TY - JOUR
AU - Wojciechowska, Sylwia
AB - Looking at the title of the book, the readers of the IJL may dismiss the publication as one aimed at cognitive linguists (semanticists). The influence of cognitive linguistics in lexicography, however, has recently been growing, as evidenced by van der Meer (1999), Moon (2004), Geeraerts (2007), Adamska-Sałaciak (2008), Atkins and Rundell (2008), Kövecses and Csábi (2014) and Ostermann (2015), the last of whom coined the term cognitive lexicography. The cognitive perspective is also adopted in some learners’ dictionaries, for example in the Macmillan Dictionary in the form of metaphor boxes, which explain the conceptual processes underlying metaphorical expressions. (Meta)lexicographers may also find the book useful in word sense disambiguation, one of the more challenging tasks in the dictionary making process (e.g. Atkins and Rundell 2008: 269-315). Admittedly, the volume is concerned exclusively with mining metaphors and metonymies in electronic corpora. It should, nonetheless, be noted that conventional metaphorical and metonymic meanings are covered in dictionaries, often occupying the first senses in dictionary entries, due to the high corpus frequency of these meanings. Metaphors and metonymies began to be widely recognised as fundamental in our language and cognition after the publication of Lakoff and Johnson’s seminal work Metaphors we Live by (1980). Since then numerous works have appeared on the topic in the field of cognitive linguistics, most of them dealing with theoretical issues. This volume stands out for its practical approach to the Cognitive Theory of Metaphor and Metonymy (henceforth CTMM), reporting on a number of empirical studies and electronic databases. Although Lakoff and Johnson (1980) drew the attention of the linguistic community to metonymy as a category distinct from metaphor, publications concerning the latter still considerably outnumber the ones dealing with the former. Notwithstanding its plentiful merits, which will be presented below, Metaphor and Metonymy in the Digital Age is no exception in this respect, as metonymy takes up much less space in it compared to metaphor. The book is an anthology, an edited collection of nine chapters written by different authors. Earlier versions of these chapters were presented at the conference and roundtable forum Building Figurative Language Repositories: Methods, Risks, and Challenges, held in Zagreb in 2016. It also includes an introduction, a metonymy and metaphor index, an author index and a subject index. Among the authors are some prominent linguists who have contributed substantially to the development of the CTMM, but also psychologists and computational linguists, which results in the volume’s interdisciplinary perspective. This is not just a random collection of chapters dealing with metaphor and metonymy, as all the contributions focus on the same aspect of the CTMM, building electronic repositories of figurative language. There are also frequent cross-references between the chapters. This is the first collection of papers by various authors entirely devoted to the trending topic of retrieving and annotating metaphors and metonymies in digital databases. The volume is divided into two parts: Part I (the first five chapters) presents new methods for metaphor and metonymy identification in natural language corpora, and the most recent digital resources as the fruits of this labour, Part II (the next four chapters) discusses the risks and challenges that accompany building and using such repositories. It should be stressed that it is not a book for a beginner in the CTMM, as none of the chapters define the terms metaphor and metonymy or other important notions in the theory such as domain and mapping. Similarly, familiarity with cognitive linguistic terminology is presupposed. Not only does the anthology require considerable knowledge of the CTMM and cognitive linguistic issues, but also presumes that computational linguistics, including corpus analysis and annotation, is not a new territory for the reader. The Introduction, written by two of the editors of the anthology, Bolognesi and Despot, catches the eye with its intriguing title Fantastic metaphors and where to find them. Choosing this title, the authors draw an analogy between the plot of the fantasy film Fantastic beasts and where to find them and the framing of the book, with metaphors compared to exotic creatures that scholars try to ‘recapture and put into the zoo – a figurative language repository or a certain theoretical framework’ (p. 1). Apart from giving an overview of the contents of the volume, the Introduction reports on metaphor identification procedures and metaphor databases that preceded the ones described in the volume, among them MIP and MIP VU (Steen et al. 2010), which gave rise to the VU Amsterdam Metaphor Corpus, the largest available corpus annotated for metaphorical uses. In a nutshell, the MIP VU method points to topical discontinuity or incongruity in discourse as the defining features of metaphor. In fact, MIP VU is adopted as the procedure for identifying metaphors by many authors in this volume, particularly by Despot et al., Bolognesi et al. and Kövecses et al. The projects presented in Part I are crucial for computational linguists for whom automatic detection of figurative language has always been a challenge. Lexicographers have also long complained about lack of methods for automatic sense differentiation in electronic corpora. Part I showcases two groundbreaking metaphor identification endeavours using Natural Language