When I was studying for my Masters in Classics, one of the texts I studied in depth was the Bellum Civile by Lucan, an epic poem from the Silver Age of Latin. In our classes we kept commenting on how Lucan’s use of color seemed distinct from other epic poems we were familiar with, such as Vergil’s Aeneid. The Bellum Civile (BC) seemed redder than other works. Not just bloodier, but redder. In addition it felt darker, with more attention being paid to darkness, shadows, and blackness. In this project I wanted to explore these observations. Is the BC redder than its predecessors? And why? Is its use of color consistent with other Silver Age works? And what about darkness? Additional I am interested in what we learn by tracing the instances of these references over the course of the text. Could it be a form of sentiment analysis? Would peaks in color terms or light/darkness help illuminate what is happening in the text, or would it show something else?
Before we look at the text itself, it is useful to know a bit about the context. Lucan (39-65 AD) was a Roman poet that wrote during the reign of the Emperor Nero in what is known as the Silver Age of Latin (18-33 AD). He is known for the epic poem the Bellum Civile, or as it is also known the Pharsalia. This poem details the events of the Roman Civil War (49-45 BC) between Julius Caesar and the Roman Senate, led by Pompey Magnus, end the fall of the Roman Republic. As such, the poem covers many battles, deaths, and suicides.
Rome’s most archetypal epic poem, the Aeneid, was written between 29-19 BC, around 84 years before the BC, during the age of Augustus, Rome’s first emperor, and depicts the events after the fall of Troy, and the fugitive Aeneas’s journey around the Mediterean and eventual settling of Italy and establishing the Roman people. It also contains many battles, both poems show trips to the underworld, however tonally they differ greatly. Mostly notably in the Aeneid human action is guided and prompted by the gods, whereas the gods have a very minor role in the BC. I wanted to compare these two texts as examples of the same genre, to see if their different tones could be reflected in their use of color.
Silver Age Latin was characterized by increasingly political and satirical themes becoming present in the literature, as speechmaking and rhetoric became increasingly more dangerous under the more tyrannical emperors such as Caligula and Nero. Neroanian literature in particular was known for it use of the grotesque in its imagery, over-the-top depictions of blood, gore, and other physical horrors. Seneca the Younger (4 BC- 65 AD) was another Silver Age author and contemporary of Lucan. By comparing his use of color with Lucan’s we could see if there were any trends from this time period, or is Lucan was doing something unique.
One of the biggest aspects of this project was assembling the materials. There was no existing dataset on color in the Bellum Civile, nor was there even a useful dataset on color words in Latin so I would know what to look for. I had to take a few approaches to building my dataset. Initially I downloaded a CSV of all of the words in William Whitacker’s Words, an online Latin dictionary. I then filtered it to just include adjectives, and then went through and deleted any non-color related terms. This way when I looked at the text I would have a reference for what types of words I was looking for. However I also wanted to record instances of light and dark, and this was not reflected in that dataset, also once I started working with the text I discovered that there were color-verbs that were also not included. So that wasn’t as useful as I had hoped.
Next I read a translation of A. S. Klein’s online translation of the BC, highlighting instances of color and darkness in the text. Then I matched those passages to the Latin and recorded the vocabulary. This helped me build the vocabulary, but was by no means a complete dataset, as while it is a pretty literal translation, this only unearthed about half of the color and light/dark instances. I then went through the Perseus Digital Library’s copy of the BC recording: the word, work, book, line, color family, light or dark, word group, and translation. Word group was mostly for the light and dark words to categorize them a bit more, such as words literally meaning light or dark, but also nature words (sun, lightning, etc.,) darkness, fire, day, night, and clarity (glittering/bright).
Once I had that dataset built, which round up being over 700 instances, I used the vocabulary I built to quickly scan Perseus’s copy of the Aeneid to create a comparison dataset. I only focused on color words in that text, and the light and dark terms were becoming unwieldy, and I decided if this was a path I wanted to go down I needed to find a way to simplify the work. Luckily I found an article “Diachronic Trends in Latin’s Basic Color Vocabulary” by Emily Gering, which had data on color in Seneca that I was able to pull out. Once I had my three datasets, I used Tableau Public to create my visualizations and dashboards.
Since I created my datasets myself I didn’t have to do any substantial cleanup on them, however I did have to do some filtering. Through conversations I had with Patrick Burns, who is a Classicist interested in digital humanities and did his PhD on the BC, I decided I should focus my work on more literal words, as opposed to some of the more metaphorical words I had recorded. While fire may work like light in a next, it isn’t the same. I filtered my data to just the color terms and words that literally meant light and darkness. Additionally I filtered out “pallidus” which can mean yellow-green, but more literally means pale. I decided this wasn’t color the way I meant in my research question.
