Understanding Metaphor through Corpora

A Case Study of Metaphors in Nineteenth Century Writing

Katie Patterson

• Publisher: Routledge
• Year: 2018
• Print ISBN: 9780815374978
• eBook ISBN: 9781351241076
• Edition: 1
• Topic: Languages & Linguistics
• Subtopic: Linguistics
• Index: Languages & Linguistics

This book introduces a unique methodology to the study of metaphor, integrating a corpus linguistic approach to explore the lexical, grammatical, semantic, and pragmatic characteristics of metaphoric instances of language. The volume questions the reliability of attempts to identify metaphor based on dichotomy and, drawing on data from a corpus of nineteenth-century writing, instead advocates for the notion that metaphoricity is context-dependent and fluid, in relation to the respective social and discourse contexts in which metaphors can be found. The book also applies Lexical Priming Theory to metaphoric language to suggest that our use of metaphor is due to unconscious behaviors, a counterpoint to perspectives that see metaphor use as part of the creative process. Taken as a whole, the volume calls for a deeper investigation of the complex web of meaning senses that contributes to our understanding of metaphor, making this key reading for students and researchers in corpus linguistics, metaphor studies, lexicography, semantics, and pragmatics.

1 Introduction

1.1 Metaphor and Corpus

For many years, metaphor research followed the standard assumption concerning the literal-metaphorical division. This dictated that metaphor was distinct from literal language, operating as a secondary or less ‘basic’ form of meaning. With the latest wave of research, metaphor has begun to take a more significant position in language research. It is acknowledged as being far more pervasive in its linguistic manifestations and its functions for example, and it is now largely accepted that metaphoric language operates along a cline or gradient, rather than in dichotomy to literal language. The last decade of metaphor research has begun to show this in particular: there has been a considerable rise in metaphor studies which rely on quantitative methods and empirical evidence. In fact, corpus linguistics is now a well-established methodology within the field of metaphor with a multitude of studies focusing on frequency, form and function (see Semino, 2017). This book is not original in that respect. What this book argues, however, is that corpus linguistic data has really been the propelling force for this change in our understanding. Furthermore, only by focusing on both quantitative and qualitative methods together can one hope to attain a more holistic approach to metaphor. With the rise in interest in metaphor comes a range of approaches, perspectives and theories. However, this influx has the potential to dilute our focus rather than expanding and enriching it. The argument for corpus linguistics as a method is that it works to unify a variety of approaches and theories by providing transferrable, empirical data.

This book provides a description of the lexical behaviour of metaphors within nineteenth century writing. The data is presented as three case studies, two of which provide a quantitative analysis semantic and structural patterning within the lexis. The third presents a qualitative discussion on other ways in which metaphor can manifest itself in language. The book draws on corpus methods and lexical analysis in order to provide data for the Lexical Priming theory (Hoey, 2005). Equally, the theory itself provides an explanation for the data: why we use language in particular structures, relationships, contexts and textual positions. Seeking out linguistic evidence is, I believe, the only way to begin to draw conclusions about the more abstract cognitive processes at work. The intention is to arrive at a better understanding of the phenomenon of metaphor: one that takes a lexical real-world approach whilst, at the same time, is able to offer an explanation for the cognitive processes that help prime us to recognise metaphor. Before outlining the specific aims of the current research, I want to begin this book by taking the briefest glimpse at how metaphor study has evolved to this enlightened empirical age of corpus methodology. The reason for doing this is to illustrate the fundamental theories that have, rather organically, led to up to this period, as well as bringing to light our need to seek evidence for more basic philosophical ideas.

