Digital and Poetic Aesthetics: Thoughts on Cross-Disciplinary Tools and Style

Although I was quite new to the digital humanities field before beginning my study of it this quarter, I have had some uncanny moments of recognition when encountering elements of form or style that intersect between 20^th century “traditional texts” and the tools and products of DH.  Of course there are many poetic texts that have been created during this century and the last that are both created and meant to be viewed electronically.  The work of Mark Amerika might serve as a nice introduction to “digital poetics;” he has worked to create and curate a range of texts that use the Internet as both medium and muse, from the comparatively old “Grammatron” that still speaks to reading and communication in the age of BuzzFeed, to the Museum of Glitch Aesthetics that augments the not-quite-real aspects of Google Earth views.  Of course, there are many other authors whose work is both digital and poetic: “Poetry for Exciteable [Mobile] Media” provides an app to make one’s “tweets” into interactive text, and “Redridinghood” uses an interactive form and feminist perspective to recreate a treasured tale for the mindset and aesthetics of the 21^st century.  I would also add webcomics that are both viewed and archived online, such as Dinosaur Comics and XKCD, to this list; both of these comics feature mouse-over text and archives that are unique to the web medium, while the latter has some seemingly infinite panels that the viewer is free to explore.

Clearly, there is a thriving group of artists and readers that seek out poetic texts in electronic form.  What I’d like to consider here, however, are the moments I’ve experienced where the form of texts that seem to either predate the concept of “digital poetics” nonetheless echoes the form of digital tools.  I had two strong moments of recognition of this sort: one came via my introduction to HTML, and the other came through the use of topic modeling tools.  The form of these tools and their products bear striking resemblance to the form of some 20^th century poetic texts (namely, those by Wallace Stevens and Amiri Baraka, as well as the Dadaists).  These cross-form encounters are striking in their own right, and give us more reason to believe that the literary world took Ezra Pound’s command to “make it new” to heart.

The strongest of these uncanny moments came as I made my way through basic HTML and CSS coding lessons via CodeCademy, a clear and enjoyable site that teaches beginner-level coding of various languages that balances the acquisition of new concepts and immediate, scaffolding practice.  Even in this beginners-level forum, and my own small foray into website-building, there was bound to be some frustration; but I found one frustrating moment to be surprisingly familiar, and I’m sure it is familiar to even higher-level programmers.

As I scanned my code looking for the misplaced or absent tag that had made my code illegible to the computer, I was reminded of my experience of reading Amiri Baraka’s poems, particularly “An Agony, As Now”.  About half a year ago, I encountered Baraka’s work for the first time while studying for my first qualifying exam.  Perhaps I had a strong reaction to this text because I first approached it near the end of a year of being in exam mode; or perhaps Baraka would be happy to claim my reaction as an essential part of his poem.  In any case, I found myself desperately marking up the text, trying to trace which parenthesis or parentheses had no match.  That might be the most absurd part of this endeavor: I knew that the poem contained a number of dangling parentheses, and that thus the entire idea of organization and perfectly nested structure had been disrupted.  But I still tried to parse out these parenthetical sets, thinking that the answer to the poem might lie within the most grammatically disrupted part of the poem.  I held out hope that the poem might be organized in its disorganization.

Clearly, I was in a strange state of mind, reverting to the type of close reading that has come to be considered faulty.  And perhaps it was only because I had reached a similar level of anxiety over “saying the right thing” as I practiced coding for the first time that I recalled this rather absurd close reading exercise that I had performed months ago.  Looking through my lines of code (which subsequent experience with HTML has made me realize were paltry in number), trying to find that one missing or extraneous punctuation mark, I was reminded of the day that I had taken close reading to what many might consider a myopic level.

