The Policy-Practice Nexus as ‘Politics of Use’: Professional Autonomy and Teacher Agency in the Classroom

Abstract

This chapter approaches the policy-practice nexus by scrutinizing the relationship between teacher agency and professional autonomy. Teacher agency has usually been researched from two different perspectives. On one side, scholars are concerned with questions of professional autonomy vis-à-vis specific accountability regimes, and apply, in the broadest sense, a governance framework. On the other, there is a more normatively grounded discussion of professional autonomy, emphasizing how teachers, due to various new forms of (neo-liberal) governance, become increasingly de-professionalized. While acknowledging both perspectives, this chapter questions the conflation of professional autonomy with teacher agency. Drawing on the concept of the ‘politics of use’ and findings from fieldwork in China, the chapter proposes a framework for conceptualizing autonomy and agency as they operate in and between systems, involving and producing different types of agents. The chapter’s findings suggest that the ways in which policy implementation processes have been conceptualized need to be reconsidered. Particular attention must be paid to the political-ideological and normative specificities of both the investigated policy system and of the investigator’s own research traditions, to ensure that policy implementation processes can be compared across a broad variety of cases.

Introduction

In a history lesson at a middle school in Beijing, the 13-year-old students get to learn about China’s “new democracy”. As a case in point, they study some details of the battle between the communist Red Army and the nationalist Guomindang in the 1930s. Zhu, the young teacher, lectures on the various Communist heroes, with the students acting as fill-ins on the heroes’ specific character qualities, at times reading passages from the textbook. As a highlight, Zhu then shows them a clip from a movie which features gruesome and very noise-intensive battle scenes, with hardly any words spoken (except for deafening death screams), and with many combatants left wounded or dead. She stops the movie towards the end of the lesson, when a Guomindang soldier is shot fatally, accompanied by appreciative grunts from the students. In the ensuing interview, I ask Zhu about her choice of material. She tells me how in her teacher training, she had learned about student-centered approaches in the classroom and the importance of developing social and emotional skills. She then reflects upon the lack of such learning approaches in China’s exam-oriented schools, and adds that by selecting this movie, she sought to transfer some of the pedagogical spirit from her teacher training into the classroom.

What does this story – which I have encountered in multiple variations during two decades of doing fieldwork – tell us? First, it illustrates the importance of observing practices. Analyzing policy and curriculum documents, textbooks, or even interviews cannot reveal how teachers realize (or resist) the curriculum on the ground. Desk research can tell us a lot about the intended curriculum, producing valuable insights into the agendas of various stakeholders, such as international organizations (e.g. OECD or UNESCO reports), governments (e.g. laws, regulations, white papers), ministries and local educational authorities (e.g. curricular specifications, guidelines), and schools (e.g. school programs); but it discloses very little about how the curriculum is enacted: the movie shown by this teacher is not included in any database of teaching material, let alone the pedagogical approaches utilized during the lesson.

This leads us, second, to the much-discussed question of policy-practice divergence, or with reference to this edited volume’s theoretical focus, to the fruitful concept of the policy-practice nexus (see e.g. Ohi, 2008; Schulte, 2018): How is policy transformed when being filtered by teacher professionalism (Evetts, 2003), and how is policy negotiated and appropriated within the micropolitics of school environments (Kelchtermans & Vanassche, 2017)? Such questions direct our attention, on one side, to the complexities and layeredness of educational practices (Wermke & Prøitz, 2022), as they need to respond to multiple and at times contradictory expectations (e.g. from parents, colleagues), norms (e.g. in the institutional or political realm), and traditions (e.g. in the form of pedagogical or professional knowledge). On the other side, these multiple processes of filtering and appropriation reveal how we need to think of ‘nexus’ in the plural: policy meets practice not just once, as for example when Teacher Zhu does her lesson planning; but practices dock onto policies both vertically (as for example when researchers, textbook authors, or school principals engage in policy translation at different levels of policy implementation) and horizontally (as for example when policy is enacted across different settings at the same level, such as at the level of the classroom).

Third, and most importantly for this chapter, this story can help us reassess the intricacies of teacher agency. On the surface, Zhu takes the liberty to digress from the textbook and teach the subject in her own way. She can thus be seen as gaining agency in designing her lessons, which especially in the Chinese context has not always been the case. But are her choices autonomous, from a pedagogical point of view? The movie she has chosen would most likely not have been included in an officially sanctioned list of teaching materials, screened and accredited by educational authorities. In that sense, her approach could be considered autonomous vis-à-vis crucial control mechanisms within the education system. Looking more closely, however, we can argue that her choice to emotionalize her history lesson in the way she did is severely compromised by her ideological-political environment. To teach and display love of the Chinese Communist Party is a constant expectation from the central government, and this expectation has recently been rendered more intrusive in the Ministry of Education’s decree of integrating Xi Jinping thought – an ideology named after the current president – at all educational levels (MOE, 2021). Thus, ironically, what looks like an increase in teacher agency does not translate into greater professional autonomy. Again, with regard to policy-practice relations, we can locate various nexuses where certain workings and enactments of policy are produced, involving agents and forces both internal and external to the education system.

