Processing issues:

1. Not all words in Stein's story share synsets, even when they share a common root (as in "regular" and "regularly").
2. Not all words in the story have synsets in the Wordnet database--that doesn't mean they aren't interesting to us, but simply that Wordnet doesn't bother categorizing certain kinds of words (such as pronouns).

Our analysis of each word token should account for:
a) its presence or absence in Wordnet. Where a word token is absent, retrieve its count and whatever information you've encoded in its attributes.
b) Where present in Wordnet with synsets, and those synsets do NOT overlap with those of other words in the database, see if there is overlap in the use of the source word itself in the wnDef , OR if it is contained in any of the synsets? Would this relationship be significant?
c) Simpler: How many of these words share a stem? (Get this from your own markup...)

@jonhoranic Start here with Wordnet:
Wordnet issue with lots of links on Pat's repo: PPH3/Eldritch#10

//repTerm[./lower-case(normalize-space()) = "gay"]

https://www.jetbrains.com/pycharm/