After the Montreal workshop we focussed on the classic close reading scenario with an emphasis on interpretation because it can be considered the one to which 3DH postulates apply the most, so we wanted to make sure we cover that first: Exploration of free annotations to sharpen the literary research question.

Developing an early prototype as a proof-of-concept for this scenario first would make it easier to transfer interface principles to the other scenarios, so we reasoned. Over the course of the last months we chose a text we deemed appropriate for the prototype and for the intended audience and populated the scenario with real data. The text we wanted to annotate needed to fulfill some basic requirements: It should be well-known, so people can relate to it. It should be complex enough, so different paths of interpretation can be pursued and it should be short enough, so people can actually read the text without spending too much time, if they want to. It should also be long enough, so visualization as a method of getting an overview really makes sense. We picked the short story In der Strafkolonie von Franz Kafka.

For this short story we created over 600 annotations in 19 different interpretation categories in Catma. In the next step we exported our Catma annotations as JSON and built a web-based demonstrator with Javascript and D3 that shows the most important interactions of the concept.

The main principles of our concept are the tripartition of the interface and the representation of annotations as glyphs. So, while we clung to the idea of glyphs (mentioned in the last article), we have abandoned the idea of a strict spatial separation between the two activity complexes research and argument. We came to the conclusion that scholarly activity is better represented by three adjustable spaces text, canvas and argument.

Here text is simply the part of the interface, where our research text can be read and annotated. For each annotation a glyph is created on the canvas in the middle of the interface. We can sort these glyphs, structure them according to different criteria and draw connections between individual or groups of glyphs. Scholars can save multiple canvasses each of them highlighting a particular aspect of the text. In the argument space on the right side of the interface these canvasses can be combined and arranged to form an argument.

Since this year’s topic of the DHd2018 conference was critical perspectives on digital humanities, our contribution put an emphasis on our design process and the accompanying design-based critical perspective we have applied in the process. We talked about how we incorporated the four methods scenarios, wireframes, prototyping and design reviews into our process and how these helped us to gain new insights and arrive at the current state of design.

Here’s a link to the early prototype that allows you to explore the interaction between annotations and glyphs:

Prototype

You can find our slides here:

Slides

These two videos show the interplay between the three parts of the interface:

Over the span of two and a half days we had an intense and productive workshop that had the goal of refining and reifying the three concepts we had developed so far over the course of the last weeks. Springboards for this process were on the one hand our four conceptual 3DH postulates: 2-way-screen, parallax, qualitative and discursive, on the other hand reflections about supporting the process of interpretation in digital tools. We specifically discussed the relevance of the article “Thinking about interpretation: Pliny and scholarship in the humanities” by John Bradley.

What is intriguing in the software “Pliny” described by Bradley, is, that scholars are very not bound in the way they organize their notes and annotations, there is no need to assign distinct categories or relations to them. Instead, these can be organized on a plane and emerging structures becoming apparent can be inscribed by encapsulating them in boxes, when the interpretation progresses.

This appears to be a way of modelling interpretative data that takes into consideration methods scholars have been using in the analog world, however, also exceeds that and opens up new possibilities enabled by the digital (in terms of interaction with and visualization of data), an approach that seems very much related to the goals of the 3DH project as well.

In our design process so far we have based our concepts on real-world scenarios fed by experiences of literature scholars in research projects and arrived at similar conclusions as Bradley: It seems counterintuitive for scholars to force them to apply structure to their annotations when they start with their process. Relations between and qualitative statements about annotations often can only be made when the process has progressed.

When we discussed the wireframes in the workshop we realized that we can differentiate two different environments or spaces of literary scholarly work: Johanna called this research and argument space. While we define typical descriptive acts of the scholarly process like annotating, collecting and commenting as research activities, we consider tasks like grouping, ordering and organizing as interpretative or at later stages  argumentative activities. Usually scholars switch between activities of either of the modes perpetually.

