So What Are You Going to Do With That?

I decided to blog about Susan Basalla & Maggie Debelius’ article So What Are You Going to Do With That? because I can relate to this. So much. I get asked the exact same question by pretty much everyone when I first mention that I study German: “German literature? What for? What kind of job can you find with a degree in German? Do any of them pay well?” All of this was expressed perfectly in the article’s first paragraph. Fortunately, the authors also gave us a useful advice: “In order to see through the fog that sometimes surrounds us in grad school, you first have to abandon some myths about postacademic careers and replace them with questions that will help you think about your skills and your potential in a more positive and productive way.” So what are these myths?

  1. No one would hire me. I have no useful skills.
  2. People who work in the business world are stupid and boring.
  3. Jobs in the business world are stupid and boring.
  4. It’s too late to change careers.
  5. I’m too old.

I agree that it’s very important to think about all of them and reflect on what it is we’re doing. It’s not true that academics do not have any useful skills: many of them simply haven’t thought about the way their skills can be an asset in the job market. Studying and knowing foreign languages, for example, is something that can be very useful wherever you are working. I worked for a research company for several years and I went from interviewer to supervisor in no time because I spoke 3-4 languages well enough to take care of various translation tasks. Pretty much everywhere, but especially in a province that’s officially bilingual, knowing more than one or two languages is always a great asset to have. If I think about this course specifically, being good with computers and digital tools is also a great asset. Technology is constantly evolving and so are the workplaces, so academics should think about how to apply the knowledge they’ve acquired to different jobs and therefore stand out. As Basalla and Debelius said, academics simply need to recognise the many talents they’ve developed as teachers and researchers. It is also true that people in the business world aren’t stupid and boring. As a matter of fact, “the postacademic world offers a greater variety of backgrounds and more room for interaction than academia.” The same goes for the jobs out there. The authors talked about how everything was boring when we were fifteen, because we simply didn’t know enough to appreciate it. And to be honest, that kind of explains the feeling I first had with this course. I saw the course name, read the course description and went through the syllabus sent by Professor Sinclair, and I thought to myself “what have I gotten myself into?” At first glance, everything looked a little boring and way too theoretical to me. However, I was reassured as soon as I attended the first class. I noticed that most people didn’t have any DH background either and that us being from different disciplines would most likely make the discussions very interesting, and I was right. The more I learned about DH, the more interesting it became. Academics simply need to do the same with the jobs out there and stop assuming everything will be stupid and boring: they just need to give it a try. It is also not too late to change careers, and you’re never too old. As the authors of this article said, there’s no shame in changing tacks. “The key to successful career changing is learning the customs and vocabulary of the field you want to enter and then articulating your value. […] Gaining new experience, investigating other uses for your skills, and keeping a foot in another career are all wise pursuits at any age.”

As mentioned earlier, the most important thing is to replace these myths with questions that will help you think about your skills and your potential in a more positive and productive way. Think about the world outside academia: the experience you’ve had, your expectations, your friends’ jobs, your pressing concerns, etc. These are all important questions to think about. Are you happy in graduate school? Are your friends who aren’t academics happier with their work? Why did you come to grad school in the first place? It’s very important to constantly keep that in mind in order to figure out what you want for yourself. Basalla and Debelius wrote that “keeping one foot outside academia, in a part-time job or a computer class, may help you adjust more quickly.” Their advice is to simply find a job and experience something new for a while, which might be very intimidating at first, but ends up being very rewarding. A friend of mine had actually been looking for a job in academia for almost a year, when she decided to abandon the idea for a while and explore other options. She ended up applying for a job in Brazil and spent 6 months in Salvador da Bahia, teaching French to future immigrants to Quebec. Basalla and Debelius ended their article on a note, which perfectly describes my friend’s experience abroad: “You’ll be a stronger and more confident candidate for having proven yourself in the outside world; plus, you won’t feel pressured to take any academic job that’s offered.” Sometimes, you simply need the right timing to find the job you wanted, and by exploring other options, you might even find out that you’re happier somewhere else. The outside world isn’t as scary as it seems, so abandon the myths, ask yourself questions and dare to try something new.

The Spatial Turn

My presentation was on the Spatial Turn, so I’m not going to repeat everything I said in class in this blog post, but I wanted to mention a few things again. First of all, I found Jo Guldi’s article “What is the Spatial Turn” very interesting and helpful overall. She gave many examples of the spatial turn in various disciplines, such as literature, history or religion. I thought the concrete contributions of the university disciplines mentioned in the different sections were very clear, but I was a little confused when I first read the introduction. She said that “to turn implies retrospection, a process of stopping in the road and glancing backwards at the way by which one has come”, which made a lot of sense. What about the Spatial turn, though? Her article was well written overall, but I somehow didn’t feel like I could explain what it was after having read the introduction.

