Research on metonymy in cognitive linguistics (henceforth CL) has grown substantially over the last 15 years, especially since Panther & Radden's (Reference Panther and Radden1999) collection of papers. This research has led to widespread recognition, in CL and elsewhere, of the conceptual nature of metonymy and its ubiquity in cognition, language and communication. This book, the first monograph entirely devoted to metonymy from a CL perspective, brings together most of this research and adds much more (especially the discussion of the possible neural basis of metonymy), by providing original theoretical analyses of many phenomena. It is a great, highly stimulating book of its sort, replete with rigorous, tightly packed discussions and a goldmine of examples, most of them from English and Polish. The critical remarks I make below are just friendly disagreements and suggestions for improvement; they do not by any means undermine my highly positive overall appraisal of the book.

Bogusław Bierwiaczonek (henceforth BB) declares in the introduction that the main aim of the book is to survey research demonstrating the pervasiveness of metonymy in language and to shed some light on the ‘possible … neural and evolutionary reasons' (1) for this ubiquity. The book definitely reaches both goals.

Chapter 1, ‘A short history of the concept of metonymy’, bears too modest a title, since it is not just a ‘short history’. The historical part proper occupies only the first seven pages out of the chapter's 60 pages. In it, BB presents and critically discusses Greek and Roman views on metonymy, structuralist formal approaches (especially Jakobson's) and their influence in some present-day proposals, and finally recent conceptual and semiotic views on metonymy (Nunberg Reference Nunberg1978, Lakoff & Johnson Reference Lakoff and Johnson1980, and Norrick Reference Norrick1981). The bulk of the chapter (from Section 1.4 onwards) is actually an insightful discussion of the various theoretical proposals on metonymy within CL since about 1993, all of which are being debated today in CL, and on almost each of which BB takes a stand. The chapter therefore constitutes both a useful introduction to metonymy theory for beginners and a stimulating critical discussion that will appeal to metonymy researchers.

The section on modern theories of metonymy (Section 1.4) analyzes Peter Koch's frame-based proposal and points out some of its limitations, and Kövecses & Radden's (Reference Kövecses and Radden1998) highly influential theory, which BB criticizes, among other things, for its reliance on a single domain or idealized cognitive model (ICM), which in his view would leave out such examples as The pork chop left without paying, and suggests to replace the single domain (matrix), ICM or frame notions in the definition of metonymy with what BB calls single integrated conceptualization, which includes both highly entrenched ICMs and online associations. In any case, Kövecses & Radden (Reference Kövecses and Radden1998: 58) made it clear that source (food) and target (customer) are connected by the specific restaurant ICM in this type of examples. BB also discusses insightfully Kövecses & Radden's attempts at constraining metonymy by means of their cognitive and communicative principles, and Ruiz de Mendoza Ibáñez & Mairal Usón's (Reference Ruiz de Mendoza Ibáñez, Usón, Radden, Koepcke, Berg and Siemund2007) own set of principles. In Section 1.5, BB continues the discussion of the role of domains in metonymy through his analysis of Croft's domain matrix and highlighting proposals (which he suggests to modify slightly) and of Ruiz de Mendoza Ibáñez's (Reference Ruiz de Mendoza Ibáñez and Barcelona2000) suggestion to reduce metonymy to only whole for part and part for whole (which BB accepts in part). Section 1.6 briefly deals with the distinction between metonymy and synecdoche (see below, on Section 1.8). Section 1.7 is an attempt at developing a typology of metonymy, where BB includes Panther & Thornburg's (Reference Panther, Thornburg, Geeraerts, Dirven and Cuyckens2007) propositional and predicational metonymies into one category, i.e. propositional metonymies, comprising sentential and predicative metonymies. The rest of the typology includes Panther & Thornburg's (Reference Panther, Thornburg, Geeraerts, Dirven and Cuyckens2007) referential and illocutionary metonymies and Barcelona's (Reference Barcelona, José, de Mendoza Ibáñez and Cervel2005) and Bierwiaczonek's (Reference Bierwiaczwonek and Kosecki2007) formal metonymies. I would add as a subtype of propositional metonymies those that guide inferencing in inferential chains (see e.g. Barcelona Reference Barcelona, Panther and Thornburg2003) or reasoning on the basis of metonymic models (Lakoff Reference Lakoff1987), whose source is otherwise not directly expressed, as in Mary's an excellent mother even though she has a demanding job as an executive (where reasoning is guided by the metonymy housewife–mother subcategory for whole mother category). Also, I find it wrong to claim that conceptual metonymies violate maxims (25); on the contrary, they are efficient prompts to derive the corresponding implicature.