In the first half of the twentieth century, there began a shift in the way philosophers viewed language. Wittgenstein was an important figure in this linguistic revolution; he argued that categories or concepts in the language should be built around “family resemblances,” rather than the necessary conditions of definitive labels which belonged to Aristotelian theories of meaning. This important claim, that categories have sets of resemblances and that these resemblances are not present in all cases, propelled other theories. Most notably, Rosch’s (1973) prototype theory furthered this idea that concepts, like language, are organic rather static, and can be described by their resemblance to a ‘best example’ or a prototype. Another crucial claim for metaphor born out of Wittgenstein’s philosophy was the idea that the meaning of a word or phrase is determined by the set of informal rules governing the use of the expression in social situations. As such, meaning can be interpreted as a consequence of our ability to follow these informal rules. Firth (1957) coined his contextual theory of meaning from these ideas, and Grice (1957, 1975) also developed these ideas with his observation that communication depends on conversational co-operation, and that language users also create meaning by exploiting these conventions at times. What these theories centre upon is the idea that meaning, metaphorical or not, can be seen as the sum of our relationship with language, and its relationship with the world.

Other theories began to influence metaphor research mid-way through the twentieth century. Most fundamental was the cognitive approach, which we will consider in Chapter 2. The roots of this influential approach grew out of ‘interaction theory,’ in which Richards (1936) put forward the view that metaphor is a pervasive phenomenon and that the functioning of metaphor is to be detected within the ordinary usage of language. Richards suggests that the underlying thought, the tenor, and the thought through which the tenor is expressed, the vehicle, interact together to create the metaphoricity. Thus metaphor is seen as a process of cognition, rather than just of expression. A breakthrough was made in Black’s (1962) development on this theory (influenced by Paul Ricoeur), which claims that metaphor takes place not only at the level of word but also at the level of the sentence. Crucially, one must rely on the context, both linguistic and situational, to decide whether a word is metaphorical or not.

Finally, towards the latter half of the twentieth century came conceptual metaphor theory which had a profound influence on studies of metaphor, and continues to presently. The theory, created by Lakoff and Johnson (1980) and heavily influenced by Reddy (1979), was also born out of the cognitive approach to linguistics, the fundamental assumption being that our most vital mechanisms of thought are entirely subconscious. Consequently, Lakoff and Johnson claim that when we come across a metaphor in spoken or written text, our understanding is not merely surface level, but is more deeply tied to the core of our world knowledge. This process of understanding draws on other mental abilities too, such as perception and reasoning. Thus metaphor is viewed not simply a characteristic of our language, but entrenched in our thoughts too. Whilst cognitive studies in metaphor have become fairly widespread since the introduction of Lakoff and Johnson’s theory, many linguists, including those working with corpora, reject the idea that every linguistic metaphor can be tied to a particular conceptual metaphor. Despite much criticism of the theory, however, Lakoff and Johnson made important claims about the way our mind organises concepts. It could be argued that their ideas helped provide the motivation for a shift to more lexically based, quantitative research, whether by seeking to provide support for, or to challenge, their pivotal cognitive claims.

A more lexical approach developed along these principles, emphasising the notion that metaphor is pervasive and fully embedded in our everyday language. This approach acknowledges that metaphoric meaning is intrinsic to our perception and understanding of the world, and just as importantly, that such metaphoric meaning is linguistic evidence of our social interaction and co-operation in the world. What is also tantamount to a lexical approach though, quite obviously, is the manifestation of metaphoricity in the lexicon. If a cognitive approach focuses on language in the mind, the lexical approach, specifically corpus linguistics, focuses on language in the world. This has led to important contributions which have refocused our approach to metaphor. Data has shown us that current systems of metaphor categorisation are unable to address the fuzzy, complex nature of metaphoric language with any definitive set of linguistic characteristics (see, for example Deignan, 2005; Partington, 2006; Philip, 2011). By providing the linguistic evidence for the ideas of Wittgenstein and other philosophers, we can argue with a high degree of certainty that metaphor is not static or definitive or objective. It exists in a variety of ways and means and, as a result, cannot be identified by sharply defined categories. Current theories, then, must address this argument in the light of the corpus data.

