Organizational routines as embodied performatives: A communication as constitutive of organization perspective

Abstract

Inquiring into how routines unfold increases our understanding of organization. This article critiques current positionings of organizational routines as practices and offers an alternative framing based on routines as communicatively constituted performatives. Two central arguments are advanced. First, present constructions of routines as comprising structurationist interpretations of Latour’s ostensive and performative are challenged and an alternative is advanced that draws from an Austinian understanding of performative as constitutive of organization. Second, bodies are brought into routines research as they are conceptualized as embodied accomplishments, extending existing research that typically neglects the body. An alternative definition of organizational routines is offered that constructs them as citational patterns of embodied conversation and textual dialectics that performatively co-orient toward an object.

Inquiring into how routines unfold increases our understanding of organization. Organizational routines are forms of recognizable and repetitive activity that practitioners regularly engage in. Indeed, much of practice can be framed as routine in some way (Pentland and Feldman, 2005; Rerup and Feldman, 2011; Wieder, 1974; Wright, 2013). Research focused on the iterative processes of organizing has a long history (e.g. Stene, 1940) and has taken a ‘dynamic turn’ in recent years (e.g. Cohen and Bacdayan, 1994; D’Adderio, 2008; Feldman and Pentland, 2008; Pentland and Feldman, 2005; Zbaracki and Bergen, 2010). This research has culminated in what Feldman and Orlikowski (2011) assert is ‘a theory of routines as practices’ (p. 1245). This theory is largely based on a structurationist (Giddens, 1984) perspective of practice applied to routines. I argue that while this work has made considerable progress in our understanding of how repetitive activity unfolds in organizations, there are gaps that themselves appear routine in such work. Some of these are acknowledged and lead Dionysiou and Tsoukas (2013: 183, 184) and Bapuji et al. (2012) to call for research that addresses how patterns of action, routines, are constituted and reconstituted (p. 1599). The purpose of this article is to critique the extant literature on routines and offer a response to Dionysiou and Tsoukas, which argues for a communicative constitution of organization (e.g. Ashcraft et al., 2009) framing of routines.

Much of the recent routines literature characterizes them as possessing both ostensive and performative aspects (e.g. Brown and Lewis, 2011; D’Adderio, 2008, 2011; Feldman, 2000; Feldman and Orlikowski, 2011; Feldman and Pentland, 2003, 2008; Feldman and Rafaeli, 2002; Hales and Tidd, 2009; Labatut et al., 2012; Lazaric and Denis, 2005; Miller et al., 2012; Pentland, 2011; Pentland and Feldman, 2005; Pentland and Rueter, 1994; Turner and Rindova, 2012; Zbaracki and Bergen, 2010) and cites Latour (1986) as the source for this social theory underpinning (e.g. Feldman and Pentland, 2003, 2008; Howard-Grenville, 2005; Pentland and Feldman, 2007). This has produced significant insight into routines as organizational phenomena, but has been slow in developing knowledge of the patterns of action that constitute and sustain them (Dionysiou and Tsoukas, 2013). I argue that far from enabling a theory of routines as practices, a structurationist reading of Latour’s (1986) notion of ostensive and performative actually prevents researchers from examining how embodied actors produce and reproduce routine activity in organizational settings.

I offer an epistemology of routines based on an ontological assumption that organization is achieved through a conversation and text dialectic (Cooren et al., 2011; Kuhn, 2008; Taylor et al., 1996; Taylor and Robichaud, 2004). In this approach, text is not limited to documents but incorporates all materiality. Conversation and text co-orientate and this process of co-orientation, relationally integrating conversation and text, is said to constitute organization (Cooren et al., 2011; Taylor and Robichaud, 2004). Applied to routines, this constructs them as constituted and reconstituted by and through communication and I locate my discussion in the recent communication as constitutive of organization (CCO) research move. CCO stances adopt views of communication as generative and constitutive, rather than merely expressive, representative, or transmitive of social realities (Ashcraft et al., 2009; Craig, 1999). CCO positions assign communication a performative quality. Consistent with this ontological framing, a CCO-inspired definition of routines constructs them as citational patterns of embodied conversation and textual dialectics that performatively co-orient toward an object. Insight is generated through reflecting upon and describing how actors collectively embody communicative processes of conversation and text. To date, CCO research has remained largely silent about how routine organization is constituted (cf. Fauré and Bouzon, 2010; Wright, 2013), and I address this gap, thus extending our understanding of CCO conceptions of organization.

The article is structured as follows. Following this introduction, I review how routines are conceptualized with a specific focus on how they are said to comprise recursively related ostensive and performative aspects. Latour’s (1986) notions of ostensive and performative are then presented, and this leads into two vignettes drawn from current routines research to illustrate my argument. Section four is a discussion of structure and structuration, detailing how they are handled in the CCO literature. This helps highlight the limitations of structurationist understandings of organization. An examination of routines as communicative constitutions follows, which includes an illustration of how a CCO approach could extend current knowledge, enhancing our comprehension of how workplace routines are constituted. Research implications are then outlined. The final section offers brief concluding comments.

