The “Hybrid” Dissertation

Last week I submitted my completed dissertation, Database Modernism: Literary Information Media, am happy to have received a Ph.D. in English Literature.  Conceiving of and sustaining an argument of that scale was a challenge, and after some rest and a return to teaching, one I think I could tackle again.  In addition to the written arguments on Joyce’s Ulysses, H.D.’s Palimpsest, and Pound’s XXX Cantos, that were printed and bound, I chose to do a digital project in place of a fourth written chapter.  In my mind, the addition of this digital component raises some interesting questions that I’ve discussed before in some conference presentations, but want to revisit now that the dissertation is complete.  This post could sound a little whiny, but my goal is only to express the unexpected challenges my dissertation presented and how they led to some important conclusions, both in terms of the project itself and how I view my academic labor and ideas of completeness.

My dissertation argues that the information overload of the early twentieth century prompted Joyce, H.D., Pound, and other modernist authors and artists to devise new methods of information organization and access that forecast media strategies that developed into the digital database.  The argument uses a comparative media approach to isolate relationships between elements of information in these print texts in order to determine where they resemble the operations of databases.  The common denominator between these print and digital platforms, my dissertation argues, is the associational logic by which individual elements of information are organized by a designer/author and then presented to a user/reader.

I have some technical experience with transitioning print texts to digital platforms from my time working on the Modernist Journals Project here at the University of Tulsa.  Because of the text encoding and OCR skills I picked up at the MJP, and because the argument of my dissertation rests on this fulcrum between print and digital media, I decided a digital project might be an interesting way to both expand the project’s scope and engage with some new questions and methodologies digital humanities has introduced into the study of literature.  I supplemented my skills with two trips and three classes at the Digital Humanities Summer Institute at the University of Victoria in British Columbia, experiences which were terrific in every possible way.  As it played out over the nearly two years I worked on the dissertation, I couldn’t be happier with the results, and I’m actually more confident in the findings now that it’s completed.  Still, I did learn some unexpected things about a dissertation that consists of traditional written arguments along with digital components.

The digital project is basically an expression of my written argument on Pound’s XXX Cantos.  In Chapter 3, I argue that the poem operates according to a dual structure of elements and relationships and that this structure resembles the node and edge structure of graph databases.  The elements, usually quotations, references, ventriloquisms, or other intertextual fragments,[1] Pound selected from the literary and historical archive and arranged on the pages of his poem according to an associational logic.  The project consists of a digital version of the 150 page XXX Cantos (which is kept secure for copyright reasons that I’ll discuss below) marked up with a custom set of tags drawn from the Text Encoding Initiative’s encoding language.   Using secondary sources (mostly Carroll F. Terrell’s A Companion to The Cantos of Ezra Pound) and a lot of my own subjective interpretation of the poem, I used TEI tags and separate bibliography, prosopography, and gazetteer files to define as many of the intertextual elements as possible.  Then I collected/created a much lighter and more agile second tagging structure to express relationships between the different elements in the poem.  I won’t recap the project in too much detail here because I did here.

What I Learned

The clearest lesson I drew from the digital project revolves around notions of completeness: writing arguments about literary texts have different timelines than completing digital projects.  My goal when I began the digital project was simply to express the written argument of Chapter 3 using digital tools.  The encoded text I ended up with, however, is not shareable, usable, or publishable in the way that a written argument is.  Most practically, I can’t disseminate the digital project in its current form because it’s a digital version of the XXX Cantos, a text that is under copyright with New Directions.  Thanks to a lucky connection, I got in touch with ND and they were more than pleasant, but politely and understandably were not comfortable with me putting my digital version of the text on the web.  In a search for alternatives, I cycled through several options,[2] but never arrived at anything that would work in the short or long term without requiring some additional training and a sizable amount of work for a non-expert like me.

As it concerns the dissertation as a discrete project signaling the end of my Ph.D., the presentation of the project is actually a moot point, however, because the larger lesson I learned from the project is that the timeline for written arguments and complete digital projects differ in ways that may have impacted how I conceived of the project at the outset.   I wrote the bulk of the dissertation while on a fellowship that was generously awarded to me by my institution.  This fellowship is designed to allow dissertators to focus as completely as possible on writing and research so she or he can finish in a reasonable amount of time.  As it was, I was right up against the original deadline having drafted and revised three chapters and more or less finished the encoding of the XXX Cantos.  Building an interface for presenting the text (which, because of copyright was conceptually and technically difficult anyway) would have required more time, expertise, and most concretely from my perspective, more funding.  The funding schedule that covered the work that I did complete provided me with relatively lavish resources, but I was not able to reach a point where the digital component was presentable because it would have required more training.  Applying for additional grants, fellowships, and other funding sources is not impossible, but the acceptance and dispersion would have operated on a timeline that exceeds my dissertation fellowship funding.