Processing (NLP) technology, MetaNet, based in Berkeley, California and its sister project and Croatian counterpart, MetaNet.HR. One may only wonder why the accounts of these two closely related projects are divided by other chapters, MetaNet covered in Chapter 1, and MetaNet.HR, in Chapter 5. In Chapter 1 by Sweetser, David and Stickles, we learn that it is no coincidence that the name MetaNet bears resemblance to FrameNet, having been largely inspired by it. MetaNet consists of two major components, a conceptual metaphor repository with ontologically organized frames, and a metaphor identification system that makes use of it to automatically extract metaphorical uses from large-scale text corpora. It is a large multilingual database providing examples from English, Spanish, Farsi and Russian which demonstrate the system’s capability of detecting the same metaphor cross-linguistically. A weakness of this system, observed by the authors themselves, is its dependence on the presence of both source domain and target domain lexical elements in the text; otherwise the system would not recognise a certain expression as metaphorical. MetaNet.HR, headed by Kristina Despot, is a metaphor repository of the Croatian language. Due to a great deal of manual annotation, unlike MetaNet, it is capable of detecting metaphorical linguistic expressions in which the source domain lexical element is not present. However, this reliance on the annotator’s knowledge and intuition, may at the same time be seen as the resource’s weakness. The project deserves credit for being the first step towards connecting potential repositories of different languages, which could then be searched and compared. Despot et al. remark that MetaNet.HR can serve as a next-generation electronic dictionary, as ‘[i]t aims at being a language specific knowledge database that schematically represents how our conceptual system is organised’ (p. 143). This, however, may sound a bit far-fetched, at least at the present moment. The last chapter of the volume (Chapter 9) relates to MetaNet (Chapter 1) and, in particular, MetaNet.HR (Chapter 5). Brdar, Brdar-Szabo and Perak, the first and the last co-authors of Chapter 5, describe in it the ontology built for MetaNet.HR, which is called the Ontological Model of Lexical Concepts and Construction. This model is to serve as a tertium comparationis for the analysis of metaphor in a multilingual repository, which MetaNet.HR aims to become. The chapter belongs to Part II of the volume, and thus focuses on the problems that concern creating figurative language databases. The authors warn against building a repository based on a metaphorical system of a single language, for it may require constant modifications to be compared with repositories of other languages. By pointing to some cross-linguistic differences, they demonstrate the usefulness of a universal ontological system. Of the five digital repositories presented in Part I, there is only one that deals with metonymy. In fact, it is solely devoted to metonymy, as the name suggests, the Córdoba Metonymy Database. As for the remaining four resources, three focus on metaphor, and only MetaNet.HR is supposed to include metonymy, as in the abstract, Despot et al. announce that ‘[t]his database lists conceptual metaphors and metonymies’ (p. 123). In the conclusion, however, the authors admit that ‘[a]t the current stage of the project, the types of figurative links between frames are primarily metaphorical and marginally metonymic’ (p. 143). Given this, the contribution in Chapter 2 about the development of the Córdoba Metonymy Database is more than welcome, especially as it written by the leader of the project, and an acknowledged expert in the CTMM, Antonio Barcelona. The Córdoba Metonymy Database is designed to serve as a tool for systematic examination of conceptual metonymy mostly in English and Spanish authentic discourse. Barcelona gives a meticulous account of both the theory and methodology behind the project. He also briefly discusses the entry model applied in the metonymy database. A substantial portion of the chapter is devoted to the description of the generic classification in the Hierarchical level field of the entry model. The author defends the traditional tripartite typology of metonymy into whole for part, part for whole, and part for part. He presents the proposals that reject the part for part category, put forward by prominent researchers in the CTMM, among them Ruiz de Mendoza, and Panther and Thornburg, and argues against them. In fact, both approaches, for and against the existence of part for part metonymies, are defensible, though not devoid of indeterminacies, and as such, a matter of individual preference. The discussion would clearly benefit from more corpus examples, rather than the clichéd ones: The ham sandwich is waiting for his bill or Wall Street is in panic. A definite advantage of the anthology is that it does not only include figurative language databases annotated by experts in the field, such as the three mentioned above (MetaNet, MetaNet.HR and the Córdoba Metonymy Database), but it also features projects which rely on crowdsourcing for data collection. Such innovatory approaches are described in Chapters 3 and 4. The method for harvesting affective metaphors demonstrated by Veale in Chapter 3 distinguishes itself from the other methods presented in the book by applying computational models for the interpretation and elaboration of metaphors. This original approach to metaphor study exploits Web n-grams to retrieve metaphors, and uses algorithms for manipulating