Networking Print in Shakespeare’s England
Influence, Agency, and Revolutionary Change
Blaine Greteman



IN EARLY MODERN ENGLAND, preachers were fond of warning their flocks that “flattery is the Devil’s invisible net, by which he catcheth and holdeth men fast in the snare.”1 Although it too has mostly remained invisible, the net described in this book is by contrast a record of the relationships through which humans act and express their agency. This is why Bruno Latour has impishly suggested that the term “network” should really be inverted to “worknet or action net”—when we act in the world, we do so by cultivating, extending, and maintaining networks.2 I’ve attempted to show that analyzing these networks can yield important historical and cultural insights, such as the fact that a gradual evolution of the printing trade led to a sudden revolution in the print network around the end of the sixteenth century. In the small new world that emerged, nearly everyone could be connected to everyone else within a few steps, and information could seem to spread like an epidemic or “public fire.” Hyperconnected hubs played an outsized role in this network’s connectivity, although they have not necessarily played such a role in our historiography—printers like Nicholas Okes and booksellers like Michael Sparke and Giles Calvert reconfigured their trades in ways that made them structurally and culturally significant. Also underrepresented in histories of the period are the texts and authors that linked disparate communities in the print network, and by turning our attention to the high-betweenness texts of female prophets like Mary Cary and Eleanor Davies we can appreciate the unusual publication strategies that made their texts important bridges. Finally, even authors at no risk of being forgotten, like John Milton, look very different when we understand the weak ties that shaped and enabled their emergence as strong voices. Like the other figures in this book, Milton may have lacked the language of network analysis, but he was sensitive to the evolving communications landscape and adept at networking strategies that allowed him to speak, with some authority, as the voice of a wider spiritual, intellectual, and national community.

But this book is an opening, rather than a final word, on the questions that network analysis can pose within the field of literary studies and book history, and I would like to conclude by pointing to ongoing projects and developments that might expand, deepen, and complicate my own work. Although this is one of the first books to develop and use digital analysis to investigate early modern networks, it could not have been written without a large and growing community of scholars who are investigating other, related networks and associations. Just naming the projects and institutions that have provided important support, feedback, and inspiration as we created the Shakeosphere website could fill a chapter: the University of Iowa’s Atlas of Early Printing, the University of Victoria’s Map of Early Modern London (MoEML), the Medici Archive, the Center for Spatial and Textual Analysis at Stanford, Early Modern Letters Online (EMLO), Six Degrees of Francis Bacon, Networking Archives, and many more. Our understanding of and ability to analyze complex, multilayered networks will become more powerful as we connect such related but distinct projects through Linked Open Data frameworks—after all, a member of the print network may also be engaged in an epistolary network, a scientific society, a particular geographical community, and an extended kinship circle. Linking projects will not only provide more context for the networks of persons, places, and things found in each of these resources, but will also lend more power to methods of network analysis such as link prediction, which may allow us to identify missing collaborators and anonymous authors by their likely association with known figures. But in order to talk about the future of this work, I also need to address a question that has become urgent as many humanities programs have come to face the dire realities of funding scarcity: Are the digital humanities a salvation, or a threat?

After rapturous early coverage, including a short-lived New York Times Book Review column called “The Mechanic Muse” that was devoted exclusively to covering forms of “distant reading,” we have arguably entered the phase of backlash and corrective.3 Timothy Brennan gleefully announced the “digital humanities bust” in the Chronicle of Higher Education in 2017, explaining that “For all its resources, the digital humanities makes a rookie mistake: It confuses more information for more knowledge.”4 Tom Eyers similarly accuses the digital humanities of “uncritical positivism,” creating an entire academic subculture that is “thrusting, confident, often garlanded with money, and finds an especially welcoming home in those largely private universities that snap at the heels of the very top tier of institutions of higher learning.”5 And Nan Z. Da has attacked the digital humanities, or “Computational Literary Studies” (CLS), for failing to deliver on their early promise, for a lack of statistical sophistication, and for “the exorbitant funding that CLS has garnered.”6 Some of these arguments—such as Da’s case that digital humanities projects are too often opaque and should make their data more openly available so others can interrogate or reproduce their results—are fair and becoming the norm, including with my own project, where underlying data are available through Github as well as through the Shakeosphere website. But others are frankly wrong in ways that are nevertheless illuminating for thinking about future development of the field, and I think they merit brief mention here.