Interplay between research environment and argument environment (by Johanna Drucker)

We understood that this circumstance has to be supported by the interface much more deliberately. Thus, for the next steps in the design process we will focus on the representation of and interaction between these spaces in the interface. What would an interface look like that supports continuous switching between these mentioned activities?

In the discussion we came up with the concept of a semantic plane that might allow us to bring these two spaces together. While we would produce annotations in the research phase that would be represented as glyphs on the plane, in the argument phase we would position and manipulate these glyphs to assign meaning to them and create  arguments that we later can publish.

Merken

Merken

Merken

Merken

Merken

We are halfway through our lecture period by now and since our first co-creation workshop in Potsdam at the end of April a lot has happened.

The concept sketches that were created by our five student groups during the first workshop were elaborated on in preparation for the next exchange between Hamburg and Potsdam.

On May 10th, Marian Dörk and I, together with the Potsdam interface design students, visited Chris Meister, Rabea Kleymann and the other members of the team to join Chris’ seminar and the accompanying exercise.

In the seminar Chris gave an introduction to the collaborative annotation tool Catma and explained to the Potsdam students, how you would use the tool with a certain literary question in mind. This introduction was meant to serve as a primer to Catma on the one hand, but also as an insight into the literary scholar’s process. Since the Potsdam students are supposed to base their visualizations on real data, i.e. narratological annotations produced in Catma, we deemed it necessary to make them comfortable with the process and the tools. The annotations they will eventually use, will be produced and made available to them by the Hamburg students via Catma.

In the exercise the interface design students presented their refined concepts in front of the Hamburg students and Chris Meister’s team. The concepts were quite diverse, in terms of narratological questions they were supposed to address, as well as media, technology and design. The images depicted below give an impression.

Sketches from short presentation in Hamburg

The predominant issues addressed with the concepts were, among others: narrative levels, advanced text search, narrative polarities and relations between objects, characters and parts of the text. After each presentation the literary scholars gave feedback on the projects. In the following three weeks the students had time to continue working on their concepts, before the two student groups from Hamburg and Potsdam got together again.

On May 31st our second co-creation workshop took place at the University of Applied Sciences Potsdam. The goal of this workshop was to sharpen the students’ concepts with respect to their ability to help answering narratological questions. In the first part of the workshop the student groups presented their current status followed each by a short discussion. In the second part all of the interface design student groups in Potsdam were assigned one or more students or researchers from Hamburg and worked on reifying their concepts. This session was also meant as a chance for the interface design students to ask questions regarding narratological analysis that they had collected over the past weeks.

In the weeks before there had been some rearrangements within the groups and some conceptual reorientations. At the moment there are six groups and their concepts are summarized in the following.

Storylines

Idea of this concept is to mark frame narratives and embedded narratives in the text and visualize their nesting on different levels of detail. These representations will be combined with contentual information, so viewers are informed about the interplay between structure and content.

Narrative Levels

Narrative levels concept

In this circular visualization ring segments represent narrative levels of different order. In the center different relations between levels are visualized, like for example, character networks or the duration and frequency of certain events.

Manual Topic Modeling

Manual topic modeling tool concept

The students developed a tool that let’s you search for the occurrence of words and let’s you manually define topics by putting words together. The distribution of these words can then be visualized. The group is currently looking into other text analysis features that might be helpful for narratological analysis.

Influences

The group wants to visualize influences on the protagonist of a novel. From a narratological perspective it could be interesting to analyze the development of the characterization of the protagonist over the course of the novel. The next step for the group will be to collect all text passages that characterize the protagonist and develop a suitable metric that can be visualized to represent the characterization. In a further step this could be related to the context surrounding a characterizing passage.

Nodes

Nodes concept

In this concept the attraction and repulsion between different entities (characters, objects, text fragments, for example) in the text is visualized in a VR environment (Google Cardboard). Attributes that influence the degree of attraction or repulsion are, for example, frequency or proximity in the text. Different entities can be pinned to get an impression, how they relate to other entities. This way it becomes possible to assume the perspective of a particular entity.

