This week's book giveaway is in the Other Open Source APIs forum. We're giving away four copies of Storm Applied and have Sean Allen, Peter Pathirana & Matthew Jankowski on-line! See this thread for details.
I'm not sure I completely understand the question, but I'll give it a shot anyway....
Just to be clear starting out: "documents" in MongoDB are actually JSON-like objects, not text documents like you'd maybe use with Microsoft Word or other word processors.
As to how you can do text analytics on fields within MongoDB documents, you can use the recently released full text search to do searching, and there are a limited number of string-processing functions within the aggregation framework, but in general I'd say that text processing isn't the best use case for MongoDB.
If you decide to do it anyway, you could certainly use MongoDB as the persistence layer of your analytics engine, but it's not going to provide you with many benefits over any other persistence engine.