Organizational routines

Routines are conceptualized as emergent, effortful accomplishments or flows (Feldman, 2000) that are ‘recognizable pattern[s] of interdependent actions, involving multiple actors’ (Feldman and Pentland, 2003: 96). From this perspective, routines are not fixed patterns of action, but sets of possible and potential patterns, enabled and constrained by the organizational, temporal, social, physical, and cognitive structures from within which actors are said to ‘enact’ their performances (Pentland and Rueter, 1994; Rerup and Feldman, 2011). Routines can be distinguished from non-routines as they exhibit recognizable reiterative patterns of practice absent in non-routine activity. Those involved in routines may have substantially different understandings about what purposes they serve (Feldman and Rafaeli, 2002). Routines unfold within and across organizational and individual contexts (Howard-Grenville, 2005), are inherently improvisational (Feldman and Pentland, 2003) and adaptive, as actors adjust their behavior in response to the actions of others contributing to their accomplishment. In the extant literature, routines are seen to comprise ostensive and performative aspects, where the ostensive corresponds to the structural and the performative to the agential of Giddens’ (1984) structuration theory. In the remainder of this section, I analyze how these have been handled in the routines literature, beginning with the ostensive.

Ostensive aspect

The ostensive aspect is the abstracted idea, generalized pattern, or road map of the routine (Pentland and Feldman, 2007) and is usefully thought of as a narrative or script (Pentland and Feldman, 2005). It shapes perceptions about what the routine is and how the set of ‘performances’ it denotes is similar to others and therefore routine-like (Feldman and Pentland, 2003). For example, a job-hiring routine within an organization will have ‘endless variations on the appropriate way to go about hiring people for different kinds of jobs, in different departments, or at different times of the year’ (Pentland and Feldman, 2005: 796–97). Yet, each of these variations identifies the activities involved as a legitimate, recognizable, and authentic representation of a job-hiring process. However, Feldman and Pentland (2003) insert a note of caution that reminds those encountering the ostensive aspect that this description cannot possibly specify in sufficient detail what actually happens when a routine unfolds. What this means is that actors should be aware that discussions focused on the ostensive are conversations about an idealized idea, not what actually happens in practice. For this reason, Feldman and Pentland (2003) say it would be a mistake to take the ostensive aspect as the whole routine.

Additionally, they (Feldman and Pentland, 2003) claim it would be wrong to promote the view that there is only one ostensive aspect. There is no single unified entity that should be interpreted as the ostensive aspect; that is why there are ‘endless variations’ of a hiring routine. It is important to talk about ostensives in the plural as there are many ostensive perspectives (Feldman and Pentland, 2008) resulting in more than one description of the hiring process. Actors have their ostensive understandings of routines framed by their situated contexts. These understandings shape how they plan, guide, and account for their actions with respect to the routine (Pentland and Feldman, 2007), to retrospectively justify their agency (Feldman and Pentland, 2003), and as a basis for discussing what actions are possible. While it is noted that ostensive narratives are used to legitimate preferred courses of action (Feldman and Pentland, 2008), how this occurs is not elaborated upon. Criticisms of how the ostensive is conceptualized are offered by Iannacci and Hatzaras (2012), who argue that Feldman and Pentland’s exposition of the ostensive is a ‘flat’ description, collapsing any possible layering of ostensive aspects (p. 5). Similarly, they (Iannacci and Hatzaras, 2012: 5) state that there is a failure to account for any pre-existing social contexts that might impact upon routines, or elements of the natural and physical world that make a difference in how they progress.

Performative aspects

A focus on ‘practice’ is said to encourage a view of organizing that recognizes unfolding activity as involving general and specific understandings, interpretations of rules, and practical judgment (Feldman and Orlikowski, 2011; Orlikowski, 2000; Schatzki, 2005, 2006). Raelin (2007) sees an epistemology of practice as essential for exploring the tacit processes practitioners evoke as they accomplish their daily organizing (p. 499). From this perspective, actors are said to improvise, innovate, and adjust over time (Orlikowski, 1996), as they adapt their agency to their contexts. The notion of ‘context’ is discussed by Latour (2005), who suggests that when actors adapt their behavior to their contexts they are contextualizing [italics added], creating their contexts through their adaptive acts. The mantra that the performative aspect is limited to the performance of the routine, executed by specific people in specific locations and at specific times (Essén, 2008; Feldman and Orlikowski, 2011; Feldman and Pentland, 2003, 2008; Rerup and Feldman, 2011), can leave the impression that material artifacts are assigned a lesser status than their human co-actors, yet recent research has attempted to assign sociomateriality an agentic quality and theoretical advancements have been made in this area (e.g. D’Adderio, 2008, 2011; Labatut et al., 2012; Leonardi, 2011; Pentland et al., 2012).