this knowledge in order to understand and generate novel deliberate metaphors. Veale presents the Web service Metaphor Magnet that provides this functionality, and additionally, is capable of retrieving affective metaphors in proper names, for example Google is another Microsoft. It also allows humans to interact with machine-generated metaphors. In the project, once metaphorical utterances were generated by Metaphor Magnet, human raters, non-experts, were asked to judge them for comprehensibility, novelty and the overall affect. The other digital metaphor repository which used crowdsourcing is VisMet, presented in Chapter 4 by Bolognesi et al. Non-experts on metaphors were exposed to a selection of visual metaphors in the VisMet Corpus 1.0 and asked to annotate them. Focusing on images rather than words, this publication is presumably not of much interest to the readers of IJL, but it is definitely worth mentioning for two reasons. Firstly, VisMet 1.0 is the first and still the only corpus of visual metaphors publicly released online. Secondly, the repository makes use of social tagging, an innovative procedure, increasingly popular among academic researchers examining large-scale collections of unstructured linguistic data for semantic information. The information retrieved through crowdsourced tags is then analysed and manually classified. By the extensive use of social tagging, Bolognesi et al. support the idea of experts working in tandem with laymen in visual metaphor interpretation. Part II of the anthology provides a critical perspective on digital repositories of figurative language. Apart from the already discussed Brdar, Brdar-Szabo and Perak (Chapter 9), there are three other contributions. Chapter 6 written by Kövecses et al. poses a challenge to the hegemony of corpus-based methods in the study of metaphor. Prepared under the leadership of Kövecses, a prominent contributor to the development of the CTMM, the chapter is definitely worth paying attention to. The authors provide an alternative approach to the ones presented in Part I, based on combining the lexical and corpus-based methods. The lexical approach has been questioned by many metaphor researchers as intuitive and methodologically unsound. Kövecses et al. defend this approach by proposing an updated version of it, which is based on dictionary information (e.g. synonyms, idioms, collocations) about lexical items related to a certain topic. One may wonder, however, about the choice of the lexicographic resources used. It seems to be a rather unsystematic selection of dictionaries for both learners and native speakers. The authors set out to find out if the lexical approach can also be considered a corpus-based method, and used in place of the latter, as dictionaries are corpus-based. By comparing metaphors and metonymies retrieved by both approaches in their study of the concept of surprise, Kövecses et al. provide a negative answer to this question. Admittedly, a considerable number of metaphors and metonymies identified by the two methods were the same. Nevertheless, the corpus-linguistic approach resulted in finding more metaphors than the lexical one. Moreover, due to the nature of dictionaries, the lexical method is capable of identifying only conventionalised metaphors, and cannot measure frequency of usage. The authors convincingly argue that both methods are needed ‘for a fuller account of metaphorical patterns’ (p. 171), stressing the role of the lexical approach in identifying metaphors that do not involve an explicit target term, but relate to the same concept, here of surprise. In Chapter 7, Panther and Thornburg, renowned for their contribution to the theory of metonymy from a pragmatic perspective, also challenge the corpus-based approach. They argue that the target meanings of indirect speech acts are ‘not coded, and therefore not compositionally computable’ (p. 176), but ‘accessible only through inference, in particular, metonymic inference’ (p. 196). The authors support their claim by means of the so-called hedged performatives, a category of speech acts in which the performative verb is ‘hedged’ by, for instance, emotive words or expressions, or modals, as in I can offer you a month’s wages. On the basis of corpus data, Panther and Thornburg demonstrate that in some cases the use of, for example modal hedges, can or must, does not affect the illocutionary meaning conveyed by the performative verb, resulting in the metonymy VIRTUALITY → ACTUALITY. In other cases, however, modal hedges cause the cancellation of the illocutionary meaning denoted by the performative verb. Consequently, currently only humans, not machines, are capable of distinguishing between such cases. Chapter 8 by Devylder, suggestively entitled Mereology in the flesh, also deals with metonymy, focusing on part-whole relations. It is another theoretical contribution to the improvement of data annotation in figurative language databases. Devylder proposes adjustments that can be applied in the coding scheme of part-whole figurative expressions cross-linguistically by distinguishing between a range of relations within the field of mereology and supporting them with examples from European and non-European languages. The author rightly differentiates between ‘part-of’ relations and ‘kind-of’ relations, partonomy and taxonomy respectively, which tend to be confused in the metonymy literature, hence in the coding schemes of figurative language databases. He further distinguishes between partonomy and three other relations: possession, containment and contiguity. While the distinctions