Externalisation and structural coupling: Applications in comparative policy studies in education

Abstract

The article presents two key concepts of sociological systems theory – externalisation and structural coupling – and applies them to explain (a) the exponential growth of international large-scale student assessments and (b) the rise of ‘policy-relevant’ educational research. The author concludes with a comparison between key concepts used in systems theory and those used in comparative policy studies. She identifies resemblances with concepts of pathways in historical institutionalism, the multiple-streams approach and the notion of punctuated equilibrium advanced in the advocacy coalition framework.

Arguably, systems are constantly exposed to external shocks and to challenges from within. According to the interpretive framework of sociological systems theory, they absorb, however, such ‘irritations’ selectively or idiosyncratically by translating them into their particular system logic. This article examines how the education system has dealt with the fast advance of (a) international large-scale assessments (ILSAs) and (b) evidence-based policy planning. These two recent ‘irritations’ serve as examples to demonstrate the usefulness of an interpretive framework that focuses on systems rather than individuals or groups as units of analysis. I draw on two key concepts – externalisation and structural coupling – to shed light on the system logic in education.

Sociological systems theory forces us to dig deeper and move beyond simply acknowledging that ILSAs primarily solidify test-based accountability with little impact on actual student learning outcomes (Verger et al., 2019) or that evidence-based policy planning needs to be seen as politics in disguise, generating, more often than not, policy-based evidence rather than knowledge-based policy decisions. An interpretive framework that focuses on how systems frame or translate current trends, such as ILSAs and evidence-based policy planning, into their own (system) logic helps us to understand and to problematise the educational meaning-making of broader trends in society.

Some common pitfalls in understanding sociological systems theory

It is important to bear in mind that Niklas Luhmann’s interpretive framework is quintessentially a sociological theory. Luhmann makes in his published work an unabashed commitment to understanding how social systems – rather than how individuals inhabiting these systems – operate. His is the sociological project of understanding social systems rather than how the ‘five billion psychological systems’ (individuals) communicate, observe, or act (Luhmann, 1990: 78, cited in King and Thornhill, 2003: 6). Confounding the two units of analysis, has apparently been

an issue that time and again causes difficulty. One can say it a hundred times, but it is always in vain. The observer is not automatically a psychic system; he is not automatically a consciousness. (Luhmann, lecture 7, winter semester 1991–1992, printed in Baecker, 2013: 105)

For systems theory, it is not the collective of individuals that make up a society, but rather the various subsystems with their respective environments: the subsystems of education, politics, law, religion and so forth. These subsystems observe and differentiate themselves from other subsystems to generate information. All the subsystems of society communicate in their idiosyncratic ways with their own (subsystem-specific) environment and thereby generate shared, societal meaning. Societies, as well as organisations and interactions, operate in an idiosyncratic manner and are, to use systems-theoretical vocabulary, self-referential or autopoietic systems (Luhmann, 1995).

To understand systems theory, readers must also kick the habit of intuitively inserting the adjective ‘national’ whenever they come across the term ‘the educational system’. For Luhmann, there is only one universal system of (world) society and only one universal economy, science, legal, political, or education subsystem or function system. ¹ This is not to downplay the differences (sometimes vast) in how these different function systems are organised at the national level. In the function system of education, for example, the normative beliefs of how schools should be organised and what should be taught in them vary considerably. Contextual or national differentiation at the organisational level is acknowledged but ultimately of only slight interest to sociological systems theory. Luhmann is primarily concerned with understanding a system’s performativity; that is, how it observes and differentiates itself from its environment or other subsystems and how it communicates that difference (Luhmann, 2006).

By definition, a system is at the same time operatively closed and cognitively open. As a result of functional differentiation in modern societies, only the subsystem of education educates, only the subsystem of politics governs, or only the legal subsystem issues laws. Concurrent with operative closure, a system observes other subsystems and communicates about its environment using its own code and logic. In a constant movement between inclusion and exclusion, a system solidifies its identity by means of boundary setting; that is, it distances and thereby differentiates itself from other systems. Interestingly, the system’s observation of the difference between itself and other systems also includes the feature of ‘second-order observation’, in that systems observe what other systems observe and therefore observe how the other systems establish difference vis-a-vis their respective environment. Theoretically, the observation of observation may continue ad infinitum, leading to the paradoxical situation that a system observes another system’s observation of the observation of other systems. As will be explained later, externalisation generates ‘additional meaning’ (German: Zusatzsinn) and thereby disrupts the endless circularity of observation and self-observation, or to use Luhmann’s terminology, has a salutary effect on the ‘pathology of the paradox’ (Luhmann, 1994).