What routines theorists call performances are carried out against a background of expectations and established norms of behavior that shape agency. The role and importance assigned to historical experience is still to be resolved; while previous iterations are acknowledged to influence current performances of routines, but not dictate them in a path-dependency sense, there are also those who see path-dependency as having a more dominant effect on how routines materialize (e.g. Becker, 2004; Becker et al., 2005; Leonardi, 2011; Pentland et al., 2012; Van Raak et al., 2007). Actors are said to have a repertoire of potential actions to draw from; the actual performance Feldman and Pentland (2003) assert, drawing from Giddens (1993) again, could always have been different. Therefore, understanding the choices not made as well as those made is essential to effectively interpret performance. Behavior though is not necessarily conscious, but inherently improvisational and intuitive (Feldman and Pentland, 2003). Ostensive descriptions do not adequately explain variability within routines (Birnholtz et al., 2007; Essén, 2008; Pentland and Feldman, 2005) as the level of analysis is too removed from the everyday actions of individuals to explain unplanned actions and deviations from ostensive scripts. However, detailed, in-depth research into the performative aspect of routines through which such assumptions can be examined is rare as most studies remain focused on the ostensive (cf. Dionysiou and Tsoukas, 2013; Hales and Tidd, 2009).

Those that do purport to investigate the performative/performances of routines (e.g. Birnholtz et al., 2007; Feldman, 2000; Pentland et al., 2010) do so from an analytic distance that means the nuances and subtleties of emerging routines are passed-over as these structuration-inspired researchers’ epistemological resources ill-afford them the capability to discern how ‘patterns are created and, since routines are repetitive, recreated’ (Dionysiou and Tsoukas, 2013: 183–84). This leads Pentland et al. (2010) to suggest that such research may not be realizable as ‘it is difficult or impossible to observe the underlying generative mechanisms of a routine’ (p. 919). Despite this methodological limitation, Feldman and Pentland (2003) seem convinced the recursive relationship between structure and agency is maintained so that ‘[p]erformances enact the ostensive aspect’ (p. 107). However, the nature of this enactment is yet to be more fully articulated. For example, Feldman and Orlikowski (2011: 1245) seem clear that ‘the development of the routine occurs through the enactment of it’ [italics in original], but do not elaborate how this manifests (cf. Bapuji et al., 2012; Dionysiou and Tsoukas, 2013). Birnholtz et al.’s (2007) study of the regeneration of Camp Poplar Grove does attempt to illustrate how a summer camp recreates itself every summer despite changes in personnel. However, as will be illustrated later, it is the routine’s abstract representation that is described rather than its practice.

In the next section, I review Latour’s idea of ostensive and performative and contrast this with how the routines literature has utilized it.

Latour’s ostensive and performative

Feldman, Pentland and colleagues use of ostensive and performative has been successful in energizing routines research, moving it away from the evolutionary economics-inspired understanding propounded by Nelson and Winter (1982). However, it is instructive to revisit Latour in its original translation, to appreciate how Feldman and Pentland’s structuration-inspired interpretation of ostensive and performative assigns meanings to these terms consistent with their structurationist ontology. It is important to remember that their understanding is but one of many possible, as within routines research it seems that structurationist framings of Latour’s ostensive and performative run the danger of becoming reified and unquestioned. What follows then, is a critique of structurationist interpretations, and through a discussion of ostensive and performative from a CCO stance an alternative reading is offered.

To the best of my knowledge, Latour (1986) only discussed ostensive and performative very briefly, covering less than 3 pages of a book chapter. He offered the following advice for conceptualizing the ostensive:

1. In principle it is possible to discover properties which are typical of life in society and could explain the social link and its evolution, though in practice they might be difficult to detect.

2. Social actors, whatever their size, are in the society defined above; even if they are active, as their name indicates, their activity is restricted since they are only parts of a larger society.

3. The actors in society are useful informants for those who seek the principles that hold society together (see 1), but since they are simply parts of society (see 2), actors are only informants and should not be relied upon too much because they never see the whole picture.

4. With the proper methodology, social scientists can sort out the actors’ opinions, beliefs, illusions and behaviour to discover the properties typical of life in society (see 1) and piece together the whole picture. (Latour, 1986: 272)

He also discussed how researchers may approach the performative:

1. It is impossible in principle to define the list of properties that would be typical of life in society although in practice it is possible to do so.

2. Actors, whatever their size, define in practice what society is, what it is made of, what is the whole and what are the parts—both for themselves and for others.

3. No assumption is necessary about whether or not any actor knows more or less than any other actor. The ‘whole picture’ is what is at stake in the practical definitions made by actors.

4. Social scientists raise the same questions as any other actors (see 2) and find different practical ways of enforcing their definition of what society is about.

   (Latour, 1986: 273)

Further, Latour (1986) criticizes social science attempts to understand society through ostensive definitions:

Society is not the referent of an ostensive definition discovered by social scientists despite the ignorance of their informants. Rather it is performed through everyone’s efforts to define it. (p. 273)

And, calls for a move that privileges the performative over the ostensive:

The result of such a continuous definition and redefinition of what collective action is about is to transform society from something that exists and is in principle knowable into something which is built equally, so to speak, by every actor and that is in principle unknowable—it involves shifting from an ostensive to a performative definition. [italics added] (pp. 276–77)

