تئوری ساختاردهی و مفاهیم واسطه ای: مشکلات و پیامدها برای تحقیقات حسابداری مدیریت

Giddens’ way of conceptualizing how structures work as both the medium for and outcome of human action – duality of structure – has been emphasized as a valuable point of departure when studying management accounting in its social context. However, we argue that in the literature there are different ways of using mediating concepts between social structure and action, whereby management accounting systems are conceptualized as both the medium for action, and human action as such. Using the often-cited article by Burns and Scapens [Burns J, Scapens RW. Conceptualizing management accounting change: an institutional framework. Management Accounting Research 2000;11(1):3–25] as an illustrative example, we discuss theoretical and methodological consequences of these different ways of conceptualizing management accounting. A main conclusion is that when management accounting is defined through concurrently referring to both ‘virtual’ structures that generate action and the situated doings of individuals, structure and action risk becoming conflated and there is a risk of drawing erroneous conclusions about structural change or stability. The paper closes with some methodological suggestions as to how these problems can be avoided.

For some 20 years now, Giddens’ theory of structuration has been proposed as a useful means of conducting alternative management accounting research (see Baxter and Chua, 2003). In particular, Giddens’ way of conceptualizing how social structures work as both the medium for and outcome of human action, has been emphasized as a valuable point of departure when studying management accounting in its social context. For example, following the early ideas of how structural elements shape the practices of accounting (e.g. Roberts and Scapens, 1985), or how accounting as such might be seen as constituting part of social structures (e.g. Macintosh and Scapens, 1990 and Scapens and Macintosh, 1996), a number of management accounting researchers have drawn upon this idea of a duality of structure in empirical research (e.g. Ahrens and Chapman, 2002, Collier, 2001, Conrad, 2005, Dirsmith et al., 1997, Granlund, 2001, Jack, 2005, Scapens and Roberts, 1993, Seal et al., 2004 and Soin et al., 2002), as well as in more conceptually oriented studies (e.g. Burns and Scapens, 2000 and Dillard et al., 2004). There has also been a rather substantial debate on how to put Giddens into work in management accounting research (see Boland, 1993, Boland, 1996, Macintosh and Scapens, 1990 and Scapens and Macintosh, 1996). Hence, a number of management accounting researchers have taken sides with Giddens and his way of dealing with the since long lasting sociological debate on the relationship between social structures and human action (see Dawe, 1970, Joas, 1987, Mouzelis, 1989, Thompson, 1989 and Urry, 1982). That is, through drawing upon Giddens’ notion of a duality of structure, researchers have allied themselves with the idea that actors draw upon structural elements when acting, “by the same token reconstituting their structural properties” (Giddens, 1984, p. 28). This, in turn, implies that social structure and human action are not viewed as two independently given sets of phenomena – on the contrary – they presuppose each other (Lawrence et al., 1997, Macintosh and Scapens, 1990 and Scapens and Macintosh, 1996). It has been argued (and we agree) that this idea of social structure and human action as intrinsically interconnected is one of the main contributions of structuration theory to the study of management accounting as a social and political phenomenon (Macintosh and Scapens, 1990 and Scapens and Macintosh, 1996). Importantly, however, by assigning our underlying arguments to structuration theory, we concurrently adhere to Giddens, 1976, Giddens, 1979, Giddens, 1981 and Giddens, 1984 conceptual distinctions between structure and action. That is, even though Giddens tries to overcome the dualism between structure and action, whereby structures are conceptualized as detached from human beings, he nevertheless argues that structure and action have fundamentally different properties (for a critique of Giddens view on this, see, e.g. Archer, 1982 and Willmott, 1986). More specifically, he proposes that human action is always situated in time-space and conducted by a subject, while structures have no specific time-space location and are “characterized by the ‘absence of a subject”’ (1976, p. 119). Furthermore, he recurrently emphasizes that structures are conceptually distinguished from action in terms of their generative characteristics (see e.g. 1979, p. 66). In this paper, we argue that although management accounting researchers often explicitly refer and adhere to Giddens original writings in these respects, the literature in the area seems to be disparate in terms of how these conceptual distinctions between structure and action are dealt with. More specifically, we argue that it is unclear as to how mediating concepts between social structure and human action are used. For example, some researchers have referred to management accounting systems as modalities (Giddens, 1984), whereby the system acts as an interpretative scheme which mediates between “the signification structure and social interaction in the form of communication between managers” (Macintosh and Scapens, 1990, p. 460) or between “structure of domination and power” (Mouritsen, 1990, p. 74). Hence, through referring to management accounting systems as modalities, characterized as rules and resources that actors draw upon in day-to-day actions, it seems evident that the concept of management accounting refers to properties