be relevant to a particular study depends on the line of inquiry, but these examples serve to highlight the fundamental material role bodies play in practices.

In underplaying what bodies offer materially, theories of social practice and their proponents have rarely considered the bodily impacts of practices in relation to health. Some practices shape bodies and have more profound health outcomes than others. As I discuss in more detail below, recent evidence shows that due to epigenetic changes and nutritional exposure in utero, obesity is now considered a heritable condition (Rhee et al., 2012; Skinner, 2014). In relation to the aims of this chapter, there are also different time scales to consider, as practices have different

bodily impacts in the moment of performance compared to the accumulated effects on bodies from performing over a lifetime. One example is smoking tobacco (or the practices that involve smoking tobacco). In the moment of performance and of inhaling cigarette smoke, nicotine immediately causes the release of dopamine, meanwhile over the lifetime of the performer, this can result in damaging bodily changes, including lung cancer (Hatsukami, Stead and Gupta, 2008). These examples show that bodies do not only provide vital materials to practices without which they could not be performed (Wallenborn and Willhite, 2014); they are also systematically affected and made by them, sometimes in profoundly detrimental ways.

Another way to conceptualise bodies relevant to theories of social practice and epigenetics is as 'human-non-human assemblages' (Bennett, 2010; Greenough, 2011) or 'extended bodies' (Wallenborn, 2013). Thinking of human bodies as assemblages is not a new idea (see Deleuze and Guattari, 1987: 4; Greenough, 2011; and Fox, 2011), but it is worth briefly reflecting on this notion in the current context. Assemblages can take many and multiple forms, as summarised by Marcus and Saka (2006):

[An assemblage] can refer to a subjective state of cognition and experience of society and culture in movement from a recent past toward a near future (the temporal span of emergence); or it can refer to objective relations, a material, structure-like formation, a describable product of emergent social conditions, a configuration of relationships among diverse sites and things... And of course, if not explicitly delineated, it can refer to all of these at once.

(2006: 102, emphasis added)

The above definition reveals the multiplicity inherent in the nature of assemblages, which assists in considering how to link theories of social practice with epigenetics. For example, the body as an assemblage might include muscles, genes, tissues and bones, together with the practice elements of meanings, materials and competences. Furthermore, there are vast numbers of microorganisms which bodies host (bacteria, viruses and parasites), as well as the things, objects, technologies or other materials required for the performance of a practice.

There is also recognition that objects and other organisms permeating bodies as part of human-non-human assemblages have agency; in Bennett's (2010: vii, xvi) terms they have 'vital materiality' or 'thing power'. In the case of any practices related to the ingestion of materials like food or drugs, these actors have a material presence in practices in interacting with bodies and create certain effects leading to varying health outcomes. As Bennett (2007: 137) argues in relation to fat:

It is more likely that an emergent causality is at work: particular fats, acting in different ways in different bodies and with different intensities, even within the same body at different times, may produce patterns of effects but not in ways that are fully predictable – for a small change in the assemblage may issue

in a significant disruption of the pattern. The agentic assemblage in which persons and fats are participants ought to be figured as a nonlinear system.

The same point applies to the vast number of microorganisms that enter or are already part of bodies, as Turnbull et al. (2007: 804) observe, 'the microorganisms that live inside and on humans (known as the microbiota) are estimated to outnumber human somatic and germ cells by a factor of ten'. In concluding this brief summary, one of the most useful aspects of thinking in terms of assemblages is that boundaries are immediately called into question; not only boundaries between organisms, but between bodily parts and their components. This type of conceptualisation allows for thinking about permeability and unsettles the illusion that human bodies – and practices – have fixed or discrete edges in the world, unconnected to life, matter and action around them (Bennett, 2010).

This section has reviewed two possible ways bodies could be thought of that are commensurate with theories of social practice and epigenetics – as materials or as assemblages. In epigenetics, bodies and their materiality play a central role. In fact, bodies are scrutinised and quantified from their phenotypic states right down to their molecular processes and genetic material. In theories of social practice, the role of bodies is less clear, leaving room for more detailed investigation and definition as this chapter initiates. Interestingly while body–environment interactions can be the central focus, in neither case are bodies themselves the key units or sites of study. When thinking about health, the strategy that makes sense to both epigenetics and theories of social practice is that of taking bodies to be the medium or pathway through which change happens. To explain this idea in more detail, it is useful to work through an example.

It is clear that global trends in deaths and illnesses said to arise from lifestyle are escalating and are unlikely to shift in the short term. In order to modify this trajectory, there is a need to reconceptualise how such health conditions arise, focusing on new understandings of wellbeing and disease from the latest thinking in both the natural and social sciences.

In the natural sciences, epigenetics is emerging at the forefront of innovation, in particular in regard to obesity. In the social sciences, theories of social practice are beginning to make a similar impact in analysing a range of health and wellbeing outcomes. In bringing these two fields together, this chapter explores new

ways to intervene. The timing is ideal as there is growing recognition in public health, and health promotion in particular, that existing approaches have not delivered desired outcomes. For example, Roberto et al. (2015: 2401) examine the 'false dichotomy that obesity is driven by either personal choice or the environment' and instead suggest 'these two competing perspectives be merged to show the reciprocal relationship'.

The first contribution from a combined practice-epigenetics perspective is the problematisation of lifestyle diseases. The main point is that while individuals perform practices, they are not the unit of study or of intervention. Instead, the proposition is that health problems arise from practice-gene-environment interactions realised through the performance of specific practices. In carrying out or performing practices, practitioners (individuals) are exposed to (and ostensibly create) a number of environmental conditions that make and remake bodies, resulting in varying states of health or disease. Therefore, interventions must target the practice, not the performer (Blue et al., 2016; Maller, 2015; Shove et al., 2012; Spurling, McMeekin, Shove, Southerton and Welch, 2013; Strengers and Maller, 2014).

There are four other aspects of how problems of obesity would be framed from this perspective, all of which are consistent with other work based on theories of social practice. These are: (1) practices are intricately linked to other practices to form bundles and complexes: no practice exists in a vacuum; (2) bodies and their parts count as material elements that are made and remade through performance and bodily processes and mechanisms, including epigenetic means; (3) there is more than just the agency of human bodies involved in practices, such that other various materials, objects, technologies and organisms also have agency; and (4) the marks of practices on bodies can be heritable such that the effects of practices performed in one generation can pass to the next, during pregnancy – all changing ideas about disease aetiology. These contributions involve thinking beyond individuals and their 'lifestyle choices' said to result in health or ill health (Blue et al., 2016; Maller, 2015). As Fox (2011: 360) writes, 'ill-health is too quickly accepted as an attribute of an individual body, rather than a wider, ecological phenomenon of body organisation and deployment within social and natural fields'.

Although bodies are not the entity of study or the only target of intervention, they are taken to be the medium or pathway through which change happens and health outcomes are realised. Blue et al. (2016) and Maller (2015) have usefully considered types of health interventions that would follow from and be consistent with a social practice framework. I expand on these types of intervention below, but with a practice-epigenetic twist. The approach resembles Wallenborn's (2013: 149) argument for a 'practice perspective combined with a body-centred analysis'. However, it is noted that:

Whilst a turn to practice makes great strides in overcoming critical, problematic dichotomies between individuals and social structures, rational actions and habits, it does not generate simple guidelines for intervention.

(Blue et al., 2016: 11)

This difficulty is in part explained by the emergent nature of practice phenomena (Shove et al., 2012). Their 'ceaseless movement and incessant rearrangement' (Schatzki, 2002: 189) presents a challenge to any attempts at control. Another point is that as practices interconnect, effective interventions can sometimes come from other quarters or be unintended outcomes of seemingly unrelated policies (Blue et al., 2016).

There are, nonetheless, methods of directing interventions at the practices of which everyday life is made. From this perspective, relevant targets include seemingly innocuous ways of going about everyday doings, so habituated they almost become invisible, or, at the very least, unremarkable. Epigenetics is demonstrating the profound impact of the everyday practices of parents and grandparents on the health of future generations 'pervad[ing] all aspects of development' (Keverne and Curley, 2008: 398).

By implication, forward-looking interventions seeking to influence the health of future populations should focus on the practices of current (and possibly previous) generation/s. At the same time, this link suggests that practice-based interventions also need to take a long-term view, as already recognised (e.g. Schatzki, 2015a; Spurling and McMeekin, 2014).

Aside from this, the key strategy to change health outcomes from a practice-epigenetic perspective is to (a) intervene in relevant practices and their elements of meanings, materials and competence and (b) orchestrate the making and breaking of links that sustain practices, as well as those between practice bundles and complexes (Blue et al., 2016; Shove et al., 2012; Spurling and McMeekin, 2014). For example, controlling fat as a material element in multiple practices causing multigenerational obesity could include: changes to practices of shopping and food consumption by developing competencies in preparing and consuming low-fat meals or developing new practices of meal preparation; changes to the meanings and materials of meals in relation to frequency, size and content; new or changed practices of policymaking to reduce the availability of fatty, sugary and highly processed foods and drinks and medical or surgical practices, such as bariatric (lap band or gastric) surgery. The latter has been shown to be effective in circumventing epigenetic changes that result in the birth of overweight children (Rhee et al., 2012). Of course, these are not new ideas. However, from a practice-epigenetic perspective, the difference is to recognise and capitalise on where and how these attempts intervene in practices, bundles and their elements and therefore better to coordinate and systematise what are usually treated as one-off, separate or unrelated interventions, directed at individuals and their behaviour at single points in time. In Spurling and McMeekin's (2014: 79) words: 'intervention should not be viewed as "external" or "one-off" but as continuous and reflexive, historical and cumulative'. From a practice-epigenetic viewpoint, intervening may also involve targeting new, or to date unrecognised material elements, such as genes, cells and molecules implicated in epigenetic modifications.

Other options would be to focus on the recruitment and defection of practitioners from practices contributing to ill health. A practice-epigenetic take on this would involve a multigenerational perspective, as suggested elsewhere (Maller and

Strengers, 2014). In public health, childhood is already recognised as a key part of the life course in creating future health outcomes, as established by the social determinants of health literature (see Wilkinson and Marmot, 2003). The importance of maternal health and early childhood development is already embedded in a range of public health interventions, particularly those aimed at reducing multigenerational health inequities (Klawetter, 2014). Similarly, from a practice-epigenetic point of view, the importance of childhood and interventions in practices associated with (or occurring during) pregnancy and child rearing should be recognised.

There is further scope to consider the recruitment and training of children as new practitioners and of the effects of such processes in creating lasting imprints and practice memories (Maller and Strengers, 2014) that matter for health. Work inspired by theories of social practice has not yet ventured into childhood or child rearing, both being areas in which the potential of practice theoretical research remains to be realised.

Finally, attempts to intervene to reduce lifestyle diseases from a practice-epigenetic outlook would be directed towards monitoring and responding to shifts in practices (Blue et al., 2016). As Schatzki (2015a: 17) observes, 'changes of all kinds constantly befall practices, arrangements and bundles, which undergo halting, irregular, not necessarily infrequent, and sometimes rapid development'. Examples might include interventions aimed at redirecting practice trajectories as well as other emergent changes arising from the inherently dynamic nature of social practices and practice-gene-environment interactions. In particular, it would make sense to undertake long-term monitoring of practices and epigenetic variations across generations as programmes and policies are introduced or modified. Finally, it is important to recognise that all policies and interventions regardless of intentions will affect the emergence, persistence and disappearance of practices (Shove, 2014), implying that interventions are an ongoing phenomenon in addition to being practices in themselves.

This chapter had two aims: to explore the territory between theories of social practice and epigenetics, focusing in particular on the relation between practices and performers' bodies; and to consider what such a focus might offer in terms of health interventions. In undertaking this theoretical exploration, I have discussed differences in the units of study and in conceptualisations of bodies, and have also identified points of connection between the two theories. A main point of intersection is that neither social practice nor epigenetics takes human bodies as the unit of study or as the site of change. Instead, in epigenetics, it is the gene (or the processes that change its expression) that is of interest, and in theories of social practice, it is practices in moments of performance and as entities. However, both strategies recognise bodies as media or substrates of change.

This discussion draws attention to the effects of practices on practitioners' bodies at a very detailed level, both in the moments of performances and over one (or

more) lifetimes. These processes are important for understanding complex health issues involving human–environment interactions such as obesity. Linking theories of social practice to epigenetics generates further related questions about the terms in which lifestyles, practices and related diseases are passed from one person to another. The mechanisms at stake may vary but such ideas question the heritability of diseases and forms of ill health currently in the NCD category.

Finally, such approaches have the potential to open up new frontiers in health promotion, health education and medicine. In particular, there is an opportunity to work with theories of social practice, drawing on a practice-epigenetic stance in conceptualising and analysing practices of childhood and child rearing. There is a strong case for this move considering the importance of early childhood development in most aspects of health and wellbeing. Furthermore, interventions based on ideas of recruitment, training and heritability generated through practice-epigenetics thinking have the potential to shift the current burden of disease and to promote greater and more equitable forms of wellbeing, although the timeframes will necessarily be long term and over several generations.

Empirical research informed by theories of social practice and epigenetics is challenging but also critical in the face of the scale and global impact of diseases now attributed to lifestyles.

Although there is new research developing in epigenetics that will have profound implications for future work on lifestyles – and more intriguingly, social practices – more needs to be done, in particular in understanding the underlying molecular mechanisms of transfer. These concerns aside, according to Bird (2007: 396), epigenetics has ‘caught the general imagination’ and there are no signs the enthusiasm it has generated will slow down any time soon. Dawkins classic text The selfish gene (Dawkins, 1978) proposed a new gene-centred view of evolution. Looking ahead, could we and should we anticipate a gene-centred view of social practices? Only time will tell.

The incorporation of material entities as integral elements of social practices, or as inextricably bundled with them, is a significant development in theories of social practice (Reckwitz, 2002a; 2002b; Schatzki, 2002; 2010a; Shove et al., 2012). As Reckwitz notes, “‘artefacts’ or “things”... necessarily participate in social practices just as human beings do’ (2002a: 208). Reflecting a similar idea, Shove et al. (2012) describe materials as one of three broad categories of element that are actively integrated when a practice is enacted. Yet materiality is diverse and additional material roles in the life of practices beyond their status as direct constituents of performance have received less attention. This chapter aims to characterise some of these more indirect material–practice relationships by exploring two examples of automated ‘machines’: central heating and fully automated factories.

My intention is twofold. First, I hope to extend the conceptual tools for positioning technologies of different kinds within accounts of practice. In so far as practice theories focus on what human bodies do, they may be limited when it comes to a wider consideration of what humans do collectively beyond their bodies; that is, through technological systems that carry out processes at varying degrees of distance, in time, space and awareness from the activity of people. Typically, these extended relationships are not recognised if technologies are only or mostly conceptualised as tool-like elements implicated in the ‘practical’ doings and sayings that define practices. Such interpretations become increasingly problematic as new forms of digitally automated and autonomous technologies come into use. A broader view is also required if we are to represent and analyse changing patterns of resource use, especially energy, which enables various forms of automation.

Second, I argue that other ways of conceptualising technologies and their dynamic relations to and within practices are important in analysing social change

more generally. In principle, materials and technologies are highly significant for how practices develop and change over time, not only as 'elements' of practice, but in other ways as well. In broad terms, technologies have dramatically altered the nature, range and qualities of the contemporary 'population' of practices by reducing and reconfiguring the contributions and qualities of human participation, how such practices are reproduced and whether and how they persist, evolve or dissolve. The concept of automation, for instance, indicates the potential for machines to 'take over' work that was previously carried out by people. In other words, some technologies appear to be important for how practices persist and change without directly 'participating', as elements, within their performance. How can these relationships be represented? And what is their significance for analysing the dynamics of practices?

To explore these questions, I start by examining how technologies have been positioned within theories of practice. The idea that practices are comprised by elements and that they change as these elements and their interrelationships change has been central to developing understandings of social change in terms of practices (Shove et al., 2012). Although valuable, I argue that such conceptualisation of materials and of their role in change is limited. In the second section, I provide an example: I consider how automated machines are powerfully implicated in modulating another mechanism by which practices change as 'populations' of practitioners change. In the third section, I extend the example of automated production to the contemporary phenomenon of fully automated factories and also draw on a contrasting example of automated central heating to conceptualise a variety of relationships between automated technologies and practices that form interconnecting constellations of practices and materials. In the fourth section, I reflect on what might be distinctive about the dynamics of such machine-practice relationships.

The idea that "'artefacts' or 'things' ... necessarily participate in social practices just as human beings do' (Reckwitz, 2002a: 208) is largely derived from the ideas of Latour and others within science and technology studies, who argue that action is 'distributed' between people and objects (or 'non-humans') such that 'implements ... are actors, or more precisely, participants in the course of action waiting to be given a figuration' (Latour, 2005: 71, original emphasis). In incorporating this line of thinking, Reckwitz (2002a: 221) remarks that 'the things handled in a social practice must be treated as necessary components for a practice to be "practiced"'. Yet, they are not the only components, as indicated in Reckwitz's much cited definition of a practice as:

[A] routinized type of behaviour which consists of several elements, interconnected to one other: forms of bodily activities, forms of mental activities, "things" and their use, a background knowledge in the form of understanding, know-how, states of emotion and motivational knowledge.

(2002b: 249)

Drawing on this general idea, Shove, Pantzar and Watson identify three broad kinds of elements (material, competence and meaning), which constitute practices when actively integrated by practitioners (Shove and Pantzar, 2005a; Watson and Shove, 2008; Pantzar and Shove, 2010b; culminating in Shove et al., 2012). These elements define the practice as an entity distinct from other activities, act as a set of 'resources' that organise the practice and are themselves constituted through instances of engagement in the practice (performances). Through this recursive interplay, practices come into being as links between elements are established through performances, evolve as new elements and/or new interlinkages are (per)formed and disappear as links are broken. That is, practices change as the elements of which they are composed change or as relations between these elements are reconfigured.

However, when viewed as 'things and their uses' or as 'things handled', material elements are conceptualised as tools or implements. Indeed, the concept of the active integration of elements, as performed by the practitioner, tends to focus on materials that are directly mobilised, attended to or manipulated in practical activities. But as Rinkinen, Jalas and Shove (2015: 1) remark, objects are encountered and engaged in multiple relations beyond 'enactment of social practices'. It seems other kinds of materials and relationships figure in the lives of practices, too.

Arguably, the category of materials, or material elements, in the plural is also problematic by its apparent reference to discrete and bounded physical entities, 'encompassing objects, infrastructures, tools, hardware and the body itself' (Shove et al., 2012: 23). Yet also referred to by Shove et al. (2012) as materiality, this category is reasonably interpreted more broadly: to represent the inherent materiality of doing and saying including structures in and on which this takes place, aspects of the environment like air and water, as well as less tangible phenomena such as sound and heat. As a dimension or aspect of practice, materiality involves, but is not necessarily synonymous with, the material entities that are present or necessary.¹ Specifically, the relationships and interactions between entities, such as bodies and technologies, might also be considered as part of the materiality of practices.