Changes occur as a result of differentiation. In fact, one of the prominent features of modern society is functional differentiation. Modern society then consists of functionally differentiated systems, all operating with their own codes, identity and modes of regulation. The function systems are closed vis-a-vis other function systems, but they are interdependent. For example, even though the education system has its own education function, its own organisations (schools), its own actors (teachers, students) and its own modes of regulation (until recently driven more by moral and legal rather than market considerations), the function system of education both depends on and contributes to the functioning of other function systems. For example, it depends on the financial resources made available in the economic system.

Three examples should suffice to illustrate the interpretive framework of systems theory. First, a historical account of modern schooling helps to illuminate the process of functional differentiation. A result of the Enlightenment and the nation-state, modern schooling introduced the new social category of a ‘school class’ (Luhmann, 2002: 119–121, cited in Mangez and Vanden Broeck, 2020), which encompassed, regardless of social class, all individuals that the state considered to be future citizens of the nation. The discovery of the child as a human species that, regardless of social background, deserves special care and attention helped suspend the earlier focus on ‘single rank categories’ (Vanderstraeten and Biesta, 2006: 162; see also Vanderstraeten, 2006), such as the education of princes, noblemen, knights and future sovereigns, taught by private tutors, or first-born sons, taught by priests. At the same time, modern schooling displaced earlier educative sites, notably the family, the community and the church, and, in fact, rendered them illegitimate. In a continuous process of professionalisation and specialisation of education, schools came to be seen as the only organisational form where literacy, numeracy and other relevant skills for living in a modern society were supposed to be taught, thereby disempowering all other educative sites, which were downgraded and framed as non-formal and unprofessional education. Second, systems differentiated themselves by pursuing specific functions and delegating other, unrelated functions to other systems. The function system of education, for example, is to educate students (meant in terms of a social category rather than in an age-specific sense). As Luhmann asserts, the education system neither needs to, nor can, generate money, nor exert power, nor produce research for society at large (2002: 14). The functions are divided among systems, and all are considered equally important. Third, in a quest to reduce uncertainty, interruption and perturbation, a system receives and translates demands for change in a selective or self-referential manner. In other words, function systems observe and react to each other, but are not able to communicate with each other because each one of them is bound by its own code and language of communication. Finally, systems are not only self-referential but also self-reflective and self-aware. Systems observe their environment and, in an effort towards ‘boundary maintenance’ (Luhmann, 2006: 38), differentiate themselves from other systems.

Externalisation and structural coupling

In this article, I apply the systems-theoretical concepts of externalisation and structural coupling to explain the proliferation of (a) international comparative studies (in the form of ILSAs) and (b) evidence-based policy planning. For the first example, I draw on the concept of externalisation to explain how the function system of education, at moments of uncertainty, is receptive or temporarily opens up to other function systems or to the generalised other (‘the world’), only to then reframe or translate these (quasi-) external impulses into its own code and logic (Luhmann, 1995, 1997, 1998). The concept of externalisation lends itself to understanding the explosive growth of ILSAs (such as the Programme for International Student Assessment (PISA), Trends in International Mathematics and Science Study (TIMSS) and Progress in International Reading Literacy Study (PIRLS)), oftentimes propagated and ‘sold’ as indispensable tools for policy learning and borrowing. For the second illustration, I use the notion of structural coupling to explain the proliferation of evidence-based policy planning in education, discussed against the backdrop of structural coupling between the function systems of science, politics and education.

ILSAs: a (quasi-) external source of authorisation for national school reforms

The spectacular growth of countries participating in ILSAs is noticeable and deserves theorising. In terms of PISA alone, 43 countries participated in 2000, 72 countries and territories in 2015 and 80 administrative entities in 2018.

In analysing the explosive growth of PISA and other international comparative tests, researchers have presented several narratives on why there is a demand for ILSAs, ranging from broad ones, such as globalisation and the political pressure on governments to be part of a larger international educational space, to more specific ones, such as an ever-increasing number of governments that are pressured into demonstrating their commitment to ‘(test-based) accountability’ by participating in ILSAs (see Addey et al., 2017; Verger et al., 2019). A new body of research has also emerged that illuminates the supply side: what do the Organisation for Economic Co-operation and Development (OECD) and the International Association for the Evaluation of Educational Achievement (IEA) undertake to continuously generate a demand for ILSAs? This is not a moot point, given the enormous financial resources that participating governments are expected to periodically allocate for the administration of the tests. As part of the supply-side scrutiny, we found, for example, that the World Bank’s Human Capital Index penalises countries that do not participate in ILSAs (Liu and Steiner-Khamsi, 2020). According to our study, countries that choose not to participate in every round of an ILSA have to expect to score 28 to 37 points, or about one full year of learning lower on the World Bank’s scale relative to countries with full ILSA participation.

