Leadership in organisational change: A post-structuralist research agenda

The organisational change literature remains dominated by macro- and microexplanatory models which tend to exclude conflict, mess and power in favour of enumerating universalistic steps or, as is the subject of this article, leadership definitions and factors for successful change. In this article, I review and question some of the mainstream literature on leadership in organisational change, drawing on Laclau and Mouffe’s political discourse theory and its mobilisation by critical leadership studies of organisational change. This article problematises change leadership as a set of multiple and changing practices, pragmatically deployed by organisational players. In exploring those avenues, I deploy a five-step ‘logics of critical explanation’ approach – specifically designed by Laclauian discourse theorists – characterising organisational change practices according to social (rules and norms), political (inclusions and exclusions) and fantasmatic (fears and hopes) logics. Rather than a set of factors or top–down causes and effects, I offer a situated and critical explanation of leadership in organisational change. This research contributes to critical explanations of organisational change politics by considering leadership as a set of changing discursive practices and by developing four situated dimensions of leadership, which build on concepts of empty and floating signifiers, to add to discussions of the role of individuals in organisational politics.


Introduction
The organisational change (OC) literature remains dominated by macro and microexplanatory models, which tend to exclude conflict, mess and power in favour of enumerating universalistic steps or, as is the subject of this article, leadership definitions and factors for successful change.
Taking an OC study focus, I problematise this status quo, drawing on Laclau and Mouffe's (1985) discourse theory and its mobilisation by some critical leadership studies (CLS) (Alvesson and Spicer, 2012;Harding, 2005;Kelly, 2014). In beginning to critically understand leadership in OC, I deploy a five-step 'logics of critical explanation' approach (Glynos and Howarth, 2007), which helps characterise the accepted norms and rules of leadership in a given organisation (social logics), its inclusions and exclusions of demands (political logics) and the hopes and fears gripped by such a concept (fantasmatic logics) in a given case. This article articulates data collected from a 9-month case study of an English local strategic partnership (LSP) formulating a project of commissioning and integration to deal with Government austerity. Based on a four-decade long genealogy of this locality, four situated dimensions of leadership are proposed to better understand the practices linked to leadership and leaders in this specific locality. In sum, I argue that leadership can be represented as a set of multiple and changing practices, pragmatically deployed by organisational subjects to re/draw alliances and, ultimately, exercise power.
The article is organised as follows. The first section highlights the lack of critique, complexity and politics of a majority of the leadership literature, leading to considering CLS and its emphasis on complexity and conflicts as constitutive of leadership. The second section outlines the article's discursive approach to leadership, emphasising how situated dimensions of leadership can add to this literature by helping to account for the complexity and conflictual aspects of leadership in OC. This is demonstrated in the third section, which mobilises data collected from the case study. This leads to concluding remarks and proposals for future leadership research.

Leadership in OC: a vast literature and a growing critical agenda
From the 1980s, the literature on OC leadership is dominated by performative models advising on where leadership can be found and distilling how it leads to successful change (Bass, 1985;Kotter, 1988;Kuipers et al., 2014 for a review). Yet, leadership in organisations continues to remain elusive, Rost (1993) claiming that two-thirds of organisational leadership studies do not actually define leadership. Furthermore, few articles -27 for the period 1990-2010 -deal empirically with leadership in OC (Ford and Ford, 2012: 3). To the point where leadership has become 'slippery' and 'understood as nearly anything' (Spicer and Alvesson, 2011: 194-195). Significantly for this article, understandings of leadership tend to remain dominated by tales of heroic individuals endowed with the likes of charisma and transactional skills (e.g. Herold et al., 2008;Yukl, 1994 for a critique). Issues of power, conflict and mess are also often sidelined by this dominant literature (Kuipers et al., 2014: 2, 33).
To address these limitations, as I suggest above, a growing group of studies, sometimes known as 'critical leadership studies', has emphasised the meaning, complexity and relational dimensions of organisational leadership (Alvesson and Spicer, 2012;Collinson, 2012;Driver, 2012;Harding, 2005). These studies find their roots in critical management studies (CMS), an eclectic school examining organisational and management issues (Parker, 2014) and mobilising, among others, discursive approaches such as Laclauian ones to analyse conflicts, the ambiguity of OC projects and the manufacturing of consent (Bridgman and Willmott, 2006;Parker and Dent, 1996;Spicer et al., 2009). Focusing on identity construction, Collinson (2006: 185) argues, for instance, that the constitution and practices of leadership, including during OC moments, involve complex identity work, interactions and negotiations with 'followers' (see also Fairhurst and Uhl-Bien, 2012;Gleeson and Shain, 2003). Often, in these studies, leadership is also understood as messy and relational. The building of informal coalitions of change leaders (Rodgers, 2006) or the identification and reconciliation of paradoxes created by the multiple realities of OC (Kan and Parry, 2004) are thus foregrounded in critical explanations of OC. Such explanations also emphasise the political, ideological (Dellagnelo et al., 2014;Harding, 2005;Sinclair, 2007;Western, 2008) and discursive aspects of organisational leadership (Alvesson and Sveningsson, 2012), demonstrating the need for de-essentialising the ontological definition of leadership (Grint, 2005). Levy and Scully (2007) have, for their part, deployed a post-structuralist reading of Gramsci's concept of power as hegemony, arguing that individuals in organisations may 'serve […] as a contemporary Modern Prince, a political agent who transforms systems through skilful analysis, building organisational capacity, the development of smart strategy, and effective leadership' (cited in Levy et al., 2009: 2). Here again, rather than a dialectical understanding, leadership is understood as implying the diverse drawing of consent by the deployment of strategies by 'Modern Princes'.