Yet the conceptualisation of materiality within practices is not my departure point. Instead, I focus on relationships that appear to matter beyond performance, at least, of single practices. This is the case with automated machines that carry on forms of 'activity' that are distanced, in different ways, from the flows of human activity, and as such would normally be absent from an account of materiality in practice. Accordingly, the potential roles of such materials in giving form and shape to practices and to the mechanisms by which they change is also not fully captured by a discussion of 'material elements'.

In this respect, Schatzki's (2002; 2010a) framing of materiality as material arrangements that are linked to, but conceptually distinct from, practices is of interest. Here, practices and material arrangements each provide a context for the other and are accordingly bundled together, persisting over time in interlinked patterns. While 'practice-arrangement nexuses' (Schatzki, 2010a: 130) or 'bundles' (Schatzki, 2012) are not dissimilar from Shove et al.'s (2012) notion of practices (Schatzki, forthcoming),

the concept of material arrangements is not restricted to the materiality of practice performances. This allows Schatzki (2012: 4) to list a number of other ways in which 'practices effect, use, give meaning to and are inseparable from arrangements while... arrangements channel, prefigure, facilitate and are essential to practices' through relations of causality, prefiguration, intelligibility intentionality and constitution (Schatzki, 2010a; 2012).

From this perspective, artefacts like machines have a certain default independence from practices, even though they are necessarily linked to them in a variety of ways. The challenge therefore comes not in thinking about whether the automated processing undertaken by machines is related to practices, but how. Specifically, are other interrelations, beyond co-constitution, salient in positioning and understanding such machines in practice theoretical accounts, especially those concerned with processes of change and stability? And how are these relations performed through bodily interactions? This is implied since, to Schatzki, as to Shove and colleagues, the actions involved in practices are 'bodily doings and sayings... that people directly perform' (Schatzki, 2002: 72).

It is worth noting before proceeding that for purposes of clarity, and despite my comments above, I continue to refer to machines as examples of 'materials' following the established understanding of this category as one of physical entities, including artefacts. More specifically, I understand machines to be examples of technologies, again understood as artefacts (Mitcham, 1994) and, in particular, ones that in any given historical period 'materialise' relatively new adaptations in design, techniques or tasks.

The account developed by Shove et al. (2012) has more to say about what is involved in the conduct, reproduction and dynamics of practices than is captured in the language of elements alone. Specifically, in addition to changes in the elements of practice and their interconnections, they note that practices also change as the populations of practitioners who sustain them change and as the connections between practices change (Shove et al., 2012; Watson, 2012).

For instance, Shove, Pantzar and Watson note that 'the contours of any one practice – where it is reproduced, how consistently, how long and on what scale – depend on changing populations of more and less faithful carriers or practitioners' (2012: 63, original emphasis). In other words, who undertakes a practice, and how, has implications for how that practice changes from within, through the creation and circulation of variety (thus the configurations of elements). In addition, if practices can be seen to colonise 'peoples' time and energy' (Shove et al., 2012: 65) these are important 'resources' by which practices connect to each other, for example, through forms of competition and collaboration. In simple terms, time spent 'energising' one practice cannot be

spent on others, with the exception of multi-tasking and blending of multiple practices (Shove et al., 2012).

It is therefore important to reflect on the ways in which forms of participation are modulated and mediated through material relations and therefore how they might figure as a source (and also an outcome) of change. For example, Shove, Watson, Hand and Ingram (2007) describe how the redesign of materials as inert as radiator fittings or varnish enables people with less-specialist skills and experience to use and apply these products. This is important for who can get involved in home improvement and hence how DIY and professional practices are reproduced. Equally, ready meals and pizzas are forms of technology that modulate and in a sense delegate the competencies involved in making dinner thus influencing who cooks at home and what it means to do so. Moreover, with the development of digital technologies, there is increasing debate, and also concern, about the redistribution of knowledge and service work to programmable machines in post-industrial economies (e.g. Ford, 2015).

Indeed, there is no shortage of narratives that explain or foretell dramatic social changes as a result of new or different technologies. Machines, as contrasted against tools, have been the focus of much of this debate (e.g. Hegel, Marx, as discussed by Heilbroner, 1967; Mumford, 1934; Illich, 1973; Schumacher, 1989). The following account is provided by Leder in an endnote to his book The Absent Body (Leder, 1990: 179–81), drawing on Tondl's (1974) categories of technology. It is characteristic of a widespread understanding of the significance of machines, but is also notable for its focus on the body.

Tondl (1974) outlined three broadly chronological stages of technological development in which the 'body-implement relationship' differs (Leder, 1990: 179). The first phase is an era of 'tools proper' that are wielded and powered by the body. Through skilled use, such tools become habitually incorporated into experience in an 'embodiment' relation, defined by a diminishing awareness of the tool itself, as a focal object. Tondl's second phase of technology is characterised by machines, which are devices powered by non-human energy sources, originally 'natural' ones such as water and animal power, and later mechanical power. People are still involved in working with machines, but in a very different way: they guide and control machines but do not provide the motive force. This places the body into a mode of 'background disappearance' (Leder, 1990: 180), having a supporting relationship to the machines whose 'needs and rhythms' provide the 'pattern for the work' instead of those of the human body. The third phase is one in which 'automated machines' carry out many of the control functions for themselves, through monitoring and regulation. Leder suggests: 'direct bodily involvement is even further reduced... primarily because it has been put out of play' (1990: 180).

In this account, three categories of technology are defined by their interplay with humans and specifically by whether the body 'powers' them and/or 'controls' them. It echoes Mumford's (1934: 10) typology of technology in which the 'essential distinction between a machine and tool lies in the degree of independence

in the operation from the skill and motive power of the operator'. As Mitcham (1994: 168) notes, this is a key shift since: 'as the machine becomes increasingly independent of direct human energy input, it becomes not just a static object but the bearer and initiator of operations or of special physical, chemical or electrical processes'. In other words, while tools are 'handled', or otherwise controlled directly and thereby 'participate' in the activities that comprise social practices, machines denote a degree of autonomy from direct bodily power/energy and intervention. As machines are developed and deployed, there are exchanges between human and technological 'participation' in work. In essence, Leder's account is of the progressive 'taking over' or 'delegation' (Latour, 2005) of previously human roles by or to technologies resulting in the effective reformulation (or even dissolution) of former production practices.

Before turning to consider this outcome in more detail in the next section, it is worth noting that such general narratives should be treated with caution and this is no exception. First, the idea that the transfer or delegation of work from people to machine (or vice versa) are ever directly or successfully achieved must be tempered. As Latour (2005: 70) warns, the term delegation should not be taken to imply that people (as Homo faber) are fully in control of what technologies do or that they deliberately hand over certain aspects of work and responsibility to them (machine or tool). Instead, and as Ruth Schwartz Cowan (1989) argues in her analysis of the changing nature of female domestic work over the twentieth century, technologies designed to automate and ease (house) work, have had unintended consequences, often introducing new forms or standards of work of their own. Moreover, not everything that technologies do is a replacement or substitute for human effort. Technologies are often useful in practice because of the way they extend the capabilities of human bodies and the possibility for human action (Kline, 2003; McGinn, 1991; Wallenborn, 2013).

Second, the distinction between machines and tools is not an essential feature of the artefacts themselves, but rather an outcome of their (changing) relationships within practices. For instance, Idhe (1993: 34, quoted in Verbeek, 2005: 117) argues that 'once taken into praxis one cannot speak of technologies "in themselves", but as the active relational pair, human-technology'. Since practices involving an artefact or its 'contexts of use' vary, technologies have more than one definition, a concept Idhe (1993: 20) describes as 'multistability'. Yet, 'contexts of use' also imply relations between technology and other materials, such as those that are stored, provide power or are powered and other mutual transformations. Thus, in addition to tool and machine, Mumford (1934) distinguishes further types of technology: utensils (pots, baskets) and apparatus (dye vat, kiln), both of which affect chemical transformations, and utilities (roads) including those that are powered (railway, electricity), to which Mitcham (1994) adds structures (buildings).

Nevertheless, it seems that through the progressive transformation of work practices, automated machines have powerfully modulated who participates and how and thereby whether certain practices persist or disappear. So let us return to the

question of what happens to the relationship between machines and practices when the machines 'take over'.

If bodily doings and sayings in dark factory production and central heating are reduced, and routinely removed, do (former) production and heating practices dissolve and disappear? Or are they still carried on by the machines? There is a distinction to be drawn here: it is fully consistent with Shove et al.'s (2012) framework and with Schatzki's theory of practices (2002; 2010a) to claim that the processing machines undertake is not in itself a practice, if there is no bodily activity that is a part of this process. This suggests that, indeed, some production practices may disappear as production tasks are automated. But this is not to say that such automated processing is not still part of a practice or sets of practices or, in Schatzki's approach, adjoined or bundled to them.

The concept of practices might be 'extended' to include the operation of machines that share or take over the same tasks as human practitioners, but which occur at some temporal or spatial distance from a range of human-enacted activities. For instance, Schatzki (2010a: 137) refers to 'the practice of warming houses'. This is a collective and dispersed definition of practice in which no single performance or practitioner carries out the practice of heating houses at any one time. Rather, it is achieved across the distributed activities of builders, engineers, planners, plumbers, safety regulators, window and insulation fitters, energy companies, as well as the efforts of inhabitants to understand and configure systems as they 'set things up' (Schatzki, 2010a: 129).

From this point of view, there might not be much difference between thinking about automated machines as part of a wider set of practices as opposed to being part of a single (but broader) practice. However, the former is preferable if we choose to work with a 'tighter' definition of practices and if we reserve this term to describe activities that are meaningful and identifiable to their practitioners. This strategy gives us more opportunity to think through what these less-than-direct relationships between practices and technologies mean for analysing change.

To conceptualise larger sets of practices and the forms of interconnection that characterise them, a number of concepts have been suggested. For instance, Shove et al. (2012) differentiate between bundles and complexes of practices, depending upon how interdependent and necessary are the interconnections between them. Kemmis, Edwards-Groves, Wilkinson and Hardy (2012) formulate inter-practice relationships in terms of ecologies of practices, while Watson (2012) refers to a 'systems of practice' approach. In contrast, Schatzki (2002; 2010a; 2015a) provides a multiply interconnected account in which practices are linked to other practices, arrangements to other arrangements and practices to arrangements, the latter forming bundles. Bundles connect to other bundles forming constellations, which together form a plenum, 'an immense maze of interconnected practices and arrangements' (Schatzki, 2015: 16).

Positioned as elements that circulate and play roles in multiple practices, materials (just as with other elements) are important in conceptualising how practices relate to one another. For instance, they might be shared or competed for. In addition,

the products of one practice often serve as inputs to others, forming the basis for sequences and other temporal connections (Shove et al., 2012; Nicolini, 2012). In such ways, 'material systems' like infrastructures coordinate and configure connections between practices in 'a trellis-like framework through and around which the combining and loosening of practice complexes occurs' (Shove et al., 2015: 10; see also Shove, 2016). In fact, reconfigurations in the ways that practices connect, bundle together or compete, is one of the key mechanisms identified by Shove et al. (2012) through which social change occurs. Conceptualising materials as elements helps to recognise and follow these dynamics. So can automated machines be positioned as elements within systems of practice? Or might such materials also play other roles in these systems?

In the case of dark factories, just as with central heating, we can identify a range of related practices in which the factories, their component machines and the inputs and outputs of their processing (raw components and products) might represent elements: in the activities of managers, engineers, mechanics, designers, marketeers, delivery drivers and so on. So while the practice of directly operating the production machines on a day-to-day basis may have disappeared, there is a whole system of other practices that remain firmly interconnected with the now fully automated machines. Below, I consider how these connections might be analysed by reference to an extended notion of the production process.

At the same time, it seems that the dark factory and its machines do not feature in associated practices in the same ways: in some, such as management practices, the machines might not have any direct or immediate material role in enacting such practices. Similarly, in the case of central heating, despite the range of practices that enable the operation of such machines, their primary significance resides in their ongoing detachment from practice. Thus, to situate automated machines as elements of potentially multiple practices is only a partial answer to the question of how to conceptualise them. Questions about the nature of these relations remain and if anything, are deepened by bringing into view the variety of practices to which central heating systems and dark factories connect. This observation is not unique to automated machines, but these examples are useful for exploring some of the varied roles that technologies may play. So before returning to the question of how to conceptualise systems (or constellations) in which central heating and dark factories are embedded, it is important to consider further what these roles may be.

Davide Nicolini

In this chapter I will discuss how a practice-based sensitivity can be used to address big issues and 'large-scale phenomena'. Examples of big issues include the nature and functioning of the financial market, large institutional arrangements, the education system, bureaucracy and the future of the planet. The topic is central to the advancement of practice-oriented studies. Practice-based sensitivities are often pigeonholed as part of microsociology and thus deemed unsuitable to deal with some of the big issues of our time and are of scarce importance outside academic circles. How we address 'large-scale phenomena' is therefore closely related to the issue of the relevance of practice-based studies and the practical uses of practice theories.

In the chapter, I will use the terms 'practice-based studies', 'practice approach' or 'practice lens' to denote a family of orientations that take practices as central for the understanding of organisational and social phenomena. Authors who embrace this orientation suggest that matters such as social order, knowledge, institutions, identity, power, inequalities and social change result from and transpire through social practices. While they all agree that the study of the social needs to start and end with social practices, they use different sensitising theoretical categories, research methods and discursive genres. As I argued elsewhere (Nicolini, 2012), I see no merit in trying to reduce this multiplicity in search of a unitary or unified theory of practice and this equally applies to a discussion of how practice-based scholars approach large-scale phenomena. Accordingly, my aim here is not to build a practice-based theory of big social phenomena or to theorise large-scale social phenomena by building upon a specific type of theory. Rather, more modestly, I am interested in the different methodological and theoretical strategies that authors in the practice-based camp have used to address large-scale social phenomena. While I do have my own view of what practices consist of,¹ I will here simply examine the question

from different practice-based vantage points and discuss how 'large phenomena' transpire amid and emerge through different theories of practice.

I will start my discussion by contrasting the position of practice-based approaches to 'large-scale phenomena' vis-à-vis the idea of macro-phenomena, levels of reality and localism. I will then critically survey some of the ways in which practice-oriented scholars have addressed 'large phenomena' and comment on their affordances and limitations.

In this section, I will examine three ways to respond to the question 'how can we understand large-scale phenomena using a practice-based sensitivity?' The first builds upon what I call textile metaphors (texture, net, fabric, weaving), the second focuses on how the global manifests in the local, the third concentrates on the practices through which large-scale phenomena are assembled. As we shall see, these broad approaches constitute what elsewhere I called 'theory method packages' (Nicolini, 2012). Although they are related in more than one way, these packages embrace slightly different ontological assumptions and propose different practical approaches to making sense, studying and representing large-scale phenomena from a practice-based perspective.

What prevented social theory so long from recognising the fundamental social importance of emotional and affective phenomena?² Since the 1980s, the exponents of the so-called affective or emotional turn (Clough and Halley, 2007) in the social sciences and humanities have been convinced of the need to alter fundamentally our understanding of the social. Of course, all this talk of turns is a means of dramatising and simplifying the situation. There does not exist some coherent block of traditional social theory that needs to be overcome, but rather a heterogeneous conglomerate of texts dating from around 1900 that later became the object of a specific style of interpretation in mainstream sociology. It is obvious that the work of certain authors from that period, such as Gabriel Tarde with his sociology of imitation (Borch and Stäheli, 2009), or Sigmund Freud and psychoanalysis (Elliott, 1992), demonstrated recognition of the fundamental social importance of affects. The dominant reception of Weber and Durkheim through Parsons and the theory

of modernisation and on through to such diverse authors as Niklas Luhmann, Jürgen Habermas, Michel Foucault and Pierre Bourdieu is, however, characterised by a systematic neglect of affectivity. Two interconnected causes can be assigned to this tendency. One reason is the widespread assumption in sociology of the identity of the social with normative orders or orders of knowledge, resulting in an understanding of affects as non-social, non-cultural phenomena occurring within individual's bodies or individual's psyche. A further reason is that modernity used to be equated to formal rationality overcoming affects.

The first reason for the anti-affective attitude of large segments of classical and contemporary social theory seems to be the identification of the social with the intersubjective force of normative rules. Since the late nineteenth century, social theory has been dominated by dualisms opposing the social to the individual and to the natural or biological (Lukes, 1973). This pair of dichotomies usually occurs in combination with a third, even more abstract dualism between the rational and the irrational. Admittedly, the social sciences have not entirely overlooked emotional, affective phenomena. But they have located them on the second, inferior pole of these pairs of opposites. Emotions were regarded as properties of the individual and thus excluded from sociological generalisation or seen as natural, biological dispositions and drives belonging to the pre-social body. In both cases they are placed beyond the rational, regular and predictable social order.

There have always existed opposing tendencies, of course, the late works of Émile Durkheim, rediscovered in the last few decades, being the prime example. In The elementary forms of religious life from 1912, Durkheim (1912) discovers that in religious communities, emotional ties do not subsist outside of the social but rather exercise a stabilising, integrating force within it. Durkheim is nevertheless doubtful that these emotional aspects of social relations can be reproduced in modern societies (Shilling, 2002).

Strangely enough, the comprehensive paradigmatic shift in the social sciences since the 1970s, the variously titled 'cultural', 'interpretive' or 'textual' turn, has not overturned the anti-affective attitude. Structuralism, poststructuralism and social constructivism have transformed our understanding of the social by substituting semiotic and discursive structures and regimes of knowledge in the place of prominence previously occupied by normative orders. The paradigm of the social is no longer religion or law, but rather language: society must be understood in analogy to language. Affects, as states of bodily excitation, then seem to stand outside of the linguistically constituted social. Emotions appear for this reason only on the margins of classical works of cultural theory like Foucault or Bourdieu. If emotions are noticed at all in such culturalist approaches, then as objects of specific discourses, as phenomena of lingual construction. This reduction of emotions to discourse, such as was promoted by social constructivism in the 1980s (Harré, 1986), made the social, cultural representation of emotions accessible to analysis by social science, but remains imprisoned in the cultural/material binary.

The second reason why social theories have so long marginalised affects and emotions is to be found in the dominant sociological understanding of modernity.

Such disparate authors as Marx, Weber, Adorno, Parsons, Foucault and Bourdieu share the common conviction, whether affirmatively or critically, that modern society is most fundamentally characterised by formal rationality. The rationalisation of action and the social spheres is seen as increasingly displacing all affective aspects of action. The affects seem more to belong to premodern or traditional societies, which are then classified as natural rather than social. In this way, the dualism between modern and traditional societies underlies a distinction between different stages of social evolution in terms of the absence or intensity of affects. Norbert Elias's (1976) major theory that the process of civilisation toward modern society is equivalent to the increase of affect control would seem paradigmatic of this type of interpretation.

The traditional social theory thus adopts the premise that modern society is characterised by what in Parson's theory of modernisation is termed 'affect neutrality', implying that in the differentiated spheres of action in modernity, emotions are largely neutralised. Many authors, from Weber to Habermas, view this alleged affect neutrality in modern society as essentially positive. The suppression of dangerous emotions looks progressive when seen from the perspective of the rationalist opposition of enlightened understanding to mere feeling (a binary that nineteenth century theories of gendered character index onto the two sexes). The reverse image of this suspicion of emotions is their latent or manifest celebration, such as held sway at the interwar Collège de Sociologie, which attacked modernity for its claim to suppress emotions, hoping in Rousseauian manner for a return of individual and collective feelings, whether sensuous, physical or what not.