Latour, then, does not consider the ostensive and performative aspects of equal analytic importance, as some form of duality as structurationists following Giddens might. Rather, he advises researchers to locate their level of analysis at the performative. He rejects any idea that what actors experience can be discerned from an analytical focus on the ostensive, the ‘macro’ in recent literatures. In the extant routines literature, Latour’s ostensive and performative are framed as dualities, as two sides of the same coin, which is clearly not his intention. In their review of actor–network theory, McLean and Hassard (2004) specifically reject any attempt to synthesize Giddens’ structure/agency duality with Latour’s views on macro/micro distinctions. Latour sees such divisions as artificial and unproductive. Boden (1994) makes a similar point when she says, ‘Society does not happen at different levels, research does’, nor does individual action build ‘toward some larger entity that “is” organization in some cumulative sense’ (p. 201 in Fairhurst and Putnam, 2004: 20). Latour’s call to focus on the performative (micro) McLean and Hassard (2004) claim, results from his conviction that the ostensive (macro) ‘is made up of the same basic connections’ as the performative and should be studied in the same way (p. 508). This lies at the heart of Latour’s thinking about ostensive/performative distinctions. They are not two sides of the same coin recursively related, but the coin looked at from two different abstracted levels of perception.

Latour’s (1986) call for researchers to shift their focus from the ostensive to the performative is clear and develops from a belief that this is necessary to better understand how organization unfolds. Studies that persist in constructing and maintaining the ostensive/performative duality drawn from structuration theory construct humans ‘enacting’ routine ‘performances’ (e.g. Feldman and Orlikowski, 2011; Howard-Grenville, 2005; Pentland and Feldman, 2005; Pentland and Rueter, 1994; Rerup and Feldman, 2011), as the conviction holds that the ostensive(s) is(are) enacted in actors’ performances. By describing such activity as an ‘enactment’ these authors construct the performative and the ostensive differently, reinforcing mind/body separations, perceiving the actors involved as disembodied objects. While Dale and Burrell (2008) convincingly argue for a definition of ‘enactment’ that denotes it as encompassing ‘the lived experience of social spaces’ (p. 109), they acknowledge difficulties exist with the term. Its association with the taking of theatrical roles means that in common parlance, and this is how I see routines scholars using the term, it more often refers to ‘[A]n actor [who] has … taken on the role they play consciously, intentionally and through choice’ (Dale and Burrell, 2008: 109). This assembles theatrical actors who ‘enact’ such parts as surface, superficial persona (Dale and Burrell, 2008). By constructing routines as embodied rather than enacted phenomena, where human bodies are seen as constitutive in their accomplishment, a move away from understandings dominated by the cognitive is achieved.

A CCO approach, consistent with Latourian views, constructs actors as ‘embodying’ routines, locating practice as accomplished by the body more so than the mind (Antonacopoulou, 2008; Cunliffe and Coupland, 2012; Merleau-Ponty, 1958). Latour acknowledges this and this is why he privileges the performative over the ostensive. It is only through this level of analysis that researchers can get closer to the bodily articulations of human and nonhuman actors, how they ‘define in practice what society is, what it is made of, what is the whole and what are the parts—both for themselves and for others’ (Latour, 1986: 273). Organizational routines are currently depicted as enacted by disembodied actors, whose bodies appear to have no significant influence on how they unfold. As will be illustrated later, a practice theory of routines needs to incorporate the embodied actor, as organizational practices are in large part bodily accomplishments.

In addition to the call to distinguish between enact and embody, McKinlay (2010) cautions against conflating performance with performative (Butler, 1993). McKinlay (2010) uses the term ‘performativity’ rather than performative, but his point is clear:

Performativ[e] is not to be confused with performance. Performance suggests an actor who consciously follows—or refuses to follow—a script … Performativ[e] is a [practice] concept that seeks to escape—or at least to reject—the dualism of structure and agency. (pp. 234–35)

‘Performance’ is consistent with the idea that the ostensive aspects of routines are scripts to be enacted. But, use of the term ‘performative’ in current routines research is problematic, as what is described are performances not performatives in any Austinian (1975), Butlerian (1993), or constitutive sense. Drawing from Austin and speech act theory, Butler (1993) identifies that it is the citationality of some utterances that gives them a performative quality, distinguishing them from other forms of communication that are not reiterative. The stance taken in this article is that organization is communicatively constituted, within this citational communication distinguishes routine practice from the non-routine, which is assumed to be constituted by non-citational communication. While Pentland et al. (2010) feel routines’ generative mechanisms are difficult or impossible to study, CCO’s origin in speech act theory offers the notion of citationality as a conceptual framing to do just that.

What follows are two vignettes drawn from existing research into the performances of routines. These are presented to illustrate the argument that what passes for practice research by routines scholars only addresses agency in the structuration sense and consequently sustains an analytical distance that prevents any in-depth study of practices, their performative constitution and reconstitution, resulting in the gaps identified by Bapuji et al. (2012) and Dionysiou and Tsoukas (2013) cited earlier.