Through the account presented in Chapter 1 of the ESTC’s history and our transformation of it into a new kind of data source, I hope I’ve already shown that the notion of the digital humanities as either a new and exciting phase in the discipline, or an existential threat to it, is something of a false dichotomy. First, the oft-repeated idea that digital humanities is some radical upstart relies on a strangely truncated view of scholarly history. Although literary studies labeled “distant reading” or “macroanalysis” are fairly new, we are well into the third or fourth generation of digital humanities work; the earlier phases of “humanities computing,” which included ambitious projects of digitization and database construction, emerged along with consumer computing in the 1960s and 70s.7 Even before that, Roberto Busa worked with Thomas Watson and IBM in the 1940s and 50s to create the searchable corpus of Aquinas’s works known as the Index Thomisticus—a work that belongs equally to the history of computing and to the tradition of indexing, annotating, and organizing that Ann Blair describes as fundamental to the humanist tradition from its inception.8 Like early computers themselves, early digital projects were costly and limited, but they laboriously laid the groundwork for much of what we do today. Incidentally, and against the idea that digital humanities scholarship today is larded with money, the analysis made possible by that early, costly work can typically now be done on a laptop, which is how I conducted all of the network analysis for this book (although sometimes the laptop needed to be left running overnight).

Nearly all humanities scholarship today, including that which openly rejects computational methods of analysis, is made possible by the digital finding aids, catalogues, and archives that have in many cases been built over previous decades. One of my goals throughout the book has been to show how such tools and data can silently inflect our knowledge of the field and to suggest that we actively need to engage, and sometimes repurpose them, to understand areas where that knowledge needs revision. As Marlene Manoff argues, library collections and “digital archives represent the entanglement of matter and meaning, content and device, and human and machine elements,” with “content produced and selected by individuals in particular social and historical contexts, but it is also shaped by multiple additional factors, including the hardware and software that enable access to and manipulation of that content.”9 The elucidation and analysis of those factors seems to me an especially appropriate goal in a book about the early modern period, which was marked by curiosity and skepticism about the ways in which telescopes, microscopes, and other instruments might reform or distort our view of the objects we study.0

To use any scholarly tool uncritically—whether it is the Short Title Catalogue of Books Printed in England, Scotland, and Ireland (STC), the British Library’s online catalogue (ESTC), or the NetworkX Python package—is to risk the kind of mistake made by the astronomers in Samuel Butler’s parodic poem, “The Elephant in the Moon,” as they cluster around a telescope and witness “two mighty armies” on the moon, “in a bloody fight engaged. . . . As by the glass ’tis clear, and plain.”11 While they rush off to publish their accounts of the celestial battle, a footboy discovers the deflating truth that flies, gnats, and a mouse have become trapped between the telescope’s lenses, and that their grotesque and distorting magnification has proven the “virtuous occasion / Of all this learned dissertation.”12 Many a learned dissertation has been written in the same way, under the influence of catalogues and archives that magnify certain categories—such as the work, the subject, and above all, the author—at the expense of others, such as the printer, the community, or the network. The analysis I’ve done offers an alternative perspective, although I’ve also tried to approach it with the footboy’s curiosity about both what he sees and why.

Of course, doing digital analysis does not exempt one from the charge of using tools uncritically or ineffectively. But in this regard, too, critiques of the digital humanities tend to rely on a narrow definition of digital scholarship that simply does not represent the field as it exists today. Eyers admits to an “inevitable flattening” of approaches before arguing that Franco Moretti’s version of “distant reading” emblematizes the entire field.13 Da states plainly that

CLS papers are all more or less organized the same way, detecting patterns based in word count (or 1-, 2-, n-grams, a 1-gram defined as anything separated by two spaces) to make one of six arguments: (1) the aboutness of something; (2) how much or how little of something there is in a larger body, (3) the amount of influence that something has on something else; (4) the ability of something to be classified; (5) if genre is consistent or mixed; and (6) how something has changed or stayed the same.14

Da’s claims led to a lively debate, hosted by Critical Inquiry, in which the scholars she cites as representatives of such arguments have defended their own work, and I will let them make their own case for the value of their methods.15 But I should add that my book and its underlying digital analysis departs completely from the kind of corpus analytics that Da and Eyers discuss, and it is hardly the only digital humanities project to do so. Indeed, the “digital humanities” is less a specific set of methodologies or a distinct field than a “tactical term,” as Matthew Kirschenbaum has put it, deployed to “get things done.”16 None of the projects I mentioned above rely on “detecting patterns based in word count” or anything like “distant reading” as it is usually described by its critics.

