- This topic has 2 voices and 4 replies.
-
Posts
-
I've gotta hand it to Google. As much as I dislike the behemoth that it is, it delivers well when it comes to researching historical topics. One of its recent creations, Ngram Viewer, is a way to search the frequency of words used in texts.I did a few searches to test it out using words which I thought were particularly popular during specific points in time. Take a look:latrine,fiddlesticks,groovy,flapperhttp://ngrams.googlelabs.com/graph?content=latrine%2Cfiddlesticks%2Cgroovy%2Cflapper&year_start=1700&year_end=2008&corpus=0&smoothing=3opium,heroin,moonshinehttp://ngrams.googlelabs.com/graph?content=opium%2Cheroin%2Cmoonshine&year_start=1500&year_end=2000&corpus=0&smoothing=3
Now that is a pretty cool application. I had fun with that for about 45 minutes without getting bored. The only question is how useful this would be for general history, it definitely has some utility if researching the history of an idea or ideas.
I don't think it has much use in too much research, unless we knew the percentage of published books that Google was using in its search. If Google was only using 50% of the books published in 1900, but 99% of the books published in 2000, the results would be skewed.
True, but Google has something like 20+ million books scanned in. Most are probably fairly old if you think about it. They have been busy scanning college libraries which don't tend to carry all the latest fiction and are largely skewed towards intellectual and research type books with a scattering of the classics.
Yeah but how can we know how many books (percentage) from 1750 are in google's database? I agree that it is quite impressive and they keep scanning more and more, but I'm sure ther is much left to go.