of social structure, i.e. represents something that generates action. Arguably, however, management accounting researchers have also used mediating concepts to denote social action as such. For example, Burns and Scapens (2000, p. 9) suggested that accounting rules and routines could be viewed as “observable, recurrent activities and patterns of interaction characteristic of a particular setting”. And drawing upon Burns and Scapens’ (2000) framework, Lukka (2007, p. 94) suggested that we need to pay attention to both rules and routines, which “jointly comprise management accounting practice”. Along the same lines, Soin et al. (2002, p. 260) argued that “much of the work process in Clearing [department] was routine and ‘countable”’ when applying the mediating concept ‘routines’ empirically. Hence, it seems as if management accounting researchers have used mediating concepts in two fundamentally different ways, namely, as social structures generating action (modalities) and as observable and recurrent social action per se (similar observations have also been made in other literatures, see, e.g. Barley and Tolbert, 1997, Becker, 2004, Pentland and Rueter, 1994 and Ravasi and van Rekom, 2003). We argue that this conceptual disparity poses a threat to theory development in the area because the different ways of using mediating concepts in ‘structuration-type’ of research have different theoretical and methodological implications, of which some may be problematic. Furthermore, it risks leading to that we as researchers become unclear as to what we mean by mediating concepts and how our results should be interpreted. And, consequently, it becomes difficult to compare and relate results from different studies to each other. As an attempt to address these issues, the aim of this paper is to discuss alternative approaches to using mediating concepts between social structures and human action, and elaborate on their respective theoretical and methodological implications. For illustrating purposes, the conceptual framework proposed by Burns and Scapens (2000) is used as a point of departure, mainly for four reasons. First, Burns and Scapens (hereafter B&S) explicitly draw upon structuration theory when formulating their fundamental ideas about the relationship between social structure and human action (duality of structure).2 Second, the B&S framework has become a piece of work that stands as a model for many of those who work within this field (see, e.g. Johansson and Baldvinsdottir, 2003, Lukka, 2007 and Soin et al., 2002; Yazdifar et al., 2008). Third, in their framework, B&S use ‘accounting rules’ and ‘routines’ as mediating concepts between social structure and human action. Finally, but not least, they seem to concurrently assign the concepts both to the structural realm and the realm of action, which brings the more general question of the distinction between structure and action to the fore. In the following sections, we describe some of the basic premises that underpin the B&S framework and analyze how the mediating notions rules and routines are conceptualized. We then highlight potential theoretical problems that arise when they are concurrently viewed as social structures generating action and as action as such. Finally, we summarize the key arguments and discuss some methodological implications for future research.

As was stated in the introduction section, structuration theory has been proposed as a useful means of understanding the social aspects of management accounting. Furthermore, since one of the main propositions of structuration theory is the duality of structure, whereby actors “draw on the modalities of structuration in the reproduction of systems of interaction, by the same token reconstituting their structural properties” (Giddens, 1984, p. 28), it was argued that it is important to make a distinction between, on the one hand, what is drawn upon and, on the other hand, social action per se (cf. Giddens, 1976). Hence, in a management accounting context we need notions both to describe the situated and recurrent management accounting practices as such, and to denote the non-situated modalities that inform those management accounting practices. However, in this respect, the management accounting area (like many other areas) seems to suffer from conceptual disparity. That is, some researchers seem to view management accounting systems as modalities drawn upon in the production and reproduction of practices ( Macintosh and Scapens, 1990 and Seal et al., 2004), while others seem to view the same systems as recurrent management accounting practices ( Lukka, 2007 and Soin et al., 2002). While this disparity can be problematic in its own right, i.e. it is unclear as to which notion one is referring, the discussion in the previous section suggests that the main problem concerns when the two different ways of conceptualizing management accounting systems are combined (mixed) in one and the same study. Again, this is so because when we use both ways of conceptualizing one and the same management accounting phenomenon, social structure and action risk becoming conflated and potentially we may draw erroneous conclusions about structural change or stability. Furthermore, this distinction between situated management accounting practices and the non-situated principles that enable and constrain those management accounting practices has important methodological implications. The argument for this is as follows. As noted above, key characteristics of management accounting practices as such are that they are observable, always situated in specific time-space locations and are conducted by a subject, while social structures have the opposite characteristics (Giddens, 