These studies ultimately highlight the partial and changing definitions and practices of leadership (Gemmill and Oakley, 1992), opening up the possibility for alternatives. Thus, they have sought to challenge the role of conflict in formulating and implementing OC, seeing struggle not as a 'barrier', as is often the case in mainstream accounts, but as the highlighting of alternatives (Knights and Murray, 1994; see also Alvesson and Sveningsson, 2003;Collinson, 2011). They have even begun to problematise the 'heroic leader' (Alvesson and Spicer, 2012;Collinson, 2011;Fairhurst, 2010;Harding, 2014;Van Knippenberg and Hogg, 2003), suggesting that 'leadership can exist without leaders' (Sutherland et al., 2014: 764), and interrogating how leaders come to occupy a symbolic role in organisations and change (Ford, 2006). Such perspectives delve into what appeals to 'followers' within change discourses, analysing how individuals are brought into hoping for the leader's 'vision' to become true, that s/he will protect the organisation against threats, or that they will themselves develop 'leadership' practices. Indeed, Calás and Smircich (1991) stress the emotional aspect of leadership in convincing, changing and charming their 'followers' while Kelly (2014) explains how leadership represents 'a space of absent presence' serving 'to create the conditions of possibility for many competing and complementary definitions, meanings and interpretations ' (pp. 905-906). Thus, leadership 'must always be described and represented by somebody or something else' (Kelly, 2014: 906).

Critically conceptualising leadership in OC research: a post-structural discursive agenda
Rather than seeking to 'discover' and teach how 'effective' leadership results in 'successful' OC -as most of the literature continues to endeavour -CLS have problematised this relationship, interrogating how and why leadership is constituted, articulated and contested in OC discourses (Gagnon and Collinson, 2014;Sinclair, 2007;Tourish et al., 2010;Western, 2008). I argue, however, that these critical studies have some limitations which could be remedied via discourse theory. For example, some remain confined in their analysis to given categories of leadership, focusing for instance on the study of leaders as actors, or leadership discourse as communicative interaction (e.g. Tourish, 2014). This is problematic because these remarks suppose that general typologies or definitions of leadership can be applied across different empirical cases. Others also underexplore identity processes concomitant with leadership practices (Gagnon and Collinson, 2014: 646). In addition, they sometimes reproduce similar performative goals to the mainstream leadership literature, such as how '[p]ost-structuralist theory offers a way of thinking that may be hugely insightful for people who are exploring how to become leaders' (Ford et al., 2008: 3; see also Ford and Harding, 2007). Finally, they limit their critique to the literature rather than empirical cases and 'lived experiences' (Kelly, 2014;Wilson, 2013: 111). This section formulates a 2-fold proposal for furthering critical understandings of leadership. First, I frame leadership as a changing and situated set of discursive practices and, second, articulating discursive concepts of empty signifier, floating signifier, subjectivity and agency, I propose four dimensions of leadership practices which could add a layer in critically explaining OC in a given context.

OC leadership as situated and changing discursive practices
I have already argued in the introduction that why and how questions of leadership in OC draw our focus to the political/hegemonic and ideological dimensions of leadership (Fairhurst, 2011: 503). Furthermore, leadership in OC should be analysed as contextually and historically constituted (Collinson, 2014;Kelly, 2014). What constitutes leadership in one organisation or for one change project, may not be discernible in another organisation, definitions and practices varying across time and space. For instance, in their review, Ford and Ford (2012) conclude that the literature on change leadership is problematic notably due to its 'single-point data collection […] [missing out on] the temporality of change' (p. 32). Thus, the focus should be on analysing OC leadership via in-depth case studies of particular organisations over time (e.g. Fairhurst, 2007), something which discourse theory combined with a logics approach can help to achieve in a critical way as I plan to illustrate. Other studies have further demonstrated this historical and contextual dimension by focusing on the emergence of particular practices (e.g. Heracleous and Barrett, 2001). Denis et al. (2010) for instance draw on five case studies of change in healthcare organisations, conducted over 8 years, to understand the dynamics of what they call collective leadership. For them, healthcare organisations are 'inherently pluralistic' spaces crisscrossed by different objectives, actors, values and ambiguous power relations, requiring detailed analyses. These authors conclude that change leadership is 'a dynamic phenomenon in which participants, roles and influences evolve over time' (Denis et al., 2010: 810). In their analysis of two leadership development programmes, Gagnon and Collinson (2014) document how such training opportunities saw senior managers 'creat[e] their own alternatives' of what leadership means (p. 661), attaching leadership to demands as varied as exclusiveness, specialness, obedience and networking (p. 662; see also Alvesson and Spicer, 2012;Kelly, 2014;Tourish, 2014: 81;Wilson, 2013: 5).
This recognition of the plurality and contextuality of leadership points to the dangers of conceptualising leadership as a single 'thing' or definitive set of practices. Leadership could instead be framed as linking together multiple demands. Such a formulation offers the possibility of apprehending the multiple aspects of leadership practices in cases of OC. Importantly, this could help draw attention to how leadership is being articulated to exclude particular demands or document the practices of redefinition and thus contestation mobilised by chosen 'leaders'. For instance, how is leadership articulated as a demand by change discourses? Are organisational subjects offered the possibility of becoming 'change leaders'? How do these practices allow linking disparate demands for change? Was leadership synonymous with management, performance, collaboration or empowerment, a particular individual? These are some of the questions I will grapple with in the case study section.
