Literature, Society and Cross-National Variation in Education Systems

IMAGINE ALL THE PEOPLE Literature, Society and Cross-National Variation in Education Systems INTRODUCTION In light of their nineteenth-century political economies, one wonders why Denmark became a leader in public, mass primary education (1814) and secondary vocational training, while Britain created public mass schooling in 1870 and chose a unitary academic track for secondary education. If industrialization drives educational development, Britain (leading the industrial revolution) should be an innovator in both early mass education and vocational training for skilled workers (Wilensky; Busemeyer and Trampusch), and backward, rural Denmark should be the laggard. Nation-building and social democratic impulses might instead be responsible for small, vulnerable Denmark’s early educational initiatives (Boli et. al.), yet the considerable diversity of Danish education (with its large vocational track) seems at odds with mandates of a universalistic, social democratic welfare state (Telhaug et. Al. 2006, Glenn, 2007). This paper sheds light on the peculiarities of education system development in Britain and Denmark by looking at the association between culture (as manifested in literature) and education policy. I argue that stark cultural differences about the purpose of education and the role of the individual in society mediated British and Danish responses to requisites for expanded education (e.g. nation-building, industrialization, and globalization). Education serves multiple purposes: schooling both develops skills and cultivates citizenship. Moreover, access to an education may be posited as an individual right and mechanism for cultivating individual capacities (for employment or self-growth), or as a collective responsibility to train citizens for societal needs (associated with economy, stability and international security). Thus the diverse educational trajectories in Britain and Denmark reflect different cultural constructions: educational institutions evolve to further individual self-growth for upper-class children in Britain, but to bolster a strong society and fulfill social needs in Denmark. More broadly, the paper contributes to the study of culture and politics with a new method and data source for assessing cultural influences on the political institutional development. I use machine learning techniques to analyze cross-national and temporal variations in large corpora of British and Danish literature between 1700 and 1920. This is supplemented with a qualitative reading of four matched sets of famous novels, allowing us to grasp the micro-foundations of how cultural views about individuals in society, attitudes toward social classes and views of political institutions inform educational policy choices. Sharp cultural differences in attitudes about the individual and society and the views of education are apparent in the diverse paths of young, male protagonists in early British and Danish fiction, contemporaneous with the period of educational system development. British novelists credit their heroes’ triumphs to individual struggle; Danish writers more frequently locate success in societal interventions that bring youth in line with social norms. Thus Charles Dickens’ winsome David Copperfield remains plucky against tremendous adversity, sweettempered in the face of foul treatment, and optimistic against all odds. Albeit truly victimized by unjust social structures, David never defines himself as a victim, battles injustices and triumphs through perseverance. In sharp contrast, the hapless Christian is the architect of his own demise in Hans Christian Andersen’s Only a Fiddler. Unlike David, Christian stumbles through life, ignoring the sage advice of his elders, and squanders his opportunities. Authors in the two countries write both novels of accountability (protagonists confront internal limits) and empowerment (they battle structures of repression). Yet even in British novels of accountability, personal struggles with individual morality rather than acquiescence to societal norms enable coming-of-age maturation (Brown; Bodenheimer, 1979). This work makes a crucial contribution by replacing thin interpretations of national culture with empirically-testable methods that allow us to evaluate empirically cultural influences on policy debates and institutional change. The endeavor constitutes exploratory research, because although literary artifacts provide a tangible mechanism for operationalizing cultural predispositions, the complicated interplay between cultural influences and policy outcomes defies easy causal claims. Fiction may reveal or construct new cultural understandings and collective social identities. Literature may legitimate (or challenge) embedded inequities, eroding or reinforcing norms of acquiescence to the dominant political order. One may make a weak or a strong argument about the high correlation between crossnational differences in the cultural depictions of education and specific choices in education reform. A weak version of this argument is that the same cultural values reflected in literature also shape education policy choices. A stronger claim is that authors use the language and topics of their craft to perpetuate ideas about education and the relationship between individual and society. I seek not to replace economic functionalism with a cultural variant, yet the findings suggest that fiction influences educational development, by tipping the