Via @pourmecoffee on Twitter, check out this interesting site, Capitol Words. Like Wordle, it removes common words and makes a word cloud of the adjectives, verbs, and nouns most used in a particular body of text. Unlike Wordle, it is dynamic. On Capitol Words, you can chose the timeframe and create a word cloud for Congress’ and your individual federal lawmaker’s most frequently uttered words. Very cool.
I used a word-frequency analysis in my thesis, and I know my committee did not consider it very scholarly. It may indeed not be very scholarly, but I do believe word frequency analysis is compelling and ultimately meaningful, especially in this context.
Once upon a time … I subjected Bush’s first four or five States of the Union to the word-frequency treatment. Links to the result here: http://radio.weblogs.com/0130494/2004/01/23.html#a55
I spent a lot of time turning the 2004 State of the Union (‘the of state union’) into what I fancied to be a form of narrative poetry. I don’t think it caught on.
I think you’re right that it is fascinating. As far as scholarly, it depends on what sort of analysis you bring to it. Simplest level is simply looking at the key words, naturally. And when you’ve got a series of documents/speeches/utterances to crunch, you can trace changes of emphasis over time. I think it would be most interesting to look at not only word frequency and emphasis but the question of framing — the meaning built by how the terminolgy is used.