Pentland et al. (2010) seek to compare how invoice processing is accomplished across four Norwegian organizations using the same software technology. They study invoice processing as networks of action and ‘analyze the patterns in terms of individual events, parts of events, and whole sequences of events’ (Pentland et al., 2010: 917). In their (Pentland et al., 2010) overview of data, they state:

The information processing actions were primarily data entry. Note that data can be entered at any time during invoice processing, even during the approval phase. For example, an approver may add a comment, or update amounts of currency information. Actions that involved transferring work, distributing work, or notification were categorized as ‘coordination’. Finally, actions that involved approving the invoice were categorized as decision-making. This process reduced a set of 31 codes down to just three theoretical codes. We use two different levels of abstraction because performances can always be made to look different with a sufficiently fine-grained coding scheme (Abell, 1987). Using the more abstract coding scheme minimizes the chance of finding differences that are an artefact of excessive detail. For our purposes, the particular choice of coding scheme is irrelevant. (pp. 923–24)

The ‘performance’ of the routine is reduced to a high-level process description in which only formal details are recorded. So, we are informed that the primary processing actions employed were data entry strokes. We are told that an approver may add comments, but none of these are shared with us. We are also left uninformed of the detail surrounding how approving the invoice, decision making, is accomplished, other than that it is. Pentland et al. (2010) actively work against incorporating any of the messiness experienced by practitioners into their work. They note that routines can always be made to look different with a sufficiently fine-grained analytical lens, and that ‘excessive detail’ is ignored-away as this does not serve their aim of theorizing on networks of action. Their focus on events also mitigates any relational understanding of routines, which are conceptualized as stand-alone incidents that happen, somehow, to result in invoices being processed. I argue that it is the very detail that Pentland et al. (2010) seem determined to avoid where the richness of routines can be found.

Birnholtz et al. (2007) focus their study on how a children’s summer camp, Camp Poplar Grove, is recreated as ‘the same’ organization every year. Their study is more practice-oriented than that of Pentland et al. (2010) as they pay attention to how a rural space is transformed every year into a recognizable place where Camp Poplar Grove hosts children for 2 weeks of outdoor fun and activity. Birnholtz et al. (2007) discuss how the staff employed for the summer must recreate Camp Poplar Grove so that it was both new and recognizable as being ‘the same old Poplar Grove’ (p. 323):

During the first days of training week, a range of physical set-up tasks were necessary in the activities areas. Tennis nets needed assembling, boats need launching, docks needed to be put in the water, and so forth. Whenever possible, groups completing these tasks had access to an experienced staff member who was not necessarily an expert in the area, but could articulate some memory of what the setup looked like in previous years. For example, an experienced member of the group assembling docks at the waterfront drew a diagram in the sand to illustrate that the assembled dock would look like a capital H. (p. 324)

Setting up activities in a recognizable way is a vital task so that the same old Poplar Grove is reconstituted every year and is familiar to staff and visitors alike who have experienced it previously. But Poplar Grove will also employ new staff and will receive children who have not visited it before and for these there is no ‘same old Poplar Grove’. What Birnholtz et al. recognize is that organization emerges through an interplay of human and nonhuman activity. They also acknowledge that new staff has access to more experienced staff who can articulate how the equipment has been assembled previously. However, the actual practice of setting up is missed because we are uninformed what citational articulations are communicated and how they differ from non-citational communications, other than that a text depicting an H was drawn on the waterfront. Similarly, no real discussion is evident of the agency of nonhumans, texts, in how ‘the same old Poplar Grove’ is accomplished. It is Camp Poplar Grove’s texts that remain largely constant over the years and bridge the gap between holiday seasons. These are opportunities missed, because rather than a script being performed, as Pentland et al. (2010) describe, what appears to have happened is that Camp Poplar is a performative, meaning it is both constituted anew and its recognizable identity is produced, reproduced, and consumed every summer. Some insight into the way experience is communicated through conversation and text dialectics that co-orients new-starters toward reconstituting the same old Poplar Grove, other than just to remark that it is done, would illuminate our understanding of how regenerative routines unfold in practice.

In the next section, I consider how the Montreal School of communication has addressed issues of structure/agency, macro/micro distinctions, demonstrating that its approach seems consistent with Latour’s original description of ostensive and performative.

Communicative constitution of organization—the Montreal School

A structurationist perspective of communication (e.g. McPhee and Iverson, 2009; McPhee and Zaug, 2009) conceives structure and agency as inseparably linked. This view is held in common by routines researchers and leads them to frame Latour’s ostensive and performative in the same way. However, any suggestion of duality is rejected by Cooren et al. (2006), from the Montreal School, who challenge the recursivity assumption held by structurationists and align themselves with the Latourian view that the ‘macro’ can be discerned through a focus on the ‘micro’: ‘structuring effects can [indeed] be identified through a bottom-up approach without resorting to any form of duality or dualism’ (Cooren et al., 2006: 533). Taylor et al. (1996) note that in communication theory, structurationist conceptualizations lead scholars to reject practice-level analytical stances. While the routines literature claims to see ‘routines as practices’ (Feldman and Orlikowski, 2011: 1245; Pentland, 2011), the level of analytical abstraction such studies employ means they often fail to uncover how routines are actually constituted (cf. Bapuji et al., 2012; Dionysiou and Tsoukas, 2013), settling instead for abstract representations.