scales toward choices that resonate with social constructions of the individual, society and class. Our story of educational policy development also has implications for contemporary political-economic concerns. Guided by concerns for equality of educational opportunity, British educational reformers sought rights and equal access to education; however, their individualistic cultural attitudes paradoxically permitted growing socioeconomic inequities. British stories justified the neglect of marginal youth, because celebrations of boys who conquer challenges with self-initiative make it easier to blame those who fail. Early Danish innovations in mass education and vocational training reflected commitment to building a strong society rather than concerns about equality among individuals; high levels of socioeconomic equality were a felicitous but fortuitous side effect of this social investment. Collectivist views and a desire to build a strong society produced a mandate to educate all the people: neglect of low-skill youth was viewed as a waste of societal resources and a threat to social fabric. THE PUZZLE OF EDUCATION SYSTEM DEVELOPMENT Education system development varies cross-nationally in the timing (and funding) of mass schooling, and in the uniformity of secondary tracks (unitary general education tracks versus systems with optional vocational alternatives). Britain and Denmark diverge in puzzling ways on each of these dimensions: The percentage of children aged 7 to 14 enrolled in primary schools was nearly the same in Britain (55 percent) and Denmark (nearly 60 percent) by 1895 (Flora and Alber); yet Denmark developed mass schooling much earlier than Britain. Denmark established “common” primary schools in 1814 with oversight by local municipalities and seven years of compulsory education (initially rural children participated less). British private, churchaffiliated schools emerged for working class children in the same period; yet, only in 1870 did a bill create comprehensive local school boards to provide primary education and enforce compulsory attendance. Denmark developed early, extensive and enduring vocational education; whereas Britain largely eliminated secondary vocational training with its general secondary education legislation in 1902. It is beyond the scope of this paper to evaluate adequately all of the reasons for crossnational divergence in education system development, and educational institutions may evolve for a variety of reasons (that may change over time). Rather I wish to show how culture mediates national responses to shifting functional needs for schooling. The following cursory review, then, is meant simply to explain why some familiar theories do not completely explain the puzzling patterns of early British and Danish schooling before the twentieth-century. Some explanations link mass education to the requisites of nation-building, social stability and paternalistic absolute monarchs (Boli et al. 1985; Soysal and Strang 1989; Ansell and Lindvall 2013; McDonagh; Greenfeld 1992). These views can account for the early innovation by the Danish absolute monarch; yet they do not explain why Denmark would develop an extensive vocational track even before industrialization.i Mass education may also be a functionalist response to industrialization, particularly in open countries with traded goods, because manufacturing workers (especially in export-sectors) need skills more than agricultural peasants and those in domestic industries (Ansell 2008; Wilensky; Anderson and Nijhuis 2012). Yet industrialization incorrectly predicts that Britain would develop schools earlier and have a larger VET track. Coordinated market economies (Denmark) may require a higher level of workforce skills than liberal market economies (Britain), and countries may use education to catch up with early industrializers (Gershenkron; Hall and Soskice 2001, Iversen and Stephens 2008; Anderson and Nijhuis 2012). Yet Denmark formed vocational programs for low-skill peasants while still an agricultural economy. Patterns of social conflict – particularly religious and class struggles – may also shape education system development (Soysal and Strang 1989, Ansell and Lindvall 2013). Britain and Denmark both had a state church and no sizable Catholic minorities. Yet Anglican/Dissenter religious wars delayed public mass education in Britain, but the Dissenter/Lutheran divide did not stop public school development in Denmark. Religious leaders preached an ethos of moderation, perhaps because the needs for a strong society would not allow for religious conflict (“Danske Skolehistorie”). Strong unions and left parties generally prefer social investment policies (Busemeyer 2014; Iversen and Stephens), and it is not surprising that workers received higher levels of education in Denmark than in Britain. Yet the educational preferences of organized labor in Britain and Denmark varied significantly. British labor and socialists supported universal secondary education in 1902 to avoid a two-tiered system, and partisan strife sharply defined education politics vis-a-vis worker education (Peterson 1973). Danish labor joined business in endorsing vocational education, and all parties sought higher levels of secondary schooling, even for working class children (Klaus Petersen). CULTURE, LITERATURE AND POLITICS Culture (embodied in literature) may also play a role in educational policy development. National cultural explanations for political differences were once a stable of political