In Section 1.8, BB comments on Peirsman & Geeraerts’ (Reference Peirsman and Geeraerts2006) prototype approach to metonymy, to which he presents four serious objections, some of which I agree with (BB ignores my prototype approach and my critique of Peirsman and Geeraerts' approach; see, e.g. Barcelona Reference Barcelona and Benczes2011 in Benczes, Barcelona & Ruiz de Mendoza Ibáñez Reference Benczes, Barcelona and Ruiz de Mendoza Ibáñez2011, a book not included in BB's references). BB offers his own alternative to Peirsman and Geeraerts' theory. In Section 1.9, he discusses quite perceptively the controversial distinction between metonymy, active zones and perspectivization (Cruse's (Reference Cruse2000) facets), and he comments on the metonymy-motivated use of certain proper names as common nouns. Finally, in Section 1.10, BB adds an element to the usual explanation for the ubiquity of metonymy, namely its usefulness in filling lexical gaps.

The discussion in all these sections is truly incisive, illuminating and stimulating, albeit occasionally confusing. I have a number of major and minor objections; given space limitation, I will only outline some of the former. In his alternative to Peirsman & Geeraerts' model (Sections 1.8, 32–40), which is based on BB's own general theory of conceptual relations, BB excludes from metonymy those mappings based on category–member (or ‘genus–species’) relations, which, following Seto (Reference Seto, Panther and Radden1999), he assigns to synecdoche, in turn excluded from metonymy; but he accepts meronymy-based metonymies. In my view, one should not confuse the static knowledge of category structure with the dynamic use of categories or category members to activate each other. And if metonymy operates within a ‘single integrated conceptualization’, both category structure and meronymy ‘in the world’ (31) constitute instances of such conceptualization. The same applies to ‘contingency’ (i.e. the absence of a conceptually necessary link between source and target) as a requirement for metonymicity: I need an aspirin (Panther & Thornburg Reference Panther, Thornburg, Geeraerts, Dirven and Cuyckens2007: 241) entails ‘I need a pain killer’ but it does not necessarily entail ‘I need any pain killer’ (i.e. any member of that category), so metonymies involving category structure are not restricted to part for part metonymies operating on ‘lower-than-basic’ subcategories (which BB calls synecdochic metonymies). According to his Principle of Minimal and Maximal Overlap (proposed inter alia as a way to distinguish metaphor from metonymy), BB claims that when ‘a lower-than-basic level category is used as a source for another lower-than-basic level category, the transfer is metonymic’ (34). But if I say This bulldog is a real poodle (given the bulldog's usual behavior, etc.) the connection is less likely to be (just) metonymic and more likely to be metaphorical. A similar principle, though applied only to the identification of degrees of prototypicality in metonymy, is proposed in Barcelona (Reference Barcelona and Benczes2011). As regards the distinction of metonymy from active zones and facets that BB insightfully discusses in Section 1.9, the notion of purely schematic metonymy I propose in that chapter might be of some use (especially as regards facets), together with the chapters by Paradis and especially Geeraerts & Peirsman in the same volume. BB objects to some aspects of my treatment of paragon names (as in Harold is a real Shakespeare) in Barcelona (Reference Barcelona, Radden and Panther2004). One of his objections is that the stereotypical model of Shakespeare (built around the property ‘immense literary talent’) underlying the use of ideal member for class is too rigid and ‘essentialist’; however, I do allow for variants of the model and for alternative, not mutually exclusive, models, which would lead to different paragonic uses of that name.

Chapter 2 is one of the chapters where BB presents some of his more innovative contributions to metonymy theory. In formal metonymy (6), also called salient part of form for whole form (27) by Barcelona and Radden in their work, ‘a part of the formal representation of a linguistic unit stands for the whole formal representation of that unit’ (61). After criticizing Radden & Panther's (Reference Radden, Panther, Radden and Panther2004a) analysis of the connection of Gosh! to God!, the chapter surveys a vast range of formal metonymies, which BB groups into writing metonymies and speech-sound metonymies. The first type includes such phenomena as letters for (dirty) words, homophonic alphabetisms (U for you), other alphabetisms (AO for Accountant Officer), and acronyms. The second type includes certain types of vowel reduction, clippings and other types of reduced word forms, phonaesthemes, certain types of rhyming slang, metonymies operating on phrases (especially generic adjective-headed NPs like the rich), sentential metonymies motivating tag questions, reduced comparatives, anaphoric ellipsis, gapping, independent subordinate clauses (especially if only-clauses), discourse- and pragmatic-based elliptical constructions, and raising constructions (BB offers an alternative to their metonymic account by Langacker). Despite my basic agreement with BB on the existence of formal metonymies, I would suggest to restrict them to pure form–form connections; in some of BB's analyses (e.g. full raising or independent if-clause constructions) the metonymies at work seem to be both formal and conceptual or only conceptual (reduced comparatives). In fact the metonymic motivation of morphosyntactic patterns is not restricted to formal metonymies (as witnessed by some of the papers in Panther, Thornburg & Barcelona Reference Klaus, Thornburg and Barcelona2009). I would also suggest not to regard discontinuous structures (like gapping or acronyms) as globally motivated by metonymy – they rather seem to work on the basis of analogy – since they do not constitute a ‘natural’ segment of the full form (this is debatable, of course, and I have myself occasionally not been consistent with this suggestion in my work).