Another moment of poetic recognition came at the hands of the topic-modeling tool discussed in my previous post.  While my classmates and I reviewed the products of our experiments with topic modeling, many of us noted the strangely poetic nature of our resulting semi-related fragments of text.  I was particularly reminded of Wallace Stevens’ poetry, which often juxtaposes strange images like glass jars and unruly forests, lamps and emperors (“Anecdote of the Jar” and “The Emperor of Ice-Cream,” respectively.)  While the topics reminded me in form of Wallace Stevens’ stark images, the topic modeling tool, as I’ve come to find in my reading, has a clear counterpart in the literary and artistic world through the concept of “found poetry” and its surrealist/Dadaist roots. As Amy Leggette explains in her post “From the Digi-Flânerie Field Guide: Approaching Data as Artifact,” cut-and-pasted found poetry was practiced well before the proliferation of computers and topic modeling, and paved the way for an expanded concept of what can be considered poetry.  It has also been used by Prof. David Hoover at NYU as a resource for “found poems” (for a brief reference to this practice, see his piece on “Hot-Air Textuality.”)

Suddenly, my insight that “then being merely without person find dreamed established some against” sounded like something Wallace Stevens would say was both validated and somewhat muted; but it does illustrate that our ideas of poetry and data have been intermingled for quite some time already. Those who work with topic modeling might argue that it is far less random of a practice than cutting up words and rearranging them sight unseen; but in the moment of searching for connection (if not meaning) between a set of words, I would argue that there are two very similar investigative and imaginative reading practices taking place.  It is a practice that has even become a fixture of pop culture, and quite literally a fixture on our refrigerators, as we see “Magnet Poetry” sets in home and office settings alike.  We are entertained by the challenge of creating a poem out of a fixed set of words and phrases, so it should come as no surprise that we would search for a poetic voice in the phrases that topic modeling creates.

Topic modeling, then, has a clear counterpart in form and philosophy within the literary realm.  But what should we make of a poem like Baraka’s, where there are formal similarities between the poem and the code, but we have no precedent for how to read these similarities; or, even worse, the precedent for this reading is the admonition not to get too hung up on details?  If someone reads Baraka’s poem as they would read code, are they doing it wrong?

In both of these scenarios, we can see similarities between the practice of reading digital and post/modern literary texts, and how easily these categories can become blurred; and I can only assume that these moments would be even stronger for programmers and social scientists who use these tools on a daily basis.  I hope that such moments of recognition will continue to happen, and that they might help to bridge the gap between fields of study, or create more “I get it!” moments as my classes grapple with odd poetic texts.  I’m currently serving as a teaching assistant in an Introduction to Literary Studies class that, in part, addresses the state of the humanities/state of the university debate; and this discourse has made me particularly attuned to the divide that is purported to exist between students in STEM fields and the rest of us (how we’ve made it this far without a snappy acronym is a debate for another day).  I have been hoping, and now truly believe, that my recognition of contemporary poetics in basic computer science is more than just a desire to find familiarity in a field outside my own.

I would not go so far as to say that Baraka intentionally draws on the form of computer code and the experience of writing within it; but I am sure that readers who are familiar with code would appreciate Baraka’s portrayal of inadequate or broken markers of containment as much as (if not more than) their counterparts in the humanities who have trained themselves to look at the forest rather than focusing on the trees (to borrow a turn of phrase in Hoover’s previously-linked article).  I imagine that they may even feel its effects on a more visceral level, sensing intuitively that “something here is not or cannot be contained,” rather than stopping at the idea that “something is being said here about containment.” Both categories of readers could eventually arrive at this knowledge, but programmers might reach it more quickly, or with a deeper affective understanding. Likewise, I don’t believe that the creators of topic modeling had Dadaist practices of creating poetry in mind; but those who work on or with the tool might have a good idea of the work that goes into finding the connection among a series of words.  And the next time that I work in an introductory literary course, I may bring the most strange and challenging poems that I can find, and see if I can create some more productively uncanny moments using the pooled resources of my interdisciplinary class.

The Interpretation of Topics: Looking for Freud’s Unconscious and the Role of the Reader through Topic Modeling

As I have acquainted myself with the tools of the digital humanities, many of my experiments through practicum assignments have led me to the uncanny feeling of having experienced the same thing before, either in my theoretical studies or in my reading of fiction.  Looking back over my quarter of experimenting with this discipline, I might ham-handedly characterize the field, or at least its relevance in my own work, as a method by which to de- and then re-familiarize myself with literature and literary studies.  I have had two major Unheimlich moments over the course of my exploration of the digital humanities; and given my dependence on the language and concepts of psychoanalysis to describe these experiences, I had better start with the Freudian one.