analysis, and the best used survey data to reveal strong, cross-national differences in cultural views toward politics (Almond and Verba). Yet national cultural arguments may be overly deterministic, difficult to falsify, insensitive to changing ideas, and tautological in that outcomes are used to justify cultural arguments (Lamont, Blyth, Schmidt). Recent work vastly improves on thin interpretations of national culture, and points to several intersections of cultural artifacts and politics. First, cultural artifacts (literature, language and other symbols) act as lenses that ascribe meaning (Wedeen 2002, 713; Geertz), and either reveal or construct perceptions of economic, social and political transformation. The first novels developed with the expanding market capitalism and Protestantism; their character-driven stories, market exchange, and political rights revealed rising cultural individualism (Watt 12-21, 60). Just as coordinated market economies relying on non-market cooperation operate differently from liberal market economies, the literature of liberal and coordinated countries reveal divergent views of capitalism (Griswold and Engelstad). Public opinion polls offer insights into contemporary cross-national cultural differences, yet these data are unavailable historically and the study of cultural artifacts helps to fill in this gap. Yet, important novels may help to construct culture and politics, because great artists have unique voices and vision (Schwarz 1983). Fiction constitutes a site for deliberating economic, political or social threats, and serves as “cultural innovators” of responses to social strain (Griswold; Swidler; Petersen and Petersen). Victorian novelists as social reformers criticized industrial capitalism (Williams), and Dickens anticipated that A Christmas Carol would have “twenty thousand time the force” of a pamphlet on child labor laws (Henderson, 2000, 140-3).ii Recognizing literary contributions does not refute other determinants of policy evolution, but merely points to the task of interpretation. Cultural predispositions may shape our perceptions of educational contributions to goals as varied as individual self-growth, nationbuilding, social control after the French revolution, and skills requisites associated with industrialization and globalization (See also Wiborg). Second, literary works legitimize or challenge contradictions in systems of governance, and reinforce or attack power relations. Institutions have multiple purposes, coercive elements and inequitable applications; and their acceptance depends on some form of legitimation. Authors may serve as organic intellectuals and justify coercive institutions, as when their stories hold victims responsible for structural failings. Thus bildungsroman convey cultural norms of appropriateness (Apol, 62). Yet fiction may also draw attention to the contradictions and cognitive dissonance inherent in governing institutions (Claybaugh 2003-4, 45-6). Changing cultural artifacts also contribute to the legitimatization of new forms of governance, as has been the case with the European Union (McNamara 2015). Education systems must reconcile contradictory aims -- to provide the differentiation of skills linked to hierarchical socioeconomic position but also to cultivate the shared values of political citizenship (Durkeim; Bordieu 1979; Bjerg et. Al, 1995, 31-2) – and literature may contribute to the legitimization of diverse systems. Third, literature has bearing on patterns of collective political action, by providing a “tool kit” of symbols, habits, and stories suggestive of strategies for action (Swidler 1986, 273-6; Lamont 2000). Scholars use careful process tracing to document how cultural ideas embodied in policy legacies shape familiar national responses to new challenges (Berman 1998; Schmidt and Thatcher). Cultural assumptions embedded in fictional works may also contribute to continuities in modes of political action, even when the goals of politics change. Yet, culture may also be mobilized in historically-contingent ways to support new political agendas, as happened with the rise of European right-wing populism (Berezin 2009), and literary language and themes may be associated with institutional change. Collective identities and definitions of self may be shifting with socioeconomic change (Greenfield 2009). Particularly important to patterns of collective political action are views of the individual in society, social group relations, and political institutions. With respect to the individual’s role in society, individual rationality varies across cultural contexts and the psychological articulation of “self” shapes patterns of engagement with society (DiMaggio and Markus, 2010, 351). Fiction influences psychological attitudes toward individual freedoms, social obligations, conformity to the social order, and individuals’ willingness to serve collectivist needs. Because education is a core institution for integrating individuals into society and “constructing the individual” (Boli et. Al 1985, 150), diverse views of individuals in society matter to the formation of education systems. Maximizing individuals’ rights to equal education suggests uniform programs; ensuring that all youths contribute to societal needs requires more diverse options. Fiction also draws or depicts cleavages among economic, regional, religious and ethnic groups and stories contribute to social memory and collective identities. Novels’ depictions of villains and blame may politicize or demobilize marginal