Writing my dissertation was not at all easy.  As this post is probably itself evidence of, I am not naturally good at expressing ideas in writing.  But I do basically know how to use the English language and have written arguments about literature.  Conceiving of and executing a project using the TEI was somewhat different because it took a lot of basic learning to just to develop a familiarity, an admission which is itself a little embarrassing because of the complexity and ease with which I see scholars using it for other projects.  Gaining familiarity with the TEI required a lot of time on my own and the DHSI classes, through which I acquired a functional understanding of how to encode a text and assemble a custom schema and tagging structure.  The rest of the project required the actual creation of the text and the encoding, all of which are complete,[3] but, again, not shareable in its current form.  In addition, as I became increasingly aware as the project developed, the encoded text isn’t all that interesting by itself.  It embodies the written argument in ways that satisfy me and my dissertation committee, but it is not a useful tool or reading aid for anyone else.  In other words, I “completed” a written argument about Pound’s XXX Cantos and I “completed” a corresponding encoding of the poem, but only the written argument is something I can share with others, and this presents some unforeseen challenges, particularly as I go on the job market.

As is, I can’t publish the project because of copyright restrictions, but also because without an interface the encoded text isn’t at all useful or illuminating anyway.  The structural and linking tags are geared toward machine reading and so don’t reveal much to a human reader of the text.  In order to make the encoding useful, it’s necessary to build an additional interfacing structure which somehow presents the metadata to a user or reader.[4]  What I’m left with is a dissertation chapter’s worth of work that isn’t useful to a public or shareable with a hiring committee.  The question of completeness also affects how I talk about the project.  I submitted a written methodology and rationale as an article for publication and the peer review feedback was largely positive, but the piece was ultimately rejected because there was no completed project to unveil.  I wrote this article while on the deadline for my dissertation, a lot of extra work that resulted in a better grasp of the digital project, but no publication on my CV and more deadline stress for the other chapters because of the lost time.

I’m also in an odd position because my written dissertation consists of an introduction and three chapters.  Because it’s customary to have a fourth chapter for a book proposal, it will be necessary for me to at least conceive of another written chapter before I approach a press.  This is not impossible, and like everyone who writes a dissertation, as I was writing, many texts presented themselves as having potential for inclusion.   This doesn’t diminish the research and writing that I could have completed during the time I was working on the digital project.

What I Really Learned

Complaining aside, I remain glad I did the project and as professional and technical training and intellectual exercise, it was all very valuable.  In addition, the questions the project raised and, in some cases, answered, will be interesting to my work going forward and for what I can bring to a literature classroom or English department.

The immediate question that presents itself to me as I compare notions of completeness between written and digital instantiations of the dissertation center on labor.  I can’t present the digital project because I don’t have the technical expertise or the time and funding to build an interface.  What I had been eliding from the production of my dissertation from the outset, and may never have considered without the challenges the digital project presented, is that I don’t have the technical expertise to print and present my dissertation either.   This is labor and expertise, that, like so much that takes place in a university and the humanities, is passed off to librarians and other information media experts.  The labor and skill required to create the object that mediates my work remain completely separate from what constitutes my sense of completion for the project.  To take a rhetorical shortcut: THAT’S PROBLEMATIC!   It’s that much more problematic considering my dissertation explicitly engages with the conceptual and material differences that arise when transitioning information between print and digital media.  The entire argument of my dissertation rests on the notion that when Joyce, H.D., and Pound began viewing print as media rather than a conduit for the transmission of their romantic interiority, it significantly changed their work.  This is a lesson I had almost failed to learn about my own work.  My understanding of the practice of writing arguments about literature had to this point completely ignored the importance of these independently fascinating and integral components of scholarly production.

I don’t want to zoom out too far from this fairly insignificant event, but for me the realization that resulted from the tension between my interaction with print and digital media stands as an identifiable instance of the value of digital humanities.  The use of digital tools and media for humanities scholarship forces us to evaluate those normalized zones of our profession.  As an early career scholar going on the  job market next year, my hope is that these situated understandings of how media objects get produced and consumed is something I could bring to a classroom of students who are themselves constantly transitioning between all manner of media.

[1] I say “usually” only because Pound’s original poetry is also mixed in with the intertextual elements and relates to the other materials in the same way.

[2] For example, providing partial access for online users that aligned with accepted fair use practices.  Or, determining which of the individual cantos appeared in periodicals before copyright took hold and publishing the encoded versions to the web.  This is tricky territory, however, because it’s not totally 100% obvious that those are out of copyright either.  I don’t have the resources to litigate and this project isn’t nearly important enough to do it anyway.  On top of all that, I have an extreme appreciation for the work that ND does and fetishize the physical copies of their modernist poets (such beautiful paperback books!).

[3] Even this notion of completness is tricky because there are certainly sections of the poem I could encode more precisely, transparently, and usefully.

[4] I’m not convinced a digital edition of the poem is all that useful (mostly because I dislike reading complex texts on computer screens).


This morning I was notified I had received a second DHSI tuition scholarship in June 2015.  This means I will be taking Text Encoding Fundamentals from June 8-12 and then XSLT: A Collaborative Approach from June 15-19.  First of all, I wanted to thank DHSI for offering these classes, but also the various organizations that contribute funds to make these scholarships possible.  Without the funding, it would likely be too expensive for someone on my budget to attend one, let alone two, classes.

