Today in mathiness

Suppose I suggested that we can think about how the plot of a work of fiction is progressing according to whether the emotional valence is rising or falling. I then asked you to think of what shape the overall plot could take, in terms of rising or falling emotional valence. You might suggest the following possibilities:

1. Rising throughout
2. Falling throughout
3. Rising then falling
4. Falling then rising
5. Rising then falling then rising
6. Falling then rising then falling.

Artsjournal.com links to this story from the MIT technology lab where scientists loaded the plots of 1,700 works of fiction into a really, really big data set, then used ‘word windows’ to check on the changes in emotional valence, so they could find out how many, and what sort of emotional arcs are most often found in fiction. With their really big data set, you’ll never guess what number they came up with, or what they are.

Six. See above.

Coda: I am not knocking digital humanities, and all the possibilities in the field. What I am questioning is the notion that big data is always helpful in understanding aspects of art, or any part of life, where the very definition of terms, and a bit of pondering, will lead you to exactly the same conclusions.