Teacher agency and professional autonomy are usually researched from two different perspectives. On the one hand, scholars are concerned with questions of professional autonomy vis-à-vis specific management forms and accountability regimes, and apply, in the broadest sense, a governance framework in order to analyze their cases (see e.g. Wermke et al., 2019). On the other hand, we can observe a more passionate discussion of professional autonomy, emphasizing how teachers, due to various new forms of (neo-liberal) governance and governmentalities, become increasingly de-professionalized (see e.g. Priestley et al., 2013). In a sense, these two strands of research can be regarded as two sides of the same coin, in that the latter is a normative response to the findings from the former. This chapter acknowledges both approaches – governance analysis and normatively framed responses – but twists both perspectives by asking the following two questions. Firstly: If we assume a weakly institutionalized education system that is vulnerable to infringements from other systems (such as from the political system) and thus can be said to possess limited autonomy – how does that impact the agency of teachers? Will their agency diminish, along with their system’s autonomy, or can it actually also increase? Secondly, and perhaps provocatively: is teacher agency always good?

The first question hence attempts to destabilize our own thinking about how education systems (and their subsystems) interact with other systems, by including socio-political contexts which deviate from what could be called the ‘prototypical’ education system of the Global North. Such a perspective can help discern policy-practice nexuses that are often hidden in ‘prototypical’ scenarios, such as the nexus linking individual teacher practices and political ideologies, as can be seen with the example of Teacher Zhu. The second question intends to make more explicit the normative connotations surrounding the concept of teacher agency. Academic discussions of ‘teacher agency’ are often framed within the emancipatory tradition of pedagogy, idealizing teacher agency as something inherently good (see e.g. Cloonan et al., 2019; Samoukovic, 2015). Rather than arguing for a removal of these underlying normative biases, this chapter aims to look into the workings of norms and values more systematically, in order to understand how norms and values co-produce the ‘agentic teacher’.

The following section will address the potential interactions between teacher agency, professional autonomy, the education system, and what I call the wider environment, including e. g. the political and economic realms. In an ensuing section, I will then zoom in on the interrelations of autonomy and agency and present a nested model of these two concepts, which takes into account both systemic and agents-based interactions. In a fourth section, and based on the approach of the ‘politics of use’ which I have discussed elsewhere (Schulte, 2018), I will address the question of whether teachers, as street-level agents of the state, can actually gain more agency when the autonomy of their schools and of the wider education system becomes restricted. I have called this process ‘side-stepping’, since the state establishes new ways to form direct alliances with teachers and circumvents their professional environments, thereby bypassing acknowledged mechanisms of quality control and accreditation. By looking more closely at the nature of agency that teachers can develop within their given contexts, I am proposing different ‘ideal types’ of teachers linked to the specific relationships between educational and political systems. In conclusion, the chapter suggests that we need to reconsider the ways in which we have conceptualized policy implementation processes. If we eclipse the political-ideological specificities of both the investigated policy system and of our own research traditions, we may unnecessarily limit our capability to compare policy implementation processes across a broad variety of cases.

Teachers, Schools, the Education System, and the State: A Complicated Ménage à Quatre

Teachers, schools, the education system, and the state are usually conceived as being embedded in a hierarchically structured system: the state sets the parameters and boundaries for the education system, which in turn produces and shapes the conditions for schools to operate, including the specifics of teacher education and training, examination and assessment procedures, school inspection etc. At the bottom of this hierarchical system, schools define the range within which teachers can meaningfully act. Depending then on the respective legal, political, and financial structures of governance, we tend to think about the entities of teachers, schools, and the education system as possessing more or less autonomy vis-à-vis the (hierarchically higher placed) entity that has the power to exert constraints.

But is autonomy merely the left-over space that is untouched by constraints? As Dworkin (2015) has pointed out, to define ‘autonomy’ entails the dilemma of reducing the concept’s complexity to the extent that it loses its theoretical power – which however has rendered ‘autonomy’ a crucial concept in the first place. Still, Dworkin convincingly argues that ‘autonomy’ cannot be simply equated with ‘freedom’ or ‘liberty’. He proposes that the concept must instead be understood as the “second-order capacity of persons to reflect critically upon their first-order preferences, desires, wishes and so forth and the capacity to accept or attempt to change these in light of higher-order preferences and values.” (Dworkin, 2015, p. 14) By redirecting the focus from the mere question of whether or not there is a constraint on freedom, to the question of capacity, we move away from an understanding of autonomy as mainly something measurable – such as having more or less autonomy – and instead link ‘autonomy’ to the extent and the ways in which people can make sense of and navigate their options, freedoms, and constraints.

