Introduction: digitizing Enlightenment

Simon Burrows and Glenn Roe

This book charts the development of a set of remarkable and interrelated intellectual and digital projects that are transforming the way we view the history and culture of the eighteenth century.¹ The featured projects include some of the best-known, most well-funded and longest established research initiatives in the emerging area of ‘digital humanities’ (DH), a field that has, particularly since 2010, been attracting a rising tide of interest from professional academics, the media, funding councils and the general public worldwide. Advocates and practitioners of the digital humanities argue that computational methods can fundamentally transform our ability to answer some of the ‘big questions’ that drive humanities research, allowing us to see patterns and relationships that were hitherto hard to discern, and to pinpoint, visualize and analyze relevant data in efficient and powerful new ways.

In this book, leading established practitioners and rising scholars in the field discuss how their projects are, both individually and collectively, helping us to revise our vision of the Enlightenment, a historical period perhaps better served for cutting-edge digital projects than any other. Divided into two parts, the book will first detail several important past and present digital projects that take the French Enlightenment period, in a broad sense, as their object of study. Individual chapters will outline each project’s institutional and intellectual history, including the techniques and methodologies they specifically developed and the sometimes-painful lessons learned in the process. Chapters will also highlight, when possible, revisions of understanding to which the various projects have given rise; how the tools and resources they have created have been used by other researchers and teachers; and prospects for the future of the projects and digital research into the period more generally as we continue to develop what Dan Edelstein’s chapter describes as an ‘ecosystem of like-minded projects’.

The second part will feature chapters from early career scholars and academic technologists working precisely at the intersection of digital methods and Enlightenment studies, an intellectual space largely forged by the projects featured in Part 1. Many of the chapters are drawn from international keynotes and invited speakers at the Digitizing Enlightenment symposium held at Western Sydney University in July 2016, and highlight current and future research methods and directions for digital eighteenth-century studies. Thus, the book will serve simultaneously as a monument to the current state of digital work in the Enlightenment period; as a guide to past techniques and lessons in digital methods; as an overview of current findings and the state of the field; and, finally, as a vision statement for future research.

The original impetus towards this book, together with the Digitizing Enlightenment symposium series, sprang from a conversation that took place in Montreal at the Society for French Historical Studies conference in April 2014. That conversation involved the panellists and chair of a session entitled ‘Digitizing the French Enlightenment: a discussion of three digital humanities projects’: Dan Edelstein, Jeff Ravel, Simon Burrows and Sean Takats. In private discussion after the panel, all agreed on the need to get together intermittently to harmonize digital projects in eighteenth-century studies, above all by adopting common digital standards and exploring how far their various projects could be made interoperable through linked data.

By historical accident, the conversations that galvanized the Digitizing Enlightenment initiative, and the response to our first international call for papers, brought together a predominantly English-speaking group of scholars.² This is reflected in the contributions to this book. Some may surmise that this is evidence of an Anglo-Saxon bias, but even at the most superficial level any such conclusion would be illusory. Our initial coalition of like-minded projects includes two Franco-American collaborations (ARTFL and the Comédie-Française Registers Project) together with the Dutch-based European-funded MEDIATE project. Some of the projects discussed in the volume, including MEDIATE and FBTEE, currently have French researchers embedded in their teams: one of them, Laure Philip, is among our contributors. Even more importantly, the research agenda described and facilitated here is dependent on collaboration, scholarship, expertise and metadata from teams and institutions in France and the European Union more generally. Already of note in the chapters that follow are the Bibliothèque nationale’s (BnF) metadata services, data@bnf, and the Consortium of European Research Libraries (CERL) Thesaurus. The Digitizing Enlightenment agenda was thus from its inception, and by its very nature, internationally collaborative.

The lack of projects originating from France, at least in the context of the current volume, is perhaps indicative of the different paths digital humanities as a field has followed in France since the 1960s. After an initial flourishing of quantitative and computational approaches in the fields of linguistics and history – not to mention early forays into stylometry and author attribution studies – digital research in France, over the past two decades, has largely been limited to the development of shared infrastructure, on the one hand, and, on the other, editorial projects often tied to short funding periods and specific authors or corpora. Obviously, a project like the BnF’s Gallica platform – which combines both editorial and infrastructural activities – has developed into a world-leading resource in these domains, and more decidedly so in the post-Google Books era of the last ten years.³ At the same time, Anglo-Saxon projects have benefited from the rising profile of digital humanities in their respective countries in order to marshal more fully the necessary resources for ensuring a certain level of sustainability, often relying on a mixture of funding models and sources, and thus allowing for projects with a longer shelf-life and potential for international impact. The environment is changing, however, in the Francophone world, as new centres of research and programmes of study are being implemented at a rapid clip. Digital projects in eighteenth-century studies will certainly follow apace, exploiting a growing shared infrastructure and opening up further potential for digital humanities in the Francophone sphere. The most recent Digitizing Enlightenment events, which will take place at the International Society for Eighteenth-Century Studies (ISECS) conference in Edinburgh and then move to Montpellier, France, in 2020, attest to this necessary opening up of the original group of projects – outlined in this volume – towards the far richer and more diverse field of digital eighteenth-century studies as it is constituted today.

