It was in the early 1990s that the European Union endorsed ‘lifelong learning’ as an instrument of policy. Several international organisations embraced the idea at around this time; an emerging global consensus saw lifelong learning as an essential element in responding to technological advances, globalised production and trade, and labour market flexibility. The EU’s position was, however, unusual, even unique, among them in that it has a strong, institutionalised, and in some respects direct, role in shaping the policies of the states which comprise it. In relation to education, this had been given a legal basis by the Maastricht Treaty in 1992, which for the first time specified that the EU activities should include ‘a contribution to education and training of quality and to the flowering of the cultures of the Member States’, and this was given an organisational basis when, in the early 1990s, the European Commission established a directorate-general with responsibility for education.Footnote 1

Although lifelong learning has never become a principal objective of EU policy,Footnote 2 it soon became a key mechanism for achieving other aims. These other aims have changed over time and have often also been in tension. Lifelong learning was first embraced as an element of competition policy. When EU expansion became a major feature in the 1990s, with it came a concern to cement democracy in member states formerly ‘behind the iron curtain’; associated with this—though also with a concern to create some European identity and to overcome problems of EU institutions being seen as distant, unresponsive and ‘bureaucratic’—was the idea that lifelong learning might foster a more engaged European citizen. Under Romano Prodi’s presidency of the European Commission (1999–2003), it was briefly enrolled in pursuit of a more democratic, complex multinational polity, engaging the EU with its citizens with a stronger role for ‘active citizenship’. (His predecessor’s Commission, beset by allegations of fraud, had ‘imploded in ignominy and mass resignation’ (Tomkins, 1999, p. 744).)

The Lisbon Strategy, adopted in 2000, however, became essentially a matter of centralising policy objectives and bolstering them with top-down implementation mechanisms (notably, such ‘soft-power’ features as indicators and benchmarks). These were operationalised through the Open Method of Co-ordination, whose prescription of discussion and consultation with social partners (or stakeholders) in shaping policy aims lent centralisation a democratic cloak. (It was seized on by some educationalists within the European Commission as a potential basis for a new and more democratic ‘European educational space’ (cf Hingel, 2001; Lawn, 2006; Nóvoa & Lawn, 2002).) Nevertheless, as it turned out, the Open Method of Co-ordination’s centralising and implementation elements proved much more salient—in education and lifelong learning, as elsewhere (cf Gold et al., 2007)—than any ‘space’ it allowed to citizens in the formulation of policy aims.

The chief reasons for this were twofold: on the one hand, the EU’s enlargement; on the other, global economic competition (or at any rate a growing fear of it). Politically, the EU expanded from 12 member states in 1995 to 27 in 2007; its population grew from roughly 350 million to over 500 million. The countries were highly diverse. Four had over 60 million people each; 13 had populations smaller than London’s (then the EU’s largest city). Ten had a recent 40-year experience of state socialist government and lacked deeply embedded democratic institutions and cultures. This generated vast challenges of governance. In parallel, politics in several more long-standing member states—where political institutions had seemed quite strong—were threatened by the growth of far-right, authoritarian populism. (The decline of the political left, and social democracy in particular—associated with industrial restructuring and the post-1990 ideological crisis of socialism—was also significant.) From around the turn of the millennium, European right-wing populism donned ‘Eurosceptic’ clothes, and—from toppling the Constitutional Treaty in 2004 to Brexit in 2016—threw up a series of existential threats to the ‘European project’.

At the same time, Europe was facing intensified economic challenges. With revolutions in information technology and capital accumulation, industrial production became increasingly mobile and markets increasingly global. Europe’s economic and political leaders saw competition from North America and Asia as posing a twin threat: structural, long-term unemployment and social exclusion. One of the attractions of lifelong learning was that it addressed both economic and social policy aims: competitiveness and inclusion. In the 1990s, Western economies generally prospered; the threat seemed relatively distant. Many felt optimism that a transformation into a ‘Europe of learning’, a ‘knowledge economy’ and a ‘learning society’ could be achieved, and clearly in any of these lifelong learning would be central (Archibugi & Lundvall, 2002; Commission of the European Communities, 2000; Green, 2002; Lundvall & Borrás, 1997).

It was in this spirit that the Lisbon Strategy set the aim of making Europe ‘the most competitive and dynamic knowledge-based economy in the world’ by 2010 (Council of the European Union, 2000). In the event, however, the globalised economy threw up unexpected—as well as anticipated—problems. The late 1990s’ ‘dot com bubble’ burst in early 2000; economic advance faltered. In 2007 and 2008 came the financial crash; a massive and global recession followed; in Europe, it threw its shadow over the following decade. Many felt its ill-effects. Some countries suffered more intensely than others—Greece’s travails stand out—but in all countries, the young were particularly hard-hit. The narrative of the previous decade had been one of hard work and educational achievement, leading to good jobs and career opportunities; the reality all-too-frequently turned out to be unemployment, or insecure, lower-skilled, jobs for which they were educationally ‘over-qualified’. The allure of the ‘gig economy’ was strongest for those with no need to earn a living from it.

Economic competitiveness has been a key aim of the nation state under neoliberalism, and one warmly embraced by the EU on behalf of its members. When, therefore, the Lisbon Strategy’s achievements began—rather soon—to fall short of its ambitions, action was taken in short order. The Commission appointed a ‘High Level Group’ chaired by a former prime minister of The Netherlands, Wim Kok. This group acknowledged the problem, but recommended, essentially, more of the same. The Lisbon Strategy was ‘even more urgent’ because ‘the growth gap with North America and Asia has widened, while Europe must meet the combined challenges of low population growth and ageing. Time is running out and there can be no room for complacency’. Its recommendation: ‘Better implementation is needed now to make up for lost time’. There should be ‘a radical improvement’ in ‘delivery’, to be achieved by sharper focus on a reduced number indicators and benchmarks, with the Commission ‘praising good performance and castigating bad performance—naming, shaming and faming’ (Kok, 2004, pp. 6, 42–43).

All bureaucracies have an appetite for policy centralisation; in few is it sated; after the financial crash of 2007–2008, little else appeared on the EU menu. Yet the politics of national inequality also mattered. In 2010, Germany accounted for nearly 20% of EU GDP; 85% of EU GDP came from the 12 countries which had been EU members before 1995. Greece, though one of these, accounted for well under 2%. When the Greek economy ran into serious trouble, it became a sacrificial lamb: the Euro was by then iconic. But the Europe 2020 strategy, inaugurated in 2010, was essentially Lisbon laced with Kok. In policy terms, the aims and top-down implementation were endorsed, but should be further intensified. Some indicators and benchmarks were promoted (to be key indicators); others were effectively demoted. Most important, however, the introduction of the European Semester enabled the Commission to oversee and intervene in national policies, particularly economic policy, on a regular basis—informed by the array of ‘data’ increasingly available, particularly through Eurostat. Of course, the Commission’s capacity to intervene is subject to the general—legal—commitment to ‘subsidiarity’, and in practice, the nature and degree of intervention has varied from country to country: it is easier for the Commission to intervene in Malta than in Germany, Estonia than in France. Nevertheless, the European Semester has become a profoundly important mechanism which, though strictly speaking focussed on economic policy, can in practice address ‘recommendations’ across a wide range of policy concerns—including lifelong learning, especially as linked to labour markets and training.

It was with this mindset and model that the EU grappled with the financial crisis and its aftermath. This book is, of course, concerned mainly with adult education, or lifelong learning, and particularly with how these shape the lives of young adults. The Lisbon objectives—and particularly the Kok readjustment—may have been driven by fear of international competition (the global or ‘neoliberal imaginary’ (Rizvi & Lingard, 2010), but this was seen as key to ‘sustainable economic growth’, providing ‘more and better jobs and greater social cohesion’. With young adults particularly hard-hit—and experience from a century and more pointed to the social impact of unemployment (Beveridge, 1909; Jahoda et al., 1972; Rowntree & Lasker, 1911)—delivering outcomes for the Union’s youth became the focus for European social and economic policies. Lifelong learning policies were no exception: what emphasis there had been on learning throughout life—from cradle to grave—was elbowed aside by the pressing demands of a displaced generation.

Origins of the Enliven Research

This was the background to the project on which this book is based. Some explanation of its nature and background is therefore in order. ‘Enliven’ is an acronym, the short title of a research project formally entitled ‘Encouraging Lifelong Learning for an Inclusive and Vibrant Europe’. The project was funded by the European Commission’s Directorate General for Research and Innovation’s Horizon 2020 programme; its main work was conducted over a three-year period beginning in October 2016.

Enliven responded to a ‘call’ for research proposals on ‘lifelong learning for young adults: better policies for growth and inclusion in Europe’. Inevitably, it reflected thinking in the Commission, and for that reason—if no other—merits some discussion. The call asserted that it was ‘obvious’ that investment in lifelong learning would ‘contribute to overcoming the economic and social crisis and meeting the Europe 2020 targets on employment, poverty reduction, education, sustainability, innovation’ and that the ‘need and markets for adult education (after initial education and training)’ were therefore ‘likely to rise in the coming years’. Yet ‘despite sustained attention over the years, adult education in Europe remain[ed] inadequate’. There were two particular concerns. First, ‘those who are more in need of adult education, such as young, unemployed, low skilled, disabled and vulnerable workers, actually benefit less … than other more advantaged groups’. At the same time, there was an imbalance between public and private investment (the costs of adult education were borne largely by enterprises, individuals and families, rather than by the public). (European Commission, 2014, p. 25).

The research specification suggested potential areas for investigation. For instance, while addressing adult education in general, it should focus particularly ‘on young adults and vulnerable groups … after entry into working life’. The relationship between public policies and the ‘dynamics of private markets’ was a concern. Programmes which improve learning outcomes, particularly for ‘young adults at risk of social exclusion and other vulnerable groups’, were to be identified, as was ‘the learning potential and innovation ability in workplaces’.

A particularly intriguing element was to ‘investigate the feasibility [of] and possibly develop an Intelligent Decision Support System (policy making modelling) for simplifying the access to information and support policy making in the different phases of the policy cycle’. This would facilitate ‘access to scientific evidence for policy making’, help correct ‘distortion of the adult education and continuing training market’, show how barriers to access could be reduced, and identify financial measures to support individuals and companies, ‘ensuring effective and fair distribution of resources, reduce[d] mismanagement and corruption’ (European Commission, 2014, pp. 25–26). Research specifications (‘calls’, in the Commission’s terminology) do not, of course, emerge from nothing. They reflect, albeit with distortions and misunderstandings, the priorities, preconceptions and prejudices of their authors—typically corporate authors—and of their time. It is, perhaps, in order to contextualise some of the emphases in this call. Many were very much responses to over-riding concerns we have already mentioned, above all, the economic crisis that engulfed Europe (and much of the world) following 2007, its impact on young people in particular, and the political challenges of European enlargement. Some, however, reflect beliefs about the causes of political, economic and social problems and where solutions to them should be sought—in short, they are ideological. Thus problems of adult education were located in market ‘distortion’—more perfect markets would, presumably, deliver better lives for excluded youth, greater global competitiveness and the other aims of lifelong learning policy. Applying ‘various psychological and behavioural sciences outcomes’ was encouraged—many in policy communities were in thrall to the counter-intuitive scientificity of behavioural economics, popularised by authors such as Thaler and Sunstein (2008) and Pinker (2011), and Europe’s leaders were no exception (Troussard & van Bavel, 2018; van Bavel et al., 2013, 2015; Zuidhof, 2019).

Much of this thinking was transmitted in an important report commissioned (in 2012) by the Directorate General for Research and Innovation and based on analysis of ‘final reports, working papers and published articles from research projects funded by the Directorate-General for Research and Innovation under the sixth and seventh framework programmes’Footnote 3 (Federighi, 2013, p. 3). A Foreword by Robert-Jan Smits, Director-General for Research and Innovation at the Commission, explained the ‘link’ between his brief and ‘continuing adult education’. It was

because investment in continuing education increases the innovation capacity of companies. More generally, of course, continuing education helps workers keep their skills up-to-date and reduces the likelihood of labour market exclusion. It plays, therefore, a fundamental role to achieve the objectives of the Europe 2020 strategy, as regards both growth and inclusion. (Federighi, 2013, p. 6)

That the Commission had ‘bought in’ to discourses of innovation as essential and beneficial hardly surprised us—they are more or less universal. Neither did its valuing of continuing adult education in terms of returns to productivity, efficiency and competitiveness: Commission common-sense had long framed even social exclusion as solved by employment and economic growth. What stood out rather more was a degree of bureaucratic vanity. ‘Policy makers’ needed to ‘understand which strategies will lead to more effective governance of adult and continuing education, to help them shape dynamic learning environments adapted to different sectors, education levels and enterprise size’. Governments (and consultancies) across the world worshipped at the mantra of ‘evidence-based policy making’; the European Commission was no exception. Yet, ironically, there was evidence in abundance. The Labour Force Survey, the Adult Education Survey, other work by Eurostat and outputs of friendly organisations (such as the OECD’s PIAAC—Programme for the International Assessment of Adult Competencies) meant the Commission was awash with information, much of it quantitative. Increasingly there was also ‘big data’: ‘large, diverse, and heterogeneous datasets, often by-products generated from business and Internet transactions, email, social media, health care facilities, and various sensors and instruments’ (Hofferth et al., 2017, p. 8).