The Atlas of Early Printing, for example, documents, maps, and visualizes the output of paper mills and printshops throughout Europe from 1450 to 1500.17 The Map of Early Modern London uses a digitized version of the 1633 “Agas map” of London to visualize literary and historical data, with layers depicting locations like chapels and playhouses, an encyclopedia of named places from literary and historical documents, and editions of topographical texts.18 The Translation and the Making of Early Modern English Print Culture project at the University of Montreal analyzes and maps the communities of translators, publishers, and printers who often played multiple roles in reproducing continental works in England.19 The Women in Book History Bibliography20 aggregates resources about female printers, publishers, and booksellers and has led to revisions in the new, more flexible ESTC 21 being developed by Brian Geiger and Carl Stahmer.21 Six Degrees of Francis Bacon mines the Oxford English Dictionary—and increasingly, the contributions of its users—to construct probabilistic maps of social connections during the early modern period.22 The English Broadside Ballad Archive reproduces, transcribes, and records around 9,000 early English Broadside ballads, as well as using metadata and text from the archive for topic modeling and other forms of analysis (such as documenting the number of ballads with tunes versus those without).23 Networking Archives uses data collected in Early Modern Letters Online and the Gale State Papers Online to visualize and analyze the connections between people in those databases.24 All of these projects offer tools for discovery that ultimately direct users back to practices of reading, researching, and writing that are neither flat, nor deterministic, nor particularly distant.

Attending a panel on digital scholarship at just about any academic conference will immediately make it clear that this field is far too big tent for “all CLS papers” to be “more or less organized the same way.” The field includes network analysis, topic modeling, and vast projects of digitization and preservation, as well as projects that do not fit into any of these categories; for example, “IDEA: Isabella d’Este Archive” describes itself as an “exercise in imagination, discovery, and critical engagement” and attempts to give users insights into the life and times of the powerful and influential Renaissance marchesa of Mantua by providing access to her texts, musical compositions, art collection, and an immersive 3D reconstruction of her palazzo.25 Such projects are truly interdisciplinary—often involving librarians, faculty, and members of the public—and they are far more representative of digital scholarship today than the straw man version of digital-humanities-as-word-counting often cited by detractors.

New databases are also being built in ways that will allow researchers to download, explore, manipulate, and analyze the data using some of the tools and methods I’ve employed, but without the more laborious data transformation and cleanup processes that were required for the ESTC MARC records. Although it does not incorporate tools of network analysis and visualization into the interface, the 4 million letters in the Medici Archive, spanning 1537–1743, could be analyzed with the tools I’ve outlined in this book, yielding new views of networks of power and patronage in early modern Tuscany and Europe. The same is true of the ambitious Universal Short Title Catalogue (USTC), which proposes to provide a “digital bibliography of early modern print culture” across Europe and which currently includes 740,000 volumes. Although it is still in development, network analysis of the USTC should eventually allow researchers to ask the same sorts of questions of the French, Dutch, Italian, or pan-European printing networks that I’ve been able to ask of anglophone printing using the data in the ESTC. Scholars who are interested in American print culture, or the networks of producers and musicians involved in twentieth-century popular music, or the names associated with Persian manuscripts, can access that data through WorldCat, which consolidates entries and metadata for 2 billion items held in libraries worldwide. Network analysis of such metadata would make it possible for researchers to ask the same sort of questions I’ve asked here: Did these networks change over time, and if so, how? Are links distributed evenly, and if not, who are the most connected figures? Who are the bridges, and why?

