It seems the text mining issue has struck a chord with our group, so I will jump in as well. Specifically, I want to refer to Sarah’s thought that the potential danger lies with the scholar interpreting the data prism could potentially collect, not with “the machine.” This allows us to do what comes natural in the humanities: critique the conclusions a scholar makes when they attempt to make use of the data that Prism “collects.” There is a well established system in place that – especially given the growing understanding of DH processes among humanities scholars – can “sift” through scholarship that uses a tool like Prism to make poorly supported conclusions.

I’d like to propose that we think about what kinds of things we can learn through Prism. Thus far we have defaulted to describe our “colors” with tongue-in-cheek phrases like “Happy passages” or “Daddy Issues passages.” I may be wrong, but I don’t think we are truly trying to “map” happiness in literature.  To make this process more tangible for our group, I think we should make it a priority to decide on a sample piece of literature and at least two possible “colors” or “expressions” that we deem valuable as crowd-interpreted data. Some of the debate about using prism to “quantify” and “mine” literary expressions or feelings may diminish if we were had a more concrete problem we’d like to solve as an example. We will thus avoid thinking about Prism as a method for mapping “happy.” Clearly there is an aspect of prism that is designed to be “hands off” and allow the crowd sourcing to be less directed and therefore more “honest.” However, many of us are simply not comfortable sending a powerful robot into the world which could be easily told to shake its metal hands and help people draw spurious conclusions before we see an example of this robot saving kittens from trees and making life better for literary scholars. Think Schwarzenegger in the original Terminator versus Schwarzenegger in Terminator 2…