Humanities Data: A Necessary Contradiction

We say that Abstracting an entity (the most vague word we've found, referring to people, places, things, events, and even other abstractions) causes you to lose details. That abstraction is contextualized to a stakeholder, since you have to have some criteria for deciding what details you keep and lose. An abstraction for a person would be different if it was for Facebook ("Name", "Age", "List of Friends") or for a doctor ("Name", "Age", "Sickness"). Abstraction is powerful because it lets us concretize real, vague things into something that a computer can understand and process ("Because computers are very stupid!").

I often discuss how an abstraction loses a lot of valuable data. One student was working with NFL games, representing them a set of properties like ("away team name", "home team name", "away team wins", "home team wins"). That representation ignores how many interceptions were made, the weather that day, the feelings that the crowd had, how much effort the players put into training - a lot of "humanist" data.

This article makes me wonder what goes through their minds. No one's complained about making these abstractions - they've never suggested we're ruining a game, a historical event, or whatever we're abstracting. But I think maybe we should encourage the students to think about what happens when you abstract something and what it means for the entity and not just the stakeholder.

Jockers is, among other things, a scholar of Irish and Irish-American writers in the 19th and 20th centuries (IIRC). Apparently, according to previous work, Irish-American authors had "gone quiet" in the first decades of the 20th century, a fact that needed explanation. One of Jockers' early results, however, was that the quiet period was an artefact of the limited set of authors studied---the "close reading" model that produces statements like, "we’ve immersed ourselves so deeply in our source material that we’re attuned to its nuances" also means that the majority of source material hasn't been examined at all. In this case, a certain subgroup of authors, male and on the east coast of the US, may have stopped producing, but that output was more than replaced by other authors.