The original vision, as imagined in Montreal and developed in the months that followed, was, and remains, ambitious, but is rapidly taking shape. As we write, the Digitizing Enlightenment symposium has become an annual international event. The first meeting, co-convened by the editors of this book at Western Sydney University in 2016, was followed by ‘DEII’ at Radboud University in the Netherlands in June 2017, which attracted scholars from Australia, Britain, France, the Netherlands and the United States. DEIII, which was held in Oxford in July 2018, boasted participants from several more countries. As mentioned above, further symposia are being scheduled beyond 2020, and we hope that they will result in further volumes in the Oxford University Studies in the Enlightenment series. More significantly, a number of the major initiatives discussed here are working towards linking their data concerning, for example, historical agents, places, books within their various projects, including Stanford’s Mapping the Republic of Letters project, Radboud’s MEDIATE project and Western Sydney’s French Book Trade in Enlightenment Europe (FBTEE) / Mapping Print, Charting Enlightenment (MPCE) projects. Our first steps in this direction are discussed in some of the chapters that follow.⁴ As a result, this book might be seen as offering a ‘state of the art’ picture of digital humanities as a discipline and its application to a historical era and field – eighteenth-century European, and particularly French, intellectual, literary, cultural and book history – where digital methods are significantly revising the knowledge landscape.

This is an important, and some might say disputable, claim about an emerging field that has attracted some controversy, particularly since ‘digital humanities’ – a catch-all term for the application of computational methods to problems traditionally associated with the humanities and now touted as an academic field in its own right – ceased to be a term used only by specialists and entered into the consciousness of a wider public. The projects discussed here are strongly implicated in this process, in which the decisive moment may have come on 16 November 2010. On that day, the front page of The New York Times discussed and reproduced an image of an award-winning student-produced data visualization from the Mapping the Republic of Letters project.⁵ At around the same time, North American university search committees for humanities positions routinely began asking for applicants to have digital humanities skills, though it was not always clear what was intended by this phrase.⁶ These developments have been accompanied, in North America at least, by vigorous and sometimes acrimonious – if not always well-informed – debate over the rigour, implications and, indeed, value of the digital humanities.

Seven years after The New York Times published the Mapping the Republic of Letters visualizations, some sceptics have gone as far as to suggest that, for all its promises and the lavish funding it has been afforded, digital humanities research has produced little of note. These concerns came together in a headline article in The Chronicle of higher education of 15 October 2017 by Timothy Brennan defiantly entitled ‘The digital humanities bust’, which has provoked numerous responses.⁷ By and large Brennan’s critique reifies two of the most prevalent arguments against digital humanities as a field that have circulated since 2010: first, that it is a ploy by the neoliberal administrative class to win grant money and render the humanities more scientific;⁸ and, second, that its fascination with ‘data’ and ‘distant reading’ is in fact antithetical to the study of literature (or art) and the specificity of the literary (artistic) object.⁹

These critiques, it should be noted, are perhaps overly specific to their North American context, and belie a pervasive tendency to conflate the entire field of digital humanities with computational text analysis and literary criticism. Furthermore, these positions are often highly myopic from a historical standpoint: they tend to focus on the past forty years or so since the so-called ‘cultural turn’ in historiography and humanities scholarship more widely (again from a predominantly North American perspective). For the most part they overlook both previous work in the digital humanities, formerly known as ‘humanities computing’, and the enduring importance of quantification in fields like literary history, cliometrics and book history. Nor do they acknowledge the significance and achievements of the quantitative and spatial turn launched by the French Annales school dating back to the 1920s, or indeed the longer-term history of literary and historical philology that flourished in the nineteenth century. All of which to say that, rather than the ‘us versus them’ mentality espoused by many of these anti-digital humanities screeds, our vision (or version) of digital humanities is more holistic and harmonizing in terms of its relationship to the wider humanities disciplines, as the following chapters will attest. It seeks to harness the unprecedented analytical power of computational methods to synergize close readings of specific objects, persons, events or texts with the rich contextual patterns that can only be revealed by distant readings of networks, social formations or entire literary fields.

But it is not our intention here to tackle Brennan’s arguments head-on, and none of the pieces in this book does so. In so far as we address his concerns, it is by offering a series of case studies and thought-pieces on the transformative potential of digital humanities when applied rigorously in projects that span an entire field. We would suggest that, particularly in the sphere of cultural history, distant readings and the patterns digital humanities reveals have the potential to shake up existing paradigms, even on a fundamental level. To acknowledge this is not tantamount to saying that we should abandon close reading of texts – far from it. Instead, digital humanities methods give us new insights into which texts we should consider culturally significant – they help us to better understand context. This point can be exemplified by examples from the work of contributors to this book. Three we might consider are the main findings of the Stanford visualization tool featured in the New York Times article and subsequent developments in humanities data visualization (Dan Edelstein and Nicole Coleman); the identification of hidden patterns of large-scale text reuse in Diderot and D’Alembert’s Encyclopédie and the ECCO database (Robert Morrissey, Glenn Roe, Clovis Gladstone and Charles Cooney); and the discoveries of the FBTEE project concerning the circulation of religious books in the Franche-Comté (Simon Burrows).

