Premise

A couple years ago, I read Henry Fielding’s The History of Tom Jones for an 18th century British novel class. One aspect about that work that interested me was how the protagonist is repeated on the edge of entering a duel or engaging in a fistfight, mostly as a result of his philandering. The framing of this dynamic in most scholarship revolves around Fielding’s abhorrence of the honor culture amongst the aristocracy in Britain, and its deleterious effect on the English soul by way of escapades abroad that included the colonization of lands and wars. That said, I found that most of his considerations about violence centered on lawlessness within England, where Highwaymen could pillage as they pleased. Fielding’s founding of the Bow Street Runners, what eventually became the Metropolitan Police Service in London, also highlighted for me what seemed an unusual preoccupation with public violence in 18th century Britain.

For this assignment, I wanted to begin interrogating my intuitions about his interest in addressing crime performing a distant reading of Henry Fielding in comparison to a small corpus of 18th century British literature.

Tools

While I played with the Natural Language Toolkit in Python, seeing how I could remove stop words, case everything in lower case, remove punctuation, I felt I’d get tangled too much in tweaking things to get on with the analysis in an initial exploration like this assignment. Using Voyant Tools not only allowed for quicker distant reading, but also serves as an excellent place to develop research questions that could be handled in detail with more flexible tools at another time. I also used n-gram as a gut check against a large corpus than I could put together by hand.

I chose line graphs and stacked bar charts for visualizations, as they are reasonably easy to reason about 😉 .

Analyses

Tom Jones

My first stop was to add a text file version of Tom Jones via Project Gutenberg, a ready at hand source to compile corpora for Voyant Tools to consume.

Here’s what a gleaned from viewing Tom Jones in Voyant Tools.

Searching with the wildcard violen* when compared to happiness, happy and joy reveals the prevalence of violence mentioned throughout Tom Jones. From memory, I can imagine the highest peaks of violence revolving around (1) when Tom saved Jenny Jones (Mrs. Waters) from Northerton attempted rape. As each of these search strings represent about 1% of unique words in the text (13,019 unique words in total), comparing the relative frequency of these terms seemed useful, without major distortions in scaling. It seems as though violen* features heavily in the text, though one might assume that it is as equally represented as happiness/joy if combining words with like sentiments. The idea of symmetry between these terms dovetails well with Fielding’s love of symmetry in plotting.

Joseph ANdrews

I have also read Joseph Andrews, another novel by Fielding, and outside of one particular scene where Highwaymen rob a carriage Joseph is on, and strip him naked, I couldn’t remember there being too many threats of violence.

Here’s what I gleaned from viewing Joseph Andrews in Voyant

Two points to make when viewing Joseph Andrews work frequency for similar strings. One is that each of these search strings make up less of the unique words in this text vs.Tom Jones (each less than .3% of the 9454 unique words). However, we do is the same sort of symmetry, with the terms connoting joy and violence showing similar parabolic shapes in the middle of the document segments.

Corpus of other 18th British Novels

At this point, I wanted to see these terms in relationship to each other in a larger corpus. I created small corpus of other 18th century British novels I’ve read over the past few years, including, in this order:

• Clarissa and Pamela by Samuel Richardson
• Evelina and Camilla by Fanny Burney
• Tristram Shandy by Laurence Sterne
• Moll Flanders and Robinson Crusoe by Daniel Defoe
• The Female Quixote by Charlotte Lennox
• The Monk by Matthew Lewis
• The Mysteries of Udolpho by Ann Radcliffe 
• Gulliver’s Travels by Jonathan Swift

Here’s what I gleaned from viewing this corpus in Voyant Tools

This corpus represents as broad cross section of published novels during this time, from the scandalous (Lewis) to the popular (Burney and Richardson) to the less canonical (Lennox and Radcliffe). It also delivers an order of magnitude higher number of words when compared to the individual Fielding novels (millions vs. hundreds of thousands). I found the line graph of violen* and the happy keywords to be a little harder to read in this instance, though it certainly seems like violence hovers below the peaks we saw in Fielding’s novels, and the trend is downward over even later part of the corpus, which includes some of the more grift spectrum of the corpus (order of the texts masters in this context). If I had to generalize, Fielding looks more concerned about violence instead of happier words,

Stacked Bar charts of Fielding Novels

Viewing these keywords across reading time in Tom Jones offered an insight into the plot structure of comedy, though I suppose this is not something that distant reading alone as uncovered by comedy as a genre.

What was more surprising is the persistent of violen* in Joseph Andrews. I wonder if this is because this novel revolves around Joseph trying to “preserve his chastity” through a series of encounters with desirous women. In the case of Tom Jones, the protagonist is having sex throughout the novel, with only occasional spikes in violence during his expulsion for his adoptive father’s estate, and the incident with Northerton and other encounters with violent men.

N-Gram Viwer

As a further point of comparison and perspective, I looked at my keywords in Google Books N-Gram Viewer in British English from 1600-1900. I’m struck by the increas in happy, happiness and violence starting in the mid 1750 (around the time Fielding was writing). Given this larger corpus, I find it hard to dismiss the null hypothesis: that Fielding wasn’t more preoccupied by violence than his peers.

Reflections

I will share three reflections having completed this exercise

1. It is definitely the case that even a cursory exploration as shown above can generate more specific research questions. For instance, does the relationship between sex-seeking (Tom Jones) vs. sex-averting (Joseph Andrews) account for the differences seen temporally in each of those texts? How is that related prescribed gender expectations of men and women in 19th century England? Some of the critiques of distant reading straw man the method as an unreflecting and totalizing effort. But there are a number of sophisticated text analysis methods to perform, not contemplated here, that could be enlightening in conjunction will closer readings of each text.
2. Some of Voyant tools are helpful, while others seem to represent spaces for further analysis. One example: the context table is hard to make sense of in the tool as opposed to the visualizations. Though I suppose Voyant Tools is aware that they aren’t the last word in text analysis
3. Call me old fashioned, but a purely distant reading of any string of characters is missing the point of what a text is. As Roland Barthes put it in the opening of S/Z:

There are said to be certain Buddhists whose ascetic practices enable them to see a whole landscape in a bean. Precisely what the first analysts of narrative were attempting: to see all the world’s stories (and there have been ever so many) within a single structure: we shall, they thought, extract from each tale its model, then out of these models we shall make a great nar- rative structure, which we shall reapply (for verification) to anyone narrative: a task as exhausting (ninety-nine percent perspiration, as the saying goes) as it is ultimately undesirable, for the text thereby loses its difference.

And this is explicit task of distant reading: to lose the difference of individual texts and align them in larger sociologies of literature. Like Barthes, I view a text as product of the interaction of a reader with the work presented. I certainly couldn’t have reasoned about the “temporal” changes in the keywords in Joseph Andrews or Tom Jones having not experienced these books through reading.