Before I develop in the next section the four dimensions of leadership practices as retroductively informed by the case study, it is important to discuss briefly how this framework may be used as a bridge between micro-and macroanalyses of phenomena such as leadership or austerity. Recasting the 'organisation' and 'OC' as discursively constituted allows departing from the widespread understanding, especially in local government, of change being the result of 'bigger' phenomena such as neo-liberalism or austerity. In contrast to this mainstream interpretation, the organisation itself can be better understood as a site or multiple sites criss-crossed by multiple demands, in which 'macro' or 'micro' levels of change give way to a flat ontology (Marston et al., 2015;Schatzki, 2005) and to the understanding of the organisation as always open to competing articulations and always-already lacking or dislocated. Change thus becomes the result of situated and competing articulations of disparate demands (e.g. neo-liberalism and austerity) as threats or/ and opportunities requiring the formulation and implementation of preferred practices or change 'solutions' (MacKillop, 2014).

A Laclauian discursive agenda for critical leadership research
This article draws on discourse theory as developed by Laclau and the Essex School of discourse (Howarth, 2000;Howarth et al., 2000;Laclau, 1990;Laclau and Mouffe, 1985) to frame organisations, leadership and change as discursively, historically and contextually constituted practices, rather than 'facts' or 'real' and permanent entities. According to (post-structualist) discourse theory, meaning, including 'leadership', is understood as the result of struggles between competing discourses seeking to hegemonise a given social order, for example, an organisation (Bridgman and Willmott, 2006;Contu and Willmott, 2005). The organisation can hence be reframed as a set of politically constituted and dislocated spaces, where different hegemonic strategies are deployed by strategically placed individuals to continuously redefine consent and alliances (Howarth, 2013;Laclau and Mouffe, 1985). I want to examine how power draws frontiers within organisations, specifically via practices of inclusion and exclusion (known as logics of equivalence and difference in discourse theory (Laclau and Mouffe, 1985: 134)) which leadership is often a case of.
Another key Laclauian concept here is that of demands. Demands are at first requests (Laclau, 2006: 655), for instance, an individual in an organisation may have a grievance or claim relating to her/his lack of participation in the decision making of the organisation. Different grievances may emerge across an organisation, relating for example to a lack of control, a desire for more training or increased pay. Laclau argues that these disparate requests become demands when articulated together by discourses via inclusion and exclusion, or logics of equivalence and difference (Laclau and Mouffe, 1985: 134). Thus, demands for better pay, greater decision-making power and training may become linked together by a project/discourse as united against a common 'enemy'. Finally, articulation accounts for the construction of meaning by the linking together of demands via these two logics whereby meaning becomes stabilised (to a point) thanks to some demands becoming sedimented around privileged/central demands (Laclau and Mouffe, 1985: 105).
Two Laclauian concepts present particular advantages for the analysis of popular but contested demands articulated by discourses: empty and floating signifiers. In cases of OC, where meaning in the organisation is being renegotiated, considering particular demands as empty or floating signifiers offer the possibility of critically explaining how and why such relations are being modified, concentrating notably on the power plays and beliefs surrounding the definition of those signifiers. Floating signifiers are signifiers which continue to see their meaning shift across context and perspectives, for instance 'corporate culture' (Angouri and Glynos, 2009: 11-12). Empty signifiers, for their part, are demands which become 'emptied' in order to symbolise a multiplicity of contradictory demands. In relation to organisational leadership, the concept of empty signifier has already been deployed. For instance, Ford et al. (2008: 10) argue that 'leadership, it would appear, is an 'empty signifier''', suggesting that leadership 'has a politically significant performative effect, […] [as] an object whose existence is impossible but which is central to that discourse of which it is a part' (p. 11). Ford and colleagues offer little explanation of the conditions of possibility for leadership to act as an empty signifier in specific organisational contexts. Instead, leadership is given an a priori status of empty signifier because of the diversity of meanings usually attached to it. Instead, this empty character should be dependent on multiple articulations, context, history or ideology (Howarth and Griggs, 2006). In her analysis of the National Health Service's (NHS) management, Harding (2005) also applies discourse theory to understand and critique 'how managers "make" organisations and at the same time make their managerial selves' (p. 264). In doing so, she argues that management functioned as an empty signifier in the NHS discourse, management being simultaneously absent but present. Harding argues that such empty signifiers are necessary to represent everything that is impossible to realise in an organisation, such as collaboration or patient care.
Emphasising how leadership as an empty signifier 'embodies' an organisation's discourse of change, Alvesson and Spicer (2012) developed a more relational and conflictual understanding of leadership where organisational struggles enable 'the manager' to continue redeploying her/his symbolic position, and thus allow other organisational subjects to continue reworking their own identities. In his article, Kelly (2014) argues in favour of studying the ideological character of leadership in language, while also paying attention to the myriad ways in which subjects and objects of language and action come to 'stand in' for, and temporarily fill, the empty centre of this seductive and endlessly adaptable signifier. (p. 607) Like Ford et al. (2008) and Alvesson and Spicer (2012), Kelly understands leadership as a 'given' empty signifier, with its 'seductive and endlessly adaptable' qualities. He stresses the need to understand how discursive practices such as 'subjects' and 'objects' come to 'stand in' for and fill the signifier 'leadership' in a given organisation. Similarly, Angouri and Glynos (2009) analyse 'corporate culture' as a floating signifier, foregrounding 'the political dimension of organisational practice' and 'suggesting that how this is fixed can only be determined through analysis of the practices under scrutiny' (pp. 4, 10). Although not focused on leadership per se, this study is key in emphasising the situated dimension of empty and floating signifiers, requiring to analyse the practices at play.