groups (Swidler 18986; Lamont 2000; Weerssen 2004; Williams). British elites feared that novels would incite working class revolt (Bratlinger). Uncle Tom’s Cabin fostered opposition to slavery and civil war. Printed books in the vernacular built “imagined communities” and nation states (Andersen). Literacy rates shaped the content of citizen rebellion during the French Revolution (Markoff 1986), and preCommunist literacy rates predicted post-Communist democratic institutions (Darden and Grzymala-Busse 2006). Literature contributes to our preferences for the appropriate institutional vehicles for policy reforms. Depictions of markets and industrialization suggest structural or individual reasons for economic malaise; stories of courts and bureaucrats influence the perceived legitimacy of legal systems and government interventions (Lacey 2001, 350). Literary narratives about schools matter to whether we blame individuals or society for stunted employment trajectories, and influence the institutional development of education systems. Thus literature may also have an impact on the different trajectories of British and Danish education reform. As a collectivist society, Denmark may seek educational expansion to meet nationalist goals; whereas, individualist Britain may view education as largely a tool for individual self-actualization. Regarding timing, collectivist Denmark may develop comprehensive, mass education earlier than individualist Britain to instill patriotism, build states, and recruit soldiers. Individualistic Britain may posit education as a vehicle for self-growth: proponents and opponents may agree that mass education fosters individual empowerment and class mobility, but they may disagree about whether giving workers rights to schooling is a positive or negative ambition. Regarding the form of schooling, individualist Britain may seek a system with equality of educational opportunity and uniform schools; in contrast, Denmark may develop strong vocational education tracks for plural societal needs. Scholars offer important insights into the relationship between literature and politics; yet some questions remain. First is an issue of causal ambiguity: Do fictional works largely reveal cultural attitudes toward political problems and solutions, or do they help to construct these attitudes, and predate policy reforms? Causal arrows probably go in both directions, but can we find evidence for the predominance of one direction over the other? Second is the issue of how cultural artifacts interact with institutional legitimacy: Do institutions have the same cultural meaning in different countries and do fictional works largely reinforce the legitimacy of these institutions, or do they point out the contradictions of systems of governance? Third is the question of how cultural artifacts foster continuity and/or change in processes of collective political action. Systematic empirical testing of observable differences in national literatures should allow us to evaluate the association between cultural predispositions embodied in literature and crossnational differences in educational development. Empirical analyses also allow us to explore some of the unresolved questions about the role of culture in politics, such as those related to causality and the specific sequencing of cultural shifts and policy reforms. If the cultural assumptions in literary works are associated with choices in educational system development, the following hypotheses should hold true: HYPOTHESES 1. Frequencies of education words should increase before major reforms (in Denmark: 1739, 1814; in Britain: 1870); thus frequencies should particularly be higher in Denmark than in Britain in the eighteenth and early nineteenth centuries. 2. Passages about education should have higher frequencies of words about individualism in Britain than in Denmark. 3. Passages about education should have higher frequencies of feeling words in Britain than in Denmark. 4. Passages about education should have higher frequencies of words about political governance in Denmark than in Britain. 5. Passages about education should have higher frequencies of words about society and collective goals in Denmark than in Britain. 6. Passages about education should have higher frequencies of upper-class words in Britain than in Denmark. 7. British education passages should include topics related to individual development and early novels should emphasize the role of schooling in cultivating upper-class sensibilities. 8. Danish education passages should include topics related to nationalism and collective identities. RESEARCH METHODS Two methods help us explore how cross-national and temporal differences in the cultural artifacts of literature are associated with variation in British and Danish educational development. The first is a quantitative analysis using machine learning techniques to explore differences in broad corpora of British and Danish literary novels, poems and plays between 1700 and 1920 (after which copyright laws limit access). The second is a qualitative discussion of four matched-pairs of British and Danish literary works. The quantitative analysis uses machine learning tools to analyze broad differences between corpora of British and Danish works. I analyze the texts with Python machine learning techniques. I use the eight major novels to derive dictionary words, by identifying the top 200 words in each and code these into the following categories: education, individualist, feeling, political, societal, and