Using Google's Ngram viewer to trace the decline of English long S

(Note: this post is from late 2010, and I found it in my drafts today. No idea why it was never published.)

-----

In case you haven't noticed, a few days ago Google put out a new tool that shows the results of keyword searches starting back as far as 1700, and yes it's as addictive as it sounds. It has also let me solve something that I was thinking of doing some looking into myself: frequency of the United States is vs. the United States are. I had suspected that the latter would have been more common during the country's early history when the country was not as quite as tight a union as it is today, and it looks like my suspicions were correct. Starting in 1800 the United States is begins to grow, and as the Civil War happens it grows at an even greater rate.

One other particularly interesting use of the tool is tracking the usage of the long s in English, as this chart shows:

http://en.wikipedia.org/wiki/File:Historical_usage_of_long_s.svg

Since the software recognizes a long s as an f, all you have to do to track this is to find a word using a long s that doesn't look like any other English word when you replace it with an f. Paradise Lost wouldn't work then as loft is a real word. Street and ftreet does though.

Also note that the tool is available for other languages besides English: French, German, Russian, Spanish, etc.