The problem was how to process the information to render it useable, an assistance in framing policy, rather than allowing it to swamp the corridors of power. This is no new difficulty: states’ collection of new forms of information and development of scientific—not least statistical—techniques to manage and analyse it stretch back centuries, even millennia: the complex inter-relationships between governing citizens (or subjects), and the techniques for collecting, handling, and measuring data about them, have been extensively discussed by historians and sociologists of knowledge (e.g. Burke, 2012; Desrosières, 1998; Prévost & Beaud, 2016). Clearly, in the ‘internet age’, radical advances should be on the cards. If Amazon and Google know our shopping preferences before we do, surely governments could find ways to do likewise to the public good? As academic policy scientists argued:

Developments in information technology offer an unprecedented opportunity to collect diverse data at fine-grained spatial and temporal scales, and present a remarkable chance to change the way social science is conducted and to greatly expand the questions that can be addressed. … This is an opportune time to rethink the primary ways in which data are collected, gathered, coded, curated, documented, archived, and disseminated …. [For instance,] spatially referenced administrative data, GPS-enabled cell phone data on the movements of individuals through their day, social media data, remotely sensed and observational data such as Google Street View, and survey data … [might be linked] to administrative data or to other characteristics of communities in which individuals live [using] … cutting edge approaches such as data trawling and web scraping [that] will produce detailed accounts of movement, social networks, and other forms of community building … (Hofferth et al., 2017, pp. 8–13).

If such data could be mastered, the scope for ‘behavioural insights’ seemed well-nigh infinite. Behavioural economics provided, or appeared to provide, a theoretical underpinning for this renewed outbreak of bureaucratic self-regard. Rather than policy-makers assuming that ‘people act ‘rationally’, making choices that lead to the best possible outcome for them’, in behavioural science literature ‘deviations from rationality in decision-making are well documented’. What was required was ‘specific empirical observations’ of behaviour in particular situations: ‘proper understanding of human behaviour requires reality checks’. This would ensure that ‘policy-makers rely on evidence, not assumptions’ (van Bavel et al., 2013, p. 8).

With this background, Federighi’s report for the EU’s Research and Innovation Directorate General was predictably bullish. ‘Policy-making processes do not make use of all the resources available’, he asserted, but their ‘building intelligent decision-support systems devices’ would simplify ‘access to information through … the policy-making process in its various phases’. (Federighi, 2013, p. 79)

Research carried out on a worldwide level has generated sufficient knowledge and know-how to foster policies of adult and continuing education which deliver the desired results. New devices, refined by research in the field of artificial intelligence, can give policy-makers easier access to available scientific knowledge and the possibility of foreseeing the impact of the policy measures that have been adopted.

Research should produce an intelligent decision-support system that facilitates the impact analysis ex-ante of the policy measures for adult and continuing education by gathering and analysing evidence, identifying and diagnosing problems, proposing possible courses of action and evaluating the proposed actions. (Federighi, 2013, p. 89; bold and italics in original)

Such was the background to the call to develop an Intelligent Decision Support System: a system which, in the policy imagination, would permit ‘policy making modelling … simplifying the access to information and support policy making in the different phases of the policy cycle’. (European Commission, 2014, p. 26)

Rethinking the Problem

We responded to the ‘call’—first of all—by reframing the problem. It was very true that adult education in Europe was ‘inadequate’. Under the label ‘lifelong learning’, it has been a key element of EU policy for 25 years—initially adopted as a response to the employment problem. The EU’s mechanisms for multinational policy co-ordination and measurement were in many respects world-leading. Yet when we designed the research in 2014, one in every four Europeans under 25 was unemployed—in some countries, it was one in two. Many were not in employment, or education, or training. An important report of 2012 had found one in five 15-year-olds across the EU lacking the literacy skills to function successfully in a modern society and 73 million adults with low levels of education and literacy: ‘if Europe achieved its current benchmark of functional literacy for 85% of 15-year-olds, this could lead to an aggregate GDP gain of EUR 21 trillion over the lifetime of the generation born in 2010’ (High Level Group on Literacy, 2012, p. 26). Clearly, Europe’s ‘educational markets’ (to adopt the term favoured by the Commission) were failing to ensure its citizens—particularly its younger citizens, though not them alone—had the education and training they required for their own prosperity and welfare. That had implications for society as a whole. Social exclusion, disaffection and long-term ‘scarring’ presented risks for economic competitiveness, social cohesion and the ‘European project’.

But reframing the problem and stressing its gravity were only the starting-point for more thorough-going Enliven reconceptualisation. Of course, research needed to review lifelong learning policies and programmes across the EU. It must examine the nature of adult participation in learning, particularly among ‘disadvantaged’ groups in society. It must look at how learning is organised in workplaces, and how this contributes to efficiency and ‘competitiveness’. Analysis must be based on evidence. But the research also had a distinctive theoretical framing. The Enliven researchers come from two main backgrounds. Most are social scientists, but of a particular kind: adult educators, informed by a common body of theory. (The Enliven team, because we were asked to develop an IDSS, also included computer scientists.) Of course, no group of scholars is theoretically homogeneous, but ‘Enliveners’ could relate to and exchange views not only by reference to the common resources of sociology, political and economic science—theories in the traditions of Weber, Marx and so forth. They also shared familiarity with major traditions in adult education theory. The ideas of Dewey, Freire, Gelpi, Jarvis, Lindemann or Tawney differ in many ways, but all have a vision of learners as active, powerful, engaged and ethically important subjects in learning processes and the social construction of knowledge.

Such premises prejudiced us to look at people not as the objects of policy but as subjects; not as subjects of the state but as intelligent, thinking citizens; not as consumers of learning products but as creative designers of learning processes; not as atomised, eternally excluded individuals but as self-governing and inventive adults entitled to contribute in educated democracies. This has, of course, a utopian dimension, but adult education has always been bound up with collective movements for human self-improvement. Self-improvement is very often against the odds: we are far from believing that people shape their own destinies untrammelled by their environment. To take a very specific case, Europe’s Youth Guarantee both empowers and constrains its beneficiaries (Dingeldey et al., 2019; Milana et al., 2020). Social theorists have long argued, and social research has shown, that structures and institutions shape our lives; they shape how we think of ourselves, what we hope for and what we think realistic and expect. Of particular importance, structures and institutions reflect, reproduce, organise and transmit the social distribution of power.

Bounded Agency

We were therefore attracted by attempts to use the notion of ‘bounded agency’ in the analysis of lifelong learning. Without engaging in a ‘genealogy’ of the idea, two important contributions stood out. Both developed the concept, deploying it with powerful explanatory effect. Evans’ (2007) focus is on the lived experience of young people. She and her collaborators used the bounded agency concept to explore ‘how young adults experience control and exercise personal agency’ in the ‘extended and multiple transitions’ they experience in youth and early adulthood (p. 88). She wants to emphasise young people as ‘having a past and imagined future possibilities, which guide and shape actions in the present, together with subjective perceptions of the structures they have to negotiate, the social landscapes that affect how they act’ (pp. 92–93). Evans’ focus is on how individual young people navigate the complex and changing environments within which their lives are led.