In my search for a text that seemed large enough to warrant study via tools often used on larger corpora, and one that had also aged enough to allow easy digital access and eliminate the threat of copyright infringement, I ultimately turned to The Interpretation of Dreams.  A hefty text that often gets boiled down to a few basic elements (often the most scandalous bits), Freud’s introductory work on psychoanalysis seemed like an ideal test case for many of the tools we would be using during the quarter.  If DH gives us the means to study literature that is either very familiar to us or vast in scope, what would it do with a text that could fit into both of these categories?

I began my study of the text with something that I thought would be easy enough to use and interpret: a word cloud of the text.  Scientific in its use of word frequency, and yet abstract or perhaps subjective in its choices of color and position, this tool seemed to have a clear correlation with dream-work and psychoanalytic interpretation.  My results, however, turned out to be a bit boring and messy, in the way that most accounts of one’s own dreams surely are to the average listener:

I created this word cloud before I learned about the magic of stop-lists, so unsurprisingly (though it did shock me in a sense), “dream” and “dreams” take up the majority of the cloud.  As I have said, however, it is probably more authentic to run an unedited version of Freud’s text through this and any other program; if we begin to pre-interpret in this way, one could argue that ultimately we will not be any closer to finding the subconscious of Freud’s text than when we began.

The results of this word cloud treatment did help me to visually analyze one hypothesis: parental relationships are not the prime focus of Freud’s work.  Both “father” and “sexual” are among the smaller words displayed here.  We have to keep in mind, however, that these are the most frequent words in Freud’s work, not all of the words that appear there; so even the relatively small appearance of these words does denote that they come up often in the text.  Not the most satisfying experiment, but somewhat fruitful nonetheless.

One tool that I did not expect to resonate with Freudian concepts, or perhaps with any concept I had encountered before, was topic modeling.  There are quite a few digital humanists that tackle the formidable challenge of explaining what might be an unexplainable tool, most notably and entertainingly Matthew Jockers’ “The LDA Buffet is Now Open.”  The idea of words being sorted into groups of ten based on how closely they are associated with each other within the original text makes basic sense.  The difficulty starts to arise when we see that the word banks change as we change the amount of banks that the computer can create; and things become odder still when we see that the word banks change even when we re-enter the same text and ask it to create the same number of word banks (topics).  Lisa M. Rhody’s piece on “Topic Modeling and Figurative Language” also does a nice job of unpacking the history, triumphs and perils of topic modeling literary texts; and she cites another piece on LDA that seems to come from a more scientific standpoint (being written by computer scientists), but has a title that reflects my sense of this program and its products the best- “Reading Tea Leaves: How Humans Interpret Topic Models.”  (I prefer to think of topic modeling results as ink blots, personally, but as I hope to prove, much of what we’re doing is learning about ourselves as readers rather than the topics or tea dregs, so my figurative language is as valid as anyone else’s.)

What are we to make of a program that sorts our data for us, and yet makes no attempt at telling us why it has chosen to create these categories, or what we should do with this information?  What, in the end, do these results tell us: do they help us to identify underlying trends within a text, or within ourselves as readers?  In the program’s original social science, non-fiction context, these problems may not have been so troubling or immediate; but now that we humanists have gotten our hands on this tool, it is time to investigate the process and products of this tool.  Thankfully, I did not tackle these questions alone; as my fellow UCSB “Intro to DH” students and I shared the results of our topic modeling of chosen texts, we all impulsively interpreted each other’s results.  How could we make sense of these categories, or could we?  After seeing my classmates’ reactions to the results, I felt more prepared to reflect on what the tool tells us about reading and ourselves, even if I’m not always certain what some topics say about the text at hand.  If you’d like to play along, here is the first batch of results that came back as I ran The Interpretation of Dreams through David Mimno’s In-Browser Topic Modeling tool:

