Friday, 14 January 2011

He said, she said.

The gang at Google have long been amazing the world with their quest to provide free access to as much of the worlds information as they possibly can. From Google Earth, to Google Search and Google everything in-between, their work has been amazing (just check out Google Labs to see what else is on the boiler..).

For a while now, Google has been indexing books from around the world, digging up some of the oldest books in existence,  scanning them, converting them to text, and making them available (licence permitting) via Google Books. An outcome of this is a database of practically every word ever published sitting on a database in Googleland, just waiting for someone to find an interesting use for them.

Enter nGrams from Google. This very interesting tool can show you the popularity of any word, or string of words published in almost any book for the past 200 years. And yes, any word (all the good ones are there too!). This led Chris Harrison, a graduate student in engineering at Carnegie Mellon in the US, to develop the above graph to show exactly what She said, and what He said – in literary sense anyway, for the past 200 years.

He analysed all text strings that began with He and She, and developed the ultimate battle of the sexes. The most popular? Both He and She were followed with “is a” more times than every other combination in history, but a closer look also shows that “He argues” whilst “She loves” really does show that Mars and Venus may be our respective homes after all.