Apparently we can teach computers how to read now.

Okay, that’s false. We can teach computers how to count words. I guess that’s cool. Right? I mean, theoretically if a word is used extensively throughout a text, it should be significant to the plot or subject. Of course, this is excluding common pronouns and prepositions and the like.

So if Sir Arthur Conan Doyle uses the word “little” more than eighty times in his first Sherlock Holmes mystery novel, and “small” or “little” over 150 times in his last novel, then we should make some kind of conclusion based on this data-driven result. But what kind of conclusion?

To me, analysis of this nature is really only useful if you, as a researcher, go into it with a pre-formed hypothesis. Just analyzing random texts like Sherlock Holmes novels or Michael Crichton novels or Shakespearean poetry is sort of useless without some kind of expectation of results.

According to Dr. John Lauden, “The Most Dangerous Game” written in 1924 by Richard Connell has some illuminating information that was revealed during an analysis of the text using his Useful Python Scripts for Texts. His results concluded that paragraphs of the story are on average quite short given the fact that it is mostly written dialog, and that you can actively separate the main sections of the story based on how it is written (narration vs. dialog). Exciting.

And so, while the abundant increase in diminutive adjective use in Holmes novels or ratio of narration to dialog in Connell’s story are interesting findings on the surface, it doesn’t really tell us anything useful. At least, nothing that my limited background in 19th century English literature can explain. Perhaps I need to be a scholar of 19th century English literature to really appreciate the nuances of adjective choice here.

Text analysis is a cool tool to use for research purposes, but given my own current interests in visual communication, I don’t really see its practicality in my own research. Especially given the fact that I would have to learn another programming language (Python) this year in addition to Ruby on Rails and markdown. Ugh, I can’t even.