1. dreams theory shall dream between material fact again further form
2. his man dream upon attention whose childhood fact could indeed
3. impression impressions same indifferent even two dreams gives different some
4. his had him friend dream little about too myself just
5. were his dreams seems such though same very since know
6. dreams psychic certain dream-life whole psychological activities problem problems series
7. dream may did yet see such content too before matter
8. her she had into out woman herself some then think
9. then being merely without person find dreamed established some against
10. dream interpretation very subject even however see given idea upon
11. waking state things place already images some three mind thoughts
12. dream patient still between another myself patients becomes while then
13. dream interpretation another here further made man meaning course found

There is much that can be said about these results (and has been said, in the context of my class discussion).  What I’d like to examine here, instead, is the process of creating and analyzing these results, in broader strokes.  Topic modeling seems, to me, to highlight two ways of thinking about what it means to read, or to interpret (to continue with my Freudian theme).  As the readers of these topics, we may see ourselves as being in the position of a psychoanalyst, or to be more accurate, the popular concept of a psychoanalyst: someone who is perceived to be all-knowing, to have the (right or wrong) answers (see important disclaimer below*).  If we seat ourselves in the analyst’s chair (rather than the analysand’s famous couch), we might see ourselves as either interpreting the free association products of the text, its author, or its historical period; or, more often than not, we may take the program itself as the object of study.  We may be trying to figure out why the topic modeling program “chose” to group these word sets together, to see what we can learn about the program itself by studying these results (symptoms, manifest dream content, and so on).  In our collective experiment, I had the feeling that if I looked at these word sets long or hard enough, I would learn more about the program that created them.  This impulse to solve the topic modeling puzzle, to figure out why the program “says” what it does, puts us as reader-analysts in the interesting position of interpreting the interpreter—an exotic-sounding paradox, until you realize that this is one of the ways that people have looked at the practice of reading.

We also might see ourselves as patients in another stereotypical vision of psychological encounters.  The practice of interpreting these topic models, depending on our approach, makes us akin to readers of a Rorschach test.  What do we make of these blots on the page (screen), grouped perhaps at random?  In this case, we are more in the position of the patient, trying to learn about ourselves via free association or acts of interpretation; and what seems like analysis of an external object is actually a test of our own mindset or assumptions.  Seeing these arguably arbitrary categories may also lead us to question what exactly we consider to be valid topics.  Some groups of words fall into topics that have clear counterparts within the humanities: there are classes on gender, politics, mental processes, and so forth.  But when categories like this get thrown in with others that seem like nonsense, what are we to make of the process by which we categorize things?  Do these categories illustrate topics and fields not yet discovered, or are they essentially white noise?

The ultimate question here for me is what, in the end, we are studying.  Are we studying the primary text (here The Interpretation of Dreams)?  Are we studying the program through which it has been run?  Are we studying ourselves and our own reaction to the program’s results, our impulse to make sense of something that seems to have no clear meaning?  These are also, of course, the same questions that we have to approach as we study literature.  Are we approaching a text with the idea of learning about the author or (to widen the field) the historical setting in which it was written?  Are we approaching the text to learn about ourselves, to see what is relevant in it to our own lives on a personal, social, or global level?  Or are we studying the text simply for the sake of hearing what they have to say, in a New Critical sense; trying to get “back to the text” through this impersonal tool?

To me, there is no one-to-one correlation between the topic modeling tool and our position as readers, which may be the most productive answer that I can give to conclude this exploration.  When the field of digital humanities presents us with (or perhaps is presented with) a tool like this, it may be that studying our approach to the tool and our reaction to it, rather than focusing on the data it creates, is the most productive approach we can muster.

*I must emphasize that this idea of the analyst as an individual who “makes sense of” the “nonsense” that his patients bring to a session of free association is a dated one, operating on the assumption that the analyst is at once an omniscient and impartial judge of the patient’s mind.

Toward a More Perfect (But Never Perfect) Timeline

Based on my first forays into the digital humanities, one of the most useful interventions that the field seems poised to make in my own research comes in the form of the visualization.  As a graduate student of English literature and die-hard close reader now grappling with concepts like distant reading for the first time, I must admit that I take some comfort in the baby steps that single-novel visualizations entail, affording me the opportunity to take a wider look at the material at hand while saving the creation of larger-scale graphs a la Franco Moretti for a later date. 

