Language, cohesion and form.

Language, cohesion and form. By Margaret Masterman, ed. by Yorick Wilks. (Studies in natural language processing.) Cambridge: Cambridge University Press, 2005. Pp. x, 312. ISBN 0521454891. $101 (Hb).

Reviewed by Aleksandar Čarapić, University of Belgrade

Language, cohesion and form brings together some of the most influential papers by Margaret Masterman (1910–1986), a pioneer in the field of computational linguistics and the founder of the Cambridge Language Research Unit. According to the editor, Yorick Wilks, the collection ‘is a posthumous tribute to Margaret Masterman’, which aims to represent ‘the influence of her ideas and life on the development of processing of language by computers, a part of what would now be called artificial intelligence’ (ix).

In addition to the editor’s preface, the collection consists of eleven chapters, which are organized into five parts. Part 1, ‘Basic forms for language structure’ (21–80), opens with Ch. 1, ‘Words’, which—using the term ‘word’ in the sense used by logicians—discusses three typical philosophers’ replies to the question ‘What is a word?’. Ch. 2, ‘Fans and heads’, is an extreme instance of M’s idea that certain kinds of logical formalism were essential for understanding the function of language. Outlining a sketch of a mathematical model of language, Ch. 3, ‘Classification, concept-formation and language’, proposes an alternative method of analyzing language.

As the opening chapter of Part 2, ‘The thesaurus as a tool for machine translation’ (81–146), Ch. 4, ‘Potentialities of a mechanical thesaurus’, deals with the thesaurus as an aid to mechanical translation (MT). It also provides examples of dictionary tree uses and outlines a mechanical translation program using a thesaurus. Ch. 5, ‘What is a thesaurus’, presents arguments for the necessity of an MT thesaurus.

Part 3, ‘Experiments in machine translation’ (147–223), opens with Ch. 6, ‘Agricola in curvo terram dimovit aratro’, which, using Roget’s Thesaurus, examines a first-stage translation from Latin into English. Ch. 7, ‘Mechanical pidgin translation’, provides ‘an estimate of the research value of word-for-word translation into a language, rather than into the full normal form of an output language’ (161). Ch. 8, ‘Translation’, presents a philosophical model of translation.

Ch. 9, ‘Commentary on the Guberina hypothesis’, opens Part 4, ‘Phrasings, breath groups and text processing’ (225–88). Ch. 10, ‘Semantics algorithms’, aims to compute semantic paragraph patterns.

Part 5, ‘Metaphor, analogy, and the philosophy of science’ (281–309), includes the final chapter, ‘Braithwaite and Kuhn: Analogy-clusters within and without hypothetico-deductive systems in science’, which, on the one hand, discusses Thomas Kuhn’s relativist conceptions of science and of a paradigm, and on the other, Richard B. Braithwaite’s account of science.

As the collection shows, M was ahead of her time because her beliefs and proposals ‘for language processing by computer have now become part of the common stock of ideas in artificial intelligence (AI) and MT fields’ (1). Some parts would not be easy to read without the commentaries of both the editor (Chs. 2, 8, and 10) and Karen Spärk Jones (Ch. 6). In short, the collection represents an important document on the development of ideas related to AI and MT and is a nice tribute to a scientist whose ideas did not get sufficient attention during her lifetime.