Since metaphor was turned into a figure of thought, its conceptual analysis has expanded across a great many disciplines (e.g. Gibbs Reference Gibbs2008). One original motivation of the radical reconceptualization of metaphor as a figure of thought, however, came from linguistics, with the observation that many lexical items were polysemous in systematically metaphorical ways. For instance, see, view, viewpoint, point of view, and so on all display ‘vision’ senses as well as ‘understanding’ senses, which led cognitive linguists to postulate an underlying metaphor in thought labelled understanding is seeing. What is more, one explanation of these ubiquitous patterns came from the claim that synchronic polysemy reflects diachronic developments and language change, to the effect that non-metaphorical senses, such as the ‘vision’ senses, pre-dated and motivated the emergence of metaphorical senses, such as the ‘understanding’ senses. This motivation was then in turn explained by the anthropological assumption that personal embodied experience of our concrete environment is basic and prior and provides a model for engaging with more abstract and unfamiliar or complex experiences, such as our psychology, relationships, time and so on. This view of metaphor has been uncontroversial in cognitive linguistics at least since the publication of Eve Sweetser's (Reference Sweetser1990) From etymology to pragmatics.

Kathryn Allan's Metaphor and metonymy: A diachronic approach is a new addition to this tradition. It presents what could be termed a case study of the way in which English-language terms for the domain of ‘intelligence’ (including its opposite ‘stupidity’) have arisen over time via metaphorical and metonymic mappings. Its inclusion of metonymy is a reflection of developments in cognitive linguistics in the 1990s, when work on metaphor could be used as a springboard for developing related views of metonymy as a figure of thought. The distinction, relation and interaction between metaphor and metonymy formed the focus of new theory formation concerning the structure and function of conceptual mappings, and it was realized that it is not just diachronic metaphor which can account for contemporary synchronic polysemy but that diachronic metonymy may play an equally important role. Allan is therefore quite justified in adopting this broader scope, which raises interesting questions about the division of labour between metaphor and metonymy as distinct processes in language change.

This academic background is usefully evoked in chapter 1, which offers an introduction to the book. The difference between metaphor and metonymy is downplayed somewhat, however, the author warning the reader that she is ‘not concerned with a narrow classification of what constitutes metaphor and metonymy so much as with an exploration of the kinds of factors that can motivate mappings of various kinds’ (p. 2). She adds:

Her book is well placed, then, in this series of Publications of the Philological Society.

The introduction continues with an adequate account of the complexities of metaphor research in contemporary linguistics. It sketches the main points of the cognitive-linguistic approach and the way it branched out from metaphor to metonymy. It then moves on to the relation with a corpus-linguistic approach and to the method of the present study, which is based on the Historical thesaurus of English and the Oxford English dictionary. A clear picture is presented of the way the data have been collected and analysed, as well as a preview of the three main groups of lexical entries comprising the intelligence corpus: terms belonging to the lexical fields of the senses, density and animals.

Chapter 2 deals with the senses as one conceptual domain that can act as a source for the metaphorical motivation of lexical meanings in the target domain of intelligence. The data comprise 157 entries relating to vision, such as wise, redewise, sightly (including 34 cases having to do with light, such as dimwit and opacity), 32 items belonging to touch (e.g. conceitful, apprehensive, nimble), 12 entries relating to taste (e.g. insipid, salted), and 6 entries pertaining to hearing (including deaf). In all, these 204 ‘sense’ words make up almost 19 per cent of the complete ‘intelligence’ corpus. The bulk (141) of these 204 ‘sense’ entries are positive in the sense that they signify cleverness, not stupidity.