One of the standout lessons from the Stanford student visualization in The New York Times is just how little of Voltaire’s correspondence seemingly came from Britain. This was not entirely a surprise to specialists, who were already aware from the volumes of his published letters that there were relatively few British correspondents. But visualized graphically and temporally alongside data on his entire correspondence network, the pattern became unavoidable. The famously Anglophile author of the Letters concerning the English nation presumably had little long-term contact with the British, at least according to this model. This raised new and unfamiliar questions concerning Voltaire and metadata drawn from the Electronic Enlightenment project¹⁰ – questions that the Mapping the Republic of Letters team set out and attempt to answer in a forthcoming paper, both by digging deeper into their own data and by applying the insights it provides to a rereading of the Lettres philosophiques and Voltaire’s relationship with the British. Ironically, the deficiencies in the now-famous visualization of Voltaire’s network led to the design of more humanities-appropriate tools that would allow flexible data models and a greater attention to uncertainty.

In a similar vein, researchers from the University of Chicago’s ARTFL Project,¹¹ have been inspired by sequence alignment techniques developed in bioinformatics to build tools that can identify patterns of text reuse across large humanities datasets. Two such datasets have proved particularly fruitful in this domain: Diderot and D’Alembert’s great eighteenth-century Encyclopédie,¹² and Gale Cengage’s massive Eighteenth Century Collections Online (ECCO) database of Anglophone print culture across the Enlightenment era.¹³ Leveraging the power of these tools, Edelstein, Morrissey and Roe’s path-breaking 2013 article in the Journal of the history of ideas revealed a hidden hierarchy of citation strategies at play in Diderot and D’Alembert’s philosophical dictionary.¹⁴ Authors would often smuggle in the contents of controversial or illegal works through the practice of loose, unreliable or simply missing citations. This meant that modern readers, aside from a few specialists, were unable to recognize the complex network of hidden citations weaved throughout the text, and, as such, were likely to assume that important but controversial works by Voltaire, Montesquieu, Locke and Helvétius, among others, had no bearing on the Encyclopédie. By identifying the dialogical layers of both seen and unseen citations at work in the Encyclopédie, we now have a more thorough understanding of the encyclopédistes’ subversive reuse of philosophical texts as well as their creative refashioning of traditional sources to undermine authority.

Turning these same algorithms loose on the 200,000-volume ECCO database,¹⁵ a data mining project of daunting proportions, the ARTFL team next sought to tackle the common literary culture of the eighteenth century through its printed output. The standout news here, however, was that the majority of identifiable ‘commonplaces’ – loosely defined as two or more similar instances of text reuse – came from a single, largely unexpected source: the Bible. In fact, over 35 million reuses, or 58.5 per cent of all reuses identified in ECCO, were flagged as originating from the Bible. The Bible’s omnipresence thus reflected its cultural dominance and familiarity, even in the age of Enlightenment.¹⁶ For eighteenth-century historians, this perhaps came as a shock. It is unlikely that any Enlightenment specialist, when confronted by a bright student asking about the best book to read in order to understand the eighteenth century, would ever hitherto have imagined giving the same reply as the celebrated Marxist historian of the seventeenth century, Christopher Hill, when asked the same thing for his era: Hill’s unexpected reply was ‘the Bible’. Enlightenment historians, particularly of France, are frequently attracted to the eighteenth century as the apparent cradle of secular and political modernity – even if we debate the accuracy of such a depiction – culminating in what François Furet and his collaborators called the French Revolutionary ‘creation of modern political culture’.¹⁷ But the commonplace evidence suggests that Enlightenment specialists, too, would do well to learn their Scriptures if they wish to better understand Enlightenment high culture and literature, at least in Protestant countries like Britain.¹⁸

Religion also features large in the FBTEE findings on the book trade in Besançon and the Franche-Comté. In a series of articles based on an industrial-scale digital survey of French publishing and bookselling, Simon Burrows has shown that it can be demonstrated that, in less than two decades before 1789, at least 50,000 copies of a single popular religious text, a devotional manual entitled L’Ange conducteur, were printed or circulated in the Franche-Comté. Since this count depends on sources that are very far from complete, it seems likely that these numbers are a significant underestimate. But, even if we use the 50,000 figure, this would seem to imply that this devotional manual penetrated into one in three or one in four households in the province.¹⁹ The actual number is probably more impressive still. This is significant. The surveys of library holdings, licensed book output by title, and book reviews in scholarly journals by the Livre et société group variously suggested that religious books comprised 8 to 25 per cent of the book market; and surveys of will inventories in the Franche-Comté had suggested that about 6 per cent of peasants – who comprised the vast bulk of the population – left books, and that 80 per cent of these usually tiny peasant libraries were exclusively or partially comprised of religious titles. Yet FBTEE had solid evidence suggesting real ownership rates in the Franche-Comté, for a single title, were five times as high. The missing evidence might double that figure. This has serious implications for our understanding of the output of religious print, popular literacy and reading practices, and popular religiosity and religious experience on the eve of the Revolution. When we compare the volume of religious titles to those for secular works traded by Besançon bookdealers, it also reinforces impressions that Enlightenment secularism took hold primarily among an urban elite, a finding that may cause us to rethink standard narratives of French secularization.²⁰