A little later, I decided to google “Spatial Turn” and found a bunch of other articles that weren’t that helpful, but I finally came across the volume The Spatial Turn: Interdisciplinary Perspectives edited by Barney Warf and Santa Arias, in which I found a definition that made much more sense to me: “Across the disciplines, the study of space has undergone a profound and sustained resurgence. Space, place, mapping, and geographical imaginations have become commonplace topics in a variety of analytical fields in part because globalization has accentuated the significance of location.” So basically, it’s a shift towards wanting to do more mapping and analysis of the physical environment in various disciplines. As mentioned in the essay’s introduction, “geography matters, not for the simplistic and overly used reason that everything happens in space, but because where things happen is critical to knowing how and why they happen.” Space does not determine social action, but it provides a context for it. I briefly talked about this during my presentation, but I wanted to mention what it was again in case you’re interested in reading more about this. I think this volume is wonderful and very diverse as well. It’s written by several people working together and it touches many different disciplines, which makes it very interesting. It is well written and easy to understand, even for someone with very little background in mapping and digital humanities overall, which is why I highly recommend it!

Demystifying Networks

I am still not 100% sure why but I often feel overwhelmed with this course. I keep up with the readings and I find them very interesting for the most part, but I somehow rarely feel confident enough to blog about them or speak up in class. I think it’s because most of it – not to say all of it – is very new to me and I often need to start with the very basics to feel like I understand a concept and/or how a tool works. I had to go over some of the readings a couple times to fully understand them, which was not the case with Scott Weingart’s Demystifying Networks. I thought his article was very informative and very clear overall. I liked that the discussion was geared “toward people with little-to-no background in networks or math”, which is exactly my case. The way he divided the article (Some Warnings, The Basics, etc.) made it very easy to follow, and he used great examples about everyday life to explain some of the concepts. For example, he said that networks can be used on any project, but should be used on far fewer, which is a little similar to using a hammer when a screwdriver would be much more useful. He also used humour, but not too much. The way the article was written kept me interested, and reading examples about things I could actually relate to made it much easier to understand the various concepts. I thought this article was a great introduction to networks.

In his blog post, Weingart explained that a network starts with “stuff”, but cannot work without “relationships”. As a matter of fact, network analysis is useless if the objects of your study aren’t interdependent. These connections not only matter: they are required. His explanations alone were very clear, but he also incorporated visual examples of networks with and without relationships. I also liked that he mentioned other words some people might use for the various terms he talked about, which helps when reading articles by different authors. The term “nodes”, for example, is something I came across a little while ago. I used to work for a research-based consulting firm, in the quality control department, and at some point I got asked to help the coding department with some of their tasks. I knew nothing about coding and had a 2-day crash course on the basics, which didn’t really help to be honest. I understood the tasks well enough to perform them, but Weingart’s article actually made the whole thing much clearer than my supervisors did at the time. He said that “Humanistic data are almost by definition uncertain, open to interpretation, flexible, and not easily definable. Node types are concrete; your object either is or is not a book.” That is exactly the problem I had at work. I had to classify various sentences under concrete nodes, often not knowing where some of these sentences would go, since they were open to interpretation and didn’t fit in any of the nodes. This is exactly what I was trying to explain to my supervisors at the time. That’s the very little background I have with nodes and networks. Some terms, on the other hand, were completely new to me. Weingart uses the word “edges”, which I had never heard in that context. By mentioning some of the synonyms, however, I was able to understand what it meant.

To better illustrate how edges work, he showed the following network and wrote: “Notice how, in this scheme, edges can only link two different types of nodes. That is, a person can be an author of a book, but a book cannot be an author of a book, nor can a person an author of a person. For a network to be truly bimodal, it must be of this form. Edges can go between types, but not among them.” This is the kind of concrete examples I was talking about earlier.


I also liked how he explained the difference between a directed edge and an undirected edge, as well as mentioning how they’re visually represented in networks. I already knew that an asymmetric relationship was represented using an arrow going from one node to another and that a symmetric relationship was represented by a simple line connecting two nodes, but I learned that both directed and undirected edges can also be weighed, which is represented by the thickness of the line, its colour or its length. It’s something you can easily notice when a different colour is used or when a line is much thicker than other ones, but I had never really thought about the length of the line actually meaning something. Then again, I’ve never really had to use network analysis for anything, so I might have simply not paid attention.