The pragmatic focus on demand- and supply-side rationales for participation in ILSAs should be complemented with longer-range theoretical reasoning. In particular, the system-theoretical concept of externalisation helps to complicate the pragmatic narratives on reasons for ILSA participation. The group of researchers in comparative education that apply the ‘externalisation thesis’ (Schriewer, 1990) to policy borrowing (i.e. to ‘other-reference’) (Luhmann, 1998: 98) and references to the ‘world’ (Luhmann, 1995: 437–477; see also Luhmann, 1990, 1997) has visibly grown over the past few years (see Steiner-Khamsi, 2004; Steiner-Khamsi and Waldow, 2012).

As with other systems-theoretical notions, the ecological perspective is also manifest in the concept of externalisation. Since systems are considered to be operatively closed but cognitively open, everything around them is environmental and therefore observable. The environment is potentially a constant source of ‘irritation’ and helps solidify self-reflection, self-representation and self-reproduction of the system. For Luhmann, autopoietic or self-referential systems are meaning-producing entities that make sense out of the (nonsensical) irritations, disturbances and noise which they experience around them:

In the context of autopoietic reproduction, the environment functions as irritation, as disturbance, as noise, and it only becomes meaningful for the system, when it can be related to the networks of decisions of the system. (Luhmann, 1992: 173, cited in Vanderstraeten, 2002: 246f.)

In the same vein, self-referential systems produce their own causes and determine what counts as internal and external causation (see Fuchs, 1988). External causes – such as the neoliberal move to outcomes-based governance and knowledge-based regulation in public administration – are effective only if they resonate internally, are rendered meaningful and are translated into the logic and language code of the system. In education, the orientation towards outcomes was translated into the shift from teaching to learning and from government to governance (Fenwick et al., 2014; Mangez and Vanden Broeck, 2020), a move that ultimately resulted in a ‘datafication’ of learning (Thoutenhoofd, 2018). Whether the class size was large or small, teachers qualified or unqualified, sufficient resources allocated to schools or not (to list only a few input indicators) was rendered irrelevant compared to the question of whether students had achieved the prescribed learning outcomes. The continuous testing of students became the primary policy tool for monitoring quality and equity in education; that is, rephrased in the terminology of systems theory, a technology for reflection in the system of education.

Arguably, educational organisations are, due to their inherent ‘technology deficit’ (Luhmann and Schorr, 1979), susceptible to non-education reference points or an ‘Ersatz Order’ (alternative order; see Vanderstraeten, 2002: 248). Schools or national education systems prepare students in the present for an unknown future and therefore need to continuously revisit their socially agreed upon criteria of what a ‘good student’ is. Today, the education system copes with its technology deficit by implementing the norms established by a collective of economists, educationalists and psychometricians who spell out the 21st-century skills which good students are supposed to demonstrate when assessed by PISA (see Schleicher, 2017). The ‘substitution of learning by data’ (Thoutenhoofd, 2018: 435) resonates with two larger irritations that the education system has experienced over the past few years: the semantics of globalisation and knowledge-based regulation, explained later in this article.

Making sense educationally of changes that occur in society is only one aspect of the self-referential operations of the education subsystem. As mentioned repeatedly, such operations presuppose an observation of the environment against which self-references are made. However, self-references are trapped in circularity. For Luhmann, ‘every type of self-reference encounters the problem of breaking out of a merely tautological circle’ and therefore ‘mere reference from a self back to itself must be enriched with additional meaning’ (Luhmann, 1995: 466). He draws on the concept of asymmetrisation and its related terms (externalisation, finalisation, ideologisation, hierarchisation and punctuation) to explain how additional meaning is generated to disrupt the ‘tautology of pure self-references’ (Luhmann, 1995: 466).

In other words, systems choose to externalise at particular moments when there is a need to break the tautological cycle of self-reference and self-affirmation. The system’s ‘internal construction of external points of reference’ (Mangez and Vanden Broeck, 2020) generates additional meaning and helps reproduce its identity. According to Luhmann (1995: 481), externalisation or the construction of such quasi-external points of reference serve as ‘regulative ideas’, which on closer examination are ‘projections’ which ‘are valid only “as if” they were valid’. The point of (system-) internally produced (quasi-) external references is key for an understanding of Luhmann’s conception of externalisation.