Taken together, these three examples, and further studies in this book, challenge us to revise received interpretations of the Enlightenment in relation to its leading actors, its elite, intellectual and popular culture, and its reading tastes and habits. So do our other contributors. It is no accident that the European eighteenth century should have produced such significant digital work. For in many ways, of all historical periods, the eighteenth century might be said to offer the perfect laboratory for applying digital technologies.²¹ It is also well served with digitized materials. For while Europe’s surviving archive of eighteenth-century print and manuscript materials is vast – particularly in comparison with previous eras – it is finite enough to dream of comprehensive datasets. Students of Britain, for example, can through pioneering digital collections consult almost the entire print archive through Gale Cengage’s magnificent twin resources, the ECCO database and the Burney collection of newspapers, as well as ancillary collections of historical newspapers and periodicals. ECCO in particular is strikingly rich, aspiring to contain a copy of at least one edition of every work produced in Britain (in any language), and of every other work printed in English worldwide. In all it contains some 200,000 works in fully OCR-searchable format, together with well-structured bibliographic metadata drawn from library MARC records. Historians of the nineteenth century do not enjoy similar luxuries. The explosion of print literature and advent of mass education in the so-called ‘second print revolution’, following the introduction of the steam press in 1814 and further print and image technologies thereafter, mean that any attempt at a comprehensive collection would be prohibitively expensive, even for the world’s leading providers of digital educational resources. Thus, when Gale Cengage set about developing its nineteenth-century collections online, it decided to split the materials thematically – thus putting discrete parts of the historical archive within reach of libraries unable to afford the whole collection (since Gale Cengage’s pricing varies by size of institution and is subject to negotiation, prices vary, but a medium- to large-sized university might expect to pay well into six figures in US dollars for ECCO).

Historians of France do not currently have easy access to such carefully structured and comprehensive collections, but they do enjoy other advantages. In recent years, a significant portion of the French Enlightenment literary output has become available in freely downloadable form through the Gallica portal at the Bibliothèque nationale de France (BnF).²² Gallica is also assembling a formidable library of manuscript materials, in part by uploading documents whenever copies of new materials are ordered from the BnF by researchers. Equally the BnF has begun supplying researchers with raw metadata through its prize-winning open data hub data@bnf. For the research community, the advantages are huge. Gallica in particular has facilitated research worldwide and levelled the playing field for researchers operating beyond France and Europe. It is not accidental that this book is being edited by two Australian-based scholars of eighteenth-century France – and three further scholars wrote their chapters while living there.²³ For these scholars, digitization, particularly by Gallica, has broken the tyranny of distance, bringing research and career opportunities undreamed of by Antipodean-based Europeanists only a few years ago.

Scholars of eighteenth-century France enjoy two further resources relevant to this book. The first, the ARTFL (American and French Research on the Treasury of the French Language) Project, a French–American collaboration hosted at the University of Chicago, is one of the oldest and better-known focal points for digital humanities research and development in North America. One of ARTFL’s main goals is to provide access to a growing collection of digitized French texts from the thirteenth to the twentieth century. Over the years, ARTFL has progressively made several important eighteenth-century resources available to the general public, including the first digital edition of Diderot and D’Alembert’s Encyclopédie, a suite of historical dictionaries, the complete works of Voltaire and Rousseau, and a host of other related content such as Grimm’s Correspondance littéraire or, most recently, the abbé Raynal’s Histoire philosophique des deux Indes.²⁴ Brought together, and made accessible using ARTFL’s open-source full-text search and analysis system PhiloLogic,²⁵ these collections have quickly become essential tools for literary and intellectual historians of the Enlightenment.

Less expansive than ARTFL, though no less impactful, the second resource we should mention is Electronic Enlightenment, a commercial subscription product available through Oxford University Press, which as of its autumn 2017 update contains over 77,000 letters and documents exchanged by members of the vast European Republic of Letters. Along with providing critical editions of these letters in digital form, Electronic Enlightenment also provides searchable relational databases of historical figures, place names, maps and manuscript images where available, all of which are linked through extensive annotated metadata. As we shall see, this metadata originally formed the heart of the Mapping the Republic of Letters project at Stanford and provides a central focus for attempts to link the various projects discussed here.

Both ARTFL and Electronic Enlightenment are, in digital terms, venerable projects. ARTFL dates back to 1981, while its 150-million-word foundational corpus of transcribed texts was prepared even earlier, between 1957 and 1977.²⁶ The Electronic Enlightenment project began in 2001 at the Voltaire Foundation in Oxford, and a first version of the database was made available in 2006. This should serve to remind us – if reminders are necessary – that the term ‘digital humanities’, in use among practitioners around the turn of the millennium and in wider currency from around 2010 – is just the latest phrase used to describe work that fits into a longer tradition of computational work in arts and humanities disciplines.²⁷ Two early practitioners of the discipline formerly known as ‘humanities computing’ have written for this collection: Robert Morrissey, director of the ARTFL Project, has already been mentioned; the second is Angus Martin, who since the 1980s has been working on a digital expansion of the print Bibliographie du genre romanesque français that he and his collaborators the late Richard Frautschi (d. 2016) and the late Vivienne Mylne (d. 1993) published in 1977. Housed in late 1980s software, this comprehensive bibliographic database is currently being transferred into more recent formats in collaboration with members of the FBTEE team, who hope to make it interoperable with FBTEE data.²⁸

