Handbook for Language Engineers (review)

Handbook for language engineers is a guide to the field of language processing aimed at linguists with no previous training in the area. The book contains a total of ten chapters (including the introduction). Each covers an important aspect of language processing. ‘Domain analysis and representation’ by Ali Farghaly and Bruce Hedin looks at how attention to ‘domains of discourse’ (e.g. product purchase e-mails) can improve the quality of natural language processing (NLP). ‘The language of the internet’ by Naomi S. Baron provides an overview of the language of the internet in the broadest sense, covering the unique character of natural language on the internet as well as other issues, such as coding systems (mark-up languages), search engines, and crosslinguistic issues. ‘Grammar writing, testing, and evaluation’ by Miriam Butt and Tracy Holloway King gives an overview of various aspects of grammar implementation, focusing on development, testing, evaluation, and documentation. ‘Ontologies’ by Natalya Noy discusses the role of ontologies in language engineering, where the term ontology is understood to mean ‘an explicit formal specification of conceptualization’ (Thomas R. Gruber, ‘A translation approach to portable ontology specifications’, Knowledge Acquisition 5.2.199–220, 1993).

‘Text mining, corpus building, and testing’ by Karine Megerdoomian looks at the role of corpora in language engineering. She discusses the nature of corpora, touching on a wide variety of topics (corpus analysis vs. text mining, tokenization and annotation, the application of corpus linguistics, tools for corpus processing, symbolic and statistical approaches). ‘Statistical natural language processing’ by Chris Callison-Burch and Miles Osborne discusses the role of statistics in NLP, beginning with a discussion of probabilistic parsing, then moving on to modeling, estimation/learning, data, and evaluation. Matthew Stone (‘Knowledge representation for language engineering’) examines how world knowledge is represented in language processing. The discussion is broadly divided into three sections, each of which covers three layers of analysis in intelligent systems: the knowledge level, the representations-and-algorithms level, and the implementation level. In ‘Speech recognition and understanding’, Jan W. Amtrup looks at the state of the art in speech recognition and speech synthesis, focusing on some of the statistical methods that have been successfully employed in the field, such as hidden Markov models. Ali Farghaly wraps up the discussion in ‘Language engineering and the knowledge economy’ by looking at the changing economics of globalization and the place of language engineering in the new economy.

The growth of the internet has created many jobs for linguists in industry and raised the profile of computational linguistics. This book should serve as an excellent guide for those interested in a practical introduction to the current state of language processing technology.