A final advantage of discourse theory for analysing leadership is that it reworks the role of the individual in a number of ways (Glynos and Howarth, 2007;Laclau, 1996). What matters for this article is that individuals are usually seen as satisfied with occupying pre-given subject positions (e.g. the dutiful employee, the 'collaborator'). Here, discourses appeal to individuals, winning over their consent by offering subject positions that fulfil the individual's longing for full identity (Cederstrom and Spicer, 2013;Stavrakakis, 2008). This affective dimension is key to understanding OC politics and how leadership practices appeal, or 'grip', individuals (Glynos and Howarth, 2007). Furthermore, when a discourse becomes threatened -dislocated -individuals may constitute their own subject positions by identifying with new demands outside given discourses.
What is still needed in organisational research are detailed cases of how and why leadership comes to operate as an empty or floating signifier for given organisational discourses. This is what I tackle now.

Four dimensions of leadership
In understanding how leadership is articulated in certain OC discourses, four dimensions of practices are proposed here. Rather than applicable to any case, these dimensions are outlined as hypothetical explanatory solutions to specific cases. Indeed, based on the retroductive framework of this article (cf. next section), theoretical issues that emerge during the research process may be resolved by iteratively articulating particular concepts, such as empty and floating signifiers, in addressing issues such as the role of individuals in change discourses or the mobilisation of particular demands, such as leadership, as empty/universal ones. Despite these four leadership dimensions being a contribution of the article, I have chosen to introduce them now rather than in the 'findings' section following the retroductive argument mentioned above where hypothesis and explanation are not clearly separable, both involving the same form of judgment by the analyst.
These four dimensions are as follows: (1) leadership as a demand tendentially emptied of meaning allowing to link together a multiplicity of contradictory demands across spaces, (2) leadership as a subject position emptied of meaning to symbolically represent a given organisational discourse, (3) leadership as a contested and thus floating demand, implying struggles and strategies to gather consent and (4) leadership as the practices of agency deployed in dislocatory contexts by individuals identifying with new and different demands to maintain the hegemony of a given discourse. These dimensions are now examined successively.
First, any consideration of leadership in OC has to examine which demands become linked to this signifier in a given organisational discourse. This requires documenting how leadership was mobilised (or not) to build a wide chain of equivalence and emphasise multiple possibilities. By (tendentially) emptying leadership of specific meaning, this signifier may become synonymous with demands as diverse as cooperation, performance, excellence or organisational pride (Gagnon and Collinson, 2014;Harding, 2005Harding, , 2014Spicer and Alvesson, 2011). Such an analysis should investigate the diverse and contradictory demands assembled under the guise of leadership and 'taught' during widespread leadership training programmes, and how/whether such demands grip the senior managers and other stakeholders. In doing so, the question of the role of key individuals in linking particular demands together across hegemonically defined spaces is brought to the fore, exploring the strategies (e.g. training or specialised groups) deployed to grip demands around leadership.
Second, some individuals may come to occupy subject positions of 'leaders' within a given organisational discourse -individuals occupying subject positions is one of the two aspects of identity in discourse theory (Laclau, 1996) -this/these individual/s 'standing in' for or symbolically representing that organisational discourse across contested and hegemonically defined spaces. Thus, if leadership has become synonymous with demands of collaboration, excellence and organisational pride (as hypothesised in the first dimension), an individual as 'leader' may also come to represent/embody those diverse demands. Until recently, analyses of identity and the role of individuals in organisations tended to focus either on the institutional structures and 'positions' made available to individuals in particular institutions, or on the behaviours, personal skills and voluntary character of identification. These approaches, however, underestimate the political and affective dimensions of organisational discourses, in which some individuals come to occupy certain positions in rendering such discourses hegemonic. Thus, understanding OC politics requires analysing the role that can be played by some individuals as empty signifiers, standing in as the signifier of a discourse. Indeed, for Laclau, 'the very notion of "individual" does not make sense' (Laclau, 2005: 34), any discursive element being the result of 'differential struggle […] [all] equally capable of expressing, beyond their differential identity, the absent fullness of the community […] none [being] predetermined per se to fulfil this role' (Laclau, 1996: 42). Laclau (1996) argues, for instance, that in 1960s and 1970s Argentina, Perón was 'very careful not to take any definitive stand in the factional struggles within Perónism' and found himself 'in ideal conditions to become the "empty signifier" incarnating the moment of universality in the chain of equivalences which unified the popular camp' (p. 55). Individuals as empty signifiers could thus produce a dynamic understanding of the role of individuals in discourses of OC.
Third, following the framing of conditions as dislocations, 'leadership' may also become the object of struggle, organisational subjects fighting over which practices should be synonymous with leadership and emphasising the changing possibilities associated with this demand. In such cases, especially frequent during OC formulation, framing leadership as a floating signifier could help analyse the negotiation and contestation of meaning. Building on previous critical research of the NHS's management, Harding (2005) for example stresses 'the complex manner by which NHS managers both absorb and resist, define and redefine, the identity of manager' (p. 269). In the same way that '[t]he manager responsible for the management of health services is thus a manager who strives to become an unattainable other, the rational manager', such a framing of leadership can lead to complex and critical discussions of the constant emptying and floating of such signifier, striving to represent demands as diverse as cooperation, performance and empowerment. Returning to Gagnon and Collinson's (2014) study of leadership training programmes 'feeding' multiple demands of what leadership ought to be, one can imagine how organisational reform occasions those designated 'leaders' or 'change leaders' to contest the meaning of leadership. Such political battles cannot be divorced from the appeal of fantasies of leadership often mobilised in organisations. Driver for instance uses psychoanalysis to 'understand the continued mystery and romance of leadership', exploring 'how leaders construct identities by drawing on existing leadership discourses […] [and] how fantasies are constructed but also, importantly, how they fail and reiterate fundamental lack' (Driver, 2012: 408).
