Arabic is among the 10 most widely spoken languages of the world. Due to this fact and to the increase of the economic and political significance of the Arabic-speaking world, it comes as no surprise that methods of natural language processing have been increasingly applied to the Arabic language in the last decade(s).

Owing to its word formation processes, Arabic poses special challenges to automatic processing of its morphology. The present book covers – as the title indicates – not only knowledge-based as well as empirical methods in the computational treatment of Arabic morphology, but also includes a part about the integration of morphological processing in larger applications.

The book consists of 15 chapters and is structured into four parts: Introduction, Knowledge-Based Methods, Empirical Methods and Integration of Arabic Morphology in Larger Applications.

The three chapters of part 1 (Introduction) set the stage: Chapter 1 (Soudi, Neumann and van den Bosch) gives an overview of the contents of the book. Chapter 2 (Habash, Soudi and Buckwalter) deals with aspects of transliteration (as opposed to transcription); in particular, it presents the transliteration standard used throughout the present volume. Finally, chapter 3 (Buckwalter) summarizes the issues one is faced with when performing Arabic morphological analysis.

Part 2 (Knowledge-Based Methods) presents four major linguistic frameworks as applied to Arabic morphology: syllable-based morphology (chapter 4, Cahill), root-and-pattern morphology (chapter 5, Al-Najem), lexeme-based morphology (chapter 6, Cavalli-Sforza and Soudi), and the use of a stem-based Arabic lexicon with grammar and lexis specifications (chapter 7, Dichy and Farhaly).

Part 3 (Empirical Methods) deals with data-driven processing models. Chapter 8 (Daya, Roth and Wintner) applies a machine-learning approach – combining several classifiers – to the problem of extracting roots of Hebrew and Arabic words. In chapter 9 (Diab, Hacioglu and Jurafsky), a support–vector–machine–based approach is used for tokenization, part-of-speech tagging and the annotation of base phrase chunks in Modern Standard Arabic text. Chapter 10 (Clark) discusses the supervised and unsupervised learning of broken Arabic plurals, centring on unsupervised learning with a memory-based model. Chapter 11 (van den Bosch, Marsi and Soudi) explores supervised memory-based learning as applied to Arabic morphological analysis and part-of-speech tagging.

Part 4 (Integration of Arabic Morphology in Larger Applications) addresses the application of Arabic computational morphology in information retrieval (chapters 12 and 13) and machine translation (chapters 14 and 15). Chapter 12 (Larkey, Ballesteros and Connell) shows the effectiveness of light stemming in Arabic information retrieval, which is far larger than that in most other languages. Chapter 13 (Darwish and Oard) addresses the adaptation of existing morphologic analysers and stemmers for use in Arabic information retrieval. Chapter 14 (Habash) discusses representational issues of morphology in the context of machine translation, and chapter 15 (Guessoum and Zantout) deals with the impact of Arabic morphological generation on the quality of machine translation output.

The book is well structured and well written. One minor comment on the transliteration scheme should however be made. It was certainly a good idea to develop a new transliteration scheme for this book; the readability in contrast to the previously widely used scheme by Buckwalter as quoted, has been increased significantly. The only unhappy choice is the symbol used for ‘ta marbuta’, as it is counter-intuitive: the same symbol is used in IPA for the pharyngeal unvoiced fricative (which is transliterated in this book as H); this may lead to confusion or at least, to an increased effort in concentration while reading.

In summary, this book is to be recommended for its broad coverage and its clear structure. It is – to this reviewer's knowledge – the first book to combine knowledge-based and empirical methods in the processing of Arabic morphology together with chapters on the integration of morphological processing in larger applications. It will thus be suitable for anyone interested in getting an overview of issues in Arabic morphology with respect to their computational treatment and current implementations thereof. Students and researchers of computational linguistics with an interest in Arabic or, in general, in languages with complex word-formation processes, particularly of the root-and-pattern variety, will certainly benefit from it, as well as Arabic language scholars wishing to learn more about the computational treatment of phenomena they are well acquainted with.