This is one of the books that I consider as a starting point / reference whenever I need to deal with a practical natural language processing (NLP) problem. I also have Natural Language Processing with Python on my shelf and it's wonderful in terms of providing a practical start for nearly any NLP problem but when the need arises to cover more ground both in terms of theory and practical pitfalls then Jurafsky & Martin is my guide.

Natural language processing is a fast-moving target and it is impossible to know about the latest developments in the field without reading recent academic articles so nobody should expect to get the same information from this book, however mastering the concepts and algoritmhs in the book will provide the reader with the necessary background to understand state-of-the-art in NLP.

Most of the exercises are very interesting but I wish they had some kind of difficulty level indicated next to them. Another criticism would be that more information on practical implementation details of the algorithms could have been given but I believe these minor criticisms does not lead to a four star rating. It is a very difficult project to give a comprehensive overview of the whole NLP field and Jurafsky & Martin achieved that. ( )

Highlights are a good intuitive description of smoothing and an excellent exposition on the Early parser...Along with Foundations of Statistical Natural Language Processing by Christopher D. Manning and Hinrich Schütze, this is the standard textbook in natural language processing. For my money, Manning and Schütze is the better book—I find the exposition clearer and the math more rigorous—but Jurafsky and Martin covers more ground and assumes slightly less mathematical background. Regardless, the two books complement each other, so if you're in this field you need both.

Later...I'm in the minority for liking Manning and Schütze better, but I like this book too, and I like the second edition much better. This edition is a thorough rewrite that improves on the original. A highlight is a very thorough chapter on speech recognition. ( )

Well written and even a little humorus in parts. The technical and mathematical descriptions are clearly written and understandable. Of the two books required for CS 401R, this was more easily used and referred to. ( )