Fourth, organisational discourses may become contested or dislocated, bringing those individuals standing in as the universal signifier of that given discourse (if this is the case, as described in the second dimension) to renegotiate the meaning of their subject position and hence of the organisational discourse as a whole. This fourth dimension relates to agency and the active identification process demonstrated by some individuals in some contexts. Indeed, if this subject position of 'leader' and the discourse it represents aim to continue mobilising a vast array of demands, and thus addressing grievances in a context of dislocation (i.e. necessary conditions for hegemony), it must sometimes renegotiate the particular demands it has become associated with. For example, these individuals as 'leaders' may in these cases strive to identify with different and novel demands (e.g. collaboration instead of performance management), rearticulating the meaning of their subject position of 'leader' and thus their own identity. In cases where the change project stricto sensu may not be implemented, OC as a discourse may still be considered a 'success' because of the ability of a given 'leader' to identify with new/different demands and thus allow a given organisational discourse to transform itself and remain hegemonic. In other words, this fourth leadership dimension may be crucial in understanding and distinguishing between the 'failure' of particular OC projects and the continued 'success' of OC discourses/regimes of practices (which is a key question in the case study).
This framing of leadership along lines of discourse theory emphasises the contradictory and disputed leadership practices in OC discourses, as illustrated by the four dimensions developed. To summarise, the first dimension of leadership offers the opportunity of discussing the drawing of equivalences between disparate demands occasioned by leadership being articulated as an empty signifier. The second dimension relates to the subject position of 'leader', which may be tendentially emptied of meaning to symbolically embody a given discourse and thus exercise power. The third dimension relates to the contestation of leadership, where the meaning of this latter is dislocated and subject to struggles, individuals striving to link leadership to contradictory demands to fulfil their individual identity lack. The fourth dimension finally allows understanding how individuals when standing in as the universal signifier of an organisational discourse may come to renegotiate the meaning of their position in dislocatory contexts, identifying with new or different demands to redeploy a given organisational discourse. This four-dimensional approach adds to CLS by offering extra ways of examining leadership practices from a non-deterministic, political and situated perspective.

Exploring discursive practices of leadership via a logics of critical explanation approach
In recent years, logics of critical explanation were formulated by post-structural researchers as a five-step methodology to help apply discourse theory to the empirical world (Glynos and Howarth, 2007). Here, I mobilise it to examine leadership practices, leading to the formulation of four leadership dimensions.
A logics approach helps analyse 'different dimensions of social reality' (Glynos and Howarth, 2007: 14) and thus different types of 'rules' governing a given system of meaning, such as higher education (Glynos and Howarth, 2007), airport expansion (Griggs and Howarth, 2013) or UK banking (Glynos et al., 2012). Thus, this approach analyses all norms, actions, identities and other discursive practices -notably leadership ones -mobilised by competing projects in exercising power over a given context, not limiting itself to 'talk and text'.
A first step problematises the phenomenon under study. This implies a longitudinal approach, implementing Foucault's genealogy, to explore the 'ignoble origins' of given discourses, allowing to understand how consent is forged over time and interrogating the 'reproduction and transformation of hegemonic orders and practices' (Glynos and Howarth, 2007;Howarth, 2000: 72-73). For instance, in this article, how was leadership formulated and became linked -or made equivalent -to changing demands over the decades to build consent? A second step of a logics approach (these steps are not successive but interlocking) consists of rendering the problematised phenomenon more intelligible, what Glynos and Howarth (2007) term 'retroductive explanation ' (p. 19). Retroduction was traditionally linked to the domain of discovery, implying 'the generation or positing of hypotheses', rather than explanation (p. 27). In this study, this step allows via iteratively articulating concepts of empty and floating signifiers, identity and discourse to make sense of leadership in the case study. Third, three types of logics are 'indispensable in helping us to explain, criticise and evaluate' problematised phenomena (Howarth, 2008). In a case of OC, contextualising interviews with social, political and fantasmatic logics allow investigating how a given reform project emerged, mobilised consent and became slowly embedded or failed. Social logics interrogate what are considered to be the rules and norms governing social practices in a given case. Political logics allow characterising how demands -identities, actions, beliefs, policies or other discursive practices -are brought in or excluded by an OC project. Finally, fantasmatic logics identify the affective dimension of leadership/change discourses, examining how demands, and particularly individuals' identities, become 'gripped' by particular discursive practices. The fourth step of a logics approach involves articulation. This is a fundamental methodological 'tool' in explaining and critiquing problematised phenomena. It also implies that theoretical concepts (the ontological) and objects of study (the empirical) cannot be considered as immune from each other. Instead, both are modified by the intervention of the researcher, as done here with leadership. Fifth, by making visible the moments of contestation, domination and excluded possibilities (i.e. the political and fantasmatic dimensions of social reality), this analytical framework opens up the space for a critique of leadership and change practices in a given context. Unlike conceptual frameworks based on pre-given categories of evaluation such as rationality or interest maximisation, a discursive and logics framework sees its normative 'bedrock' being constituted in situ, according to the context and practices under study.

