Story Core: UVM Researchers Use Data Mining To Identify Emotional Story Arcs
Finding and mapping typical story arcs is nothing new, but now a group of researchers at the University of Vermont's Computational Story Lab has used data mining to look at the emotional arcs contained in thousands of stories. They've identified six basic core arcs that form the emotional foundation of complex narratives.
Andrew Reagan, a mathematics Ph.D. candidate at the Computational Story Lab, led the research team that identified the six emotional story arcs and joined VPR to discuss his team’s findings.
On how the research started:
"We got into this by looking at happiness and originally were interested in measuring words. So we started this this tool called the hedonometer and using a survey approach, we were able to measure the happiness of the most used words and language.
"Out of that, words and sentiment are important, but what we're really hoping to get out of the group is the story lab. Stories are big with us and they're what we use to share experiences, so this has kind of been a goal of ours for a while."
On using "sentiment analysis" to find the core emotional story arcs:
"The basic idea is ... given a bunch of text we can measure the happiness of the individual words. We've done this by having a lot of people rate the most used words, and when we look at a piece of text we kind of take a telescopic view of it. We're looking at 10,000 words at a time and we're taking the average happiness of all the words in a group of text to get a feel for the temperature of the sentiment and text."
"We're looking at 10,000 words at a time and we're taking the average happiness of all the words in a group of text to get a feel for the temperature of the sentiment and text." - Andrew Reagan, mathematics Ph.D. candidate at UVM's Computational Story Lab
The six core emotional story arcs:
"The rags-to-riches story was the most common, the second most common was tragedy. Rags-to-riches [is] a rise and tragedy [is] a fall. We found four more that are pretty common: Man in a hole, which is a homage to Kurt Vonnegut, who inspired this line of research. The other ones are Icarus, which is a rise and then a fall; the Cinderella story, that's a rise, fall and then rise again; and the Oedipus story, which is a fall, rise and then a fall."
"There's been a lot of theories in scholarly literature about what the most common story types are. And we're measuring something slightly different, which is the emotional trajectory."
On the purpose of identifying these emotional arcs:
"There's been a lot of theories in scholarly literature about what the most common story types are. And we're measuring something slightly different, which is the emotional trajectory. It’s kind of three levels. You have the plot, the sequence of events and then the structure. Then produced by those is the emotional experience of the reader.
"We're studying this to answer those questions in a data-driven way by looking across lots of books. You couldn't, for example, read all 1,700 books in detail [and] catalog how the fortune or the emotion of the characters [were] portrayed and then group them, so we’re using a data science approach to do that."