Narratological Cards

Narratological cards concept

In this concept a set of visual modules allows users to analyze a narrative in a physical way by putting together cards that represent different narratological features, like narrators or narrative levels, their position in the narrative and their interconnections.

Our hope is that this close collaboration between the two student groups over the whole course of the semester continues in such a fruitful way and will eventually lead to visualization tools that are truly user-centered and oriented towards the needs of narratologists.

Merken

Merken

Merken

Merken

Merken

Merken

Even when you deal with relative frequency you can get what look like wild variations. These often are not indicative of something in the time, but indicate a small sample size. The diachronic datasets often have far fewer books per year in the early centuries than later so the results of searches can vary. One book with the search pattern can appear like a dramatic bump in early years.

There are also problems with claims made about data. There is a “real world” from which we then capture (capta) information. That information is not given but captured. It is then manipulated to produce more and more surrogates. The surrogates are then used to produce visualizations where you pick what you want users to see and how. All of these are acts of interpretation.

What we have are problems with tools and problems of data. We can see this in how women are represented datamining, which is what this talk is about. She organized her talk around the steps that get us from the world to a visualization. Her central example was Matt Jocker’s work in Macroanalysis on gender that seemed to suggest we can use text mining to differentiate between women and men writing.

World 2 Capta

She started with the problem of what data we have of women’s writing. The data is not given by the “real” world. It is gathered and people gathering often have biased accounting systems. Decisions made about what is literature or what is high literature affect the mining downstream.

We need to be able to ask “How is data structured and does it have problems?”

Women are absent in the archive – they are getting erased. Laura thinks these erasures sustain the illusion.

Capta 2 Data or Data Munging

She then talked about the munging of data – how it is cleaned up and enriched. She talked about how Matt Jockers has presented differences in data munging.

The Algorithms

Then she talked about the algorithms, many of which have problems. Moritz Hardt arranged a conference on How Big Data is Unfair. Hardt showed how the algorithms can be biased.

Sara Hajian is another person who has talked about algorithm unfairness. She has shown how it shows prestigious job ads to men. Preferential culture is unfair. Why Big Data Needs Thick Data is a paper that argues that we need both.

Laura insisted that the solution is not to give up on big data, but that we need to keep working on big data to make it fair and not give it up.

Data Manipulation to Visualization

Laura then shifted to problems with how data is manipulated and visualized to make arguments. She mentioned Jan Rybicki’s article Vive la différence that shows how ideas about writing like a man and like a woman don’t work. Even Matt Jockers concludes that gender doesn’t explain much. Coherence, author, genre, decade do a much better job. That said, Matt concluded that gender was a strong signal.

Visualizations then pick up on simplifications.

Lucy Suchman looks at systems thinking. Systems are a problem, but they are important as networks of relations. The articulation of relations in a system is perfomative, not a given. Gender characteristics can be exaggerated – that can be the production of gender. There are various reasons why people choose to perform gender and their sex may not matter.

There is also an act of gender in analyzing the data. “What I do is tame ambiguity.”

Calculative exactitude is not the same as precision. Computers don’t make binary oppositions; people do. (See Ted Underwood, The Real Problem with Distant Reading.) Machine learning algorithms are good at teasing out loose family resemblances, not clear cut differences and one of the problems with gender is that it isn’t binary. Feminists distinguished between sex vs. gender. We now have transgender, cisgender … and exaggerated gender.

Now that we look for writing scales we can look for a lot more than a binary.

Is complexity just one more politically correct thing we want to do? Mandell is working with Piper to see if they can use the texts themselves to generate genders.

It is also true that sometimes we don’t want complexity. Sometimes we want simple forceful graphics.

Special Problems posed by Visualizing Literary Objects

Laura’s last move was to  then looked at gender in literary texts and discuss the problem of mining gender in literary texts with characters. To that end she invoked Blakey Vermeule, Why Do We Care About Literary Characters? about Miss Bates and marriage in Austen’s Emma.

Authors make things stand out in various ways using repetition which may through off bag-of-words algorithms. Novels try to portray the stereotypical and then violate it – “The economy of character.