I applied for these two classes because they will provide practical skills I will need to complete the TEI of Ezra Pound’s A Draft of XXX Cantos project I am working on for my dissertation.  As I’ve outlined in other posts, this project is part of a hybrid digital and written chapter of my dissertation in which I attempt to enact some of the textual dynamics I argue Pound employs in the XXX Cantos.  The TEI version of the XXX Cantos I have created thus far is a specific interpretation of how the text works.  The coding structure I’ve developed encodes specific types of information that I’ve gathered under the conceptual umbrella of intertextual references.  The chief goal is to use TEI to encode associational relationships between informational elements and use these tags to represent how the text generates procedurally.  In other words, what is on the page of the XXX Cantos is a result of informational processing that “calls” another associationally related piece of information to the page.

In order to do this, I have settled on tags that indicate person names, place names, ventriloquisms and allusions to other texts, quotations, and then tried to block off the different sections of Pound’s text that are associated with these intertextual sources.  In addition to the structural tags for line numbers, page breaks, Canto breaks, and so forth, these are the tags I settled on for now:

<persName key=”OVI1” reg=”Ovid”>Ovidio</persName>
<cit type=”quotation” text=”The Odyssey” author=”Homer” part=”12.200-205”>… </cit>

The other half of this argument is that Pound was very focused on textual dynamics as they materialize on the page as an interface.  This is drawing from my previous class at DHSI, Digital Humanities Databases in which I learned about the components of databases, the most essential of which is a relay between the user interface and the informational infrastructure.  This is to say that the relationship between my dissertation’s argument about information, interface, and modernist textual effects in the work of H. D., Joyce, and Pound is tightly raveled with the practical skills I have been and will be able to acquire at DHSI.

Developing the encoding structure and methodology for my TEI of the XXX Cantos has proven a bit more troublesome than I anticipated.  The tags I have listed above have been changed back and forth between other options multiple times as I’ve read the TEI P5 Guidelines and the Gentle Introduction.  These guides are essential tools and I would be nowhere without them, but I need to describe the goals of my project to someone and work with them to arrive at a way to achieve them.  I have had some difficulty determining which tags to use, which validation schema to insert into my header, and how to make the cleanest, most translatable version of the text that I can.  This is all toward the goal of using languages such as XSLT to process the TEI encoded text I’ve created.  In this sense then, I’ve chosen classes for 2015 that will help me solidify my technical foundation as well as teach me the skills and concepts I will need to complete the project.


A view of the ocean in Victoria

This year’s Digital Humanities Summer Institute in Victoria, British Columbia was my first.  From the time I applied for a scholarship to when I returned home on June 8, I couldn’t help but notice the well-established and inviting community surrounding the institute.  The information on the conference and the individual classes gave me a real sense of the organizers’ mission and their commitment to promoting the digital humanities defined broadly.  That said, and as the press dedicated to examining it attests, simply promoting big tent digital humanities is problematic, or at the very least, some people have some problems with it.

It’s my feeling that the challenges DHSI poses to its detractors is what make it so effective and valuable.  DHSI has a lot in common with the average humanities conference (as far as my limited experience can serve as a guide), but it also differs in important ways because attendees are taking classes, acquiring skills, returning to their home institutions with concepts, complete or in-process projects, and experiential knowledge about the realities of doing work that claims space under the dh canopy.  So, like at other humanities conferences, everyone gets exposed to new ideas and engages in dialog in ways that having everyone in the same place at the same time has for so long facilitated (though twitter seems to have gone some distance in trivializing the importance of presence at a conference, and this is a dh question).  To this familiar conference setting, DHSI adds the classes that are comprised of different mixtures of theory and heavy-duty computing (mine, Digitial Humanities Databases, was a great mix of the two).  Breaking out into the classes reminded me the seminar model used by conferences like the Modernist Studies Association which groups attendees into topic-driven roundtable-style sessions which are more enjoyable and rewarding than 90 minute panel sessions.  My experience at the 2012 MSA in Las Vegas was amazing; I was in a seminar on modernism and spectacle and I contributed a paper that interpreted Samuel Beckett’s Film and compared it with Bioshock the first person shooter.  

Still, this seminar structure is unlike the DHSI class because where I left MSA with new ideas and new email addresses and twitter handles, I left DHSI with a database I built in MySQL using command line.  There were moments in the course when my ears perked up as it was revealed to me that MySQL language, developed in the ’70’s (let’s historicize MySQL discourse, I’m comfortable with that), and is based on mathematician grade symbolic logic set theory.  It just so happens I’m wildly interested in set theory (though I have no practical knowledge of it) because of some recreational reading I’ve done on Kurt Godel and his incompleteness theorem.  I wanted to ask questions about whether or not one could make a database that took the incompleteness theorem into account, or if the creators of MySQL language found Godel’s thinking to be a problem for the potential coherence and completeness of databases.   I wanted to see the database, to read it, close read it as a matter of fact, and suss out the places where it may contradict itself or expose the ideologies it unknowingly promotes.