As a corollary, reference is an act of observation that serves the system to distinguish itself from the other. In the same vein, systems also observe themselves from a quasi-external perspective to better communicate their distinct logic to themselves and to others. Both self-reference and other-reference (German: Fremdreferenz) are intrinsically linked to the act of observation and self-referentiality. As Luhmann asserts (Luhmann, 1998: 98), ‘[t]he system reproduces itself in the imaginary space of its references and uses the differences between self- and other-references as an impetus to reproduce itself’.

Given the principles of self-determination, self-referentiality and the interdependence of systems, the questions of when and how externalisation occurs are relevant. In policy-borrowing research, we use the term ‘reception’ to label this particular moment of selective or system-specific adoption and we use the term ‘translation’ to denote the act of system-specific understanding or sense-making.

It is important to bear in mind that sociological systems theory does not attempt to be an action theory. Quite to the contrary, Luhmann finds regulatory science, steering, or ‘the theory of planning’ to be ‘in a desolate state’ (Luhmann, 1997: 41; see also Luhmann, 2017; Luhmann, 2019: chapter 11) and diametrically opposed to his notion of ‘self-steering’ of the system. Against the backdrop of self-determined and interdependent systems, we have a keen interest in comparative policy studies in understanding how decisions in national educational systems are made and justified. Given that only one function system of education exists, researchers in policy studies often draw their attention to processes of secondary differentiation or segmentation to use national education systems or organisations as their units of analysis. As pointed out astutely by Armin Nassehi (2005: 179), any system-theoretical analysis of organisations must begin with why and how questions, such as why organisations tend to describe themselves as rational and efficient. Dissecting self-descriptions of organisations provides important clues on how a system operates and how it justifies its decisions.

As expected with any coherent interpretive framework, including sociological systems-theory, a multitude of research questions open up when we investigate, for example, how policy actors make use of ILSAs: at what particular moments do systems externalise, or observe and reference others? Whom or rather what do they choose as their object of observation? What do they do with the observation or reference? How do they (back-) translate it to fit their own (system) logic? To apply the system-theoretical concepts mentioned above to national education systems, policy-borrowing researchers have pursued empirical studies on externalisation. They have investigated the particular moments at which national education systems seek out (additional) information and generate additional meaning by making a reference (labelled ‘reception’ in policy-borrowing research) to other subsystems, other national education systems, or to the ‘world’ (broadly defined), as well as how they frame, or fill the meaning of the reference (labelled ‘translation’ in policy-borrowing research) in a system-specific manner. Meant to pique the reader’s curiosity, these questions merely represent a small sample of questions that arise when the concepts of system self-referentiality, operative closure, cognitive opening and externalisation are applied to concrete examples from comparative policy studies.

Policy-borrowing researchers that take their inspiration from sociological systems theory tend to be sensitive to boundary work, which is produced by organisations or policy actors, respectively. Even though only one system of education, one system of politics and one system of science exist, each with its own code, logic and mode of regulation, a secondary differentiation of politics is divided into national segments. At particular moments in a given context, national policy actors use references to experiences in other national education systems (e.g. Finland, Singapore), other function systems (e.g. economy, science), or the world (e.g. best practices, international standards) as a way to generate leverage for their own national reforms. These types of externalisation become tools of quasi-external, self-induced certification of national reform efforts. For this reason, these researchers have made it their intellectual project to understand the rapid growth of standardised international comparisons against the backdrop of what is debated, contested, or at stake internally; that is, in a local policy context. As a corollary, the focus lies on exploring the ‘socio-logic’ in policy reception and translation processes (Schriewer and Martinez, 2004: 33).

Methodologically, the conceptual focus on externalisation implies an empirical inquiry into how, when and by whom (explicit) references are made. In comparative policy studies, in particular, the investigation entails analyses of which countries are referenced, when a reference is made and how the reference is interpreted. These items of analysis include whose publications (including the publications of which organisations) are cited as authoritative texts, which education systems are considered as sources for emulation, within which policy context a reference to ILSA results is made and how policy analysts in a given country make sense of ILSA results. Several system theorists use the rapid growth of standardised international comparison as an opportunity to understand why PISA resonates and how PISA results are translated or interpreted in various policy contexts. We do not assume that PISA a priori has a salutary effect on national school reform, but we rather analyse why, how and when national policy actors ‘open up’, or are receptive to the system’s ‘irritations’ caused by PISA, TIMSS, or by any other ILSA and how they subsequently translate these external impulses into the language of the organisation or the national policy context.

