Derek pointed me to the style.org site which also parses political speeches. Their preferred graphic is not the tag cloud but a labeled bar chart.
From top to bottom, each bar represents a sentence; the length of each bar is the length of each sentence. Further, the user can specify word pairs for comparison. Here the red bars are sentences containing the word "freedom"; the blue bars, "security".
It's a good illustration of the "small multiples" principle in constructing comparative graphics.
However, the choice of dimensions is perplexing. I'd be much more interested in the timing of mentions of those words, rather than which sentence they appeared in. I also find the length of each sentence to be irrelevant.
Here's one concept that brings out the point better. It uses less space and voluntarily gives up some of the data (the sentence structure).