In summary, the strict dualism between the social and the biological or individual, in conjunction with the theory that modernity is distinguished by affect neutrality, whether benign or problematic, is the reason for the exclusion of affects and emotions from social theory. The affective is confined to the individual, the biological or corporeal or banished to premodern societies. It has all the trappings of the outside, constituting the inside of affectively neutral society.

A practice theory perspective on affects is concerned neither to denigrate nor to celebrate the affects. The presence of affects in a society cannot be simply ignored. Whether they are menacing or benign they are a constitutive part of the social life which incessantly produces them. The diagnosis of modernity as affect neutral is therefore false. The affects may be shaped differently in some modern institutions but they do not vanish altogether. Practice theory avoids the cultural/material binary by approaching every social order, conceived as an arrangement of practices, as its own affective order; every social practice is then affectively tuned in a particular way and has, as such, a built-in affective dimension (see also Seyfert, 2011).

What then is the particularity of a practice theory perspective on affects? I would suggest three underlying principles of such a perspective: (1) affects are not subjective, but social; (2) affects are not properties, but activities; (3) affects are states of

physical arousal, of pleasure or displeasure, directed at some definite person, object or idea.

The practice theory approach brings about a fundamental change of perspective on affectivity. It approaches affects not – as the terms emotion or feeling have traditionally suggested – as interior properties of individuals only accessible to an introspection plumbing the depths of the psyche, but places them instead on the level of social practices themselves. Affects are then properties of the specific affective ‘attunement’ or mood of the respective practice.³ As soon as a person is competent to perform a practice and is ‘carried away’ by it, she incorporates and actualises its mood. This state of affairs is ordinarily obscured by an individualist bias in ordinary language. When, for example, an early twenty-first century person falls in love, it only seems to be an individual feeling (or even a basic anthropological constant). In fact, the feeling is embedded and informed by the bundle of practices called love that developed around the late eighteenth century in Western culture. ‘Love’ is a set of routine behaviours dependent on specific cultural patterns (the uniqueness of the beloved, fascination for their otherwise mundane seeming qualities) and containing a peculiar set of affects: physical desire, longing in the absence of the beloved, the existential pain of love unrequited or lost. These routine behaviours are clearly closely tied to discursive practices and fields, such as romantic novels and films in which the codes and affects of love are publicly represented for imitation.

It is apparent here how the affective practice of falling in love is both cultural and material at once. It is cultural in that it depends on the specific cultural, historical and local knowledge schemata referred to above and with which participants in the practice think, feel and remember in a certain way. The affect must be cultural since it is oriented on specific entities in the world, which first become desirable or repellent within the framework of certain systems of interpretation. The practice is at the same time material to the extent that the affects inherent to it are real states of physical excitement, manifesting in measurable physical reactions or at least in the subjective fact of physical feelings which can only be ignored with the greatest effort. Falling in love involves such physical facts, the cultural origins of which are no longer visible to the lover. The affective structure of a practice thus explodes the inside/outside binary (see Burkitt, 1999), by being internal and external at the same time. It consists in external, public, intelligible signs of emotion and internal, subjective states of physical excitement.⁴

The bundle of practices called love is manifestly affective. However, on closer inspection, all social practices turn out to be affectively attuned in one way or another, including the allegedly neutral ones. Academic practices in the modern fields of the natural sciences or humanities turn out to involve attitudes of curiosity and attentiveness to the nature of things. Affects enter into economic practices in the modern market economy, associated with success or victory in competition or the joy of creative work. The question is why this affectivity not only occurs in special cases of social practices but is a general phenomenon. Two main structural properties of practices can be said to require the presence of affects: motivation for the practice and the focussing of attention.

Social practices 'interpellate' the subject in a certain way. But how can the subject be interpellated in this way and thus participate in the practice?²⁵ The answer is that the practice must entail a specific motivation to perform it. From this perspective, it is not the individual who comes to the practice with their own 'psychological' motivation, but rather the practice itself of which the motivation is already an integral part. Motivation is where affects come into play; there must be some affective incentive to participate in the practice. This can be a positive desire, a defensive incentive to avoid displeasure or a combination of the two. Intrinsically creative work in late modernity operates characteristically on the basis of a built-in motivation for creative enjoyment, whereas a serf's labours will tend to be motivated by the threat of corporal punishment for disobedience. Generally, I would state that motivation without a process of being affected by something is not thinkable. However, we must leave aside the problematic 'individualist' connotation of the term motivation in favour of the praxeological idea of motivations 'embedded into' social practices themselves. Of course, this does not exclude the possibility that under certain circumstances individuals do not fully participate in the practice as they lack being successfully 'interpellated' into the underlying motivation structure. The possible failure to participate in a practice thus is not only a question of lacking corresponding skills and interpretations, but also of lacking corresponding desires and fascinations.

Just as every practice must have its particular built-in motivation structure, so too must the directedness of sense perception be organised in a specific way. The fundamentally volatile human sensory attention is guided by a practice toward phenomena relevant to the practice and deflected from irrelevant ones, which then drop out of the radar altogether. Of course, bundles of practices have diverse possibilities of how to direct attention. Above all, material artefacts are such means of influencing attention. A specific architecture and interior design, for instance, subtly directs the attention structure of the subjects in certain ways. Media technology succeeds in doing this in different ways. There might also exist explicit norms of attention, for instance at work or in school classes. However, in our context it is crucial that perceptual attention is laden with affects (see Ciompi, 1997) so that the affectiveness built in practices succeeds in directing attention in a profound way. In academic practices, for instance, it can be the curiosity or interest for certain themes that guides attention; in body practices, it can be the sense of shame at one's own impurity. Any perception involving strong positive or negative affects will enter the focus of sensuous attention, while entities with weaker or neutral affective significance will remain beneath perceptual attention, like white noise.

Clearly then, a practice theory cannot conceive the affects as qualities or properties, but must regard them as dynamic processes and relations. Within a practice, people can be affected in specific ways by other people, by things and ideas. Affectivity is therefore always a relation between different entities. For this precise reason, I prefer the term affect to that of emotion. The term affect may have the disadvantageous connotations of some spontaneous pre-cultural force. Yet the term emotion seems more problematic still, suggestive as it is of something a person 'has'. The concept

of the affect, with its verb form 'to affect', is more fully the description of processes, of affecting and being affected (see Ott, 2010, and in the background Deleuze and Guattari, 1987). As such, the question must always be: who is affected by whom or what? From the point of view of the human subject, it can be said that the affectivity of a practice comprises specific stimulations attached to other people, things or ideas. These latter are interpreted within the practice in some typical way and so can become triggers for feelings of pleasure or displeasure. An affect is therefore always directed at something of significance, always intentional in the phenomenological sense, in that it involves the interpretation of some entity as desirable, revolting, etc. In contrast, the same entities trigger entirely different states in people within other practices or leave them indifferent. A basic capacity for reactions of desire and aversion, pleasure and displeasure, would seem to enter the human body at the point of birth. But the direction this desire or aversion tends, the objects it is formed by, is a question of the social practices. Luc Ciompi (1997) suggests we should assume the existence of five basic feelings – joy, interest, fear, rage or sadness – which list I would take on board rather as a heuristic catalogue than as a fundamental anthropological constant.

In the new debate, especially authors working in a Deleuzian framework such as Brian Massumi (2002) have preferred the term affect over emotion, because it denotes a disruptive force that breaks through cultured routines. Where does this disruptive element fit into the analysis being proposed here? At this point, we must back up a bit. Practice theory indeed presupposes that practices constitute social orders and as such are undergirded by a structure of social reproduction and repetition. Yet at the same time, practices also always harbour the potential for novelty, surprise and experiment, which can modify or transform the practice from within. Therefore, on the account I am proposing here, affects are not per se anarchic and disruptive, but rather among the main ingredients in culturally standardised, routine bundles of practices. However, there is always the chance that new and different acts of affecting will emerge from within social practices and explode their normality. These non-routine affective events are indeed no rarity, although it is in every case questionable whether they will constitute their own regular practice or disappear again. Such a non-routine affective event can occur on the individual level – some idiosyncrasy, viewed in some cases as pathological, a unique phobia or fetish – or on the level of collective spontaneous excitement, be it a state of pleasure, panic or other.

A practice theory perspective on affects is thus able to assign the affect discourses their proper place. The 'thematization' or 'problematisation' of feelings has been itself a theme of intense discussion in cultural theory since the 1970s, which tended to regard feelings as 'culturally constructed'. As elaborated above, this kind of culturalist theoretical reductionism inevitably falls short of the mark by tending to confuse feelings with the semantics of feeling and so neglecting the physical facts of affects. The practice theoretical perspective in no way excludes such a discourse analysis of emotions and affects, but rather preserves them within the more comprehensive framework of a concept of practice-discourse formations. If we want to trace the

ramifications of affects as formed in specific discourses at specific times, then we cannot view them in isolation, but must analyse the connections between discursive and non-discursive practices. Especially in modernity, affects are often not implicitly and traditionally present in social practices, but rather amplified, weakened or even brought into existence by being treated as a theme or problem on the level of discourse. These discursive practices always already exist in interconnection with non-discursive practices, oscillating between the discursive and the non-discursive, becoming physically internalised and causing real effects in the body.

Any theory of social practices would end up in a cul-de-sac were it strictly to oppose practice and discourse, doings and sayings, as though they were two separate orders of things. If practices are grasped instead as a meshwork of doings and sayings, then discursive formations are just social practices, neither more nor less; they are practices of representation in which objects in the world are represented, imagined and evaluated, with the aid of media technologies. However, depending on the context, these practices of representation are linked to other practices, into which their contents are 'translated' (Latour). For example, the discourse of psychology is translated into a therapeutic practice. The exact form of translation between discursive and non-discursive practices is then a question of empirical research. One of the best examples for the relevance of affect discourses in the context of a practice-discourse formation is the previously mentioned modern discourse of love at work in higher literature and film, as well as in popular and academic, psychological literature, and coupled with practices of dating and mating, sexuality and partnership.

Now, the affective orientations instituted by social practices and practice-discourse formations can invest meaning in different 'affective objects' that can be roughly divided into three classes: persons, things and ideas. Where sociology has dealt empirically with emotions (see, e.g. the interactive sociology of Erving Goffman (1971) or Arlie Hochschild (1983)), it has concentrated on inter-subjective emotions, on emotions directed at other people, such as envy, affection or hate. Affects directed at other human beings, be they present or absent, are characteristic of many social practices. The affect can be directed at whole groups, be they, once again present or imagined, at a labour force, a crowd in a football stadium, an ethnic group. Beside humans, affective inclinations can also be attached to pure ideas, to transcendent entities, such as in religious practices or abstract entities, such as with an aversion toward 'the system' or trust in 'the market'.

Above all, however, the relations between people and inanimate things have an enormous and generally underestimated significance for social practice and its affects (see Knorr Cetina, 2001). The artefacts involved in a practice are often not at all affect-neutral but rather intensely charged with negative, or, in modernity more often, positive qualities. The care in engaging with tools, the glamorous fascination of a piece of clothing, the fear or enthusiasm at finding oneself in a metropolis, are all examples of this. An appropriate framework for analysing affects therefore needs to keep this interobjective dimension in view in order to compensate for the pre-dominance of the intersubjective in sociology. Within the sphere of artefacts two special constellations can be highlighted which, particularly in modernity, operate

as prominent generators of affect: atmospheres of place and reflexive, symbolic or imaginary artefacts.

The three versions of discourse analysis considered in this chapter approach language from what surrounds it. Instead of focusing on the properties of language as such (e.g. grammar, syntax), they treat language as an element of sayings and texts, which are embedded in social life. They theorise language as a social phenomenon and examine grammar and syntax only in so far as the latter either engage with fundamentals of social existence (as in Halliday's 1994 systematic functional linguistics) or reflect or implicate social phenomena.

As indicated, moreover, these three versions of discourse analysis highlight activity and set aside construals of language and discourse as abstract structures. These features reflect the ideas of two philosophers, John Austin and Ludwig Wittgenstein. Consider what is perhaps the most prominent definition of discourse: language-in-use. Because 'use' refers to activities, this definition treats discourse as an element or feature of activity. This way of thinking reflects Austin's contention that speaking is acting. Another widely cited notion of discourse is Gee's concept of Discourse:

I use the term 'Discourse' ... for different ways in which we humans integrate language with nonlanguage [sic] 'stuff,' such as different ways of thinking, acting, interacting, valuing, feeling, believing, and using symbols, tools, and objects in the right places at the right times so as to enact and recognize different identities, give the material world certain meanings, distribute social goods in a certain way, make certain sorts of meaningful connections in experience ...

(1999: 13)

This definition reflects an intuition that lies behind Wittgenstein's coinage of the term 'language game'. Wittgenstein (2009: §23) wrote: 'The word "language-game" is used here to emphasise the fact that the speaking of language is part of an activity, or of a form of life'. This quote claims that speaking, which is an activity, is part of broader activities. As suggested by the expression 'form of life', these broader activities encompass both further acts of speaking and non-linguistic doings (doings that do not use words, phrases, or sentences). Wittgenstein's intuition, in other words, is that linguistic activity is interwoven with non-linguistic activity in human life. This is exactly the thought expressed in the quote from Gee.

Discourse theorists advance different conceptions of the interwovenness of linguistic activity with other activity. Since each of the three theorists considered here

works with a notion of practices, these different conceptions can be described as alternative understandings of the relation of language or discourse to practices.

Scollon nearly submerges discourse in practices. His starting point is the idea of mediated action (e.g. Wertsch, 1998). According to this idea, human activity is always mediated by (meaningful) entities distinct from the actor. Material objects figure prominently among such mediators, as does language: when a person speaks or writes, language mediates her activity, that is, it is a means whereby her action is accomplished. For Scollon, moreover, a practice is a repeated action: the practice of x exists when actions of x-ing have been sufficiently repeated to be recognisable as x-ings. The crystallisation of repeated x-ings as the practice of x also coordinates with the development of knowing how to x in the bodies of those who x (Bourdieu's habitus). In addition, actions, in the locales in which they are performed, for instance classrooms, stores, kitchens and airport terminals, are usually performed as part of combinations or sequences of actions. When each of the (types of) actions involved is a practice, the resulting bundlings of practices in locales are called nexuses of practice.

Gee, meanwhile, holds that discourse is language-in-use or strings of spoken or written sentences. As indicated, he adds to these the idea of Discourse, which denotes ways of combining language with non-language stuff. The notion of Discourse is particularly important for the project of analysing language as a social phenomenon, for according to Gee, society is composed basically of Discourses (2014: 128). Identifying a Discourse, however, reveals little about the particular configuration of language and non-language stuff that it embraces. The notion of practice performs some of this work. A practice, or game (as Gee Wittgensteinianly prefers), is a 'socially recognized [sic] and institutionally or culturally supported endeavour that usually involves sequencing or combining actions in certain specific ways' (Gee, 2014: 32). Examples are mentoring a student, lecturing before a class and playing Yu-Gi-Oh. Like non-discursive actions, discourses as strings of sentences or as language-in-use are components of practices. Practices, in turn, are concrete forms of Discourse.

Fairclough propagates related views. He distinguishes discourse and practice in the singular from discourses and practices in the plural. His conception of these phenomena, however, shifts. Sometimes, for instance, 'practices' refers to types and tokens of action (e.g. Fairclough, 2015: 61). At other times the plural form is not used and 'practice' more or less means situated action. In this usage, the expressions 'discursive practice' and 'social practice' mean speaking and writing as situated, respectively, in processes of text production, distribution and consumption and in social institutions and organisations (Fairclough, 1992: 66–73). At still other times (e.g. Chouliaraki and Fairclough, 1999), practices are analysed as entities that combine (1) material activity (i.e. non-discursive doings), (2) discourse (i.e. semiotic entities such as language and images as used in activity), (3) social relations and processes and (4) mental phenomena. This third view offers a rich conception of practices that treats discourse as part of practices. As I discuss below, Chouliaraki and Fairclough annex this account of discourse to Bourdieu's theory of field and habitus

and hold (1999, 104–5) that it provides the account of this phenomenon missing and needed in Bourdieu.

Like practice theorists, the three discourse analysts under discussion disagree about whether the distinction between discursive and non-discursive action is fundamental (cf. Giddens, 1984: xxii). Scollon joins Giddens, Gherardi and Shove, Pantzar and Watson in upholding what might be called a ‘monolithic’ theory that portrays all activity as of one basic sort (e.g. mediated action in Scollon). By contrast, Gee and Chouliaraki and Fairclough join Reckwitz, Taylor, Bourdieu and myself in treating the discursive/non-discursive distinction as fundamental and in construing practices as composed of actions of both sorts.¹ In my opinion, the difference between discursive and non-discursive actions is fundamental to conceptualising practices and social life. Acknowledging it upholds the importance of language and facilitates a more nuanced understanding of what action accomplishes, for example, the varied differences and contributions that doings and sayings make to social existence. Consider, for instance, Gee’s idea that when people speak or write they at the same time ‘build seven things or seven areas of “reality”’ (2014: 32). These areas are significance, practices, identities, relationships, politics (the distribution of social goods), connections (between things/topics) and sign systems and knowledge. Non-linguistic doings contribute to these building tasks, too, although they do so in different ways and to different degrees than sayings do. Making the discursive/non-discursive difference fundamental facilitates explorations of convergences, divergences and entanglements in these contributions.

However, the conceptions of practices promulgated by practice-minded discourse theorists lack a key element: organisation. Practices are organised sets of doings and sayings. The only organisations that Scollon and Gee recognise are sequences, combinations, repetitions and co-occurring repetitions of action. These phenomena are patterns and, thus, organisations only of the weakest, empiricist sort. What Scollon and Gee neglect is the organisation of what informs action, in particular, the organisation of what informs the actions that compose practices.

What organises practices in this way informs both the saying and the doings that compose given practices. This insight is fundamental to a proper understanding of practices. I noted that Chouliaraki and Fairclough distinguish discursive from non-discursive activities. They hold, further, that discourses are organised by what, following Foucault (see section below), they call ‘orders of discourse’. Orders of discourse are socially ordered images, types of language and non-verbal forms of communication that structure particular social practices or fields (in Bourdieu’s sense of field; Chouliaraki and Fairclough, 1999: 106, see also 50–1, 56). Orders of discourse, however, structure only the discursive dimension of practices and fields. They do not also structure the ‘material activities’, which according to Chouliaraki and Fairclough also compose practices: ‘an order of discourse is a socially structured articulation of discursive practices which constitutes the discursive facet of the social order of a social field’ (1999: 114). Thus, although orders of discourse organise practices, they do not inform both the sayings and the doings that compose practices. Below I return to orders of discourse and to the idea that the discursive component

of practices has a distinct organisation. At this point I want to focus on the organisation that is common to doings and saying.