The focus on the idiosyncrasies of a system and its national forms of organisation brings a fascinating phenomenon to light that at first sight appears to be contradictory: despite the widespread rhetoric of learning from ‘best performing’ school systems, no universal consensus has been established on why some school systems do better than others on tests such as PISA. On the contrary, great variations exist in terms of how national governments, media and research institutions explain Finland’s, Shanghai’s, or Singapore’s ‘success’ in PISA or TIMSS. However, a pattern can be discerned in these varied, sometimes diametrically opposed, explanations, which is best captured by the term ‘projection’ (Waldow, 2017; Waldow and Steiner-Khamsi, 2019). The reception and translation of ILSA results reflect controversial policy issues in a country’s own policy context, rather than the actual organisational features of the league leaders.

‘Finnish success’ is a good case in point. A long list of explanations is given for why Finnish students do well on ILSAs. Depending on what the controversial policy issue is for which policy actors seek an (internally induced) quasi-external source of authority, Finland’s success is alternately attributed to its strong university-based teacher education system, the system of comprehensive schooling with minimal tracking of students, or the nurturing environment in schools where students are ironically exposed to few high-stakes standardised tests.

The same applies for the league leaders themselves: depending on the timing, notably whether the positive results are released at the end or the beginning of a reform cycle, the policy actors tend to take credit for the positive results or, on the contrary, belittle the success with the explanation that the students performed well for all the wrong reasons, including private tutoring, being in a stressful school environment and learning to the test (see Waldow and Steiner-Khamsi, 2019).

The science–politics–education nexus: the emergence of policy-relevant educational research

A much-discussed phenomenon in sociological systems theory is structural coupling. Structural coupling per se is nothing out of the extraordinary. Perhaps a banal yet illustrative example, the legal system and the economic system are structurally coupled through sales contracts (Andersen, 2000, 2013). Despite their structural coupling, both systems operate autonomously and preserve their own logic and their own code. For example, the legal system communicates in the legal code: it frames or translates all phenomena in terms of the code just/unjust. The code of the education system is more complex and difficult to determine because students are prepared for the future, but the future cannot be observed today. Nevertheless, the binary construction of learning/not learning is part of the primary code (Luhmann and Schorr, 1979), followed by selection or gate keeping as a secondary code

In educational research, two recent forms of structural coupling have become objects of intense scrutiny: the coupling between the education and economic system and the coupling between the science and political system, discussed in this section of the article. As mentioned above, systems absorb complexity and irritations by making sense of them in a system-logical, meaningful manner. As a result of structural coupling, the education system has become economised, and, vice versa, the (knowledge) economy nowadays takes into account ‘effective years of schooling’ (one of the indicators of the Human Capital Index) as a predictor of economic productivity (see World Bank, 2018, chapter 3). Similarly, as a result of structural coupling between the science and political systems, an additional type of knowledge production has emerged in education: so-called policy relevant research, also known as applied or ‘mode 2 research’ (Nowotny et al., 2003). In contrast, mode 1 research represents the traditional type of knowledge production, also known as foundational research or basic research. Unsurprisingly, policy analysts and politicians are enamoured with mode 2 research. In fact, the political system nowadays tends to explain and justify its political decisions with a recourse to ‘evidence’. Even though the mutual transformation process – the politicisation of science and the scientisation of politics – is generally acknowledged, more empirical studies have offered investigations of the politicisation of science, also known as agenda-driven research, than of the scientisation of politics.

The changing nature of the relationship between politics and science has preoccupied comparative policy studies for a while. One of the early, important comparative studies exploring the interpenetration of the two function systems was the research project ‘The Role of Knowledge in the Construction and Regulation of Health and Education Policy in Europe: Convergences and Specificities among Nations and Sectors’, abbreviated as Knowandpol, and funded in the Sixth Framework Programme of the European Commission (e.g. Fenwick et al., 2014; Maroy, 2012). Among other foci, the Knowandpol research project examined the changing role of the state in the wake of new public management reforms and knowledge-based regulation. The changes were substantial. The role of the state has changed from being a provider of educational programs to a standard setter and monitor of learning outcomes. As a result, a multitude of providers, including businesses, have nowadays entered the school market as providers of educational programs. Knowledge-based regulation also enlarged the radius of individuals contributing to policy-relevant educational knowledge. The open-access policies that governments and research councils have put in place in recent years need to be seen as an early indication of changes that occurred in knowledge production and sharing.