The instructors made no attempt to exclude these kinds of questions, and following some of the other classes’ twitter feeds showed that they were asking theoretical questions about digital topics.  (The feminist dh class in particular was performing what seemed like close readings of computer programs question whether or not they contained gender biases, which I found fascinating, but never followed  up on.)   Our focus, and we were pressed for time because there is a lot to cover in five six hour sessions, was squarely on the building of databases.  We did talk a lot about the types of queries databases, as a logic of information organizations, allow for, and we had some hilarious conversations about how databases and digital culture differ from document culture.  This distinction revealed the intertial influence print exerts on digital technology and our tendency to think of the internet and databases as documents.  This is all to say that the class wasn’t not theoretical, but the theory parts of the course were geared toward understanding the affordances a database offers (what can a database do that a document can’t and how might this help me as a humanist?) and how to ask programmers better questions when we need help as our projects develop.  I found this shift in thinking exhilarating.  We weren’t usurping issues of interpretation with graphs, maps, or trees (and we’d still have to interpret those anyway).  We were learning how things work and shedding light on how our own disciplines were similar and different.

My own research on Pound, Joyce, and H. D. requires a lot of close reading interpretation.  For modernist literary studies, these authors are often the close reading gauntlets everyone has to test themselves against.  On the other hand, in addition to being interpreted as poetry and novels always are, The Cantos, Ulysses, and Palimpsest themselves use daunting amounts of inter and intratextual quotation.  This is a dynamic I had originally tried to approach with archive theory, intertextuality, and even museum studies, but for various reasons I found myself wanting to express situations these suites of ideas didn’t address.  Eventually, with the help of my dissertation advisor, I came around to the idea that these texts may be forecasting the textual effects of more contemporary information presentation methods (e. g. Vannevar Bush’s 1945 description of the memex has a lot in common with The Cantos and The Wake).  Without getting into the specifics, I shifted from questions about what does this or that passage mean (though I am still very much on the hook for that type of work) to why is this passage here on the page before me?  How does this text work?

Thus my experience in the database class resembled my dissertation in that I had to set aside the “what does it mean?” questions in order to try and understand the “how does it work?” ones.  The interpretive questions do not disappear.  If anything, they are more persistent and grounded because this approach forces me to integrate the physical realities of each works’ composition and the ways it generates meaning into my interpretation of it.  I’m not pretending I have a coherent revelatory takeaway from this shift, but if this is what dh is, then the fears over machine reading solutionism or whatever alleged takeover of humanities departments computers are staging, are unrecognizable to me.  Overall, the experience of looking at information differently thanks to my class on databases really reminded me of the shift in my thinking that eventually became my dissertation idea.

Wake DiagramThe genesis of my dissertation idea resulted from some thinking with my committee about the overlap between database as a textual form and literary modernism.  The early exploratory readings included Lev Manovich’s book, The Language of New Media and the then newly released How We Think: Digital Media and Contemporary Technogenesis by N. Katherine Hayles.  Manovich has many interesting things to say about how databases operate, but of primary interest to my project as it relates to modernist literature is how he opposes the database to narrative.  Though it has evolved into an investigation of comparative textual interfaces as much as database at this point, the main argument of my dissertation will demonstrate how the associational logic of the database operates in the works of H. D., Pound, and Joyce.  After reading Manovich and then Hayles’ critique of his use of Saussure in theorizing the database, an interesting side project emerged that I’ll try and cover in this blog post.

Manovich defines the database, via the computer science discipline, simply as “a structured collection of data” (218).  However, he complicates this broad definition: “the data stored in a database is organized for fast search and retrieval by a computer and therefore, it is anything but a simple collection of items…The user’s experience of such computerized collections is, therefore, quite distinct from reading a narrative” (218).  The non-sequential characteristic of the database is less a technological substrate than a method of organization and this is essential since it allows for “fast search and retrieval” (218).

The search and retrieval function of the database is what interrupts the narrative logic because items do not relate to one another by advancing a linear story or creating an image, that is, the text doesn’t proceed according to a causal logic, but by the associative rules of the database (there is a great deal more to say about this difference). Manovich uses Ferdinand de Saussure’s linguistic theory from Course in General Linguistics as a model to describe how a database retrieves information through associational relations with a given query. Here is Manovich on his application of Sassure’s twin concepts of paradigm and syntagm:

Elements in the syntagmatic dimension are related in praesentia, while elements in the paradigmatic dimension are related in absentia.  For instance, in the case of a written sentence, the words that comprise it materially exist on a piece of paper, while the paradigmatic sets to which these words belong only exist in the writer’s and reader’s minds.

In other words, in Saussure’s formulation, the syntagmatic is the material that is actually present on the page and the paradigmatic is that information that is potentially (but not) present through associational relationships in the writers’ and readers’ minds.  Manovich feels that the database “reverses this relationship.  Database (the paradigm) is given material existence, while narrative (the syntagm) is dematerialized.  Paradigm is privileged, syntagm is downplayed.  Paradigm is real; syntagm is virtual.”