The collaborative efforts involved in such a data recovery are typical of digital humanities projects, which are seldom undertaken by single individuals. The new model of humanities scholarship that they exemplify also encourages making data available on an open access and knowledge transfer basis (though some expensive projects necessarily rely on subscription models – Electronic Enlightenment and, for access to its main corpus, ARTFL are the standout examples from projects discussed here). Such exchange of data and information is a practical necessity if we are to fully realize the opportunities digitizing (the) Enlightenment can provide, since the projects discussed here frequently treat the same people, places, temporalities and texts from different angles. Across the next decade, as we learn to align our datasets along linked data principles, it is probable that the best digital humanities scholarship will come not from individual projects, but rather from our attempts to interrogate the intersections and interstices between our various datasets, to uncover and explore patterns of behaviour, activities, cultural production and relationships that were not visible before. The chapters that follow reveal what is already being achieved on the level of individual projects exploring both data and metadata. They also give hints of what lies around the corner as we move to a world of linked open data in the digital humanities. We present it as a catalogue of what Digitizing (the) Enlightenment has achieved, and a blueprint for what it still aspires to achieve.

Hence our exploration begins with a discussion of a number of leading digital eighteenth-century projects, how they came about, what they aimed to achieve, and their key findings in the words of their instigators and their close associates. These chapters and the projects they describe deal with a variety of the most significant cultural and intellectual artefacts of the period, namely the leading compendium of scientific knowledge and philosophical thought, Diderot and D’Alembert’s Encyclopédie; the major literary and intellectual correspondences available; the Republic of Letters as a transnational network of intellectual exchange and dissemination; the theatre as a social phenomenon and its reception; and the book trade and history of reading in all of their historical and cultural complexity. As we hope to demonstrate, there are significant overlaps between these projects, in terms of personnel and institutions, certainly, but also, and more significantly, in terms of their importance as discrete entry points into the larger literary and cultural fields of Enlightenment Europe. The Encyclopédie, to take just one example, becomes, thanks to its complicated publication and reception history,²⁹ a useful starting point for thinking about the dissemination of ideas, state censorship, authorship and authority, the illegal book trade, and a host of other germane subjects, all of which are touched upon in some manner or other by the following projects. Imagining a general interface that could enable this sort of transversal scholarly exploration, in as frictionless a manner as possible, is thus an inevitable outcome of this book. As digital humanities methods and practices become more prevalent and the barrier to entry less costly, it will fall to established teams and projects to more efficiently link and bring together the various threads that bind them. This volume is a first step, we hope, in that direction.

Our opening chapter, by Robert Morrissey and Glenn Roe, deals with the oldest of the projects featured in the book. It discusses the ARTFL Project’s undertaking of the first digital edition of Diderot and D’Alembert’s monumental Encyclopédie (1751-1772). Begun in the mid-1990s, when digitization protocols and text encoding specifications were far from standardized, the ARTFL Encyclopédie has become over the past twenty years a shining example of the flexibility and tenacity needed to sustain a large-scale digital edition. In addition, the Encyclopédie database has in more recent times become an almost ideal test bed for experimenting with so-called ‘big data’ computational methods developed in the fields of information retrieval and machine learning. Conceived itself as an information technology designed to constrain what Ann Blair has called the ‘information overload’ of the early modern era,³⁰ the Encyclopédie adopts an aesthetic dimension that embraces abundance and copia while at the same time promoting a posture of critical engagement and reduction. As Morrissey and Roe demonstrate, the ‘hermeneutical aesthetic’ put into play by Diderot and D’Alembert perhaps finds its fullest elaboration only within the digital medium, allowing users to engage with this seminal text in new ways and thus ensuring its relevance for the information age.

A second chapter by Nicholas Cronk gives a summary history of the Electronic Enlightenment (EE) project. Begun in 2001 as a project to digitize existing critical editions of Enlightenment correspondence collections, and starting with the Voltaire Foundation’s editions of Voltaire’s and Rousseau’s letters, EE has since become the digital clearing house for more than 77,000 letters by over 10,000 correspondents, making it the most wide-ranging online collection of edited correspondence of the early modern period. Cronk details here the various editorial and technical decisions that were necessary to realize such an ambitious project, which included a new business model for the sharing and online dissemination of copyrighted material such as scholarly notes and apparatus. More than just a simple digital edition of letters, however, EE has over time developed a data-rich interface linking people and ideas across Europe, the Americas and Asia from the early seventeenth to the mid-nineteenth century. Here again the transformative nature of digital remediation is important to underscore: print editions of correspondence are decidedly one-sided in terms of information exchange and interactivity. By bringing multiple scholarly witnesses of correspondences together into one digital interface, EE not only makes available valuable primary sources that would otherwise remain unconnected on library shelves, but also links scholarly annotation, critical apparatus and related metadata in ways that print editions simply cannot. Thus, the digital reconstruction of the Enlightenment Republic of Letters enacted by EE is also, importantly, a first step in reimagining critical editing practices for the digital age.