However, the understanding of autonomy as the capacity for second-order, critical reflection has its practical limitations: From what vantage point can we assess a reflection to be ‘critical’, when the nature of critical thinking is itself highly dependent on the context in which an individual has been socialized? Can the reflections by Teacher Zhu, who was introduced at the beginning of this chapter, be considered ‘critical’, since she calls into question teaching and testing practices, and adapts her teaching accordingly? Also, can first and second-order thinking be clearly distinguished from one another in empirical reality? The latter question is particularly relevant with regard to teachers: If reflection is a deeply engrained as well as widely expected part of teachers’ everyday activities, can this activity then still be considered higher-order, or would we rather have to add a third-order level of reflection – namely a level from which individuals such as teachers can reflect upon the very figure of the ‘reflective practitioner’ (in the sense of Schön, 1983)? Moreover, if we think of individuals as being embedded in multiple ways – socially, emotionally, politically, professionally – how can such a capacity for higher-order reflection develop independently, despite the many interdependencies that characterize social and professional lives?

The answer to these questions lies in incorporating, rather than ignoring, these interdependencies. Second- (or third-) order reflection does not take place in a vacuum but is bound by norms, which again are produced by (and in turn keep alive) social-cultural, emotional, political, professional, etc. normative systems. This means that the capacity for reflection does not develop in spite of, but because of these system’s interactions with individuals (and groups of individuals). Depending on whose and what kind of autonomy we have in mind, these interactions will then by categorized on a spectrum between (illegitimate) interference and (welcome) support. To return to the case of teachers: If we see teachers primarily as professional facilitators of learning, teachers’ capacity for second-order reflection would then be expected to take place with reference to professional norms (developed within the education system). Any interference that suggests or prescribes other primary references (such as to cultural or political norms) would consequently be labeled as a breach of autonomy. If, however, we were to consider teachers primarily as, say, political or religious agents, the contrary would be the case: Reflection oriented towards political or religious norms would be the autonomous default situation, whereas reference to other norms would mean encroachment.

These latter, so far hypothetical cases – teachers as political or religious agents – highlight the importance of environment, including the question of which environment serves as primary reference for conceptions of autonomy. From this perspective, autonomy has little to do with pure freedom, or the “comparative absence of regulation”, as claimed by Priestley et al. (2015, p. 144), even though this might be the perception of involved agents (e.g., regarding the so-called ‘freedom of teaching’). On the contrary, autonomy usually entails densely regulated systems (such as that of education or teaching), whose agents do not just endure, but engage in mechanisms of self-governance and (internal) control (cf. Mausethagen & Mølstad, 2015).

Inside these systems, autonomy within the professional work of teachers can assume different shapes. As has been shown in empirical studies, teachers can be autonomous in relation to different aspects of their work, such as lesson planning, choice of teaching methods, learning assessment etc. (Dieudé & Prøitz, 2022), as well as in relation to different domains (such as educational, social, developmental, administrative) and different levels (classroom, school, profession) (Wermke et al., 2019). Autonomy thus becomes a concept that is highly practice-related, and develops in relation to (sub-)systems of regulation. These system-specific regulations do not simply constrain autonomy, but they actually enable autonomy to emerge: Strictly speaking, there would be no teacher autonomy if it wasn’t for an educational system that defines and refines the rules, and hence creates the space for teachers that then comes to be understood as ‘autonomy’. Archer comes to a similar conclusion when she notes that low autonomy entails the difficulty

to pursue goals which have been arrived at within that sphere; instead, institutional operations are defined externally by the party which constrains its services. It is not interdependence as such which results in loss of autonomy but rather […] the emerging capacity of one part to direct and organize the other in accordance with its own operations. (Archer, 1979, p. 62; my emphasis)

Following Archer’s explication, we can place the organization of the education system on a scale between ‘heteronomy’ and ‘autonomy’: On one end of the scale, all organization is determined by the rules, norms, and laws of the ‘other’ (hetero); empirically, it would be very difficult to find a pure heteronomous education system, since the mere existence of a system already entails a certain extent of autonomy. On the other end of the scale, organization is completely driven by the laws of the ‘self’ (auto), molding the respective system into a distinct system with specific tasks and rules. It is from this vantage point – autonomy through specialization, or differentiation – that also Luhmann (2017, p. 114; emphasis in original) has approached the concept of autonomy:¹

Autonomy is grounded in the specifics of system-building operations and their structural condensates. […] The dependence on environment cannot be eliminated, on the contrary [the environment] needs to be seen as the precondition for these systems to exist, and it determines the direction of potential differentiations. We therefore define autonomy as the operative closure of the system, and [we define] operative closure of the system as the autopoietic reproduction of the system’s elements through the network of precisely these elements. Therefore, we can understand the school as a social system, but not […] as a “micro-cosmos” of society within society. […] This [perspective] is absolutely compatible with legal regulations and financial dependencies as long as these are not used as sources of power in order to oppress pedagogical intentions and replace them by something else.