Rubenson and Desjardins (2009) employ bounded agency a little differently. Their concern is why, in some ‘welfare state regimes’—a notion they take from Esping-Andersen (1989, 1990)—participation in adult education is much higher than in others. In their view, ‘although individuals have a degree of agency with regard to their learning behaviours, they are also bounded by structures and contexts and by features of the self that constrain choices’ (Rubenson & Desjardins, 2009, p. 192). Public policy, they suggest (which is represented for them by different types of welfare state regimes) not only ‘directly affect[s] the contextual (structural) conditions that individuals face (on the job, in civil society, at home)’ but also shapes, indirectly, ‘individuals’ subjective rationality and view (disposition) of their opportunity structure’. Drawing on Sen (2000), they see these subjective or ‘dispositional barriers’ as important constraints on ‘a person’s capability and hence freedom to participate’ in adult education.

This perspective places policy at a premium. People’s agency operates within particular institutional frameworks. The origins of these institutions are complex, but in some sense they can be regarded as the outcome of public policy—conscious and accidental, historic and current. Just as important, these institutions also play a part in shaping our subjectivity—our agency. This ‘interplay between structure and individual’ generates a ‘conditioning of values and subjective rationality’ (Rubenson & Desjardins, 2009, p. 196). Such ‘bounded agency’ suggests, for our purposes, that the structure of social and educational institutions not only provides incentives or barriers to adults who do wish to participate in education but also affects whether and how they wish—develop a desire—to do so. Here the capability approach plays in. The decision to participate is a function not only of the adult education available (and related factors such as how difficult it is to take part) but also of the subjective processes by which a person formulates the desire and takes action—makes the decision.

Individuals’ capabilities can play a part in ‘defining structural conditions’, notably ‘situational ones such as job and family’ but also institutional ones through forms of collective action (Rubenson & Desjardins, 2009, p. 196). This ‘structure-agency interaction’ means that public policy can influence adults’ take-up of learning. For our purposes, it is important that adult education itself ‘can be instrumental in fostering capabilities’ (p. 196): adults who take part in education develop broader capabilities, including those which encourage and enable them to learn and study more. But more broadly, they argue that the ‘Nordic welfare state’ has a number of features that encourage adult participation in education. In short, these are: first, close integration of adult education and active labour market policies, based on ‘the full employment concept’; second, highly developed corporatist industrial relations structures, allowing strong trade union voice and influence; third, civil society encompassing a strong ‘publicly supported sector of adult popular education in the form of folk high schools and adult education associations’; and fourth, an ‘emphasis on equity’ that has ‘a deep impact’ on funding regimes (p. 198).

The ‘major difference between Nordic and non-Nordic countries’, according to Rubenson and Desjardins (2009), is ‘not the existence of barriers to participation but the conditions that allow a person to overcome these’. They question how useful it is to try ‘to understand barriers by focusing solely on how individuals interpret the world’; instead, we should analyse the interaction between structural factors and individual dispositions. A welfare state regime ‘can affect a person’s capability to participate through the way it constructs structural conditions and helps individuals overcome both structurally and individually based barriers’ (p. 203). In short, the structure of a ‘welfare state regime’ is vital in generating (or not generating) capability; and in certain circumstances—such as the Nordic example—adult education itself seems to play a part in this.

The bounded agency concept throws explanatory light on a widespread finding of recent comparative research in lifelong learning: the persistence, pervasiveness and diversity of institutional and social structures across the nations of Europe, and their significance in structuring differences within and between social groups in participation in learning and motivation to learn (Blossfeld et al., 2014; Boeren et al., 2012; Saar et al., 2013). While others (Busemeyer & Trampusch, 2011; Desjardins, 2017; Hall & Soskice, 2001) have used the language of ‘political economy’, in Enliven, we have generally referred to ‘welfare state regimes’. Perhaps, we have tacitly associated welfare states with popular and social movement activity and with the historical legacy of adult education as associated with democratic social movements. The difference in language, however, is more normative than analytical: Nordic welfare states may have the features Rubenson and Desjardins enumerate, but several European ‘welfare state regimes’ in our typology do not—indeed, in some welfare can seem very much at a discount.

The Nordic example also shines light on why the concept of bounded agency is useful. From a ‘behavioural’ viewpoint, Nordic welfare states appear as relatively fixed structures; they are contexts which shape people’s decisions. In general, statistical models such as those advocated by behavioural economists and psychologists will be based on how people have behaved within the institutional frameworks they set. Yet these frameworks—labour markets, trade unions, industrial relations systems, adult education institutions and so forth—are not simply ‘given’. They are the outcomes of ‘agency’ by people and by the organisations and movements people create. Nordic welfare states reflect the social and political histories of Scandinavian countries; the welfare state regimes or political economies of other countries similarly reflect the historical interaction of various social forces. For our purposes, this illustrates that while institutions must for some purposes be taken as ‘given’ or ‘fixed’, they are themselves the product of agency—the activity, over generations, of people and their organisations.

This is not, of course, in itself a very original insight. ‘Men make their own history’, Marx famously wrote, ‘… but under circumstances … given and transmitted from the past’ (1969/1852, p. 398). ‘Contingency theory’ in the sociology of organisations and management stems from the insight that structures influence, even determine, the behaviour of organisations and the people who work in them (Burns & Stalker, 1961; Child, 1972; Woodward, 1958). For neoliberal economics, markets bring efficiency, but almost more importantly, they shape morality (Bowles, 1998; Hayek, 1944; Rodrigues, 2013). Nevertheless, bounded agency is an important corrective to over-enthusiastic adoption of behavioural approaches (particularly in economics and psychology) for modelling and accurately predicting human behaviour—particularly when combined with increasingly sophisticated computer-based technologies. ‘Real world’ events periodically provide cautionary tales of the limitations of behaviourally based models, however technologically sophisticated; the 2007 financial crash is a case in point (Mackenzie, 2011; MacKenzie & Spears, 2014)—as, more recently, the Covid-19 pandemic has proved.

Making and Analysing Policy

Public policy has been a central concern of the Enliven research, as of its funders. Politicians and civil servants, of course, regard themselves—too often, themselves alone—as ‘policy-makers’. Enliven researchers had to work with, and deliver results to, civil servants for whom this self-image was sheer common sense. They looked to us for findings, insights and technologies so they could make better decisions. For us common sense cannot be taken for granted. Assumptions about how policy is formulated and implemented, and who plays a part, have now been explored, empirically and theoretically, over several decades (cf Ham & Hill, 1984; Lipsky, 1980; Sabatier, 1986). Our own normative theoretical roots told us we should think of citizens as active subjects in policy processes, not as mere objects of policy made by those appointed to govern them. Of course, some voices and interests have more traction than others in policy processes, and over the years, scholars have developed a range of concepts and models to explore this: policy communities, policy networks, policy cycles and so forth. A principal concern has been to map the myriad ‘policy actors’, the parts they play and their relative significance. For us, this was made more complex by the multinational and multi-level nature of governance in the European Union.

We approached this challenge from two directions. From one, we developed and tested an Intelligent Decision Support System (IDSS). From the other, we sought to investigate policy processes analytically, working with the concept of ‘policy trails’. These were very different approaches, and there was a continual—though generally productive—tension between them. The first involved a close collaboration between social scientists and computer scientists; the second was very much a matter of social scientific inquiry.