Arguably, ad hoc expert commissions, which review policies and make recommendations to the government, lend themselves to the study of coupling between the science and politics systems. According to Weingart and Lentsch, these government-appointed commissions have experienced three distinct shifts over the past 70 years (2008: 207 ff.). During the early period of scientific policy advice (1950s to 1970s), the ad hoc expert commissions insisted on being autonomous and independent from government. As a corollary, their reports amassed foundational studies (mode 1 scientific knowledge) that policy actors could or could not use, respectively. In a second phase, the commissions became increasingly politicised (1970s to 1990s) because they were charged with the task of producing policy-relevant scientific knowledge. In the current, third phase, governments in many countries have experienced a shift from ‘knowledge-based legitimacy’ to ‘participation-based legitimacy’. Governments are under pressure to democratise scientific policy advice by (a) providing open access to reviews and expertise; (b) expanding the definition of ‘experts’ (including, nowadays, both producers and users/consumers); and (c) insisting that knowledge products are useful; that is, they provide a clear foundation for stop/go policy decisions.

It is important to point out that the democratisation of expertise has led in practice to a (pseudo) rationalisation or scientification of political decisions (Maasen and Weingart, 2005) and is driven by, and at the same time exacerbated by, governance by numbers (Grek, 2008; Ozga, 2009) and steering at a distance (Rose and Miller, 1992). The facade of rationality has been thoroughly dismantled in policy studies and includes critics who shed doubts on whether governance by numbers is less political or more rational than other modes of regulation. Finally, a few scholars have examined the impact that this particular type of structural coupling has had on the relations between different levels of bureaucracies within the state. Most recently, Piattoeva et al. (2018: 115ff.) convincingly argued that the governments of Brazil, China and Russia (the objects of their study) resort to ‘governance by data circulation’ in an effort to reach out to district authorities.

In one of our empirical bibliometric network analyses, we examined whether government officials in Norway produce and use scientific knowledge in their publications when they present the rationale for a school reform (Steiner-Khamsi et al., 2019). Norway is a good case in point to study the phenomenon because it has experienced an academisation of the government-appointed ad hoc commissions, known as Norwegian Public Commissions (Norges offentlige utredninger or NOU in Norwegian). These commissions or NOUs are in charge of reviewing past reforms and proposing recommendations for future reforms to the Ministry of Education and Research (Christensen and Holst, 2017). Norway also stands out as the country where, according to the Eurydice Report (European Commission/EACEA/Eurydice 2017: 10), ‘evidence is used at all stages of the policy process’.

First, we found that the number of academics appointed to serve on such commissions has dramatically increased over the past 20 years. Our observations are in concert with a larger study on expert commissions that Christensen and his associates conducted (Christensen and Holst, 2017; Christensen and Hesstvedt, 2018). They examined the composition of expert commissions or ad hoc commission of various line ministries over a period of nearly 50 years. What they found is astounding: the interest group representatives in the NOUs have dropped sharply, by more than half, over the past four decades. Over the same time period, the share of academics on ad hoc commissions has dramatically surged. In the 1970s, academics accounted for only 7% of commission members. In the first decade of the new millennium, the share of academics across line ministries in Norway increased to an average of 26%. Today, the three largest groups serving on education ad hoc commissions are civil servants, followed by academics and, finally, interest groups (7.9%). The linear increase of academics from 1972 to 2009 and with explosive growth since 2010 has led Christensen and Hesstvedt (2018) to suggest that more research is necessary to understand the ‘expertisation’ in NOUs in greater detail.

Second, our bibliometric analysis of the 2006 reform indicates that both the Ministry of Education and Research and its appointed NOUs apparently feel compelled to substantiate their policy decisions and policy recommendations, respectively, with ample evidence, as manifested in the long list of references published in the White and Green Papers. In every reform period (reforms of 1996, 2006 and 2020), the number of references drastically increased (Baek et al., 2018; Steiner-Khamsi et al., 2019). The White and Green Papers prepared for the 2020 period more than doubled their use of references compared to the previous reform period. The 2006 reform averaged 96 references per White or Green Paper, compared to 234 references on average for the 2020 reform. Strikingly, the pressure to back up policy recommendations or decisions with evidence did not exist during the first school reform period of 1996. The few references included in the 1996 White and Green Papers were embedded in the actual texts; that is, not listed in a separate reference section.