The motivation of the assumed metaphorical mappings from sense to intelligence is discussed in a separate section, dealing with each of the senses in turn. Arguments are related to the discussion by Sweetser (Reference Sweetser1990), but also to plain common sense. For vision, Allan draws on the conventionalized conceptual metaphor in thought knowing/understanding is seeing. This is an example of the more general position that

I have highlighted a number of expressions in this view because they raise a fundamental question about the status of the mapping between seeing and intelligence as metaphorical: when physical experience cannot be separated from conceptualization, it is contiguous to it, and part of the same conceptual domain. Metaphor, however, is defined in cognitive linguistics as invoking a separate conceptual domain that needs a cross-domain mapping. The same problem holds for the situation where humans have access to knowledge and understanding through the senses; and when, as a result, process and result are linked, we are looking at a classic process-for-result metonymy rather than some metaphor, which is based on a cross-domain mapping in which there are non-literal correspondences between two conceptual domains. In other words, the question arises why the senses are taken as a conceptual source domain in a metaphorical mapping to intelligence, and not as a conceptual domain in a metonymic one.

This question is also addressed by the author, in a later section in the same chapter. Before, Allan discusses data from Proto-Indo-European in an attempt to check the general validity of the claim that the senses are used ‘metaphorically’ in order to structure intelligence. Her motivation here is to critically engage with the paradigmatic work by Sweetser, who has claimed that ‘Vision verbs commonly develop abstract senses of mental activity’ (1990: 33). According to Allan, however, Sweetser has been selective in her collection and analysis of the data. In her own analysis, Allan finds that the abstract and concrete meanings often co-occur, reflecting the generally perceived contiguous connection between the senses and intelligence mentioned above instead of some sort of experiential and conceptual priority of the senses which could then function as a more basic source domain for the mapping of conceptual structures onto the target domain of intelligence. Her conclusion to this section is rather damaging to the classic cognitive-linguistic position:

This raises a problem which is addressed in the next section, in which Allan invokes a cognitive-linguistic proposal by Chris Johnson (Reference Johnson, Hiraga, Sinha and Wilcox1999). He has suggested that children's acquisition of metaphorical meanings of sense-intellection terms such as see does not proceed via metaphorical mapping from some prior concrete to some later abstract sense. Instead, he argues, acquisition of metaphorical ‘intellect’ senses works via their differentiation from concrete ‘sense’ meaning in a stage which follows upon a more basic stage in which the two types of senses were still conflated, in one monosemous lexical unit. If this is how word senses in acquisition arise, Allan argues, it might also account for her alternative findings to Sweetser's original picture, and it leads her to search for more evidence for this differentiation hypothesis in non-Indo-European languages. She consults sources on Proto-Afroasiatic and on Austronesian and comes to the conclusion that conflation between concrete senses and abstract intellect seems to be a general pattern, indeed, vision occupying a privileged position in that it is the clearest case. As a result, Allan resurrects the old parallel between ontogeny and phylogeny as an interesting point of view for further research into the conceptual connection between the senses and the intellect, concluding that the connection between them ‘must be re-evaluated’ (p. 64).

In discussing where this revaluation might go, however, Allan does not want to go as far as to decide that we are dealing with a case of metonymy instead of metaphor here – which would place her in a different camp than Sweetser. Instead, she maintains that the discussion in cognitive linguistics about the nature and function of different types of mappings, within and between conceptual domains, is unresolved and needs more attention. This is true. Yet some pointers to different groups of cases and their implications on the basis of her own data might have been useful, for both historical linguists and cognitive semanticists, who typically do not have access to this particular historical point of view.

This is particularly important since the next chapter presents a data set, from the domain of ‘density’, whose motivation ‘provides an interesting parallel to the senses data, since the mappings involved in each are motivated by metonymy’ (p. 89). This comment is not a slip but returns in two other places in the chapter: ‘and in this sense it can be compared to the senses mapping explored in the previous chapter’ (p. 93); ‘Although, as with the senses mapping, the connection between density and intelligence is motivated by metonymy, . . .’ (p. 94). Where in chapter 2 the author was not prepared to shift the analysis from a metaphorical to a metonymic footing, this has occurred implicitly in chapter 3. The demarcation and interaction of the two types of mapping that are the topic of the book is obfuscated. Moreover, this shift in position about the analysis of the senses entails another critique of the position of Sweetser: it does not just question the alleged priority of the non-metaphorical senses, as was concluded at the end of chapter 2, but also their non-metaphorical status. Given the academic tradition within which the book is positioned from the start, I would have liked to see a more explicit engagement with these points.