Moreover, the future of digital scholarship will be defined not by the infinite proliferation of sites organized to investigate such questions, but by their integration with one another. With the advent of the Semantic Web, in which data in web pages can be structured and tagged in ways that make it directly readable by computers, Linked Open Data allows researchers to conduct more powerful forms of analysis, working at scale. As Alan Liu pointed out as early as 2012,

[E]arly digital humanities projects . . . scaled up more or less accidentally, with the result (to the best of my knowledge) that every one of them ran up against the same virtual supersonic barrier. The barrier took the form of the following impasse: either a project retains established practices of scholarly quality-control (e.g., hierarchically-organized editing teams led by authorities, whose work is peer-reviewed in the traditional way by other authorities) . . . or a project uses some combination of algorithmic means and crowd sourcing to take the brakes off . . . whereupon quality-control no longer meets the standards of scholarship.26

Liu suggested that one way to address this issue was for digital humanities projects to engage more rigorously with theories of scale, and I’ve tried to model that in the opening chapter of this book, drawing on existing work in network analysis to discuss acceptable amounts of data loss. But another way to address scalability issues is Linked Data, which can allow for bigger, more robust analysis across datasets that have been hand-curated.

Carl Stahmer, who works with both the ESTC and the English Broadside Ballad Archive, sums up the possibilities this way:

Linked Data provides the technological framework for the creation of such a common interface without pulling or duplicating data from external sources and storing it in a local silo. A Linked Data architecture would take user queries from the common interface, run them against existing and reputable external sources, aggregate the results, including combining with local records, and return this aggregation to the user. It would also offer a mutually accessible endpoint or API to its own locally created records that others could query and likewise aggregate. It would, in other words, operate as part of a “share and share alike” community of bibliographers.27

A Linked Data architecture, in short, allows for the efficient distribution of scholarly labor, making it possible to scale up—through aggregation—without compromising local expertise and quality control. Several projects have begun to demonstrate the possibilities. Amsterdam’s Rijksmuseum, for example, released as Linked Open Data the digital images and full metadata for over 350,000 objects in its collection.28 This data can be explored, curated, and analyzed in connection with records from 3,500 other museums, galleries, libraries, and archives through the Europeana Foundation and various interfaces in the Europeana Labs, including a SPARQL endpoint that allows users to connect data to outside sources like VIAF, Shakeosphere, and Wikidata.29

The Advanced Research Consortium has similarly created the Big Data Visualization Application (BigDIVA) to allow users to browse, search, and visualize data from member nodes of the consortium, rather than viewing results of individual catalogues in list form. The BigDIVA visual interface is appealing, but as is the case with the Europeana Labs, the project’s most remarkable innovation is that it provides a single endpoint through which users can explore, and pivot between, resources as diverse as 18th Connect (which aggregates nearly a million texts and other digital objects from 60 federated websites on eighteenth-century literature and culture), Nines. org (which does the same thing for nineteenth-century literature), MESA (aggregating 112,000 medieval texts and digital objects from 32 sites), and ModNets (nearly 100,000 modernist texts and digital objects from 69 sites). A completely realized Linked Data architecture would go further by allowing even deeper interoperability. The locations of printshops mentioned on one website that specializes in print culture, for example, could be mapped onto another partner website that specializes in period-specific cartography; a project mapping the connections between artists could link to the museums holding their works; an epistolary network could be combined with a print network, a network of members in a scientific society, and a biographical database that allowed users to learn more about any of the figures in these overlapping communities.

My Shakeosphere project has experimented with these sorts of functions as part of the Linked Data for Libraries (LD4L) initiative.30 LD4L is a collaboration between Cornell, Harvard, the Library of Congress, Stanford, and the University of Iowa—organized to pilot the use of Linked Data to improve the discovery, use, and understanding of scholarly information. Shakeosphere served as a proof of concept tool. Shakeosphere’s data can be retrieved and manipulated with the SPARQL query language, which allowed us, for example, to match the locations of printshops and bookstalls in our database to places identified by the Map of Early Modern London. The record for the 1609 edition of Shakespeare’s Troilus and Cressida, As It Was Acted by the Kings Maiesties Servants at the Globe, which was sold “at the Spred Eagle In Paules Church-Yeard,” contains links to three MoEML locations—“The Globe,” “St. Paul’s Cathedral,” and “St. Paul’s Church Yard”; users who find the record in Shakeosphere can pivot through our site to view the locations in MoEML, or vice versa.