Finally, we also explored the political translation process; that is, the transfer from scientific knowledge (produced by expert commissions) to political knowledge (produced by government) by asking: from all the references listed in the Green Papers (produced by the government-appointed expert commission), how many are also shared, that is, cited in the White Papers (produced by the Ministry of Education and Research)? In other words, we examined the political translation of expert knowledge. The finding was unexpected: the Ministry of Education and Research uses surprisingly few texts compiled by its expert commissions. Of the 232 texts that the two 2006 White Papers cited, only 22 of them were referenced in the four commission reports (9.5%). This means that 90.5% of the commissions’ body of knowledge was lost in (political) translation. The disregard of knowledge amassed in the Green Papers is not to be underestimated. The question then becomes: which texts constitute the 22 citations that the White and Green Papers have in common?

A closer examination revealed that 20 of the 22 citations were from a single commissioned report entitled In the First Row (NOU, 2003). The importance of this particular Green Paper can be explained by its specific mandate, the composition of the commission and its focus on a particular kind of knowledge: policy-relevant knowledge. The commission was tasked with producing an overall analysis of the education system and recommendations regarding how to improve the quality of education. Strikingly, the report focuses a great deal on recommendations from the OECD and the DeSeCo project (Definitions and Selections of Competencies) and makes a case for a competency-based curriculum. In addition, it reconfirms the need for a national testing system that would allow for monitoring quality improvement in schools. The global language of the OECD is not to be overheard, and the report comes across as an ‘indigenisation’ or a national translation of OECD education policies. Both recommendations of the Green Paper (NOU, 2003) – the shift toward competency or outcomes-based curriculum reform and the introduction of test-based accountability – carry features of school reforms that the OECD has propelled globally, including in Norway, and merit the label ‘global education policy’ (Mundy et al., 2016), or in this particular case, OECD education policy.

In terms of systems theory, the finding is remarkable for two reasons. Similar to the governments who selectively draw on ILSA results to make a case for their own national reform, as discussed above and in other publications (e.g. Waldow and Steiner-Khamsi, 2019), the Ministry of Education and Research of Norway established numerous ad hoc commissions yet only ‘listened’ to, or rather cited, the sources from the one (OECD-oriented) report that provided a quasi-external stamp of approval for introducing a fundamental, competency-based curriculum reform. The finding is also relevant for understanding what evidence-based and knowledge-based regulation means in practice. In the case of the two relevant government-issued White Papers for the 2006 reform, the Ministry of Education and Research consulted or cited publications in a self-referential manner. The majority of references (61%) were either policy documents or policy-relevant research (known as the Institute Sector), sponsored by the government.

This second example, which is intended to illustrate the concept of the structural coupling of two or more function systems, is manifested in the emergence of knowledge-based regulation (structural coupling between science and politics) and, if education is added to the mix, in the proliferation of policy-relevant educational research (structural coupling between politics, education and science).

Interpretation and application of a theory

In 1988, Stephan Fuchs credited Niklas Luhmann with presenting a new theory of social systems of ‘universal range and application’ which ‘is the only general social theory that can claim to introduce a new paradigm in the field’ (Fuchs, 1988: 21). The times have changed. Not only have the social sciences survived the era during which ‘grand narratives’ fell out of fashion and came under attack for their ‘totalising narrative’ (Lyotard, 1979), but we are witnessing today, at least in comparative education, a rekindled interest in social theory. By way of a disclaimer, but also to be situated in the broader theory debates, justifying how one presents and applies a theory is necessary nowadays. As sketched below, this article reflects the assertion that theories are continuously refined and that comparing elements of system-theoretical thought with concepts in other social theories is useful for helping to advance the theory debate within sociological systems theory and within comparative policy studies.

Clearly, theories serve to explain (new) phenomena. With every new application, a theory becomes more meaningful for the research community. For example, one would be hard-pressed to find the term ‘externalisation’ explained in detail in one particular location of Luhmann’s 60 books and 400 articles (Khan, 2014). Compared to and derived from the fundamental system-theoretical principles of system closure, self-referentiality and interdependence, his discussion of the concept of externalisation is only brief and interspersed throughout his writings. Externalisation as a means to break the tautology of self-referentiality is sometimes also explained in the context of asymetricisation (Luhmann, 1995). Despite the scant discussions of the concept per se, the ‘externalisation thesis’ (Schriewer, 1990) is extensively used in comparative education. I adopted the term and applied it to comparative policy studies because it helps to explain how organisations at particular moments (that deserve empirical scrutiny) observe and selectively adopt (quasi-) external impulses, such as global education policies, best practices or international standards, and translate them into their own logic. Similarly, the operation of generating ‘additional meaning’ (German: Zusatzsinn) to disrupt the tautology of self-referentiality is creatively applied in policy-borrowing research to examine the salutary effects of externalisation on protracted conflict. Finally, the preoccupation with national systems is uncommon in systems theory as it captures a secondary, organisational segmentation which is not treated with a priority. Despite all these disclaimers, policy-borrowing research is informed by system-theoretical thought.

