As Todd Rowland said the ideas of n-grams and Markov chains are easy to implement in Mathematica -- see
http://mathematicaforprediction.wordpress.com/2014/01/25/markov-chains-n-gram-model-implementation/ .

As Todd and Hector already mentioned, there are many built in tools that can help you with language processing. For example, you can find out how to combine Markov Chains, Graph and Networks, Statistics and even notions of Entropy to analyze the text of Alice in Wonderland and build a program that generates text with similar statistical properties. Read about it in the Wolfram Blog post Centennial of Markov Chains where you can download the complete notebook of that research. By the way, analysis in that article is similar but extended version of original paper by Andrey Markov presented for the Royal Academy of Sciences in St. Petersburg about legendary Russian poem by Pushkin called “Eugene Onegin”. Here is list of some articles on the subject worth reading due to originality and extreme popularity online:

• Centennial of Markov Chains
• The Longest Word Ladder Puzzle Ever
• 25 Best Hangman Words
• Code Length Measured in 14 Languages
• Word Facts
• The Distance between “Zero” and “Hero”: Exploring Synonym Chains with Mathematica

Also note large section on the subject at the Demonstrations Project: Linguistics

IMAGES: typical networks appearing in linguistic analysis  - click to see related article

I don't know of any projects that are open source like the Stanford project, but you can get pretty far with just WordData, which is built into Mathematica, and the ideas of n-grams and Markov chains, which are easy to implement.  Also ExampleData has a few texts which are immediately available.  See the linguistic guide in the documentation for ideas: Linguistic Data