Putnam and McPhee (2009: 192) identify a structurationist framing of communication as deductive and that sees actors drawing on structural rules [italics added] during their face-to-face interrelations. Again, a similarity exists with the rule-based view of routines that persists in some quarters (e.g. Cyert and March, 1992; Lewin et al., 2011). Cooren and Fairhurst (2009: 119) charge that structurationist descriptions of communicative flows tend to vacillate between human and organizational agency (Taylor and Van Every, 2000). Structurationists conflate the two so that what is presented as practice research is commonly scholarship removed from the messiness and complexity of unfolding practice as researchers do not have the conceptual tools to handle these kinds of data (see Pentland et al., 2010). What we get instead is a focus on ‘parts’ of the routine, or the routine as ‘entity’ (Parmigiani and Howard-Grenville, 2011; Rerup and Feldman, 2011). And, a baffling attempt to link an evolutionary economics perspective with practice: ‘Although grounded in evolutionary economics and thus paying relatively less attention to individual agency in routines, many of the ideas it advances remain consistent with a practice perspective’ (Parmigiani and Howard-Grenville, 2011: 416). Structuration theory is presented as a theory of practice (Rerup and Feldman, 2011), when it is more a theory of formal process as it offers little in the way of actionable advice that researchers can employ to get them closer to describing how routines are actually accomplished (Stones, 2005).

Critiques of structurationist conceptions of communication argue that structure does not dictate—from the top down—what social actors actually do in the messy worlds of organizations (Cooren and Fairhurst, 2009). Cooren and Fairhurst (2009) assert that any reference to structure or structures ‘is completely useless to explain what is happening in … specific situation[s]’ [emphasis in original] (pp. 133–34). In their view, ‘the term “structure” is a conceptual hodgepodge that does not explain anything’ (Cooren and Fairhurst, 2009: 136). They challenge lingering notions that there are any overarching structures transcending the local accomplishments of actors. What Cooren and Fairhurst (2009) point to instead are ‘strings of associations’ that relationally link actors, both human and nonhuman, in ongoing co-orientation as routines unfold. The outcomes routines produce are both the planned and unintentional products of these associations. This is a performative framing of practice as it acknowledges that ‘hybrid action[s]’ (Latour, 1993) of human and nonhuman associations construct organization through their ongoing relational entwinement.

Rather than seeing the relationship between structure and agency as simply recursive, Cooren and Fairhurst (2009) note that ‘[T]here is no structure overarching human interactions … mak[ing] human actors do things’ (p. 137). What constitutes routines, from a Montreal School CCO standpoint, are combined human and nonhuman acts, some of which transcend the immediate ‘here and now’ and become the ‘there and then’, and some of which transcend the ‘there and then’ and influence the ‘here and now’ (Cooren and Fairhurst, 2009). Cooren and Fairhurst (2009: 138, 139) argue that the relationship between the ‘here and now’ and the ‘there and then’ is not one of uncomplicated mutual reinforcement, but a relational mangling (Pickering, 1993) of hybrid associations where human participants are acted upon as much as they act. What typically transcend the ‘here and now’ are immutable texts as these are generally, but not always, constants in a routine whereas the human actors may not be (Cooren and Fairhurst, 2009; Hardy, 2004). To merely assert that a relationship is recursive fails to account for the kind of complexity endemic within organizational routines. Structurationist perspectives, founded on an assumption of a recursive relationship between ostensive and performative, cannot explain how routines can both enact ostensive[s] and provide for instances when ‘specific performances of routines start to constitute the ostensive aspects of … routines, they may become different from what was imagined’ (Rerup and Feldman, 2011: 582). Birnholtz et al.’s (2007) Camp Poplar Grove changed and yet remained the same, but these researchers did not get close enough to the practices involved to observe how the texts and citational articulations more experienced staff shared with newcomers produced this.

The next example draws from Bapuji et al.’s (2012) actor-network theory-inspired analysis of a towel-changing routine in a hotel. This work, while not based on a framing of routines as communications, has a relational perspective that other published works lack. Indeed, Bapuji et al. (2012) assert that routines ‘emerge[s] from interactions between actors’(p. 1586) and that to understand them ‘it is necessary to examine the interactions occurring in [a] routine[s]’ (p. 1587). Following Emirbayer (1997), I avoid the term ‘interaction’ in favor of ‘interrelating’ as this denotes a more relational ontology. By recognizing interrelationships, Bapuji et al. (2012) move away from the entity approach of most routines researchers and open the door to framing them as constituted through conversation and text dialectics.

Drawing from a blend of interviews and a field experiment, the research team was interested in how housekeeping staff members decided whether or not to change guests’ towels based on where in the bathroom they were left (Bapuji et al., 2012). The field experiment involved making changes to bedrooms on one floor, the treatment group. This change involved inserting a basket in the bathroom for towels to be changed and asking guests to put the towels they wanted to keep on hooks on the bathroom door. Rooms on another floor, the control group, were left unchanged with guests asked to place the towels they wanted to re-use back on the towel rack and those they wanted changed in the bath (Bapuji et al., 2012). For both groups, a message was placed in the bathroom informing guests of the system for re-using towels. When comparing the two groups, Bapuji et al. (2012) found that guests recycled nearly 30% more towels in the treatment group than the control group. This example holds several important insights for advancing a CCO perspective of routines.

