While thinking about some examples of text mining for a report I am currently writing, I stumbled on a neat idea for legal texts, especially court filings and rulings. Just parse out all references to laws and their individual articles or parse all reference cases and analyze the cocitations. With relatively little work you should be able to get first results.
That would then be "Lawyers who liked this law also liked:" :-) I'd guess, that this could actually be a nice research tool if you start with some good starting, ermm, laws.
I mailed my friends at Juleg, who work on the lawsearch search engine, and immediately got an answer, that they're already working on a similar idea. Cool!