1984 and Sewell, 1992). As we see it, this distinction implies that we as researchers, by using appropriate research methods, need to capture both the situated doings of individuals and the non-situated principles underlying those actions. True, as proposed by Giddens, 1979 and Giddens, 1984 and others (e.g. Scapens and Macintosh, 1996), different emphasis may be put on the way in which structure and action presuppose each other by means of methodological bracketing. That is, in a specific analysis, primacy can be given to either the agent as a skillful and conscious actor, as in the analysis of strategic conduct, or to institutions as “chronically reproduced rules and resources” (Giddens, 1984, p. 375), as in institutional analysis. In the former case, the structural properties are assumed to be given in a specific context, thereby concentrating the analysis upon how actors use their practical and discursive consciousness about how to go on in social encounters (1979). In the latter case, “the skills and awareness of actors” (1979, p. 375) are placed in suspension, thereby giving primacy to the ways in which modalities act as the very medium for the reproduction of social systems (p. 95). However, as stated by Giddens, regardless of the form of methodological bracketing; “there is no clear-cut line that can be drawn between these, and each, crucially, has to be in principle rounded out by a concentration upon the duality of structure” (1984, p. 288). This is so, Giddens argue, because the distinction between them is methodological, not substantive (1979). Hence, regardless of which methodological approach is used to study the duality of structure in a management accounting context, we need to distinguish the situated doings of individuals from the underlying principles guiding and constraining those (inter-) actions. Accordingly, on the one hand we must collect data on situated recurrent (inter-) action, where a nomenclature is needed in order to make sense of and categorize the management accounting practices that can be observed. On the other hand, we also need to identify and understand the non-situated principles (including management accounting structures) which enable and constrain the situated practices. This, in turn, implies that we preferably should use several, complementary methods of collecting data. The premise is (again) that while participative observation is crucial for the identification of recurrent management accounting action, the recurrent (inter-) actions themselves often reveal very little about the non-situated and virtual principles that guide this action (cf. the fictive strip of interaction presented above where the two costing approaches underpinning the actors’ behavior were never explicitly spelled out in the conversation). Accordingly, we also need to use various types of non-observational sources of data such as interviews, training manuals and minutes in order to seek actors’ interpretations/explanations of their actions. As was discussed above, the use of multiple sources of data is also important in order to avoid the problems associated with viewing a change in behavioral patterns alone as an indicator of structural change. That is, ideally we should compile evidence of structural change (or stability) which is independent of the data from which the patterned social actions (scripts) are derived (Barley and Tolbert, 1997). For example, as suggested by Burns and Scapens (2000) and others (Seal et al., 2004) management accounting ‘rules’, in terms of manuals, procedures, documents and the like can and should be observed, since these may embody structural principles. As argued by Seal et al. (2004, p. 78): Traces of modalities in the accounting domain include formal documents such as business plans, budgets and financial contracts that have encoded more general institutional principles in specific social settings. Indeed, in some cases the social structures implicated in the (re)production of practice are only tacitly grasped by actors which implies that they implicitly draw upon the stock of knowledge about ‘how to go on’ but cannot discursively give an account for their actions. Accordingly, it is unlikely that the underlying principles have been articulated in formal documents or can be articulated in interviews. In these cases, we may thus be reduced to using only participative observation. This, in turn, implies that social structures must be abstracted solely from what is expressed and done in specific time-space locations. However, and this is important, from an analytical perspective we must not confuse what is said and done with what is drawn upon. Again, regular practice per se is not a rule, since a rule is not a generalization of what people say or do (Giddens, 1979, p. 67). A rule is a ‘virtual’ memory trace that allows the binding of time and space, through its enabling and constraining characteristics. Hence, and to conclude, we suggest that future management accounting research should explicitly state how mediating concepts are defined and used, in order to avoid further conceptual disparity. In doing so, we propose that researchers need to choose research designs appropriate to make the distinction between situated management accounting practices per se and the non-situated modalities that enable and constrain those practices. We also believe that it is important to fill both categories of management accounting concepts with empirically grounded meaning and content beyond the existing and constructive contributions of Burns and Scapens (2000) and others ( Macintosh and Scapens, 1990 and Seal et al., 2004). In these ways, we may hopefully further develop our understanding of management accounting as a social phenomenon.