class words. After stemming the corpora and taking out stop words, I explore temporal and cross-national variations in word frequencies. I calculate frequencies of education words in the each corpus, and construct snippets of (50-word) texts around these words. I then calculate frequencies within these education passages of words referencing individualism, feelings, political institutions, collectivities, and class. LDA unsupervised topic modeling is used as a check on the analysis of word frequencies, to investigate whether topics in education snippets also concern more individualistic issues in Britain and collectivist ones in Denmark. (See Appendix for lists of words and description of methods.) I also provide a close reading of eight major Danish and British works with coming-ofage themes in four literary periods that are also moments of educational expansion: mid 1700s, around 1800, mid-1800s and around 1900. Danish and British authors in each period write in the same aesthetic genre, are inspired by one another and believe that they are writing similar works. Yet they differ cross-nationally in depictions of schooling, the individual’s role in society, and institutions to solve social problems. This qualitative analysis of major novels and plays generates hypotheses about quantitative differences in the word frequencies and topics of the Danish and British corpora. If culture plays a role in educational system development, then thematic differences in major British and Danish literary works should appear in advance of critical education reforms (or non-reforms). Especially relevant are themes depicting education, the individual versus society and political institutions. The cases of Denmark and Britain were selected to represent archetypal dimensions on many explanatory variables of cross-national divergence in education (and broader social and economic policy). Britain is a classic LME whereas Denmark is a classic CME; Britain has a liberal welfare regime whereas Denmark has a SD welfare regime; Britain is the leader of industrialization, Denmark is one of the laggards. Both have protestant super-majorities with few Catholics. Yet the outcomes seem to contradict directly these dominant explanatory variables: DK is one of the earliest innovators of mass education; Britain comes quite late. Denmark has extensive VET; UK has very small VET sector. In selecting works for the broader corpora, the Danish corpus of over 500 works is drawn from the Archive of Danish Literature, a website for “classic Danish literature,” and is complimented by leading authors from other lists of great works. The corpus of over 300 British works is constructed from multiple online lists, with full text files provided by HathiTrust.iii because the full-text files in HathiTrust were often not first editions, it was necessary to alter manually the dates of the works to reflect their initial publication. Timing of publication is crucial for establishing the sequential relationship between cultural artifacts and education reform. I chose works by leading authors who appear on multiple national lists of major works and write about young men coming of age. All works had a major impact among educated elites (who made public policy) but these works (predating the twentieth-century explosion of popular culture paperbacks) were also read by mass readers in the nineteenth-century after the expansion of literacy. In most cases, the matched authors knew one another and believed that they were writing in the same genre as their counterpart. The works are chosen from important literary periods that encompass periods of education reform in at least one of the countries: the prose fiction and novels of the early 1700s when the Danish king issued a decree for schools in every town, the romantic period from around 1770 to 1820 when Denmark enacted public mass education and Britain created private schools, the realist novels of the mid-1800s that anticipated British mass education reform in 1870, and the modernist novels between 1870 and 1920 anticipating the secondary education school reforms. British works include Daniel Defoe’s Robinson Crusoe, Samuel Coleridge’s play Remorse, Charles Dickens’ David Copperfield, and Thomas Hardy’s Jude the Obscure. The Danish works include Ludvig Holberg’s Niels Klim’s Journey Under the Ground, Adam Oehlenschläger’s play Hakon Jarl, Hans Christian Andersen’s Only a Fiddler, and Henrik Pontoppidan’s Lucky Per. Holberg is the father of Danish literature, Oehlenschläger wrote the Danish national anthem, HC Andersen has an international reputation, and Pontoppidan won the Nobel Prize. I also mention contributions of these and other authors to education debates, although my extensive historical documentation is beyond this paper’s scope. QUANTITATIVE FINDINGS Analyses of the British and Danish corpora using machine learning techniques reveal significant and largely predicted differences. First, as expected, the frequency of education words in the Danish corpus is much higher than in the British corpus until the latter nineteenth-century and the gap is particularly large from 1770 to 1820 (See Graph One). Graph Two shows that frequencies of education words rose in both countries in advance of major education acts. Denmark had high word frequencies in the early 1700s (before the king’s decree in 1739), followed by declining frequencies, and then a rise again in advance of the 1814 mass education act. Thereafter, education words declined again and remained low over the