Case study
This article critically explores leadership practices in an English County Council, anonymised as 'Internshire County Council' (ICC) and its LSP, anonymised as 'Internshire Together'. ICC is an upper-tier and semi-rural English local authority. It is divided into seven District Councils (DC) and collaborates with these latter, as well as 19 other local organisations (public, private and voluntary), via the medium of its LSP. This locality was selected because it was formulating an OC project, Integrated Commissioning 2012 (IC 2012), in a tumultuous context. Furthermore, Internshire portrayed itself and was nationally perceived as 'successful' in its past reforms. This was a locality that had transformed itself, from one of the most politically unstable in the country in the 1980s and 1990s, to 'four-star ' andexcellent authority under New Labour's (1997-2010) performance regime. Data were collected from three types of sources. First, semi-structured interviews with 33 key players were conducted between November 2011 and May 2013. Second, between September 2011 and April 2012, observations of participants in the organisation were compiled into field notes. Third, documents spanning 40 years of OC from 1974 to 2013 were systematically compiled into a documentary archive.
Understanding this project and how leadership was renegotiated requires problematising their murky origins (step 1 of the logics approach). From the mid-1970s, this authority's nascent corporate centre, spearheaded by corporate managers and later the Chief Executive's Department (CED), mobilised shifting national and local demands (e.g. government reforms, economic recession, local political instability) as dislocatory conditions requiring change. Yet, between 1974 and 2010, similar solutions of corporate planning, performance management, centralisation and unification were proposed and accepted. This hegemonisation did not however take place without conflict. Grievances from councillors, DC, other partners and even County officers demanded more participation and equality. These were however successively muted via strategies of corporate training, the creation of multiple tasked groups and chairs, or simply being excluded. Thus, progressively, and despite changing governments, shifting economic conditions or new organisational demands, this corporate centre became and remained hegemonic.
From 2010, events such as the financial crisis, strenuous austerity measures and new government agendas of localism and commissioning became mobilised locally by these old grievances, challenging 'taken-for-granted' corporate practices of change and policymaking. The 'role of the State at the local level' had been 'fundamentally chang[ed]' (CED1 manager), leaving Internshire Together 'wondering […] what to do now' (DC ChiefExec1; ICC Councillors 1 and 2). A representative from a partner organisation also believed that policies such as 'the localism agenda' offered organisations outside the County Council the opportunity to have 'more control […] over funding and priorities', stressing that that dispersal would 'be a benefit to the locality' (LSP Officer1). Councillors too believed that 'the job […] of a councillor now [was] changing', becoming one of 'understand[ing] different agencies' and 'bringing everything together for those people out there' (ICC Councillor 2 and 3). This new context of freedom for local players to decide their 'own destiny' (CED3 manager) was framed by corporate managers as creating complexities (CED4 manager), 'local authorities were now left with their partners to actually think "what are we trying to do?"' (CED5 manager). Localism was disputed, some interviewees arguing that Districts were 'in a better place than County in terms of relationship within the localities' because they were 'far more local' (DC ChiefExec1).
From 2010, this corporate centre formulated a new project for the County Council and the partnership. Entitled 'Integrated Commissioning 2012' (IC 2012), this project framed conditions such as austerity, localism or partnership disputes as dislocations or 'pressures' requiring two 'solutions': the integration of partnerships and priorities, and the move to the vaguely defined 'commissioning' of services, which echoed demands of collaboration, local delivery and privatisation. It was argued at the time that this project would help Internshire Together remain an excellent locality, working better in 'hard' rather than 'soft' partnerships and achieving 'more with less' resources. Partners were for instance asked to draft 'commissioning plans', or planning documents outlining which three or four priorities they would work towards, how they planned to achieve them and how they would collaborate with other partnerships. Although the initial principles of IC 2012 were adopted by all partners in October 2011, by the summer of 2012, the project was a failure: few plans had been accepted and voices across Internshire were claiming it was 'business as usual'. Despite this failure however, the corporate centre, and in particular the Chief Executive, was successful in renewing its control over the organisation and the LSP. Understanding this paradox requires examining how and why leadership practices were renegotiated during this phase, mobilising the dimensions outlined supra.
The corporate centre had succeeded in hegemonising County leadership practices during the 1990s to 2000s, particularly with the then Chief Executives (CEO 2, 1976(CEO 2, -1991CEO 3, 1991-1994and CEO 4, 1994 linking together an increasing array of demands from officers and councillors via strategies of training and specialised groups. Similar strategies were deployed within Internshire Together. First, the corporate centre continued to widen the meaning of leadership (Dimension 1). For instance, alliances with specific individuals were built by repositioning them as 'strategic leaders'. The Leaders of Tomorrow programme selected 26 managers 'to ensure our future senior officers are well placed to work together in leading public service improvements', leading to what was hailed as 'a new breed of public sector manager' (CEO 4 writing in Source 16, 2009: 6). Leadership included diverse practices of collaboration, management, policy formulation in tasked groups, commissioning and waste reduction. Second, the centre reinforced the link between leadership and excellence. One particular project, Total Place, illustrates this well. Total Place included a series of pilots conducted in a dozen of English localities in 2009 to consider issues of savings, notably via the pooling of resources, and delivering better services, especially for complex issues such as social deprivation. Internshire was selected to pilot Total Place. The Chief Executive was instrumental in framing the project as also one of leadership: The aim is to make ours the best place in which to live and work -everything we do revolves around this ambition. However, this is no easy challenge, and requires a focus and a drive across a partnership covering city and county. There is certainly a leadership role for local government but the key is taking our partners with us. The old ways of working in the public sector have to change and we are determined to be in the forefront, showing how it can be done and how citizens and taxpayers benefit. (CEO 4, quoted in Source 18, 2010: 7) In a context where the old strategic partnership model was looking exhausted and mobilised growing criticisms locally, Total Place was here framed by the Chief Executive as a means of resuscitating the purpose of partnership working. What was proposed was a 'single offer' model of drugs and alcohol misuse services, the specific project piloted (CEO 4, in Source 12, 2010: 14). In this process, leadership was mobilised as equivalent to collaboration between partners in achieving excellence, the key 'ambition' of Internshire. Total Place was replacing strategic partnership working with the notion of 'places' (Source 12, 2010). By extension, the corporate centre was reframing itself as a vanguardist in what was pitched as an ineluctable evolution. The Chief Executive argued that '[t]he Government [was] looking for strong leadership from local authorities to take Total Place forward.
