2017-02-04 - Questions 3 (Dylan)
Questions from Week 3’s Readings (Owens, Flanders & Jannidis)
These readings both explored how the field of digital humanities is a delicate balance. As we make strides in bringing textual media into the digital world, we must be cautious of pitfalls along the way.
Owens: Defining Data for Humanists
Owens’ short essay suggested an interpretation of data as both human-made and computer-processable. Data sets constructed by scholars will always have a subjective quality to them; decisions must be made about what data is selected, how it’s organized, etc. However, the result is (hopefully, given proper formatting) able to be processed by a computer, and analyzed in the most strict, literal way. This dual nature is tricky. Owens mentions Franco Moretti, whose ideas about “distant reading” caused much controversy among the literary community. His detached, data-driven method flew in the face of close reading (I checked out this NYT article on Moretti, as well as this critical article, although, ironically enough, I only skimmed the latter). The main criticism of Moretti seems to be that literature is far too complex to be quantified or organized into neat boxes (although that’s heavily simplifying it.) It’s a fair point. Perhaps it may be dangerous to believe that, just because we can analyze and quantify some aspects of some texts–such as certain word counts or character interactions–we can form vast overarching literary theories with this data.
Flanders & Jannidis
This reading connects well with the idea mentioned above, particularly in the final paragraph:
“Obviously there is a conflict of values here: the computer-science perspective makes us look for a good general description applicable to all entities, while the humanities perspective makes us look for those features which make this entity special: models conceal when they reveal (McCarty, 2005:52). As digital humanists we feel the pull in both directions, and simple solutions would only sacrifice one side to the other.”
This is the dual nature of digital humanities. This is also why human-computer interactions are so fascinating. Humans design things (data sets, code, models) to be read by computers. These creations have a touch of human subjectivity and individuality to them; everyone has their own coding style, everyone has their own ideas about what’s important when crafting a data set. Computers, on the other hand, don’t have any emotions at all, nor have they any need of them. This also happens to be why people are much easier to trick than computers, and why it’s not quite accurate to use computer metaphors to talk about the human brain. I agree with Flanders and Jannidis that navigating these different perspecitves will not be simple.
(Although, I must say, I disagree with their formatting choices–I kept having to re-read lines due to eyeskip.)