Models are useful things to think with. They represent something in miniature. Like an architect’s model, they can be realistic approximations of the world, or, like a ship in a bottle, they can be whimsical and eccentric. When we think about data-driven research through the prism of modeling, we are reminded of the constructedness of our knowledge. Models are not the real world. They are representations of it. They involve craft, but also creativity. Data modeling isn’t something rote and mechanical. As I hope you’ll see, it involves a great deal of imagination.
Philosophers of science refer to models as tools of “surrogative reasoning.”[1] What matters is not their exact fidelity to the world, but their ability to enable what Gabriele Contessa calls “valid inferences.”[2] Models are meant to be useful, even the whacky ones. They help us solve some problem, give us an understanding of something, or make the world more navigable. But it is important to remember that you are always still relying on a representation of the world, not the world as it actually is.
In this section I am going to walk you through the six steps that I have identified for constructing a computational model of texts.[3] I find these useful, but by no means universal. Others may have other models of how best to model. I have found however that this can be very helpful for people just starting out to help organize their thinking and reflect in the steps of the process. I am going to stay at a very general conceptual level before diving into the details of how to process texts using specific commands. I think this conceptual level is actually more important. Tools will change, but how we frame questions and derive answers remains constant. It is the bedrock of research. The final point to make is not to think of models as singular and monolithic. No one model can approximate the true complexity of the world. We will always want multiple perspectives.
[1] Contessa, Gabriele. “Scientific Representation, Interpretation, and Surrogative Reasoning.” Philosophy of Sci- ence, vol. 74, no. 1, 2007, pp. 48–68.
[2] Contessa 54.
[3] For a fuller treatment, see Andrew Piper, “Think Small: On Literary Modeling,” PMLA 132.3 (2017): 651-658.
The Fish and the Painting 