Although Bapuji et al. (2012: 1588) assert that ‘individuals us[ed] tools in order to carry out tasks’, their evidence actually assigns a much more active role to tools (texts) and demonstrates textual agency (Cooren, 2004, 2008) in stark relief. The aim was to co-orientate guests toward making clearer decisions around the re-use of towels. As face-to-face contact between guest and housekeeping staff did not typically occur, the hotel had to communicate with guests through texts: ‘we decided to employ another unique object that guests could use in order to convey their intention to receive towels’ (Bapuji et al., 2012: 1594). A pre-existing hook on the bathroom was identified as the place where towels intended for re-use should be placed, while a basket was placed in the room for towels to be changed:

[W]hile the housekeeping staff felt that the guests’ intentions were clear in the treatment group, guests themselves felt that the procedure in the treatment group better communicated to them the intentions of the hotel, which enabled them to act as per the procedure outlined in the sign. In the words of one interviewee: ‘The wicker basket and door hook may be only an investment of five dollars for the hotel, but, for me it’s an impression of, “wow they have a system set up here”’. (Bapuji et al., 2012: 1598)

Texts, in the form of a wicker basket and door hook, communicated a clear intent on the part of the hotel that guests responded to, or in CCO vocabulary, entwined with human actors in co-oriented practice as the routine unfolded. Bapuji et al. (2012) do not really explore how humans and nonhumans interrelate (interact) other than stating that they do. A CCO lens on routines would lead researchers to actively seek out such relationality as these are likely to be the associations that make a difference in how routines emerge. A CCO framing would identify texts as both the medium of particular unfoldings of routines and as transcending individual iterations (Hardy, 2004). Texts possess the potential to live beyond the immediate practice and constitute what Latour sees as the ostensive aspect as they represent typifications, not specifications. However, a focus on texts that excludes human conversation is unlikely to get close enough to discern practice in any performative sense (Latour, 1986).

A further issue Bapuji et al. (2012) raise but do not pursue is that of routines as embodied phenomena that require bodies and not just hands or minds to accomplish. They (Bapuji, 2012) report:

Another guest felt that he would follow the instructions on the treatment group sign because the overall set-up communicated to him that the hotel had strong intentions to follow through on the procedure outlined. In his words: ‘I would say that if they put a basket out there for you, they’re trying to accommodate you, they’re giving you that service, as opposed to saying, just throw it in the bathtub, which is taking up your own space, or invading what’s already yours’. (p. 1598)

These guest’s words highlight that the hotel bathroom is a space where human bodies constitute and experience their realities and that such places are not mere neutral containers, but themselves act, and in a routines sense repeatedly act, in how organization emerges. Indeed, citational space could itself be considered a text that matters in how routines unfold. In body-less routines research, if actors are portrayed, they are seen as disembodied; they do things, but we are not informed how it is that they do. In the messiness of practice, actors’ bodies both reiteratively do and communicate, and communication changes bodies. The body is ignored in current routines research as emphasis is placed on steps or stages a routine is said to go through, whereas, framing routines as communications invites the body into such research. However, researchers can only get close to bodies constituting routines if they observe the fine-grained, nuanced moves of which routines are made.

Research implications

Routines are practices that have an essential citational quality integral to their form. Citationality denotes routines as reiterative performatives (Butler, 1993) and distinguishes them from non-routine practice. A CCO framing of routines articulates their condition as well as their mode of production as citational communications. Definitionally, a CCO conceptualization constructs routines as citational patterns of embodied conversation and textual dialectics that performatively co-orient toward an object. I have argued that absent from current conceptions is the notion of embodiment and a clearer expression of performative, both of which are needed to supply researchers with a vocabulary for describing how it is that routines are communicatively constituted. Therefore, any claim toward a theory of routines as practices must be premature, as to date a considerable body of existing research had neglected to theorize how they are actually constructed and sustained.

Routines as performatives refers to how they are constituted and reconstituted; it rejects the notion that they are disembodied performances of some higher-level script (McKinlay, 2010), as this diverts attention away from the embodied articulations of conversation and text, and toward abstract representations. Routines as performatives foreground the strings of associations that relationally link actors, both human and nonhuman, in ongoing co-orientation. Latour’s (1986) notion of performative needed unpicking, clarifying, and developing. Its attraction to those interested in how experience unfolds is clear from the routines literature, but it requires the fuller articulation that CCO supplies for it to translate into actionable knowledge. CCO equips researchers with an ontological framing and the epistemological wherewithal to discern how performative routines unfold through citational practice. To move toward a clearer performative conceptualization of routines that speaks of how they are constituted, it is necessary to adopt a CCO research approach. I set out below three areas where future research could focus.