Elsewhere (1996, 2002) I have argued that the organisations common to doings and sayings are made up of rules, teleological-affective structures and both practical and general understandings. Practice organisations are teleological, normative and affectual structures, in the context, and out of a knowledge, of which humans who are brought up to act for ends, to heed normativity and to be affected emotionally proceed in their lives. The organisation of a practice also determines which doings and sayings belong to it. To say that practice organisations pertain to sayings as much as to doings implies, among other things, that sayings are intentional, oriented to ends, parts of tasks and projects and variously emotional, that they are carried out in light of rules and that they, to varying degrees, articulate general understandings. People do not just do, but also say, things when carrying on a given practice by way of pursuing ends, carrying out tasks or sets thereof and being imbued by particular emotions and general understandings. As people proceed through different spheres of life – work, family, religion, recreation, provision, etc. – they carry out practices through both sayings and doings.

Reckwitz's account of practice-discourse complexes contains an alternative conception of the common organisation of doings and sayings. Reckwitz construes practices (2002b: 249–50) as routinised types of behaviour (persisting blocks of bodily activity, mentality, background knowledge, emotion, motivation and things and their use). He treats discourses (2006: 43), moreover, as practices of the production of regulated representations, or practices of representation for short. By 'representations', he means textual and visual presentations (Darstellungen) of objects, subjects and contexts (which constitute these as meaningful entities in the first place). Practices of representation include those of speaking, writing, science, painting, filmmaking, sculpture, and the like. Note that this list can be usefully expanded to include practices (in Reckwitz's sense) in which representations are consumed, thus such practices as those of listening, reading, looking at and watching. Reckwitz claims that what holds practices and discourses together in complexes are orders of knowledge. These orders are composed of forms of implicit knowledge such as 'know-how, interpretive knowledge of routinised attributions of sense, and complexes of culturally modelled affects and motivations' (Reckwitz, 2008: 202). Such orders imbue practices and discourses, giving them their form. Discourses, in turn, produce knowledge orders and make them explicit (Reckwitz, 2008: 205). Knowledge orders, finally, are organised by cultural codes. These codes are networks of meaning patterns (Schemuster), 'systems of central differences and classifications' (Reckwitz, 2006: 44), for use in describing the world and dealing with it. These differences and classifications inform know-how, interpretation and forms of motivation.

Practice theories connect what people say (and do) to the organisation of social practices. Practices, in turn, fill out the social context in which people proceed. Like,

therefore, the three discourse analyses considered in the previous section, practice theories approach language as a social phenomenon. Their general account also bears some resemblance to M. M. Bakhtin's (1986: 60, 78) theses that people always speak in specific speech genres and that such genres are determined by the different functional spheres of activity and communication in society, for example, science, the technical, commentary, business and everyday life. Each of these spheres, Bakhtin claimed, develops its own types of utterance, which are constituted by typical speech situations, typical themes, typical expressivities and typical addressees. These typicalities reflect the conditions and goals of the spheres in which the genres they help compose develop. Relatedly, a speaker's choice of utterance type reflects not just thematic considerations and the concrete speech situation, but also the nature of the activity-communication sphere in which s/he speaks. Note that speech genres, like Fairclough's orders of discourse, concern discursive activity alone. Oriented primarily toward discourse, both theorists overlook the common organisation of sayings and doings.

To illustrate what it is to analyse sayings by reference to organisations that are common to doings and sayings, consider Hanks's notion of the participatory frameworks that belong to communication practices (practices such as conversations in which sayings take the lead). His notion is based on Erving Goffman's (1981) idea of a speaker's footing. A speaker's footing is her relationship to her own words. Direct, indirect and quoted speech exemplify different possibilities. Goffman proposed that the status of speaker be replaced by a triumvirate of roles: the animator (the person who makes the sounds), the author (the person who selects the words and phrasings) and the principal (the person who is responsible for the statements and opinions expressed). This proposal underlies Hanks's idea (1996: 207) that communicative practices carry participant frameworks that are composed of communication roles open to those who speak or are the addressees of speech. According to my account, these communication roles, like roles more generally (see Schatzki, 2002), are reference points for the apportionment of a practice's organisation: which tasks, projects and ends, for example, are acceptable for or enjoined of a participant in, say, educational practices depends on the role(s) she occupies in them. In, for instance, the linked practices carried on by the animator of a speech act (e.g. Mr Grey, a teacher), the author of that act (e.g. Ms Wiggins, a bureaucrat in the superintendent's office) and its principal (e.g. Dr Lee, the superintendent), the tasks, projects and ends that normatively fall to Mr Grey as animator differ from those that normatively fall to Ms Wiggins and Dr Lee as author and principal, respectively. Such role-based apportionment can also affect which non-linguistic doings people carry out: different non-linguistic doings are acceptable for or enjoined of the person (animator) who speaks at a parent teacher association meeting and the person (author) who earlier typed the text on a computer. Finally, the apportionment of acceptable and enjoined tasks, projects and ends among people who occupy particular communication roles can be infected by the wider roles that these individuals assume in the practices involved. For example, the doings that are acceptable for or enjoined of animators will vary partly depending on whether the animator is a school child

quoting a friend, a teacher announcing a central school office policy or a drama coach acting in a play.

Any position that grants equal footing to sayings and doings is obliged to examine sayings as such. I will not say much on this topic. Sayings are a kind of activity. They are activities in which something is said. The idea, however, of saying something is ambiguous. What a person says in saying something is, first, the words, sentences and strings of words and sentences she utters. These words and sentences, as vocables, are texts. They are texts, of course, that usually perish in the event. What a person says in saying something is, second, what she says. If someone says, 'the sky is blue' or 'bring me an apple', what she says in this second sense is that the sky is blue and to bring her an apple. Saying something in this sense is sometimes construed as assuming a relationship (e.g. asserting, wanting) to a propositional content. When what someone says in the second sense is so understood, sayings can be described as representings – as Reckwitz suggests in construing (linguistic) discourses as 'sign-using practices viewed from the point of view of their production of representations' (2008: 203). Talk of representations (or of propositions) as opposed to representings raises a host of issues debated by philosophers for decades. This is not the place to take them up.

Sayings as a general category of doing have been well analysed in speech act theory (Austin, 1975; Searle, 1969, 1985). As especially Searle's version demonstrates, to treat sayings as doings is to subsume one's account of speech into one's account of action. For instance, since practice theory ties actions to practices and practice organisations, in treating sayings as activities, it holds that they are elements of practices subject to practice organisations. In this context, it is crucial to emphasise Wittgenstein's insight (2009: §23), affirmed in speech act theory, that sayings can be actions of countless sorts, for example, asserting, denying, explaining, asking, complaining, describing, insulting, bothering, ordering, remonstrating, begging, celebrating and so forth. Emphasising multiplicity makes clear that what people are typically about in speaking is, not speaking as such, but performing an action to which the use of language is useful or crucial.

Under 'sayings' I include acts of writing. Writing is a very different activity than speaking, but they share several key features. One is the use of language. A second is the saying of things. Reckwitz would add, controversially, that they both produce representations. What both incontrovertibly engender are texts, that is, collections of meaningful words and sentences, although these texts are usually evanescent in the one case and more durable in the other.

Sayings as activities qua token speech acts must be distinguished from what Bakhtin called 'utterances', types of which he called 'speech genres'. Utterances are not like particular acts of asserting, asking or ordering that a person can perform without ado by vocalising some words. For Bakhtin (1986: 71) held that the boundaries of an utterance are given by a change in speaking subjects, and this claim implies that an utterance can encompass a series of actions, all performed by the same person. From my point of view, Bakhtin's utterances can be conceptualised as tasks, which can be elements of teleoaffective structures and which people

pursue by carrying out one or more actions, primarily sayings but also possibly non-discursive doings.

To conclude this sketch of elements of a practice theoretical account of sayings and texts, I point out that some phenomena that transpire within practices, or rather within bundles (see below), are best approached through bodies of investigation and theory that are different from practice theory but compatible with its ontology. A phenomenon transpires 'within' a bundle when it is essentially composed of or dependent on components of the linked practices and arrangements that compose a bundle. When either situation obtains, the bundle forms a constitutive context in which the phenomenon transpires, although what this more specifically involves can vary. Examples of phenomena that transpire within bundles in this sense are interactions; the dissemination of knowledge; power and domination; experience, including aesthetic experience; and the constant adjustments that actors make to the world. I mention all this because another example is understanding and interpreting texts and doings/sayings. Understanding is present within bundles both as action and as condition, and to theorise it, practice theory might be advised to draw on analyses of a philosophical sort. A further example is conversation, which is a type of interaction that transpires within practices. To theorise it, practice theorists might draw on the established body of work called 'conversational analysis' (see Nicolini, 2012). A final example is texts, which can circulate within and among practices. A sort of approach that practice theory might appropriate in analysing texts is text analysis: the analysis of their vocabulary, grammar, cohesion and structure (see Fairclough, 1992: 75ff).

Practices – organised manifolds of doings and saying – connect to material arrangements – composed of bodies, artefacts, living creatures and things of nature – to form practice-arrangement bundles (e.g. Schatzki, 2002). Such bundles, in turn, connect to other bundles to form wider constellations of practices and arrangements. Social life transpires within these bundles and constellations; all social phenomena consist of sectors, slices or aspects of bundles and constellations. Bundles and constellations, moreover, assume diverse shapes and sizes. When, for example, bundles connect, the resulting constellations are typically larger (in the sense of spatial extension) than the original bundles; repeated connections that are not matched by significant decouplings result in very large constellations such as those in which governments or economic systems consist. Interconnected constellations, finally, blanket the globe and extend under the earth and into space. Taken together, bundles and constellations form one gigantic nexus of practices and arrangements, what I dub the 'plenum of practice' (Schatzki, 2016d).

I wrote in the first section that some discourse analysts neglect the organisation of practices. They also do not analyse this wider plenum. For instance, Gee, in addition to writing that society is composed of Discourses, claims that interactions among Discourses determine both the workings of society and much of history

(2014: 128). This is not, however, an idea he develops. Similarly, Scollon envisions practices forming nexuses (networks of repeatedly linked practices), imagines a form of analysis – nexus analysis – that would systematically and ethnographically study intersecting cycles of discourse (e.g. Scollon and Scollon 2004: 29; see Nicolini's chapter in the present volume), and analyses communities of practice as objectified nexuses, that is, as nexuses of practice that participants treat in discourse as bounded communities (2001: 155–6, 170). Like Gee, however, Scollon does not develop these ideas further, although some of his remarks suggest that he is not so much disinterested in social ontology as an advocate of a Garfinkelian view of the local occasionality of the social. Chouliaraki and Fairclough (1999), by contrast, are very concerned to connect discourses and practices to major issues in social theory. As suggested, moreover, they appropriate Bourdieu's notion of a field and, like him, construe fields as the principal container of practices. This is not, however, a matter they elucidate.

These observations are not criticisms. Discourse analysis is interested in discourse, not in social ontology, and it is not obliged to fill out the ontologies that it affirms or toward which its concepts point. At the same time, issues exist about the role of sayings, texts and discourses in the plenum of practices. One such issue is how sayings, texts and discourses contribute to the construction of the plenum, that is, to linkages among bundles and constellations.

Elsewhere I have written that practices and arrangements are linked by relations such as causality (e.g. activities effecting and responding to objects and arrangements), constitution (activities and objects/arrangements being essential for each other), intentionality (e.g. the directedness of activities and mental conditions toward objects and arrangements), intelligibility (practices making the entities that compose arrangements meaningful) and prefiguration (the bearing of arrangements on future courses of action). The resulting bundles, too, are related to one another in various ways, including via common actions, common organisational elements or common material entities; chains of action; common motivating events; participants in one bundle being intentionally directed to other bundles; overlapping, orchestrated or mutually referring places and paths; orchestration of (i.e. mutual dependencies among) actions, material entities and organisational elements of different types in different practice-arrangement bundles; and physical connections and causality. It is out of relations such as these that practices link with arrangements and bundles form constellations.

Language, sayings and texts play a myriad of roles in these relations. Space considerations require that my discussion of these be schematic. Sayings can motivate people to intervene in or to respond to the world and also, as events, cause events that befall material arrangements; elements of practice organisation, like people's directedness toward entities, are articulated, that is, given (explicit) content in the terms of language; sayings and texts contribute to what makes sense to people to do and the intelligibility of things in the world; sayings and texts can indicate events and matters, to which people in different practices and bundles react; sayings and texts in one bundle can be about other bundles, the components of other bundles,

or events occurring in them; and sayings can be links in chains of action as well as contribute to orchestration. Clearly, language, sayings and texts greatly contribute to the relations by virtue of which the plenum of practices is a mass of interconnected practices and arrangements. Collectively, they are a key component and means of connections among bundles.

Sayings and texts contribute to this interconnectedness in other ways. Texts, for instance, travel among bundles, among other things, disseminating ideas, topics, motivations, self-understandings and focuses of attention, establishing intentional directness among bundles and leading to individual or joint actions. An example is what Reckwitz (2006: 67) calls 'interdiscourses'. Interdiscourses are discourses, carrying descriptive-normative representations of subjects, that cross different social fields and diverse practices, thereby establishing representations of subjects as unified wholes. Closely related to the circulation of texts is what Fairclough calls 'intertextual chains', which are 'series of types of texts which are transformationally related to each other in the sense that each member of the series is transformed into one or more of the others in regular and predictable ways' (1992: 130). An example is a speech becoming a press release becoming both a webpage and an item on the evening news and in these forms peregrinating among bundles. Just like texts, sayings, too, both the activity and what is said, can give people ideas, shape their motivations, direct them to particular events and phenomena, lead them to respond and contribute to what actions they subsequently perform, individually, connectedly or collectively. In these ways, sayings contribute to the evolution of practices and bundles and, more broadly, to the course of history, in predominantly minuscule and occasionally large ways.

Sayings are also links, or parts of links, in what Bakhtin called 'chains of utterances'. Bakhtin (1986: 91) pointed out that utterances are full of echoes and reverberations of prior utterances. People quote others, appropriate words or phrases that they have heard or read, absorb ideas expressed in others' words and are motivated or oriented by what others have said. What's more, their utterances, explicitly or sotto voce, refute, affirm, supplement, rely on, presuppose and take what others say into account. Utterances also anticipate possible responses and, of course, become part of the stock of utterances to which subsequent utterances 'respond' in the ways just mentioned. By virtue of all this, utterances form chains, each link in which responds to previous links. These chains circle within bundles and constellations and pass through and between different bundles and constellations, thereby connecting them into larger bundles and constellations.

Finally, types of sayings such as explaining or describing peregrinate through the practice plenum, appearing in large numbers of practices and bundles and thereby forming a commonality among them. Reckwitz (2006: 66) develops this idea by noting complex dispersed practices that show up in various domains of modern life, that is, as parts of different constellations in the overall modern plenum. An example he mentions is the practice of experimentation, which marks postmodern culture in different domains of contemporary Western life.

Practices, bundles and constellations are replete with sayings and texts. These sayings and texts link up with other elements of bundles and constellations, connect

bundles and constellations while also threading through them and make a considerable difference to the evolution of the latter. They are also thoroughly imbricated with the non-linguistic doings that are just as widely distributed through bundles and constellations. This distribution of sayings and texts gives concrete sense to the idea that discourse pervades the plenum of practice. Discourses as language-in-use and as strings of words and sentences are everywhere: connected, connecting, spreading, moving through and always making a difference.

An additional form of language-mediated connectivity in the practice plenum is captured in Kristeva's (1986) concept of intertextuality (foreshadowed by Bakhtin's notion of utterance chains). Intertextuality comprises the ways texts – written and spoken – are content-fully interconnected through links and explicit as well as implicit cross-references among their concepts, themes, topics, statements, stances, claims and orientations toward subject matters (see, e.g. Fairclough, 1992: chap. 3). As a significative net through which texts in different bundles connect, it links bundles and is an important dimension of the plenum of practices.

This section makes a start on investigating whether the practice plenum exhibits large discursive or linguistic phenomena. Given the variety of such phenomena that theorists have ascribed to social life, my discussion can only broach the topic. I just affirmed the existence of one sort of large discursive phenomenon: the intertextuality that pervades the practice plenum. The present section will consider two other putative sorts.

The first is Foucault's notion of discourse, which exemplifies the idea of constellations made up of sayings alone (see also Lyotard, 1988). Before turning to genealogy, Foucault (e.g. 1976) theorised discourses as discursive practices – arrays of statement-making sayings (Dreyfus and Rabinow, 1982: 44–8) – that are organised by multiple sets of rules pertaining to the objects, concepts, theories and discursive infrastructures (e.g. presuppositions) that characterise or are bound up with these arrays. The discourses that Foucault was particularly interested in were those of the human sciences (e.g. linguistics, natural history, demography, medicine, psychology). Discourses, accordingly, are different ways of structuring areas of knowledge (cf. Fairclough, 1992: 3).

Foucault's discourses illustrate what William James (1979) called 'vicious abstractionism'. Discourses are specious substantialising abstractions (extractions) from the plenum of practices. The statement-making sayings that compose Foucault's discourses always happen amid doings and material set-ups (offices, libraries, laboratories) that help compose bundles of which the sayings are likewise part. On a practice theoretical approach to sayings, it is crucial that the sayings involved be resolutely treated as components of these bundles and that these bundles be taken as the entities in which social phenomena consist. Foucault, by contrast, took these sayings, together with what is said in them, as composing distinct entities, which can be analysed and explained. He did not deny that sayings occur amid doings and

material set-ups but failed to treat sayings essentially as parts of entities (practices) that also comprise the latter. Of course, it is true, as the notion of intertextuality indicates, that sayings are significantly connected to one another. They are significantly connected, however, as activities that are components of practices. It is not surprising that Foucault's analyses of discourses gave way to analyses of apparatuses (dispositifs; cf. 1980) in his subsequent genealogical studies. A discourse is composed of statement-making sayings alone, whereas an apparatus embraces sayings, non-discursive behaviours, and architectures.

A better way of construing discourses is exemplified in Hajer's (1995) notion of a discourse coalition, which is a group of individuals and organisations who work under the aegis of a particular set of ways of talking and thinking. The environmental discourse coalition, for example, coalesces around such concepts as sustainability, clean energy, saving the Earth, conservation and concern for the future, which it uses in particular ways in speech and writing. The discourse involved can thus be understood as a set of concepts together with their spoken and written (and thought) use in certain constellations of practices and arrangements.

This notion of discourse is unproblematic. It treats discourses such as environmental discourse as patterns in the use of particular concepts in certain interconnected bundles of practices and arrangements, for example, those of environmental activism, academic research, media reporting and politics. There is no attempt to isolate these concepts and uses from the acts and bundles concerned and to treat them as distinct discourse formations.