Simply put, such an understanding of autonomy proposes that autonomy exists when the system can do its own thing – ‘its own thing’ consisting e.g. of this system’s rules, norms, rationales, and routines; while the system’s legitimacy and, hence, existence derives from its capacity to produce such rules, and generate a sufficient extent of specificity, in relation to other systems, in order to be recognized as a distinct system. While we may intuitively think of ‘autonomy’ as a right or entitlement, a systemic perspective highlights how ‘autonomy’ is also a “burden […] simply because no other functional system can fulfill the function of another [system]. The state can introduce compulsory education and cover the costs of schools and universities through tax revenues; as an organization of the political system, it cannot itself educate” (Luhmann, 2017, p. 116).

This dialectical approach towards autonomy makes also sense in light of professionalization, such as teacher training: ‘autonomous’ teachers are certainly not those who have not undergone any professional training, but who on the contrary are able to enact their professionally acquired skills and competences within the protected but regulated space of the educational system. To be sure, more recent calls for ‘decolonializing’ education and thereby ‘unlearning’ established ways to teach and learn may suggest otherwise (cf. Caruso & Maul, 2020). However, from a Luhmannian perspective, such developments hardly mean that teachers step out of the system; but rather that, through a partial opening of the system to the outside world, some rules are changed in such a way that autonomy can be exercised in new and different ways. Why systems open up is an essential question for understanding change, and will be taken up in the following section, when addressing the interrelations between autonomy and agency.

Autonomy and Agency: Same, Same, But Different?

If we take this dialectic approach towards autonomy seriously, we need to dismiss the antagonistic conception of autonomy (i.e., schools/teachers versus the state) that has been pervading much of the literature surrounding the pros and cons of neoliberalization, auditing, managerial control, and so on (Forrester, 2000; Helgøy et al., 2007). Likewise, to define autonomy as the scope of decision-making (vis-à-vis control mechanisms; cf. Wermke et al., 2019) does not do full justice to the multiple (potential and actualized) relationships between agents and their environments.² This section of the chapter therefore intends to pick up where Wermke et al. (2019, p. 310) have left, who explicitly concede that their conceptualization takes place “at the price of complexity reduction [… and] excludes other themes related to the question, such as teacher empowerment, the structure of teacher agency and also issues of teacher self-determination”.

How is agency related to autonomy – and how does this make a difference for teachers? Existing attempts at capturing ‘teacher agency’ are problematic for several reasons. Firstly, and frequently, professional autonomy and teacher agency are used interchangeably, with no clear distinction between the two concepts, at times outright conflating the two terms (e.g. Lundström, 2015). Secondly, many definitions of teacher agency are normative and somewhat instrumentalist, clarifying what and how a teacher should be, and how this can be achieved, such as in the definition by Toom et al. (2021, p. 2):

[P]rofessional agency […] in addition to being a teacher’s core capability in the sense that it offers a key for active and skilful teacher learning, also provides understanding of the dynamics of the preconditions for such learning in their work. Yet, professional agency embodies a capacity that allows teachers to learn actively and skilfully, regulate their own learning, learning competencies needed in their work, develop professionally, promote students’ and colleagues’ learning, as well as innovate and promote change in schools.

A third approach turns against this instrumentalization of teacher agency, and instead views agency as “an emergent ‘ecological’ phenomenon dependent upon the quality of individuals’ engagement with their environments” (Priestley et al., 2015, p. 136).³ Environments, in turn, consist of “a configuration of influences from the past, orientations towards the future and engagement with the present” (ibid., p. 137), which individual teachers, due to their diverse life histories, expectations, and actual choices, navigate differently. Interestingly, Priestley et al. note that teachers’ self-perception of agency it not necessarily congruent with actual agency: teachers may feel to possess agency when “they simply go with the flow” (ibid., p. 144).

This observation points to a weak spot in Priestley’s et al. conceptualization: From what vantage point can it be assessed whether agency is real or not, if the relationships between agents and environment are only insufficiently defined, and if autonomy is simply conceptualized as the absence of regulation? Scrutinizing these diverging measurements of agency, Moore (2016, p. 1) distinguishes, on the one hand, between a “feeling of agency” as a “lower level non-conceptual feeling of being an agent”, and, on the other, a “judgment of agency” as a “higher level conceptual judgment of agency” which uses background beliefs and contextual knowledge when assessing an action. While this compensates for the somewhat lacking clarity of Priestley et al. regarding (internally) perceived and (externally) observed agency, it does not solve the problem of the contextual embeddedness of judgment, as noted above with reference to critical higher-order reflection: Depending on the context in which agent and observer are located, both the content and extent of agency can be interpreted very differently. Moore attempts to escape this relativism by assuming an “objective reality” from which “the sense of agency can be quite divorced” (ibid., p. 2). But who is to draw the line between an objective and a subjective reality of a teacher who is then observed to develop either a real or a false sense of agency? Has Teacher Zhu, from the beginning of this chapter, developed a real or a false sense of agency when diversifying her teaching methods? One can easily fall prey to circular reasoning when addressing this question.

