A dialectical approach to the politics of learning in a major city police organization

In this paper we develop a dialectical approach to the organizational politics of learning, exploring complexity, tensions and asymmetries. Turning this kaleidoscopic lens on our empirical setting, a major city police organization, we mix the blue light of police vehicles into Driver’s (2002) ‘fluorescent’ light of office workplaces, fragmenting the brightness of ‘Utopian sunshine’ and the darkness of ‘Foucauldian gloom’ perspectives on organizational learning, and making visible a wider spectrum of political colours of learning. We identify four interdependent political modalities of learning: empowering, coercive, insurgent and palliative and explore how they interplay in complex and contradictory ways. We note that, whilst mainstream and critical literatures tend to focus on organizational learning as, respectively, empowering and coercive, and to a lesser extent insurgent, much of the politics of learning in our study converges in the palliative modality, where the emphasis is on learning-to-cope (rather than learning-to-thrive, learning-to-comply or learning-to-resist). We show that the palliative modality of learning is in many ways an outcome of the dynamic and complex engagement between the other three modalities. We discuss the implications of our findings for a more nuanced understanding of learning as political, and of the relationship between organizational learning and power.


Introduction
In this paper we explore tensions and dynamic complexities of the politics of learning in a large public service organization. Our theoretical aim is to contribute a more nuanced understanding of the relationship between learning and power (Collien, 2018;Contu and Willmott, 2003;Harman, 2012) by developing a dialectical approach to the organizational politics of learning. We contribute, in particular, to the strand of work in which attempts have been made to move beyond the simple political dichotomization of organizational culture of learning as constitutive of either 'Utopian sunshine' or 'Foucauldian gloom' (Coopey, 1998;Driver, 2002). Proponents of the first of these contrasting perspectives (Utopian sunshine) see organizational learning as a path to empowered agency freeing employees from the shackles of structural power and top-down power flows characteristic of a traditional bureaucracy. Proponents of the second perspective (Foucauldian gloom) understand organizational learning culture itself as a form of ideological power -a 'workplace nightmare for employees, in which they are exploited in even more devious ways. . . to carry their organizations to competitive success or simply serve the interests of those in power' (Driver, 2002: 34). We argue that a dialectical approach, in bringing together elements from each perspective, can help foster the development of an organizational learning theory that more fully reflects the richness and complexity of organizational life, as Driver (2002: 34) advocates.
Splitting into sides that privilege opposing elements of binary divides (between, for example, power and resistance, compliance and resistance, free agency and structural determinism) is a common feature in analyses of power, and dialectical approaches have been used elsewhere in management and organization studies to address this issue and explore the richness of power relations (e.g. Benson, 1977;Bristow et al., 2017;Collinson, 2005;Mumby, 2005;Ybema et al., 2019). We bring the dialectical focus on complexity, tensions and contradictions in power dynamics to the analysis of organizational learning. We do so by starting from the view of learning practices as embedded in power relations that enable, facilitate or obstruct them (Contu and Willmott, 2003;Coopey and Burgoyne, 2000) to approach learning in organizations as inherently and inevitably political. We develop this into a dialectical perspective that enables us to bring into view the complex and contingent nature of the politics of learning, and the ways in which different political modalities of learning interplay with each other in multiple and often contradictory ways.
The empirical context of our research enables us to do the above in several ways. Firstly, our chosen organization is in the midst of what they regard as a major transformation, in which centrally-planned change towards a learning culture is a key element. As Mastio and Dovey (2019: 800) point out, turbulent times are an important source of intra-organizational tensions, as 'the power interests which the organizational arrangements serve become demystified'. In our case, by challenging established norms and conventions, the rhetorics and processes of change expose everyday aspects of learning and power relations that would otherwise be more taken-for-granted, making it easier to recognize and trace them. Moreover, we recognize that change is itself constituted through power-resistance relations (Thomas and Hardy, 2011), which in our case grapple explicitly as well as implicitly over the role and practices of learning, adding layers of 'dynamic complexity' (Guiette and Vandenbempt, 2016) to the analysis of the politics of learning. Learning and change are also inextricably linked, learning being the means and the product of change, and change necessitating and exacerbating learning (Bosma et al., 2016).
Secondly, we focus on a police organization with the associated political complexities typical of public service organizations. The latter tend to operate in rapidly changing and highly complex contexts, which require them to continuously respond to multiple and often conflicting stakeholder demands, values and expectations (Finger and Brand, 1999;Hartley, 2006;Hartley and Skelcher, 2008;Hartley et al., 2017;Hoggett, 2006). Moreover, police in particular being an important part of the disciplinary machinery of the state, which often takes the shape of coercive control (Bowling et al., 2019;Das and Marenin, 2000;Marenin, 1985), enables us to explore some of the added tensions and paradoxes of learning and power in such disciplinary contexts.
Thirdly, this paper is part of a broader engaged scholarship (Cunliffe and Scaratti, 2017;Van de Ven, 2007) project, which means that, as researchers, we are not only deeply embedded but also actively implicated in the production of the politics of learning of which we write. This makes us acutely aware of the multiplicity of contingent and contradictory ways in which learning, including our own, is inherently political. Tracing some of these shifting political modalities in the context of dynamic complexity then becomes a matter of asking why, to what ends, how, for whose benefit and for what kinds of benefit learning takes place, what consequences it has for different groups of people, and what implications it has for the power relations within which it takes place. In examining these issues through a kaleidoscope-like dialectical lens, we mix the blue flashing light of police vehicles into Driver's (2002) 'fluorescent' light of office workplaces, the brightness of 'utopian sunshine' and the 'Foucauldian gloom' perspectives, making a wider spectrum of political shades and colours of learning and their consequences visible. This enables us to address the following questions: (1) what political modalities of learning are at play in our empirical case and what tensions and asymmetries emerge through their interplay, and (2) what implications does this configuration of modalities have for understanding and theorizing the relationship between learning and power, in the context of a policing organization and more generally?
Below, we first discuss our theoretical perspective on the politics of learning and apply it to existing literature on organizational learning and power. Next, we consider the empirical context of our research, followed by a discussion of our methods. We then proceed to the analysis and discussion of our findings and conclude the paper with a summary of our contributions and their implications.