The second putative sort of large discursive phenomenon to be examined in this section is Fairclough's orders of discourse. Like intertextuality, discourse orders exemplify the idea of abstract discursivity.² As discussed, orders of discourse are composed of images, types of language and forms of non-verbal communication that are associated with particular social fields. (An earlier (1992: 124–30) version of the notion held that orders of discourse embrace genres, activity types, styles and discourses.) Such orders constitute potentials that acts of speaking and writing etc. can draw on. They are abstract structures because, as wholes, they are never encountered in experience, although elements of them are encountered when used in acts of communication. According to Chouliaraki and Fairclough, orders of discourse 'order' and 'regulate' communicative acts and interactions (principally, speaking and writing) in two ways: they structure, that is, enable and constrain these acts and interactions (1999: e.g. 63) and use of their elements is a means through which relations of power as forms of control are realised (1999: e.g. 144–5). I will set power and control aside because an adequate discussion of them is too ramified for the present context.

Chouliaraki and Fairclough claim that orders of discourse are associated with fields à la Bourdieu. As features of large phenomena of this alleged type, orders of discourse themselves qualify as a kind of large discursive formation. On my account, accordingly, they are features of constellations of practices and arrangements. To simplify matters, I will henceforth bracket images and non-verbal forms of communication and focus on types of language alone.

Types of language encompass various elements, which for simplicity's sake I will summarise as words, ways of using words and ways of constructing discourses qua strings of words and sentences. Examples of types of language are eyewitness accounts, storytelling and literary narrative (1999: 56–7). Types of language – that is, combinations of words, ways of using words and ways of constructing discourses – presumably enable communication in providing both the terms and the techniques for putting together spoken and written texts. I doubt, however, that specific sets of language types, as opposed to language types at large, enable acts of speaking and writing in particular bundles and constellations (e.g. those of education), including any that constitute fields (e.g. the educational field). In the first place, types of language are mobile and circulate among bundles and constellations, both through lines of communication and by virtue of people participating in different bundles. This especially holds of language types that employ non-technical words, but it also holds of some types that utilise technical ones (e.g. 'vocational learning', 'active learning techniques'). The elements of types, moreover, can circulate within and among bundles independently of types. Non-technical words, for instance, are remarkably mobile, as are some technical ones (e.g. 'inflation' and 'neurotic'). In addition, both within and across bundles, words, ways of using them and ways of constructing discourses can (re)combine and form a multitude of types. And the distribution of, especially, natural language types and their elements is contingent and often haphazard; sayings anytime can, and occasionally do, import types or elements thereof into bundles or constellations where they have not hitherto appeared. Consequently, it is not distinct sets of language types that enable speaking, writing and communicative interaction in particular constellations such as those of education. Rather, language types in general achieve this.³ Finally, it is true that bundles and constellations sometimes exhibit 'typical' language types (cf. Bakhtin's speech genres). Typicalities, however, enable nothing: they are simply patterns in language use laid down in past usage.

Now, nothing can be spoken or written that is not spoken or written in words, in a certain way, and as part of some discourse. When, consequently, a person speaks or writes, s/he must employ some type of language, either an existing one or a revised (or even new) one for whose development the existing spectrum of types was the starting point. Either way, she is constrained by the current spectrum. Again, however, the mobility of words (especially those of natural language) and of ways of using words or constructing discourses suggests that language types in general, and not distinct collections thereof, constrain speaking and writing in particular bundles and constellations. Likewise again, the repetition (typicality) of certain language types in particular bundles and constellations constrains nothing, in this case not just because typicalities are just patterns, but also because some sayings in many, if not most, bundles and constellations fail to use typical words or to exemplify typical ways of using words or constructing discourses – simply too many things go on in any bundle. This is even true in highly regulated environments such as missile control rooms and courts of law. In sum, specific sets of language types do not order particular constellations – e.g. fields, if they exist – by enabling and constraining acts of communication in them.

Another, possibly more promising way of construing how collections of language types (along with images and non-verbal forms of communication) order and regulate bundles is to ascribe them normative force. Normativity is a contested notion, but under one interpretation for types of language to carry normative force is for combinations of words, ways of using them, and ways of constructing discourses to be acceptable for or enjoined (prescribed, required, expected) of people. On this interpretation, such combinations enjoy this status by virtue of people knowing about them, unreflectively using or exemplifying them and sanctioning activity that conforms or fails to conform. This is how teleoaffective structures possess normative force on my account. I do not know whether Fairclough accepts that discourse orders regulate bundles normatively or, if he does, whether he affirms that their normative regulation can be construed thus (see, however, Fairclough, 2015: 68). Regardless, when normativity is so understood, some bundles and constellations are governed by distinctive sets of (normative) language types. It is true that the mobility of language types as well as the mobility and re-combinability of their elements, together with the fact that most of what people say is acceptable in the bundles in which they say it, indicate (1) that broad ranges of language types are acceptable in most bundles (this has only become truer in the modern world) and (2) that an indefinitely large common range of types is acceptable in the bundles and constellations where a given dialect of some natural language predominates. At the same time, bundles and constellations vary in the language types acceptable or enjoined in them in so far as, for example, they utilise technical vocabulary or slang (as in professional or subcultural bundles), embrace official procedures (as in governmental or religious bundles), exhibit command structures (as in military bundles) or are associated with particular localisable ethnic groups (cf. Gumperz's 1982 communication styles). These bundles are all contexts in which peer pressure, control by rule or command or the situations in which people act require the use of particular types. Thus, although the bundles in which a given dialect of some natural language holds sway more or less share an indefinitely large set of normative discourse orders, some of these bundles possess additional orders of technical, slang, official, ritual or ethnic discourse alongside this large set. These latter bundles – but these alone – carry distinct semantic spaces (Taylor, 1985a) that are tied, inter alia, to the languages used in them. Note, however, that these additional distinct orders that govern certain less extensive constellations do not thoroughly regulate speaking and writing in them. This contrasts with the indefinitely complex orders of natural language that comprehensively, although weakly, regulate speaking and writing in bundles dominated by particular dialects of natural language.

In sum, language types and discourse orders, like sayings and texts, travel and are distributed throughout bundles and constellations. This holds of many technical or specialised types and orders, too. Types and orders also (1) govern the linguistic articulation of the intelligibility of human activity and the world in which it proceeds and (2) make intertextuality possible. Meaning, however, suffuses bundles. Although

sayings and texts are weakly regulated semantically wherever a given natural language dialect dominates, only in certain bundles and constellations are sayings and texts semantically regulated – sometimes strongly, but always incompletely – by distinct language types and discourse orders. So distinct orders do exist in the practice plenum but lack the comprehensive structuring significance that Chouliaraki and Fairclough attribute to them. Grammar and syntax are different matters.

To conclude, discursivity pervades the plenum of practices in two ways. First, sayings and texts, and thus language and concepts, exist throughout and constantly circulate through the plenum. They thereby connect and thread through bundles and constellations. Certain bundles and constellations are also incompletely ordered normatively by distinct sets of language types and discourse orders. Second, sayings and texts give human life linguistic conceptual content. Through such content, sayings and texts are intertextually linked and the practice plenum is suffused by articulated significance. Language is an immensely important and in some sense omnipresent part of social existence. Its structuring significance for social life lies partly in contributing to the hanging together of bundles and constellations and partly in providing an articulatory and intertextual potential for social life generally and overall.⁴

The majority of the (new) approaches that are currently discussed under the heading of the practice turn have evolved in very close connection with empirical

research: this is true of ethnomethodology (Garfinkel, 1967), of Goffman's (1967) naturalistic observations and analyses of interactions and not least of Bourdieu's ethnography and his praxeology of Kabyle (1977) and later French society (1984). The practice turn, in other words, is at the same time a turn that develops theories based on empirical research. Notwithstanding this fact, especially in the German debate, practice theory approaches are seen mainly as new developments in theory formation. This trivialises the critical and reflexive arguments that such approaches articulate against the conventional understanding of theory, in particular, against what one might call with Bourdieu (2000) the 'scholastic' understanding of theory.²

'Scholastic' theories tend to universalise the social experiences of theoreticians and intellectuals and to impute these experiences to the members and practitioners that they study. This creates a picture (or rather: caricature) of a social world that consists entirely of talking heads without bodies, which are exclusively engaged in mental activities, i.e. in thinking, arguing and communicating with each other. Thus, according to Bourdieu, 'scholastic' understandings of theory are hallmarked by a predisposition to mentalism.

'Scholastic' theories also tend to confuse the theoretical models of reality that they have constructed with the foundations of that reality. This categorical 'scholastic' fallacy is found especially in social theories that are marked by realistic and substantialist understandings of underlying social structures, systems, rules, norms or other analytical concepts. The normativist functionalism of Talcott Parsons may serve as an instructive example here. Parsons substantialises norms and values and depicts them as discrete and independent entities juxtaposed with social action.³

By critiquing such 'scholastic' perspectives and confusions, praxeological approaches focus squarely on analysing the relationships between practices of researching and theorising on the one hand, and the practices studied on the other; that is the methodological twist that they give to 'questions of social theory'. Theory is not explained so much by reference to relations between socially situated theoreticians and their objects. Rather, theory is conceptualised as the outcome and product of empirically accessible and more or less self-contained theoretical practices and their relations to the practices they study and theorise, i.e. whose practical logic they at the same time discern and necessarily fall short of. Accordingly, the praxeological programme aims at a specific, reflexive and revised relationship between theory and empirical research (Schmidt, forthcoming a).

Implementation of this programme takes place via a process of praxeologisation – an epistemic and methodological process that can create interesting tensions and interactions between empirical perspectives and theoretical tools of perception (Schmidt, 2016). Praxeologisation involves describing objects, phenomena, processes and connections of interest as effects and consequences of social practices, thereby interpreting them in a new way. This methodological understanding of praxeologisation contrasts praxeology with realist views of social practices. Views that claim that the social is really composed of practices (and not of social actions, social structures, networks of humans and non-humans, etc.) tend to incite unproductive ontological debates. Praxeologisation proposes to replace the definitional,

declarative, anthropological, ontological or normative assumptions that are inevitable in such debates with a methodological procedure. This procedure at once sets up a contrast with realism and secures a realness strength of the praxeological constructs (Schmidt, 2016).

The praxeological critique of 'scholastic' and inherently mentalistic approaches has nonetheless also brought about a certain skewedness or one-sidedness in practice theory: for the reverse side of the sensitivity shown by praxeological analyses to the material carriers, conveyors and elements of social practices is a neglect (typical of many empirical studies in practice sociology) of mental features that are bound up with practices. These phenomena include sense-making and reflexive, analytical and theoretical forms of knowledge, which I will elaborate on in more detail below.

Inasmuch as these aspects are explored at all, they are usually treated as irreducible capabilities that reside in a private mental interior of a subject and cause visible external behaviour. This assumption is in tension with praxeological epistemology, since it implies that 'the subject' and its subjectivity (and not practices or the situational interplay of practice components) are sources of reflexivity and sense, creativity and transformation.⁴ Praxeologising and empiricising (1) the mental, cognitive and reflexive components of practices, (2) reflexive practices (i.e. theorising, analysing, reflecting, etc.); and (3) theoretical modes of knowledge, amounts to treating them as overtly public and shared activities not unique to an individual subject. A desideratum in praxeology is making both the practical uses of theoretical knowledge and the production of analytical models – i.e. theoretical and analytical practices – the object of empirical-praxeographic research. This perspective aims to clarify precisely how reflexive, analytical and theoretical knowledge processes contribute to practices. In what follows, I will explain and develop this perspective using an instructive case: the new analysis and knowledge practices that have been adopted in professional football.⁵

As comparison of historical and current recordings of televised football matches shows, the cultural practice of football has greatly changed since the 1950s. The game is played more quickly today; players appear much more athletic and the game more intense; and the formations and patterns made by teams on the pitch are fundamentally different.

From a sociological and practice theoretical point of view, these changes derive from innovations in all the core elements of the bundle of practices that is professional football – in other words, changes in materials, in cultural meanings and in forms of knowledge and knowing-how (Shove et al., 2012). New materials (shoes, clothing, lawns, training and playing devices, medical technology, physiologically optimised bodies, but also – as we shall see – GPS-supported tracking systems and thermal imaging cameras, software, etc.) combine with new cultural meanings (such as the currently dominant image of the team as a flexible work group that is adapted to the contingencies of the market place and consists of team players who can be

freely substituted) and, not least, with new developments in forms of knowledge and know-how.

The bodily practice of game-playing is at the heart of professional football as well. What 22 players do during a match can hardly be described using conventional sociological conceptions of social action of a methodological individualist sort. Such concepts reduce game-playing to a mere aggregate result of the actions, choices and mentalities of individual players. Playing a match, however, does not involve planned reflexive action or rational choice, but rests on a practical embodied form of knowing-how. The players are not authors of individual moves that they mentally devise in advance.⁶ Instead, movements develop out of the continually changing match figuration created by all the players on the pitch.⁷ What dominates, in other words, is the physical practice of playing; analytical distance from events and the forms of knowledge and reflexivity associated with this are secondary.

Nevertheless, in the last 30 years, the cultural practice of football has developed a cognitive and epistemic dimension and its own analytical reflexivity: quantification, digital match analysis and associated forms of knowledge have fundamentally changed the practice. These innovations in practice-specific thought and analysis have had an impact not only on how the game is played, but also on how it is understood, interpreted and commented on. These new forms of knowledge have influenced the critical sense-making of coaches, journalists and audiences. They have led to football being examined, discovered and analysed anew.

The production of match analysis data in German professional football has increased rapidly since the 1990s in connection with the increasing importance of sports reporting. This can be seen in longer broadcast times, the development of special-interest programmes on private TV stations and Pay TV and greater coverage of sports in all national newspapers. Match data supplemental to audio-visual and print reporting increases the newsworthiness and accountability of football matches and enhances the marketing opportunities for professional football.

In the 2011–12 season, the German Football League commissioned a private service provider to produce match data for all matches of the first and second Bundesliga leagues.⁸

These data are made available to the clubs. The service provider is also allowed to market the data and sell them to online, TV and print media, as well as to sports betting providers.

Match data are mainly generated by a process known as scouting and writing: several scouts watch a match on television and in the stadium. They are linked by headset to a writer, who codes the match events verbally described by the scouts through mouse clicks or through briefly touching a touch screen containing the most important pre-defined event categories. Collaborative classification and coding thus takes place while the match is played. Because of the time pressure involved,

it requires practised, routine understanding and skill, rather than much cognitive weighing up.

This method analyses or translates the match into a sequence of countable events. It sets aside team play and team behaviour and mainly focuses on passes, tackles, shots on goal and dead ball situations, i.e. events that can be assigned to individual players in order to record their performance. This practice of gathering statistics thus yields a specific interpretation of, and distinctive perspective on, the match. What is remarkable here is that many decisive aspects and dimensions of the play cannot be captured. For instance, the way in which the defenders' coordinated legwork ensures that the ball cannot be passed to the opposing forward necessarily escapes this analytic practice: it cannot register that, owing to careful and anticipatory playing behaviour, something has not happened.

Event-orientated coding creates and works with classifications whose relation to the game-as-played is necessarily arbitrary – despite the pre-defined catalogue and despite the training of scouts and writers – because in football, a sport which is comparatively short on cleanly demarcated events, most of the incidents requiring coding are not clearly delimited from each other or classified within the game itself.⁹ Rather, they blend into each other in the ongoing process of playing. But although identifying and differentiating meaningful units is a difficult task, it must be carried out in real-time practice.

Club managers and competent fans in particular tend to be critical of the methods used in these scouting and writing processes and in the production of event data. Above all, they critically point to the 'subjectivity' of such processes (meaning their arbitrariness), relying as they do on the football competence of scouts and coders. Within this ethnotheory of subjectivity, tracking processes¹⁰ are considered 'more objective'. These processes are designed to reduce the contribution of human observers (and their 'subjectivity') and rely instead on the technical-visual registering of play and players' movements that turns them into data and quantifies them.

Tracking processes transpose the problem of identifying and interpreting meaningful units from match observation to the processing and analysis of data. These analytic processes, for example, automatically calculate the distances covered by players and their positions and distances from each other and they determine and visualise the top and average speeds of the teams and individual players. These data on movement and positioning must then be interpreted. In spite (or because) of the sometimes elaborate graphics – for instance, the widely used heat maps, which show the distribution of a given player's presence on the pitch throughout an entire match – new ways of reading the displays so that they make sense in match and player analysis are constantly being explored and discussed.

At the same time, the search continues for data-mining algorithms that are capable of automatically recognising patterns in the large volumes of data generated by tracking and that are therefore capable of extracting analytically relevant knowledge. The main purpose here is to produce new techniques for long-term monitoring of the athletic and tactical abilities of squads and individual players (Schlipping, 2014).

These considerations have given us an empirical-praxeological understanding of analytic activities and knowledge that distances itself from pervasive mentalistic concepts of reflexivity and the reflecting subject. Reflexivity can be described as practice-specific wholes of public and fundamentally observable practices of analysing, reflecting, theorising, etc. These are practical activities that express mental activities and to whose performers mental activities are ascribed.

We can even go a step further and claim that using mentalist concepts of the subject and ethnotheories of subjectivity in the conceptualisation of reflexive or analytic practices and reflexivity in practices only complicates efforts to grasp social practices and bundles of practices in their procedural logic and transformational dynamics. To clarify this, I would like to outline a few moves that praxeology makes to distance itself from such an understanding of the subject. These moves ultimately enable a transition from a subjectivist to a praxeological epistemology.¹⁷

According to Elisabeth Anscombe (1975), a student of Wittgenstein's, we must consider 'the subject' (which knows itself as 'I') as an illusion that is deeply rooted in the Indo-Germanic languages. This illusion becomes problematic from an epistemological perspective when, as so often happens, the subject is cited as a cause in sociological descriptions of sequences of events or processes (on this point, see also Elias, 1984). Praxeological epistemology emphasises that it is not necessary to posit acting subjects as causes, authors or conceptual points of departure in the attempt to describe and apprehend the meaningfulness, reflexivity and dynamics of change of social practices. This epistemological orientation is the result of various conceptual praxeological moves that remove the subject from the focus of analytical attention. This decentring is needed – so the consensus argues – because when 'the subject' occupies the centre of attention, it blocks analytical access to collective structures of meaning, shared implicit knowledge systems, public patterns of meaning and the logics of implementation of social processes.

Although there have been careful and detailed expositions of things within and as part of social practices (Reckwitz, 2002b; Schatzki, 2002), there is rather less analysis of the range and variety of material relationships involved or of precisely how material entities figure in what people do.

To date, the most significant difference is between discussions of material elements which are treated as being integral to the conduct of a practice (Shove et al., 2012) and material arrangements amidst which practices transpire (Schatzki, 2010a). For Shove, Pantzar and Watson, the material elements of car driving might reasonably encompass the road network, a system of petrol stations and the steering wheel itself. All are accorded the same material status. Meanwhile, Schatzki's concept of material arrangements, amidst which practices transpire, does not distinguish between things which are directly, routinely or only distantly and occasionally implicated in the conduct of a practice. This is not in itself a problem. In both cases broad brush representations of 'material' are sufficient and consistent with the similar but not equivalent ambitions of the authors involved.