As we have seen above, we were asked to develop an IDSS ‘for simplifying the access to information and support policy making in the different phases of the policy cycle’ (European Commission, 2014, p. 26). Developing an IDSS is a matter of engineering design; in order to begin, our computer scientists needed answers to apparently straightforward questions. Who would use the IDSS? What policy problem did they want to address? What policy interventions had worked? It soon became clear that we required a clear and bounded area on which to focus: ‘lifelong learning’ was both too large and too vague. Discussion—much favoured among critical social scientists—about the complexities of what ‘policy’ comprises, who is involved in making it, how they interact, who defines what ‘works’ or ‘succeeds’ and so forth, muddied the waters. In effect, much social scientific research on policy was beside the point so far as IDSS design was concerned—though it took time to grasp this reality.

Our computer scientists needed to know who the ‘end-users’ of the IDSS would be; the answer, of course, was ‘policy-makers’; if this begged the question, it was also a basis for practical progress. For largely pragmatic reasons, we found a boundary for our IDSS by focussing on young people not in employment, education or training: that section of humanity to whom the label ‘NEET’ is now commonly attached. The team adopted this focus after some discussion but without great controversy. It seemed a relatively bounded issue. Being NEET had become a clear problem and policy priority, across Europe; researchers had worked on it; there had been many interventions of various kinds to address the problem. We anticipated that many of the interventions—having been funded by the EU—would have been subject to formal evaluations. Altogether, it was a matter of key public concern, and the sources of data about interventions seemed likely to be rich. In the event, while the former presumption proved valid, the latter was more problematical. While there had been many interventions and a good number of evaluations, we were to discover that the consistency and quality of the latter are very variable: of that, more in later chapters.

Nevertheless, we pressed forward to develop the IDSS. Consultations with a range of potential ‘end users’ helped our computer scientists identify the essential features needed and design the user interface. However, the poor documentation of cases—often spread inconsistently through several databases and sparse even when in a single document—and the enormous variation in how interventions were described, made large-scale comparison virtually impossible. The team tried to overcome these problems by generating a common repository (after considerable labour, this contained 222 items) and a similarity measure to index and retrieve relevant policy and evaluation documents. Their strategy drew theoretically on Schardt et al. (2007)’s use of the PICO (Patient problem, Intervention, Comparison and Outcome) framework to perform meta-analyses of clinical evaluations. The IDSS incorporated a matching algorithm based on the similarity measure so that ‘most similar’ interventions could be retrieved from the knowledge base; it also used natural language processing techniques to semi-automate the task of maintaining the knowledge base.Footnote 4

Policy Trails

Policy trails, first developed for an earlier European research project on lifelong learning, seemed to us to provide a theoretical basis for fruitful empirical investigation of policy, particularly in the European setting. With contemporary society being highly ‘globalised’ and interconnected, the power of nation states has reduced in many, though by no means, all ways (Dickinson, 2016); policies often emerge through complex processes in ‘transnational and globalized work spaces’ (Rizvi & Lingard, 2010, p. 22), before being filtered through national contexts. Late twentieth and early twenty-first century globalisation has been closely associated with neo-liberal perspectives on economy and business, and with the organisation of public services through forms of ‘new public management’. Public policies today are therefore often intimately bound up with the activities of private enterprises, which both ‘deliver’ many services and have an interest in shaping how this is done. The boundary between private and public sector is therefore blurred and constantly in flux. As Wiesel and Model (2014, p. 201) argue, ‘longitudinal analyses of the manoeuvring of multiple actors in and around individual organizations’ is needed as a way of examining the ‘alignment of interests’ and ‘to fully grasp the complex interplay through which potential conflicts are muted or amplified’.

European multi-level governance blurs other boundaries: public policy is formulated and carried through in an interplay of EU, national and local governments. EU agencies (the Commission, executive agencies, parliament, courts and so forth) must work with the governments of countries varying in size from Germany (83 million) to Malta (440,000). All have constitutions, laws and structures of internal government—often complex in themselves. North Rhine-Westphalia, for instance, the most populous of Germany’s 16 states (Länder), is larger than 21 EU members—including Belgium and the Netherlands, which it borders. Bremen, Germany’s smallest state, is half as big again as Malta. ‘Policy’ is key to organising—governing—Europe; yet how policy is made, what policy consists of, who does what with a policy, and who and what are affected, remain complex and in some respects open questions.

The challenge was to find an approach which, in relation to EU policy and its role, provides a framework for providing ‘penetrating’ understanding of ‘messiness and complexity’, while delivering the simplicity and neatness which allow analysis to be applied in (and lessons to be generalised from) a range of differing contexts. A pointer was the concept of ‘governance’ (Holford & Van der Veen, 2006, pp. 27–29), which ‘gained ground in order to capture the changing relationship between actors in the policy process’ (Cort, 2014, p. 128). This highlights policy processes, increasingly transcending national boundaries, and the growing national and supra-national influence in education and lifelong learning of non-governmental actors (individuals and organisations) and private sector interests. The idea that policy can be ‘trailed’—or mapped and followed across a landscape—recognises ‘the complex nature and consequences for both political structures and agency of the prominent multi-scalar model of lifelong learning governance within the EU’ (Melo et al., 2015, p. 11). Policy trails acknowledge that policy is shaped, negotiated and contested, in formulation and in application, and that those engaged in this shaping include not only ‘the state’ (itself far from simply unitary) but also individual, organisational and collective actors from teachers to trade unions, and from bureaucrats to businesses. From the perspective of lifelong learning in Europe, they allow for the importance of actors at different levels and from different sectors.

The idea of a policy trail incorporates a spatial metaphor: policy moves across terrain, a range of social, political and economic environments. The methodology of studying a policy trail entails adopting a position of seeking to map ‘the terrain the policy travels through rather than the policy itself, analysing how the policy is shaped through its journey’ (Holford & McKenzie, 2013, p. 1). Thus we know that policy in lifelong learning is strongly influenced by fields such as business and economic policies (hence the prominence of ‘competitiveness’) and social policy (hence social cohesion, social inclusion and the like). In each of these, a range of actors share (but also struggle over and through) language, ideas, modes of working, common sense. Further, a policy is enacted not only in a ‘policy’ field—by ‘policy makers’—but also in fields of practice. So ‘non-policy actors’ play a part in shaping policies, and particularly how they are carried into effect.

In practice, mapping policy trails involved developing contextualised policy case studies, covering the adoption, or non-adoption, of policy options, contributions by different actors in varying ways to shaping policies, and so forth. Trails can be mapped using various methods, including analysis of documents and interviews with actors (or those excluded de facto or de jure from being actors). They open, as it were, the messiness of policy to social scientific analysis, investigating the effects of policy when it connects locations and people at different levels. In the context of their work with enterprises, Melo et al. (2015, p. 13) found the policy trail approach valuable not only for gaining ‘insights on the enterprise governance environment from a broader perspective than that of the enterprise itself but also for pinpointing the moments when the enterprise plays an influential role as a governance actor in its public sphere’. Policy trailing is also particularly useful for investigating the interaction between policy agendas and how they actually play out ‘on the ground’.