Hayles points out that “[t]his influential formulation, despite its popularity is seriously flawed…in semiotics, the alternative choices of the paradigm interact with the inscribed word precisely because they are absent from the page, although active in the reader’s imagination as a set of related terms” (180).  The presence/absence polarity Hayles insists on is does not apply to databases:

databases are not paradigmatic in this sense.  In a relational database configured as columns and rows, the data values of a row constitute the attributes of a given record, while the columns represent the kinds of attributes itemized for many different records.  In neither the rows nor the columns does a logic of substitution obtain; the terms are not synonyms or sets of alternative terms but different data values. (180)

As Hayles points out, a direct reversal of the textual dynamic of narrative doesn’t bring textual media to the database.  I am proposing that this debate between Manovich and Hayles is an extremely fruitful way to begin a discussion about how different media present information.  However, rather than seeing narrative and database as punctually opposed terms, I would shift the discussion the bundle of activities that take place when the reader/user interacts with books and screens as interfacial technologies.   Switching to an analysis of interface helps because the primary task of an interface is precisely the negotiation of the presence and absence of information that Saussure, Manovich, and Hayles so carefully describe.  Rather than opposites, the two interfaces afford and demand different activities from their creators and users.


fig. 1: associational constellation of ‘enseignement’

Saussure says that the syntagmatic terms on the page have paradigmatic terms swirling around them: “Any given term act a the centre of a constellation, from which connected terms radiate ad infintum.” Saussure is describing associations that radiate outward from a stable word on a page, and he provides a diagram of how this works for the word “enseignement” (fig. 1).  It is important to underscore that Saussure is describing how associations operate for a page of print, but it is also worth mentioning that he is expressing the impulse to represent the absent elements swirling around what is on the page.  The digital interface, which presents the reader/user with the information in a database or even from a book as in Google Books, take advantage of the relationship between the paradigmatic and syntagmatic by anticipating associational relations between textual elements.  In other words, the information is made present based on the types of relations Saussure insists are only virtual ones.  This anticipation of the reader/user’s associations is the textual work that digital interfaces do, they realize Saussure’s desire to identify related elements through textual relationships.  It is as if digital interfaces mobilize the relational network Saussure describes with the “constellation” of connected terms.  Instead of radiating outward, a digital interface brings the associated elements to the interface and, consequently, to the reader/user.

To further complicate this argument, my feeling is that several modernist authors use the print page interface to forecast the type of textual effects digital interfaces realize.  These authors see beyond the horizon of the book (or more precisely the stable word on the page) toward the type of operations and textual effects digital interfaces perform.  As an example, I’ll use Finnegans Wake because this won’t appear in my dissertation and I want to work through this idea.

I would argue that The Wake, through portmanteau words, symbolic clusters surrounding characters (HCE=>Finn Macool=>Fish=>Humpty Dumpty, etc.), and the repetition of events throughout history (the Judeo-Christian Fall, Humpty Dumpty’s fall, HCE’s sexual transgression, etc.) is making these associative relations present; the text, simultaneously operates on multiple scales to perform the search and retrieval function Manovich argues is fundamental to the associational logic of the database.  Through associational channels based on a word, character, or event that takes place in the text, “other” textual elements are “searched” for and then “retrieved” and made present to the textual interface.  To be clear, I’m not arguing Saussure’s syntagmatic/paradigmatic schema applies.  I am arguing that Joyce is experimenting with exactly the type of associational “constellations” the Saussure diagrammed in fig. 1.  These effects are difficult to identify because they appear on the stable print page (though even a cursory look at genetic analyses of The Wake resemble a search and retrieval dynamic).  The reader/user is not dealing with an absence/presence polarity because the assocational elements are present on the page (though of course, more associations are always hovering).  The question is how and why they appear there and what type of logic of information presentation they exhibit. (In an entirely different post, I would argue that interface analysis of any kind should privilege questions about how and why information is present over that information’s hermeneutic meaning.)  Here is a portmanteau word,  I snatched from The Wake:

fig. 2: meanderthalltale word cloud

fig. 2: meanderthalltale word cloud

“[M]eanderthalltale” (FW 19.25) is an agglutination of at least the five elements I’ve represented in fig. 2.  What I’m describing is fundamentally different from what Saussure diagrams in fig. 1 because the associations he describes are strictly linguistic relations within a solitary language (French).  The types of association Joyce uses work in a similar, but much less rigorous fashion.  They include aural polyglottal puns, thematic associations, word disfigurement, orthographic correspondences, etc.  However, my point is that the impulse is exactly the same: Joyce is trying to account for the associational elements swirling around a given word.  However, instead of maintaining the opposition between the presence of what is on the page and the associational elements, Joyce anticipates the associations and makes them present, retrieves them to the interface, to create a new type of writing which exhibits many of the characteristics of the database aesthetic Manovich describes.  The entirety of The Wake proceeds in this fashion.  In the same way that associational elements are made present in the portmanteau words, characters and events such as HCE and his sexual misconduct, function to bring other similar characters and situation to the text as if when Joyce thought of the character and the story he also thought of all the other similar characters and stories from throughout literary history.  Thus, based on these associational relations, these elements become part of The Wake.

In The Work of Art in the Age of Technological Reproducability, Walter Benjamin says “The history of every art form shows critical epochs in which a certain art form aspires to effects which could be fully obtained only with a changed technical standard, that is to say, in a new art form.”  The argument of my dissertation is that modernism is one of these “critical epochs,” but different from the previous ways we’ve discussed these artistic aspirations.  A few authors aspire to textual effects that directly influence our contemporary digital culture.  The “new art” form these authors aspire to through the book is an engineering of a new type of interface that could only be fully realized with the dynamism and creator/user interactivity digital technology provides.