However, this language of elements and arrangements is of limited value if we want to know how and why specific patterns of production and demand arise and are engendered by correspondingly specific conjunctions of practice. Warde's (2005) observation that things, including energy and other resources, are consumed in the course of practice provides the starting point for a more differentiated account. The statement that 'the enactment of any one practice (for example, cooking a meal or travelling to work) typically depends on the prior existence and availability of a range of energy sources (gas, electricity, oil), infrastructures (grids, pipes, roads) and devices (cookers, cars, bicycles)' suggests that objects can be classified, in advance, under one or another of these ready-made headings (Shove and Walker, 2014: 50). This is a rather literal account.

A more subtle approach, and one that I develop here, is to distinguish between the roles that materials play in the enactment of any one practice. This is in keeping with those who view objects not as isolated entities, but as always integrated within and always inseparable from more extensive assemblages (Appadurai, 1986; Ingold, 2007; Intronla, 2013; Shove, Walker and Brown, 2014). It is also consistent with Rinkinen et al's (2015) method of characterising 'object relations' in daily life. Rather than taking objects to have a fixed status, Rinkinen, Jalas and Shove adopt a relational approach, distinguishing between the various ways in which people describe and engage with the materials involved in keeping warm in winter. Although I work with a more bounded and also more pragmatic view of things, the shared proposition is that materials are defined, constituted and positioned with respect to each other through their role within specific practices.

This method makes it possible to show how things switch between roles and to recognise that things which have a background or infrastructural relation to certain practices may be more directly engaged in the conduct of others. These theoretical moves are important, but they do not prevent me from appropriating concepts and insights from disciplines and fields which define resources, artefacts and infrastructures in other ways and which focus on them for different reasons. As well as picking out useful points of connection, the next few paragraphs provide a reminder both of the complexity of the material world and the specialisation of academic research dealing with infrastructures, devices and artefacts or resources.

If power is understood at the most basic level as acting with effect, then practice theory can be understood as essentially being all about power. Indeed, all of the relations comprising the social are constituted and reproduced through the actions of humans (amidst the many non-human entities also involved in those actions). All sorts of human action have effect in this way, whether in how the repetitive timing of eating reproduces fundamental shared social rhythms (Southerton, 2009) or in how the consequences of day traders' routinised actions (Schatzki, 2010b) cumulatively reshape financial markets. In accounting for both social change and the reproduction of social stability as the result of human action, practice theory is inherently about power, if power is seen as capacity to act with effect.

Of course, emphasising the power inherent in the actions of individual humans is only part of the story. Practice theory is perhaps best understood for its emphasis on the shaping of human action by relations and phenomena external to the person performing any such action. This is so to the extent that practice theory is sometimes cast as denying human agency or problematising the possibility of social change. While such claims reflect a profound misrepresentation on both points, they highlight the extent to which practice theory is centrally about the shaping as well as possibility of action.

With a focus on the shaping of action, we move closer to the second basic understanding associated with seeing power as object, that is, an interpretation of power as the capacity to direct or influence the actions of others. Understanding that action is always an effect of diverse relationships implies the shaping of action from

elsewhere. This starting point for approaching the shaping of action remains rather one-sided, attending to the heterogeneous phenomena that share in how action is shaped, rather than how power is wielded to shape it. However, leaving aside the question of how to identify who or what wields influence, practice theory is replete with resources for understanding the shaping of (the possibilities for) human action. This is perhaps clearest in relation to the roles of rules.

An emphasis on the role of rules in the shaping of human action represents one of the most significant points of commonality between the range of scholars identified as key protagonists in the intellectual history of practice theory, reflecting shared roots in Wittgenstein's work. However, just what is encompassed by the concept of the rule varies. For Schatzki, rules are 'explicit formulations, principles, precepts and instructions that enjoin, direct or remonstrate people to perform specific actions' (Schatzki, 2010b: 79). Meanwhile for Giddens, what Schatzki refers to here as rules are formulated rules, 'codified interpretations of rules rather than rules themselves' (1984: 21). Rules – or more broadly the normativity of practices, however understood – are both the grounds for and limits upon the possibility of meaningful and practicable action by practitioners.

Among the ways in which the shaping of individual action is conceptualised, rules are easiest to grasp. Particularly in formalised or codified form, rules can look like means of exercising power in a conventional sense: after all, laws are prime examples of codified rules. Indeed, for Schatzki, rules 'are formulations interjected into social life for the purpose of orienting and determining the course of activity, typically by those with the authority to enforce them' (2010b: 79). However, rules as apparent means of power are situated amidst a great range of other ways in which action is constituted and influenced. In Reckwitz's (2002b) 'ideal type' practice theory, it is the conventionalised assembly of the diverse elements and their interconnections which constitute the pattern reproduced in the performance of a practice and in the action of individual practitioners. The routes through which power might be considered to be exercised are still more obscure in Shove, Pantzar and Watson's model of practices as composed by the relations between meanings, competences and materials, even if rules and other means of normativity run through accounts of how practitioners integrate these elements in moments of performance (2012). This model has provided the basis for attempts to reconceptualise possible targets for intervention (Shove et al., 2012: 152–63), but it has little to say about the means through which power operates.

At first sight, other theorists' work appears more amenable to developing analyses framed in terms of power relations. Within Giddens' theory of structuration (1984), practices are the medium through which recursive relations between moments of human action and social structures constitute one another. Giddens invokes a vocabulary of power that is absent in more recent articulations of practice theory. For example, starting from his analysis of action, he identifies the role of allocative resources (capabilities) and authoritative resources ('types of transformative capacity generating command over persons or actors' (Giddens, 1984: 33)) leading to consideration of the structural dimensions of social systems, in signification, domination and legitimisation.

While Giddens' work offered routes for articulating practices with the workings of power, Bourdieu (1984) provides the most compelling account of the systematic reproduction of unequal distributions in relation to practice, through the concepts of habitus, capital and field. The meanings of these concepts, their relations to each other and the relations of each and all of them to practice, are somewhat unfixed over Bourdieu's work. Moreover, the concepts – particularly that of habitus – cover aspects of what other theorists would consider part of practices, representing the socialised norms and tendencies of conduct guiding actions and dispositions, along with the ways in which social relations become embodied to persons in capacities, dispositions and ways of thinking. However, the concept of habitus enables an appreciation of social difference, which a focus on practices as the principal unit of analysis obscures, and it does so without resorting to individualism. While Bourdieu might be considered to hollow out the concept of practice and to omit relations which others take to be central to an understanding of practice, he draws out concepts which facilitate the conceptualisation of the production and reproduction of unequal distributions, including of those things which constitute the capacity to act. These differences, and the processes through which they come about and are maintained, constitute the grounds of systematic social differences, as reified into concepts of class, for example.

So, it is clear that practice theory can indeed be understood as being all about power. Practice theory demonstrably offers an understanding of how capacities to act with effect are constituted through its account of the relational, and profoundly social, grounds for action – understood as the performance of practice. However, it has not yet been shown to account for the ways in which some practices and practitioners are able deliberately to affect the conduct of practices and practitioners elsewhere. Yet, in enabling one to grasp the different phenomena and relations which shape and influence patterns of action, practice theory should be able to account for means of executing power which involve shaping or directing the action of 'others'. Practice theory must be in a position to cast distinctive light on, say, how inequality results from uneven distributions of the capacities to act, as explored by Walker (2013). However, it is harder to grasp how power is executed in the directing of another's action, in authority over others, or in the core of what it takes to understand and tackle the effects of power in the world.

This is unsurprising. The above discussion treads a line through various articulations of practice theory guided by a heuristic understanding of power as object. This understanding is in tension with the fundamental ontological commitments of practice theory. An account of action which shows it to be both enabled and shaped by a distributed and heterogeneous range of phenomena and relations has little or no space for recognising specific instruments of power which direct action. In its basic expression in action, power is rendered a relational, socially constituted effect. Yet, observable phenomena in the social world – powerful institutions, patterns of domination, the reproduction of social elites and of hegemonic ideologies – demand some means of understanding, if practice theory is indeed an account of the social. Developing such a position depends on looking for different ways of

understanding power, as effect rather than object. In attempting to grapple with power while understanding it as an effect, the next step is to turn to Foucault.

Turning to Foucault to help theorise power is not an unusual move, but it is somewhat ironic. He disavowed both the analysis of the phenomenon of power and elaboration of the foundations of such analysis as the goal of his work (Foucault, 1982: 777). Fully comprehended as effect rather than object, power escapes analysis. What can be analysed are power relations, which are always agonistic. Some people and institutions are systematically advantaged by their position amidst these power relations and can use those relations to pursue their own ends, which can include shifting their location amidst power relations further to enhance relative advantage. But ultimately, no one person or entity has control of those relations. To understand those relations, we need to 'trace them] down to their actual material functioning' (Dreyfus and Rabinow, 1982: 186).

Questions of how institutions such as states or markets structure fields of action across space and time, far beyond the immediate reach of practitioners (including the situations that look most like the exercise of power conventionally understood), can be approached through the framing of governmentality. Foucault's own working through of governmentality is as an analytic representation of specific historical processes. In his 1978 lectures (Foucault, 1991), the concept is developed while accounting for the shift in governing he identifies in sixteenth century Europe, from defining the purpose of rule to be the retention of territory to it being the governing of population. Governmentality is initially an account of this specific process, of

the ensemble formed by the institutions, procedures, analyses and reflections, the calculations and tactics that allow the exercise of this very specific albeit complex form of power, which has as its target population, as its principal form of knowledge political economy, and its essential technical means apparatuses of security.

(Foucault, 1991: 101)

This characterises the rise and spread of government as the purpose of the state: a process of governmentalisation (Foucault, 1991: 101). However, numerous scholars have further developed Foucault's underlying ideas under a burgeoning field of governmentality studies (Burchell, Gordon and Miller, 1991; Dean, 2009).

This vein of work has done much to unpick the means – the rationalities, techniques and apparatuses – through which conduct is conducted (Gordon, 1991). But how is the conduct of conduct effected? That is, what is distinctive (rather than the same) about the practices of governing or of corporate influence? What characterises those practices which have influence over the performance of other practices?

If Foucault enables us to consider 'the conduct of conduct' in ways that are consistent with practice theory's ontological commitments, there could be a route for taking practice theory's engagement with power beyond accounting for capacities to act and the distributed range of relations which converge in shaping those capacities. Can we also understand how certain practices are distinctively capable of orchestrating, disciplining and shaping practices conducted elsewhere?

'Basically power is ... a question of government' (Foucault, 1982: 789). The potential of Foucault's account of how conduct is conducted here may seem limited given that his focus is not on power but on governing. Clearly, power operates in, on and through practices in many ways other than through the actions of government as formally understood. However, for Foucault, the term government is not restricted to formal institutions of state, but is used in a more general sense to mean shaping the conduct of others, to 'structure the possible field of action of others' (1982: 790). All scales of social phenomena are governed, from the self to the national and beyond. Governing can be understood as those actions and means through which the conduct of other people is more or less deliberately conducted, throughout social situations. With governing so understood, an account of how it is practised promises to fill out an account of power using resources already identified within practice theory literature.

However, for present purposes it makes sense to focus upon the power relations that act over space and time with the involvement of identifiable formal institutions – such as those of the state – that are in a position of relative dominance. Foucault himself acknowledges the value of such institutions as an empirical focus for an analytics of power relations, recognising that they 'constitute a privileged point of observation' (1982: 791). Institutions are sometimes considered something of a stretch for practice theory thanks to the putative difficulty it has in dealing with social phenomena which can be understood as large scale. However, the flat ontology of practice theory does not mean denying the scale of institutions or other large social phenomena. Rather, it entails recognition that such scale is produced and reproduced through practices. As Schatzki explains,

all social phenomena... are slices or sets of features of the plenum of practices and arrangements, differing simply in the continuity, density and spatial-temporal spread and form of the practices, arrangements and relations that compose them. It follows that all social phenomena – large or small, fleeting or persistent, micro or macro – have the same basic ingredients and constitution.

(2015b)

The observation that social phenomena have the same basic ingredients and constitution in whatever realm of the social or whatever apparent scale of social

phenomena, means that the practices of ministerial offices, cabinet rooms and corporate board rooms mostly have the same characteristics as the practices of domestic life or leisure pursuits. They too are comprised of meanings, rules, competences, embodied knowledges, materials, spaces and more, brought together through largely routinised and mundane patterns of action. Increasingly, the lines of practice theory discussed above are being brought to bear upon institutional situations and into articulation with approaches which are well established in analyses of such settings.

Indeed, some lines of inquiry associated with practice theory, broadly defined, have been developed by authors who focus upon institutions and organisations. Much of this work is concerned with conceptualising learning and knowing as processes that are situated, ongoing and generally collective in character, even in work places that are thought to be highly rationalised. For example, questions about how someone becomes competent as a member of a profession or work place have been addressed with reference to concepts of shared engagement, enterprise, repertoires and histories of learning (Wenger, 1998). More broadly, the notion of a community of practice, associated with Lave and Wenger (1991), has been used to represent and influence the workings of more or less identifiable organisations and institutions such as schools or hospitals.

Orlikowski (2002) follows Lave (1988) and Suchman (1987) in understanding people comprising organisations as 'purposive and reflexive, continually and routinely monitoring the ongoing flow of action' (Orlikowski, 2002: 249). Whereas practice-based studies of organisations typically concentrate on particular individuals or spatially proximal work groups, Orlikowski writes about globalised processes of product development as necessarily collective arrangements distributed across geographically separated situations and moments of practices: in short, she focuses on 'organisational knowing' rather than individuals' knowledge.

She identifies practices that are part of belonging to and sustaining the group – of sharing identity and interaction, but also of the doing of the work, in spatially and temporally distributed locales, to common purpose. Orlikowski consequently develops an account of a corporate organisation's capacity to act as that is constituted through the widely distributed, ongoing and situated practices of people comprising the organisation. In accounts like these, institutions are shown to take form as distinctive social phenomena through shared, collective, predominantly tacit ways of shaping, enabling, disciplining and aligning a multitude of largely mundane practices.

While Foucault acknowledges the value of an empirical focus on institutions, he goes on to identify certain problems with such a method. These are problems which are typically associated with practice theory-informed studies of institutions, such as those discussed above. Key here is the risk that an analysis of the practices comprising institutions will focus upon practices which are essentially reproductive of that institution. While fundamental to understanding the phenomenon of an institution, reproductive practices – such as informal social interaction, bodily engagement with technologies and so on – often lack any direct relation to the ways in which institutions act external to themselves. That the practices comprising work, even in spaces

of state or corporate power, are of essentially the same emergent, relational character as practices in more obviously 'everyday' situations, is another important corrective to conventional rationalist accounts of the workings of state or economy. However, while an understanding of the ways in which institutions reproduce themselves is clearly pertinent to an understanding of how institutions operate, stressing the sameness of the practices involved fails to account for how power is done.

Foucault's means of avoiding the problem of focusing too exclusively on the reproductive functions of institutions is to approach them from the standpoint of power relations, rather than vice versa (1982: 791). Rather than being distracted by the mass of institutional activity that is common to social life in institutions and elsewhere, his method is to follow the technologies and apparatuses of governing. These are the mechanisms – administrative, institutional and physical – which enable the exercise of power. This approach has been used to good effect. For example, Dean (2009) examines the distinctive 'technologies of performance' characterising neo-liberal governance – the targets, audits and indicators through which the actions of agencies are shaped and policed. These complement parallel 'technologies of agency', which are the means through which responsibility is shifted from central government to increasingly fragmented and diverse agencies of governing.

However, in the context of this discussion, research that focuses on the technologies of governing is often limited in that it does not connect to an understanding of practices within institutions or show how these articulate with, constitute and operate relevant technologies and apparatuses. It therefore fails to characterise the practices which enable the conduct of conduct and the accumulation of the necessary resources to act in such ways. Governing over space, as is the case with institutions identified with the nation state or a multinational corporation, is only possible through the marshalling, coordination and harnessing of countless practices, which provide financial resources (e.g. through the multitude of practices that generate and gather taxes or profits), information (e.g. through census) or threat of force (e.g. through the armed forces and police).

This is because the embodied action at the core of all performances of practices can only be spatially and temporally immediate. As a result, the extension and amplification of action can only happen through intermediation. Such intermediation can rarely, if ever, be accomplished without depending on other practices as well as on technologies and more. Appreciating the ability of some practices to orchestrate and align others makes it possible to account for the appearance of institutional hierarchy and scale and for differential capacities to act, while retaining a flat ontology. Clearly, governing technologies must articulate with the practices of governing which rely upon them as means of influence and as means of shaping the conditions of possibility and thus the actions of others. It is this conjunction which is important and which helps specify what is distinctive about the practices of governing.

The sociology of translation provides further resources and means of developing connections between Foucauldian approaches to governmentality and practice theory-informed understandings of institutions. In recent years, a number of authors coming from different starting points, but often inspired by the work of

Bruno Latour, have sought to conceptualise practices and the properties of large organisations. Rose and Miller recognise the need, in investigating the problematics of government, to 'study the humble and mundane mechanisms by which authorities seek to instantiate government' (Rose and Miller, 1992: 183) from techniques of calculation and computation and devices like surveys and means of data presentation, to aspects of professions and details of buildings. While principally dealing with Foucauldian governmentality, Rose and Miller turn to Latour's account of power in pursuing this project. Latour (1987; 1984) sees power as an effect of the composition and alignment of heterogeneous relations, rather than seeing it as an explanation of an actor's successful composition of that network of relations. The power of a given social actor is an effect of its location in networks of relations through which that actor can shape the actions and calculations of others. In working with the sociology of translation as a means of interrogating modes of neoliberal governing, Rose and Miller (1992) focus on inscription devices as means of making stable, mobile, comparable and combinable vast ranges of data involved in governing; and the ways that modes of representation so achieved work in enabling centres of calculation – the nodes of networks which aggregate and re-represent the flows of inscriptions so produced, as a means of acting over distance.

In writing about action over distance, Nicolini (2012: 179) also introduces Latour's work, doing so as a means of overcoming the limits he identifies within practice theory – particularly the work of Schatzki. While recognising fundamental differences between Schatzki and Latour, Nicolini nevertheless sees the potential for linking the two approaches. Indeed, Schatzki (2015b) has himself engaged with Latour's work in the cause of better understanding the constitution of large social phenomena. While restating points of difference with actor-network theory (ANT) – whereas ANT sees the social as comprised only by associations, Schatzki sees it as comprised of practices and arrangements – Schatzki recognises commonalities in the flat ontologies underlying both approaches. Based upon such commonalities, he appropriates concepts from ANT to help illuminate how large social phenomena like governments, corporations or universities can emerge from the plenun of practices and arrangements.

The attraction of Latour's work and ANT more broadly, both for Rose and Miller (1992) coming from a governmentality tradition and in part for Schatzki (2015b) and Nicolini (2012), is ANT's capacity to account for large phenomena and action over distance without recourse to explanation at any level other than that defined by actions. Concepts like centres of calculation help move governmentality-oriented understandings of the technologies of neoliberal governance towards the practices comprising it. For Nicolini and Schatzki, such concepts offer ways of accounting for how practices have effect over time and space.