In Enliven, the ‘point of entry’ to the trail was education and training programmes: we sought to uncover the flow of policy relating to the organisations and to identify those aspects of public lifelong learning policy with which the organisation interacted. To achieve this, we interviewed informants in each country involved in forming and enacting policy related to two programmes: one funded under the EU’s Youth Guarantee initiative (supporting disadvantaged young people into employment), the other funded under the EU’s Upskilling Pathways programme (helping disadvantaged young people overcome barriers as part of the transition into education, employment or training. We treated these as focusing on ‘employment’ and ‘empowerment’, respectively, though accepting that these labels were rough-and-ready. Our interest was in young people (under 30) who do not continue into academic or general upper secondary or tertiary education: low qualification levels reduce opportunities to access secure employment but may also result from disadvantage.

We examined how policy in the relevant area (employment, empowerment) flows through to specific providers active in these areas, how this affects the programmes designed, how educators and support staff working on the programmes experience their involvement, and the opinions and experiences of the learners themselves. We combined insights from interviews, documents and contextual data and tried to be alive to interactions and mutual influences among different ‘actors’ in each policy trail. Informants included policy makers, regulators, funders at the policy level, managers of education and training programmes, facilitators, teachers and trainers working on the programmes, and the adult learners themselves.

The policy trails provided evidence for several aspects of our inquiry: for example, we used them to study the experiences of disadvantaged learners and how funding schemes work. Thus they showed—in relation to particular groups of young people, and programme and provider settings—the dispositional, situational (including financial) and institutional barriers adult learners had to overcome, the problems they encountered, and their successes; this helped us to understand the processes of active interpretation involved when policy is linked to practice.

The Book

The results of a project as large and rich as Enliven cannot be confined to a single book, even quite a long one. Members of the research team have already written about many of our findings and used them to develop new ideas and theories.Footnote 5 Nevertheless, in this book, we attempt to give some overall sense of the purpose, shape, methods, findings and implications of the research.

Our book is divided into four parts. Part I introduces the work. It comprises the present chapter and one other. In Chap. 2, Bounded Agency in Policy and Action: Empowerment, Agency and Belonging, Sharon Clancy and I explore theories of bounded agency and what we think they offer the study of lifelong learning for young adults. We do this partly by theoretical discussion (enriched by the recent work of two Enliven colleagues (Boyadjieva & Ilieva-Trichkova, 2017, 2018a, 2018b, 2021)) but also through a qualitative case study of a personalised youth employability project which places the young people’s personal experience centrally as ‘experts’ in their own lives. Theoretically, we discuss the relationships between such concepts as agency, empowerment and belonging. The case study illustrates how important belonging can be to adult learning, particularly among the most socially and politically excluded young adults, for whom social isolation and mental ill-health can compound pre-existing social and structural barriers.

Part II, comprising seven chapters, digs deeper into the nature and effect of lifelong learning policies, how they are made, and how this shapes what they contain and how they are implemented. Focussing especially on young adults, it addresses questions such as how the EU’s lifelong learning policies have conceptualised ‘vulnerability’ and ‘exclusion’ over the last quarter century, how policies and financial systems are used to co-ordinate European adult education, and who participates in adult learning (and why they do or do not). It discusses limitations in the data currently available (and used) for policy-making in adult education, and what can be learned from an attempt to strengthen policy-making using Artificial Intelligence.

Chapter 3, Vulnerability in European Lifelong Learning Policies 1992–2018: Seeing Young People as a Problem to be Fixed?, takes a dynamic view of the nature of vulnerability. Through critical discourse analysis drawing on a corpus of 68 European lifelong learning documents over 25 years, Concepción Maiztegui-Oñate, Triin Roosalu, Alvaro Moro-Incaurtieta and Marti Taru explore the conceptual foundations and rationale of lifelong learning policies that treat young people as a vulnerable group and show how lifelong learning has tried to address the challenges of vulnerability among young people. The chapter shows how young people, especially in vulnerable situations, have been constructed primarily as a category that is, or should be, economically active—needing lifelong learning to equip them with the right skills for the labour market. This approach has overshadowed other visions—of citizenship and social participation—implicit in lifelong learning. The chapter also shows that human capital perspectives dominate European lifelong learning policies, underpinning a narrow view that misunderstands the causes of vulnerability and generates incomplete policy objectives for socially excluded youth.

In Chap. 4, Participation in Adult Learning: System Characteristics and Individuals’ Experiences, Ellen Boeren, Sofie Cabus and Alan Mackie build on the body of research that compares the macro- and micro-level factors underlying participation and inequality in adult education and training. Several ‘system characteristics’—the organisation of education and labour markets, modes of production within firms, the quality of child care, the level of economic development, the cost of courses—play a key role. In particular, they restrict young adults with low levels of education from access to lifelong learning. This chapter focuses on what advantages an inclusive policy on lifelong learning has for society as a whole. However, although structural barriers mean that the opportunity (or desire) to take part in lifelong learning is not equally spread, small percentages of low-educated adults do so; from them we can learn how barriers to access can be lowered. The chapter therefore analyses initiatives under the EU’s Youth Guarantee and Upskilling Pathway programmes in nine countries. The countries represent different welfare regimes and take significantly different approaches to adult education: their levels of provision vary, as do their active labour market policies and other system characteristics. Learners’ experiences, however, and staff experiences, are broadly similar across programmes and countries. Participants’ motivation and confidence levels are key to success; young adults with low levels of education need individual support.

Chapter 5, Participation of Vulnerable Young Labour Market Groups in Job-Related Training: The Effect of Macro-Structural and Institutional Characteristics, by Ellu Saar, Eve-Liis Roosmaa and Liisa Martma, focuses on why two groups of disadvantaged young adults (the low-educated and those working in low- or medium-skilled occupations) participate—or do not—in job-related, non-formal education and training. Their analysis suggests that, across occupational groups, job content is the core mechanism that keeps young adults away from training. They also find that institutional context is important: when unemployment rates are high, inequality in participation between occupational groups increases. The chapter shows that during the post-2008 recession, investment in active labour market policies and similar measures reduced the ‘training gap’ between high-skilled occupations on the one hand, and low- and medium-skilled occupations on the other. When labour market conditions are more favourable, strong employment protection legislation appears to reduce inequalities in job-related, non-formal education; during the economic crisis, however, it seems to have increased them.

In Chap. 6, Gender Gaps in Participation in Adult Education in Europe: Examining Factors and Barriers, Rumiana Stoilova, Ellen Boeren and Petya Ilieva-Trichkova investigate the barriers which prevent men and women from participating equally in adult education. More specifically, they ask what factors and barriers play out differently—in relation to lifelong learning—between (a) men and women, and (b) women from different class and ethnic backgrounds. Building on previous research and using data from the European Social Survey and the Adult Education Survey, they confirm that social origin presents a significant structural barrier, affecting how likely men and women are to participate in adult education. They also draw attention to the role of intersectionality: of multiple simultaneous disadvantages including gender (for women), social origin (for those whose parents have only basic or lower levels of education), and having an ethnic minority or migrant background. There are also differences between the probability of men and women receiving financial support to participate from their employers—this tends to be lower for women. However, women are more likely to receive financial support from public institutions. The negative effects of ethnic minority status are stronger for women than men. They also conclude that Mediterranean and post-socialist welfare regimes are less able to reduce the effects of intersectionality than social-democratic regimes.