Obviously, there is a great deal more to develop here with regard to The Wake and with modernism more generally.  I am also in the process of learning to discuss the interface as it relates to different textual media and so am reading and working with the significant number of books now appearing on these topics.

Works Cited

Benjamin, Walter. “The Work of Art in the Age of Technological Reproducibility.” Illuminations. Ed. Hannah Arendt. N. p., 1969. Print.

Hayles, N. Katherine. How We Think: Digital Media and Contemporary Technogenesis. University Of Chicago Press, 2012. Print.

Joyce, James. Finnegans Wake. New York: Penguin Books, 1999. Print.

Manovich, Lev. The Language of New Media. Cambridge (Mass.); London: The MIT Press, 2001. Print.

Saussure, Ferdinand de. Course in General Linguistics (Open Court Classics). N. p., 2008. Print.

The post below represents the first draft of a paper I wrote in the tradition of Raymond Williams’ Keywords.  This was for Dr. Ben Peters’ Digital Media Keywords class at the University of Tulsa.  Dr. Peters and the other members of the class did provide me with notes for revision which I have not taken into account for this version only because other projects are demanding my attention.  I have included a short postscript about possible ways to add to the piece.  My interest in interface stems from my dissertation project which examines how Modernist authors such a Pound, Joyce, and H. D. experimented with textual effects in ways that I feel forecast digital media.  Thus the thrust of this piece is to bind together the operations of print and digital interfacial effects.

Keyword: Interface

According to the Oxford English Dictionary Online, the first appearance of a term in the interface family was in 1837 by geologist James Dwight Dana who used “interfacial” to describe how surfaces align in crystal formation (“interface” a). The next use of interface listed in OED is not until 1882 (this time as a noun) referring to “a face of separation … between two contiguous portions of the same substance” (“interface” n). In the twentieth and twenty-first centuries, uses of interface and the terms related to it diversified and now refer to any type of intermediacy such as interpersonal liaisons, commerce, transition between systems, and most ubiquitously, any kind of screen-based human activity such as selecting a television program or using a computer. In addition to these early uses of the term, Peter Schaefer, in a chapter from The Long History of New Media, “Interface: History of a Concept 1868-1888,” looks to the late nineteenth-century where he finds that “the concept of the interface reveals places of overlap between communication history and the history of science” (164). It’s valuable to preserve the correlation between communication and science that Schaefer highlights, and connect those early zones of overlap to the diverse uses of interface in contemporary discourse. This inclusive approach will help illuminate the array of technical and conceptual issues that prompt thinkers in different fields to deploy the term to describe those places where systems meet and interact.

Schaefer’s account of the early uses of interface centers around James and William Thomson, two brothers interested in the physical sciences and communication technology. Both Thomson brothers used interface to describe the areas where different substances come into contact, but it was William who began theorizing the behavior of energy within systems. James Clerk Maxwell, with help from Thomson, proposed a thought experiment called Maxwell’s Demon, which, as Schaefer describes it, consists of “an enclosed space divided into two sections that are kept separate except for a small opening. Both sections are filled with gas molecules, and a tiny creature is stationed at the opening between the two chambers. This creature only will allow the faster, warmer molecules on one side and the slower, colder molecules on the other” (165).

Fig. 1

Fig. 1

The Demon’s role in this system resembles an interface because it operates the division between the two chambers which sorts the hot from the cold particles. Thomson’s other interest was communication technology: he was involved in increasing signal velocity and reliability in transatlantic telegraph networks. Though technology has changed dramatically since the telegraph networks Thomson worked to improve, similar issues of information delivery present themselves in the current digital age where the primary role of the interface is to provide a user with relevant information. While Maxwell’s Demon is just a thought experiment, it’s active–rather than passive–role in the sorting of particles is an apt metaphor for the the ways interfaces are programmed to function as a filter between the user and information systems. Thus, between its uses in communication technology and the physical sciences, the early story of interface as a term forecasts the ways it’s used well over a century later.

The uses of interface that refer to people using screen-based technologies almost invariably refer to the transfer of information from where it is stored to a user. Lev Manovich, in his 2001 book, The Language of New Media, uses the term “cultural interface” which “describe[s] a human-computer-culture interface–the ways in which computers present and allow us to interact with cultural data.” (69). Tellingly, Manovich’s primary concern is not “what, essentially, is an interface?” Instead, he focuses on the processes (“the ways”) a computer interface performs to present information to the user. Theorizing the interface as a bundle of activities or events is helpful because such an understanding can be applied to screen-based technologies, but also provides an opportunity for comparative analysis with other media, such as print, which also function to provide individuals with information. Rather than defining interface as a thing the parts of which could be disassembled and labeled, it is more important to view it, as the Thomson brothers did, as a concept that allows us to think through the complex circumstances that arise at the point where different entities meet. To confine my own discussion to media studies I will focus on analyzing how different interfacial processes negotiate the problems, issues, challenges, whatever we want to call them, that arise when a reader/user requests information.

