Pun Salad

This is a fun book written by "statistician and journalist" Ben Blatt, who brings the power of Python's Natural Language Toolkit to analyze the writings of various famous (and some deservedly unfamous) authors. For example: people (like Steven King) will tell you to avoid adverbs, especially -ly adverbs; there are better ways to communicate. (Sometimes the advice is in joke form: use adverbs sparingly. Heh!)

But how well do authors follow that advice? Are there differences between authors? (Yes: Hemingway and Toni Morrison used relatively few. J. K. Rowling and E. L. James use a lot.) How about different books by the same author? (Also yes, and in many cases an author's more critically-acclaimed works have fewer adverbs than his or her others.) This is all presented with bar charts and tables. Cool!

Other queries: are there significant differences in male and female writing styles? How about between Americans and Brits? Do authors have "favorite" words? (Well, see the title.) How well do authors follow their own writing advice? Can you determine authorship of anonymous or pseudonymous works by crunching word use? (Yes, this was done convincingly for The Federalist Papers, showing that Alexander Hamilton got a little too enthusiastic in claiming authorship of some of them.)

And more.

Blatt doesn't seem to have carried out his investigations in any systematic way, just looking for answers to the questions that occur to him. Although he's not kidding about the "statistician" part—his Harvard degree is in Applied Math—there's not a lot of worry or discussion about whether the results he teases out of the data are significant. That's fine, it's still fun.