One thing that was repeated throughout the article and that I also mentioned earlier is that networks aren’t always helpful. Depending on what you’re working on, network analysis might not be the way to go and it’s important to keep that in mind. Weingart wrote that, first of all, many digital humanists often use networks incorrectly. They often try to link too many things at the same time, which ends up creating a dense network where almost everything is connected to almost everything else. This is obviously not very helpful. In his response to one of the readers’ comments, Weingart wrote: “Don’t sacrifice understandability for the desire to squeeze everything into one picture.” This is why network analysis works much better when the network is sparse, meaning when most nodes are only connected to a small percentage of other nodes. It is as a matter of fact very important to know which data to keep, and which data to cut. Wanting to put everything together is the perfect example of when network analysis can be used, but shouldn’t. However, as Weingart said, “given that humanistic data are often uncertain and biased to begin with, every arbitrary act of data-cutting has the potential to add further uncertainty and bias to a point where the network no longer provides meaningful results. The ability to cut away just enough data to make the network manageable, but not enough to lose information, is as much an art as it is a science.”

Topic Modeling, Figurative Texts, and Conclusion

This will likely be my final post for this seminar.  As such, I wanted to wrap up some commentary that bears relevance to the conclusion of my Plotto-“Bartleby” research, as well as ethical and methodological implications for digital textual analysis which we have discussed over the term.

The final stage of my project concerns the introduction of topic modeling via MALLET over a singular text that has been manually encoded in TEI-XML and then separated by those tagged divisions, each acting as a separate document for MALLET’s topic modeling algorithm.

Literary scholar Lisa Rhody notes in her blog posting, “Some Assembly Required” that seeking to gain meaning from using topic modeling over literary texts presents an obvious hazard: topic modeling approaches assume that ambiguity within a text is lessened inherently by the author’s desire to present their meaning clearly.  As students of literary studies, we know full well that the ambiguity of figurative language is prized in many textual genres – in Rhody’s case, it’s poetry.

Therefore, as Rhody goes on to suggest, if topic modeling is to be used to glean (new) meanings from texts, there is a need to contextualize, to closely read – at least at some points – the texts being modeled.  By using Plotto to determine plot clauses and conflict situations in “Bartleby,” I believe this is one way of doing so.  Another means of contextualization that I introduced into the analysis of the topics yielded from MALLET is that I used the topic weights within my Plotto-“Bartleby” model’s subdivisions themselves to decide on topic labels.  The entire time though, I was conscious of a growing wariness.  By prefiguring or handling the data beforehand in such a way, I was risking confirmation bias of the topic modeling results.  That is to say, if I consciously divide “Bartleby” in Plotto subdivisions (in this case, according to a more literal reading of the story), I could just be re-affirming my suspicions about the content of those subdivisions when I chose the labels for the topics generated by MALLET.  (Here are my notes on my topic modeling experimentation and choices.)

The hazard of this potentiality is somewhat mollified by a few different aspects to my methodology.  Over-interpreting topic modeling results is a much-discussed problem, but if I keep in mind that pre-modeling the text is a subjective approach, and that data from programmatic analysis should speak for itself before I speak for it, I can reduce the possibilities of confirmation bias where they exist.  For instance, from the increased stratification from my results (which I will link below for viewing) in Plotto subdivisions with large amounts of text, I can immediately tell that my Plotto clauses and conflicts can be further refined.  The topics generated in those subdivisions and my assumptions about them do still tend to line up.  But topics weighted in significant proportion within subdivisions where I did not expect those topics to occur could yield new insight into the content and operation of those subdivisions.  In that line, I argue that the most successful results of my experiment occur where those clause and conflict subdivisions appear to be less stratified.

The most obvious case is the denouement, where the lawyer comes to the prison for the last time, only to find Bartleby dead, and where he goes on to lament the loss and the enigma of Bartleby that remains.  Clause and conflict happen to coincide here – I have labelled this as clause ‘C’ (13. Comes finally to the blank wall of enigma) and combined conflict situations 1411 and 1355 (“A is haunted all his days by an act which he committed in an effort to help an employee, AX.  A, when the death of employee, AX, brings a philosophical investigation to an end, finds his skepticism at the blank wall of enigma.”).  This small section (and thus part of the reason for less notable topics within it) has been identified by MALLET as almost entirely containing a topic I have labelled “Bartleby, humanity, and death.”  The generated topic’s word list consists of, “bartleby, dead, died, letters, man, ah, clerk, report, rumor, curiosity, dining, yard, turnkey, sleeping, strange, life, eyes, humanity, death.”  Make of that what you will.