[…] Local government […] being offered a change it must grasp' (Source 12, 2010). With Total Place, the Chief Executive was identifying 'the' key project for Internshire to continue excelling.
Total Place was also mobilised by the Chief Executive to renegotiate his own subject position, becoming the flagship of Internshire at the national level, 'showcasing' the locality's excellence when negotiating with Government (Dimension 2): When it switched towards Total Place, it was on the back of trying to think that the emphasis should be on localities and outcomes. And again, [Intern]shire always trying to be piloting nationally, always trying to be out there, trying to put influences in and given credit for that. They put a lot of time and effort into Whitehall.
[…] clearly, [ex-County Leader's first name] had aspirations with the Local Government Association and also [CEO 4]. [CEO 4] wants to be seen nationally … because obviously now he's a seasoned chief exec' with a huge amount of experience. (DC ChiefExec2) Further illustrating the militating position of the Chief Executive, in November 2009, he reported to the partnership Executive that he 'detected a significant shift in thinking in Whitehall' and that Internshire's Total Place pilot was 'well positioned to influence national thinking' (IT Executive, 2009: 4), another illustration of Dimension 2. Internshire could be playing an important role nationally, soon reaching its goal of influencing local issues on a national scale (e.g. ICC Councillor4). By March 2010, he was explaining in a widely read specialised local government journal that 'government had listened' to local demands such as 'making greater use of pooled budgeting' (Source 12, 2010: 14). Thus, alongside the leadership positions offered to all partnership organisations, leadership also represented the particular demands of the Chief Executive, actively articulating his position, framing projects locally, negotiating with Government and symbolising the changing aspirations of this locality.
The IC 2012 project failed. Although several conditions can be outlined in explaining this failure (Author, 2014), this article focuses on a particular one: the definition and practices of leadership. IC 2012's failure can be explained by the Chief Executive renegotiating his subject position as 'leader', distancing himself from IC 2012 to identify with new demands such as localism and voluntary partnership. Crucially, what was framed at the time by some corporate managers as 'the decision' of the Chief Executive not to push further the project is key in differentiating the failure of IC 2012 from that of the corporate management regime as a whole and thus in reaching a more complex understanding. Here, the Chief Executive played a key role in severing IC 2012 from the regime. Indeed, here the four dimensions, rather than being 'discovered' are iteratively constructed in situ to make sense of a puzzle: why did IC 2012 fail but the wider corporate management discourse appeared to continue to exercise power, albeit in new forms and echoing new demands? Once again, the critical role of the researcher is important here in juggling between theory and case study so as to devise the best possible explanation (MacKillop, 2016).
The link between the Chief Executive and the success of IC 2012 was already discussed by corporate managers during interviews in February and March 2012, one of them explaining that 'if there [was] a major push' then the project would work, illustrating how this was the case for the Sports and Physical Activity commissioning plan, the exemplar commissioning project having benefited from 'a particular political agenda pushing it forward' (CED6 manager). Another one explained in November 2011 that the Assistant Chief Executive 'work[ed] very closely with [CEO 4]', this latter 'championing it without being involved in the detail' and 'very much leading the process' (CED7 manager). He was reported by a County lead member to be leading conversations with health (ICC Councillor1). A corporate presentation to the Corporate Management Team in October 2012 confirmed this important and multifaceted role played by the Chief Executive, stating that '[m]ost Integrated Commissioning progress has been made where there has been strong intervention from the Chief Executive or central teams, and where external factors (money, structural change) have forced a different approach' (ICC, 2012). Thus, in a first instance, the Chief Executive's support of the project was considered by local participants as crucial in determining the success or failure of the project.
Follow-up interviews from September 2012 illustrated how the Chief Executive was slowly withdrawing his support from IC 2012, severing that link between his position and this particular project (Dimension 4). This active role of the Chief Executive in renegotiating the regime's demands is illustrated by several examples. For instance, one corporate manager interviewed in September 2012 was particularly frustrated, explaining that he had highlighted to the 'Chief Executive and others' that there was 'some good progress' but stressed that the Chief Executive 'could get them to do it' (CED1 manager). This manager reported that 'the answer [from the Chief Executive] [had] been […] "let's not panic, this is progress"' (ICC, 2012). This manager and others illustrated the key role played by the Chief Executive in framing this project as synonymous, or not, with himself and the locality. It also illustrated how far he was ready to go in convincing or demanding consent for this new project. A key article was about to be presented by corporate managers to the Chief Executive, the future of the project being framed by one corporate manager as depending on whether this article would have to be 'rip[ped] up' or not following the meeting (CED2 manager). This further illustrated this individual's influence in determining the future of this project. ' Alongside these redefinitions of commissioning, the Chief Executive became associated with other projects of reform (Dimension 4). These illustrated the new negotiated approach to policymaking in Internshire, addressing some of the grievances of Districts and other players and previous understandings of leadership. The new negotiated attempts at change were seen as a success in collaboration across Internshire Together, highlighting how the corporate centre was addressing some of the old grievances linked to the lack of equality and negotiation (CED1 and 2 managers).