Chapter 3, ‘Metonymy in morphology’, is an excellent critical survey of recent research on the topic: derivation (especially -er nouns in English and similar nouns in Polish); major and minor types of conversion; compounding, both endo- and exocentric (including a suggestive alternative to Langacker's classical analysis of jar lid); and onomastics (included in the chapter because most anthroponyms and troponyms involve conversion or compounding). I find questionable, however, the treatment of bahuvrihis (e.g. redbreast) as endocentric because their conceptual structure is more complex than BB suggests. I have suggested elsewhere (Barcelona Reference Barcelona2008) that the overall metonymy motivating these compounds is characteristic property (having a red breast) for (bird) category.

Chapter 4, ‘Metonymy in pragmatics’, is likewise an insightful critical discussion of research in the field, including Panther & Thornburg's (e.g. Reference Panther, Thornburg, Panther and Radden1999, Reference Panther and Thornburg2003) work on metonymy and pragmatic inferencing (but I miss at least a mention of Gibbs (Reference Gibbs1994), who inspired their work on indirect speech acts). The chapter includes BB's interesting analysis of propositional metonymies for love and of blending and metonymy in the illocutionary force of the past if-only construction.

Chapter 5, ‘Metonymy in semantics’, is a fascinating survey replete with detailed, perceptive analyses of the role of metonymy mostly in lexical semantics, which demonstrates BB's profound knowledge of lexical semantic theory. The discussion in the chapter is framed by BB's own theory of conceptual relations presented in Chapter 1 (and his exclusion of taxonomy-based tranfers from metonymy) and among other topics it deals with metonymy-motivated polysemy and with the connection between metonymy and interlexemic sense relations. An interesting notion is that of metonymygenerators (i.e. lexemes that tend to generate polysemous senses by means of metonymy).

The survey of polysemy includes synecdochic and metonymic extensions in general, eponymous uses of names (which then stand for non-personal categories: a colt based on the name of the inventor Samuel Colt), and verb polysemy. The topic of metonymy and sense relations is discussed in great detail, with interesting discussion of hyperbole and litotes, metonymy and partial synonymy, meronymy-based metonymy, antonymy- and complementarity-based metonymy, synesthesia-based metonymy, and other issues. Apart from a few confusing passages, the only objection I would voice is BB's continued rejection of any role to category–member metonymies in lexical semantics (which is congruent with his position in Chapter 1).

Chapter 6, ‘Metonymy in the embodied mind’, is one of the fortes of the book. It is a systematic analysis of recent research in neuroscience and related areas in search of a plausible hypothesis on the embodiment of metonymy. The core of that hypothesis and of the Neural Theory of Metonymy, whose foundations BB attempts to lay in this book, is the blending of what Damasio calls image spaces (corresponding, according to BB, to metonymic targets) with what that neuroscientist calls disposition spaces (corresponding to metonymic sources); the joint processing of these spaces probably takes place in the middle layers of the prefrontal cortex. BB claims that this blending process is mainly constrained by his ‘Principle of Correspondence’ (‘make your dispositional space fit your image space as closely as possible’ (252)), which is a re-formulation of three principles earlier put forward by Ruiz de Mendoza Ibáñez & Mairal Usón (Reference Ruiz de Mendoza Ibáñez, Usón, Radden, Koepcke, Berg and Siemund2007). BB has worked hard at reading most of the relevant neuroscience literature and makes frequent connections in the chapter to data and analyses in the previous chapters. He modestly presents his claims as mere ‘suggestions’ that he hopes will stimulate further research, especially by neurolinguists and other neuroscientists. A minor critique here is that despite his statement that linguistic synesthesia is only metonymy-based, this may only be the case for real synesthetes; but it is probably metaphorical (with a metonymic basis) for non-synesthetes.

Chapter 7, ‘Summary and prospects for future research’, is a clear brief summary of BB's main proposals and also a suggestion for future research.

On the whole this is a great book to be recommended to both experts and newcomers to metonymy research. It does not constitute yet a complete theory of metonymy, but it is a great step in that direction. BB exhibits in general an astounding command of the literature, independent thinking and keen analytical skills. The book is very well written, although a final thorough revision would have eliminated its relatively few formal errors.