1.2 Lexical Priming

Evidence of linguistic patterning in corpus data must say something not only about our individual use of language, but about our use as a society (if our corpora are representative of a given society). Evidence of repetitive patterns in metaphoric language, then, say something about our collective use of metaphor. A metaphorical turn of phrase would not be understood without some shared principles. To explain this better, let us return full circle to Wittgenstein, who acknowledges the influence of society on our language use. Language in the world is the product of linguistic acts of individuals, but these individuals behave in accordance with their previously acquired knowledge. In turn, their acquired knowledge is the product of their encounters with external language. In other words, whilst every metaphor can be said to have a life of its own (Deignan, 2005), language is ultimately a social tool, used to communicate and be understood. It is a prevalent part of what it is to be human. Certain social norms or conventions must be obeyed in order to create meaningful utterances. Crucially, part of that social norm or convention referred to by Wittgenstein is built up from prior language use. This, I believe, should be seen as the key to understanding metaphoric meaning.

Prior context plays a fundamental role in our understanding of anything. Moreover, it creates a strong expectation about how we use language. This includes the sense in which a word or phrase will be used, and thus can determine the activation of a metaphoric sense. Prior context concerns a full range of lexical characteristics, both within and outside of the specific utterance. Clues are found from studying lexical associations (Deignan, 2005) and syntagmatic structures, which are often more tightly constrained in metaphor (Hanks, 2010) and even pragmatic inferences (Philip, 2011). Finding evidence for these conventional patterns amongst metaphoric language, one can draw conclusions about our expectations of those patterns. More than this though, using this evidence, we can claim something about our understanding of the concept ‘metaphor’. This is where cognitive and lexical approaches come together.

Hoey’s (2005) theory of Lexical Priming provides an explanation for the pervasiveness of the concepts collocation and colligation (see Sinclair, 1991), and accounts for our motivation to conform to expectations. Drawing on and expanding upon psycholinguistic literature (cf. Hoey, 2005: 8; Pace-Sigge, 2013: Chapter 2), Hoey’s theory claims that every time we encounter a word we subconsciously note the patterns this word tends to form with other words in certain contexts, so that, eventually, as a result of the cumulative effects of our encounters with this word, it becomes “part of our knowledge of a lexical item that it is used in certain combinations in certain kinds of text” (Hoey, 2005: 10). These patterns are manifest in grammar and lexis, but also in more secondary aspects, such as semantic association and pragmatic association. They are more prevalent than structured rules: they are encountered psychologically and created through repetition. When we re-use a lexical item, we are then likely to reproduce these combinations in their respective contexts in our own language production. Importantly for this research, these primings or expectations are dependent upon a community, genre and time, and have the ability to change.

Hoey’s theory is adopted as a theoretical tool for analysing metaphoric language. Within this book, metaphoric and non-metaphoric instances of a single lexical item will be analysed in order to determine how far the instances (and thus senses) avoid each other’s patterns of use and meaning. This in turn will determine the extent to which we as language users are primed to understand and recognise metaphoric senses as distinct from non-metaphoric, non-figurative senses. In this book metaphoric meaning is discussed from the perspective of the language users as much as of the text itself, and the findings contribute to the idea that metaphoricity is not inherent within the language. Rather, metaphoricity should be seen as a fluid concept, dependent on language users and their relationships and experiences with language, both individually and as a collective whole.

1.3 Aims of This Book

There are two main aims to this book. The first aim is to show how far the theory of Lexical Priming is applicable to metaphoric language. So far there has been little attention paid to figurative language and in how far priming can account for its usage. Similar research by Hoey (2005) and Tsiamita (2009) looked at polysemy, and found that two distinct senses of a word or item tend to avoid each other’s primings (as claimed in Hoey’s drinking problem hypothesis, 2005). In relation to a pervasive phenomenon such as metaphor, wherein analysis of metaphoric behaviour and subsequent identification of metaphoric language remains creatively ‘unrestricted’ and largely problematic, Hoey’s (2005) theory provides an explanation for what drives us as language users to identify such a phenomenon. The introduction of an extended theory involving our psychological associations with language offers an explanation for how we recognise conventional norms and creative exploitations in relation to metaphor.