A possible way out of this circular argument is to adopt a nested model of agency (see Fig. 3.1), which places both individual and collective agency within the context of professional autonomy, which again interacts with the educational system, the latter embedded in a wider environment consisting of other systems, such as the political, economic etc. Each of these embeddings, or interfaces, can be considered a potential nexus in which policy-practice relations are being negotiated and enacted. While (individual/collective) agency is most closely connected to professional autonomy, as this is where professional and personal identities are being formed, the nested model also allows for other relationships, visualized by the darker and lighter links in Fig. 3.1, which represent stronger, routinized relations and weaker, shifting relations, respectively. Accordingly, teacher agency forms nexuses with organizational arrangements within the education system, as well as with political, religious etc. requirements and narratives.

Fig. 3.1

Interdependency relations between teacher agency, professional autonomy, the educational system, and other systems

System approaches such as the one developed by Luhmann have not been particularly interested in the workings of agency, since they view agents mainly as communicative elements executing the logics of a system. However, drawing on feminist studies (e.g. Abrams, 1999), there might be a way to bring a systems-based perspective on autonomy and an agent-based concept of agency together. Feminists in particular have been concerned with questions of individual agency and empowerment on the one side and structural constraints on the other. Similar to Sen’s (2003) development of the capability approach, feminists have been struggling with the dilemma of choice: namely, with the fact that there is no free choice, since there are always mechanisms (cultures, beliefs, upbringing etc.) that pre-structure our seemingly “free” choices.

Agency, against this background, implies the power of self-definition (as opposed to structure and pre-definition) and, based on this self-definition, the capacity for action; or according to Sherwin et al. (1998, p. 12), it captures the “ideal of informed choice”. Autonomy, they continue, constitutes something more than “actively choosing”: It denotes “a more comprehensive notion of freedom where not only is the immediate choice uncoerced but the circumstances that structure that choice are also free of the coercive dimension of oppression.” To be sure, feminist literature, as a body of theories for social change, and system theory, which is mainly oriented towards accurately describing and explaining processes of systemic interaction, are positioned very differently when it comes to conceptualizing oppression. Empowerment and freedom from oppression are moral imperatives in feminist research; while Luhmann’s (2017, p. 114) “sources of power in order to oppress pedagogical intentions” only call into question the autonomy of the education system, without however passing a moral judgment on how such an encroachment is to be related to values. The closest Luhmann comes to connecting inter-system interaction with questions of legitimacy, is when he raises the question of “which possibilities of intervention the owner [i.e., the state, which provides infrastructure and resources to the education system; BS] has. Under older (and more small-scale) conditions, a pietist king such as Christian VI of Denmark could result in a pietist orientation of school-based instruction. Nowadays such developments are conceivable, if at all, only under an ideologically oriented one-party regime.” (Luhmann, 2017, p. 118) Hence, from a Luhmannian perspective, the state as a pedagogue is an interesting exception to the rule.

As different as these responses to ‘oppression’ may look, they share the view on autonomy as an interconnected concept: in contrast to notions of autonomy in the liberal tradition, which references the disconnected, authentic, proactive autonomous self, both system and feminist theorizing propose a relational, situated, reactive autonomous entity (see e. g. the discussion in Abrams, 1999). In a sense, the feminists’ entangled woman is the equivalent to system theory‘s networked element; both are marked by social reproduction or, in Luhmann’s terms, autopoesis. Despite these overlaps in thinking about autonomy, the differences regarding agency are pronounced and important: While feminist theory works towards raising individual and collective awareness of one’s own situatedness in order to override (parts of) the system’s workings, system theory reduces individual agency largely to “the attribution of decision rights to the communication roles of alter and ego” (Blaschke, 2015, p. 466), resulting in proxy agency (that of the system) rather than individual and/or group-based agency.

Both perspectives, however, can be used to draw a distinction between autonomy and agency. As exemplified in Fig. 3.1, a relational, nested approach allows for diverging types of interaction between, on the one side, agency and different environments, and, on the other, autonomy and its environments. For example, while the relationship between professional autonomy and the wider environment (beyond the educational system) may impact on teachers’ social status or entail forms of political recognition, this environment’s relationship with individual or collective agency may assume forms of ideological influence or moral engagement. Since agency is embedded within professional autonomy, it is within this nexus of agency and autonomy that identities are formed and spaces are created for enacting these identities. Consequently, such a differentiated approach towards agency and autonomy also opens up for diverging directions of agency and autonomy: a high amount of agency (whether perceived or observed) does not necessarily translate into extensive autonomy; conversely, strong autonomy does not automatically lead to high levels of agency. To return to the case of history teacher Zhu at the beginning of this chapter: Zhu may have developed a considerable extent of agency when designing her history lessons by using quasi-propaganda films not sanctioned by the education system. However, this agency has emerged within a space that decreasingly operates according to the rules and norms of professional autonomy (cf. e.g. Evetts, 2009). Zhu could be considered, as we will discuss in the next section, a ‘zealous teacher’, marked by high agency, low professional autonomy, and located in a weak educational system with an intrusive state.