Metadata from Electronic Enlightenment was also foundational for Stanford’s Mapping the Republic of Letters (MRofL) project, which is discussed by Dan Edelstein in the next chapter. Dan outlines the major issues involved in the aggregation and visualization of correspondence networks across Europe and North America, aimed at approximating the historical transnational Republic of Letters geospatially by joining together several major datasets and developing new tools designed specifically for the task. Work on this project has been widely discussed in the media, leading eventually to the now-famous student visualization that was published in The New York Times in November 2010. But, as this chapter demonstrates, the road to this watershed moment was far from straight and in no way lacking in obstacles. As with many digital humanities projects, people are often the intangible element and primary resource that no amount of grant money, technological gadgetry or institutional prestige can replace. This was certainly the case for MRofL, and is equally valid for the other projects outlined in this book. In a lesson for posterity, Edelstein modestly recounts how he and his team benefited from incredible luck, modest financial assistance and fortuitous encounters, all of which allowed them to ‘stumble’ into a digital humanities project that transformed their careers.

Keeping with the theme of early modern correspondences, the next chapter deals with the Cultures of Knowledge (CofK) project led by Howard Hotson at Oxford. Since 2009, CofK has used a variety of research methods to reassemble and understand correspondence networks across the European Republic of Letters over a 250-year time span. One of the key outcomes of the CofK project was the development of a digital union catalogue of sixteenth-, seventeenth- and eighteenth-century correspondence, called Early Modern Letters Online (EMLO). Subsequent phases of the project focused on metadata aggregation, systems development and community building necessary to turn EMLO into a major resource in the field. Hotson then outlines a future vision for both CofK and EMLO that will involve defining new capabilities for EMLO on the basis of eight years of experience and dialogue with a wide range of partners, including the MRofL project at Stanford. As with many of the projects featured in this volume, the promise of linked open data (LOD) looms large for CofK researchers, who plan to launch Early Modern People, Early Modern Places and Early Modern Dates as free-standing LOD resources that will be capable ultimately of being used and populated independently of EMLO as well as integrated and interlinked with other relevant linked open data resources.

In ‘The Comédie-Française Registers Project: questions of audience’, Jeffrey S. Ravel explains how the Franco-American Comédie-Française Registers Project (CFRP) is giving new insights into theatre audiences and the reception of drama across the eighteenth century. At the same time, Ravel explores the changing nature of audience in terms of scholarly communication and publishing given the ubiquity of the web and many digital projects’ recourse to ‘crowdsourcing’ or citizen science today. Indeed, the blurring of the boundaries between humanities scholarship and its audience in the digital age reminds Ravel of the earlier efforts to assert cultural authority that characterized the early modern theatre. This process of audience building has left the CFRP team uneasy at times, as they have struggled to identify the audience they hoped to address. But the tensions and uncertainties around the project have also been transformative, as the team grapple with the meanings of scholarship in an age of social media. It is both disconcerting and liberating, Ravel argues, to consider ones’ disciplinary and professional boundaries in a mediated environment that incessantly expands them. After narrating the history of the CFRP project, its origins, personnel and funding sources, and discussing potential future directions and collaborations for the project, Ravel ends the chapter on a cautionary note. This narrative of seamless project success, it seems, also belies endless stops and starts, failures and triumphs, and a general anxiety that appears to be an inevitable part of any digital humanities project.

The final three chapters in this section describe three closely related projects which explore the production, dissemination and reception of print culture. Angus Martin and Richard Frautschi’s chapter ‘Towards a new bibliography of eighteenth-century French fiction’ discusses of the fifty-year history and development of the Bibliographie du genre romanesque français database (MMF2) from its inception as an analogue era project to recent efforts to recover the data and reformat it for publication. Much more than a mere data recapture project, Martin’s lucid description of the promise and pitfalls of bibliographic research and the changes it has undergone over the past half century is an invaluable lesson for those of us who may take digital resources for granted. The history of MMF2 is thus, in many ways, also the history of information technologies as they develop first computationally and then move to the wider web. In this new context of information overload, the value of bibliographic knowledge and the precise research methods used to establish it are as relevant as ever, and perhaps even more so, in this, the age of Google.

The value of the MMF2 data is also one of the themes explored in Simon Burrows’ chapter, ‘The FBTEE revolution: mapping the Ancien Régime book trade and the future of historical bibliometric research’. It offers an overview of the history and methods of the French Book Trade in Enlightenment Europe (FBTEE) project and its latest incarnation, Mapping Print, Charting Enlightenment. This, as Burrows explains, is an industrial scale ‘big data’ project to trace the contours and trajectory of French print culture in the twenty years before the French Revolution. Although many of the sources it uses have been extensively mined by previous historians, Burrows shows that by amalgamating, enriching and analyzing this data as a whole, important new insights and patterns emerge. In particular, Burrows emphasizes the supreme importance of religious works as a proportion of printed outputs, and relative marginality of themes which have dominated much recent writing.

While Burrows’ chapter focuses on the supply side of the print industry, Alicia Montoya’s allied project on Middlebrow Enlightenment: Disseminating Ideas, Authors, and Texts in Europe, 1665-1830 (MEDIATE) is concerned with the larger literary field of the eighteenth century and the relational nature of literary valuation and status in the ‘Enlightenment’. Montoya’s chapter thus outlines several lessons learned over the relatively short lifespan of MEDIATE, the most recent of our featured digital projects. The aim of MEDIATE is to study the circulation of books in eighteenth-century Europe, focusing on the movement of books and ideas traditionally associated with the concept of ‘Enlightenment’, by building an open access database housing data from a corpus of catalogues of private libraries sold at auction in the Dutch Republic, France and the British Isles between 1665 and 1830. In ultimately problematizing the notion of which ‘Enlightenment’ might be worthy of digitizing, Montoya seeks to understand which ideas and books were embedded into the cultural field of the eighteenth century by way of the notion of the ‘middlebrow’, or forms of literature that situated themselves between elite forms and more popular ones. By focusing on the various definitional issues and problems encountered in thinking through the technical and conceptual nuts and bolts of their first project year, Montoya offers us insights into how to manage shifting expectations and the inevitable setbacks that are all too common in digital humanities projects.