The second aim of the book is to determine the extent to which corpus linguistic methods and lexical priming enrich our understanding of metaphor. Evidence of the theory brings to light reasons for why we use metaphoric language in particular ways, and similarly why we recognise metaphors in use. To summarise the main aims in the form of research questions, these are as follows:

1. 1 Can the theory of Lexical Priming account for how we recognise and understand metaphoric language?
2. 2 What can corpus linguistic methods and Hoey’s theory of lexical priming add to our current understanding of metaphor from a linguistic perspective?

This book sets out to explore these questions by means of descriptive case studies of how metaphor manifests itself in nineteenth century writing. The investigation is based on a corpus-driven lexical analysis of three target keywords,¹ identified in a 49-million-token corpus of nineteenth century writing assembled by the author. The keywords examined are cultivated, flame and grew.

By applying the Lexical Priming theory to metaphor, metaphor is explored from a psychologically-motivated perspective, whereby characteristics or patterns found amongst metaphoric instances of an item are the result of our expectations, or primings. This approach explains what other metaphor theories appear to have missed so far: namely that metaphoric uses of language, alongside their literal, non-figurative counterparts, must be discussed firstly, in relation to meaning as an extended unit, and secondly, as meaning existing within the language users and their collective mental lexicon. As metaphoric uses of a lexical item are shown to avoid the primings of the non-metaphoric uses of that same item (as has been shown to be the case with polysemy (by Hoey, 2005, and Tsiamita, 2009)), I argue that metaphoric senses have, to an extent, a fixed set of choices in terms of grammar and lexis.

1.4 Structure of This Book

Following on from this introduction, the second chapter focuses on both the categorisation and the identification of metaphoric language. Firstly, in 2.1, the traditional assumptions for what constitutes a metaphor in both thought and language is presented. Metaphor is discussed largely within a cognitive framework. I aim to trace the cognitive foundations which have influenced, and continue to influence, metaphor research today. The shift of viewpoint from metaphor traditionally seen as a dichotomy, to the now widely accepted cline theory will also be outlined here, along with notions of conventionality. These ideas lay the foundations for the current study. The second section (2.2) introduces the issue of metaphor identification, specifically from a lexical perspective. The focus here is on identifying the textual manifestations of metaphoricity (key terms such as collocation and colligation will be discussed in relation metaphor study). Finally, an outline of corpus linguistics and its advantages as a methodological approach to identifying and studying metaphoric language is presented. An account of some problems with both metaphor categorisation and identification methods is also outlined, preparing the way for the proposed method in the next chapter.

Chapter 3 then adopts a usage-based approach, to explore how these manifestations of metaphoric meaning help to shape and develop our thought, both consciously and unconsciously. It develops on from both the theoretical and methodological discussions presented in the previous chapter and adds a socio-linguistic dimension to this research. The emphasis here is on the point at which a novel and ‘convention-exploiting’ metaphor becomes re-used and even expected within a community. Hanks’ theory of norms and exploitations (2004) and Hoey’s Lexical Priming theory (2005) are offered up as alternatives to current approaches to identifying metaphoric characteristics within the language. We return to the notion of conventionality and discuss it in rela...

Table of contents

• Cover
• Title
• Copyright
• Dedication
• Contents
• List of Figures
• List of Tables
• Acknowledgements
• 1 Introduction
• 2 Metaphor: Theories and Approaches
• 3 A New Theory for Metaphor
• 4 Corpus Methodologies and Metaphor
• 5 Metaphor and Conventionality Part I: Semantic Evidence of Priming
• 6 Metaphor and Conventionality: Structural Evidence of Priming
• 7 Metaphor and Other Forms of Creativity: Polysemy, Simile, Metonymy and Their Relationship to Metaphor
• 8 Conclusion
• Index