Chapter 7, Adult Education as a Pathway to Empowerment: Challenges and Possibilities, outlines a theoretical framework for conceptualising the role of adult education in individual empowerment from a capability approach perspective and provides empirical evidence on how adult education can contribute to individuals’ empowerment. Pepka Boyadjieva and Petya Ilieva-Trichkova argue that adult education should be regarded as both a sphere of and a factor for empowerment. They see empowerment through adult education as embedded in the institutional structures and socio-cultural contexts available and as having both intrinsic and instrumental value. They also argue that the process of empowerment through adult education is not linear or unproblematic, so that only in some cases do the benefits of adult education lead to empowered agency. Adult education’s empowerment role is revealed in expanded agency; this enables individuals and social groups to gain power over their environment as they strive towards individual and societal well-being. Using both quantitative and qualitative data, the chapter shows that participation in non-formal adult education can be a means of empowering individuals through increasing their self-confidence and their capacity to find employment and control their daily lives.

The media actively contribute to the production of knowledge, and the knowledge produced and circulated in the media contributes to governing adult education and learning. This is the focus of Chap. 8, Governing Adult Learning through Influencing Public Debate: How the Media use PIAAC Data in Denmark, Italy and the United Kingdom. The authors, Marcella Milana, Sandra Vatrella, Gosia Klatt, Palle Rasmussen and Anne Larson, examine how visible the Programme for the International Assessment of Adult Competencies (PIAAC) is in widely read newspapers in Denmark, Italy and the United Kingdom, and how those papers use PIAAC data. Their findings are based on analysis of 83 articles, editorials and opinion pieces, following a common protocol. They show that the information about PIAAC conveyed in the newspapers tends to confirm general standards for adult learning. Data generated through PIAAC, and its ‘implicit’ benchmarking of Level 3 in adults’ performances, support European standard setting in adult learning; and when national governments ask international organisations to do ‘objective’ evaluations, the latter enhance their power. The media contribute to knowledge production, and this is one way they may exert influence. How PIAAC data are used by different newspapers varies not only between countries but also according to the role each newspaper plays within national debates.

In Chap. 9, Policy, Practice and Praxis: Computer-Aided Decision Support to Enable Policy Making in Lifelong Learning, Sharon Clancy and Claire Palmer explore some of the issues that arise in applying Artificial Intelligence to policy-making in lifelong learning. As we have seen, the Enliven project was asked to develop a prototype Intelligent Decision Support System (IDSS) to this end. This focused on young people Not in Employment, Education or Training (NEETs). To support new developments (particularly those aiming to benefit groups disadvantaged in terms of gender, ethnicity, culture and other factors), the IDSS provides a repository of information about existing projects. This means that new initiatives can be based on knowledge of what measures been used and which have worked, and that they can be evaluated against suitable criteria. The chapter reflects on issues involved in the collaborative, cross-disciplinary and inter-sectoral work required, describes how the Enliven IDSS developed, and provides insights into what was learned in the development process. It also discusses the limited nature of the information available and the ways in which computer-based responses to complex social problems are constrained.

Part III is devoted to the pivotal role of workplace learning in organisational life and in learning during what has been called ‘emerging adulthood’ (Arnett, 2014). In this period of life (roughly from late teens to early thirties), young people need access to gainful work not only for income but for learning and development. Some workplaces provide rich opportunities to learn; others restrict it. Organisations shape workplace learning potential, and how they allocate work signals how their agency is used to develop learning potential. The chapters also explore the interaction of individual and organisational bounded agency in workplace learning, and the role of early career workers’ agency in youth-led social movement organisations.

In Chap. 10, The Interplay of Organisational and Individual Bounded Agency in Workplace Learning: A Framework Approach, Günter Hefler and Ivana Studená set out the framework that Enliven researchers used to investigate the interplay of ‘organisational’ and individual agency in workplace learning. The framework guided the research on 17 organisational case studies across three economic sectors and nine countries, and it underpins all the chapters in Part III. Agency lies at its centre, enabling us to understand why individuals take up learning opportunities in different dimensions of their lives over the life course. The concept is, of course, often deployed in the study of individuals’ behaviour, but Hefler and Studená argue for treating agency as relational: actors’ choices are bound—enabled and restricted—by their environments. They also explain how the framework goes beyond individual agency, treating the organisation as a specific type of actor with its own agency: whether workplace learning actually occurs is the outcome of interaction between individual and ‘organisational’ agency. Some of those who find that their workplace offers limited opportunities will no doubt seek a more learning-conducive job; others will look to learn outside work—policies should promote meaningful learning across society. But poor workplace organisation is a key barrier to making lifelong learning a reality for all.

Chapter 11, Working and Learning in the Retail Sector: A Cross-Country Comparative View, by Ulrik Brandi, Jolien De Norre, Triin Roosalu, Maaris Raudsepp and Alesia Khadatovich, compares workplace learning opportunities in the retail sector in Belgium, Denmark and Estonia, analysing conditions for workplace learning as experienced by early career workers. Across Europe, the retail sector provides early career workers with a first entry into gainful employment, yet also employs a large proportion of young adults on a part-time and/or temporary basis. Using in-depth case studies, the chapter shows that workplace design can limit early career workers’ opportunities for day-to-day workplace learning. However, there are important variations within and across the cases studied. Early career workers respond in different ways when retail organisations limit their meaningful learning and career development. Their responses are shaped by their evolving life structures and organisational settings. They may find working in retail ‘right’ for the time being even though, overall, it creates only restrictive workplace learning opportunities. Given the size and dominant role of the retail sector, employers enjoy considerable leeway and can often deviate from the labour standards generally expected in a given country—an important consideration when we seek to understand the role of organisational agency.

Chapter 12 examines Organisational and Individual Agency in Workplace Learning in the European Metal Sector. Vassil Kirov, Ana Isabel Estevez-Gutierrez, Iciar Elexpuru-Albizuri, Fernando Díez, Lourdes Villardón-Gallego and Maite Aurrekoetxea-Casau compare the evolution of day-to-day, informal, workplace learning and work-based training arrangements in Bulgaria and Spain’s Basque region. Although Bulgaria and Spain are respectively located in the post-communist and Mediterranean welfare groups, the Basque region’s advanced, industrialised economic model, requiring efficient skills provision, makes it rather different from other Spanish regions. In the Basque country, institutions successfully link vocational education with labour market needs, but Bulgaria’s state educational system is poor at delivering skills. The co-operative sector is important in the Basque region, and the chapter studies two co-operatives in depth. Having developed in a global value chain, Bulgarian companies (both domestic and subsidiaries of multinationals) have recently introduced some in-house training to cope with a deficit of qualified labour. Using qualitative methods and working on under-researched countries, the authors go beyond what relatively static macro-data show about lifelong learning in the sector, revealing how organisational and individual agency provide space for informal workplace learning and what outcomes this has for early career workers.

