19/12/2010

Lennon Was Wrong

Look:


That's from Ngram, the tool that searches for words in books archived in Google Books. I predict it's going to be a runaway success because it provides a tool of quantitative social research for the very lazy person. Here's a more sociolinguistic example:


(Note, however, that the denominator for the two is different. For "Negro" it is all single words, for "African American" all two-word combinations.)

Here are two from the German corpus, starting with a surprise result (I would have expected more Hitler and less Brandt):




Added: Serious discussion of the program's shortcomings at LanguageLog, including in the comments.

No comments: