This highly intellectual book, based on a doctorate, is not for the fainthearted. For the reader prepared to take the time to digest its contexts, it is, however, rewarding and places our understanding of general linguistics today in a complex philosophical context. The text concentrates on the period from the election of Michel Bréal to the first Chair of General Linguistics at the Collège de France in 1864 to the 1916 publication of Saussure's Cours de Linguistique Générale, but also explains the origin of the powerful ‘vie du langage’ metaphor and its relevance to contemporary linguistics. This review outlines the main areas covered to help potential readers decide whether the book is of interest to them; the validity of its contents is unassailable.

The book opens with an examination of the generalised use of metaphor in science with references to Kant, Popper and particularly Schleicher, which may lead readers to a reassessment of accepted scientific ‘facts’. The starting point is Foucault's concept of épistème which determines what was thinkable at a particular point in time. Various areas of scientific endeavour are examined in terms of their impact on scientific methodology. Since what we now call Creationism ruled supreme, scientists had to find ways around the problem, hence Cuvier's theory of natural catastrophes to explain change. Lamarck, in the path of Buffon, then developed the concept of evolution and what had been a static vision of the world became dynamic. This led to naturalism not only in the natural sciences but also in linguistics, to which must be added the influence of Comte's positivism.

Naturalism along with Darwinism, led to the idea of the birth, development, and ultimate death of organisms, of which language was an example, and genealogical trees were created for various language groups (Schleicher). A disturbing aspect of this is how ‘scientific research’ gave a pseudo-theoretical basis for racism. There were, however, major theoretical problems with this theory, one being the inability to prove a common origin to all languages, essential to evolutionism. Henry also pointed out that, contrary to evolutionist principles, creoles prove that there can be cross fertilisation between languages. He therefore tried to dissociate linguistics both from research into the origin of language and from the natural sciences, distinguishing between the study of langage, which belongs to extra linguistic spheres such as anatomy or physiology, and the study of individual langues, the proper subject of linguistics.

This period was marked by numerous important debates. Was evolution due to nature or history? Which was the best classification model, the typological one based on morphology or the genealogical one based on change? Dauzat thought they should be combined, and they have certainly survived the test of time. Bréal, for his part, wondered how contemporary comparativists should deal with the works of the Anciens while Dauzat developed the idea of primary and secondary laws in order to solve some of the problems which had emerged from the previously set up phonetic laws (and laws were seen as essential for the establishing of linguistics as a science). But are they real laws? And are any of them language universals?

Such debates led linguistics to pastures new: Bréal, who coined the term ‘semantics’, moved away from phonetics to meaning, and studied the intellectual rules which correct the inadequacies of language (e.g. such as one word being replaced by two to avoid ambiguity). Darmesteter, author of the expression ‘la vie des mots’, saw language as belonging both to the physical and the moral world in its expression of thought. Studying words and their change of meaning led to an appreciation of the importance of metaphor, to which Henry added the concept of semantic association. Saussure, for his part, saw analogy both as a source of creativity (new combinations being formed) and as an organiser of linguistic material into categories, in which role it acts as a unifying factor. And in the debate as to the validity of using written language as a source of data, Saussure maintained the spoken and written language were two different systems of signs, with the latter the servant of the former. With his roots in comparative linguistics and faced with a negation of the whole subject area in the context of the continuity vs. discontinuity debate, he set up a list of linguistic antinomies: logic vs. life, synchrony vs. diachrony, langue vs. parole, evolution vs. state, construct vs. ontology, universals vs. individuals, which were to help shape modern linguistics. It is also at this time that the role of collective intelligence in determining language and of language as a mould of thought came to the fore.

Sociology finally broke the deadlock on the nature of linguistics. Durkheim was the foundation on which to rebuild the principle of linguistics as a science, but as a social rather than a natural science. He distinguished two types of social phenomena (fait social) acting as constraints on individuals: codified or prescriptive laws and commonly accepted un-codified rules (e.g. the concept of inappropriate behaviour). In the human sciences both theoretical and a-theoretical rules operate, the latter corresponding to speakers’ instinctive knowledge. This led to studies on what speakers think about their language in contrast to the views held by linguists. Thus sociology ‘cured’ linguists of their naturalistic view of language. As Henry put it ‘une langue ne naît pas’ but is the result of continuous change making it impossible to state, for example, when French started and Latin stopped. By the end of the 19th century a naturalistic approach to linguistics was dead. Or was it? The author maintains that the debate on the nature of language and linguistics, the origins of language, and the links between linguistics and biology are still current, and that the metaphor of la vie du langage is unlikely to completely disappear, since it is now part of what she calls – metaphorically – our ‘cultural genes’. An intriguing thought. But this work is also interesting from a broader point of view, since many of the debates referred to have been revisited in recent times: debates on the innate nature of language, the existence of rules and possible universals, and the use of ‘scientific’ statistics have marked the second half of the 20th century, while debates as to whether certain linguistic situations are to be seen in diglossic or variationist terms are reminiscent of the continuities vs. discontinuities debate.

In conclusion, this is an impressively detailed work that goes well beyond its historical interest.