First, a fundamental principle of CCO is that organization emerges through an interplay of conversation and text as actors, both human and nonhuman, co-orientate, producing organization. Future routines research could focus on how conversation and text interrelate (Cooren et al., 2011; Kuhn, 2008; Taylor et al., 1996; Taylor and Robichaud, 2004). Bapuji et al. (2012) give some hint as to how this would work in an actual study; however, their actor–network theory approach lacks the depth CCO research pursues. CCO researchers assume that routines are constituted through communication and text and therefore actively seek out those instances when conversation or text, or an entwinement of the two, makes a difference in how specific routines emerge. Co-orientation does not denote an objective but an ‘object’, suggesting that through dialogic conversations and textual agency organization as relational object is accomplished (Cooren et al., 2011; Taylor and Robichaud, 2004). The ‘object’ of a routine could be pre-formed, but co-orientation also holds the possibility that it will emerge in a different form to how it was envisaged as the routine progresses (Groleau, 2006). This is important as it allows for political maneuvering and exercises of power when the ‘object’ of co-orientation is compromised, contested, fudged, or simply forgotten (Wright, 2013) to be integrated into routines research, both of which are currently underrepresented. Conversation and text interrelating and co-orientation offer routines researchers a vocabulary for handling how routines are constituted.

Second, I have been critical of how the routines literature treats Latour’s (1986) notions of ostensive and performative. Structurationist interpretations run the danger of becoming reified as the understanding of ostensive and performative among routines researchers. Iannacci and Hatzaras’ (2012) recent critical realist analysis of ostensive is a much-needed break from this. Current routines research overplays rather than problematizes an assumed ostensive/performative (‘macro/micro’) recursivity and consequently reveals little about how local and dis-local may interrelate. The Montreal School’s perspective of CCO is skeptical about such distinctions. Practitioners experience the local ‘here and now’, which can and in some instances certainly will be influenced by the ‘bearing down’ of the dis-local ‘there and then’ (Cooren and Fairhurst, 2009). And, local practices can ‘scale up and along’ to shape the ‘there and then’. But, to say this relationship is one of straightforward recursivity oversimplifies complex interrelating. It would be fruitful for routines researchers to discern what texts ‘bear down’ on practice, how this occurs, when this occurs, and whether this influence is uniform over different iterations of the same routine. Similarly, a challenge exists to unpick which aspects of conversation and textual co-orientation transcend the ‘here and now’ and re-fashion the ‘there and then’, how this is done, why this is done, and what effects it has? The only way researchers can achieve this is by getting close to actors accomplishing routines to observe and to listen to how they feel their localized and dislocalized practice is shaped.

Finally, I have made the point several times that structuration theory does not appear to supply routines researchers with a conceptual toolkit that allows them to make sense of performative experiences. In consequence, their focus lies in abstract representation rather than practice. Conversely, CCO offers a suite of constructs researchers can draw from to help them interpret and understand how it is that organization emerges from communicative acts. Not all communication progresses a routine in the same way; deontism (Cooren and Taylor, 1997; Taylor, 2006; Von Wright, 1951) refers to communicational moves that oblige, permit, or forbid action. Signs requesting guests to place the towels they plan to re-use on a hook on the bathroom door (Bapuji et al., 2012) are deontic texts. Teleaction (Cooper, 2009; Cooren et al., 2006) allows for those not present to have their views represented in performative practice. Actions approving the invoice in Pentland et al.’s (2010) study see the approver representing more than just himself or herself. Approval signifies agreement of a greater authority that is represented by someone who impacts directly on the routine. Immutable mobiles (Cooren et al., 2007; Latour, 1987) acknowledge that nonhuman texts are not all of the same hue, some exist prior to the routine, some change during the unfolding of it, and some do not. The set-up of the docks in Birnholtz et al.’s (2007) study involved texts, some of which were immutable and some of which were mutable and changed as the human actors assembled the dock such that it would resemble a capital H. And finally, it is worth remembering that it is not just speaking mouths and writing hands that communicate; the body through gesture, posture, movement, spatial awareness, signs, and looks also communicates (e.g. Hindmarsh and Pilnick, 2002, 2007). Organizational routines are embodied performatives, constituted communications whose investigation enhances our understanding of how repetitive practice contributes to organization.

Conclusion

This article argues for a framing of organizational routines as communicatively constituted constructions. To do this, present conceptualizations based on a structurationist ontology that result in claims toward a theory of routines as practices have been critiqued. Central to this criticism has been two interrelated themes. First, that present understandings of routines as comprising recursively related ostensive and performative aspects, a structure and agency duality, have produced inadequate descriptions and explanations of routines as practices leading to recent recognition that extant knowledge of how routines are constituted and reconstituted remain underdeveloped. Clarifying performative, by drawing on its speech act theory roots and distinguishing it from performance, helps construct routines as citational performatives. Second, current research sees routines as accomplished by disembodied actors who are seen to enact some overarching script. Taking human bodies as constitutive in the reiteration of routines constructs them as bodily accomplishments, involving humans and nonhumans in a conversation and text dialectic co-oriented toward an evolving object. Studying routines as practices has proved beyond structurationist researchers who seem both disinclined and ill-equipped to describe how reiterative routine experiences unfold in and between organizations. A CCO lens, on the other hand, offers a vocabulary for articulating the communicative strings of associations that assemble and re-assemble constituting routines.