I wrote about a paper analysing some 500 billion words from 5 million words scanned into Google Books. As part of the study, the authors worked with Google to create an real-time browser that could search the publicly available database, and the Ngram Viewer was born. This rapidly blossomed into a game to find the most insightful or amusing “ngrams” (“n-gram” means a set of words where n is the number of words; “apple” is a 1-gram, while “never gonna give you up” is a 5-gram). Here’s a short FAQ from me on common problems encountered with the search, and some of my favourites:

This is amazing. A doctor manages to get a hoax “buttock cupping”