Furthermore, I consider it useful to translate theoretical concepts of one theory (here sociological systems theory) into the language of other interpretive frameworks. This intellectual effort may not only be insightful for newcomers or sceptics of sociological systems theory but also contributes to engendering a debate within the community of system theorists. Clearly, several systems-theoretical concepts may also be found in other interpretive frameworks. Explaining how they are used (similarly/differently) in the other interpretive frameworks may help sharpen the understanding of how they are used in sociological systems theory.

Comparing systems-theoretical explanations of globalisation with neo-institutional thoughts on world polity and world culture has a long history (e.g. Holzer et al., 2015; Schriewer, 2012; Stichweh, 2000). Therefore, reiterating the debate is not necessary here. Nevertheless, a few comments on how systems-theoretical notions are related to other prominently used interpretive frameworks in policy studies may be useful.

A good case in point is the concept of ‘pathways’, used in historical institutionalism. It is similar to how some scholars with affinity to system theory use the terms system logic and, applied to national contexts, ‘socio-logic’ (Schriewer and Martinez, 2004: 33). Antoni Verger and his associates trace the pathways to privatisation in different countries (Verger et al., 2016). They identify six different pathways, ranging from pathway one (e.g. in Chile and the United Kingdom), where the state was systematically restructured along market lines, and services previously provided by the public sector were outsourced to the private sector, to pathway six, labelled ‘privatisation by catastrophe’, where policy actors use chaos during a catastrophe to advance their privatisation agenda (e.g. in New Orleans after Hurricane Katrina and the aftermaths of wars in El Salvador and Iraq). Verger et al. (2016) intended to demonstrate that the global neoliberal wave of privatisation has encountered varied degrees and variants of privatisation that already existed in various national contexts. The receptiveness or resistance towards the new neoliberal wave of privatisation can only be adequately captured against the backdrop of past adaptations of privatisation. National policy actors buy into global education policies, such as privatisation in education, for different reasons. In policy-borrowing research, we use the term reception, rather than pathways, to explore why national policy actors engage with privatisation and other global school reforms.

Similarly, the timing of policy change is a well-known unit of analysis in the multiple-streams approach, formulated by John Kingdon (1984) and taken up by other scholars, including Zahariadis (2014) and Jones et al. (2016). Kingdon coined the term ‘policy window’ to identify favourable conditions for policy change (Kingdon, 1984). He found that the convergence of the following three streams is likely to produce change: the problem stream (recognition of a problem), the policy stream (availability of solutions) and the political stream (new developments in the political realm, such as recent change in government). It is important to point out, however, that the multiple-streams approach does not take into account the processes of transnational policy problems and policy solution transfer (see critique of Baek, 2019).

The free access to and surplus of global public goods (i.e. databases, toolkits, technical reports and case studies) propelled and funded by international organisations (EU, Global Partnership for Education, OECD, World Bank, etc.) exacerbates transnational policy borrowing. Thus, the policy stream tends to be available to politicians and decision makers at all times in the form of best practices, international standards or lessons learned from other educational systems. In fact, the pressure to borrow is great to the extent that policy analysts are frequently placed in the awkward position of having to retroactively define the local problem that fits the already existing global solution or reform package (see Radtke, 2009; Steiner-Khamsi, 2016).

Finally, the concept of punctuated equilibrium – related to the systems-theoretical concept of operative closure – is widely used in policy studies (Baumgartner and Jones, 1993; Jones and Baumgartner, 2005). Policies remain uncontested and policy-making is idle, resting in an equilibrium stage, until they are challenged and profoundly destabilised. As a corollary, punctuated equilibrium scholars draw their attention to agenda setting. How do policy actors manage to prevent problems put on an agenda for reform, and inversely, how do they manage to generate a coalition that helps them to generate reform pressure? The latter focus is systematically pursued in Paul Sabatier’s Advocacy Coalition Framework (1988; see also Weible, 2008). In a similar vein, system theorists consider systems to be operatively closed yet open for change, or in a punctuated equilibrium. Therefore, the attention is drawn to the moments of change, reform or ‘interruptions in relations of interdependence’ (Schriewer and Martinez, 2004: 31).

Even though several system-theoretical concepts strike a chord with well-known interpretive frameworks used in comparative policy studies, the strength of Luhmann’s theory lies with its internal coherence and universal claim, enabling researchers to explain a host of new phenomena, including the proliferation of international large-scale student assessments and policy-relevant educational research.