Before I started creating my visualizations I sketched mock-ups of my dashboards. I planned on having one dashboard examining color, and a second dashboard examining light and darkness. However I was realizing in this stage that I didn’t have enough data or questions around light and darkness to build a whole dashboard focused on that alone. So I tried to find a way to combine all my visualizations on one dashboard.
My actual visualizations were pretty easy to create, because at this point I was very familiar with my data and my questions. Most of the effort at this point was on tweaking the design of my dashboards, which I will cover in the next two sections.
Design and UX Testing
Since my data was concerned with color and darkness, some of the initial design choices around the colors used in my visualizations were simple. I would use intuitive colors representing the colors of the data, only using a light grey instead of white since I had a white background. I made one dashboard with all of my visualizations, as at this point I was still thinking of this as one narrative, with most of the space being devoted to the comparison of the three authors’ use of color, with a few smaller visualizations focused on what is happening specifically in the BC.
For my first I did a think-aloud with a colleague who is working on an MA in Digital Humanities. While he did read my visualization in the order I anticipated, and got the main points I was trying to make, there were a few issues with the design that were raised that I addressed. I initially chose bold colors for the design, but he felt that since I was focusing on red that maybe the other colors (yellow, green, blue and purple) should be more muted. This way the readers are able to really focus on the red and the point would stand out more. The heatmap for the use of light and darkness in the text also felt tacked on, and wasn’t as integrated into the story I was trying to tell. Our session also generated an interesting conversation about what I meant by the grotesque, and I realized while I had a little bit of contextualization I could use more on the visualization.
My second pass at my dashboard got rid of the visualization about the use of light and dark, and instead replaced it with a chart of the vocabulary I was tracking and what color families they were in. I also changed the layout, putting the visualization of Seneca and Vergil below my BC chart, which gave me more space for a longer paragraph on Lucan and Silver Age Latin.
My second UX session was with another coworker that had done some DH work before. She appreciated the introductory paragraph since she isn’t familiar with Latin poetry. She brought up good concerns though about space. There was so much on the dashboard that it was a little crowded and that made some of the labels unclear. She also thought the book by book and vocabulary visualizations were very interesting. She suggested including the frequency of the terms as it would help people connect the visualizations to the text. She liked seeing the words, especially as someone who hasn’t studied Latin as a way to connect with the data.
With her feedback I decided to split my visualizations across two dashboards. At its heart there were two different stories I was telling color, color as a whole in three different bodies of work, and then a closer look at how color was working in the BC. Since I only had 2-3 visualizations on each of my dashboards I had more space to talk about what is happening in the graphics. I know the visualizations should speak for themselves, and I think they do, but for users who don’t know about the topic a little more explanation would be needed.
The feedback I got from class seemed to support this breaking up of the dashboard, and focused on tightening some of the design choices on the individual visualization. I changed the color of the combined black and white bar graph to reflect that the two weren’t separated, and changed the order of the stacked bar graphs. I put red on the bottom to highlight its consistency, followed by black, white, yellow, green, blue and purple. This way the colors I was talking about were closer together.
All of the feedback I got during the process of creating my data and visualization, and during the process of designing my dashboard really helped me refine my topics and questions. I think I created a concise narrative, which could be expanded with more data.
As I expected, my visualizations showed that Lucan used more red than Vergil. Going into this project I was expecting that perhaps Vergil instead used more greens (part of them poem is about Aeneas “discovering’ Italy and establishing what would become the Roman people), but that wasn’t as dramatic difference. What I didn’t expect, probably because I am less familiar with his work, was that Lucan’s use of red seemed to match Seneca’s. Rather than Lucan’s use of red demonstrating how he was doing something unique in his poem, it seems to suggest that it is a sign of what was happening in the literary scene at large at the time. Seneca and Lucan were both involved in a conspiracy to kill the Emperor Nero, which when discovered prompted them both to committ suicide. They had similar political views and perhaps that is what was coloring their work.
When looking closer at the Bellum Civile I was surprised to see that red was the only color in each book, and that it had the most distinct terms. So even though there are more instances of black, it is more concentrated in certain books, and is split between two main terms. Red also had three verbs, the only other color that had a verb was white. I think it is interesting that red then can be an active state.
These findings make me want to go back to the text and perform a closer reading of the passages where they appear. I am curious to see how the colors are being used and what they are describing. Maybe even though Seneca and Lucan are using red in similar amounts they are using it differently.
I would also like to explore my abandoned topic of light and darkness in the text. This would require building a larger dataset for the Aeneid. I am not sure I would need to for Seneca, as what I am interested in is how those terms are used across the books. Since the Aeneid has a similar structure that would be easier to compare. Along that thought I would also like to do a color breakdown of the Aeneid by book to see how color is distributed in the text.
While I think the visualizations are powerful, I think they could be strengthened if they were incorporated into an essay exploring these themes. Rather than living on as a poster or a digital dashboard, I think this presentation would give me the space to explore the visualizations in depth and bring in scholarship and specific examples.