Information Scale

Different interfaces manage information scale in different ways. Books, file cabinets, library catalogs, and search engines all provide information to users and, while they perform tasks unique to the demands placed on them, they are designed to present a reader/user with a consumable helpings of data. Compare a book to a file cabinet, for instance: the linear way a book is organized is different from the alphabetical organization of the folders in the cabinet (granted, the cabinet will almost certainly go A-Z or chronologically; the point is no one would read it A-Z). This is a significant difference and both are designed for different types of information storage. However, if a book did not progress successively or if a file cabinet wasn’t organized according to some structure, they would cease to be useful ways to store and present information. In their separate ways, each method solves a similar problem. Naturally, these organization methods are even more important for larger quantities of information. To increase the scale, think of the stacks in McFarlin Library (TU’s library) scrambled: if there were no structure to the arrangement of those books it would take an unfathomable amount of time to find anything. As commonplace an observation as it is, these organizational methods are designed to make finding and using information convenient and effective, but on extremely large scales these types of organizational structures are necessary to make information accessible in the most basic sense.

Fig. 2

Fig. 2

Alexander Galloway, in his book The Interface Effect, examines complex network visualizations such as this image of the internet like Fig. 2  from Wikipedia. Galloway’s argument is that this is a meaningless representation because each “map” or image of the internet appears, for all practical purposes, identical to the reader/user because she could never process that much information at once. To scale back down, I could say the same about the visual representation of a book. To the left is an image (Fig. 3) from a website that sells the complete text of a novel, here James Joyce’s Ulysses


Fig. 3

comprised of some 265,000 words, on a single poster sheet. While the poster is an interesting example of the fetishization of information as object, it is ridiculous to think that a book from this perspective is informatically meaningful to a human. And, as with network maps of the internet, all books would look more or less identical from this from this view.

The book, as an information delivery technology, solves this problem by its imposition of linear word, sentence, and page order. The reader moves from the beginning of the book to the end by reading each word. Even books arranged alphabetically like an encyclopedia or dictionary, or ones that experiment with non-linear organization (perhaps like Friedrich Nietzsche’s aphoristic style of philosophy, Walter Benjamin’s Arcades Project, or Mark Danielewski’s Only Revolutions) present their readers with a consumable datastream of linear words and sentences. For nonlinear methods of information organization such as library catalogs, databases, internet searches, the datastream is different, but the requirements remain the same. A reader/user requires an interface that presents large amounts of data in a way they can comfortably digest it. These issues of scale foreground an important aspect of the interface: because we’re so familiar with it, we don’t even notice that a book is an interfacial technique with its own methods for efficient and portable delivery of information. The same is true of the Encore or Classic Catalog search methods on the McFarlin Library’s homepage. We only notice the interface when it can’t deliver a book we’re looking for from out of the nearly million items housed in the building, to say nothing of the massive interdisciplinary databases and full text holdings remotely accessible.

Proximity v. Presence

The difference between proximity and presence is the same as when I am searching for my phone and it turns out to be in my pocket. During the search, I’m extremely close to my phone, but without the awareness of its nearness, it’s necessary to continue looking until the knowledge that it was in my pocket the entire time makes it present to me. Of course, the difference between proximity and presence is more complex when it comes to searching for and retrieving information. For instance, if I’m searching for a particular item in McFarlin and I stand at the center of the lower level stacks I’m relatively near in space to that item. On the other hand, if we again imagine McFarlin’s holdings scrambled, it doesn’t make much of a difference how close or far I am from the item, it will take a long time to find it. To get more perverse, imagine the stacks of McFarlin are in perfect order except for the one book I need. That book is not checked out, it hasn’t been stolen; it’s been misfiled. This book could be as close as the next shelf or filed in the bottom level with the government documents and, if so, may never be seen again. This is not to ignore the importance of proximity in access to information, particularly when it comes to non-digital archives. Living in Tulsa, it is much easier for me to poke around in the R. A. Lafferty papers in McFarlin Special Collections than it would be for someone interested in Lafferty living in Perth, Australia. Once digitized, however, the importance of that spatial disparity is dramatically reduced.

Like print archives, digitally stored information trivializes physical distance and increases the importance of addressability, the ability to locate items where they should be, but on a much more extreme scale. To use a specific example, if I am sitting on the intermediate level in the McFarlin stacks looking for a journal article in the Modern Language Association International Bibliography (MLAIB) and McFarlin has a print version on the bottom floor in the periodicals section, but the MLAIB has a linked full text .pdf version located on a server in Dublin, Ireland. It’s faster and easier to view the .pdf physically located in Dublin than it is to stand up, go downstairs, and locate the print version. The mind boggling speed and scope with which digital communication technologies deliver information creates the illusion that everything is immediately and universally available to the user.
Computers, or the internet, provide individuals with information from all over the globe and throughout time just like any archive or library. Standing in a physical archive puts one in close proximity to information, but without an interface it becomes difficult to access. Because a user can with comparable ease access information on a local memory device or on a server anywhere in the world, it seems as if all information is available. Still, the radical compression of data in digital storage comes with a tradeoff that seems to represent a break with print and other media. Media theorist Wolfgang Ernst takes great pains to demonstrate that the information stored in digital devices is so compressed that human perception is not equipped to interact with it. Without an interface programmed to locate, retrieve, and then translate that information into an understandable format it is inaccessible even if it’s in the user’s hand. This is the same polarity between proximity and presence, but because the digitized information’s ontology is so drastically different than the user’s, simply searching for and finding it as one might for a misfiled book would still never make it present. A byproduct of the increased capacity digital storage can boast of is that it puts the user at a further, and without the proper interface, perhaps insuperable, remove from the original. This underscores the significant difference between proximity and presence, and demonstrates how different interfacial tools are designed in response to similar challenges of information presentation.


