Because compounds are simply juxtaposed lexical items without any inflectional or other functional material overtly indicating the relationship between the constituents, the mechanisms by which they are freely produced and understood have intrigued linguists for generations. How, for example, do we readily understand that whereas olive oil is oil produced by compressing olives, baby oil is oil intended for rubbing on babies? The question is hard enough when the interpretation does at least involve the entities named, in this case olives, oils and babies, but it is even more difficult when the contribution of the component parts to the meaning of the whole is less obvious, for example in the compound egghead meaning ‘intellectual’. This, however, is the task Réka Benczes has set herself: to uncover the semantics of the group of constructions she calls ‘creative compounds’, namely those English noun–noun compounds that involve, to varying extents, the processes of conceptual metaphor and metonymy. These include examples where one element is metaphorical, e.g. puppy love and jail bird, or metonymical, e.g. white-collar worker and gaslight; some where both elements are affected, e.g. chicken hawk (‘a person who now advocates war but who once took special measures to avoid military service’), phone neck (‘neck and shoulder pain caused by holding a phone for long periods’), firedog and acidhead (‘an LSD user’); others where the relationship between the elements is metaphorical, e.g. clubfoot, or metonymical, e.g. doorknob; and finally, those where the whole compound is metonymical, e.g. hunchback. The book undoubtedly makes a contribution to this area of linguistic enquiry, and so, to some extent, the author succeeds. However, when one delves into the contents in more detail, there is a lack of precision and explicitness that makes for a frustrating read.

The book is number 19 in John Benjamins' Human Cognitive Processing series and is based on the author's PhD thesis. Her central argument is that compounds previously regarded as unanalysable exceptions, i.e. creative compounds, do in fact exhibit systematic semantic patterns and, furthermore, that this systematicity can best be described within the framework of cognitive linguistics. The first, introductory chapter poses two questions whose answers lead to this thesis: firstly, ‘In what sort of theoretical framework can the semantics of metaphorical and/or metonymical noun–noun combinations be systematically analysed?’ and, secondly, ‘What regularities can be uncovered behind the cognitive processes that yield [these] compounds; and the . . . compounds themselves?’ (pp. 3–4). The rest of the book is divided into two parts, part I (chapters 2–4) addressing the first of these questions and part II (chapters 5–9) addressing the second.

Part I, then, covers theory and past approaches, starting with a survey, in chapter 2, of some twentieth-century studies of nominal compounding. The author pays particular attention to the wide, even chaotic, variety of taxonomies that have been proposed, to the ways in which different authors have analysed the semantic headedness of compounds and to the relatively scant attention paid in the literature to those regarded as exocentric, i.e. with no overt syntactic and/or semantic head. These aspects are of particular relevance to the study because many of the combinations included in Benczes' category of creative compounds are elsewhere classed as exocentric. In addition, she revisits the reasons, first elaborated by Matthews (Reference Matthews1961), for the failure of transformationalist and generativist grammars to explain compound semantics. Then, having established that other schools offer no systematic account, particularly of creative compounds, the author moves on in chapter 3 to outline some principles of cognitive linguistics: the most significant of these, at least for this study, are conceptual metaphor, conceptual metonymy and blending theory. Benczes argues that cognitive linguistics in general, but most especially blending theory, is able to shed new light on noun–noun compounds, so that they ‘have finally become analysable’ (p. 41). She outlines the few previous compound studies within the cognitive linguistic paradigm as well as Sweetser's (Reference Sweetser, Janssen and Redeker1999) similar work on adjective–noun combinations. Finally, part I concludes with a discussion, in chapter 4, of the notions of compositionality and transparency. Benczes accepts Taylor's (Reference Taylor2002) argument that compositionality is a gradient phenomenon, but takes issue with Dirven & Verspoor's (Reference Dirven and Verspoor1998) proposal that compounds can be placed on a cline according to the transparency of their meanings; instead she argues for a cline of creativity.

Part II consists of the detailed semantic analysis of seventy-eight creative compounds, whereby the author applies her hypotheses to actual data. The position she wishes to defend, developed theoretically in part I, has three main tenets: firstly, that the processes involved in forming creative compounds exhibit systematicity, so that the compounds can be categorised according to which process acts on which part of the compound; secondly, that the metaphors and metonymies involved in forming creative compounds are established and pervasive throughout the language; and thirdly, that the semantic relations between the constituents of creative compounds are the same as those between the elements of compounds generally. Chapters 5–8 are organised on the basis of the first of these tenets: chapter 5 looks at compounds where either the modifier, the other constituent (referred to as the profile determinant), or both, are metaphorical; chapter 6 deals with combinations where the relation between the elements is metaphorical; chapter 7 discusses compounds that involve metonymy in any of these ways; and chapter 8 covers those with elements of both metaphor and metonymy. Finally, chapter 9 summarises the patterns found, hypothesises about why people form creative compounds, discusses degrees of creativity and restates the thesis.