A dialectical approach to organizational politics of learning
We view organizational learning as part of organizational dynamics that comprise individual, group and organizational phenomena (Crossan et al., 1999;Vince 2001), and of which power and politics are a pervasive and endemic feature (Lawrence et al., 2005;Vince, 2001). From this perspective, organizational learning processes are always embedded in power relations, from which they cannot be separated, and which can enable, facilitate or obstruct them in a variety of ways (Contu and Willmott, 2003;Coopey and Burgoyne, 2000). Power and politics thus cannot be just assumed to be a problem that gets in the way of learning (Argyris, 1990;Senge, 1990), rather they can be thought of as productive and constitutive of learning, as well as disciplinary in relation to it, in a Foucauldian sense (Foucault, 1991). Our argument is that the embeddedness of organizational learning in power relations means that it makes sense to think of learning itself as inherently and inevitably political, and that this makes it important, for the purposes of developing a nuanced understanding of organizational learning processes, to critically and reflexively consider the politics of learning and their implications. By organizational politics of learning we mean dynamic configurations of political modalities, comprising the ends to which and the means through which the power of organizational learning discourses and practices is operationalized, exercised, contested and transformed.
Conceptualizing organizational politics of learning in the above way acknowledges their embeddedness in broader discourses and structures, their pervasiveness, multiplicity, and multifaceted and dynamic character. It thus lays down the basis for a dialectical exploration of their complexity, tensions and contradictions. The dialectical tradition has a long history of diverse thinkers including Hegel, Marx, Horkheimer, Adorno, Lukács, Gramsci, Bakhtin and Giddens, and is based on the notion of social order as 'premised on tensions and contradictions that underlie apparent cohesion and that point to potential social change and transformation' (Mumby, 2005: 22). Our own understanding of dialectics draws in particular on Adorno's (1973) negative dialectics and Benson's (1977) dialectical view of organizations, especially as developed in relation to organizational power-resistance relations by Mumby (2005). A key feature of this dialectical approach is the refusal to seek Hegelian transcendence or grand synthesis as the resolution to contradictions, choosing instead 'the more difficult path of keeping tensions and contradictions in constant play' (Mumby, 2005: 22). Mumby treads such a path in relation to the control-resistance dualism, which, he argues, has tended to characterize organizational studies of resistance and power. The literature, he contends, has been split into two camps, one privileging the role of hegemonic, ideological and normative power of organizational cultures and structures that reduces individual agency to merely reproducing managerially-defined realities, and the other privileging the free, active and creative agency of subjects 'operating from a pristine, authentic space of resistance' (Mumby, 2005: 37). Instead, Mumby (2005: 38) proposes a dialectical study in which 'social actors are neither romanticized nor viewed as unwitting dupes but rather are seen as' always engaging in context-specific, ongoing, tension-filled struggles over meaning that often draw 'on larger discourses that are articulated in ways that may be coherent and contradictory'. Collinson (2005Collinson ( : 1422 uses a similar approach to question persistent dualisms in leadership research (including one between leaders as powerful subjects and followers as passive objects) and instead explore coexisting, interrelated dialectics of leadership that 'are mutually reproducing features' of 'the complex, shifting dynamics of leadership'. Bristow et al. (2017) and Ybema et al. (2019) both draw on Mumby (2005) -the former to explore tensions, contradictions, overlaps and alliances of academic resistance and compliance 'in the cauldron of forces' that constitute the UK Higher Education (Bristow et al., 2017(Bristow et al., : 1187, and the latter to challenge the dichotomy between 'change agents' and 'change recipients' by examining the complexity of individual change positionings in a police organization. In this paper we develop a dialectical approach to organizational politics of learning as a means of crafting a more nuanced dialogue between the two sides of the 'Utopian sunshine' and the 'Foucauldian gloom' dualism (Coopey, 1998;Driver, 2002). This involves moving away from 'either-or' thinking about the binary opposites and instead embracing a 'both-and' orientation (Collinson, 2005;Mumby, 2005). This orientation, 'in which the individual and the system are each viewed as constitutive elements' (Collinson, 2005(Collinson, : 1420, emphasizes the complex, contingent, interrelated and dynamic nature of contradictory aspects of organizational politics, and focuses on exploring the ways in which they struggle with each other but also mutually constitute each other (Bristow et al., 2017;Collinson, 2005;Mumby, 2005;Ybema et al., 2019). Accordingly, our interest in this paper is not only in identifying different political modalities of learning simultaneously at work in our chosen organizational setting, but more significantly also in exploring the interplay, tensions, alliances and asymmetries between them within the contradictory and changing context of UK policing.
In terms of identifying political modalities, Driver (2002) helps to distil at least three major recurring political modalities from the existing literature on organizational learning, and think of them in terms of the different 'shades' of politics -'Utopian sunshine', 'Foucauldian gloom' and the middle-ground position Driver proposes as 'fluorescent light'. The first of these is prevalent in the foundational writings on organizational learning, such as those of Senge (1990), that have set the tone and impetus for much of the mainstream academic work and organizational programmes of change. In such work learning tends to be treated as a kind of utopian liberating force freeing individuals from bureaucratic structural and procedural constraints (Driver, 2002) with a capacity to transform organizations into progressive and egalitarian places (Dixon, 1998). In this utopia, empowered employees pursue personal fulfilment and transcendental humanistic values (Kofman and Senge, 1995), with minimal formal managerial control (Goh, 1998) and through highest levels of participation in the shaping of organizational agendas (Bhatt, 2000). This vision evokes highly enticing imagery of the 'positive ideal' (Driver, 2002) of organizational learning operating in what can be thought of as a learning-to-thrive mode for everyone. Learning is presented as a generative, transformative and empowering organizational force (Dixon, 1994;Senge, 1990) that can evolve societies as well as human consciousness to the next level (Gozdz, 2000). To put it differently, this view of learning is evocative of an empowering politics of learning that is positively transformative of organizational power relations and constitutive of progressive and emancipatory organizational change.
By contrast, within much of critical scholarship the politics of change towards an organizational culture of learning are typically presented in much darker tones, which Coopey (1998) and Driver (2002) have dubbed 'Foucauldian gloom'. Critics have argued that the utopian rhetoric of introducing a learning culture tends to result, in practice, in rather dystopian realities for employees, who are manipulated to work harder without the extra rewards because they are misled to believe that to do so is in their own best interests (Easterby-Smith, 1997). (This argument is also reminiscent of the effects of Lukes' (2005) abstract forms of power in producing 'false consciousness' among the less powerful.) From this perspective, rather than acting as an instrument of empowerment, a centrally-imposed organizational culture of learning becomes a form of ideological power and normative control through which employees are exploited by their employers -a kind of 'weapon of the powerful few against the powerless many' (Driver, 2002: 48). Instead of becoming egalitarian places of distributed power, organizations driven by the imperative to learn can instead turn into psychic prisons, in which employees are subtly coerced to learn how to more effectively achieve organizational goals without being allowed to question them or their underlying assumptions (Schein, 1999). By contrast to its utopian counterpart, this critical, dystopian vision calls up the imagery of learning dominated by the interests of organizational elites (Coopey, 1995;Easterby-Smith, 1997, Snell andChak, 1998) and therefore operating in what can be called a learning-tocomply mode in relation to its subjects. Rather than being generative, positive and transformative, learning is tightly controlled and performative  because it is limited to what is acceptable and not disruptive of organizational norms and conventions (Coopey, 1995;Field, 2011;Schein, 1999). In other words, this perspective is evocative of a coercive politics of learning that reproduces rather than challenges existing power relations and power asymmetries by aligning with dominant vectors of power in the processes of organizational change.
In response to the above polarized, dichromatic views of learning, Driver (2002) has called for more dialogue between the two perspectives and has attempted to synthesize a middle-ground position that would incorporate elements of organizational learning hopes and critiques into what she calls a 'fluorescent light' (referencing office workplaces) view. Such an approach arguably enables a more nuanced analysis of the politics of learning capable of considering how contrasting political modalities of learning interact within specific organizational contexts. One outcome of this is the increased visibility of the bottom-up participatory learning agency that can be resistant as well as compliant in relation to the top-down organizational imperatives and structural and ideological forms of power. For example, Harman (2012: 275) takes a Foucauldian perspective to analyse everyday learning of a manager, but instead of painting a 'monolithic view of power' through which workplace learners are passively acted upon in a top-down manner, 'provides the analytic space for re-presenting workplace learners as active in the ongoing negotiation' of their identities. Taylor et al. (2010) explore the power, politics and resistance surrounding the rise and fall of a work-based learning initiative, a university for the UK National Health Service. Collin et al. (2011) consider the relationship between learning and the subtle ways in which power is exercised and resisted in a hospital operating theatre. Huzzard (2000) uses organizational learning as a means to analyse labour relations at Ericsson Infocom from a bottom-up perspective, teasing out the relationship between learning and (union-based) resistance.
This strand of scholarship thus opens up a way for imagining a political modality of learning that is much less dystopian than the coercive modality and yet politically distinct from its empowering counterpart. Here, learning can be thought of as operating in a learning-to-resist mode in relation to the dominant vectors of power (which can be oriented in a number of ways in relation to organizational change (Thomas and Hardy, 2011)). Drawing on a view of resistance as a form of power that opposes, subverts and reconfigures other organizational forces (Bristow et al., 2017;Fleming and Spicer, 2007) and that therefore can be not just disruptive, but also generative (Thomas and Davies, 2005), transformative (Thomas and Hardy, 2011) and productive (Courpasson et al., 2012), allows for a wide spectrum of images of learning in this mode. For instance, it can manifest as more mundane and micro-level, such as an ongoing search for ways to dismantle dominance in organizational learning processes (Collien, 2018), or more radical and revolutionary (Griffin et al., 2015;Örtenblad, 2002), invoking utopias of its own kind, such as a radical imagining of an 'unmanaged organization' (Izak, 2015). Such utopias have a different political undercurrent to the empowerment utopias typically found in mainstream literature. Whereas mainstream notions of empowerment involve a delegation, a giving of control, which implies a top-down initiative and benevolent action to the end result of more democracy and equality (Duvall, 1999), critical utopias are more about the taking of control implying bottom-up initiative and action. This strand of scholarship therefore evokes an insurgent politics of learning that is more revolutionary than devolutionary in relation to existing power relations.
In this paper we contribute to the above literature by further exploring the complexities of organizational politics of learning. We will show through our dialectical analysis how all three of the political modalities reviewed above are at work simultaneously in our police organization, how their interplay and struggle with each other in the context of broader power relations creates tensions and asymmetries, and how these lead to a fourth political modality we call 'palliative'. In this way, our kaleidoscope-like dialectical lens mixes the blue flashing light of police vehicles into Driver's (2002) 'fluorescent' light of office workplaces, the brightness of 'Utopian sunshine' and the 'Foucauldian gloom' perspectives, making a wider spectrum of political colours of learning and their consequences visible. Below, we start this work by discussing the specific nature of our chosen organizational setting and how, together with our research approach, it has shaped our thinking about organizational politics of learning, before proceeding to the analysis of our findings.