In order to circle back around to telegraph networks and Maxwell’s Demon, the overlap between science and communication technology where Schaefer started us out, it seems worthwhile to mention that many of the differences between the print and digital information management methods I described above are differences in technology rather than new solutions to previously unsolved and archaic problems. Rather than thinking of digital memory and transmission technologies as solutions to analog problems, it is more appropriate to view these technological innovations on a continuum of ever increasing efficiency and portability. While a digital storage device hardly resembles the holdings of a library, it essentially performs the same task more efficiently. Interface design, while it has certainly benefitted from these digital innovations, is asked to solve the same problem that Thomson and Maxwell were laboring over the Maxwell’s Demon: how could we program a division between systems that would automatically perform the desired sorting tasks. Google’s high-powered search algorithms and our humbler Encore McFarlin search bar perform this filtering by responding to specific queries and presenting relevant information to the user. At the same time, it’s important to remember the active role the demon and the interface play in this dynamic. Rather than the passive conduits each interface operates according to rules that do good in that they give us what we want, but they also have a vested interest, whether political, commercial, or socially normative, in presenting certain items and omitting others.

Another draft of this paper would include two more sections to deal with issues that are implicitly or explicitly raised here: the first section would discuss digital ontology and how that may affect how we view the difference between print and digital storage capacities. Looking at the technology this way raises further questions about whether or not magnetic hard drives holding electrical charges and then translate that information into text or images is fundamentally different than pre-digital tech (Jonathan Sterne’s discussion of the difference between the digital and analog would prove very useful). This also seems to be an interesting way to reconnect with Schaefer’s account of the early overlaps between the physical sciences and communication technology. The second section would try and suss out the differences between organizational techniques like the LOC or Dewey Decimal system for libraries and the interfaces such as card catalogs and Encore-type search bars. It’s not clear to me that these organizational methods are separate from the interface for a variety of reasons, but I am not comfortable pronouncing they are identical. At any rate, the above represents an incomplete version of what should eventually become a longer paper.

Works Cited

Galloway, Alexander R. The Interface Effect. Polity, 2012. Print.
“interface, n.”. OED Online. December 2013. Oxford University Press. 1 March 2014
“interfacial, adj.”. OED Online. December 2013. Oxford University Press. 1 March 2014
Schaefer, Peter. The Long History of New Media. N. p. Web. 19 Mar. 2014.

Image Credits
Fig. 1:’s_demon.svg
Fig. 2:
Fig. 3:


I am in the very beginning stages of a new digital project in which I will put the skills I have learned as a member of the Modernist Journals Project staff to use on A Draft of XXX Cantos.  I will use ABBYY 11 to OCR A Draft of XXX Cantos (thanks to TU Copy for those scans).  The Finereader files ABBYY creates will allow me to create viewable and searchable PDFs of the text for the future.  For now, I am most interested in the text files ABBYY creates.  I will be pulling the full text of these early cantos into Oxygen XML Editor and then using TEI markup language to create new ways of searching, arranging, and most importantly, sharing the text.  The project will end up being part of a dual format chapter in my dissertation tentatively entitled Database Modernism.  My dissertation examines the role of what I’m calling database aesthetics as well as textual and formatting experimentation in modernist literature.  Using TEI to attach metadata to Pound’s Cantos is an attempt to enact the text’s interfacing ambitions, close the gap between print and digital ontologies, and describe how these innovations of literary modernism on both the formal and content levels forecast late twentieth and early twenty-first century media culture.


UW-Madison | Digital Humanities | Research Network


News & Information for Students Admitted to the UW-Madison Information School MA Programs


researching, writing, triathloning


Manuscripts, materiality, method: a research network

Text and Medium: Intro to Digital Humanities

ENGL 2393, The University of Tulsa

Posts on the Penman

An Iowa Joycean's Musings on the Hilarious, Irreverent Irish Genius

an online business card, entry point, and space for musings

The Digital HuMannist

English PhD student wields laptop, wants to talk about it.

Ragman's Circles

Talking about what the matter was

Modernist Fragmentation and After

International Postgraduate Conference, Princeton University, September 29-30 2016

This and That Continued

Jenny Diski - Writing and stuff.

Parrots Ate Them All

A blog about Stevie Smith and the art of the aphorism.

New Readia

WordPress site of Nicholas M Kelly, PhD, Lecturer in Rhetoric at the University of Iowa, literature, media, and digital culture scholar

Stewart Varner

scholarship, libraries, technology