Novels are performing both bias and the analysis of bias – they can create and unmask biases. How is text mining going to track that.

In A Matter of Scale, Jockers talks about checking confirmation bias to which Flanders replies about how we all operate with community consensus.

The lone objective researcher is an old model – how can we analyze in a community that develops consensus using text mining? To do this Laura Mandell believes we need capta open to examination, dissensus driving change, open examination of the algorithms and then how visualizations represent the capta.

The Canadian Society for Digital Humanities 2016 conference was held this year in Calgary, Alberta. Milena Radzikowska presented a paper on “Materializing Text Analytical Experiences: Taking Bubblelines Literally” in which she showed a physical system designed to materialize a Bubblelines visualization. (Bubblelines is a tool in the Voyant suite of tools.) In here talk she demonstrated the materialization filling tubes with different coloured sand for the words “open” and “free” as they appeared in a text. She talked about how the materialization changed her sense of time and visualization. Read more about the conference in Geoffrey Rockwell’s conference report.

In mid-March this year, I was contacted by Prof Christoph Meister of Universität Hamburg, with whom I had previously collaborated on the re-branding of the European Association for Digital Humanities (EADH). He wanted a logo for the 3DH project.

In the course of the following two weeks, I engaged in an intensive email exchange with the 3DH team and Profs Johanna Drucker and Geoffrey Rockwell, both of whom are visiting professors in Hamburg this summer term as part of the 3DH project. By the end of the month, we had worked out a logo design that everybody considered a success.

The following timeline is a collage of discussion fragments, logo sketches and drafts that passed back and forth in an ad-hoc collaboration conducted entirely via email; the timeline seeks to document the main ideas that guided the collaboration and to capture a sense of the process by which we arrived at the final design as displayed on this site now.

…

In his initial message to me, Christoph emphasised the need for the project to have a visual identity. On 14 March 2016, he wrote:

I’ve just started a new research project for which we need a visual identity – and this time that’s doubly important as the project itself is about visualization.

Moreover, the design needed to reflect the ambitions of the project as articulated in Johanna’s latest book:

[I’m looking for] a suitable logo idea that can serve to highlight what Johanna so aptly emphasizes in “Graphesis”: the importance of thinking about visualizations as a genuine epistemic and explorational device rather than a mere representational instance of ‘data’.

To illustrate what a successful design might look like, he cited a drawing reproduced in the book:

I came across Johanna’s mention of Kandinsky’s “From Point and Line to Plane” on p.35 in “Graphesis” and was immediately attracted by Fig.98 in the right hand margin: For me the vertical line touching the upper border signifies a subtle transgression of the idea of visually supported dualism as it pulls the reference plane within the square and that of the perceiver notionally situated outside the square (if you wish, the discourse plane) together and brings them into contact.

I accepted the job and set to work.

Identity design is in large measure typography, and typography is often a good starting point when creating a new logo. For this project I decided to start from the genre of decorative typefaces known as ‘shaded,’ as these invoke a sense of 3-dimensionality.

I also noticed a possible connection between the design brief and a peculiarity of the 3DH acronym. I knew enough of Johanna and Geoffrey’s work to understand that the phrase ‘epistemic and explorational’ in Christoph’s brief related to their conviction that interfaces and visualisations should be re-conceived to facilitate interpretation as their primary affordance; at the same time the D in the 3DH seemed to invite, if not require, interpretation due to its indeterminacy. I wrote on 16 March 2016 at 12:04 hrs:

As design briefs go, the above requirement is definitely one of the tougher assignments I’ve seen.¶ I wonder if it could be solved through a bit of playfulness. Let’s start with the project name: 3DH. There’s a quibble to be had from this name as to whether the 3D or the DH part should be the privileged reading: the D is ambiguous, therefore in need of interpretation.