The contested public services and the police conundrum
There is persisting over-reliance on private sector organizations for theorizing and empirical research into organizational learning and, as critical commentators have argued, this can be conceptually limiting (Olejarski et al., 2019;Rashman et al., 2009). Organizational learning plays a major role in public sector organizations, and in turn the latter constitute an important, distinctive context for the study of the former (Rashman et al., 2009). This distinctiveness includes differences in purpose, catalysts, motivations and key actors between private and public sectors (Hartley, 2005;Rashman et al., 2009). One central difference is said to be that rather than being driven by profit as the ultimate bottom line, the purpose of public sector organizations is to maximize public value, which involves an intricate balancing of multiple and competing stakeholder interests (Benington and Moore, 2011;O'Flynn, 2007;Rashman et al., 2009). This includes the interests of politicians, who have formal control over public services, attracting their close scrutiny, as well as the interests of and scrutiny by the broader public to whom public sector organizations are ultimately accountable (Hartley and Skelcher, 2008).
Another distinguishing feature of public sector organizations is therefore their inherent political complexity (Hartley and Skelcher, 2008;Hoggett, 2006). This goes hand-in-hand with exposure to rapid and relentless change imposed on the sector by shifting government policies as a result of broader political, economic and social changes (Brodtrick, 1998;Rashman et al., 2009). Such externally-imposed changes periodically introduce new paradigms of public management, demanding shifts (such as from 'traditional' public administration to new public management, and more recently to 'networked' or 'citizen-centered' governance) in structures, cultures and thinking about the very purpose of public sector organizations (Hartley, 2006). However, rather than completely replacing each other, these paradigms end up coexisting as competing layers of societal and organizational politics 'with particular circumstances or contexts calling forth behaviours and decisions related to one or the other conception of governance and service delivery' (Hartley, 2006: 29). However great the pressures and the operational challenges generated by such dynamic complexity, public service organizations cannot simply choose to cease to exist in the same way as would be possible in the private sector. They are bound by the mandate to provide certain products and services (Brodtrick, 1998) and so have to continue in the grip of the contradictions they face. In practice, this means that their struggles with their publicly and politically contested purpose and values are translated into several arenas, including what can be thought of as intensive on-theground politics. This is often referred to as 'street-level bureaucracy' (Lipsky, 2010), in which public sector employees engage on the daily basis 'to contain the unresolved (and often partially suppressed) value conflicts and moral ambivalence of society' (Hoggett, 2006: 175). Public sector organizations can therefore be said to be saturated with both macro-and micro-political processes in ways that are arguably more convoluted and ambivalent and yet also more externally visible and scrutinized than private sector organizations.
In the current times of austerity, public sector organizations are increasingly required to do more with less, which intensifies the imperative for learning (Olejarski et al., 2019) and, as in other challenging times, increases political scrutiny and demands for reform (Dekker and Hansen, 2004). Given also that the complexity of organizational change exacerbates organizational politics (Buchanan and Badham, 2008), there is an ongoing need to set public sector research within its political context (Ferlie et al., 2003) and develop organizational learning frameworks that can account more fully for the changing, complex and contested nature of the public services (Rashman et al., 2009). For those who, like us, want to explore the dialectics and dynamic complexity of the politics of organizational learning, public sector organizations are therefore a good place to start.
Police organizations share all the characteristics of public sector organizations discussed above but are also distinct in a number of ways, which means that they face their own additional specific complexities and challenges. Police are 'inherently and inescapably political' as an institution concerned with the organization of coercion and consent (Bowling et al, 2019:15), the rise of modern policing historically entwined with the rise of the modern state and its disciplinary apparatus (Crawford, 2012;Foucault, 1991;Johnson, 2014). In democratic contexts, contemporary police forces remain key actors in the political economy of modern state ruling, within which they act as a channelling and redistributive mechanism for the changing configurations of diverse interests and agendas (Bowling et al., 2019). Police are a state agency disciplined by the powerful, but also part of the broader political processes and conflicts in which they can stake their own claims and on which they can leave their own mark through the exercise of power and politics that policing work entails (Bowling et al., 2019;Martin and Bradford, 2019). This includes their potential to actively shape elements of their political environment and systems of government within which they operate, ranging from the making and maintenance of democracies (Aitchison and Blaustein, 2013;Bayley, 1969) to campaigning on police resources and priorities (Bowling et al., 2019) and resisting government-imposed reforms (Leishman et al., 1995). In contemporary democratic contexts, policing therefore involves juggling a 'dynamic set of balances' between competing and often conflicting interests and values (Marenin, 2000). These include the fundamental dilemma between obedience to law and responsiveness to community demands (Bradford et al., 2013), which is unresolvable on a permanent basis and is subject to temporary compromises led by shifting government and organizational policy. This means that 'every policing arrangement is problematic. . . and must balance coercion and consent, autonomy and responsiveness (Brogden, 1982), the desire for general order versus protection of particular interests (Marenin, 1982) and the need for control of behaviour and the desire of citizens to be left alone' (Marenin, 2000: 312). No form of policing fundamentally solves the conflicts of interests between the state, police and all segments of the public, as each swing of policy disadvantages some groups in relation to others. Moreover, to tackle these 'value conflicts and moral ambivalence' as 'street-level bureaucrats' (Hoggett, 2006;Lipsky, 2010), police officers are vested by the state with specific powers. These include the use of significant symbolic, coercive and sometimes deadly force, 'mandated as [police] are with the value laden task of maintaining order in an ever-changing society' (Martin and Bradford, 2019:7).
Viewed through a dialectical lens that brings together the three political modalities of learning (empowering, coercive and insurgent) we discussed in the previous section, police organizations present an interesting conundrum. On the one hand, one could say that police officers need to be empowered to learn in order to effectively partake in the unpredictable street-level politics rather than simply (and blindly) learning-to-comply with the letter of the law; that empowering learningto-thrive for all is exactly what should happen in situations in which lives may be at stake. On the other hand, unlimited police empowerment in the context in which lives are often at stake is, for many, the stuff of nightmares rather than dreams. From this perspective, police are already sufficiently -or perhaps even too -empowered simply by means of being police, and emphasis should necessarily be placed on controlling their influence through learning-to-comply, which, though coercive, could act to limit coercion elsewhere. This dilemma places immediate constraints on the extent to which a police force can become a 'learning organization' enjoying the empowering 'Utopian sunshine'. It also brightens up the 'Foucauldian gloom' perspective with much more ambivalent light, in which coercion and compliance are politically and morally multifaceted. Furthermore, the insurgent politics of learning-to-resist are charged with the same ambivalence in the police context, playing off the corresponding ambivalence of coercion.
In the empirical part of our paper, to which we turn next, we explore the tensions and contradictions involved in the politics of learning in policing context further. We do this with the recognition that, whilst the specific dilemmas faced by the police may be unique, the analysis of these dilemmas has a broader appeal. This is not just because they represent the complexities of public sector organizations, but also because, due to their unique political and societal remit, police act as a reflection of the society in which they are embedded, holding up a mirror to its conflicts, tensions and contradictions (Hartley et al., 2017). It is therefore pertinent to explore the politics of organizational learning in a policing context. In the sections that follow, we show how this focus makes visible a wider spectrum of political modalities of learning that interact with each other in complex and often contradictory ways.