While Latour's ontology has no place for understanding practices as the basic stuff of the social, it is not difficult to see how practice-based accounts of the activity comprising institutions mesh with accounts of particular forms of association that appear to be crucial in enabling action in one locale to shape action over distance in another (or in many) locales. Inscriptions are outcomes of particular,

normalised practices – practices of inscription. Similarly, the forms of calculation that characterise centres of calculation depend upon routinised and standardised processes of data storage and manipulation, which are performed and reproduced through more or less institutionalised practices of filing, archiving, etc. As Schatzki (2015b) indicates, it is not difficult to recast sites of association and alignment (as seen within ANT) as 'bundles' of practices. In turn, and as Rose and Miller (1992) discuss, particular modes of inscription and calculation are the stuff of technologies of governing in a Foucauldian sense. Latour's work consequently promises a means of developing connections between a focus on practices of governing and an understanding of the technologies and apparatuses through which governing is enacted.

In their engagements with Latour, neither Nicolini nor Schatzki directly addresses power. Indeed, Schatzki quickly moves to reduce any sense that one site has determinative influence, given that 'the progression of social affairs is thoroughly contingent' (Schatzki, 2015b: 8). Accepting this, there is nevertheless the problem that some sites, some organisations and some people are clearly situated in systematically advantageous positions amidst the associations, arrangements and alignments comprising social life, such that they have distinctive capacity to act purposively in ways which shape action over distance and across locales of action. The challenge is to develop concepts and methods that can help grasp how arrangements and associations of practices and the heterogeneous flows they are bound with are produced through, and reproduce, systematic inequities in capacities to act, including to act in ways which shape others' capacities to act.

One possibility is to consider the ways in which Latourian sites, such as centres of calculation, relate to the properties of organisations and institutions of governing. Such institutions comprise the ordering and stability necessary for the complex orchestration of practice that provides both the means and purpose of governing. They do so through aligning and disciplining the performance of key practices through other practices such as objective setting, managing, disciplining and incentivising. But perhaps more distinctively, such institutions are characterised by the extent to which capacities to accomplish governing are solidified and sedimented into relatively durable properties of the institution. Means and functions of practice are delegated to technologies and more or less codified procedures. Buildings, information infrastructures, divisions of labour and hierarchical institutional relationships between people and more are technologies for effectively aggregating the means of and the means to power. As Rose and Miller put it: 'powers are stabilised in lasting networks only to the extent that the mechanisms of enrolment are materialised in various more or less persistent forms – machines, architecture, inscriptions, school curricula, books, obligations, techniques for documenting and calculating and so forth' (1992: 183–4). Such materialised features of institutionalisation are part of the means through which practices are ordered and aligned, enabling those institutions to have effect, however constrained by the inescapable contingency of social life. These features also underpin an institution's capacity to accumulate the means of extending the capacity to act, for example, in the form of money or information.

At the same time, those features also underlie the obduracy of the practices of governing.

In sum, the conduct of conduct happens through practices which, while made of the same stuff as other practices, have distinctive characteristics not least resulting from the ways in which they are aligned over time. Concepts from both governmentality and ANT help to draw those working with practice theory towards a recognition that not all practices are 'the same' and that only some enable the aggregation and alignment of the resources necessary to assemble, maintain and exert some degree of control via technologies of governing. Practice theory is well equipped to describe and specify practices which have this potential, and in so doing address the missing links in understanding the processes of governing as constituted, reproduced and enacted through practices all of the way through.

This chapter has not argued that practice theory provides the best or only way of understanding power. Much of power is performed through immediate interpersonal interaction, in the details of speech, bodily conduct and human interaction. Practice theory clearly can have things to say here, but power relations at this level are probably more amenable to analysis in terms of conversation analysis or through an ethno-methodological approach, with or without an underpinning in practice theory. The chapter has been more concerned with relating practice theoretical concepts to apparently powerful large scale phenomena, like corporations and governments. Practice theorists have repeatedly shown that such large phenomena are comprised through practices and the arrangements they produce and reproduce. But questions about the ability of a corporation to shape actions and accumulate resources or about the ways in which international tax laws shape trade are often better approached through other means. For better or worse, economic theory, or theories of political economy, could not do the work they do if they refused to reify power relations and if instead power relations were always analysed through the multiple practices from which they are an effect.

Like any other approach, practice theory is not going to be able to give an all-encompassing account of power. However, the discussion above has demonstrated that it has distinctive contributions to make as part of a range of related strategies that shed light on how power exists as an effect of collective activity and its consequences. Ultimately, power relations and their consequences only exist through connections between moments of the performance of practices. Showing how this works out with reference to the reality of large phenomena like companies, economies, states and ideologies calls for a wide range of conceptual tools. In identifying affinities between practice theory, Foucauldian governmentality and Latourian ANT, this chapter highlights what appear to be especially promising options.

But what does practice theory stand to gain in pursuing an understanding of power? To return to the imperatives noted at the start of this chapter, if practice theory can account for all aspects of the social, it should be able to account for power as

a pervasive aspect of the social. Second, if practice theory is to make a difference, it must be able to provide an account of power with which it is consistent. Change is likely to entail and come through changes in power relations and purposive change will involve engaging in and with existing dominant power relations. In addition, engaging with questions of power provides a means of developing and advancing practice theory.

In summary, practice theory has within it a largely unspoken account of power. It is unspoken because within practice theory, power is ubiquitous. A practice (as entity) shapes human action (as performance). While the practice as entity is only the effect of performances, any one performance is substantially shaped by the practice as entity. Human action is therefore always influenced from elsewhere: it is the effect of relationships which are arguably always power relations (relations shaping action and the capacity to act), however diffused and distanced. In turn, power relations are always and only the effect of the performance of practices, in concert with their arrangements. Further, power relations never result only from distinct, specifiable, moments of practice, but are effects of the ordering and the churn of innumerable moments of practices. This explains why practice theory does not tend to focus on power as a separate or distinct property of the social. As Barnes said: 'talk of practices is talk of powers' (2001: 28).

However, not much is gained by noticing that power is ubiquitous to practice. The more significant challenge is to understand power as integral to and an effect of distributed practices, while also accounting for distinct social phenomena which can be meaningfully understood as powerful. It is clearly relevant to point out that the practices taking place in a multinational's global HQ are shaped by embodied knowledge and tacit routine, just as much as the practices of a domestic kitchen or amateur sports club. But the further ambition is to be able to account for the qualities of the corporate HQ that make it distinct from those other sites of practice and for how these arise, while recognising that practices are made of the same basic stuff. The answers lie in understanding how practices are related to each other across different sites – hence the importance of conceptualising the connections and nature of relationships between practices (Shove et al., 2012; Nicolini, 2012; Schatzki, 2015b). Tangling with questions about connections between practices takes on a sharper edge when the problem is that of explaining how some actors and sites come to be loci of a disproportionate capacity for shaping action elsewhere.

Existing thinking in cognate fields is of value in responding to this challenge. In most respects, an account of power that is compatible with practice theory can be fully encompassed by Foucault's account of governing and of power relations. Concepts from his work, not least in relation to governmentality, provide a means of ordering an analysis of practices in relation to the doing of governing (and being governed) which is sensitive to what it is in the flow of practices that is pertinent in understanding particular power relations. However, Foucault's own work and the work comprising more recent governmentality studies does not develop an account of how practices and their performance relate to and are anchored in the action of technologies and apparatuses of governing. In

seeking to address this question, I turned to concepts from Latourian ANT that have already been used in describing how practices connect and act upon one another over distance (Nicolini, 2012). The processes of alignment and aggregation that characterise classic Latourian concepts like inscriptions, mobiles and centres of calculation come close to bridging between analyses of technologies and apparatuses of governing and accounts of practice. Working through these affinities and identifying tensions and gaps between practice theory and related approaches helps specify methods of revealing and showing how certain practices act upon others.

A practice theoretically compatible approach to power relations casts new light on the processes underpinning and effecting them. As outlined above, the challenge is in essence one of explaining exactly what flows between moments of performance, of revealing the dynamics within and between those flows and of showing how they are distinctively aggregated and aligned to serve distinct purposes. In taking up this challenge, practice theory can make a distinctive contribution to understanding the existence and operation of power in the social and can do so by focusing on how practices relate to and align with each other so as to enable and perpetuate the capacity to act and to act at a distance to shape conduct in other spaces and times.

Thanks to the book's editors and Torik Holmes for excellent advice and comments; to fellow contributors to this volume for discussion of an early draft at the Lancaster workshop. The ideas this chapter started from took the form of a manifesto which was productively taken apart at the Windermere workshop Demanding ideas in 2014.

Practice theory is increasingly prominent in social scientific analysis. Hence, its virtues and its defects are being subjected to greater scrutiny. As part of that exercise, we examine the conceptualisation of widespread cultural understandings in the social scientific use of practice theory. Schatzki's (1996; 2002) schema of practice components has proved generative, particularly as adapted by Warde (2005) and Shove et al. (2012). However, virtually no attention has been paid to Schatzki's (2002) concept of 'general understandings', which is one potentially relevant instrument to address this challenge.¹ In The site of the social, Schatzki (2002) introduces the category of general understandings into his schema of components of practice, an addition to his earlier tripartite model of 'practical understandings', 'rules' and 'teleoaffective structures' (1996). In this chapter, we seek to open up issues posed by general understandings as a category within a schema of generic practice components. The concept of general understandings promises to deal with broad cultural conceptions which transcend the boundaries between 'integrative practices' (Schatzki, 2002).

General understandings might include such things as concepts, values and categories. Specific candidates might include: collective concepts, such as nation, state, economy or organisation; membership categories, such as ethnicity or gender; fundamental, culturally structuring concepts, such as animal/human or private/public and diffuse but culturally significant understandings, such as notions of convenience, cosmopolitanism or authenticity. Of course, many of these kinds of general understandings are the objects of whole fields of social scientific inquiry. Furthermore, we claim no originality in conceiving such objects of analysis in terms of practice. The movement away from substantialist ontologies in social science has seen the widespread reformulation of fixed social entities and collective concepts, such as

nation or gender, as processes of enactment and (specific) categories of practice (e.g. Brubaker, 1996; Butler, 1993). We do not presume to add insight into the operations of specific kinds of general understandings (e.g. membership categories, values, etc.). This chapter, rather, aims to cast light on the theoretical issues posed by the concept of general understandings in the spirit of Schatzki's formulation.

The chapter proceeds by first examining Schatzki's use of the category of general understandings and its relation to 'teleoaffective structures' (section 2), 'practical understandings' and Schatzki's concept of 'practical intelligibility' (section 3). We explore how general understandings might serve cultural analysis or play a part in a model of culture-in-practice (section 4). In section 5, we identify three putative general features of the diffusion, persistence and actuation of general understandings through illustrative examples. First, they may have their origins either in discourse or in practices. Second, they may display intimately connected tacit and discursive elements. Third, they circulate between integrative practices through typical channels, processes and mechanisms. We conclude by summarising the core features and effects of general understandings and their potential explanatory power.

General understandings, in Schatzki's (2002) formulation, are common to many practices, condition the manner in which practices are carried out and are expressed in their performance. Schatzki gives the example of the Shaker view of labour as a sanctification of the earthly sphere, which conditioned how labouring practices were carried out, as well as being explicitly formulated in doings and sayings (2002: 86). The simplest way to approach general understandings, then, is as ideational elements common to multiple practices. That general understandings are formulated in both doings and sayings gestures towards a central feature of the category: that it sits across the boundary between the discursive and the non-discursive. The category thus immediately inveigles itself into two thorny problematics for practice theory: how to conceptualise the relation between practices and discourse (see Schatzki, this volume) and how to understand the tacit (see Collins, 2001; 2010; Rouse, 2001; 2007a).

Schemas which lack analytical differentiation between diffuse cultural understandings and practice-specific understandings risk obscuring important dynamics and processes (Welch and Warde, 2015). Pellandini-Simányi (2014) usefully suggests two different forms of change in ideational elements common to multiple practices, which we might think of as two ideal-typical processes.² One form of change is when an element, such as an ideal of masculinity, is affected by changes within one practice – for example a work practice – which in turn has knock-on effects on the ideal of masculinity in other practices, say, domestic cooking. A second form of change is when general understandings are connected as elements in a more or less coherent axiology, or 'cosmology' as Pellandini-Simányi (2014) has it. Schatzki (2010b) alludes to such configurations of general understandings in the context of his discussion of Eliade's work on ritual. Schatzki observes that cosmological general

understandings establish, for Eliade, 'basic features of religious man's being-in-the-world' and that this 'complex of general understandings informs the teleological organisation of religious man's life (2010b: 151).³ Changes in one element of such a complex of general understandings may trigger changes in other elements of that complex as well. The example Pellandini-Simányi (2014: 139) gives is that of the 'ethical vision of gentility', which contained (in our terms) the general understandings of masculinity and of luxury. As masculinity became increasingly defined in terms of rationality, previous notions of opulence were challenged and the ideal of luxury came to be redefined in rational terms, as a controlled expression of taste. A focus on general understandings thus foregrounds the issue of how to articulate analytically specific practices, say cooking, with wider configurations like domestic organisation.

Nicolini plausibly suggests general understandings constitute 'external understandings' of the overall project in which the practice is engaged (2012: 167). How should we think about that externality? Where, external to the focal practice, does the general understanding lie, or rather, come from? The external here hints at a supra-practice level which threatens to undo the 'high order ontological sameness' of Schatzki's flat ontology of the social (2011: 4). Thus Caldwell (2012) suggests that the notion of 'general understandings' potentially leads Schatzki back to the position from which his work aims to escape: 'the "social" conceived as "general understandings" may presuppose an object of enquiry that goes beyond the practices in which it is carried or enacted' (2012: 291). However, Caldwell evinces a widely held misunderstanding that Schatzki's general position amounts to a kind of 'Il n'y a rien en dehors du pratique'. Caldwell assumes either that general understandings presuppose a distinct ontological level (perhaps implied by the relative coherence that complexes of general understandings may assume) or that they are carried in discursive formations exhibiting forms of organisation other than that of integrative practices and therefore debarred from Schatzki's ontology. To take the latter possibility first, as Schatzki makes clear in his contribution to this volume, discursive formations exhibiting their own forms of organisation are entirely compatible with his ontology. Schatzki shows that various forms of discursive organisation intersect with integrative practices in the form of sayings and texts incorporated into and carried by them. It is an empirical matter how more or less integrated those discursive formations are with specific configurations of practices. As for the former, no distinct ontological level is necessarily implied by the category of general understandings. The 'external' or 'beyond' in which general understandings subsist simply is configurational, whether found specifically in discourse⁴ or in heterogeneous assemblages of practice and discourse. A number of ontologically compatible concepts of such assemblages are already common conceptual currency, such as Foucault's 'dispositif', Boltanski and Thévenot's (2006) 'orders of worth' or Hajer's (1995) 'discourse coalitions'.

Schatzki (2002), however, offers a new configurational concept to 'illustrate' general understandings, although he does not develop it, that of 'teleoaffective regime'. This refers to something beyond the teleoaffective structure characteristic of specific

practices. He gives three examples of teleoaffective regimes among the Shakers: a religious faith in salvation through Shaker existence and belief that the Shaker's lived order was the kingdom of God on earth, governing hierarchies through which Shaker life was administered and a commitment to communal property and living (2002: 28). Schatzki's (2002) fleeting deployment of teleoaffective regime in his account of the Shakers underscores how a focus on general understandings often foregrounds the configuration of sets of practices. General understandings are the property of integrative practices. But to articulate convincingly the role of these central beliefs in the Shakers' cultural life and social organisation, Schatzki has recourse to a configurational concept that expresses organisation beyond the integrative practices that subend it. It seems that teleoaffective regime here does the conceptual work at the level of the sociocultural group that teleoaffective structures cannot, given the latter 'are not equivalent to collectively willed ends and projects (e.g. the general will or the we-intentions of a group)', but are the property of individual practices (Schatzki, 2002: 81). Teleoaffective regimes by contrast enjoin common ends. They articulate teleology and affect across practices, conditioning the teleoaffective structures of the practices they govern and subsequently becoming instantiated in situated activity through their performance. However, each of the Shaker teleoaffective regimes – belief in the Shaker order as the kingdom of God on earth and commitment to the hierarchical and communistic principles – could itself be conceptualised in terms of a general understanding. Hence the relation between general understandings and teleoaffective regimes as conceptual categories in Schatzki's (2002) account is therefore somewhat obscure.⁵ Nevertheless, it might be profitable to specify further the category of teleoaffective regime even if the category of general understandings does much of the conceptual work.

General understandings, then, may inform the normative ordering of the teleoaffective structure of practices and may partake of axiologies which subend the teleoaffective structures of multiple individual practices. General understandings may invoke or adjudicate normative controversy in the proper pursuit of practice, both in the sense of justifications and conventions discursively deployed in situations of contention (Boltanski and Thévenot, 2006) and in the broader sense in which the use of categories is both descriptive and normative (Bowker and Star, 2000; Jayyusi, 2013).

We now address how general understandings relate, first, to the practice component 'practical understandings' and second, 'practical intelligibility' (Schatzki, 2002). The former allows us to distinguish clearly between the pre-reflexive or tacit aspects of general understandings and the tacit aspects of practical understandings. The latter informs the role of general understandings in situated performance.

Schatzki's (2002) category of 'practical understandings' refers to abilities germane to the practical procedures of practices, the sense of how to go on with an activity,

identify it as such-and-such activity or respond appropriately: 'a skill or capacity that underlies activity' (Schatzki, 2002: 79). The most common forms of practical understandings are: 'knowing how to X, knowing how to identify X-ings, and knowing how to prompt as well as respond to X-ings' (Schatzki, 2002: 77). For example: 'Shaker medicinal herb production practices were linked by an inter-dependent pool of practical understandings of grinding, macerating, drying, storing, mixing, labelling, feeding, and printing labels' (Schatzki, 2002: 78).

While practical understandings are components of practice, 'practical intelligibility', for Schatzki, is a property of individuals, not practices: 'it is always to an individual that a specific action makes sense' (2002: 75). 'Practical intelligibility determines what it is that a person does next in the flow of conduct' (Schatzki, 2010b: 114). Practical understanding executes what practical intelligibility selects to do. In Schatzki's account, to the extent that practices determine practical intelligibility, they do so by moulding 'mental conditions' of individuals, such as those formed by the learning of practices or the ends and projects that individuals pursue, as well as affectivity (2002: 75 fn. 220, 81). General understandings inform practical intelligibility and govern activity by conditioning practical intelligibility or the normative form that practical intelligibility can assume. For example, European tourists in, say, Mexico, may decide to forego eating at the local MacDonald's in favour of eating traditional Mexican food, because their general understanding of authenticity informs them that this accords with authentic cultural experience.

Practical understanding for Schatzki does not therefore mean a sort of general know-how lying behind most or all of human behaviour that exhibits sensitivity to context, such as Bourdieu's (1977) habitus or 'sens pratique' or Giddens's (1984) 'practical consciousness'. By contrast, practical understandings are the property of particular practices. For Bourdieu and Giddens, it is these general know-hows that determine what people do in specific contexts. For Schatzki (2002), however, practical understandings do not govern activity or determine what it makes sense for people to do: that aspect of activity instead is governed by practical intelligibility.