These chapters on large, well-funded, high-profile projects are intended to contextualize the second part of the book, which concerns digital methods and innovations. This section will discuss methods and technologies developed by the various projects described in Part 1, including interventions from some of the digital designers involved, as well as new research from emerging scholars in the field, who have in some cases achieved impressive and important results on limited budgets. They focus on the technical challenges faced by researchers and developers working at the intersection of digital methods and Enlightenment studies, and how they were addressed and overcome. At the same time, they address important further questions, questions that have troubled digital humanities sceptics and practitioners alike. How can digital approaches generate new knowledge and new research questions? What is the epistemological status of digital models and visualizations in terms of scholarly evidence and argumentation? Are methods and techniques developed in the computer and information sciences applicable to humanities questions and datasets? How can digital humanities projects ensure their results are open and, ideally, reproducible? What are the significant decisions, compromises, conceptual challenges and lessons learned along the way?

Hence Nicole Coleman in ‘Seeking the eye of history: the design of digital tools for Enlightenment studies’ discusses the design and evolution of the Palladio visualization tool, the application Jo Guldi describes as ‘what the historian sees when she closes her eyes’. She explains how Palladio grew out of the shortcomings of previous visualization tools to help answer actual research questions in the humanities. The trigger for the Stanford team to become actively engaged in the design of tools for historical research was, ironically, the prize-winning data visualization featured in the New York Times article that we have already discussed. It was first produced in 2009 by students in Jeff Heer’s data visualization class at Stanford specifically for the Mapping the Republic of Letters project, and offered exciting new possibilities for producing knowledge. The visualization let historians explore seventeenth- and eighteenth-century correspondence in a novel way: early modern communication networks, previously only imagined in the minds of historians, could now be seen graphically as pulsing arteries connecting the centres of Enlightenment thought. Yet there were many things that could not be captured in a single data visualization, and so the chapter explains how decisions were taken to extract the most effective graphic views of data in prototype tools and bring them together in one application – Palladio. This new tool is designed to support a distinctively humanistic mode of enquiry by providing an environment that enables thinking through data. Coleman argues that, since historians work with fragments of information from the past, not complete datasets, the discipline does not offer models against which to test hypotheses. Instead, tools need to support scholars in building an understanding of the historical material through working with it. Palladio thus seeks to externalize what is a very internal, individual, thought process.

Our second chapter in this section, Elizabeth Bond and Robert Bond’s ‘Topic modelling the French pre-Revolutionary press’, applies topic modelling methodologies to evaluate the merits of this popular digital humanities technique for the study of Enlightenment France. Recent studies of digitized eighteenth-century newspaper collections have employed topic models to validate well-known, established trends in the historiography, and to potentially uncover other trends historians might have previously dismissed. The sources under scrutiny here are readers’ letters to the editors of provincial newspapers, or affiches, in the 1770s and 1780s, historical sources which have frequently been dismissed by historians as mere advertisers with little political or high cultural content or engagement with public affairs. This chapter argues on the contrary that readers of affiches penned letters on myriad topics – from the arts, philanthropy and belles-lettres to scientific observations, medical reports and product advertisements. With the assistance of a team of students, hundreds of letters were prepared in a text format suitable for topic modelling. The evidence presented here suggests that the Enlightenment was a largely practical project guided by the desire to improve society and be useful to humanity. But it also helps to challenge established historical opinion about the (assumed) marginality of current affairs to the affiches and to highlight overlooked trends.

Katherine McDonough’s chapter, ‘Putting the eighteenth century on the map: French geospatial data for digital humanities research’, discusses geospatial mapping, historical gazetteers and geographic information systems (GISs), subjects that lie at the heart of several projects discussed in this book. McDonough explores the challenges and opportunities arising from attempts to map and visualize Enlightenment geographic space, particularly in the context of France, a country for which we still lack a historical gazetteer. To address this gap in knowledge, she and her collaborators attempt to leverage the reference sources at the heart of the Enlightenment project (like Diderot and D’Alembert’s Encyclopédie) and use these sources to unpack spatial questions about other sources. These spatial questions matter for the eighteenth century, and the early modern period more broadly. Mapping is, however, a far from neutral activity, and the decisions we take about representing geographic space through digital media shape not only the questions we can ask, but also the analyses we can make and the answers we receive. Faced with such challenges, and in the light of other national mapping projects, how should we best map Enlightenment France?