My first exposure to the research potential of these (relatively) micro-level visualizations came last year courtesy of Liz Shayne, UCSB’s Transcriptions research assistant and fellow Literature and the Mind enthusiast.  Her social network map of Faulkner’s Light in August illustrates who speaks to whom in the novel, and which connections would be left if Joe Christmas were removed from it.  Her project lets us simultaneously take a more distant and intimate/interactive look at the novel.

Liz’s project got me thinking about what sort of visualization would prove most useful to my study of memory and the 20^th century novel.  Luckily, there is a clear counterpart to mapping the social aspect of a novel: mapping the temporal.  Timelines are, of course, nothing new to lit studies: we are all grateful to timelines that have been made to illustrate the plot points of particularly difficult novels, and those made to map the historical and biographical events surrounding the production of a work (as included in every Oxford World Classic novel, Norton Anthology, etc.)

While these are invaluable visual resources, they do not seem to me to capture the same sense and depth of the novel as social network visualizations.  Something about their linearity seems limiting, as if one were to make a list of character interactions and then leave them unmapped.  Modernist novels (and the people who study them) in particular seem increasingly aware of their place within global and literary history.  Therefore, it becomes vital to think about the time in two senses: the time span of the basic plot, and the time covered through flashbacks and flash-forwards of both personal and historical scope.  It seems safe to say that every novel is tainted by shades of both the past and present, and an ideal mapping of time would reflect these shifts in perspective. 

Even before I began thinking about timelines in particular, I knew that Mrs. Dalloway would be the ideal test case for my forays into mapping.  One of Woolf’s many geometric and fragile images (as in “The leaden circles dissolved in the air”) might serve as a proto-theory on both the use value and pitfalls of character maps.  (The passage is too beautiful and apropos not to quote, and so I have included in full below*.)  While Woolf’s writing lends itself to visualization, thanks to her distinctive style that is often discussed in terms of visual art (often being called impressionist or post-impressionist), it also challenges the would-be visualization by demanding a rigorous and yet fluid system of mapping.

One could argue that by setting her novel within the scope of a single day, yet covering over half a lifetime of memories and events, Woolf makes the concept of creating a timeline just as obvious of an exercise as the creation of a social network map.  There are countless publications that try to grasp the temporal experience of Woolf’s novels through written analysis, and students of the DH field has already begun to have a go at her most famous text.  I came across a visualization of Mrs. Dalloway that makes impressive strides toward an ideal mapping system for the modern novel, courtesy of a group of students at Georgia Tech University. http://www.tiki-toki.com/timeline/entry/14025/Digital-Mapping-of-Mrs.-Dalloway/#vars!date=1923-06-15_10:00:00!

The students’ most valuable innovation here is their color-coding method of indicating whether each event takes place in the present, or is a recollection of the past.  They appear to have collectively mapped the majority of the novel’s plot points, and used their results to conclude that the narrative and the characters within it become more present-oriented as the novel reaches its close.  The tool used here, www.tiki-toki.com , is both flexible and remarkably user-friendly; but as far as my experimenting has found, it does not allow the timeline at the bottom to “zoom out” enough to make it immediately clear to the viewer that their hypothesis is correct.  I would venture to say, however, that they reached this conclusion more as a result of creating the map rather than viewing it, in much the same way that my own process of writing up data for my social network of the novel led me to some conclusions before the visualization was even finished.

Like all visualization tools, the Tiki-Toki program seems limited in at least one respect.  While each point on the timeline can be given any color, title, or image, it still remains a two-dimensional visualization, linear in the most basic sense.  While we can see that the narrative frequently shifts from the past to the present, we cannot see exactly how deep into the past (or at all into the future) the novel delves.  With this system, someone’s recollection of that morning looks as “past-y” to the viewer as another’s thought of a morning from the previous century. 

I find it particularly important to visually represent how far into the past the narrative goes in this novel because of the novel’s strong association with World War I.  While Septimus Smith is touted as the poster-boy of shell shock (commonly understood as the interruption of the present moment by frequent flashbacks), I’ve always found that Septimus is much more oriented toward the present moment than most other characters in the novel.  In fact, it is the other characters or the omniscient narrative voice that think of the past when they see Septimus, rather than him thinking of the war himself.  To me, the ideal temporal map would indicate how far back the novel’s view travels into the past or future, and at what moments it does so, to visually corroborate or challenge theories like this one.