It is far less common for general understandings directly to shape practical understandings. Think, for example, of the relationship between the Shakers' general understanding of the sanctification of labour and the practical understandings of their herb production practices, such as grinding, macerating or drying. While the general understandings inform the teleoaffectivity of herb production practices, it leaves their practical procedures undisturbed. The Shakers are highly atypical of commonplace contemporary social conditions, in the extent to which their core religious understandings suffused and orientated all their everyday practices. But even in this situation, unusually strongly governed by general understandings, it was only those specifically religious practices, where their religious general understandings were intrinsic, as it were, to the practice, that general understandings directly shaped practical understandings. By contrast, there is contingency between their general understandings and herb production practices, even in the light of the teleological imperative of the belief in the sanctification of labour.

We have to signal some disquiet with this account of practical intelligibility, which potentially has voluntaristic and individualistic implications for an account of social action which sit uneasily with stronger pragmatic and practice theoretical understandings of human activity. This potential is the more so, we suggest, in the context of an account of general understandings, where the latter might be misconstrued as anterior drivers of the subject's activity (Whitford, 2002). Having explicated the place of general understandings within Schatzki's schema, we now turn to the functions of the category for sociological analysis more broadly.

General understandings – all the more so when we invoke notions of complexes of general understandings, such as cosmologies or teleoaffective regimes – raise the question of how to understand the relation between culture and action. As noted above, the example of the Shakers is highly atypical of contemporary sociocultural milieu and thus of the place of general understandings within them. The Shakers were an unusually homogeneous sociocultural group, among whom foundational general understandings invested all practices. Furthermore, the sociocultural group was coextensive with a complex of practices: an orthopraxy animated by an orthodoxy of general understandings.

Despite the simplicity of the case, the question remains of how the Shakers' general understandings related to situated activity; or, how does this resolve the problem of, as it is commonly phrased, culture in action (Swidler, 1986). Schatzki invokes the Shakers' belief in the sanctification of labour in the context of evidence for the Shakers' propensity to toil hard for long hours. No doubt individual Shakers' engagement in their labours was invested with a sense of higher purpose and this served a strong motivating function. However, the danger would be to fall back into an account of social action where general understandings serve as anterior motivation for action and where declarative cultural commitments provide explanation (Whitford, 2002; Lizardo, 2012).

Instead, general understandings may contribute to a model of culture in practice: an alternative to both a Parsonian account of culture as internalised propositional knowledge or a Geertzian account of culture as externalised sign system (Biernecki, 2000; Lizardo and Strand, 2010; Warde, 2015). Biernecki argues that cultural meaning is generated by the ties between 'an order of representations and an order of practice that connects representation to a context of social exchange' such that 'we can concentrate analysis on variation in those types of ties' (2000: 302). That is to say, exploration of culture in practice might profitably proceed by addressing empirical questions regarding the relationships between general understandings and configurations of practice. We offer examples of such accounts below.

Invoking representation here suggests that we should specify the relation between the discursive and the non-discursive aspects of general understandings. General understandings sit across the discursive and the non-discursive divide and may exhibit pre-reflexive or tacit aspects. There is no need to establish a phenomenological

boundary between the discursive and the tacit (Rouse, 2001; 2007a; Collins, 2001; 2010). As Biernacki argues, a pragmatics of the relation between representation and practice moves one away from 'a purely discursive notion of culture without... counterposing "corporeal" practice in a binary opposite that is inaccessible' (2000: 308). Furthermore, the notion of discursive practice holds that linguistic expression and understanding is integral to practical competence (Rouse, 2007a).

Sayer usefully articulates how values – a particular kind of general understanding – combine conceptual, pre-reflexive and affective components:

Values are 'sedimented' valuations of things (including persons, ideas, behaviours, practices, etc.) that have become attitudes or dispositions, which we come to regard as justified. They merge into emotional dispositions, and inform the evaluations we make of particular things, as part of our conceptual and affective apparatus.

(Sayer, 2012: 171)

Values are both pre-reflexive 'moral intuitions' (Haidt, 2007; cf. Vaisey, 2009), which inform the cognitive and affective dispositions through which individuals respond to their environment, and they are deployed in discursive practice in justification and contention (Boltanski and Thévenot, 2006). The example of values illustrates how general understandings may combine the tacit and the discursive, the linkage of which we explore further in the following section.

If we follow this line of reasoning, the category of general understandings does not, contra Caldwell (2012), suggest a model of the social that undermines the social ontology of practice. Rather, general understandings may inform a framework of culture-in-practice consistent with the social ontology of practice theories. The tight enmeshing of Shaker orthopraxy in teleoaffective regimes might be misleading when studying the functioning of general understandings in contemporary milieus with which sociological inquiry is commonly concerned. Let us therefore look to contrasting examples. Multiple existing examples of broadly practice theoretical accounts, even if not framed as such, address the emergence and entanglement of general understandings within contexts of practice. We will briefly examine three exemplary cases. In this section, we examine different examples of how general understandings are articulated with configurations of practice, or of culture in practice, in order to identify some general processes and mechanisms which might sensitise future inquiry.

Central questions for sociology include where general understandings come from and what social processes and mechanisms convey them between different domains or configurations of practice. We begin with two ideal-typical processes concerning translation from discourse to praxis, on the one hand, and from praxis to discourse on the other.

Taylor (2004) offers examples of the first process in the emergence of the three 'modern social imaginaries': the public sphere, the market economy and the sovereign people. These imaginaries are fundamental cultural constituents of Western modernity: widely shared, pre-reflexive, background understandings that make possible certain common practices. Modern social imaginaries are congruent with secular 'meta-topical spaces' present to their members as 'framework[s] that exists prior to and independent of their actions' (Taylor, 2002: 115). However, the contents of the modern social imaginaries do not originate at this non-discursive level. Rather, they began life, in Taylor's account, as theory that gradually 'infiltrates and transmutes' an existing social imaginary, whose horizon they breach (Taylor, 2004: 109). Taylor is alert to 'the spectre of idealism', however, and thus grounds these processes in the practices that carry those general understandings (imaginaries), which 'are the essential condition of the practice making the sense that it does to the participants' (2004: 31–2). As Taylor has it, 'if the understanding makes the practice possible, it is also true that it is the practice that largely carries the understanding' (2004: 25).

Taylor suggests two kinds of processes through which transference comes about: first, a 'theory may inspire a new kind of activity with new practices, and in this way form the imaginary of whatever groups adopt these practices' (2004: 109). An example is the theological innovation that the first Puritan churches formed around the notion of covenant, which came to influence the civil structures of the American colonies. Second, a change may come about with a reinterpretation of a practice. The novel concept of popular sovereignty in the American Revolution found a propitious home in the existing institutions of elected legislatures in the colonies, such that: 'older forms of legitimacy are colonized ... with new understandings of order, and then ... transformed ... without a clear break' (Taylor, 2004: 110).

This latter example draws our attention to how, as Brubaker puts it when discussing ethnonational identities, general understandings 'need ecological niches in which to survive and flourish' (Brubaker, 2002: 185). Also, as Biernacki (2000) notes, in the context of Communist ideology in Eastern Europe, the life world may prove barren ground for ideological innovations: while some aspects of state ideology, such as production as the creator of value, became 'natural givens among working populations', other ideological aspects, such as collective authority over the generation of wealth 'were never accepted as genuine or taken for granted' (Biernacki, 2000: 306). A belief, furthermore, 'may appear specious if it is not an implicit organising principle of practice' (Biernacki, 2000: 308).

In the second, reverse, process, general understandings arise unbidden from practice. Swidler (2001) draws on an analysis of the formation and unification of the gay community in San Francisco from the early 1970s onward, and asks how, in the space of just a few years, did the adoption of one of a proliferating number of 'identities' become a crucial feature of membership in the wider lesbian and gay community? Notably, the community's diversity was enacted in the Lesbian/Gay Freedom Day Parade 'with floats, contingents, and marchers representing a panoply of more

or less flamboyant identities' (Swidler, 2001: 91). From this condition, without any discursive coordination, a general understanding was formulated

that diverse identities did not split the community, but united it; that organisers should not aspire to create a single unified organisation to represent the community; that the discovery and public assertion of new identities was part of the community building project.

(Swidler, 2001: 91)

In the parade itself, 'groups apply to have a float included in the parade ... and the more diverse their identity displays the more successful, exciting, and newsworthy the parade is' (Swidler, 2001: 92), anchoring the general understanding that the 'community' was composed of multiple identity groups. Swidler (2001) argues that the general understanding was precipitated from the practices involved in setting up the parade.

Our second illustrative case concerns issues of national identity. Anderson (1991) famously saw nations as imagined communities. Imagined communities exemplify the mutual dependency of representation and practice in the successful establishment of a general understanding of national belonging, including their pre-reflexive and affective aspects. In Anderson's account, the discursive and symbolic aspects of nationality are not sufficient in themselves to anchor the understanding of national membership or citizen obligation. Rather, he finds the basis of perceived solidarity in the practice of reading the morning newspaper, a routine shared by anonymous millions within the national boundaries, which enacts the core principle of that community. However, such diffuse general understandings of the existence of a nation – or, indeed of the public sphere, economy, society or authenticity – extend beyond the immediate background understanding that makes sense of particular practices. In such vein, Taylor speaks not just of modern social imaginaries but of the social imaginary, the 'inarticulate understanding of our whole situation' (2004: 25). This 'background' does not have clear limits, but nor is it limitless and despite being largely unstructured, it is neither inaccessible nor in-articulable.

Billig's (1995) analysis of nation and nationalism gives some substance to a claim that the broad background transcends the boundary between the articulated-discursive and the pre-reflexive, tacit, diffused and habituated. Billig's concern is not with expressly nationalist mobilisation, but with the sentiments lying behind and within the political arrangements of well-established Western polities which require no regular appeals for positive commitment or support for activities promoting national consciousness. How can a mostly dormant idea, with which most members of the population are not engaged, be credibly and effectively mobilised on occasion for political purposes by Western governments? Billig's answer is that national identity is reproduced in a habitual and mundane fashion as a type of nationalism specific to the West: banal nationalism. Billig writes:

[T]he ideological habits which enable the established nations of the West to be reproduced... are not removed from everyday life... Daily, the nation is indicated, or 'flagged', in the lives of its citizenry. Nationalism, far from being an intermittent mood in established nations, is the endemic condition.

(Billig, 1995: 6)

It is an ideological form which grounds an effective call to action or sacrifice when circumstances require it. But it is not consciously, intentionally, mindfully present in the populations affected by it. Rather, it lies in the background, signalled by cues in the environment, in commentaries on sport and cultural performances, in newspaper columns and stories. It is present and perceptible, but beneath the threshold of conscious attention. As Billig puts it:

[N]ationhood provides a continual background for political discourses, for cultural products, and even for the structuring of newspapers. In so many little ways, the citizenry are reminded of their national place in a world of nations. However, this reminding is so familiar, so continual, that it is not consciously registered as reminding. The metonymic image of banal nationalism is not a flag which is being consciously waved with fervent passion; it is the flag hanging unnoticed on the public building.

(Billig, 1995: 8)

The un-waved flag, a symbolic object to which no attention is paid, is yet the means by which nationhood is 'constantly flagged'. Billig employs an analysis of the sports pages which 'day after day, invite "us", the readers, to support the national cause' (Billig, 1995: 11). The regular and routine use of 'we' and 'us' – in international sports coverage or news broadcasts is a form of banal appeal to a political identity which mostly exists at the preconsciously level, presuming an acceptance of allegiance and obligation upon which people neither reflect nor meditate. Nations, nationhood and national identity are sometimes topics attracting deliberate and discursive attention, as explanations and justifications of behaviour, general understandings expressed and contested in an explicit manner. However, they are probably much more often present and effective as tacit background to understandings and experience of other practices like spectatorship at sports events, listening to radio reports on issues of international relations or encountering people with a different coloured skin on the bus to work. Traces of them are all around in the environment. Their taken-for-granted ubiquity is the source of their strength. In Billig's terms, banal nationalism has 'the powers of an ideology which is so familiar that it hardly seems noticeable' (1995: 12). Thus, nation stands as an exemplary instance of a general understanding which operates across the divide between the discursive on the one hand, and the pre-reflexive, dispositional and affective on the other.

A third illustrative case of general understandings in circulation concerns the cultural understanding of authenticity. The concept of authenticity plays a prominent

role in modern Western cultural history and we have a number of authoritative accounts of its translations. Lindholm notes:

[T]here are two overlapping modes for characterising any entity as authentic: genealogical or historical (origin) and identity or correspondence (content). Authentic objects, persons and collectives are original, real, and pure; they are what they purport to be, their roots are known and verified, their essence and appearance are one.

(Lindholm, 2007: 2)

Such understandings of authenticity are carried in diverse practices. Tourists seek authentic cultural experiences. Food connoisseurs seek authentic dishes. Brand managers regard the authenticity of their brands as of foremost importance. The commonality among this diversity is that these disparate practices all in some way relate to identity, whether cultural, organisational or personal.

The concept of authenticity demonstrates how general understandings may both subend and precipitate from broad cultural shifts and become instantiated in diverse practices thus creating novel configurations. It is possible to follow various translations of the idea of authenticity: from Protestant religiosity to Romantic literary innovation, to a pre-reflexive, background condition of modern identity (Taylor, 1989); from a value of the Modernist artistic avant-garde to a commonplace of modern consumer culture (Trilling, 1972; Orvell, 1989; Holt, 2002) and from the ideal of the anti-bourgeois individual to a buzzword of contemporary work place 'employee engagement' (Berman, 1970; Honneth, 2004). Space does not allow a detailed exposition but instances of authenticity illustrate several social processes and mechanisms by means of which general understandings may be conveyed from domain to domain, from the extra-mundane to the mundane and between the discursive and pre-reflexive. Its progress poses the always empirical questions of to what extent, and in what ways, do the general understandings condition those practices in which they are taken up and in what ways are they conditioned by them?

One feature of this story involves mediation by new material modes of communication. For example, the development in the eighteenth century of printing technology and the emergence of 'print capitalism' saw secular literary genres develop, disseminating Romanticism to a new reading public and a novel cultural concern with 'the authenticity of the selves who wrote such works' (Sinanan and Milnes, 2010: 2). In addition to material affordances, the diffusion of general understandings requires space or need for novel cultural understandings, a 'demand for intelligibility' which exerts 'pressure for greater explanation' (Boltanski and Chiapello, 2005: 103). Cultural dislocation is one regular explanation given for the fertility of the idea of authenticity, for example in the association of authenticity with commodities in the context of the cultural dislocations of nineteenth century urbanisation and industrialisation (Orvell, 1989). Another push comes from specific groups of cultural innovators and intermediaries who are often the Träger

of general understandings and the translators of understandings across domains. Thus, as well as artistic avant-gardes, practitioners of commercial communications have played a key role in the translation of authenticity from, for example, 1960s counterculture to mainstream consumer and commercial culture (Frank, 1997; Welch, 2012).

To conclude our excursion into authenticity, the above cited genealogies demonstrate historical translations of authenticity from explicit discursive and symbolic articulation to pre-reflexive background understanding and back again. For example, the radical reflexivity of the modern self becomes ground for the Romantic cult of the artist; Modernism gives explicit voice to the notion of authenticity, which is diffused into wider culture by novel forms of media, finding receptive cultural niches at anomic moments of social and economic transformation.

As with Billig's 'banal nationalism', authenticity figures as diffuse background understanding, sedimented in cognitive and affective dispositions and activated and transmitted by diverse discursive, symbolic and material elements of popular and consumer culture. It became a pre-reflexive background understanding to the concept of identity, which itself came under intense scrutiny in the scholarship of the later twentieth century. It is central to commonplace modern axiology and informs the teleoaffective structures of diverse practices across multiple domains, from the religious to the artistic to the commercial; and from cultural production to technologies of the self (Foucault, 1988). Its ubiquitous taken-for-granted-ness enables its transposition into novel practical and discursive contexts, as for example with the prima facie unlikely instance of the identity of brands and organisations (Holt, 2002; Welch, 2012).

These examples have highlighted the point that general understandings may be implicated across diverse domains and illustrated the diverse processes and dynamics through which general understandings are translated, diffused and instantiated. A genealogy of general understandings will necessarily trace a contingent historical path where understandings are articulated within different institutional contexts.

We have drawn attention to how general understandings operate both discursively and pre-reflexively. Thus, for example, contemporary seekers of authentic experience do not have to resort to discursive articulations in order to recognise and embody authenticity. The same claim could be made for nationality. Also, general understandings may be articulated in the 'sayings' of various practices and in their discursive form link practices together in ways unachievable through non-discursive performances. They may partake of teleoaffective formations that reign over complexes of practice, sociocultural groups, professions, cultural domains or other slices of praxis and thus offer insight into larger configurational phenomena. The concept potentially therefore contributes to the practice theoretical analysis of larger scale social and cultural phenomena; it might stretch beyond local phenomena or situated

activity to which analysis practice theory is all too often relegated (see Nicolini, this volume; Schatzki, 2016b). It offers the analytical affordances of practice theory without debarring traditional sociological concerns with 'macro' phenomena, providing a route through which the characteristics of the 'macro' can be generated from a flat ontology of practice. Common to the illustrations above is the articulation of macro-sociological concerns (regarding modernity and nationalism, for example) with micro-sociological contexts, or put in other ways, of the level of discursive formation with that of situated activity or of the level of the institutional with that of affective, embodied dispositions.

We suggest that general understandings exhibit at least three functions. First, general understandings may have an integrating or organising function, as is demonstrated by Schatzki's (2002) analysis of the Shakers, for whom general understandings integrated practice into an overarching cultural formation or by Swidler's (1986) case of the organisation of the San Franciscan gay and lesbian community through diverse identities. Billig's (1995) 'banal nationalism' demonstrates a second possible function – justification, which sits across the pre-reflexive responses of affective disposition and discursive contention. Third, general understandings have a function of enabling intelligibility, as when general understandings animated the everyday practice of the Shakers or when the quest for authentic self-expression motivates and orientates artistic endeavours. General understandings inform practical intelligibility, which governs individual conduct (Schatzki, 2002), albeit always through the 'normative accountability' of practices (Rouse, 2007a: 529). They are among the shared presuppositions which ground the intelligibility of social practices (Rouse, 2007a: 517).

There are several things that a theory of practice might want to do with a notion of general understanding. First, general understandings permit the analysis of large scale phenomena. Second, the concept presents a way to explicate the role of culture in practice, accounting for how very general ideas are incorporated into practice; it captures how such understandings are transmitted, translated and appropriated by practices, how they inform and shape practices and in turn how they are themselves conditioned by practices. General understandings are experienced, articulated and negotiated in situated and embodied activity and thereby transpose the cultural to situated activity. They encompass the pre-reflexive and thus help us understand forms of transmission of practice without discursive articulation as well as the actuation of dispositions and orientation of action through environmental cues. Third, the concept might also show how adjacent and distant practices might borrow from and change one another. As general understandings inform multiple practices, they could also help us to answer questions about the role of cultural intermediaries in transferring meanings and understandings and about the effect of intermediation on practices. Lastly, general understandings mediate discursive formations and practices, linking a focal practice to the discourses which inform and condition it and in the process become practical categories of identification, justification and evaluation.