may be controversial in view of the traditional classifications of metonymic relations, they are well-documented. Moreover, it is unarguable that establishing clear categorisation criteria will help to keep inter-rater disagreement in coding to a minimum. The book can definitely be recommended not only to cognitive linguists but also to lexicographers, especially as increasing cooperation can be observed between both disciplines. Analysing figurative senses of words, some authors in this book use dictionaries as the sources of information, especially Veale and Kövecses et al. The former also refers to the well-known lexicographer Patrick Hanks and his Theory of Norms and Exploitations (Hanks 2013). Other lexical resources are also employed, for example Sketch Engine and FrameNet. In fact, MetaNet.HR is described as the Croatian version of FrameNet, as it defines semantic roles within certain frames. Finally, lexicographers might find the presented semi-automated and fully automated methods of retrieving figurative senses from corpora, for instance, the ones used in MetaNet and Metaphor Magnet, applicable in the process of word sense disambiguation. Metonymy researchers, however, might be a little disappointed due to the marginalisation of metonymy in the volume despite the title which would suggest that both phenomena are treated on an equal footing. As has already been observed, of the five digital repositories of figurative language presented in Part I, only one deals with metonymy, the Córdoba Metonymy Database. The theoretical Part II offers a more even distribution with two chapters devoted to metonymy identification, by Panther and Thornburg, and Devylder. Kövecses et al. also examines metonymies, but the sections on metonymies are much shorter than the ones on metaphor. Despite this objection, the anthology deserves praise for a number of reasons. It contributes to the development of metaphor research in computer science, which is crucial as metaphor studies still remain at the periphery of NLP (Veale p. 78). Another obvious advantage of the book is the innovative methodology with a variety of procedures for mining metaphor and metonymy, both manual and automated, top-down and bottom-up, performed by experts and non-experts alike. The volume demonstrates that far from being mutually exclusive, these methods should work together to create comprehensive repositories of figurative language. Other strong points include: the diversity of languages analysed, rich statistical data, a multitude of illustrative examples, and a thorough theoretical background, providing an extensive bibliography of metaphor and metonymy. It is hoped that the anthology will stimulate further work on cross-linguistic digital repositories of figurative language. References Adamska-Sałaciak A. 2008 . ‘Prepositions in Dictionaries for Foreign Learners: A Cognitive Linguistic Look.’ In Bernal E. , DeCesaris J. (eds). Proceedings of the XIII EURALEX International Congress, EURALEX 2008 , Barcelona : Universitat Pompeu Fabra , 1477 – 1485 . Google Scholar Google Preview OpenURL Placeholder Text WorldCat COPAC Atkins B. T. S. , Rundell M. . 2008 . The Oxford Guide to Practical Lexicography . New York : OUP . Google Scholar Google Preview OpenURL Placeholder Text WorldCat COPAC Geeraerts D. 2007 . ‘Lexicography.’ In Geeraerts D. , Cuyckens H. (eds). The Oxford Handbook of Cognitive Linguistics , Oxford : OUP , 1160 – 1174 . Google Scholar Google Preview OpenURL Placeholder Text WorldCat COPAC Hanks P. 2013 . Lexical Analysis: Norms and Exploitations . Cambridge, MA : MIT Press . Google Scholar Crossref Search ADS Google Preview WorldCat COPAC Kövecses Z. , Csábi S. . 2014 . ‘ Lexicography and Cognitive Linguistics.’ Revista Española de Lingüística Aplicada /Spanish Journal of Applied Linguistics 27 . 1 : 118 – 139 . Google Scholar OpenURL Placeholder Text WorldCat Lakoff G. , Johnson M. . 1980 . Metaphors we Live by . Chicago – London : The University of Chicago Press . Google Scholar Google Preview OpenURL Placeholder Text WorldCat COPAC Meer G. van der. 1999 . ‘ Metaphors and Dictionaries: The Morass of Meaning, or How to Get Two Ideas for One.’ International Journal of Lexicography 12 . 3 : 195 – 208 . Google Scholar Crossref Search ADS WorldCat Moon R. 2004 . ‘ On Specifying Metaphor: An Idea and its Implementation.’ International Journal of Lexicography 17 . 2 : 195 – 222 . Google Scholar Crossref Search ADS WorldCat Ostermann C. 2015 . Cognitive Lexicography: A New Approach to Lexicography Making Use of Cognitive Semantics. (Lexicographica. Series Maior 149.) Berlin: de Gruyter. Steen G. J. , Dorst A. G. , Herrmann J. B. , Kaal A. A. , Krennmayr T. , Pasma T. . 2010 . A Method for Linguistic Metaphor Identification. From MIP to MIPVU. Amsterdam: John Benjamins Publishing Company. © 2020 Oxford University Press. All rights reserved. For permissions, please email: journals.permissions@oup.com This article is published and distributed under the terms of the Oxford University Press, Standard Journals Publication Model (https://academic.oup.com/journals/pages/open_access/funder_policies/chorus/standard_publication_model)
TI - Marianna Bolognesi, Mario Brdar and Kristina Despot (eds). 2019. Metaphor and Metonymy in the Digital Age: Theory and Methods for Building Repositories of Figurative Language.
JF - International Journal of Lexicography
DO - 10.1093/ijl/ecaa023
DA - 2021-02-16
UR - https://www.deepdyve.com/lp/oxford-university-press/marianna-bolognesi-mario-brdar-and-kristina-despot-eds-2019-metaphor-PDic563a3a
SP - 151
EP - 155
VL - 34
IS - 1
DP - DeepDyve
ER -