This is the first book in an interesting new computer science series on computer applications entitled “Explorations in LOGO” and published by MIT Press.

As the foreword tells us, this is an introductory textbook on linguistics and is devoted to the study of structures and patterns in meaning, spelling, sound, and style. An outstanding novel contribution is its systematic utilization of the programming language LOGO as a pedagogical tool, which allows an active learning style that encourages a playful exploration. Well-chosen examples study generative grammars that describe gossip, jokes, poems, new words, business documents, and various other chunks of language. Of course, the main focus is on the English language, but there are also lots of illustrations from other well-known languages such as French, Italian, German and Spanish, and even from less-used ones such as Basque, Estonian, Polish, and Romanian.

The book contains the following sections: Foreword; Acknowledgements; Introduction; Part I--The Grammar of Big Things: Sentences, Poems and Stories (chapters 1–6); Part II--The Grammar of Little Things: What is in a Word?--Morphology, Orthography, and Phonology (chapters 7–11); Epilogue; Appendix A--The LOGO Language; Appendix B--Some Roots, Prefixes and Suffixes; Appendix C--Strategies and Commentary for Selected Explorations; Bibliography; Source Note; and Index.

Chapter 1 introduces some basic ideas of natural language modeling through a computer language like LOGO, giving as a first example a simple grammar for gossip. Chapter 2 gives grammars a for simple poetry. Chapters 3 and 4 develop more complex and flexible grammar models. Chapter 5 explains the well-known controversy between competence and performance. Chapter 6 proposes various models emphasizing a human-machine dialogue.

Chapter 7 explores the spelling of suffixes and chapter 8 continues with the exploration of various other prefixes and suffixes and gives a short but vivid history of English. Chapter 9 gives an accurate overview of Indo-European languages and investigates the meaning of words. Chapter 10 explores the relations between sounds and spellings. Playing with the creation of new words, Chapter 11 studies some phonological grammars for natural language words.

The epilogue shows a 12-stanza poem created by the author with the assistance of some LOGO computer programs. The source notes are interesting, and the bibliography is comprehensive and up-to-date.

Although the book is not a LOGO textbook, the reader gradually learns, by means of a carefully selected set of examples, all the needed programming skills. The examples are well balanced between two categories: etudes, which are completely coded and explained, and explorations, which are independent, open-ended experiments, hints for which are given at the end of the book.

Appendix A gives some implementation details of various LOGO dialects. As elsewhere in the book, for a more complete description of the LOGO language and of the LOGO tools used, the reader is referred to Harvey’s anthological book [1]. Still, I feel that some of the more sophisticated programming techniques used, such as procedural parameters and metaprogramming, should have been explicitly mentioned and more thoroughly explained, as they usually tend to be less well understood.

Nevertheless, besides its linguistic relevance, the book can also be used as a valuable reference book for LOGO applications, possibly in an AI course focusing on natural language processing. Here I must mention that, although the book gives a fairly extensive treatment of the generative role of grammar models, it (seemingly purposefully) does not yet cover the topics of parsing and semantic analysis.

Many elevating cultural comments are spread throughout the book, such as relevant historical facts, clear etymological notes, and many adequate quotations (such as “A good model is worth a thousand facts”).

The book uses a very suggestive iconographic notation for LOGO procedures, illustrating them as box-shaped machines that may have some input hoppers and an output spout. The layout is attractive and the reviewer spotted only a few minor slips and errors (e.g., the first source notes do not point back to the corresponding pages and there is no such Romanian word as “aiales”).

The book also has a companion disk, which contains the tools described in appendix A and may be ordered in four versions (Apple, Terrapin, Macintosh, and IBM).

This book being the first in a new series, I should perhaps add some general comments about its motivating philosophy. According to G. Polya, an efficient learning process has three phases: first, an exploration (either directly with reality or indirectly with models); second, a formalization of concepts; and last, an internal assimilation stage. Computer systems, and especially powerful AI languages such as LOGO, offer unique opportunities for a fundamentally new kind of exploration with working models; this book is an outstanding proof of this assertion. I look forward to the following items of this promising series.