Counting Tagged Words by Plot Section

We included counts of our tagged words for each plot section to show that some of the stories had only four sections. The number of words per section is not necessarily indicative of anything, though it generally correlates with the length of the story. "Old Nurse's Story" is the longest, while "John Charrington's Wedding" is the shortest. The length of the story is an important consideration when considering richness, simply because of how it is calculated.

Except for The Old Nurse’s Story, the “Haunting” section of each ghost story contains by far the largest amount of scary words per story. Both John Charrington’s Wedding and The Shadow in the Corner contain eight times the number of scary words in this section than in their introductions! Again, with the exception of The Old Nurse’s Story, the remaining four follow a general trend, building gradual suspense between the “Introduction” and “Protagonist” sections, displaying climactic scary word clusters around “Haunting” and “History,” and then significantly tapering off the at the “Resolve.” This is almost identical to the traditional dramatic structure chart, but with different plot points. These general trends may reflect the structure of the genre, but may also reflect our preconceived ideas about structure, as we were very conscious of those tropes as we did markup.

Scariness Value by Story

In the “Number of Scary Words per Scariness Rating per Story” graph, we see that most stories contain mostly scary words rated 1 or 2 and relatively few intensifiers. Including 0 and intensifier ratings, John Charrington’s Wedding contains the least amount of possible scary words at 93, while Walnut-Tree House contains the most with a whopping 583 possible scary words. The same holds for true scary words—words with scariness ratings of 1-3—with John Charrington’s Wedding coming in at 80 and Walnut-Tree House at 476. The Old Nurse’s Story has 101 scale-3 scary words, making it the Victorian ghost story with the highest number of scariest words, while John Charrington’s Wedding again comes in last place with only 23.

While we have not taken length of story into the data, it's interesting to see that there are consistently more scale-1 and scale-2 words per story. Fear, then, is more about key moments of suspense, rather than maintaining spookiness throughout.

We also see a general antipathy toward intensifiers. Whether this was on the part of the authors or the taggers is an issue we would like to address. Our markup, while complete, lacks some of the objectivity that a second pass reader could offer.

Scariness by Part of Speech

The scariest type of words per story, according to part of speech, are nouns, with a combined total of 99 scale 3 words in all five ghost stories. The least scariest type of words are adverbs and modifiers, with combined totals of 4 and 1 respectively. Adjectives were a close second for scariest word type, with a combined total of 90 scale 3 scary words, though we had originally suspected they would take the cake. The most action oriented ghost story was The Old Nurse’s Story, with 26 scale 3 verbs, while the most passive was John Charrington’s Wedding with only 1.

Richness of Expression of Scariness

Based on our WordNet analysis and the resulting vocabulary richness ratios, the story with the greatest diversity in vocabulary is John Charrington’s Wedding. We found this particularly interesting in light of the fact that John Charrington’s Wedding is the shortest of the ghost stories in our corpus. It is tempting to think that a longer story would include more variation in scary words, but there seems to be no direct correlation between the richness of the expression of scariness and the length of the story. The story with the lowest vocabulary richness ratio is The Shadow in the Corner, which is squarely in the middle in terms of length with respect to the other stories.

Our data indicates that lexical diversity in scary words is most likely a factor of literary choices on the part of the writer. For instance, the two ratios of vocabulary richness within the closest range of one another are those associated with The Old House in Vauxhall Walk and Walnut-Tree House, both written by Charlotte Riddell. Some authors are prone to repeating certain words, perhaps for emphatic effect, whereas other authors incorporate a wider range of chill-inducing vocabulary, engendering suspense by keeping the reader on their toes.

The richest story is John Charrington's Wedding, which we've judged to be one of the least suspenseful reads. The ghost is not revealed until the end: the story is sweetly sad rather than terrifying. It can be read as the most narratively complex, and perhaps the most literary story we include. It also fails to conform to the narrative tropes we assigned to it, completely forgoing any substantive resolution for a more broad, interpretive conclusion. This leads us to believe that richness of synsets may correlate positively with what we call "literary value," a quality which we can't really encode.

Were we to expand our data to include ghost stories of canonical literary value, like those from Dickens, we hypothesize the richness of language ratio would be comparatively high. This hypothesis also leads to historicist questions of class and gender politics, as vocabulary richness may be indicative of education level of the audience or author.


From our current data, we find that female authors writing Victorian ghost stories typically built suspense in two ways: (1) With a standardized plot structure; and (2) by using scary adjectives to describe scary nouns. In the Victorian ghost story plot, action gradually rises until the ghost appears, where it climaxes and then slowly dissolves as the protagonist finds some way to bring peace the situation. Despite the genre-specific tropes of "Haunting," "History," and "Protagonist," this more or less the pattern of most stories in the Western tradition. At a syntactic level, the norm seems to be to focus on the ghost or haunted house itself as the object of fear, increasing tension through describing its features and, to a lesser extent, by what the object actually does. This perhaps made the stories more accessible to middle-class readers, the authors relying more heavily on sensory, visceral details than deeply intellectual content. This does not imply that these stories were "dumbed down," but rather that they were meant to be more physically frightening, creating a sense of urgency and tangibility in the ethereal.

While it’s hard to declare objectively which of our five stories is indeed the scariest, we can make a few modest measurements. Walnut-Tree House has the highest number of scary words, but the words The Old Nurse’s Story uses are scarier. John Charrington’s Wedding is by far the shortest text, but displays the greatest increase in scariness from its “Introduction” to “Haunting” sections relative to its size. Perhaps he’d like to step out of his grave and claim the prize?

[an error occurred while processing this directive]

Creative Commons License Victorian Ghost Stories by Abigail Drabick, David Wade, Gabi Keane, Kaylen Sanders is licensed under Creative Commons Attribution-NonCommericial 3.0 United States