Here are my overall results in poster form: Plotting in Reverse: Plotto – The Master Book of All Plots and “Bartleby, the Scrivener”

As I note in the conclusion linked above, the larger project would next be to programmatize Plotto plot analysis, first by making the text more digitally accessible, and then attempting algorithm-oriented approaches to modeling in this way, or other related ways.  I believe that more traditional interventions of digital analysis such as this hold at least two benefits: they comfort institutional fears of quantification, and they showcase the possibilities of a more humanistic computing of texts.

The drama remains in the bits.  And the plot thickens.

My previous posts on Plotto for context:

How to Construct a Critical Narratological Algorithm

Plotto Structure: “Bartleby, the Scrivener: A Story of Wall-street”

Plotto, Bartleby, and Markup


Plotto, Bartleby, and Markup

I had been considering a markup scheme for this joint “Bartleby”-Plotto project for some time.  My initial attempt was mostly me figuring out how Plotto structuring works in the first place.  While I was learning that, I designed how I envisioned a manual XML schema might function for my reverse-work with Plotto.

Ultimately, instead of hand coding “Bartleby,” I used CATMA, a free, web-based markup and analysis tool coming out of an initiative at the University of Hamburg.  So, all of the XML was handled automatically for me.  I imported a cleaned up version of my primary text, created a tagset based on some of my thoughts, having gone through learning Plotto and applying/theorizing a structure for “Bartleby.”

Next up, I started to consider where exactly in the text each masterplot clause and conflict situation might sit.  See here for a description of my work on creating a Plotto plot model for “Bartleby,” which includes definitions of those terms, “masterplot” and “conflict situation.”

While I worked I kept a detailed log of the process of my markup decisions, as they would be key later on in my work.  Having exported an XML of the text and markup from CATMA, I separated each masterplot clause and conflict situation in order to treat them as separate texts for the next part of my project: topic modeling these text segments.  As I noted, the reasoning for my markup decisions will be key in this next step, to see if there is any correlation between the topics generated from these text segments and the Plotto descriptions of these plot segments.

Plotto Structure: “Bartleby, the Scrivener: A Story of Wall-street”

In one of my last posts I presented an overview of William Wallace Cook’s Plotto: The Master Book of All Plots.

Plotto is a book intended to aid writers by guiding them through a process that will allow them to construct a skeleton of a plot.  And at that point it’s up to them to proceed with the tough work of fiction writing.  After an intensive review of the book, and putting myself through a few Plotto tutorial sessions, I was able to work Plotto in reverse for the purposes of my final research project for this class.

How does one do that exactly?  Glad you asked.  During the process of a creating a Plotto plot structure for Herman Melville’s “Bartleby, the Scrivener: A Story of Wall-street”, I kept a detailed log.

But before I get to that, here’s a brief (re)introduction to how Plotto works, as well as some information regarding the notation you will see in the document linked below.

Cook allows his readers to construct plots via three key means, each of which can be used in conjunction, it just depends on your initial ideas for a story that you bring to Plotto.

First up are the “Masterplot clauses.”  These three clauses – A, B, and C – are what Cook considers an overarching structure to a work.  Through some practice, it’s quite arguable that there can be multiple of each, particularly of B.

The A clause “establishes the protagonist (in general terms).”
The B clause “originates and continues the action.”
The C clause “continues and resolves the action.” (Cook xiv)

You can pick your clauses from a pre-categorized list, and then there is a guide on how to proceed into the next sections of Plotto plot construction.

Second up are “Conflict situations.”  Conflicts are usually complex, and Cook allows for this complexity in a few ways.  Conflicts are categorized hierachically in this order from top-down: B clause number, conflict situation number, variation, and subclause.  Each conflict listed also has a set of “lead-up” conflicts and “carry-on” conflicts to choose from.  The conflict itself is listed as a body of text, potentially subdivided into subclauses and features symbols for character names.  (*, **, *** refer to subclauses and would be marked as such in a conflict situation subdivided in this way.)  Characters are also often ‘changed’ or ‘transposed’ for the purpose of a suggested conflict.

