Readers of this roughly 300-page work will be treated to a time capsule of the deep, experimental, and tentative thoughts of linguist and philosopher Margaret Masterman, and her influence on research in the pre-dawn of computational linguistics. An amalgam of Wittgensteinian picture theory of truth, semantic networks, breath groups, and a critique on Kuhnian paradigms does make for interesting, if not thought-provoking, reading. It can, however, only really be appreciated by philosophers and linguists dedicated to viewing their craft from a historical perspective.

The book, a posthumous tribute to Masterman, the founder of the Cambridge Language Research Unit (CLRU), contains 11 papers from vastly different subject areas, and of differing degrees of readability, some accompanied by commentaries to bring perspective to the uninitiated. Masterman, we learn from the editor of the volume, liked footnotes, diagrams, and notations. The editor, Yorick Wilks, notes in the preface that the papers “were in some cases only internal working papers of the CLRU, and not published documents” (page x).

Part 1 discusses “basic forms of language structure” in three separate papers: “Words,” “Fans and heads,” and “Classification, concept-formation and language.” Even though Masterman was very critical of mathematical and logical formulas a la Chomsky being applied to linguistic investigation, she applied a host of arcane formalisms that at times made it very hard for the reader to appreciate her work. A case in point is her notion of lattices and fans, somewhat derived from Wittgenstein’s notion of concepts. A fan is a construct depicting a word by its set of uses, represented as spokes in a fan-like structure. Fans are too simplistic, by not being recursive, and lattices are too restrictive, assuming a neatness of words and things that does not exist.

Part 2 investigates “the thesaurus as a tool for machine translation” in papers titled “The potentialities of a mechanical thesaurus” and “What is a thesaurus?” Masterman broke new ground in computational lexicography by emphasizing the usefulness of thesauri and dictionaries in her research. For her, these resources were constructs put together by their creators, and had a mind of their own based on their patterns and generalizations about the words they contained.

Part 3 features “experiments in machine translation,” in three papers: “‘Agricola in curvo terram dimovit aratro,’” “Mechanical pidgin translation,” and “Translation.” For Masterman, translation is a process of transferring meaning from one language into another via an interlingua as an abstract representation of language or as an existing form of language, such as Esperanto. In her assessment, the use of an interlingua is much more expedient than the direct translation from one language into another.

Part 4 addresses “phrasings, breath groups and text processing” in “Commentary on the Guberina hypothesis” and “Semantic algorithms.” Based on the work of Guberina, a Yugoslavian therapist for the deaf, Masterman developed a theory that a breath group was the spoken equivalent to punctuation in written language. Subsequently, she tried to tie repetition in language to the structure of breath groups, and ultimately to Greek rhetorical figures.

In Part 5, Masterman attempts to defend Braithwaite (her husband) against Kuhn’s theory of paradigms, in “Braithwaite and Kuhn: analogy-clusters within and without hypothetico-deductive systems in science.” Interestingly enough, the chapter provides a discussion of empirical evidence for scientific metaphors, a theme that was addressed in “Translation,” where metaphor as a process is a focal point in investigating language use. For Masterman, words were a priori metaphorical, and became concrete while being used in context, an interpretation that carries over to her investigation of semantic primitives and interlinguas.

Three things make Masterman stand out in the group of linguists of her time: first, her work in semantic primitives, leading up to her conceptualization of semantic networks as the leading paradigm of knowledge representation in artificial intelligence (AI); second, her belief in the primacy of meaning over grammar, which ultimately would be the central focus for computers to understand the meaning of sentences in particular, and language in general; and, third, her valuable and, considering the state of the art of computing, ingenious experiments for machine translation with semantic primitives and thesauri to analyze the semantic structure of texts.

Masterman’s conviction that meaning, not grammar, should be the basis of understanding language ran counter to the time’s prevalent dogma of structural linguistics, and the Chomskyan primacy of syntactic structure. Furthermore, her enthusiasm for Wittgenstein and his thoughts on language did not put her in the popular mainstream of linguistic research. Even more iconoclastically, she followed her instinct to disregard the edict by Bar-Hillel [1], another great in his time, that machine translation was impossible (for example, how can machines ever understand a sentence like “The pen is in the box”?). Masterman was able to conduct significant experiments on primitive machines to experiment with word sense disambiguation, semantics, and thesauri.

The selection of papers in this volume shows Masterman to be an “outside-of-the-box” thinker, in 21st century parlance, who liked to get to the bottom of things. Never having published a book to present her framework, albeit incomplete, Masterman had no chance to make a mark outside of the group of people privileged to work with her. What she did was leave behind scientific papers that are truly gems of original ideas, and provocative theses on linguistic study and philosophical thought; above all, she was asking the right questions.