Methods
To address our aim of developing a more nuanced understanding of the politics of organizational learning, we draw on an empirical study of a large city police force in the UK employing many thousands of officers and staff, to which we will refer as 'CPF' (an abbreviation of 'City Police Force'). Like other police organizations in the UK and elsewhere, CPF is having to respond to significant political, economic and social changes and corresponding shifts in policy. Faced with austerity and budget cuts, it is under the ongoing imperative to continue making financial savings whilst addressing the growing complexity of police work and rising demands on police time (as crime and protecting the public become more sophisticated and boundaries between policing and social services, also under growing budget constraints, become more blurred). At the same time, CPF is under pressure -not just from the government and its own ambition but also through the growing scrutiny from traditional and social media -to become more trusted as a police service by fulfilling its remit in reducing crime and protecting increasingly diverse communities. In response to these pressures, CPF is currently in the midst of what they consider to be a major organizational transformation, in which a move towards a culture of learning is a key element.
This paper draws on a subset of data from a two-year engaged scholarship ethnography (Cunliffe and Scaratti, 2017;Van de Ven, 2007) conducted at CPF. The role of our research team has been to work in collaboration with CPF to explore the practices and understandings of organizational learning, and to assist them in both a facilitating and a critical capacity as they undergo the transformation. This role has meant that we have been deeply embedded in CPF and implicated in its organizational politics of learning. In many ways, we have been allied with the driving forces of change towards a learning culture, especially where it involves the questioning of more problematic ways of working, such as the propensity to resort to blame and coercion when things go wrong (Tomkins et al., 2020). On the other hand, the influence of critical scholarship has played an important role in our research, making it sensitive to insidious, diverse and clashing operations of power and giving us the impetus to explore the richness of contrasting perspectives and agendas. Like King and Land (2018), we have sought to channel the democratizing force of participative, engaged research into crafting a version of 'critical performativity' (Spicer et al., 2009(Spicer et al., , 2016 in which research becomes 'affirmative, caring, pragmatic, potential focused and normative' (Spicer et al., 2009: 537). This channelling took the form of situated knowledge co-construction through ongoing dialogue (Cunliffe and Scaratti, 2017) with our police colleagues.
Working in-depth with a single organization in a specific national context enabled us to immerse ourselves into its situated complexity, explore a multiplicity of interweaving and contradictory perspectives and practices through multiple sources of data, and become engaged in conversations with our CPF colleagues that were meaningful to them and us. It also put us at a significant risk of 'going native' as we experienced very closely the immensity of challenges police officers and staff face on a daily basis and the care and motivation to help others many of them bring to their work. Holding on to the values of critical performativity allowed us to combine reflexive distancing (Ybema and Kamsteeg, 2009) and critical questioning with empathy and care in relation to our CPF colleagues. This was a precarious and tension-filled balance, but it helped to open up spaces for dialogue and collective critical reflection within CPF. In these spaces different and dissenting voices could be heard and different assumptions surfaced and questioned, shaping CPF approaches to organizational learning. This also meant that the validity of our study was co-constructed with our participants through ongoing, reflexive and recursive 'thinking out loud' 'about researcher concerns, safeguards and contradictions' (Cho and Trent, 2006: 327). In other words, we embraced a holistic and processual approach to validity, combining the more transactional validity techniques of reflexive member checking and triangulation with the more transformative validity aims of organizational and political action (Cho and Trent, 2006: 333).
Engaged research implies a rebalancing of power relations historically tipped in favour of researchers over the researched by moving to a co-production of research agendas and knowledge (Van de Ven, 2007). For us this has meant that the scope and focus of our project, within the overall theme of organizational learning, have been constantly evolving and (re)negotiated. Abduction has been a fundamental vehicle in this process of co-creation, through which a succession of research strands, problems and theoretical lenses have emerged. This paper in particular has developed out of the research sub-theme on 'empowering learning' and draws on data from that sub-theme. This includes primarily 32 in-depth, semi-structured interviews with senior leaders, officers and staff from both the centralized functions and front-line policing, incorporating response teams, neighbourhood policing and community support, investigations and specialist safeguarding functions, including child protection and mental health. These interviews were conducted face-to-face in various locations where our participants work, and audio-recorded, fully transcribed 1 and stored securely in an anonymous format. The semi-structured format has been an important feature in terms of knowledge co-construction, as we used a set of pre-planned questions as a loose guide to keep us anchored in the previously negotiated direction, whilst the free-flow aspect of the interviews enabled new ideas to emerge around the guide. Our understanding of the politics of learning within CPF is also informed by more informal conversations and observations we have made as part of our ongoing work with CPF, including our participation in board and working group meetings, in a new organizational learning network that we helped to create, and the time we spent in police headquarters and stations and out on shift with a response team. This work has meant frequent contact and collaboration with police colleagues, with some of whom we have had regular contact over an extended period of time, and many others whom we met once or occasionally. We have taken extensive notes to capture these data, reconstructing them from memory immediately after the occasions on which they emerged. We have used these notes extensively to structure conversations within the research team as we shared our interpretations of the data and developed our abductive analysis (see below). Finally, our document data, which include multiple strategy, policy and other organizational documents, some of which we have co-created with our police colleagues, were used as an important source of official CPF discourses of organizational learning and the associated broader discourses (e.g. of austerity and 'doing more with less') which helped to contextualize our other data.
Our participants represent a wide range of seniorities (from trainee constables to top policing ranks), functions and specialisms, and we have also purposefully sought to include as many female participants as possible to counter-balance the persisting male dominance in the police. This has been an important part of our aim to understand multiple views, interests and agendas. Our analysis took the form of a multi-stage, iterative, abductive and reflexive thematic approach. This meant that our analysis and understanding evolved gradually through ongoing dialogue and repeated cycling between theory and literature, our data and our own and our police colleagues' interests, adding, negotiating and refining layers of individual and collective understanding. The aim of this process was to look for patterns, overlaps and tensions to draw out themes (and the relationships between them) that would develop validity through resonance across our multiple points of reference. Throughout data collection and analysis we have been constantly aware of how our varying political and identity positionings (that differed within our own team as well as for each researcher situationally), including those of 'ignorant newcomer', 'managerial representative' and 'interested interlocutor' described by Ybema et al. (2019) but also 'expert' and sometimes even worryingly 'hero' (expected to solve the organization's manifold dilemmas) elicited a wide variety of responses and narratives from our participants and knowledge co-producers. We have sought to use this richness to convey the complexity of the politics of learning within CPF.
In the findings section below, we discuss the themes that have emerged through the above process in terms of political modalities of learning, the tensions and asymmetries between them, and the dynamic complexity that emerges through their interaction. Due to the highly sensitive topic and context of our work within a single organization, we have to be extremely careful about making individuals unidentifiable. Therefore, we purposefully blur the boundaries between our data sources and take special care of anonymity and non-traceability with direct quotations, only identifying the speakers as belonging to one of three broad groups (senior leaders, police staff and frontline officers). On the other hand, when citing our participants, we often do so at length in order to convey their voices and experiences as much as possible in their own words.

Findings: A kaleidoscope of political modalities
This is the most unique job in the world. . . It is what you might do for two different situations that are the same situation, effectively, will be completely different every time. . . There is no approved method for anything. It is all judgment-based. Everyone will do it differently. As long as they act within the law, within procedure, and what they do is proportionate, legal, accountable, necessary, then it's fine. But it's always going to be different, every single time. And what someone might think is the right way to do something might be the wrong way. And someone who thinks the wrong way to do it might be the right way. It is a very upside down, strange job. (Frontline officer) Our findings paint a rich picture of different and contrasting shades of political modalities of learning intermixing as part and parcel of the dynamic complexity characterizing CPF. We have been told about, shown and ourselves experienced the complexities of learning in the policing context in almost every interaction with our police colleagues. In this kaleidoscope of organizational learning, all three modalities we derived from the literature (empowering, coercive and insurgent) are simultaneously visible. They interact with each other in convoluted ways, overlapping, contradicting and struggling with each other in a dynamic arrangement of several competing and contrasting dialectics. The organizational transformation acts as an important ideological and structural backdrop to this complexity. It embodies the tensions between discourses of austerity and community responsiveness translating into budget cuts, the need to make savings and restructure and, on the other hand, the imperative to empower officers and staff in order to foster a culture of learning. As a result, the three political modalities of learning are mixed in such a way as to give rise to a fourth, which we call the palliative modality and in which the emphasis is on learning-to-cope. Below, we discuss each aspect of these findings in turn with reference to our data, and starting with an overview of the empowering, coercive and insurgent modalities of learning.