Less State Equals More Agency: Does It, and for Whom?

From the perspective of this nested approach, teachers can be considered agents with potentially multiple roles and connections. Depending on how the education system interacts with other systems, autonomy and agency can unfold differently within these interactions. Since agents do not mechanically execute predefined tasks but imbue their actions with meaning and values, their agency is tightly connected to their political, socio-cultural, economic etc. environments. Elsewhere, I have called this enactment (of e. g. the curriculum) the ‘politics of use’. ‘Politics’, in this concept, means a very broad practice determining “which and whose political values will be put into use when implementing policy” (Schulte, 2018, p. 634):

When policies, reforms, and new curricula are put into use in the classroom, they become necessarily imbued with normative conceptions and values. Whose values the politics of use mobilizes depends both on the teachers’ previous training and socialization, and on the school’s/ subsystem’s autonomy vis-à-vis other sectors, including the state. It can be assumed that the more teacher education is aligned with the objectives of the state, and the more in-service teachers are directly and continuously exposed to state narratives, the less likely it is that organizational levels and actors in between will interfere in the process of policy implementation. Thus, in the case of minimal autonomy of the subsystem and maximum exposure to state narratives, teachers will try to align policy implementation with what they perceive to be the state’s interests. This results in an implementation short-cut in which centrally released policies can jump various levels of implementation: intermediary actors and organizations are being side-stepped. (ibid., p. 630)

However, a strong state and a weak education system are not the sole determinants for teacher actions but need to be related to these teachers’ spaces of autonomy and agency. If teachers were simply agents of the state (or partisans challenging the state), we would again get caught in a dichotomous and rather one-sided approach that we have already criticized above. Instead, teachers, when implementing the curriculum, are differently positioned to resort to norms and rules on which they can base their actions (see Fig. 3.2): At a higher level, with regard to the interrelations between educational and other systems, spaces of autonomy and possibilities of agency depend on the extent to which the education system as a whole can assert itself against other systems, such as the political one (strong vs. weak educational systems). At an organizational and institutional level, these spaces and possibilities are shaped by the extent to which, on the one hand, educational environments such as the school can define and prescribe their own professional rules and norms (autonomy vs. heteronomy); and on the other hand, to which self-definitions (including e. g. ideas about professional ethos or teaching philosophies) harmonize with the organizational and institutional environments (agency vs. proxy-agency).

Fig. 3.2

Teachers in a strong vs. weak educational system (with non-intrusive vs. intrusive state)

Correspondingly, in Fig. 3.2, we can identify different ideal types of teachers as they emerge in a field between the poles of autonomy/heteronomy and agency/proxy-agency.⁴ In the upper right quadrant, we can locate teachers who both possess a considerable amount of agency and can rely on an environment (such as the school) that operates according to the rules of the specific system (the education system in this case). As Fig. 3.2 shows, the relative educational autonomy of the organizational environment can be found in two different settings: On the left side, we would assume the organizational environment to be embedded in an education system that acts autonomously, with little political interference, and educational norms and rules as primary reference; on the right side, the education system as such is subject to political interference which attempts to replace educational norms by political ones, but as a system it still generates environments that can operate by their own rules. Agentic teachers emerge differently within the two settings: In the setting to the left, the ‘professional teacher’ denotes the maximum overlap between self-definition and professional norms, within an environment that strongly protects these norms. In contrast, the ‘renitent teacher’ on the right side, while equally aligning self-understanding with professional norms, is forced to act in an environment that is vulnerable to forces that attempt to dismantle precisely these norms and replace them by political ideologies.

The teacher types in the lower left quadrants constitute the exact opposites: Their agency is severely limited, meaning that they have no self-determination in developing their identities as teachers; and they lack the support of an autonomous environment, resulting in constant exposure to infringements from outside the education system. In the setting to the left, the policy and practice of ‘scripted lessons’, i. e. ready-made lessons that can be taught in a copy-and-paste fashion, are a good example of low autonomy/low agency. In politically intrusive settings (setting to the right), teachers can be easily degraded to puppet agents, with the state pulling the strings. The remaining quadrants – lower right and upper left – are marked by divergent extents of autonomy/agency. A highly autonomous, strongly protected educational environment which however grants little agency to its teachers (lower right quadrant) reduces them to mere executors of the logics of the system: to technocrats. If the system of which the environment is part is not even able to assert itself against political encroachment (setting to the left), these technocrats become disengaged proxy-agents. Finally, as represented in the upper left quadrant, teachers can develop a high amount of agency even when their educational environment is not facilitating these teachers’ alignments with professional norms and values. In strong educational systems, these agents without routinized links to professional norms can be called ‘unorthodox teachers’; while in weak educational systems with an intrusive state, these agents connect to values and ideologies outside their system, becoming ‘zealous teachers’ – like Teacher Zhu from the beginning of this chapter.