Chapter 3 includes three main subgroups under the rubric of density: 34 terms for wood (e.g. compounds from block, such as blockhead), 12 for earth (e.g. derivations from clod, such as clod-head and mud-headed) and 16 for food (e.g. words deriving from meat, beef and mutton: for instance, beef-witted, beef-brained), with 18 general (thick, including thick-skulled) as well as 9 miscellaneous terms completing the data set. All of these cases share the following characteristics: they conceptualize intelligence in terms that are more common, familiar, and somewhat worthless (or without much value) and formless. Additional interpretations are then needed to see why as thick as two planks makes sense: if one's skull is as thick as that, it may be hard to get some sense into a person. The author offers interesting interpretations here of specific sets of cases, aiming to show that ‘attempting case studies of this kind that focus on particular mappings, either with a source or target concept as a starting point, gives the opportunity to gain some sense of the complexities involved in metaphor’ (p. 117).

A third source domain frequently used to conceptualize intelligence and stupidity is the one of animals, which is the subject of chapter 4. Here, too, the point is to show that metaphorical mappings onto one of the core characteristics of people, intelligence, are not just simple projections from self-evident properties of animals, but that additional conceptual work as well as cultural background knowledge are at play. This is then shown to hold for all major groups within the data, including terms related to mammals (39), birds (36), insects (15) and fish (9).

The descending order in frequencies from mammals through birds and insects to fish is argued to reflect an underlying principle by which humans tend to compare themselves to those animals that are most like them. In a section invoking the Great Chain of Being and its use by Lakoff & Turner (Reference Lakoff and Turner1989), Allan asserts that ‘The association of people with animals in general tends to be derogatory, and often indicates some quality perceived as “less than human”’ (p. 138). This is why most animal metaphors tend to be derogatory, as when sheep and donkeys are used to say something about somebody's lack of intelligence.

This argument sits well in a section that makes connections with other cultural influences on language change, such as the medieval tradition of the bestiary, the classics and the Bible. In that context, speculations are presented about a possible rise in animal metaphorizing over the centuries, which is supported by the relatively more recent dates of these metaphorical mappings of intelligence than for the other two groups. The dates themselves, however, turn out to be somewhat problematic, which makes the author duly cautious about her suggestions.

The latter part of the chapter then presents an inventory of the particular animals found. Detailed discussions of specific cases go into the varied motivations for using, for instance, donkeys and asses, sheep and rams for the metaphorical conceptualization mostly of stupidity. Most of these present acceptable interpretations along the lines of common sense, folklore and cultural history.

Chapter 5 offers a brief conclusion to the study. It presents a clear summary of the differences between the three groups of source domains, and repeats the idea that metaphor and metonymy may be hard to distinguish because they might be located on one continuum. Methodological reflections are then offered on the various approaches informing the study, including the issues of a corpus, diachrony and interdisciplinarity. This conclusion is followed by an extensive appendix offering all of the data that are part of the intelligence corpus, including lexical entries from other source domains than the three most popular ones analysed in this book.

The value of this study lies in its detailed presentation of specific cases in the context of similar and related cases. What is on offer here is a goldmine for historically interested metaphor scholars and metaphorically minded philologists. However, I do not feel that the author has succeeded in giving ‘an account of some of the factors that can be involved in the motivation for mappings of different kinds of metaphor and metonymy’ (p.180), because that would have required a more ambitious and distanced approach to some of the fundamental theoretical issues in this area. The conclusion does not offer a summary of this more general account either. This is particularly the case for the differentiation between metaphor and metonymy themselves, which might have afforded a solid starting point for becoming more precise about the way in which various factors relate to these two distinct kinds of mappings, and why. It would also have given some interesting fuel to the more general debate about what counts as a metaphorical mapping and why. In that respect, the title of the book might have been better if it had profiled the book's major focus on the figurative conceptualization of intelligence in the history of the English language.