I proposed a series of four typographic markers based on this idea, using the colours black and red to delimit the 3D and the DH groupings, with the fourth piece in the series separating the two colours in a diagonal division running through the letter D:

Johanna responded to this design by bringing up the concept of parallax, the displacement in the apparent position of an object viewed along different lines of sight, which she had discussed previously in some of her published work. In this work she asserts that through visualisations implementing the concept, it will be possible for value, identity, and relation of temporal events to be ‘expressed as a set of conditions, rather than givens’. She wrote on 16 March 2016, 13:24 hrs:

I’m wondering if the concept of parallax could be built in here to go “beyond representational concepts of visualization”.¶ Unfortunately, most diagrams of parallax are pretty reductive. But if you could imagine the 3DH logo you’re playing with constructed from two points of view or scales and have them not match but still relate–sort of like extending that diagonal slice through the D in the fourth version of the logo, but so that it refracts the letters. I would suggest lightening the design as well so it is not quite so solid/architectonic.

She illustrated her idea with a number of hand-drawn and scanned sketches, collated into a single PDF.

A diagram consisting of conceptual space mapped out by thin lines and inhabited by typographic elements at comparatively small size is in danger of looking ethereal and anaemic at small scales, especially when displayed next to more conventional logotypes.

On 21 March 2016 at 15:19 hrs, Johanna conceded that her diagrammatic approach would be susceptible to the scaling issue, yet she suggested that there might be a way to merge our separate approaches into in a single design:

I wonder if we can work with that ambiguity and an indication of non-identity or non-similarity between the two meanings of the “D” in the acronym. That could introduce the parallax issue in some way.

I frankly didn’t know how to act on this suggestion, so I tried a variation on the D that faces two ways. I understood that both Johanna and Geoffrey were opposed to anything ‘Cartesian’, so central perspective was out of the question. My suggestion made use of a roughly cobbled-together axonometric projection, about which I wrote, on 22 March 2016 at 11:48 hrs:

Attached as well is another take on the same idea expressed in a 3×5 pixel font rendered 3D in axonometric projection (channelling my inner Max Bill here). The piece turns on the ‘ambiguity’ of the D again, as the letter associates with the 3 in its orientation but associates with the H in its colour.

Christoph was intrigued by the piece, and he wrote on 22 March 2016 at 12:53 hrs:

it creates a weird Escher-like paradoxical n-dimensionality that loops onto itself and makes it, how shall I put it, “performative” in that you simply cannot stop re-processing the image.

He encouraged me to pursue the idea further, but Johanna, as I had anticipated from her critique of my initial offering as too ‘solid/architectonic’, was unconvinced. She urged a change of approach on 22 March 2016 at 13:07 hrs:

We might consider using the positive/negative space instead of closing the forms

Playing on positive/negative space, unlike the parallax idea, was something I knew how to handle. I wasn’t very keen on the idea because it seemed to offer less scope for the play on the letter D, but I pursued the idea anyway, resulting in a few iterations that struck me as nicely done but showing little relevance to the design brief.

Meanwhile, Johanna had been at work trying to bring about the merger of our separate starting ideas that she had hinted at. She wrote on 23 March 2016:

I’m going back to your sliced “D” idea and seeing if I can play with some parallax in it.

She supplied a sketch with two drawings:

Johanna’s drawings were a welcome occasion to drop the ‘negative space’ idea. They reminded me that a few days earlier my preoccupation with ‘shaded’ typefaces had led me to look at fonts constructed as impossible objects, a typographic genre often associated with the name of the Dutch artist M.C. Escher, who may have done most to popularise such objects. I wrote on 24 March 2016 at 22:29 hrs:

I returned to a few recent Escher-inspired retail typefaces and examined them under the aspect of whether the two yoked-together perspectival components of the D might be coloured different to convey the ‘ambiguity’ of the letter. I put this through a few iterations until it occurred to me that I could vectorise Johanna’s D sketch and use it in the same fashion.¶ Which I did as a rough and ready first cut.¶ Please find the whole series also included in the zip. Don’t worry about the grey/white/yellow colour scheme just yet.¶ I think we have a candidate here.

The attachment included two pieces that would form the basis of the eventual design.