nineteenth-century. This decline reflects the successful realization of campaigns for instruction in the Danish language and for mass education, making schooling a less crucial social issue. Authors were instead more concerned about industrialization and the role of workers in the industrial project. (My unpublished charts show industrial words to surge during this period). In Britain, the frequency of education words is quite high until 1740 during a period when literacy expands dramatically among the middle-class. Word frequencies then decline to a low point around 1800 to 1820, when Britain fails to pass mass education legislation although churches begin to build schools. The inadequacies of voluntary church schools become a big theme for Victorians and education words then rise in advance of the 1870 reform. Second, within passages about education, frequencies of individualism words (Graph Three) and feeling words (Graph Four) are consistently higher in Britain than in Denmark. Individualism is comparatively higher Denmark from 1720 until 1770, when the enlightenment inspired new liberal values (Watts), drop off when anxieties about social unrest and statebuilding goals motivated educational innovation, and climb slightly toward the end of the nineteenth-century with modernization. These findings offer support that education is more about individual self-development in Britain than in Denmark. Third, education passages have nearly equal references to national governing entities (Graph Five) from 1720 to 1820in both countries, which we did not expect given our expectations that nation-building drove education reform. Denmark does have more references than Britain during the period of 1820-1870, when it lost Schleswig-Holstein. The marked drop off in the latter nineteenth-century of governing words in Denmark perhaps reflects the rise of the “Danish model” of industrial cooperation, which replaced governmental policymaking with negotiated agreement (Martin and Swank 2012). Graph Six, however, shows that Denmark has a marked advantage over Britain in references to societal goals and entities: the biggest difference is the sense of a strong society rather than of a strong nation state. The increase in British references to societal goals after 1870 (and decline of individualism words seen in Graph Three) may reflect a need for greater collectivism after mass suffrage, and a search for order with anxieties about industrialization, globalization, technological change and world war, all themes of modernist literature (Crosthwaite 2010). Graph Seven shows that Britain has more references to upper class individuals than Denmark in the education snippets. These findings further document greater support for mass education in Denmark than in Britain, because snippets of text surrounding education words in British literature have greater emphasis on upper-class individuals than in Denmark. It is possible, of course, that something in structure of the two languages skew the results, although one then should expect one language to have elevated frequencies across categories (not the case). Graph Eight gives us added confidence in the findings by showing that the frequencies of the British word “give” and the Danish word “giver” (both stemmed) track quite closely. The high frequencies of “give” in both countries the early eighteenth-century probably reflect stronger commitments to duty and obligation, and “give” drops in both with modernization (Greenfield). Finally, I offer added confidence in these findings with unsupervised topic modeling techniques (although space constraints prevent full reporting of these). I request that five unspecified topics should be identified within British and Danish snippets of 500 words surrounding “school” and “education.” Unsupervised topic modeling produces probabilistic results (each run delivers a different organization of words), yet repeated trials converged on similar topics in each country and demonstrated stark differences between Britain and Denmark. Table One presents results from one such trial for 1720-1770 (which I chose to report in advance). The British topics largely concerned individual self-development of upper class youth. The first topic linked schooling to the development of wit, genius and capabilities. The second suggested that good ladies would lend their hand to the poor. The third suggested that masters would receive education to become great, worldly, and fit for public matters. The fourth and fifth emphasize the development of good nature, love, reason in the cultivation of gentlemen. In contrast, the Danish snippets referred to nationalism, foreign influences, and religion, consistent with authors’ campaign against instruction in Latin and concerns about building national identities. The first and third Danish topics concerned learning the meaning of truth and cultivating wisdom for those lost or wild. The second and fifth referred to the issue of schooling in translated Latin and Greek philosophy (which was a hotly debated theme in this period). The fourth references the following the king, God and church in learning as the right way to develop love and wisdom. These themes become even stronger in 1770-1820. Topic modeling provides added confidence that education in eighteenth-century Britain was to cultivate individual self-improvement among the upper-class before mass education reform, whereas Danish authors already linked schooling to collective entities and goals. (See online Appendix for additional findings.)