There seem to be two ways to create this more nuanced visualization that I’ve cooked up in my head.  The first of these is to give each moment of the past a lighter or deeper shade of blue, to reflect how far back in time the narrative has traveled.  Perhaps the truly present moments could be rendered white or clear, the furthest moments in the past could be given, the deepest blue, and moments where the view shifts to the future (a category not represented on the current map) could be colored the deepest shade of green.  This might get us closer to the fluidity of Woolf’s stream-of-consciousness narrative style, and do something to alleviate the divisive nature of the time plot as it stands now.  The other option, one that unfortunately cannot be completed using the Tiki-Toki program alone, is to trade linear timeline for time plots, with an x-axis to represent the basic narrative plot (here, one day in June from mid-morning to beyond midnight), and an added y-axis to represent how far backward or forward in time the current perspective reaches at each moment.  Both of these methods do lend a third dimension to the plot- either by way of color, or by way of height.  I’m still unsure whether it would be most fitting to divide this x-axis in terms of measured time (hours in a day, years in a decade, etc.) or in terms of the length of the novel itself (chapters, sets of 50 pages, etc.); should we be mapping time in the strictest sense on this axis, or with an eye to how it is experienced by the novel, its readers and characters?

Of course, even the most thorough map, whether temporal or social, will not give the viewer a “complete” picture of the novel at hand.  Social networks are limited by the necessity of defining (and thus perhaps limiting) what kind of link an “edge” denotes; some of the most significant connections can take place between characters that only speak once, or never speak at all (as in a map of Mrs. Dalloway).  Even as I seek to improve the idea of a temporal map, I never intend to perfect it; and I do not think this is a bad thing. 

In my experience, visualizations may be most useful of all when we find them faulty.  When we create a visualization, or even look at one that we find to be incomplete (as I found some moments of Mrs. Dalloway that I found temporally important unmapped), both of these encounters help us to understand the text at hand in two equally meaningful ways: by showing us data that we do not expect (such as links between characters of which we weren’t aware), and by not showing us the data that we do expect (such as links between characters who we know to be intimately connected).  These graphs present us with new knowledge, and remind us of the knowledge we already have; they de-familiarize and then re-familiarize.  As I continue to tinker with the idea of a three-dimensional time plot, I intend to keep this in mind, to remind myself of the value of seeing a text in a new, and even an incomplete or distorted, way. 

 

*As promised, I will give the final words to Woolf.  After a productive lunch and letter-drafting session, the influential and politically conscious Lady Bruton thinks about her tenuous connection between herself and her guests: “And [Mr. Dalloway and Mr. Whitbread] went further and further from [Lady Bruton], being attached to her by a thin thread (since they had lunched with her) which would stretch and stretch, get thinner and thinner as they walked across London; as if one’s friends were attached to one’s body, after lunching with them, by a thin thread, which (as she dozed there) became hazy with the sound of bells, striking the hour or ringing to service, as a single spider’s thread is blotted with rain-drops, and, burdened, sags down.  So she slept.

And Richard Dalloway and Hugh Whitbread hesitated at the corner of Conduit Street at the very moment that Millicent Bruton, lying on the sofa, let the thread snap; snored.”  (Woolf 112)

Modern Literature, Mind, and Digital Humanities

The following blog posts will chart some of the major lines of thought I have followed this quarter in my “Introduction to Digital Humanities” course.  I’ve found that this field, its tools, and the questions that it poses all resonate with the material I study (trauma theory, psychoanalysis, cognitive science, and 20th century literature dealing with memory and perception), often in uncanny ways.  As something of an outsider to the field, I have often been startled by moments where the experience of delving into DH mirror and speak to my experiences of reading theory and literature that I would have assumed to be far afield.  Not only have I found tools to digitally/visually map out material of interest to me (as seen in my first blog post), but I have also had exciting moments of recognition when the act of using these tools mirrors the act of reading modern literature.  I hope to chart these moments of de-familiarization and  re-familiarization through my reflections here as the course draws to a close.