Using the same types of query, the developers at Wikidata matched 20,405 Shakeosphere IDs with their own and linked them to the other resources queried by the data clearinghouse. To take just one random example, the English antiquary Andrew Ducarel’s Shakeosphere ID is linked via Wikidata to an image on Wikipedia, a record in Early Modern Letters Online, an ID and set of records in the UK National Archives, a file and ID in the National Portrait Gallery, VIAF, the Cambridge Alumni Database, the National Library of Wales, and the Dictionary of Art Historians. To demonstrate the ways that such cross-linkage could aid scholarly inquiry, I also worked with James Lee and several students at Grinnell College to create a system that would allow users to toggle between his Global Renaissance project, which does topic modeling and mapping of full texts, and the Shakeosphere project, which identifies people involved in the production of those texts. We published the results of this collaboration in Cultural Analytics, where we described a “methodology we call ‘Linked Reading’ that embeds topic models and vector space models within historical book networks, allowing us to explore both the language associated with race in texts that mention Moors and the extensive networks of booksellers, printers, and publishers who produced these texts.”31 By pivoting between both datasets, we were able to make two key observations: first, that the language of race ranged across multiple topics including “trade, geography, and militarization,” and second that the network of printers, publishers, and booksellers involved in producing early editions of Othello was particularly invested in these topics.32 Both datasets informed each other, making the collective conclusions we were able to draw from them far more meaningful for understanding Othello’s place in the history of early modern race than if we would have been working in isolation. As digital projects have proliferated in recent years, so have the opportunities for this sort of data sharing and collaboration, which will be integral to the next phase of digital scholarship in the humanities.

Finally, contrary to arguments that the digital humanities offer shortcuts and easy answers, our paper on the Linked Reading project concluded that “At every stage, the process of Linked Reading involved interpretive work, as the researchers puzzled over results, debated them, read, recalibrated tools, puzzled over the relationship between our findings and the existing scholarly and primary literature, debated some more, and finally agreed on an argument that we think can make sense of it all.”33 This has also been consistent throughout the process of researching and writing this book, and it is what makes me optimistic about the future of humanities research using network analysis and other digital methods. If it were true that digital scholarship threatened “the displacement of a critical humanities praxis with one that announces its resistance to interpretation and to engaging with virtually every canon of existing interpretive thought,” then this book wouldn’t have been nearly as difficult, frustrating, or fun to write.34

Networking Print in Shakespeare’s England began with the idea that the labor of cultural production matters and that network analysis offers a promising tool to learn more about the figures and communities responsible for that labor. However capacious an author’s vision, in other words, it can only be realized through the efforts of a whole universe of other figures, including publishers, booksellers, and printers, who have their own motives and constraints. It has been more fascinating than I could have predicted to learn about printers like Nicholas Okes and booksellers like Michael Sparke, who reconfigured their respective businesses in ways that made them hubs in the network and serious dangers to the Laudian establishment. It has been a revelation to read the works of Mary Cary and to piece together the publishing practices that made her works, along with those of Eleanor Davies, unusually powerful bridges in the print network before both women slipped into obscurity. And I am hopeful that by turning attention to the weak ties that enabled Milton’s strong poetic voice I have offered a model that will be compelling for other canonical figures. But the most gratifying aspect of the project was the way it merged theory and practice, requiring collaboration, discussion, and debate at every stage.

As John Paget contemplated his own personal conclusion in the 1630s, the aging, exiled, leader of the English reformed church in Amsterdam paused to remark on the miracle of life, emblematized by the body’s system of veins and arteries: “a curious network & woven together . . . of subtle threads & thousands of them, more fine & small then hairs . . . & all these held and upheld in their severall functions by the finger of God, extending his quickening power unto every one of them.”35 This is yet another case of a “network” used as a metaphor, deriving from the visual similarity of a net to the venous system. But as it imagines quickening power pulsing through that net, sustaining the action of the whole, it also demonstrates the capacity of metaphor to advance our understanding. Through the collective action of thousands of threads, work gets done. And even as we’ve gained the capacity to quantify and analyze such networks, their curiosity and subtlety has not been diminished.