This batch of pieces was well received. Johanna wrote on 24 March 2016 at 22:45 hrs:

Oooohhhhh! I am really loving these. I have my favorites, but will hold off until others weigh in. SUPER!!! We are really getting close, I think. Elegant, too!!

Christoph wrote on 25 March 2016 at 06:04 hrs:

Wow, Rudolf,¶ this is really a leap forward!

Things took a curious turn at this juncture: Johanna never came back to name her favourites among the Escher batch, whereas Christoph and I focused on the version of the design using Jeremia Adatte’s Bron Black typeface. Oddly, we both shared the concern that the 3 character looked like Homer Simpson‘s face, and that it required a modification to its shape. I also developed an obsession with searching for ‘impossible object’ typefaces and, telling myself I was doing my due diligence, went through as many such typefaces as I could find. And it occurred to me that we could bake the yin and yang motif into the design by not just making the D the location where the two colours cross over from one to the other; in addition, both of the other characters could have their respective main colour counterpointed by a small included segment of the opposite colour.

The work now seemed nearly completed. Christoph wrote on 26 March 2016 at 1:36 hrs:

I think we’re about to reach design freeze!

Johanna agreed, writing on 26 March 2016 at 15:38:

This has been REALLY fun! And so fast!

However, still unresolved was the question of what the colours would be. One possibility, perhaps the obvious one, was to rely on shading, which is conventionally used to evoke the physicality of a three-dimensional object in two dimensions; we could render the design in a local colour and a corresponding shaded hue.

Christoph had another idea. He wrote on 29 March 2016 at 12:48 hrs:

I have to consider internal politics and strategy: I would appreciate if we could either use Hamburg University’s color scheme (see https://www.uni-hamburg.de/) or one that resembles that of the City of Hamburg (which includes blue: see http://www.hamburg.de/). These are my funders who I need to get on board as co-owners and I want to make sure that they, too, will be able to identify with our project.

I objected to the adoption of Hamburg’s colour scheme, but Christoph insisted and asked for a draft of the logo, so he could attach it to a mailing to the project funders at the end of the week. In response to this request I started to look for ways to colour the design red and blue.

We seemed to have arrived at the end of the process now. Yet by that point I also nurtured a growing sense of dissatisfaction. It vexed me that we were merely going to apply a minor tweak to a typeface and to play a game with colours that seemed too clever by half while failing to state the design’s basic idea with any clarity or forcefulness.

Why weren’t we using Johanna’s D drawing, which was our own original creation? If anything, I wanted that drawing back! With due apologies for the very late about-face, I lamented on 30 Mar 2016 at 16:27 hrs:

instead of building further on the piece with Johanna’s unique drawing, uniquely connected to the project, we went for a generic, commercially available revival of a nineteen-seventies typeface

I attached a few revisions of the earlier piece.

To accommodate Christoph’s wish for Hamburg styling, I adopted the colour scheme specified in the branding guidelines of the City of Hamburg [PDF] and modified the typography. Throughout this project, I had been using the Futura typeface as a nod to the Bauhaus aesthetic, following Christoph’s mention of Kandinsky during the earliest stage of the collaboration. As Universität Hamburg’s branding guidelines specify TheSans of Lucas de Groot’s wonderful Thesis family of typefaces, I happily switched from the geometric sans serif to the humanistic sans serif.

I was apprehensive of the response to my about-face, as the proposal second-guessed what very much seemed like a done deal. However, both Johanna and Christoph supported the change right away. Johanna wrote on 31 March 2016 at 00:56 hrs:

I really love these […] the larger D in the center with the real dimensionality to it is terrific.

Christoph concurred and wrote on 31 March 2016 at at 06:54 hrs:

Escher-Drucker it shall be. It’s leaner, less self-absorbed and elegantly accentuates the dynamic D as a perceptual and intellectual axis.

This version was adopted, then, and made it into the mailing.

With the site coming online in early April, I implemented the logo and the Hamburg colours in a lightly modified version of the content management system’s GeneratePress presentation layer.

And this is how it all got that way.