The next layer of the modeling process is what I call the conceptual layer. It is here where you specify the conceptual proxies that you will be using to test an existing theory. Social scientists refer to these as “theoretical constructs.” These are ideas that are not directly observable in the world (like prestige or intelligence), but that we want to know something about and that we will use some form of observation to better understand. For example, at the most general level Lancashire and Hirst are interested in understanding the relationship between mental illness and creativity. They then specify these ideas through their core concepts of “age-related cognitive decline” and “vocabulary richness.” The former is one kind of mental illness and the latter is one aspect of creativity. These then are their theoretical constructs that they are interested in better understanding. While the theoretical and conceptual layers are very close to each other, by going through this exercise it will help you see where the very top of your ladder of generality lies and how you are beginning to specify your question (and what you are losing in the process). Age-related cognitive decline will imply that time is a fundamental feature that will need to be captured, while vocabulary richness suggests we will be focusing on semantic issues related Christie’s texts rather than other possible formal or historical features (like narrative or her relationship to the genre of detective fiction). I.e. your construct may have intrinsic qualities that will need to be captured in your observation or measurement phase.
Notice how the process of specification that belongs to building a model is also an act of realization. By making your research framework more specific and concrete, you make it more potentially implementable. Many, many questions in the humanities are extremely abstract. I actually think this serves an important purpose of idea exploration and expansion. If I ever write an autobiography it will be called How I learned to stop worrying and love Adorno. But these levels of abstraction and fluidity do not lend themselves well to empirical validation (which of course was Adorno’s point). If you want to explore and open our minds to new ideas, have at it. But this is not a good approach if you want to prove something about how the world works. For that, you will need to simplify things to make your concepts more approximate to the world and also to make your ideas more actionable. One of the things that data-driven research aims to do is be more impactful on the world around us. Humanists love to hate the word “impact.” But I’m going to embrace it because it forces us to ask, why am I doing this and whom is it for? To answer those questions, to be impactful, specification and, <gasp>, reduction are really important. Think of everything we have learned from fruit flies about the nature of life and you will begin to see the point.
Finally, notice how we haven’t even begun to talk about measurement or tools. I can’t emphasize this point enough. Don’t be the proverbial hammer in search of nails. Instead, your job as a data-driven humanist is to develop questions first and then build the appropriate tools to try to answer your question.
The Fish and the Painting 