Explorations for linguists 'in' love

This massive yet reasonably priced book provides a very useful guide to major parts of the "cognitive" strand of linguistics, and I have no doubt that every cognitive linguist will want it as a reference book. It provides a comprehensive introduction to a great deal of work by the main American players of the past decade or so (George Lakoff, Ronald Langacker, Michael Tomasello, Leonard Talmy, Adele Goldberg, William Croft). It devotes 150 pages to general issues, 300 to semantics and 300 to grammar - probably a fair reflection of the research interests and activities of this group of scholars and their followers, who constitute the mainstream of cognitive linguistics.

First, what kind of student could use it? The first chapter is clearly aimed at complete novices in linguistics who do not know what a morpheme is or what an asterisk at the start of a sentence means, and later chapters continue to explain basic concepts. But it is hard to believe that novices could follow the brief discussions of "double raising" or of universal grammar. The level of difficulty and detail is generally much more appropriate to second or third-year undergraduates or postgraduates.

Two qualities are pre-eminent in a good textbook: clarity and reliability. Unfortunately, clarity is too often marred by inconsistencies here. To give just two examples, the authors tell us that the preposition "in" is evidence for a metaphorical view of love as a container in "George is in love", but an earlier diagram for "The vase is in bits" shows the vase containing the bits, not the other way round as we might expect if " in " means "in"; maybe this is true to published research, but students need an explanation. Similarly, one of the two main distinctive characteristics of cognitive linguistics is the attempt to unify analysis at different levels (the "generalisation commitment"), but this claim is rightly retracted by the admission that every other kind of linguist shares this goal, too. As far as reliability is concerned, the account of cognitive linguistics seems accurate, but other approaches are too often misrepresented. For instance, the authors wrongly accuse generative grammar of eliminating redundant information, ignoring all the discussion of "redundancy rules" in phonology and morphology, which are designed precisely to allow redundant information to be included in lexical entries.

In summary, the book contains a lot of information about work in the mainstream of cognitive linguistics, but it is not suitable for beginners, and students should be warned of inconsistencies and inaccuracies.

Richard Hudson is emeritus professor of linguistics, University College London.