As corroborated throughout the genealogy of this project, the multiplicity and diversity of practices and projects associated with this particular Chief Executive illustrated how his leadership position was constantly being renegotiated. Testimonies of local players account for this, illustrating in the 1980s and 1990s how CEO 4 deployed an organic and collaborative approach to the training of officers and councillors, but also how he became the 'apex' of change when important reforms such as Total Quality Management were implemented. Interviewees argued that CEO 4 did not subscribe to the idea of telling others what commissioning meant (CED1 manager), but on some occasions however, he clearly 'pushed' particular items on the agenda to improve the chances of this locality being recognised nationally (DC ChiefExecs 2 and 3). Furthermore, despite drawing frontiers between this locality and its neighbours and the Government, this Chief Executive personally negotiated with these 'enemies' on several occasions (e.g. Multi-Area Agreement, Health and Wellbeing Board). These multiple and contradictory practices of leadership, with the Chief Executive constantly redefining what his position and the organisation at large was about, illustrate the situated, multifaceted and iterative dimensions of leadership in Internshire.
Thus, although the IC 2012 project failed, the Chief Executive renegotiated his position of leader and the overarching leadership practices embodying this locality. He did so by identifying with other demands such as locality management and third sector commissioning, and thus redefining his embodiment of the organisation and by extension what this locality aspired to.

Conclusion
This article has set out some proposals for discursively analysing and critically explaining OC leadership, building on political discourse theory and CMS. The discursive and logics approach is applied throughout the article, from the manner in which the problem is formulated (leadership as discursive practices articulated to exercise power), how it is examined (e.g. formulation of several problematisations such as what was the discourse hegemonising this organisation? How did it deal with resistance? How did it rearticulate or reimagine itself from 2010 and build new alliance via articulating new demands in new ways?), to the leadership dimensions proposed to critically analyse the case of Internshire (how may we understand how a given OC project failed and yet an established discourse/discursive regime survived? By formulating different dimensions of leadership). Thus, the article demonstrates what a discursive approach to leadership research can look like, from theoretical framing to the articulation of empirical data and critical analysis. Nevertheless, each study will necessarily be different and specific to each researcher and their background and values, making for interesting new findings.
Building on CLS which have emphasised the political, contextual, complex and ideological dimension of leadership in OC, two proposals were made. It was argued that leadership should be understood as a set of changing and situated discursive practices. It was also suggested that leadership should be problematised within given sites, articulating discursive concepts of empty signifier, floating signifier, subjectivity and agency, to expose its diverse mobilisation in renewing and negotiating consent. This dynamic understanding of the articulation of leadership opens up the possibility for a situated critique of leadership practices as well as developing a discursive understanding of the role of individuals in such practices, thus elaborating on some of the comments raised in the literature Rhodes, 2004, 2006;Fairclough, 2005). By revisiting the concept of leadership in organisational studies, this article has sought to understand how particular leadership practices constitute, transform and sometimes fail OC discourses. General and static definitions of leadership are unhelpful in grappling with the diversity of practices linked to leadership in a given organisation. Building on the critical literature of leadership in change and organisations at large, dimensions of leadership practices were formulated retroductively and in situ, which allowed analysing how leadership brings forth issues of ideology, politics, subjectivity and agency. With the help of Laclauian discourse theory and a logics approach, change and its leadership can be analysed as the constant mobilisation of strategies of inclusion and exclusion (political logics), drawing up of rules, values and norms (social logics) and mobilising individual fears and desires (fantasmatic logics) to hegemonise those spaces and manage grievances. This standpoint opens up the possibility of analysing organisations, change and leadership as non-necessary, dislocated and political enterprises (Cederström and Willmott, 2007;Parker, 2002;Parker and Dent, 1996). Furthermore, this study and its dimensions have emphasised the role and subjectivity of the researcher in formulating tools to explain empirical puzzles such as why do OC projects fail.
The discursive framework proposed has added a layer of understanding to how and why change is formulated and implemented in a given organisation. Where some approaches may entertain an artificial 'disconnect' between theory and empirical data, this research has offered a more flowing argument, where methodology and data form part of a single, to-ing and fro-ing explanatory movement. With the four dimensions, I add to current understandings of leadership by further interrogating and reworking what leadership may represent in a given site and how 'it' is pragmatically articulated by organisational discourses in addressing conflict and resistance. Leadership can coincide with a multiplicity of practices. As such, one should be prepared to articulate and interrogate current theoretical categories in order to explain as fully as possible leadership in a given context. I believe that these dimensions of leadership have helped conceptualise further how change discourses gather consent and which role individuals may play in such processes, without however adopting a deterministic approach (Cederström and Willmott, 2007: 2). Particularly, the second and fourth leadership dimensions discursively address the role of individuals in dislocatory contexts, deploying concepts of empty and floating signifiers in doing so (Howarth, 2013: 272-273).
Contrary to studies framing individuals as either free and powerful agents, or as empty shells (Badham et al., 2003;Driver, 2009), a more complex understanding of individuals can be achieved.
In summary, this article has demonstrated that leadership can be understood as a constantly changing performance of organisational power, addressing grievances by mobilising different practices under the appealing and hollowed-out demand of leadership. Furthermore, following Spicer and Alvesson's (2011) recommendations, research in leadership should avoid normatively discussing the benefits or 'dark' side of leadership (Conger, 1990). Instead, as illustrated by the four leadership dimensions devised here, leadership should be analysed as a set of discursive practices and studied to make sense of the intricacies and pragmatism of power plays deployed across organisations (although this performativity should remain questioned; cf. Parker, 2014). It is by articulating critical methodologies such as discourse theory, combined with a logics approach and in-depth case studies, that CLS can continue their enquiry into power in organisations.