For example, conflict 19a can be broken down in this way:


Conflict Number: 19

Variant: (a)

Lead-ups: (117) (24a, b, c)

Conflict Text: A falls in love with B, and renounces wealth, which he was to inherit by marrying BX

Carry-ons: (223) (43-**)

Third and lastly are “Character Symbols” as notated above (A, B, BX, where is A is a “male protagonist”, B is a “female protagonist”, and BX is a “mysterious female person, or one of unusual character” (Cook 23)).  There are approximately 54 character symbols.  At the back of a book is an entire guide consisting of listings of character combinations for plots, lists of suggestions for interactions of those combinations, which then lead back to suggested conflict situation numbers.  (For the purposes of the exercise linked below you can consider it as initially guided by ‘A and AX’ [male protagonist and a man of mystery], which you will see me use briefly to look at ‘A tries to solve the mystery of a stranger AX’ which is the second subclause (**) of the conflict situation listed in this character symbol guide: 1393.)

And when finished pushing around the results of those three structuring mechanisms, it’s time to actually use your imagination in the literary sense and get writing.  In my case, I’m describing a story instead of telling one, and as you’ll read, some complications arise along the way in moving backwards through Plotto.  Like all elements of Plotto, Cook qualifies in his introduction that nothing should be taken as set in stone.  Characters, lead-ups, carry-ons, conflict text, conflict subclauses, all can be changed up as the imagination so suits it.  As much as I tried to follow Plotto’s suggestions as strictly as possible for the sake of my research, I definitely did a bit of changing and substituting of text in order to describe “Bartleby” using this model.

One further note, green text indicates a successful line of research, orange a point of questioning, and red, a dead end.

But without further ado, I give you my notes:  “Bartleby” as Structured Through “Plotto”

Works Cited

Cook, William Wallace. Plotto: The Master Book of All Plots. 2nd ed. New York, New York: Tin House, 2011.

geospatial stop animation

Here is my first attempt at stop animation. It is basically a sequence of about 370+ pictures. One photo should only vary SLIGHTLY from the photo that follows, otherwise, as you can see, the result is rather choppy. I wasn’t expecting to have a perfect result on the first go, and some of the choppiness was actually intentional, but I was surprised how hard it is to frame a photo exactly the same way 370+ times.
I know some of you read German… I’m not sure if the message (and humour) comes through if you don’t… Very fun to experiment with, even though it didn’t work well enough to be part of my presentation.


In the summer of 2010, I was in London, and I went to the British Library and fell upon this conference (and exhibition) of Maps.   Very interesting… You should see the video: Click here: Money–World Map-viewer

“N.A.S.A. – Money feat. David Byrne, Chuck D, Ras Congo, Seu Jorge & Z-Trip (Grant Phabao Remix) – Worldmapper Video Edition”
All maps taken from


(If I remember correctly, the presenters concluded with saying that Maps should be represented not by land mass but in terms of demonstrating e.g. world populations, poverty, etc. The video, representing the different types of maps,  should speak for themselves in the youtube video posted above. )

Geospatiality: Discipline and Pundit

After reading Anne Knowles’ articles on Looking at… and/or Seeing at… Gettysburg, as well as Jo Guldi’s “The Spatial Turn”, I have to say that GIS may be on the brink of becoming just another approach (a digital one at that) at observing the coordination of space and power relations from above. Far from disapproving or disregarding her work altogether, Michel Foucault’s panoptic view in Discipline and Punish seems a well suited metaphor of her study, reminding us all of  the relentless and pervasive inclination of disciplinary societies to observe and normalize. I don’t mean to sound disrespectful, but viewing The Battle of Gettysburg from above, and in this context, makes me think of the game of Risk from a high definition, 3-D, highly interactive, and overpowering visual effects perspective. I reiterate, I think Professor Knowles brings innovative ideas not only to DH, but also to other disciplines that historically have been involved in the writing process of a national identity. Is for this reason that I would have to side with Michel de Certeau’s on this one, for his postmodernist perspective best conveys the need to descend from the highest point and observe at ground level the true narrative being composed by the interaction between individuals and the space they occupy. Presently, there are plenty of narratives and episodes that take place on a daily basis, and to view them (forgive the neologism) geospatially will allows the opportunity to better understand the current state of the nation: corruption, hunger, poverty, violence, racism, crime, etc., (take you pick). I truly believe that we must look at our past in order to understand the present, and in the case of wars and battles, both past and present, I think that Professor Knowles has her heart in the right place, but not in the right time. One does not have to look far back to understand that in 2013, just as in 1863, conflicts have a way of representing the best opportunity to [re]write history in the name of national identity.