Such a conceptualization can explain why teachers, within one and the same socio-political system, can develop highly different identities – that is, develop different ways of calibrating their self-definitions with spaces for agency and autonomy. This contradicts conceptualizations of teachers as exclusively professional educationists. But it also calls into question an overly politicized view of the educational system as the state’s stooge, in contrast to much of the critical literature on education. For example, Apple (2003, p. 1) writes that the educational system, “as inherently part of a set of political institutions, […] will constantly be in the middle of crucial struggles over the meaning of democracy, over definitions of legitimate authority and culture, and over who should benefit the most from government policies and practices.” Empirically, and in contrast to these claims, most present-day societies, even autocratic ones, have become sufficiently differentiated to also feature education systems with distinct rationales and routines. Only in the case of a complete overlap between political and educational system – which arguably is the case merely in totalitarian societies – would teachers act as direct agents of the state.

Yet, as the example of Teacher Zhu and other fieldwork observations show, politics does matter, and it matters to a particularly large extent in the education system. The reason for this is to be found in the nature of the different subsystems within the education system: Apart from the legal and administrative management of education in the form of laws and regulations, education, on the one hand, materializes in a school system, and is implemented, on the other hand, in the form of an instructional system. While school systems are strongly intertwined with both legal and administrative systems and have thus been displaying considerable inertia when responding to changes, instructional systems are much more dependent on ad hoc, face-to-face communication and interaction, and are hence more prone to change (Vanderstraeten, 2003). This means that teachers, despite their reliance on teacher training and their knowledge of regulations, need to decide rather spontaneously what kind of class interaction is pedagogical, or ‘good’, and what is unacceptable, or ‘bad’. Mostly, these daily operations occur unproblematically, and contribute to the (partial) independence of classroom instruction. What is expected of the system and how its agents actually operate constitutes, to a certain extent, a process of “loose coupling” (Gaus & Drieschner, 2014).

How are processes of loose coupling to be understood when related to spaces and enactments of autonomy and agency? To answer this question, we need to be able to distinguish ‘loose coupling’, which would be situated within the education system (or instructional subsystem), from ‘interference’, which would point to a larger degree of porousness, or weakness, of the education system. ‘Loose coupling’ occurs when teachers make active choices, in congruence both with their self-understanding as teachers and with the constraints and options that characterize the specific situations in which they need to act; it thus happens within the reflection processes typical of the instructional situation. ‘Interference’, in contrast, constitutes a situation in which this reflection process is interrupted by forces that are external to the education-instructional system and that are beyond the control of educational agents (such as teachers). In a different context – namely with regard to inter-national rather than inter-system interactions – Schriewer (2014, p. 92), drawing upon a system-theory approach, has pointed to the centrality of “interruptions in relations of interdependence” and “externalization” when it comes to breaking up, and to some extent, disturbing processes of reflection and self-reference:

Such interruptions typically take the form of the reflection and communication process opening itself to its external environment, however selectively this may be done, for it is through the incorporation of “supplemental meaning”, as extractable from external points of reference, that circular self-reference becomes amenable to specification (Luhmann 1995a, 466). (Schriewer, 2014, p. 93)

Externalization is a powerful concept to account for change: If the education system (and the instructional sub-system) were a forever self-referential, autopoetic system, any change would be an impossibility. Hence, a certain degree of porousness is necessary for a system to undergo any kind of change. This becomes the case when existing modes of reflection are no longer considered sufficient to handle educational/instructional situations, and “supplemental meaning” needs to be fetched from outside the indigenous system in order to find adequate solutions elsewhere. For example, Teacher Zhu, perceiving the present-day, exclusive focus on exams to be detrimental to educational and pedagogical objectives, externalized to ideologies outside the education system in order to restore what she conceives as the pedagogical spirit.

However, as pointed out in the previous section, system theory is interested in intra-/inter-system communication, not in questions of autonomy/agency from the perspective of these systems’ agents. Connecting the very useful concept of ‘externalization’ with those of autonomy and agency as developed above, we can establish that pure instances of ‘loose coupling’, without any kind of ‘interference’, are located in the upper right quadrants of Fig. 3.2, as these denote the cases in which teachers can retain both their professional autonomy and their agency. While in a strong educational system, without an intrusive state, such a teacher may (perhaps tautologically) be called ‘professional’, in societies such as the Chinese one, ‘renitent’ teachers would fulfill an equivalent function, as they would enact their self-defined teacher identities by drawing on professional norms and routines, however threatened these norms and routines might be. In all the other quadrants, teachers inadvertently experience some form of infringement: Either other systems (such as the political) override the distinctive rules of the educational profession; or these rules are in fact enforced, but at the expense of the teachers’ self-defined identities.

