4.5.1 Categorized Collections of Metaphors
4.5.2 Annotations of Specific Constructions or Target expressions
4.5.3 Full-Text Annotations
4.5.4 Summary of Annotated Data Available for Public or License-Based Use
4.5.5 Additional Annotated Materials
5 Knowledge Acquisition and Metaphor
5.1 WordNet and other Lexical Ontologies
5.2 Extracting Knowledge from Free Text
5.2.1 Simile Patterns
5.2.2 Categorization Patterns
5.2.3 Arbitrary Relations
5.3 Conceptual Metaphors
5.4 Summary
6 Statistical Approaches to Metaphor
6.1 Association Measures and Selectional Preferences
6.2 Supervised Classification
6.3 Clustering
6.4 The Topical Structure of Text
6.5 Vector Space Models
6.6 Concreteness
7 Applications of Metaphor Processing
7.1 Metaphor in Culture, Communication, and Education
7.1.1 Common Metaphors in Cultures and Discourse Communities
7.1.2 Metaphor as a Framing Device in Political Communication
7.1.3 Metaphor Comprehension and Use as a Skill
7.2 Creativity Applications in NLP
7.2.1 Creative Writing Tools
7.2.2 Creative Information Retrieval of Figurative Possibilities
7.2.3 Narrative Possibilities: From Metaphor to Metamorphosis
Preface
The aim of this book is to introduce metaphor research to the wider NLP community, and to survey the state-of-the-art in computational methods in a way that may also be helpful to those approaching metaphor from a perspective that is not principally informed by work in Artificial Intelligence (AI). We focus on the history, methods, and goals of past research into this fascinating phenomenon in the hope of making metaphor a more accessible topic of future research, thereby pushing it further up the NLP community’s wait-list. Our treatment will provide a condensed history of metaphor research that introduces the main theories of metaphor that survive, in one form or another, in contemporary analysis. Our coverage will include the main AI contributions to the field, which are modern attempts to give algorithmic form to views on metaphor that range from the ancient to the contemporary. And, just as contemporary AI research has taken on a distinctly web-colored hue, we shall explore the role of the Web in metaphor research, both as a source of data and as a computational platform for our metaphor-capable NLP systems. Computational linguistics and AI alike have each embraced statistical models as a means of improving robustness, exploiting rich veins of user data, and reducing a system’s dependence on hand-crafted knowledge and rules. Metaphor research offers no exception to this trend, and so our book will also explore the role of statistical approaches in the analysis of metaphorical language. Since such approaches are ultimately only as good as the data over which they operate, we shall also focus on the contributions of corpus linguistics to the construction of annotated metaphor corpora. Finally, we shall draw these strands together to offer an application-oriented view of metaphor, asking whether there is a killer application for metaphor research, and whether (and how) computational approaches to metaphor can help advance not only the field of NLP, but other fields as well, such as the social sciences and education.
The ultimate goal of this book is not to make you believe, as we do, that metaphor is the very soul of language, though the growing field of metaphor research is always eager to welcome new converts. We will consider this book a success if readers take away a desire to address metaphor head on, in some form or another in their research, and find in this book the necessary tools to make this engagement a practical reality.
Tony Veale, Ekaterina Shutova, and Beata Beigman Klebanov
January 2016
CHAPTER 1
Introducing Metaphor
Language would be a dull and brittle thing without metaphor. It is metaphor and its figurative kin—simile, analogy, blending, irony, understatement, hyperbole, and the like—that lend language its vitality and elasticity. It is metaphor and its kin that allow us to suggest much more than we actually say, and to invent new ways of saying it, when conventional language shows us its limits. It is metaphor and its kin that allow us to communicate not just information, but also real feelings and complex attitudes. Metaphor does not just report the result of personal insights, but also prompts and inspires listeners to have these insights for themselves. Each metaphor is a concise but highly productive way of saying that communicates a new and productive way of seeing.
But what exactly are metaphors and where do we look for them? Although metaphors are products of our faculty for creative thinking, metaphors in the wild can range from the scintillating to the banal. Just as repeated usage dulls the blade of a trusty knife, or repeated telling robs a once-funny joke of its ability to raise a laugh, repetition takes the bloom off a metaphor and turns it from an eye-catching flower into just another piece of the undergrowth. So metaphors are everywhere, in language, in film, in music, or in any system of signs that allows us to express ourselves creatively. These metaphors range from the novel to the conventional, and indeed some are so conventional as to escape our attention altogether. Nonetheless, even highly conventionalized metaphors retain a spark of the creativity that forged them, and in the right hands this spark can be fanned into a roaring fire. Consider the following example, which was used as the title of a popular science book by German rocket scientist Wernher von Braun: “I Aim For The Stars!” Since the heavens are filled with stars in any direction we care to look, von Braun could not have intended to use the word “aim” literally, which is to say “I look and/or point in