Empowering modality
Empowering learning is an important part of the official politics of learning in CPF, in as far as it is stated as a prominent objective in the current CPF strategic vision and people strategy documents, which are acting as blueprints for the organizational change. These strategic documents draw on the broader discourses, ideologies and policies of austerity and community responsiveness. They emphasize the need to 'do more with less' by ensuring that the change initiatives deliver on their promise to empower people. This entails, the documents state, developing a culture of learning (balanced with managing risk and accountability), in which staff and officers are empowered to act more autonomously and creatively so that they have the support and capability to respond to the rapidly evolving and increasingly challenging operational demands. Empowering people to make decisions in order to do their job, thrive and make a difference is also presented as a fundamental part of the strategic goal of nurturing a healthier work environment and making CPF a more attractive place to work, where staff and officers are engaged and take pride in their work. Altogether, CPF strategy creates a vision of organizational learning basking in the light of 'Utopian sunshine' (Coopey, 1998;Driver, 2002), under which learning is fundamental to achieving a positive change for CPF, the individuals within it and for the broader public the organization serves.
Far from being empty rhetoric, we have worked with many police staff and officers who are trying to translate the above vision into operational reality. These attempts include a number of different initiatives, from a fast-track programme to empower new kinds of (empowering) leadership, to new technologies of information capture and sharing, an initiative to give response officers control and responsibility over their own crime investigations, an initiative to shift from blame to learning in appropriate situations (Tomkins et al., 2020), and a new organization-wide network of organizational learning 'champions'. The latter initiative, for example, to which we have contributed, aims to regularly bring together those who are interested in organizational learning from different parts of the organization in an informal environment in which rank distinctions become less noticeable and in which it is safe to voice dissenting opinions.
Our frontline interviewees have told us that at least some of the initiatives are having resonance in terms of empowering learning on the ground. For example, some of the new technologies can be experienced as giving the officers more control over their own learning, such as when the recently introduced body-worn cameras for response officers provide opportunities for autonomous reflection: You can watch it back. And sometimes. . . the camera might have picked up potentially something that you've not seen. And. . . in the future, you might think, well, I remember this happened and I didn't see it. So, it opens your mind to different things. And you see different things. And you need to learn from it, really. (Frontline officer) In addition to the central top-down initiatives, local leadership practices also sometimes embody empowering politics of learning, although this can vary widely between teams. The practice of debrief in particular can be an empowering learning experience: The best things about the debrief was it wasn't rank-orientated, the inspector would get a bashing from the PC [Police Constable]. And they would tell him what they thought, and it was great because the inspector would sometimes say, all right, I'll take that on the chin. My fault, I'll take that on board. . . It wasn't the whole 'sir, I think. . .' Because you can't learn like that. (Frontline officer)

Coercive modality
We found that the strategic imperative to empower learning through organizational transformation continually comes up against pervasive and well-entrenched coercive politics of learning within CPF. This entrenchment is partially due to the sometimes quasi-military nature of policing work, which can involve command-and-control situations requiring the learning of discipline and obedience without the luxury of opportunity for questioning the instructions. This is reflected in the ongoing operational and symbolic importance of hierarchy and ranks that persist through the organizational transformation, and the lingering 'old school' leadership styles that even the more 'empowering' leaders have to occasionally employ. The coercive modality of learning is also deeply entrenched in CPF due to policing being a very thoroughly (externally as well as internally) regulated context and potentially becoming increasingly so -the case of changing government policy and legislation making police powers more transparent and accountable to, and therefore controllable by the broader public. For our police colleagues, this translates into the perennial need to learn-to-comply with the changing regulations, which is exacerbated in the times of organizational change: The proliferation of regulations, policies and guidelines means in practice that CPF remains to a great degree process-driven, as it is often safer for staff and officers to learn-to-comply with the prescribed process than face the high risks of doing things differently: There is always the nervousness that, that's all very well, but IOPC [Independent Office for Police Conduct] may come in and take a different view. And we see them running gross misconduct and I could lose my job. So, actually maybe I'll follow the process. You know, it takes a bold cop to say, I have this process, but what's in front of me is slightly different and I'm going to do that. . ., and I feel safe doing so. (Senior leader) The CPF training provision, in particular, being used as a key means for communicating and normalizing compliance with regulations, is a common arena for the playing out of the coercive politics of learning. Many of our frontline interviewees described training as centrally-imposed, compliancefocused, procedural and often not reflective of their genuine learning and development needs: The PDD [Professional Development Day] days are just, 'oh, by the way, we have now decided you are going to do this. And this is how you are going to do it'. PowerPoint. Done. Okay, fine. (Frontline officer) Moreover, training can work to institutionalize coercive learning in CPF by emphasizing the dire consequences of failing to learn-to-comply. As one frontline officer told us: The training that I had. . . it was like, 'if you mess up, you're going to lose your job'. That's how I felt. So. . . I don't want to mess up, so you're proper cautious about things. They kind of said, 'you really need to cover your ass'. And it would take me a couple of hours just to write a crime report, where it shouldn't take that long. And the reason it took me that long is because I was so worried about messing up. (Frontline officer) This can create coercive and punitive dynamics of fear and blame evocative of the dystopian 'Foucauldian gloom' (Coopey, 1998;Driver, 2002), within which learning-to-comply also becomes learning to keep quiet about things that have gone wrong: I hear a lot of stories where officers or perhaps staff as well have put their head above the parapet. . ., explained how something hasn't gone well and then they've been punished for it. Or they've made a mistake. . . and have been too scared to own up to that because they feel they get punished and they have got punished. . . I do continuously hear stories of that is how people are being treated. So why would they put themselves at risk if. . . managers, senior officers are punishing them, penalizing them for putting their hands up and saying that I've done something wrong? (Police staff) There is also the sense of learning being seen as performative , especially where it relates to meeting targets in an increasingly monitored environment: A lot [is] monitored. Exactly what you are doing. How much you've done. Numbers. It's a generalization, but when I first started, it was more to do with how you did it. Whereas now, it tends to be more how much you do of it. . . The camera is on all day. Everything is monitored. And CCTV cameras. (Frontline officer) The above also highlights the double-edged role of new technology, often experienced less as empowering and more as an 'electronic Panopticon' (Lyon, 1993) on the frontline.

Insurgent modality
Most of the manifestations of the insurgent politics of learning we have come across in our work with CPF have been of the micro-level and micro-emancipatory variety. This could be partially because, as in the case of Ybema et al. (2019), we were sometimes seen as senior management representatives and therefore potentially exposed to particular kinds of narratives at the expense of others. However, based on what we have learnt from the project overall, it seems likely that our experiences have been mostly representative of the nature of insurgent learning in CPF, and that this nature is largely due to the prevailing strength of the coercive politics of learning as discussed in the previous sub-section.
It could be tempting to dismiss the importance of 'micro-insurgent' learning we encountered as 'decaf resistance' (Contu, 2008), but it plays a notable role in the politics of learning in CPF. Particularly significant in this regard are the views of the central staff, many of whom see the prominent position given to organizational learning in the CPF strategy and organizational transformation plans as an outcome of a (still ongoing) bottom-up movement to which they are contributing. They told us that they have had to take control and build up an agenda and a vision for organizational learning through small steps in working together with other interested people within and outside the organization: This is driven bottom-up, which I'm not saying is a bad thing. But this is groups of people within [CPF] and within other forces and within the College [of Policing] and that sort of thing, realizing that it's important. Arranging a meeting, getting together, finding out what each of us are doing, realizing that there are gaps, bringing other people into their meetings. So, we're kind of joining the dots because there is no Chief Officer lead for Organizational Learning nationally. There isn't one. (Police staff) It's almost like I visualize myself on Mars, just with rocky terrain. And you're walking around and eventually you come across another person, that's what it feels like, and that they're organizational learning, they represent another person doing organizational learning. At the moment there's nothing built yet. (Police staff) For those on the frontline, the insurgent politics of learning look somewhat different, as officers are more concerned with gaining more control over the learning agendas that are imposed on them through training, and also with getting the frontline voices and learned experiences heard by those further up the CPF hierarchy, and by politicians, media and the public. In this context, learning to speak out despite the personal risks it poses becomes important: The insurgent politics of learning on the frontline are also often engaged in negotiating the obstacles that come up against frontline experiences becoming 'integrated' and 'institutionalized' (Crossan et al., 1999) into broader organizational learning, with the officers becoming frustrated when they do not see impact and change as an outcome of their learning: Sometimes you feel like your voice isn't heard when you say things to, I don't know, your supervisor. All they tend to say is, 'I'll pass on the message'. You never find out the result. It's always, 'I'll just pass it on, I'll send an email to a superior'. That's what it is, but you never really see the impact and the change. (Frontline officer) Learning-to-resist in CPF therefore can be positioned in a variety of ways in relation to organizational change (Thomas and Hardy, 2011), sometimes acting as a creative force driving change, sometimes being directed at reshaping and re-routing change, and sometimes trying (often in vain) to oppose and challenge change: If you don't listen to the people that actually do the job, and you keep implementing change. Change, change, change, change. But don't listen to the people that are facing the repercussions of that change, then nothing is going to happen, is it? (Frontline officer)