If mapping is not a neutral activity, nor are the choices surrounding database structures, ontologies and protocols. As a relative newcomer to the digital humanities, recruited to the mature and rapidly expanding FBTEE project, Laure Philip reflects critically upon these issues and the significance of decisions taken at the database conceptualization and data entry phases. In a chapter entitled ‘The illegal book trade revisited: an insight into database protocols and pitfalls’, she demonstrates the difficulty of decoding and interpreting documentary sources, of recording complex real-world transactions in inconsistent and often cryptic ways, and of capturing them in digital formats. In the process, she offers sobering reminders of the impact of practical decisions and research practices on the outcomes we generate, as well as the ways in which the digital frequently renders transparent to other scholars what was hidden or assumed by analogue methods. She also makes the case for considering the outputs of data entry on complex archival sources as significant historical research and indeed historical interpretation in their own right. Her thoughts are worthy of reflection by anyone considering historical database research, and any scholar or assessor seeking to evaluate historical digital projects.

If our own choices are not ‘neutral’, neither is our available source material free from biases, as any student of History 101 knows. In the case of the Enlightenment salons, for example, we generally lack formal attendance registers, and thus are forced to rely mainly on recollections in memoir sources to compile our lists of salon attendees, which are far from complete and probably biased in favour of well-known and elite guests. Nevertheless, Melanie Conroy and Chloe Summers Edmondson argue in their chapter, ‘The empire of letters: Enlightenment-era French salons’, that digital prosopological and network analysis of surviving lists can tell us a great deal of significance about the evolving nature of salons, their attendees and the relationship with Enlightenment correspondence networks, whose memberships overlapped substantially with those of the salons. Confirming the conclusions of other explorations of ‘the Enlightenment network’ by the Mapping the Republic of Letters team,³¹ Conroy and Edmondson demonstrate the relative marginality of scientific in comparison with literary authors, whilst adding an important chronological nuance by demonstrating that the proportion of attendees with political and philosophical writings rose significantly as the eighteenth century progressed. Further, the striking degree of convergence between the great philosophes’ correspondence networks and salon membership establishes that both can be seen as manifestations of a single French Enlightenment network, dominated by aristocratic and other elite groups. That similar patterns are discernible from digital analysis of multiple datasets validates both research findings and methodologies.

In our final chapter, ‘Opening new paths for scholarship: algorithms to track text reuse in Eighteenth Century Collections Online (ECCO)’, Clovis Gladstone and Charles Cooney, based at the ARTFL Project in Chicago, describe their attempts to identify patterns of large-scale text reuse in Gale Cengage’s ECCO database. Given the size of this collection, as well as the state of the data in terms of its OCR output, identifying shared passages that are relatively short, repeated and rhetorically significant is a non-trivial computational task. The scope and scale of the ECCO dataset thus represented a major hurdle in terms of both computational expense and the evaluation of the matching algorithms. The chapter explains how Gladstone and Cooney met these challenges – compounded along the way by the overwhelming preponderance of the Bible in terms of overall reuse – by adapting several known sequence alignment algorithms borrowed from bioinformatics and developing new ones for refining matches. The resulting database of textual reuses and its search interface make an entire literary culture available in a form and on a scale previously unimagined.³² As they carefully lay out in their case study on the reception and reuse of Lucretius in the eighteenth century, this sort of ‘big data’ approach can transform fundamentally the questions we ask of our data as well as the manner in which we can now explore millions of textual relationships with the click of a search button. Digging into a dataset such as ECCO can thus offer us new perspectives from which to view and understand eighteenth-century print culture as a whole, provided – as with all the projects discussed here – we unearth more than we obscure.

In fact, we would contend that the rigours of digital humanities work render transparent what was frequently obscure in traditional scholarship. The possibilities for digitally recording and presenting both the original data and how we have interpreted it or enriched it expose each step and item in the preparation and analysis of data to potential scrutiny and reinvestigation. Since, as this implies, the gathering and curation of historical data are, to a certain extent, both an interpretative and a creative process, digital humanists in general, and the contributors to this volume in particular, showcase an acute methodological awareness. This is doubly necessary because most digital humanities projects are by their very nature team (ad)ventures involving participants with differing skill sets. Team members need to communicate effectively in order to translate research aspirations into digital tools, structures and standard protocols suitable for recording in consistent digital form, housing and meaningfully analyzing the residues of the past while maintaining, as far as possible, the integrity of their original sources.

Thus, on many levels, the contributors to this volume are describing, showcasing and envisioning a new, collaborative, technologically empowered, methodologically transparent, interdisciplinary form of humanities scholarship. It is collaborative not merely within multi-member projects, but between and across them. Much like collaborators on Diderot and D’Alembert’s Encyclopédie or citizens in the enlightened Republic of Letters imagined themselves as participants in a common and revolutionary knowledge project, contributors to Digitizing Enlightenment are in many ways participating in the reinvention and transformation of scholarly practices in the humanities at the dawn of the digital age. This common vision involves more than breaking with the traditional humanities paradigms based on the single-author research monograph and the systematic pooling of research data and metadata. The co-operative future will be collaborative by pragmatism as well as instinct, for the most valuable and enduring work in the field of digital humanities will be that which produces distinctive and problem-focused research outputs, and interoperable, user-friendly digital resources built on linked open data principles wherever possible. The chapters that follow reveal how much has already been achieved in this direction and the challenges yet to be overcome. At the same time, they give some exhilarating glimpses of possible futures for our chosen field of Enlightenment studies and for humanities scholarship more generally. We hope our readers will share in our excitement, for the digital humanities revolution is only just beginning.