The book's strength lies in the extent to which the examples analysed illustrate the author's claims. Benczes succeeds in demonstrating that creative compounds can be classified according to the element(s) involved in conceptual metaphor or metonymy and that they can be analysed in terms of patterns found elsewhere in compounding and in English more generally. For example, in armchair, although the modifier, arm, is metaphorical, the relation between arm and chair is the same part–whole relation found in nonmetaphorical compounds such as corridor train or picture book. Similarly, in car bait (‘a vehicle, monitored by the police, that is used to tempt a car thief into stealing it’), where the relation between the elements is metaphorical in that the car resembles bait, the conceptual metaphor is a version of PEOPLE ARE ANIMALS which is commonplace, as exemplified by expressions such as play possum, eat like a pig, rat on someone, etc. For many of the examples analysed in the book, discussion in the text is supplemented by a diagram illustrating how components of different mental spaces might map onto each other in producing or processing the compound in question. Although some of these seem slightly superfluous, they certainly do make the relevant points crystal clear. Finally, the book has good back material: particularly useful are the appendix, listing the compounds in the sample with their sources and meanings, and the separate index of the conceptual metaphors and metonymies mentioned.

At a superficial level, the book is easy to read, but problems emerge when trying to engage with the text in more depth. There are quite a number of typographical errors, some of which impede understanding, as well as an occasional tendency to confuse hypothesis with fact. In addition, there is the considerable difficulty of deciphering the terminology: although the author comments on the terminological confusion in the field, she does little to improve the situation. For example, Benczes introduces the term ‘creative compound’ on p. 6 to refer to ‘metaphorical and/or metonymical noun–noun combinations’ and, in a footnote, adds that metaphor and metonymy can turn up in endocentric as well as exocentric compounds. However, at various other points, creativity seems to be identified with exocentricity, e.g. ‘the main difference between endocentric compounds such as apple tree and exocentric compounds such as hammerhead is not transparency of meaning, but creativity’ (p. 184). Reading the book for the first time, I found that I was constantly having to crossreference backwards and forwards to check whether I had misunderstood the term and, if so, on which occasion. Furthermore, the term itself is perhaps an unfortunate choice since ‘creativity’ has already been coined with a variety of meanings in the morphological literature (Bauer Reference Bauer2001: 62ff.). In particular, it has sometimes been used in contrast to productivity, but Benczes seems to use it more like a synonym of productivity, e.g. ‘creative compounds are based upon productive patterns’ (p. 190). It would have helped me very much to have these and other key terms more clearly defined at first occurrence and consistently differentiated throughout the book, perhaps even backed up by a short glossary.

A possibly more serious problem lies in the selection and analysis of the example compounds. Although the author explains that her aim was to provide a ‘heterogeneous database’ (p. 11), there is no description of how this was actually achieved, other than by using a variety of sources, so the reader has no real idea of whether the sample is representative and, if so, of what. Not all of the logically possible types are included in the proposed taxonomy, but it is not clear whether the omissions are intended to suggest that those types rarely if ever occur. Furthermore, some of the analyses seem implausible or frankly mistaken, even allowing for the fact that, as Benczes points out (p. 140), the method is inherently subjective. For example, the suggestion that belly button involves mapping the image of an upper garment onto the image of the upper body seems unnecessarily to complicate the matter; the navel is metaphorically a button simply because of its size and shape. Another unconvincing analysis is that of beanpole family (‘a family whose living members come from many generations, but with few members in each generation’), where the author fails to realise that the beanpole element is a metaphor not for the family members but for the family tree. In other cases, the classification seems inconsistent. For example, although armchair is listed simply as having a ‘metaphor-based modifier’, other part–whole compounds, such as pendulum clock, are categorised as having a ‘metonymy-based relation between the constituents’. By this reasoning, it seems that armchair should actually fall into the class ‘metonymy-based semantic relationship and metaphor-based modifier’ – but this class is inexplicably absent from the taxonomy, even though other composite classes are included.

Overall, the book will be of interest to those working in the field of compound semantics. There is undoubtedly some merit in trying to develop an inclusive theory that accounts more comprehensively for the rich variety of combinations found in the language. Furthermore, cognitive linguistics does seem to be a promising approach to these constructions, which represent conceptual combination in the absence of an elaborated syntactic structure. Given all this, it is a shame that a certain lack of attention to detail makes the book a less satisfying read than it would otherwise have been.