Tensions between the modalities
As is already evident from our discussion above, the empowering, coercive and insurgent political modalities of learning do not just coexist in CPF side-by-side but intermix in complex ways. There are two persistent tensions that emerge through their interaction, both of which spin their own dialectics. The first tension can be traced back to the fundamental policing conundrum of trying to steer a course between compliance and responsiveness (Bradford et al., 2013). It translates into a contradictory and dialectical relationship between coercive and empowering politics of learning, which is one of the key threads running through our data. The tension often boils down to the understanding that empowering learning is deeper and more transformative, but coercive learning is easier, quicker and safer: I'm quite passionate about what I think organizational learning should be as opposed to it's not a series of [prescriptive] messages, which again a lot of people wish to default to. So, if something goes wrong, the solution to it isn't telling people 'get it right'. The solution is understanding why it went wrong and helping create the environment in which those circumstances don't repeat. Just telling someone not to do it again is not learning. But telling people not to do it again is quick, easy and provides a figure, whereas understanding what went wrong, trying to create an environment in which it won't go wrong, doesn't produce a quick win or a tangible figure. (Police staff) Being a process-driven organization makes it difficult, but if you're a people-driven organization and you empower people to go and talk about vulnerable people with other professional people, you can probably do it. But we are slightly uncomfortable with it being a little woolly. (Senior leader) The second tension has to do with the dialectical nature of the relationship between compliance and resistance (Bristow et al., 2017;Mumby, 2005;Ybema et al., 2019) and translates into the same kind of relationship between coercive and insurgent modalities of learning, where learningto-comply and learning-to-resist continuously struggle with each other but also mutually constitute each other. For example, this can take the shape of insurgent learning becoming gradually edited through the layers of hierarchy until it becomes compliant with and co-constructive of the dominant narrative, as in the following quotation: We did a report. . . and I wrote the analysis. . . And every time it went up a level, it would get polished, as I called it. And the kind of raw of what people were saying was the problem would get polished out a little bit. So, by the time the report ended, it wasn't really reflective of what the people were saying at the bottom. It was a nice sanitized version of something. . . So, the truth of the actual learning never really gets out because somebody thinks that the raw truth of what the problem is, you can't handle the truth, so we'll just make it something different. (Police staff) The tension between insurgent and coercive modalities can also take the shape of learned cynicism, which represents a contradictory mix of resistance through dis-identification with the dominant culture or change, and compliance through continuing to reproduce the very same things 'at a cynical distance' (Fleming and Spicer, 2003): I'm cynical but lots of my cynicism is born out of the fact that I've been here before. . . And people would say, 'it will be different this time'. And well, it wasn't different the other four times. But I plod on because there's a value to it. (Police staff) Police officers, by their nature, I think are very cynical. And it's very easy to be cynical when we see the same faces every day. The same people. The same crimes. It is very easy to be cynical. Very, very easy. And yes. I think a lot of people do become disillusioned after a while. Especially when things start to get a bit challenging, when we're asked more and more things to do with less and less people. It's the biggest issue that I know of, from my perspective. And I think every other officer will probably tell you the same. . . Yes, it is very easy to become cynical and disillusioned with the way things are at the moment. (Frontline officer)

Palliative modality
The frontline officer in the above quotation points to the impact of a particular aspect of organizational change within CPF -the reorganization and cost saving in response to discourses of austerity and externally-imposed budget cuts. This aspect of transformation operates in a contradictory relationship with the discourses of empowering learning, on the one hand acting as a justification and an impetus for the latter (learning being presented as a means of managing to do more with less), and on the other hand placing constraints upon it. This contradiction is experienced both on the frontline and by central staff: Fewer resources, less money, less time. More is expected of you and you're expected to learn more as well. I don't always think it's the best environment for learning or retaining information because it's so hectic.

(Frontline officer)
There are costs that can be cut. But it's supposed to be a drive to improving the professionalism of policing, and a central core of most definitions of professional organizations is investment and training and learning and continuing development. So, there's a bit of a conflict there in my opinion. (Police staff) We found that this contradictory impact of change casts a long shadow over the palette of political modalities of learning in CPF, tinting their mixing shades darker. In particular, many response officers, who feel the brunt of the combination of dwindling officer numbers, wider geographical areas to cover due to the reorganization and the growing demand for police time, are experiencing so much strain and pressure that making a positive difference and thriving are seen as unrealistic goals under the current conditions: You struggle to make a difference when you can't do your job effectively. And that's what a lot of officers are facing now. The biggest thing is we are almost being set up to fail, in a way. . . We can't do the job to the best of our ability. We're failing ourselves. We're failing our victims. Yes, it's very demoralizing. (Frontline officer) Finding time for learning-to-comply with the changing regulations, or even for insurgent learningto-resist, let alone for empowering learning-to-thrive can seem like an almost unimaginable luxury: I've been doing everything so fast time, some of the things that I could do with a bit of help and. . . teaching on, there simply doesn't feel like there's always the time to fill it in because of the pressures of everything that's required of us now. (Frontline officer) We're trying to do too much with nothing. And we don't have time to learn or breathe or eat food, do you know what I mean? They don't have time to eat. I'm not joking, they don't have time to eat. (Frontline officer) However, rather than learning grinding to a halt as the above quotation could be taken to indicate, what we see from our data overall is the blending of the empowering, coercive and insurgent political modalities of learning together in such a way as to give rise to a fourth -'blue and flashing'political modality of learning. We call this modality palliative to refer to a politics of learning that is aimed at alleviating pressing problems without resolving the underlying cause, or (as in medicine) relieving pain without curing the underlying conditions. The word 'palliative' derives from Latin palliare, meaning 'to cloak', and so it conveys a sense of some of the deeper tensions and contradictions becoming cloaked, smoothed over or temporarily patched up when palliative learning is operationalized. In palliative modality, the aim is to find and maintain a pragmatic balance between the contradictory power relations, and the focus shifts correspondingly from learning-tothrive, learning-to-comply and learning-to-resist to learning-to-cope, which involves learning to 'make it work' whilst 'holding on for dear life' (as in palliative care) on the edge on non-survival. To quote a response team leader: I know exactly what's going on in the streets and our systems and how we make it work. And that's what we're doing, making it work, because it's not broken but it almost is. It's on the edge. . . I'm not a politician but the cuts that we had, tremendous amounts of cuts. And because we made it work then, they thought well, do it again. And it's just got too far, too deep. . . And it's not sustainable. People are going off sick, people are stressed, people are strained. Every single officer is working beyond their capacity and. . . you have this revolving door where people are leaving the job with all the experience, so. . . you're left with very young in service police officers, who are coming out of a training school which is totally different to the one I came out of. It's broken almost. We're just making it work and we're just holding on for dear life and we're just making it work somehow. (Frontline officer) The palliative modality is in some ways evocative of the notions of 'adaptive' and 'single-loop' learning (Argyris andSchön, 1974, 1978;Senge, 1990) because it is not directed at resolving the sources of the tensions that set it in motion, and indeed it may produce organizational learning outcomes that on the surface look adaptive. However, we want to emphasize the extent to which, somewhat ironically, palliative politics of learning involve ongoing critical reflexivity, questioning of what is truly important and creativity in the reworking of the norms and assumptions underlying the other political modalities whilst trying to achieve a balance between them. It is therefore far from being as simple as 'a thermostat that learns when it is too hot or too cold and turns the heat on or off' (Chiva et al., 2010: 116) and more like the Niebuhr prayer in placing an emphasis on learning 'the wisdom to know the difference' between what can be changed or challenged and how, what must be complied with, and what can be ignored. Sometimes, this can take the more straightforward form of learning where to put limits on conflicting demands, as in the example below: When you're learning and. . . Say you're filling out a little book, they will tell you to fill out every single detail. You need to put all this stuff in, cross everything out, underline everything. You have to fill out all forty pages. . . Where in actual life, you know that at times you can't physically do that. So, you fill out the bits you need to and that's all that matters. It goes to court and they're happy with it, because they know what's been written and what's not been written. Whereas in the training room. If things were done as they were literally taught, no-one would be out and about policing, they would all be spending their time doing paperwork, because no-one would have time to do everything so thoroughly as in the training. But that's how they have to train us. (Frontline officer) At other times, it involves making use of the 'overlaps and alliances' (Bristow et al., 2017) implicit in the tensions between conflicting modalities. We have come across examples of makeshift learning solutions that employ an alignment of coercive and insurgent politics to palliative ends. In one case, a probationary frontline officer developed (in personal time) an interactive PDF that presented the entire foundation training curriculum in a format that was quickly and easily accessible when out on street duty, thus taking more control in enabling himself and close colleagues to comply with regulations in pressurized situations.
Yet at other times, the palliative modality involves working out a contingent deployment of each of the other three modalities depending on the minutiae details of a specific situation. One example of this concerns the new requirement for response officers to carry out their own crime investigations -a top-down initiative that is meant to be empowering but that in practice has adverse implications for response officers' already significant workloads. One response sergeant described to us how this leads to police constables learning to take more control in more serious incidents (where they are not expected to do so) and resist dealing with less serious incidents (where they are expected to take the responsibility) in an attempt to make their workloads more manageable: It is to do with the amount of workload they all carry. If you took responsibility for something then you're going to carry on investigating it and therefore you'll get there and there will be this sort of 'no, I don't want to do it'. 'No, you can do it'. So, a bit of shuffling around, at which point, we [sergeants] shouldn't really have to turn up to those incidents, but we do because no one's taken responsibility for it. So, it's the inverse of what should happen. You go to a serious incident where we [sergeants] are required to turn up, you get there and there's nothing for us to do. When we go to a slightly serious incident, you should be able to rely on your team. But because they're all stressed about their workloads, nobody actually wants to take any sort of responsibility for it. . . They will try and just slope their shoulders and hope someone else will deal with it. (Frontline officer) As the above quotation indicates, the palliative modality may cloak or patch up some tensions, but they can resurface again in new ways. Individuals can learn-to-cope with the challenging conditions they face, but this learning can have a range of outcomes (Örtenblad, 2002). These can range from unexpected initiative or empowerment, to more efficient compliance, learned shuffling of responsibilities, learned helplessness and learned workarounds that, whilst they make things temporarily work, can lead far from a healthy work environment where staff and officers are engaged and take pride in their work.