The adult learning sector is particularly diverse, and evidence on how employees in adult learning organisations access workplace learning is very limited. In Chap. 13, Work and Learning in the Adult Education Sector: A Cross Country Comparative View, Sharon Clancy, Günter Hefler, Francesca Rapanà, Eva Steinheimer and Ivana Studená, therefore fill an important gap. The chapter focuses on learning by early career teaching staff, using evidence from eight case studies across four countries (Austria, Italy, Slovakia and the UK) and covering different sub-fields of adult education—from organisations providing basic skills training to those specialising in management courses in the corporate sector. The research reveals that despite employees’ high levels of skill, conditions for young or early career teaching staff vary significantly across and within sub-fields. In general, adult learning teachers’ jobs have a strong emphasis on self-directed learning and professionalism. Novice teachers’ career pathways are characterised by poorly structured career opportunities. However, how conducive the work available is to learning varies significantly. This cannot be attributed to the type of services provided alone: it reflects organisational agency. How work and HR practices are organised makes a real difference to teachers’ learning experiences in adult education, not only early in their careers but beyond.

Despite the undeniable legacy of labour movements for adult learning worldwide, the interplay of industrial relations with lifelong learning remains outside mainstream lifelong learning research. Learning from activism is, moreover, usually informal and therefore rarely recognised. In Chap. 14, Maite Aurrekoetxea- Casaus, Edurne Bartolomé Peral, Günter Hefler, Ivana Studená and Janine Wulz focus on activism, not only as an important component of the industrial relations system but also as a major learning source for individuals, organisations and society. Young workers who feel they lack support from existing employee interest organisations may try to create their own. Based on case vignettes of social movement organisations in three different countries with highly diverse industrial relations systems (Austria, Spain’s Basque Region and Slovakia), the chapter presents a framework for analysing and comparing the position of novel social movement organisations within countries’ specific industrial relation systems. Each social movement organisation was founded because of particular challenges that the national system did not adequately address. Learning from activism enables young people employed in workplaces unfavourable to learning—or unemployed—to compensate for what a better workplace might have offered. Youth-led social movement organisations are thus laboratories producing important knowledge and practical skills; they challenge established organisations, including trade unions; and they renew and enliven industrial relations structures for representing interests and developing strategy.

Workplace learning is the dominant source of learning for adults, but opportunities are closely linked to the type of job an individual has. In addition, while some people make good use of the opportunities at hand, others do not. Learning opportunities do not, therefore, translate automatically into learning: individuals also need to apply themselves, to take advantage of them. In Chap. 15, Günter Hefler, Denisa Fedáková, Eva Steinheimer, Ivana Studená and Janine Wulz present a novel approach to investigating the role of individual agency in workplace learning. They study the individual agency of early career employees in three sectors (Retail, Metals and Adult Education) across nine countries. They develop accounts of 71 workers’ learning, constructing between three and eight individual learning biographies across 17 organisations. Their approach means they can investigate workplace learning as embedded both in the organisational context of the job and in an individual’s wider life structure. The chapter shows that when an individual’s agency in workplace learning is considered in isolation from its context, agency in workplace learning cannot be properly explained. Other areas of life add to and/or limit individuals’ learning opportunities and agency in learning: gainful work is an integral part of an individual’s life structure, and always interacts with what happens in other parts of life.

Part IV comprises a single, concluding, chapter. In it, Chap. 16, Pepka Boyadjieva, Sharon Clancy, Günter Hefler, Ivana Studená and I—the five editors of the book—reflect on Adult Education, Learning Citizens, and the Lessons of Enliven. We begin by emphasising the importance of a rich adult education and how this resonates with the Universal Declaration of Human Rights and the body of international law developed in the wake of the Second World War. This envisaged adults as citizens, playing an active part in society and government, and saw this is one reason why adult education matters. We discuss how viewing adults as agentic citizens—while also objects of state policy—relates to the notion of bounded agency. We summarise important lessons of the Enliven project and discuss their implications. These include the need for a rich and broad education throughout the lives of adults and the limiting effect (especially for the ‘socially excluded’) of framing the education for adults through the language of ‘markets’—we advocate listening to the voices of those most excluded. They also include the importance of work, workplaces and labour markets to adult learning, and how organisationally stunted workplaces impoverish their workers’ learning.

The Message

A project as wide-ranging as Enliven can be reduced neither to a single nor to a simple message. At one level, our conclusion is that adult learning is complicated; as in other areas of social life, many of its problems are intractable; easy solutions seldom work. More profoundly, yet also oversimply, we argue the profound value of adult learning to civil society and democracy, to community and organisational life, to fulfilled and healthy lives. Our findings reflect central traditions of adult education and—for which we should be thankful—a succession of important policy statements from UNESCO (e.g. Delors et al., 1996; Faure et al., 1972; UNESCO, 1947). They will surprise no-one who has been an educator of adults.

Yet they contrast sharply with the approaches that have dominated over the past three decades. Policies from most international organisations have implied that adult learning can be stripped of its richness and complexity, and ‘targeted’ at key economic and social problems (OECD, 1996, 2019). Foremost among these problems have been training and retraining for an ever-changing world of work, particularly—and ironically, given that it is typically labelled ‘lifelong’ learning—for young adults seen to have ‘failed’ in school, to have been born into a deprived neighbourhood or social group, or grown to ‘working age’ during an economic ‘downturn’.

The Enliven perspective is informed by wider visions. Our advisory board often reminded us of the breadth and depth inherent in the idea of Bildung; one way of framing our message is to argue that European adult learning should reappropriate this and similar concepts. Many countries have their own traditions, which could inform and enrich what the EU does. Few Britons, for instance, know the meaning of Bildung, but the British Ministry of Reconstruction argued a century ago that adult education ‘is an inseparable aspect of citizenship, … should be both universal and lifelong’, and ‘spread uniformly and systematically over the whole community’ (1919, p. 5). Yet:

One of the oddities of international lifelong learning over the past couple of decades has been the European Union’s enthusiastic distancing of itself from what Europe has contributed – through education – to western civilisation. That this has coincided with the EU’s coming of age as an educational agent makes it all the odder. (Holford, 2017, p. 176)

We propose reimagining—reappropriating—adult learning as central to democratic society, rather than as servant of business competitiveness or a tool to address social exclusion. Adult education enables communication and knowledge development among citizens entitled equally to voice their views, to influence public policies and to enjoy full, rich and healthy lives. It is a vital way in which communities—local, national, European, global—can collectively think through and generate creative responses to the problems facing the human race today; it builds social capital, civil society, tolerance and community resilience. Europe’s response to the Covid-19 pandemic has shown the fragility of societies robbed of embedded educational institutions for adults. Yet at the same time, it has shown how adults, allowed the freedom—and given the resources—to do so, can invent new collective ways of learning and use them to shape their own worlds.