Book Reviews: Errors and Intelligence in Computer-Assisted Language Learning: Parsers and Pedagogues by Trude Heift and Mathias Schulze

As is immediately clear from its name, Computer-Assisted Language Learning (CALL) is a field in which computers and language play key roles. To conclude from this that CALL would be a prominent area of computational linguistics, however, would be hugely mistaken. Handbooks and journals of computational linguistics, including this one, do not usually have articles on CALL. Nerbonne (2003) is a rare exception. Conversely , journals devoted to CALL (e.g., CALL, ReCALL, CALICO) contain many articles that readers of this journal would consider far outside the domain of computational linguistics. They are devoted, for instance, to the use of email or chatting in language teaching. Nevertheless, I think there is every reason for computational linguists to be interested in CALL. And the book under review is a very good starting point, for the part of CALL that is likely to be of most interest to computational linguists that is the topic of this book. Often this is called intelligent CALL (ICALL), but the authors prefer to call it parser-based CALL. Parser-based CALL is informed on the one hand by second-language acquisition theories and teaching practice, on the other by natural language processing and artificial intelligence. The purpose of this book is to give an overview of the relevant issues and the state of the art in each of these issues in a way that is accessible to practitioners of all of these fields. Computational linguists without a background in language acquisition will therefore get a good foundation in all the relevant domains to become fully updated on (the most interesting part of) CALL, but they will have to put up with (or skip) basic introductions to, for instance, the theory of formal grammar. The book consists of a brief introduction, four chapters (which the authors consistently call " parts ") giving an overview of the current state of parser-based CALL, and a conclusion with an outlook. The introduction identifies the subject field and gives an overview of the structure of the book. Chapter 2, " NLP in CALL, " is by far the longest (74 pages). It starts with an argument about the highly interdisciplinary nature of the field and the influence of different currents of psychology. Three sections (37 pages together) are devoted to an outline of the history of computational linguistics and an overview of parsing. The final section (30 pages) is an overview of parser-based CALL projects. …