Discussion
Our dialectical analysis paints a colourful picture of the politics of learning in the politically charged and evolving context of CPF. The tensions and complexities of the politics of learning in CPF defy an either-or classification into empowering 'Utopian sunshine' or coercive 'Foucauldian gloom' (Coopey, 1998;Driver, 2002). Instead, as seen through the dialectical 'both-and' lens (Collinson, 2005;Mumby, 2005), they kaleidoscopically fragment and combine into shards of sunlight, gloom, fluorescence and flashing emergency blue lights, which intermix and interplay with each other in complex and contradictory ways. The dialectical approach to the politics of learning is thus helpful in articulating more fully the richness and complexity of organizational life, as Driver (2002: 34) advocates, in a number of ways. Firstly, our dialectical analysis explores the coexistence and interplay of four distinct but contingent and interrelated political modalities of learning in our research setting. Learning in CPF is at once empowering, coercive, insurgent and palliative, manifesting alternately and simultaneously as learning- to-thrive, learning-to-comply, learning-to-resist and learning-to-cope. Secondly, we articulate and examine persistent and ongoing tensions between the four modalities, notably between empowering and coercive, and between coercive and insurgent modalities. Thirdly, we draw attention to the asymmetries between the modalities with the balance between them being tipped heavily towards palliative learning. Fourthly, we note how the complexity of the politics of learning within CPF mirrors the complexity of CPF itself, as the latter absorbs and reflects the conflicting and evolving broader political and societal pressures and expectations. These are heaped upon CPF as a major policing organization and on its staff and officers as 'street-level bureaucrats' (Hoggett, 2006;Lipsky, 2010) and conduits of state-sanctioned (deadly) force. Below, we discuss these aspects in more detail in relation to the theoretical and practical implications of our paper.

Theoretical implications
We contribute a dialectical perspective on the politics of learning to organizational learning literature, helping to articulate a more nuanced understanding of the relationship between learning and power (Collien, 2018;Contu and Willmott, 2003;Harman, 2012). Having started with the view of learning practices as embedded in power relations that enable, facilitate or obstruct them (Contu and Willmott, 2003;Coopey and Burgoyne, 2000), we have conceptualized learning in organizations as inherently and inevitably political. Developing this into a dialectical approach similar to those used to analyse power dynamics elsewhere in management and organization studies (Benson, 1977;Bristow et al., 2017;Collinson, 2005;Mumby, 2005;Ybema et al., 2019) has helped to avoid the trap of either-or dichotomization of the politics of learning as 'Utopian sunshine' or 'Foucauldian gloom' (Coopey, 1998;Driver, 2002). Seen through a kaleidoscope-like dialectical lens, organizational learning is neither simply liberating empowering sunshine nor simply disciplinary coercive gloom but rather a dynamically complex (Guiette and Vandenbempt, 2016) and contingent amalgam of conflicting and contrasting political modalities that interplay with each other in multiple and often contradictory ways.
The dialectical approach thus enables the exploration of the 'texture and nuance' (Mumby, 2005: 38) of the organizational politics of learning. Texture and nuance are added, firstly, through the acknowledgement of the contingency and therefore context-specificity and embeddedness of the politics of learning. In other words, if the politics of learning are contingent and dynamically complex, then their precise configuration and effects depend on the organizational setting within which they take place and the broader political, economic and societal context within which they are located. In our case, our research setting arguably exacerbates the political complexity of learning. We have focused on a large police organization, which embodies the manifold political and operational challenges and dynamic complexities characteristic of public sector organizations (Hartley and Skelcher, 2008;Hoggett, 2006;Rashman et al., 2009). This organization also faces its own additional challenges associated with the paradox of police compliance and responsiveness (Bradford et al., 2013). Within the 'inherently and inescapably political' (Bowling et al., 2019:15) policing context, the politics of learning involve juggling a 'dynamic set of balances' between competing and often conflicting interests and values between the state, police and different publics (Hartley et al, 2019;Marenin, 2000). Moreover, our organization is in the midst of a transformation, which adds political complexities of change (Mastio and Dovey, 2019;Thomas and Hardy, 2011;Ybema et al., 2019) and both exposes and complicates its politics of learning. This amplified political complexity of learning underscores the need for more studies of public service organizations in organizational learning, where they are currently under-researched (Olejarski et al., 2019;Rashman et al., 2009). The context-specificity and contingency of the politics of learning also call for further research to include diverse organizational settings and contexts more generally.
The dialectical approach also contributes texture and nuance to the analysis of organizational learning by drawing attention to the interplay between more agentic and more structural and abstract aspects of power-resistance relations in the politics of learning. The latter are neither wholly free-wheeling and liberating in the Utopian tradition of Senge (1990), Kofman and Senge (1995), Goh (1998), Bhatt (2000), etc. nor completely constrained by totalizing and disciplinary Coopey, 1995;Easterby-Smith, 1997, Snell andChak, 1998) abstract forms of power where learners are inescapably locked into Lukes' (2005) style false consciousness. Our dialectical perspective enables an orientation 'in which the individual and the system are each viewed as constitutive elements' (Collinson, 2005(Collinson, : 1420. From this perspective, organizational learners are seen as 'neither romanticized nor. . . as unwitting dupes' (Mumby, 2005:38) but as always engaging in context-specific, tension-filled struggles over meaning in which learners draw 'on larger discourses that are articulated in ways that may be coherent and contradictory ' Mumby (2005: 38). 'Street-level bureaucracy' (Hoggett, 2006;Lipsky, 2010) is a helpful concept in highlighting the extent of tensions and contradictions between such larger discourses and their impact on daily lives of organizational learners in public services. Conversely, it is also helpful in emphasizing the active role of public service learners in reworking, 'on the ground', such tensions and contradictions and thus co-shaping the configuration of the conflicting discourses that constitute them. This is very notable at CPF, where the political complexity of learning emerges out of police officers', staff' and senior leaders' continuous working through the tensions between the discourses of law enforcement, community responsiveness, austerity and public accountability and transparency. The attempts to instil an empowering learning culture in this historically command-andcontrol law enforcement organization bombarded with extensive and changing external regulations and in the context of the 'doing more with less' imperative both respond and contribute to the above tensions but never fully resolve them.
The dialectical approach also contributes texture and nuance by enabling the analysis of the specific configuration of intermixing political modalities of learning in a particular organizational setting. Through this analysis, we have identified four distinct but contingent and interrelated political modalities of learning at work within our research setting: empowering, coercive, insurgent and palliative. The empowering modality embodies discourses of workplace empowerment through learning typical of foundational organizational learning literature such as Senge (1990). At CFP it is embedded in the strategic vision, top-down change initiatives and some local leadership practices, which are directed at being positively transformative of organizational power relations and constitutive of progressive and emancipatory organizational change. The empowering modality aspires to the 'positive ideal' (Driver, 2002) of organizational learning operating in a learning-tothrive mode for everyone, and seductively intertwines discourses of learning and health (Tomkins and Pritchard, 2019). By contrast, the coercive modality that is often the problematic and 'gloomy' object of critical organizational learning literature Easterby-Smith, 1997;Schein, 1999), at CPF is concerned with ensuring that officers and staff learn-to-comply with the relevant rules, regulations, processes and hierarchies. This modality embodies the broader discourses of obedience to law, police transparency and accountability and also the command-andcontrol and disciplinary cultures and structures at CPF. This modality is mostly directed at the reproduction rather than questioning of the existing power relations and asymmetries, often through the disciplining medium of formal training, but also through the fear of blame and punitive action. Yet again by contrast, the insurgent modality operates in a learning-to-resist mode in relation to the dominant vectors of power (which can be oriented in a number of ways in relation to organizational change (Thomas and Hardy, 2011)), and within CPF is concerned with struggles over the role and meaning of organizational learning and over the control of learning agendas. Finally, the palliative modality emerges from our research in the form of officers and staff learning-to-cope by finding and maintaining, as street-level bureaucrats (Hoggett, 2006;Lipsky, 2010), a pragmatic balance between the contradictory power relations and competing discourses in the very challenging context of dynamic complexity. We show that the palliative modality of learning is in many ways an outcome of dynamic and complex engagement between the other three modalities in the context of organizational change; it is the political shade of learning that transpires and dominates when the other colours in the kaleidoscope of political modalities are combined. Or, to mix the metaphors somewhat, the palliative modality is the envelope within which other political modalities are folded, like complex origami, in ways that are contingent on specific contexts and situations. Therefore, although the palliative modality may evoke the notions of 'adaptive' and 'single-loop' learning (Argyris andSchön, 1974, 1978;Senge, 1990) in not being directed at resolving the sources of the tensions that set it in motion, we show that it involves ongoing critical reflexivity, questioning of what is truly important and creativity in the reworking of the norms and assumptions underlying the other political modalities whilst trying to achieve a balance between them.
The above underscores the final two ways in which the dialectical lens contributes texture and nuance, that is by exploring 'tensions, contradictions, overlaps and alliances' (Bristow et al., 2017) and also asymmetries between political modalities of learning within their specific organizational configuration. The four political modalities of learning we have identified within CPF do not operate in monolithic or uniform ways or in isolation form each other but intermix and interact in complex, contradictory and asymmetrical ways. In addition to (and within) the dominant palliative dialectics discussed above there are two persistent tensions that emerge through the interplay between the first three modalities. Both of these tensions spin their own political dialectics in which opposing elements struggle with each other but also mutually constitute and enable each other: one between empowering and coercive learning (embodying the compliance/responsiveness policing dilemma (Bradford et al., 2013), and the other between coercive and insurgent learning (embodying compliance/resistance dialectics (Bristow et al., 2017;Mumby, 2005;Ybema et al., 2019)). Neither of these tensions is resolvable, both being fundamental to the political dynamics of organizational learning within CPF. Moreover, our conclusion is also similar with regards to the role of each of the four political modalities of learning -that not only each of them plays a part, but that each of them is crucial to 'making CPF work'. These points lead us to consider the practical implications of our findings.