Such a differentiated approach towards autonomy and agency, as they operate in different contexts and draw on different strategies of externalization (or change), is also useful for distinguishing ‘agency’ from ‘empowerment’, or even from some sort of positive force contributing to grassroots democracy in favor of students. In some cases, and depending on the respective norms and values of the interacting systems, teacher agency may be conducive for student empowerment. For example, the “renitent” and “unorthodox” teachers in Fig. 3.2 may be imagined as agents who, sometimes in spite of all odds, nourish a sense of democracy or civic awareness amongst students. In many other cases, teacher agency may just as well exacerbate practices of disempowerment and oppression. As also Imants and Van der Wal observe, agency “should not be treated a priori as a positive factor for reform and development” but can instead result in (to the external observer) “inadequate teaching practices or beliefs about teaching” (Imants & Van der Wal, 2020, p. 4). Even though Imants and Van der Wal have in mind teachers who resist educational reforms and development, their observation is equally valid regarding teachers who overzealously respond to political (or other) ideologies.

Conclusion: The political in Policy Implementation – And Policy Implementation Research?

This chapter has presented a nested approach towards autonomy and agency, taking into consideration, on the one hand, the interaction between different, distinct systems when spaces of autonomy and enactments of agency become operative, and, on the other, paying attention to how different degrees of autonomy/agency, when contextualized in specific educational and political environments, allow for different types of teachers and different forms of (non-)interference. It thus attempts to reconcile a systemic perspective with that of individual and collective agency. It also relativizes, on one side, the view that processes of policy implementation are to be seen as hierarchical mechanisms of policies ‘trickling down’ from the top to the bottom; and, on the other, the perspective that policy implementation is to be understood as largely a process of appropriation and indigenization on the ground. While the latter approach is in a sense a truism – there can be no policy implementation without some sort of local processing of the respective policy – it has so far been insufficiently conceptualized, since policy appropriation has been mainly subsumed under the rather generic concept of local agency, without however clarifying how the concept of agency can be understood in relation to different systemic and agentic constellations.

This chapter has argued that such a differentiated approach is necessary in order to better understand, and assess, the ramifications of teacher agency and professional autonomy in diverse contexts. Contrary to an understanding of autonomy as the absence of regulation, in which teachers then develop real or false agency – as maintained by Priestley et al. (2015) – the chapter emphasizes the highly regulated and specialized nature of autonomous spaces, where rules and norms are the prerequisites for building and maintaining autonomy, and for creating spaces for agency in alignment with professional norms. Depending on how sharply demarcated the borders are between the education system and other systems, and on how subsystems such as the school are able to operate according to the educational system’s principles of autonomy, teachers have different options for developing various forms of agency, which in turn are fed by different norms and values. The chapter hence attempts to complement Wermke et al.’s (2019) conception of autonomy as emerging between decision-making and control, by systematically addressing the question of whose rules, norms, and values are being enacted in processes of decision-making and control.

Ultimately, the policy-practice nexus is deeply political, if we understand politics – in line with our argument above, with reference to the politics of use – as a practice of signaling and enacting particular norms and values. Norms and values are (co-)produced both individually/collectively, in processes of social interaction and (self-)definition, and systemically, in processes of specialization and differentiation. Therefore, any kind of (inter-)action, including that of teachers, will need to resort to those norms and values that are, firstly, compatible with the respective agents’ environments; secondly, available to these agents as resources in specific (long-term and short-term) situations; and, thirdly, desirable to agents as meaningful instruments of legitimizing their actions.

In conclusion, we may pose the question of why much of the literature on teacher agency and professional autonomy has tended to neglect this political dimension. A straightforward answer could be that empirical research on teachers and school systems in less democratic contexts has not (yet) succeeded in theory-building: While adding to our knowledge about what is happening in these contexts, findings from these studies have not been sufficiently brought into a conversation with concepts and theories developed elsewhere. As a tentative and perhaps provocative conclusion, I would like to put forward a different explanation: part of the reason for this neglect may also lie in our own ideological blindfoldedness regarding conceptions of agency and autonomy. As already critically noted by Abrams (1999), autonomy as a concept has been largely developed within the liberal tradition, treating relative values such as authenticity and freedom as if they were absolute truths. Likewise, it may be argued that ideas revolving around ‘agency’ have been blended largely with normative conceptions, fed by the (latent or explicit) conviction that agency is a desirable objective in itself. Much of the normatively framed debates on education today are deeply entrenched in the emancipatory tradition of pedagogy (see e.g. Cloonan et al., 2019; Samoukovic, 2015). In this tradition, teacher agency is often conflated with learner autonomy (Benson, 2007), and is therefore considered inherently good. However, research on cultural and political contexts characterized by traditions and rationales that contrast starkly with, or have moved away from, the emancipatory perspective reveals that more agency for teachers, or even more ‘participation’ for students, can result in increased indoctrination, rather than more autonomy or empowerment (Schulte, 2019). Therefore, an analytical, rather than normative, approach towards agency and its interrelationship with autonomy in diverse systems can help denormalize our own ways of thinking about agency and autonomy, and it can enhance our capability to compare policy implementation processes across a broad variety of cases.