1. Thomas Taylor, The Progresse of Saints to Full Holiness, Described in Sundry Apostolical Aphorisms (London, 1630), 296. See also Edward Leigh, Annotations of Five Poetical Books of the Old Testament (London, 1657), 152, and William Gearing, A Bridle for the Tongue (London, 1663), 217.

2. Bruno Latour, Reassembling the Social: An Introduction to Actor-Network-Theory (Oxford: Oxford University Press, 2005), 132.

3. “Introducing the Mechanic Muse,” New York Times Sunday Book Review (26 June 2011): BR4.

4. Timothy Brennan, “The Digital-Humanities Bust,” Chronicle of Higher Education (15 October 2017).

5. Tom Eyers, Speculative Formalism: Literature, Theory, and the Critical Present (Evanston: Northwestern University Press, 2017), 34–35. See also Daniel Allington, Sarah Brouillette, and David Golumbia, “Neoliberal Tools (and Archives): A Political History of Digital Humanities,” Los Angeles Review of Books (1 May 2016).

6. Nan Z. Da, “The Computational Case Against Computational Literary Studies,” Critical Inquiry 45 (2019): 603.

7. See Alan Liu, “The State of Digital Humanities: A Report and a Critique,” Arts and Humanities in Higher Education (2011); and Patrick Svensson, “Humanities Computing as Digital Humanities,” DHQ 3.3 (2009).

8. Roberto Busa, “The Annals of Humanities Computing: The Index Thomisticus,” Computers and the Humanities 14 (1980) 83–90. Ann M. Blair, Too Much to Know: Managing Scholarly Information Before the Modern Age (New Haven: Yale University Press, 2010).

9. Marlene Manoff, “Human and Machine Entanglement in the Digital Archive: Academic Libraries and Socio-Technical Change,” Portal: Libraries and the Academy 15 (2015): 516–517; see also Marlene Manoff, “The Materiality of Digital Collections: Theoretical and Historical Perspectives,” Portal: Libraries and the Academy 6 (2006): 311–325.

10. See Joanna Picciotto, Labors of Innocence in Early Modern England (Cambridge, MA: Harvard University Press, 2010), 323–399.

11. Samuel Butler, “The Elephant in the Moon,” The Collected Works of Samuel Butler, vol. 3, ed. René Lamar (Cambridge: Cambridge University Press, 1928), 3:56–57.

12. Ibid., 358–358.

13. Eyers, Speculative Formalism, 34.

14. Da, “Computational Case,” 605.

15. See also Ted Underwood’s Distant Horizons (Chicago: Chicago University Press, 2019), which offers a model of statistical corpus analytics that departs in overt ways from the one Da describes, since “we have recently graduated from measuring variables to framing models of literary concepts. . . . Instead of starting with, say, the frequency of connective words, quantitative literary research now starts with social evidence about things that really interest readers of literature, like audience, genre, character, and gender” (xii).

16. Matthew Kirschenbaum, “Digital Humanities As/Is a Tactical Term,” in Debates in the Digital Humanities, ed. Matthew K. Gold (Minneapolis: University of Minnesota Press, 2012), 416.










26. Liu, “The State of the Digital Humanities,” 20.

27. Carl Stahmer, “The Universal Short Title Catalogue,” Spenser Review 45.1 (2015).

28. Chris Dijkshoorn et al., “The Rijksmuseum Collection as Linked Data,” Semantic Web Journal 9 (2018): 221–230.

29. To see how this works, visit the “Linked LODer” demonstration site Selecting a place name on the map loads available content sourced from the Place-name Database of Ireland, Europeana, DBpedia, the National Library of Ireland’s Longfield Maps collection, and the Irish Historic Towns Atlas project.

30. See For Shakeosphere’s role as a proof of concept in LD4L, see

31. James Jaehoon Lee, Blaine Greteman, Jason Lee, and David Eichmann, “Linked Reading: Digital Historicism and Early Modern Discourses of Race Around Shakespeare’s Othello,” Journal of Cultural Analytics (26 January 2018).

32. Ibid.

33. Ibid.

34. David Golumbia, “Death of a Discipline,” Differences: A Journal of Feminist Cultural Studies 25 (2014): 158–159.

35. John Paget, Meditations of Death (Dort, 1639), 238–239.