Practical implications
The context-specificity and contingency of the politics of learning highlighted by our dialectical approach imply that the specific configuration of political modalities of learning and the effects of their interplay are likely to differ in different organizational and national contexts. However, we propose that the four political modalities of learning we identified are likely to be of relevance for studying and engaging in the politics of learning across settings, and especially in police and public service organizations in democratic contexts. In such contexts, it is almost impossible to imagine a police organization (or indeed a public sector organization, or even a private sector organization) freed from the imperative to learn-to-comply with regulations, whether externally or internally imposed. It is similarly hard to imagine a police organization, or any organization aiming to create public value through street-level bureaucracy (Hoggett, 2006;Lipsky, 2010) giving up completely on the notions of empowerment and learning-to-thrive, or being so process-driven and totalizing that it leaves no need and no scope for learning-to-resist. And, if we can therefore expect all three of the above political modalities of learning to be almost necessarily present, at least within the public sector context but probably all sectors, in the crafting of organizational learning, then we can also expect the fourth, in the form of learning-to-cope with the tensions and contradictions produced by the competing agendas of the other political modalities.
What is possible to imagine, however, is a situation or context in which organizational change is channelled differently in relation to the politics of learning, in which political modalities of learning are weighted differently in relation to each other (in which their palette of colours is mixed somewhat differently, and in which their complex origami is folded in a different way), and in which other as yet unidentified political modalities of learning come into play. In this regard, understanding the role of the different political modalities of learning and their configuration in relation to each other within specific settings can be a starting point for reflecting and acting upon the asymmetries between them and perhaps even the sources of some of their tensions. It is possible, and we would say politically imperative, to imagine a context without externally-imposed budget cuts combined with the high demands and expectations that are tipping the politics of learning within CPF so heavily towards learning-to-cope. Given that organizations across the public sector are facing similar contextual challenges and dilemmas (Granter et al., 2019;McCann et al., 2015;Olejarski et al., 2019), the heavy reliance on palliative learning that we see within CPF might be more broadly symptomatic of the malaise of the age. This issue, we suggest, as well as searching for and developing alternatives, is something for future research and organizational and political action to address.

Conclusion
Our key contribution is the development of a dialectical approach to the organizational politics of learning. It has enabled us to articulate more fully the richness and complexity of the relationship between learning and power characterizing organizational life within our research setting. In so doing, we have addressed the two questions posed in the beginning of the paper: (1) what political modalities of learning are at play in our empirical case and what tensions and asymmetries emerge through their interplay, and (2) what implications does this configuration of modalities have for understanding and theorizing the relationship between learning and power, in the context of a policing organization and more generally? In terms of the first question, we have identified four interdependent political modalities of learning: empowering, coercive, insurgent and palliative, and explored tensions and asymmetries between them. We have noted that, whilst mainstream and critical literatures tend to focus on organizational learning as, respectively, empowering and coercive, and to a lesser extent insurgent, much of the politics of learning in our study converges in the palliative modality, where the emphasis is on learning-to-cope (rather than learning-to-thrive, learning-to-comply or learning-to-resist). We have shown that the palliative modality of learning is in many ways an outcome of the dynamic and complex engagement between the other three modalities in the context of organizational change and clashing broader discourses and policies.
In terms of the second question, we note some limitations of our research -namely that in choosing to focus in-depth on a single organization in a particular national context, we have sacrificed a degree of breadth and generalizability. We cannot speak for all organizations in all contexts, but as with other ethnographic studies we can and have offered 'theoretical generalization' (Fine et al., 2009:613), 'suggesting new interpretations and concepts' and 're-examining earlier concepts or interpretations in new and innovative ways' (Orum et al., 1991:13). In particular, we have drawn out several implications of viewing organizational learning through our kaleidoscope-like dialectical lens, which avoids the trap of either-or dichotomization of the politics of learning as utopian or dystopian (Coopey, 1998;Driver, 2002) and enables the exploration of their 'texture and nuance' (Mumby, 2005: 38). The implications include: understanding the politics of learning as complex, contradictory, contingent, context-specific and embedded in organizational and broader power relations; drawing attention to the interplay between more agentic and more structural and abstract aspects of power-resistance relations in the politics of learning; the need for further research into the role and configurations of the four political modalities of learning across organizational and national contexts; and potential for future organizational and political action in terms of exploring and addressing the asymmetries between the modalities. Overall, through our dialectical approach organizational learning emerges as neither a panacea nor a curse, but as at once an object, a means and an arena of complex organizational politics. Within the latter, different political modalities struggle with each other in diverse ways that both embody and co-construct broader relations of power.